Delivering a STINGing Blow to Small Cell Lung Cancer via Synergistic Inhibition of DNA-Damage Response and Immune-Checkpoint Pathways

Immunotherapies targeting the inhibitory immune checkpoints CTLA4 and PD-1/PD-L1 have rapidly swept to the forefront of clinical practice for several solid tumors since the first FDA approval in 2011. Early results have been especially promising in tumor types associated with high somatic mutation burden, such as melanoma and non–small cell lung cancer (NSCLC), as somatic mutations are thought to lead to immunogenic neoantigens that can be recognized and targeted by the adaptive immune system. Small cell lung cancer (SCLC) is a highly aggressive neuroendocrine neoplasm with high somatic mutation burden linked to heavy tobacco exposure (1). SCLC has been treated similarly for nearly four decades, is associated with very poor long-term survival, and is classically associated with immune-mediated paraneoplastic syndromes suggesting potential for a primed antitumor immune response (1). Immune-checkpoint blockade (ICB) was therefore eagerly investigated as a promising new paradigm for patients with advanced SCLC, and, in August 2018, nivolumab (an anti–PD-1 agent) became the first drug approved by the FDA for SCLC since topotecan 11 years prior. However, results from the first wave of trials were generally less dramatic than in other malignancies with similar mutation burden, with rare durable responses seen in SCLC. In a recent landmark study, the addition of atezolizumab, a PD-L1 inhibitor, demonstrated the first survival improvement over standard-of-care platinum doublet in decades. However, the magnitude of improvement in overall survival was 2 months, durable responses remained very rare, and the overall response rate was not improved (2). It is therefore likely that SCLC harbors robust immunosuppressive mechanisms that overcome high tumor mutation burden and hinder responses to ICB.

Dysregulation of genomic integrity and the cell cycle via biallelic inactivation of TP53 and RB1 is a near-ubiquitous feature of SCLC, likely contributing to a high response rate to DNA-damaging platinum and topoisomerase inhibitor–based chemotherapy in first line and leading to the hypothesis that SCLC cells may be especially dependent on the remaining mechanisms of DNA-damage response (DDR) and cell-cycle arrest. Indeed, SCLC cells overexpress DDR pathways, and compounds targeting these pathways, either alone or in combination with chemotherapy, are actively being investigated in early-phase clinical trials (3). Among others, inhibitors of CHK1, which is an especially critical mediator of cell-cycle arrest following DNA damage in cells with aberrant TP53, and the PARP family of proteins, which regulate multiple avenues of DNA repair following DNA damage, are a focus of ongoing studies. However, whether DDR inhibition (DDRi) might potentiate ICB in SCLC had not been explored. In this issue, Sen and colleagues leverage immunocompetent murine models of SCLC to identify a dramatic synergistic effect between DDRi and ICB treatment and provide strong evidence for an underlying mechanism involving innate immune signaling via a well-described cytoplasmic DNA-sensing innate immune pathway (Fig. 1).

The authors found that treatment of human SCLC cell lines with either CHK1 or PARP inhibitors led to appreciable upregulation of the immunosuppressive signal PD-L1, a finding that was recapitulated with genetic knockdown of CHK1 or PARP (4). To study the immune response to CHK1 and to PARP inhibition (PARPi), Sen and colleagues used genetically engineered mouse models of SCLC generated via deletion of the tumor suppressor genes Rb1 (RB), Trp53 (p53), and Rbl2 (p130) in lung epithelium (RPP model). They compared the effect of CHK1 inhibition (CHK1i) on the growth of RPP cells allografted into the flank of mice with a permissive immunocompetent (B6/129 F1 hybrid offspring) versus an immunodeficient (nude) genetic background. The growth-delaying effect of prexasertib, a CHK1 inhibitor, was more prominent in the immunocompetent system, suggesting a role for adaptive immunity in potentiating the effect of CHK1i. This effect was reproduced in primary tumors in the autochthonous system, where CHK1i was also associated with increased tumor-infiltrating CD8⁺ T cells. Motivated by these findings, Sen and colleagues asked whether PD-L1 inhibition (i.e., ICB) could synergize with CHK1i or PARPi in their syngeneic flank tumor system. Notably, ICB and PARPi had minimal single-agent activity and CHK1i had only modest activity in that system. However, when ICB was combined with CHK1i, deep and durable responses were observed. ICB and PARPi also displayed considerable synergism in reducing tumor volume at a single time point in the autochthonous system. Studies of the post-treatment tumor microenvironment established that the combination of DDRi and ICB led to remodeling of the adaptive immune cell compartment, with generally higher relative frequencies of CD8⁺ T cells and memory/effector T cells and decreased frequencies of naïve, exhausted, and regulatory T cells. Antibody depletion experiments showed that CD8⁺ T cells were required for the synergizing effect between DDRi and ICB. A similar synergy between ICB and DDRi was observed in other mouse SCLC model–derived cell lines, demonstrating that these effects were not cell line–specific.

Finally, they sought to delineate a mechanism by which combined DDRi and ICB led to rapid and sustained responses. Sen and colleagues showed that DDRi led to increased cytoplasmic DNA resulting in the engagement of the cGAS/STING pathway, which senses cytoplasmic foreign pathogenic or self DNA (Fig. 1). Activation of STING results in phosphorylation of IRF3 and transcription of IRF3 targets such as IFNβ and key proinflammatory cytokines. cGAS/STING pathway components were shown to be required for both IFNβ and PD-L1 upregulation upon DDRi. Most importantly, when cGAS or STING was knocked down in RPP cells prior to implantation, synergism between DDRi and ICB was abrogated, strongly supporting a central role for this pathway in the induction of an effective antitumor adaptive immune response upon treatment with DDRi and ICB.

The study by Sen and colleagues contributes substantially to efforts to advance immunotherapy in SCLC. Their results suggest the existence of multiple phases of immunosuppression in SCLC, wherein a basal PD-L1–independent state of immunosuppression (i.e., primary resistance to ICB) can be converted into a PD-L1–dependent state that is sensitive to ICB via activation of tumor cell innate immune signaling pathways. These results may explain, at least in part, why PD-L1 expression has not proven to be a predictive biomarker in SCLC. Whether DDRi could also overcome acquired resistance to ICB is not addressed experimentally by Sen and colleagues but merits investigation. Their findings are also reminiscent of other recent exciting preclinical studies showing that ICB can be strongly potentiated by activating tumor cell innate immune signaling. Inhibition of the epigenetic regulator LSD1 leads to double-stranded RNA (dsRNA) production from promiscuous transcription of endogenous retroviral elements and sensitization to ICB in a melanoma model (5). LSD1 inhibition suppresses tumor growth in SCLC even in immunodeficient models (6), and combining LSD1 inhibition with ICB might augment immune responses in SCLC. Similarly, suppression of the RNA-editing enzyme ADAR1 leading to increased sensing of dsRNA by cytosolic sensors also activated IFN responses and synergized with ICB (7). Taken together, these studies suggest a new form of synthetic lethality between ICB and disruption of pathways upon which tumor cells rely to suppress innate immune signaling.

Discoveries made by Sen and colleagues also illustrate key features of SCLC immuno-oncology research. Because of their compatibility with fully immunocompetent hosts, murine-derived tumor models, such as those that are already well established and widely used in the SCLC field, are likely to gain further traction. Importantly, murine SCLC models harbor a far lower somatic mutation burden than typically seen in the human disease, and the rapid tumor regressions observed suggest that synergism between ICB and DDRi has potential even in SCLC with lower mutational burden.

Most notably, the study by Sen and colleagues provides a compelling rationale for rapid translation to clinical studies testing DDRi with ICB in patients with SCLC, and patients with SCLC are being included in some such trials that are already under way. The findings are likely to be relevant beyond SCLC, as three new independent studies in other cancer types, including ovarian cancer, also support synergy between PARPi and ICB mediated through activation of the STING pathway (8–10). An important unanswered question is the degree to which DDRi might also promote proinflammatory signals in non-neoplastic cells and increase rates of serious ICB toxicity. Looking ahead, numerous studies are under way testing the combination of DDRi and genotoxic chemotherapy for a variety of cancers including SCLC, and the combination of platinum doublet and ICB, although a modest improvement over platinum doublet alone, is likely to become the standard of care for newly diagnosed patients with extensive stage disease. It is therefore tempting to imagine preclinical studies and clinical trials testing a “quadruple” (platinum doublet, ICB, and DDRi) regimen for SCLC. After decades without major changes to treatment options and outcomes, the advent of ICB has indeed reshaped the SCLC treatment landscape, but the accompanying improvement in outcomes has been small relative to other diseases. Further studies to establish a new generation of immunotherapy for patients with SCLC are therefore desperately needed.

D. MacPherson reports receiving commercial research grants from Roche and Janssen. No potential conflicts of interest were disclosed by the other author.

J.B. Hiatt acknowledges support from NIH training grant T32HL007093. D. MacPherson acknowledges support from NCI R01CA200547 and U01CA235625.