INTRODUCTION

Cytokine targeting drugs such as TNFoblocking agents (herein after referred to as "TBA") are increasingly used in the treatment of various inflammatory diseases. The first indications in which such TBA were approved are rheumatoid arthritis and Crohn's disease. Rheumatoid arthritis (RA) is a chronic, progressive, debilitating auto-immune disease of largely unknown etiology that affects approximately 1 % of the population (1 ). RA is characterized by chronic inflammation of the synovium, which ultimately leads to joint damage, pain and disability (2). The clinical spectrum of RA is heterogeneous, ranging from mild to severe, with variability in secondary organ system involvement. Disease heterogeneity is further illustrated by the current variation in treatment response rates. First line treatment is usually initiated with so called disease-modifying anti-rheumatic drugs (DMARDs), such as methotrexate (MTX). Approximately 30% of patients display a suboptimal response or intolerance to traditional DMARDs (3). In these patients, second line treatment is initiated with "biologies", agents that block molecules or cells thought to be instrumental to disease progression, such as tumor necrosis factor-a (TNFa) and interleukin-1 (IL-1 ) or B and T-cells. There are indeed nine biologic agents currently available, each with overlapping or unique mechanisms of action (4). The response rates to such treatments vary widely, with a great number of patients remaining refractory to treatment or demonstrating only partial improvement (5). The incomplete understanding of drug mechanisms of action together with disease heterogeneity means that there are no methods of identifying patient suitability for the various biologies prior to the initiation of the treatment. Establishing a rational basis on which to select patients for specific biologies would help patients to be treated more efficiently; those that would be likely to respond would initiate the biologic in question whereas those unlikely to respond could be provided with another treatment.

In absence of reliable literature on efficacy and safety of biologies and given the percentage of patients that do not respond or experience severe adverse effects, the destructive nature of RA, and the societal costs of inefficacious biological treatments, there is a strong need to make predictions on success before starting the therapy. A clinically or radiographic-based test will most probably assess conditions too late for protecting joints from irreversible destruction. Ideally, a molecular biomarker signature as a predictor for therapy responsiveness should be obtained prior to the start of therapy in a readily available bio-sample, such as peripheral blood. Given the systemic nature of RA and the communication between the systemic and organ-specific compartments, the peripheral blood may not directly have implications for the understanding of disease pathogenesis, but it is especially suitable to analyze gene expression profiles that provide a framework to select clinically relevant biomarkers. Furthermore, blood-based tests remain less invasive for the patients than synovial tissue-based tests. Ultimately, this may lead to a personalized form of medicine, whereby the best suited therapy will be applied to an individual patient.

The same is true for other inflammatory diseases for which TBA have been approved or for which preliminary results indicate that TBA might be a useful treatment. For the moment, TBA have been approved, in addition to the treatment of RA or Crohn's disease, for the treatment of ankylosing spondylitis, psoriatic arthritis, plaque psoriasis, and ulcerative colitis (see notably FDA labels of infliximab and etanercept). In addition, preliminary results suggest that TBA may be useful in the treatment of several other inflammatory diseases, such as vasculitis (notably Behcet's disease, Churg-Strauss vasculitis, polyarteritis nodosa, and giant cell arthritis); Wegener's granulomatosis; sarcoidosis; adult-onset Still's disease, polymyositis/dermatomyositis, and systemic lupus erythematosus (SLE) (31 and 32).

In all cases, only a proportion (although sometimes a high proportion) of patients treated with TBA display a clinical response to the treatment (see notably FDA labels of infliximab (Remicade®) and etanercept (Enbrel®)). For all diseases in which TBA may be useful, it would thus be very helpful to be able to predict the capacity of a subject to respond or not to TBA treatment.

A very powerful way to gain insight into the molecular signatures underlying pathophysiological processes has arisen from DNA microarray technology, which allows the identification of the fraction of genes that are differentially expressed in blood and tissue samples among patients with clinically defined disease. These differentially expressed genes may provide insight into biological pathways contributing to disease and represent classifiers for early diagnosis, prognosis, and response prediction. Several pitfalls were experienced using this multistage and relatively expensive technology, which highly depends on perfectly standardized conditions. Factors that might influence the sensitivity and reproducibility range from sample differences, variation in amount and quality of starting RNA material, amplification and labeling strategies and dyes, to probe sequence and hybridization conditions. In addition, the lack of standardized approaches for normalization and usage of data analysis algorithms could influence the outcome. Furthermore, most microarray studies are not prospectively planned and often do not have detailed protocols, but rather tend to make use of existing samples. Therefore, verification of results is an essential step in microarrays studies and quality criteria have to be set.

Several groups have explored the possibility of identifying molecular traits (single nucleotide polymorphisms, gene expression etc.) capable of classifying patients according to their response to treatment based on retrospective analyses of biological samples (synovium or peripheral blood) collected at treatment baseline. In particular, much interest has been paid to the TNFa-blocking agent Infliximab, with the first report several years ago by T Lequerre and colleagues of gene expression-based prediction of response to therapy (6). Since then, several other groups have similarly reported on large-scale gene expression analyses of peripheral blood as a means to predict response to Infliximab (7) (8) (9). All of these studies reported on differentially expressed genes and combinations thereof for the prediction of response to therapy.

These studies provided important proof of concept for the prediction of response to Infliximab at baseline of therapy. Nevertheless, as with all studies of this kind, the use of microarray technology, measuring thousands of genes simultaneously in relatively small cohorts of patients, runs the risk of over-fitting data, leading to false positive results. Moreover, the mono-centric nature of these studies may limit the relevance of the genes identified to a wider and more demographically varied population. The present invention overcomes these drawbacks by combining information from multiple existing studies. This approach can increase the reliability and generalizability of results. Quantitative approaches in which individual studies addressing a set of related research hypothesis are statistically integrated and analyzed to determine the effectiveness of interventions (meta-analysis) showed the broad utility of applying meta-analytic approaches to genome-wide data for the purpose of biological discovery. Meta-analysis were already used to identify genes differentially expressed between two groups, to compare results obtained on different microarray platforms (cross-platform classification), to identify overlaps between samples from heterologous datasets, to identify co-expressed genes or to reconstruct gene networks.

Meta-analyses of multiple gene expression microarray datasets provide discriminative gene expression signatures that are identified and validated on a large number of microarray samples, generated by different laboratories and microarray technologies. Predictive models generated by this approach are better validated than those generated on a single data set, while showing high predictive power and improved generalization performance.

In the present invention, the meta-analysis was performed according to the stepwise approach in conducting meta-analysis on microarray datasets (1 -identify suitable microarray studies; 2-extract data from studies (this step also involved getting additional information from the authors of selected studies); 3-prepare the individual datasets; 4-annotate the individual datasets; 5-resolve the many-to-many relationship between probes and genes; 6-combine the study-specific estimates; 7-analyze, present, and interpret results) described in Ramasamy et al. (10).

Sixty one genes differentially expressed between future responders and non responders to Infliximab therapy have thus been identified. Furthermore, two individual genes have been found to be highly correlated to the Infliximab responsive or non- responsive phenotype of tested subjects, and several combinations of a minimum number of genes are proposed as being predictive of the primary (week 14 and week 22) response to anti-TNF treatment in RA patients. These combinations comprise genes that are known to be involved in inflammatory or immune processes rather than in the metabolism pathway of Infliximab, which clearly gives a rational for their general usefulness for predicting TNFoc-blocking agents (TBA) responsive or non-responsive phenotype of subjects suffering from other inflammatory diseases, notably those for which TBA have been approved or have been shown to be useful in preliminary studies.

DETAILED DESCRIPTION OF THE INVENTION

The invention thus relates to a method for the in vitro diagnosis or prognosis of a cytokine targeting drug (hereafter referred to as CyTD, such as aTNFoblocking agent, hereafter referred to as TBA) responding or non-responding phenotype, comprising: (a) determining from a biological sample of a subject suffering from an inflammatory disease an expression profile comprising or consisting of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or (iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s),

(b) comparing the obtained expression profile with at least one reference expression profile, and

(c) determining the CyTD responding or non-responding phenotype from said comparison.

The invention also relates to a method for designing a CyTD treatment for a subject suffering from an inflammatory disease, said method comprising:

(a) determining from a biological sample of said subject an expression profile comprising or consisting of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes

MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or (iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s),

(b) comparing the obtained expression profile with at least one reference expression profile,

(c) determining the CyTD responding or non-responding phenotype from said comparison, and

(d) designing the dose of CyTD treatment according to said identified CyTD responding or non-responding phenotype. The invention is also drawn to a method of treatment of a subject suffering from an inflammatory disease with a CyTD, comprising: (a) determining from a biological sample of the said subject the presence of a CyTD responding or non-responding phenotype using a method according to the invention, and

(b) adapting the CyTD treatment in function of the result of step (a).

Said adaptation of the CyTD treatment may consist in:

a reduction or suppression of the said CyTD treatment if the subject has been diagnosed as CyTD non-responding, or

the continuation of the said CyTD treatment if the subject has been diagnosed as CyTD responding.

The invention also refers to a new use of a CyTD in the treatment of an inflammatory disease, comprising the steps of:

(a) determining from a biological sample of a subject suffering from an inflammatory disease the presence of a CyTD responding or non-responding phenotype using a method according to the invention, and

(b) determining the dose of CyTD to administer with respect to the result of step (a). Optionally; the dose of CyTD determined in step (b) is administered to the subject.

The invention thus relates to a CyTD, for use in treating an inflammatory disease, wherein the CyTD is administered to a subject suffering from said inflammatory disease who has been diagnosed and/or prognosed as responsive using a method according to the invention.

The invention also relates to the use of a CyTD for preparing a drug for the treatment of an inflammatory disease in subjects suffering from said inflammatory disease who have been diagnosed and/or prognosed as responsive using a method according to the invention.

In all the present description, an "inflammatory disease" refers to a disease involving uncontrolled inflammation processes leading to body damages, and includes any disease generally considered as inflammatory diseases by those skilled in the art. Advantageously, said inflammatory disease is known to involve, at least in some cases, a pathogenic inflammatory cytokine (I L-1 , I L-6, IL15, I L-17, I L-18, I L-23 or TNF-oc) secretion. The methods according to the invention then permit to diagnose the presence of such a pathogenic inflammatory cytokine secretion in a tested subject, to predict his/her capacity to respond to a CyTD treatment, and thus to adapt his/her treatment in view of his CyTD responding/non-responding phenotype. Even more advantageously, said inflammatory disease is known to involve, at least in some cases, a pathogenic TNF-oc secretion. The methods according to the invention then permit to diagnose the presence of such a pathogenic TNF-oc secretion in a tested subject, to predict his/her capacity to respond to a TBA treatment, and thus to adapt his/her treatment in view of his TBA responding/non-responding phenotype. Such inflammatory diseases may be of autoimmune or non-autoimmune origin. Non limiting examples of inflammatory diseases for which the methods and kits according to the invention are useful, in particular for determining a TBA responding or non-responding phenotype, include rheumatoid arthritis (RA), Crohn's disease, ankylosing spondylitis, psoriatic arthritis, plaque psoriasis, ulcerative colitis, vasculitis (notably Behcet's disease, Churg-Strauss vasculitis, polyarteritis nodosa, and giant cell arthritis); Wegener's granulomatosis; sarcoidosis; adult-onset Still's disease, polymyositis/dermatomyositis, and systemic lupus erythematosus (SLE). An advantageous group of diseases for which the methods of the invention are useful are those in the treatment of which TBA have been approved: rheumatoid arthritis (RA), Crohn's disease, ankylosing spondylitis, psoriatic arthritis, plaque psoriasis, ulcerative colitis. The methods according to the invention are particularly useful for RA-suffering patients for determining a TBA responding or non-responding phenotype.

For rheumatoid arthritis (RA), since first line treatment is usually initiated with so called disease-modifying anti-rheumatic drugs (DMARDs), the invention also refers to a method of treatment of an RA-suffering subject, comprising the steps of:

(a) administering a therapeutic dose of a DMARD to the said subject suffering from RA,

(b) determining from a biological sample of the said RA-suffering subject the presence of a CyTD responding or non-responding phenotype using a method according to the invention, and

(c) determining the dose of CyTD to administer with respect to the result of step (b).

Thus the invention also refers to a combination of a DMARD and a CyTD for the treatment of RA, comprising the steps of:

(a) administering a therapeutic dose of a DMARD to a subject suffering from RA,

(b) determining from a biological sample of the said RA-suffering subject the presence of a CyTD responding or non-responding phenotype using a method according to the invention, and

(c) determining the dose of CyTD to administer with respect to the result of step (b). Optionnally; the dose of CyTD determined in step (c) is administered to the subject. In a preferred embodiment, the DMARD is methotrexate (MTX).

By "cytokine targeting drug" or "CyTD", it is meant any molecule neutralizing a cytokine signalling, notably by binding to and neutralizing the cytokine or its receptor. Such a binding and neutralizing molecule may notably be an antibody or a fragment thereof specific for said cytokine or cytokine receptor, cytokine receptor antagonists, or any other molecule, such as a recombinant protein, binding to and neutralizing said cytokine or cytokine receptor. Said CyTD preferably targets an inflammatory cytokine such as IL-1 , IL-6, IL-15, IL-17, IL-18, IL-23 or TNF-oc or a receptor of such inflammatory cytokines. Molecules targeting IL-1 signalling include monoclonal antibodies to IL-1 , such as Canakinumab (commercial name Maris®), a human anti-IL- 1 β monoclonal antibody; antagonists of IL-1 receptor such as anakinra (commercial name Kineret®), and a fusion protein between lgG1 Fc portion and ligand-binding domains of human IL-1 RI and IL-1AcP such as Rilonacept (nom commercial Arcalyst™) Molecules targeting II-6 signalling notably include Tocilizumab, an anti-IL- 6R monoclonal antibody. Molecules targeting 11-15 signalling notably include HuMax-IL- 15 (AMG 714), an anti-IL-15 monoclonal antibody. Molecules targeting 11-17 signalling notably include AIN457, an anti-IL-17A monoclonal antibody. In all the present description, a preferred embodiment of a CyTD is a "TNFoblocking agent" or "TBA".

By "TNFoblocking agent" or "TBA", it is herein meant a biological agent which is capable of neutralizing the effects of TNFa. Said agent is a preferentially a protein such as a soluble TNFa receptor, e.g. Pegsunercept, or an antibody. In a further preferred embodiment, the said agent is a monoclonal antibody. In an even further preferred embodiment, the said agent is selected in the group consisting of Etanercept (Enbrel®), Infliximab (Remicade®), Adalimumab (Humira®), and Certolizumab pegol (Cimzia®). In an even more preferred embodiment, the said agent is Infliximab.

In a particularly preferred embodiment of any method according to the present invention, the inflammatory disease is rheumatoid arthritis and the CyTD is Infliximab, a particular TBA.

According to the present invention, a "CyTD responding phenotype" is defined as a response state of a subject to the administration of a CyTD. A "response state" means that the said subject (referred to as a CyTD responding subject or a responding subject or a responsive subject: for the purpose of this application, these terms are similar) responds to the treatment, i.e. that the treatment is efficacious in the said subject. The definition of response is an improvement in clinical symptoms. The quantification of such response is made according to ACR20, ACR50, ACR70 criteria (1 1 ) and/or EULAR criteria at weeks 14 or 22 or change in DAS28 >1 .2. Even more preferred is EULAR response criteria at 14 weeks. These criteria (31 ) have been established by organizations regrouping the professionals in the field (ACR: American College of Rheumatology; EULAR: European League Against Rheumatism). These criteria are thus well known to the skilled person in the art and need not be detailed here.

In contrast, a "CyTD non-responding phenotype" refers to the absence in said subject (referred to herein as a CyTD non-responding subject or a non responding subject or a non-responsive subject: these terms should be construed in the context of this application as having the same meaning) of a state of response, meaning that said subject remains refractory to the treatment.

In a preferred embodiment of any of the above-described in vitro methods of diagnosis/prognosis according to the invention, the said subject is an RA-suffering subject. An "RA-suffering subject" is a subject fulfilling the American College of Rheumatology (ACR) criteria for RA (1 1 ). In one further embodiment, the said subject is not treated with a CyTD; in another further embodiment, the said subject is treated with a CyTD.

It will easily be conceived that when the said subject is not treated with a CyTD, the methods of the invention permit a prognosis of the responsiveness/non responsiveness of the said subject. Thus, in this embodiment, the method of the invention allows the person skilled in the art to prognose (i.e. to identify) the subjects susceptible of responding to the CyTD treatment. This is important because of the destructive nature of RA and the societal costs of inefficacious biological treatments. Moreover, since this embodiment of the invention allows for identification of non responsive subjects before any treatment is initiated, the risks for one treated subject to encounter severe adverse effects are greatly diminished.

When the subject according to the invention is treated with a CyTD, the methods of the invention are useful for diagnosing if a subject responds to the said CyTD, and whether the said subject would thus benefit from a continuation of the said treatment. Moreover, they are useful for diagnosing subjects who are not responding to the treatment, i.e. who are refractory to the CyTD, and should thus swiftly shifted to another therapy. In regard of the debilitating nature of RA, this achievement is crucial. In particular, the methods of the invention allow for a diagnosis at week 14 or 22 after the beginning of the CyTD treatment.

In the present description, what is described for CyTD also particularly applies to TBA, which is a preferred embodiment of a CyTD in any method or kit according to the invention. A "biological sample" may be any sample that may be taken from a subject, such as a serum sample, a plasma sample, an urine sample, a blood sample, a lymph sample, or a biopsy. Such a sample must allow for the determination of an expression profile comprising or consisting of: (i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes

MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or (iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s).

Preferred biological samples for the determination of an expression profile include samples such as a blood sample, a plasma sample, a lymph sample, or a biopsy. Preferably, the biological sample is a blood sample. Indeed, such a blood sample may be obtained by a completely harmless blood collection from the patient and thus allows for a non-invasive diagnosis of a CyTD responding or non-responding phenotype.

By "expression profile" is meant the expression levels of a group of genes comprising or consisting of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes

MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or (iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s), since these expression profiles have been demonstrated to be particularly relevant for assessing the responding/non responding phenotype of a subject. In a most preferred embodiment, the expression profile for diagnosing if the subject is responding at week 14 or week 22 after the beginning of the CyTD treatment comprises or preferably consists of the genes IL2RB, S100A9, and CASP5, or of the gene S100A9 or Equivalent Expression Profile thereof, and optionally one or more housekeeping gene(s). In another most preferred embodiment, the expression profile for diagnosing responsiveness at week 14 or week 22 after the beginning of the CyTD treatment (in particular a TBA treatment) comprises or preferably consists of the genes MAPK14 and S100A9, or Equivalent Expression Profile thereof, and optionally one or more housekeeping gene(s). In yet another most preferred embodiment, the expression profile for diagnosing responsiveness at week 14 or week 22 after the beginning of the CyTD treatment (in particular a TBA treatment) comprises or preferably consists of the gene S100A9, or Equivalent Expression Profile thereof, and optionally one or more housekeeping gene(s). In still another most preferred embodiment, the expression profile for diagnosing responsiveness at week 14 or week 22 after the beginning of the CyTD treatment (in particular a TBA treatment) comprises or preferably consists of the genes MAPK14 and GNLY, or Equivalent Expression Profile thereof, and optionally one or more housekeeping gene(s).

The determination of the presence of a CyTD responding or non-responding phenotype is carried out thanks to the comparison of the obtained expression profile with at least one reference expression profile in step (b).

The term "Equivalent Expression Profile" herein refers to expression profiles of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, wherein the addition, deletion or substitution of some of the genes does not change significantly the reliability of the test and is considered as an "acceptable expression profile". As an example, the addition or substitution of some of the genes of the sets described in the present invention by other genes belonging to the same metabolic pathway should also be considered as an equivalent expression profile. For example S100A8 is equivalent to S100A9, and any above mentioned (i) or (ii) expression profile in which S100A9 is replaced by S100A8 (i.e. expression profiles comprising or consisting of the genes MAPK14 and S100A8; 8 or the 61 genes of Table 2 in which S100A9 is replaced by S100A8; or the gene S100A8; or the genes S100A8, IL2RB, and CASP5; or the genes S100A8, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A8, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C) should be considered as Equivalent Expression Profiles of above mentioned (i) and (ii) expression profiles and thus as included in the scope of the present invention. More generally are considered as Equivalent Expression Profiles, any expression profile wherein some of the genes are replaced by genes belonging to the same biological network, such as described in figures 2 and 3 below.

The term "Acceptable Expression Profile" herein refers to an expression profile which is capable of correctly classifying at least 60% of the analyzed samples, preferably 65%, and more preferably 70%, has a sensitivity and specificity of at least 60% preferably 65%, and more preferably 70. The sensitivity value is defined as the ratio of the number of patients actually clinically responding to the CyTD treatment and classified as responding using the test according to the invention amongst all patients treated with the CyTD. Specificity measures the proportion of patients actually clinically not responding to the CyTD treatment which are correctly identified using the test according to the invention amongst all patients treated with the CyTD.

By "Best Expression Profile" is meant an expression profile which is able to correctly classify at least 80% of the analyzed samples, has either a sensitivity or a sensitivity of at least 80%. Although the lists of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C,

have been determined as the Best Expression Profiles to assess responsiveness/non responsiveness, an Equivalent Expression Profile such as defined above, still permits to assess responsiveness, with an acceptable reliability. In particular embodiments, sublists of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes

S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C,

still permit to assess responsiveness with a good reliability and should be considered as Acceptable Expression Profiles.

While the expression profile used for determining the CyTD (notably TBA) responsive or non-responsive phenotype may comprise and not only consist of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s), it is preferred that the expression profiles consists or essentially consists of one of the above described (i), (ii) or (iii) expression profiles, optionally with one or more housekeeping gene(s), meaning that no more than 50, 40, 30, 25, 20, preferably no more than 15, preferably no more than 10, preferably no more than 9, 8, 7, 6, 5, 4, 3, 2, or 1 genes that are not a gene belonging one the above described (i), (ii) or (iii) expression profiles or a housekeeping gene are included in the expression profile.

By "housekeeping genes", it is meant genes that are constitutively expressed at a relatively constant level across many or all known conditions, because they code for proteins that are constantly required by the cell, hence, they are essential to a cell and always present under any conditions. It is assumed that their expression is unaffected by experimental conditions. The proteins they code are generally involved in the basic functions necessary for the sustenance or maintenance of the cell. Non-limitating examples of housekeeping genes that may be used in methods of the invention include: HPRT1 (hypoxanthine phosphoribosyltransferase 1 ),

- UBC (ubiquitin C),

YWHAZ (tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, zeta polypeptide),

- B2M (beta-2-microglobulin),

GAPDH (glyceraldehyde-3-phosphate dehydrogenase),

FPGS (folylpolyglutamate synthase),

DECR1 (2,4-dienoyl CoA reductase 1 , mitochondrial),

PPIB (peptidylprolyl isomerase B (cyclophilin B)),

- ACTB (actin 3),

PSMB2 (proteasome (prosome, macropain) subunit, beta type, 2), GPS1 (G protein pathway suppressor 1 ),

- CANX (calnexin),

NACA (nascent polypeptide-associated complex alpha subunit), - TAX1 BP1 (Taxi (human T-cell leukemia virus type I) binding protein 1 ), and

PSMD2 (proteasome (prosome, macropain) 26S subunit, non-ATPase, 2). Preferably, the number of housekeeping genes used for normalizaztion in methods according to the invention is comprised between one and five with a preference for three.

The determination of the presence of a responsive or non responsive phenotype is carried out thanks to the obtained expression profile with at least one reference profile in step (b). A "reference expression profile" is a predetermined expression profile, obtained from a biological sample from a subject with a known particular response state. In order to have this comparison meaningful and robust, the "reference expression profile" is determined by the training of an algorithm resulting from the following steps:

the collection of the expression profiles of biological samples obtained from responsive and non-responsive patients,

the training of an algorithmon the expression profiles of the aforementioned patients for classifying said expression profiles as responsive and non- responsive phenotypes.

The "comparison of an obtained expression profile with a reference expression profile" can be understood as the application of the adjusted algorithm on the obtained expression profile. In particular embodiments, the reference expression profile used for comparison with the test sample in step (b) may have been obtained from biological samples from CyTD responsive subjects ("CyTD responsive reference expression profile" or "responsive reference expression profile"; as used herein these expressions are synonymous), and/or from biological samples from CyTD non-responsive subjecst ("CyTD non- responsive reference expression profile" or "non-responsive reference expression profile" ; as used herein these expressions have the same meaning).

Preferably, at least one reference expression profile is a CyTD responsive reference expression profile. Alternatively, at least one reference expression profile may be a CyTD non-responsive reference expression profile. More preferably, the determination of the presence or absence of a CyTD responsive phenotype is carried out by comparison with at least one responder and at least one non-responder reference expression profiles. The diagnosis or prognostic may thus be performed using one responsive reference expression profile and one non-responsive reference expression profile. Advantageously, to get a stronger diagnosis or prognostic, said diagnosis or prognostic is carried out using several responsive reference expression profiles and several non-responsive reference expression profiles. The comparison of a tested subject expression profile with said reference expression profiles, which permits prediction of the tested subject's clinical response based on his/her expression profile, can be done by those skilled in the art using statistical models or machine learning technologies. The PLS (Partial Least Square) regression is particularly relevant to give prediction in the case of small reference samples. The comparison may also be performed using Support Vector Machines (SVM), logistic regression, Linear Discriminant Analysis, Random Forests, k-NN (Nearest Neighbour) or PAM (Predictive Analysis of Microarrays) statistical methods.

The expression profile may be determined by any technology known by a man skilled in the art. In particular, each gene expression level may be measured at the genomic and/or nucleic and/or proteic level. In a preferred embodiment, the expression profile is determined by measuring the amount of nucleic acid transcripts of each gene. In another embodiment, the expression profile is determined by measuring the amount of protein produced by each of the genes.

The amount of nucleic acid transcripts can be measured by any technology known by a man skilled in the art. In particular, the measure may be carried out directly on an extracted messenger RNA (mRNA) sample, or on retrotranscribed complementary DNA (cDNA) prepared from extracted mRNA by technologies well-know in the art. From the mRNA or cDNA sample, the amount of nucleic acid transcripts may be measured using any technology known by a man skilled in the art, including nucleic microarrays, quantitative PCR, and hybridization with a labelled probe.

In a preferred embodiment, the expression profile is determined using quantitative PCR. Quantitative, or real-time, PCR is a well known and easily available technology for those skilled in the art and does not need a precise description.

In a particular embodiment, which should not be considered as limiting the scope of the invention, the determination of the expression profile using quantitative PCR may be performed as follows. Briefly, the real-time PCR reactions are carried out using the TaqMan Universal PCR Master Mix (Applied Biosystems). 6 μΙ_ cDNA is added to a 9 μΙ_ PCR mixture containing 7.5 μΙ_ TaqMan Universal PCR Master Mix, 0.75 μΙ_ of a 20X mixture of probe and primers and 0.75μΙ water. The reaction consisted of one initiating step of 2 min at 50 deg. C, followed by 10 min at 95 deg. C, and 40 cycles of amplification including 15 sec at 95 deg. C and 1 min at 60 deg. C. The reaction and data acquisition can be performed using the ABI 7900HT Fast Real-Time PCR System (Applied Biosystems). The number of template transcript molecules in a sample is determined by recording the amplification cycle in the exponential phase (cycle threshold or C _T ), at which time the fluorescence signal can be detected above background fluorescence. Thus, the starting number of template transcript molecules is inversely related to C _T . The level of expression of a gene is measured using the "AACT method", briefly a gene is normalized by the value of one or a group of reference/housekeeping genes and/or by a reference sample such as a pooled sample or a commercially available reference such as the qPCR Human Universal Reference cDNA, random primed ; Ozyme ; ref 639654. In another preferred embodiment, the expression profile is determined by the use of a nucleic microarray.

According to the invention, a "nucleic microarray" consists of different nucleic acid probes that are attached to a substrate, which can be a microchip, a glass slide or a microsphere-sized bead. A microchip may be constituted of polymers, plastics, resins, polysaccharides, silica or silica-based materials, carbon, metals, inorganic glasses, or nitrocellulose. Probes can be nucleic acids such as cDNAs ("cDNA microarray") or oligonucleotides ("oligonucleotide microarray"), and the oligonucleotides may be about 25 to about 60 base pairs or less in length.

To determine the expression profile of a target nucleic sample, said sample is labelled, contacted with the microarray in hybridization conditions, leading to the formation of complexes between target nucleic acids that are complementary to probe sequences attached to the microarray surface. The presence of labelled hybridized complexes is then detected. Many variants of the microarray hybridization technology are available to the man skilled in the art.

In a preferred embodiment, the nucleic microarray is an oligonucleotide microarray comprising or consisting of oligonucleotides specific for:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s).

Preferably, the oligonucleotides are about 50 bases in length. It is acknowledged that the nucleic acid microarray or oligonucleotide microarray of the invention encompasses the microarrays specific for an Equivalent Expression Profile as defined above.

Suitable microarray oligonucleotides specific for any gene of Table 2 may be designed, based on the genomic sequence of each gene (see Table 2 Genbank accession numbers), using any method of microarray oligonucleotide design known in the art. In particular, any available software developed for the design of microarray oligonucleotides may be used, such as, for instance, the OligoArray software (available at http://berrv.enqin.umich.edu/oliqoarrav/), the GoArrays software (available at http://www.isima.fr/bioinfo/goarrays/), the Array Designer software (available at http://www.premierbiosoft.com/dnamicroarray/index.html), the Primer3 software (available at http://frodo.wi.mit.edu/primer3/primer3 code.html), or the Promide software (available at http://oliqos.molqen.mpq.de/). In another embodiment, the expression profile is determined by the use of a protein microarray.

In a particular embodiment of a method according to the invention, said method may further comprise determining at least one additional parameter useful for the diagnosis. Such "parameters useful for the diagnosis" are parameters that cannot be used alone for a diagnosis but that have been described as displaying significantly different values between responsive subjects and subjects who are clearly refractory and may thus also be used to refine and/or confirm the diagnosis according to the above described method according to the invention. They may notably include relevant clinical parameters depending on the inflammatory disease. For rheumatoid arthritis (RA), such clinical parameters include an assessment of the subject's pain, duration of morning stiffness, the number of swollen joints, the number of painful joints etc. Preferably, the parameters useful or diagnosis are determined from a non invasive biological sample of the subject. In particular, for RA, they may be selected from standard biological parameters specific for RA. According to the invention, "standard biological parameters specific for RA" are biological parameters usually used by clinicians to monitor the efficacy of a treatment of RA. These standard biological parameters specific for RA or autoimmune diseases usually comprise serum or plasma concentrations of particular proteins which are well known of those skilled in the art. The said standard biological parameters specific for RA can be determined by tests which include the Antinuclear Antibody test (ANA test), C-Reactive Protein test (CRP test), Erythrocyte sedimentation rate (ESR test), Cyclic Citrullinated Peptide Antibody test (CCP test), and the Rheumatoid Factor test. These tests are well known to the person skilled in the art and not be detailed here. They may be used on their own or in combination.

Such additional parameters may be used to confirm the diagnosis obtained using the expression profile comprising or consisting of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and and optionally one or more housekeeping gene(s).

The invention further concerns a kit for the in vitro diagnosis of a CyTD responsive or non responsive phenotype, comprising at least one reagent for the determination of an expression profile comprising, or consisting of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB,

KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and and optionally one or more housekeeping gene(s).

By "a reagent for the determination of an expression profile" is meant a reagent which specifically allows for the determination of said expression profile, i.e. a reagent specifically intended for the specific determination of the expression level of the genes comprised in the expression profile. This definition excludes generic reagents useful for the determination of the expression level of any gene, such as Taq polymerase or an amplification buffer, although such reagents may also be included in a kit according to the invention.

In a preferred embodiment of a kit according to the invention, said kit is dedicated to the in vitro diagnosis of a CyTD responsive or non responsive phenotype based on expression profiles comprising or consisting of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB,

KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s).

By "dedicated", it is meant that reagents for the determination of an expression profile in the kit of the invention essentially consist of reagents for determining the expression level of the above (i), (ii) or (iii) expression profiles, optionally with one or more housekeeping gene(s), and thus comprise a minimum of reagents for determining the expression of other genes than those mentioned in above described (i), (ii) or (iii) expression profiles and housekeeping genes. For instance, a dedicated kit of the invention preferably comprises no more than 50, 40, 30, 25, 20, preferably no more than 15, preferably no more than 10, preferably no more than 9, 8, 7, 6, 5, 4, 3, 2, or 1 reagent(s) for determining the expression level of a gene that does not belong to one of the above described (i), (ii) or (iii) expression profiles and that is not a housekeeping gene.

Such a kit for the in vitro diagnosis of a CyTD responsive or non responsive phenotype may further comprise instructions for determination of the presence or absence of a responsive phenotype. Such a kit for the in vitro diagnosis of a responsive phenotype may also further comprise at least one reagent for the determining of at least one additional parameter useful for the diagnosis such as standard biological parameters. In particular, the said reagent is useful for performing of any of the following tests: the Antinuclear Antibody test (ANA test), C-Reactive Protein test (CRP test), Erythrocyte sedimentation rate (ESR test), Cyclic Citrullinated Peptide Antibody test (CCP test), and the Rheumatoid Factor test.

In any kit for the in vitro diagnosis of a responsive phenotype according to the invention, the reagent(s) for the determination of an expression profile comprising, or consisting of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB,

KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s), preferably include specific amplification primers and/or probes for the specific quantitative amplification of transcripts of: (i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and and optionally one or more housekeeping gene(s), and/or

a nucleic microarray for the detection of:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes

MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and

(ii),

and optionally one or more housekeeping gene(s).

The determination of the expression profile may thus be performed using quantitative PCR and/or a nucleic microarray, preferably an oligonucleotide microarray, and/or protein microarrays.

In addition, the instructions for the determination of the presence or absence of a CyTD (notably TBA) phenotype preferably include at least one reference expression profile, or at least one reference sample for obtaining a reference expression profile. In a preferred embodiment, at least one reference expression profile is a responsive expression profile. Alternatively, at least one reference expression profile may be a non responsive expression profile. More preferably, the determination of the level of responsiveness is carried out by comparison with both responsive and non-responsive expression profiles as described above.

The invention is also directed to a nucleic acid microarray comprising or consisting of nucleic acids specific for:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or (ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s).

Said nucleic acid microarray may comprise additional nucleic acids specific for additional genes and optionally one or more housekeeping gene(s), but preferably consists of a maximum of 500, 400, 300, 200 preferably 100, 90, 80, 70 more preferably 60, 50, 45, 40, 35, 30, 25, 20, 15, 10, or even less (for instance 9, 8, 7, 6, 5, 4, 3, 2 or 1 ) distinct nucleic acids.

In a preferred embodiment, said nucleic acid microarray comprises no more than 50, 40, 30, 25, 20, preferably no more than 15, preferably no more than 10, preferably no more than 9, 8, 7, 6, 5, 4, 3, 2, or 1 distinct nucleic acids specific for a gene that does not belong to one of the above described (i), (ii) or (iii) expression profiles and that is not a housekeeping gene.

Advantageously, said microarray consists of nucleic acids specific for:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or

(ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iii) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s). In a preferred embodiment, said nucleic acid microarray is an oligonucleotide microarray comprising or consisting of oligonucleotides specific for:

(i) the gene MAPK14; or the genes MAPK14 and S100A9; or the genes MAPK14 and GNLY; or the 61 genes of Table 2, or (ii) the gene S100A9; or the genes S100A9, IL2RB, and CASP5; or the genes S100A9, IL2RB, KLRK1 , HCK, and GNLY; or the genes S100A9, IL2RB, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C, or

(iv) Equivalent Expression Profile of anyone of the expression profiles of (i) and (ii),

and optionally one or more housekeeping gene(s).

Having generally described this invention, a further understanding of characteristics and advantages of the invention can be obtained by reference to certain specific examples and figures which are provided herein for purposes of illustration only and are not intended to be limiting unless otherwise specified.

DESCRIPTION OF THE FIGURES Figure 1 : The two most significantly enriched GO terms for the top differentially expressed 8 genes are: "Positive Regulation of interferon gamma production" (A) and "Positive regulation of natural killer cell mediated cytotoxicity" (B) (p-value <5x10 ^"3 ). Figure 2: Pathway analysis for MAPK14 and S100A9 using PathGen

Figure 3 : The top differentially expressed 8 genes issued from Ingenuity® System analysis belong to a single biological network (Cell-to-cell signalling and interaction, haematological system development and function, immune cell trafficking) In Figure 3, symbols represent different types of molecules;

upward triangles: phosphatase,

downward triangles: kinase,

- horizontal ellipses: transcription regulator,

vertical ellipses: transmembrane receptor,

rhombuses: enzyme,

horizontal rectangles: ligand-dependent nuclear receptor,

vertical rectangles: g-protein coupled receptor,

- simple circles: other types of molecules,

double circles: complex group,

Shaded shapes represent the molecules that are part of the 61 list.

The types of line represent the type of interaction:

plain arrows: means the first molecule directly acts on the second one (such as but not limited to activation)

dotted arrows: means the first molecule indirectly acts on the second one A line ending with a short segment indicates inhibition

A simple line represents interaction (such a but not limited to protein-protein interaction) EXAMPLES

Example 1 : meta-analysis of two datasets (Bienkowska et al.Q) and Julia et al.(8)) Materials and Methods

In this example, the materials and methodologies used in subsequent examples 1 to 3 are described.

Data Identification and Data Extraction: Studies were selected on the basis that they had been performed on RA patients naive to biologies who had started therapy with Infliximab and measurement of their response to treatment was available at 14 or 22 weeks. Large scale gene expression information had to be available at baseline (prior to treatment). Following the steps described in Ramasamy et al. (10), we identified four studies that matched our research criteria: Lequerre et al. (6), Sekiguchi et al. (7), Bienkowska et al. (9) and Julia et al. (8). The expression data, the phenotypes and the annotation data were all downloaded from GEO (GSE3592, GSE8350, GSE12051 and GSE15258 respectively).

All four studies identified "Gene expression signatures of response to anti-TNF therapy". Interestingly, however, no two publications used the same approach and this can partly explain the lack of overlap between the reported signatures. To make the four studies more comparable, we contacted the authors to obtain additional individual information such as the DAS28 at baseline and week 14 or week 22 to use a single definition of response (EULAR criteria - with "moderate" and "good" responders considered as responders) or detail of treatment to ensure that only Infliximab-treated patients were analyzed. We therefore reclassified patients as responders based on the EULAR definition at week 14 and week 22 and performed a binary analysis of good and moderate responders versus non-responders. This binary grouping is particularly suited for the identification of non responders. The final dataset is summarized in Table 1 and was the most homogeneous data we could obtain.

Study Bienkowska et Lequerre ef

Julia et al. Sekiguchi ef al. Reference al. al.

whole

Sample type whole blood PBMCs whole blood

blood

Table 1 : Summary of datasets after alignment of clinical criteria

Data Quality and Processing: Data from Bienkowska et al. was the only data for which we downloaded the raw .CEL files and processed them using our internal protocols (normalization was performed using GC-RMA in refiner array by GENEDATA Expressionist® (Genedata AG, Basel, Switzerland)). Six chips were flagged with quality issues due to increased distortion. However, due to the limited number of arrays available for Infliximab samples, we included them in our analysis.

Data from Lequerre et al., Julia et al. and Sekiguchi et al. were all downloaded as expression matrices, which correspond to expression values after normalization. The data by Lequerre et al. included technical replicates; we averaged the technical replicates and excluded the control samples from the analysis.

To translate probe information into gene information we used the NCBI's Geneld as provided in the respective annotation files found on GEO. When multiple probes were available, we selected probes with the same NCBI Gl number across platforms. When probes differed in terms of Gl, they were selected randomly. Therefore for each gene, only one probe contributed to the analysis.

Statistical Analysis: Because of the limited number of probes on the arrays used by Sekiguchi et al. and the major difference introduced by the sample source in the study by Lequerre et al. (PBMCs as opposed to whole blood), we initially performed a metaanalysis only on the Bienkowska et al. and Julia et al. data sets. We then performed the meta-analysis adding the data from Sekiguchi et al. and lastly, the data from Lequerre et al. Thus three meta-analyses were performed (examples 1 , 2, and 3 respectively).

The statistical analysis was performed in R using the MetaArray package (D Ghosh and H Choi, 2009). This package implements the latent variable model described in (H Choi, 2005) as well as the integrative correlation (12). For the probability of expression (POE) estimation we used the EM estimation method. Probes were filtered on 100% presence and based on an average correlation >-0.2. Individual probes contribution was estimated using the t-test as implemented in the multtest library in R. We set the threshold for significance at 2.5 based on visual inspection of the distribution plots. The list of probes identified using this approach was then further evaluated in the individual data sets to assess their predictive performance. This was done in Genedata Expressionist® using the Fisher LDA predictor and leave-one-out cross validation. The assessment of biological processes enrichment was done in Genedata Expressionist® (Genedata AG, Basel, Switzerland) using the Gene Ontology Fisher's Exact Test function. The list of significant probes was compared to the list of probes that were tested (the common probes). Pathway analysis was performed using PathGen (available at http://dna.cs.byu.edu/pathgen).

Results

The initial meta-analysis of the two datasets (i.e. Bienkowska et al. and Julia et al.) was performed on 4022 probes. Some heterogeneity between the two studies could be due to remaining clinical differences such as severity of disease, use of mono- versus bi- therapy or to probe differences due to the different platforms used (Affymetrix versus lllumina microarrays). Based on the t-test of the POE from the merged datasets, we ranked the genes based on their significance. Table 2 provides the gene symbol and direction of the 61 most differentially expressed genes between responders and non responders. Interestingly this gene list overlaps only by two genes with the genes identified by Julia et al. and by one gene identified by Bienkowska et al. To assess the discriminatory performance of a group of genes we selected the top 8 genes and applied them to the individual data sets. The performance can be found in Table 3. This eight gene set performs extremely well in the dataset of Julia et al.: by LOO, 7 out of 44 samples were misclassified giving an overall error rate of 84%, the PPV was as high as 97%.

Gene Symbol Name Accession Nb Direction

NM_000878.2

IL2RB interleukin 2 receptor, beta Increased

(SEQ ID NO :1 ) NM_002965.3

S100A9 S100 calcium binding protein A9 Decreased

(SEQ ID NO :2)

killer cell lectin-like receptor NM_007360.2

KLRK1 Increased subfamily K, member 1 (SEQ ID NO :3)

NM_001 172129.1

HCK hemopoietic cell kinase Decreased

(SEQ ID NO :4)

NM_006433.3

GNLY granulysin Increased

(SEQ ID NO :5)

NM_001336.3

CTSZ cathepsin Z Decreased

(SEQ ID NO :6)

NM_001 142272.1

ARF5 ADP-ribosylation factor 5 Decreased

(SEQ ID NO :7)

U3 small nucleolar NM_021645.5

UTP14C Increased ribonucleoprotein, homolog C (SEQ ID NO :8)

DR1 -associated protein 1 (negative NM_006442.3

DRAP1 Decreased cofactor 2 alpha) (SEQ ID NO :9)

NM_138392.3

SHKBP1 SH3KBP1 binding protein 1 Decreased

(SEQ ID NO :10)

NM_002161.4

IARS isoleucyl-tRNA synthetase Increased

(SEQ ID NO :1 1 )

ATPase, H ⁺ transporting, lysosomal NM_001039367.1

ATP6V1 E1 Decreased

31 kDa, V1 subunit E1 (SEQ ID NO :12)

Ribonuclease P/MRP 21 kDa NM_024839.1

RPP21 Increased subunit (SEQ ID NO :13)

NM_001039465.1

SFRS5 splicing factor, arginine/serine-rich 5 Increased

(SEQ ID NO :14)

chromosome 5 open reading frame NM_175921.4

LOC285636 Increased

51 (SEQ ID NO :15)

NM_032536.2

NTNG2 netrin G2 Decreased

(SEQ ID NO :16)

ASF1 anti-silencing function 1 NM_014034.2

ASF1A Increased homolog A (SEQ ID NO :17)

NM_001 130455.1

DYSF dysferlin Decreased

(SEQ ID NO :18)

eukaryotic translation initiation NM_031992.1

EIF4H Increased factor 4 H (SEQ ID NO :19) NM_006590.2

USP39 ubiquitin specific peptidase 39 Increased

(SEQ ID NO :20)

Anaphase promoting complex NM_013367.2

ANAPC4 Increased subunit 4 (SEQ ID NO :21 )

NM_018283.1

NUDT15 nudix Decreased

(SEQ ID NO :22)

NM_014408.3

TRAPPC3 trafficking protein particle complex 3 Increased

(SEQ ID NO :23)

signal recognition particle receptor, NM_021203.3

SRPRB Increased

B subunit (SEQ ID NO :24)

NM_023079.3

UBE2Z ubiquitin-conjugating enzyme E2Z Increased

(SEQ ID NO :25)

NM_002902.2

RCN2 reticulocalbin 2 Increased

(SEQ ID NO :26)

NM_004251.3

RAB9A member RAS oncogene family Increased

(SEQ ID NO :27)

Shwachman-Bodian-Diamond NM_023248.1

SBDS Increased syndrome (SEQ ID NO :28)

NM_015293.2

(SEQ ID NO :29)

NM_033071.2

spectrin repeat containing, nuclear (SEQ ID NO :30)

SYNE1 Increased envelope 1 NM_133650.2

(SEQ ID NO :31 )

NM_182961.2

(SEQ ID NO :32)

NM_005091.1

PGLYRP1 peptidoglycan recognition protein 1 Decreased

(SEQ ID NO :33)

carbohydrate kinase domain NM_018210.2

FLJ10769 Increased containing (SEQ ID NO :34)

NM_005926.2

MFAP1 microfibrillar-associated protein 1 Increased

(SEQ ID NO :35)

NM_0021 15.2

HK3 hexokinase 3 Decreased

(SEQ ID NO :36) myeloid/lymphoid or mixed-lineage

NM_006818.3

MLLT1 1 leukemia (trithorax homolog, Increased

(SEQ ID NO :37)

Drosophila); translocated to, 1 1

NM_003272.3

GPR137B G protein-coupled receptor 137B Increased

(SEQ ID NO :38)

NM_001040034.1

(SEQ ID NO :39)

CD63 CD63 Decreased

NM_001780.4

(SEQ ID NO :40)

NM_152334.2

TARSL2 threonyl-tRNA synthetase-like 2 Increased

(SEQ ID NO :41 )

NM_032646.5

TTYH2 tweety homolog 2 Increased

(SEQ ID NO :42)

NM_016098.2

BRP44L brain protein 44-like Increased

(SEQ ID NO :43)

NM_015942.3

MTERFD1 MTERF domain containing 1 Increased

(SEQ ID NO :44)

NM_001 1361 1 1.1

CASP5 caspase 5 Decreased

(SEQ ID NO :45)

NM_153005.1

(SEQ ID NO :46)

RIOK1 RIO kinase 1 Increased

NM_031480.2

(SEQ ID NO :47)

NM_006493.2

CLN5 ceroid-lipofuscinosis, neuronal 5 Increased

(SEQ ID NO :48)

zinc finger CCCH-type containing NM_014153.2

ZC3H7A Increased

7A (SEQ ID NO :49)

NM_02461 1.4

(SEQ ID NO :50)

NARG2 NMDA receptor regulated 2 Increased

NM_001018089.1

(SEQ ID NO :51 )

NM_016454.2

TMEM85 transmembrane protein 85 Increased

(SEQ ID NO :52)

NM_015456.3

COBRA1 cofactor of BRCA1 Increased

(SEQ ID NO :53) NM_012289.3

(SEQ ID NO :54)

KEAP1 kelch-like ECH-associated protein 1 Decreased

NM_203500.1

(SEQ ID NO :55)

leucine-rich repeats and calponin NM_032773.2

LRCH3 Increased homology (CH) domain containing 3 (SEQ ID NO :56)

NM_001031726.2

chromosome 19 open reading (SEQ ID NO :57)

C19orf12 Increased frame 12 NM_031448.3

(SEQ ID NO :58)

NM_002642.3

phosphatidylinositol glycan anchor (SEQ ID NO :59)

PIGC Increased biosynthesis, class C NM_153747.1

(SEQ ID NO :60)

NM_032564.3

DGAT2 diacylglycerol O-acyltransferase 2 Decreased

(SEQ ID NO :61 )

NM_019082.2

DDX56 DEAD box polypeptide 56 Increased

(SEQ ID NO :62)

NM_024946.2

NIP30 NEFA-interacting nuclear protein Increased

(SEQ ID NO :63)

NM_001315.2

(SEQ ID NO :64)

NM_139012.2

(SEQ ID NO :65)

MAPK14 mitogen-activated protein kinase 14 Decreased

NM_139013.2

(SEQ ID NO :66)

NM_139014.2

(SEQ ID NO :67)

NM_001 127257.1

solute carrier family 39 (zinc (SEQ ID NO :68)

SLC39A10 Increased transporter), member 10 NM_020342.2

(SEQ ID NO :69)

NM_000671.3

ADH5 alcohol dehydrogenase 5 Increased

(SEQ ID NO :70)

NM_015325.1

KIAA0947 KIAA0947 Increased

(SEQ ID NO :71 ) glioma tumor suppressor candidate NM_015710.4

GLTSCR2 Increased region gene 2 (SEQ ID NO :72)

NM_001 1351 10.1

DnaJ (Hsp40) homolog, subfamily (SEQ ID NO :73)

DNAJA3 Increased

A, member 3 NM_005147.4

(SEQ ID NO :74)

family with sequence similarity 134, NM_178126.3

FAM134C Increased member C (SEQ ID NO :75)

Table 2: Ranked list of first 61 differentially expressed probes. The top genes are the most significant ones.

Table 3: Discriminatory performance of the gene set based on the top 8 genes the data of Julia et al.

The top 8 most significantly differentially expressed genes (IL2RB, S100A9, KLRK1 , HCK, GNLY, CTSZ, ARF5, and UTP14C) have all been associated to the same biological network ("Cell-to-cell signalling and interaction, haematological system development and function, immune cell trafficking") by an Ingenuity® Systems analysis (Figure 3).

Furthermore, the top 5 most significantly differentially expressed genes (IL2RB, S100A9, KLRK1 , HCK, GNLY) are particularly relevant to RA physiopathology.

IL2RB (lnterleukin-2 receptor subunit beta)

IL2RB gene encodes the beta subunit of the IL2R, which is present in the moderate and high affinity forms of the receptor required for signal transduction from IL2. Several studies showed that SNPs in IL2RB gene (especially rs743777) are significantly associated with RA (13).

It has been recently proposed that thymic selection of an auto-reactive T-cell repertoire is an important risk factor for rheumatoid arthritis (RA) (14). The IL2RB gene is thus an attractive candidate for RA because of the key role played by IL2 in T cell activation and regulation.

S100A9 (S100 calcium-binding protein A9)

S100A9 gene codes for the S100A9 protein also called myeloid-related protein 14 (MRP14) or calgranulin B that belongs to the S100 family of proteins. S100 proteins are known to be associated with several pathological conditions such as cystic fibrosis (15), cancer (16) and inflammatory diseases (17).

S100A9 has also been shown to induce neutrophil degranulation (which may account for the tissue damage in RA) via a MAPK-dependent mechanism (18).

A recently published study showed that S100 proteins are associated with RA inflammation and auto-antibody production (19).

KLRK1

KLRK1 gene encodes the natural killer group 2D (NKG2D) protein which belongs to the family of activating NK cell receptors (NKRs).

In RA patients, a significant proportion of CD4 ⁺ CD28 ^" T cells were shown to express NKG2D which stimulated auto-reactive responses against RA synoviocytes (20).

HCK

HCK gene encodes the tyrosine-protein kinase HCK (hematopoietic cell kinase) predominantly expressed in hematopoietic cell types.

HCK has been shown to be involved in functional pro-inflammatory responses such as the FcR-mediated respiratory burst (21 ), neutrophil migration (22) and neutrophil degranulation (23) that play important roles in the early phases of RA physiopathology Furthermore, it has also been shown that HCK mediates IL-2 signaling in human monocytes (24). GNLY

GNLY gene codes for Granulysin, a saponin-like protein present in cytotoxic granules of cytolytic T cells and NK cells and released upon antigen stimulation.

Granulysin is a chemoattractant for T lymphocytes, monocytes and other inflammatory cells and induces the expression of a number of cytokines, including RANTES, MCP-1 , MCP-3, MIP-1 a, IL-10, IL-1 , IL-6 and IFN-a (25).

Granulysin is also involved in several diseases including infection, cancer, transplantation, autoimmunity, skin and reproductive maladies (26). Example 2: meta-analysis of three datasets (Bienkowska et al. Julia et al, and Sekiguchi et al.)

A second meta-analysis was performed after merging the dataset of Sekiguchi et al. This dramatically reduced the number of genes being tested, bringing it down to 290. Keeping the same threshold for the test statistic as previously (2.5) only three of the 61 previously described (see Example 1 ) genes were significant: IL2RB, S100A9 and CASP5. However, these genes widen the prediction of the Infliximab treatment to a 22 weeks follow-up. Besides the relationship between IL2RB and S100A9 and physiopathology of RA (see above), CASP5 gene also has a relevant significance in this auto-immune disorder.

CASP5

CASP5 gene encodes the Caspase 5 enzyme that proteolytically cleaves other proteins at an aspartic acid residue. It is an inflammatory caspase that plays a role in the immune system (27) and in the complex process of cellular apoptosis (28).

The aberrant decrease in apoptosis or increased cell cycle activity of fibroblast-like or macrophage-like synoviocytes is responsible for the synovial hyperplasia and contributes to the destruction of cartilage and bone in RA patients thus suggesting a potential role of apoptosis-related caspases in the physiopathology of RA (29). Example 3: meta-analysis of four datasets (Bienkowska et al. Julia et al,

Sekiguchi et al., and Lequerre et al.)

Further adding the data by Lequerre et al., and thus leading to a total number of 103 samples analysed in 179 genes identified only two of the 61 previously described (see Example 1 ) genes that are significantly increased in non responders: MAPK14 et S100A9 (Figure 2). The combination of these two genes is thus able to predict response to Infliximab treatment on a biological sample (whole blood or PBMCs) at week 14 or week 22.

MAPK14

MAPK14 gene encodes the mitogen-activated protein kinase 14, a member of the MAP kinase family. MAP kinases act as an integration point for multiple biochemical signals, and are involved in a wide variety of cellular processes such as proliferation, differentiation, transcription regulation and development. The MAP kinase pathway is particularly influencing the biosynthesis of pro-inflammatory cytokines TNFa and I L-1 β both at a translational and a transcriptional level (30). This pathway might thus be a key anti-inflammatory target for RA treatment.

Example 4: quantitative PCR confirmation of the efficiency of signatures obtained from meta-analysis of microarrav data

Patients and methods

Patients

Peripheral blood samples from 40 patients, with rheumatoid arthritis and classified as responders and non responders to Infliximab treatment at week 14 according to the EULAR definition of response (Good and moderate responders were classified as responders and poor responders were classified as non responders), were identified and analyzed by RT-PCR. The clinical and demographic variables of the patients are summarized in Table 4.

Table 4: Patient clinical characteristics

Methods

Peripheral blood was collected into RNA PAXgene® tubes (PreAnalytix, Switzerland) and RNA from whole blood was extracted according to the methods described in Julia et al. (8). Samples underwent quality control and were dried using speed vacuum prior to the RT and PCR steps. The real-time PCR reactions was carried out on a total of 40 samples using the TaqMan Universal PCR Master Mix (Applied Biosystems). 6 μΙ_ cDNA was added to a 9 μΙ_ PCR mixture containing 7.5 μΙ_ TaqMan Universal PCR Master Mix, 0.75 μΙ_ of a 20X mixture of probe and primers and 0.75μΙ water. The reaction consisted of one initiating step of 2 min at 50 deg. C, followed by 10 min at 95 deg. C, and 40 cycles of amplification including 15 sec at 95 deg. C and 1 min at 60 deg. C. The reaction and data acquisition was performed using the ABI 7900HT Fast Real-Time PCR System (Applied Biosystems). The "AACT method" was used as a measure of gene expression. Post extraction steps and analyses were carried out at TcLand Expression IS013485 laboratory.

For 7 out of the 1 1 genes, probes were custom designed and the sequence information can be found in Table 5a. For 3 out of the 1 1 genes, probes were ordered from Applied Biosystems directly and are referenced in Table 5b.

Table 5a : Sequence information for the 7probes that were custom designed.

Table 5b: Probe information for the three genes for which probes were ordered from Applied Biosystems repositoryResults Following the results from the meta-analysis, Taqman probes were ordered through applied or designed in house. After our internal quality control steps 8 out of the 1 1 genes of the present invention were tested on the RA samples. The 40 samples, that represent a subset of the original samples used for microarray analysis in the study of Julia et al., were analyzed by RT-PCR following our internal protocols. The following statistical analysis has been performed: Identification of individually differentially expressed genes between the two groups of responders versus non responders (Table 6). The selection criteria used was a significant t-test of at least 0.05. Additionally the discriminatory power of the combination of 1 , 2, 5 and 8 genes listed in the invention was assessed using a logistic regression and classification rates were provided (Table 7).

Table 6: P-values and ΔΔΟΤ for Responder and Non Responder groups. Genes underlined and in italic were confirmed as being discriminant (P-value<0.05). Model Sensitivity Specificity Error

Rate

S100A9 0% 100% 16%

MAPK14, S100A9 83% 96% 5%

IL2RB, S100A9, KLRK1 ,HCK and GNLY 83% 96% 6%

IL2RB, S100A9, KLRK1 , HCK, GNLY, 100% 100% 0% CTSZ, ARF5 and UTP14C

Table 7: Classification performance using logistic regression of claimed list gene combinations.

Conclusion

The above results confirm that the particular combinations (signatures) of genes identified as predicting the response to infliximab of RA-suffering patients based on meta-analysis of microarray data obtained in several independent studies are actually efficient (high sensititity and specificity, low error rate) for predicting the response to influximab treatment of a validation group of RA-suffering patients.

Example 5: Prediction of the response to Infliximab at week 14 of included patients using different sub-combinations of genes

Patients

The same patients as in Example 4 have been studied. Results

The identification of an optimal subgroup of genes that together best discriminate between the responders and non responders was then performed. A logistic regression was applied and the classification error was used to identify the optimal subset of genes.

The most discriminating gene on its own is MAPK14 (Table 7), the optimal two-gene combination is MAPK14 with GNLY. Model Sensitivity Specificity Error

Rate

MAPK14 50% 97% 1 1 %

MAPK14, GNLY 100% 100% 0%

Table 7: Classification performance using logistic regression of claimed list gene combinations.

Conclusion

These results show that MAPK14, already identified as significant in Example 3 by meta-analysis of microarray data obtained from 4 independent studies, is actually highly predictive, even alone, of response to Infliximab treatment in RA-suffering patients. Moreover, in addition to the combination of genes S100A9 and MAPK14, already identified as useful in Example 3 and confirmed by qPCR experiments in Example 4, the combination of genes MAPK14 and GNLY (GNLY being already identified as significant in Example 1 ) is also highly predictive of response to Infliximab in RA-suffering patients, all patients tested being correctly classified using this combination.

Globally, the results presented in Examples 1 to 5 support the tight correlation of two individual genes, MAPK14 and S100A9, or equivalent correlated genes, with the Infliximab responsive or non-responsive phenotype of RA-suffering subjects. In addition, combination of one or both of these genes with a small number of other genes found associated to Infliximab responsive or non-responsive phenotype of RA-suffering subjects (notably the 61 genes of Table 2, and more particularly genes GNLY, IL2RB, S100A9, KLRK1 , HCK, CTSZ, ARF5, and UTP14C), permit a highly sensitive and specific (very low error rate) prediction of the Infliximab responsive or non-responsive phenotype of RA-suffering subjects.

Finally, we note that genes found to be highly correlated to the Infliximab responsive or non-responsive phenotype of RA-suffering subjects are not genes that might be involved in the metabolism of Infliximab, but rather genes that may be associated to the disease or its underlying dysfunctions themselves. In particular, many genes found to be highly correlated to the Infliximab responsive or non-responsive phenotype of RA- suffering subjects are known to be involved in inflammatory or immune processes. This clearly gives a rational for the extension of the methods of the invention for the prediction of the TBA responsive or non-responsive phenotype of subjects suffering from other inflammatory diseases, in particular those involving pathogenic TNFoc secretion and even more particularly those for which TBA have been approved or have been shown in preliminary studies to be useful.

Example 6: Correlation between gene expression and inflammatory response measured by DAS28, CRP and ESR in rheumatoid arthritis patients

Clinical parameters measuring inflammation at baseline such as DAS28, C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) do not discriminate between responders and non responders (respective p-values for t-test are 0.7792, 0.4839 and 0.1755). To assess whether the gene expression levels are correlated to the clinical parameters of inflammation, a correlation coefficient was computed (Table 8). Table 8 indicates that the genes in the signature are not correlated to the clinical parameters and thus add independent information. Only the gene expression levels of IL2RB correlate negatively to CRP (correlation coefficient = -0.35, p-value 0.03), please note this p-value is uncorrected for multiple testing.

Table 8: Correlation coefficient of gene expression versus clinical parameters indicate lack of correlation between gene expression levels and clinical parameters of inflammation. These results support the claim that our gene expression signature provides independant discriminatory power over clinical variables.

REFERENCES

1 . Lee DM, Weinblatt ME. Rheumatoid arthritis. Lancet. 2001 Sep 15;358(9285):903-1 1 .

2. Choy EH, Panayi GS. Cytokine pathways and joint inflammation in rheumatoid arthritis. N Engl J Med. 2001 Mar 22;344(12):907-16. 3. Kooloos WM, de Jong DJ, Huizinga TW, Guchelaar HJ. Potential role of pharmacogenetics in anti-TNF treatment of rheumatoid arthritis and Crohn's disease. Drug Discovery Today. 2007;12(3-4):125-31.

4. Isaacs JD. Antibody engineering to develop new antirheumatic therapies. Arthritis Res Ther. 2009;1 1 (3):225.

5. Hetland ML, Christensen IJ, Tarp U, Dreyer L, Hansen A, Hansen IT, et al. Direct comparison of treatment responses, remission rates, and drug adherence in patients with rheumatoid arthritis treated with adalimumab, etanercept, or infliximab: results from eight years of surveillance of clinical practice in the nationwide Danish DANBIO registry. Arthritis Rheum. 2010 Jan;62(1 ):22-32.

6. Lequerre T, Gauthier-Jauneau AC, Bansard C, Derambure C, Hiron M, Vittecoq O, et al. Gene profiling in white blood cells predicts infliximab responsiveness in rheumatoid arthritis. Arthritis Res Ther. 2006;8(4):R105.

7. Sekiguchi N, Kawauchi S, Furuya T, Inaba N, Matsuda K, Ando S, et al. Messenger ribonucleic acid expression profile in peripheral blood cells from RA patients following treatment with an anti-TNF-alpha monoclonal antibody, infliximab. Rheumatology (Oxford). 2008 Jun;47(6):780-8.

8. Julia A, Erra A, Palacio C, Tomas C, Sans X, Barcelo P, et al. An eight-gene blood expression profile predicts the response to infliximab in rheumatoid arthritis. PLoS One. 2009;4(10):e7556.

9. Bienkowska JR, Dalgin GS, Batliwalla F, Allaire N, Roubenoff R, Gregersen PK, et al. Convergent Random Forest predictor: methodology for predicting drug response from genome-scale data applied to anti-TNF response. Genomics. 2009 Dec;94(6):423-32.

10. Ramasamy A, Mondry A, Holmes CC, Altman DG. Key issues in conducting a meta-analysis of gene expression microarray datasets. PLoS Med. 2008 Sep 30;5(9):e184.

1 1 . Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum. 1988 Mar;31 (3):315-24.

12. Parmigiani G, Garrett-Mayer ES, Anbazhagan R, Gabrielson E. A cross-study comparison of gene expression studies for the molecular classification of lung cancer. Clin Cancer Res. 2004 May 1 ;10(9):2922-7. 13. Barton A, Thomson W, Ke X, Eyre S, Hinks A, Bowes J, et al. Rheumatoid arthritis susceptibility loci at chromosomes 10p15, 12q13 and 22q13. Nat Genet. 2008 Oct;40(10):1 156-9.

14. Goronzy JJ, Weyand CM. Developments in the scientific understanding of rheumatoid arthritis. Arthritis Res Ther. 2009;1 1 (5):249.

15. Lorenz E, Muhlebach MS, Tessier PA, Alexis NE, Duncan Hite R, Seeds MC, et al. Different expression ratio of S100A8/A9 and S100A12 in acute and chronic lung diseases. Respir Med. 2008 Apr;102(4):567-73.

16. Cheng P, Corzo CA, Luetteke N, Yu B, Nagaraj S, Bui MM, et al. Inhibition of dendritic cell differentiation and accumulation of myeloid-derived suppressor cells in cancer is regulated by S100A9 protein. J Exp Med. 2008 Sep 29;205(10):2235-49.

17. Lim SY, Raftery M, Goyette J, Hsu K, Geczy CL. Oxidative modifications of S100 proteins: functional regulation by redox. J Leukoc Biol. 2009 86(3): 577-87.

18. Simard JC, Girard D, Tessier PA. Induction of neutrophil degranulation by S100A9 via a MAPK-dependent mechanism. J Leukoc Biol. 2010 Jan 26.

19. Chen YS, Yan W, Geczy CL, Brown MA, Thomas R. Serum levels of soluble receptor for advanced glycation end products and of S100 proteins are associated with inflammatory, autoantibody, and classical risk markers of joint and vascular damage in rheumatoid arthritis. Arthritis Res Ther. 2009;1 1 (2):R39.

20. Groh V, Bruhl A, El-Gabalawy H, Nelson JL, Spies T. Stimulation of T cell autoreactivity by anomalous expression of NKG2D and its MIC ligands in rheumatoid arthritis. Proc Natl Acad Sci U S A. 2003 Aug 5;100(16):9452-7.

21 . Paul R, Obermaier B, Van Ziffle J, Angele B, Pfister HW, Lowell CA, et al. Myeloid Src kinases regulate phagocytosis and oxidative burst in pneumococcal meningitis by activating NADPH oxidase. J Leukoc Biol. 2008 Oct;84(4):1 141 -50.

22. Fumagalli L, Zhang H, Baruzzi A, Lowell CA, Berton G. The Src family kinases Hck and Fgr regulate neutrophil responses to N-formyl-methionyl-leucyl-phenylalanine. J Immunol. 2007 Mar 15;178(6):3874-85.

23. Mocsai A, Ligeti E, Lowell CA, Berton G. Adhesion-dependent degranulation of neutrophils requires the Src family kinases Fgr and Hck. J Immunol. 1999 Jan

15; 162(2): 1 120-6.

24. Bosco MC, Curiel RE, Zea AH, Malabarba MG, Ortaldo JR, Espinoza-Delgado I. IL-2 signaling in human monocytes involves the phosphorylation and activation of p59hck. J Immunol. 2000 May 1 ;164(9):4575-85. 25. Deng A, Chen S, Li Q, Lyu SC, Clayberger C, Krensky AM. Granulysin, a cytolytic molecule, is also a chemoattractant and proinflammatory activator. J Immunol. 2005 May 1 ;174(9):5243-8.

26. Krensky AM, Clayberger C. Biology and clinical relevance of granulysin. Tissue Antigens. 2009 Mar;73(3):193-8.

27. Martinon F, Tschopp J. Inflammatory caspases and inflammasomes: master switches of inflammation. Cell Death Differ. 2007 Jan;14(1 ):10-22.

28. Kurokawa M, Kornbluth S. Caspases and kinases in a death grip. Cell. 2009 Sep 4;138(5):838-54.

29. Morel J, Audo R, Hahne M, Combe B. Tumor necrosis factor-related apoptosis- inducing ligand (TRAIL) induces rheumatoid arthritis synovial fibroblast proliferation through mitogen-activated protein kinases and phosphatidylinositol 3-kinase/Akt. J Biol Chem. 2005 Apr 22;280(16):15709-18.

30. Korb A, Tohidast-Akrad M, Cetin E, Axmann R, Smolen J, Schett G. Differential tissue expression and activation of p38 MAPK alpha, beta, gamma, and delta isoforms in rheumatoid arthritis. Arthritis Rheum. 2006 Sep;54(9):2745-

31 . Fransen J, van Riel PLCM. The Disease Activity Score and the EULAR response criteria. Clin Exp Rheumatol. 2005 23(5 Suppl 39): S93-9.

32. Lorenz HM et al. Arthritis Res. 2002;4 Suppl 3:S17-24

33. Atzeni F et al. Autoimmun Rev. 2007 Sep;6(8):529-36.