"Present and future of immunotherapy in Neuroendocrine Tumors"

While in the past the field of cancer therapy was dominate by surgery and radiotherapy associated with chemotherapy [1, 2], in the last decades research has provided new treatment strategies, such as target therapies. Target therapy, understood as use of drugs or other substances that targets specific molecules to arrest the growth and spread of cancer cells [3], has sharply revolutionized the outcomes of different types of cancer.

Currently, another intriguing weapon available in cancer therapy is represented by immunotherapy, based on the stimulation of the immune system against cancer cells through the introduction of cytokines and antibodies (passive immunotherapy) or the introduction of vaccines and immune cells themselves (active immunotherapy) [4]. Immune evasion mechanisms have a pivotal role for tumor cell proliferation and growth [5, 6]. The Cytotoxic T-Lymphocyte Antigen 4 (CTLA-4) [7] and the programmed death-1 (PD-1) and programmed death-1 ligand (PD-L1) [8] represent a key point regarding escape from immune surveillance by cancer cells, and this reason has led to the development of antibodies against these molecules. These antibodies, named ‘immune checkpoint inhibitors’ (ICIs), have opened a new era in oncology [9, 10]. ICIs proved one of the most effective therapeutic approaches for several cancers. However, many patients are unresponsive to ICIs [11] and these immunotherapies can produce serious non-specific systemic inflammation and autoimmune side effects [12]. Therefore, the study of the tumor microenvironment has become fundamental [13] with the aim of identifying biomarkers to select the patients who can benefit most from these treatments.

Immunotherapy, so promising in many neoplasms, still does not have a precise role in the treatment of neuroendocrine neoplasms (NENs).

NENs constitute a heterogeneous group of rare tumors, which arise from enterochromaffin cells and can present throughout the body, however, in most cases, they are localized in the gastrointestinal tract, pancreas and thorax [14]. A rate amounting around 11–22% of all NENs are defined as unknown primary origin when their primary tissue of origin has not been identified with standard diagnostic work-up. Delineating the primary site of origin has important implications for selecting the appropriate treatment and overall prognosis. The small bowel, followed by the lung and pancreas are the most prevalent primary sites of origin in case of NEN with unknown primary. [15]

NENs display a highly variable biological features, clinical course and prognosis, making prediction of survival difficult. The most recent WHO 2019 classification [16], based on the WHO 2010 classification and extending the WHO 2017 classification [17], has established the importance of classifying gastroeneteropancreatic (GEP) NENs considering the primary site localization, the morphological differentiation, and the grading. According with these affirmations, GEP-NENs are graded into three different categories taking into account the Ki67 proliferation index. The real novelty of the last WHO 2019 classification is the G3 category, characterized by a Ki67 > 20% and a well differentiation morphology, extended to all GEP-NENs. Furthermore, it worth considering neuroendocrine carcinomas (NECs) represented by a Ki67 proliferation index > 20% and poorly differentiated.

The cornerstone of treatment in well-differentiated neuroendocrine tumors (NETs) is based on surgery, local ablative treatments, antisecretory and antiproliferative drugs, such as somatostatin receptor ligands (SRLs), peptide receptor radionuclide therapy (PRRT), and target therapies, while high-grade NECs benefit mostly of platinum-based cytotoxic chemotherapy [18]. However, new therapeutic strategies are being studied for the treatment of NENs, such as immunotherapy. In particular promising results have been especially observed in the treatment of two of most aggressive NENs with ICIs [19]: Merkel cell carcinoma (MCC), a rare very aggressive NET of the skin, and small cell lung cancer (SCLC). Few other types of neuroendocrine neoplasms are beginning to show interesting results [20].

In this article, we provide an overview on the current knowledge about immunotherapy applied to NENs, evaluating future perspectives in this setting.

We performed a literature search by MEDLINE (PubMed database) and we also considered the trials registered on clinicaltrials.gov to identify potentially relevant articles on immunotherapy with ICIs and NENs of any grade and primary site. The search was last updated 31 January 2021. The search strategy included the following terms “immunotherapy” OR “immune checkpoint inhibitor” OR “immune checkpoint blockade” OR “spartalizumab” OR “pembrolizumab” OR “toripalimab” OR “nivolumab” OR “ipilimumab” OR “atezolizumab” AND:

Only articles published in English were considered. Additional studies were identified by reviewing the references of all selected articles. The methods of potentially eligible studies were assessed independently by three reviewers (MA, AD, FN).

The incidence of NENs has risen in the last few decades. Especially, as for prevalence, GEP-NETs are the second neoplasm of the gastro-intestinal tract [21]. GEP-NENs arise in secretory cells of the gastrointestinal tract (GI-NENs) and pancreas (panNENs) [22, 23]. As for treatment, in well-differentiated GEP-NETs the first choice is local surgical treatment, when applicable. In case of metastatic NETs or surgical approach unenforceable, antiproliferative and antisecretory therapy with SRLs are recommended [18, 24]. Moreover, other strategies such as PRRT has been demonstrated effective in prolonging progression-free survival (PFS) especially in patients with GI-NETs [25]. Everolimus, sunitinib, and chemotherapy are further treatments investigated and approved for GEP-NENs. As regards GEP-NECs, cytotoxic chemotherapy platinum-based with etoposide is the most common available approach [26, 27]. It’s worth considering the poor prognosis of these carcinomas with a median of about 7.5 months [28]. Furthermore, the therapeutic landscape for NENs has been evolving constantly and several clinical trials are ongoing to characterize immunotherapy as an emerging therapeutic strategy. In particular, ICIs, using anti–PD-1, anti–PD-L1 and anti-CTLA-4 agents, has been investigated in GEP-NENs in the last few years. Several studies have evaluated the expression of these checkpoints molecules in NETs of different grades and in NECs [29, 30]. The potential a given NEN to respond to ICI is still largely unknown. Immunohistochemical (IHC) evaluation of PD-L1 expression in the tumor microenvironment and its role in predicting response to ICIs is a very burning topic. Moreover, a high tumor mutational burden (TMB) has been found associated to an increased benefit of ICIs [31‐33] and was advocated as a possible useful biomarker to select potential responders to these drugs.

The importance of the immune microenvironment in patients with NETs has been established in the last years [34]. Immune cell infiltration is only documented in GEP-NETs, and overall, it appears to be higher in panNETs rather than in midgut NETs, perhaps as a result of the higher TMB of panNETs [35, 36]. Morever, in another series of NENs, potentially immunosuppressive regulatory T cells (Treg) were present in 55% of intermediate/high-grade tumours, whereas only 16% of low-grade metastatic NENs showed intratumoral Treg (P = 0.02). [37]. Nevertheless, more knowledge about the composite immune landscape of these heterogeneous tumors needs to be gained in order to clarify the prognostic implication that have these NENs features. [38].

Furthermore, even prognostic evidences linked with the expression of immune checkpoint molecules PD-L1 and PD-1 are still unclear and debated across different studies [39‐42]. There are some data in the literature showing that PD-L1 expression was significantly correlated to a higher WHO grade of NENs [39, 42] and also with a poor PFS and overall survival (OS). On the contrary, others authors have found that PD-L1 expression did not correlate with grade or prognosis [40, 43]. In fact, recently, the expression of PD-L1 was investigated in a large cohort of patients with G3 GEP-NEN and only 10% of these tumors expressed PD-L1 and those lesions, with positive PD-L1 immunoreactivity in tumoral cells, were all poorly differentiated cases [43]. In this study no correlation was found between PD-L1 immunoreactivity and clinical parameters evaluated, such as age, sex, primary site, PFS and OS. Anyway, the design of the study did not include treatment intervention with ICIs, therefore the clinical outcome and the response to therapy could not be correlated with the expression of PD-L1 in this series [43].

As reported in different studies, TMB is typically low in well-differentiated lesions (NET G1 or G2) and higher in poorly differentiated carcinomas (G3), [34, 44].

Globally, all these studies have aroused new perspectives to initiate clinical trials aimed to identify the efficacy of ICIs in NENs.

Evidence so far available for ICIs in GEP-NETs is definitively not as robust as for other tumors such as SCLC or MCC. We, therefore, had to take a further look to what the near future might reserve for ICIs in GEP-NET and NEC through an in-depth research on the clinicaltrials.gov database, considering only phase II or III studies, evaluating every Food and Drug Administration (FDA) approved ICIs molecule and not approved ones as well. Our research yielded 42 results, of which 19 were deemed relevant for our purpose; the results are summarized in Table 1; the considered drugs are used either in monotherapy, or in association with another ICIs, with Tirosin Kinase Inhibitors (TKI)/other target therapy, or with SRLs, or PRRT, or chemotherapy.

A total of nine different ICIs drugs are used throughout the 19 Randomised Controlled Trials (RCTs). Such numbers highlight the growing attention gathering around NENs and ICIs, in response to the need of stronger evidences supporting such therapy.

Some of the reported trials have, however, already produced initial results, most of them in the form of abstracts/interim analyses: we retrieved seven initial reports from seven different RCTs, the majority of which showing promising results.

Four analyses report a clinical benefit rate (CBR) > 30% [51‐54]. The first [51] is derived from NCT03043664, which investigates pembrolizumab plus SRLs in 22 patients with low/intermediate grade GEP-NETs. With this schedule around 40% of patients achieved stable disease (SD). Relevant antitumor activity was also described for avelumab, tested in 29 patients [52] deriving from NCT03352934, with either NEC or NET (mostly GEP), that obtained a disease control rate (DCR) of 32%. Nivolumab combined with temozolomide showed promising results in an interim analysis [53] from NCT03728361: out of 12 patients (7 GEP-NETs and 5 pulmonary carcinoids) 25% showed partial response (PR), 67% SD and 8% progressive disease (PD), although the median follow-up was short (approximately 7 months) and the cohort very limited. Lastly, according to an interim analysis [54] from NCT03365791, spartalizumab, administered with LAG525 (an antibody to LAG-3), achieved an astounding 86% of clinical benefit rate at 24 weeks in a GEP-NET cohort. With the exception of nivolumab plus temozolomide, the majority of the above mentioned analyses reported a good to an excellent tolerability as well.

More modest results are reported in another interim analysis [55] of NCT03074513, which aimed at defining the activity of atezolizumab plus bevacizumab in panNET and extra panNETs, achieving a confirmed objective response rate of 20% and 15% (95% CI 3–38%), respectively. This last association was also well tolerated. Worse results were, instead, reported in 14 patients with progressive NECs treated with pembrolizumab alone (partial results from NCT03136055). Pembrolizumab was deemed ineffective in this small cohort [56]. Similarly, durvalumab in combination with tremelimumab was almost ineffective the abstract [57] derived from NCT03095274: a global cohort of 123 patients was divided in subgroups considering the primary origin and grade (GI NETs, panNETs, GEP NECs and lung carcinoids, 123 patients in total) where a CBR > 30% at median follow-up of 10.8 months was observed only in the panNET group. Nor safety or tolerability concerns were registered.

As already mentioned, the results achieved in these RCTs should be considered partial because derived from studies including often a too low number of patients. However, considering the overall promising preliminary results observed, final results are eagerly waited for.

Lung NENs are a group of rare and heterogenous pulmonary tumors classified into four different histological groups following the 2015 WHO lung NEN classification [58]. Well-differentiated lung NENs are typical (TC) and atypical (AC) carcinoids often characterized by indolent clinical behavior. On the other hand, pulmonary high-grade NENs are SCLC and large-cell neuroendocrine carcinoma (LCNEC). Briefly, TC is the most frequent lesion among the four histological groups and have a metastatic potential in up to 15% of patients with a median time to recurrence of 4 years. AC presents metastatic spread in up to one half of cases and has a median time to recurrence of 1.8 years. As for advanced lung NET, median survival from diagnosis is about 6–7 years [59]. Concerning treatment, it worth considering a different approach based on the histological group and patients tailored. Since now, few trials have been designed for lung NETs specifically, therefore the therapeutic management largely derives from studies in patients with GEP-NENs [60]. In this view, the emerging immunotherapy strategies might represent a new base to change the therapeutic scenario of this group of NENs. Particularly, immunotherapy has become a paradigm shift in the treatment of SCLC. The presence of lymphocyte infiltration in lungs NENs have not been fully investigated so far. Researchers have obtained discrepant results as for the association between the host immune response to lymphocyte infiltration and the prognosis of lung NENs, especially in terms of OS and PFS [29].

MCC is a rare skin malignant tumor first described in 1972 [79], which neuroendocrine origin has been hypothesized only in 2012 [80]. Although being a rare neoplasm, it is responsible for the higher number of skin cancer-related death per year after melanoma. About 80% of the cases is driven by the infection of Merkel cell polyomavirus (MCPyV) [81], therefore being the major risk factor, while other risk factors for the development of MCC include UV irradiation, and immunodeficiency conditions.

Since patients showing a particularly strong immune response have been found to have more favorable prognoses [82], the possibilities of immunotherapy for MCC appear to be greater than in other types of neoplasms, including other NENs.

Indeed, therapy with ICIs for MCC is certainly proceeding at a faster pace than for the majority of other NENs, and evidence from literature concerning MCC is more robust.

The position of immunotherapy in the treatment algorithm of metastatic/advanced MCC is beginning to gain more and more consideration, since chemotherapy, despite showing satisfying response rates, cannot guarantee a durable response [83]. Considering that some chemotherapy regimens can further worsen the impairment of the patient’s immune system, and given the pivotal role of such factor when speaking of MCC, attention towards different approaches is steadily growing, and ICIs are drawing much of such attention.

As said, the risk of developing MCC is increased from five to 50-fold in persons with proven T-cell dysfunction, such as HIV-positive patients, solid organ transplant recipients, patients affected by hematological malignancies, and other immune system-impairing situations. Indeed, immunocompromised individuals represent approximately 10% of all the MCC patients [82, 84, 85]. Furthermore, immunosuppression represents a dramatic negative prognostic factor for mean survival, with patients showing any degree of immune system dysfunction showing an OS at three years approximately half of immunocompetent individuals (40% vs 74%) [86]. Conversely, better outcomes have been observed for patients showing particularly strong immune response features, such as highly CD8⁺ lymphocytes infiltrated tumors and MCCs showing no primary (MCC of unknown primary), which have higher tumor mutational burden and drive strong immune response [87, 88].

It comes of no surprise, though, due to its peculiar carcinogenesis, that MCC is indeed a highly immunogenic neoplasm.

As a matter of fact, it has been observed that both the infection by MCPyV and the DNA damage from UV radiation exposure (and the consequent high tumor mutational burden) lead to an increased response by the patient’s immune system [89].

MCPyV-positive MCCs express virus-related oncoproteins that eventually will drive unregulated growth [90]. Such proteins cause the patient’s immune system response in terms of intratumoral infiltration of CD8 + lymphocytes, the presence of MCPyV-specific T cells in the peripheral blood stream and IgG antibodies against said antigens [91, 92].

Similar pattern has been shown also in virus-negative MCCs, where the significant mutational burden caused by UV-radiation leads to production of neo-epitopes and proteins that are recognized by the patient immune system [93]. The presence of the aforementioned markers of immune system activation are all predictors of better prognosis [82].

Where it is true that a robust immune system activation is a favorable prognostic factor, it is also true that MCC can develop methods to escape the patient’s immune system. For example, MCCs thwart NK and T-cell–mediated oncolysis is by downregulating Toll-like receptor 9 MICA/MICB and MHC. A pivotal role in escaping lymphocyte-mediated eradication is also represented by PD-1 signaling: the tumor microenvironment, through the action of type II interferons, cause the overexpression of PD-L1 by the cancer cells, contributing to a decrease in the ability of inducing tumor cell death. Moreover, CD8 + T-cells infiltrating the tumor frequently overexpress PD-1 protein in response to persistent viral or tumor antigen exposure, as a marker of T-cell exhaustion [94].

PD-1 signaling appears to play a fundamental role in MCC; moreover, the highly immunogenic nature of this tumor provides a credible rationale for treatment with ICIs. To date, the only two ICI molecules FDA-approved for MCC are avelumab and pembrolizumab. [95, 96]; nevertheless, studies and trials concerning different ICIs are starting to populate the scientific panorama and initial evidence is available also for them eligible papers are summarized in Table 2.

Immunotherapy, at this time, finds its precise place in the treatment of SCLC and MCC. In these two forms of aggressive and high-grade NENs, ICIs have obtained important and encouraging results, which in the future will still be implemented by the ongoing trials regarding combined treatments among different ICIs and other molecules.

On the contrary, with regard to the well-differentiated forms of NENs, the results obtained so far have been disappointing and, unfortunately, immunotherapy cannot be considered a promising therapy at the moment. ICIs have no impact on clinical practice in the treatment of differentiated NENs and do not yet have a precise role in the therapeutic sequence of these tumors.

Considering the low TMB and the scarce tumor-infiltrating lymphocytes that characterize NETs, these results, obtained so far, are not surprising. This picture does not mean that in the future a better outcome cannot be obtained in patients with NETs as well. However, the most important aspect will be to study strategies that can make NETs more susceptible to response to ICI and, thus, enhance the effectiveness of these treatments. Therefore, the combination of conventional therapy, target therapy and immunotherapy deserves attention and warrant to be explored [114]

A sequential chemotherapy, possibly inducing an increase in TMB and tested before immunotherapy, could be a hypothesis deserving more consideration. Radiotherapy or PRRT, that precedes the start of ICI and induces an inflammation at tumoral level, with an increase of tumor-infiltrating lymphocytes, could be another approach to explore. Equally essential will be the identification of biomarkers useful for selecting patients potentially responsive to this type of treatment. In fact, responses to immunotherapy were also reported in tumors presented with low PD-L1 expression and low TMB, underlying that a screening of patients based only on these two biomarkers could exclude potential responders to ICIs [32, 33].

Some intrinsic characteristics of NENs make difficult to study all these hypotheses, since the large heterogeneity of NENs in their biology and clinical behavior, and their relative rarity, combined with the habit of conducting studies on non-homogeneous case-series by primitive site or grading, do not allow to obtain univocal and established data on this topic. Until we have clarified all these aspects, it will be difficult for ICI to revolutionize the therapeutic approach of well-differentiated NENs.