Background Cyclophosphamide (CYC) and mycophenolate (MMF) are presently the mainstay of treatment for systemic sclerosis-related interstitial lung disease (SSc-ILD). Despite the widespread use of these two immunosuppressant agents for SSc-ILD, CYC and MMF have markedly different purported mechanisms of action. To our knowledge, no studies have directly linked the gene expression and gene network overlap profiles to the adverse event profiles of these two drugs.

Objectives To compare the gene network overlap profiles, gene expression profiles and adverse event profiles of CYC and MMF to gain insight into how these two immunosuppressants are effective in SSc-ILD and to better understand why the effects of CYC and MMF may vary in different patients with SSc-ILD.

Methods Gene network interactions for CYC, MMF and other drugs were estimated by examining drug-gene associations (via literature text mining using drug-gene co-citations for >8000 compounds from 24 million abstracts in the 2014 PubMed database). Reported adverse event data for CYC, MMF and other marketed drugs (including drug combinations) were obtained from the FDA Adverse Event Reporting System (FAERS) Database. Adverse event and gene network data were converted to drug-effect and drug-gene spectra. Profile comparisons and hierarchical clustering (Spotfire) of these spectra were employed to assess medicine adverse event and gene network overlap similarity. Rat gene expression profiles (bone marrow) for CYC and MMF and human gene expression data (skin) for CYC in SSc patients were obtained from GEO (NCBI). Functional analysis was performed using DAVID (NIH).

Results The literature text mining search yielded 29,843 CYC gene-related abstracts and 14,953 MMF gene-related abstracts. Examination of the gene network overlap data and the drug adverse event data for CYC and MMF revealed that that these two drugs have strikingly different profiles. The gene network and effect spectra of CYC were most similar to those of cisplatin, carboplatin, epirubicin, and melphalan. The gene network and effect spectra of MMF were most similar to tacrolimus and azathioprine (Figure 1). Rat gene expression profiles (bone marrow) of CYC and MMF were also examined. Functional enrichment analysis of the expression data showed that both compounds affect the regulation of apoptosis and cell death, as well as cell cycle and inflammatory response. Interestingly, MMF treatment also affects the expression of genes involved in lung development and regulation of blood vessel size.

• Download figure
• Open in new tab
• Download powerpoint

Conclusions While CYC and MMF remain the most commonly prescribed drugs for SSc-ILD, these agents have significantly different gene network overlap, adverse event and gene expression profiles. Moreover, while the gene network overlap of CYC is most similar to other alkylating/alkylating-like agents, the gene network overlap of MMF is most similar to immunosupppressants which block T cell activation by calcineurin pathways. These findings have important implications for developing cause-effect models for these two drugs, and may potentially have utility for tailoring treatment for SSc-ILD based on patient-related factors. Understanding the cause-effect models for CYC and MMF is furthermore important for the design of new agents to treat this devastating clinical dimension of SSc.

Disclosure of Interest None declared