Clonal Evolution of Lymphoid Malignancies Following the Mutagenic Impact of Radiotherapy

Introduction

Ionizing radiotherapy (RT) is a mainstay in the local control of lymphoid malignancies. It introduces DNA double-strand breaks (DSB) that result in cell death unless repaired prior to cell division. Resulting mutations are measurable as mutational signatures with whole genome sequencing (WGS). Importantly, detection of a therapy-induced signature in bulk WGS is predicated by the expansion of a single cell bearing therapy-associated mutations that sweeps to clonal dominance. This concept can be leveraged to track cancer evolution relative to treatment (Diamond et al. Blood 2023). The influence of RT, which has been linked with the ID8 indel signature, on patterns of relapse remains understudied, in part, due to the difficulties in isolating RT-induced lesions from those induced by genotoxic chemotherapies, which are often co-administered.

Methods

We reworked mmsig for indel signatures (github.com/UM-Myeloma-Genomics) and used 80x WGS to characterize the indels of 514 total WGS samples including 56 relapsed/refractory (RR) and 100 newly diagnosed (ND) large B-cell lymphoma (LBCL) and 266 ND and 98 RR multiple myeloma (MM; 327 patients). Finally, we validated and refined the indel landscape of chemoRT by exposing TP53-mutant lymphoma cells (DHL-4) to cisplatin and RT, then performing 30x bulk WGS of 15 single-cell-derived colonies.

Results

Of 514 total WGS, 29 and 16 samples were from MM and LBCL patients exposed to prior RT. Of these 45 cases, 41 (91%) also had prior mutagenic chemo (platinum/melphalan). Indel burden was increased in RR cases compared to ND (LBCL: median 467 vs 1106, p<0.001; MM: 527 vs 591, p<0.001) but there was no difference in indel counts between chemoRT- and chemo-treated cases (p = 0.51). We performed indel and single base substitution (SBS) signatures extraction, decomposition, and fitting for all samples. SBS signatures included platinum- (SBS31, SBS35, E_SBS37) and melphalan-induced (SBS99) mutagenesis in 36/56 (64%) and 71/107 (66%) of exposed samples, respectively. We deconvoluted 7 indel signatures including ID8. ID8 was seen in 18 RR cases (of 154 total, 11.6%) and not in any at baseline. While presence of ID8 was strongly correlated with prior RT (p<0.001), it was also observed in 10 unexposed samples. All of these RT-unexposed cases had evidence of platinum/melphalan mutagenesis.

As chemo-induced ID8 has not been reported, we performed a single cell expansion experiment wherein DHL-4 cells were exposed daily to platinum or RT. Single cells from days corresponding to a sub-lethal and a half-lethal dose were expanded, and 3 resulting colonies from each time were subjected to 30x WGS (12 colonies). An unexposed expansion was performed as control (3 colonies). RT and platinum expansions had a higher indel burden (mean, 4247) compared to control (mean, 3065; p = 0.004). Where all RT expansions had evidence of ID8 mutagenesis, only sub-lethal-dose platinum harbored ID8, implying a dose threshold must be met, but that DSB from chemotherapy can indeed cause ID8. Indel profiles of RT and sub-lethal-dose platinum colonies had cosine similarity > 0.95.

To investigate the impact of RT on tumor evolution, we focused on 2 ID8+ cases with RT exposure, but no chemo mutagenesis. One LBCL case had an irradiated neck node prior to CD19 CART. At relapse, 9 months later, an inguinal biopsy was ID8+. One MM case had RT of a clavicular plasmacytoma and a relapse 7 years later with ID8+ iliac crest marrow. Given the requisite expansion of a single cell bearing a unique catalogue of RT-induced indels required for ID8 detection in bulk WGS, this constitutes evidence that, in both cases, a single cell that survived RT seeded systemic relapse (single cell accelerated model; Landau et al. Nat Comm 2020). Finally, ID8-associated deletion motifs were identified in driver genes across these samples including in SOCS1, ACTB, TNFAIP3, and RB1, among others, implicating chemoRT in the potential development of refractoriness.

Discussion

ID8 is strongly associated with RT in LBCL and MM. Yet, we show, for the first time, that it can be induced via DSB from chemo. Dissection of therapy-related mutational signatures can dissect the patterns of relapse after each clinical intervention. Here, we see that even a single resistant tumor cell from an RT-exposed lesion can seed systemic relapse. further underscoring the need to eradicate disease in entirety to avoid propagation of resistant subclones.