Local skin gene expression reflects both local and systemic skin disease in patients with systemic sclerosis

Sir, The extent of skin disease in patients with SSc is typically measured using the modified Rodnan skin score (mRSS). Skin involvement at 17 sites is scored from normal to severely thickened (0–3) and the sum of all 17 sites is defined as the mRSS (0–51). It has been shown that an increase in mRSS (mRSS ≥20) correlates with internal organ complications and increased risk of mortality, and a decrease in mRSS correlates with improved survival [1, 2]. Skin is easily biopsied and provides insights into other sites of fibrosis and information about therapeutic response [3]. We have shown previously that the expression of certain biomarker genes in skin biopsies taken from the mid-forearm of SSc patients correlates highly with the mRSS [4, 5]. However, it remained unclear whether the success of such an approach represented a particularly accurate molecular quantification of local forearm skin disease or a manifestation of altered gene expression known to occur in all the skin of SSc patients [6].

RNA purified from dorsal mid-forearm skin biopsy specimens from 42 patients with SSc [7] and five healthy controls (HCs) was analysed using a custom nanostring (Nanostring nCounter) constructed to measure 17 genes selected from previous gene expression studies [4, 5, 8]. All skin samples and clinical data were collected under approval with written informed consent. The study was approved by the Institutional Review Board (H31479) at Boston University Medical Center. There were 18 housekeeping genes included for normalization purposes. Patient characteristics were as follows: average age, 51 years (s.d. 11), HCs 44 years (s.d. 15); female, 71% (HCs 80%); average mRSS, 22 (s.d. 12). The average disease duration was 28 months (s.d. 25), with 100% having diffuse cutaneous skin involvement as defined by skin involvement proximal to the forearms or legs and 81% classified as early dcSSc (≤36 months from the first non-Raynaud’s symptom). Gene expression was transformed (log2) and any difference between categories of forearm scores were assessed using a one-way analysis of variance, with a post-test for linear trend. The mRSS was treated as a continuous variable and correlated (Pearson’s) with gene expression.

All 17 putative biomarker genes were expressed significantly higher in SSc patients than HCs [false discovery rate (FDR) corrected P < 0.05], with the exception of WIF1, which was expressed at a significantly lower level in SSc patients (FDR corrected P < 0.05). Expression of genes was significantly different across groups based on forearm score (Fig. 1). TGF-β-associated genes (CTGF, SERPINE1, ADAM12, THBS1, COMP) were consistently different across groups (P < 0.0001) and correlated moderately to strongly with both the local score and the mRSS (0.80 ≥ r ≥ 0.40). THBS1 correlated strongly (r = 0.76, P < 0.0001), while COMP correlated moderately (r = 0.63, P < 0.0001) with the local skin score (Fig. 1A). These same genes also correlated with the mRSS, although not to the same extent as the forearm score (Fig. 1A). Conversely, CCL2 and IL13RA1 correlated more strongly with the mRSS (r = 0.65, P < 0.0001; Fig. 1B) than the forearm score (r = 0.47, P = 0.0005; difference in means P < 0.01). Macrophage-associated genes MRC1, MS4A4A, CD14, CD163 and SIGlLEC1, although elevated, correlated only weakly with the local skin score and mRSS [representative CD14 and CD163 r ≤ 0.36, P = 0.003 (shown in Fig. 1C), difference in means P < 0.05]. Several other genes examined correlated moderately [THY1 (r = 0.42), IGFBP3 (r = 0.50) and IFITM3 (r = 0.42)] or weakly [WIF1 (r = −0.33) and IFI44 (r = 0.2)] with the local skin score.

Fig. 1.

Forearm and mRSS scores correlate with gene expression

Genes were assessed for differences between local forearm scores (analysis of variance); significance is represented as *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001. Annotated on the graphs are R² values for the forearm skin score (post-test for linear trend) and mRSS (Pearson’s). Dotted lines indicate mean and s.d. of healthy controls for each gene. Genes shown: thrombospondin-1 (THBS1), cartilage oligomeric protein (COMP), chemokine ligand 2 (CCL2), IL-13 receptor alpha 1 (IL13RA1), cluster of differentiation 14 (CD14) and cluster of differentiation 163 (CD163). mRSS: modified Rodnan skin score.

The different relationships between local gene expression and local vs systemic assessment of the skin (mRSS) suggest a hierarchy of molecular events. TGF-β is highly implicated directly in skin fibrosis, and expression of TGF-β-regulated genes (THBS1, COMP) correlated most highly with the local skin score. Thus these genes may represent molecular events closest to local clinical manifestations. This is unsurprising, as in previous studies COMP and THBS1 expression correlated more strongly with mRSS in lesional vs non-lesional skin [4]. The expression of other genes (CCL2, IL13RA1) that correlate more closely with the mRSS than the local forearm skin score may reflect more generalized immune dysfunction that is driving systemic pathogenesis. IL13 is an immune system–derived mediator of fibrosis and regulates CCL2 [5]. Using CCL2 as a marker of early disease is further supported here by the observation that CCL2 was the only gene in which the majority of patients with a forearm skin score of zero had gene expression >1 s.d. of HCs. Altered gene expression that does not correlate with the skin score (CD14, CD163) may represent molecular events earlier in the pathogenic cascade. This is consistent with earlier work suggesting that CD14 is a prognostic biomarker for SSc [8]. Alternatively those genes that do not correlate with clinical scores might represent epiphenomena without direct roles in pathogenesis. Since the forearm skin score correlated strongly with the mRSS in this cohort (r = 0.86), up-regulated expression of some genes in SSc skin, such as THBS1 and COMP, could potentially be used as biomarkers for clinical trials of drugs delivered by local topical application.

Rheumatology key message

• Local forearm and systemic disease in patients with SSc is reflected in gene expression.

Funding: This work was supported by a National Institutes of Health National Institute of Arthritis and Musculoskeletal and Skin Diseases Center of Research Translation Grant (1P50AR060780) and by the Boston University Clinical and Translational Science Institute, National Institutes of Health National Center for Advancing Translational Sciences Clinical and Translational Science Award (UL1-TR000157).

Disclosure statement: The authors have declared no conflicts of interest.