The rationale for including immune checkpoint inhibition into multimodal primary treatment concepts of head and neck cancer

Medical need for improvement of definitive treatment in locally advanced SCCHN

Patients with SCCHN completing radiotherapy-based treatment remain at considerable risk for local relapse within the radiation field, regional recurrence in the neck and hematogenous spread of tumor cells with the potential to form distant metastases. Primary radiochemotherapy (RCT) applied concurrently with cisplatin still cures less than 40 % of patients [1], and in case of disease recurrence after RCT, the 2-years survival rate is below 20 %. Furthermore, the addition of chemotherapy to radiation improves locoregional control at the cost of severe acute and late morbidity [2] but does not reduce the risk of distant metastases [1, 3].

Immune evasion in SCCHN: hideout, defense, camouflage and balanced immune destruction

There are several ways for SCCHN to evade recognition by the adaptive immune system, as schematically illustrated in Fig. 2. Early tumors may grow in a hideout, because they display neither a significant level of apoptosis or necrosis nor inflammatory signals and thus do not elicit any danger signal in the tissue (Fig. 2a). Although danger signals might subsequently be produced during the progression of the disease, the lymphocytic infiltrate may be confined to the rim of the tumor tissue with no infiltration into the tumor itself (Fig. 2b). Secretion of indoleamine 2, 3-dioxygenase (IDO) is among the major defense mechanisms used by tumors to prevent lymphocytic infiltration [10]. Tumors frequently also counterbalance infiltration by lymphocytes by down-regulation of their MHC molecules, thereby avoiding the presentation of peptides from intracellular proteins to T cells which results in an effective camouflage (Fig. 2c). As schematically depicted in Fig. 2d, in tumors with extensive inflammatory cell infiltration a delicate balance between immune destruction and immune evasion may exist which is based on immunosuppressive mechanisms including high expression of IDO and PD-L1 as well as recruitment of FoxP3+ regulatory T cells (Treg) [11]. Representative histological images from SCCHN tumor sections exemplifying the above-mentioned types of immune evasion are presented in Fig. 3.

A number of both genetic and environmental mechanisms allow such immune escape and have been described in SCCHN (for a recent detailed review see [12]), including selection of poorly antigenic cancer cell subsets, disturbances in MHC class I- and class II-mediated antigen presentation [13–15], expression and secretion of immunosuppressive cytokines [16], expression of the pro-apoptotic Fas ligand to induce activation-induced cell death in T cells [17], and recruitment of immunosuppressive immune cell subsets into the tumor [18]. More recently, evidence is increasing that expression of immune checkpoint components that may limit T cell responses also occurs frequently in these tumors [18–20].

Interference of immune checkpoints with resistance to RCT

Deregulated expression of immune checkpoint proteins has already been linked to poor efficacy of RCT in several tumor models. High expression of PD-L1 in tumor cells and stromal lymphocytes accompanied by low CD8+ T cell infiltration has recently been identified as a poor prognostic biomarker in patients with stage III non-small cell lung cancer (NSCLC) receiving cisplatin-based RCT [45]. In addition, tumor control by neoadjuvant or concurrent RCT was observed to be inefficient in patients with esophageal squamous cell carcinomas displaying elevated immunostaining for PD-L1 in neoplastic and adjacent non-malignant esophageal epithelium [46]. Preclinical studies in a variety of syngeneic mouse models of cancer [47] have also demonstrated that expression of PD-L1 can be induced by radiation itself and that such upregulation impairs both local tumor control and protection against tumor re-challenge [47]. It is therefore not surprising that blocking antibodies directed to these immune checkpoints were able to significantly enhance the immunogenic effects of radiotherapy [27, 48, 49].

Potential biomarkers for patient selection for immune checkpoint blockade

Precise biomarkers to identify patients who benefit from immune checkpoint inhibition alone or in combination with RCT still have to be established. Current data in multiple cancers reveal that verification of PD-L1 overexpression by immunohistochemistry is associated with improved clinical outcome of anti-PD-1 therapy. However, the presence of robust responses in some patients with low or undetectable expression of PD-L1 complicates the issue of PD-L1 as an exclusive predictive biomarker [54]. In the Keynote-012 trial of SCCHN, an elevated expression of PD-L1 and the presence of an interferon-γ expression signature were associated with improved progression-free survival [39, 55]. The same signature had previously been established as predictive signature for outcome after pembrolizumab in metastatic melanoma [56] and its predictive value was also demonstrated in advanced gastric cancer [57]. In addition, patients with large immune cell infiltration in tumor tissue and high mutational load were more likely to benefit from immune checkpoint blockade in bladder [58] and colorectal cancer [59]. Taken together, patients with PD-L1+ tumors displaying immune-related gene expression signatures, including genes regulating T cell functions, the antigen presentation machinery and IFN-γ signaling, are most likely to benefit from immune checkpoint inhibition.

Rational development of combination regimens

While there is much excitement around the phenomenon of a radiation-induced anticancer immune response and combining radio(chemo)therapy with immunotherapy, numerous questions remain to be addressed in clinical trials. A major challenge is to identify not only the optimal immune checkpoint inhibitor as partner for a given radiotherapy schedule but also the best chronological sequence for their combined application. Preclinical evidence can serve as guidance in treatment schedule and clinical trial development. As outlined above, danger signals induced by radiation lead to the recruitment of immune cells into the tumor. However, cells of the immune system are also vulnerable to radiation, as their exposure to ionizing radiation induces apoptosis in mature natural killer (NK) cells as well as T and B cells. Since radiotherapy is generally delivered in daily fractions, re-irradiation of the tumor might therefore damage infiltrating immune cells that display cytolytic activity themselves or might significantly reduce the capacity of DCs to activate effector T cells. In addition, DCs may find a hostile environment for T cell activation in draining lymph nodes, which represent their natural surrounding for interaction with T cells, as draining lymph nodes are systematically included into the radiation field in SCCHN. Conversely, however, the induction of immunogenic cell death (ICD) by each daily radiotherapy fraction might transiently generate a favorable milieu for immune activation within the tumor tissue, which may vanish again after termination of radiotherapy. In support of the latter, concurrent application of an anti-PD-L1 antibody together with fractionated radiotherapy significantly improved tumor control in a xenograft model [47]. In contrast, fractionated radiotherapy followed by delayed application of anti-PD-L1 was completely ineffective in enhancing the local efficacy of radiotherapy [47]. Certainly, more studies will be needed to address this important issue.

Besides the optimal time schedule also the optimal type and dosing of chemotherapy has to be established, if desired to be included into the treatment regimen. Significant differences in the ability of chemotherapeutic drugs to induce ICD have been reported previously [32]. Cisplatin which is an essential component of current state-of-the-art RCT regimens in SCCHN does not induce ICD [60] despite its presumed identical mechanism of action to that of oxaliplatin, a potent inducer of ICD. This has been attributed to the lack of calreticulin exposure after cisplatin treatment [60]. However, radiotherapy is a potent inducer of calreticulin exposure [30], and recent studies have shown that combining cisplatin with compounds that induce calreticulin exposure leads to full-scale ICD [60]. Thus, potentiation of ICD by cisplatin could still represent one of its major mechanism of action when cisplatin is administered concurrently with radiotherapy.

Taxanes including docetaxel and paclitaxel which are also common combination partners of radiotherapy in locally advanced SCCHN are known to modulate antitumor immune responses as well [61]. Similar to cisplatin, paclitaxel does not induce ICD. However, concurrent paclitaxel treatment was shown to significantly enhance radiation-induced ICD in breast cancer cell lines [62]. Similarly, docetaxel treatment itself did not induce ATP or HMGB1 secretion by tumor cells. However, calreticulin exposure of tumor cells after docetaxel treatment was observed which significantly enhanced tumor cell killing by antigen-specific CD8+ cytotoxic T cells [61].

Intriguingly, investigations on the immunogenic effects of chemotherapeutic drugs revealed also their direct interference with immune checkpoint expression. Treatment of DCs in vitro with platinum-based compounds strongly enhanced their potential to activate T cells which was caused by downregulation of PD-L2 in DCs [63]. This effect was mediated by inactivation of STAT6, the transcriptional regulator of PD-L2, and occurred also in tumor cells resulting in their enhanced recognition by T cells [63].

Overall, these preclinical observations provide a sound rationale for investigating immune checkpoint inhibitors with radiotherapy alone as well as in combination with standard cisplatin-based as well as taxane-based RCT. In addition, there is accumulating data that the efficacy of cetuximab-based regimens in treatment of recurrent/metastatic SCCHN is not only based on the inhibition of EGFR signaling pathways but also on the activation of Fcγ receptor-positive NK cells leading to DC maturation and activation of cytotoxic T cells [64]. A combination of immune checkpoint inhibitors with cetuximab-based bioradiation protocols might therefore also represent a very attractive chemotherapy-free concept for improvement of primary treatment of locally advanced SCCHN.

Toxicity of combination regimens

The toxicity of radiotherapy is mostly occurring directly at the irradiation site. Mucositis, xerostomia and swallowing dysfunctions are common side effects in radiotherapy of head and neck cancers. Clearly, the extent of early and late toxicity is dependent on the radiation technique and the applied dose: Intensity-modulated radiotherapy, which conforms closely to the tumor volume, avoids or minimizes exposure to unaffected tissue and thereby significantly reduces local side effects of irradiation [65]. On the other hand, addition of concurrent chemotherapy to radiotherapy not only improved efficacy of the treatment in locally advanced SCCHN but also increased both the toxicity and the spectrum of adverse events as compared to radiotherapy alone [66]. The toxicity of immunotherapy is dependent on the administered agent and dosage. In previous clinical trials, immune checkpoint blockade immunotherapy presented acceptable toxicity. Even occasional severe toxicity was manageable through treatment interruption or involvement of immunosuppressive drugs. During ipilimumab treatment approximately 60 % patients showed immune-related adverse events, of them 10–15 % being grade 3–4 [67]. The blockade of PD-1/-L1 showed less severe ir-AEs in previous phase I studies [68]. Diarrhea and skin rash were the most common immune-related adverse events after ipilimumab, other adverse effects included enterocolitis, hypothyroidism, hypophysitis and neuropathies [68]. The most common adverse events reported for both nivolumab and pembrolizumab were mild fatigue, rash, pruritus and diarrhea, which could be usually managed without dose interruption or discontinuation [68]. First toxicity data from a phase I study of the combined application of ipilimumab with radiotherapy for treatment of metastatic melanoma (NCT01497808, [48]) argue against an exacerbating toxicity profile of the combined regimen. No dose-limiting toxicities, defined by the study protocol as any treatment-related grade ≥4 immune-related toxicity or grade ≥3 non-immune related toxicity experienced during study treatment or within 30 days after the last injection of ipilimumab were observed [48]. Considering the different kinds and acceptable adverse events, the combinatorial treatment of radiotherapy and immune checkpoint inhibitors seems feasible for SCCHN patients.