Methods of detecting osteoarthritis and predicting progression thereof

Abstract

Provided herein are methods and biomarkers useful for detecting and diagnosing osteoarthritis and predicting the progression of osteoarthritis in subjects. The diagnoses and predictions of prognosis may be used to develop treatment plans for subjects. Also included are methods of treating subjects and administering pharmaceuticals based on the diagnosis and prognosis predictions.

Description

SEQUENCE LISTING

A Sequence Listing accompanies this application and is incorporated herein by reference in its entirety. The Sequence Listing was filed with the application as a text file on Feb. 5, 2016.

INTRODUCTION

Osteoarthritis is a prevalent chronic disease that represents a large and growing global health burden of large unmet need with respect to diagnostics, prognostics and therapeutics. Based on data from the Global Burden of Disease 2010 Study, musculoskeletal conditions are the second greatest cause of disability, as measured by years lived with disability (YLDs) worldwide and across most regions of the world; the main contributors are low back pain (83.1 million YLDs), neck pain (33.6 million YLDs) and osteoarthritis (17.1 million YLDs) with osteoarthritis of the knee accounting for 83% of this total. Globally, osteoarthritis of the knee affects 251 million individuals, and back and neck pain (likely largely also attributable to osteoarthritis) currently affect 964 million people worldwide. In the US, according to the Centers for Disease Control, osteoarthritis affects an estimated 26.9 million US adults (estimates from 2005, up 28% from the estimated 21 million US adults impacted in 1990). As the US population continues to age and struggle with obesity, the incidence and prevalence of the disease is expected to continue to grow. Consequently, the annual cost of osteoarthritis to the US, estimated to be $89.1 billion in 2001, is anticipated to continue to grow.

Demographics and baseline characteristics are poor predictors of OA progression including age, sex, body mass index, knee pain, general bone mineral content, and joint space width at baseline. In a systematic literature review, it was noted that 25-75% of painful knees cannot be diagnosed as OA by x-ray. Bedson J and Croft P R, BMC Musculoskelet Disord 9:116 (2008). Moreover, knee pain has been reported to have only a 23% sensitivity and 88% specificity for the diagnosis of radiographic OA. Hart et al., Ann Rheum Dis 50(7):467-70 (1991). Bedson et al concluded that, “The results of knee x-rays should not be used in isolation when assessing individual patients with knee pain.” Bedson J and Croft P R, BMC Musculoskelet Disord 9:116 (2008). Our goal was to develop a better means of diagnosing and predicting progression of knee osteoarthritis.

SUMMARY

Methods of diagnosis and predicting progression of osteoarthritis, and in particular knee osteoarthritis, are provided herein. In one aspect, methods of diagnosing osteoarthritis by measuring biomarkers are provided. The methods include measuring the level of a biomarker in a sample from the subject. The biomarker may be at least one biomarker selected from the group consisting of CRAC1 (CRTAC1), A2AP, A1BG, A2GL, AACT, ACTG, AMBP, APOB, APOE, B2MG, C1QC, C1R, C1RL, C4BPA, C4BPB, CD14, CD44, CERU, CFAB, CFAH, CFAI, CILP1, C1S, CNDP1, CO2, CO4B, CO5, CO6A3, CO8B, CO8G, CO9, coll3, COMP, CTX1a, CTX1b, CTX2, CTXi, CXCL7, ECM1, FA12, FA5, FBLN1, FBLN3, FCGBP, FCN3, FETUA, FINC, GELS, HA, HABP2, haptoglobin, HEMO, HEP2, HGFA, HRG, hyaluronan, IC1, ITIH1, ITIH4, KNG1, LAMA2, LUM, LYAM1, MASP1, PCOC1, PGCA, PHLD, PLF4, PLMN, PRG4, RET4, SAMP, SHBG, TENX, TETN, THBG, TIMP1, TSP1, TSP4, VTDB, VTNC, ZA2G, ZPI, or any combination thereof. The levels of the biomarker in the subject are then compared to the levels of the biomarker in a control subject or a reference level of the biomarker. The subject can then be diagnosed with osteoarthritis if the expression of any of the biomarkers is altered as compared to the reference level.

In another aspect, methods of predicting progression of osteoarthritis by measuring the expression levels of biomarkers in a sample from a subject are also provided. The biomarker may be at least one biomarker selected from the group consisting of A1BG, A2AP, A2GL, AACT, ACTG, AFAM, ANT3, APOB, APOH, B2MG, C1QC, C1R, C1RL, C4BPA, C4BPB, CD14, CD163, CD44, CERU, CFAB, CFAH, CFAI, C1S, CO2, CO4B, CO5, CO6A3, CO8B, coll3, CRAC1 (CRTAC1), CTX2, CXCL7, DOPO, ECM1, FA5, FA12, FBLN1, FCGBP, FCN3, FETUA, FINC, GELS, HABP2, haptoglobin, HEMO, HEP2, HGFA, HRG, hyaluronan, ITIH4, KLKB1, KNG1, LUM, LYAM1, PGCA, PHLD, PLF4, PLMN, PRG4, RET4, SAMP, TENX, TETN, THBG, THRB, TIMP1, TSP1, TSP4, VTDB, VTNC, or combinations thereof. The level of the biomarker in the sample is compared to a reference level of the biomarker. The comparison is then used to predict the progression of the osteoarthritis. A significant alteration in the level of any of the biomarkers as compared to the reference level is predictive of progression of osteoarthritis or indicative of risk of osteoarthritis progression.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a flow chart showing how the biomarkers could be used to assist in diagnosing and predicting the progression of knee osteoarthritis and how this would be translated into a treatment plan.

FIG. 2 is a dot plot showing the coefficients of variation for each of the tested peptides and those colored blue and clustering near the log based line were included for further analysis and those in red were not included in the dataset.

DETAILED DESCRIPTION

Methods of diagnosing or predicting progression of osteoarthritis in a subject are provided. The methods all rely on detecting or determining the level of at least one biomarker or combinations of biomarkers in a sample from a subject. In some cases, the subject has knee pain or has already been diagnosed with knee osteoarthritis. The subject may also be diagnosed with, or suspected of having osteoarthritis in another joint other than the knee, such as the hip, back, hand, elbow, shoulder, neck or other joint in the subject. Suitably, the subject is a human, but subjects may include other non-human mammals such as domesticated animals.

Thus, the present methods permit the diagnosis and personalization of therapy or a treatment plan, wherein a subject's biomarker profile is predictive of, or indicative of, a diagnosis of osteoarthritis or risk of progression of osteoarthritis. The methods disclosed herein related to osteoarthritis can be used in combination with assessment of conventional clinical factors or measures, such as age, sex, body mass index or radiographic parameters; this is analogous to the practice for diagnosis or prognosis of rheumatoid arthritis by measuring rheumatoid factor and/or anti-cyclic citrullinated peptide and considering them in conjunction with morning stiffness, joint swelling and/or radiographic features of joint disease, etc. In this manner, the methods of the present disclosure permit a more accurate evaluation of osteoarthritis both at the level of diagnosis and of prognosis of progression of the disease.

In some embodiments, the method includes determining the levels of the biomarkers provided herein in Table 2a in a sample from a subject diagnosed with or suspected of having osteoarthritis. Biomarker levels in some instances may be normalized against the levels of all proteins in the sample, or against a reference or normalization protein(s) in the sample as discussed and exemplified in the Examples. The following set of peptides may be used as normalization peptides in the methods provided herein: TSP1, CNDP1, FA5, SHBG, PLF4, C1QC, ADIPO, APOA4, ACTG, CD14, K2C1, CBG, CHLE, FA11 or any combinations thereof. The level of the biomarkers is indicative of the prognosis for the progression of osteoarthritis in the subject or may be used for the diagnosis of osteoarthritis and may be used to develop a treatment plan or determine the effectiveness of a particular treatment.

In some embodiments, the methods disclosed herein further comprise measuring the level of at least one normalization peptide from a protein selected from TSP1, CNDP1, FA5, SHBG, PLF4, C1QC, ADIPO, APOA4, ACTG, CD14, K2C1, CBG, CHLE, FA11 or any combination thereof in a sample from a subject and normalizing the level of the biomarker in the sample from the subject and the reference level of the biomarker to the level of the normalization peptide in the sample and the reference prior to comparing the level of the biomarker in the sample to the reference level of the biomarker. In some embodiments, the normalization peptide comprises at least one of the sequences of Tables 12 or 13.

The early stages of osteoarthritis are characterized by pain in the affected joint and changes to the cartilage that lines the ends of the bones and cushions the joints. The changes may include thinning of the cartilage layer. Usually the pain in the early stages is well-managed with rest and topical or non-prescription ingestible anti-inflammatory pharmaceutical agents. Moderate osteoarthritis often involves pain with standing as well as when moving and may show bone thickening along joint margins. Treatment involves increased use of anti-inflammatory and anti-pain medications including NSAIDS and steroids. Progression to later stages may restrict the ability of the affected subject to move and to work and may require treatment with stronger pain medications such as opioids, injectable corticosteroids, lubrication injections, physical therapy and joint replacement surgery. Identifying patients likely to progress to a more severe form of the disease would aid medical professionals in determining the appropriate treatment options for individuals with osteoarthritis earlier in the disease course, before disability occurs and when the disease is more likely to be favorably impacted by treatment.

The methods of the present disclosure can also be used to assist in selecting appropriate courses of treatment and to identify patients that would benefit from a particular course of therapy. As shown in FIG. 1, if a subject is demonstrated, via evaluation of the biomarkers provided herein, to be likely to have non-progressive disease then the treatment plan on the left can be pursued which is non-invasive and includes non-pharmacologic therapy. If instead the subject is determined to be likely to have progressive osteoarthritis, then more aggressive treatment options can be pursued including injections or treatment with pharmaceuticals and surgery. Thus, the levels of the particular biomarkers described herein provide insight into which treatment regimens will be most effective for the subject. This information can be used to generate treatment plans for the subject to prolong an active, pain-free lifestyle and minimize side effects, adverse reactions or therapy related toxicity. Methods of developing a treatment plan for a subject with osteoarthritis are also provided herein. Treatment plans may be developed as shown in FIG. 1 using the predictions of the likelihood of progression of osteoarthritis. Methods of monitoring a treatment plan are also provided. The methods may be performed on a recurring basis in order to determine whether a particular treatment plan is effective in reducing and maintaining reduction of at least one symptom of osteoarthritis.

In some embodiments, the methods of the present disclosure may further include administering an anti-inflammatory or anti-pain therapeutic to the subject if the subject is diagnosed with osteoarthritis or predicted to have non-progressive or progressive disease. Suitable anti-inflammatory therapeutics are known to those skilled in the art and may include, without limitation, nonsteroidal anti-inflammatory drugs (NSAIDs), disease-modifying osteoarthritis drugs (DMOADs), disease-modifying antirheumatic drugs (DMARDs), corticosteroids, and hyaluronans. Several classes of DMARDs may be used in accordance with the present invention including, but not limited to, traditional DMARDs such as methotrexate, hydroxycholorquine, sulfasalazine, leflunomide, cyclophosphamide and azathioprine; biologics such as anti-IL-1 therapeutics, anti-TNF therapeutics, metalloproteinase inhibitors, p38 inhibitors, abatacept, adalimumab, anakinra, certolizumab pegol, etanercept, infliximab, golimumab and rituximab; and JAK inhibitors such as Tofacitinib. Suitable anti-pain therapeutics include, without limitation, non-opioid analgesics (e.g., acetaminophen), nonsteroidal anti-inflammatory drugs (NSAIDs), opioid analgesics, and co-analgesics and most likely in future, nerve growth factor inhibitors. Subjects having stable or non-progressive disease may be treated with topical or ingestible pain or anti-inflammatory medications. Subjects identified as having progressive destructive or aggressive disease likely to lead to joint destruction may be referred for injectable lubricant or biologic agent procedures, stronger pain medications such as opioids, bone-acting agents such as calcitonin, bisphosphonates and hormonal therapies, physical therapy, arthroscopic surgery, osteotomy, fibulectomy or joint replacement surgery.

Methods of treating osteoarthritis in a subject are provided. The methods of treating osteoarthritis may include administering a therapeutically effective amount of an anti-inflammatory or anti-pain therapeutic to the subject provided that the levels of at least one of the biomarkers listed in Table 2A in a sample from the subject was determined to be modified (increased or decreased) as compared to the reference level as shown in Table 2A or greater than/less than the threshold values reported in Table 2A, 2C or 2D to diagnose osteoarthritis or indicate the subject's disease is likely to progress.

In some embodiments, the age, gender and/or body mass index of the subject are also used in making the prediction of progression or diagnosis. In some embodiments described herein, diagnostic and prognostic performance of the biomarkers and/or other clinical parameters such as demographics including sex, age, BMI and cohort were assessed utilizing logistic regression to compute p-values and confidence intervals. These statistics were then used to calculate a Benjamini-Hochberg FDR threshold. A biomarker was considered a significant biomarker if the FDR passed 10%. Knee-level analysis required a paired evaluation and the generalized estimating equation method was used to account for the correlation structure and the significance of the biomarker was assessed by a Wald statistic. The statistical analysis used is described in the Examples section. Methods for assessing statistical significance are well known in the art and thus other methods may be used. In some aspects of the invention, a p-value of less than 0.05 constitutes statistical significance.

As used herein, the term “subject” and “patient” are used interchangeably and refer to both human and non-human animals. The term “non-human animals” as used in the disclosure includes all vertebrates, e.g., mammals and non-mammals, such as non-human primates, sheep, dog, cat, horse, cow, chickens, rodents, guinea pigs, amphibians, reptiles, and the like. Preferably and in some embodiments, the subject is a human patient. The subject may be a human patient with knee pain or already diagnosed with or suspected of having osteoarthritis.

The biomarkers of the present disclosure include proteins and genes encoding the proteins. The biomarkers analyzed are provided in Table 2B along with an indication of the commonly used abbreviations for each marker. Such biomarkers include the entire protein or peptide portions of the protein. As shown in the Examples, peptides from each of these proteins were identified as useful in the methods provided herein. The biomarker peptides used in the examples are shown in Table 2A. These biomarkers may be used alone in the methods or in combinations as described below.

Fragments and genetic variants of biomarkers are also encompassed by the present invention. “Fragment” is intended to include a portion of the amino acid sequence and hence a portion of the protein encoded thereby. A fragment or a biomarker peptide will generally encode at least 7, 8, 9, 10, 12, 15, 17, 20, 22, 25, 30 or more contiguous amino acids, or up to the total number of amino acids present in a full-length biomarker. “Variant” is intended to mean substantially similar sequences. Generally, variants of a particular biomarker of the invention will have at least about 80%, 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or more amino acid identity to that biomarker as determined by amino acid alignment programs.

A “biomarker” is a protein or glycan whose level in a sample is altered compared to that of a normal or healthy sample or is indicative of a condition. The biomarkers disclosed herein are proteins or glycans whose levels correlate with osteoarthritis and can be used to predict the progression of the disease as well as diagnose the disease.

In particular embodiments, the methods for predicting progression of or diagnosing osteoarthritis in a subject include collecting a patient body sample. The sample may or may not include cells. In particular, the methods described herein may be performed without requiring a tissue sample or biopsy. “Sample” is intended to include any sampling of cells, tissues, or bodily fluids in which a level of a biomarker can be detected. Examples of such samples include, but are not limited to, blood, serum, urine, synovial fluid, saliva, or any other bodily secretion or derivative thereof. Blood can include whole blood, plasma (citrate, EDTA, heparin), serum, or any derivative of blood. Samples may be obtained from a patient by a variety of techniques available to those skilled in the art. Methods for collecting various samples are well known in the art. In some embodiments, the sample is serum, plasma, urine, or synovial fluid. In some embodiments, the sample is serum depleted of at least 7 major serum proteins. In some embodiments, the serum proteins depleted are selected from the group consisting of albumin, IgG, IgA, transferrin, haptoglobin, anti-trypsin, fibrinogen, alpha 2-macroglobulin, IgM, apolipoprotein AI, apolipoprotein AII, complement C3, and transthyretin.

Any methods available in the art for detecting the level of biomarkers are encompassed herein. The level of a biomarker of the invention can be detected using a peptide corresponding to the biomarker. “Measuring an expression level of” is intended to mean determining the quantity or presence of a biomarker (i.e., peptide) in a sample for at least one of the biomarkers of Table 2a. Thus, “measuring an expression level of” encompasses instances where a biomarker is determined not to be detectable due to failure to be produced, or due to production below the detection limit of the assay; “measuring an expression level of” also encompasses low, normal and high levels of detection. Measuring an expression level also includes instances where a marker is degraded or is more stable in a person with osteoarthritis or with progressive disease and is not limited to production of new peptide or the timing of peptide production.

Methods suitable for “measuring an expression level of” biomarkers are known to those of skill in the art and include, but are not limited to, ELISA, immunofluorescence, FACS analysis, Western blot, magnetic immunoassays, mass spectroscopy, gel electrophoresis, antibody-based microarrays and non-antibody-based microarrays or combinations of these methods. In the past, the gold standard for detection of growth factors and cytokines in blood was the use of ELISAs; however, multiplex technology and mass spectroscopy offer attractive alternative approaches for protein-based analyses. The advantages of multiplex technology compared to traditional ELISA assays are conservation of patient sample, and significant savings in cost, time and labor. In some embodiments, the biomarker is measured using an antibody-based capture method. In some embodiments, the biomarker is measured using mass spectrometry.

Several multiplex platforms currently exist. The Luminex bead-based systems are the most established, being used to detect circulating cytokines and growth factors in both mice and humans. This method is based on the use of microparticles that have been pre-coated with specific antibodies. These particles are then mixed with sample and the captured analytes are detected using specific secondary antibodies. This allows for up to 100 different analytes to be measured simultaneously in a single microplate well. The advantages of this flow cytometry-based method compared to traditional ELISA assays are in the conservation of patient samples as well as significant savings in terms of cost and labor. An alternative, plate-based system is produced by Meso Scale Discovery (MSD). This system utilizes its proprietary Multi-Array® and Multi-Spot® microplates with electrodes directly integrated into the plates. This enables the MSD system to have ultra-sensitive detection limits, high specificity, large dynamic range, and low background signal. Another plate-based multiplex system is the SearchLight Plus CCD Imaging System produced by Aushon Biosystems. This novel multiplexing technology allows for the measurement of up to 16 different analytes simultaneously in a single microplate well. The assay design is similar to a sandwich ELISA where the capture antibodies are pre-spotted into individual wells of a 96-well plate. Samples or standards are added which bind to the specific capture antibodies and are detected using Aushon's patented SuperSignal ELISA Femto Chemiluminescent Substrate. Still another method is SomaLogic which is a bead-based technology for multiplex quantification of proteins or protein fragments.

The term “probe” refers to any molecule that is capable of selectively binding to a specifically intended target biomolecule, for example, a protein encoded by or corresponding to a biomarker. Probes can be synthesized by one of skill in the art, or derived from appropriate biological preparations. Probes may be specifically designed to be labeled. Examples of molecules that can be utilized as probes include, but are not limited to, RNA, DNA, proteins, antibodies, and organic molecules. Detectable labels include, but are not limited to, any heterologous addition to the probe that can be used to detect the selectively bound probe. Examples of detectable labels include fluorescent and radiological labels as well as labels that can be detected because they have a specific binding partner (ligand/receptor interaction) such as biotin/avidin or a nucleic acid tag that may act as a barcode to bind and specifically identify a larger nucleic acid.

As used herein the term “predicting progression” or “a prediction of progression” refers to providing a probability-based analysis of risk for osteoarthritis progression in a particular subject. The prediction of progression of osteoarthritis is not a guarantee or absolute, only a statistically probable indication of the disease state of the subject. The term prediction of a “diagnosis” or “diagnosing” of osteoarthritis refers to providing a probability-based analysis of an osteoarthritis diagnosis in a particular subject. The prediction of a diagnosis of osteoarthritis is not a guarantee or absolute, only a statistically probable indication of the disease state of the subject.

The level of the biomarker in the sample from the subject is compared to a reference level of the biomarker. The reference level may be determined empirically such as illustrated in the Examples, by comparison to the levels found in a set of samples from subjects with known clinical outcomes or known to have or not have osteoarthritis. Alternatively, the reference level may be a level of the biomarker found in samples, such as serum samples, which becomes a standard and can be used as a predictor for new samples. The level of the biomarker in the sample from the subject may be increased or decreased (i.e., “altered”) as compared to the reference level. The Examples and Tables provide information regarding how each biomarker is altered to indicate a diagnosis or to predict progression.

The predictive methods described herein may be combined to provide increased significance of the results, i.e. increased AUCs. For example, the levels of multiple markers may be determined in a sample from the subject and the results may have additional statistical or predictive power via the combination. The levels may be compared to the reference levels and a diagnosis or a prediction of risk of progression made. Several exemplary combinations are provided below and in the Examples, but any combination of 2, 3, 4, 5, 6, 7, 8, 9, 10 or more of the biomarkers may provide a prediction with increased accuracy and thus be beneficial.

Diagnosis

In some embodiments, the invention relates to a method of diagnosing osteoarthritis comprising measuring an expression level of at least one biomarker selected from the group consisting of CRAC1 (CRTAC1), A2AP, A1BG, A2GL, AACT, ACTG, AMBP, APOB, APOE, B2MG, C1QC, C1R, C1RL, C4BPA, C4BPB, CD14, CD44, CERU, CFAB, CFAH, CFAI, CILP1, C1S, CNDP1, CO2, CO4B, CO5, CO6A3, CO8B, CO8G, CO9, coll3, COMP, CTX1a, CTX1b, CTX2, CTXi, CXCL7, ECM1, FA12, FA5, FBLN1, FBLN3, FCGBP, FCN3, FETUA, FINC, GELS, HA, HABP2, haptoglobin, HEMO, HEP2, HGFA, HRG, hyaluronan, IC1, ITIH1, ITIH4, KNG1, LAMA2, LUM, LYAM1, MASP1, PCOC1, PGCA, PHLD, PLF4, PLMN, PRG4, RET4, SAMP, SHBG, TENX, TETN, THBG, TIMP1, TSP1, TSP4, VTDB, VTNC, ZA2G, ZPI, or any combination thereof in a sample from a subject; comparing the level of at least one biomarker in the sample to a reference level of the biomarker; and diagnosing the subject with osteoarthritis if the level of at least one biomarker is altered as compared to the reference level. In some embodiments, such methods further comprise developing a treatment plan for the subject if the subject is diagnosed with osteoarthritis.

In some embodiments, the biomarker is a combination of at least two of CRAC1, COMP, CO6A3, SHBG, PCOC1, CO8G, LUM, ACTG, CO5, A2AP, CO2, FA5, CERU, KNG1, HPLN1, CD14, CERU, CTX1a, CTX1b, VTNC, ZPI and haptoglobin and the diagnosis of osteoarthritis includes the presence of an osteophyte or a bone anabolic response. A combination of all the listed biomarkers or only two, three, four, five, six, seven, eight, nine, ten or more may also be used. In some embodiments, the biomarker is a combination of at least one of CRAC1, CXCL7, C4BPA, COMP, LUM, CO5, TIMP1, C4BPA, PCOC1, A2AP, CO2, FA5, HRG, CO6A3, VTDB, KNG1, HPLN1, CD14, CERU, CTX1a, CTX1b, PLF4, TETN, TSP1, PHLD, C4BPB, CFAI, SAMP, CO8B, ECM1, TSP4, CILP, APOE, IHA, CTX2, CTXi, hyaluronan and haptoglobin and the diagnosis of osteoarthritis includes the presence of worsening joint space narrowing indicative of cartilage or meniscal abnormality. A combination of all the listed biomarkers or only two, three, four, five, six, seven, eight, nine, ten or more may also be used. In some embodiments, the biomarker is a combination of at least one of CRAC1, SHBG, COMP, CO8G, PCOC1, CO6A3, LUM, CO5, A2AP, CO2, FA5, FBLN3, VTDB, KNG1, HPLN1, CD14, CERU, CTX1a, CTX1b, CTX2, CTXi, CFAH, TSP4 and haptoglobin and the diagnosis of osteoarthritis includes the progression of the Kellgren-Lawrence (KL) grade where progression is defined by a joint KL grade/score increasing to indicate a higher, i.e. worse, grade. A combination of all the listed biomarkers or only two, three, four, five, six, seven, eight, nine, ten or more may also be used. In some embodiments, the biomarker includes CRAC1, KNG1 and/or haptoglobin for diagnosis of osteoarthritis.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CRAC1 (CRTAC1), CXCL7, CO8G, ACTG, CD44, CERU, CFAH, CFAI, CO6A3, CO8G, COMP, FINC, HRG, KNG1, PLF4, PRG4, SAMP, TSP4, and any combination thereof. A combination of all the listed biomarkers or only two, three, four, five, six, seven, eight, nine, ten or more may also be used. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1) and CXCL7. In some embodiments, the biomarkers measured comprise CXCL7 and CO8G. In some embodiments, the biomarkers measured comprise CO8G and CRAC1. In still further embodiments, the biomarkers measured comprise CRAC1 (CRTAC1), CXCL7, and CO8G.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CRAC1 (CRTAC1), A2AP, ACTG, APOE, C1QC, C4BPB, CD14, CFAI, CO5, CO6A3, CO8G, coll3, CXCL7, FA5, FCGBP, FINC, GELS, HA, HEMO, KNG1, PCOC1, TENX, VTDB, or any combination thereof. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CRAC1 (CRTAC1), A2AP, ACTG, APOE, C1QC, C4BPB, CFAI, CO5, CO6A3, CO8G, CXCL7, FA5, FCGBP, FINC, GELS, PCOC1, TENX, or any combination thereof. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1), A2AP, ACTG, APOE, C1QC, C4BPB, CFAI, CO5, CO6A3, CO8G, CXCL7, FA5, FCGBP, FINC, GELS, PCOC1, and TENX. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CD14, coll3, HA, HEMO, KNG1, VTDB or any combination thereof. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CRAC1 (CRTAC1), COMP, CO6A3, CO8G, CXCL7, or any combination thereof. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1), COMP, CO6A3, CO8G, and CXCL7.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CRAC1 (CRTAC1), CXCL7, C4BPA, COMP, LUM, CO5, TIMP1, or any combination thereof. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1) and CXCL7. In further such embodiments, the biomarkers measured further comprise C4BPA and/or COMP. In still further such embodiments, the biomarkers measured further comprise LUM, CO5, and/or TIMP1.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CRAC1 (CRTAC1), SHBG, COMP, CO8G, PCOC1, CO6A3, LUM, or any combination thereof. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1) and SHBG. In further such embodiments, the biomarkers measured further comprise COMP and/or CO8G. In still further such embodiments, the biomarkers measured further comprise PCOC1, CO6A3, and/or LUM. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1) and COMP. In further such embodiments, the biomarkers measured further comprise CO6A3 and/or SHBG.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CRAC1 (CRTAC1), C4BPA, LUM, CO5, PCOC1, CXCL7, COMP, or any combination thereof. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1) and CXCL7. In further such embodiments, the biomarkers measured further comprise C4BPA and/or LUM. In still further such embodiments, the biomarkers measured further comprise CO5, PCOC1, and/or COMP.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CRAC1 (CRTAC1), PCOC1, CO8G, LUM, COMP, CO6A3, CO5, ACTG or any combination thereof. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1) and PCOC1. In further such embodiments, the biomarkers measured further comprise COMP and/or CO8G. In still further such embodiments, the biomarkers measured further comprise LUM, CO6A3, and/or CO5. In some embodiments, the biomarkers measured comprise CRAC1 (CRTAC1) and COMP. In further such embodiments, the biomarkers measured further comprise CO6A3 and/or PCOC1. In still further such embodiments, the biomarkers measured comprise CO8G, ACTG, and/or CO5.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of A2AP, CO2, COMP, FA5, CO5, CRAC1 (CRTAC1), SHBG, or any combination thereof. In some embodiments, the biomarkers measured comprise A2AP and CO2. In further such embodiments, the biomarkers measured further comprise COMP and/or FA5. In still further such embodiments, the biomarkers measured further comprise CO5, CRAC1 (CRTAC1), and/or SHBG.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of A2AP, FA5, CO2, CO5, COMP, SHBG, CO6A3, or any combination thereof. In some embodiments, the biomarkers measured comprise A2AP and FA5. In further such embodiments, the biomarkers measured further comprise CO5 and/or CO2. In still further such embodiments, the biomarkers measured further comprise COMP, SHBG, and/or CO6A3. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of A2AP, CRAC1 (CRTAC1), CO2, COMP, FA5, CO5, or any combination thereof. In some embodiments, the biomarkers measured comprise A2AP and CO2. In further such embodiments, the biomarkers measured further comprise COMP and/or CRAC1 (CRTAC1). In some embodiments, the biomarkers measured comprise A2AP, CRAC1 (CRTAC1), and CO2. In further such embodiments, the biomarkers measured further comprise FA5. In still further such embodiments, the biomarkers measured further comprise COMP and/or CO5.

In some embodiments, the biomarkers measured comprise a peptide sequence listed in Table 2A or 2C. In some embodiments, the subject is diagnosed with osteoarthritis if the level of at least one biomarker is altered as shown in Table 2A or 2C.

Prognosis

In some embodiments, the invention relates to a method of predicting progression of osteoarthritis comprising measuring an expression level of at least one biomarker selected from the group consisting of A1BG, A2AP, A2GL, AACT, ACTG, AFAM, ANT3, APOB, APOH, B2MG, C1QC, C1R, C1RL, C4BPA, C4BPB, CD14, CD163, CD44, CERU, CFAB, CFAH, CFAI, C1S, CO2, CO4B, CO5, CO6A3, CO8B, coll3, CRAC1 (CRTAC1), CTX2, CXCL7, DOPO, ECM1, FA5, FA12, FBLN1, FCGBP, FCN3, FETUA, FINC, GELS, HABP2, haptoglobin, HEMO, HEP2, HGFA, HRG, hyaluronan, ITIH4, KLKB1, KNG1, LUM, LYAM1, PGCA, PHLD, PLF4, PLMN, PRG4, RET4, SAMP, TENX, TETN, THBG, THRB, TIMP1, TSP1, TSP4, VTDB, VTNC, or combinations thereof in a sample from a subject; comparing the level of the biomarker in the sample to a reference level of the biomarker; and predicting the progression of the osteoarthritis, wherein altered levels of any of the biomarkers as compared to the reference level is indicative of progression of the osteoarthritis. In some embodiments, such methods further comprise developing a treatment plan for the subject based on the prediction of progression of the osteoarthritis.

In some embodiments, the biomarker is a combination of at least two of PLF4, CXCL7, ANT3, AACT, THRB, ITIH4, CO8B, PLMN, PRG4, C4BPA, C4BPB, A2AP, LYAM1, CO8G, KLKB1, hyaluronan and haptoglobin and the prediction of osteoarthritis progression includes osteophyte growth or a bone anabolic response. A combination of all the listed biomarkers or only two, three, four, five, six, seven, eight, nine, ten or more may also be used. In some embodiments, the biomarker is a combination of at least two of PGCA, APOH, SAMP, AACT, CFAH, PHLD, TSP1, THRB, HRG, CO4B, FCN3, CD44, TSP4, TETN, FINC, ECM1, HEMO, CD163, CERU, TIMP1, A1BG, THBG, A2GL, FBLN1, CO5, B2MG, FETUA and haptoglobin and the prediction of osteoarthritis progression includes worsening joint space narrowing. A combination of all the listed biomarkers or only two, three, four, five, six, seven, eight, nine, ten or more may also be used. In some embodiments, the biomarker is a combination of at least one of CFAH, SAMP, TSP1, HEP2, C1R, APOB, FINC, PGCA, AACT, KNG1, A2AP, CO6A3, HGFA, CO2, PRG4, DOPO, CD44, CERU, VTDB, and haptoglobin and the prediction of osteoarthritis progression includes progression by the Kellgren-Lawrence (KL) grading scale (progression is defined as joint KL score increasing to a higher or worse grade). A combination of all the listed biomarkers or only two, three, four, five, six, seven, eight, nine, ten or more may also be used. In some embodiments, the biomarkers include CO8B, haptoglobin and/or PLF4 and the prediction relates to the progression of osteoarthritis.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of ACTG, ANT3, CD44, CERU, CFAH, CFAI, CO8B, CXCL7, CO6A3, CRAC1 (CRTAC1), FINC, haptoglobin, HRG, KLKB1, PLF4, PRG4, SAMP, TSP4 and any combination thereof. A combination of all the listed biomarkers or only two, three, four, five, six, seven, eight, nine, ten or more may also be used. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of PLF4, CFAH, and ANT3. In still further embodiments, the biomarkers measured comprise PLF4, CFAH, and ANT3. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of ANT3, CD14, CD163, CD44, CERU, CFAH, CO8B, coll3, CRAC1 (CRTAC1), CTX2, CXCL7, haptoglobin, HEMO, HRG, KLKB1, LYAM1, VTDB, or any combination thereof. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of ANT3, CFAH, CO8B, CRAC1 (CRTAC1), CXCL7, HRG, KLKB1, LYAM1, or any combination thereof. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CFAH, CO8B, CRAC1 (CRTAC1), HRG, or any combination thereof. In some embodiments, the biomarkers measured comprise CFAH, CO8B, CRAC1 (CRTAC1), and HRG. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of ANT3, CO8B, CXCL7, KLKB1, LYAM1, or any combination thereof. In some embodiments, the biomarkers measured comprise ANT3, CO8B, CXCL7, KLKB1, and LYAM1. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from CD14, CD163, CD44, CERU, coll3, CTX2, haptoglobin, HEMO, VTDB or any combination thereof. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of ACTG, ANT3, CD44, CERU, CFAH, CFAI, CO8B, FINC, HRG, KLKB1, PLF4, PRG4, SAMP, TSP4 or any combination thereof. In some embodiments, the biomarkers measured comprise ACTG, ANT3, CD44, CERU, CFAH, CFAI, CO8B, FINC, HRG, KLKB1, PLF4, PRG4, SAMP, and TSP4.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of ANT3, CERU, CFAH, CO8B, FINC, HRG, PLF4, PRG4, SAMP, TSP4 or any combination thereof. In some embodiments, the biomarkers measured comprise ANT3, CERU, CFAH, CO8B, FINC, HRG, PLF4, PRG4, SAMP, and TSP4. In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of ACTG, ANT3, CD44, CFAI, CO8B, KLKB1, PLF4, or any combination thereof. In some embodiments, the biomarkers measured further comprise ACTG, ANT3, CD44, CFAI, CO8B, KLKB1, and PLF4.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of PGCA, APOH, SAMP, AACT, CFAH, PHLD, TSP1, THRB, or any combination thereof. In some embodiments, the biomarkers measured comprise SAMP and AACT. In further such embodiments, the biomarkers measured further comprise PGCA and/or APOH. In still further such embodiments, the biomarkers measured further comprise CFAH, PHLD, TSP1, and/or THRB. The prognosis includes joint space narrowing.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CFAH, SAMP, TSP1, HEP2, C1R, APOB, FINC, HEP2, PGCA, or any combination thereof. In some embodiments, the biomarkers measured comprise CFAH and SAMP. In further such embodiments, the biomarkers measured further comprise TSP1 and/or APOB. In still further such embodiments, the biomarkers measured further comprise HEP2, C1R, FINC, HEP2, and/or PGCA. The prediction includes KL grade.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of PLF4, CXCL7, ANT3, AACT, THRB, ITIH4, CO8B, PLMN, or any combination thereof. In some embodiments, the biomarkers measured comprise PLF4 and AACT. In further such embodiments, the biomarkers measured further comprise CXCL7 and/or ANT3. In still further such embodiments, the biomarkers measured further comprise THRB, ITIH4, CO8B, and/or PLMN. The prediction includes osteophyte growth.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of TSP1, CFAH, THRB, HRG, APOH, AACT, PHLD, AACT, or any combination thereof. In some embodiments, the biomarkers measured comprise TSP1 and CFAH. In further such embodiments, the biomarkers measured further comprise THRB and/or APOH. In still further such embodiments, the biomarkers measured further comprise HRG, AACT, PHLD, and/or AACT. The prediction includes joint space narrowing.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CFAH, TSP1, SAMP, APOB, AACT, C1R, or any combination thereof. In some embodiments, the biomarkers measured comprise CFAH and SAMP. In further such embodiments, the biomarkers measured further comprise TSP1 and/or AACT. In still further such embodiments, the biomarkers measured further comprise APOB and/or C1R. The prediction includes the KL grade.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CO8B, PLF4, PRG4, ANT3, C4BPA, CXCL7, C4BPA, or any combination thereof. In some embodiments, the biomarkers measured comprise CO8B and PLF4. In further such embodiments, the biomarkers measured further comprise PRG4 and/or ANT3. In still further such embodiments, the biomarkers measured further comprise C4BPA, CXCL7, and/or C4BPA. The prediction may include osteophyte growth.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CO4B, FCN3, CO8B, FINC, PGCA, TSP4, TETN, or any combination thereof. In some embodiments, the biomarkers measured comprise CO8B and CO4B. In further such embodiments, the biomarkers measured further comprise FINC and/or FCN3. In still further such embodiments, the biomarkers measured further comprise PGCA, TSP4, and/or TETN. The prognosis may include joint space narrowing.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of A2AP, KNG1, HGFA, PRG4, AFAM, DOPO, FINC, CO8B, or any combination thereof. In some embodiments, the biomarkers measured comprise KNG1 and HGFA. In further such embodiments, the biomarkers measured further comprise A2AP and/or CO8B. In still further such embodiments, the biomarkers measured further comprise PRG4, AFAM, DOPO, and/or FINC. The prediction includes a KL grade determination.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of CXCL7, C4BPA, C4BPB, A2AP, ITIH4, PLMN, HRG, or any combination thereof. In some embodiments, the biomarkers measured comprise CXCL7 and C4BPA. In further such embodiments, the biomarkers measured further comprise C4BPB and/or A2AP. In still further such embodiments, the biomarkers measured further comprise ITIH4, PLMN, and/or HRG. The prediction includes osteophyte growth.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of PGCA, CO4B, TENX, FCN3, C4BPA, TSP1, CO8B, HRG, or any combination thereof. In some embodiments, the biomarkers measured comprise CO4B and PGCA. In further such embodiments, the biomarkers measured further comprise TENX and/or C4BPA. In still further such embodiments, the biomarkers measured further comprise FCN3, TSP1, CO8B, and/or HRG. The prognosis may include joint space narrowing.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of KNG1, HGFA, A2AP, FA5, TSP1, PGCA, TENX, FINC, or any combination thereof. In some embodiments, the biomarkers measured comprise KNG1 and HGFA. In further such embodiments, the biomarkers measured further comprise A2AP and/or PGCA. In still further such embodiments, the biomarkers measured further comprise FA5, TSP1, TENX, and/or FINC. The prediction includes KL grade prediction.

In some embodiments, the biomarkers measured comprise at least two biomarkers selected from the group consisting of C4BPA, C4BPB, CXCL7, LYAM1, A2AP, TSP1, FINC, or any combination thereof. In some embodiments, the biomarkers measured comprise C4BPA and CXCL7. In further such embodiments, the biomarkers measured further comprise C4BPB and/or LYAM1. In still further such embodiments, the biomarkers measured further comprise A2AP, TSP1, and/or FINC. The prediction includes osteophyte growth.

In some embodiments, the biomarkers measured comprise a peptide sequence listed in Table 2A, 2C or 2D. In some embodiments, a prediction of progression of the osteoarthritis in the subject is made if the level of at least one biomarker is altered as compared to the reference level as shown in Table 2A or 2D. In some embodiments, a diagnosis of osteoarthritis or a prediction of progression of the osteoarthritis in the subject is made if the level of at least one biomarker is greater or less than the threshold values shown in Table 2A or 2C. In some embodiments, the biomarkers comprise the sequences listed in the Tables included herein. The various embodiments described herein may be combined or used individually.

The present disclosure is not limited to the specific details of construction, arrangement of components, or method steps set forth herein. The compositions and methods disclosed herein are capable of being made, practiced, used, carried out and/or formed in various ways that will be apparent to one of skill in the art in light of the disclosure that follows. The phraseology and terminology used herein is for the purpose of description only and should not be regarded as limiting to the scope of the claims. Ordinal indicators, such as first, second, and third, as used in the description and the claims to refer to various structures or method steps, are not meant to be construed to indicate any specific structures or steps, or any particular order or configuration to such structures or steps. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., “such as”) provided herein, is intended merely to facilitate the disclosure and does not imply any limitation on the scope of the disclosure unless otherwise claimed. No language in the specification, and no structures shown in the drawings, should be construed as indicating that any non-claimed element is essential to the practice of the disclosed subject matter. The use herein of the terms “including,” “comprising,” or “having,” and variations thereof, is meant to encompass the elements listed thereafter and equivalents thereof, as well as additional elements. Embodiments recited as “including,” “comprising,” or “having” certain elements are also contemplated as “consisting essentially of” and “consisting of” those certain elements.

Recitation of ranges of values herein are merely intended to serve as a shorthand method of referring individually to each separate value falling within the range, unless otherwise indicated herein, and each separate value is incorporated into the specification as if it were individually recited herein. For example, if a concentration range is stated as 1% to 50%, it is intended that values such as 2% to 40%, 10% to 30%, or 1% to 3%, etc., are expressly enumerated in this specification. As another example, if it is stated that the biomarkers measured comprise “at least one (or any other number)” biomarker selected from a particular group it is intended that values such as “at least two,” “at least three,” “at least four,” etc. (up until the maximum allowed by the statement) are expressly enumerated in the specification. These are only examples of what is specifically intended, and all possible combinations of numerical values between and including the lowest value and the highest value enumerated are to be considered to be expressly stated in this disclosure. Use of the word “about” to describe a particular recited amount or range of amounts is meant to indicate that values very near to the recited amount are included in that amount, such as values that could or naturally would be accounted for due to manufacturing tolerances, instrument and human error in forming measurements, and the like. All percentages referring to amounts are by weight unless indicated otherwise.

No admission is made that any reference, including any non-patent or patent document cited in this specification, constitutes prior art. In particular, it will be understood that, unless otherwise stated, reference to any document herein does not constitute an admission that any of these documents forms part of the common general knowledge in the art in the United States or in any other country. Any discussion of the references states what their authors assert, and the applicant reserves the right to challenge the accuracy and pertinence of any of the documents cited herein. All references cited herein are fully incorporated by reference, unless explicitly indicated otherwise. The present disclosure shall control in the event there are any disparities between any definitions and/or description found in the cited references.

The following examples are meant only to be illustrative and are not meant as limitations on the scope of the invention or of the appended claims.

EXAMPLES

For this project, candidate prognostic and diagnostic biomarkers in non-depleted (normal serum) and depleted serum (serum treated to deplete one or more abundant serum proteins as described more fully below) were evaluated by mass spectrometry. We tested a multiple reaction monitoring (MRM) panel developed on the basis of three discovery proteomics experiments: in synovial fluid, urine and depleted serum. Below we describe the results for a cohort of 124; approximately two-thirds of the subjects were selected on the basis of knee osteoarthritis with either knee OA progression or stability (non-progression) and the remaining one-third of subjects were controls without knee osteoarthritis.

Subjects and Methods

Cohorts:

Subjects were selected from two cohorts, the Prediction of Osteoarthritis Progression (POP) cohort, and the Genetics of Generalized Osteoarthritis (GOGO) cohort. Kraus et al., Ann Rheum Dis 68(11):1673-9 (2009); Kraus et al., Osteoarthritis Cartilage 15(2):120-7 (2007). In brief, POP was a single site study at Duke with 159 subjects recruited on the basis of symptomatic radiographic knee OA of at least one knee; a total of 138 subjects (87%) returned for 3-year follow-up. Knee synovial fluid (from both knees when possible), serum (2 hour post-prandial) and urine (second morning void) were obtained at each evaluation. GOGO was a multi-site (7 sites) study with 2728 subjects recruited on the basis of two siblings with hand OA (defined as 3 joint radiographic KL≥2 grade distributed bilaterally). Knee, hip and spine radiographs were obtained. A total of 1329 subjects from 4 sites returned for follow-up at a mean 3.8 years (range 1.4-6.5 years). Serum (two hour post-prandial) and urine (second morning void or time recorded if other than this) were obtained at each evaluation. The demographics are shown in Table 1 for the patient subsets used for each aspect of biomarker discovery and validation.

Phenotypes

Phenotypes were based on 2 features from knee radiographs (joint space narrowing (JSN) and osteophyte (OST)) evaluated at 2 levels (knee based and person-based). JSN, indicative of cartilage and/or meniscal extrusion loss (cartilage and/or meniscal catabolism), and osteophyte, indicative of bone formation at the margins of the joint (joint tissue anabolism), were graded on a scale 0-3 using a standardized atlas with 0 being normal and 1, 2, and 3 representing increasing severity. Altman R D and Gold G E, Osteoarthritis Cartilage 15 Suppl A:A1-56 (2007). The medial and lateral compartments of the knee were graded for JSN (yielding a maximum score of 6 per knee joint); the four margins of the knee were graded for OST (yielding a maximum score of 12 per knee joint). Progression was defined as a one unit change over time in these variables at a knee level or person level. A Diagnosis of OA was defined as any OA represented by a score of greater than or equal to 1 for these variables at a knee level or person level. In addition, a half-century old definition of OA, the Kellgren Lawrence grade, scored on a 0-4 scale, was also evaluated for a one unit change (defining progression) or any OA (defined as KL≥2). Kellgren J H and Lawrence J S, Ann Rheum Dis 16(4):494-502 (1957). Controls for each phenotype were defined as a knee or person that did not meet the minimal definition.

TABLE 1
Patient demographics for sample sets used in experiments.
C = control	Number for JSN	Number for OST	Mean Age,		Mean BMI,
NP = non-progressor	Person knee	Person knee	SD (range)	Gender	SD (range)
P = progressor	C/NP/P	C/NP/P	years	% female	kg/m2
SF Proteomics	1/12/10	3/5/15	67 ± 12	74%	30.4 ± 5.8
(N = 23)	9/16/21	9/16/21	(43-80)		(23-47)
Urine	14/16/15	16/9/20	62 ± 12	76%	29.3 ± 6.6
Proteomics	32/36/22	37/23/30	(41 to 81)		(18.6-50.0)
(N = 45)
Urine ELISA	47/32/33	40/24/48	65 ± 10	59%	28.7 ± 6.6
(N = 118)	109/65/52	94/66/66	(41-86)		(18.6-61.7)
Serum MRM	4/16/20	3/11/26	63 ± 12	78%	30.0 ± 6.1
Validation	13/38/29	10/30/40	(42-80)		(20-50)
(N = 40)
Serum MRM	50/34/38	41/37/54	64 ± 10	82%	27.6 ± 6.1
and ELISA	116/69/61	98/72/76	(41-86)		(18.6-61.7)
(N = 124)
C = control'
NP = non-progressor;
P = progressor;
SF = synovial fluid;
MRM = multiple reaction monitoring

Statistical Methods

Two classes of methods were used, inferential methods and predictive methods, depending on the structure of the data. For person-level analysis, logistic regression was used to compute p-values and confidence intervals. Covariates included age, sex, BMI, and cohort. The effect of a biomarker was added to a model containing these covariates and a likelihood-ratio test was used to assess the significance of the biomarker after accounting for the covariates. A biomarker was considered significant if it surpassed a Benjamini-Hochberg FDR threshold of 10%. We defined person level phenotypes as follows: a person-level control as both knees normal; a person level osteoarthritis diagnosis as at least one knee with osteoarthritis; a person level knee osteoarthritis progressor as having at least one knee that has progressive osteoarthritis over time; and a person level non-progressor as having neither knee progressing over time. For a knee-level analysis the dependence arising from paired observations must be considered. We used the generalized estimating equation (GEE) method to account for the correlation structure. A biomarker was added to a model containing base covariates and its significance was assessed by a Wald statistic.

We evaluated the capability of the biomarkers described herein to predict or diagnose osteoarthritis based on its separate features consisting of: cartilage and/or meniscal extrusion or loss (reflected in radiographic joint space narrowing), an anabolic repair response (reflected in radiographic osteophyte presence and growth), and the Kellgren-Lawrence grade of disease (reflecting both osteophyte and joint space narrowing).

Predictive models were used to assess discrimination through the AUC. We used feature selection coupled with ridge regression, a form of penalized regression, for all models implemented in the glmnet R package. Penalized regression is often used for predictive models to constrain the size of coefficients to lessen the effects of overfitting the data. Feature selection consisted of selecting the 8 markers with the lowest p-values, which is a simple but effective method for the numbers of peptides in the current data set. Leave-one-out cross-validation was used in which selecting tuning parameters and carrying out feature selection was repeated at each iteration of cross-validation to mimic the process of fitting a model to new data. In sum, all multimarker AUCs have been properly cross-validated. For knee-level (paired) analysis, the leave-one-out cross-validation scheme was modified to a leave-sample-out scheme so that the test set was independent of the training set. Of note, an AUC 0.80≈sensitivity and specificity of 75% (depending on clinical context) and odds ratio 9-10. Qoronfleh et al., Drug Discovery World Winter:19-28 (2011).

The mean and SD values of ELISA results are natural log values for two of the serum markers that had skewed distributions (Hyaluronan and Ceruloplasmin), and all the urine biomarkers. None of the other serum markers were log transformed. Some values are negative because they are natural log transformations of values less than 1. In particular, all of the urine values were normalized prior to the analysis. Two separate ways of normalizing the urine values were tested: 1) by dividing the urine biomarker level by the creatinine value and taking the natural log [urine 1 variation], and 2) by dividing the urine biomarker level by the creatinine level and the cystatin level and taking the natural log [urine 2 variation]. In this case it is possible for some values to be negative because the natural log of a fraction is <0.

Of note, for the MRM analyses, the same amount of heavy labeled peptide was spiked into depleted and non-depleted serum; in retrospect, approximately one third this amount would have been more appropriate for non-depleted serum as it would have more closely approximated amounts of the biomarkers targeted for quantification in the non-depleted serum.

Results

1. Technical Validation

To assess analytical variation across all of the targeted analytes, a cost effective strategy for quality control (QC) was adopted a priori for this project. This consisted of creating a representative quality control sample by pooling equal portions of all patient samples within a particular matrix (i.e. synovial fluid, urine or serum). This approach allows for any matrix-specific interferences to be included in the technical variation calculations.

The analytical measurements were made from this QC sample many times throughout the analysis of the cohort samples. This allowed the measurement of the reproducibility of the quantification for the exact species of interest at the concentration levels where they are found in the sample. The closer an absolute signal is to its limit of detection or lower limit of quantification, the lower the expected reproducibility (or greater the expected variability). For this study we defined the acceptable upper limit of variability of repeated measurements of any analyte within a sample to be 30% relative to the standard deviation. This limit was based on our expectation that any real biological or treatment group dependent variation would achieve this level of variability or higher; this threshold is consistent with a range of coefficients of variation (CVs 20-35%) for proteomics studies deemed acceptable in a recent summary of a workshop held at the National Institutes of Health with representatives from the multiple communities developing and employing targeted mass spectrometry assays. Carr et al., Mol Cell Proteomics 13(3):907-17 (2014). FIG. 2 (above) plots CVs for each of the 147 targeted analytes representing 99 proteins (as a ratio of the heavy to light peptides) across all analyses of the QC sample; of note this does not include the 6 Alcohol Dehydrogenase control peptides. The blue dots represent peptides accepted into the final MRM panel of 146 analytes; the red dots represent the peptides rejected from the final dataset for analysis. As expected, the CVs (blue dots) tend to increase at lower ratios because their quantity approaches the lower limit of quantification for these analytes. The red dots show no correlation to ratio, as their quantification is likely confounded by matrix interferences.

2. Proteomics Results

For this project, candidate prognostic and diagnostic biomarkers were identified by three means: through discovery proteomics experiments in synovial fluid and urine by liquid chromatography mass spectrometry (LC/MS); and a discovery proteomics experiment in depleted serum (serum depleted of the 14 most abundant serum proteins. Based on this work, we selected a potential list of 155 peptides corresponding to 110 proteins for MRM assay development. The MRM assay was evaluated in synovial fluid (pooled samples from 3 progressor and 3 non-progressor knee OA patients) and test sera (3 progressor and 3 non-progressor knee OA patient samples for both depleted and non-depleted serum). Of the original 155 peptides, a total of 146 peptides from 99 proteins were detectable in these test samples and selected for final validation in depleted and non-depleted serum of 124 patients by MRM. Additionally, ELISA based analyses were used to evaluate some prognostic and diagnostic biomarker candidates when commercial ELISA kits were available for a biomarker of interest.

As described above, a final total of 146 peptides (99 proteins) were evaluated in this study by mass spectrometry. Below are listed the results from serum proteomic analysis of the non-depleted serum and the MARS14 depleted serum. MARS14 depleted serum is serum after it has been subjected to a column absorption to remove 14 of the most abundant serum proteins such that the abundance of these proteins does not interfere with the measurement and analysis of other serum proteins. The 14 depleted proteins include the following: albumin, IgG, IgA, transferrin, haptoglobin, anti-trypsin, fibrinogen, alpha 2-macroglobulin, IgM, apolipoprotein AI, apolipoprotein AII, complement C3, and transthyretin. To date, more extensive statistical analyses have been performed on the non-depleted serum because results were stronger than with the depleted serum; in general results are consistent between the two. We also experimented with the use of a normalization peptide (actin) in one case (diagnosis of knee level Osteoarthritis) and it showed a slight improvement in AUCs. Further normalization peptides are reported below. A total of 19 of 97 of the panel of OA proteins were predicted to be involved in the pathogenesis of OA and might have the potential to be ‘direct biomarkers’ of OA; 1 of these 19 was associated with development of a connective tissue disorder; and 16 of the 19 were linked to the process of post-translational modification, protein degradation and synthesis in OA cartilage. A summary of the markers and corresponding peptides for the proteomic results discussed below is shown in Tables 2A, 2B, 2C and 2D.

TABLE 2A
Summary of Markers and Corresponding Peptides;
lower case marker names in peptide sequence
column indicate markers that were detected by
ELISA. All markers listed were found to have a
statistically significant association (p-
value < 0.05) with osteoarthritis diagnosis,
prediction of osteoarthritis progression, or
both.
				Diagnos	Prognos
			SEQ	(/ug	(/ug
	Acces-		ID	total	total
	sion	Peptide	NO:	pro-	pro-
Marker	No.	Sequence(s)	X	tein)	tein)
CRAC1	Q9NQ79	GVASLFAGR	1	>	>
(CRTAC1)		SSPYYALR	2	>	>
A1BG	P04217	IFFHLNAVALG	3	>
		DGGHYTCR
A2AP	P08697	SPPGVCSR	4	>
		LCQDLGPGAFR	5	<
A2GL	P02750	VAAGAFQGLR	6	>
		ALGHLDLSGNR	7	>
AACT	P01011	ADLSGITGAR	8		>
		NLAVSQVVHK	9		>
		EQLSLLDR	10		>
ACTG	P63261	VAPEEHPVLLT	11	<
		EA PLNPK
AFAM	P43652	VNCLQTR	12
AMBP	P02760	AFIQLWAFDAV	13
		K
ANT3	P01008	ATEDEGSEQK	14		<
APOB	P04114	LAIPEGK	15	<	>
		WNFYYSPQSSP	16
		DK
APOE	P02649	LQAEAFQAR	17	>
APOH	P02749	ATFGCHDGYSL	18		>
		DGPEEIECTK
B2MG	P61769	VEHSDLSFSK	19
C1QC	P02747	VVTFCGHTSK	20	<,>
C1R	P00736	NIGEFCGK	21
		GLTLHLK	22		>
		GYGFYTK	23		>
C1RL	Q9NZP8	GSEAINAPGDN	24
		PAK
C4BPA	P04003	LSLEIEQLELQ	25	>
		R
		GVGWSHPLPQC	26	>	>
		EIVK
C4BPB	P20851	SQCLEDHTWAP	27	>
		PF
		PICK
CD14		cd14
CD163		cd163
CD44	P16070	YGFIEGHVVIP	28	>
		R
		cd44
CERU	P00450	HYYIAAEEIIW	29	<
		NY
		APSGIDIFTK
		cerulo
		EYTDASFTNR	30
		DIASGLIGPLI	31
		ICK
		GAYPLSIEPIG	32
		VR
		EVGPTNADPVC	33
		LAK
CFAB	P00751	QLNEINYEDHK	34
CFAH	P08603	CLPVTAPENGK	35		>
CFAI	P05156	HGNTDSEGIVE	36	>	>
		VK
		AQLGDLPWQVA	37	>
		IK
CILP1	O75339	IVGPLEVNVR	38
C1S	P09871	LLEVPEGR	39		>
CNDP1	Q96KN2	ALEQDLPVNIK	40
CO2	P06681	SSGQWQTPGAT	41	<
		R
		DGNDHSLWR	42		>
CO4B	P0C0L5	LVNGQSHISLS	43
		K
CO5	P01031	GIYGTISR	44	>	>
		TLLPVSKPEIR	45	>	>
		IIHFGTR	46	>
		FSYSSGHVHLS	47
		SENK
		SYFPESWLWEV	48
		HL VPR
CO6A3	P12111	EVQVFEITENS	49	>
		AK
		LLPSFVSSENA	50
		FYLS PDIR
CO8B	P07358	GILNEIK	51	>	>
CO8G	P07360	QLYGDTGVLGR	52	>	>
CO9	P02748	FTPTETNK	53
coll3		coll3
COMP	P49747	NALWHTGDTES	54	>
		QVR
		SSTGPGEQLR	55	>
		SNPDQADVDHD	56
		FVGDAC
		DSDQDQDGDGH
		QDSR
CTX1a		ctx1a_ctx1b
CTX1b		ctx1a_ctx1b
CTX2		ctx2
CTXi		ctxi
CXCL7	P02775	NIQSLEVIGK	57	>	>
DOPO	P09172	VISTLEEPTPQ	58
		CPT SQGR
ECM1	Q16610	FCEAEFSVK	59	>
FA12	P00748	CLEVEGHR	60
FA5	P12259	SEAYNTFSER	61	>
		EFNPLVIVGLS	62	>
		K
FBLN1	P23142	TGYYFDGISR	63
FBLN3	Q12805	NPCQDPYILTP	64	>
		ENR
		ADQVCINLR	65	>
FCGBP	Q9Y6R7	VTASSPVAVLS	66
		GH SCAQK
FCN3	O75636	TFAHYATFR	67
FETUA	P02765	HTLNQIDEVK	68
		FSVVYAK	69		>
FINC	P02751	EYLGAICSCTC	70		>
		F GGQR
		IGDTWSK	71		>
GELS	P06396	GGVASGFK	72	>
HA		HA
HABP2	Q14520	FCEIGSDDCYV	73
		G DGYSYR
		GQCLITQSPPY	74	>
		YR
hapto-		Hapto
globin
HEMO	P02790	QGHNSVFLIK	75
		hemopexin
HEP2	P05546	NFGYTLR	76		>
		FTVDRPFLFLI	77		>
		YEHR
HGFA	Q04756	YIPYTLYSVFN	78
		PS DHDLVLIR
HRG	P04196	DSPVLIDFFED	79	>
		TER
		GGEGTGYFVDF	80
		SVR
		YWNDCEPPDSR	81		>
		GEVLPLPEANF	82
		PS FPLPHHK
		SSTTKPPFKPH	83
		GSR
hyaluronan		hyaluronan
IC1	P05155	LVLLNAIYLSA	84
		K
ITIH1	P19827	VTFQLTYEEVL	85
		K
ITIH4	Q14624	FKPTLSQQQK	86		>
KLKB1	P03952	VSEGNHDIALI	87		>
		K
KNG1	P01042	LDDDLEHQGGH	88	>
		VLDHGHK
		kinno
LAMA2	P24043	TPYNILSSPDY	89
		VGVTK
LUM	P51884	ILGPLSYSK	90	>
		VANEVTLN	91	>
		SLEDLQLTHNK	92
LYAM1	P14151	AEIEYLEK	93
MASP1	P48740	TGVITSPDFPN	94
		PYPK
PCOC1	Q15113	TGGLDLPSPPT	95	>
		GASLK
PGCA	P16112	VSLPNYPAIPS	96	<	>
		D ATLEVQSLR
		EVVLLVATEGR	97
PHLD	P80108	FGSSLITVR	98	>	>
PLF4	P02776	ICLDLQAPLYK	99	>	>
PLMN	P00747	HSIFTPETNPR	100		>
PRG4	Q92954	ITEVWGIPSPI	101		>
		DTVFTR
		DQYYNIDVPSR	102		>
RET4	P02753	LIVHNGYCDGR	103		>
SAMP	P02743	AYSDLSR	104	>	>
SHBG	P04278	IALGGLLFPAS	105	<
		NLR
TENX	P22105	TVTVEDLEPGK	106
TETN	P05452	TFHEASEDCIS	107	>
		R
THBG	P05543	NALALFVLPK	108
		AVLHIGEK	109
THRB	P00734	NPDSSTTGPWC	110		>
		YTTDPTVR
TIMP1	P01033	GFQALGDAADI	111	>	>
		R
TSP1	P07996	FVFGTTPEDIL	112	>	>
		R
TSP4	P35443	DVDIDSYPDEE	113	>
		L PCSAR
		AVAEPGIQLK	114
VTDB	P02774	vitd_
		binding
		ELPEHTVK	115
		VLEPTLK	116
		LCDNLSTK	117
		SCESNSPFPVH	118
		PG TAECCTK
		SLGECCDVEDS	119
		T TCFNAK
VTNC	P04004	QPQFISR	120	<
ZA2G	P25311	DIVEYYNDSNG	121
		SHVLQGR
ZPI	Q9UK55	VVNPTLL	122	>

TABLE 2B
Peptide key
Protein name	Biology
A1BG	P04217	Alpha-1B-glycoprotein	Plasma protein
A2AP	P08697	Alpha-2-antiplasmin	Acute phase serine protease inhibitor
A2GL	P02750	Leucine-rich alpha-2-	Plasma protein
		glycoprotein
AACT	P01011	Alpha-1-antichymotrypsin	Acute phase serine protease inhibitor
ACTG	P63261	Actin, cytoplasmic 2	Cytoskeleton component
AFAM	P43652	Afamin	Vitamin E binding protein
AMBP	P02760	Protein AMBP	Inter-alpha-trypsin inhibitor (a serpin)
ANGT	P01019	Angiotensinogen	regulator of blood pressure, body fluid
			and electrolyte homeostasis
ANT3	P01008	Antithrombin-III	serine protease inhibitor in blood
			coagulation
APOB	P04114	Apolipoprotein B-100	major protein constituent of
			chylomicrons, LDL and VLDL
APOE	P02649	Apolipoprotein E	binding, internalization, and catabolism
			of lipoprotein particles
APOH	P02749	Beta-2-glycoprotein 1	Heparin sulphate binding plasma
			protein
B2MG	P61769	Beta-2-microglobulin	Component of the class I MHC
BTD	P43251	Biotinidase	release of biotin from biocytin
C1QC	P02747	Complement C1q	Complement pathway
		subcomponent subunit C
C1R	P00736	Complement C1r	Complement pathway
		subcomponent
C1RL	Q9NZP8	Complement C1r	Mediates the proteolytic cleavage of
		subcomponent-like protein	HP/haptoglobin in the ER
C1S	P09871	Complement C1s	Complement pathway
		subcomponent
C4BPA	P04003	C4b-binding protein alpha	Complement pathway
		chain
C4BPB	P20851	C4b-binding protein beta	Complement pathway
		chain
CD44	P16070	CD44	HA cell surface binding receptor
CERU	P00450	Ceruloplasmin	ferroxidase activity and iron transport
			across the cell membrane
CFAB	P00751	Complement factor B	Complement pathway
CFAH	P08603	Complement factor H	Complement pathway
CFAI	P05156	Complement factor I	Complement pathway
CILP1	O75339	Cartilage intermediate	Cartilage protein
		layer protein 1
CNDP1	Q96KN2	Beta-Ala-His dipeptidase	Serum metaloproteinase
CO2	P06681	Complement C2	Complement pathway
CO4B	P0C0L5	Complement C4-B	Complement pathway
CO5	P01031	Complement C5	Complement pathway
CO5A1	P20908	Collagen alpha-1(V) chain	Type V fibular collagen
CO5A1	P20908	Complement component	Complement pathway
		C6
CO6	P13671	Complement component	Complement pathway
		C6
CO6A3	P12111	Collagen alpha-3(VI) chain	Collagen VI, cell binding
CO8B	P07358	Complement component	Complement pathway
		C8 beta chain
CO8G	P07360	Complement component	Complement pathway
		C8 gamma chain
CO9	P02748	Complement component	Complement pathway
		C9
CO9A3	Q14050	Collagen alpha-3(IX) chain	Type IX collagen
COMP	P49747	Cartilage oligomatrix	Cartilage ECM protein
		protein
COOA1	Q17RW2	Collagen alpha-1(XXIV)	Fibular collagen XIV
		chain
CRAC1	Q9NQ79	Cartilage acidic protein 1	ECM protein found in cartilage, bone
CRTAC			and lung
CSPG2	P13611	Versican core protein	ECM protein binds HA
CXCL7	P02775	Platelet basic protein	stimulates DNA synthesis, mitosis,
			glycolysis, cAMP accumulation,
			prostaglandin E2, and synthesis of HA
			and GAG
DOPO	P09172	Dopamine beta-	Conversion of dopamine to
		hydroxylase	noradrenaline
ECM1	Q16610	Extracellular matrix	Involved in endochondral bone
		protein 1	formation as negative regulator of bone
			mineralization
F13B	P05160	Coagulation factor XIII B	Glycoprotein involved in blood
		chain	coagulation
FA12	P00748	Coagulation factor XII	Serine proteinase involved in blood
			coagulation
FA5	P12259	Coagulation factor V	Blood coagulation, Hemostasis
FBLN1	P23142	Fibulin-1	ECM protein
FBLN3	Q12805	EGF-containing fibulin-	Fibulin-3, ECM protein, May function
		like extracellular matrix	as a negative regulator of chondrocyte
		protein 1	differentiation
FCGBP	Q9Y6R7	IgGFc-binding protein	May be involved in the maintenance of
			the mucosal structure
FCN3	O75636	Ficolin-3	May function in innate immunity
FETUA	P02765	Alpha-2-HS-glycoprotein	influences the mineral phase of bone
FINC	P02751	Fibronectin	ECM multiple roles
FREM2	Q5SZK8	FRAS1-related	Cell membrane adhesion protein
		extracellular matrix protein
		2
GELS	P06396	Gelsolin	Plasma protein
HABP2	Q14520	Hyaluronan-binding	Plasma serine proteinase
		protein 2
HEMO	P02790	Hemopexin	Binds heme and transports it to the liver
HEP2	P05546	Heparin cofactor 2	Blood coagulation, Chemotaxis,
			hemostatsis
HGFA	Q04756	Hepatocyte growth factor	Activates hepatocyte growth factor
		activator
HPLN1	P10915	Hyaluronan and	Link protein, cartilage ECM HA
		proteoglycan link protein 1	binding protein
HRG	P04196	Histidine-rich glycoprotein	Angiogenesis, Blood coagulation,
			Chemotaxis, Fibrinolysis, Hemostasis
IC1	P05155	Plasma protease C1	SERPING1 inhibits C1 of complement
		inhibitor
IRK4	P48050	Inward rectifier potassium	Potassium channel
		channel 4
ITIH1	P19827	Inter-alpha-trypsin	Protease inhibitor, Serine protease
		inhibitor heavy chain H1	inhibitor, binds HA
ITIH4	Q14624	Inter-alpha-trypsin	Acute phase protein
		inhibitor heavy chain H4
KIZ	Q2M2Z5	Centrosomal protein	Centrosomal protein
		kizuna
KLKB1	P03952	Plasma kallikrein	Serine protease
KNG1	P01042	Kininogen-1	Protease inhibitor, Thiol protease
			inhibitor, Vasoactive, Vasodilator
LAMA2	P24043	Laminin subunit alpha-2	Cell ECM binding
LUM	P51884	Lumican	Cartilage ECM protein
LYAM1	P14151	L-selectin	rolling of leukocytes
MASP1	P48740	Mannan-binding lectin	Lectin complement pathway
		serine protease 1
MAST3	O60307	Microtubule-associated	Cytoplasmic enzyme
		serine/threonine-protein
		kinase 3
PCOC1	Q15113	Procollagen C-	Type 1 collagen fibril formation
		endopeptidase enhancer 1
PGBM	P98160	Perlecan	ECM protein
PGCA	P16112	Aggrecan core protein	Cartilage ECM HA binding protein
PGRP2	Q96PD5	N-acetylmuramoyl-L-	May play a scavenger role by digesting
		alanine amidase	biologically active peptidoglycan
PHLD	P80108	Phosphatidylinositol-	hydrolyzes the inositol phosphate
		glycan-specific	linkage in proteins
		phospholipase D
PLF4	P02776	Platelet factor 4	Released during platelet aggregation
PLMN	P00747	Plasminogen	Blood coagulation, Fibrinolysis,
			Hemostasis. Tissue remodeling
PRG4	Q92954	Proteoglycan 4	Lubricin, cartilage ECM protein
PRLD2	Q8N945	PRELI domain containing	Mitrochondrial
		2
PZP	P20742	Pregnancy zone protein	inhibit all four classes of proteinases
RET4	P02753	Retinol-binding protein 4	Retinol transporter protein
RTN4	Q9NQC3	Reticulon-4	neurogenesis
RUNX2	Q13950	Runt-related transcription	osteoblastic differentiation and skeletal
		factor 2	morphogenesis
SAMP	P02743	Serum amyloid P-	Can interact with DNA and histones
		component	and may scavenge nuclear material
			released from damaged circulating cells
SEPP1	P49908	Selenoprotein P	Selenium transport and ECM
			antioxidant
SHBG	P04278	Sex hormone-binding	androgen transport protein
		globulin
SPTA2	Q13813	Spectrin alpha chain, non-	Ca depended cytoskeletal re-
		erythrocytic 1	organization
TENX	P22105	Tenascin X	Anti-adhesive ECM glycoprotein
TETN	P05452	Tetranectin	Plasma protein
THBG	P05543	Thyroxine-binding	Major thyroid hormone transport
		globulin	protein in serum.
THRB	P00734	Prothrombin	Acute phase, Blood coagulation,
			Hemostasis
TIMP1	P01033	Tissue inhibitor	MMP inhibitor
		metalloproteinase 1
TNR6C	Q9HCJ0	trinucleotide repeat	miRNA silencing
		containing 6C
TSP1	P07996	Thrombospondin 1	Cell/cell/matrix binding
TSP4	P35443	Thrombospondin-4	Mediates cell/cell and cell/matrix
			adhesion
VTDB	P02774	Vitamin D-binding protein	Vitamin D transporter protein
VTNC	P04004	vitronectin	Cell/matrix adhesion factor
ZA2G	P25311	Zinc-alpha-2-glycoprotein	Stimulates lipid degradation in
			adipocytes
ZPI	Q9UK55	Protein Z-dependent	Blood coagulation, Hemostasis
		protease inhibitor

TABLE 2C
Diagnostic Data (OA = Osteoarthritis, C = Control)
Diagnostic Data
phenotype (SEQ ID NO: X)	OAmean	OAsd	Cmean	Csd
MRM markers; non-depleted Serum; knee
Knee_JSN_Pheno_Q9NQ79_CRAC1_GVA	−5.3467	0.31328	−5.5677	0.29133
SLFAGR (1)
Knee_JSN_Pheno_P02775_CXCL7_NIQSL	−1.579	0.26191	−1.8924	0.78391
EVIGK (57)
Knee_JSN_Pheno_P04217_A1BG_IFFHLN	0.01795	0.2559	−0.0399	0.22048
AVALGDGGHYTCR (3)
Knee_JSN_Pheno_P02776_PLF4_ICLDLQ	−2.6781	0.31999	−2.9236	0.74256
APLYK (99)
Knee_JSN_Pheno_P05452_TETN_TFHEAS	−2.7375	0.23026	−2.824	0.23499
EDCISR (107)
Knee_JSN_Pheno_P01033_TIMP1_GFQAL	−5.5478	0.39529	−5.7389	0.421
GDAADIR (111)
Knee_JSN_Pheno_P49747_COMP_NALW	−5.2222	0.39632	−5.4563	0.48604
HTGDTESQVR (54)
Knee_JSN_Pheno_P04003_C4BPA_LSLEI	−0.6158	0.24333	−0.732	0.21636
EQLELQR (25)
Knee_JSN_Pheno_P07996_TSP1_FVFGTT	−3.9111	0.33405	−4.1246	0.70016
PEDILR (112)
Knee_JSN_Pheno_P80108_PHLD_FGS_S_LI	−3.9109	0.31233	−4.0202	0.295
TVR (98)
Knee_JSN_Pheno_P16112_PGCA_VSLPN	−8.5077	0.60466	−8.3027	0.59736
YPAIPSDATLEVQSLR (96)
Knee_JSN_Pheno_P04003_C4BPA_GVGW	−0.6333	0.30499	−0.7658	0.27236
SHPLPQCEIVK (26)
Knee_JSN_Pheno_P01031_CO5_GIYGTIS	−2.6335	0.19845	−2.727	0.18412
R (44)
Knee_JSN_Pheno_P51884_LUM_ILGPLSY	−2.2931	0.18757	−2.381	0.17292
SK (90)
Knee_JSN_Pheno_P01031_CO5_TLLPVSK	−2.6148	0.19708	−2.7026	0.18569
PEIR (45)
Knee_JSN_Pheno_P07360_CO8G_QLYGD	−4.6571	0.35419	−4.8008	0.38448
TGVLGR (52)
Knee_JSN_Pheno_P20851_C4BPB_SQCLE	−2.4987	0.25857	−2.6078	0.24415
DHTWAPPFPICK (27)
Knee_JSN Pheno_P01031_CO5_IIHFGTR	−2.4834	0.2052	−2.5725	0.19137
(46)
Knee_JSN_Pheno_P05156_CFAI_HGNTDS	−2.1353	0.23576	−2.2203	0.2542
EGIVEVK (36)
Knee_JSN_Pheno_P02743_SAMP_AYSDL	−1.172	0.26186	−1.3063	0.39507
SR (104)
Knee_JSN_Pheno_Q9NQ79_CRAC1_SSPY	−5.3076	0.32576	−5.4371	0.30832
YALR (2)
Knee_JSN_Pheno_Q15113_PCOC1_TGGL	−5.0009	0.19706	−5.0763	0.17589
DLPSPPTGASLK (95)
Knee_JSN_Pheno_P12259_FA5_SEAYNTF	−4.6902	0.23862	−4.7772	0.27625
SER (61)
Knee_JSN_Pheno_P51884_LUM_VANEVT	−1.3903	0.22821	−1.4664	0.22184
LN (91)
Knee_JSN_Pheno_P04278_SHBG_IALGGL	−4.7876	0.75918	−4.5138	0.65685
LFPASNLR (105)
Knee_KL_Pheno_Q9NQ79_CRAC1_GVAS	−5.36	0.30055	−5.634	0.28483
LFAGR (1)
Knee_KL_Pheno_Q9NQ79_CRAC1_SSPY	−5.2986	0.30834	−5.5093	0.30855
YALR (2)
Knee_KL_Pheno_P07360_CO8G_QLYGDT	−4.6643	0.36247	−4.8469	0.3723
GVLGR (52)
Knee_KL_Pheno_P49747_COMP_NALWH	−5.2483	0.40041	−5.5027	0.51099
TGDTESQVR (54)
Knee_KL_Pheno_P07358_CO8B_GILNEIK	−4.337	0.36071	−4.4394	0.33841
(51)
Knee_KL_Pheno_P04004_VTNC_QPQFIS	−2.5841	0.32518	−2.4986	0.31075
R(120)
Knee_KL_Pheno_P04278_SHBG_IALGGL	−4.7795	0.74084	−4.415	0.62606
LFPASNLR (105)
Knee_KL_Pheno_P02775_CXCL7_NIQSLE	−1.6465	0.43393	−1.8899	0.80079
VIGK (57)
Knee_KL_Pheno_Q15113_PC0C1_TGGLD	−5.0083	0.20352	−5.0934	0.14771
LPSPPTGASLK (95)
Knee_KL_Pheno_P12111_CO6A3_EVQVF	−5.7888	0.27059	−5.9159	0.24476
EITENSAK (49)
Knee_KL_Pheno_P51884_LUM_ILGPLSY	−2.3063	0.18911	−2.3918	0.16554
SK (90)
Knee_KL_Pheno_P02743_SAMP_AYSDLS	−1.195	0.2826	−1.317	0.41746
R(104)
Knee_KL_Pheno_P01031_CO5_IIHFGTR	−2.5003	0.21274	−2.5762	0.17348
(46)
Knee_KL_Pheno_P51884_LUM_VANEVT	−1.4005	0.22997	−1.4781	0.21611
LN (91)
Knee_KL_Pheno_P01031_CO5_GIYGTISR	−2.6535	0.21307	−2.7265	0.15017
(44)
Knee_KL_Pheno_P04003_C4BPA_LSLEIE	−0.6435	0.25492	−0.7256	0.18876
QLELQR (25)
Knee_OST_Pheno_Q9NQ79_CRAC1_GVA	−5.353	0.30661	−5.5993	0.28726
SLFAGR (1)
Knee_OST_Pheno_Q9NQ79_CRAC1_SSP	−5.2892	0.30785	−5.4886	0.31058
YYALR (2)
Knee_OST_Pheno_P04004_VTNC_QPQFIS	−2.5963	0.31787	−2.4944	0.32087
R(120)
Knee_OST_Pheno_P49747_COMP_NALW	−5.2442	0.40004	−5.467	0.50102
HTGDTESQVR (54)
Knee_OST_Pheno_P04278_SHBG_IALGG	−4.7799	0.74413	−4.4746	0.65493
LLFPASNLR (105)
Knee_OST_Pheno_P12111_C06A3_EVQV	−5.7842	0.27689	−5.9018	0.24009
FEITENSAK (49)
Knee_OST_Pheno_Q15113_PCOC1_TGGL	−5.0092	0.20737	−5.078	0.15463
DLPSPPTGASLK (95)
MRM markers; non-depleted Serum; person
Knee_JSN_person_Pheno_P00450_CERU_	0.65572	0.24983	0.76035	0.27569
HYYIAAEEIIWNYAPSGIDIFTK (29)
Knee_JSN_person_Pheno_P01031_CO5_GI	−2.6381	0.20909	−2.7346	0.16563
YGTISR (44)
Knee_JSN_person_Pheno_P01031_CO5_II	−2.4887	0.21661	−2.5783	0.17261
HFGTR (46)
Knee_JSN_person_Pheno_P01031_C05_TL	−2.6187	0.20754	−2.7103	0.16802
LPVSKPEIR (45)
Knee_JSN_person_Pheno_P01033_TIMP1_	−5.5604	0.40065	−5.7497	0.42285
GFQALGDAADIR (111)
Knee_JSN_person_Pheno_P02743_SAMP_	−1.1745	0.27542	−1.3229	0.39912
AYSDLSR (104)
Knee_JSN_person_Pheno_P02775_CXCL7_	−1.5899	0.27108	−1.924	0.83168
NIQSLEVIGK (57)
Knee_JSN_person_Pheno_P02776_PLF4_IC	−2.6915	0.33412	−2.9413	0.78317
LDLQAPLYK (99)
Knee_JSN_person_Pheno_P04003_C4BPA_	−0.6459	0.33432	−0.7677	0.2194
GVGWSHPLPQCEIVK (26)
Knee_JSN_person_Pheno_P04003_C4BPA_	−0.6165	0.25436	−0.7484	0.18988
LSLEIEQLELQR (25)
Knee_JSN_person_Pheno_P04217_AlBG_I	0.00845	0.27983	−0.0351	0.17265
FFHLNAVALGDGGHYTCR (3)
Knee_JSN_person_Pheno_P04278_SHBG_I	−4.7677	0.77275	−4.501	0.62726
ALGGLLFPASNLR (105)
Knee_JSN_person_Pheno_P07360_CO8G_	−4.6695	0.36108	−4.8047	0.38572
QLYGDTGVLGR (52)
Knee_JSN_person_Pheno_P07996_TSP1_F	−3.9283	0.35544	−4.1322	0.73241
VFGTTPEDILR (112)
Knee_JSN_person_Pheno_P12259_FA5_SE	−4.6923	0.24286	−4.7873	0.27747
AYNTFSER (61)
Knee_JSN_person_Pheno_P20851_C4BPB_	−2.5112	0.27892	−2.6064	0.21459
SQCLEDHTWAPPFPICK (27)
Knee_JSN_person_Pheno_P49747_COMP_	−5.2389	0.40342	−5.4677	0.49665
NALWHTGDTESQVR (54)
Knee_JSN_person_Pheno_P51884_LUM_IL	−2.2943	0.18871	−2.3925	0.16777
GPLSYSK (90)
Knee_JSN_person_Pheno_P51884_LUM_V	−1.3877	0.23105	−1.4815	0.21506
ANEVTLN (91)
Knee_JSN_person_Pheno_Q15113_PCOC1_	−4.9982	0.19726	−5.0915	0.16919
TGGLDLPSPPTGASLK (95)
Knee_JSN_person_Pheno_Q9NQ79_CRAC	−5.3647	0.31833	−5.5753	0.28924
1_GVASLFAGR (1)
Knee_JSN_person_Pheno_Q9NQ79_CRAC	−5.3165	0.32877	−5.4439	0.30529
1_SSPYYALR (2)
Knee_KL_person_Pheno_P01031_CO5_GI	−2.6531	0.21059	−2.7391	0.14588
YGTISR (44)
Knee_KL_person_Pheno_P01031_CO5_IIH	−2.4986	0.21063	−2.5922	0.1707
FGTR (46)
Knee_KL_person_Pheno_P02743_SAMP_A	−1.1947	0.27677	−1.3369	0.44546
YSDLSR (104)
Knee_KL_person_Pheno_P02775_CXCL7_	−1.6407	0.42587	−1.9424	0.85113
NIQSLEVIGK (57)
Knee_KL_person_Pheno_P04278_SHBG_I	−4.7405	0.75939	−4.4535	0.59653
ALGGLLFPASNLR (105)
Knee_KL_person_Pheno_P05156_CFAI_A	−2.3816	0.24298	−2.5067	0.27195
QLGDLPWQVAIK (37)
Knee_KL_person_Pheno_P05156_CFAI_H	−2.1467	0.24629	−2.247	0.24236
GNTDSEGIVEVK (36)
Knee_KL_person_Pheno_P07360_CO8G_Q	−4.6628	0.36307	−4.8793	0.36711
LYGDTGVLGR (52)
Knee_KL_person_Pheno_P12111_CO6A3_	−5.7924	0.26658	−5.927	0.25429
EVQVFEITENSAK (49)
Knee_KL_person_Pheno_P12259_FA5_SE	−4.7013	0.24808	−4.8058	0.27986
AYNTFSER (61)
Knee_KL_person_Pheno_P49747_COMP_N	−5.2524	0.39475	−5.5326	0.5373
ALWHTGDTESQVR (54)
Knee_KL_person_Pheno_P51884_LUM_IL	−2.3066	0.18753	−2.4043	0.16517
GPLSYSK (90)
Knee_KL_person_Pheno_P51884_LUM_V	−1.3978	0.22966	−1.497	0.21235
ANEVTLN (91)
Knee_KL_person_Pheno_Q15113_PCOC1_	−5.0045	0.20633	−5.116	0.11555
TGGLDLPSPPTGASLK (95)
Knee_KL_person_Pheno_Q9NQ79_CRAC1_	−5.365	0.30117	−5.6649	0.27366
GVASLFAGR (1)
Knee_KL_person_Pheno_Q9NQ79_CRAC1_	−5.2981	0.31394	−5.5437	0.28344
SSPYYALR (2)
Knee_KL_person_Pheno_Q9UK55_ZPI_VV	−3.0199	0.13919	−3.0653	0.15335
NPTLL (122)
Knee_OST_person_Pheno_P02775_CXCL7_	−1.6482	0.43797	−1.8835	0.8009
NIQSLEVIGK (57)
Knee_OST_person_Pheno_P04004_VTNC_	−2.5885	0.33293	−2.4911	0.29601
QPQFISR (120)
Knee_OST_person_Pheno_P07360_CO8G_	−4.6736	0.37267	−4.8264	0.36566
QLYGDTGVLGR (52)
Knee_OST_person_Pheno_P12111_CO6A3_	−5.7868	0.27448	−5.9182	0.2382
EVQVFEITENSAK (49)
Knee_OST_person_Pheno_P49747_COMP_	−5.2438	0.39886	−5.5084	0.51353
NALWHTGDTESQVR (54)
Knee_OST_person_Pheno_P63261_ACTG_	−4.1926	0.28264	−4.0727	0.29266
VAPEEHPVLLTEAPLNPK (11)
Knee_OST_person_Pheno_Q15113_PCOC1_	−5.0081	0.2103	−5.0927	0.13171
TGGLDLPSPPTGASLK (95)
Knee_OST_person_Pheno_Q9NQ79_CRAC	−5.357	0.3055	−5.6367	0.27386
1_GVASLFAGR (1)
Knee_OST_person_Pheno_Q9NQ79_CRAC	−5.2931	0.3177	−5.5175	0.28595
1_SSPYYALR (2)
Knee_OST_person_Pheno_Q9UK55_ZPI_V	−3.0144	0.12597	−3.0693	0.1705
VNPTLL (122)
MRM markers; depleted Serum; person; actin
Knee_JSN_person_Pheno_P02776_PLF4_IC	0.01106	0.56919	−0.0167	0.52339
LDLQAPLYK (99)
Knee_JSN_person_Pheno_P06396_GELS_G	0.01656	0.24249	−0.0285	0.2353
GVASGFK (72)
Knee_JSN_person_Pheno_P08697_A2AP_S	0.20908	0.55165	−0.3041	0.41444
PPGVCSR (4)
Knee_JSN_person_Pheno_P08697_A2AP_L	−0.1708	0.72306	0.24376	0.29575
CQDLGPGAFR (5)
Knee_JSN_person_Pheno_P04114_APOB_	−0.0027	0.32783	0.01013	0.3457
LAIPEGK (15)
Knee_JSN_person_Pheno_P02649_APOE_L	0.08819	0.4891	−0.1192	0.48267
QAEAFQAR (17)
Knee_JSN_person_Pheno_P02747_C1QC_	−0.0366	0.91229	0.08887	0.78024
VVTFCGHTSK (20)
Knee_JSN_person_Pheno_P05156_CFAI_A	0.01703	0.2518	−0.0191	0.24696
QLGDLPWQVAIK (37)
Knee_JSN_person_Pheno_P06681_CO2_SS	−0.1684	0.76581	0.25677	0.27432
GQWQTPGATR (41)
Knee_JSN_person_Pheno_P01031_CO5_GI	−0.0594	0.43213	0.09809	0.26792
YGTISR (44)
Knee_JSN_person_Pheno_P12111_CO6A3_	0.05255	0.2943	−0.0797	0.23417
EVQVFEITENSAK (49)
Knee_JSN_person_Pheno_P07360_CO8G_	0.05261	0.30651	−0.0672	0.31322
QLYGDTGVLGR (52)
Knee_JSN_person_Pheno_Q9NQ79_CRAC	0.08623	0.38824	−0.1257	0.34708
1_SSPYYALR (2)
Knee_JSN_person_Pheno_Q9NQ79_CRAC	0.09636	0.33552	−0.1408	0.33742
1_GVASLFAGR (1)
Knee_JSN_person_Pheno_P12259_FA5_SE	−0.0749	0.41656	0.11694	0.26129
AYNTFSER (61)
Knee_JSN_person_Pheno_Q12805_FBLN3_	0.05341	0.36555	−0.0956	0.29821
ADQVCINLR (65)
Knee_JSN_person_Pheno_P01042_KNG1_	0.0016	0.76436	0.0537	0.40236
LDDDLEHQGGHVLDHGHK (88)
Knee_JSN_person_Pheno_Q15113_PCOC1_	0.04065	0.28654	−0.0545	0.28022
TGGLDLPSPPTGASLK (95)
Knee_JSN_person_Pheno_P80108_PHLD_F	0.06571	0.6118	−0.1345	0.52321
GSSLITVR (98)
Knee_JSN_person_Pheno_P02743_SAMP_	0.04589	0.26942	−0.075	0.30542
AYSDLSR (104)
Knee_JSN_person_Pheno_P05452_TETN_T	0.02835	0.28053	−0.0469	0.25301
FHEASEDCISR (107)
Knee_JSN_person_Pheno_P01033_TIMP1_	0.04351	0.34846	−0.052	0.19604
GFQALGDAADIR (111)
Knee_JSN_person_Pheno_P49747_COMP_	−0.1264	0.55906	0.20054	0.42498
SSTGPGEQLR (55)
Knee_JSN_person_Pheno_P35443_TSP4_D	0.07752	0.49144	−0.1112	0.35704
VDIDSYPDEELPCSAR (113)
Knee_JSN_person_Pheno_Q12805_FBLN3_	0.06224	0.39736	−0.1043	0.27103
NPCQDPYILTPENR (64)
Knee_KL_person_Pheno_P02776_PLF4_IC	0.01508	0.56529	−0.0374	0.51299
LDLQAPLYK (99)
Knee_KL_person_Pheno_P06396_GELS_G	0.01122	0.24343	−0.0334	0.2305
GVASGFK (72)
Knee_KL_person_Pheno_P23142_FBLN1_	0.01083	0.28993	−0.0576	0.25805
TGYYFDGISR (63)
Knee_KL_person_Pheno_P08697_A2AP_S	0.15955	0.57776	−0.3861	0.22829
PPGVCSR (4)
Knee_KL_person_Pheno_P08697_A2AP_L	−0.1189	0.68515	0.28116	0.2684
CQDLGPGAFR (5)
Knee_KL_person_Pheno_P02649_APOE_L	0.07706	0.46658	−0.1739	0.52368
QAEAFQAR (17)
Knee_KL_person_Pheno_P02747_C1QC_V	−0.0097	0.88292	0.07314	0.81026
VTFCGHTSK (20)
Knee_KL_person_Pheno_P16070_CD44_Y	0.01844	0.2285	−0.0494	0.20504
GFIEGHVVIPR (28)
Knee_KL_person_Pheno_P08603_CFAH_C	0.02996	0.26309	−0.0792	0.20569
LPVTAPENGK (35)
Knee_KL_person_Pheno_P06681_CO2_SS	−0.1153	0.72103	0.29553	0.25057
GQWQTPGATR (41)
Knee_KL_person_Pheno_P01031_CO5_GI	−0.0458	0.42048	0.12717	0.22113
YGTISR (44)
Knee_KL_person_Pheno_P12111_CO6A3_	0.03803	0.28928	−0.0966	0.22546
EVQVFEITENSAK (49)
Knee_KL_person_Pheno_Q9NQ79_CRAC1_	0.0889	0.37549	−0.2157	0.32057
SSPYYALR (2)
Knee_KL_person_Pheno_Q9NQ79_CRAC1_	0.09367	0.3324	−0.2277	0.30339
GVASLFAGR (1)
Knee_KL_person_Pheno_P12259_FA5_SE	−0.0563	0.40341	0.14727	0.23078
AYNTFSER (61)
Knee_KL_person_Pheno_Q12805_FBLN3_	0.04385	0.35046	−0.1311	0.30663
ADQVCINLR (65)
Knee_KL_person_Pheno_Q15113_PCOC1_	0.04197	0.27951	−0.0952	0.28438
TGGLDLPSPPTGASLK (95)
Knee_KL_person_Pheno_P01033_TIMP1_	0.02734	0.33457	−0.0504	0.17746
GFQALGDAADIR (111)
Knee_KL_person_Pheno_P49747_COMP_S	−0.08	0.57743	0.21684	0.31976
STGPGEQLR (55)
Knee_KL_person_Pheno_P35443_TSP4_D	0.06604	0.47366	−0.1577	0.34252
VDIDSYPDEELPCSAR (113)
Knee_KL_person_Pheno_Q12805_FBLN3_	0.04932	0.37969	−0.1385	0.26628
NPCQDPYILTPENR (64)
Knee_OST_person_Pheno_P06396_GELS_	0.01502	0.24089	−0.035	0.23657
GGVASGFK (72)
Knee_OST_person_Pheno_P08697_A2AP_	0.14605	0.57584	−0.2877	0.39295
SPPGVCSR (4)
Knee_OST_person_Pheno_P08697_A2AP_	−0.1351	0.70192	0.26055	0.27125
LCQDLGPGAFR (5)
Knee_OST_person_Pheno_P63261_ACTG_	−3.0367	0.27282	−2.9771	0.25759
VAPEEHPVLLTEAPLNPK (11)
Knee_OST_person_Pheno_P02747_C1QC_	0.01463	0.89969	0.01413	0.78594
VVTFCGHTSK (20)
Knee_OST_person_Pheno_P06681_CO2_SS	−0.1157	0.72985	0.2422	0.34386
GQWQTPGATR (41)
Knee_OST_person_Pheno_P01031_CO5_GI	−0.0524	0.42324	0.11753	0.24644
YGTISR (44)
Knee_OST_person_Pheno_P12111_CO6A3_	0.04953	0.26305	−0.1016	0.28306
EVQVFEITENSAK (49)
Knee_OST_person_Pheno_Q9NQ79_CRAC	0.08684	0.38283	−0.1715	0.33148
1_SSPYYALR (2)
Knee_OST_person_Pheno_Q9NQ79_CRAC	0.09341	0.33665	−0.1849	0.31822
1_GVASLFAGR (1)
Knee_OST_person_Pheno_P12259_FA5_SE	−0.067	0.40428	0.14172	0.25122
AYNTFSER (61)
Knee_OST_person_Pheno_Q12805_FBLN3_	0.04391	0.35403	−0.1082	0.31055
ADQVCINLR (65)
Knee_OST_person_Pheno_Q15113_PCOC1_	0.03371	0.28161	−0.0608	0.2897
TGGLDLPSPPTGASLK (95)
Knee_OST_person_Pheno_P05452_TETN_	0.01949	0.27559	−0.0453	0.2601
TFHEASEDCISR (107)
Knee_OST_person_Pheno_P49747_COMP_	−0.0702	0.58158	0.15842	0.3793
SSTGPGEQLR (55)
Knee_OST_person_Pheno_P35443_TSP4_D	0.0712	0.47712	−0.1384	0.35618
VDIDSYPDEELPCSAR (113)
Knee_OST_person_Pheno_Q12805_FBLN3_	0.046	0.3822	−0.1073	0.28853
NPCQDPYILTPENR (64)
ELISA markers; knee; GEE
Knee_JSN_Pheno_cd14_serum	2575.92	697.218	2248.74	613.423
Knee_JSN_Pheno_1HA_serum	5.71988	0.85584	5.11415	0.75264
Knee_JSN_Pheno_cerulo_urine_urine1	0.5915	0.79293	1.01494	0.92315
Knee_JSN_Pheno_kinno_serum	294.282	121.894	399.274	235.284
Knee_JSN_Pheno_cerulo_urine_urine2	−8.792	1.27149	−8.0153	1.54441
Knee_JSN_Pheno_hapto_serum	1333.67	757.365	1072.78	642.54
Knee_JSN_Pheno_ctx1a_ctx_1b_urine2	−15.333	2.20557	−14.262	2.50326
Knee_JSN_Pheno_ctx1a_ctx1b_urine1	−5.9539	1.36704	−5.2837	1.54433
Knee_JSN_Pheno_vitd_binding_serum	939.199	593.361	1196.78	705.038
Knee_KL_Pheno_kinno_serum	295.263	119.804	441.7	259.637
Knee_KL_Pheno_cd14_serum	2546.19	715.028	2169.76	514.811
Knee_KL_Pheno_1HA_serum	5.63111	0.81308	5.01795	0.7993
Knee_KL_Pheno_cerulo_urine_urine2	−8.6707	1.29058	−7.9629	1.63572
Knee_KL_Pheno_ctx1a_ctx_1b_urine2	−15.234	2.22103	−14.065	2.56318
Knee_KL_Pheno_ctx1a_ctx_1b_urine1	−5.8997	1.38261	−5.1219	1.56123
Knee_KL_Pheno_vitd_binding_serum	960.501	602.213	1263.09	725.301
Knee_OST_Pheno_kinno_serum	287.479	113.997	436.257	249.333
Knee_OST_Pheno_cerulo_urine_urine2	−8.6634	1.28494	−8.0895	1.61516
Knee_OST_Pheno_1HA_serum	5.61496	0.81508	5.11291	0.83713
Knee_OST_Pheno_ctx1a_ctx1b_urine2	−15.23	2.21429	−14.263	2.55727
Knee_OST_Pheno_ctx1a_ctx1b_urine1	−5.9052	1.38361	−5.2423	1.55556
Knee_OST_Pheno_cd14_serum	2525.14	699.29	2262.88	613.837
Knee_OST_Pheno_ctx2_urine1	−5.0633	0.67926	−5.3241	0.66827
Knee_OST_Pheno_vitd_binding_serum	969.878	616.484	1204.18	702.73
ELISA markers; knee; person
Knee_JSN_person_Pheno_cd14_serum	2602.29	697.192	2162.75	564.587
Knee_JSN_person_Pheno_cerulo_urine_urin	0.57459	0.78943	1.10486	0.92437
e1
Knee_JSN_person_Pheno_cerulo_urine_urin	−8.8113	1.25771	−7.8761	1.56237
e2
Knee_JSN_person_Pheno_1HA_serum	5.69895	0.82885	5.06371	0.77171
Knee_JSN_person_Pheno_kinno_serum	294.293	122.788	412.385	243.775
Knee_JSN_person_Pheno_ctx1a_ctx1b_urin	−15.324	2.21534	−14.121	2.52557
e2
Knee_JSN_person_Pheno_ctxi_urine2	−12.991	1.55032	−12.262	1.85502
Knee_JSN_person_Pheno_ctx1a_ctx1b_urin	−5.9456	1.37228	−5.1932	1.55878
e1
Knee_JSN_person_Pheno_ctxi_urine1	−3.613	0.77049	−3.2866	0.82608
Knee_JSN_person_Pheno_vitd_binding_seru	946.698	594.172	1224.92	721.843
m
Knee_KL_person_Pheno_kinno_serum	296.711	124.85	455.818	265.551
Knee_KL_person_Pheno_cd14_serum	2530.04	710.539	2149.83	508.765
Knee_KL_person_Pheno_1HA_serum	5.61727	0.80017	4.97511	0.83057
Knee_KL_person_Pheno_cerulo_urine_urine	−8.6167	1.33156	−7.9782	1.65722
2
Knee_KL_person_Pheno_vitd_binding_seru	981.978	620.479	1258.48	725.875
m
Knee_KL_person_Pheno_ctx1a_ctx1b_urine	−5.8193	1.4108	−5.1917	1.60826
1
Knee_OST_person_Pheno_kinno_serum	291.557	123.07	452.669	256.211
Knee_OST_person_Pheno_cd14_serum	2520.46	709.306	2223.94	573.96
Knee_OST_person_Pheno_1HA_serum	5.601	0.79819	5.05754	0.87099
Knee_OST_person_Pheno_ctx1a_ctx1b_urin	−5.8403	1.42891	−5.2499	1.55116
e1

TABLE 2D
Prognostic Data (P = Progressor, NP = non−Progressor)
Prognostic Data
Phenotype (SEQ ID NO: X)	Pmean	Psd	NPmean	NPsd
MRM markers; nondepleted serum; knee; GEE
Knee_JSN_Pheno_P08603_CFAH_CLPVT	−1.8733	0.18394	−1.9774	0.25529
APENGK (35)
Knee_JSN Pheno_P02749_APOH_ATFGC	0.21648	0.18367	0.10653	0.27925
HDGYSLDGPEEIECTK (18)
Knee_JSN Pheno_P04196_HRG_YWNDC	−2.3987	0.25856	−2.4794	0.35586
EPPDSR (81)
Knee_JSN_Pheno_P01011_AACT_NLAVS	0.06855	0.2493	−0.0534	0.25394
QVVHK (9)
Knee_JSN_Pheno_P01011_AACT_ADLSG	0.00169	0.23843	−0.1136	0.23088
ITGAR (8)
Knee_JSN_Pheno_P02753_RET4_LIVHN	−1.328	0.26527	−1.4141	0.29284
GYCDGR (103)
Knee_JSN_Pheno_P02743_SAMP_AYSDL	−1.1083	0.2406	−1.2255	0.26865
SR (104)
Knee_JSN_Pheno_P00734_THRB_NPDSS	−1.2674	0.3453	−1.399	0.36795
TTGPWCYTTDPTVR (110)
Knee_JSN_Pheno_P01011_AACT_EQLSL	0.64266	0.24252	0.52223	0.25274
LDR (10)
Knee_JSN_Pheno_Q14624_ITIH4_FKPTL	−1.1111	0.21404	−1.1857	0.24724
SQQQK (86)
Knee_JSN_Pheno_P07996_TSP1_FVFGTT	−3.8418	0.26133	−3.9693	0.37682
PEDILR (112)
Knee_JSN_Pheno_P16112_PGCA_VSLPN	−8.3507	0.5464	−8.6397	0.6233
YPAIPSDATLEVQSLR (96)
Knee_KL_Pheno_P08603_CFAH_CLPVT	−1.8514	0.15151	−1.9889	0.26845
APENGK (35)
Knee_KL_Pheno_P80108_PHLD_FGSSLI	−3.8693	0.223	−3.9931	0.33846
TVR (98)
Knee_KL_Pheno_P02743_SAMP_AYSDL	−1.0849	0.2601	−1.2417	0.27982
SR (104)
Knee_KL_Pheno_P01031_CO5_TLLPVSK	−2.5743	0.17657	−2.6595	0.21099
PER (45)
Knee_KL_Pheno_P01031_CO5_GIYGTIS	−2.586	0.17701	−2.6822	0.22119
R (44)
Knee_KL_Pheno_P04003_C4BPA_GVGW	−0.6009	0.26167	−0.7042	0.35265
SHPLPQCEIVK (26)
Knee_KL_Pheno_P02753_RET4_LIVHNG	−1.318	0.28743	−1.405	0.27131
YCDGR (103)
Knee_KL_Pheno_P02749_APOH_ATFGC	0.21968	0.20833	0.12004	0.27085
HDGYSLDGPEEIECTK (18)
Knee_KL_Pheno_P07996_TSP1_FVFGTT	−3.818	0.24999	−4.0312	0.51382
PEDILR (112)
Knee_KL_Pheno_P02765_FETUA_FSVV	0.47776	0.1871	0.42755	0.18126
YAK (69)
Knee_KL_Pheno_P00734_THRB_NPDSST	−1.2604	0.36827	−1.4032	0.36302
TGPWCYTTDPTVR (110)
Knee_KL_Pheno_P00736_C1R_GLTLHLK	−1.5768	0.17668	−1.6834	0.23557
(22)
Knee_KL_Pheno_P01011_AACT_NLAVS	0.09174	0.25744	−0.0404	0.25316
QVVHK (9)
Knee_KL_Pheno_P00736_C1R_GYGFYT	−2.4711	0.1838	−2.5691	0.21992
K(23)
Knee_KL_Pheno_P09871_C1S_LLEVPEG	−2.3873	0.16142	−2.4663	0.17408
R(39)
Knee_KL_Pheno_P04114_APOB_LAIPEG	−1.1379	0.23634	−1.2751	0.2705
K(15)
Knee_KL_Pheno_P01011_AACT_ADLSGI	0.0238	0.24725	−0.0937	0.22724
TGAR (8)
Knee_KL_Pheno_P05546_HEP2_NFGYTL	−0.3486	0.23541	−0.4828	0.22968
R(76)
Knee_KL_Pheno_P05156_CFAI_HGNTDS	−2.0777	0.19407	−2.1776	0.26138
EGIVEVK (36)
Knee_KL_Pheno_P02751_FINC_EYLGAI	−1.5794	0.37426	−1.7745	0.50563
CSCTCFGGQR (70)
Knee_KL_Pheno_P02751_FINC_IGDTWS	−1.9724	0.28416	−2.1392	0.44287
K(71)
Knee_OST_Pheno_P02776_PLF4_ICLDLQ	−2.6274	0.32402	−2.8927	0.54107
APLYK (99)
Knee_OST_Pheno_P01011_AACT_EQLSL	0.63078	0.24078	0.5271	0.24313
LDR (10)
Knee_OST_Pheno_P01011_AACT_ADLS	−0.0125	0.22315	−0.1117	0.22687
GITGAR (8)
Knee_OST_Pheno_P01008_ANT3_ATEDE	5.43145	0.39014	5.62562	0.4794
GSEQK (14)
Knee_OST_Pheno_P01011_AACT_NLAV	0.04359	0.24761	−0.0522	0.24691
SQVVHK (9)
Knee_OST_Pheno_P02775_CXCL7_NIQS	−1.5495	0.28583	−1.7827	0.54969
LEVIGK (57)
Knee_OST_Pheno_Q9NQ79_CRAC1_SSP	−5.2482	0.31894	−5.3319	0.29197
YYALR (2)
Knee_OST_Pheno_Q14624_ITIH4_FKPTL	−1.1228	0.21444	−1.2061	0.25212
SQQQK (86)
Knee_OST_Pheno_P00747_PLMN_HSIFT	−0.9833	0.22029	−1.0623	0.23109
PETNPR (100)
Knee_OST_Pheno_P07996_TSP1_FVFGT	−3.8992	0.3198	−4.0688	0.58725
TPEDILR (112)
Knee_OST_Pheno_P01031_CO5_GIYGTIS	−2.6246	0.18727	−2.6943	0.23962
R (44)
Knee_OST_Pheno_P00734_THRB_NPDSS	−1.3017	0.35587	−1.4473	0.37127
TTGPWCYTTDPTVR (110)
Knee_OST_Pheno_P04003_C4BPA_GVG	−0.6339	0.29939	−0.7455	0.37048
WSHPLPQCEIVK (26)
MRM markers; nondepleted serum; knee; person
Knee_JSN_person_Pheno_P04196_HRG_Y	−2.3884	0.28937	−2.5389	0.3736
WNDCEPPDSR (81)
Knee_JSN_person_Pheno_P08603_CFAH_	−1.8864	0.19324	−2.0163	0.29792
CLPVTAPENGK (35)
Knee_JSN_person_Pheno_P00734_THRB_	−1.2579	0.34881	−1.4555	0.39541
NPDSSTTGPWCYTTDPTVR (110)
Knee_JSN_person_Pheno_P02749_APOH_	0.21115	0.20133	0.09183	0.30327
ATFGCHDGYSLDGPEEIECTK (18)
Knee_JSN_person_Pheno_P07996_TSP1_F	−3.8365	0.27116	−4.0282	0.40996
VFGTTPEDILR (112)
Knee_JSN_person_Pheno_P01011_AACT_	−0.0089	0.24667	−0.1122	0.22344
ADLSGITGAR (8)
Knee_KL_person_Pheno_P08603_CFAH_	−1.8612	0.15652	−2.0084	0.27151
CLPVTAPENGK (35)
Knee_KL_person_Pheno_P04003_C4BPA_	−0.5958	0.26047	−0.7189	0.35023
GVGWSHPLPQCEIVK (26)
Knee_KL_person_Pheno_P00734_THRB_	−1.2649	0.35943	−1.4086	0.36133
NPDSSTTGPWCYTTDPTVR (110)
Knee_KL_person_Pheno_P02743_SAMP_	−1.0876	0.27091	−1.2492	0.26573
AYSDLSR (104)
Knee_KL_person_Pheno_P07996_TSP1_F	−3.8132	0.25135	−4.0374	0.51953
VFGTTPEDILR (112)
Knee_KL_person_Pheno_P01033_TIMP1_	−5.4698	0.473	−5.666	0.36542
GFQALGDAADIR (111)
Knee_KL_person_Pheno_P01011_AACT_	0.0924	0.26356	−0.0471	0.24199
NLAVSQVVHK (9)
Knee_KL_person_Pheno_P01011_AACT_	0.02557	0.24965	−0.1008	0.21807
ADLSGITGAR (8)
Knee_KL_person_Pheno_P00736_C1R_GL	−1.5842	0.19482	−1.6865	0.22542
TLHLK (22)
Knee_KL_person_Pheno_P00736_C1R_GY	−2.474	0.19643	−2.572	0.20978
GFYTK (23)
Knee_KL_person_Pheno_P04114_APOB_L	−1.1494	0.23247	−1.2804	0.26849
AIPEGK (15)
Knee_KL_person_Pheno_P01011_AACT_	0.662	0.25482	0.54442	0.24088
EQLSLLDR (10)
Knee_KL_person_Pheno_P05546_HEP2_F	−1.0844	0.30497	−1.2083	0.25962
TVDRPFLFLIYEHR (77)
Knee_KL_person_Pheno_P05546_HEP2_N	−0.3694	0.24506	−0.4742	0.2235
FGYTLR (76)
Knee_OST_person_Pheno_P02776_PLF4_I	−2.6369	0.35736	−2.9409	0.55151
CLDLQAPLYK (99)
Knee_OST_person_Pheno_P02775_CXCL7_	−1.556	0.31036	−1.8293	0.58207
NIQSLEVIGK (57)
Knee_OST_person_Pheno_P07358_CO8B_	−4.2578	0.31545	−4.5185	0.43523
GILNEIK (51)
Knee_OST_person_Pheno_P01008_ANT3_	5.44905	0.40304	5.69097	0.46458
ATEDEGSEQK (14)
Knee_OST_person_Pheno_P02751_FINC_	−1.6448	0.41689	−1.8881	0.56808
EYLGAICSCTCFGGQR (70)
Knee_OST_person_Pheno_P02751_FINC_I	−2.0429	0.42336	−2.2089	0.36074
GDTWSK (71)
Knee_OST_person_Pheno_P07996_TSP1_	−3.8963	0.33583	−4.1239	0.64185
FVFGTTPEDILR (112)
Knee_OST_person_Pheno_P07360_CO8G_	−4.6125	0.3262	−4.7935	0.43203
QLYGDTGVLGR (52)
Knee_OST_person_Pheno_Q92954_PRG4_	−4.4216	0.33795	−4.6091	0.37538
DQYYNIDVPSR (102)
Knee_OST_person_Pheno_Q92954_PRG4_	−4.087	0.38455	−4.3242	0.46606
ITEVWGIPSPIDTVFTR (101)
Knee_OST_person_Pheno_P03952_KLKB1_	−1.9461	0.24557	−2.097	0.31923
VSEGNHDIALIK (87)
Knee_OST_person_Pheno_P04196_HRG_	−2.4096	0.26503	−2.5521	0.35484
YWNDCEPPDSR (81)
Knee_OST_person_Pheno_P08603_CFAH_	−1.9136	0.21584	−2.0406	0.29389
CLPVTAPENGK (35)
Knee_OST_person_Pheno_P04003_C4BPA_	−0.6346	0.32245	−0.804	0.33093
GVGWSHPLPQCEIVK (26)
Knee_OST_person_Pheno_P06681_CO2_D	−3.1113	0.23486	−3.2284	0.23677
GNDHSLWR (42)
ELISA markers; knee; GEE
Knee_JSN_Pheno_hapto_serum	1606.88	817.954	1107.57	624.568
Knee_JSN_Pheno_cd44_serum	150.65	27.9062	167.947	45.6712
Knee_JSN_Pheno_hemopexin_serum	1550.7	222.741	1419.74	270.341
Knee_JSN_Pheno_cd163_serum	821.961	284.381	881.345	302.904
Knee_JSN_Pheno_vitd_binding_serum	809.959	554.359	1047.53	607.144
Knee_JSN_Pheno_lcerulo_serum	6.58372	0.6484	6.80215	0.64846
Knee_JSN_Pheno_1HA_serum	5.63002	0.93682	5.79114	0.78678
Knee_JSN_Pheno_tbg_serum	15.2226	5.21314	16.5391	4.14235
Knee_JSN_Pheno_ctx_lbeta_urine1	−3.0961	0.84019	−3.2716	0.86223
Knee_JSN_Pheno_ctx1a_ctx1b_urine1	−6.0929	1.3862	−5.8491	1.3537
Knee_JSN_Pheno_cd14_serum	2581.92	660.618	2571.06	730.345
Knee_JSN_Pheno_comp_serum	1624.31	586.104	1666.57	614.388
Knee_JSN_Pheno_hapto_urine_urine1	−1.8209	1.93942	−2.104	1.75651
Knee_JSN_Pheno_hapto_urine_urine2	−11.306	2.02314	−11.512	1.87493
Knee_JSN_Pheno_ctx1a_ctx1b_urine2	−15.535	2.27897	−15.179	2.15288
Knee_JSN_Pheno_ctx_lbeta_urine2	−12.538	1.23799	−12.603	1.30309
Knee_JSN_Pheno_ctx2_urine1	−5.007	0.89173	−5.1513	0.59994
Knee_JSN_Pheno_ctx2_urine2	−14.449	1.04709	−14.485	1.13908
Knee_JSN_Pheno_ctxi_urinel	−3.6442	0.78839	−3.5905	0.75005
Knee_JSN_Pheno_ctxi_urine2	−13.087	1.59755	−12.921	1.54536
Knee_JSN_Pheno_coll3_serum	23.7345	3.40834	23.5759	3.41947
Knee_JSN_Pheno_cerulo_urine_urine2	−8.767	1.30124	−8.8112	1.25825
Knee_JSN_Pheno_kinno_serum	290.915	142.942	297.069	102.493
Knee_JSN_Pheno_cerulo_urine_urine1	0.67532	0.8622	0.52831	0.73697
Knee_KL_Pheno_lcerulo_serum_serum	6.51632	0.69437	6.87555	0.60683
Knee_KL_Pheno_hapto_serum	1579.83	865.189	1144.89	625.329
Knee_KL_Pheno_cd44_serum	155.42	35.625	167.649	42.8249
Knee_KL_Pheno_hapto_urine_urine2	−10.901	2.05767	−11.662	1.90504
Knee_KL_Pheno_vitd_binding_serum	905.41	563.381	983.619	618.757
Knee_KL_Pheno_ctx_lbeta_urine2	−12.416	1.4272	−12.628	1.20552
Knee_KL_Pheno_ctx2_urine1	−5.2374	0.91413	−5.0238	0.64339
Knee_KL_Pheno_tbg_serum	16.1841	4.51185	16.4486	4.63061
Knee_KL_Pheno_cd14_serum	2661.92	691.486	2499.09	722.077
Knee_KL_Pheno_ctx1a_ctxlb_urine2	−14.695	2.6064	−15.428	2.04488
Knee_KL_Pheno_ctxi_urine2	−12.666	1.85633	−13.046	1.42449
Knee_KL_Pheno_cerulo_urine_urine1	0.73544	0.86068	0.67915	0.82822
Knee_KL_Pheno_hapto_urine_urine1	−1.7817	1.86541	−2.1445	1.86244
Knee_KL_Pheno_ctx2_urine2	−14.334	1.12639	−14.478	1.10132
Knee_KL_Pheno_cerulo_urine_urine2	−8.3612	1.43515	−8.7819	1.22296
Knee_KL_Pheno_cd163_serum	890.667	308.31	827.306	270.548
Knee_KL_Pheno_comp_serum	1805.5	696.916	1603.06	549.065
Knee_KL_Pheno_kinno_serum	306.848	150.277	290.86	106.542
Knee_KL_Pheno_1HA_serum	5.59659	0.86984	5.64353	0.79587
Knee_KL_Pheno_hemopexin_serum	1589.28	253.919	1477.24	372.807
Knee_KL_Pheno_ctx1a_ctx1b_urine1	−5.6508	1.58585	−5.9889	1.29875
Knee_KL_Pheno_coll3_serum	24.4897	3.53911	23.6109	3.59543
Knee_KL_Pheno_ctxi_urine1	−3.5684	0.91432	−3.5921	0.69628
Knee_KL Pheno ctx lbeta urine1	−3.2978	0.91008	−3.169	0.85943
Knee_OST_Pheno_1HA_serum	5.47246	0.88898	5.76204	0.70844
Knee_OST_Pheno_hapto_serum	1386.54	779.479	1103.49	611.724
Knee_OST_Pheno_cd163_serum	794.197	276.997	892.772	285.564
Knee_OST_Pheno_cd44_serum	158.26	37.0897	172.328	44.3459
Knee_OST_Pheno_tbg_serum	15.7036	5.16053	16.9976	4.04741
Knee_OST_Pheno_cerulo_urine_urine2	−8.8172	1.30041	−8.5072	1.26027
Knee_OST_Pheno_ctx1a_ctx1b_urine1	−6.0859	1.30181	−5.7273	1.44766
Knee_OST_Pheno_ctxi_urine1	−3.6674	0.79932	−3.5154	0.72453
Knee_OST_Pheno_comp_serum	1565.03	476.409	1736.8	679.457
Knee_OST_Pheno_hemopexin_serum	1546.7	357.536	1481.31	359.671
Knee_OST_Pheno_ctx1a_ctx1b_urine2	−15.493	2.15187	−14.962	2.26187
Knee_OST_Pheno_ctxi_urine2	−13.101	1.60025	−12.784	1.49467
Knee_OST_Pheno_cerulo_urine_urine1	0.61404	0.8256	0.77457	0.8674
Knee_OST_Pheno_vitd_binding_serum	1004.24	643.216	934.067	589.759
Knee_OST_Pheno_cd14_serum	2535.59	667.835	2514.55	734.427
Knee_OST_Pheno_ctx_lbeta_urine2	−12.619	1.34067	−12.536	1.23706
Knee_OST_Pheno_ctx2_urine2	−14.451	1.0706	−14.383	1.10635
Knee_OST_Pheno_kinno_serum	291.233	113.953	283.484	114.836
Knee_OST_Pheno_hapto_urine_urine2	−11.556	2.16059	−11.38	1.86644
Knee_OST_Pheno_ctx2_urine1	−5.0195	0.72808	−5.1065	0.63031
Knee_OST_Pheno_lcerulo_serum	6.75596	0.61035	6.83172	0.67827
Knee_OST_Pheno_hapto_urine_urinel	−2.0969	1.86438	−2.0368	1.86182
Knee_OST_Pheno_coll3_serum	23.707	3.64127	24.2162	3.52646
Knee_OST_Pheno_ctx_lbeta_urine1	−3.177	0.88927	−3.2504	0.90828
ELISA markers; knee; person
Knee_JSN_person_Pheno_hapto_serum	1603.35	829.077	996.543	480.362
Knee_JSN_person_Pheno_cd44_serum	152.974	28.6622	171.635	47.738
Knee_JSN_person_Pheno_lcerulo_serum_s	6.55994	0.64513	6.88947	0.63355
erum
Knee_JSN_person_Pheno_vitd_binding_ser	824.727	563.488	1075.84	606.629
UM
Knee_JSN_person_Pheno_hemopexin_seru	1529.95	229.779	1415.63	301.975
m
Knee_JSN_person_Pheno_cd163_serum	814.817	288.557	887.965	320.78
Knee_JSN_person_Pheno_tbg_serum	15.2671	4.85939	16.9759	4.25425
Knee_JSN_person_Pheno_hapto_urine_urin	−1.6798	1.87231	−2.2666	1.73415
e1
Knee_JSN_person_Pheno_hapto_urine_urin	−11.113	1.984	−11.68	1.81184
e2
Knee_JSN_person_Pheno_1HA_serum	5.61905	0.89867	5.78457	0.75377
Knee_JSN_person_Pheno_cd14_serum	2642.04	662.952	2560.2	739.359
Knee_JSN_person_Pheno_ctx2_urine1	−4.9732	0.82227	−5.2162	0.63565
Knee_JSN_person_Pheno_ctx_lbeta_urine1	−3.0907	0.88169	−3.313	0.90923
Knee_JSN_person_Pheno_ctx2_urine2	−14.351	1.11176	−14.602	1.10084
Knee_JSN_person_Pheno_ctx_lbeta_urine2	−12.469	1.28896	−12.694	1.26058
Knee_JSN_person_Pheno_cerulo_urine_uri	0.6837	0.90088	0.46547	0.656
ne1
Knee_JSN_person_Pheno_cerulo_urine_uri	−8.6944	1.39072	−8.9319	1.11398
ne2
Knee_JSN_person_Pheno_ctx1a_ctx1b_urin	−6.0306	1.43529	−5.8607	1.32367
e1
Knee_JSN_person_Pheno_ctxi_urine2	−13.005	1.6566	−12.977	1.4598
Knee_JSN_person_Pheno_comp_serum	1679.54	599.743	1633.04	649.199
Knee_JSN_person_Pheno_ctx1a_ctx1b_urin	−15.409	2.3739	−15.236	2.07455
e2
Knee_JSN_person_Pheno_ctxi_urine1	−3.6268	0.81623	−3.5992	0.73473
Knee_JSN_person_Pheno_coll3_serum	23.6763	3.52326	23.7218	3.58698
Knee_JSN_person_Pheno_kinno_serum	291.262	138.289	297.648	105.434
Knee_KL_person_Pheno_hapto_serum	1600.51	862.418	1130.33	601.341
Knee_KL_person_Pheno_lcerulo_serum_se	6.58129	0.69665	6.87527	0.62344
rum
Knee_KL_person_Pheno_vitd_binding_seru	863.701	542.634	1040.08	651.991
m
Knee_KL_person_Pheno_cd44_serum	157.132	40.4298	167.18	40.7173
Knee_KL_person_Pheno_hapto_urine_urine	−10.809	2.00864	−11.637	1.93665
2
Knee_KL_person_Pheno_ctx_lbeta_urine2	−12.378	1.37454	−12.639	1.23375
Knee_KL_person_Pheno_tbg_serum	16.0036	4.84965	16.8026	4.64742
Knee_KL_person_Pheno_cd14_serum	2636.39	682.689	2477.81	723.986
Knee_KL_person_Pheno_hapto_urine_urine	−1.6427	1.80641	−2.1936	1.86267
1
Knee_KL_person_Pheno_cd163_serum	863.405	311.188	828.269	274.802
Knee_KL_person_Pheno_ctxi_urine2	−12.642	1.7623	−12.988	1.4886
Knee_KL_person_Pheno_ctx1a_ctx1b_urin	−14.701	2.49009	−15.301	2.14653
e2
Knee_KL_person_Pheno_ctx2_urine2	−14.272	1.10587	−14.488	1.16679
Knee_KL_person_Pheno_cerulo_urine_urin	−8.3301	1.46313	−8.7435	1.26323
e2
Knee_KL_person_Pheno_comp_serum	1794.71	730.993	1613.72	541.515
Knee_KL_person_Pheno_ctx2_urine1	−5.1635	0.88338	−5.0676	0.67814
Knee_KL_person_Pheno_cerulo_urine_urin	0.77872	0.94037	0.68227	0.79315
e1
Knee_KL_person_Pheno_hemopexin_seru	1563.53	274.345	1474.19	372.59
m
Knee_KL_person_Pheno_ctxlbeta_urine1	−3.2369	0.88412	−3.2204	0.87136
Knee_KL_person_Pheno_coll3_serum	24.4696	3.42393	23.524	3.72776
Knee_KL_person_Pheno_kinno_serum	300.616	147.734	295.027	115.194
Knee_KL_person_Pheno_ctx1a_ctx1b_urin	−5.6758	1.5659	−5.8843	1.34559
e1
Knee_KL_person_Pheno_1HA_serum	5.66124	0.89792	5.59917	0.76519
Knee_KL_person_Pheno_ctxi_urine1	−3.5329	0.8692	−3.5709	0.72404
Knee_OST_person_Pheno_hapto_serum	1322.97	762.45	1107.02	566.902
Knee_OST_person_Pheno_ctxi_urine1	−3.658	0.80638	−3.393	0.69443
Knee_OST_person_Pheno_hemopexin_seru	1550.32	394.039	1417.66	240.545
m
Knee_OST_person_Pheno_ctx1a_ctx1b _uri	−6.0124	1.35475	−5.5033	1.53767
ne1
Knee_OST_person_Pheno_tbg_serum	16.1968	4.99907	17.3638	4.31623
Knee_OST_person_Pheno_ctxi_urine2	−13.038	1.6066	−12.509	1.54218
Knee_OST_person_Pheno_cerulo_urine_uri	−8.754	1.31091	−8.2981	1.39088
ne2
Knee_OST_person_Pheno_ctx1a_ctx1b _un	−15.355	2.20092	−14.61	2.43526
ne2
Knee_OST_person_Pheno_comp_serum	1590.05	483.543	1772.73	757.186
Knee_OST_person_Pheno_ctx_lbeta_urine2	−12.637	1.33736	−12.307	1.18124
Knee_OST_person_Pheno_cd14_serum	2592.97	695.535	2380.79	727.765
Knee_OST_person_Pheno_cerulo_urine_uri	0.62306	0.85367	0.83754	0.8564
ne1
Knee_OST_person_Pheno_ctx2_urine2	−14.458	1.06195	−14.214	1.25659
Knee_OST_person_Pheno_1HA_serum	5.56184	0.86152	5.67931	0.66442
Knee_OST_person_Pheno_kinno_serum	301.275	116.42	271.699	136.201
Knee_OST_person_Pheno_vitd_binding_ser	1018.49	635.383	918.883	609.159
um
Knee_OST_person_Pheno_coll3_serum	23.8769	3.69506	23.7534	3.49101
Knee_OST_person_Pheno_cd44_serum	161.529	39.9551	167.546	43.4789
Knee_OST_person_Pheno_cd163_serum	834.289	292.216	837.768	287.367
Knee_OST_person_Pheno_ctx_lbeta_urine1	−3.2442	0.9187	−3.183	0.81553
Knee_OST_person_Pheno_hapto_urine _un	−11.442	2.13675	−11.4	1.75876
ne2
Knee_OST_person_Pheno_hapto_urine _un	−2.0095	1.86826	−2.2442	1.8478
ne1
Knee_OST_person_Pheno_lcerulo_serum_s	6.78475	0.6028	6.76521	0.72645
erum
Knee_OST_person_Pheno_ctx2_urine1	−5.0812	0.75831	−5.0836	0.7278

2a. Non-Depleted Serum Proteomics (Analysis Using Calculated Ratios)2a.1. Progression Analysis

The most significant results are summarized in Tables 3-4 below; these are the results on which the multimarker AUC calculations are based. These Tables list the AUCs achieved in ROC curves for the biomarker alone—AUC_BM, and the AUC for the full model achieved for the biomarker with demographics (age, gender, BMI and cohort)—AUC_full, and their corresponding p values.

In brief, markers were identified that could identify JSN progression modestly (best single biomarker AUC 0.65; multimarker AUC 0.55) and OST more strongly (best single biomarker AUC 0.67; multimarker AUC 0.61). Considering only the biomarker (peptide) capability and prediction of knee level progression, 6 peptides achieved AUC≥0.65 for JSN progression (PGCA, APOH, AACT ×3 peptides, and PHLD), 2 peptides for OST progression (PLF4 and CSCL7), and 6 peptides for KL progression (CFAH, SAMP, HEP2 ×2 peptides, C1R, APOB). The multimarker AUCs for person level progression were somewhat stronger than for knee level progression (Table 3 compared with Table 4) with multimarker AUC 0.67 for OST progression. Considering only the biomarker (peptide) capability and prediction of person level progression, the highest single biomarker AUC was achieved for CXCL7 with AUC 0.70 for prediction of OST. AUC_BM in the tables below refers to the AUC with the biomarker alone; AUC_full in the tables refers to the AUC calculated when the biomarker and the age, gender and BMI of the subject were considered.

TABLE 3
Top 10 peptides for prediction of knee level Progression from
non-depleted serum. Multimarker AUCs based on top 8 peptides.
	Peptides		Peptides		Peptides
	predicting		predicting		predicting
	OST (SEQ		JSN (SEQ		KL (SEQ
	ID NO: X)		ID NO: X)		ID NO: X)
OST AUCs and	Multi-	JSN AUCs and	Multi-	KL AUCs and	Multi-
p values for	marker	p values for	marker	p values for	marker
peptides	AUC = 0.61	peptides	AUC = 0.55	peptides	AUC = 0.50
0.67BM/0.69full	PLF4 (99)	0.62BM/0.70full	PGCA (96)	0.67BM/0.82full	CFAH (35)
(p = 0.001/0.002)		(p = 0.019/0.076)		(p = 0.003/0.004)
0.67BM/0.67full	CXCL7 (57)	0.62BM/0.73full	APOH (18)	0.65BM/0.79full	SAMP (104)
(p = 0.0099/0.013)		(p = 0.029/0.016)		(p = 0.009/0.018)
0.62BM/0.63full	ANT3 (14)	0.61BM/0.70full	SAMP (104)	0.62BM/0.80full	TSP1 (112)
(p = 0.012/0.006)		(p = 0.011/0.027)		(p = 0.016/0.029)
0.62BM/0.65full	AACT (8)	0.65BM/0.71full	AACT (10)	0.66BM/0.79full	HEP2 (76)
(p = 0.008/0.003)		(p = 0.029/0.044)		(p = 0.021/0.138)
0.63BM/0.65full	AACT (10)	0.64BM/0.71full	AACT (8)	0.65BM/0.80full	C1R (22)
(p = 0.01/0.002)		(p = 0.021/0.023)		(p = 0.022/0.035)
0.63BM/0.62full	THRB (110)	0.63BM/0.65full	AACT (9)	0.65BM/0.77full	APOB (15)
(p = 0.031/0.053)		(p = 0.01/0.002)		(p = 0.019/0.073)
0.61BM/0.63full	AACT (9)	0.63BM/0.71full	CFAH (35)	0.63BM/0.76full	FINC (71)
(p = 0.017/0.008)		(p = 0.015/0.021)		(p = 0.039/0.74)
0.59BM/0.62full	ITIH4 (86)	0.66BM/0.69full	PHLD (98)	0.66BM/0.80full	HEP2 (77)
(p = 0.0397/0.040)		(p = 0.065/0.075)		(p = 0.061/0.173)
0.59BM/0.62full	CO8B (51)	0.59BM/0.68full	TSP1 (112)	0.61BM/0.77full	PGCA (96)
(p = 0.061/0.063)		(p = 0.040/0.139)		(p = 0.072/0.439)
0.60BM/0.62full	PLMN (100)	0.64BM/0.71full	THRB (110)	0.62BM/0.77full	FINC (70)
(p = 0.046/0.044)		(p = 0.068/0.041)		(p = 0.044/0.33)

These multimarker AUCs are based on the top 8 peptides and cross-validated as described in the statistical methods. BM is the AUC for the biomarker alone; full is the AUC for the biomarker plus demographics (age, gender, and BMI).

TABLE 4
Top 8 peptides for prediction of person level Progression from non-depleted serum.
	Peptides		Peptides		Peptides
	predicting		predicting		predicting
	OST (SEQ		JSN (SEQ		KL (SEQ
	ID NO: X)		ID NO: X)		ID NO: X)
OST AUCs and	Multi-	JSN AUCs and	Multi-	KL AUCs and	Multi-
p values for	marker	p values for	marker	p values for	marker
peptides	AUC = 0.67	peptides	AUC = 0.50	peptides	AUC = 0.57
0.69BM/0.70full	CO8B (51)	0.63BM/0.70full	TSP1 (112)	0.68BM/0.69full	CFAH (35)
(p = 0.003/0.007)		(p = 0.020/0.038)		(p = 0.005/0.005)
0.70BM/0.71full	PLF4 (99)	0.66BM/0.70full	CFAH (35)	0.62BM/0.64full	TSP1 (112)
(p = 0.004/0.005)		(p = 0.025/0.015)		(p = 0.014/0.044)
0.66BM/0.67full	PRG4 (101)	0.69BM/0.69full	THRB (110)	0.67BM/0.67full	SAMP (104)
(p = 0.017/0.059)		(p = 0.025/0.027)		(p = 0.007/0.038)
0.64BM/0.66full	PRG4 (102)	0.59BM/0.68full	HRG (81)	0.65BM/0.68full	APOB (15)
(p = 0.024/0.055)		(p = 0.054/0.011)		(p = 0.027/0.128)
0.64BM/0.71full	ANT3 (14)	0.62BM/0.68full	APOH (18)	0.64BM/0.67full	AACT (9)
(p = 0.018/0.019)		(p = 0.048/0.029)		(p = 0.015/0.049)
0.64BM/0.66full	C4BPA (26)	0.62BM/0.67full	AACT (8)	0.64BM/0.66full	AACT (8)
(p = 0.029/0.122)		(p = 0.064/0.042)		(p = 0.017/0.061)
0.71BM/0.73full	CXCL7 (57)	0.66BM/0.67full	PHLD (98)	0.66BM/0.68full	C1R (22)
(p = 0.006/0.007)		(p = 0.054/0.080)		(p = 0.032/0.083)
0.63BM/0.66full	C4BPA (25)	0.60BM/0.67full	AACT (9)	0.66BM/0.67full	C1R (23)
(p = 0.051/0.187)		(p = 0.079/0.059)		(p = 0.031/0.095)

These multimarker AUCs are based on the top 8 peptides and cross-validated as described in the statistical methods. BM is the AUC for the biomarker alone; full is the AUC for the biomarker plus demographics (age, gender, and BMI).2a.2. Diagnostic Analysis—

By multimarker cross-validated AUCs, markers were identified that could diagnose quite strongly at a knee level and person level, all definitions of Osteoarthritis including JSN (AUC 0.71 knee level, 0.66 person level), OST (AUC 0.70 knee level and person level) and KL grade (0.77 knee level and 0.74 person level). Considering only the biomarker (peptide) capability and prediction of knee level diagnosis, 2 peptides achieved AUC≥0.65 for a JSN diagnosis (CRTAC1 (also denoted as CRAC1) and CO5), 2 peptides for an OST diagnosis (CRTAC1 ×2 peptides), and 3 peptides for a KL based diagnosis (CRTAC1 ×2 peptides and SHBG). Considering only the biomarker (peptide) capability and prediction of person-level diagnosis, 5 peptides achieved AUC≥0.65 for a JSN diagnosis (CRTAC1, C4BPA, LUM, CO5 and PCOC1), 1 peptide for an OST diagnosis (CRTAC1 peptides), and all 8 top peptides for a KL based diagnosis (CRTAC1 ×2 peptides, PCOC1, CO8G, LUM, COMP, CO6A3 and CO5). The highest single biomarker AUCs for diagnosis were achieved for CRTAC1 (peptide GVASLFAGR) for all definitions of osteoarthritis with AUCs for the biomarker alone ranging from 0.67-0.71 and AUCs ranging form 0.80-0.88 with addition of demographics (age, gender and BMI). One Q9NQ79_CRAC1 (CRTAC1 encoded) peptide, corresponding to Cartilage Acidic Protein 1 (aliases include ASPIC and CEP-68), an extracellular matrix protein found in cartilage, bone and lung, passed a false discovery rate (FDR) threshold (that accounts for multiple testing) of <0.01 as a diagnostic of osteoarthritis based on OST or KL grade (knee and person level) and for JSN (knee level).

TABLE 5
Top 8 peptides for prediction of knee level Diagnosis from non-depleted serum.
	Peptides		Peptides		Peptides
	diagnosing		diagnosing		diagnosing
	OST (SEQ		JSN (SEQ		KL (SEQ
	ID NO: X)		ID NO: X)		ID NO: X)
OST AUCs and	Multi-	JSN AUCs and	Multi-	KL AUCs and	Multi
p values for	marker	p values for	marker	p values for	marker
peptides	AUC = 0.70	peptides	AUC = 0.71	peptides	AUC = 0.77
0.71 BM /0.82 full	CRAC1 (1)	0.68 BM /0.85 full	CRAC1 (1)	0.74 BM /0.88 full	CRAC1 (1)
(p = 1.60E−06/0.0001)		(p = 1.45E−05/0.001)		(p = 7.91E−07/6.77E−05)
0.66 BM /0.80 full	CRAC1 (2)	0.61BM/0.83full	CXCL7 (57)	0.67 BM /0.86 full	CRAC1 (2)
(p = 5.77E−05/0.0007)		(p = 0.0002/0.004)		(p = 6.61E−05/0.0006)
0.62BM/0.77full	COMP (54)	0.63BM/0.83full	C4BPA (25)	0.65BM/0.83full	SHBG (105)
(p = 0.005/0.036)		(p = 0.001/0.029)		(p = 0.003/0.049)
0.63BM/0.77full	CO6A3 (49)	0.63BM/0.83full	COMP (54)	0.64BM/0.84full	COMP (54)
(p = 0.015/0.149)		(p = 0.002/0.019)		(p = 0.003/0.017)
0.63BM/0.77full	SHBG (105)	0.63BM/0.83full	LUM (90)	0.62BM/0.85full	CO8G (52)
(p = 0.009/0.0997)		(p = 0.003/0.054)		(p = 0.004/0.016)
0.59BM/0.76full	PCOC1 (95)	0.65BM/0.84full	CO5 (44)	0.63BM/0.83full	PCOC1 (95)
(p = 0.016/0.245)		(p = 0.008/0.0515)		(p = 0.006/0.118)
0.57BM/0.77full	CO8G (52)	0.61BM/0.83full	TIMP1 (111)	0.64BM/0.83full	CO6A3 (49)
(p = 0.052/0.203)		(p = 0.004/0.018)		(p = 0.015/0.120)
0.58BM/0.76full	LUM (90)	0.64BM/0.83full	C4BPA (26)	0.62BM/0.83full	LUM (90)
(p = 0.059/0.602)		(p = 0.005/0.047)		(p = 0.007/0.153)

These multimarker AUCs are based on the top 8 peptides and cross-validated as described in the statistical methods. Data above in bold denote results passing an FDR multiple testing threshold of 0.05 to 0.01. BM is the AUC for the biomarker alone; full is the AUC for the biomarker plus demographics (age, gender, and BMI).

TABLE 6
Top 8 peptides for prediction of person level Diagnosis from non-depleted serum.
	Peptides		Peptides		Peptides
	diagnosing		diagnosing		diagnosing
	OST (SEQ		JSN (SEQ		KL (SEQ
	ID NO: X)		ID NO: X)		ID NO: X)
OST AUCs and	Multi-	JSN AUCs and	Multi-	KL AUCs and	Multi-
p values for	marker	p values for	marker	p values for	marker
peptides	AUC = 0.70	peptides	AUC = 0.66	peptides	AUC = 0.74
0.74 BM /0.84 full	CRAC1 (1)	0.67BM/0.80full	CRAC1 (1)	0.76 BM /0.88 full	CRAC1 (1)
(p = 1.86E−06/5.43E−05)		(p = 0.0002/0.015)		(p = 8.05E−07/1.42E−05)
0.69BM/0.82full	CRAC1 (2)	0.66BM/0.78full	C4BPA (25)	071/0.85 full	CRAC1 (2)
(p = 0.00019/0.0009)		(p = 0.002/0.014)		(p = 8.37E−05/7.31E−05)
0.64BM/0.78full	COMP (54)	0.65BM/0.76full	LUM (90)	0.66BM/0.81full	PCOC1 (95)
(p = 0.002/0.026)		(p = 0.004/0.038)		(p = 0.002/0.044)
0.64BM/0.77full	CO6A3 (49)	0.67BM/0.77full	CO5 (44)	0.65BM/0.82full	CO8G (52)
(p = 0.0096/0.099)		(p = 0.007/0.047)		(p = 0.003/0.010)
0.61BM/0.77full	PCOC1 (95)	0.66BM/0.77full	PCOC1 (95)	0.65BM/0.80full	LUM (90)
(p = 0.017/0.208)		(p = 0.006/0.101)		(p = 0.008/0.139)
0.59BM/0.77full	CO8G (52)	0.60BM/0.78full	CXCL7 (57)	0.65BM/0.81full	COMP (54)
(p = 0.032/0.129)		(p = 0.001/0.012)		(p = 0.002/0.015)
0.63BM/0.77full	ACTG (11)	0.63BM/0.76full	COMP (54)	0.65BM/0.80full	CO6A3 (49)
(p = 0.031/0.180)		(p = 0.006/0.052)		(p = 0.011/0.097)
0.63BM/0.76full	CO5 (44)	0.64BM/0.78full	CO5 (45)	0.66BM/0.80full	CO5 (44)
(p = 0.095/0.799)		(p = 0.0002/0.015)		(p = 0.030/0.449)

These multimarker AUCs are based on the top 8 peptides and cross-validated as described in the statistical methods. Data above in bold denote results passing an FDR multiple testing threshold of 0.05 to 0.01. BM is the AUC for the biomarker alone; full is the AUC for the biomarker plus demographics (age, gender, and BMI).2b. Depleted Serum Proteomics (Using Original Ratios)2b.1. Progression Analysis

TABLE 7
Top 8 peptides for prediction of knee level Progression from
depleted serum. Multimarker AUCs based on top 8 peptides.
Peptides predicting	Peptides predicting	Peptides predicting
OST (SEQ ID NO: X)	JSN (SEQ ID NO: X)	KL (SEQ ID NO: X)
Multimarker	Multimarker	Multimarker
AUC = 0.50	AUC = 0.50	AUC = 0.50
knee level/0.61	knee and	knee and
person level	person level	person level
CXCL7 (57)	CO4B (43)	KNG1 (88)
C4BPA (25)	PGCA (97)	A2AP (4)
C4BPB (27)	FCN3 (67)	CO6A3 (50)
C4BPA (26)	CD44 (28)	A2AP (5)
ITIH4 (86)	TSP4(113)	HGFA (78)
PLMN (100)	TETN (107)	CO2 (41)
A2AP (5)	FINC (71)	PRG4 (102)
LYAM1 (93)	ECM1 (59)	DOPO (58)

2b.2. Diagnostic Analysis—

TABLE 8
Top 8 peptides for prediction of knee level Diagnosis from
depleted serum. Multimarker AUCs based on top 8 peptides.
Peptides diagnosing	Peptides diagnosing	Peptides diagnosing
OST (SEQ ID NO: X)	JSN (SEQ ID NO: X)	KL (SEQ ID NO: X)
Multimarker	Multimarker	Multimarker
AUC = 0.74	AUC = 0.67	AUC = 0.75
knee level/0.77	knee level/0.77	knee level/0.89
person level	person level	person level
A2AP (4)	A2AP (4)	A2AP (4)
A2AP (5)	A2AP (5)	CO2 (41)
CO2 (41)	CO2 (41)	CRAC1 (1)
FA5 (61)	FA5 (61)	CRAC1 (2)
CO5 (44)	CRAC1 (1)	FA5 (61)
CRAC1 (1)	COMP (55)	CO6A3 (49)
CRAC1 (2)	CRAC1 (2)	CO5 (44)
CERU (33)	HRG (79)	COMP (55)

TABLE 9
Top 8 peptides for prediction of person level Diagnosis from depleted serum
and using actin normalization. Multimarker AUCs based on top 8 peptides.
	Peptides		Peptides		Peptides
	diagnosing		diagnosing		diagnosing
	OST (SEQ ID		JSN (SEQ ID		KL (SEQ ID
	NO: X)		NO: X)		NO: X)
	Multimarker		Multimarker		Multimarker
	AUC = 0.77		AUC = 0.77		AUC = 0.89
	Multimarker		Multimarker		Multimarker
OST p	AUC = 0.81	JSN p	AUC = 0.78	KL p	AUC = 0.88
values for	with actin	values for	with actin	values for	with actin
peptides	normalization	peptides	normalization	peptides	normalization
7.93E−06	A2AP (4)	1.33E−07	A2AP (4)	7.83E−11	A2AP (4)
3.97E−05	A2AP (5)	5.89E−05	A2AP (5)	5.02E−06	CRAC1 (1)
4.76E−05	CRAC1 (1)	7.05E−05	CO2 (41)	1.83E−05	CO2 (41)
0.000368059	CRAC1 (2)	0.000385554	CRAC1 (1)	1.99E−05	A2AP (5)
0.000584044	CO2 (41)	0.000614719	COMP (55)	4.39E−05	CRAC1 (2)
0.001053574	FA5 (61)	0.003020118	CRAC1 (2)	0.000735498	COMP (55)
0.007678923	CO6A3 (49)	0.003279263	FA5 (61)	0.001071963	FA5 (61)
0.00822894	CO5 (44)	0.009168162	CO6A3 (49)	0.0036707	FBLN3 (64)

3. ELISA Biomarker Results

A total of 18 separate Osteoarthritis-related biomarkers were evaluated in this study:

• • Serum: CD44, CD163, collagen 3, COMP, haptoglobin, hemopexin, kininogen, ceruloplasmin, hyaluronan, TBG, Vitamin D, TSG6, CD14
  • Urine: CTXIbeta, CTXIalpha, (CTX1alpha/CTXIbeta), CTX2, ceruloplasmin, haptoglobin

The most significant results are summarized Table 10 below; these are the results for which the biomarker alone or the full model (biomarker, age, gender, BMI and cohort) achieved p<0.05. This Table lists the AUC achieved in ROC curves for the biomarker alone—AUC_BM, and the AUC for the full model achieved for the biomarker with demographics (age, gender, BMI and cohort)—AUC_full, and their corresponding p values.

In brief, progression markers for JSN include sHaptoglobin (knee and person level, (s) indicates serum measured), sCD44 (knee and person level), sHemopexin (knee level), and sCeruloplasmin (person level). We did not identify any strong progression markers for OST but we identified several strong markers of progression based upon KL grade change, including sHaptoglobin (knee and person level), sCD44 (knee level), and sCeruloplasmin (knee and person level). Of these, sHaptoglobin is the strongest progression marker.

Some strong diagnostic markers were identified for JSN (knee and person level) including sKininogen, sHyaluronan, sCD14, uCeruloplasmin and uCTX1alpha/CTX1beta ratio ((u) indicates urine measured). In addition, some strong diagnostic markers were identified for OST (knee and person level) including sKininogen, sCD14, and uCTX1alpha/CTX1beta ratio. Of these sKininogen and sHyaluronan are extremely strong diagnostic markers.

TABLE 10
Summary of most significant ELISA Results for Osteophyte (OST), Joint Space
Narrowing (JSN), and Kellgren Lawrence (KL) grade progression and diagnostic phenotypes at a
knee (black) and person-based level (red).
Biomarker
s = serum,
u = urine	PROGRESSION	DIAGNOSIS
(Relative amount	AUC	AUC for	AUC for	AUC for	AUC for	AUC for
in progressors)	for OST	JSN	KL	OST	JSN	KL
sHaptoglobin	0.62BM/	0.70BM/	0.68BM/	0.69BM/	0.59BM/
(higher)	0.63full	0.71full	0.82full	0.78full	0.85full
	(p = 0.05	(p = 0.002/	(p = 0.023/	(p = 0.0013/	(p = 0.045/
	6/0.036)	0.003)	0.021)	0.179)	0.289)
		0.75BM/	0.70BM/
		0.75full	0.68full
		(p = 0.001/	(p = 0.011/
		0.002)	0.015)
sCD44		0.59BM/	0.57BM/
(lower)		0.71full	0.81full
		(p = 0.051/	(p = 0.15/
		0.021)	0.033)
		0.59BM/
		0.65full
		(p = 0.048/
		0.050)
sHemopexin		0.65BM/
(higher)		0.71full
		(p = 0.018/
		0.04)
sKininogen				0.67BM/	0.60BM/	0.65BM/
				0.82full	0.85full	0.87full
				(p = 0.000003/	(p = 0.001/	(p = 0.00004/
				0.00012)	0.073)	0.0023)
				(0.68BM/	(0.61BM/	(0.67BM/
				0.82full)	0.79full)	0.84full)
				(p = 0.00004/	(p = 0.001/	(p = 0.00009/
				0.0005)	0.065)	0.002)
sHyaluronan	0.62BM/			0.67BM/	0.71BM/	0.71BM/
(higher)	0.65full			0.79full	0.88full	0.86full
	(p = 0.099/			(p = 0.004/	(p = 0.001/	(p = 0.001/
	0.033)			0.133)	0.016)	0.073)
					(0.72BM/	(0.72BM/
					0.80full)	0.82full)
					(p = 0.00007/	(p = 0.00016/
					0.046)	0.069)
sCD14				0.63BM/	0.65BM/	0.67BM/
(higher)				0.77full	0.85full	0.85full
				(p = 0.031/	(p = 0.008/	(p = 0.0008/
				0.157)	0.013)	0.043)
				(0.64BM/	(0.70BM/	(0.67BM/
				0.78full)	0.80full)	0.82full)
				(p = 0.018/	(p = 0.0002/	(p = 0.003/
				0.128)	0.001)	0.055)
sCD163	0.58BM/
	0.68full
	(p = 0.43/
	0.046)
sCeruloplas		0.63BM/	0.65BM/		0.59BM/	0.65BM/
min (lower)		0.66full	0.82full		0.83full	0.82full
		(p = 0.032/	(p = 0.019/		(p = 0.051/	(p = 0.019/
		0.054)	0.002)		0.078)	0.002)
			0.62BM/
			0.64full
			(p = 0.051/
			0.025)
uCeruloplas					0.63BM/
min					0.85full
(lower)[urine1]					(p = 0.007/
					0.064)
					(0.66BM/
					0.82full)
					(p = 0.002/
					0.017)
uCeruloplas				0.60BM/	0.65BM/	0.63BM/
min				0.79full	0.86full	0.85full
(lower)[urine2]				(p = 0.027/	(p = 0.004/	(p = 0.013/
				0.130)	0.129)	0.152)
					(0.68BM/	(0.61BM/
					0.82full)	0.81full)
					(p = 0.001/	(p = 0.035/
					0.031)	0.465)
uCTX1alpha/				0.62BM/	0.62BM/	0.64BM/
CTX1beta				0.79full	0.84full	0.85full
(higher)[urine1]				(p = 0.019/	(p = 0.017/	(p = 0.013/
				0.149)	0.63)	0.232)
				0.61BM/		(0.61BM/
				0.79full		0.82full)
				(p = 0.045/		(p = 0.041/
				0.556)		0.779)
uCTX1alpha/				0.62BM/	0.63BM/	0.65BM/
CTX1beta				0.80full	0.85full	0.85full
(higher)[urine2]				(p = 0.038/	(p = 0.022/	(p = 0.025/
				0.146)	0.418)	0.165)
					(0.65 BM/	(0.62BM/
					0.80full)	0.82full)
					(p = 0.010/	(p = 0.053/
					0.257)	0.518)
uCTX1					(0.63BM/	0.64BM/
(higher)[urine2]					0.79full)	0.85full
					(p = 0.028/	(p = 0.050/
					0.338)	0.224)
sVitamin D			0.58BM/		0.60BM/	0.62BM/
binding			0.62full		0.82full	0.83full
protein			(p = 0.209/		(p = 0.022/	(p = 0.0085/
(lower)			0.052)		0.797)	0.413)

AUCs not underlined are knee-based results and AUCs underlined are Person-based results. BM is the AUC for the biomarker alone; full is the AUC for the biomarker plus demographics (age, gender, BMI and cohort).Conclusions

More diagnostic than prognostic biomarkers were discovered. Non-depleted serum worked as well or better than depleted serum, therefore, the validation does not necessarily require serum depletion of the most abundant proteins. By proteomics, our strongest prognostic biomarkers were for OST and included serum CO8B and serum PLF4. By proteomics, our strongest diagnostic biomarker was serum CRTAC1 (encoding Cartilage Acidic Protein 1), which diagnosed all phenotypes of OA and passed a 1% FDR rate.

By ELISA, serum Haptoglobin was the strongest predictor of progression and predicted JSN indicative of cartilage loss. Immunoaffinity depletion of high abundance plasma proteins is frequently employed to enhance detection of lower abundance proteins in both shotgun and targeted proteomic analyses. MARS columns afford highly repeatable and efficient plasma protein depletions and a global enrichment in non-target plasma proteins of 2-4 fold. Tu et al., J Proteome Res 9(10):4982-91 (2010). We have identified serum haptoglobin to be a strong marker of knee osteoarthritis progression but it is ordinarily depleted from blood biospecimens (by both the MARS-7 and MARS-14 Human protein depletion columns by Agilent) prior to mass spectrometry analysis. Future analyses of non-depleted serum will permit addition of haptoglobin to the panel of analytes surveilled. By ELISA, serum Kininogen and serum Hyaluronan were the strongest diagnostic markers of knee OA predicting most strongly OST and JSN, respectively.

Normalization

We developed a methodology to select peptides capable of acting as normalization peptides in multiple reaction monitoring (MRM) and mass spectrometry analyses to control for intensity loading and variation of efficiency of Mars-14 protein depletion of serum or other biospecimen. We identified 14 proteins that could serve as normalization controls that are superior or equivalent to the standard methodology of normalizing to mean signal intensity. Specifically, we identified 4 normalization peptides from 4 proteins that are superior to mean intensity normalization (the standard methodology) and 21 peptides from 14 proteins that are equivalent or slightly better than standard methodology for normalization. Details are provided below. For MRM, these candidates can be run to evaluate and control for potential technical variation related to MARS-14 depletion. They might also be used in analyses of non-depleted samples to control for variation introduced by sample processing.

Table 11 below lists the proteins that are depleted by the MARS-14 column. Depletion efficiency varies across samples. This introduces variability in sample results. The standard practice is to normalize signal intensities for each peptide of interest with overall signal intensity of all peptides. Obviously, when the sample depletion has been variable, the data normalization to overall intensity will introduce variability and error in the results.

TABLE 11
List of blood proteins depleted by MARS-14.
	MARS Hu-14 proteins	Primary Protein Name
1	albumin	ALBU_BOVIN
		ALBU_HUMAN
2	IgG	IGHG3_HUMAN
3	IgA
4	transferrin	TRFE_HUMAN
5	haptoglobin	HPT_HUMAN
6	antitrypsin	A1AT_HUMAN
7	fibrinogen	FIBA_HUMAN
8	alpha2-macroglobulin	A2MG_HUMAN
9	alpha1-acid glycoprotein	—
10	IgM	—
11	apolipoprotein AI	APOA1_HUMAN
12	apolipoprotein AII	—
13	complement C3	CO3_HUMAN
14	transthyretin	TTHY_HUMAN

One method we devised was to normalize based on total mean signal intensity using all signals except those emanating from any residual amounts of the proteins in the above list of proteins (that should have been depleted but that are generally depleted with varying efficiencies)—‘targeted mean total intensity’ normalization.

Our goal was to find a peptide or protein that could serve as a normalization control, i.e. a “housekeeping protein” that would eliminate the need to normalize to total mean intensity or targeted mean total intensity described above. For a normalization peptide to benefit the analysis, it should produce better signals than normalization in the standard way using mean overall intensity. Therefore, ideally, we would like to find a normalization peptide or protein that is superior to either of these methods. The test of superiority is to compare qq plots for the sample data normalized by the targeted mean total intensity (our refinement of standard practice) vs normalized to the candidate normalization protein or peptide.

TABLE 12
Normalization peptides (total of 4 peptides) that are superior to normalization by mean
intensity.
	Peptide		Superior or
	Teller		Equivalent to
Modified Peptide Sequence	Proba-	Primary Protein	Standard Method of
(SEQ ID NO:)	bility	Name	Normalization
FVFGTTPEDILR (112)	1	TSP1_HUMAN	Superior
ALEQDLPVNIK (40)	1	CNDP1_HUMAN	Superior
SEAYNTFSER (61)	0.89	FA5_HUMAN	Superior
IALGGLLFPASNLR (105)	0.99	SHBG_HUMAN	Superior

TABLE 13
Normalization peptides (total of 21) that are slightly better or equivalent to
normalization by mean intensity.
	Peptide		Superior or
	Teller		Equivalent to
Modified Peptide Sequence	Proba-	Primary Protein	Standard Method
(SEQ ID NO:)	bility	Name	of Normalization
ICLDLQAPLYK (99)	1	PLF4_HUMAN	Equivalent
FQSVFTVTR (123)	1	C1QC_HUMAN	Equivalent
IFYNQQNHYDGSTGK (124)	1	ADIPO_HUMAN	Equivalent
EWVAIESDSVQPVPR (125)	0.96	CNDP1_HUMAN	Equivalent
SVVLIPLGAVDDGEHSQNEK	1	CNDP1_HUMAN	Slightly Better
(126)
LVPFATELHER (127)	0.66	APOA4_HUMAN	Slightly Better
VAPEEHPVLLTEAPLNPK (11)	0.84	ACTG_HUMAN	Equivalent
FTGSQPFGQGVEHATANK	1	TSP1_HUMAN	Equivalent
(143)
EFNPLVIVGLSK (62)	0.74	FA5_HUMAN	Equivalent
VLSIAQAHSPAFSCEQVR (128)	0.94	CD14_HUMAN	Equivalent
SITLFVQEDR (129)	0.99	TSP1_HUMAN	Equivalent
AEAESLYQSK (130)	0.99	K2C1_HUMAN	Equivalent
NALWHTGNTPGQVR (131)	0.95	TSP1_HUMAN	Equivalent
AIHLDLEEYR (132)	1	CNDP1_HUMAN	Equivalent
AGTLDLSLTVQGK (133)	0.99	TSP1_HUMAN	Slightly Better
EENFYVDETTVVK (134)	0.9	CBG_HUMAN	Equivalent
DNNSIITR (135)	0.61	CHLE_HUMAN	Equivalent
VVLSSGSGPGLDLPLVLGLPL	1	SHBG_HUMAN	Equivalent
QLK (136)
DNCQYVYNVDQR (137)	0.99	TSP1_HUMAN	Equivalent
LFLGALPGEDSSTSFCLNGLW	0.6	SHBG_HUMAN	Equivalent
AQGQR (138)
HNEVWHLVGITSWGEGCAQR	0.77	FA11_HUMAN	Equivalent
(139)

Overall, the following numbers of peptides were identified as normalization controls from 14 proteins: TSP-1 (6), CNDP1 (4), FA5 (2), SHBG (3), and one each for PLF4, C1Qc, ADIPO, APOA4, ACTG, CD14, K2C1, CBG, CHLE and FA11. Three additional peptides identified in our human specimen analyses also cover 3 of these proteins, including the following:

76969249 515.7786 2
(SEQ ID NO: 140)
GPDPSSPAFR
TSP1_HUMAN Thrombospondin-1 OS = Homo sapiens GN = THBS1 PE = 1 SV = 2;
(SEQ ID NO: 141)
76967646 520.31244 2HITSLEVIK PLF4_HUMAN
Platelet factor 4 OS = Homo sapiens GN = PF4 PE = 1 SV = 2;
76968171 522.26917
(SEQ ID NO: 142)
2LDVDQALNR
SHBG_HUMAN Sex hormone-binding globulin OS = Homo sapiens
GN = SHBG PE = 1 SV = 2

Additional analyses have identified two other potential normalization peptides in the Carbonic anhydrase 1 protein:

CAH1_HUMAN (peptide 8856058)
(SEQ ID NO: 144)
GGPFSDSYR
Carbonic anhydrase 1 - Homo sapiens (8848161)
(SEQ ID NO: 145)
GGPFSDSYR

Multimarker Analysis I Using Serum Biomarkers and Clinical Covariates

Analysis was performed to identify biomarkers that could add value for disease classification over and above clinical parameters. Serum biomarkers previously selected by the literature (E biomarkers) could modestly increase classification of Control vs. Disease (C v D). In contrast, the novel serum biomarkers (M biomarkers) discovered in the study could significantly increase this classification.

C v D (Control v Disease)

Clinical covariates could classify C v D moderately well (AUC, ca. 0.78). Biomarkers previously selected by the literature and measured by ELISA (E biomarkers) could increase classification (AUC, 0.81). In contrast, the novel M biomarkers discovered in the study could significantly increase classification (AUC, 0.97). Adding both E and M biomarkers to clinical covariates increased the AUC to 0.99.

N v P (Non-Progressor v Progressor, Person)

Clinical covariates were not significant. Adding E+M biomarkers improved classification to AUC=0.69.

PO1 v PO2 (Non-Progressor v Progressor, Osteophyte)

For clinical covariates, only gender was significant and classification was AUC=0.65. No E biomarkers were selected by the analysis. Addition of M biomarkers improved classification to AUC=0.72.

PJ1 v PJ2 (Non-Progressor v Progressor, JSN)

Clinical covariates were not significant. No M biomarkers were selected by the analysis. Adding E biomarkers improved classification to AUC=0.69.

PK1 v PK2 (Non-Progressor v Progressor, KL)

For clinical covariates, only cohort was significant for classification. No M biomarkers were selected by the analysis. Adding E biomarkers resulted in classification AUC=0.65.

General Methods

Biomarker Selection

For each outcome, and each set of biomarkers (class E and class M, separately), biomarker selection was performed using the lasso selection method with the R package “glmnet”, and selecting the tuning parameter using the built in 10-fold cross-validation.

ROC Analysis

For each outcome, following biomarker selection, the observations were randomly split in half to generate a training and a test set. For the training set, two logistic regression models were fit: one with clinical parameters only (CP Training), and the other with clinical parameters+ biomarkers (CP+B Training). Coefficients from the logistic regression training models were applied to the respective test sets (CP Test and CP+B Test). ROC AUC was calculated for each of the training and test sets. This process was repeated 50 times for C v D and 10 times for all other analyses and the average AUC (+/−sd) was calculated. A summary of the results is shown below in Tables 14-18.

TABLE 14
ROC AUC
Clinical		CP Training	CP Test	CP + B	CP + B Test
Outcome	Biomarkers	Set	Set	Training Set	Set
C v D	E	0.802 (0.049)	0.763 (0.059)	0.946 (0.028)	0.807 (0.075)
C v D	M	0.822 (0.043)	0.792 (0.047)	1.00 (0.000)	0.974 (0.020)
C v D	E + M	0.844 (0.050)	0.793 (0.056)	1.000 (0.000)	0.990 (0.011)

The numbers are the mean AUC over 10 random splits with standard deviations given in parenthesis.

TABLE 15
ROC AUC
Clin-
ical	Bio-	CP
Out-	mark-	Training		CP + B	CP + B Test
come	ers	Set	CP Test Set	Training Set	Set
N v P	E	0.677 (0.087)	0.441 (0.103)	0.938 (0.074)	0.641 (0.125)
N v P	M	0.689 (0.056)	0.504 (0.061)	0.884 (0.033)	0.662 (0.085)
N v P	E + M	0.725 (0.127)	0.538 (0.080)	0.930 (0.154)	0.686 (0.070)

The numbers are the mean AUC over 10 random splits with standard deviations given in parenthesis.

TABLE 16
ROC AUC
Clinical		CP		CP + B	CP + B
Out-	Bio-	Training	CP	Training	Test
come	markers	Set	Test Set	Set	Set
PO1 v	E	no E	no E	no E	no E
PO2		biomarkers	biomarkers	biomarkers	biomarkers
		selected	selected	selected	selected
PO1 v	M	0.720	0.587	0.916	0.721
PO2		(0.057)	(0.083)	(0.052)	(0.115)
PO1 v	E + M	no E	no E	no E	no E
PO2		biomarkers	biomarkers	biomarkers	biomarkers
		selected	selected	selected	selected

The numbers are the mean AUC over 10 random splits with standard deviations given in parenthesis.

TABLE 17
ROC AUC
		CP		CP + B	CP + B
Clinical	Bio-	Training		Training	Test
Outcome	markers	Set	CP Test Set	Set	Set
PJ1 v PJ2	E	0.681	0.459	0.971	0.686
		(0.063)	(0.094)	(0.040)	(0.127)
PJ1 v PJ2	M	no M	no M	no M	no M
		biomarkers	biomarkers	biomarkers	biomarkers
		selected	selected	selected	selected
PJ1 v PJ2	E + M	no M	no M	no M	no M
		biomarkers	biomarkers	biomarkers	biomarkers
		selected	selected	selected	selected

The numbers are the mean AUC over 10 random splits with standard deviations given in parenthesis.

TABLE 18
ROC AUC
		CP		CP + B	CP + B
Clinical	Bio-	Training		Training	Test
Outcome	markers	Set	CP Test Set	Set	Set
PK1 v PK2	E	0.700	0.549	0.836	0.646
		(0.053)	(0.102)	(0.076)	(0.094)
PK1 v PK2	M	no M	no M	no M	no M
		biomarkers	biomarkers	biomarkers	biomarkers
		selected	selected	selected	selected
PK1 v PK2	E + M	no M	no M	no M	no M
		biomarkers	biomarkers	biomarkers	biomarkers
		selected	selected	selected	selected

The numbers are the mean AUC over 10 random splits with standard deviations given in parenthesis.Additional Information

For the C v D analysis, 19 M biomarkers were selected by the lasso method. Additional analysis was performed to identify the M biomarkers in the selection path and to test the chain of biomarkers in the path. Using 10× repeat halves sampling, it was observed that the first two biomarkers, CRAC1 (SEQ ID NO: 1) and A2AP (SEQ ID NO: 4) from depleted samples, gave an AUC of 0.948. R-scripts containing the outputs from the logistic regression analyses contain additional information for the biomarkers used in the models. The relative p-values in each model may be used to select the more significant biomarkers for that model whereby the lower p-values indicate more significant biomarkers in the model.

I. Analysis for C/D Outcomes

Y=1 if D; Y=0 if C;

There are totally 126 observations with no missing in Y (89 with Y=1, 70.6%). Clinical covariates: gender, age, bmi.

I.1. Analysis Using Only Clinical Covariates.

We fit a logistic regression, and found that age and bmi are significant with p-values 0.0027 and 0.0002, respectively. The AUC (i.e. area under the ROC) is 0.7756.

I.2. Analysis Using Clinical Covariates and E Markers.

There are 19 E markers and 96 observations with complete E markers (63 with Y=1, 65.6%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 6 E markers: HA, kinno, vitd_binding, coll3, hemopexin, cd14. The AUC is 0.9192.

To evaluate the value of added E markers for prediction of C/D outcomes, we randomly split the data into half training and half testing datasets, and did this 50 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 6 E markers. Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Tables 14-18. The numbers are the mean AUC over 50 random splits with standard deviations given in parenthesis.

I.3. Analysis Using Clinical Covariates and M Markers.

There are 238 M makers and 110 observations with complete M markers (77 with Y=1, 70.0%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 19 M markers: TENX (SEQ ID NO: 106), FCGBP (SEQ ID NO: 66), C4BPB (SEQ ID NO: 27), A2AP (SEQ ID NO: 4), APOE (SEQ ID NO: 17), C1QC (SEQ ID NO: 20), CO6A3 (SEQ ID NO: 50), CRAC1 (SEQ ID NO: 1), FA5 (SEQ ID NO: 61) from depleted samples; and GELS (SEQ ID NO: 72), CXCL7 (SEQ ID NO: 57), ACTG (SEQ ID NO: 11), CFAI (SEQ ID NO: 37), CO5 (SEQ ID NO: 46), CO6A3 (SEQ ID NO: 49), CO8G (SEQ ID NO: 52), CRAC1 (SEQ ID NO: 1), FINC (SEQ ID NO: 70), PCOC1 (SEQ ID NO: 95) from nondepleted samples. The AUC is 1.

To evaluate the value of added M markers for prediction of C/D outcomes, we randomly split the data into half training and half testing datasets, and did this 50 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 19 M markers (here we use logistic regression with ridge penalty since some M markers may be highly correlated). Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 14.

I.4. The Selection Path for M Markers and Associated AUC.

The selected 19 M markers in order are: CRAC1 (SEQ ID NO: 1) (nondepleted), A2AP (SEQ ID NO: 4) (depleted), CO8G (SEQ ID NO: 52) (nondepleted), CXCL7 (SEQ ID NO: 57) (nondepleted), CO5 (SEQ ID NO: 46) (nondepleted), FCGBP (SEQ ID NO: 66) (depleted), PCOC1 (SEQ ID NO: 95) (nondepleted), CFAI (SEQ ID NO: 37) (nondepleted), CO6A3 (SEQ ID NO: 50) (depleted), GELS (SEQ ID NO: 72) (nondepleted), C4BPB (SEQ ID NO: 27) (depleted), CO6A3 (SEQ ID NO: 49) (nondepleted), ACTG (SEQ ID NO: 11) (nondepleted), CRAC1 (SEQ ID NO: 1) (deplated), FINC (SEQ ID NO: 70) (nondepleted), C1QC (SEQ ID NO: 20) (depleted), TENX (SEQ ID NO: 106) (depleted), APOE (SEQ ID NO: 17) (depeleted), FA5 (SEQ ID NO: 61) (depleted).

To evaluate the value of top k (k=1, 2, . . . , 19) selected M markers for prediction of C/D outcomes, we randomly split the data into half training and half testing datasets, and did this 10 times. The mean AUCs for testing data are given by: 0.8944272 0.9478328 0.9428793 0.9447368 0.9346749 0.9524768 0.9521672 0.9517028 0.9572755 0.9571207 0.9643963 0.9659443 0.9664087 0.9721362 0.9724458 0.9752322 0.971517 0.976161 0.977709. This suggests that the top 2 selected M markers can already improve the prediction ability sufficiently.

I.5. Analysis Using Clinical Covariates, E and M Markers.

There are 84 observations with complete E and M markers (54 with Y=1, 64.3%). The AUC based on clinical covariates plus the selected 6 E markers and 19 M markers is 1. To evaluate the value of added E and M markers for prediction of C/D outcomes, we randomly split the data into half training and half testing datasets, and did this 50 times. The results are summarized in Table 14.

II. Analysis for N/P Outcomes

Y=1 if P; Y=0 if N;

There are totally 89 complete observations with P/N (66 with Y=1, 74.2%). Clinical covariates: cohort id, gender, age, bmi.

II.1. Analysis Using Only Clinical Covariates.

We fit a logistic regression, and found that none of them are significant. The AUC is 0.5975.

II.2. Analysis Using Clinical Covariates and E Markers.

There are 63 complete observations (47 with Y=1, 74.6%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 6 E markers: cd163, hapto, coll3, hemopexin, ctx2, cd14. The AUC is 0.8484.

To evaluate the value of added E markers for prediction of P/N outcomes, we randomly split the data into half training and half testing datasets, and did this 10 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 6 E markers. Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 15. The numbers are the mean AUC over 10 random splits with standard deviations given in parenthesis.

II.3. Analysis Using Clinical Covariates and M Markers.

There are 77 complete observations (59 with Y=1, 76.6%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 4 M markers: CO8B (SEQ ID NO: 51), CFAH (SEQ ID NO: 35), CRAC1 (SEQ ID NO: 1), HRG (SEQ ID NO: 81) from depleted samples. The AUC is 0.8013.

To evaluate the value of added M markers for prediction of P/N outcomes, we randomly split the data into half training and half testing datasets, and did this 10 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 4 M markers. Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 15.

II.4. Analysis Using Clinical Covariates, E and M Markers.

There are 54 complete observations (41 with Y=1, 75.9%). The AUC based on clinical covariates plus the selected 6 E markers and 4 M markers is 0.8949. To evaluate the value of added E and M markers for prediction of P/N outcomes, we randomly split the data into half training and half testing datasets, and did this 10 times. The results are summarized in Table 15.

III. Analysis for P/O Outcomes

Y=1 if PO2; Y=0 if PO1;

There are totally 82 complete observations with PO1/PO2 (54 with Y=1, 65.85%). Clinical covariates: cohort id, gender, age, bmi.

III.1. Analysis Using Only Clinical Covariates.

We fit a logistic regression, and found that only gender is significant. The AUC is 0.6548.

III.2. Analysis Using Clinical Covariates and E Markers.

There are 61 complete observations (41 with Y=1, 67.2%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. None of the E markers are selected.

III.3. Analysis Using Clinical Covariates and M Markers.

There are 71 complete observations (47 with Y=1, 66.2%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 5 M markers: LYAM1 (SEQ ID NO: 93) from depleted samples, KLKB1 (SEQ ID NO: 87), CXCL7 (SEQ ID NO: 57), CO8B (SEQ ID NO: 51), ANT3 (SEQ ID NO: 14) from nondepleted samples. The AUC is 0.8555.

To evaluate the value of added M markers for prediction of P/O outcomes, we randomly split the data into half training and half testing datasets, and did this 10 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 5 M markers. Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 16.

IV. Analysis for P/J Outcomes

Y=1 if PJ2; Y=0 if PJ1;

There are totally 74 complete observations with PJ1/PJ2 (40 with Y=1, 54.1%). Clinical covariates: cohort id, gender, age, bmi.

IV.1. Analysis Using Only Clinical Covariates.

We fit a logistic regression, and found that none of them are significant. The AUC is 0.6279.

IV.2. Analysis Using Clinical Covariates and E Markers.

There are 52 complete observations (27 with Y=1, 51.9%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 6 E markers: hapto, vitd_binding, cd44, hemopexin, cerulo_serum, ctx2. The AUC is 0.9185.

To evaluate the value of added E markers for prediction of P/N outcomes, we randomly split the data into half training and half testing datasets, and did this 10 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 6 E markers. Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 17.

IV.3. Analysis Using Clinical Covariates and M Markers.

There are 66 complete observations (37 with Y=1, 56.1%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. None of the M markers are selected.

V. Analysis for P/K Outcomes

Y=1 if PK2; Y=0 if PK1;

There are totally 89 complete observations with PK1/PK2 (31 with Y=1, 34.8%). Clinical covariates: cohort id, gender, age, bmi.

V.1. Analysis Using Only Clinical Covariates.

We fit a logistic regression, and found that only cohort id is significant. The AUC is 0.7269.

V.2. Analysis Using Clinical Covariates and E Markers.

There are 63 complete observations (18 with Y=1, 28.6%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 1 E marker: hapto. The AUC is 0.7605.

To evaluate the value of added E markers for prediction of P/N outcomes, we randomly split the data into half training and half testing datasets, and did this 10 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 1 E marker. Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 18.

V.3. Analysis Using Clinical Covariates and M Markers.

There are 77 complete observations (28 with Y=1, 36.4%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. None of the M markers are selected.

Multimarker Analysis II Using Serum Biomarkers and Clinical Covariates

Analysis Based on M Markers: M141-M257

I. Analysis for C/D Outcomes

Y=1 if D; Y=0 if C;

I.0. Selection Based on M Markers Only.

There are totally 118 observations with no missing in Y (83 with Y=1, 70.3%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 5 M markers: CXCL7 (SEQ ID NO: 57), CO6A3 (SEQ ID NO: 49), CO8G (SEQ ID NO: 52), CRAC1 (SEQ ID NO: 1), COMP (SEQ ID NO: 54) from nondepleted samples.

I.1. Analysis Using Clinical Covariates and M Markers.

To evaluate the value of added M markers for prediction of C/D outcomes, we randomly split the data into half training and half testing datasets, and did this 50 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 5 M markers (here we use logistic regression with ridge penalty since some M markers may be highly correlated). Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 19 below:

	TABLE 19
		Training Data	Testing Data
	Clinical + M markers	0.908 (0.032)	0.851 (0.047)

I.2. Analysis Using Clinical Covariates, E and M Markers.

To evaluate the value of added E and M markers for prediction of C/D outcomes, we randomly split the data into half training and half testing datasets, and did this 50 times. The results are summarized in Table 20 below:

	TABLE 20
		Training Data	Testing Data
	Clinical + E + M markers	0.969 (0.022)	0.893 (0.038)

II. Analysis for N/P OutcomesY=1 if P; Y=0 if N;II.0. Selection Based on M Markers Only.

There are totally 83 observations with no missing in Y (62 with Y=1, 74.7%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 10 M markers: PLF4 (SEQ ID NO: 99), CO8B (SEQ ID NO: 51), CERU (SEQ ID NO: 29), CFAH (SEQ ID NO: 35), FINC (SEQ ID NO: 70), HRG (SEQ ID NO: 81), PRG4 (SEQ ID NO: 101), SAMP (SEQ ID NO: 104), TSP4 (SEQ ID NO: 113) from nondepleted samples.

II.1. Analysis Using Clinical Covariates and M Markers.

To evaluate the value of added M markers for prediction of P/N outcomes, we randomly split the data into half training and half testing datasets, and did this 50 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 10 M markers (here we use logistic regression with ridge penalty since some M markers may be highly correlated). Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 21 below:

	TABLE 21
		Training Data	Testing Data
	Clinical + M markers	0.902 (0.078)	0.726 (0.080)

II.2. Analysis Using Clinical Covariates, E and M Markers.

To evaluate the value of added E and M markers for prediction of P/N outcomes, we randomly split the data into half training and half testing datasets, and did this 50 times. The results are summarized in Table 22 below:

	TABLE 22
		Training Data	Testing Data
	Clinical + E + M markers	0.937 (0.100)	0.739 (0.100)

III. Analysis for P/O OutcomesY=1 if PO2; Y=0 if PO1;III.0. Selection Based on M Markers Only.

There are totally 76 complete observations with PO1/PO2 (50 with Y=1, 65.8%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. It selects 7 M markers: PLF4 (SEQ ID NO: 99), KLKB1 (SEQ ID NO: 87), CO8B (SEQ ID NO: 51), ANT3 (SEQ ID NO: 14), ACTG (SEQ ID NO: 11), CD44 (SEQ ID NO: 28), CFAI (SEQ ID NO: 37) from nondepleted samples.

III.1. Analysis Using Clinical Covariates and M Markers.

To evaluate the value of added M markers for prediction of PO1/PO2 outcomes, we randomly split the data into half training and half testing datasets, and did this 50 times. Based on the training data, we fit two logistic regression: (i) using the clinical covariates only; (ii) using the clinical covariates plus the selected 7 M markers (here we use logistic regression with ridge penalty since some M markers may be highly correlated). Then, we applied the fitted scores to both training and testing data to compute AUC. The results are summarized in Table 23 below:

	TABLE 23
		Training Data	Testing Data
	Clinical + M markers	0.900 (0.052)	0.776 (0.059)

III.2. Analysis Using Clinical Covariates, E and M Markers.

No E markers were selected.

IV. Analysis for P/J Outcomes

Y=1 if PJ2; Y=0 if PJ1;

IV.0. Selection Based on M Markers Only.

There are totally 70 complete observations with PJ1/PJ2 (50 with Y=1, 54.3%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. No M markers were selected.

V. Analysis for P/K Outcomes

Y=1 if PK2; Y=0 if PK1;

IV.0. Selection Based on M Markers Only.

There are totally 83 complete observations with PK1/PK2 (30 with Y=1, 36.1%). We conducted lasso selection using the R package “glmnet”, and selected the tuning parameter using the built-in 10-fold cross-validation. No M markers were selected.

Multimarker Analysis III Using Serum Biomarkers

Methods

We generated further multi-marker models based on the markers we identified. Inferential methods and predictive methods were used depending on the structure of the data. For a person-level analysis, logistic regression was used to compute p-values and confidence intervals. Covariates included age, sex, BMI, and cohort. The effect of a biomarker was added to a model containing these covariates and a likelihood-ratio test was used to assess the significance of the biomarker after accounting for the covariates. A biomarker was considered significant if it surpassed a Benjamini-Hochberg FDR threshold of 10%. For a knee-level analysis the dependence arising from paired observations must be considered. We used the generalized estimating equation (GEE) method to account for the correlation structure. A biomarker was added to a model containing base covariates and its significance was assessed by a Wald statistic.

Predictive models were used to assess discrimination through the AUC. We used feature selection coupled with ridge regression, a form of penalized regression, for all models implemented in the glmnet R package. Penalized regression is often used for predictive models to constrain the size of coefficients to lessen the effects of overfitting the data. Feature selection consisted of selecting the top 8 markers with the lowest p-values, which is a simple but effective method for the numbers of peptides in the current data set. Leave-one-out cross-validation was used in which selecting tuning parameters and carrying out feature selection was repeated at each iteration of cross-validation to mimic the process of fitting a model to new data. In sum, all multimarker AUCs have been properly cross-validated. For knee-level (paired) analysis, the leave-one-out cross-validation scheme was modified to a leave-sample-out scheme so that the test set was independent of the training set.

Results

A summary of the results is shown in Tables 24-29.

Depleted Serum—Multimarker Analyses; Dx=Diagnosis (Predict Knee OA Diagnosis); Px=Prognosis (Predict Knee OA Progression).

TABLE 24
		p value adjusted
	p value peptide	for age, gender,
	alone	BMI and cohort
JSN (SEQ ID NO: X)	(pval_pep_only)	(pval_base_full)
knee JSN Dx depleted	AUC = 0.77
A2AP (4)	9.26E−05	0.046613974
A2AP (5)	2.42E−05	0.142711257
CO2 (41)	9.76E−05	0.078527954
COMP (55)	0.001556863	0.124035832
FA5 (61)	0.002710746	0.209278011
CO5(44)	0.009185471	0.865395357
CRAC1 (1)	0.018818294	0.011011104
SHBG (105)	0.014916307	0.803569832
knee JSN person Dx depleted	AUC = 0.80
A2AP (4)	3.31E−07	0.008051178
A2AP (5)	2.81E−05	0.292406073
CO2(41)	8.72E−05	0.162905105
COMP (55)	0.001303611	0.180539523
CRAC1 (1)	0.003128843	0.006690407
FA5 (61)	0.005274008	0.367965374
CO5(44)	0.022968897	0.863307127
CRAC1 (2)	0.024934734	0.051344408
knee JSN Px depleted	AUC = 0.41
CO4B (43)	0.033584489	0.054265435
FCN3 (67)	0.076630893	0.020166429
CO8B (51)	0.017661819	0.011132744
FINC (70)	0.040090816	0.045404495
PGCA (97)	0.065006579	0.105883213
TSP4 (113)	0.076843539	0.02082132
FINC (71)	0.051256407	0.091823154
TETN (107)	0.05530156	0.041645638
knee JSN person Px depleted	AUC = 0.39
PGCA (97)	0.026293441	0.046827678
CO4B (43)	0.023475468	0.015862788
TENX (106)	0.039475824	0.033732372
FCN3 (67)	0.18678107	0.107948938
C4BPA (25)	0.064638445	0.030266734
TSP1 (112)	0.13133037	0.18210438
CO8B (51)	0.113256944	0.07456324
HRG (79)	0.1438498	0.062444406

	TABLE 25
			p value
			adjusted for
			age, gender,
		p value peptide	BMI and
		alone	cohort
	KL (SEQ ID NO: X)	(pval_pep_only)	(pval_base_full)
	knee KL Dx depleted	AUC 0.82
	A2AP (4)	7.71E−05	0.00276514
	CO2 (41)	5.90E−05	0.002320056
	A2AP (5)	8.93E−06	0.062148737
	FAS (61)	0.000470793	0.058126283
	COMP (55)	0.001082738	0.12106214
	CO5 (44)	0.000562536	0.229466647
	CRAC1 (1)	0.004460634	0.001449587
	CRAC1 (2)	0.033778439	0.009978404
	knee KL person Dx depleted	AUC = 0.89
	A2AP (4)	1.50E−07	0.000135815
	A2AP (5)	9.64E−05	0.140576823
	CRAC1 (1)	0.000106179	6.44E−05
	CO2 (41)	0.00023168	0.024410373
	CRAC1 (2)	0.001782088	0.00062291
	COMP (55)	0.006752536	0.186567751
	FA5 (61)	0.00570514	0.108817433
	CO5 (44)	0.019947104	0.406134156
	knee KL Px depleted	AUC = 0.43
	A2AP (4)	0.038108974	0.38563753
	KNG1 (88)	0.020411121	0.391748644
	HGFA (78)	0.037729146	0.03308354
	PRG4 (102)	0.128488782	0.333151487
	AFAM (12)	0.120643826	0.844521484
	DOPO (58)	0.13172217	0.264987362
	HNC (71)	0.068344045	0.133741621
	CO8B (51)	0.062727522	0.03532693
	knee KL person Px depleted	AUC = 0.39
	KNG1 (88)	0.052000668	0.315327177
	HGFA (78)	0.054701983	0.056927485
	A2AP (4)	0.10641075	0.563510596
	FAS (61)	0.140793944	0.673707755
	TSP1 (112)	0.148443778	0.390538592
	PGCA (97)	0.135584259	0.183730223
	TENX (106)	0.149867446	0.055623532
	FINC (71)	0.164143337	0.229058614

TABLE 26
		p value adjusted
	p value peptide	for age, gender,
	alone	BMI and cohort
OST (SEQ ID NO: X)	(pval_pep_only)	(pval_base_full)
knee OST Dx depleted	AUC = 0.66
A2AP (5)	3.35E−05	0.022657938
A2AP (4)	0.001538871	0.021017782
FA5 (61)	0.000257061	0.006428595
CO2(41)	0.002198765	0.036511365
CO5(44)	0.001100287	0.024126793
COMP (55)	0.004918696	0.096093869
SHBG (105)	0.005156474	0.10532715
CO6A3 (49)	0.008366857	0.086848008
knee OST Dx person depleted	AUC = 0.82
A2AP (5)	2.09E−05	0.017678836
A2AP (4)	6.34E−05	0.01810586
CO2 (41)	0.000414071	0.024973599
FAS (61)	0.001001698	0.014610139
CRAC1 (1)	0.002263888	0.007041871
CO5 (44)	0.006050591	0.091904821
COMP (55)	0.009743019	0.152299519
CRAC1 (2)	0.02066252	0.028329279
knee OST Px depleted	AUC = 0.49
CXCL7 (57)	0.003370215	0.003050257
C4BPA (25)	0.003682397	0.005664358
C4BPB (27)	0.003909035	0.00385861
C4BPA (26)	0.006334513	0.007556145
A2AP (4)	0.038641621	0.029667601
ITIH4 (86)	0.041117212	0.033451744
PLMN (100)	0.042208729	0.03218861
HRG (83)	0.033182707	0.01429709
knee OST person Px depleted	AUC = 0.54
C4BPA (25)	0.011047566	0.011004835
C4BPB (27)	0.014086851	0.010277587
C4BPA (26)	0.011958263	0.012290566
CXCL7 (57)	0.013978919	0.015727039
LYAM1 (93)	0.071115015	0.024706069
A2AP (4)	0.126628387	0.024897652
TSP1 (112)	0.101650546	0.03394991
FINC (71)	0.099794158	0.146642438
Non-Depleted Serum - multimarker analyses;
Dx = diagnosis (predict knee OA diagnosis);
Px = prognosis (predict knee OA progression)

TABLE 27
		p value adjusted
	p value peptide	for age, gender,
	alone	BMI and cohort
JSN (SEQ ID NO: X)	(pval_pep_only)	(pval_base_full)
knee JSN Dx nondepleted	AUC = 0.71
CRAC1 (1)	1.45E−05	0.001334661
CXCL7 (57)	0.000220209	0.004033964
C4BPA (25)	0.001092182	0.029609534
COMP (54)	0.001921796	0.019441399
LUM (90)	0.003127373	0.053918866
CO5 (44)	0.007557254	0.051506718
TIMP1 (111)	0.003709593	0.017775522
C4BPA (26)	0.005168882	0.046877637
knee JSN person Dx nondepleted	AUC = 0.66
CRAC1 (1)	0.00026204	0.014921219
C4BPA (25)	0.002059188	0.013830002
LUM (90)	0.003509795	0.038433173
CO5 (44)	0.007188587	0.047295814
PCOC1 (95)	0.006323501	0.10111207
CXCL7 (57)	0.001410986	0.011528214
COMP (54)	0.005892446	0.051513052
CO5 (45)	0.010430159	0.054225697
knee JSN Px nondepleted	AUC = 0.55
PGCA (96)	0.018643447	0.076340059
APOH (18)	0.029178909	0.015538341
SAMP (104)	0.011280848	0.026954743
AACT (10)	0.028545301	0.044498503
AACT (8)	0.020730449	0.02340926
AACT (9)	0.015384508	0.02138342
CFAH (35)	0.030877701	0.008199734
PHLD (98)	0.065419311	0.075331479
TSP1 (112)	0.040324672	0.139428301
THRB (110)	0.153796326	0.031114401
knee JSN person Px nondepleted	AUC = 0.45
TSP1 (112)	0.020309869	0.038161024
CFAH (35)	0.025234989	0.015467182
THRB (110)	0.025437875	0.026717885
HRG (81)	0.05431721	0.01105114
APOH (18)	0.048284307	0.028696354
AACT (8)	0.064355392	0.042526395
PHLD (98)	0.054354097	0.079726494
AACT (9)	0.079176188	0.058988993

TABLE 28
		p value adjusted
	p value peptide	for age, gender,
	alone	BMI and cohort
KL (SEQ ID NO: X)	(pval_pep_only)	(pval_base_full)
knee KL Dx nondepleted	AUC = 0.77
CRAC1 (1)	7.91E−07	6.77E−05
CRAC1 (2)	6.61E−05	0.000611469
SHBG (105)	0.002780496	0.049018578
COMP (54)	0.003119022	0.016939944
CO8G (52)	0.004439784	0.01618559
PCOC1 (95)	0.005867991	0.118372629
CO6A3 (49)	0.014832263	0.120147034
LUM (90)	0.006899035	0.152788954
knee KL person Dx nondepleted	AUC = 0.74
CRAC1 (1)	8.05E−07	1.42E−05
CRAC1 (2)	8.37E−05	7.31E−05
PCOC1 (95)	0.001985962	0.044267004
CO8G (52)	0.003214569	0.010196682
LUM (90)	0.007570064	0.138934545
COMP (54)	0.002099819	0.015488264
CO6A3 (49)	0.0110763	0.097029625
CO5 (44)	0.03014129	0.449289979
knee KL Px nondepleted	AUC = 0.43
CFAH (35)	0.003413853	0.003720699
SAMP (104)	0.008874549	0.01757112
TSP1 (112)	0.016396723	0.029173098
HEP2 (76)	0.021317466	0.138072515
C1R (22)	0.022324877	0.035342549
APOB (15)	0.019423193	0.073310656
FINC (71)	0.038839588	0.741216237
HEP2 (77)	0.061398201	0.173258732
PGCA (96)	0.071967152	0.438930963
FINC (70)	0.044013606	0.329880745
knee KL person Px nondepleted	AUC = 0.57
CFAH (35)	0.004647535	0.004744558
TSP1 (112)	0.014072412	0.044301897
SAMP (104)	0.007169693	0.038267836
APOB (15)	0.026721333	0.127645336
AACT (9)	0.014781295	0.049356758
AACT (8)	0.017046792	0.061084614
C1R (22)	0.031947667	0.082794302
C1R (23)	0.030621922	0.095021007

TABLE 29
		p value adjusted
	p value peptide	for age, gender,
	alone	BMI and cohort
OST (SEQ ID NO: X)	(pval_pep_only)	(pval_base_full)
knee OST Dx nondepleted	AUC = 0.70
CRAC1 (1)	1.60E−06	0.00013426
CRAC1 (2)	5.77E−05	0.0007071
COMP (54)	0.004838815	0.036180336
CO6A3 (49)	0.015033517	0.149650103
SHBG (105)	0.009303617	0.099766058
PCOC1 (95)	0.016164645	0.244959617
CO8G (52)	0.052018531	0.202674481
LUM (90)	0.058692947	0.601716961
knee OST Dx person nondepleted	AUC = 0.70
CRAC1 (1)	1.86E−06	5.43E−05
CRAC1 (2)	1.86E−06	5.43E−05
COMP (54)	0.002304261	0.025690528
CO6A3 (49)	0.009634809	0.098819568
PCOC1 (95)	0.016841954	0.207971822
CO8G (52)	0.031580042	0.128396316
ACTG (11)	0.030554232	0.180275486
CO5 (44)	0.095059051	0.79862693
knee OST Px nondepleted	AUC = 0.61
PLF4 (99)	0.001094118	0.001590437
CXCL7 (57)	0.009976718	0.012612223
ANT3 (14)	0.011884039	0.006241664
AACT (8)	0.00821959	0.002747701
AACT (10)	0.010576976	0.002088226
THRB (110)	0.030760839	0.053273043
AACT (9)	0.016720482	0.0084448
ITIH4 (86)	0.039793857	0.04003038
CO8B (51)	0.061264758	0.062748503
PLMN (100)	0.046320809	0.044378072
knee OST person Px nondepleted	AUC = 0.67
CO8B (51)	0.00274445	0.007382756
PLF4 (99)	0.003710526	0.005376922
PRG4 (101)	0.016806589	0.059964262
PRG4 (102)	0.024390935	0.054916368
ANT3 (14)	0.017572544	0.019097753
C4BPA (26)	0.029402337	0.122776322
CXCL7 (57)	0.005698	0.006885551
C4BPA (25)	0.051311885	0.186676691

Claims

1. A method of treating a subject with progression of osteoarthritis, comprising: measuring the expression levels of a group of biomarkers in a sample from the subject, wherein the biomarkers measured comprise the complement C1r subcomponent (C1R) peptide of SEQ ID NO: 23, the cartilage acidic protein 1 (CRAC1 or CRTAC1) peptide of SEQ ID NO: 1 or SEQ ID NO: 2, the kininogen-1 (KNG1) peptide of SEQ ID NO: 88, the phosphatidylinositol-glycan-specific phospholipase D (PHLD) peptide of SEQ ID NO: 98, and the vitamin D-binding protein (VTDB) peptide of SEQ ID NO: 116, SEQ ID NO: 118, or SEQ ID NO: 119;predicting progression of osteoarthritis in the subject, wherein altered levels of any of the biomarkers measured as compared to a reference level is indicative of progression of osteoarthritis; andselecting and administering a therapeutic for treatment of osteoarthritis to the subject identified as having progression of osteoarthritis based on the prediction.

2. The method of claim 1, wherein the biomarkers measured further comprise at least one biomarker selected from the group consisting of antithrombin-III (ANT3), complement factor H (CFAH), and platelet factor 4 (PLF4).

3. The method of claim 2, wherein the biomarkers measured comprise at least two biomarkers selected from the group consisting of ANT3, CFAH, and PLF4.

4. The method of claim 3, wherein the biomarkers measured comprise ANT3, CFAH, and PLF4.

5. The method of claim 3, wherein the at least two biomarkers are selected from the group consisting of the ANT3 peptide of SEQ ID NO: 14, the CFAH peptide of SEQ ID NO: 35, and the PLF4 peptide of SEQ ID NO: 99.

6. The method of claim 3, wherein the levels of the at least two biomarkers are altered in the subject as follows: ANT3 level is decreased, CFAH level is increased, or PLF4 level is increased, as compared to the reference level.

7. The method of claim 1, wherein the biomarkers measured further comprise at least one biomarker selected from the group consisting of alpha-1B-glycoprotein (A1 BG), alpha-2-antiplasmin (A2AP), leucine-rich alpha-2-glycoprotein (A2GL), alpha-1-antichymotrypsin (AACT), actin, cytoplasmic 2 (ACTG), afamin (AFAM), apolipoprotein B-100 (APOB), beta-2-glycoprotein 1 (APOH), beta-2-microglobulin (B2MG), complement C1q subcomponent subunit C (C1QC), complement C1r subcomponent-like protein (C1RL), c4b-binding protein alpha chain (C4BPA), c4b-binding protein beta chain (C4BPB), cluster of differentiation 14 (CD14), cluster of differentiation 163 (CD163), cluster of differentiation 44 (CD44), ceruloplasmin (CERU), complement factor B (CFAB), complement factor I (CFAI), complement C1s subcomponent (C1S), complement C2 (CO2), complement C4-B (CO4B), complement C5 (CO5), collagen alpha-3(VI) chain (CO6A3), complement component C8 beta chain (CO8B), complement component C8 gamma chain (CO8G), collagen alpha-1(III) chain (COL3A1), collagen type II C-telopeptide fragments (CTX2), platelet basic protein (CXCL7), dopamine beta-hydroxylase (DOPO), extracellular matrix protein 1 (ECM1), coagulation factor V (FA5), coagulation factor XII (FA12), fibulin-1 (FBLN1), IgG Fc-binding protein (FCGBP), ficolin-3 (FCN3), alpha-2-HS-glycoprotein (FETUA), fibronectin (FINC), gelsolin (GELS), hyaluronan-binding protein 2 (HABP2), haptoglobin, hemopexin (HEMO), heparin cofactor 2 (HEP2), hepatocyte growth factor activator (HGFA), histidine-rich glycoprotein (HRG), hyaluronan, inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), plasma kallikrein (KLKB1), lumican (LUM), L-selectin (LYAM1), aggrecan core protein (PGCA), plasminogen (PLMN), proteoglycan 4 (PRG4), retinol-binding protein 4 (RET4), serum amyloid P-component (SAMP), tenascin X (TENX), tetranectin (TETN), thyroxine-binding globulin (THBG), prothrombin (THRB), tissue inhibitor metalloproteinase 1 (TIMP1), thrombospondin-1 (TSP1), thrombospondin-4(TSP4), and vitronectin (VTNC).

8. The method of claim 7, wherein the biomarkers measured comprise FA5.

9. The method of claim 1, further comprising: measuring the level of at least one normalization peptide from a protein selected from beta-Ala-His dipeptidase (CNDP1), sex hormone-binding globulin (SHBG), adiponectin (ADIPO), apolipoprotein A4 (APOA4), keratin, type II cytoskeletal 1 (K2C1), corticosteroid-binding globulin (CBG), carboxylic ester hydrolase (CHLE), coagulation factor XI (FA11), and carbonic anhydrase 1 (CAH1) in the sample from the subject; andnormalizing the levels of the biomarkers measured in the sample from the subject prior to comparing the levels of the biomarkers to the reference level.

10. The method of claim 9, wherein the normalization peptide is selected from the group consisting of the CNDP1 peptide of SEQ ID NO: 40, and the SHBG peptide of SEQ ID NO: 105.

11. The method of claim 1, wherein the sample is serum, plasma, urine, or synovial fluid.

12. The method of claim 11, wherein the sample is serum depleted of at least 7 major serum proteins.

13. The method of claim 12, wherein the serum proteins depleted are selected from the group consisting of albumin, IgG, IgA, transferrin, haptoglobin, anti-trypsin, fibrinogen, alpha 2-macroglobulin, IgM, apolipoprotein AI, apolipoprotein AII, complement C3, and transthyretin.

14. The method of claim 1, wherein the biomarkers are measured using an antibody-based capture method or mass spectrometry.

15. The method of claim 1, wherein the therapeutic for treatment of osteoarthritis is an anti-inflammatory or anti-pain therapeutic comprising a nonsteroidal anti-inflammatory drug (NSAID).