cD11c as a Transcriptional Biomarker to Predict Response ... ... cD11c as a Transcriptional Biomarker

Embed Size (px)

Text of cD11c as a Transcriptional Biomarker to Predict Response ... ... cD11c as a Transcriptional...

  • CliniCal pharmaCology & TherapeuTiCs | VOLUME 87 NUMBER 3 | MaRch 2010 311

    articlesnature publishing group

    Tumor necrosis factor (TNF), a cytokine known to be produced by monocytes,1,2 promotes the synthesis of proinflammatory cytokines, stimulates endothelial cells to express adhesion molecules, and accelerates the synthesis of metalloproteinas- es.1 Systemic TNF neutralization in rheumatoid arthritis (RA) responders is associated with a remarkable suppression of local and systemic inflammation, as well as with a reduction in bone destruction1,3 and radiological progression of joint destruction.4 Thus, TNF blockade inhibits inflammatory mechanisms and local tissue destruction initiated by cells of the monocytic lineage.5

    TNF-blocking agents lead to the control of RA activity in 60–80% of patients when used in combination with meth- otrexate (MTX). Gene profiling at baseline shows that levels of different small subsets of transcripts are suitable for reliably predicting response to either infliximab/MTX in MTX-resistant RA subjects6 or etanercept in combination with various disease- modifying antirheumatic drugs.7

    Approximately 40% of RA patients either do not respond at all or only inadequately respond to anti-TNF monotherapy,8–10 and predictors of outcome for monotherapies are currently not available. This is a major drawback, considering the risk involved in exposing nonresponders to the sometimes serious side effects of TNF blockade11–15 (http://jama.ama-assn.org/ cgi/reprint/296/18/2203.pdf; http://jama.ama-assn.org/cgi/ reprint/296/18/2203-a.pdf) and the substantial costs of this treatment.

    Assessing gene-expression changes in circulating blood cells6,7,16 is an attractive approach for the following reasons: the blood transcriptome constitutes a source of potential markers; response to factors undetectable in the serum may be measured in circulating immune cells; and commercial microarrays, at affordable costs and with good reproducibility, allow measure- ment of blood transcriptional profiles on an unprecedented scale using mainstream microarray technology. As a result, a

    1Department of Rheumatology and clinical Immunology, charité University hospital, humboldt University of Berlin, Berlin, Germany; 2Deutsches Rheumaforschungszentrum, cell Sorting Unit, Berlin, Germany; 3Laboratory of Functional Genomics, charité-Universitätsmedizin Berlin, Berlin, Germany; 4abbott Bioresearch center, Worcester, Massachusetts, USa; 5Department of Orthopedics, Experimental Rheumatology Unit, University hospital Jena, Jena, Germany. correspondence: B Stuhlmüller (bruno.stuhlmueller@charite.de)

    Received 6 July 2009; accepted 16 October 2009; advance online publication 23 December 2009. doi:10.1038/clpt.2009.244

    cD11c as a Transcriptional Biomarker to Predict Response to anti-TNF Monotherapy With adalimumab in Patients With Rheumatoid arthritis B Stuhlmüller1, T Häupl1, MM Hernandez1, A Grützkau2, R-J Kuban3, N Tandon1, JW Voss4, J Salfeld4, RW Kinne5 and GR Burmester1

    We performed transcription profiling using monocytes to identify predictive markers of response to anti–tumor necrosis factor (anti-TnF) therapy in patients with rheumatoid arthritis (ra). several potential predictors of response were identified, including CD11c. Validation in samples from independent cohorts (total of n = 27 patients) using reverse transcription–pCr confirmed increased expression of CD11c in responders to adalimumab (100% sensitivity; 91.7% specificity, power 99.6%; α = 0.01). pretherapy CD11c levels significantly correlated with the response criteria as defined by the american College of rheumatology (aCr) (r = 0.656, P < 0.0001). however, CD11c was neither predictive of response to methotrexate (mTX) alone (n = 34) nor to mTX in combination with adalimumab (n = 16). Clinical responders revealed a reset to a normal expression pattern of resident/inflammatory monocyte markers, which was absent in nonresponders. Therefore, an analysis of key cell types identifies potentially predictive biomarkers that may help to restrict the use of adalimumab to therapy responders. larger studies, including studies of monotherapy with other drugs, are now needed to confirm and validate the specificity of CD11c for anti-TnF biologics.

    mailto:bruno.stuhlmueller@charite.de http://www.nature.com/doifinder/10.1038/clpt.2009.244

  • 312 VOLUME 87 NUMBER 3 | MaRch 2010 | www.nature.com/cpt

    articles

    number of studies have been carried out on the blood profiles of patients with autoimmune diseases,17–19 including RA,6,7,16 in an attempt to discover diagnostic and prognostic biomarker signatures. As shown by Batliwalla et al., 21 of 52 genes overex- pressed in peripheral blood mononuclear cells of RA patients were monocyte specific.19 The expression levels of these mono- cyte-related genes showed a significant correlation with disease activity in patients who were studied just before starting therapy with a new disease-modifying antirheumatic drug, either MTX or an anti-TNF agent. Furthermore, monocyte activation seems to contribute to flare-ups of RA after pregnancy, as shown by comparative analysis of peripheral blood before and after birth and by comparison with the specific expression signatures of normal blood cells.20

    A number of studies on the monocyte lineage in RA have focused on the activation of these cells and the appearance of particular subsets in the periphery. The CD14+/CD16+ mono- cyte subset appears to be increased in RA patients.21 CD14+/ CD16+ cells have features that are similar to those of activated tissue macrophages; they express altered levels of chemokine receptors and adhesion molecules and have an enhanced capac- ity for transendothelial migration. Both subsets (CD16+ and CD16−) of monocytes are able to differentiate into dendritic cells in vitro. The binding of small immunoglobulin G–rheumatoid factor immune complexes to CD16+ monocytes could enhance TNF induction in tissue macrophages.

    In the present study, it appeared advantageous to investigate enriched monocytes, given their prominent involvement in the processes of RA pathology that are particularly responsive to anti-TNF.22 By taking this approach, we sought to focus on immune functions of this cell type instead of using whole blood samples and to identify single biomarkers that would be suffi- cient for the prediction of anti-TNF response in RA. In addition, to investigate drug specificity and the effect of comedication, we tested the informative value of the predictive biomarkers also in patients treated either with MTX alone or with MTX in combination with adalimumab.

    Results Gene-expression profiling and microarray analysis In an initial exploratory analysis using microarrays, monocytes from seven patients with RA (study DE011; adalimumab anti- TNF therapy without disease-modifying antirheumatic drug

    application) were examined before and during adalimumab monotherapy and compared with monocytes from seven age- matched normal donors.

    The group of seven patients constituted a representative RA cohort and demonstrated the expected correlations among clini- cal parameters before and after anti-TNF treatment (Table 1). This RA group included two nonresponders to adalimumab therapy (RA patients 4 and 6) as assessed by the American College of Rheumatology (ACR) improvement criteria (ACR score

  • CliniCal pharmaCology & TherapeuTiCs | VOLUME 87 NUMBER 3 | MaRch 2010 313

    articles

    protein; Supplementary Table S1 online; for disease activ- ity score (28 joints); see Supplementary Figure S1 online) were generally consistent with the status of nonresponders. Hierarchical clustering of clinical parameters further confirmed the identification of these two patients as nonresponders (see Supplementary Figure S2 online). The data from this group of

    seven “non-placebo-treated” patients were therefore screened for candidate biomarkers associated with a response to adali- mumab treatment.

    To focus on therapeutically relevant genes, we identified 51 genes by selecting for differential expression in ≥70% of the 49 pairwise comparisons between patients with RA (n = 7) and

    −2.0 2.0

    212671_s_at

    a

    203097_s_at 204160_s_at 221737_at 209710_at 205624_at 206207_at 211734_s_at 212314_at 207815_at 217753_s_at 219269_at 203231_s_at 219410_at 202110_at 200655_s_at 200087_s_at 202157_s_at 202233_s_at 209146_at 218545_at 200061_s_at 218728_s_at 201193_at 205042_at 214512_s_at 202922_at 209193_at 217722_s_at 201889_at 221011_s_at 213309_at 219905_at 208051_s_at 217970_s_at 221434_s_at 206790_s_at 212386_at 200642_at 215726_s_at 218190_s_at 213883_s_at 218345_at 220532_s_at 201690_s_at 204777_s_at 39248_at 212999_x_at 202378_s_at 214305_s_at 219862_s_at 212897_at 204750_s_at 214807_s_at 203300_x_at 210042_s_at 221042_s_at 205114_s_at 212335_at 200745_s_at 207008_at 211506_s_at 208982_at 211995_x_at 216016_at 208637_x_at 209009_at 210732_s_at 200772_x_at 202910_s_at 208918_s_at 209020_at 202664_at 210895_s_at 213506_at 208161_s_at 218486_at 202950_at 210184_at 214953_s_at 201693_s_at AFFX-M27830_5_at

    R A

    _4 R

    A _6

    R A

    _2 R

    A _3

    R A

    _5 R

    A _7

    R A

    _1

    1.1

    214953_s_at 207008_at 211506_s_at 208982_at 210184_at 218486_at 206207_at 201889_at 217722_s_at 219269_at 219410_at

    b −2.0 2.0

    R A

    _4 R

    A _6

    R A

    _2 R

    A _5

    R A

    _3 R

    A _7

    R A

    _1

    1.1

    210184_at 201889_at 219410_at

    c −2.0 2.0

    R A

    _4 R

    A _6

    R A