PD-L1 expressing circulating tumour cells in head and neck cancers

Head and neck cancer (HNC) patients often present with advanced metastatic disease. Whilst there have been improvements in the management of locoregional disease, distant metastatic spread remains a challenge in the field [1‐3]. Palliative chemotherapy is platinum based and for patients who progress after first line treatment or are refractory, therapeutic options are limited. Numerous agents including cetuximab, paclitaxel, gemcitabine and docetaxel have been assessed prospectively in the treatment of platinum refractory patients and the time to progression ranged from 2 to 6 months [4]. These systemic treatments produce a significant degree of morbidity and new therapeutic options are therefore a need in these patients. Once there is an established role in metastatic disease, translation into the curative setting is appropriate.

The programmed cell death-1/programmed cell death-1 ligand (PD-1/PD-L1) pathway has shown to play a crucial role in tumour immune invasion. Recent literature suggests that PD-L1 over expression in solid tumour types has shown direct tumour protection. Recent studies have shown that antibodies targeting PD-1/PD-L1 have significant anti-tumour activity with a much lower toxicity profile and are currently being investigated in a number of tumour types [5, 6]. Pembrolizumab (previously MK-3475) is a highly selective, humanized igG4 (kappa) isotype monoclonal antibody designed to block PD-1 interacting with ligands, PD-L1 and PD-L2, thereby allowing the immune system to target and destroy the tumour. Pembrolizumab was the first anti PD-1 antibody to be approved by the FDA [6].

In the 2014, American Society of Clinical Oncology (ASCO) meeting, it was reported that in a majority (77.9%) of pre-treated HNSCC patients, PD-L1 is expressed in the tumour, defined by ≥1% stained cells in the tumour microenvironment [7]. In the Keynote 012 trial presented at ASCO 2015, tumour shrinkage was found in 57% of patients, and overall response of 24.8%, comprised of 26.3% in HPV-negative and 20.6% HPV-positive patients [5, 8]. The Keynote 012 study indicated that Pembrolizumab was twice as effective as cetuximab with durable responses in patients which has not been seen previously in HNC. Pembrolizumab was also well tolerated in these patients with low rates of adverse effects. 86% of the responsive patients enrolled in the Keynote 012 study continued to receive treatment highlighting the acceptable safety profile [7, 8].

Metastatic sites have shown unique genomic alterations, which can be quite different from the primary site [9, 10]. Invasive procedures are currently required to biopsy these metastatic sites, some of which may be inaccessible. Other studies have shown that these biopsies may not be representative of all of the metastatic disease [11]. An alternative approach used in other cancer types is the analysis of blood samples for circulating tumour cells (CTCs) as a form of “liquid biopsy” [12‐14]. These rare tumour cells in circulation represent the “transient” cancer cell population that have the propensity to metastasize to distant sites. Recent reports have shown how CTCs may provide complementary information to identify candidate therapeutic targets and drug resistance mechanisms [9, 12, 15]. Moreover, CTCs represent cells from the primary and metastatic sites, thereby possibly providing a more comprehensive overview of the tumour burden of an individual patient. CTCs in the blood of HNC patients provide an opportunity to identify patients “at-risk” of developing overt metastasis in due course. More importantly, the analysis of these metastatic seeds in circulation may reveal important information for systemic therapy targeting metastatic disease [9]. CTCs are currently being investigated as predictive biomarkers for HER-2 targeted therapies [16]. A similar strategy could be used for immune checkpoint blockade therapies such as PD-L1.

We report the case study of a 56 year old man who had been diagnosed with a supraglottic SCC and treated with chemoradiotherapy. The patient was assessed for CTCs by the FDA-approved CellSearch (Janssen Diagnostics) and spiral microfluidics platform. Single CTCs were detected by both platforms. Using the spiral approach, CTC clusters were identified expressing PD-L1. We propose the notion that CTC PD-L1 assessment may be an avenue to identify patients who would be suitable candidates for anti-PDL1 therapy.

A 56 year old Caucasian male with a background history of Crohn’s disease, treated with azathioprine, presented at the end of 2013 with a supraglottic T3N2b squamous cell carcinoma which was treated with upfront chemoradiotherapy, utilising Cisplatin. In October 2014, a CT scan showed a 20 mm right jugulodigastric node. Lung lesions were observed which were queried to be fungal/distant metastasis by the MDT clinic. Biopsied specimen of the lung lesions confirmed moderately differentiated SCC. Progressive disease to the lungs and pelvic bone were observed whilst on Taxol and Carboplatin, given weekly. Having limited options, a trial of infusional 5-FU (fluorouracil) was given at the end of 2015. The patient died in February 2016.

The predictive value of PD-L1 expression in primary tissues is limited [9]. Furthermore, there is weak correlation between matched primary tumour and distant metastasis, suggesting that primary tumour is not an adequate surrogate for determining PD-L1 expression at metastatic sites [21]. Importantly, this highlights the fact that a single core biopsy may not suffice to determine tumour PD-L1 expression [9]. The FDA-approved CellSearch technology has shown clinical significance of EpCAM-positive CTCs [22]. However, there has been a shift in the field to maker-independent technologies to capture a greater proportion of CTCs in circulation in an unbiased fashion [18, 23]. By the use of established [9, 22] and spiral microfluidics technology [17, 24] this study aimed to capture the tumour cells in circulation of this patient and characterize the PD-L1 expression.

The patient presented with CTCs by enrichment using both CellSearch and spiral technologies. Importantly, the spiral technology was able to enrich for CTC clusters which are rarely detectable by CellSearch [25]. These microemboli/tumour cell clusters have shown an increased metastatic potential compared to single cells [26, 27]. Moreover, the cluster had a mid-high PD-L1 expression compared to the panel of known head and neck cancer cell lines. The PD-L1 + CTCs indicate that the patient had tumour cells in circulation with the capacity to block the immune system. These could be a potential targets for PD-L1 therapies [9]. The development of an immune-score is desirable for metastatic HNC patients.

In the 2016 AACR meeting, in locally advanced HNC patients, PD-L1+ CTCs were associated with shorter progression free survival (PFS) and overall survival (OS), and were proposed to “select and monitor patients for PD-1 checkpoint inhibitors” [28, 29]. These studies as well as this study, demonstrate that PD-L1 is expressed in HNSCC tumours and CTCs and may contribute to the tumours ability to evade the immune system. Moreover, that PD-L1 may be used as a biomarker for predicting responders from non-responders in lieu of the cost burden of such therapies.

This case report highlights the potential of using of CTCs to (i) identify patients ‘at-risk’ of developing metastasis (ii) identifying HNC patients that are likely to benefit from anti PD-L1 therapy and (iii) the development of a CTC PD-L1 immune-score for HNC. Further studies are warranted comparing patient tumour and CTC PD-L1 expression to develop predictive biomarkers.