SE1930321A1 - A method to feed component in perfusion culture - Google Patents

Abstract

This invention relates to a method for rationally designing the concentrations of nutrient(s) in cell culture media for use in cell cultures in perfusion mode, e.g., cell cultures for the production of polypeptide production, by selecting an appropriate target cell specific consumption rate of the nutrient(s).

Description

Stockholm, Datum: 2019-10-08 Page 1 of 14 Att: Patent - och RegistreringsverketBox 5055 10242 Stockholm Fax: 08 66 60 286 svENsK PATENTANSÖKAN Titel på uppfinningen: A method to feed component in perfusion culture UPPFINNARE OCH SÖKANDEVéronique Chotteau Adress: Långsjövägen 5, 13133 Nacka, Sweden Hubert SchwarzAdress: Studentbacken 21, Lgh 1503, 11557 Stockholm, Sweden Liang ZhangAdress: Körsbärvägen, 114 23, Stockholm, Sweden Description of the invention Large molecules and assemblies such as polypetides, proteins, enzymes, viral vectors,viruses, polysaccharides, polymers, exosomes, mRNA, siRNA, need to bemanufactured by cells due to their complex and/or large structure. These molecules orassemblies are used for applications such as human therapy, animal therapy,diagnostics, support for analytical methods, support for the biomanufacturing usingcells, food, etc... In the text, 'large molecules' stands for large molecules andassemblies such as polypetides, proteins, antibodies, enzymes, viral vectors, viruses,polysaccharides, polymers, exosomes, mRNA, siRNA. For the production of largemolecules, living cells are often used. This is typically carried out in a vessel, such as a bioreactor, in which the cells are cultured in a culture medium providing the Stockholm, Datum: 2019-10-08 Page 2 of 14 nutrients and other necessary factors to the cells.

Operation in perfusion mode consists in applying a continuous or semi-continuous orinterrnittent renewal of the culture medium and additives. This results in renewal ofthe necessary nutrients and removal of the by-products, and is therefore creating afavorable environment for the cell health, cell metabolism, cell growth, andpotentially for the product of interest. Perfusion mode allows as well increasing thecell density to very high levels, potentially higher than 200 x 106 cells/mL [Clincke,M. F., C. Molleryd, Y. Zhang, E. Lindskog, K. Walsh and V. Chotteau (2013). "Veryhigh density of CHO cells in perfusion by ATF or TFF in WAVE bioreactor. Part I.Effect of the cell density on the process." Biotechnology Progress 29(3): 754-767].

In animal cell processes, bi-products can potentially have inhibitory effects on the cellgrowth and/or the productivity of the product of interest and can affect the productquality, such as the glycosylation [Hassell, T., S. Gleave and M. Butler (1991)."Growth inhibition in animal cell culture. The effect of lactate and ammonia." App_lBiochem Biotechnol 30(1): 29-41]. For instance, sugar and amino acid availability inmammalian cell culture media are usually higher than the cell needs for their energy production, assimilation in thebiomass and production of the product of interest. This excess of sugar and/or aminoacid can induce an inefficient high uptake of these nutrients, resulting in theproduction of inhibitory levels of waste metabolites such as lactate from sugar uptakeand/or ammonia from amino acid uptake. Examples of sugar are glucose, galactose,fructose, etc. and examples of amino acids are glutamine, asparagine, etc. Limitedfeeding of glucose or glutamine are known to generate low cell specific intake ofthese nutrients and subsequently more efficient cell metabolism in fed-batchprocesses. However if one desires to apply the same principle in perfusion culture, theonly method, which exists today to determine how these nutrients should be fed in theculture, is to select arbitrary feeding concentration by trial and error. Importantfactors are that these nutrients have to be fed in amount satisfying for theperformances of the run, where a) too high level will generate production of by-products with unfavorable effects on these performances, and b) too high level will beunnecessary high in the medium where the component concentrations need to beminimized in order to avoid medium precipitation, while c) too low level will result inunfavorable effects related to limitation of this nutrient such as glucose limitation can cause unfavorable glycosylation. Here performances are satisfying typically in terms Stockholm, Datum: 2019-10-08 Page 3 of 14 of by-products derived from glucose such as lactate, as Well cell health, production of product of interest and its quality.

The herein invention presents a novel method for the delivery or feed of one orseveral nutrients in a culture performed in perfusion mode. In this method, thesenutrients can be delivered as part of the medium or as additives fed separately from the medium. A substrate or component delivered to the perfusion culture can be defined by its concentration (e.g. given in millimolar/ liter or g/L) and its flow rate (e. g. given in milliliter/minute or liter/minute). For simplicity, the method is first describedto feed glucose When this substrate is part of the medium fed into the perfusion process. The principle of the method is that the operator selects the target of the cell specific consumption rate, called qtarget at Which the cells Will use this substrate. glucose»The cell specific consumption rate of a component is the rate at Which the cells aretaking up this component from the culture medium, calculated per cell. In perfusionprocess, it is desirable to maintain the level of a given component constant in theculture, so the variation With time of this component is ideally zero. Furthermore,since the cells have a tendency to consume higher amounts of glucose When thissubstrate is delivered more abundantly, in the present invention, We design a method to deliver glucose such that the residual concentration of glucose in the bioreactor is a targetresidual' target given target, glucose Tesi-dm The glucose l is taken as for instance equal to zero. This can be changed to a given level different from zero, and this is presented in the next paragraph. Given that the variation of glucose in the bioreactor is zero and targetresidua target g lucose , the concentration that glucose l is zero, in order to obtain the desired q of lucose to be fed into the bioreactor, lucoselN lucosem , has to be as followsg g g glucosem _ target * X/D [eq 1] _ qglucoseWhere X is the cell density, D is perfusion or reneWal rate of the medium in the bioreactor. This can also be Written as glucoselN _ target /CSPR [eq 2] _ qgiuwseWhere CSPR is the cell specific perfusion rate and is equal to D / X.The application of eq 1 and eq 2 are counter intuitive and against the Way a personexperimented in developing the cell culture processes in the field of biologics manufacturing Would norrnally select the glucosem. This person Would select a Stockholm, Datum: 2019-10-08 Page 4 of 14 concentration to feed a given nutrient based on his experience and knowledge of thecells and the process and would then measure (i.e. monitor) the cell specificconsumption rate of glucose in an experiment as information about the performancesof the experiment. Surprisingly, application of eq 1 and eq 2 is an efficient way to deliver nutrients in perfusion as illustrated in the examples Section. target In case glucoseresidual is different from zero, eq 1 and eq 2 are replaced by eq 3 and eq 4, as follows glucoselN = ( qgfgffste * X/ D ) + glucosefgsråïal [eq 3]glucosem = ( qgfgffste /CSPR ) + glucosefgsråïal [eq 4] targetglucose The selection of q can be made by trying a given value during several days in perfusion according to eq 3 and eq 4 and studying if the performances are satisfyingin terms of by-products derived from glucose such as lactate, as well cell health, production of product of interest and its quality. When glucose is fed following eq 3 targetglucose and eq 4 with q between 0.9 and 1.3 picomol/cell/day, the production of lactate is remaining at a level, which does not damage the cells and the process performances, as well as glucose is provided in amounts sufficient so that the glycosylation is not affected. For qgllgfjste equal to 0.8 picomol/cell/day or lower, the glycosylation might be affected.

Eq 3 and 4 can be generalized for any nutrient or cofactor taken up by the cells, as follows: nutrientm = ( qäTtfl-eetnt * X/ D ) + nutrientïfåïal [eq 5] nutrientm = ( qäTtfl-eetnt / CSPR ) + nutrientïfåïal [eq 6] where nutrientm is the concentration of the nutrient fed during the perfusion,:m is the selected cell specific consumption rate of the nutrient and nutrientíísrl-gclïal is the concentration of the nutrient in the bioreactor.

In case the animal cells can use different nutrients for a comparable usage and these are fed to the culture, a variant of eq 5 and 6 can be used as follows. Suppose that the target target cell specific nutrient consumption rate qnum-entj is selected for this type ofnutrient. This selection can be based on the production of by-products by the cells, oron the performances of cell growth, cell viability, productivity of product of interest, CÉC.

Stockholm, Datum: 2019-10-08 Page 5 of 14 When N nutrients of comparable usage are fed instead of one nutrient, it is imposed target It that the sum of the N target cell specific nutrient consumption rates is qnutn-ent T . follows for nutrient i (i = 1, 2, ..., N): . IN I target * . target N target _nutrlenti ( qnutrientj X / D ) + nutrlentlflresidual and 2i=1 qnutrientjtargetqnutrient] [eq 7]. lf lf . É Énutrtentf” = ( qnïTtfl-eent i / CSPR ) + nutrtentiïfieciual andN target _ targeti=1 nutrientj _ nutrient] [eq Where nutrientf” is the concentration of the nutrient i fed during the perfusion, target num-entj is the selected cell specific consumption rate of nutrient i and target ïlutïleïltl-'Tesl-dua l is the concentration of nutrient z' in the bioreactor. The selection of target num-entJ- has to take into account as Well the transport rate of the nutrient into the cells in case this transport is limited.

The principle given in eq 1 to 8 is applicable to any component of the medium or anadditive or a feed fed during a perfusion culture, for the selection of the concentrationof this component. These components can be, but not restricted to, amino acid, growthfactor, horrnones, metal co-factors, salts, surfactant, anti-foam, lipid, fatty acid, proteins, etc. ..

Stockholm, Datum: 2019-10-08 Page 6 of 14 Exemplary embodiments Embodiment 1 In Embodiment l, a culture of HEK293 cells producing human erythropoietin(rhEPO) is performed in a bioreactor of 200 mL working volume where the cells aregrown in suspension. This cell line expresses rhEPO since the gene of humanerythropoietin has been inserted in the cells by recombinant technology. The purposeof this culture is to produce EPO, which is thus the product of interest.

To initiate the culture, the bioreactor is filled with culture medium. This mediumcontains the components necessary for the cells to grow and produce rhEPO, such assugar, amino acids, vitamins, salts, buffer, metal traces, organic components. Thecells are inoculated at day 0 from a stock culture maintained in shake flasks. Thebioreactor is equipped with a cell retention system. This allows performing perfusionoperation. The culture medium is continuously removed from the culture while freshmedium is automatically added so that the culture level is maintained.

To initiate the experimental work, a preliminary perfusion test run was performed inmini-bioreactor, mini-EPO#l;04, using feeding medium (from the shelf) containing27.7 mM glucose. In the bioreactor during the run mini-EPO#l;04, the glucoseconcentration was measured to vary between 5 and 17 mM (see Figure lb), and thelactate concentration (see Figure ld) was measured to exceed 20 mM at several daysdespite trial to remove the lactate with a high cell specific perfusion rate CSPR (seeFigure le). The viability (see Figure lc) decreased with time. In mini-EPO#2;08, the target present invention was applied and q g m C O s e at levels given in Figure lg was applied using eq l. The residual concentration of lactate decreased significantly from this strategy, reaching much lower lactate concentrations at highest level of 16 mM (see target glucose . The culture was started in batch mode and Figure ld) resulting from lower q targetglucose the medium renewal started at day 4. q values of 1.5 and then l.3 picomol/cell/day resulted in maintaining the lactate concentration around 15 mM - E Ewh1le values of q gg? f s e of l.2 and then l picomol/cell/day (at days l5 to 22 and days 23 to end respectively) resulted in lower concentrations of lactate. Notice that applying the qgllgfjšecaused these lower concentrations of lactate. It is importantto notice that this was not an effect of the medium renewal since the perfusion rate was l reactor volume / day at days 8 to l7 and days 23 to 30 in mini-EPO#2;08 run and at Stockholm, Datum: 2019-10-08 Page 7 of 14 day 7 to day 14 in run n1ini-EPO#1;04 however the lactate concentration Was much higher in run n1ini-EPO#1;04.

Stockholm, Datum: 2019-10-08 Viable cell density (10^6 cells/mL) C100š.å 90:a.E>soE9N'UEså_.E:ID.IIIUå Fi 311111: 1. Periíxsiorl process aïíiísíííí Page 8 of 14 .x\míní_EPo¿¿1,.O4 fimínäpoflzos šg -mini-EPo#1;04 fimini-sPoßzos1817ââÉ 14å 13E 12.e 11š 102 afl 1§ sw iå 3G å0 .. 4040 1'ime(days)Time(clays) , 30 -~mini-Er>o::1;04 fimini-aPouzos-1 25fxv-:ÄEå 20âE'3c 108'\“mini-EPOfl1;O4 fimini-EPOHZOS å 50U2. . . . . 0 I0 10 20 30 40 0 10 20 30 40Time (days) Time (days)240 rèmíníEpofloll 3 +mini-EPo#1;04 fimini-EPoflzßs -X~1L-r0n#6;11c5220 '200180 ' g160 fimini-sPoflzos å 2140 å120 E'ä100 »-80 .å 1 ::60 å40 30 ' 0 .' L, . . . .40 0 10 20 30 40Time (days) Time (days)33 » q_g|ucosetarget-m|n|-EPO#Z,08 EÉ šE 2 Äcellspecificglucoseuptake rate- lE mini-EPO#Z;08 2åE:ä9.”g 1äO03ulD'0 . . . . 20 Time (days) 40 method míní-EPßšä 1 ië-'l :and in 0 nan appkyíng targfli (Lglucose 1110111011niíïii-fiPíÅäiaëïgüä (_21) Viable 0011 asured in bie>reacl;orç 1 (f) vizalïsilíty; (d) ízactzats cfoncantfatiøz: rneasalred in i>iofeaa>tor; (le) ÅÉSPR; (i) perfusicn: ram; ( appïied targe: ggízicfose 20,10 0011 speciñc: 1111101140 tc ibr glucusf: Stockholm, Datum: 2019-10-08 Page 9 of 14 Embodiment 2 In Embodiment 2, a culture of Chinese Hamster Ovary (CHO) cells producing amonoclonal antibody (mAb) is perforrned in a bioreactor of 200 mL working volumewhere the cells are grown in suspension. This cell line expresses mAb since the geneof antibody has been inserted in the cells by recombinant technology. The purpose ofthis culture is to produce mAb, which is thus the product of interest.

To initiate the culture, the bioreactor is filled with culture medium. This mediumcontains the components necessary for the cells to grow and produce mAb, such assugar, amino acids, vitamins, salts, buffer, metal traces, organic components. Thecells are inoculated at day 0 from a stock culture maintained in shake flasks. Thebioreactor is equipped with a cell retention system. This allows performing perfusionoperation. The culture medium is continuously removed from the culture while freshmedium is automatically added so that the culture level is maintained.

Perfusion runs were performed in the same conditions except for the sugars fed ascarbon source. For this glucose or mannose or galactose were used alone or incombination. Glucose or mannose or galactose can be used by the cells for acomparable usage however they are not completely similar. Here we used our newmethod of targeted feeding for these sugars, fed alone or simultaneously into the culture. Here the information of the need in carbon source was taken from the target knowledge of target for glucose feeding, i.e. qmrbon smmej = 1.2 pmol/ cell/ day.

Another information is that the transport of galactose is limited to e.g. 0.4 targetgalactose pmol/cell/day and therefore q cannot be larger than 0.4 pmol/cell/day in this example. Eq 7 and 8 thus are applied as follows: N target I target target + target I target I 1 2 i=1 nutrientj qglucose galactose qmannose carbon source_Tpmol/cell/dayand 1N _ target * targetglucosel- - qgmwse X/D ) + glucoseresl-dual and[N _ ÉCLTQBÉ ÉCLTQBÉmannosel- - ( qmannose * X / D ) + mannose Tesi-dual and1N _ target * target target galactosel- -( qgalacwse X / D ) + galactose Tesi-dual where qgalactose 5 0 4 pmol/ cell/ day Stockholm, Datum: 2019-10-08 Page 10 of 14 target t tgalactose arga I 12 carbon source_T For instance, if one select q = 0.4 pmol/cell/day and q pmol/cell/day, then qffgfiíâse has to be 0.8 pmol/cell/day if galactose and mannose are fed.

In the runs, different combinations, 1 to 10, of sugars according to Table 1 wereapplied during 5 or 6 days before switching to a new combination. Figure 2 to 4shows the cell density, the Viability, the cell specific glucose consumption rate, thelactate production rate, the cell specific consumption rate of the carbon sources andthe glycosylation pattem. These results demonstrate that by applying the designedfeed strategy given by eq 7 and 8, a defined profile for the perfusion operation wereobtained. In particular, this gave in this example a tool to profile the glycosylation in avery efficient way while obtaining a very low lactate production (as demonstrated by the low cell specific lactate production rate) and a Very high cell Viability.

Table 1: Conditions for the targets qflïrtfl-eetntl- and qíâígíï: sourcej all expressed inpicomol/cell/dayConditionfif Åïlffše fßlåíise ÅÉÉÉÉÉOSQ Éïïííisoume;1 1.2 1.22 0.9 0.3 1.23 0.8 0.4 1-24 0.4 0.8 1-25 1.2 1-26 0.9 0.3 1.27 0.45 0.45 0.3 1.28 0.8 0.89 0.4 0.4 0.810 0.55 0.25 0.8 Stockholm, Datum: 2019-10-08 Page 11 of 14 ' \~'- ~.\-.-.~;-.-.\-.-.-.\-.-.~:-:- - -.-~~~Y~-~¿“' ___.~x-.\:_¿>___ wmy _ \\. - x ¿w.ë\\___\\+\--._.¿. \_\\~-.~~ \“\\_ \\\' ~ '« \\§~ Figure 2: Cell density and Viability in conditions 1 to 10 of Table 1. Conditions 1 to are marked by the coloured areas and the numbers 1 to 10.

Stockholm, Datum: 2019-10-08 Page 12 of 14 w \\šš\\xšššx\n\\ 7411/11; _. flflfflfw,»vw/ulf ~\\~\\~\\~~\\~~\\~ xfimfl hä; w _~ ~~~~~= :w- S -\: wmn §\<«-\-,-_:-NN _\m\_\..-_.

Mg»- eü"«"°"^5(~.šßwzswaw s..\.çsxwwyt šße Figure 3: Cell specific glucose consumption rate qglucose and cell specific lactateproduction rate qlactate (Where a negative qlactate indicates a production and apositive qlactate indicates a consumption) in conditions 1 to 10 of Table 1. Conditions 1 to 10 are marked by the coloured areas and the numbers 1 to 10.

Page 13 of 14 Stockholm, Datum: 2019-10-08 .w m n» ................. _. \\\ \ _ ................................ _ .> z _, šššššššš \ Figure 4: Cell specific consumption rate of all the carbon source, cell specific niAbproduction rate (qp) and glycosylation pattern given the G0, G1, G2, high niannoseand fucolysation in conditions 1 to 10 of Table 1 Stockholm, Datum: 2019-10-08 Page 14 of 14 Embodiment 3 In Embodiment 3, a culture of Chinese Hamster Ovary (CHO) cells producing amonoclonal antibody (mAb) is perforrned in a bioreactor of 200 mL working volumewhere the cells are grown in suspension. This cell line expresses mAb since the geneof antibody has been inserted in the cells by recombinant technology. The purpose ofthis culture is to produce mAb, which is thus the product of interest.

To initiate the culture, the bioreactor is filled with culture medium. This mediumcontains the components necessary for the cells to grow and produce mAb, such assugar, amino acids, vitamins, salts, buffer, metal traces, organic components. Thecells are inoculated at day 0 from a stock culture maintained in shake flasks. Thebioreactor is equipped with a cell retention system. This allows performing perfusionoperation. The culture medium is continuously removed from the culture while freshmedium is automatically added so that the culture level is maintained.

In this example, the purpose is to feed the amino acid as efficiently as possible inperfusion mode. An aim is to push the perfusion rate to low level in order to reducethe volume of medium renewal. For this, a perfusion run A is performed in which theCSPR is minimised: it is progressively decreased while ensuring that the culture hasstill satisfactory performances for instance in terms of cell specific productivity ofmAb. The observation of the amino acids in run A show that some amino acids are almost completely while others are not completely consumed. The cell specific E Emye - used consumption rates of these amino acids are taken to be the values of qnutn-ent l in eq 3 and eq 4 where nutríent_í stands for the each amino acid at a time. Thedesigned feeds of eq 3 and 4 are then used to determine the amount of amino acid tobe fed during the perfusion run. A new run B is then performed using this feed. Theamino acids are now all almost completely consumed. By having them almostcompletely consumed, the process is more efficient. Another aspect is that in order tobe able to decrease the perfusion rate the medium and feeds are often concentrated(using a smaller amount of liquid to reconstitute the liquid medium from the powdermedium than indicated by the manufacturer) however, this concentration can easilyprovoke a precipitation of the medium or feed, so here using eq 3 and eq 4 is a way to identify how the amino acid level can be optimal in the medium or feed.

Claims (8)

Claim

1. l. A method of producing a molecule of interest in perfused cell culture, wherein saidperfused cell culture is such that the spent medium is partially or completely replaced byfresh culture medium, consisting in one or several fed mixtures, delivered mixed orseparately, and at least some portion of the cells are retained, comprising: (1) cells at concentration X; (2) culture perfused at flow rate D with a desired fresh culture medium, comprising baseline-adjusted concentrations of a number N of nutrients according to the forrnulas target * . target N target I targetnutrienu. X/ D ) + nutrzent and 2 q - 11v _nutrlenti _ ( q i,residual i=1 qnutrientj nutrient] INti wherein N is an integer between l and 20, nutrien is the concentration of the íth nutrient targetnutrientj nutrientl- in said fresh medium where i is an integer varying from l to N, q is the selected target of the cell specific consumption rate of nutrientl- given as mass per cell per target time unit, nutrienti Tesi-dm l is the selected residual concentration of nutrientl- in the culture,target num-entj is a selected cell specific consumption rate for this type of nutrient, which is the targetnutrientj sum of all the q for i varying from l to N; target targetnutrient] > nutrientj target Mest-dual are selected such that the cell wherein said q and/or nutrient culture is maintained in conditions allowing expression of said molecule of interest, cellsurvival, cell growth, concentration of by-product under a desired concentration, desiredquality attribute of the said molecule of interest, reduction of said flow rate D and/or absenceof precipitation in said fresh culture medium; and wherein, in case N is equal to or larger than 2, the said nutrients nutrientl- consist ofcomponents used in a comparable way by the cells, such as, but not limited to, using at leastone similar intracellular biochemical pathway, using at least one similar transport mechanism into the cells, interacting at least with one similar extracellular receptor and/or involving at least one similar intra-cellular interaction.

2. The method according to claim 1 wherein the said molecule of interest is a polypeptide, aprotein, an antibody, an enzyme, a Viral Vector, a Virus, a polysaccharide, a polymer, anexosome, an mRNA, a siRNA, or any combination of these, or assemblies derived from these,or assemblies derived from these and associated with a small molecule of a size less than 1000 kDa.

3. The method according to claim 1 wherein said nutrients are components taken up by thecells from the medium and selected from the group consisting of sugar, glucose, galactose,mannose, fructose, lactose, maltose, dextrose, glucosamine, pyruvate, lactate, insulin,epiderrnal growth factor, insulin-like growth factor, amino acids, asparagine, aspartic acid,alanine, arginine, cysteine, cystine, glutamine, glutamic acid, glutamate, glycine, proline,serine, tyrosine, isoleucine, leucine, Valine, histidine, lysine, methionine, phenylalanine,threonine, tryptophan, omithine, glutathione, growth factors, horrnones, Vitamins, biotin,pantothenate, choline, folic acid, myo-inositol, nicotinamide, pyridoxal, pyridoxine,riboflavin, thiamine, Vitamin B12, Vitamin K, a-tocopherol acetate, flavin adeninedinucleotide, phosphate, hydrocortisone, putrescine, linoleic acid, thioctic acid, lipoic acid,pluronic, camosine, ethanolamine, oxalacetic acid, xanthine, hypoxanthine, metal co-factors,salts, surfactants, lipids, fatty acids, anti-oxidants, peptides, nucleotides, alpha-ketoglutarate,citrate, isocitrate, oxoglutarate, succinyl-CoA, succinate, fumarate, malate, oxaloacetate and proteins.

4. The method according to any one of claims 1, 2 or 3 wherein said integer N is equal to 1;wherein the culture is perfused at flow rate D with a desired fresh culture medium, comprising a baseline-adjusted nutrient concentration nutrientm according to the forrnula . IN _ target . target - target -nutrtent - ( qnutrient * X / D ) + nutrLentTesl-dual where1n qnutn-ent 1s a selected ratetarget of said nutrient given as mass per cell per time unit and nutrient is a selected residual target nutrient and/or sa1d residual concentration of said nutrient in the culture; wherein said q . f fnutrtent mye residual are selected such that the cell culture is maintained in conditions allowing expression of said molecule of interest, cell survival, cell growth, concentration of by-productunder a desired concentration such as, but not limited to, lactate and ammonia, desired quality attribute of the said molecule of interest, absence of precipitation in said fresh culture medium and/or transport of said nutrient into the cells. Claim

5. The method according to claim l wherein said integer N is between l and 3 and the target carbon source T is the Selected said nutrients are glucose, galactose and/or mannose, wherein q cell specific consumption rate of carbon sources, which is the sum of the selected rates of target targetglucose > galactose target glucose, galactose and mannose, q and qmannose respectively, according to the forrnulas N target target target target _ target i=1 nutrient_i :qglucose galactose +qmannose _ carbonsource_T and 1N _ target * targetglucosel- - qgmwse X/D)+glucoseresidual and targetresidual [N I target mannosel- ( qmannose * X / D ) + mannose and targetgalactose targetresidual 9 galactosef” =( q * X / D ) + galactosetarget Wherein qcarbon source T is between 0.00001 pmol/cell/day and 4 pmol/cell/day;target wherein q gm CO s e is between 0 pmol/cell/day and 4 pmol/cell/day;target Wherein qgalactose is between 0 pmol/cell/day and 4 pmol/cell/day; wherein se is between 0 pmol/cell/day and 4 pmol/cell/day; target wherein glucose Tesi-dm l is between 0 mmol/ liter and 20 mmol/liter; target Tesi-dual is between 0 mmol/liter and 20 mmol/liter; wherein galactose target wherein mannose Tesi-dm l is between 0 mmol/ liter and 20 mmol/ liter; wherein the molecule of interest is a protein of which a desired profile of the quality attribute - - - target targetof glycosylation can be obtained by selecting the values of q C mb On smmej, glucose,target targetgalactose and qmßlïmßSß'

6. The method according to any one of claims l, 2, 3, 4, 5 or 8, wherein the cells areeukaryotic selected from the group consisting of mammalian cells, human cells, aVian cells, insect cells and plant cells.

7. The method of claim 6, wherein the cells are selected from the group consisting of: CHO,CHO-DBX11, CHO-DG44, CHO-S, CHO-Kl, Vero, BHK, HeLa, COS, MDCK, HEK-293,HEK-293T, HEK-293S, HEK-293F, L293, NIH-3T3, W138, BT483, Hs578T, HTB2, BT20,T47D, NSO, CRL7030, HsS78Bst cells, PER.C6, SP2, SPO, hybridoma, MRC-5, MDCK,WI-98, CAP, EB66, AGEl.CR, CR, Trichoplusia ni, Spodoptera Frugiperda, SF9, SF2l, Hi5, mesenchymal stem cells, induced pluripotent stem cells, embryonic stem cells, primary cells,Nicotiana tabacum, BY2, Nicotiana benthamiana, Oriza sativa, Arabidopis thaliana, and Daucus carota.

8. The method according to claim l wherein said integer N is between l and 20 and the saidnutrients are N amino acids selected from asparagine, aspartic acid, alanine, arginine,cysteine, cystine, glutamine, glutamic acid, glycine, proline, serine, tyrosine, isoleucine, leucine, Valine, histidine, lysine, methionine, phenylalanine, threonine and tryptophan; target Wherein qnutrient T is the selected target of cell specific consumption rate of the N selected amino acids, which is the sum of the selected target of cell specific consumption rates of the target N selected amino acids, qnutn-entj, where integer i is a integer Varying from l to N; wherein targetnutrient_T is between 0.0000l pmol/cell/day and l pmol/cell/day; target Wherein qnutrient i is between 0 pmol/cell/day and l pmol/cell/day; wherein the selected residual amino acid concentrations are between 0 mmol/ liter and 20 mmo l/ liter; target target targetnutrient_T > qnutrientj ”est-dual or a combination of these wherein said q or nutrient parameters are selected such that the cell culture is maintained in conditions allowing a cellspecific perfusion rate CSPR, which is equal to D / X, to be reduced to a Value equal to orsmaller than 30 picoliter per cell per day reactor Volume per day, and/or the residualconcentration of by-product lactate in perfused culture can be maintained under 30millimol/ liter, and/or the residual concentration of by-product ammonia in perfused culturecan be maintained under 6 millimol/ liter, in absence of precipitation in said fresh culture medium.