Acquired Immunodeficiency Landscapes Via Integrated Noninvasive Viral & Immune Profiling in HIV-Associated Lymphoproliferative Disease

Background: People living with human immunodeficiency virus (PLWH) have a higher risk of lymphoproliferative disease and latent virus reactivation due to immune dysregulation including dysfunctional surveillance of Epstein-Barr virus (EBV). The EBV latency patterns differ in variant lymphoma histologies in PLWH, and lytic gene expression is associated with EBV replication and lymphoma development. Despite significant progress in HIV and lymphoma treatments, patients with HIV-associated lymphomas continue to have inferior overall survival compared to immunocompetent individuals. Methods for tracking viral infection and the immune state of PLWH are lacking, and the genomic landscape in HIV infection is less well characterized. Comprehensive viral and host genomic profiling using cell-free DNA (cfDNA) and tumor tissues allows the evaluation of viral abundance, viral and host mutations, and immune response in PLWH.

Methods: We studied 246 serial plasma and 20 tumor samples from 51 HIV-associated diffuse large B cell lymphoma (HIV/DLBCL) and 10 HIV-associated Burkitt lymphoma (HIV/BL) patients. Patients were treated with SC-EPOCH-RR (NCT000019253). Tumor EBV status was determined by Epstein-Barr-encoded small RNA (EBER) ISH. Tumor, cfDNA, and PBMC were profiled by VirCAPP-Seq targeting 180 viral species (Garofalo, Blood 2019), CAPP-Seq targeting 186 B-cell lymphoma genes (Newman, Nature Medicine 2014; Alig, Nature 2024), and TCR repertoires by SABER (Shukla, Blood 2020).

Results: Clinical characteristics: 73% of HIV/DLBCL patients were stage IV (Ann Arbor), 49% were germinal center B cell subtype, 24% were tumor EBER+, and 45% were on antiretroviral therapy (ART) pretreatment (preTX).

Viral profiling: In HIV/DLBCL samples, all preTX plasma had EBV load (cfEBV) above healthy controls (mean=0.1, SD=0.3 copies/ml) and 47% had cfEBV > 100copies/mL (cfEBV+) by VirCAPP-Seq. Tumor EBER+ cases had ~100x higher cfEBV than EBER- (p<0.001). The cfEBV was similar between HIV/DLBCL and HIV/BL. Patients with preTX ART had lower HIV and Anellovirus loads (all p<0.05). EBV and Anellovirus loads decreased on therapy (post-C2, all p<0.01) but slightly increased at the last follow-up compared to post-C2. We observed 3 patterns of EBV mutation in preTX plasma: 1) Lytic mutational subgroup: lytic phase gene mutations in all EBER+ and some EBER- cases. 2) Latent phase subgroup: fewer altered lytic genes in some EBER- cases. 3) Undetected mutation or low EBV genome depth (<10x) subgroup: all are EBER- cases. EBER+ cases had more coding variants in EBNA-1, EBNA-3, and select lytic genes than EBER-.

Immune & mutational profiling: Baseline CD4 count was higher in ART-treated patients (mean 306 vs 145 cells/l, p<0.05), and slightly but not significantly higher in EBER- cases. TCR clonotypes and Gini coefficient were higher in patients with baseline CD4 >100 cells/l (all p<0.05), indicating T cell clonal expansion. TCR Gini coefficient slightly increased and Chao1 index slightly decreased at the end of C2, suggesting lower TCR diversity after chemotherapy. Recurrent mutations in HIV/DLBCL plasma samples were TP53, CREBBP, KMT2D, STAT3, HIST1H1E, and SGK1. Mutations of MYC, TP53, and CCND3 were enriched in HIV/BL plasma samples.

Clinical response: 82% of HIV/DLBCL achieved interim complete response/unconfirmed complete response (CR/CRu) after 2 or 3 cycles, and 18% had progression or death (PD) < 1 year. All HIV/Burkitt patients achieved CR/CRu after 3 or 4 cycles. Lymphoma and cfEBV mean allele frequency decreased (all p<0.01) with fewer detected mutations post-C2. PreTX cfEBV were ~100x higher in HIV/DLBCL patients experiencing PD (p<0.01), with 90% in the lytic mutational subgroup. Both EBER+ and cfEBV+ were associated with inferior PFS (all p<0.05). Patients with baseline CD4 > 100 cells/l had superior PFS (p<0.05). Patients with a lower preTX TCR diversity had a higher risk of PD.

Conclusions: Integrated non-invasive profiling of a large cohort of PLWH revealed distinct virome and immunogenomic features across different lymphoma histologies and clinical responses. Elevated cfEBV, tumor EBER+, EBV lytic gene mutations, lower CD4 count and lower preTX TCR diversity indicate worse clinical response. These findings may help personalize therapy for improved long-term outcomes, and our approach provides potential monitoring for global immune states in immunosuppressed individuals.