Abstract
Introduction: Recent studies linking cancer genomics and immunity have reinforced the concepts that some mutations trigger T cell effector responses and that the likelihood of an immunogenic mutation increases with increasing mutation load. Importantly, these data highlight the potential utility of such markers in identifying patient subsets likely to respond to cancer immunotherapies. This study investigated the clinical impact of mutation load and its association with T cell gene expression in newly diagnosed patients with follicular lymphoma (FL).
Methods: We used clinical and genomic data from FL patients (n = 249; 216 with follow-up information) with evaluable pre-treatment tumor tissue who were treated in a randomized study of rituximab maintenance vs observation (PRIMA; ClinicalTrials.gov ID: NCT00140582). We estimated mutation load per megabase (Mb) as a proxy for neoantigen formation using FoundationOne Heme (Foundation Medicine, Inc). We quantified expression of T cytotoxic effector genes (GZMA, GZMB, PRF1, IFNG, EOMES, CD8A) as a surrogate for pre-existing immunity (and the inflammatory state of the tumor) using TruSeq (Illumina, Inc) RNA seq (n = 142). We used Cox regression to examine associations between these markers and progression-free survival (PFS), adjusting for the FL International Prognostic Index, age, sex, treatment arm and response to induction therapy. Pvalues were calculated for exploratory purposes.
Results: The mutation load estimate among newly diagnosed patients with FL was highly variable (range, 0-33 mutations/Mb; Q1: 4.2; median: 6.6; Q3: 10.0). Patients with > 15 mutations/Mb (n = 19) were considered to have a high probability of neoantigen formation, and the remaining patients were stratified into mutation-low (< 6.6 mutations/Mb; n = 112) or mutation-mid (≥ 6.6 mutations/Mb and ≤ 15 mutations/Mb; n = 85) groups. The 3-year PFS in patients with high mutation load was 83% compared with 66% for mid-mutation load and 68% for low-mutation load groups, but mutation load was not independently prognostic in either the rituximab (P = .13) or observation (P = .66) arms. Of note, 92% of FL patients with high mutation load (n = 12/13) also had high T-effector gene expression compared with 49% of those with midlevel (n = 24/49) and 44% of those with low mutation load (n = 35/80) (P = .001). Mutation load was also associated with benefit from rituximab maintenance: FL patients with low mutation load experienced a significant benefit from rituximab maintenance (HR, 0.29 [95% CI, 0.15-0.54]; P < .001), whereas no statistically significant benefit was seen among FL patients with medium (HR, 0.81 [95% CI, 0.43-1.5]; P = .51) or high mutation load (HR, 0.29 [95% CI, 0.026-3.3]; P = .32). Importantly, the T/NK gene signature was prognostic as a continuous predictor (P = .008) and clearly separated 2 large groups of FL patients into an "inflamed" subset (T-effector signature high; n = 74) and an "uninflamed" subset (T-effector signature low; n = 75), with longer PFS seen in the "inflamed" FL subset (PFS HR, 0.39 [95% CI, 0.21-0.70]; P = .002). T-effector gene expression may be particularly useful for identifying the immunologically primed FL subset among patients with low/mid mutation load: there was a trend in 3-year PFS in 84.4% vs 56.6% for T-effector-high vs T-effector-low among low-mutation load patients (P = .002) and 76.2% vs 58.3% of mid-mutation load patients (P = .17), respectively. The subset of inflamed (T-effector signature high) FL tumors also demonstrated high expression of IDO1, which similarly correlated with longer PFS (HR, 0.25 [95% CI, 0.14-0.45]; P < .001), and a strong correlation was observed between IDO1 and IFNG expression (R2 = 0.61; P< .001). This is consistent with an interplay of pro- and anti-inflammatory immunity, wherein pro-inflammatory IFNγ drives the clinical outcome.
Conclusions: Collectively, our results suggest that mutation load and T-effector gene expression may help identify immunologically distinct lymphoma subsets appropriate for modern immunotherapies.
Disclosures
Bolen: Genentech, Inc.: Employment. McCord:Genentech, Inc.: Employment. Frampton:Foundation Medicine: Employment, Equity Ownership. Bourgon:Genentech, Inc.: Employment; F Hoffman-La Roche: Other: Shareholder. Punnoose:Genetech, Inc.: Employment. Szafer-Glusman:Genentech, Inc.: Employment. Xerri:Novartis: Honoraria. Salles:Gilead: Honoraria, Research Funding; Janssen: Consultancy, Honoraria; Celgene: Consultancy, Honoraria; Mundipharma: Honoraria; Roche/Genentech: Consultancy, Honoraria, Research Funding; Amgen: Consultancy, Honoraria; Novartis: Consultancy, Honoraria. Venstrom:Genentech: Employment.
Uncover Cancer Genomics by Jointly Analysing Aneuploidy and Gene Expression