Developing androgen receptor targeting for salivary gland cancers

Salivary gland cancers (SGCs) are rare malignancies that can arise from major (parotid, submandibular, sublingual) or minor glands present throughout the upper aerodigestive tract. These tumors encompass a diverse set of diseases representing more than 20 different histologic subtypes with distinct clinical behaviors, genomic/molecular profiles, and responsiveness to drug therapy [1–3]. Similar to other rare diseases, the lack of prospective drug trials focused on singular SGC tumor types makes evidence-based management a challenge.

In this issue, Fushimi et al. move the field a step forward by reporting the results of a prospective, biomarker selected SGC clinical trial evaluating combined androgen blockade (CAB; leuprorelin acetate plus bicalutamide) in androgen receptor positive (AR+) SGC patients [4]. AR is most commonly a pathologic feature of salivary duct carcinomas (SDCs) as 70%–89% of this histology are AR+ by immunohistochemistry [5–7]; a smaller subset of adenocarcinomas also may express AR (∼20%) [8]. Indeed, 34 of the 36 enrolled patients on the trial were SDC patients (the other two were adenocarcinoma, NOS). Recent studies have reported X chromosome gain (AR is located on Xq11-12) in a significant number of SDC tumors (37% and 58% [9, 10]), although this cytogenetic aberration did not always correlate to AR protein expression [10], leaving the mechanism by which AR is overexpressed uncertain. Importantly, AR expression in SDCs correlates to upregulation of an AR transcriptional signature [11] that is required to drive the proliferation of several SDC cell lines in vitro [10, 12]. These observations argue that AR is not simply expressed in SDCs, but biologically critical for the oncogenic phenotype of these cancer cells, nominating it as a rational therapeutic target.

Even before these biologic insights, treating physicians have extrapolated from prostate cancer therapeutic paradigms to empirically treat AR+ SGC patients with antiandrogens alone or in combination with androgen deprivation therapy (ADT; known as CAB when combined with an antiandrogen), reporting major response rates ranging from 18% to 65% in larger retrospective series [9, 13]. The work of Fushimi et al. now provides the first prospective evidence regarding the efficacy of AR-targeting in this disease [4]. The authors note that the trial was conceived without a standard biostatistical plan given the lack of historical data necessary for formulating statistical assumptions and hypotheses, as well as the investigators’ uncertainty regarding the sample size that could be feasibly accrued. These pragmatic concessions aside, the overall results of the trial are impressive as CAB produced a 41.7% major response rate, including four (11.1%) complete responses. Moreover, 70.0% had evidence of tumor regression, a 75.0% clinical benefit rate was observed, and only five patients (13.9%) had progression of disease as the best overall response, all indicators of fairly broad benefit with AR-targeting in this population. It is difficult to ascertain how these results compare to the efficacy of cytotoxic chemotherapy regimens that are also typically used to treat these patients, although presumably CAB would be less toxic. An ongoing European Organisation for Research and Treatment of Cancer (EORTC) randomized phase II trial will address this question by comparing the efficacy of CAB to standard chemotherapy in the first-line setting for recurrent/metastatic AR+ SGC patients (NCT01969578). Additionally, a US National Clinical Trials Network (NCTN) study evaluating the second generation antiandrogen enzalutamide (NCT02749903) as well as another Canadian trial (NCT02069730) will provide greater prospective insight into the efficacy of AR-targeting for these patients.

With a median progression-free survival of 8.8 months, acquired resistance to CAB presumably was an issue for some of the patients on the study. While it is not known what therapies patients received subsequent to study removal, there are published case reports of responses with the androgen depleting CYP17A1 inhibitor abiraterone in ‘castrate-resistant’ SDC patients (those who progressed on CAB) [14, 15], suggesting an important biologic parallel with castrate-resistant prostate cancer (CRPC) in which tumor growth may remain AR-dependent even after failure of first-line hormonal therapy. It is not known if some of the AR-dependent mechanisms of resistance delineated in CRPC, such as AR amplification and AR point mutations, are relevant for SDCs. Alternative splicing of the AR transcript resulting in constitutively active isoforms missing the exons bound by AR-targeting agents has been implicated as another mechanism of hormone resistance in CRPC. Interestingly, the splice variant AR-V7, which has been associated with resistance to enzalutamide and abiraterone in prostate cancer [16], is also present in SDC tumors [10, 11]. Analysis of the SDC genomic landscape revealed other alterations that could impact CAB therapy outcomes, including genes in the AR axis [e.g. forkhead box protein A1 (FOXA1), fatty acid synthase (FASN)] and other biochemical signaling pathways (e.g. PTEN/PI3K/AKT and HER2) that have been shown to modulate AR function in prostate cancer [11, 17–19]. Among the enrolled patients, the authors report the discovery of HER2 amplification as well as mutations in PIK3CA, AKT1, N-/K-/H-RAS, and BRAF [4]. While the sample size of the study is too small to clearly delineate how these alterations may have impacted outcomes with CAB, these findings are consistent with the previously published observation that more than 60% of molecularly profiled SDCs possess potentially actionable genetic mutations [3, 9, 11, 20], highlighting the unique utility of interrogating the SDC genome to help patients potentially identify experimental treatment options beyond AR-targeting.

The field of developmental therapeutics for salivary malignancies in the last decade has made significant strides with greater scientific interest for understanding the biology of these diseases as well as the development of an increasing number of SGC-specific clinical trials. Fushimi et al.’s accomplishment of developing and completing a 36-patient AR+ SGC trial not only provides valuable evidence regarding the efficacy of an AR-targeted approach, but more importantly demonstrates the feasibility of evaluating molecularly targeted therapies in a biomarker selected subset of a rare disease population. This success should only further encourage the development of biologically informed therapeutic trial options for rare cancer patient populations in the current genomic era.

Disclosure

The author has declared no conflicts of interest.