Chronic Graft-Versus-Host Disease Immunoprofiling Reveals T Cell Clonal Dynamics That Correlate with Disease Activity: A Novel Molecular "Biomarker"?

Chronic graft-versus-host disease (cGVHD) is the main cause of morbidity and transplant-related mortality following allogeneic hematopoietic stem cell transplantation (alloHSCT), however effective treatment options are limited. Lack of objective surrogates (biomarkers) for treatment response has hindered progress in this respect. T cells are considered the major effectors of cGVHD, yet the respective repertoires are insufficiently charted. Here, we investigated the dynamic architecture of T-cell repertoires in cGVHD by exploiting next-generation sequencing (NGS), aiming to uncover immunogenetic signatures linked with cGHVD occurrence and response to treatment.

We analyzed 53 blood samples from 15 patients with hematological malignancies who underwent alloHSCT, with an intended bias towards patients who developed cGVHD (n=12). The remaining 3 patients had no cGVHD (control group). Patients with cGVHD were analyzed at cGVHD onset and/or prior to a new line of treatment (n=17 samples) as well as at clinically relevant timepoints following treatment: (i) partial response (PR, n=8), (ii) complete response (CR, n=3), (iii) stable disease (SD, n=18), (iv) progression (PD, n=1). Treatment modalities involved corticosteroids, mycophenolate myfetil, extracorporeal phototherapy, ruxolitinib and ibrutinib. Patients with no cGVHD were analyzed at 3 months (+3mo) and 6 months (+6mo) post-alloHSCT (n=6 samples). Starting material was PB mononuclear cells. TRBV-TRBD-TRBJ gene rearrangements were RT-PCR amplified and subjected to paired-end NGS and detailed bioinformatics analysis. Only productive TRBV-TRBD-TRBJ rearrangements were evaluated (n= 13,059,730, 89.9% of filtered-in sequences, median 219,063/sample). For repertoire characterization, clonotypes (i.e., TRB rearrangements with identical TRBV gene usage and amino acid complementarity-determining region 3 sequence) were considered (median 9,725 distinct clonotypes/sample). The 10 most frequent clonotypes/sample were defined as "major". To purge T cell clones expanded secondary to viruses common in the alloHSCT setting, we compared all major clonotypes against the GenBank TRB sequence database (n=18,402), as well as an extensive TR repertoire dataset from a previous study by our group, profiling tri-virus-specific (CMV, EBV, BK) T cell products generated from immunocompetent donors (n=947,298).

Overall, we identified 289 unique major clonotypes; 38 were excluded due to match within the tri-VST database (no match within GenBank). All cases with cGVHD displayed significant clonal T cell expansions both pre- and post-treatment (overall median cumulative frequency of the 6 most expanded T cell clonotypes/sample 32.4%). However, clonality tended to decrease in PR samples compared to pre-treatment (19.8% vs 33.9%, respectively), although not reaching statistical significance possibly due to small sample size (p=0.06). Patients with no cGVHD, on the other hand, consistently displayed a clonality decrease overtime (31.8% at +3mo vs 18.4% at +6mo, p=0.02). Importantly, patients with no GVHD displayed TR repertoire reconstitution with few major T cell clones of the +3mo timepoint persisting at +6mo (median 20%). In clear contrast, cGVHD T cell repertoires were dominated by clones which persisted overtime (median 60%). In fact, repertoire persistence was most evident in SD (median 66.7%) and significantly lower in PR and CR (33.3% and 10.0%, respectively, p<0.05), suggesting that the persisting T cell clones are implicated in cGVHD pathogenesis. Notably, repertoire comparisons across patients in our cohort revealed 6 "public" clonotypes [5 clonotypes shared among a single pair of patients (Pt1 and Pt2) and 1 clonotype shared among the same pair plus an additional patient, Pt3], suggesting the existence of a common antigenic trigger.

In conclusion, NGS immunoprofiling in cGVHD reveals expanded T cell clones with clonal dynamics that correlate with clinical response, indicating a causal relationship to cGVHD pathogenesis. Identification and longitudinal tracking of T cell clones associated with cGHVD could serve as a molecular surrogate marker for disease activity, with evident benefits for cGVHD monitoring and evaluation of response to various treatments.