WO2015017673A1

WO2015017673A1 - Methods for producing influenza vaccine compositions

Info

Publication number: WO2015017673A1
Application number: PCT/US2014/049192
Authority: WO
Inventors: Weidong Cui; Ashish BEZAWADA; Guangyu Zhu; Phuong TRAN; Loleta CHUNG
Original assignee: Medimmune, Llc
Priority date: 2013-08-01
Filing date: 2014-07-31
Publication date: 2015-02-05

Abstract

Provided are methods for producing influenza viruses. Typically, influenza virus production includes one or more purification processes. Provided are modified purification methods which can improve purification efficiency and increase viral yield without negatively impacting viral potency or stability.

Description

METHODS FOR PRODUCING INFLUENZA VACCINE COMPOSITIONS

Field The technology relates in part to methods for purifying influenza viruses. Background

Each year numerous individuals are infected with different strains and types of influenza virus. Infants, the elderly, those without adequate healthcare, and immunocompromised persons are at special risk of death from such infections. Compounding the problem of influenza infections is that novel influenza strains evolve readily, thereby necessitating the continuous production of new vaccines. Numerous vaccines capable of producing a protective immune response specific for such different influenza viruses have been produced for over 50 years and include, e.g., whole virus vaccines, split virus vaccines, surface antigen vaccines and live attenuated virus vaccines. However, while appropriate formulations of any of these vaccine types are capable of producing a systemic immune response, live attenuated virus vaccines have the advantage of being also able to stimulate local mucosal immunity in the respiratory tract. A vaccine comprising a live attenuated virus that is capable of being quickly and economically produced and that is capable of easy storage/transport is thus quite desirable. In certain instances, a vaccine capable of being stored/transported at refrigerator temperatures (e.g., approximately 2-8 °C) is desirable. Typically, influenza vaccines are propagated in embryonated hen eggs. Although influenza virus grows well in hen eggs, the production of vaccine is dependent on the availability of such eggs. Because the supply of eggs must be organized, and strains for vaccine production selected months in advance of the next flu season, the flexibility of this approach can be limited, and often results in delays and shortages in production and distribution. Additionally, influenza vaccines produced in eggs are typically subjected to several purification processes which can have a negative impact on yield and/or can significantly increase production time. Therefore, methods to improve influenza vaccine production efficiency (e.g., purification efficiency) are desirable to achieve, for example, higher yields and faster production. Summary

Provided herein are methods for making an influenza virus composition comprising a) clarifying a viral harvest comprising influenza viruses by filtration, thereby producing a clarified viral harvest; b) concentrating the clarified viral harvest, thereby producing a concentrated viral harvest; c) subjecting the concentrated viral harvest to centrifugation, thereby producing a further clarified viral harvest; and d) sterilizing by sterile filtration the further clarified viral harvest, thereby producing a sterilized viral harvest. Also provided herein are methods for making an influenza virus composition comprising: a) clarifying a viral harvest comprising influenza viruses by filtration, thereby producing a clarified viral harvest; b) subjecting the clarified viral harvest to centrifugation, which centrifugation comprises continuous zonal centrifugation performed over a sucrose density gradient, where the sucrose density gradient is generated by combining a volume of a 60% (w/w) sucrose composition and a volume of a 10% (w/w) sucrose composition, where the volume of the 60% (w/w) sucrose composition is equal to or greater than the volume of the 10% (w/w) sucrose composition; thereby producing a further clarified viral harvest; and c) sterilizing by sterile filtration the further clarified viral harvest, thereby producing a sterilized viral harvest. Certain embodiments are described further in the following description, examples, claims and drawings.

Brief Description of the Drawings The drawings illustrate embodiments of the technology and are not limiting. For clarity and ease of illustration, the drawings are not made to scale and, in some instances, various aspects may be shown exaggerated or enlarged to facilitate an understanding of particular embodiments.

FIG. 1 shows a flow diagram of an existing live attenuated influenza virus monovalent bulk (LAIV-MB) manufacturing process (left) and a modified LAIV-MB manufacturing process (right).

FIG. 2 shows a tangential flow filtration (TFF) set up.

FIG. 3 presents equations useful for calculating transmembrane pressure (TMP). FIG. 4 shows a summary of process parameters.

FIG. 5 shows a summary of results from TFF concentration.

FIG. 6 shows a characterization of TFF concentrated clarified harvest fluid (CHF).

FIG. 7 shows permeate flow per unit area vs. time using a GE Healthcare HF cartridge. FIG. 8 shows permeate flow per unit area vs. time using a Spectrum Labs single-use HF cartridge.

FIG. 9 shows flux vs. TMP curves for A/South Dakota/6/07. FIG. 10 shows flux vs. TMP curves for B/Malaysia/2506/04. FIG. 1 1 shows flux vs. TMP curves for A/Uruguay/716/07. FIG. 12 shows flux vs. TMP curves for B/Florida/4/2006.

FIG. 13 shows a summary of optimal TMP values at different shear rates and corresponding flux rates.

FIG. 14 shows average flux and TMP vs. shear rate.

FIG. 15 shows a summary of certain process development studies.

FIG. 16 shows recovery and material balances after TFF concentration for a supplemental characterization study (phase 1 ).

FIG. 17 shows recovery and material balances after TFF concentration for a supplemental characterization study (phase 2).

FIG. 18 shows impurity profiles for a supplemental characterization study (phase 1 ). FIG. 19 shows impurity profiles for a supplemental characterization study (phase 2). FIG. 20 shows TFF experimental conditions and average permeate flux.

FIG. 21 shows a contour plot of shear rate, TMP and average permeate flux. FIGS. 22A and 22B show potency assay data for GE HF TFF processes. FIG. 23 shows a summary of process parameters for certain pilot-scale studies.

FIG. 24 shows impact of loading flow rate on virus recovery in an ultracentrifugation process. FIG. 25 shows certain operational parameters for concentration and ultracentrifugation.

FIG. 26 shows doses per batch comparison for A/Victoria strain (TFF vs. non-TFF). TFF batches include the two shown on the right.

FIG. 27 shows doses per batch comparison for B/Wisconsin strain (TFF vs. non-TFF).

FIG. 28 shows doses per batch comparison for A/California strain (TFF vs. non-TFF).

FIG. 29 shows impurity removal data for TFF and non-TFF batches for A/Victoria strain. FIG. 30 shows impurity removal data for TFF and non-TFF batches for A/California strain.

FIG. 31 shows impurity removal data for TFF and non-TFF batches for B/Wisconsin strain.

FIG. 32 shows filterability and potency change of cold adapted influenza virus (CAIV) after pre- clarification through 47 mm pre-filters.

FIG. 33 shows a filterability comparison between clarification processes with and without pre- clarification filtration. FIG. 34 shows a potency change comparison between clarification processes with and without pre-clarification filtration.

FIG. 35 shows a summary of average potency of five CAIV strains before and after pre- clarification filtration and clarification filtration.

FIG. 36 shows potency and virus recovery data for 8 micrometer clarification batches compared to previous commercial batches. FIG. 37 shows filters and operation conditions used for a cross-flow microfiltration (CF-MF) study.

FIG. 38 shows permeate flux of CF-MF using a GE 0.45 micrometer hollow fiber cartridge without permeate control.

FIG. 39 shows permeate flux of CF-MF using a GE 0.45 micrometer hollow fiber cartridge with permeate control at 45 LMH.

FIG. 40 shows permeate flux of CF-MF using a Pall 0.65 micrometer hollow fiber cartridge without permeate control.

FIG. 41 shows potency of pooled harvest fluid (PHF) and clarified harvest fluid (CHF) from a hollow fiber CM-MF process. FIG. 42 shows permeate flux of flat sheet cassette TFF (Sartorius 0.45 micrometer flat sheet with permeate control at 3 psi.

FIG. 43 shows permeate flux of flat sheet cassette TFF (Millipore 0.65 micrometer flat sheet with permeate control at various pressures).

FIG. 44 shows potency of pooled harvest fluid (PHF) and clarified harvest fluid (CHF) from a CF-MF process using flat sheet cassettes. FIG. 45 shows permeate flux of a Pall KLEENPAK 0.65 micrometer capsule with no permeate control in a CF-MF process using ca A Uruguay/716/07.

FIG. 46 shows permeate flux of a Pall KLEENPAK 0.65 micrometer capsule with permeate control at 150 LMH in a CF-MF process using ca A/Uruguay/716/07.

FIG. 47 shows potency of pooled harvest fluid (PHF) and clarified harvest fluid (CHF) from a Pall KLEENPAK capsule TFF process. FIG. 48 shows virus potency and filtration throughput of the Millipore MILLISTAK+ DOHC and C0HC filtration process.

FIG. 49 shows a summary of linear flux and corresponding flow rates used in certain

MILLISTAK+ DOHC filtrations.

FIG. 50 shows a summary of filtration throughput at a flux of 250 LMH when filtration end pressure reached 30 psi.

FIG. 51 shows a summary of potency change after MILLISTAK+ DOHC depth filtration at a flux of 250 LMH and at a filtration end pressure of 30 psi.

FIG. 52 shows an effect of flux on filtration throughput and potency recovery of MILLISTAK+ DOHC depth filtration. FIG. 53 shows throughput of MILLISTAK+ DOHC filtrate on a 0.8/0.45 micrometer

(SARTOPORE 2) filter.

FIG. 54 shows potency of monovalent bulk (MVB) stored in a 125 mL bottle and a 1 L bag at 2- 8°C over a period of 14 days.

FIG. 55 shows filtration throughput after filtering through a MILLISTAK+ DOHC depth filter at 250 LMH.

FIG. 56 shows potency data for filtration through a MILLISTAK+ DOHC depth filter at 250 LMH. FIG. 57 shows sucrose and PBS volume of three gradient buffer compositions.

FIG. 58 shows volume of 60% and 10% sucrose and total time for rotor speed maintained at 35,000 rpm.

FIG. 59 shows a sucrose gradient generated from gradient buffer 1 (GB 1 ) with 0, 1 , 3, 5 and 12 hour run times at 35,000 rpm on a Hitachi CP40Y ultracentrifuge. FIG. 60 shows sucrose gradients generated from GB 1 , GB 2 and GB 3 with 1 hour run times at 35,000 rpm on a Hitachi CP40Y ultracentrifuge.

FIG. 61 shows sucrose gradients generated from GB 1 , GB 2 and GB 3 with 3 hour run times at 35,000 rpm on a Hitachi CP40Y ultracentrifuge.

FIG. 62 shows total volume of 60% sucrose recovered with different gradient buffers used and total ultracentrifuge run time.

FIG. 63 shows sucrose gradient profiles for certain batches using GB 1 and GB 3.

FIG. 64 shows sucrose concentration of centrifuge fractions using GB 3 (1.2 L 60% sucrose, 1 .6 L 10% sucrose and 0.4 L PBS).

FIG. 65 shows ultracentrifugation process times.

FIG. 66 shows sucrose gradient concentration of GB 1 (1.5 L 60% sucrose, 1 .3 L 10% sucrose and 0.4 L PBS) and total centrifuge run time.

FIG. 67 shows sucrose gradient concentration of GB 2 (1.35 L 60% sucrose, 1.45 L 10% sucrose and 0.4 L PBS) and total centrifuge run time.

FIG. 68 shows sucrose gradient concentration of GB 3 (1.2 L 60% sucrose, 1 .6 L 10% sucrose and 0.4 L PBS) and total centrifuge run time. FIG. 69 shows a flow diagram of an existing influenza virus production method (left) and an influenza virus production with certain optional modifications outlined in dashed boxes (right).

FIG. 70 shows CHF concentration and loading flow rate for various strains.

FIG. 71 shows loading flow rate, volume and total process time of concentrated CHF, and the percentage of virus lost in the flow-through.

FIG. 72 shows the impact of loading flow rate on virus lost in the flow-through.

FIG. 73 shows volume and fluorescent focus assay (FFA) titer of concentrated CHF and centrifuge flow-through.

Detailed Description

Provided herein are methods for producing influenza viruses. Typically, influenza virus production includes one or more purification processes. For example, influenza viruses in a viral harvest (e.g., from eggs or cells) may be purified using methods such as filtration and/or centrifugation. Such purification methods reduce or substantially eliminate contaminants (e.g., cellular debris, bioburden, host cell proteins and/or host cell nucleic acid) present in a viral harvest. However, such methods can be costly and time consuming, and/or can have a negative impact viral yield, potency, and/or stability. Provided herein are modified purification methods which can improve purification efficiency and increase viral yield without negatively impacting viral potency or stability.

Influenza virus and vaccines

Provided herein are methods for the production of influenza viruses suitable as vaccines, including live attenuated influenza vaccines, such as those suitable for administration in an intranasal vaccine formulation. Influenza viruses are made up of an internal ribonucleoprotein core containing a segmented single-stranded RNA genome and an outer lipoprotein envelope lined by a matrix protein. Influenza A and influenza B viruses each contain eight segments of single stranded negative sense RNA. The influenza A genome encodes eleven polypeptides. Segments 1 -3 encode three polypeptides, making up a RNA-dependent RNA polymerase. Segment 1 encodes the polymerase complex protein PB2. The remaining polymerase proteins PB1 and PA are encoded by segment 2 and segment 3, respectively. In addition, segment 1 of some influenza strains encodes a small protein, PB1-F2, produced from an alternative reading frame within the PB1 coding region. Segment 4 encodes the hemagglutinin (HA) surface glycoprotein involved in cell attachment and entry during infection. Segment 5 encodes the nucleocapsid nucleoprotein (NP) polypeptide, the major structural component associated with viral RNA. Segment 6 encodes a neuraminidase (NA) envelope glycoprotein. Segment 7 encodes two matrix proteins, designated M1 and M2, which are translated from differentially spliced mRNAs. Segment 8 encodes NS1 and NS2, two nonstructural proteins, which are translated from alternatively spliced mRNA variants.

The eight genome segments of influenza B encode 1 1 proteins. The three largest genes code for components of the RNA polymerase, PB1 , PB2 and PA. Segment 4 encodes the HA protein. Segment 5 encodes NP. Segment 6 encodes the NA protein and the NB protein. Both proteins, NB and NA, are translated from overlapping reading frames of a biscistronic mRNA. Segment 7 of influenza B also encodes two proteins: M1 and M2. The smallest segment encodes two products, NS1 which is translated from the full length RNA, and NS2 which is translated from a spliced mRNA variant. Typically, influenza virus vaccines are produced in embryonated hen eggs (e.g., specific pathogen free (SPF) embryonated hen eggs) using strains of virus selected based on empirical predictions of relevant strains. Often, reassortant viruses are produced that incorporate selected hemagglutinin and neuraminidase antigens in the context of an approved attenuated, temperature sensitive, and/or cold-adapted master strain. Following culture of the virus through multiple passages in hen eggs, influenza viruses are recovered and, optionally, inactivated, e.g., using formaldehyde and/or beta-propiolactone; or are used in live attenuated vaccines.

However, production of influenza vaccine in this manner has several significant concerns. For example, contaminants remaining from the hen eggs can be highly antigenic and/or pyrogenic, and can frequently result in significant side effects upon administration. Thus, certain methods include purification methods that reduce such contaminants and/or replacement of some or all of egg components with animal free media. Virus strains designated for vaccine production typically are selected and distributed months in advance of the next flu season to allow time for production and inactivation of influenza vaccine. Thus, improvements in production efficiency and/or stability at certain temperatures (e.g., refrigerator temperature of about 2-8 °C), are desirable.

Recombinant and reassortant vaccines also may be produced in cell culture (e.g., using a vector system described, for example, in U.S. patent no. 8,012,736) using any appropriate type of host cell. Host cells can be prokaryotic cells such as E. coli, or eukaryotic cells such as yeast, insect, amphibian, avian or mammalian cells, including human cells. Host cells may include, for example, Vero (African green monkey kidney) cells, BHK (baby hamster kidney) cells, CHO cells, Hep-2 cells, HeLa cells, LLC-MK2 cells, primary chick kidney (PCK) cells, Madin-Darby Canine Kidney (MDCK) cells, Madin-Darby Bovine Kidney (MDBK) cells, human diploid lung fibroblast cell lines (e.g., MRC-5 and WI-38), human retinoblastoma cell lines, fetal rhesus lung cell lines (e.g., FRhl_2), human kidney cell lines (e.g., PER.C6 and 293 (293T)), and COS cells (e.g., COS1 , COS7 cells). In certain instances, reassortant influenza A and/or influenza B viruses can be produced in cells using an eight plasmid system from cloned cDNA (see e.g., U.S. patent no. 8,012,736). Such reassortants are optionally further amplified in hen eggs. Typically, cell cultures are maintained in a system, such as a cell culture incubator, under controlled humidity and C0₂, at constant temperature using a temperature regulator, such as a thermostat to insure that the temperature does not exceed 35 °C. Such cell culture methods can be modified using methods described herein in whole or part.

In some embodiments, the influenza viruses correspond to one or more influenza B viruses. In some embodiments, the influenza viruses correspond to one or more influenza A viruses. In certain embodiments, the methods include producing recombinant and/or reassortant influenza viruses capable of eliciting an immune response upon administration, e.g., intranasal administration, to a subject. In some embodiments, the viruses are inactivated prior to administration. In some embodiments, live-attenuated viruses are administered. In certain embodiments, viruses include an attenuated influenza virus, a cold adapted influenza virus, a temperature sensitive influenza virus, or a virus with any combination of these desirable properties. In some embodiments, an influenza virus incorporates an influenza B/Ann

Arbor/1/66 strain virus, e.g., a cold adapted, temperature sensitive, attenuated strain of B/Ann Arbor/1/66. In some embodiments, an influenza virus incorporates an influenza A/Ann Arbor/6/60 strain virus, e.g., a cold adapted, temperature sensitive, attenuated strain of A/Ann Arbor/6/60. In some embodiments, viruses are artificially engineered influenza viruses incorporating one or more substituted amino acids which influence certain biological properties of a donor strain, e.g., ca A/Ann Arbor/6/60 or ca B/Ann Arbor/1/66. Such substituted amino acids may correspond to unique amino acids of ca A/Ann Arbor/6/60 or ca B/Ann Arbor/1/66, e.g., in an A strain virus: PB1³⁹¹ (K391 E), PB1⁵⁸¹ (E581 G), PB1⁶⁶¹ (A661 T), PB2²⁶⁵ (N265S) and NP³⁴ (D34G); and, in a B strain virus: PB2⁶³⁰ (S630R); PA⁴³¹ (V431 M); PA⁴⁹⁷ (Y497H); NP⁵⁵ (T55A); NP¹¹⁴ (V1 14A); NP⁴¹⁰ (P410H); NP⁵⁰⁹ (A509T); M1¹⁵⁹ (H159Q) and M1¹⁸³ (M 183V). Such amino acid substitutions are described, for example, in U.S. Patent No. 8,722,059 and U.S. Patent No. 8,1 14,415. Similarly, other amino acid substitutions at any of these positions resulting in temperature sensitivity, cold adaptation and/or attenuation are contemplated.

Certain A or B viruses may already have the recited residues at the indicated positions. In such instances, the substitutions can be made such that the resulting virus will have all of the above substitutions.

Reassortant viruses may be produced by introducing vectors including the six internal genes of a first viral strain selected for its favorable properties regarding vaccine production, in combination with the genome segments encoding the surface antigens (HA and NA) of a selected, e.g., pathogenic strain. Such reassortants are sometimes referred to as 6:2 reassortants. In some instances, seven complementary gene segments (i.e., 6 internal genes and 1 surface antigen) of a first strain are introduced in combination with either an HA or NA encoding segment. Such reassortants are sometimes referred to as 7:1 reassortants. In certain instances, an HA segment can be selected from a pathogenically relevant influenza A strain (e.g., H1 , H3) or influenza B strain. In certain instances, the HA segment can be selected from an emerging pathogenic influenza strain such as an H2 influenza strain (e.g., H2N2), an H5 influenza strain (e.g., H5N1 ) or an H7 influenza strain (e.g., H7N7). Similarly, the NA segment can be selected from a pathogenically relevant or emerging pathogenic influenza A strain or influenza B strain, and may be selected from any NA subtype (e.g., N1 , N2, N3 , N7). In certain embodiments, the internal gene segments are derived from the influenza B/Ann Arbor/1/66, A/Ann Arbor/6/60 or other suitable master strain. In some embodiments, the master strain is selected from the group consisting of A/Ann Arbor/6/60, B/Ann Arbor/1/66, PR8,

B/Leningrad/14/17/55, LEN-B14/5/1 , B/USSR/60/69, B/Leningrad/179/86, B/Leningrad/14/55 and B/England/2608/76. In some embodiments, the master strain is derived from a strain selected from the group consisting of A/Ann Arbor/6/60, B/Ann Arbor/1/66, PR8, B/Leningrad/14/17/55, LEN-B14/5/1 , B/USSR/60/69, B/Leningrad/179/86, B/Leningrad/14/55 and B/England/2608/76. The master strain may be derived from any of the above strains by the introduction of one or more amino acid substitutions that confer a desirable phenotype such as attenuation, temperature sensitivity and/or cold-adaptation, as describe above and as described, for example in U.S. patent no. 8,354, 1 14.

The terms "temperature sensitive," "cold adapted" and "attenuated" are known in the art. For example, the term "temperature sensitive" ("ts") indicates that the virus exhibits a 100 fold or greater reduction in titer at a higher temperature, e.g., 39°C relative to a lower temperature, e.g., 33°C for influenza A strains, and that the virus exhibits a 100 fold or greater reduction in titer at a higher temperature, e.g., 37°C relative to a lower temperature, e.g., 33°C for influenza B strains. For example, the term "cold adapted" ("ca") indicates that the virus exhibits a higher growth rate at a lower temperature, e.g., 25°C within 100 fold of its growth at a higher temperature, e.g., 33°C. For example, the term "attenuated" ("att") indicates that the virus replicates in the upper airways of ferrets but is not detectable in lung tissues, and does not cause influenza-like illness in the animal. Growth indicates viral quantity as indicated by titer, plaque size or morphology, particle density or other measures known in the art.

Vaccine production

Influenza vaccine production typically includes multiple manufacturing steps including, for example, co-infection, reassortment, selection and cloning of reassortants, purification and expansion of reassortants, harvesting, purification of a viral harvest, stabilization, and potency/sterility assays. Various aspects of vaccine production are described, for example, in U.S. patent no. 7,262,045; U.S. patent no. 8,247,207; U.S. patent no. 8,012,736; U.S. patent no. 7,465,456; U.S. patent no. 8,354, 1 14; U.S. patent no. 7,601 ,356; U.S. patent no. 8,357,376; U.S. patent no. 8,202,726; U.S. patent no. 8,093,033; and U.S. patent no. 7,510,719, each of which is incorporated by reference in its entirety. Certain aspects of the methods described herein may or may not be included in a given production series. Thus, while in certain embodiments, multiple steps and/or compositions herein are performed or exist, in other embodiments, one or more steps are optionally omitted or changed (e.g., in scope, order, placement). Certain embodiments comprise

steps/methods/compositions that are known in the art. Therefore, appropriate conditions, sub- steps, step details, and the like, for such known steps can be determined for producing viruses, virus solutions, compositions, and the like. Certain embodiments include parameters presented in ranges or given as approximate values (e.g., using the term "about", which, unless otherwise indicated, refers to a value that is no more than 10% above or below the value being modified by the term). Certain aspects of influenza vaccine production and modifications thereto are described in detail below.

Purification of a viral harvest Viruses (e.g., reassortant influenza viruses) grown in eggs or host cells may be harvested (i.e., removed from the eggs or host cells) and subjected to one or more purification processes which may include, for example, clarification, concentration, centrifugation and/or sterilization. Certain aspects of viral purification can be modified to increase production efficiency (e.g., higher yield, faster production, less waste, and the like). Such modified aspects of viral purification are described herein.

In some embodiments, purification of a viral harvest comprises a) subjecting a concentrated viral harvest to centrifugation, thereby producing a clarified viral harvest; and, optionally, b) sterilizing by sterile filtration the clarified viral harvest, thereby producing a sterilized viral harvest. A viral harvest herein generally comprises influenza viruses. In certain embodiments, a viral harvest is initially clarified before or during a concentration step.

In some embodiments, purification of a viral harvest comprises a) concentrating a viral harvest, thereby producing a concentrated viral harvest; b) subjecting the concentrated viral harvest to centrifugation, thereby producing a clarified viral harvest; and, optionally, c) sterilizing by sterile filtration the clarified viral harvest, thereby producing a sterilized viral harvest. In certain embodiments, a viral harvest is initially clarified before or during concentration.

In some embodiments, purification of a viral harvest comprises a) concentrating a viral harvest, where the viral harvest optionally is a clarified viral harvest, thereby producing a concentrated viral harvest; and b) subjecting the concentrated viral harvest to centrifugation, thereby producing a clarified viral harvest. In some embodiments, purification of a viral harvest comprises a) clarifying a viral harvest comprising influenza viruses, thereby producing a clarified viral harvest; b) concentrating the clarified viral harvest, thereby producing a concentrated viral harvest; c) subjecting the concentrated viral harvest to centrifugation, thereby producing a further clarified viral harvest; and, optionally, d) sterilizing by sterile filtration the further clarified viral harvest, thereby producing a sterilized viral harvest.

In some embodiments, purification of a viral harvest comprises a) clarifying a viral harvest comprising influenza viruses, thereby producing a clarified viral harvest; b) subjecting the clarified viral harvest to centrifugation, which centrifugation comprises continuous zonal centrifugation performed over a sucrose density gradient, where the sucrose density gradient is generated by combining a volume of a 60% (w/w) sucrose composition and a volume of a 10% (w/w) sucrose composition, where the volume of the 60% (w/w) sucrose composition is equal to or greater than the volume of the 10% (w/w) sucrose composition; thereby producing a further clarified viral harvest; and, optionally, c) sterilizing by sterile filtration the further clarified viral harvest, thereby producing a sterilized viral harvest.

Clarification In some embodiments, a viral purification process comprises an initial clarification of a viral harvest. Methods useful for the initial clarification of a viral harvest include, but are not limited to, centrifugation, dialysis, and membrane filtration, which includes, but is not limited to, methods such as single pass, dead-end, direct flow filtration (DFF) in which liquid flows directly through the filter medium, depth filtration, and crossflow or tangential flow filtration (TFF) in which liquid flows tangential to (along) the surface of the membrane. Membranes for use in filtration applications are available from commercial sources. Certain methods for the initial clarification of a viral harvest, and modifications thereto, are described herein in Example 2.

In some embodiments, a viral harvest is clarified by filtration. Filtration typically involves use of membranes which generally are defined by the size of the material they remove from a solution. For example, from the smallest to largest pore size, filtration membranes include reverse osmosis membranes, nanofiltration membranes, ultrafiltration membranes, and microfiltration membranes. Filtration using such membranes separates molecules according to their molecular weight by using membranes with specific pore sizes. For example, filtration with reverse osmosis membranes that have pore sizes less than 0.001 micrometers generally is intended for separation of molecules that have a molecular weight less than 200 Daltons.

Filtration with nanofiltration membranes that have pore sizes from 0.001 - 0.008 micrometers, inclusive, generally is intended for separation of molecules that have a molecular weight from 200 Daltons to 15 kilodaltons (kD, kDa) inclusive. Filtration with ultrafiltration membranes that have pore sizes from 0.005 - 0.1 micrometers, inclusive, generally is intended for separation of molecules that have a molecular weight from 5 kDa - 300 kDa, inclusive. Filtration with microfiltration membranes that have pore sizes from 0.05 - 3.0 micrometers, inclusive, is intended for separation of molecules that have a molecular weight from 100 kDa - 3000 kDa and larger. Accordingly, membrane-filtration can separate molecules of interest (e.g., viruses) from other cellular components based on size exclusion by utilizing membranes that have a particular Molecular Weight Cut-Off (MWCO) that is determined by the pore size of the membrane. The MWCO, also called Nominal Molecular Weight Limit (NMWL) or Nominal Molecular Weight Cut- Off (NMWCO), is the kilodalton size designation for the filtration by membranes. The MWCO is defined as the molecular weight of the molecule that is 90% retained by the membrane.

Because, for example, molecules of the same molecular weight can have significantly different shapes, the MWCO may not be an exact metric, but is nevertheless a useful metric and is commonly employed by filter manufacturers. Membranes may be used as flat sheets or in a spirally wound configuration, for example. Hollow fibers may also be used depending on the type of filtration method. Any number of potential membrane materials may be used including, but not limited to, regenerated cellulose, polyether sulfone (which may or may not be modified to alter its inherent hydrophobicity), polyvinylidene fluoride (PVDF), and ceramic and metal oxide aggregates, as well as polycarbonate, polypropylene, polyethylene and PTFE (TEFLON®). In some embodiments, combinations of filtration methods and membrane types may be used. The capacity of certain filters, columns, etc., comprising separation membranes can be adjusted depending on the volume and/or concentration of material being processed.

In some embodiments, clarification of a viral harvest comprises use of one or more filter species. A filter species may be distinct from another filter species based on pore size, membrane material, filter manufacturer, membrane area, layers of membrane, filter capacity and the like or a combination thereof. In some embodiments, clarification of a viral harvest comprises use of at least two filter species. For example, clarification of a viral harvest may comprise use of at least three filter species, at least four filter species, at least five filter species, at least six filter species or more. In some embodiments, clarification of a viral harvest comprises use of at least three filter species. In certain embodiments, one or more filter species is a pre-filter. A pre-filter generally is used in a filtration process prior to the use of one or more other filter species (e.g., downstream filters), and can remove certain cell debris components from a viral harvest (e.g., host cell debris). A pre-filter may have a pore size that is larger than one or more downstream filters. In some embodiments, a pre-filter has a pore size ranging from about 3 microns to about 20 microns. For example, a pre-filter may have a pore size of about 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14, 15, 16, 17, 18, 19 or 20 microns. In some embodiments, a pre-filter has a pore size of about 8 microns. In some embodiments, a pre-filter has a pore size of about 10 microns.

In some embodiments, use of a pre-filter increases filtration throughput for a viral harvest clarification process. Filtration throughput refers to the passage of a solution (e.g., viral harvest fluid) through a filter (e.g., one or more clarification filters such as the one or more clarification filters downstream of a pre-filter) for a certain duration, flow rate and/or volume of solution passaged before filtration slows or ceases due to, for example, filter clogging. For example, filtration throughput may be increased when a pre-filter is used relative to filtration throughput when a pre-filter is not used. In some embodiments, filtration throughput is increased by at least about 1 .5-fold. For example, filtration throughput may be increased by at least about 2-fold, 2.5- fold, 3-fold, 3.5-fold, 4-fold or more. In some embodiments, filtration throughput is increased by at least about 3-fold.

In some embodiments, one or more other filter species is used after a pre-filter to remove other cell debris components, bacteria, other bioburden, and the like from a viral harvest. In some embodiments, the one or more other filter species have pore sizes that are smaller than a pre- filter. In some embodiments, the one or more other filter species are selected from filters having pore sizes ranging from about 0.2 microns to about 3.0 microns. For example, the one or more other filter species may have pore sizes of about 0.3, 0.4, 0.45, 0.5, 0.6, 0.7, 0.8, 0.9, 1 .0, 1.1 , 1.2, 1 .3, 1 .4, 1 .5, 1.6, 1.7, 1.8, 1 .9, 2.0, 2.1 , 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.8, 2.9 or 3.0 microns. In some embodiments, one or more other filter species have a pore size of 1 .2 microns. In some embodiments, one or more other filter species have a pore size of 0.8 microns. In some embodiments, one or more other filter species have a pore size of 0.45 microns. In some embodiments, one or more other filter species comprise one or more membrane layers. In some embodiments, one or more other filter species comprise two membrane layers (e.g., paired filters). For example, one or more other filter species may comprise two membrane layers, each having a different pore size (e.g., 0.8 microns and 0.45 microns). In some embodiments, the one or more other filter species comprise one or more filters having a pore size of 1 .2 microns and one or more filters having two membrane layers, each having a pore size of 0.8 microns and 0.45 microns.

In some embodiments, clarification of a viral harvest comprises use of one or more membrane filters. Membrane filters (sometimes referred to as screen filters) generally have pores of a certain size that allow certain particles to pass through. In some embodiments, clarification of a viral harvest comprises use of one or more depth filters. Depth filters generally include filters that comprise a porous filtration medium (e.g., fibers, or fibrous materials) to retain particles throughout the medium, rather that just on the surface of the medium. Such filters often can retain a large mass of particles before becoming clogged. In some embodiments, a depth filter is a stacked depth filter. In some embodiments, clarification of a viral harvest comprises use of a combination of one or more membrane filters and one or more depth filters. In some embodiments, a depth filter is used prior to a membrane filter. In some embodiments, a depth filter is used after a membrane filter. In some embodiments, a depth filter is used after a first membrane filter and before a second membrane filter. In some embodiments, one or more depth filters is used as a pre-filter. In some embodiments, one or more depth filters is used in combination with one or more paired membrane filters (e.g., 0.8/0.45 micron filter).

In some embodiments, filtration throughput is increased when a depth filter is used relative to filtration throughput when a depth filter is not used. In some embodiments, filtration throughput is increased by at least about 1.5-fold. For example, filtration throughput may be increased by at least about 2-fold, 2.5-fold, 3-fold, 3.5-fold, 4-fold or more. In some embodiments, filtration throughput is increased by at least about 3-fold.

Concentration

In some embodiments, a viral purification process comprises concentration of a viral harvest. For example, a viral purification process may comprise concentration of a clarified viral harvest, such as a viral harvest clarified by a filtration process described above. Methods useful for concentrating a viral harvest (e.g. a clarified viral harvest) include, but are not limited to, dialysis, tangential flow filtration (TFF), ultrafiltration (UF) and diafiltration (DF; continuous or

discontinuous). TFF may incorporate both UF, which can be used to concentrate, and DF, which can be used to exchange buffers. Tangential flow filtration (TFF), sometimes referred to as crossflow filtration, is a process whereby a feed stream passes parallel to the membrane face as one portion passes through the membrane (permeate) while the remainder (retentate) is recirculated back to the feed reservoir. The use of TFF, in certain instances, may result in additional purification by the fractionation process that washes smaller molecules (e.g., contaminants) through a membrane and leaves larger molecules of interest (e.g., virus) in the retentate. A viral purification process may incorporate the use of any suitable TFF system known in the art and any TFF components (e.g., cartridges) by various manufacturers. Non- limiting examples of certain TFF systems and components useful for concentrating a viral harvest are described herein in Example 1.

In some embodiments, a TFF process comprises use of a hollow fiber cartridge (i.e., a filter membrane composed of a collection of hollow fibers (e.g., polysulphone)). In some

embodiments, the hollow fiber cartridge has a pore size ranging from about 500 kD to about 750 kD. In some embodiments, the hollow fiber cartridge has a pore size of about 500 kD. In some embodiments, the hollow fiber cartridge has a pore size of about 750 kD. In some

embodiments, the hollow fiber cartridge has a membrane with a nominal internal diameter (ID) of about 0.5 mm and pore size of about 500 kD. During TFF concentration, in some embodiments, product flows tangentially across the surface of the hollow fiber filter membranes at a defined flow rate. As the product solution flows through the hollow fibers, the inlet and outlet pressures are controlled to provide a constant differential pressure. This differential pressure enables concentration whereby waste and impurities (which generally are smaller than 500 kD or smaller than 750 kD) pass through the pores and enter the waste stream (permeate) while virus particles (which are typically bigger than 500 kD or 750 kD) are retained in the product solution. Over a period of time, waste material (which is generally less than 500 kD or 750 kD) held in a separate bioprocess container increases in volume. Virus particles (which are larger than 500 kD or 750 kD) are retained (as the retentate) in a separate bioprocess container and subsequently may be loaded onto an ultracentrifuge, for example, once the concentration process is complete.

A TFF process typically includes several operational parameters, some of which may be modified to achieve an optimal concentration process. Certain operational parameters are described below and non-limiting examples of modifications thereof are described herein in Example 1.

Shear rate (s^"1) is the ratio of velocity and distance. Shear rate can be controlled, and an increased shear rate typically ensures the efficiency of the filter is maintained over the lifetime of a concentration process. An optimized shear should prevent filter blockage and thus ensures effective concentration times are maintained, although shear may be dependent on the limits of certain equipment and changes in the nature of the product. The shear rate for a hollow fiber (HF) cartridge, for example, can be calculated based on the flow rate through the fiber lumen as follows:

Shear rate = [ 4q/ 7ir³], where

q = flow rate through the fiber lumen (cm³ /sec), and

r =fiber radius (cm)

In some embodiments, a TFF process is performed using a shear rate ranging from about 8,000 s^"1 to about 22,000 s^' In some embodiments, a TFF process is performed using a shear rate ranging from about 10,000 s^"1 to about 16,000 s^' For example, a TFF process may be performed at a shear rate of about 1 1 ,000 s^"1; 12,000 s^"1; 13,000 s^"1; 14,000 s^"1; or 15,000 s^'

Transmembrane pressure (TMP) is the average applied pressure from the feed to the filtrate side of the membrane. An optimal TMP generally ensures the rate of concentration is maximized and controlled within an acceptable timeframe and within certain physical limits of the equipment, and prevents damage to the filter or the virus that is being concentrated. TMP may be measured as pounds per square inch (psi) or pounds per square inch gage (psig) and can be calculated as follows:

TMP = [(P_in + P_ret)/2] - P_perm, where

P_in = inlet pressure

Pret = retentate pressure

Pperm = permeate pressure In some embodiments, a TFF process is performed using a transmembrane pressure (TMP) ranging from about 10 psig to about 20 psig. For example, a TFF process may be performed at a TMP of about 1 1 psig, 12 psig, 13 psig, 14 psig, 15 psig, 16 psig, 17 psig, 18 psig, or 19 psig. Flux (filtrate flux rate) is the volume of the permeate flowing through the defined filter membrane area during a given time and is expressed as LMH (liters per square meter per hour). In some embodiments, a TFF process is performed at a filtrate flux rate of at least about 25 LMH. For example, a TFF process may be performed at a filtrate flux rate of about 30 LMH, 40 LMH, 50 LMH, 60 LMH, 70 LMH, 80 LMH, 90 LMH, 100 LMH, 150 LMH, 200 LMH or more.

Load factor is defined as the ratio of feed volume to filter surface area and is expressed as L/m² (liters per square meter). In some embodiments, a TFF process is performed using a load factor ranging from about 50 L/m² to 100 L/m² of clarified viral harvest per square meter. For example, a TFF process may be performed at a load factor of about 55 L/m², 60 L/m², 70 L/m², 80 L/m², or 90 L/m².

In certain embodiments, TFF systems may be run so as to maintain a constant filtrate flux rate (i.e., flux) or to maintain a constant transmembrane pressure (TMP). In some embodiments, flux and/or TMP may be regulated, for example, to prevent membrane fouling.

In some embodiments, a viral harvest (e.g., clarified viral harvest) is concentrated (e.g., by a TFF process) at least about 2-fold (e.g., 200 L clarified viral harvest concentrated to 100 L clarified viral harvest). For example, a clarified viral harvest may be concentrated about 3-fold, 4-fold, 5-fold, 6-fold, 7-fold, 8-fold, 9-fold, 10-fold, 20-fold, 50-fold, 100-fold or more.

In some embodiments, the volume of clarified viral harvest that can be purified is greater relative to the volume of clarified viral harvest that can be purified in a method that does not comprise concentrating a clarified viral harvest. In some embodiments, certain starting volumes of a viral harvest (e.g., clarified viral harvest) are subjected to concentration (e.g., by a TFF process). For example, 100 L, 150 L, 200 L, 250 L, 300 L, 350 L, 400 L, 450 L or more clarified viral harvest may be concentrated in a viral purification method provided herein.

In some embodiments, a viral purification process comprises concentrating a viral harvest prior to centrifugation. In some embodiments, viral yield is increased relative to viral yield of a method that does not comprise concentrating a clarified viral harvest prior to centrifugation. For example, a viral yield may be increased al least about 2%, 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% or more. In some embodiments, the amount of clarified viral harvest (i.e., the number of virus particles in a viral harvest) subjected to centrifugation is greater (e.g., less viral harvest is wasted or discarded) relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation. For example, the amount of clarified viral harvest subjected to centrifugation may be at least about 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45% or 50% greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise

concentrating the clarified viral harvest prior to centrifugation. In some embodiments, all or substantially all (e.g., about 90% or greater) of the clarified viral harvest (i.e., virus particles in a viral harvest) is subjected to centrifugation.

Centrifugation

In some embodiments, a viral purification process comprises centrifugation of a viral harvest. In some embodiments, a clarified viral harvest is subjected to centrifugation. In some

embodiments, a concentrated viral harvest is subjected to centrifugation. Centrifugation may include continuous zonal centrifugation, which may also be referred to as ultracentrifugation, continuous flow zonal centrifugation, continuous flow zonal ultracentrifugation, and the like. Any centrifuge device suitable for the methods described herein may be used. Non-limiting examples of certain centrifugation devices, processes and modifications thereto are described herein in Example 3.

Centrifugation may be performed at any temperature, rotor speed and/or duration suitable for virus purification. For example, centrifugation may be performed at room temperature or below. In some embodiments, centrifugation may be performed at about 2 °C to about 25 °C. In some embodiments, centrifugation may be performed at about 2 °C to about 14 °C. For example, centrifugation may be performed at about 3 °C, 4 °C, 5 °C, 6 °C, 7 °C, 8 °C, 9 °C, 10 °C, 1 1 °C, 12 °C, or 13 °C. In some embodiments, centrifugation performed at a speed of about 25,000 RPM to about 50,000 RPM. In some embodiments, centrifugation performed at a speed of about 30,000 RPM to about 40,000 RPM. For example, centrifugation may be performed at a speed of about 31 ,000 RPM, 32,000 RPM, 33,000 RPM, 34,000 RPM, 35,000 RPM, 36,000 RPM, 37,000 RPM, 38,000 RPM, or 39,000 RPM. In some embodiments, centrifugation has a run time of at least about 6 hours. For example, centrifugation may have a run time of about 7 hours, 8 hours, 9 hours, 10 hours, 1 1 hours, 12, hours, 13 hours, 14 hours, 15 hours or longer. In some embodiments, centrifugation has a run time of at least about 9 hours. In some embodiments, centrifugation has a run time of at least about 12 hours.

In some embodiments, a method comprises loading a viral harvest (e.g., concentrated viral harvest, clarified viral harvest) into a centrifuge device at a particular loading flow rate. In certain instances adjusting the loading flow rate (e.g., decreasing) can prevent product loss during a centrifugation process. In some embodiments, the loading flow rate is lower relative to a loading flow rate for centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation. For example, the loading flow rate may be less than about 200 mL/min, 190 mL/min, 180 mL/min, 170 mL/min, 160 mL/min, 150 mL/min, 140 mL/min, 130 mL/min, 120 mL/min, 1 10 mL/min or 100 mL/min. In some embodiments, the loading flow rate ranges from about 120 mL/min to about 160 mL/min. In some embodiments, the loading flow rate ranges from about 140 mL/min to about 180 mL/min. In some

embodiments, the loading flow rate is about 180 mL/min, 170 mL/min, 160 mL/min, 150 mL/min, 140 mL/min, 130 mL/min, or 120 mL/min.

In some embodiments, centrifugation comprises continuous zonal centrifugation. In some embodiments, centrifugation is performed over a sucrose density gradient. In some

embodiments, the sucrose density gradient is a 0% to 100% sucrose gradient. In some embodiments, the sucrose density gradient is a 0% to 90% sucrose gradient. In some embodiments, the sucrose density gradient is a 0% to 80% sucrose gradient. In some embodiments, the sucrose density gradient is a 0% to 70% sucrose gradient. In some embodiments, the sucrose density gradient is a 0% to 60% sucrose gradient. In some embodiments, the sucrose density gradient is a 10% to 100% sucrose gradient. In some embodiments, the sucrose density gradient is a 10% to 90% sucrose gradient. In some embodiments, the sucrose density gradient is a 10% to 80% sucrose gradient. In some embodiments, the sucrose density gradient is a 10% to 70% sucrose gradient. In some embodiments, the sucrose density gradient is a 10% to 60% sucrose gradient. In some embodiments, a sucrose density gradient is generated using two different sucrose concentrations. In some embodiments the sucrose is in a buffer. In a specific embodiment, the sucrose is in a phosphate buffer (e.g., a phosphate-glutamate buffer, or PBS) In some embodiments, a sucrose density gradient is generated using two different sucrose

concentrations and a buffer without sucrose. In some embodiments, a sucrose density gradient is generated using 1 ) a sucrose concentration of 60% and 2) a sucrose concentration of 10%. In some embodiments, a sucrose density gradient is generated using a volume of a 10% sucrose (w/w) composition that is greater than the volume of a 60% sucrose (w/w) composition. In some embodiments, a sucrose density gradient is generated using equal or substantially equal volumes of a 60% sucrose (w/w) composition and a 10% sucrose (w/w) composition. In some embodiments, a sucrose density gradient is generated using a volume of a 60% sucrose (w/w) composition that is greater than the volume of a 10% sucrose (w/w) composition. For example, a sucrose density gradient may be generated where the volume of a 60% sucrose (w/w) composition is at least about 1 .1 , 1 .2, 1.3, 1.4, 1.5, 1 .6, 1.7, 1 .8, 1 .9, 2.0 or more times greater than the volume of a 10% sucrose (w/w) composition. In some embodiments, a sucrose density gradient is generated where the volume of a 60% sucrose (w/w) composition is at least about 1 .1 times greater than the volume of a 10% sucrose (w/w) composition. In some embodiments, a sucrose density gradient is generated using volumes of a 60% sucrose (w/w) composition, a 10% sucrose (w/w) composition and buffer (e.g., PBS) at a ratio of 1 .3-1.6 to 1 .2- 1.5 to 0.4, respectively. In some embodiments, a sucrose density gradient is generated using volumes of a 60% sucrose (w/w) composition, a 10% sucrose (w/w) composition and PBS at a ratio of 1.5 to 1 .3 to 0.4, respectively.

In some embodiments, a viral harvest (e.g., a clarified viral harvest, a further clarified viral harvest) is collected from the sucrose density gradient at certain gradient coordinates. In some embodiments, a viral harvest (e.g., a clarified viral harvest, a further clarified viral harvest) is collected from the sucrose density gradient at gradient coordinates between about 30% to about 55% sucrose. In some embodiments, a viral harvest (e.g., a clarified viral harvest, a further clarified viral harvest) is collected from the sucrose density gradient at gradient coordinates between about 34-36% to about 48-50% sucrose. In some embodiments, a viral harvest (e.g., a clarified viral harvest, a further clarified viral harvest) is collected from the sucrose density gradient at gradient coordinates between about 35% to about 49% sucrose.

Pooling and/or Dilution of Peak Fractions In some embodiments, centrifuge peak fractions are pooled and/or diluted as described, for example in U.S. Patent No. 8,247,207. Peak fractions may be identified, for example, by a hemagglutinin assay. In some embodiments, sucrose concentration is determined for a peak fraction or peak fraction pool according to, for example, a refractive index (Rl) reading. Peak fractions or peak fraction pools also may be sampled for potency. Fractions or pools with certain sucrose concentrations and/or potencies may be diluted by addition of a buffer, in certain embodiments. Buffers can be sterile and/or cold (e.g., 2-8°C) and may include, for example, a phosphate buffer. For example a phosphate-glutamate buffer (PBG buffer; e.g., at pH 7.2) may be used to dilute peak fractions or peak fraction pools. In certain embodiments, PBG buffer components may be added to a peak fraction or peak fraction pool to achieve final

concentrations of about 0.01 M to 1 .0 M sucrose, about 0.01 M to 1 .0 M phosphate, and about 0.001 M to 0.01 M glutamate. For example, PBG buffer components may be added to a peak fraction or peak fraction pool to achieve a final concentration of about 0.2 M sucrose, about 0.1 M phosphate, and about 0.005 M glutamate. Dilution of a peak fraction or peak fraction pool may be about a 1 :2 dilution, a 1 :3 dilution, a 1 :4 dilution, a 1 :5 dilution, 1 :6 dilution, a 1 :7 dilution, a 1 :8 dilution, a 1 :9 dilution, or a 1 :10 dilution, for example. In certain embodiments, a diluted centrifuge peak fraction or a diluted centrifuge peak fraction pool can be sampled for potency and/or bioburden, as described, for example in U.S. Patent No. 8,247,207. In some

embodiments, dilution may be performed simultaneously with or prior to addition of a stabilizer, as described below. In certain embodiments, dilution is performed before sterilization, as described below.

Sterilization, stabilization and blending In some embodiments, a viral harvest (e.g., further clarified viral harvest) is sterilized after centrifugation. Sterilization can be performed as a terminal filtration step and/or using one or more other sterilization methods. Methods useful for the sterilization of vaccine components (e.g., viruses) include, but are not limited to, irradiation, filtration, chemical treatment, and other suitable procedures. In some embodiments, a further clarified viral harvest is sterilized by filtration. Filtration methods useful for sterilization include, but are not limited to, single pass, dead-end, direct flow filtration (DFF) and tangential flow filtration (TFF), some of which are described above, using, for example, one or more sterilization grade filters (e.g., pore size of about 0.2 microns). In some embodiments, a method provided herein further comprises combining a sterilized viral harvest with a stabilizer. In some embodiments, the sterilized viral harvest is combined with a stabilizer to obtain a final concentration of 6-8% sucrose weight/volume (w/v), 1 -2% arginine w/v, 0.05-0.1 % monosodium glutamic acid w/v and 0.5-2% gelatin hydrolysate. In some embodiments, the final concentration is 6.84% sucrose weight/volume (w/v), 1.21 % arginine w/v, 0.094% monosodium glutamic acid w/v, and 1 % gelatin hydrolysate. In some

embodiments, the sterilized viral harvest is combined with a stabilizer to obtain a final concentration of 6-8% sucrose weight/volume (w/v), 1-2% arginine w/v, and 0.5-2% gelatin hydrolysate. In some embodiments, the final concentration is 6.84% sucrose weight/volume (w/v), 1.21 % arginine w/v, and 1 % gelatin hydrolysate. In some embodiments, a viral harvest is combined with a stabilizer (or certain components of a stabilizer) prior to, during or after sterilization.

In some embodiments, a method provided herein further comprises blending a sterilized viral harvest with at least one other sterilized viral harvest, thereby producing a blended viral harvest. In some embodiments, a sterilized viral harvest is blended with two other sterilized viral harvests, thereby producing a trivalent blended viral harvest. A trivalent blended viral harvest may comprise two influenza A viruses and one influenza B virus, or may comprise one influenza A virus and two influenza B viruses. In some embodiments, a sterilized viral harvest is blended with three other sterilized viral harvests, thereby producing a quadrivalent blended viral harvest. A quadrivalent blended viral harvest may comprise two influenza A strains and two influenza B strains; three influenza A strains and one influenza B strain; or one influenza A strain and three influenza B strains. Vaccine formulations

The methods provided herein can be useful for the production of liquid vaccines (e.g., live attenuated influenza virus vaccines), and formulations thereof, that are substantially stable at temperatures ranging from 4°C and 8°C. For example, liquid vaccine formulations produced by the methods herein are substantially stable at temperatures ranging from 2-8°C or at 4°C for a period of at least 1 month, or at least 2 months, or at least 3 months, or at least 4 months, or at least 5 months, or at least 6 months, or at least 9 months, or at least 12 months, or at least 18 months, or at least 24 months, or at least 36 months, or at least 48 months, in that there is an acceptable loss of potency (e.g., influenza virus potency loss) at the end of such time, for example, a potency loss of between 0.5-1.0 logs or a potency loss of less than 10%, or less than 20%, or less than 30%, or less than 40%, or less than 50%, or less than 60%, or less than 70%, or less than 80%, or less than 90%. In some embodiments, a liquid vaccine formulation produced by the methods herein has a potency loss of less than 1 .0 logs when stored for a period of 3 months at 4°C to 8°C. In some embodiments, a liquid vaccine formulation produced by the methods herein has a potency loss of less than 1.0 logs when stored for a period of 6 months at 4°C to 8°C. In some embodiments, a liquid vaccine formulation produced by the methods herein has a potency loss of less than 1 .0 logs when stored for a period of 12 months at 4°C to 8°C. In some embodiments, a liquid vaccine formulation produced by the methods herein has a potency loss of less than 1 .0 logs when stored for a period of 3 to 12 months at 4°C to 8°C. In some embodiments, a liquid vaccine formulation produced by the methods herein has a potency loss of less than 1 .0 logs when stored for a period of 6 to 12 months at 4°C to 8°C. In some embodiments, a liquid vaccine formulation produced by the methods herein has a potency loss of less than 1 .0 logs when stored for a period of 3 to 6 months at 4°C to 8°C. Viral potency (and potency loss) can be measured, for example, by TCID₅₀ or

Fluorescent Focus Assay (FFA)).

In some embodiments, liquid vaccine formulations comprise live influenza viruses. For instance, formulations may comprise one or more of the following: an attenuated influenza virus, a cold-adapted influenza virus, a temperature-sensitive influenza virus, an attenuated cold- adapted temperature sensitive influenza virus, an influenza A virus, and an influenza B virus.

In some embodiments, liquid vaccine formulations comprise one or more stabilizers which may include, for example, one or more of the following: arginine (e.g., 0.5-1 %, 1 -2%; 1 %; 1 .2%; 1.5%, 0.75-2%); poloxamer; sucrose (e.g., 2-8%; 2%; 6-8%; 3%; 4%; 5%; 6%; 7%, or 8%); hydrolyzed gelatin (e.g., 1 %; 0.5-2%; 1.5%; 0.5%; 0.75%); and glutamate (e.g., 0.05-0.1 %, 0.02-0.15%, 0.03%, 0.04%, 0.06%, 0.02-0.3%, or 0.094%). Certain formulations also may comprise one or more buffers such as, for example, one or more of the following: phosphate buffer (mono or dibasic or both) (e.g., 10-200mM, pH 7-7.5; 100 mM, pH 7.2; 100 mM, pH 7- 7.3); potassium phosphate (e.g., at least 50 mM, or at least 100mM, or at least 200mM, or at least 250mM); and histidine buffers (e.g., 25 - 50 mM histidine, pH 7-7.5; 50- 100mM histidine, pH 7-7.5; at least 50 mM histidine, or at least 100mM histidine, or at least 200mM histidine, or at least 250mM histidine). In some embodiments, vaccine formulations comprise one or more of the following in the final formulations: sucrose: 6-8% weight/volume (w/v); arginine monohydrochloride 1 -2% w/v;

glutamic acid, monosodium monohydrate 0.05-0.1 % w/v; gelatin hydrolysate, porcine Type A (or other sources) 0.5-2% w/v; potassium phosphate dibasic 1-2%; and potassium phosphate monobasic 0.25-1 % w/v. In some embodiments, vaccine formulations comprise one or more of the following: sucrose: 6.84% weight/volume (w/v); arginine monohydrochloride 1 .21 % w/v; glutamic acid, monosodium monohydrate 0.094 w/v; gelatin hydrolysate, porcine Type A (or other sources) 1 % w/v; potassium phosphate dibasic 1 .13%; and potassium phosphate monobasic 0.48% w/v. In some embodiments, vaccine formulations comprise all of the following: sucrose: 6.84% weight/volume (w/v); arginine monohydrochloride 1.21 % w/v; glutamic acid, monosodium monohydrate 0.094% w/v; gelatin hydrolysate, porcine Type A (or other sources) 1 % w/v; potassium phosphate dibasic 1 .13%; and potassium phosphate monobasic 0.48% w/v. In some embodiments, vaccine formulations comprise all of the following (within 10% variation of one or more component): sucrose: 6.84% weight/volume (w/v); arginine monohydrochloride 1.21 % w/v; glutamic acid, monosodium monohydrate 0.094% w/v; gelatin hydrolysate, porcine Type A (or other sources) 1 % w/v; potassium phosphate dibasic 1.13%; and potassium phosphate monobasic 0.48% w/v. In some embodiments, vaccine formulations comprise all of the following (within 10% variation of one or more component): sucrose: 6.84% weight/volume (w/v); arginine monohydrochloride 1.21 % w/v; gelatin hydrolysate, porcine Type A (or other sources) 1 % w/v. In such embodiments, formulations are in a buffer (e.g., a potassium phosphate buffer (pH 7.0-7.2)). In some embodiments, vaccine formulations may comprise trace amounts of EDTA. In some embodiments, vaccine formulations may comprise no EDTA.

FluMist^®

Numerous types of influenza vaccine can be produced. For example, FluMist^® is a live, attenuated vaccine that protects children and adults from influenza illness. In certain

embodiments, the methods and compositions herein may be adapted to, or used with, production of FluMist^® vaccine. In some embodiments, the methods and compositions herein are adaptable to production of similar or different viral vaccines and their compositions. FluMist^® vaccine strains typically contain, for example, hemagglutinin (HA) and neuraminidase (NA) gene segments derived from the wild-type strains to which the vaccine is addressed along with six gene segments, PB1 , PB2, PA, NP, M and NS, from a common master donor virus (MDV), also referred to herein as a donor strain or backbone strain. Influenza A strains of FluMist^® can include, for example, MDV-A as the master donor virus. MDV-A was created by serial passage of a wild-type A/Ann Arbor/6/60 (A/AA/6/60) strain in primary chicken kidney tissue culture at successively lower temperatures (see e.g., Maassab (1967) Nature 213:612-4). MDV-A replicates efficiently at 25°C (ca, cold adapted), but its growth is restricted at 38°C and 39°C (ts, temperature sensitive). Additionally, this virus does not replicate in the lungs of infected ferrets (att, attenuation). The ts phenotype is believed to contribute to the attenuation of the vaccine in humans by restricting its replication in all but the coolest regions of the respiratory tract. The stability of this property has been demonstrated in animal models and clinical studies. In contrast to the ts phenotype of influenza strains created by chemical mutagenesis, the ts property of MDV-A does not revert following passage through infected hamsters or in shed isolates from children (see e.g., Murphy & Coelingh (2002) Viral Immunol. 15:295-323).

Clinical studies in over 20,000 adults and children involving 12 separate 6:2 reassortant strains have shown that these vaccines are attenuated, safe and efficacious (see e.g., Belshe et al. (1998) N. Engl. J. Med. 338:1405-12; Boyce et al. (2000) Vaccine 19:217-26; Edwards et al. (1994) J. Infect. Dis. 169:68-76; Nichol et al. (1999) JAMA 282:137-44). Reassortants carrying the six internal genes of MDV-A and the two HA and NA gene segments of a wild-type virus (i.e., a 6:2 reassortant) consistently maintain ca, ts and att phenotypes (see e.g., Maassab et al. (1982) J. Infect. Dis. 146:780-900). Certain systems and methods described previously are useful for the rapid production in cell culture of recombinant and reassortant influenza A and B viruses, including viruses suitable for use as vaccines, including live attenuated vaccines, such as vaccines suitable for intranasal administration (e.g., FluMist^®). Certain methods provided herein, are optionally used in conjunction with or in combination with such cell culture methods involving, e.g., reassortant influenza viruses for vaccine production to produce viruses for vaccines in a more stable, consistent and efficient manner. Examples

The examples set forth below illustrate certain embodiments and do not limit the technology. In some instances, methods presented in the examples utilize certain materials and/or equipment, however such methods may be performed using any other suitable materials and/or equipment. Accordingly, the methods herein are not limited by the particular materials and equipment described below.

Example 1: Concentration of clarified harvest fluid by tangential flow filtration (TFF)

This example describes certain improvements to the purification process for influenza viruses. Improved purification methods are described for a live attenuated influenza virus monovalent bulk (LAIV-MB) manufacturing process, however such methods may be applied to any influenza virus manufacturing process. Improvements include introduction of a tangential flow filtration (TFF) step, as illustrated in FIG. 1 and described in detail below.

Several studies of varying scale were performed to evaluate inclusion of a TFF process in a LAIV-MB manufacturing process and are described in further detail below. Introduction of an additional step in a viral purification method can present risks to overall viral yield, potency and/or stability (e.g., product loss in media, wash buffer and/or on equipment, product precipitation, deactivation of live viruses, and the like). For example, shearing as part of a TFF process may compromise the integrity, potency, and/or stability of a virus. Live viruses can be particularly vulnerable to such processes. Additionally, manipulation of a viral harvest in the presence of allantoic fluid can present certain challenges (e.g., due to the viscosity of allantoic fluid) which may negatively affect viral yield, potency and/or stability. However, the studies described below demonstrated that the introduction of TFF was beneficial to a LAIV-MB manufacturing process in that viral recovery increased by 15% or greater (without substantially affecting viral potency and/or stability), harvest fluid processing capability increased and ultracentrifugation process time decreased.

Study #1: Concentration of clarified harvest fluid of cold-adapted influenza virus using tangential flow filtration (TFF) with hollow fiber cartridges The purpose of this study was to compare the effectiveness and robustness of tangential flow filtration (TFF) in concentrating a clarified harvest fluid (CHF) in the manufacturing process for cold-adapted influenza vaccine - monovalent bulk (CAIV-MB). A TFF concentration step was introduced prior to an ultracentrifugation step, in an approximately 10-fold scaled down model of an existing CAIV-MB manufacturing process. During a typical TFF process, product-containing fluid is passed tangentially to the filter membrane at a fixed shear rate. The difference in pressure between the inlet and outlet can be controlled to provide a constant driving force for the filtration to occur. Depending on the choice of the pore size of the filter membrane, the product is collected either in permeate or in retentate.

In this study, the concentration of CHF by TFF with a hollow fiber (HF) membrane cartridge of pore size of 500 kD and a fiber ID of 0.5 mm was investigated. Two types of HF cartridges were tested (e.g., single use HF cartridges from Spectrum Labs and HF cartridges from GE

Healthcare). Each pass of the CHF through the HF cartridge resulted in a portion of the impurities that were smaller than 500 kD to pass through the filter pores as permeate, while the virus particles (i.e., bigger than 500 kD) were retained in the retentate. Thus, each pass through the filter module resulted in a concentration of virus particles present in the CHF. The process was continued until a desired concentration of CHF volume was obtained. The process parameters of shear and transmembrane pressure (TMP) were not optimized for this study. However, to minimize the possibility of fouling the cartridge, a shear rate of 10,000 ±1 ,000 s^"1 was chosen. The TMP was chosen as 20 ± 2 psig based on prior TFF studies. To test for process consistency with strain variation, four different CAIV strains were used (e.g., A/Uruguay/716/07, A/South Dakota/6/07, B/Florida/4/2006 and B/Malaysia/2506/04).

Observations and results from this study are presented below.

Materials and equipment

Materials used in this study included the following: 1X SP (Sucrose Phosphate) Buffer

(HYCLONE, Logan, UT, Cat No. SH3A1796.01 ); 20 L Labtainer™ BIOPROCESS Container (HYCLONE, Logan, UT, Cat No. SH30709.03); 50 L Flexboy^® Assembly (Stedim Biosystems, Cat No. FBP50141 ); Polycarbonate Connector, ½ " HB non-valved body (Qosina, Edgewood, NY, Cat. No. MPX17803); Polycarbonate Connector, 3/8" HB non-valved body (Qosina, Edgewood, NY, Cat. No. MPC17006T03); Polycarbonate Connector, ½ " HB non-valved Insert (Qosina, Edgewood, NY, Cat. No. MPX22803-M); Polycarbonate Connector, 3/8" HB non- valved Insert (Qosina, Edgewood, NY, Cat. No. MPX22603-M); Polysulfone MPX Cap Body with Lock, ½ " ID (Qosina, Edgewood, NY, Cat. No. MPXK32003); Polysulfone MPX Cap Body with Lock, 3/8" ID (Qosina, Edgewood, NY, Cat. No. M PC2206T03M); Platinum Cured Silicone Tubing, Masterflex™ l/P 73 Tubing, 3/8" ID x ½ " OD, (Cole Parmer Instrument Co., Vernon Hills, IL, Cat. No. 96410-73); Pharmed^® BPT, l/P 82 Tubing, ½ " ID x 3/4" OD, 0.125" wall thickness (Saint Gobain Performance Plastics, Cat. No. AY242038); ca A/Uruguay/718/07 (Lot nos. PD-08Jul08, PD-22Jul08 & PD-21Aug08); ca A/South Dakota/6/07 (Lot no. PD-01Aug08); ca B/Malaysia/2506/04 (Lot no. PD-07 Aug08); ca B/Florida/4/2006 (PD-14Aug08);

SARTOCLEAN CA, Sterile filter Capsule (Sartorius, Edgewood, NY, Cat No. 5621304E0-OO); Hollow Fiber Module, 500-kD, 0.5 mm Fiber ID (GE Health Care, Piscataway, NJ, Cat. No. UFP- 500-C-5A, S/N: 91982101 153); and Hollow Fiber Modules, 500-kD, 0.5 mm Fiber ID, (Spectrum Labs, Rancho Dominguez, CA, Cat. No.: M6-500S-100-01 S, M7-500S-100-01 N & M8-500S- 300-01 N).

Equipment used in this study included the following: Bio Safety Cabinet (Baker Co., Stanford, MN, Model: STERILGARD III Advance); Refrigerator (Thermo Forma, Marietta, OH Model: 3787); Pipet aid (VWR International, Brisbane, CA, Cat. No.: 14006-026); Flex Stand System (GE Healthcare, Piscataway, NJ, Cat. No.: FS01 S); Pressure Gauge, 0-60 psig (Anderson Instrument Co., Fultonville, NY, Cat. No.: 3004300); Weigh Balance, 0-35 kg (Sartorius,

Edgewood, NY, Model: EB35EDE-1 ); Peristaltic Pump (Spectrum Labs, Rancho Dominguez, CA, Model: KROSFLO, MINIKROS Pilot system); and Viscometer, 0.2 - 5cP (Cambridge Viscosity, Medford, Ma, Model: VISCOLAB 450).

Methods

Each CAIV strain was propagated and harvested essentially as described in U.S. patent no. 8,247,207 to provide the clarified harvest fluid (CHF) material for TFF experimental runs that were performed. The TFF set up included a peristaltic pump, a Flex Stand, an HF cartridge with appropriately sized permeate and retentate lines connected to appropriate containers and analog pressure gauges to measure pressure at the inlet and outlet of the HF cartridge (FIG. 2). The feed and retentate pressure were monitored and

controlled. The permeate pressure was neither monitored nor controlled. Because the permeate pressure was equal to atmospheric pressure, the permeate pressure was zero and could be neglected from Equation 1 to give Equation 2 for calculating TMP. Equations 1 and 2 are presented in FIG. 3. Thus, the term TMP as used for this study refers to the average of the feed and retentate pressures.

Evaluation of HF cartridges

HF cartridges from GE Healthcare and Spectrum Labs were prepared and evaluated according to manufacturers' instructions. For example, for the evaluation of HF cartridges from GE Healthcare, the following procedure was followed for the preparation and use of the HF cartridges.

Flush HF cartridge 1 ) Flush the HF cartridge with clean water using a shear rate of 10,000 sec^"1 or greater with the permeate line closed and retentate line fully open in order to establish the cross flow before the permeate line is fully opened. Adjust the TMP to 20 ± 2 psig.

2) Gradually open the permeate valve to the fully open position while maintaining the TMP at 20 ± 2 psig. Flush the HF cartridge with clean water until filtrate volume is 2 mL/cm² or greater of surface area.

Module integrity testing

1 ) Close the filtrate side of the wetted module and system.

2) Use clean air or nitrogen gas to pressurize the retentate side of the module and system up to 5 psig.

3) If the pressure drop is greater than 0.5 psig/min, then there is a leak in the system connections. Tighten all connections, pressurize HF cartridge to 5 psig and check for pressure drop again.

4) After assuring system integrity, open the filtrate line.

5) If the pressure drop is less than 0.5 psig/min, the membrane is integral. Release pressure before use.

Equilibration with 1X sucrose phosphate (sp) buffer 1 ) Following the successful integrity testing of the HF module, connect a 1X SP buffer line to the TFF system.

2) Rinse and equilibrate the HF cartridge with 1X SP buffer following the same procedure as clean water flush step.

Procedure for concentration of CHF

Disconnect the 1X SP buffer line from the system and connect the CHF reservoir in its place.

1 ) With permeate valve closed and the retentate valve fully open, re-circulate the CHF through the HF cartridge using a shear rate of 10,000 to 1 1 ,000 sec^"1.

2) With the permeate line connected to a permeate collection vessel, gradually open the permeate valve to fully open position, while simultaneously adjusting the retentate valve, as necessary, to maintain the TMP at 20 ± 2 psig.

3) Monitor the permeate flow at the appropriate periodic intervals of time. The ultrafiltration is stopped when the desired concentration of CHF is achieved as indicated by the total amount of permeate collected.

Samples of permeate and retentate were collected and a Hemagglutination Assay (HA) was performed to confirm that the virus particles were present in the retentate only and not in the permeate. A negative plate reading on the HA from permeate, indicated the absence of the virus leakage. The retentate was passed downstream for ultracentrifugation if no virus particles were detected in the permeate.

Process parameters

To test the robustness of the process, four cold adapted strains of virus were used to evaluate four different TFF concentration runs with the 500-kD GE Healthcare and Spectrum Labs HF cartridges. A shear rate of 10,000 ± 1 ,000 sec^"1 and a TMP of 20 ± 2 psig were used. 3 to 5 fold concentrations of clarified harvest fluid (CHF) were performed. Loading volume per unit membrane area ranged from 58 L/m² to 160 L/m² depending on the volume of CHF that was available for each run. The table presented in FIG. 4 summarizes the process parameters for the six runs that were performed in this study. Sampling and analysis

Samples were withdrawn from the retentate bags using a 10 ml. syringe connected to the Luer- lock port on each bag. Before collecting the samples, the bags were inverted at least 10 times to ensure the contents were well mixed. Samples (15-30 ml.) were withdrawn from the bags using a 10 ml. syringe. 9 ml. of each sample collected was stabilized with 1 ml. of 10X SP before aliquotting 1 ml. into a 2 ml. CRYOVIAL. The samples were stored at -80 °C and submitted for potency testing (potency was measured by Fluorescent Focus Assay (FFA) analysis, essentially as described in U.S. patent no. 7,262,045).

Results and discussion

Concentration of CHF 3-5 fold by TFF resulted in little to no change in the density and viscosity of the concentrated CHF from that of the pre-concentrated CHF. The table presented in FIG. 5 summarizes the results of the TFF concentration step for the six runs in this study. The table presented in FIG. 6 characterizes the TFF concentration step. The concentration in volume is the ratio of the volume of CHF to that of TFF retentate. The concentration in potency is the ratio of potency in FFU/mL of the TFF retentate to that of potency in FFU/mL of the CHF. Although some impurities (e.g., ovalbumin and dsDNA) in the CHF may also concentrate together with CAIV virus, it typically does not impact the subsequent ultracentrifugation process. The recovery is the ratio of the product of potency in FFU/mL and volume of the TFF retentate to that of the product of potency in FFU/mL and volume of the CHF, expressed as a percentage. The HF cartridges from both manufacturers provided similar virus recoveries. The virus recovery from TFF concentration process is 83 to 100%. FIG. 7 and FIG. 8 show permeate flow rates versus time for the GE Healthcare and Spectrum Labs cartridges, respectively. Shear rates below 10,000 sec^"1, may have contributed to fouling of the filter membrane as evidenced by lower permeate flow rates and consequently longer processing time for B/Florida/4/2006 -PD14Aug08 (see FIG. 4 and FIG. 8). Accordingly, shear rates of 10,000 sec^"1 or greater (e.g., 10,000 sec^"1 to 18,000 sec^"1 or greater) may be necessary to prevent fouling and obtain high permeate flow rates.

Study #2: Characterization of process control parameters for concentration of clarified harvest fluid of live attenuated influenza virus by tangential flow filtration (TFF)

In study #1 described above, very high to complete product recoveries (85 - 100%) were observed when CHF of different LAIV strains were concentrated by TFF with HF cartridges of pore size 500 kDa. The little to no product losses observed with 500-kDa membrane pore size led to the selection of the 500-kDa pore size membranes for further process development of the TFF concentration procedure. Along with high product recoveries, another design objective of the TFF concentration process was the optimization of the filtrate flux rate. Optimizing filtrate flux rate can minimize the membrane area and operating time. For a given feed concentration, the filtrate flux rate can depend on certain operating parameters, for example: trans-membrane pressure (TMP), shear rate (s^"1) and temperature of the process fluid. In this study, ranges of the above operating parameters were investigated.

TMP provides the driving force for filtration to take place across the filter membrane.

Consequently, filtrate flux typically increases with increased TMP. However, the filtrate flux typically levels off after a certain TMP value. A typical flux vs. TMP curve is shown in FIG. 9. The first part of the curve where the flux increases with the TMP is the pressure dependent regime. The "plateau" phase of the curve where the flux is practically unaffected by increase in TMP is the pressure independent regime. In certain instances, the TFF process is operated at pressures corresponding to the initial onset of the pressure-independent region i.e., the TMP corresponding to the "knee" of the Flux vs. TMP curve. The filtrate flux rate also can be influenced by shear rate, which is a measure of the tangential or cross-flow. Higher cross-flow typically provides higher turbulence and shear that can improve the mass transfer of solutes from the membrane surface back to the bulk liquid. Temperature typically influences fluid viscosity and density and solute diffusivity with higher the operating temperatures often resulting in higher filtrate flux rate and vice versa.

In a preliminary parameter study (study #2A, described below), optimum TMP values for maximizing filtrate flux rates at various shear rates were determined at room temperature. In a second study (study #2B described below) non-optimal values of the operating parameters were evaluated to determine an acceptable range on the TMP and shear rate for regular operations.

Materials and equipment

Materials used for study #2A included the following: 1 X SP (sucrose-phosphate) buffer, (100 mM potassium phosphate and 200 mM sucrose) (HYCLONE, Logan, UT, cat no.

SH3A1796.01 ); 5 L Labtainer™ BIOPROCESS container (HYCLONE, Logan, UT, cat no.

SH30712.01 ); Polysulfone connector, 1/4" HB non-valved insert (Qosina, Edgewood, NY, cat. no. MPC22004T39M); Polysulfone connector, 3/8" HB non-valved insert (Qosina, Edgewood, NY, cat. no. MPC22006T39M); Polysulfone sealing cap with lock, 1/4" ID (Qosina, Edgewood, NY, cat. no. MPCK32039); Polysulfone in-line hose barb, 3/8" ID (Qosina, Edgewood, NY, cat. no. MPC17006T39); Polysulfone in-line hose barb, 1/4" ID (Qosina, Edgewood, NY, cat. no. MPC17004T39); Platinum cured silicone tubing, Masterflex™ L/S 24 tubing, (Cole Parmer Instrument Co., Vernon Hills, IL, cat. no. 96410-24); Platinum cured silicone tubing,

Masterflex™ L/S 36 tubing, (Cole Parmer Instrument Co., Vernon Hills, IL, cat. no. 96410-36); ca A/South Dakota/6/07 (lot no. PD-09Oct08); ca B/Malaysia/2506/04 (lot no. PD-07Oct08); ca A/Uruguay/718/07 (lot no. PD-17Oct08); ca B/Florida/4/2006 (lot no. PD-08Aug08);

SARTOPORE 2, 0.45/0.2μηι, sterile filter capsule (Sartorius, Edgewood, NY, cat no.

5441307H8G-OO); and MINIKROS Sampler, hollow fiber module, 500 kD, 0.5 mm fiber ID, 120 fiber count, 2/3 FL and 245 cm² (Spectrum Labs, Rancho Dominguez, CA, cat. no.: M5-500S- 220-01 N).

Materials used for study #2B included the following: 1 X PBS (phosphate buffered saline) buffer, (8.4 mM sodium phosphate, dibasic, 1.6 mM potassium phosphate, monobasic, 150 mM sodium chloride) (HYCLONE, Logan, UT, cat. no. SH3A1798.01 ); 5 L Labtainer™ BIOPROCESS container (HYCLONE, Logan, UT, cat no. SH30712.01 ); Polysulfone connector, 1/4" HB non- valved insert (Qosina, Edgewood, NY, cat. no. MPC22004T39M); Polysulfone connector, 3/8" HB non-valved insert (Qosina, Edgewood, NY, cat. no. MPC22006T39M); Polysulfone sealing cap with lock, 1/4" ID (Qosina, Edgewood, NY, cat. no. MPCK32039); Polysulfone in-line hose barb, 3/8" ID (Qosina, Edgewood, NY, cat. no. MPC17006T39); Polysulfone in-line hose barb, 1/4" ID (Qosina, Edgewood, NY, cat. no. MPC17004T39); Platinum cured silicone tubing, Masterflex L/S 73 tubing, (Cole Parmer Instrument Co., Vernon Hills, IL, cat. no. 96410-73); ca B/Brisbane/60/2008 (lot no. PD-10Mar10); ca A/Uruguay/716/07 (lot no. PD-17Mar10); ca A/California/07/09 (lot no. PD-24Mar10); ca A/Uruguay/716/07 (lot no. PD-30Mar10); ca A/California/07/09 (lot no. PD-06Apr10); SARTOPORE 2, 0.45/0.2μm, sterile filter capsule (Sartorius, Edgewood, NY, cat no. 5441307H8G-OO); MiniKros^® Plus, hollow fiber module, 500 kD, 0.5 mm fiber ID, 320 fiber count, and 1050 cm² (Spectrum Labs, Rancho Dominguez, CA, cat. no.: M7-500S-100-01 N); and MiniKros^® Plus, hollow fiber module, 500 kD, 0.5 mm fiber ID, 950 fiber count, and 3100 cm² (Spectrum Labs, Rancho Dominguez, CA, cat. no.: M7-500S- 300-01 N).

Equipment used for both study #2A and study #2B included the following: Biosafety cabinet (Baker Co., Stanford, MN, model: STERILGARD III Advance); Pipet aid (VWR International, Brisbane, CA, cat. no.: 14006-026); FlexStand™ system (GE Healthcare, Piscataway, NJ, cat. no.: FS01 S); Pressure gauge, 0-60 psig (Anderson Instrument Co., Fultonville, NY, cat. no.: 3004300); Weighing balance, 0-35 kg (Sartorius, Edgewood, NY, model: EB35EDE-1 );

Peristaltic pump (Spectrum Labs, Rancho Dominguez, CA, model: KROSFLO, MINIKROS Pilot system).

Methods

The TFF set up included a peristaltic pump, a FLEXSTAND, a hollow fiber (HF) cartridge with permeate and retentate lines connected to appropriate containers and analog pressure gauges to measure the pressure at the inlet and outlet of the HF cartridge. Each CAIV strain was propagated in specific pathogen-free (SPF) eggs. At the end of secondary incubation the inoculated eggs were de-capped and harvested. The harvested egg allantoic fluid was pooled to make the Pooled Harvest Fluid (PHF). The PHF was filtered using a SARTOPORE 2, 0.2/0.45-μηι filter to obtain the CHF. This process was used for the different CAIV strains (ca A/South Dakota/6/07, ca B/Malaysia/2506/04, ca B/Florida/4/2006 ca

A/Uruguay/716/07, ca A/California/07/09 and ca B/Brisbane/60/08) to provide CHF material for the TFF experimental runs that were performed.

Study #2A MINIKROS Sampler HF modules from Spectrum Labs, Inc. were used for this study. The modules were made up of 120 fibers with a fiber length of 13.9 cm with a membrane surface area of 245 cm². Four cold-adapted (ca) influenza strains, ca A/South Dakota/6/07, ca

B/Malaysia/2506/04, ca A/Uruguay/716/07 and ca B/Florida/4/2006 were used in this study.

Before each TFF optimization run, the HF cartridge was flushed with 500 ml. or greater (2 mL/cm² or greater of surface area) of clean water. The clean water flush was performed at a shear rate of 14,000 s^' After at least 500 mL of clean water was flushed through the HF cartridge, the feed flow rate was increased to correspond to a shear rate of 16,000 s^' The feed line was then disconnected from the clean water vessel and connected to the CHF bag. The optimization of the TMP was performed at flow rates corresponding to four different shear rates: 16,000 s^"1, 14,000 s^"1, 12,000 s^"1 and 10,000 s^' At each shear rate, the TMP was increased in 5 psig increments from 10 psig to 25 psig. At each TMP setting the permeate (filtrate) flow was allowed to stabilize for at least 5 minutes before recording the permeate flux data. The permeate flux rate was expressed as LMH (liters per square meter per hour).

Study #2B

Before each TFF optimization run, the HF cartridge was flushed with 4 L or greater (2 mL/cm² or greater of surface area) of clean water. The clean water flush was performed at a shear rate of 10,000 s^' The cartridge was then tested for integrity and flushed with at least 2 void volumes of 1X PBS buffer solution to equilibrate the cartridge with that buffer.

Subsequently, the entire TFF set up was moved to a 5 ± 3 °C environment or left at room temperature. With the permeate valve completely closed, the CHF was then re-circulated through the TFF set up and the required TMP was adjusted. The permeate valve was gradually opened while simultaneously restricting the retentate flow to maintain the desired TMP. The concentration process was continued until a 4- to 5-fold reduction in volume of the CHF was obtained. Subsequently, the concentration process was stopped and the system was flushed with 2 void volumes of 1X PBS buffer. The CHF/retentate bag was then disconnected from the system and the TFF concentration process was considered complete.

Results and discussion From study #2A, the permeate (filtrate) flux data collected for each CAIV strain at different shear rates were plotted against the TMP values (FIGS. 9-12). Higher shear rates provided higher permeate flux rates and vice versa. As discussed above, the inflection point or the "knee" on the Flux vs. TMP curves marks the onset of the pressure-independent region and may be an optimal TMP value, because there is no appreciable increase in the flux beyond that value. The optimal TMP values along with the corresponding flux values at each shear rate for each strain are summarized in the table presented in FIG. 13.

Also shown in the table presented in FIG. 13 are the average values of the TMP and the flux across the four strains. In FIG.14, the average TMP and flux values are plotted against the shear rate to depict the overall trend. In the shear rate range that was evaluated, the best flux rates were observed when the concentration process was operated at a shear rate range of 14,000 s^"1 to 16,000 s^"1 and a TMP range of 15-16 psig (the region indicated by the shaded area in FIG. 14, i.e., the range of TMP values (y-axis) for the TMP vs. shear rate curve that span the shaded region (along the x-axis), corresponding to the leveled region of the flux vs. shear rate curve). When the process was operated at higher shear rates the onset of the pressure independence of flux occurred at relatively lower TMP values than when the process was operated at lower shear rates. The overall trend observed was that high shear rates contributed to high flux rates. However, shear rates higher than 16,000 s^"1 were not evaluated because they require large pumps and piping and these additional pump and piping requirements offset the gains in filtrate flux rates realized at high shear rates. Therefore, the highest shear rates evaluated in the TFF process development studies were limited to values of 15,000 ± 1000 s^' Thus, this study demonstrated a shear rate in the range of 14,000 s^"1 to 16,000 s^"1 and a TMP in the range of 15 psig to 20 psig achieved high filtrate flux rates (e.g., 25 LMH or greater). In certain instances, a shear rate of 14,000 s^"1 or greater and a TMP of 15 psig achieved high filtrate flux rates.

The preliminary parameter study was followed by a pilot-scale study. In the pilot-scale study, outlined in FIG. 15, the TFF concentration process was operated at a shear rate of 14,000 s^"1 and a TMP of 15 psig. The first run was performed at room temperature (i.e., 18 to 20°C) and the average flux observed was 95 LMH. The other four runs were performed in a 2 to 8°C refrigerator. The lowest, highest and average fluxes for the four runs were 33, 85 and 59 LMH, respectively, for the 2 to 8°C refrigerator runs. The feasibility study conducted before the preliminary parameter study represents a departure from the shear rate and TMP values described above, and was performed using low shear rates (10,000 ± 1000 s^"1) and a high TMP of 20 psig. However, even at the low shear rates, filtrate flux rates were comparable to flux rates achieved in the pilot-scale run performed at room temperature. This indicated that the TMP has a greater effect on the filtrate flux rate than shear rate.

In order to further evaluate how filtrate flux is affected by low TMP values, low shear rates and low temperatures, a supplementary characterization study (Study #2B described above) was performed (FIG. 15). In the first part of the supplementary characterization study, the effect of temperature was evaluated - each batch of the CHF for the three LAIV strains was divided into two arms and the TFF concentration process was performed on both arms using the same operating parameters except for operating temperature: one arm was performed in a refrigerator (2 to 8 °C) and the other arm was performed at room temperature (18 to 20 °C). In general, the filtrate flux rates were lower at the colder temperature but the degree of decrease in filtrate flux rates did not show any specific pattern and may have been dependent on the LAIV strain.

In the second part of the supplementary characterization study, two runs (runs 7 and 8) were performed in a refrigerator and at a shear rate of 14,000 s^"1 or greater and a TMP of 15 psig (FIG. 15). These two runs in addition to run 5 of the pilot-scale study completed a data set that can be compared with the three runs (runs 1 , 3 and 5) in the cold temperature arm in the first part of the supplementary characterization study. A decrease in filtrate flux was observed at the lower shear rate of 10,000 s^"1 and at the lower TMP (10 psig). However, complete loss of filterability from fouling was not observed even in cold temperature conditions. The lowest average flux rate observed was 28 LMH. The results also indicated that a decrease in filtrate flux at low shear rate and low TMP values varied with the virus strain.

Additional details regarding the supplementary characterization study such as recovery, material balances and impurity profiles are presented in FIGS. 16-19.

Conclusions

The effects of shear rate and TMP on the filtrate flux rate for the TFF concentration of LAIV were evaluated in a preliminary parameter study. High flux rates were observed at a shear rate range of 14,000 to 16,000 s^"1 and a TMP range of 15 to 20 psig. Higher shear rates (greater than 16,000 s^"1) may contribute to higher flux rates but require larger pumps and piping to deliver the flow required to generate high shear rates. For the majority of the LAIV strains tested, increasing the TMP above 15 psig at shear rates of 14000 s^"1 or greater provided little to no benefit to filtrate flux rate. Therefore, a shear rate greater than 14,000 s^"1 and a TMP of 15 psig can generate high filtrate flux rates.

When the TFF process was operated at a 14,000 s^"1 shear rate and a TMP of 15 psig at the pilot-scale (5- to 10-fold lower CHF volumes compared to manufacturing volume), 3- to 4- fold concentrations in volume of CHF were achieved in less than 2 hours of processing time for all the LAIV strains tested. The load factor for the pilot-scale runs varied between 50-100 L of CHF per square meter of the HF filter surface area. In a supplementary characterization study, a decrease in filtrate flux rate was observed at low shear rates (10,000 s^"1), low temperatures (5- 12 °C), and low TMP values (10 psig). However, complete loss of filterability from severe fouling was not observed for any of the LAIV strains evaluated.

In summary, the following ranges on the process control parameters of a TFF concentration process were evaluated on a wide variety of LAIV strains and were found to result in filtrate flux rates that were 25 LMH or greater: shear rate of 10000 to 15000 s^"1; TMP of 10 to 20 psig; load factor of 50 to 100 L of CHF per square meter of the HF filter surface area; and ambient temperature of 5 to 20 °C.

Study #3: Evaluation of a 500 kDa hollow fiber cartridge for concentration of a clarified harvest fluid of live attenuated influenza virus

The tangential flow filtration (TFF) procedure detailed above uses disposable hollow fiber (HF) cartridges from Spectrum Labs for concentration of clarified harvest fluid (CHF) of live attenuated influenza virus (LAIV). In addition, alternate disposable HF cartridges were evaluated. In this study, disposable HF cartridges manufactured by GE Healthcare were evaluated for their suitability for the concentration of CHF.

Materials and equipment Materials and equipment used in this study included the following: 500 kDa, Polysulfone, 0.5 mm fiber ID, 1400 cm², 60 cm, hollow fiber cartridge (GE Healthcare, Piscataway, NJ, P/N: RTPUFP-500C-4X2MS); 500 kDa, Polysulfone, 0.5 mm fiber ID, 290 cm², 60 cm, hollow fiber cartridge (GE Healthcare, Piscataway, NJ, P/N: UFP-500C-3X2MA); 500 kDa, Polysulfone, 0.5 mm fiber ID, 140 cm², 30 cm, hollow fiber cartridge (GE Healthcare, Piscataway, NJ, P/N: UFP- 500C-3MA); Platinum cured silicone tubing, Masterflex™ l/P 73, 3/8" ID (Cole Parmer

Instrument Co., Vernon Hills, IL, Cat. No. 96410-73); Platinum cured silicone tubing,

Masterflex™ L/S 24, ¼" ID (Cole Parmer Instrument Co., Vernon Hills, IL, Cat. No. 96403-24); Flex Stand™ System (GE Healthcare, Piscataway, NJ, Cat. No. FS01 S); Milligard^® Opticap^® XL 5 1 .2 μπΊ filter (Millipore, Billerica, MA, cat. No. KW19A05HH1 ); SARTOPORE 2 0.8/0.45 μπι filter (Sartorius Stedim, Goettingen, Germany, Cat. No. 5445306G0-OO); 50 L Labtainer™ with quick connects (HYCLONE, Logan, UT, Cat No. SH30709.04); 10 L Labtainer™ with quick connects (HYCLONE, Logan, UT, Cat No. SH30709.02); 5 L Labtainer™ with quick connects (HYCLONE, Logan, UT, Cat No. SH30709.01 ); 2 L Polyethylene terephthalate glycol (PETG) modified bottle (Nalgene, Rochester, NY, Cat No. 2019-2000); 1 L Polyethylene terephthalate glycol (PETG) modified bottle (Nalgene, Rochester, NY, Cat No. 2019-1000); 1 X Phosphate buffered saline (PBS) without Ca²⁺ and Mg²⁺, pH 7.2 (Invitrogen, Grand Island, NY, Cat. No. 20012043); Peristaltic pump (Spectrum Labs, Rancho Dominguez, CA, Model: KROSFLO, MINIKROS; Pilot system); Peristaltic pump (Watson Marlow, Wilmington, MA, Model: 520Di); 2 to 8 °C 3-door refrigerators (Nor Lake Scientific, Hudson, Wl, Model: NSPR803WWG/0);

KrosFlo^® disposable pressure transducers (Spectrum Labs, Rancho Dominguez, CA, P/N: ACPM-499-03N); SciPres^® pressure monitor (Intelligent Bioprocessing Systems, Middleton, Wl, Cat. No. 080-690); and SCIPRES 3/8" barb pre-calibrated disposable pressure flow cells (SCILOG, Middleton, Wl, Cat. No. 080-694PSX-5).

Methods

Pooled harvest fluid (PHF) of four LAIV strains, A/Uruguay/716/07 (H3N2), A/California/07/09 (H1 N1 ), B/Florida/04/06 (Yamagata lineage), and B/Brisbane/60/08 (Victoria lineage), was produced and then clarified by filtering through a 1.2-μηη filter followed by a 0.8/0.45-μηι filter.

The clarified harvest fluid (CHF) was concentrated using a TFF process as described below:

1 ) Inside a refrigerator set at 2-8 °C, a GE HF cartridge (0.5 mm ID, 500 kDa

ultrafiltration membrane) was affixed to a GE Flex Stand™ with the inlet and outlet lines connected to the CHF/retentate bag and the permeate line connected to the permeate collection vessel. A peristaltic pump was used to pump CHF from the CHF/retentate bag or 1 X PBS from a bag or bottle through the HF cartridge. The HF cartridge was equilibrated using 1 X PBS (0.5 mL/cm² or greater of surface area).

2) CHF was circulated at the set shear rate and TMP. Once the flow was fully established, the permeate valve was gradually opened to a fully open position. The TMP was re-established to the set value by adjusting the retentate valve after the permeate valve was fully open.

3) Permeate flow was monitored using a flow meter (or a weight scale) and the inlet and retentate pressures were monitored using pressure gauges or sensors.

4) The concentration process continued until the CHF was concentrated by about 4 to 7- fold. After concentration, the pump flow was stopped and the inlet tube was connected to a bag or bottle containing 1X PBS and the cartridge and tubing were flushed with 1X PBS (0.5 mL/cm² or greater of surface area). The PBS flush was combined with the retentate.

5) Samples were taken from the retentate and permeate for a potency assay.

All TFF concentration processes were conducted in 2-8 °C refrigerators. Three sizes of 500 kDa GE HF cartridges with surface area of 140, 290, and 1400 cm² were used in the

experiments. Shear rates from 8000 to 18000 s^"1 and TMP from 9 to 18 psi were evaluated at load factors 200 to 31 1 L/m² and concentration factors of 4.7 to 7.2. In certain instances, the average permeate flux was evaluated by setting the TFF process parameters to shear rates at 8000 s^"1, 10000 s^"1, 14000 s^"1, 16000 s^"1, or 18000 s^"1 and the TMPs at 9 psi, 10 psi, 12 psi, 15 psi, or 18 psi. Results and discussion

FIG. 20 presents a table summarizing experimental conditions of the TFF processes and the corresponding average permeate fluxes. FIG. 21 shows a contour plot of shear rates, TMPs, and average permeate fluxes. FIGS. 22A and 22B present a table summarizing potency assay data of the GE HF TFF process.

In a total of 24 TFF processes, the CHF of four different LAIV strains (A/Uruguay/716/07, A/California/07/09, B/Florida/04/06, and B/Brisbane/60/08) was concentrated 4.7 to 7.2- fold with shear rates ranging from 8,000 to 18,000 s^"1 and TMPs ranging from 9 to 18 psi. An average permeate flux of 50 LMH or greater was observed for 22 of the TFF processes. An average permeate flux of 41 and 48 LMH was observed during the concentration of

B/Brisbane/60/08 at a shear rate of 8,000 s^"1 and TMP of 10 psi (run 1 ), and during the concentration of B/Florida/04/06 at a shear rate of 10,000 s^"1 and TMP of 9 psi (run 2), respectively.

The results, summarized in FIG. 21 , show that when the TFF process is operated at a shear rate of 10,000 s^"1 or greater, an average permeate flux of 50 LMH or greater was achieved with a TMP ranging from 9 to 18 psi. This indicates that 200 L of CHF can be concentrated 4 to 7- fold within 2.6 hr using a 1.15 m² HF cartridge. FIG. 21 also shows an optimal operation region with shear rate ranging from 1 1 ,000 to 18,000 s^"1 and TMP ranging from 10 to 14 psi, in which an average permeate flux of 58 LMH or greater, was obtained. When operated at shear rate of 16,000 s^"1 and TMP of 12 or 13 psi, the highest average permeate flux of 64 LMH or greater was achieved. Certain parameters, e.g., optimal shear rate, can vary depending on the TFF materials used. For example, column packing density and/or membrane pore density can vary amongst TFF cartridge manufacturers, and parameters such as shear rate can be adjusted accordingly.

The potency assay results showed that the potency of the permeate from all the 24 TFF processes was less than 3.3 log-io FFU/mL (below the assay detection limit), which indicated that no virus leaked through the GE HF membrane.

Conclusion 500 kDa GE HF cartridges were tested for concentration of four LAIV strains,

A/Uruguay/716/07, A/California/07/09, B/Florida/04/06, and B/Brisbane/60/08, with the shear rates ranging from 8,000 to 18,000 s^"1 and TMPs ranging from 9 to 18 psi. The results showed that the permeate flux was 50 LMH or greater when the TFF process was operated at shear rates of 10,000 s^"1 or greater and at TMPs ranging from 9 to 18 psi. Leakage of virus through the HF membrane was not observed. Therefore, HF cartridges from different manufacturers are suitable for the concentration of CHF.

Study #4: Scale-up In this study, the TFF process described above was applied to larger volumes of CHF.

Equipment and material engineering studies Engineering studies to test a disposable TFF tubing assembly (e.g., for use with a SCIPURE 200 TFF system) were performed. High vibrations occurred when the feed pump (Watson Marlow 720 series pump that is rated for a maximum flow rate of 25 LPM with 19 mm ID tubing) was operated at flow rates of 16.3 LPM or greater, which corresponds to a shear rate of 12000 s^"1 or greater for the GE HF cartridge. Additionally, at high rotations per minute of the pump rollers, the pump tubing experienced significant wear. Therefore, operational durability of the tubing was tested at a flow rate of 16.3 LPM and 15 psi TMP.

The acceptance criteria for selection of tubing are a minimum operation time of 6 hours without damage or imminent signs of damage, animal derived component free (ADCF) manufacture and low degree of spallation. From this study Pure Weld^® tubing was chosen and displayed no imminent signs of damage on the outer or inner wall after 6.5 hours of operation and had a low level of tubing degradation (spallation). Furthermore, to minimize vibrations, the process was best operated at flow rates that corresponded to shear rates of 1 1 ,000 ± 1 ,000 s^"1 and TMP of 13 ± 1 psi. This indicates that while the TFF column can be run over a wide range of conditions, the actual run conditions may be limited by other equipment limits. For example, use of alternative tubing and/or pump system could expand and/or shift the operating range.

Pilot-scale studies A pilot-scale study was performed to evaluate the scalability of the TFF process described above with GE HF cartridges using the SciPure^® 200 system. A total of four runs with four LAIV strains (ca B/Brisbane/60/2008 (Victoria lineage), ca A/Victoria/361/201 1 (H3N2), ca

B/Wisconsin/01/2010 (Yamagata lineage) and ca A/California/07/2009 (HIN1 )) were performed with CHF volumes ranging from 34 to 49.5 L. FIG. 23 presents a table summarizing the results. No virus was detected in the permeate samples (which confirms that the membrane was integral and adequately sized), and a 10-fold concentration was performed in less than 2 hours of processing time for all four runs.

Characterization of concentrated CHF loading flow rate into an ultracentrifuge Several studies were conducted to evaluate the impact of loading flow rate of concentrated CHF on virus recovery in a sucrose-gradient ultracentrifugation process, the results of which are described below and summarized in FIG. 24, and FIGS. 70-74.

Methods

A Hitachi CP40Y continuous-flow ultracentrifuge with a 3.2 L CT40C rotor-core was utilized. Operation of the CP40Y and 3.2 L CT 40C rotor core were conducted according to

manufacturer's instructions.

The stationary rotor was completely filled with approximately 3.2 L of PBS through the bottom port of the ultracentrifuge. A portion of the PBS was then displaced with 1 .6 L of 10% sucrose followed by 1 .2 L of 60% sucrose. After the 60% sucrose was pumped into the rotor, the top and bottom tubing were clamped off, and the rotor was immediately accelerated to 7000 rpm. When the rotor reached 7,000 rpm, PBS was pumped into the rotor at 100 -120 mL/min through the bottom port of the ultracentrifuge. Once the centrifuge tubing pressure was stable, the flow rate was increased and the rotor was accelerated to 35,000 rpm.

The CHF that was concentrated approximately 3 or 5-fold in volume by TFF was loaded into the ultracentrifuge with the flow rates ranging from 75 to 275 mL/min as shown in FIG. 70. The flow-through was collected when concentrated CHF was pumped into the

ultracentrifuge. After all the concentrated CHF was pumped into the ultracentrifuge, the virus banding started with PBS flow at 100 -120 mL/min for one hour. At the end of the banding, the bottom inlet line of the centrifuge was clamped, and the rotor was decelerated to 7000 rpm under normal brake. Once the speed reached 7000 rpm, the rotor was stopped in free coasting mode. Once the rotor completely stopped, the sucrose gradient was offloaded at 100 mL/min from the bottom port of the ultracentrifuge. The sucrose gradient was collected into three pools according to certain cut-off densities. The sucrose gradient with a density greater than 1.2276 g/cm³ (greater than 49.2% sucrose solids) was directed into a high sucrose density pool (P1 ) biotainer. The virus peak within a density range from 1 .1525 to 1 .2276 g/cm³ (equivalent to 34.8% - 49.2% sucrose solids) was collected into a virus peak pool (P2) biotainer. The sucrose gradient with density less than 1.1525 g/cm³ (less than 34.8% sucrose solids) was collected in a low sucrose density pool (P3) biotainer. The purified and concentrated virus in the P2 biotainer was further processed to produce a diluted centrifuge pool (DCP) and monovalent bulk (MB).

Results

The amount of virus that was captured in the sucrose gradient ranged from 97.5% to 76.1 % when the loading flow rate of the TFF concentrated CHF varied from 75 to 275 mL/min (the amount of virus lost in the flow-through ranged from 2.5% to 23.9%, respectively). Less virus was lost and, thus, higher virus recovery was observed at lower loading flow rates. A virus recovery of 90.0% or greater was achieved when a loading flow rate ranging from 120 to 160 mL/min was applied. For example, the total time to process 13.4 L of concentrated CHF at 140 mL/min was 1.6 hour, and 96.8% of virus capture in the sucrose gradient was achieved. At the same loading flow rate, a 6.4-hour centrifugation would be able to process 160 L of CHF after 3-fold concentration by TFF, which, under normal operating conditions without TFF process, would require 9.2 -1 1 .1 hours of ultracentrifugation process with a flow rate range of 240 - 290 mL/min. Thus, loading concentrated CHF at lower loading flow rate could shorten the ultracentrifugation process time without compromising virus capture.

The percentage of virus loss was calculated base on the amount of virus loaded into the ultracentrifuge and the amount present in the flow-through (FIG. 73). The virus lost in the flow- through ranged from 2.5% to 23.9% as the loading flow rate of the concentrated CHF varied from 75 to 275 mL/min, respectively. Increasing the loading flow rate led to a higher loss of virus in the flow-through as indicated by a linear fit of the data with 95% confidence (FIG. 72). More than 16% of virus was lost in the flow-through when the flow rate was higher than 200 mL/min. Therefore, a lower loading flow rate was chosen to achieve higher virus capture in the sucrose gradient during centrifugation of the

concentrated CHF.

Following ultracentrifugation, the centrifuge pool was diluted and filtered through a 0.2 μηη filter to obtain the monovalent bulk (MB). The impurity content of the MB solution met in-process control (IPC) specifications according to current IPC limits for CHF that has been concentrated to 5.3-fold by TFF concentration, without affecting the IPC specifications for the impurity content of the MB. In certain instances, a maximum 5 fold concentration limit is applied for a TFF process (e.g., 200 L CHF maximum volume concentrated to minimum TFF retentate volume of 40L).

The operational parameters for TFF concentration were selected based on results from previous studies and are shown in FIG. 25. TFF engineering and validation runs (full-scale batches)

As described above, development studies were conducted to evaluate the suitability for introducing a TFF process and reducing centrifuge flow rate for an existing monovalent bulk manufacturing process. Full-scale engineering and validation runs also were conducted which further established the benefits of a TFF concentration step at full-scale. One engineering batch was run followed by three consistency batches (i.e., process validation). All TFF batches were maximized for harvest volume to ensure the TFF process consistently performed with increased harvest volumes as intended for commercial production. The monovalent bulk (MB) lots for the full-scale runs included 2012/2013 influenza strains: A/Victoria/361/201 1 , B/Wisconsin/1/2010 and A/California/07/2009.

The engineering and validation batches demonstrated the following:

• The engineering batch confirmed the process performs as intended in the commercial scale production environment and provided the confidence that the process can be taken forward for validation.

• The TFF process validation batches demonstrated that it was possible to consistently produce monovalent bulks that met pre-determined specifications and quality attributes when compared to historical data obtained from monovalent bulks produced without the TFF process.

• All process parameters were consistent and reproducible and within their acceptable ranges.

Routine and non-routine QC testing

Full release testing was performed on the drug substance (DS) consistency batches including characterization and stability tests. All three consistency lots met certain acceptance criteria, in- process control limits, and in-process control test specifications. In addition, all three lots of bulk drug substance met the acceptance criteria for all the MB release tests. The results from the characterization tests were comparable to certain non-TFF derived lots and certain tests performed on non-TFF strains. During the engineering and validation batches, in-process samples at a number of process stages were collected and analyzed for impurities such as ovalbumin, total protein and DNA in addition to product recovery based on potency analysis. This testing was performed and verified that the introduction of the TFF step does not adversely impact the down-stream purification steps. In addition to process validation, a separate comparability study was performed to review specific performance attributes of the virus that were additional, non-release attributes which were not examined as part of the process validation. Results were compared to those from clinical trial material and additional characterization testing performed on equivalent strains of non-TFF batches, and it was confirmed that the TFF process did not adversely affect the drug substance.

Engineering run and process results summary

The successful execution of the engineering and validation runs confirmed that TFF can be scaled up with similar yield improvements and process impurity removal as demonstrated for smaller batches. The number of doses created for each virus strain (TFF and non-TFF batches) is shown in FIGS. 26-28. The figures show the mean and range profiles for the number of doses generated for each non-TFF batch compared with individual TFF batches for each strain. Data for impurity performance from PHF to monovalent bulk for each TFF batch is compared with commercial non-TFF batches of the same strain in FIGS. 29-31 .

The results for the number of doses generated for the TFF batches when compared to commercial non-TFF batches demonstrate an increase in the number of available doses for A/Victoria, A/California and B/Wisconsin strains (70%, 22% and 6% yield improvements respectively).

Certain results presented in this Example were obtained using a 500 kD hollow fiber TFF cartridge. Similar results (data not shown) were obtained using a 750 kD hollow fiber TFF cartridge. Example 2: Modified clarification methods

This example describes certain improvements to the purification process for influenza viruses. Improved purification methods are described for a live attenuated influenza virus monovalent bulk (LAIV-MB) manufacturing process, however such methods may be applied to any influenza virus manufacturing process. Improvements include clarification using modified filtration methods described in detail below. Pre-filter for improved clarification of an influenza virus

The manufacture of a refrigerator-stable, liquid formulation of an influenza virus drug substance often includes a sequence of downstream processing steps: clarification filtration,

ultracentrifugation, virus peak pool dilution, sterile filtration, and final freezing of the drug substance. The drug substance is also called the monovalent bulk (MB), which contains a single strain of the cold adapted, live-attenuated, influenza virus (CAIV).

Certain clarification filtration processes include filtering pooled virus harvest fluid (PHF) through a 1 .2 μηι filter (e.g., Milligard^®) followed by a 0.8 μηι /0.45 m filter (e.g., SARTOPORE 2). A single filtration rig can include two-10 inch 1.2 μηη filters (1 .6 m² total effective filter area) in parallel and one-20 inch 0.8 μηη /0.45 μηη filter (1 .2 m² total effective filter area) in succession. At a production scale of 22,500 eggs, for example, filter clogging near a production volume of 80 to 90 liters (L) can occur for certain CAIV strains. To increase batch scale and to improve clarification performance, various pre-filters were screened for their ability to filter an estimated target volume of 160 L through a single filtration rig (equivalent to filterability of 100 L/m²). This study was designed to first assess the performance of a pre-filter on pre-clarification filtration of CAIV, and then to evaluate the effectiveness of using the pre-filter in facilitating the filterability (e.g., clarification throughput) of CAIV in the subsequent clarification filtration process, and the potency change after clarification filtration through the 1 .2 μηη filter followed by the 0.8 μηη /0.45 μηι filter.

In this investigation, four pre-filters of differing pore size (POLYGARD CN 10 μηι, SARTOPURE PP2 8 μηη, SARTOPURE PP2 20 μηη, and stainless steel 42 μηη mesh) were evaluated for their ability to aid in clarification filtration of CAIV to achieve a filterability rate of 100 L/m² on five virus strains (ca A/South Dakota/6/07, ca A/Mississippi/4/08, ca A/Uruguay/716/07, ca

B/Florida/4/2006 and ca B/Bangladesh/3333/07).

Materials and equipment

Materials used in this study included: Gibco 1X Phosphate Buffered Saline (PBS) without Ca2+ and Mg2+, pH 7.2 (Invitrogen, Grand Island, NY, Cat No.: 20012043); 10X Sucrose-phosphate (2 M sucrose-1 M potassium phosphate, HYCLONE Ltd, Logan, UT, Cat. No.: SH3A1797.01 ); PUMPSIL tubing with 1.6 mm I.D. and 2.4 mm wall thickness (Watson Marlow, Wilmington, MA, Cat. No.: 913.A016.024); PUMPSIL tubing with 4.8 mm I.D. and 2.4 mm wall thickness (Watson Marlow, Wilmington, MA, Cat. No.: 913.A048.024); MASTERFLEX platinum-cured silicone tubing L/S size 16 (Cole Parmer, Vernon Hills, IL, Cat. No.: 96410-16); MASTERFLEX platinum- cured silicone tubing LIS size 36 (Cole Parmer, Vernon Hills, IL, Cat. No.: 96410-36); 2 mL CRYOVIALS (Nalgene, Rochester, NY, Cat. No.: 5012-020); ¾" x ¼" Sanitary barbed adapter (Cole Parmer, Vernon Hills, IL, Cat. No.: 30515-01 ); ¼" x 3/16" Reducer (Cole Parmer, Vernon Hills, IL, Cat. No.: 30703-52); 3/8" to 1/4" Reducer (Qosina, Edgewood, NY, Cat. No.: 60019); 3/16" ID male luer INTEGRA lock (Cole Parmer, Vernon Hills, IL, Cat. No.: 45503-08); Female luer lock to barb connector, 3/16" ID (Qosina, Edgewood, NY, Cat. No.: 1 1540); 1-L

Polyethylene terephthalate glycol modified (PETG) bottle (Nalgene, Rochester, NY, Cat. No.: 2019-1000); ca A/South Dakota/6/07, Batch number: 141900666A; ca A/Uruguay/716/07, Batch number: 141900675A; ca A/Mississippi/4/08, Batch number: 2000018430; ca B/Florida/4/2006, Batch number: 141900641A; and ca B/Bangladesh/3333/07, Batch number: 2000018547.

Equipment used in this study included: Luer pressure sensor flow cell, (SCILOG

BIOPROCESSING Systems, Middleton, WC, Cat. No.: 080-699PSX); 3/8" Barb pressure sensor flow cell, (SCILOG BIOPROCESSING Systems, Middleton, WC, Cat. No.: 080-694PSX);

SCIPRESS monitor, (SCILOG BIOPROCESSING Systems, Middleton, WC, Cat. No.: 080-690);

Watson Marlow pump (Watson-Marlow Bredel Inc., Wilmington, MA, Model 520 Di/R pump);

HACH 21 OOP turbidity Meter (HACH Company, Loveland, CO, Part number: 46500-00); and -80 °C freezer (Revco Technologies, Asheville, NC, Model No.: UL T2586-9-D35).

Filters used in this study included: Milligard^® 1.2 μηη Optiscale™ disposable capsule filter (Millipore Corporation, Billerica, MA, Cat. No.: SW19A47HH3); SARTOPORE 2 0.8/0.45 m SARTOSCALE disposable capsule filter (Sartorius Stedim Biotech, Goettingen, Germany, Cat. No.: 5445306GS-FF); Milligard^® Opticap^® XL2, (Millipore Corporation, Billerica, MA, Cat. No.: KW19A02HH1 ); SARTOPORE 2 MIDICAP, (Sartorius Stedim Biotech, Goettingen, Germany, Cat. No.: 5445306G8-00); Polygard^® CN 10.0 μηι Nominal Optiscale™ disposable capsule filter (Millipore Corporation, Billerica, MA, Cat. No.: SN1 HA47HH3); SARTOPURE PP2 8 μπι Nominal SARTOSCALE disposable capsule filter (Sartorius Stedim Biotech, Goettingen, Germany, Cat. No.: 5595301 PS-FF); SARTOPURE PP2 20 μπι Nominal SARTOSCALE disposable capsule filter (Sartorius Stedim Biotech, Goettingen, Germany, Cat. No.:

5595320PS-FF); and Stainless Steel Mesh 42 μπι (Plastok (M&F) Ltd, Birkenhead, UK, Stainless steel mesh 42 μηη, wire diameter 0.036 mm).

Methods

Propagation of virus-infected allantoic fluid (VAF) was performed according to previous methods. The following process steps were repeated for each CAIV strain.

Filter preparation

Each filter was wetted and equilibrated before use in the clarification filtration. Each filter capsule assembly was completely bled before filtration.

Pre-clarification filtration

The pre-clarification filtration rig assembly included a tubing assembly, a pressure monitoring SCIPRESS pressure sensor, and a pre-filter. Filters were first wetted with purified water and then equilibrated with 1X PBS individually. Buffer was drained out of the filter before each filtration. Air bubbles were removed through the filter vent port by briefly tapping the rig assembly. CAIV clarification filtration at a linear flux rate of up to 6 L/min/m² (LPM/m²) typically results in minimal potency loss of 0.1 log₁₀FFU/mL or less. Thus, 5 LPM/m² was used in this study. PHF was pumped by a Watson Marlow Bredel pump through the pre-clarification filtration rig assembly at a linear flux rate of 5 LPM/m² and the filtrate was collected in a 1-L PETG bottle. Pressure was monitored upstream of the pre-filter during the filtration process. A constant flow method was used to assess the filterability of the pre-clarification fluid in this study. The filtrations proceeded until the differential pressure plateaued or reached 30 psi.

Clarification filtration

The clarification filtration rig included a tubing assembly, a pressure monitoring SCI PRESS pressure sensor, and a 1 .2 μηι filter (Milligard^®) followed by a 0.8/0.45 μηι filter (SARTOPORE 2). Filtrate obtained from the clarification filtration was pumped by a Watson Marlow Bredel pump through the pre-clarification filtration rig assembly at a linear flux rate of 5 LPM/m²

(equivalent to 6.9 mL/min through the 1.2 μηη 47 mm filters, or 500 mL/min through the 0.1 m² capsule filters) and the clarified filtrate was collected in a 1 -L PETG bottle as the clarified harvest fluid (CHF). Pressure also was monitored upstream of the 1.2 μηη filter during the filtration process. The filtrations proceeded until the differential pressure plateaued or reached 30 psi.

Potency Analysis

Samples taken from the filtration process were stabilized with 10X SP (diluted to 1 X SP) and then frozen in 1 mL aliquots in a -80 °C freezer. Viral potency was analyzed using a

Fluorescent Focus Assay (FFA) with six replicates read per sample.

Effect of p re-filters on CAIV pre-clarification filtration

FIG. 32 presents a table summarizing potency change and filterability of PHF after pre- clarificationfiltration through the four pre-filters chosen for this study. Potency data are presented in FIG. 35.

Among the five CAIV strains, the range in potency change for each pre-filter after pre- clarification filtration was 0 to -0.2 log™ FFU/mL for 10 μηι (POLYGARD CN), +0.1 to -0.2 log™ FFU/mL for 8 μπι (SARTOPURE PP2), -0.1 to -0.2 log™ FFU/mL for 20 μπι (SARTOPURE PP2), and 0 to -0.1 log™ FFU/mL for stainless steel 42 μηη mesh filter. The positive increase in potency was due to assay variation. The overall potency changes for all four pre-filters among the five CAIV strains were similar and were all -0.2 log-io FFU/mL or less after pre-clarification filtrations. Despite the similarity in potency change post pre-clarification filtration among the pre-filters, the filterability of CAIV in each pre-filter varied. The filterability of the five CAIV strains on the 10 μηη pre-filter (POLYGARD CN) ranged between 185 to 314 L/m². At the lowest filterability of 185 L/m² (ca B/Florida/4/2006), the minimum effective filter area (EFA) needed was 0.86 m² (= 160 L divided by 185 L/m²) based on a 160-L batch size. Thus, three 10-inch capsules (0.42 m² EFA per capsule) were needed to cover the required EFA due to the lack of a greater EFA in this capsule format.

Filterability using the 8 μηι pre-filter (SARTOPURE PP2) ranged between 164 to 354 L/m². Based on a batch size of 160 L, the minimum EFA needed to achieve filterability at this scale was 0.98 m² (= 160 L divided by 164 L/m²). Two 10-inch capsules (0.6 m² EFA per capsule) or one 20-inch capsule (1 .2 m² EFA per capsule) provided sufficient filter area to cover the minimum required EFA. A bigger capsule with 1.6 m² EFA per capsule (30-inches) also accommodated the increasing batch size. Compared to the 10 μηη pre-filer (POLYGARD CN), fewer 8 μηη pre-filters were required based on a 160-L batch size.

The minimum filterability determined using the 20 μηη pre-filter (SARTOPURE PP2) was 196 L/m² and the minimum EFA required to filter 160-L was 0.82 m² (= 160 L divided by 196 L/m²). Thus, a similar number filters as the 8 μηη pre-filter (SARTOPURE PP2) were needed to process this batch size. Other filter configurations could be determined based on the results as demonstrated above (e.g., determine minimum filterability for a given filter and calculate a minimum EFA for a given batch size).

The stainless steel 42 μηη mesh has the largest pore size compared to other pre-filters. No pressure drop across this pre-filter was observed throughout filtration. Thus, the absolute filterability on this pre-filter could not be determined.

Effect of using pre-filter on subsequent clarification filtration of CAIV through a 1.2 μηι filter (MILLIGARD) followed by a 0.8/0.45 μπι filter (SARTOPORE 2) FIG. 33 presents a table summarizing filterability of the five CAIV strains after clarification filtrations through a 47 mm 1.2 μηη filter (Milligard^®) followed by a 47 mm 0.8/0.45 μηη filter (SARTOPORE 2) with and without a preceding pre-clarification step. The results among the five CAIV strains used demonstrated that pre-clarification filtrations using a 10 m pre-filter (POLYGARD), an 8 μηι pre-filter (SARTOPURE PP2), and a 20 μηι pre-filter (SARTOPURE PP2) consistently increased the CAIV clarification filterability (e.g., clarification throughput, filtration throughput) by up to 140%, 184%, and 128%, respectively. Consequently, the targeted filterability of 100 L/m² was achieved and was exceeded in all subsequent clarification filtrations for the five CAIV strains. However, using the stainless steel 42 μηη mesh in pre-clarification filtration did not consistently improve the filterability in subsequent clarifications. This indicated that the stainless steel 42 μηη mesh may have a pore size too large to remove some particulates in the PHF that can potentially clog the clarification filters.

Compared to the clarification filterability without pre-clarification, the decrease in filterability after the pre-clarifications using stainless steel 42 μηη mesh might be due to variations in the filter membrane.

Overall potency change results across five CAIV strains shown in FIG. 34 indicated that adding a pre-clarification filtration step preceding the clarification step resulted in an overall potency change of -0.1 log-io FFU/mL or less after clarification filtration; whereas the potency change was -0.2 log-ιο FFU/mL or less after a filtration process without a preceding pre-clarification.

Conclusions

The results from this study demonstrated that adding a pre-clarification filtration step using a pre-filter having a pore size of 8 μηι to 20 μηι (e.g., POLYGARD CN 10 μηι, SARTOPURE PP2 8 m or SARTOPURE PP2 20 μηη) preceding an existing clarification process consistently improved the CAIV clarification filterability to 100 L/m² or greater (equivalent to filtering a 160 L batch size through a single filtration rig); while the pre-clarification filtration using stainless steel 42 m mesh did not facilitate the filterability of CAIV in the subsequent clarification filtration step In addition, overall potency change was -0.2 log-io FFU/mL or less across the five CAIV strains on all four pre-filters tested and -0.1 log-io FFU/mL or less after the subsequent clarification filtrations. Such potency change was comparable to that of an existing clarification filtration in which a potency change of -0.2 log-io FFU/mL or less was detected. Additional pre-filter studies using an 8 pm pre-filter (SARTOPURE PP2)

Clarification filtration performance was investigated for certain TFF engineering and validation batches (described in Example 1 ). Intermediate potency samples from pooled harvest fluid (PHF) and clarified harvest fluid (CHF) samples were collected from each of the batches for which the 8μηι filter (SARTOPURE PP2 8 μηη) was used in the process validation. The results are presented in FIG. 36, which also includes a virus recovery assessment across the clarification filtration process by using the volume of material generated at the PHF and CHF process stages. Virus recovery data from these batches was compared with commercial material from each strain for which intermediate potency testing was completed. This allowed a direct comparison with the virus recovery during clarification both with and without the 8 μηη filter in place.

The results showed that virus recoveries for A/Victoria and A/California strains were equivalent, both with and without the 8 μηη filter in place (i.e., the addition of the filter had no negative impact). The virus recovery for the A/Victoria process validation batch was towards the upper end of the recovery range observed during commercial batch manufacture whereas the for the A/California strain the recovery was on the lower end of commercial batch virus recovery.

The recovery observed for the B/Wisconsin batch was slightly lower than that achieved during the commercial batches. The lower recovery may have been due to a slow down in filtration flow rate observed for the B/Wisconsin strain when the CHF exceeded 150 kg. Thus, a filter swap out option may be used for B/Wisconsin strains when the CHF volume exceeds 150 kg (equivalent to 125 L/m²). The study of clarification filtration performance during the TFF engineering and validation batches, with and without an 8 μηη filter, demonstrated that there is no adverse impact on virus recovery. The filter removed a significant quantity of blood particles and therefore provided protection to the downstream clarification filters. Alternative filter systems for clarification of pooled harvest fluid (PHF)

In an existing process for the manufacture of cold-adapted live-attenuated influenza virus (LAIV), pooled harvest fluid (PHF) is clarified by filtering through two 10-inch 1.2-μηι filters (e.g., Milligard^® (Millipore)) in parallel (filtration area of each filter is 0.8 m²) followed by a 0.8/0.45-μηι filter (e.g., Sartopore^® 2 (Sartorius); 1.2 m² filtration area). In certain instances, the 1.2-μηη filters clogged at a filtration volume as low as 80 L (equivalent to a throughput of 50 L/m²) when processing certain LAIV strains. As described above, addition of a pre-filter with a pore size of 8 m to 20 m (e.g., POLYGARD CN 10-μηι (Millipore) filter; SARTOPURE PP2 8-μηι or 20-μηι filter (Sartorius)) prior to an existing clarification filtration process greatly improved overall clarification to above 100 L/m² without causing additional virus potency loss after filtration.

This study was designed to evaluate different types of filters and filtration methods to reduce the number of filters used in the LAIV clarification process and improve filtration throughput. A series of tests was conducted to evaluate different filters, including hollow fiber cartridges, flat sheet cassettes, cross flow filtration capsules, and depth filters, for their capability to clarify PHF with high throughput and high virus potency recovery.

Materials and equipment

Materials used in this study included: Gibco 1X phosphate-buffered saline (PBS) without Ca²⁺ and Mg²⁺, pH 7.2, cat. No. 20012043 (Invitrogen, Grand Island, NY); 10X Sucrose-phosphate (2 M sucrose-1 M potassium phosphate), cat. No. SH3A1797.01 (HYCLONE, Logan, UT);

PUMPSIL tubing with 1.6 mm I.D. and 2.4 mm wall thickness, cat. No. 913.A016.024 (Watson Marlow, Wilmington, MA)I PUMPSIL tubing with 4.8 mm I.D. and 2.4 mm wall thickness, cat. No. 913.A048.024 (Watson Marlow, Wilmington, MA); MASTERFLEX platinum-cured silicone tubing L/S size 16, cat. No. 96410-16 (Cole Parmer, Vernon Hills, IL); MASTERFLEX platinum-cured silicone tubing L/S size 36, cat. No. 96410-36 (Cole Parmer, Vernon Hills, IL); 2 mL

CRYOVIALS, cat. No. 5012-020 (Nalgene, Rochester, NY); 3/4" x ¼" Sanitary barbed adapter, cat. No. 30515-01 (Cole Parmer, Vernon Hills, IL); ¼" x 3/16" Reducer, cat. No. 30703-52 (Cole Parmer, Vernon Hills, IL); 3/8" to 1/4" Reducer, cat. No. 60019 (Qosina, Edgewood, NY); 3/16" ID Male luer INTEGRA lock, cat. No. 45503-08 (Cole Parmer, Vernon Hills, IL); Female luer lock to barb connector, 3/16" ID, cat. No. 1 1540 (Qosina, Edgewood, NY); 1 -L Polyethylene terephthalate glycol modified (PETG) bottle, cat. No. 2019-1000 (Nalgene, Rochester, NY); ca A/Uruguay/716/07, batch number: 141900675A; ca A/California/07/2009, batch number:

141900764A; ca A/Perth/16/2009, batch number: ZO015-PD 1 1 Marl 0. seed; and ca

B/Brisbane/60/2008, batch number: 14190099A. Equipment used in this study included: Weighing scale, cat. No. TE1502S (Sartorius, Inc., Edgewood, NY); Refrigerator set at 5 ± 3 °C, model NSPR803WWG/0 (Nor-Lake Scientific, Hudson, Wl); Biosafety cabinet, model STERILGARD III Advance (Baker Co., Stanford, MN); Watson Marlow peristaltic pump, model 505Di/RL (Watson Marlow Inc., Wilmington, MA);

MINIKROS pilot system, peristaltic pump, cat. No. Sym3 01 1 01 n (Spectrum Laboratories INC., Rancho Dominguez, CA); Luer pressure sensor flow cell, cat. No. 080-699PSX (SCI LOG

BIOPROCESSING Systems, Middleton, WC); SCIPRESS monitor, cat. No. 080-690 (SCILOG BIOPROCESSING Systems, Middleton, WC); 3/8" Barb pressure sensor flow cell, cat. No. 080- 694PSX (SCILOG BIOPROCESSING Systems, Middleton, WC); HACH 21 OOP turbidity Meter, part number 46500-00 (HACH Company, Loveland, CO); and -80 °C Freezer, model No.

ULT2586-9-D35 (Revco Technologies, Asheville, NC).

Filters used in this study included: Millipore Milligard^® 1.2-μηη OptiScale™ disposable capsule filter, cat. No.: SW19A47HH3 (Millipore Corporation, Billerica, MA); Sartorius SARTOPORE 2 0.8/0.45-μηι SARTOSCALE disposable capsule filter, cat. No. 5445306G-FF; (Sartorius Stedim Biotech, Goettingen, Germany); GE XAMPLER laboratory scale microfiltration cartridges, model No. CFP-4-E-3X2MA, polysulfone, pore size 0.45 μηη, fiber ID 1 mm, membrane area 0.023 m², flow path length 60 cm (GE Healthcare Bio-Sciences, Piscataway, NJ); Pall MICROZA hollow fiber microfiltration module, part No. UJP-0047R, polyvinylidene difluoride, pore size 0.65 μηη, fiber ID 1 .1 mm, membrane area 0.02 m², flow path length 31.4 cm (Pall Corporation, Covina, CA); Sartorius SARTOCON Slice 200, cat. No. 3081860602W-SG, stabilized cellulose based membrane, pore size 0.45 μηη, filtration area 0.02 m² (Sartorius Stedim Biotech, Goettingen, Germany); Millipore PELLICON 2 microfiltration module, cat. No. P2DVPPV01 , Hydrophilic PVDF, pore size 0.65 μηη, filtration area 0.1 m² (Millipore Corporation, Billerica, MA); Pall Super^® membrane LV Centramate™ cassette, cat. No. PSM80C12P2, polyethersulfone, pore size 0.8 μηι, filtration area 0.02 m² (Pall Corporation, Covina, CA); Pall KLEENPAK capsule (pleated cross flow), polyethersulfone, pore size 0.65 μηη, filtration area 0.06 m² (Pall

Corporation, Covina, CA) (development product under validation, no cat. No. available);

Millipore MILLISTAK+DOHC disposable capsule filter, cat. No. SG3J017A03, cellulose fibers with inorganic filter aid, pore size 9.00-0.55 μηη, filtration area 23 cm² (Millipore Corporation, Billerica, MA); Millipore MILLISTAK+COHC disposable capsule filter, cat. No. MC0HC23HH3, cellulose fibers with inorganic filter aid, pore size 2.5-0.2 μηι, filtration area 23 cm² (Millipore Corporation, Billerica, MA). Methods

Propagation of virus-infected allantoic fluid (VAF) was performed according to previous methods. Filtration was performed as described below. Filter preparation

Each filter was wetted as recommended by each filter manufacturer and equilibrated using 1X PBS before use in clarification filtration. Each filter capsule assembly was completely bled before filtration.

Direct flow clarification filtration (DFF)

The clarification filtration rig included a tubing assembly, a pressure monitoring SCI PRESS pressure sensor, a 1 .2-μηι filter (Milligard^®) followed by a 0.8/0.45 μηι filter (SARTOPORE 2) serving as a control, or a depth filter (MILLISTAK+DOHC or MILLISTAK+COHC). Pooled harvest fluid (PHF) was pumped through the clarification filtration rig assembly at a linear flux of 5 L/min/m² (LPM/m²) (equivalent to 6.9 mL/min through the 1.2-μηι 47 mm filters (Milligard^®), or 1 1 .5 mL/min through the 23-cm² depth filters (MILLISTAK+)) using a Watson Marlow Bredel pump and the clarified filtrate was collected in a 1 -L PETG bottle as the clarified harvest fluid (CHF). The filtrations were performed until the differential pressure plateaued or reached 30 psi.

Cross-flow microfiltration (CF-MF)

FIG. 37 presents a table summarizing three different filter formats and operating conditions used in this study for evaluating the performance of cross-flow microfiltration in the clarification of PHF. Filter choices were based, in part, on the availability of suitable small-scale configurations for initial evaluation. Selection of filter pore size was based, in part, on the largest available pore size for the chosen filter. Operating conditions used in this study were selected based, in part, on the manufacturer's recommendation. For the CF-MF process without permeate control, a pressure monitoring SCIPRESS pressure sensor was placed upstream of the filter inlet, downstream of the retentate outlet, and downstream of the permeate outlet. The PHF was pumped using a Watson Marlow Bredel pump or the MINIKROS pilot system peristaltic pump through the inlet of the microfiltration filter at the chosen operating condition, and clarified filtrate was collected at the permeate outlet into 1-L PETG bottles as clarified harvest fluid. The filtrations were performed until the permeate flux plateaued or until the PHF was exhausted. The transmembrane pressure (TMP) and permeate flux were monitored and measured during each filtration process.

To alleviate membrane fouling and reduce the swiftness of permeate flux decline, permeate control was also examined in this study. For the CF-MF process that uses permeate control, a Watson Marlow pump was placed downstream of the permeate outlet to control the permeate flow at a constant flow rate of 150 mL/min (equivalent to 9 LMH). Clarified filtrate was collected at the permeate outlet into 1 -L PETG bottles as clarified harvest fluid. The filtrations were performed until an air bubble was observed at the permeate outlet.

Potency analysis

Samples taken from the filtration process were stabilized with 10X SP (to a final concentration of 1X SP) and then frozen in 1-mL aliquots and stored in a -80 °C freezer. Viral potency was analyzed using a fluorescent focus assay (FFA), with six and twelve replicates read per sample.

Performance of hollow fiber cartridges in PHF clarification filtration

The CF-MF performed on the 0.45-μηη (GE) and the 0.65-μηη (Pall) hollow fiber cartridges using PHF of ca A/Uruguay/716/07 showed that permeate flux declined quickly in the filtration process when permeate flow was un-restricted (FIG. 38 and FIG. 40). Within the first 2 minutes, the permeate flux of the 0.45-μηι filter (GE) dropped from 352 to 147 L/m²/hour (LMH) at a decline rate of 205 LMH/min, while the permeate flux of the 0.65-μηη filter (Pall) dropped from 597 to 203 LMH at a decline rate of 253 LMH/min. The rapid decline of permeate flux indicated a fast blinding of the hollow fiber membrane without restriction of the permeate flux. In some instances, a high and steady permeate flux was employed to obtain a high throughput from the filtration process. FIG. 39 shows that a steady permeate flux of about 47 LMH was achieved on the 0.45-μηη hollow fiber cartridge (GE) when the permeate flux was controlled at 45 LMH by a peristaltic pump. Similarly, a steady permeate flux was expected for the 0.65-μηι hollow fiber cartridge (Pall) with controlled permeate flow.

While a potentially high throughput can be obtained from the CF-MF using hollow fiber cartridge, high potency losses were observed after CF-MF filtration (FIG. 41 ). When operated at a low steady permeate flux, the 0.45-μηη hollow fiber (GE) CF-MF process showed a 1 .4 log₁₀ FFU/mL potency loss. Less potency loss was observed for the CF-MF using the 0.65-μηη hollow fiber (Pall) process, which was 0.2 log-io FFU/mL higher than that of DFF filtration through a 1 .2 m filter (Milligard^®) followed by a 0.8/0.45 μηι filter (SARTOPORE 2), i.e., control filtration.

Flat sheet cassettes

In the CF-MF studies using PHF of ca A/Uruguay/716/07, permeate control at 3 psi was applied on a 0.45-μηη cassette (Sartorius), while the permeate pressure was adjusted at different levels (0 to 4 psi) during the processes using 0.65-μηη (Millipore) and 0.8-μηι (Pall) flat sheets to achieve a steady permeate flux. However, an unstable inlet pressure was encountered when operating the 0.8-μηι flat sheet (Pall) and the pressure constantly reached 30 psi in less than 4 minutes even after lowering the flow rate gradually to 30 mL/min. FIGS. 42 and 43 show permeate flux for the 0.45-μηι (Sartorius) and 0.65-μηι (Millipore) flat sheet CF-MF processes. Permeate flux continuously declined during both processes despite operating with permeate control. In addition, the initial flux was too low and in 30 min the flux dropped from 62 to 36 LMH and from 38 to 25 LMH in the CF-MF process using the 0.45-μηι (Sartorius) and 0.65-μηι (Millipore) flat sheet cassettes, respectively.

FIG. 44 presents a table showing potency data for a CF-MF process using flat sheet cassettes. Compared to control filtration (DFF through a 1.2-μηη filter (Milligard^®) followed by a 0.8/0.45-μηι filter (SARTOPORE 2)), a further 0.1 and 0.2 log FFU/mL potency loss was observed from the CF-MF process using 0.45-μηι (Sartorius) and 0.65-μηι (Millipore) flat sheet cassettes, respectively.

Cross flow disposable capsule

A small-scale 0.06-m² 0.65-μηη cross flow disposable capsule (KLEENPAK, Pall) was used for this study (a 0.50 m² capsule also is available). Similar to the hollow fiber cartridges using ca A/Uruguay/716/07, permeate flux declined quickly when permeate control was not applied as shown in FIG. 45. Compared to the CF-MF using the hollow fiber cartridge, however, a significantly higher initial flux (1875 LMH) was obtained using the capsule. FIG. 46 shows that permeate flux decline was alleviated by applying permeate control at 150 LMH, and a more steady permeate flux was achieved by controlling permeate at a lower flux (the first flux decline shown in FIG. 46 was due to kinked tubing). FIG. 47 presents a table showing the same potency loss (e.g., 0.1 log-io FFU/mL from PHF to CHF) for ca A Uruguay/716/07 from a CF-MF process using a cross flow filtration capsule (KLEENPAK) compared to control filtration (i.e., DFF through a 1.2-μηι filter (Milligard^®) followed by a 0.8/0.45-μηι filter (SARTOPORE 2)). Comparable potency loss to the control filtration process also was observed for other strains, including ca A/Perth/16/2009, ca

A/California/07/2009, and ca B/Brisbane/60/2008 (FIG. 47). Therefore, the cross flow filtration capsule (KLEENPAK) was able to clarify the PHF with high throughput and low potency loss. A sufficiently high and steady flux may be achieved by optimizing the permeate flux control.

Depth filter

Direct flow filtration was performed using depth filters (Millipore MILLISTAK+DOHC (pore size range: 9.00-0.55-μηι) and COHC (pore size range: 2.5-0.2-μηι)) for ca A/Uruguay/716/07 and ca B/Brisbane/60/2008, respectively. FIG. 48 presents a table showing throughput and potency data of DFF depth filtration processes using a 9.0-0.55 μηη depth filter (Millipore

MILLISTAK+DOHC) and a 2.5-0.2 μπι depth filter (Millipore MILLISTAK+COHC).

A substantial increase in throughput was obtained from the depth filtration processes: a 3.9-fold increase using a 9.0-0.55 μηη depth filter and a 1.7-fold increase using a 2.5-0.2 μηη depth filter compared to a control DFF (control filtrations). Specifically, tests with ca A/Uruguay/716/2007 showed that throughput of the 9.0-0.55 μηι depth filter increased to 428 L/m² compared to an existing filtration process (1 1 1 L/m²) while no further potency loss was observed (e.g., -0.1 log-io FFU/mL after the 9.0-0.55 μηη depth filtration vs. -0.2 log-io FFU/mL after the current filtration). Filtration of ca B/Brisbane/60/2008 through the 2.5-0.2 μηη depth filter resulted in higher potency loss (-0.3 logio FFU/mL) compared to an existing filtration process (-0.1 logi₀ FFU/mL) and less throughput improvement (e.g., 170 L/m² on the 2.5-0.2 μηη depth filter vs. 99 L/m² on control filters). Thus, while the filtration process using a 9.0-0.55 μηη depth filter retained comparable virus potency recovery to that of the control filtration, there was a further 0.2 log₁₀ FFU/mL potency loss with the filtration process using the 2.5-0.2 μηη depth filter compared to control filtration. The lower virus potency recovery observed following the filtration process using the 2.5-0.2 μηη depth filter may have been due to smaller pore size.

Because the 9.0-0.55 μηη depth filter achieved a high filtration throughput and low virus potency loss, it was chosen for further testing as a potential substitute to the current 1.2-μηη filter (Milligard^®) for clarifying LAIV PHF.

Conclusions

The results from this study showed that hollow fiber cartridges and flat sheet cassettes were less suitable for use in clarifying LAIV PHF either because of high potency loss or low flux associated with these filtration processes. A low potency loss and high and stable flux was achieved with cross flow filtration using a Pall KLEENPAK disposable capsule. Additionally, depth filtration using a 9.0-0.55 μηη depth filter was identified as a potential candidate to substitute for a 1 .2-μηη clarification filter based on its higher throughput and improved potency recovery.

Use of depth filtration for clarifying pooled harvest fluid of live-attenuated influenza virus

In an existing manufacturing process, clarification of pooled virus harvest fluid (PHF) of live- attenuated influenza virus (LAIV) is performed by filtering PHF through a 1.2-μηη filter (e.g., Milligard^®) followed by a 0.8-μΓη/0.45-μηΊ filter (e.g., SARTOPORE 2). The 1 .2-μηι membrane filter clarifies PHF by removing large size particles, such as egg shells, membrane debris, blood cells, and large aggregates, while the 0.8/0.45-μηΊ filter is used for microbial contaminant (bioburden) removal. Two 10-inch 1 .2-μηη filters are generally used for a 30,000-egg batch production and an additional two 10-inch 1.2-μηη filters are often required for some LAIV strains due to the relatively low throughput.

In order to accommodate higher production capacity and reduce the number of filters used in the clarification process, improved filtration is sought. To this end, a number of studies were performed to evaluate various options (as described above). A 9.0-0.55 μηη depth filter achieved higher filtration throughput and similar virus potency recovery, while maintaining ease of operation as well as scale-up compared to the existing clarification process described above. This study further evaluated the performance of the 9.0-0.55 μηη depth filter by measuring filtration throughput and potency recovery at an increased filtration scale (i.e., from 23 cm² to 1 100 cm²). Pooled harvest fluids (PHF) of six live-attenuated influenza virus (LAIV) strains, A/California/07/2009, A/Uruguay/716/07, A/Perth/16/2009, B/Brisbane/60/2008,

B/Florida/4/2006, and B/Malaysia/2506/14, were clarified. A mini capsule filter of 23 cm² and lab scale pod filters of 270, 540, and 1 100 cm² filtration areas were evaluated for filtration throughput and potency recovery at flux ranges of 100- 300 LMH. A 1.2-μηη membrane filter (Milligard^®) was used in this study as a control for comparison (filtration area: 17.7 cm²). The filtrate of the depth filtration was filtered through a 0.8/0.45-μηι filter (SARTOPORE 2) and the corresponding throughput and potency recovery also were determined. The stability of monovalent bulk (MVB) produced in subsequent steps following the depth filtration was assessed at 2-8 °C in 125-mL bottles and 1 -L bags for a period of 14 days.

Materials and equipment

The following materials were used for this study: 1-mL U-100 Insulin syringe, fitted with 28 gauge needle, cat. No.: 329424 (Becton Dickinson, Franklin Lakes, NJ); 1X Phosphate buffered saline (PBS) without Ca²⁺ and Mg²⁺, pH 7.2, cat. No. 20012043 (Invitrogen, Grand Island, NY); 10X Sucrose phosphate (2 M sucrose-1 M potassium phosphate), cat. No. SH3A1797.01 (HyClone^®, Logan, UT); HyPure™ WFI quality water, cat. No. SH30221.24 (HyClone^®, Logan, UT); PUMPSIL tubing with 4.8 mm ID and 2.4 mm wall thickness, cat. No. 913.A048.024 (Watson Marlow, Wilmington, MA); MASTERFLEX platinum- cured silicone tubing L/S size 16, cat. No. 96410-16 (Cole Parmer, Vernon Hills, IL);

MASTERFLEX platinum-cured silicone tubing L/S size 15, cat. No. 96410-15 (Cole Parmer, Vernon Hills, IL); MASTERFLEX platinum-cured silicone tubing L/S size 36, cat. No. 96410-36 (Cole Parmer, Vernon Hills, IL); 2 mL CRYOVIALS, cat. No. 5012-0020 (Nalgene, Rochester, NY); Barbed to sanitary adapter, 3/8" to 1 -1/2", cat. No. EW-31513-15 (Cole Parmer, Vernon Hills, IL); Quick-disconnect fittings, polycarbonate, 3/8" ID, cat. No. EW-31305-05 (Cole Parmer, Vernon Hills, IL); Quick-disconnect fittings, polycarbonate, 1/4" ID, cat. No. EW-31305-00 (Cole Parmer,

Vernon Hills, IL); Female luer lock to barb connector, cat. No. 1 1540 (Qosina, Edgewood, NY); 1/4" x 1/8" Reducer, cat. No. 30703-52 (Cole Parmer, Vernon Hills, IL); 1 -L Polyethylene terephthalate glycol modified (PETG) bottle, cat. No. 2019-1000 (Nalgene, Rochester, NY); 2-L Polyethylene terephthalate glycol modified (PETG) bottles, cat. No.: 2019-2000 (Nalgene, Rochester, NY); 125-mL Polyethylene terephthalate glycol modified (PETG) bottle, cat. No. 2019-0125 (Nalgene, Rochester, NY); 125-mL BIOTAINER PC bottle, cat. No. 3030-42 (VWR International, Brisbane); 1 -L BioProcess Container™, cat. No. SH30658.15 (HyClone^®, Logan, UT); 5-L BioProcess Container™, cat. No. SH30709.01 (HyClone^®, Logan, UT); 10-L

BioProcess Container™, cat. No. SH30712.02 (HyClone^®, Logan, UT); 50-L BioProcess Container™, cat. No. SH30712.04 (HyClone^®, Logan, UT); 60% Sucrose in PBS, cat. No.

SH3A1800.01 (HyClone^®, Logan, UT); 10% Sucrose in PBS, cat. No. SH3A1799.01 (HyClone^®, Logan, UT); Centrifuge Diluent Phosphate Buffer (CDPB), cat. No. SH3A 1801.01 (HyClone^®, Logan, UT); 10-mL Pipettes, cat. No.: 53283-708 (VWR International, Brisbane, CA); Lancet (LIFESCAN, Model: One Touch, FinePoint™); A/California/07/2009, batch number:

141900764A; A/Uruguay/716/07, batch number: 141900675A; A/Perth/16/2009, batch number: 141600471 ; B/Brisbane/60/2008, batch number: 141900699A; B/Florida/4/2006, batch number: 141900641A; and B/Malaysia/2506/14, batch number: 141900442A.

Equipment used in this study included: Weighing scale, cat. No. EB35EDE-1 (Sartorius, Inc., Edgewood, NY); Refrigerator set at 5 ± 3 °C, model NSPR803WWG/0 (Nor-Lake Scientific, Hudson, Wl); Biosafety cabinet, model STERILGARD III Advance (Baker Co., Stanford, MN); Watson Marlow peristaltic pump, model 505Di/RL (Watson Marlow Inc., Wilmington, MA); Luer pressure sensor flow cell, cat. No. 080-699PSX (SCILOG BIOPROCESSING Systems, Middleton, WC); SCIPRESS monitor, cat. No. 080-690 (SCILOG BIOPROCESSING Systems, Middleton, WC); 3/8" Barb pressure sensor flow cell, cat. No. 080-694PSX (SCILOG

BIOPROCESSING Systems, Middleton, WC); Refrigerator Incubator, model No. 2005 (VWR International, Brisbane, CA); -80 °C Freezer, model No. UL T2586-9-D35 (Revco Technologies, Asheville, NC); Pipette aid, cat. No.: 14006-026 (VWR International, Brisbane, CA); Hitachi CC40 continues flow ultracentrifuge (Hitachi, Japan); Automatic Centrifuge Offloading System (Medlmmune, Inc., Santa Clara, CA); WAVE MIXER, model MIXER 20/50P (Wave Biotech, Bridgewater, NJ, Model); Egg incubator, model NMC2500 (Natureform Inc., Jacksonville, FL); Egg candler, cat. No.: N4130 (FIBREOPTIC LLLUMINATOR, FIBREOPTIC LIGHTGUIDES, Australia); and Egg puncher, model No.: E90 (Glas Col, Terre Haute, IN).

Filters used in this study included: Millipore Milligard^® 1.2-μηη OptiScale™ disposable capsule filter, cat. No.: SW19A47HH3 (Millipore Corporation, Billerica, MA); Millipore Millistak+^® D0HC disposable capsule filter, cat. No. SG3J017A03, cellulose fibers

with inorganic filter aid, pore size 9.00-0.55 μηη, filtration area 23 cm² (Millipore Corporation, Billerica, MA); Millipore Millistak+^® D0HC disposable capsule filter, cat. No. MD0HC027H1 , cellulose fibers with inorganic filter aid, pore size 9.00-0.55 μηη, filtration area 270 cm² (Millipore Corporation, Billerica, MA); Millipore Millistak+^® D0HC disposable capsule filter, cat. No.

MD0HC054H1 , cellulose fibers with inorganic filter aid, pore size 9.00-0.55 μηη, filtration area 540 cm² (Millipore Corporation, Billerica, MA); Millipore Millistak+^® D0HC disposable capsule filter, cat. No. MD0HC01 FS1 , cellulose fibers with inorganic filter aid, pore size 9.00-0.55 μηη, filtration area 0.1 1 m² (Millipore Corporation, Billerica, MA); Sartorius SARTOPORE 2 0.8/0.45- μηη 300 capsule filter, cat. No. 5441306G5-OO (Sartorius Stedim Biotech, Goettingen, Germany); Sartorius SARTOPORE 2 0.8/0.45-μΓη MIDICAP filter, cat. No. 5441306G8-OO (Sartorius

Stedim Biotech, Goettingen, Germany); and Sartorius SARTOPORE 2 0.45/0.2-μηι 150 capsule filter, cat. No. 5441307H4-OO (Sartorius Stedim Biotech, Goettingen, Germany).

Methods Propagation of LAIV in embryonated chicken eggs was performed according to previous methods. Filtration was performed as described below.

Filter preparation

A 1 .2-μηι filter (Milligard^®) and a 0.8/0.45-μηι filter (SARTOPORE 2) were prepared according to previous methods. 9.0-0.55 μηη depth filters (Millistak+^® DOHC) were wetted as referenced in the filter manufacturer's specification and equilibrated using 1 X PBS before being used in the clarification filtration. Each filter assembly was completely bled before filtration.

Clarification filtration using a 1.2-pm filter (control)

The clarification filtration rig included a tubing assembly, a pressure monitoring SCI PRESS pressure sensor and a 1.2-μηι filter (Milligard^®). The 1.2-μηι filter membrane was used as a control in this study to determine the throughput of an existing clarification filtration process.

Pooled harvest fluid (PHF) was pumped through the clarification filtration rig assembly at a flux of 250 L/m²/hour (LMH) (equivalent to 7.4 mL/min through the 1.2-μηι 47 mm filters (Milligard^®)) using a Watson Marlow Bredel pump, and the filtrate was collected in a 1 -L PETG bottle as the clarified harvest fluid (control-CF). The filtrations were performed until the differential pressure plateaued or reached 30 psi.

Clarification filtration using a 9.0-0.55 pm depth filter The clarification filtration rig included a tubing assembly, a pressure monitoring SCI PRESS pressure sensor and a 9.0-0.55 μηη depth filter. The depth filters (pod filters) included three ports, one inlet, one vent, and one outlet; while the capsule filter has one inlet and one outlet. A pilot pod filter holder was used for the operation of pod filters larger than 1 100 cm². At the start of filtration, the vent port was opened while the outlet port was closed. PHF was pumped through the clarification filtration rig assembly at a designated flux (outlined in FIG. 49) using a Watson Marlow Bredel pump. Different fluxes were evaluated within a flux range of 100-300 LMH. Upon filling the entire filter and bleeding all bubbles through the vent port, the vent port was closed and the outlet port was opened. Filtrate was collected into a HyClone^® bag as clarified harvest fluid (DOHC-CF). Filtrations were performed until the differential pressure plateaued or reached 30 psi.

Clarification filtration using a 0.8/0.45-um filter

Virus filtrate after filtration using a 9.0-0.55 μηη depth filter (DOHC-CF) was pumped using a Watson Marlow Bredel pump through a clarification filtration rig that included a tubing assembly, a pressure monitoring SCI PRESS pressure sensor and a 0.8/0.45-μηι filter (SARTOPORE 2). The clarified harvest fluid (CHF) was collected in a HyClone^® bag. Filtrations were performed until the differential pressure plateaued or reached 30 psi.

Monovalent bulk (MVB) production and stability

For A/California/07/2009 and B/Brisbane/06/2008, CHF was loaded into a Hitachi CC40 ultracentrifuge at 100 mL/min while 240 mL/min was used to load A/Perth/16/2009. About 100 ml. and 800 ml. of MVB were transferred into a 125-mL PC bottle and a 1-L HyClone^® bag, respectively. Both the bottle and the bag were stored in a refrigerator at 2-8 °C for stability study over a 14 day period. Potency analysis

Samples taken from the filtration process were stabilized with 10X SP (to a final concentration of 1X SP) and then frozen in 1-mL aliquots and stored in a -80 °C freezer. Viral potency was analyzed using a fluorescent focus assay (FFA) with twelve replicates read per sample over two days.

Impact of depth filtration filter size on filtration throughput and potency recovery

The filtration throughput of LAIV using 9.0-0.55 μηη depth filters is summarized in the table presented in FIG. 50. Throughput values were expressed as averages of the experimental results (shown in FIG. 55) measured when filtration differential pressure reached 30 psi, except those for A/Uruguay/716/07. The filtration throughput tracked closely across the various tested filter areas and formats (from the 23 cm² mini capsule to the 270, 540, and 1 100 cm² pod filters). Compared to control filtration using a 1.2-μηη filter (Milligard^®), the throughput of a 9.0-0.55 μηη depth filter was higher with a 1.8-fold increase for B/Malaysia/2506/04 and more than a 3-fold increase for all other LAIV strains. On average, the throughput improvement was 3.4-fold.

The corresponding average virus potency change (experimental results shown in FIG. 56) after the 9.0-0.55 μηι depth filtration (Millistak+^® D0HC) at a flux of 250 LMH and a maximum differential pressure of 30 psi is shown in FIG. 51. The potency change after depth filtration was similar across different filter areas and formats. The potency drop after depth filtration was 0.2 log-ιο FFU/mL or less, except for the filtration of A Uruguay/716/07 using the 540 cm² filter and B/Malaysia/2506/04 using the 270 cm² filter, which showed a potency drop of 0.3 log₁₀ FFU/mL. Overall, the potency drop after depth filtration was similar to that after filtration using a 1.2-μηη filter (Milligard^®).

These results indicated that filtration throughput and potency recovery remained consistent during a filtration scaling up process using 9.0-0.55 μηη depth filters; while similar potency recovery and higher filtration throughput were achieved compared to that of 1 .2-μηη filtration. Effect of flux on filtration throughput and potency recovery after 9.0-0.55 pm depth filter

FIG. 52 shows that throughput of depth filtration using a 9.0-0.55 μηη depth filter (Millistak+^® D0HC) was consistently greater than 200 L/m² at a flux range of 100-300 LMH across different filter sizes; while the potency change was -0.2 log-io FFU/mL or less after each filtration.

Therefore, a flux in the range of 100-300 LMH can be used for depth filtration.

Throughput and potency recovery after 0.8/0.45-pm filtration

FIG. 53 shows that the throughput of filtering D0HC-CF through a 0.8/0.45-μηι filter

(SARTOPORE 2) at a flux of 300 LMH was greater than 300 L/m2 and the potency change was -0.1 log™ FFU/mL or less. The data indicated that the 1.2 m² 0.8/0.45-μηι filter (SARTOPORE 2) was sufficient.

MVB stability following the use of a 9.0-0.55 pm depth filter in the clarification process The clarified virus fluids of A/Perth/16/2009, A/California/07/2009, and B/Brisbane/06/2008 were loaded into a CC40 ultracentrifuge after filtration through a 9.0-0.55 μηι depth filter followed by a 0.8/0.45-μηι filter (SARTOPORE 2). The stability of monovalent bulk (MVB) produced by this process was assessed over a 14-day period at 2-8 °C in two types of containers: 125-mL PC bottle (used in an existing MVB process) and 1-L HyClone^® bag (an alternative container to the 125-mL PC bottle).

FIG. 54 summarizes the potency of MVB during a 14-day testing period in the 125-mL PC bottle and in the 1 -L HyClone^® bag. The data showed a similar potency drop for the MVB stored in the 125-mL PC bottle and in the 1-L bag for each LAIV strain.

Conclusions This study demonstrated that a higher throughput (3.4-fold on average) and similar potency recovery (-0.2 log-io FFU/mL or less potency drop) were achieved in pooled harvest fluid (PHF) clarification using 9.0-0.55 μηη depth filters compared to a 1.2-μηη membrane filter.

Filtration throughout and potency recovery remained consistent during the filtration scale-up process from a 23 cm² mini capsule to a 1 100 cm² pod at flux ranges of 100-300 LMH.

Filtration throughput was greater than 300 L/m² and a 0.1 log-io FFU/mL or less in potency loss was observed after subsequent filtration of the depth filtrate through the 0.8/0.45-μηι filter. The monovalent bulk (MVB) produced following the use of depth filtration in the clarification process showed the same stability when stored in a 125-mL PC bottle and in a 1 -L bag at 2-8 °C for a period of 14 Days. Therefore, a 9.0-0.55 μηη depth filter can be substituted for the 1 .2-μηη membrane filter in the clarification of PHF.

Example 3: Sucrose gradient optimization This example describes certain improvements to the purification process for influenza viruses. Improved purification methods are described for a live attenuated influenza virus monovalent bulk (LAIV-MB) manufacturing process, however such methods may be applied to any influenza virus manufacturing process. Improvements include optimization of a sucrose gradient described in detail below. The ultracentrifugation process sometimes used in the manufacture of cold adapted influenza virus (CAIV) monovalent bulk typically uses a sucrose gradient to concentrate and isolate the virus from clarified harvest fluid (CHF). The sucrose gradient decays over time during the centrifugation process due to the diffusion of sucrose from high concentration to low

concentration. Thus, the volume of the CHF load into the ultracentrifuge is limited. To extend the centrifuge run time window, a study evaluating the diffusion rate of a 60% (w/w) sucrose solution was conducted. Three different sucrose gradient compositions were tested in this study.

Materials and equipment

Material used in this example included: Phosphate Buffered Saline (PBS) pH 7.2 without Calcium and Magnesium, MEDI part No. 4101086 (HYCLONE, Logan, UT, Cat. No. 20012- 043); 60% Sucrose in PBS, MEDI Part No. 4101084 (HYCLONE, Logan, UT, Cat. No.

SH3A1800.01 ); 10% Sucrose in PBS, MEDI Part No. 4101083 (HYCLONE, Logan, UT, Cat. No. SH3A1799.01 ); PUMPSIL platinum-cured silicone tubing with 4.8 mm ID and 2.4 mm wall thickness (Watson Marlow, Wilmington, MA, Cat. No. 913.A016.024); 3/16 x 5/16" tubing (Cole- Parmer, Vernon Hills, IL. Cat. No. 95802-09); 1/8" x 1/4" Polypropylene tubing (Cole-Parmer, Vernon Hills, IL, Cat. No. 95875-02); Deionized water (DIW) (Medlmmune Vaccines, Santa

Clara, CA); 10 L Stedim FLEXBOY bags (Stedim, Concord, CA, Product No. FBP10M01 ); 125 mL Polyethylene terephthalate glycol modified (PETG) bottles (Nalgene, Rochester, NY, Cat. No. 2019-0125); 2 L Polyethylene terephthalate glycol modified (PETG) bottles (Nalgene, Rochester, NY, Cat. No. 2019-2000); MASTERFLEX platinum-cured silicone tubing US 24 (Cole-Parmer, Vernon Hills, IL, Cat. No. 96410-24); 1/4" Polypropylene T-connector (Cole- Parmer, Vernon Hills, IL, Cat. No. K-30610-30); and 1/4" Male and female quick-disconnect fittings (Cole-Parmer, Vernon Hills, IL, Cat. No. K-31305-00).

Equipment used in this example included: Hitachi large scale continuous flow ultracentrifuge (Hitachi, Japan, Model CP40Y); Leica AR600 automatic refractometer (Leica Microsystems Inc., Buffalo, NY, Model AR600); Biosafety cabinet (Baker Co., Stanford, MN, Model STERILGARD III Advance); Peristaltic pump (Watson Marlow Inc., Wilmington, MA, Models 505Di/RL);

MASTERFLEX EASYLOAD II pump (Cole-Parmer, Vernon Hills, IL, Model No. 77521-40);

Refrigerator (Forma Scientific, Marietta, OH, Model 3775); Weighing scale ULTIMA II (Sartorius, Inc., Edgewood, NY, Cat. No. TE1502S); and 3 ½" Sanitary glycerin-filled pressure gauge (Cole-Parmer, Vernon Hills, IL, Cat. No. K-68802-12).

Methods

Operation of the CP40Y and 3.2 L CT40C rotor core was conducted according to field training and the manufacturer's equipment manual.

Gradient buffer (GB) compositions

Three compositions of gradient buffers (GB1-3) were used to generate sucrose gradient profiles at different centrifuge run times. The volumes of 10% sucrose, 60% sucrose and PBS that were used to make GB1 , GB2 and GB3 are described in the table presented in FIG. 57. GB 1 was produced using 60% sucrose, 10% sucrose and PBS at a ratio of 1.5:1 .3:0.4; GB 2 was produced using 60% sucrose, 10% sucrose and PBS at a ratio of 1.35:1.45:0.4; and GB 3 was produced using 60% sucrose, 10% sucrose and PBS at a ratio of 1.2:1 .6:0.4.

Loading sucrose gradient and centrifuge rotor acceleration The stationary rotor was completely filled with approximately 3.2 L of PBS (total possible volume) at 250 - 300 mL/min through the bottom port of the ultracentrifuge. A portion of the PBS was then displaced with 10% sucrose at 250 - 300 mL/min followed by 60% sucrose at 100 - 200 mL/min. After 60% sucrose was pumped into the rotor, the inlet tubing was clamped off, and the rotor was immediately accelerated to 7,000 rpm. When the rotor reached 7,000 rpm, PBS was pumped into the rotor at 100 - 150 mL/min through the bottom port of the

ultracentrifuge. Once the centrifuge tubing pressure was stable, the flow rate was increased to 150 - 200 mL/min then the rotor was accelerated to 35,000 rpm.

PBS was pumped into the bottom port of the ultracentrifuge at 200 - 300 mL/min while rotor speed was maintained at 35,000 rpm for 0, 1 , 3, 5 or 12 hours as described in the table presented in FIG. 58. For 3, 5 and 12 hour-run times, PBS was re-circulated until a total time rotor spinning at 35,000 rpm was reached. After spinning the rotor at 35,000 rpm for each of the times listed below, the rotor was decelerated to 7,000 rpm under normal braking then coasted from 7,000 rpm to a complete stop. For certain experiments (1 -4 and 6-9, FIG. 58), a development tubing rig with short tubing length was used to load the gradient buffer and offload the sucrose gradient. The empty tubing was connected to the bottom port of the ultracentrifuge before the start of offloading. For the 12 hour-run (experiment 5), a centrifuge tubing assembly that mimics an existing tubing rig was used.

Offloading the sucrose gradient Once the rotor was completely stopped, the sucrose gradient was pumped out from the bottom port of the ultracentrifuge at 100 mL/min. The sucrose gradient was collected at 100 ml. per fraction in 125 ml. bottles and tested for refractive index (index - temperature compensated (TC)) and solids - TC to determine the sucrose concentration. Diffusion of sucrose in the gradient based on total centrifuge run time

FIG. 59 shows sucrose gradients generated from GB 1 (60% sucrose, 10% sucrose and PBS at a ratio of 1 .5:1 .3:0.4 (e.g., 1 .5 L of 60% sucrose, 1.3 L of 10% sucrose and 0.4 L of PBS)) with 0, 1 , 3, 5, and 12 hour total run times at 35,000 rpm. For the 12-hour run the inlet line was filled with PBS, which mixed with the sucrose upon offloading. This resulted in a low sucrose concentration for the first fraction collected. The data showed the 60% sucrose continued to diffuse over time during the ultracentrifuge operation from 0 hour to 12 hour run times. The peak sucrose concentration recovered was 56% after 12 hours with the rotor spinning at 35,000 rpm (FIG. 66). Additionally, the concentration of the sucrose gradient front recovered depended on the concentration of initial 60% sucrose which can be varied from 58% to 63% solids - TC. Thus, the concentration of sucrose gradient front recovered can be higher or lower based on the starting concentration of 60% sucrose. In this study, the 60% sucrose ranged from 62% to 63% solids - TC; and the concentration of sucrose gradient front recovered was around 62%. FIGS. 60 and 61 show sucrose gradients generated from GB 1 , GB 2 and GB 3 when rotor speed was maintained at 35,000 rpm for 1 and 3 hours. The data indicated that different initial volumes of 60% sucrose resulted in similar gradient profiles across the concentration ranges 35% to 49% used for selection of peak virus fractions (FIGS. 66, 67, 68). The sucrose gradient that was generated from GB 1 had more than 300 ml. of 60% sucrose (60% solids - TC or higher) recovered compared to GB 3 after 1 and 3 hour runs. This may be because GB 1 had a greater volume of 60% sucrose (1 .5 L) compared to GB 3 (1.2 L of 60% sucrose). This result also indicated that the diffusion rate of 60% sucrose was constant for the gradient compositions tested. Thus, a greater volume of 60% sucrose was recovered if the initial volume of 60% sucrose loaded into the ultracentrifuge was higher.

Correlation of the sucrose diffusion rate and different initial volumes of 60% sucrose

In FIG. 62, the volume of 60% sucrose concentration recovered (60% solids-TC or higher) is plotted against the total run time at 35,000 rpm for the three gradient buffer compositions (GB 1 , GB 2, GB 3). The data indicated that the greater the volume of 60% sucrose used in the gradient composition, a higher volume of 60% sucrose was recovered after the same hour of run time (1 and 3 hour run times). For the gradient profiles using GB 1 (60% sucrose, 10% sucrose and PBS at a ratio of 1.5:1 .3:0.4 (e.g., 1 .5 L of 60% sucrose, 1 .3 L of 10% sucrose and 0.4 L of PBS)), the diffusion rate of the sucrose gradient front was estimated by the volume of the fractions that had 60% sucrose concentration. The correlation coefficient of the regression line was greater than 0.95 indicating the gradient front movement within the rotor fit well to the linear model. Thus, the diffusion rate of the 60% sucrose may be estimated as approximately constant (at approximately 95 mL/hour, y = 913.56 - 94.915x).

In FIG. 63, sucrose gradient profiles of a 30,000 egg batch (UK-300597) and a 15,000 egg batch (UK-300639) from an existing manufacture (FIG. 65) using GB 3 (60% sucrose, 10% sucrose and PBS at a ratio of 1.2:1 .6:0.4 (e.g., 1 .2 L of 60% sucrose, 1 .6 L of 10% sucrose, 0.4 L of PBS)) are compared to the 12 hour-run using GB 1 (60% sucrose, 10% sucrose and PBS at a ratio of 1.5:1 .3:0.4 (e.g., 1 .5 L of 60% sucrose, 1 .3 L of 10% sucrose and 0.4 L of PBS)). The sucrose gradient profile from an existing manufacture (batch 300597) had a front offloading sucrose concentration at approximately 49% compared with 56% recovered from the

experimental run after spinning the rotor at 35,000 rpm for 12 hours. These results indicated that increasing the initial 60% sucrose volume allowed a higher sucrose concentration recovered in the front offloading of the sucrose gradient. For the 12 hour-experimental run, the gradient concentration at the tail-end was lower due to loading PBS instead of loading CHF. Additionally, the sucrose concentration of the offloading front from the 12-hour experimental run was approximate 56% solids - TC, which was comparable to 52% solids - TC from the 9-hour commercial run (batch 300639). The data also showed that the same total number of fractions were in the range of 35% - 49% solids - TC that were available for peak virus pooling (FIGS. 64, 65, 66). The results showed that a change in gradient composition compensated for the decay in the sucrose gradient over time and extended the ultracentrifugation process time while still maintaining a comparable sucrose concentration offloading front and similar virus peak pooling volume. Thus, the quantity of the CHF load into the ultracentrifuge can be increased.

Conclusions

The results demonstrated that as the centrifuge run time was extended, the percent sucrose concentration recovered from the front offloading fractions was reduced. Gradient buffer compositions with different volumes of 60% and 10% sucrose were studied at different centrifuge run times. Three different gradient compositions: GB 1 (60% sucrose, 10% sucrose and PBS at a ratio of 1.5:1 .3:0.4), GB 2 (60% sucrose, 10% sucrose and PBS at a ratio of 1.35:1.45:0.4), and GB 3 (60% sucrose, 10% sucrose and PBS at a ratio of 1.2:1 .6:0.4) were used to generate sucrose gradient profiles. The results showed the diffusion of 60% sucrose was constant for the three gradient compositions tested. This indicated that a greater volume of 60% sucrose was recovered if the initial volume of 60% sucrose loaded into the ultracentrifuge was higher. The study demonstrated that increasing the initial 60% sucrose volume (e.g., from 1.2 L to 1.5 L (GB 1 )) can compensate the decay of sucrose over time, extend the centrifugation run time to 12 hours and still maintain a similar sucrose gradient profile compared to a 9-hour centrifugation run time using GB 3.

Example 4: Examples of embodiments A1. A method for making an influenza virus composition comprising subjecting a concentrated viral harvest comprising influenza viruses to centrifugation, thereby producing a clarified viral harvest.

A2. The method of embodiment A1 , wherein the viral harvest is initially clarified by filtration prior to or during concentration.

A3. A method for making an influenza virus composition comprising:

a) concentrating a viral harvest comprising influenza viruses, thereby producing a concentrated viral harvest; and b) subjecting the concentrated viral harvest to centrifugation, thereby producing a clarified viral harvest.

A4. The method of embodiment A3, wherein the viral harvest is initially clarified by filtration prior to or during the concentrating in (a).

A5 The method of any one of embodiments A1 to A4, further comprising after centrifugation sterilizing by sterile filtration the viral harvest, thereby producing a sterilized viral harvest. A6. A method for making an influenza virus composition comprising subjecting a clarified and concentrated viral harvest comprising influenza viruses to centrifugation, thereby producing a further clarified viral harvest.

A7. A method for making an influenza virus composition comprising:

a) clarifying a viral harvest comprising influenza viruses by filtration, thereby producing a clarified viral harvest;

b) concentrating the clarified viral harvest, thereby producing a concentrated viral harvest; and

c) subjecting the concentrated viral harvest to centrifugation, thereby producing a further clarified viral harvest.

A8. The method of embodiment A6 or A7, further comprising sterilizing by sterile filtration the further clarified viral harvest, thereby producing a sterilized viral harvest.

A9. The method of any one of embodiments A3 to A8, wherein the concentrating comprises use of a tangential flow filtration (TFF) process.

A10. The method of embodiment A9, wherein the TFF process comprises use of a hollow fiber cartridge. A1 1 . The method of embodiment A10, wherein the hollow fiber cartridge has a pore size ranging from about 500 kD to about 750 kD.

A12. The method of embodiment A1 1 , wherein the hollow fiber cartridge has a pore size of about 500 kD. A13 The method of embodiment A1 1 , wherein the hollow fiber cartridge has a pore size of about 750 kD. A14 The method of any one of embodiments A9 to A13, wherein the TFF process is performed using a shear rate ranging from about 10,000 s^"1 to about 16,000 s^'

A15 The method of any one of embodiments A9 to A14, wherein the TFF process is performed using a transmembrane pressure (TMP) ranging from about 10 psig to about 20 psig.

A16 The method of any one of embodiments A9 to A15, wherein the TFF process is performed using a load factor ranging from about 50 L to about 100 L of clarified viral harvest per square meter. A17 The method of any one of embodiments A9 to A16, wherein the TFF process is performed at a filtrate flux rate of at least about 25 LMH.

A18. The method of any one of embodiments A1 to A17, wherein the viral harvest is

concentrated at least about 2-fold.

A19 The method of embodiment A18, wherein the viral harvest is concentrated at least about 4- fold.

A20. The method of embodiment A19, wherein the viral harvest is concentrated at least about 5-fold.

A21. The method of embodiment A20, wherein the viral harvest is concentrated at least about

6- fold. A22. The method of embodiment A21 , wherein the viral harvest is concentrated at least about

7- fold.

A23 The method of embodiment A22, wherein the viral harvest is concentrated at least about 10-fold. A24 The method of embodiment A23, wherein the viral harvest is concentrated at least about 20-fold. A25 The method of embodiment A24, wherein the viral harvest is concentrated at least about 50-fold.

A26 The method of embodiment A25, wherein the viral harvest is concentrated at least about 100-fold.

A27. The method of any one of embodiments A1 to A26, wherein viral yield is increased relative to viral yield of a method that does not comprise concentrating the viral harvest prior to centrifugation. A28. The method of embodiment A27, wherein the viral yield is increased at least about 2%. A29. The method of embodiment A28, wherein the viral yield is increased at least about 5%. A30. The method of embodiment A29, wherein the viral yield is increased at least about 10%.

A31 . The method of embodiment A30, wherein the viral yield is increased at least about 15%.

A32. The method of embodiment A31 , wherein the viral yield is increased at least about 20%. A33. The method of embodiment A32, wherein the viral yield is increased at least about 50%.

A34. The method of embodiment A33, wherein the viral yield is increased at least about 70%.

A35. The method of any one of embodiments A1 to A34, wherein at least about 100 L of viral harvest is concentrated.

A36. The method of embodiment A35, wherein at least about 150 L of viral harvest is concentrated. A37. The method of embodiment A36, wherein at least about 200 L of viral harvest is concentrated.

A38 The method of embodiment A37, wherein at least about 400 L of viral harvest is concentrated.

A39. The method of any one of embodiments A1 to A38, wherein the amount of viral harvest subjected to centrifugation is greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

A40. The method of embodiment A39, wherein the amount of viral harvest subjected to centrifugation is at least about 10% greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

A41 . The method of embodiment A39, wherein the amount of viral harvest subjected to centrifugation is at least about 20% greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

A42. The method of embodiment A39, wherein the amount of viral harvest subjected to centrifugation is at least about 40% greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

A43. The method of any one of embodiments A1 to A42, wherein all or substantially all of the viral harvest is subjected to centrifugation. A44. The method of any one of embodiments A1 to A43, wherein the centrifugation is performed at about 2 °C to about 25 °C.

A45. The method of embodiment A44, wherein the centrifugation is performed at about 2 °C to about 14 °C. A46. The method of any one of embodiments A1 to A45, wherein the centrifugation is performed at a speed of about 30,000 RPM to about 40,000 RPM. A47. The method of any one of embodiments A1 to A46, comprising prior to or during centrifugation loading the concentrated viral harvest into a centrifuge device at a particular loading flow rate.

A48. The method of embodiment A47, wherein the loading flow rate is lower relative to a loading flow rate for centrifugation in a method that does not comprise concentrating the viral harvest prior to centrifugation.

A49. The method of embodiment A48, wherein the loading flow rate is less than about 200 mL/min.

A50. The method of embodiment A49, wherein the loading flow rate is less than about 160 mL/min.

A51 . The method of embodiment A50, wherein the loading flow rate is less than about 150 mL/min.

A52. The method of embodiment A51 , wherein the loading flow rate is less than about 130 mL/min. A53. The method of embodiment A52, wherein the loading flow rate is less than about 120 mL/min.

A54. The method of embodiment A53, wherein the loading flow rate is less than about 100 mL/min.

A55. The method of embodiment A47 or A48, wherein the loading flow rate ranges from about 120 mL/min to about 160 mL/min. A56. The method of embodiment A47 or A48, wherein the loading flow rate ranges from about 140 mL/min to about 180 mL/min.

A57. The method of any one of embodiments A1 to A56, wherein the centrifugation comprises continuous zonal centrifugation.

A58. The method of embodiment A57, wherein the continuous zonal centrifugation is performed over a sucrose density gradient. A59. The method of embodiment A58, wherein the sucrose density gradient is a 0% to 100% sucrose gradient.

A60. The method of embodiment A58, wherein the sucrose density gradient is a 0% to 60% sucrose gradient.

A61 . The method of embodiment A58, wherein the sucrose density gradient is a 10% to 60% sucrose gradient.

A62. The method of any one of embodiments A58 to A61 , wherein the sucrose density gradient is generated using equal or substantially equal volumes of a 60% sucrose (w/w) composition and a 10% sucrose (w/w) composition.

A63. The method of any one of embodiments A58 to A61 , wherein the sucrose density gradient is generated using a volume of a 60% sucrose (w/w) composition that is greater than the volume of a 10% sucrose (w/w) composition.

A64. The method of embodiment A63, wherein the volume of the 60% sucrose (w/w) composition is at least about 1.1 times greater than the volume of the 10% sucrose (w/w) composition.

A65. The method of embodiment A63, wherein the sucrose density gradient is generated using volumes of a 60% sucrose (w/w) composition, a 10% sucrose (w/w) composition and PBS at a ratio of 1.3-1.6 to 1 .2-1.5 to 0.4, respectively. A66. The method of embodiment A63 or A65, wherein the sucrose density gradient is generated using volumes of a 60% sucrose (w/w) composition, a 10% sucrose (w/w) composition and PBS at a ratio of 1.5 to 1 .3 to 0.4, respectively. A67. The method of any one of embodiments A62 to A66, wherein the centrifugation has a run time of at least about 9 hours.

A68. The method of any one of embodiments A62 to A66, wherein the centrifugation has a run time of at least about 12 hours.

A69. The method of any one of embodiments A58 to A68, wherein after centrifugation the viral harvest is collected from the sucrose density gradient at gradient coordinates between about 35% to about 49% sucrose.

A70. The method of any one of embodiments A1 to A69, wherein after centrifugation the viral harvest is diluted with a buffer.

A71. The method of embodiment A5 or any one of embodiments A8 to A69, wherein after centrifugation and before sterilizing, the viral harvest is diluted with a buffer. A72. The method of embodiment A70 or A71 , wherein the buffer is a phosphate buffer.

A73. The method of any one of embodiments A7 to A72, wherein the clarifying in (a) comprises use of at least two filter species.

A74. The method of embodiment A2 or A4, wherein the initial clarification comprises use of at least two filter species.

A75. The method of embodiment A73, wherein the clarifying in (a) comprises use of at least three filter species.

A76. The method of embodiment A74, wherein the initial clarification comprises use of at least three filter species. A77. The method of any one of embodiments A73 to A76, wherein the filter species comprise at least one pre-filter.

A78. The method of embodiment A77, wherein the pre-filter has a pore size ranging from about 3 microns to about 20 microns.

A79 The method of embodiment A77, wherein the pre-filter has a pore size of about 8 microns.

A80. The method of embodiment A77, A78 or A79, wherein the pre-filter functions as a pre-filter for one or more other filter species.

A81 . The method of any one of embodiments A77 to A80, wherein filtration throughput is increased when a pre-filter is used relative to filtration throughput when a pre-filter is not used. A82. The method of embodiment A81 , wherein filtration throughput is increased by at least about 1 .5-fold.

A83. The method of embodiment A81 , wherein filtration throughput is increased by at least about 3-fold.

A84. The method of any one of embodiments A80 to A83, wherein the one or more other filter species have pore sizes ranging from about 0.2 microns to about 3.0 microns.

A85. The method of embodiment A84, wherein the one or more other filter species are selected from filters having pore sizes of about 0.8-3.0 microns and 0.2-1.0 microns.

A86. The method of embodiment A84 or A85, wherein the one or more other filter species are selected from filters having pore sizes of about 1 .2 microns, 0.8 microns and 0.45 microns. A87. The method of any one of embodiments A73 to A86, wherein at least one filter is a depth filter.

A88. The method of embodiment A87, wherein the depth filter is a stacked depth filter. A89. The method of embodiment A87 or A88, wherein filtration throughput is increased when a depth filter is used relative to filtration throughput when a depth filter is not used.

A90. The method of any one of embodiments A1 to A89, further comprising after centrifugation combining the viral harvest with a stabilizer.

A91. The method of embodiment A90, wherein the viral harvest is combined with a stabilizer to obtain a final concentration of 6-8% sucrose weight/volume (w/v), 1-2% arginine w/v, 0.05-0.1 % monosodium glutamic acid w/v and 0.5-2% gelatin hydrolysate.

A92. The method of embodiment A90, wherein the viral harvest is combined with a stabilizer to obtain a final concentration of 6-8% sucrose weight/volume (w/v), 1-2% arginine w/v, and 0.5- 2% gelatin hydrolysate.

A93. The method of embodiment A91 , wherein the final concentration is 6.84% sucrose weight/volume (w/v), 1 .21 % arginine w/v, 0.094% monosodium glutamic acid w/v, and 1 % gelatin hydrolysate. A94. The method of embodiment A92, wherein the final concentration is 6.84% sucrose weight/volume (w/v), 1 .21 % arginine w/v, and 1 % gelatin hydrolysate.

A95. The method of any one of embodiments A1 to A94, wherein the influenza virus composition is a refrigerator-stable influenza virus composition

A96. The method of embodiment A95, wherein the influenza virus composition exhibits a potency loss of less than 1 .0 log over a 6 to 12 month period when stored at 4°C to 8°C.

A97. The method of any one of embodiments A1 to A96, wherein the influenza viruses comprise live influenza viruses.

A98. The method of any one of embodiments A1 to A97, wherein the influenza viruses comprise reassortant influenza viruses. A99. The method of embodiment A98, wherein the reassortant influenza viruses comprise hemagglutinin and/or neuraminidase antigens in the context of an attenuated and/or

temperature sensitive and/or cold adapted master strain.

A100. The method of embodiment A99, wherein the master strain is selected from the group consisting of A/Ann Arbor/6/60, B/Ann Arbor/1/66, PR8, B/Leningrad/14/17/55, LEN-B14/5/1 , B/USSR/60/69, B/Leningrad/179/86, B/Leningrad/14/55 and B/England/2608/76.

A101 . The method of embodiment A99, wherein the master strain is derived from a master strain selected from the group consisting of A/Ann Arbor/6/60, B/Ann Arbor/1/66, PR8,

B/Leningrad/14/17/55, LEN-B14/5/1 , B/USSR/60/69, B/Leningrad/179/86, B/Leningrad/14/55 and B/England/2608/76.

A102. The method of any one of embodiments A1 to A101 , further comprising after

centrifugation blending the viral harvest with at least one other viral harvest, thereby producing a blended viral harvest.

A103 The method of embodiment A5 or A8, further comprising blending the sterilized viral harvest with at least one other sterilized viral harvest, thereby producing a blended viral harvest.

A104. The method of embodiment A102 or A103, wherein the viral harvest is blended with two other viral harvests, thereby producing a trivalent blended viral harvest. A105. The method of embodiment A102 or A103, wherein the viral harvest is blended with three other viral harvests, thereby producing a quadrivalent blended viral harvest.

A106. The method of embodiment A105, wherein the quadrivalent blended viral harvest comprises two influenza A strains and two influenza B strains.

A107. The method of embodiment A105, wherein the quadrivalent blended viral harvest comprises three influenza A strains and one influenza B strain. A108. The method of embodiment A105, wherein the quadrivalent blended viral harvest comprises one influenza A strain and three influenza B strains.

A109. The method of any one of embodiments A1 to A108, which comprises formulating the viral harvest, whereby an influenza virus composition suitable for intranasal administration is produced.

A1 10. The method of any one of embodiments A1 to A108, which comprises formulating the viral harvest, whereby an influenza virus composition suitable for administration to a human is produced.

B1. A method for making an influenza virus composition comprising:

b) subjecting the clarified viral harvest to centrifugation, which centrifugation comprises continuous zonal centrifugation performed over a sucrose density gradient, wherein the sucrose density gradient is generated by combining a volume of a 60% (w/w) sucrose composition and a volume of a 10% (w/w) sucrose composition, wherein the volume of the 60% (w/w) sucrose composition is equal to or greater than the volume of the 10% (w/w) sucrose composition; thereby producing a further clarified viral harvest; and, optionally,

c) sterilizing by sterile filtration the further clarified viral harvest, thereby producing a sterilized viral harvest.

B2. The method of embodiment B1 , wherein the sucrose density gradient is a 0% to 60% sucrose gradient.

B3. The method of embodiment B1 , wherein the sucrose density gradient is a 10% to 60% sucrose gradient.

B4. The method of embodiment B1 , B2 or B3, wherein the volume of the 60% sucrose (w/w) composition is at least about 1.1 times greater than the volume of the 10% sucrose (w/w) composition. B5. The method of embodiment B1 , B2 or B3, wherein the sucrose density gradient is generated using volumes of a 60% sucrose (w/w) composition, a 10% sucrose (w/w) composition and PBS at a ratio of 1.3-1.6 to 1.2-1.5 to 0.4, respectively.

B6. The method of embodiment B5, wherein the sucrose density gradient is generated using volumes of a 60% sucrose (w/w) composition, a 10% sucrose (w/w) composition and PBS at a ratio of 1.5 to 1 .3 to 0.4, respectively.

B7. The method of any one of embodiments B1 to B6, wherein the centrifugation in (c) has a run time of at least about 9 hours.

B8. The method of any one of embodiments B1 to B6, wherein the centrifugation in (c) has a run time of at least about 12 hours.

B9. The method of any one of embodiments B1 to B8, wherein the further clarified viral harvest in (c) is collected from the sucrose density gradient at gradient coordinates between about 35% to about 49% sucrose.

B10. The method of any one of embodiments B1 to B9, wherein after centrifugation the viral harvest is diluted with a buffer.

B1 1 . The method of any one of embodiments B1 to B9, wherein after centrifugation and before sterilizing, the viral harvest is diluted with a buffer.

B12. The method of embodiment B10 or B1 1 , wherein the buffer is a phosphate buffer.

B13. The method of any one of embodiments B1 to B12, further comprising after (a), concentrating the clarified viral harvest, thereby producing a concentrated viral harvest.

B14. The method of embodiment B13, wherein the concentrating comprises use of a tangential flow filtration (TFF) process. B15. The method of embodiment B14, wherein the TFF process comprises use of a hollow fiber cartridge.

B16. The method of embodiment B15, wherein the hollow fiber cartridge has a pore size ranging from about 500 kD to about 750 kD.

B17. The method of embodiment B16, wherein the hollow fiber cartridge has a pore size of about 500 kD. B18. The method of embodiment B16, wherein the hollow fiber cartridge has a pore size of about 750 kD.

B19. The method of any one of embodiments B14 to B18, wherein the TFF process is performed using a shear rate ranging from about 10,000 s^"1 to about 16,000 s^'

B20. The method of any one of embodiments B14 to B19, wherein the TFF process is performed using a transmembrane pressure (TMP) ranging from about 10 psig to about 20 psig.

B21. The method of any one of embodiments B14 to B20, wherein the TFF process is performed using a load factor ranging from about 50 L to about 100 L of clarified viral harvest per square meter.

B22. The method of any one of embodiments B14 to B21 , wherein the TFF process is performed at a filtrate flux rate of at least about 25 LMH.

B23. The method of any one of embodiments B13 to B22, wherein the clarified viral harvest is concentrated at least about 2-fold.

B24. The method of embodiment B23, wherein the clarified viral harvest is concentrated at least about 4-fold.

B25. The method of embodiment B24, wherein the clarified viral harvest is concentrated at least about 5-fold. B26. The method of embodiment B25, wherein the clarified viral harvest is concentrated at least about 6-fold.

B27. The method of embodiment B26, wherein the clarified viral harvest is concentrated at least about 7-fold.

B28. The method of embodiment B27, wherein the clarified viral harvest is concentrated at least about 10-fold. B29. The method of embodiment B28, wherein the clarified viral harvest is concentrated at least about 20-fold.

B30. The method of embodiment B29, wherein the clarified viral harvest is concentrated at least about 50-fold.

B31 . The method of embodiment B30, wherein the clarified viral harvest is concentrated at least about 100-fold.

B32. The method of any one of embodiments B13 to B31 , wherein viral yield is increased relative to viral yield of a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

B33. The method of embodiment B32, wherein the viral yield is increased at least about 2%. B34. The method of embodiment B33, wherein the viral yield is increased at least about 5%. B35. The method of embodiment B34, wherein the viral yield is increased at least about 10%. B36. The method of embodiment B35, wherein the viral yield is increased at least about 15%.

B37. The method of embodiment B36, wherein the viral yield is increased at least about 20%. B38. The method of embodiment B37, wherein the viral yield is increased at least about 50%. B39. The method of embodiment B38, wherein the viral yield is increased at least about 70%.

B40. The method of any one of embodiments B13 to B39, wherein the amount of clarified viral harvest subjected to centrifugation in (b) is greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

B41. The method of embodiment B40, wherein the amount of clarified viral harvest subjected to centrifugation in (b) is at least about 10% greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

B42. The method of embodiment B40, wherein the amount of clarified viral harvest subjected to centrifugation in (b) is at least about 20% greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

B43. The method of embodiment B40, wherein the amount of clarified viral harvest subjected to centrifugation in (b) is at least about 40% greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

B44. The method of any one of embodiments B13 to B40, wherein all or substantially all of the clarified viral harvest is subjected to centrifugation in (b).

B45. The method of any one of embodiments B1 to B44, wherein the centrifugation in (b) is performed at about 2 °C to about 25 °C.

B46. The method of embodiment B45, wherein the centrifugation in (b) is performed at about 2 °C to about 14 °C.

B47. The method of any one of embodiments B1 to B46, wherein the centrifugation in (b) is performed at a speed of about 30,000 RPM to about 40,000 RPM. B48. The method of any one of embodiments B13 to B47, comprising in (b) loading the concentrated viral harvest into a centrifuge device at a particular loading flow rate.

B49. The method of embodiment B48, wherein the loading flow rate is lower relative to a loading flow rate for centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

B50. The method of embodiment B49, wherein the loading flow rate is less than about 200 mL/min.

B51 . The method of embodiment B50, wherein the loading flow rate is less than about 160 mL/min.

B52. The method of embodiment B51 , wherein the loading flow rate is less than about 150 mL/min.

B53. The method of embodiment B52, wherein the loading flow rate is less than about 130 mL/min. B54. The method of embodiment B53, wherein the loading flow rate is less than about 120 mL/min.

B55. The method of embodiment B54, wherein the loading flow rate is less than about 100 mL/min.

B56. The method of embodiment B48 or B49, wherein the loading flow rate ranges from about 120 mL/min to about 160 mL/min.

B57. The method of embodiment B48 or B49, wherein the loading flow rate ranges from about 140 mL/min to about 180 mL/min.

B58. The method of any one of embodiments B1 to B57, wherein the clarifying in (a) comprises use of at least two filter species. B59. The method of embodiment B58, wherein the clarifying in (a) comprises use of at least three filter species.

B60. The method of embodiment B58 or B59, wherein the filter species comprise at least one pre-filter.

B61 . The method of embodiment B60, wherein the pre-filter has a pore size ranging from about 3 microns to about 20 microns. B62. The method of embodiment B61 , wherein the pre-filter has a pore size of about 8 microns.

B63. The method of embodiment B60, B61 or B62, wherein the pre-filter functions as a pre-filter for one or more other filter species. B64. The method of any one of embodiments B60 to B63, wherein filtration throughput is increased when a pre-filter is used relative to filtration throughput when a pre-filter is not used.

B65. The method of embodiment B64, wherein filtration throughput is increased by at least about 1 .5-fold.

B66. The method of embodiment B64, wherein filtration throughput is increased by at least about 3-fold.

B67. The method of any one of embodiments B63 to B66, wherein the one or more other filter species have pore sizes ranging from about 0.2 microns to about 3.0 microns.

B68. The method of embodiment B67, wherein the one or more other filter species are selected from filters having pore sizes of about 0.8-3.0 microns and 0.2-1.0 microns. B69. The method of embodiment B67 or B68, wherein the one or more other filter species are selected from filters having pore sizes of about 1 .2 microns, 0.8 microns and 0.45 microns.

B70. The method of any one of embodiments B58 to B69, wherein at least one filter is a depth filter. B71 . The method of embodiment B70, wherein the depth filter is a stacked depth filter.

B72. The method of embodiment B70 or B71 , wherein filtration throughput is increased when a depth filter is used relative to filtration throughput when a depth filter is not used.

B73. The method of any one of embodiments B1 to B72, further comprising after (b) or (c) combining the viral harvest with a stabilizer. B74. The method of embodiment B73, wherein the sterilized viral harvest is combined with a stabilizer to obtain a final concentration of 6-8% sucrose weight/volume (w/v), 1 -2% arginine w/v, 0.05-0.1 % monosodium glutamic acid w/v and 0.5-2% gelatin hydrolysate.

B75. The method of embodiment B73, wherein the sterilized viral harvest is combined with a stabilizer to obtain a final concentration of 6-8% sucrose weight/volume (w/v), 1 -2% arginine w/v, and 0.5-2% gelatin hydrolysate.

B76. The method of embodiment B74, wherein the final concentration is 6.84% sucrose weight/volume (w/v), 1 .21 % arginine w/v, 0.094% monosodium glutamic acid w/v, and 1 % gelatin hydrolysate.

B77. The method of embodiment B75, wherein the final concentration is 6.84% sucrose weight/volume (w/v), 1 .21 % arginine w/v, and 1 % gelatin hydrolysate. B78. The method of any one of embodiments B1 to B77, wherein the influenza virus composition is a refrigerator-stable influenza virus composition

B79. The method of embodiment B78, wherein the influenza virus composition exhibits a potency loss of less than 1 .0 log over a 6 to 12 month period when stored at 4°C to 8°C.

B80. The method of any one of embodiments B1 to B79, wherein the influenza viruses comprise live influenza viruses. B81 . The method of any one of embodiments B1 to B80, wherein the influenza viruses comprise reassortant influenza viruses.

B82. The method of embodiment B81 , wherein the reassortant influenza viruses comprise hemagglutinin and/or neuraminidase antigens in the context of an attenuated and/or

temperature sensitive and/or cold adapted master strain.

B83. The method of embodiment B82, wherein the master strain is selected from the group consisting of A/Ann Arbor/6/60, B/Ann Arbor/1/66, PR8, B/Leningrad/14/17/55, LEN-B14/5/1 , B/USSR/60/69, B/Leningrad/179/86, B/Leningrad/14/55 and B/England/2608/76.

B84. The method of embodiment B82, wherein the master strain is derived from a master strain selected from the group consisting of A/Ann Arbor/6/60, B/Ann Arbor/1/66, PR8,

B85. The method of any one of embodiments B1 to B84, further comprising after (b) or (c) blending the viral harvest with at least one other viral harvest, thereby producing a blended viral harvest.

B86. The method of embodiment B85, wherein the viral harvest is blended with two other viral harvests, thereby producing a trivalent blended viral harvest.

B87. The method of embodiment B85, wherein the viral harvest is blended with three other viral harvests, thereby producing a quadrivalent blended viral harvest.

B88. The method of embodiment B87, wherein the quadrivalent blended viral harvest comprises two influenza A strains and two influenza B strains. B89. The method of embodiment B87, wherein the quadrivalent blended viral harvest comprises three influenza A strains and one influenza B strain.

B90. The method of embodiment B87, wherein the quadrivalent blended viral harvest comprises one influenza A strain and three influenza B strains. B91 . The method of any one of embodiments B1 to B90, which comprises formulating the viral harvest, whereby an influenza virus composition suitable for intranasal administration is produced.

B92. The method of any one of embodiments B1 to B90, which comprises formulating the viral harvest, whereby an influenza virus composition suitable for administration to a human is produced.

The entirety of each patent, patent application, publication and document referenced herein hereby is incorporated by reference. Citation of the above patents, patent applications, publications and documents is not an admission that any of the foregoing is pertinent prior art, nor does it constitute any admission as to the contents or date of these publications or documents.

Modifications may be made to the foregoing without departing from the basic aspects of the technology. Although the technology has been described in substantial detail with reference to one or more specific embodiments, those of ordinary skill in the art will recognize that changes may be made to the embodiments specifically disclosed in this application, yet these modifications and improvements are within the scope and spirit of the technology. The technology illustratively described herein suitably may be practiced in the absence of any element(s) not specifically disclosed herein. Thus, for example, in each instance herein any of the terms "comprising," "consisting essentially of," and "consisting of" may be replaced with either of the other two terms. The terms and expressions which have been employed are used as terms of description and not of limitation, and use of such terms and expressions do not exclude any equivalents of the features shown and described or portions thereof, and various modifications are possible within the scope of the technology claimed. The term "a" or "an" can refer to one of or a plurality of the elements it modifies (e.g., "a reagent" can mean one or more reagents) unless it is contextually clear either one of the elements or more than one of the elements is described. The term "about" as used herein refers to a value within 10% of the underlying parameter (i.e., plus or minus 10%), and use of the term "about" at the beginning of a string of values modifies each of the values (i.e., "about 1 , 2 and 3" refers to about 1 , about 2 and about 3). For example, a weight of "about 100 grams" can include weights between 90 grams and 1 10 grams. Further, when a listing of values is described herein (e.g., about 50%, 60%, 70%, 80%, 85% or 86%) the listing includes all intermediate and fractional values thereof (e.g., 54%, 85.4%). Thus, it should be understood that although the present technology has been specifically disclosed by representative embodiments and optional features, modification and variation of the concepts herein disclosed may be resorted to by those skilled in the art, and such modifications and variations are considered within the scope of this technology.

Certain embodiments of the technology are set forth in the claim(s) that follow(s).

Claims

What is claimed is:

1. A method for making an influenza virus composition comprising subjecting a concentrated viral harvest comprising influenza viruses to centrifugation, thereby producing a clarified viral harvest.

2. The method of claim 1 , wherein the viral harvest is initially clarified by filtration prior to or during concentration.

3. The method of claim 1 or 2, further comprising after centrifugation sterilizing by sterile filtration the viral harvest, thereby producing a sterilized viral harvest.

4. The method of claim 1 , 2 or 3, wherein the viral harvest is concentrated by a process comprising tangential flow filtration (TFF) process.

5. The method of claim 4, wherein the TFF process comprises use of a hollow fiber cartridge, wherein the hollow fiber cartridge has a pore size of about 500 kD or about 750 kD.

6. The method of claim 4 or 5, wherein the TFF process is performed using one or more parameters selected from the group consisting of:

a) a shear rate ranging from about 10,000 s^"1 to about 16,000 s^"1;

b) a transmembrane pressure (TMP) ranging from about 10 psig to about 20 psig;

c) a load factor ranging from about 50 L to about 100 L of clarified viral harvest per square meter; and

d) a filtrate flux rate of at least about 25 LMH.

7. The method of any one of claims 1 to 6, wherein the amount of viral harvest subjected to centrifugation is greater relative to the amount of clarified viral harvest subjected to centrifugation in a method that does not comprise concentrating the clarified viral harvest prior to centrifugation.

8. The method of any one of claims 1 to 7, wherein the centnfugation is performed using one or more parameters selected from the group consisting of:

a) a temperature of about 2 °C to about 25 °C;

b) a speed of about 30,000 RPM to about 40,000 RPM; and

c) a run time of at least about 9 hours.

9. The method of any one of claims 1 to 8, comprising prior to or during centrifugation loading the concentrated viral harvest into a centrifuge device at a loading flow rate that is lower relative to a loading flow rate for centrifugation in a method that does not comprise concentrating the viral harvest prior to centrifugation, wherein the loading flow rate is selected from the group consisting of: less than about 200 mL/min; less than about 160 mL/min; less than about 150 mL/min; less than about 130 mL/min; less than about 120 mL/min; less than about 100 mL/min; about 120 mL/min to about 160 mL/min; and about 140 mL/min to about 180 mL/min.

10. The method of any one of claims 1 to 9, wherein the centrifugation comprises continuous zonal centrifugation, and is performed over a sucrose density gradient selected from the group consisting of: a 0% to 100% sucrose gradient; a 0% to 60% sucrose gradient; and a 10% to 60% sucrose gradient.

1 1 . The method of claim 10, wherein the sucrose density gradient is generated according to a process selected from the group consisting of:

a) using equal or substantially equal volumes of a 60% sucrose (w/w) composition and a 10% sucrose (w/w) composition;

b) using a volume of a 60% sucrose (w/w) composition that is greater than the volume of a 10% sucrose (w/w) composition, wherein the volume of the 60% sucrose (w/w) composition is at least about 1 .1 times greater than the volume of the 10% sucrose (w/w) composition;

c) using volumes of a 60% sucrose (w/w) composition, a 10% sucrose (w/w) composition and PBS at a ratio of 1.3-1.6 to 1.2-1.5 to 0.4, respectively; and

d) using volumes of a 60% sucrose (w/w) composition, a 10% sucrose (w/w) composition and PBS at a ratio of 1.5 to 1 .3 to 0.4, respectively.

12. The method of claim 10 or 1 1 , wherein after centrifugation the viral harvest is collected from the sucrose density gradient at gradient coordinates between about 35% to about 49% sucrose.

13. The method of any one of claims 2 to 12, wherein the initial clarification comprises use of at least two filter species, wherein:

a) the filter species comprise at least one pre-filter having a pore size ranging from about 3 microns to about 20 microns;

b) the pre-filter functions as a pre-filter for one or more other filter species; and c) filtration throughput is increased by at least about 1.5-fold when a pre-filter is used relative to filtration throughput when a pre-filter is not used.

14. The method of claim 13, wherein the one or more other filter species are selected from filters having pore sizes of:

a) about 0.2 microns to about 3.0 microns;

b) about 0.8-3.0 microns;

c) 0.2-1 .0 microns; and

d) about 1.2 microns;

e) about 0.8 microns;

f) about 0.45 microns.

15. The method of claim 13 or 14, wherein at least one filter is a depth filter, and wherein filtration throughput is increased when a depth filter is used relative to filtration throughput when a depth filter is not used.

16. The method of any one of claims 1 to 15, wherein the influenza virus composition is a live refrigerator-stable influenza virus composition, which composition exhibits a potency loss of less than 1 .0 log over a 6 to 12 month period when stored at 4°C to 8°C.

17. The method of any one of claims 1 to 16, wherein the influenza viruses comprise reassortant influenza viruses comprising hemagglutinin and/or neuraminidase antigens in the context of an attenuated and/or temperature sensitive and/or cold adapted master strain.

18. The method of claim 17, wherein the master strain is selected from or is derived from the group consisting of: A/Ann Arbor/6/60, B/Ann Arbor/1/66, PR8,

B/Leningrad/14/17/55, LEN-B14/5/1 , B/USSR/60/69, B/Leningrad/179/86,

B/Leningrad/14/55 and B/England/2608/76.

19. The method of any one of claims 1 to 18, further comprising after centrifugation blending the viral harvest with at least one other viral harvest, thereby producing a blended viral harvest.

20. The method of claim 19, wherein the viral harvest is blended with three other viral harvests, thereby producing a quadrivalent blended viral harvest, which quadrivalent blended viral harvest comprises a combination of influenza A and influenza B strains selected from the group consisting of:

a) two influenza A strains and two influenza B strains;

b) three influenza A strains and one influenza B strain; and

c) one influenza A strain and three influenza B strains.

21 . A method for making an influenza virus composition comprising:

a) clarifying a viral harvest comprising influenza viruses by filtration, wherein the clarifying comprises use of a pre-filter having a pore size of 8 microns and two other filter species having pore sizes of 1 .2 microns and 0.8/0.45 microns, thereby producing a clarified viral harvest;

b) concentrating the clarified viral harvest using a tangential flow filtration (TFF) process, wherein the TFF process is performed at a shear rate of 15,000 s^"1 and a transmembrane pressure of 15 psig; thereby producing a concentrated viral harvest; c) subjecting the concentrated viral harvest to centrifugation, which centrifugation comprises continuous zonal centrifugation performed over a 0% to 60% sucrose density gradient, wherein the sucrose density gradient is generated by combining a volume of a 60% (w/w) sucrose composition, a volume of a 10% (w/w) sucrose composition and PBS at a ratio of 1.5 to 1 .3 to 0.4, respectively; thereby producing a further clarified viral harvest; and, optionally,

d) sterilizing by sterile filtration the further clarified viral harvest, thereby producing a sterilized viral harvest.