Small-cell lung cancer (SCLC) globally accounts for 13–15% of all lung malignancies. It is a highly aggressive neuroendocrine tumor, characterized by rapid growth and early tendency to widespread metastasis; stage IV disease represents over 70% of new diagnoses. Clinical onset is often associated with heavy symptomatic burden and rapid decline of overall health [1].

Chemo- and radiotherapy still represent the mainstay of treatment and an initial high responsiveness to such treatments is often observed [2, 3]. Recurrence, however, occurs very early in most cases, leading to a very dismal prognosis and a 5-year overall survival (OS) of 14.7–27.3% and 2.8% for early-stage (LD) and extended disease (ED), respectively [1, 4, 5].

Unfortunately, during the last three decades, life expectancy for SCLC patients has not improved, resulting in SCLC being defined as a recalcitrant cancer [6, 7].

In this disappointing scenario, there is a strong rationale for testing immune checkpoint inhibitors (ICIs), drugs that have changed the paradigm of treatment of non-small cell lung cancer (NSCLC) and other solid tumors in the latest years [8] (Table 1).Tab1

Summary of immune-modulating drugs and their targets

Epidemiological, biological and clinical features of SCLC suggest a potential efficacy of ICIs.

First of all, SCLC has a strong association with smoking status and exposure to cigarette smoking is a predictive factor for responsiveness to ICIs in NSCLC [9].

SCLC also harbors a high load of non-synonymous somatic mutations, so called Tumor Mutational Burden (TMB) [10]. This feature potentially results in the release of tumor neoantigens able to elicit an adaptive immune response against the tumor cells [11].

The capacity of SCLC to elicit immune response is also suggested by the presence of auto-immune paraneoplastic syndromes in about 20 to 40% of cases [12]. Tumor-enhanced immunity and neurologic paraneoplastic syndromes have been associated with better prognosis. In particular, in a recent study, median OS of SCLC patients without paraneoplastic syndromes was 9.5 months, versus 18 months for the patients with Lambert-Eaton syndrome [13, 14]. Even when a clinically overt paraneoplastic syndrome is not diagnosed, the mere presence of auto-antibodies is related to better outcome, reflecting the ability to elicit a humoral immune response [15].

On the other hand, there are specific clinical features of SCLC that may potentially limit the usefulness and the benefit of ICIs. First of all, SCLC is a rapidly progressive disease, requiring rapid tumor shrinkage with chemotherapy. Moreover, the majority of SCLC patients are symptomatic and require steroids and this is particularly true in case of superior vena cava syndrome and brain metastases [16, 17]. Chronic steroids are a known limitation for ICIs treatment [18].

For all these reasons, up to now, clinical data on the efficacy of monotherapy with ICIs in this disease are not so promising, in spite of a sound biological background. The antibodies used as immunotherapeutic agents belong to different IgG isotypes (Table 1). This may result in different activities since IgG1 are known to have stronger binding affinity to Fcϒ receptors compared to IgG2-3-4, thus able to mediate more effective antibody dependent cell-mediated cytotoxicity (ADCC). Despite pharmacological rationale, there are no demonstrated clinical differences among different isotypes; a reason can be found in the mechanism of action in relation to the immune target, since the action of anti-PD-1 antibodies can be independent from Fcϒ receptors [19].

Increasing evidence is available about molecular characterization and key pathways explaining specific features of immune-related microenvironment and key pathways responsible for the development of chemo-resistance.

In the manuscript we review molecular rationale for immunotherapy treatment, for synergism with chemotherapy and for other potential combination treatment including immunotherapy. We also summarize clinical evidence available and provide future potential perspectives.

Cytotoxic drugs can induce an immunogenic cell death, leading to the generation of molecular signals that promote the uptake of dying cancer cells’ debris by antigen presenting cells (APC), and the cross-presentation of tumor antigens to T cells. Multiple molecular mechanisms induced by cancer cells, such as downregulation of major histocompatibility complex antigen expression, induction of an immunosuppressive milieu and negative regulation of cytotoxic T-cells via checkpoint inhibition, can inhibit this response. Combining ICIs with chemotherapy may disrupt these escape pathways and efficiently restore the anti-tumor activity of the immune system [20, 21]. In SCLC, however, the level of evidence in this field is still scarce and incomplete; a more comprehensive knowledge of the molecular basis of the mechanisms of resistance to chemotherapy and immunotherapy and of the expected activity of different chemo-immunotherapy combinations is needed.

SCLC cells are characterized by ubiquitous loss of tumor protein p53 (TP53) and Retinoblastoma 1 (Rb1), the main gatekeepers of G1-S transition [11]. This results into tumor cells arrest upon DNA damage at G2-M checkpoint with subsequent imbalance in expression and interaction of many DNA-damage response (DDR) proteins (Fig. 1) [22].Fig1Fig. 1

Molecular landscape of SCLC. SCLC cells are characterized by ubiquitous loss of TP53 and Rb1 (dotted lines), the main G1-S cellular cycle checkpoints. SCLC cells depend on G2-M cell cycle checkpoint, that may be influenced by Aurora kinase A over-expression, characterizing the Myc-driven “variant” subtype of SCLC) and by Chk1-WEE1 axis. Chk1 is activated by Ataxia telangiectasia Mutated (ATM)/Ataxia telangiectasia and Rad-3 related protein (ATR) pathway upon chemo-induced DNA double strand break. After its activation, Chk1 can induce G2 cell cycle arrest through the phosphorylation of WEE1. Activated Chk1 can also up-regulate PD-L1 expression through the activation of the Signal Transducer and Activator of Transcription 1–3 (STAT1–3) mediated regulation of Interferon regulatory factor 1 (IRF1). Signaling pathways involving Enhancer of zeste homolog 2 (EZH2), an epigenetic modifier inducible both by immunotherapy and cytotoxic agents, also seem crucial in SCLC. EZH2 activity is required for the acquisition of an immunosuppressive phenotype, down-regulating antigen presentation process (resistance to immune-therapy), and also for an enhanced chemo-resistance property, through the inhibition of Schlafen family member 11 (SLFN11), a negative regulator of homologous repair machinery (HRM)

A body of evidence has been gathered over the years on the role of the tumor immune microenvironment (TME), i.e. the milieu of lymphocytes, monocytes and other immune cells intertwined with cancer cells, in neoplastic initiation and progression. The composition of the TME differs across time and stages even in cancers with same histology and it is one of the determinants of tumor characteristics and outcome of NSCLC patients [42].

An early study focusing on the interaction between SCLC cells and their TME showed how SCLC tumor cell lines were able to inhibit activated CD4+ T-cells [43]. The inhibitory activity did not require a direct cell-to-cell contact, but was mediated by cytokine secretion by tumor cells (IL-15 in particular) that caused a de novo functional differentiation of lymphocytes towards a T-regulatory immunophenotype (FOXP3+ CD4+ T-cells). Another study has analyzed FOXP3+ infiltrate in archival biopsies from patients with SCLC and the ratio of FOXP3+ turned out to be an independent indicator of poor prognosis in these patients [43].

The histological assessment of SCLC TME was the focus of another study that evaluated the prognostic role of CD45 (a pan-inflammatory cell marker) positive immune cells [44]. The extent of CD45+ infiltrate was predictive of a longer OS independently from clinical parameters such as stage and performance status [45].

Increasing evidence has indicated that TME is able to modulate the PD-1/PD-L1 axis, promoting the innate tendency of cancer cells to escape immune-surveillance [46]. Data on distribution of PD-L1 expression in SCLC across stages are very limited; in patients with advanced disease the level of PD-L1 expression seems to be lower than in earlier stages [47, 48] and also than in NSCLC [49].

A retrospective study conducted in ED-SCLC and LD-SCLC patients treated with a multimodal approach, including surgery for early stage, showed an association between CD8+ tumor infiltrating lymphocytes (TILs) and PD-L1 expression on tumor cells, whereas FOXP3+ infiltrate showed a positive correlation with PD-L1 positive tumor-infiltrating T cells [48]. Furthermore, a stronger infiltration of FOXP3+ TILs characterized early stage disease and was associated with a better prognosis in LD-SCLC patients, shedding a new light upon the controversial role of the T regulatory subset of TILs even in this malignancy [48, 50, 51].

In order to increase the therapeutic role of ICIs in SCLC, biological rationale supports the potentiality of combining ICIs with a number of non-chemotherapy agents with the aim to obtain synergism and subsequently improve both the percentage of patients who benefit from immunotherapy and the duration of clinical benefit (Table 5).Tab5

Ongoing clinical trials of immune-checkpoint inhibitors combined with non-cytotoxic agents

A first strategy concerns the idea that immune-tolerance mechanisms are redundant and that inhibiting more immune-suppressive targets may enhance anti-tumor activity. This is the most explored strategy and studies with the combination of nivolumab and ipilimumab have been described before.

On the other side, new drugs are under evaluation with the aim of actively promoting immune-response in combination with anti-PD1/PD-L1 antibody. For example, Utomilumab is a fully human IgG2 agonist monoclonal antibody targeting CD137, a co-stimulatory receptor expressed on activated immune cells (effector and regulatory T cells, NK cells and dendritic cells), causing an enhanced cytotoxic T-cell and NK-cell activity [75] and triggering antitumor response [76] (Fig. 2). In this case, rationale for synergism is strong: anti-PD-1/PD-L1 disrupts the PD1/PD-L1 interaction, thus avoiding tumor-induced anergy of tissue infiltrating lymphocytes, while utomilumab can boost the anti-tumor activity of different effector white blood cells.Fig2Fig. 2

New combination strategies. Mechanisms of action of drugs that are being studied for new combination strategies in small-cell lung cancer. Panel a: utomilumab triggers CD137, a co-stimulatory receptor expressed on activated immune cells and it is studied in combination with avelumab; trilaciclib is a CDK4/6 inhibitor and it is studied with platinum/etoposide and atezolizumab; SGI110 contrasts the role of EZH2, by interfering with DNA methylation and it is under evaluation in combination with durvalumab. Panel b: another promising strategy is to associate immune checkpoint inhibitor, such as Ipilimumab, to immune stimulatory agents. INCAGN01876 is a monoclonal antibody that activates Glucocorticoid-induced TNF-receptor-related protein (GITR), a T cell co-stimulatory receptor involved in the immunological synapsis able to enhance T cell responsiveness to weakly immunogenic tumor-associated antigens. INCAGN01949, another antibody that targets and stimulates OX40, a T cell co-stimulatory receptor that potentiates TCR signalling

Immune-related toxicity represents a major concern in SCLC. Autoimmune disorders are indeed frequent in SCLC patients, who may develop autoimmune diseases as paraneoplastic syndromes [12]. In this scenario, the relation between immune-related toxicities and treatment response might be intriguing, although data are scanty since patients with autoimmune disorders were excluded from clinical trials. To address this issue, retrospective series mainly involving NSCLC and melanoma patients have been described [81, 82]. Patients with active or inactive autoimmune disease have been treated with anti-PD1/anti-PD-L1 or anti-CTLA4. An autoimmune disease flare, mostly low grade and rarely requiring systemic corticosteroids, has been reported by about 20% of patients and this did not affect treatment outcome [81, 82]. Overall, the risk of immune-related adverse events was higher among patients with pre-existing autoimmune conditions, but the toxicity had no impact on survival [82]. No cases of paraneoplastic auto-immune syndromes were included in these series [81, 82].

In the CheckMate 032 trial with combined nivolumab and ipilimumab, the most frequent adverse events were increased lipases and diarrhea [49]. A peculiar, although rare, toxicity, was limbic encephalitis and aseptic meningitis across all treatment arms, while rash and hypothyroidism, mainly low-grade, were more frequently reported in the nivolumab-ipilimumab combination arms [49]. Rash and hypothyroidism were also the most common irAEs observed in the IMpower133 trial in the chemotherapy plus atezolizumab arm [54].

Pulmonary toxicity from the association of ICIs with chest radiotherapy may also be an issue. However, in the PACIFIC study, investigating durvalumab after chemo-radiation in stage III NSCLC, there were no differences in the incidence of grade 3 and 4 pneumonitis between the durvalumab and the placebo group [83].

In our experience, treatment with second-line nivolumab in a SCLC patient who had previously received thoracic radiotherapy for limited disease showed an exceptional clinical and radiological response. In the same patient, treatment was interrupted after 6 doses due to the occurrence of pneumonitis. The patient experienced single-site progression and received radiotherapy on a peri-pancreatic lymph-node. After the radiotherapy, he experienced further response on liver lesions and a relapse of immune-related pneumonitis, seven months after the completion of nivolumab treatment [84]. This experience shows how complicated are the effects of immune-modulation induced by cancer treatments and that the administration of radiotherapy also after ICIs and at distant-site may elicit immune-related adverse events.

Several trials have included correlative studies in order to find potential predictive markers of response.

In a trial combining ipilimumab 10 mg/kg with carboplatin and etoposide, the relation between baseline positivity of autoantibodies and clinical outcomes was evaluated. Patients with any positive autoimmune antibody (anti-SOX2, anti-Hu, anti-Yo, anti-VGCCA, anti-VGPCA, anti-nuclear, anti-neutrophil cytoplasmic antibodies) showed a trend for a prolonged survival (18.5 versus 17 months, p = 0.144), a significantly longer median progression free survival (8.8 versus 7.3 months, p = 0.036) and a trend for a higher response rate (p = 0.066) [85].

Differently from NSCLC trials, tumor PD-L1 expression in the Checkmate 032 was not predictive of ICIs efficacy in patients with SCLC [49]. Given this finding, samples were further analyzed: whole exome sequencing was performed and the tumor mutation burden was defined as the total number of non-synonymous somatic mutations [86]. Patients harboring higher tumor mutational load (defined as higher than the upper tertile of the mutation distribution of the study population) experienced an enhanced efficacy from the treatment, especially when the combination was administered.

Due to the limited availability of adequate tissue, there is an increased interest to use blood-based tests through cell free tumor DNA profiling. A blood-based surrogate of tissue-based tumor mutation burden evaluation has been shown to be a potential predictive tool for advanced NSCLC patients treated with atezolizumab [87]. Differently from NSCLC patients, patients with SCLC treated with atezolizumab plus carboplatin and etoposide showed a benefit in terms of OS and PFS, regardless of blood-based tumor mutational burden [54].

A retrospective study has evaluated tissue mutational burden (defined as total number of nonsynonymous mutations) of 120 patients with SCLC of all stages and the association with PD-L1 expression both on tumor and on immune cells [88]. Tissue mutational burden had no particular relationship with tumor expression of PD-L1, whereas there was a positive correlation with PD-L1 expression on immune infiltrate (p = 0.04). Gadgeel et al. have studied PD-L1 expression of cells confined in the tumor stroma of patients receiving pembrolizumab as a maintenance treatment after first line chemotherapy [60]. The stromal interface was considered PD-L1 positive if PD-L1 membrane-stained cells surrounding the tumor nests were identified at low power magnification. Patients with PD-L1 expression at the stromal interface had longer median PFS and median OS than patients with no expression (6.5 versus 1.3 months and 12.8 versus 7.6 months respectively). Exploratory analysis performed in the SCLC cohort of Keynote 158 has shown the potential of the PD-L1 combined score, i.e. the ratio of PD-L1 positive cells, including tumor cells, lymphocytes and macrophages, to the total number of tumor cells [67]. This PD-L1 score was able to define a subset of pre-treated ED-SCLC patients who achieved a better ORR (35.7% versus 6%), 1-year PFS (28.5% versus 8.2%) and 1-year OS (53.1% versus 30.7%) while on pembrolizumab.

The systemic treatment of SCLC represents a major challenge for medical oncologists and immunotherapy has a great appeal and solid biological rationale.

Initial clinical experiences confirm the potentialities of ICIs for this aggressive disease and indicate the need for reliable predictive biomarkers. Preliminary data suggest that predictive biomarkers of ICIs efficacy might be disease-specific and that findings validated in NSCLC cannot be translated in SCLC. In fact, a different evaluation score for PD-L1 expression has been suggested.

Responsiveness to immunotherapy is related to clinical disease course and to the host, but also to biological features of the disease. The study of molecular mechanisms at the basis of chemo-resistance and aggressiveness of the disease may help in understanding also the immune-resistance mechanisms and in individuating new combinations treatment strategies with the aim of improving clinical benefit of immunotherapy.

In addition to combining ICIs with chemotherapy and immunotherapy, new therapeutic approaches, specifically addressing molecular pathways involved in SCLC growth and chemo-resistance, need to be explored in order to contribute to improve the outcome of SCLC patients, commonly recognized as an unmet clinical need.