WO2011025919A1 - Mirna biomarkers of lung disease - Google Patents

Abstract

This application describes miRNAs that may be used as serum or plasma biomarkers for characterizing lung disease in a patient. These miRNA biomarkers may be used alone or in combination with other markers for the diagnosis, prognosis, or monitoring of diseases such as lung cancer. In one embodiment, a method for characterizing lung cancer in a patient with a lung tumor or lesion comprising the steps of: measuring the level of a miRNA in a serum sample, and determining reduced or elevated levels of the miRNA in the sample, thereby characterizing lung cancer in the patient.

Description

MIRNA BIOMARKERS OF LUNG DISEASE

[001] This application claims priority to U.S. Provisional Application No. 61/237,972, filed August 28, 2009, which is incorporated by reference in its entirety.

[002] Work described in this application was partially funded by the

Federal government under National Cancer Institution/NIH Grant No.

1R43CA141786-01. Accordingly, the Federal government may have certain rights in this invention.

[003] In the United States, lung cancer has the highest incidence and the highest mortality rate of all cancers (http://seer.cancer.gov/statistics/). Lung cancer is divided into two major classes: small cell lung cancer (SCLC) and non- small cell lung cancer (NSCLC), the former affecting 20% of patients and the latter 80%. NSCLC consists of three major subtypes: adenocarcinoma, squamous cell carcinoma (SCC), and large cell carcinoma, with adenocarcinoma and squamous cell carcinoma accounting for the vast majority (Sekido et al., Biochim Biophys Acta, 1378:F21-59; Forgacs et al., Pathol Oncol Res, 7:6-13 (2001)). An estimated 215,000 new cases of lung cancer and approximately 162,000 deaths from lung cancer occurred in the United States during 2008 (Jemal et al., J Natl Cancer Inst, 100:1672-94 (2008)). The high mortality rate of lung cancer has been attributed in part to the fact that more than 75% of lung cancer patients are diagnosed with regional or distant metastasis at first presentation, and in part to the lack of highly effective therapies for lung cancer. Although the current treatment options for lung cancer are limited, it has been shown that patients who are diagnosed at the earliest stage have a better chance of 5-yr survival (50%) than those in later stages (<10%) (see, e.g., http://seer.cancer.gov/statistics;

Horner et al. (eds). SEER Cancer Statistics Review, 1975-2006, National Cancer Institute. Bethesda, MD). Consequently, extensive efforts have been directed to the detection of lung cancer at an early stage so that medical intervention may improve survival.

[004] Risk factors for lung cancer include age and tobacco consumption by smoking (first hand and second hand). Approximately 90% of all lung cancer cases occur in smokers. Additionally, significant differences exist in both incidence and mortality of lung cancer between males and females, suggesting the possibility of inherent biological processes between the two sexes. These differences are among the most consistently reported significant risk factors in lung cancer (Visbal et al, Ann Thorac Surg, 78:209-15 (2004)).

[005] Several screening studies designed to detect lung cancer at an earlier stage have been undertaken. Many of these studies have used two major imaging technologies: chest X-ray and computed tomography (CT). A chest X-ray may reveal a mass in the lung that may be biopsied for verification and

classification of lung cancer. Improvements in imaging technologies, most notably the development of spiral computed tomography (spiral CT), have enabled clinicians to detect smaller lung tumors in pre-symptomatic individuals. The Early Lung Cancer Action project (ELCAP), designed to screen symptom-free high-risk smokers using chest X-ray and low dose CT, showed that CT greatly improved the detection of smaller, potentially curable lesions (Henschke et al., Lancet, 354:99-105 (1999)). However, the major disadvantages were the rate of false positives and high cost (average $600) that were difficult to justify as a screening paradigm. Id.; see also Henschke et al., N Engl J Med, 355:1763-71 (2006).

[006] These studies have generated an intense discussion within the scientific community centering on the design of the studies, particularly the lack of randomization, and the conclusions drawn with concerns of high cost and high false positive rate (Henschke et al., 1999; Gould et al., N Engl J Med, 356:743- 747 (2007)). Between 25 and 60 percent of spiral CT scans of smokers and former smokers show benign abnormalities (Swensen et al., Am J Respir Crit Care Med, 165:508-13 (2002)). Imaging screening of asymptomatic patients does not distinguish benign from early malignant lesions. It has also been reported that repeated exposure to low dose CT scans may expose patients to potentially harmful levels of radiation (Brenner et al., Radiology, 231 :440-5 (2004)).

[007] At present, there is still a need for clinically relevant markers for noninvasive detection of lung disease including cancer, monitoring response to therapy, or detecting lung cancer recurrence. It is also clear that such assays must be highly specific with reasonable sensitivity, and be readily available at a reasonable cost. Circulating biomarkers offer an alternative to imaging with the following advantages: 1) they are found in a minimally-invasive, easy to collect specimen type (blood or blood-derived fluids), 2) they can be monitored frequently over time in a subject to establish an accurate baseline, making it easy to detect changes over time, 3) they can be provided at a reasonably low cost, 4) they may limit the number of patients undergoing repeated expensive and potentially harmful CT scans, and/or 5) unlike CT scans, biomarkers may potentially distinguish indolent from more aggressive lung lesions (see, e.g., Greenberg and Lee, Opin PuIm Med, 13:249-55 (2007)).

[008] Existing biomarker assays include several serum protein markers such as CEA (Okada et al., Ann Thorac Surg, 78:216-21 (2004)), CYFRA 21-1 (Schneider, Adv Clin Chem, 42:1-41 (2006)), CRP (Siemes et al., J Clin Oncol, 24:5216-22 (2006)), CA-125 (Schneider, 2006), and neuron-specific enolase and squamous cell carcinoma antigen (Siemes et al., 2006). Low sensitivity and specificity, with a significant number of false positive results due to benign pulmonary diseases have limited the application of these assays.

[009] Circulating nucleic acids such as DNA and mRNA have also been evaluated as possible diagnostic markers for lung cancer. These studies are based on the observations that circulating nucleic acids show differentia! expression that is suggestive of cancer. (See, e.g., Bremnes et al., Lung Cancer, 49:1-12 (2005); Johnson et al., Cell, 120:635-47 (2005); Yanaihara et al., Cancer Cell, 9:189-98 (2006); Chen et al., Cell Res, 18:997-1006 (2008); Fabbri et al., Cancer J, 14:1-6 (2008); Garofalo et al., Oncogene, 27:3845-55 (2008); Mitchell et al., Proc Natl Acad Sci, 105:10513-8 (2008); Schickel et al., Oncogene,

27:5959-74 (2008); Weiss et al., Ann Oncol, 19:1053-9 (2008); and Yu et al., Cancer Cell, 13:48-57 (2008).) The origin of free DNA in circulation is not completely understood, but they are thought to represent the stable remaining fraction from damaged (apoptotic, necrotic) tumor cells (Jahr et al., Cancer Res, 61 :1659-65 (2001); Bianchi, Placenta, 25 Suppl A:S93-S101 (2004)).

[010] We herein describe methods for detecting, diagnosing, or monitoring lung disease by measuring miRNAs from serum or plasma.

[011] In some embodiments, the invention relates to the detection or monitoring of lung diseases such as small cell lung cancer or non-small cell lung cancer by detecting miRNAs from serum or plasma. The methods of the invention include detection of biomarkers that can be used to diagnose disease and/or evaluate the prognosis or aggressiveness of a lung disease. Further, the methods may be used to characterize the progression of a lung disease. In certain embodiments, the methods of the invention may be used to determine whether a lung tumor or lesion in a patient is cancerous or benign. The patients tested using the methods of the invention may also be tested using other known methods in the art.

[012] In certain embodiments of the invention, the diagnosis or prognosis may be achieved by measuring the amount of a miRNA that is present in elevated or reduced levels in the serum or plasma of a subject with lung disease. In some instances, one serum or plasma miRNA may be detected (e.g., amplified and measured) to characterize lung disease, while in other embodiments, two or more miRNAs are detected from serum or plasma. Some embodiments include detecting a pair of miRNAs. In some instances, one miRNA in the pair is elevated in serum or plasma of patients with lung disease or lung cancer, and the other miRNA in the pair is reduced. In other circumstances, both miRNAs in the pair can be elevated or both reduced. In certain embodiments, non-miRNA

biomarkers such as protein markers may also be measured. Some embodiments of the invention relate to diagnosis or prognosis of lung cancer, or determining the type of lung cancer in a patient. In some embodiments, the patient has previously been screened for lung disease.

[013] Additional embodiments of the invention are discussed throughout this application. Other objects, features, and advantages of the present invention will become apparent from the following detailed description. Any embodiment discussed with respect to one aspect of the invention applies to other aspects of the invention as well and vice versa. The embodiments in the Example section are understood to be embodiments of the invention that are applicable to all aspects of the invention.

[014] It should be understood, however, that the detailed description and the specific examples, while indicating specific embodiments of the invention, are given by way of illustration only, since various changes and modifications within the spirit and scope of the invention will become apparent to those skilled in the art from this application.

BRIEF DESCRIPTION OF THE DRAWINGS

[015] The following drawings form part of the present specification and are included to further demonstrate certain aspects of the present invention. The invention may be better understood by reference to one or more of these drawings in combination with the detailed description of specific embodiments presented herein.

[016] Figure 1 is a schematic outline of the study design for the mouse model for lung cancer. One hundred fifteen A/J mice were administered with an oral gavage of benzo(a)pyrene at the indicated dose at time (0) and at 10 weeks post the first gavage. Ten control mice received an oral gavage of 200 μl cottonseed oil with no carcinogen at week/day 0. The number of mice sacrificed at each time point is shown below the arrows. The remaining mice were bled every two weeks until 18 weeks and every four weeks thereafter. At the end of 34 weeks, all remaining mice were sacrificed.

[017] Figure 2A and Figure 2B show temporal changes of miR-183 and miR-182 (Figure 2A) and miR-365 and miR-141 (Figure 2B) plasma miRNA expression in mice that developed lung tumors and in control mice.

[018] Figure 3 shows the distribution of area under the curve (AUC) values for the Linear Discriminate Analysis (LDA) classifier as a function of the number of biomarkers, as described in Example 5. This procedure determines the optimal number of biomarkers to use for the classifier. Solid black circles represent results from the training set; Open triangles represent results from the test set. The vertical dashed line is where performance levels off. The maximum mean AUC is where the number of features = 9.

[019] Figure 4 shows performance estimates of training and test data from the Linear Discriminate Analysis of Example 5. SENS, sensitivity; SPEC, specificity, NPV, negative predictive value; PPV, positive predictive value. Six differential miRNA pairs were used: miR-142-5p and miR181d; miR-142-3p and miR181d; miR-142-3p and miR-422a; miR-142-5p and miR-422a; miR-92 and miR-27b; and miR-24 and miR-27a.

[020] Figure 5 shows human precursor miRNA (pre-miRNA) sequences (SEQ ID NOS 1-323, respectively, in order of appearance), as provided by

Release 13.0 of the miRBase::Sequences Database (http://microrna.sanger.ac.uk; Griffiths-Jones et al., Nucleic Acids Research, 2008, 36, Database Issue, D154- D158; Griffiths-Jones et al., Nucleic Acids Research, 2006, 34, Database Issue, D140-D144; Griffiths-Jones, Nucleic Acids Research, 2004, 32, Database Issue, D109-D111). The names of mature miRNAs from Tables 1-5 are also provided. EXEMPLARY EMBODIMENTS

[021] In certain aspects, the methods of the invention provide assays for amplifying and measuring the amount of a miRNA in a serum or plasma sample, thereby characterizing a lung disease.

[022] To assist in understanding the present invention, certain terms are first defined. Additional definitions are provided throughout the application.

[023] As used herein, the term "microRNA" (miRNA or miR) includes human miRNAs, mature single stranded miRNAs, precursor miRNAs (pre-miR), and variants thereof, which may be naturally occurring. In some instances, the term "miRNA" also includes primary miRNA transcripts and duplex miRNAs.

Unless otherwise noted, when used herein, the name of a specific miRNA refers to the mature miRNA of a precursor miRNA. For example, miR-122a refers to a mature miRNA sequence derived from pre-miR-122. The sequences for particular miRNAs, including human mature and precursor sequences, are reported in the miRBase::Sequences Database (http://microrna.sanger.ac.uk (version 15 released April 2010); Griffiths-Jones et al., Nucleic Acids Research, 2008, 36, Database Issue, D154-D158; Griffiths-Jones et al., Nucleic Acids Research, 2006, 34, Database Issue, D140-D144; Griffiths-Jones, Nucleic Acids Research, 2004, 32, Database Issue, D109-D111). For certain miRNAs, a single precursor contains more than one mature miRNA sequence. In other instances, multiple precursor miRNAs contain the same mature sequence. In some instances, mature miRNAs have been re-named based on new scientific consensus. For example, miR-213, as used herein, refers to a mature miRNA from pre-miR-181a- 1 , and is also called miR-181a^*. Other miRNAs that have been re-named include miR-189 (also called miR-24*), which comes from pre-miR-24-1 ; miR-368 (also called miR-376c); and miR-422b (also called miR-378^*). The skilled artisan will appreciate that scientific consensus regarding the precise nucleic acid sequence for a given miRNA, in particular for mature forms of the miRNAs, may change with time. MiRNAs detected by assays of this application include naturally occurring sequences for the miRNAs.

[024] The term "characterizing" is used herein to encompass detection and prognosis, and it includes detection of a miRNA for making diagnostic or prognostic determinations or predictions of disease. In some instances, the characterization will identify whether a subject has a lung disease such as cancer, or will determine the disease state. Additionally, detection of a miRNA according to the methods herein includes measuring the amount of a miRNA that can used to distinguish patients with lung cancer from patients having other lung diseases, or determine whether a patient with a lung tumor has cancer. In other

circumstances, "characterizing" includes detection of a miRNA for determining the stage or aggressiveness of a disease state such as lung cancer, or determining an appropriate treatment method for lung disease.

[025] The use of the word "a", "an" or "the" when used in conjunction with the term "comprising" in the claims and/or the specification may mean "one," but it is also consistent with the meaning of "one or more," "at least one," and "one or more than one."

I. SAMPLES

[026] Serum is typically the fluid, non-cellular portion of coagulated blood. Plasma is also a non-cellular blood sample, but unlike serum, plasma contains clotting factors. In some embodiments, serum or plasma samples are obtained from a human patient previously screened for lung disease using another diagnostic method. In various embodiments, serum or plasma samples are obtained from patients that have tested positive for a tumor or lung lesion. In certain embodiments, the patient has undergone imaging detection, e.g., by chest X-ray or CT scan. In other embodiments, the methods involve detection of miRNA in patients with a positive imaging result for lung disease. In some circumstances, samples are obtained from patients that have a lung tumor or lesion. The tumor or lesion may have been detected by chest X-ray or CT scan, or by other imaging or detection methods known in the art. Additional embodiments include

measuring miRNA in samples from patients previously or currently undergoing treatment for a lung disease. In additional embodiments, the sample is from a patient suspected of having lung cancer or at risk of developing lung cancer. The volume of plasma or serum obtained and used for the assay may be varied depending upon clinical intent.

[027] One of skill in the art will recognize that many methods exist for obtaining and preparing serum samples. Generally, blood is drawn into a collection tube using standard methods and allowed to clot. The serum is then separated from the cellular portion of the coagulated blood. In some methods, clotting activators such as silica particles are added to the blood collection tube. In other methods, the blood is not treated to facilitate clotting. Blood collection tubes are commercially available from many sources and in a variety of formats (e.g., Becton Dickenson Vacutainer^® tubes-SST™, glass serum tubes, or plastic serum tubes).

[028] In some methods, the blood is collected by venipuncture and processed within three hours after drawing to minimize hemolysis and minimize the release of miRNAs from intact cells in the blood. In some methods, blood is kept on ice until use. The blood may be fractionated by centrifugation to remove cellular components. In some embodiments, centrifugation to prepare serum can be at a speed of at least 500, 1000, 2000, 3000, 4000, or 5000 X G. In certain embodiments, the blood can be incubated for at least 10, 20, 30, 40, 50, 60, 90, 120, or 150 minutes to allow clotting. In other embodiments, the blood is incubated for at most 3 hours. When using plasma, the blood is not permitted to coagulate prior to separation of the cellular and acellular components. Serum or plasma can be frozen after separation from the cellular portion of blood until further assayed.

[029] Before analysis, RNA may be extracted from serum or plasma and purified using methods known in the art. Many methods are known for isolating total RNA, or for specifically extracting small RNAs, including miRNAs. The RNA may be extracted using commercially-available kits (e.g., Perfect RNA Total RNA Isolation Kit, Five Prime-Three Prime, Inc.; mirVana™ kits, Ambion, Inc.).

Alternatively, RNA extraction methods for the extraction of mammalian

intracellular RNA or viral RNA may be adapted, either as published or with modification, for extraction of RNA from plasma and serum. RNA may be extracted from plasma or serum using silica particles, glass beads, or diatoms, as in the method or adaptations described in U.S. Publication No. 2008/0057502. II. miRNA MARKERS FOR LUNG DISEASE

[030] Certain embodiments of the invention provide serum or plasma miRNAs as markers for lung disease. In some embodiments, miRNAs that are present at elevated levels in the serum and/or plasma of patients with lung disease are used as markers. In other embodiments, miRNAs that have reduced levels are used as markers. In some embodiments, more than one miRNA from serum or plasma will be used as markers. When more than one miRNA biomarker is used, the miRNAs may all have elevated levels, all have reduced levels, or a mixture of miRNAs with elevated and reduced levels may be used.

[031] The terms "reduced levels" or "elevated levels" refer to the amount of a miRNA in a serum or plasma sample from a patient compared to the amount of the miRNA in serum or plasma from a cohort or cohorts that do not have the lung disease that the patient is being tested for. For instance, a miRNA that has reduced levels in the sera of lung cancer patients is present at lower amounts in lung cancer patient sera than in serum from a donor who does not have lung cancer (e.g., patients with benign tumors or normal patients). For certain miRNAs, elevated levels in a patient serum or plasma sample indicates presence or prognosis for a lung disease. Other miRNAs are present in reduced levels in patients with lung disease.

[032] Lung disease includes cancer and benign conditions. Lung cancer refers to malignant tumors of the lung, and can be classified as small ceil or non- small cell lung cancer. In some embodiments, non-small cell lung cancer can be further characterized as adenocarcinoma, squamous cell carcinoma (SCC), and large cell carcinoma. In addition, cancers can be classified based on X-ray or CT scanning results, aggressiveness, pathology, and measurements of non-miRNA biomarkers, as well as other methods known in the art. In certain aspects, the lung cancer is classified by TNM principles (T-primary tumor, N-regional lymph nodes, M-distant metastasis) and/or stage 0, IA, IB, HA, HB, HIA, IHB or IV. (See, e.g., Lababede et al., Chest, 115:233-235 (1999).) In some embodiments, the methods described herein can be used to characterize a lung disease in a patient with at least 75, 80, 85, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% sensitivity. The degree of sensitivity indicates the percentage of patients with a disease who are positively characterized as having the disease. In additional embodiments, the methods have at least 75, 80, 85, 90, 91 , 92, 93, 94, 95, 96, 97, 98, or 99% specificity (e.g., the percentage of non-diseased patients who are correctly characterized). The assay parameters can be adjusted to optimize for both sensitivity and specificity.

[033] In some cases, the level of the miRNA marker will be compared to a control to determine whether the level is reduced or elevated. The control may be an external control, such as a miRNA in a serum or plasma sample from a subject known to be free of lung disease. The external control may be a sample from a normal (non-diseased) subject or from a patient with benign lung disease. In other circumstances, the external control may be a miRNA from a non-serum sample like a tissue sample or a known amount of a synthetic RNA. The external control may be a pooled, average, or individual sample; it may be the same or different miRNA as one being measured. An internal control is a marker from the same serum or plasma sample being tested, such as a miRNA control. See, e.g., US Publication No. US 2009/0075258, which is incorporated by reference in its entirety.

[034] Table 1 lists miRNAs that have elevated or reduced levels in serum from patients with lung disease. These miRNAs may be used in accordance with the invention. Some of the miRNAs are useful for characterizing lung cancer, including distinguishing the type of cancer and/or distinguishing cancer from benign lung disease. In addition, some miRNAs may be used to predict the aggressiveness or outcome of lung cancer.

Table 1. miRNAs with elevated or reduced levels in serum from patients with lung cancer. Levels are of miRNA in lung cancer patients compared to patients with benign tumors or lesions.

[035] In certain embodiments, one serum miRNA is used to detect, diagnose, characterize, or monitor lung disease and/or lung cancer. In other embodiments, more than one miRNA is used as a marker. In additional embodiments, two or more miRNAs are used to characterize lung disease. In certain embodiments, at least 2, 3, 4, 5, 6, 7, 8, 9, or 10 miRNAs are detected in the methods of the invention. In certain methods, miRNAs that have reduced levels in serum or plasma from patients with lung disease are used as biomarkers. In other embodiments, a miRNA with elevated levels in serum or plasma can be used as a biomarker. In certain embodiments, the patient has a lung tumor or lesion. In additional embodiments, the patient has previously been screened for lung disease.

[036] In some embodiments, a miRNA for diagnosing lung cancer is chosen from miR-375, miR-499, miR-22, miR-122a, miR-206, miR-103, miR-24, miR-26a, miR-498, miR-205, miR-222, and let-7c.

[037] In certain embodiments, a miRNA is chosen from let-7a, let-7b, let- 7c, let-7d, let-7e, let-7f, let-7g, miR-106a, miR-106b, miR-125a, miR-126, miR- 130b, miR-132, miR-133a, miR-133b, miR-140, miR-142-3p, miR-143, miR-146a, miR-150, miR-151 , miR-155, miR-15a, miR-15b, miR-16, miR-181a, miR-181 b, miR-181d, miR-186, miR-18a, miR-190, miR-191 , miR-195, miR-197, miR-19b, miR-202, miR-206, miR-20a, miR-210, miR-214, miR-22, miR-221 , miR-222, miR- 223, miR-23a, miR-24, miR-25, miR-26a, miR-26b, miR-27b, miR-30a-5p, miR- 30b, miR-30c, miR-30d, miR-30e-3p, miR-320, miR-324-3p, miR-324-5p, miR- 335, miR-342, miR-345, miR-346, miR-361 , miR-365, miR-374, miR-378, miR- 382, miR-422a, miR-432, miR-485-3p, miR-486, miR-496, miR-502, miR-584, miR-638, miR-660, miR-92, and miR-93.

[038] In additional embodiments, a miRNA is chosen from let-7f, let-7g, let-7i, miR-106b, miR-126, miR-126^*, miR-140, miR-142-3p, miR-142-5p, miR- 143, miR-145, miR-150, miR-15a, miR-15b, miR-181a, miR-181b, miR-181d, miR- 202, miR-214, miR-27a, miR-27b, miR-30e-5p, miR-320, miR-324-3p, miR-340, miR-342, miR-345, miR-374, miR-378, miR-422a, miR-486, miR-518b, and miR- 92.

[039] In other embodiments, at least one miRNA is chosen from let-7b, let- 7c, let-7d, let-7e, miR-10a, miR-10b, miR-130b, miR-132, miR-133b, miR-139, miR-143, miR-152, miR-155, miR-15b, miR-17-5p, miR-193, miR-194, miR-195, miR-196b, miR-199a^*, miR-19b, miR-202, miR-204, miR-205, miR-206, miR-20b, miR-21 , miR-210, miR-214, miR-221 , miR-27a, miR-27b, miR-296, miR-29a, miR- 301 , miR-324-3p, miR-324-5p, miR-339, miR-346, miR-365, miR-378, miR-422a, miR-432, miR-485-3p, miR-496, miR-497, miR-505, miR-518b, miR-525, miR-566, miR-605, miR-638, miR-660, and miR-93. [040] In additional embodiments, a miRNA is chosen from miR-106a, miR- 106b, miR-126^*, miR-142-3p, miR-15b, miR-181c, miR-182, miR-26b, miR-30b, miR-30e-5p, miR-422b, let-7i, and let-7g.

[041] In further embodiments, a miRNA is chosen from miR-24, miR-92, miR-142-3p, miR-142-5p, miR-181d, miR-27a, miR-27b, miR-422a, miR-29b, miR-15a, miR-106b, miR-126, miR-140, and miR-202. In some embodiments, 2, 3, 4, 5, 6, 7, or 8 of these miRs can be used to distinguish patients with lung cancer from patients with benign lung tumors or lesions. In additional

embodiments, 2, 3, 4, 5, 6, 7, or 8 miRNAs are chosen from miR-24, miR-92, miR- 142-3p, miR-142-5p, miR-181d, miR-27a, miR-27b, and miR-422a.

[042] In certain embodiments, a miRNA for characterizing lung cancer vs. non-cancer samples is chosen from miR-15b, miR-182, miR-15a, miR-30b, miR- 26b, miR-106b, let-7g, miR-142-3p, miR-301 , miR-181c, miR-126, miR-346, miR- 422b, and miR-92. Non-cancer samples include samples from subjects with benign lung tumors or lesions, or from normal subjects.

[043] Certain embodiments include a method for characterizing lung disease and/or lung cancer in a patient comprising the steps of measuring the level of a miRNA in a serum sample, wherein the miRNA is chosen from let-7b, let-7c, let-7d, let-7e, miR-10a, miR-10b, miR-130b, miR-132, miR-133b, miR-139, miR-143, miR-152, miR-155, miR-15b, miR-17-5p, miR-193, miR-194, miR-195, miR-196b, miR-199a^*, miR-19b, miR-202, miR-204, miR-205, miR-206, miR-20b, miR-21 , miR-210, miR-214, miR-221 , miR-27a, miR-27b, miR-296, miR-29a, miR- 301 , miR-324-3p, miR-324-5p, miR-339, miR-346, miR-365, miR-378, miR-422a, miR-432, miR-485-3p, miR-496, miR-497, miR-505, miR-518b, miR-525, miR-566, miR-605, miR-638, miR-660, and miR-93; and determining reduced or elevated levels of the miRNA in the sample, thereby characterizing lung disease or lung cancer.

[044] Table 2 lists miRNAs that have elevated or reduced levels in plasma from patients with lung disease. These miRNAs may be used in accordance with the invention. Table 2. miRNAs with elevated or reduced levels in plasma from patients with lung disease. Levels are of miRNA in lung cancer patients compared to patients without lung cancer.

[045] In some embodiments, a single plasma miRNA may be used to characterize lung cancer. In other embodiments, one of the miRNAs from Table 2 may be used to characterize lung cancer, either alone or in combination with one or more additional miRNA markers.

[046] In certain embodiments, the methods distinguish lung cancer from benign lung disease. In some circumstances, at least one measured miRNA is elevated in the serum or plasma of lung cancer patients compared to patients with benign conditions or no disease. In some circumstances, at least one measured miRNA is reduced. In certain embodiments at least one measured miRNA is elevated and at least one miRNA is reduced. In other embodiments at least two elevated miRNAs or at least two reduced miRNAs are measured. [047] In certain embodiments, one of the following miRNAs is used in combination with at least one other miRNA biomarker to determine whether a patient has lung cancer: let-7a, let-7b, let-7d, Iet-7f, let-7g, let-7i, miR-101 , miR- 106a, miR-106b, miR-125a, miR-126, miR-126*, miR-130b, miR-132, miR-133b, miR-140, miR-142-3p, miR-142-5p, miR-145, miR-146a, miR-146b, miR-148b, miR-150, miR-151 , miR-15a, mιR-15b, miR-181a, miR-181b, miR-181d, miR-185, miR-186, miR-190, miR-191 , miR-193a, miR-199a*, miR-202, miR-210, miR-214, miR-222, miR-23a, miR-24, miR-26a, miR-26b, miR-27a, miR-27b, miR-29b, miR- 301 , miR-30a-5p, miR-30b, miR-30c, miR-30d, miR-30e-5p, miR-320, miR-324- 3p, miR-326, miR-335, miR-340, miR-342, miR-345, miR-346, miR-34a, miR-374, miR-375, miR-378, miR-422a, miR-422b, miR-425, miR-486, miR-496, miR-518b, miR-660, miR-7, miR-92, miR-93, miR-98, miR-99a, or miR-99b.

[048] In other embodiments, one of the following miRNAs is used in combination with at least one other miRNA biomarker to determine whether a patient has lung cancer or to distinguish iung cancer from benign lung disease: miR-422a; miR-29b; miR-92; miR-142-5p; miR-142-3p; miR-181d; miR-27b; miR- 378; miR-27a; miR-30e-5p; miR-181a; miR-126; miR-342; miR-140; miR-15a; miR-324-3p; miR-374; miR-486; miR-518b; miR-106b; miR-145; miR-150; miR- 191 ; miR-345; miR-126*; miR-148b; miR-214; miR-320; let-7g; let-7i; miR-146a; miR-15b; miR-185; miR-186; miR-23a; miR-24; miR-30a-5p; miR-340; miR-34a; miR-101 ; miR-132; miR-181b; miR-199a*; miR-202; miR-222; miR-422b; miR- 660; miR-7; or miR-93.

[049] In other embodiments, one of the following miRNAs is used in combination with at least one other miRNA biomarker to determine whether a patient has lung cancer: let-7g, miR-106b, miR-126, miR-126^*, miR-132, miR- 140, miR-142-3p, miR-146a, miR-150, miR-15a, miR-15b, miR-181a, miR-181b, miR-181d, miR-214, miR-24, miR-30a-5p, miR-320, miR-342, miR-345, miR-374, miR-422a, miR-422b, miR-486, or miR-92.

[050] Some embodiments of the invention relate to amplifying and measuring at least a pair of miRNAs from serum. Table 3 includes pairs that may be used to characterize lung disease. These pairs may be used in combination with other lung disease biomarkers. Table 3. miRNA pairs measured in serum samples.

[051] In certain embodiments, the pair of miRNAs is chosen from miR-202 and miR-29b; miR-142-5p and miR-422a; miR-24 and miR-27a; miR-27b and miR-422a; miR-140 and miR-422a; miR-185 and miR-93; miR-126 and miR-181d; miR-142-3p and miR-422a; miR-30e-5p and miR-345; miR-29b and miR-422a; miR-324-5p and miR-422a; miR-374 and miR-422a; miR-140 and miR-345; miR- 23a and miR-27a; miR-29b and miR-378; miR-30e-5p and miR-422a; miR-30e-5p and miR-92; miR-660 and miR-92; miR-106b and miR-422a; let-7g and miR-422a; miR-101 and miR-92; miR-126 and miR-92; miR-126^* and miR-422a; miR-126^* and miR-92; miR-150 and miR-29a; miR-15a and miR-92; miR-30e-5p and miR- 324-3p; miR-126 and miR-378; miR-126 and miR-422a; miR-132 and miR-29b; miR-142-3p and miR-181a; miR-142-3p and miR-378; miR-148b and miR-92; miR-181b and miR-29b; miR-26a and miR-422a; miR-15a and miR-422a; miR- 146a and miR-27b; miR-148b and miR-422a; let-7i and miR-422a; miR-340 and miR-422a; miR-143 and miR-150; miR-15a and miR-378; miR-15b and miR-422a; miR-27b and miR-92; miR-30e-3p and miR-422a; miR-30b and miR-422a; miR- 142-3p and miR-92; miR-27a and miR-422a; miR-142-5p and miR-342; miR-142- 5p and miR-223; miR-27b and miR-378; miR-186 and miR-27a; miR-106b and miR-324-3p; miR-34a and miR-518b; miR-27a and miR-92; miR-30a-3p and miR- 422a; miR-181d and miR-29b; miR-15b and miR-191 ; miR-192 and miR-422a; miR-422a and miR-576; miR-30c and miR-422a; miR-142-3p and miR-145; miR- 142-3p and miR-181d; miR-181d and miR-27b; miR-142-5p and miR-186; miR- 150 and miR-29c; miR-200c and miR-422a; miR-185 and miR-92; miR-148a and miR-422a; miR-34a and miR-422a; miR-32 and miR-422a; miR-214 and miR-566; miR-206 and miR-422a; miR-214 and miR-518b; miR-142-3p and miR-320; miR- 331 and miR-422a; miR-26b and miR-422a; miR-142-5p and miR-345; miR-27b and miR-324-3p; miR-30a-5p and miR-30e-5p; miR-29b and miR-92; miR-191 and miR-27a; miR-140 and miR-181d; miR-422a and miR-660; miR-126 and miR- 181a; miR-126^* and miR-378; miR-15a and miR-181d; let-7f and miR-422a; miR- 181a and miR-27b; miR-29b and miR-324-3p; miR-132 and miR-30e-5p; miR-214 and miR-422b; miR-140 and miR-378; miR-29b and miR-30a-5p; let-7a and miR- 422a; miR-142-5p and miR-191 ; miR-24 and miR-27b; miR-126* and miR-181d; miR-181d and miR-30e-5p; miR-142-3p and miR-191 ; miR-181c and miR-29b; miR-181d and miR-27a; miR-148b and miR-378; miR-15a and miR-320; miR- 199a^* and miR-422a; miR-146a and miR-27a; miR-142-5p and miR-181a; miR- 15a and miR-486; miR-27b and miR-342; miR-181a and miR-27a; let-7g and miR- 342; miR-29b and miR-345; let-7e and miR-422a; miR-29b and miR-422b; miR- 142-3p and miR-202; miR-214 and miR-422a; miR-142-5p and miR-145; miR- 422a and miR-497; miR-29b and miR-433; miR-140 and miR-92; miR-142-5p and miR-92; miR-142-3p and miR-181b; miR-30e-5p and miR-486; miR-196b and miR-422a; miR-222 and miR-29b; miR-142-3p and miR-518b; miR-29b and miR- 3Od; miR-29b and miR-361 ; miR-342 and miR-34a; miR-106b and miR-181d; miR-340 and miR-92; miR-486 and miR-7; miR-21 and miR-422a; miR-106b and miR-92; let-7i and miR-92; miR-101 and miR-486; miR-181a and miR-29b; miR- 142-5p and miR-150; miR-142-3p and miR-342; miR-29b and miR-518b; miR- 142-5p and miR-181d; miR-193b and miR-29b; miR-146b and miR-422a; miR-15b and miR-92; miR-142-5p and miR-23a; miR-27a and miR-378; miR-197 and miR- 29b; miR-342 and miR-374; miR-422a and miR-487b; miR-142-3p and miR-99b; miR-29b and miR-320; miR-340 and miR-378; miR-29b and miR-486; miR-23a and miR-422a; miR-222 and miR-27b; miR-142-5p and miR-24; miR-148b and miR-181d; miR-29b and miR-342; miR-152 and miR-422a; miR-7 and miR-92; miR-210 and miR-422a; let-7c and miR-422a; miR-27b and miR-518b; miR-145 and miR-374; miR-27a and miR-324-3p; miR-140 and miR-186; miR-185 and miR-486; miR-422a and miR-496; miR-181a and miR-199a^*; miR-422a and miR- 432; miR-181d and miR-23b; miR-125b and miR-422a; miR-145 and miR-29b; miR-142-5p and miR-146a; miR-126 and miR-320; miR-191 and miR-374; miR- 126 and miR-145; miR-150 and miR-29b; miR-133b and miR-422a; miR-142-5p and miR-324-3p; miR-27b and miR-345; miR-374 and miR-378; let-7b and miR- 422a; miR-346 and miR-432*; miR-142-5p and miR-518b; miR-142-3p and miR- 328; miR-181a and miR-374; let-7g and miR-181d; miR-106b and miR-378; miR- 30a-5p and miR-422a; miR-339 and miR-422a; let-7i and miR-378; miR-142-5p and miR-93; and miR-224 and miR-422a. [052] In certain embodiments, the pair of miRNAs is chosen from let-7a and miR-181a; let-7a and miR-181d; let-7b and miR-150; let-7b and miR-181d; let-7b and miR-92; let-7c and miR-150; let-7c and miR-181d; let-7c and miR-92; let-7e and miR-378; let-7f and miR-181a; let-7f and miR-181d; let-7f and miR-342; let-7f and miR-92; let-7g and miR-150; let-7g and miR-181a; let-7g and miR-181d; let-7g and miR-342; let-7g and miR-92; let-7i and miR-486; let-7i and miR-92; miR-106b and miR-150; miR-106b and miR-181d; miR-106b and miR-92; miR- 125a and miR-142-3p; miR-125a and miR-374; miR-126^* and miR-181d; miR- 126^* and miR-30a-5p; miR-126^* and miR-92; miR-126 and miR-146a; miR-126 and miR-150; miR-126 and miR-181a; miR-126 and miR-181d; miR-126 and miR- 342; miR-126 and miR-92; miR-130b and miR-142-3p; miR-132 and miR-142-3p; miR-132 and miR-214; miR-140 and miR-150; miR-140 and miR-30a-5p; miR-140 and miR-345; miR-140 and miR-92; miR-142-3p and miR-146a; miR-142-3p and miR-150; miR-142-3p and miR-151 ; miR-142-3p and miR-181a; miR-142-3p and miR-181b; miR-142-3p and miR-181d; miR-142-3p and miR-186; miR-142-3p and miR-210; miR-142-3p and miR-22; miR-142-3p and miR-23a; miR-142-3p and miR-24; miR-142-3p and miR-30a-5p; miR-142-3p and miR-342; miR-142-3p and miR-345; miR-142-3p and miR-425; miR-142-3p and miR-486; miR-142-3p and miR-92; miR-142-3p and miR-99b; miR-142-5p and miR-30a-5p; miR-143 and miR-223; miR-143 and miR-486; miR-150 and miR-15b; miR-150 and miR-214; miR-150 and miR-29b; miR-150 and miR-374; miR-150 and miR-576; miR-15a and miR-181a; miR-15a and miR-181b; miR-15a and miR-181d; miR-15a and miR-210; miR-15a and miR-30a-5p; miR-15a and miR-342; miR-15a and miR- 345; miR-15a and miR-486; miR-15a and miR-92; miR-15b and miR-17-5p; miR- 15b and miR-181a; miR-15b and miR-181d; miR-15b and miR-24; miR-15b and miR-342; miR-15b and miR-92; miR-16 and miR-486; miR-16 and miR-92; miR- 181a and miR-214; miR-181a and miR-26a; miR-181a and miR-26b; miR-181a and miR-30b; miR-181a and miR-30c; miR-181a and miR-374; miR-181a and miR-98; miR-181b and miR-214; miR-181b and miR-374; miR-181d and miR-214; miR-181d and miR-26b; miR-181d and miR-30b; miR-181d and miR-30c; miR- 181d and miR-374; miR-181d and miR-432; miR-181d and miR-496; miR-181d and miR-638; miR-181d and miR-98; miR-193a and miR-422a; miR-195 and miR- 486; miR-199a^* and miR-92; miR-20a and miR-92; miR-214 and miR-342; miR- 214 and miR-422b; miR-214 and miR-92; miR-24 and miR-374; miR-26a and miR-342; miR-26a and miR-92; miR-26b and miR-342; miR-26b and miR-92; miR- 27a and miR-30a-5p; miR-27b and miR-30d; miR-29b and miR-30a-5p; miR-30b and miR-342; miR-30b and miR-92; miR-30c and miR-342; miR-30c and miR-92; miR-320 and miR-98; miR-342 and miR-374; miR-346 and miR-422a; miR-346 and miR-518b; miR-346 and miR-566; miR-374 and miR-92; miR-422a and miR- 496; miR-422a and miR-638; miR-422a and miR-98; miR-432 and miR-92; miR- 496 and miR-92; and miR-7 and miR-92.

[053] In other embodiments, the pair of miRNAs is chosen from miR-142- 5p and miR181d; miR-142-3p and miR181d; miR-142-3p and miR-422a; miR-142- 5p and miR-422a; miR-92 and miR-27b; and miR-24 and miR-27a. In additional embodiments, the pair of miRNAs is chosen from miR-106a and miR-422b; miR- 126* and miR-26b; miR-106a and miR-26b; miR-15b and miR-30a-5p; let-7a and miR-26b; let-7g and miR-106a; miR-126^* and miR-15b; miR-126^* and miR-30b; and miR-106a and miR-106b.

[054] In some embodiments, the pair is measured in a serum sample. Optionally, one or more additional miRNAs are measured.

[055] In additional embodiments, certain miRNA pairs may be used to characterize lung disease in female or male patients (Table 4). In certain embodiments, the methods detect sex-specific miRNA biomarkers.

Table 4: miRNA pairs for characterizing lung disease in female or male patients

[056] Some embodiments of the invention relate to amplifying and measuring two or more miRNAs from plasma. One of the following miRNA plasma biomarkers may be used in combination with at least one other miRNA: miR-10b, miR-192, miR-206, miR-101 , miR-205, miR-16, miR-151 , miR-137, miR- 215, miR-181a, miR-218, miR-126^*, miR-125b, miR-326, miR-100, miR-31 , miR- 197, miR-222, miR-191 , miR-200c, miR-186, miR-145, miR-155, miR-29c, let-7c, miR-181c, miR-125a, miR-134, miR-181d, let-7b, miR-127, miR-146a, miR-139, miR-152, miR-190, miR-30e-5p, miR-106b, miR-10a, miR-132, miR-148a, miR- 213, miR-29a, miR-375, miR-133b, miR-15a, miR-107, miR-148b, miR-19a, miR- 106a, miR-130a, miR-17-3p, miR-18a^*, miR-195, miR-20b, miR-301 , miR-339, miR-410, miR-188, miR-193a, let-7g, let-7i, miR-140, miR-181b, miR-25, miR-328, miR-133a, miR-150, miR-17-5p, miR-21 , miR-214, miR-370, miR-383, miR-130b, miR-199a, miR-212, miR-221 , miR-27b, miR-30e-3p, miR-338, miR-361 , miR-141 , miR-142-5p, miR-30a-3p, miR-30a-5p, miR-451 , miR-142-3p, miR-146b, miR-15b, miR-18a, miR-210, miR-296, miR-323, miR-362, let-7a, miR-196b, miR-223, miR- 29b, miR-324-5p, miR-376a, miR-379, miR-491 , let-7d, miR-126, miR-182, miR- 185, miR-204, miR-23a, miR-27a, miR-324-3p, miR-342, miR-34c, miR-382, miR- 425, miR-432^*, miR-103, miR-193b, miR-196a, miR-199a^*, miR-199b, miR-28, miR-30d, miR-330, miR-423, miR-433, miR-485-5p, miR-20a, miR-23b, miR-26a, miR-30b, miR-30c, miR-320, miR-345, miR-422b, miR-335, miR-365, miR-486, miR-24, miR-26b, miR-331 , miR-340, miR-34a, miR-374, miR-452, miR-483, miR- 512-5p, let-7e, miR-32, miR-422a, miR-424, miR-432, miR-485-3p, miR-487b, miR-496, miR-505, miR-7, miR-202, miR-369-3p, miR-495, miR-502, miR-511 , miR-516-3p, miR-517c, miR-92, miR-93, miR-99a, or miR-99b. [057] Table 5 includes pairs of miRNAs that may be used to characterize lung cancer from plasma samples. Optionally, one or more additional miRNAs are measured.

Table 5. miRNA pairs measured in plasma samples

[058] Other miRNAs and groups of miRNAs that can be used in the methods of the invention will be apparent from the Examples described herein. III. METHODS TO MEASURE THE LEVEL OF A MIRNA

[059] Many methods of measuring the levels or amounts of miRNAs are contemplated. Any reliable, sensitive, and specific method can be used. In some embodiments, a miRNA is amplified prior to measurement. In other embodiments, the level of miRNA is measured during the amplification process. In still other methods, the miRNA is not amplified prior to measurement.

A. Amplification Reactions

[060] Many methods exist for amplifying miRNA nucleic acid sequences such as mature miRNAs, precursor miRNAs, and primary miRNAs. Suitable nucleic acid polymerization and amplification techniques include reverse

transcription (RT), polymerase chain reaction (PCR), real-time PCR (quantitative PCR (q-PCR)), nucleic acid sequence-base amplification (NASBA), ligase chain reaction, multiplex Iigatable probe amplification, invader technology (Third Wave), rolling circle amplification, in vitro transcription (IVT), strand displacement amplification, transcription-mediated amplification (TMA), RNA (Eberwine) amplification, and other methods that are known to persons skilled in the art. In certain embodiments, more than one amplification method is used, such as reverse transcription followed by real time quantitative PCR (qRT-PCR) (Chen et al., Nucleic Acids Research, 33(20):e179 (2005)).

[061] A typical PCR reaction includes multiple amplification steps, or cycles that selectively amplify target nucleic acid species: a denaturing step in which a target nucleic acid is denatured; an annealing step in which a set of PCR primers (forward and reverse primers) anneal to complementary DNA strands; and an elongation step in which a thermostable DNA polymerase elongates the primers. By repeating these steps multiple times, a DNA fragment is amplified to produce an amplicon, corresponding to the target DNA sequence. Typical PCR reactions include 20 or more cycles of denaturation, annealing, and elongation. In many cases, the annealing and elongation steps can be performed concurrently, in which case the cycle contains only two steps. Since mature miRNAs are single-stranded, a reverse transcription reaction (which produces a

complementary cDNA sequence) may be performed prior to PCR reactions. Reverse transcription reactions include the use of, e.g., a RNA-based DNA polymerase (reverse transcriptase) and a primer.

[062] In PCR and q-PCR methods, for example, a set of primers is used for each target sequence. In certain embodiments, the lengths of the primers depends on many factors, including, but not limited to, the desired hybridization temperature between the primers, the target nucleic acid sequence, and the complexity of the different target nucleic acid sequences to be amplified. In certain embodiments, a primer is about 15 to about 35 nucleotides in length. In other embodiments, a primer is equal to or fewer than 15, 20, 25, 30, or 35 nucleotides in length. In additional embodiments, a primer is at least 35 nucleotides in length.

[063] In a further aspect, a forward primer can comprise at least one sequence that anneals to a miRNA biomarker and alternatively can comprise an additional 5' non-complementary region. In another aspect, a reverse primer can be designed to anneal to the complement of a reverse transcribed miRNA. The reverse primer may be independent of the miRNA biomarker sequence, and multiple miRNA biomarkers may be amplified using the same reverse primer. Alternatively, a reverse primer may be specific for a miRNA biomarker.

[064] In some embodiments, two or more miRNAs are amplified in a single reaction volume. One aspect includes multiplex q-PCR, such as qRT-PCR, which enables simultaneous amplification and quantification of at least two miRNAs of interest in one reaction volume by using more than one pair of primers and/or more than one probe. The primer pairs comprise at least one amplification primer that uniquely binds each miRNA, and the probes are labeled such that they are distinguishable from one another, thus allowing simultaneous quantification of multiple miRNAs. Multiplex qRT-PCR has research and diagnostic uses, including but not limited to detection of miRNAs for diagnostic, prognostic, and therapeutic applications.

[065] The qRT-PCR reaction may further be combined with the reverse transcription reaction by including both a reverse transcriptase and a DNA-based thermostable DNA polymerase. When two polymerases are used, a "hot start" approach may be used to maximize assay performance (U.S. Patent Nos.

5,411 ,876 and 5,985,619). For example, the components for a reverse

transcriptase reaction and a PCR reaction may be sequestered using one or more thermoactivation methods or chemical alteration to improve polymerization efficiency (U.S. Patent Nos. 5,550,044, 5,413,924, and 6,403,341).

B. Detection of mi RNAs

[066] In certain embodiments, labels, dyes, or labeled probes and/or primers are used to detect amplified or unamplified miRNAs. The skilled artisan will recognize which detection methods are appropriate based on the sensitivity of the detection method and the abundance of the target. Depending on the sensitivity of the detection method and the abundance of the target, amplification may or may not be required prior to detection. One skilled in the art will recognize the detection methods where miRNA amplification is preferred.

[067] A probe or primer may include Watson-Crick bases or modified bases. Modified bases include, but are not limited to, the AEGIS bases (from Eragen Biosciences), which have been described, e.g., in U.S. Patent Nos.

5,432,272, 5,965,364, and 6,001 ,983. In certain aspects, bases are joined by a natural phosphodiester bond or a different chemical linkage. Different chemical linkages include, but are not limited to, a peptide bond or a Locked Nucleic Acid (LNA) linkage, which is described, e.g., in U.S. Patent No. 7,060,809.

[068] In a further aspect, oligonucleotide probes or primers present in an amplification reaction are suitable for monitoring the amount of amplification product produced as a function of time. In certain aspects, probes having different single stranded versus double stranded character are used to detect the nucleic acid. Probes include, but are not limited to, the 5'-exonuclease assay (e.g., TaqMan™) probes (see U.S. Patent No. 5,538,848), stem-loop molecular beacons (see, e.g., U.S. Patent Nos. 6,103,476 and 5,925,517), stemless or linear beacons (see, e.g., WO 9921881 , U.S. Patent Nos. 6,485,901 and 6,649,349), peptide nucleic acid (PNA) Molecular Beacons (see, e.g., U.S. Patent Nos.

6,355,421 and 6,593,091), linear PNA beacons (see, e.g. U.S. Patent No.

6,329,144), non-FRET probes (see, e.g., U.S. Patent No. 6,150,097), Sunrise™ /AmplifluorB™probes (see, e.g., U.S. Patent No. 6,548,250), stem-loop and duplex Scorpion™ probes (see, e.g., U.S. Patent No. 6,589,743), bulge loop probes (see, e.g., U.S. Patent No. 6,590,091), pseudo knot probes (see, e.g., U.S. Patent No. 6,548,250), cyclicons (see, e.g., U.S. Patent No. 6,383,752), MGB Eclipse™ probe (Epoch Biosciences), hairpin probes (see, e.g., U.S. Patent No.

6,596,490), PNA light-up probes, antiprimer quench probes (Li et al., Clin. Chem. 53:624-633 (2006)), self-assembled nanoparticle probes, and ferrocene-modified probes described, for example, in U.S. Patent No. 6,485,901.

[069] In certain embodiments, one or more of the primers in an

amplification reaction can include a label. In yet further embodiments, different probes or primers comprise detectable labels that are distinguishable from one another. In some embodiments a nucleic acid, such as the probe or primer, may be labeled with two or more distinguishable labels.

[070] In some aspects, a label is attached to one or more probes and has one or more of the following properties: (i) provides a detectable signal; (ii) interacts with a second label to modify the detectable signal provided by the second label, e.g., FRET (Fluorescent Resonance Energy Transfer); (iii) stabilizes hybridization, e.g., duplex formation; and (iv) provides a member of a binding complex or affinity set, e.g., affinity, antibody-antigen, ionic complexes, hapten- ligand (e.g., biotin-avidin). In still other aspects, use of labels can be

accomplished using any one of a large number of known techniques employing known labels, linkages, linking groups, reagents, reaction conditions, and analysis and purification methods.

[071] MiRNAs can be detected by direct or indirect methods. In a direct detection method, one or more miRNAs are detected by a detectable label that is linked to a nucleic acid molecule. In such methods, the miRNAs may be labeled prior to binding to the probe. Therefore, binding is detected by screening for the labeled miRNA that is bound to the probe. The probe is optionally linked to a bead in the reaction volume.

[072] In certain embodiments, nucleic acids are detected by direct binding with a labeled probe, and the probe is subsequently detected. In one embodiment of the invention, the nucleic acids, such as amplified miRNAs, are detected using FlexMAP Microspheres (Luminex) conjugated with probes to capture the desired nucleic acids. Some methods may involve detection with polynucleotide probes modified with fluorescent labels or branched DNA (bDNA) detection, for example.

[073] In other embodiments, nucleic acids are detected by indirect detection methods. For example, a biotinylated probe may be combined with a streptavidin-conjugated dye to detect the bound nucleic acid. The streptavidin molecule binds a biotin label on amplified miRNA, and the bound miRNA is detected by detecting the dye molecule attached to the streptavidin molecule. In one embodiment, the streptavidin-conjugated dye molecule comprises Phycolink® Streptavidin R-Phycoerythrin (PROzyme). Other conjugated dye molecules are known to persons skilled in the art.

[074] Labels include, but are not limited to: light-emitting, light-scattering, and light-absorbing compounds which generate or quench a detectable

fluorescent, chemiluminescent, or bioluminescent signal (see, e.g., Kricka, L., Nonisotopic DNA Probe Techniques, Academic Press, San Diego (1992) and Garman A., Non-Radioactive Labeling, Academic Press (1997).). Fluorescent reporter dyes useful as labels include, but are not limited to, fluoresceins (see, e.g., U.S. Patent Nos. 5,188,934, 6,008,379, and 6,020,481), rhodamines (see, e.g., U.S. Patent Nos. 5,366,860, 5,847,162, 5,936,087, 6,051 ,719, and

6,191 ,278), benzophenoxazines (see, e.g., U.S. Patent No. 6,140,500), energy- transfer fluorescent dyes, comprising pairs of donors and acceptors (see, e.g., U.S. Patent Nos. 5,863,727; 5,800,996; and 5,945,526), and cyanines (see, e.g., WO 9745539), lissamine, phycoerythrin, Cy2, Cy3, Cy3.5, Cy5, Cy5.5, Cy7, FluorX (Amersham), Alexa 350, Alexa 430, AMCA, BODIPY 630/650, BODIPY 650/665, BODIPY-FL, BODIPY-R6G, BODIPY-TMR, BODIPY-TRX, Cascade Blue, Cy3, Cy5, 6-FAM, Fluorescein Isothiocyanate, HEX, 6-JOE, Oregon Green 488, Oregon Green 500, Oregon Green 514, Pacific Blue, REG, Rhodamine Green, Rhodamine Red, Renographin, ROX, SYPRO, TAMRA,

Tetramethylrhodamine, and/or Texas Red, as well as any other fluorescent moiety capable of generating a detectable signal. Examples of fluorescein dyes include, but are not limited to, 6-carboxyfluorescein; 2',4',1 ,4,-tetrachlorofluorescein; and 2^l,4^l,5',7',1 ,4-hexachlorofluorescein. In certain aspects, the fluorescent label is selected from SYBR-Green, 6-carboxyfluorescein ("FAM"), TET, ROX, VICTM, and JOE. For example, in certain embodiments, labels are different fluorophores capable of emitting light at different, spectrally-resolvable wavelengths (e.g., 4- differently colored fluorophores); certain such labeled probes are known in the art and described above, and in U.S. Patent No. 6,140,054. A dual labeled fluorescent probe that includes a reporter fluorophore and a quencher fluorophore is used in some embodiments. It will be appreciated that pairs of fluorophores are chosen that have distinct emission spectra so that they can be easily

distinguished. [075] In still a further aspect, labels are hybridization-stabilizing moieties which serve to enhance, stabilize, or influence hybridization of duplexes, e.g., intercalators and intercalating dyes (including, but not limited to, ethidium bromide and SYBR-Green), minor-groove binders, and cross-linking functional groups (see, e.g., Blackburn et al., eds. "DNA and RNA Structure" in Nucleic Acids in Chemistry and Biology (1996)).

[076] In further aspects, methods relying on hybridization and/or ligation to quantify miRNAs may be used, including oligonucleotide ligation (OLA) methods and methods that allow a distinguishable probe that hybridizes to the target nucleic acid sequence to be separated from an unbound probe. As an example, HARP-Iike probes, as disclosed in U.S. Publication No. 2006/0078894 may be used to measure the quantity of miRNAs. In such methods, after hybridization between a probe and the targeted nucleic acid, the probe is modified to

distinguish the hybridized probe from the unhybridized probe. Thereafter, the probe may be amplified and/or detected. In general, a probe inactivation region comprises a subset of nucleotides within the target hybridization region of the probe. To reduce or prevent amplification or detection of a HARP probe that is not hybridized to its target nucleic acid, and thus allow detection of the target nucleic acid, a post-hybridization probe inactivation step is carried out using an agent which is able to distinguish between a HARP probe that is hybridized to its targeted nucleic acid sequence and the corresponding unhybridized HARP probe. The agent is able to inactivate or modify the unhybridized HARP probe such that it cannot be amplified.

[077] In an additional embodiment of the method, a probe ligation reaction may be used to quantify miRNAs. In a Multiplex Ligation-dependent Probe Amplification (MLPA) technique (Schouten et al., Nucleic Acids Research 30:e57 (2002)), pairs of probes which hybridize immediately adjacent to each other on the target nucleic acid are ligated to each other only in the presence of the target nucleic acid. In some aspects, MLPA probes have flanking PCR primer binding sites. MLPA probes can only be amplified if they have been ligated, thus allowing for detection and quantification of miRNA biomarkers. IV. EXAMPLES

[078] The following examples illustrate various embodiments of the invention and are not intended to limit the scope of the invention.

[079] The examples described herein include the use of qRT-PCR, which includes real-time monitoring of PCR products during the exponential phase instead of by an end-point measurement. The threshold cycle (C_t) measurements in the examples refer to the number of cycles it takes to reach a pre-defined point in the fluorescent signal.

[080] The examples also describe the use of Receiver Operator

Characteristic (ROC) analysis. A Receiver Operator Characteristic curve is a graphical plot of the sensitivity vs. specificity for a binary classifier system as its discrimination threshold is varied. ROC analysis provides a tool to select possibly optimal models and to discard suboptimal ones independently from (and prior to specifying) the class distribution. Numerous examples of ROC analysis are present in the literature, including applications in selecting and applying

biomarkers for the diagnosis of disease (Pepe et al., Biometrics 62:221-229 (2006); Dodd et al., Biometrics 59:614-623 (2003)). ROC analysis captures the continuum of sensitivity and specificity, but it can be summarized as a single quantity, i.e., the area under the curve (AUC) of the ROC. The AUC is closely related to the nonparametric Wilcox test and summarizes the separation of classes over all the thresholds. Advantages of the ROC technique include (1) it does not assume a parametric form of the class probability as required in the logistic regression method, (2) it is adaptable to outcome-dependent samplings, e.g. the case-control design, which are widely used in medical studies, and (3) it is relatively straightforward to assign different 'costs' to false positives and false negatives (Dodd 2003; Pepe 2006).

Example 1. Mouse Model for Early Detection of Lung Cancer

[081] The A/J mouse model was used to chemically induce lung tumors and monitor the miRNA expression profile in plasma. Benzo[a]pyrene (Sigma- Aldrich, St. Louis, MO, USA; cat. no. 48564) served as a carcinogen to

chemically induce lung tumors in mice. A/J mice (67 male and 68 female) were purchased from The Jackson Laboratory (Bar Harbor, Maine, USA; Stock Number 000646) and sent to Perry Scientific Inc. (San Diego, CA, USA), at the age of six weeks. Perry Scientific performed all the animal related experiments including carcinogen administration, animal monitoring, blood withdrawal and plasma processing.

[082] To obtain baseline data, two weeks prior to carcinogen

administration, 125 mice were bled via orbital sinus and 10 mice were sacrificed by isofluorane inhalation and bled out by cardiac puncture. On day 0, mice in the experimental group were dosed with carcinogen via oral gavage at a dose equivalent to 20 μM in 200 μl (250 mg/kg body weight) in cottonseed oil (Sigma- Aldrich, cat. no. CIlQl). Ten control mice (5 males and 5 females) received an oral gavage of 200 μl cottonseed oil with no carcinogen. A second carcinogen dose (250 mg/kg body weight) was administered to the test group at 10 weeks post the first gavage, with the control animals receiving a second gavage of cottonseed oil as described above. Mice were sacrificed at specific time points (Figure 1), and the remaining mice were bled every two weeks (Figure 1). At termination, animals were weighed, bled via cardiac puncture, and lung lobes collected. Control mice were bled throughout the study and sacrificed at week 34 post-carcinogen administration.

[083] Blood samples were processed individually into plasma. For plasma preparation, blood samples were collected into BD Vacutainer® K2EDTA tubes (Becton, Dickinson and Company; Franklin Lakes, NJ, USA; cat. no. 367841). Blood was centrifuged at 2,000 x g for 10 minutes, after which the plasma layer was aspirated and put into a fresh tube and centrifuged at 2,500 x g for 10 minutes. The resulting plasma was immediately stabilized by the addition of 2X denaturing buffer from the m/Λ/ana™ miRNA Isolation Kit (Ambion Inc., Austin, TX, USA; cat no. AM1560 ) to achieve a final concentration of 1X. Plasma and buffer were mixed by vortexing and immediately frozen at -80 ⁰C until shipment to Asuragen on dry ice. Cell pellets were also frozen at -80 ⁰C and shipped to Asuragen with plasma samples.

[084] Plasma RNA was purified using the organic extraction of the mirVana PARIS™ Kit (Ambion, Inc.; Part No. AM1556), with the following modifications. Following the addition of acid phenokchloroform and vortexing, samples were incubated on ice for 5 min then centrifuged at 13,000 x g for 10 min at 4 ⁰C. The aqueous layer was removed, extracted with chloroform, and centrifuged again. The aqueous layer was removed from the second extraction, and 3M NaOAc (1/10 volume), glycogen (5 mg/ml), and 100% ethanol (1.5 volume) were added to the samples. Lysate/ethanol mixtures were passed through a mirVana PARIS™ filter cartridge, and filters were washed once with 650 μl of Wash 1 buffer and twice with 650 μl of Wash 2/3 buffer. RNA was eluted with two aliquots of nuclease free water (50 μl) and stored at -80 ⁰C.

[085] One half of the lung lobes from each mouse were fixed in 10% buffered formalin overnight, embedded in paraffin (FFPE), and processed for hematoxylin and eosin (H&E) stained slides. The other half were snap frozen at - 80 °C. FFPE lung sections were step-sectioned at 5 mm intervals. The following sectioning guidelines were used: If no grossly visible tumors were present in either lung, three random 5 x 5 x 5 mm sections from each lung were taken for snap- freezing and for FFPE preparation. If grossly visible tumors were present, one 5 x 5 x 5 mm section of one tumor from each lung was snap-frozen and the other half was prepared for FFPE. If only one lung had tumors, random regions from the uninvolved lung were snap-frozen and FFPE prepared. For histopathology, the lung lobes preserved in formalin were sent to Pacific Pathology Inc. (San Diego, CA, USA) where they were embedded in paraffin, sectioned, stained with hematoxylin and eosin, and prepared into slides.

[086] H&E-stained slides were sent to Asuragen and analyzed by a board certified pathologist and classified as normal, hyperplasia, adenoma, or

adenocarcinoma of the lung. Table 6 shows the number of visible tumors observed in each sacrificed animals, and the pathological diagnosis in each of the lung lobes.

Table 6. Lung tumors and pathological diagnosis from experimental and control mice. SCC, squamous cell carcinoma.

[087] miRNA expression was evaluated in plasma samples from mice with pathology-confirmed adenocarcinoma of the lung and from control mice that did not receive carcinogen. Control mice were age matched to the mice with adenocarcinoma. Expression levels of 329 miRNAs, were determined by qRT- PCR using TaqMan® MicroRNA Assays (Applied Biosystems) specific for each miRNA. A subset of the miRNAs (170) have identical mature sequences in both mice and humans and the remainder are mouse homologues of corresponding human miRNAs. Reverse transcription (RT) reaction components (Table 7) were assembled on ice prior to the addition of RNA template, and included ~1 ng plasma RNA per each 10 μl reaction. RT reactions were incubated in a 384-well GeneAmp® PCR System 9700 (Applied Biosystems) at 16 ⁰C for 30 minutes, then at 42 ⁰C for 30 minutes, then at 85 ⁰C for 5 minutes. RT reactions were then frozen at -20 ⁰C. Table 7. Reverse transcription reaction components.

[088] PCR components (Table 8) were assembled on ice prior to the addition of cDNA (2 μl) from the RT reaction. Reactions were incubated in an ABI PRISM™ 7900HT Fast Real-Time PCR system (Applied Biosystems) at 95 ⁰C for 1 minute, then for 50 cycles at 95 ⁰C for 5 seconds and 60 ⁰C for 30 seconds. Results were analyzed with the 7900HT Fast Real-Time PCR system SDS V2.3 software (Applied Biosystems).

Table 8. PCR components.

[089] Table 9 shows the average Ct data and ddCt values for cancer vs. control mice from the qRT-PCR experiment. miRNAs with p-values of 0.1 or less are shown.

Table 9. miRNA expression in plasma samples from mice with pathology- confirmed adenocarcinoma of the lung and from control mice that did not receive carcinogen.

[090] In addition, a temporal study was performed to examine miRNA expression in animals that had been diagnosed with lung adenocarcinomas. Figure 2A and Figure 2B show the changes in expression of selected miRNAs over time, prior to the detection of tumors. This figure shows that differential expression of miRNAs can be detected as early as 10 weeks after carcinogen administration-at least 6 weeks prior to the detection of tumors. These results demonstrate that miRNA can be used to detect lung cancer at an early stage and are suitable for screening for lung cancer.

Example 2. q RT-PC R Identification of miRNA Biomarkers for the Detection of Lung Cancer

[091] The expression of 12 miRNAs were evaluated in serum samples from eight lung cancer patients and eight normal patients. Blood from lung cancer patients and normal donors was collected into BD Vacutainer® Plus Plastic Serum Tubes. All blood was collected at the time of diagnosis but prior to any medical intervention, such as tumor resection or treatment. Blood was centrifuged at 1 ,000 x g for 10 min and serum transferred to a fresh tube and immediately frozen at -80 ⁰C. Serum RNA was extracted using the organic extraction of the mirVana PARIS™ Kit (Ambion, Inc., Austin, TX, USA; Part No. AM1556), with the following modifications. Following the addition of acid phenohchloroform and vortexing, samples were incubated on ice for 5 min then centrifuged at 13,000 x g for 10 min at 4 ⁰C. The aqueous layer was removed, extracted with chloroform, and centrifuged again. The aqueous layer was removed from the second extraction, and 3M NaOAc (1/10 volume), glycogen (5mg/ml), and 100% ethanol (1.5 volume) were added to the samples.

Lysate/ethanol mixtures were passed through a mirVana PARIS™ filter cartridge, and filters were washed once with 650 μl of Wash 1 buffer and twice with 650 μl of Wash 2/3 buffer. RNA was eluted with two aliquots of nuclease free water (50 μl) and stored at -80 ⁰C.

[092] miRNA expression levels were determined by qRT-PCR using TaqMan® MicroRNA Assays (Applied Biosystems; Foster City, CA, USA) specific for each miRNA. Reverse transcription (RT) reactions were performed using the TaqMan® MicroRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA; cat. no. 4366597). Reaction components listed in Table 10 were assembled on ice prior to the addition of RNA template. All reaction components were as provided by the manufacturer (Applied Biosystems; Foster City, CA, USA) as a kit with multiple components. Serum RNA (1 ng total RNA per reaction) was added and mixed. RT reactions were incubated in a 384-well GeneAmp® PCR System 9700 (Applied Biosystems) at 16 ⁰C for 30 minutes, then at 42 ⁰C for 30 minutes, then at 85 ⁰C for 5 minutes. RT reactions were then frozen at -20 ⁰C.

Table 10. Reverse transcription reaction components.

[093] PCR components (Table 11) were assembled on ice prior to the addition of cDNA (4 μl) from the RT reaction. Reactions were incubated in an ABI PRISM™ 7900HT Fast Real-Time PCR system (Applied Biosystems) at 95 °C for 1 minute, then for 50 cycles at 95 ⁰C for 5 seconds and 60 ⁰C for 30 seconds. Results were analyzed with the 7900HT Fast Real-Time PCR system SDS V2.3 software (Applied Biosystems).

Table 11. PCR components.

[094] Table 12 provides average dCt values for cancer and normal samples for each miRNA. In addition, ddCt values represent the difference in expression between cancer and normal samples.

Table 12: Expression of miRNAs in serum from patients with and without lung cancer.

[095] In an expanded panel, the expression of 180 miRNAs were evaluated in serum samples from sixteen lung cancer patients and from twelve individuals with benign lung conditions (Table 13). In Table 13, Diff stands for differential status of the tumor; MOD stands for moderate; AC for

adenocarcinoma; SCC for squamous cell carcinoma; R for right; RUL for right upper lobe; RML for right mid lobe; RLL for right lower lobe; LUL for left upper lobe; and LLL for left lower lobe.

Table 13. Histopathological and patient information for lung cancer and benign specimens.

[096] Blood from lung cancer patients and comparator donors was collected into BD Vacutainer® Plus Plastic Serum Tubes and processed as described above for the initial human samples.

[097] Table 14 provides differential expression data for miRNAs in the form of average dCt values for cancer and benign samples. In addition, ddCt values represent the difference in expression between cancer and benign samples.

Table 14. Expression of miRNAs in serum from patients with lung cancer or benign lung conditions.

[098] Pairs of miRNA biomarkers were also evaluated for their abilities to distinguish serum from patients for these samples. Un-normalized qRT-PCR data were used to calculate dCt values for each miRNA pair that was evaluated. The dCt values of the various miRNA pairs were analyzed using Receiver Operator Characteristic (ROC) analysis to identify the diagnostic miRNA pairs for lung cancer. In this example, ROC curves were used to evaluate the capacities of individual and combinations of biomarker candidates to distinguish patient classes. The abundance(s) of biomarker candidates were measured, and the measurements were used to develop classifiers whose true positive and false positive rates were plotted in a ROC curve. The AUC was then calculated for each classifier. An ideal classifier has an ROC AUC value of 1, and classifiers were ranked based upon how close their ROC AUC is to 1. [099] Table 15 includes miRNA pairs that distinguish patients with lung cancer from those with benign lung conditions. The miRNA pairs in Table 15 are listed in order of decreasing ROC AUC. Serum miRNA biomarker pairs for classifying patients with cancer vs. patients with benign lung conditions are presented. Mean values represent the average difference in Gt values between the two miRNAs in the pair. In Table 15, "Benign mean" represents the mean of data from 12 subjects with benign lung conditions; "Cancer mean" is the mean for data from 16 lung cancer patients; "Benign SD" is the standard deviation for data from patients with benign lung conditions; "Cancer SD" is the standard deviation for data from lung cancer patients; "Benign vs Cancer Assoc." refers to the p- value associated with the diagnosis of benign/lung cancer; ROC stands for receiver operating characteristic; and AUC is area under curve.

Table 15. microRNA biomarker pairs - benign lung and lung cancer patient sera.

[0100] The prevalence of serum biomarkers in paired analysis from Table 15 are shown in Table 16.

Table 16: Prevalence of serum biomarkers in pairs

[0101] The data in Tables 15 and 16 show miRNAs and miRNA biomarker pairs that can distinguish patients with benign conditions of the lung from patients with lung cancer in serum samples, irrespective of patient sex or cancer type. These miRNAs are useful for diagnosis of lung cancer. The serum samples from cancer patients were from patients with early stage I through stage III lung cancer. These results show that miRNAs are suited to detect lung cancer at an early stage and are effective for screening patients and distinguishing patients with benign lung conditions from those with lung cancer.

Example 3. Differentially Expressed miRNAs in Serum from Male and

Female Lung Cancer Patients

[0102] While the serum miRNAs in Tables 15 and 16 are able to distinguish patients with benign lung conditions from patients with lung cancers, further analysis of the data from Example 2 revealed that certain miRNAs pairs were better for distinguishing those patient groups within a specific sex. In particular, a number of miRNA pairs had an ALJC ROC of 1.00 (Table 17). These pairs showed a significant improvement of the ROC AUC when distinguishing female patients with benign lung conditions from female lung cancer patients. Similarly, an improvement in AUC was also observed for specific miRNA pairs when distinguishing male patients with benign lung conditions from male patients with lung cancers (Table 18). The miRNAs and miRNA pairs in Tables 17 and 18 are useful for diagnosing lung cancer in female and male patients respectively. These miRNAs can be used in combinations of two or more. Table 17. microRNAs pairs that are differentially expressed in serum of female patients with benign lung conditions as compared to female lung cancer patients.

Table 18. microRNA pairs that are differentially expressed in serum of male patients with benign lung conditions as compared to male lung cancer patients.

Example 4. miRNA Markers That Distinguish Lung Cancer from Cancer-free Subjects With or Without Benign Conditions

[0103] In an additional study, serum miRNAs that distinguish lung cancer patients from cancer free human subjects with or without benign conditions are identified. Serum samples were obtained from patients with benign conditions, normal patients, and lung cancer patients (Table 19). All procedures and data analysis were performed as described above.

Table 19. Histopathological and patient information for lung cancer and benign specimens.

[0104] Table 20 provides differential expression data for miRNAs in the form of average dCt values for cancer and other samples (other=benign and normal). Table 21 shows differentially expressed miRNA biomarker pairs that distinguish samples from lung cancer patients from normal patients and patients with benign tumors. Such miRNA and biomarker pairs can be used for screening and diagnosis of lung cancer. Table 20. Expression of miRNAs in serum from patients with lung cancer or from patients with benign or normal conditions. Other=benign and normal cases.

Table 21. Differentially expressed miRNA biomarker pairs that distinguish lung cancer patients from normal patients with no lung tumors and patients with benign lung tumors. Other=benign and normal cases.

[0105] Table 22 shows miRNAs that can be used in combination with other miRNA biomarkers to characterize lung disease, as well as the prevalence of those biomarkers in pairs from Table 21.

Table 22: Prevalence of serum biomarkers in paired analysis of cancer vs.

(normal and benign) patients.

Example 5. Lung Cancer Classifier Development

[0106] An initial training set of 14 miRNAs were selected from pairs in Table 15 to further verify their diagnostic potential. These miRNAs were shown above to have diagnostic potential for distinguishing patients with benign lung conditions from patients with lung cancer, as evidenced by the high AUC ROC scores for miRNA biomarker pairs that include these miRNAs. From a group of 20 benign lung samples and 34 lung cancer samples (Table 23), a set of training samples (12 benign, 16 lung cancer) was chosen from patients that were age and sex matched. qRT-PCR was performed as described above in Example 2.

Table 23. Histopathological and patient information for lung cancer and benign specimens.

[0107] To verify a subset of the 14 miRNAs in the training set, a separate test sample set was selected from Table 23 that was composed of 8 benign and 18 LuCa samples. miRNA biomarker selection and classifier evaluation was generated by performing 25 repetitions of 5-fold cross validation on the training samples, and measuring the AUC ROC values as a function of the number of miRNA biomarkers (features). A range of biomarkers was used in order to determine the optimal number of features for classification. Five-fold cross- validation is a process where the training set is subdivided into an 80% sample set and a 20% sample set, and feature selection and classifier training is performed on the 80% training samples. Performance is measured by classifying the remaining 20% samples not used for feature selection and classifier training. This process was repeated 5 times so that all samples were in the 20% test set once and only once. One iteration of 5-fold cross-validation produced a single AUC value. Five-fold cross-validation was repeated 25 times in order to achieve a better sampling of combinations of 80% and 20% sets, and to identify an empirical distribution of AUC values for a given number of biomarkers in the classifier. Top biomarkers according to a simple Welch's t-test were used as inputs into a Linear Discriminate Analysis (LDA) classifier for training. Biomarkers were defined as pairs of miRNAs (differential expression of two miRNAs), where the expression value of one miRNA is subtracted from the expression value of another miRNA. All possible pairs were investigated. Figure 3 shows the distribution of AUC values for the LDA classifier as a function of the number of biomarkers. This procedure determines the optimal number of biomarkers to use for the classifier, and performance expectations moving forward. A classifier of 6 differentially expressed miRNA pairs (miR-142-5p and miR181d; miR-142-3p and miR181d; miR-142-3p and miR-422a; miR-142-5p and miR-422a; miR-92 and miR-27b; miR-24 and miR-27a) was identified as the optimal classifier for diagnosis of lung cancer in this experiment. This classifier is marked by the vertical dotted line in Figure 3. A cutoff for the LDA classifier was set so that the continuous valued classification probabilities were converted into calls diagnosing patients as having lung cancer (classifier-positive) or benign (classifier-negative) tumors. For simplicity, a cutpoint of 0.5 was chosen as the threshold. This value produced a near maximum of sensitivity and specificity. Performance estimates of training and test data using a cutoff of 0.5 of the LDA classifier trained on six biomarker pairs is provided in Figure 4. The sensitivity (SENS), specificity (SPEC), negative predictive value (NPV) and positive predictive value (PPV) are as indicated.

[0108] Any of the six pairs identified as part of the classifier in this experiment may be used to diagnose lung cancer. Predictive accuracy may increase with the use of more than one pair from the optimal classifier set.

Example 6. q RT-PC R for evaluation of microRNA expression in human lung cancer patient plasma RNA

[0109] To identify miRNA biomarkers for lung cancer in human plasma, qRT-PCR was used to compare levels of miRNAs present in the plasma of patients with lung cancer (n=14) to the levels in patents without lung cancer (n=12). Patient information and lung cancer pathology information are shown in Table 24.

Table 24. Human Lung Cancer and Normal Patient Information and Tumor Pathology. TNM stage (Sobin and Wittekind, 2002)

[0110] For plasma preparation, whole blood (10 ml) was collected from each donor into a BD Vacutainer® K2EDTA tube (Becton, Dickinson and

Company; Franklin Lakes, NJ, USA). Tubes were inverted 8-10 times

immediately after blood collection and then centrifuged for 10 minutes at 2,000 x g within two hours of collection. Plasma was transferred to a new tube using a serological pipette and frozen at -80 ⁰C until it was subjected to RNA isolation.

[0111] Plasma RNA was purified using the organic extraction of the mirVana PARIS™ Kit (Part No. AM1556; Applied Biosystems/Ambion; Austin, TX, USA), with the following modifications. After thawing plasma on ice, an equal volume of 2X denaturing solution from the mirVana PARIS™ Kit was added and the mixture was incubated on ice for five minutes. An equal volume of acid phenol:chloroform was added, then the mixture was vortexed for one minute, and incubated on ice for five minutes. Tubes were centrifuged at 13,000 x g for 15 minutes at 4 ⁰C, and the aqueous phase was removed to a fresh tube. Glycogen (5 mg/ml), 3M NaOAc (1/10 volume), and 100% ethanol (1.5 volume)! were added to the samples. Samples were mixed by inversion and incubated at room temperature. Lysate/ethanol mixtures were passed through a mirVana PARIS™ filter cartridge, and filters were washed once with 650 μl of Wash 1 buffer and twice with 650 μl of Wash 2/3 buffer. RNA was eluted with nuclease free water (50 μl, heated to 95 ⁰C), by centrifugation at 10,000 x g for 1 minute. [0112] qRT-PCR reactions were performed using TaqMan® MicroRNA Assays (Applied Biosystems; Foster City, CA, USA) specific for each individual miRNA. Reverse transcription reactions were assembled on ice prior to the addition of RNA template (Table 25). Next, 0.25 μl of plasma RNA template per reaction was added and mixed. RT reactions were incubated in a 384-well GeneAmp® PCR System 9700 (Applied Biosystems) at 16 ⁰C for 30 minutes, then at 42 ⁰C for 30 minutes, then at 85 ⁰C for 5 minutes, and then were frozen in a -20 ⁰C freezer.

Table 25. Reverse transcription reaction components.

[0113] PCR components were assembled on ice prior to the addition of the cDNA from the RT reactions (Table 26). PCRs were incubated in an ABI

PRISM™ 7900HT Fast Real-Time PCR system (Applied Biosystems) at 95 ⁰C for 1 minute, then for 50 cycles at 95 ⁰C for 5 seconds, and then at 60 ⁰C for 30 seconds. Initial data analysis was done using the 7500 Fast System SDS V2.3 software (Applied Biosystems). Table 26. Real-time PCR components.

[0114] The qRT-PCR data were initially assessed for outliers. All miRNAs in a given sample with raw Ct values of 50 were eliminated from further analysis. All data from samples with fewer than 150 miRNAs that had raw Ct values less than 50 were eliminated. The average raw Ct for 50 miRNAs that were detected in each plasma sample was calculated for each individual sample. The average Ct for a given sample was subtracted from the raw Ct values for each miRNA in the corresponding sample to produce a dCt for each miRNA that was detected. These normalized measures were used to identify miRNAs that were present at significantly different levels in the plasma samples from normal donors and lung cancer patients.

[0115] The average dCt values for each miRNA in the normal donor and lung cancer patient samples were calculated. The average dCt values for the lung cancer patient samples were subtracted from the average dCt values for the normal donors to determine the variance in the levels of the various miRNAs between the two patient sets. The student t-test was then used to evaluate the capacity of the various miRNAs to distinguish the plasmas of lung cancer patients and normal donors. [0116] Table 27 provides differential expression data for miRNAs in the form of average dCt values for cancer and normal samples. In addition, ddCt values represent the difference in expression between cancer and normal samples.

Table 27. miRNA Biomarkers Differentially Expressed Between Plasma Samples of Lung Cancer Patients and Plasma Samples of Normal Patients.

Example 7. QRT-PCR for Evaluation of MicroRNA Expression in Human Lung Cancer Patient Plasma RNA

[0117] A second set of plasma samples, isolated from a distinct set of lung cancer and normal patients (Table 28), was used to perform an additional comparison of miRNAs expressed in lung cancer and normal patients. In Table 28, TNM (tumor, node, metastasis) stage is described in Sobin and Wittekind, New Jersey: John Wiley & Sons, 2002. The following abbreviations are used in Table 28: adenocarcinoma (ADCA); bronchoalveolar (BA), squamous cell carcinoma (SCC), and not available (NA). Table 28. Human Lung Cancer and Normal Patient Information and Tumor Pathology. .

[0118] To assess the expression of miRNAs, real-time RT-PCR detection with TaqMan® Micro RNA Assays (Applied Biosystems, Foster City, CA, USA) was used to screen RNA isolated from seven normal human plasma and nine lung cancer patient plasma samples.

[0119] Plasma RNA was isolated as described above in Example 6. qRT- PCR reactions were performed using TaqMan® MicroRNA Assays (Applied Biosystems; Foster City, CA, USA) specific for each individual miRNA. Reverse transcription reactions were assembled on ice prior to the addition of RNA template (Table 29). Next, 0.25 μl of plasma RNA template per reaction was added and mixed. RT reactions were incubated in a 384-well GeneAmp® PCR System 9700 (Applied Biosystems) at 16 ⁰C for 30 minutes, then at 42 ⁰C for 30 minutes, then at 85 ⁰C for 5 minutes, and then were frozen in a -20 ⁰C freezer.

Table 29. Reverse transcription reaction components.

[0120] PCR components were assembled on ice prior to the addition of the cDNA from the RT reactions (Table 30). PCRs were incubated in an ABI

PRISM™ 7900HT Fast Real-Time PCR system (Applied Biosystems) at 95 ⁰C for 1 minute, then for 50 cycles at 95 ⁰C for 5 seconds and then at 60 ⁰C for 30 seconds. Initial data analysis was done using the 7500 Fast System SDS V2.3 software (Applied Biosystems).

Table 30. Real-time PCR components

[0121] The qRT-PCR data were initially assessed for outliers. All miRNAs in a given sample with raw Ct values of 50 were eliminated from further analysis. All data from samples with fewer than 150 miRNAs that had raw Ct values less than 50 were eliminated. The average raw Ct for 50 miRNAs that were detected in each plasma sample was calculated for each individual sample. The average Ct for a given sample was subtracted from the raw Ct values for each miRNA in the corresponding sample to produce a dCt for each miRNA that was detected. These normalized measures were used to identify miRNAs that were present at significantly different levels in the plasma samples from normal donors and lung cancer patients.

[0122] The average dCt values for each miRNA in the normal donor and lung cancer patient samples were calculated. The average dCt values for the lung cancer patient samples were subtracted from the average dCt values for the normal donors to determine the variance in the levels of the various miRNAs between the two patient sets. The student t-test was then used to evaluate the capacity of the various miRNAs to distinguish the plasmas of lung cancer patients and normal donors.

[0123] Table 31 provides dCt values for normal and cancer patients, as well as ddCt values for cancer vs normal samples for each miRNA. Table 31. miRNA Biomarkers Differentially Expressed Between Plasma Samples of Lung Cancer Patients and Plasma Samples of Normal Patients.

Example 8. Combinations of miRNAs that Distinguish Plasma from Lung Cancer Patients and Plasma from Normal Patients

[0124] The un-normalized qRT-PCR data, generated as described in Example 6, were used to calculate dCt values for each pair of miRNAs that was quantified. The dCt results for the various miRNA pairs in the lung cancer and normal donor plasma samples were analyzed using Receiver-Operator

Characteristic (ROC) analysis to determine which miRNA pairs have the capacity to distinguish plasma samples from lung cancer patients and those from normal patients. One miRNA pair (let-7c:miR-326) accurately classified the 14 lung cancer patient samples and 12 normal donor samples that were analyzed (Table 32). Twenty additional miRNA pairs accurately classified all but one of the 26 samples (ROC AUC >0.98) and 221 miRNA pairs had an ROC AUC score of at least 0.93 (Table 32). 166 independent miRNAs were included at least once in the 221 miRNA pairs of Table 32 (Table 33). Numerous miRNAs were used in multiple biomarkers pairs, indicating the strength of their variable levels in the plasmas of lung cancer patients and normal donors. Table 32. Biomarker pairs that can be used to identify plasma from lung cancer patients.

Table 33. Biomarkers that can be used in combination to identify plasma from lung cancer patients, and the prevalence of each miRNA.

Example 9. QRT-PCR for Evaluation of MicroRNA Expression in Human Lung Cancer Patient Serum RNA

[0125] To further assess the selectivity and sensitivity of certain miRNAs and combinations of miRNA, an additional set of serum samples, isolated from a distinct set of lung cancer patients and normal subjects (Table 34), was used to compare miRNAs expressed in serum from lung cancer patients and normal subjects. Histopathological analyses of lung lobe sections from cancer patients and normal subjects were performed by a board certified pathologist to determine a pathological diagnosis.

Table 34. Histopathological and patient information for serum samples. ADCA, adenocarcinoma; SCCA, squamous cell carcinoma.

[0126] First, the expression of 180 miRNAs in serum RNA was evaluated from the 30 lung cancer patients and 20 normal subjects (Table 34). Blood collection, serum RNA extraction, and qRT-PCR were performed as described in Example 2 above. Table 35 shows the mean dCt and ddCt values following PCR amplification of miRNAs from the serum samples. dCt values were calculated as the difference between the Ct of a specific miRNA and the mean Ct for the top 50 expressed miRNAs as a normalizer. The miRNAs are arranged in increasing t- test p-value.

Table 35. Expression of miRNAs in serum from patients having lung cancer and from normal subjects.

[0127] Based on analysis of differentially expressed miRNA pairs, 27 miRNAs were selected for further validation in serum samples from an additional 55 lung cancer patients and from 75 cancer-free, presumed normal individuals (Table 36).

Table 36. Lung cancer patient, tumor pathology, and normal subject information. TNM Stage (Sobin and Wittekind, 2002); AJCC Stage (Greene, 2002); ADCA, adenocarcinoma; ADCA/BA, adenocarcinoma of the bronchoalveolar; SCCA, squamous cell carcinoma; NA, not applicable.

[0128] The 27 miRNAs were evaluated for their ability to distinguish lung cancer patient samples from normal subject samples shown in Table 36. Table 37 shows the mean Ct values and standard deviation for each miRNA for the cancer and the normal group. The dCt values represent the difference of the Cts between the cancer and the normal group. The miRNAs are arranged in increasing t-test p-value. Twenty-four of the 27 miRNAs were differentially expressed at a high statistical significance level, having t-test p-values of less than 0.001. These twenty-four miRNAs can be used for diagnosing lung cancer.

Table 37. Expression of miRNAs in serum from patients having lung cancer or from normal subjects.

[0129] Receiver Operator Characteristic (ROC) analysis was used to identify the miRNA pairs having the capacity to distinguish serum samples of lung cancer patients from those of normal patients. Select differentially-expressed pairs that distinguished normal patients from cancer patients are shown in Table 38 and are arranged in decreasing AUC ranging from 0.99 to 0.74.

Table 38. miRNA biomarker pairs that can be used to identify serum from lung cancer patients.

[0130] In addition to the differential paired analysis of miRNA (Table38), several miRNA biomarker based classifier models that discriminated lung-cancer from normal (cancer free) subjects were evaluated. Classifier models were designed with data obtained from the samples described in Table 34. When applied to the unrelated samples shown in Table 36, the models demonstrated good performance in discriminating lung cancer patients from normal subjects.

[0131] Specifically, the simple Welch's t-test and Relief (Kira and Rendel, Proc 10^th Natl Conf on Artificial Intelligence, MIT Press, pp. 129-134 (1992)) methods were used to select features (i.e., miRNAs) as inputs, and the Linear Discriminant Analysis (LDA) and Linear Support vector machines (LSVM) were used as the classifiers during training (Burges, Data Mining and Drug Discovery 2:121-167 (1998)). Table 39 shows representative examples of classifier models that distinguished lung cancer from cancer free subjects in both training and test sets.

Table 39. Performance of classification models and feature selection in training set and test set of normal subjects and lung cancer patients.

[0132] The model's performance was estimated on training data by performing 25 repetitions of 5-fold cross validation on the training samples and measuring the ROC AUC values as previously described in Example 5. The frequency of appearance of miRNA pairs in each of the classification models is shown in Table 40.

Table 40. Prevalence of paired biomarkers in classifier generation. The classification models are as indicated.

[0133] The unique miRNAs that were most frequently associated with the classifier models from Table 40 and that distinguished lung cancer from cancer free subjects are shown in Table 41. These miRNAs can be used to distinguish patients with lung cancer from cancer free subjects, and therefore can be used in the characterization or diagnosis of lung cancer.

Table 41 : miRNAs that can be used to distinguish patients with lung cancer from cancer free subjects

[0134] Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.

[0135] All references cited herein are incorporated herein by reference in their entirety. To the extent publications and patents or patent applications incorporated by reference contradict the invention contained in the specification, the specification will supersede any contradictory material.

Claims

What is claimed is:

1. A method for characterizing lung cancer in a patient with a lung tumor or lesion comprising the steps of:

a. measuring the level of a miRNA in a serum sample, wherein the miRNA is chosen from let-7b, let-7c, let-7d, let-7e, miR-10a, miR-10b, miR- 130b, miR-132, miR-133b, miR-139, mιR-143, miR-152, miR-155, miR-15b, miR-17-5p, miR-193, miR-194, miR-195, miR-196b, miR-199a^*, miR-19b, miR-202, miR-204, miR-205, miR-206, miR-20b, miR-21 , miR-210, miR- 214, miR-221 , miR-27a, miR-27b, miR-296, miR-29a, miR-301 , miR-324- 3p, miR-324-5p, miR-339, miR-346, miR-365, miR-378, miR-422a, miR- 432, miR-485-3p, miR-496, miR-497, miR-505, miR-518b, miR-525, miR- 566, miR-605, miR-638, miR-660, and miR-93; and

b. determining reduced or elevated levels of the miRNA in the sample, thereby characterizing lung cancer in the patient.

2. The method of claim 1 , wherein the lung cancer is small cell or non- small cell lung cancer.

3. The method of claim 1 , wherein the patient has previously been screened for lung disease.

4. The method of claim 1 , wherein the patient is suspected of having lung cancer or is at risk of developing lung cancer.

5. The method of claim 3, wherein the screening was by CT scan or chest x-ray.

6. The method of claim 1 , further comprising amplifying the miRNA.

7. The method of claim 6, wherein the amplification is by quantitative reverse transcriptase polymerase chain reaction.

8. The method of claim 1 , further comprising amplifying, measuring, and determining reduced or elevated levels of a second miRNA.

9. A method for characterizing lung disease in a patient comprising the steps of:

a. measuring the level of a first miRNA and a second miRNA in a serum sample, wherein the first miRNA is chosen from let-7a, let-7b, let-7d, let-7f, let-7g, let-7i, miR-101 , miR-106a, miR-106b, miR-125a, miR-126, miR-126^*, miR-130b, miR-132, miR-133b, miR-140, miR-142-3p, miR-142-

5p, miR-145, miR-146a, miR-146b, miR-148b, miR-150, miR-151 , miR-15a, miR-15b, miR-181a, miR-181b, miR-181d, miR-185, miR-186, miR-190, miR-191 , miR-193a, miR-199a^*, miR-202, miR-210, miR-214, miR-222, miR-23a, miR-24, miR-26a, miR-26b, miR-27a, miR-27b, miR-29b, miR- 301 , miR-30a-5p, miR-30b, miR-30c, miR-30d, miR-30e-5p, miR-320, miR- 324-3p, miR-326, miR-335, miR-340, miR-342, miR-345, miR-346, miR- 34a, miR-374, miR-375, miR-378, miR-422a, miR-422b, miR-425, miR-486, miR-496, miR-518b, miR-660, miR-7, miR-92, miR-93, miR-98, miR-99a, and miR-99b; and

b. detecting reduced or elevated levels of the first and second miRNAs in the sample, thereby characterizing lung disease in the patient.

10. The method of claim 9, wherein the first miRNA is chosen from let-7g, miR-106b, miR-126, miR-126^*, miR-132, miR-140, miR-142-3p, miR-146a, miR- 150, miR-15a, miR-15b, miR-181a, miR-181b, miR-181d, miR-214, miR-24, miR- 30a-5p, miR-320, miR-342, miR-345, miR-374, miR-422a, miR-422b, miR-486, and miR-92.

11. The method of claim 9, wherein the first and second miRNAs are a pair chosen from Table 3.

12. The method of claim 9, wherein the lung disease is cancer.

13. The method of claim 9, wherein the patient has a lung tumor or lesion.

14. The method of claim 9, wherein the patient has previously been screened for lung disease.

15. The method of claim 12, wherein the cancer is small cell or non-small cell lung cancer.

16. The method of claim 9, further comprising amplifying the first miRNA and the second miRNA.

17. The method of claim 16, wherein the amplification is by quantitative reverse transcriptase polymerase chain reaction.

18. The method of claim 9, wherein the patient is male.

19. The method of claim 9, wherein the patient is female.