More than 40% of pts DLBCL will not respond to first-line immunochemotherapy. The therapeutic response of DLBCL is rooted in its heterogeneous genetic landscape and its lymphoma microenvironment (LME). In this regard, the roles of cancer-associated fibroblasts (CAF) and tumor-associated macrophages (TAMs) in lymphoma progression and therapy response have been started to be elucidated (Kotlov ... Cerchietti. Cancer Discovery 2021). We analyzed (by transcriptomics and multiplex imaging) the LME cellular composition and heterogeneity in paired samples from diagnostic and relapsed (RR) DLBCLs (n = 10) and early and late murine lymphoma progression models (A20, Myc + Bcl2, Ezh2 mutant + Bcl6, and Setd2 mutant + Bcl2). We found that, in both human and mouse DLBCLs, the heterogeneity of the LME decreases as a function of lymphoma progression. RR-DLBCLs are characterized by a significant decrease in tumor CD3+ T-cell infiltration, lower T-cell diversity, and expansion of a subpopulation of CAFs upregulating the fibroblast growth factor receptor FGFR1 (CAFFGFR1). Thus, we investigated the role of CAFFGFR1 in pre-clinical models of RR-DLBCL biology and therapy. We first proved that lymphoma cells (murine and human, primary and relapsed, ABC- and GCB-DLBCLs) do not express FGFR1 and do not respond to its pharmacological inhibition or genetic knock-down in cell cultures. We then tested the anti-lymphoma effect of the novel selective oral FGFR1 inhibitor SSR128129E (FGFR1i) in two engrafted syngeneic mouse models DLBCL (A20 and Setd2+/-/Bcl2OE) implanted subcutaneously and intrasplenically, respectively, in immunocompetent mice. FGFR1i (oral, 10 mg/kg) significantly decreased tumor burden in both models (p<0.05). Analysis of the LME at the end of the treatment showed that the LME of FGFR1i treated mice had significantly increased TAMs (p<0.005). There were no differences in the proportions of other LME cell populations, including CAFs. To further investigate the role of TAMs concerning FGFR1i in RR-DLBCL harboring CAFFGFR1-rich LMEs, we established two RR-DLBCL patient-derived xenografts (PDX) models. Treatment of RR-DLBCL PDX-1 and PDX-2 with CHOP confirmed the chemorefractory nature of the models; however, both models were highly responsive to oral 10 mg/kg FGFR1i (p<0.05, both models vs. vehicle). To understand how the FGFR1i affected distinct CAF and TAM subpopulations, we conducted single-cell RNA-sequencing in LME from these PDXs. We found that FGFR1i-treated mice had a significant increase in inflammatory-like CAFs characterized by the expression of several monocyte/macrophage chemotactic factors, including CCL2, CCL7, CXCL3, and CXCL12, which agrees with the increase in the LM-TAM in the syngeneic models. In FGFR1i mice, we found a sub-set of TAMs that expressed high levels of classical and alternative complement receptors critical for phagocytosis, suggesting that this process can be relevant for the anti-lymphoma effect. To test this notion, we assessed the effect of FGFR1i in a murine lymphoma model depleted of macrophages using liposomal clodronate. We found a significant decrease in the anti-lymphoma effect of FGFR1i when TAMs were depleted (p<0.05). This suggests that phagocytic TAMs are important for the response to FGFR1i, a concept we are testing in additional DLBCL models. In conclusion, our results suggest that CAFFGFR1s could constitute a novel LME-resident therapeutic vulnerability in RR-DLBCL and that FGFR1i may require phagocytic TAMs for its anti-lymphoma effect. Given the FDA approval of several FGFR1/2 inhibitors for solid tumors, our results bare translational potential to benefit RR-DLBCL patients in the near term.

No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution