AFFILIATIONS
1 Broad Institute of MIT & Harvard, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
2 Broad Institute of MIT & Harvard, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States
3 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
4 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
5 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
6 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
7 Broad Institute of MIT & Harvard, Cambridge, Massachusetts, United States
8 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
9 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
10 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
11 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
12 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
13 Broad Institute of MIT & Harvard, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States
14 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
15 Harvard Medical School, Boston, Massachusetts, United States
16 Broad Institute of MIT & Harvard, Cambridge, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States, Massachusetts General Hospital, Boston, Massachusetts, United States
17 Brigham & Women's Hospital, Boston, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States
18 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
19 Broad Institute of MIT & Harvard, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
ABSTRACT
Despite major advances in our understanding of the immunobiology of graft-versus-host disease (GvHD), the molecular determinants of allorecognition remain poorly defined. Rare clinical presentations may therefore provide unique insights into disease pathogenesis. Linear cutaneous GvHD along Blaschko lines is an uncommon manifestation whose striking distribution has long been hypothesized to reflect cutaneous mosaicism but this has never been tested at the molecular level.
A 60-year old man with relapsed Hodgkin lymphoma after multiple prior lines of therapy, including ABVD, autologous HCT, brentuximab vedotin, and nivolumab, underwent reduced-intensity conditioning allogeneic HCT from a 10/10 HLA-matched unrelated donor with a nonpermissive HLA-DP mismatch. Approximately one year later, he developed cutaneous GvHD along Blaschko lines involving the trunk and upper extremities. To identify the antigenic determinants underlying this distinctive pattern of GvHD, paired affected and unaffected skin biopsies were obtained and enzymatically separated into epidermis and dermis prior to whole exome and RNA sequencing, along with matched donor and recipient germline DNA as references.
Because post-transplant skin contains both donor and recipient cells, accurate assignment of genomic and transcriptomic features required deconvolution of tissue chimerism. We therefore applied two orthogonal approaches, mitochondrial DNA heteroplasmy and CIBERSORTx deconvolution of bulk RNA-seq. Both confirmed that the specimens were predominantly composed of recipient-derived cells, with donor contribution enriched in affected relative to unaffected samples (P=0.044). Immune cell deconvolution showed increased CD8-associated transcripts in the affected epidermis, consistent with localized cytotoxic T-cell infiltration.
We systematically evaluated 3 potential sources of antigens associated with the linear pattern: (i) viral antigens associated with subclinical localized skin infection, (ii) minor histocompatibility antigens (mHAgs) encoded by genes selectively expressed along the Blaschko lines, and (iii) neoantigens generated by somatic mosaicism.
We first evaluated viral reactivation using PathSeq and VirPy but detected no transcripts from viruses with skin tropism, including HSV1/2 and VZV. We then sought to identify candidate mHAgs by integrating published RNA-seq datasets from chronic GvHD, fetal skin, and wound healing to define transcriptional programs potentially associated with lesional skin. Sequential filtering for expression restricted to affected skin, recipient-specific nonsynonymous polymorphisms, and HLA binding prediction prioritized 15 candidate mHAgs across 4 genes (SAA1, OLFML1, OLFML2B and GBP4) for functional testing. Finally, we investigated somatic mosaicism at both the genetic and epigenetic level. While no copy number variants were detected in any sample, WES identified a 26 bp deletion in TJAP1 present only in affected epidermis sample. The mutation generated a frameshift resulting in a premature stop codon, yielding a truncated 113 amino acid protein with 7 predicted neoepitopes downstream of the frameshift across 4 HLA-B/C alleles (HLA-B*35:01, -B*41:02, -C*04:01, and -C*17:03).
To investigate epigenetic causes of mosaicism, we analyzed transposable element (TE) expression, as retrotransposons have been implicated in developmental mosaicism underlying Blaschko-like skin patterns across mammals. Using the tools Salmon and SQuIRE, we identified 19 transposable elements selectively enriched in affected samples. Among them, a SINE TE present at the junction between exon 4 and subsequent intron of SMARCAD1 emerged as the strongest candidate. The resulting neojunction encoded a predicted HLA-B*41:01-restricted epitope. Functional testing of all predicted mHAgs, TJAP1-, and TE-derived peptides by IFNγ ex vivo ELISPOT using patient PBMCs showed no reactivity to candidate mHAgs, whereas robust T-cell responses were observed against both the TJAP1-derived peptide pool (117 SFC/106 PBMCs) and, even more prominently, against the TE-derived epitope (148 SFC/106 PBMCs).
Overall, these findings provide the first molecular evidence that somatic mosaicism can drive tissue-specific allorecognition in chronic GvHD. Developmental mosaicism may thus represent a previously unrecognized source of clinically relevant antigenic determinants, with implications for tissue-specific immune recognition across immune-mediated diseases.