Differentially methylated regions in T cells identify kidney transplant patients at risk for de novo skin cancer

Abstract

Background: Cutaneous squamous cell carcinoma (cSCC) occurs 65-200 times more in immunosuppressed organ transplant patients than in the general population. T cells, which are targeted by the given immunosuppressive drugs, are involved in anti-tumor immune surveillance and are functionally regulated by DNA methylation. Prior to kidney transplantation, we aim to discover differentially methylated regions (DMRs) in T cells involved in de novo post-transplant cSCC development.

Methods: We matched 27 kidney transplant patients with a future de novo cSCC after transplantation to 27 kidney transplant patients without cSCC and studied genome-wide DNA methylation of T cells prior to transplantation. From 11 out of the 27 cSCC patients, the DNA methylation of T cells after transplantation was also examined to assess stability of the observed differences in DNA methylation. Raw methylation values obtained with the 450k array were confirmed with pyrosequencing.

Results: We found 16 DMRs between patients with a future cSCC and those who do not develop this complication after transplantation. The majority of the DMRs were located in regulatory genomic regions such as flanking bivalent transcription start sites and bivalent enhancer regions, and most of the DMRs contained CpG islands. Examples of genes annotated to the DMRs are ZNF577, coding for a zinc-finger protein, and FLOT1, coding for a protein involved in T cell migration. The longitudinal analysis revealed that DNA methylation of 9 DMRs changed significantly after transplantation. DNA methylation of 5 out of 16 DMRs was relatively stable, with a variation in beta-value lower than 0.05 for at least 50% of the CpG sites within that region.

Conclusions: This is the first study demonstrating that DNA methylation of T cells from patients with a future de novo post-transplant cSCC is different from patients without cSCC. These results were obtained before transplantation, a clinically relevant time point for cSCC risk assessment. Several DNA methylation profiles remained relatively stable after transplantation, concluding that these are minimally affected by the transplantation and possibly have a lasting effect on post-transplant cSCC development.

Keywords: Cutaneous squamous cell carcinoma; DNA methylation; Epigenetics; Non-melanoma skin cancer; Solid organ transplantation; T lymphocytes.

Fig. 1

The genomic characteristics of the CpG sites within each DMR. a CpG island content for all regions together and the individual DMRs separately, the array content is given as reference. The color represents the CpG island content of each CpG site within that region according to the legend below the graph. b Primary T cell-specific chromatin state according to the 15-state model of the ROADMAP epigenomics reference data [16] for all regions together and the individual DMRs separately, the array content is given as reference. The color represents the primary T cell-specific chromatin state of the CpG sites within that region according to the legend below the graph

Fig. 2

Stability of the 16 DMRs. a Mean difference in beta-value per region between pre-transplant and post-transplant samples. The difference is calculated per CpG site for each individual patient and is then averaged over all CpG sites per region for all 11 cSCC patients together. b Percentage of CpG sites that show a Δbeta-value of less than 0.05 presented per region. The numbers within each bar represent the number of stable CpG sites from the total sites within that region

Fig. 3

Mean difference in beta-value per patient between pre-transplant and post-transplant sample. The difference was calculated per CpG site for each individual patient and was then averaged over all CpG sites per patient

Fig. 4

Methylation values on the array and by pyrosequencing of six CpG sites within two DMRs. a DMR 2 (r = 0.95; p < 0.0001). b DMR 3 (r = 0.88; p < 0.0001). The CpG sites correspond to the CpG sites within the DMRs (Table 4)