Adenoid cystic carcinoma (ACC) is among the most lethal salivary gland tumors, with no treatments for metastatic disease that prolong survival. We examined tissue from 28 primary and metastatic ACC deposits obtained from 21 patients for infiltrating immune cells and PD-L1/PD-L2 expression and determined mRNA profiles of over 1,400 oncogenic and immune-related genes. We also assessed the effect of chemoradiation on immune mediators in a patient who had serial biopsies available. Most tumors expressed PD-L2 but had few infiltrating immune cells. Lack of immune-cell infiltrate was associated with expression of genes in the β-catenin/Wnt and PI3K pathways. Additionally, certain transcripts linked to growth and invasion were differentially expressed among primary and metastatic deposits. Chemoradiation appeared to increase CD8+ effector T cells, decrease regulatory T cells, and promote a systemic humoral response. These data suggest a potential role for PD-L2 inhibition and immune modulation as treatment for patients with ACC. Cancer Immunol Res; 4(8); 679–87. ©2016 AACR.

Immune checkpoint blockade, and specifically inhibition of the programmed-death-1 (PD-1) pathway, has proven clinical benefit in the treatment of melanoma, non–small cell lung cancer, and renal cell carcinoma, and preliminary studies have demonstrated efficacy in an increasing number of other tumor types (1–11). As the use of these agents expands, and a variety of other immunotherapy agents are developed (12), significant value has been shown in predicting which patients will respond to PD-1 pathway blockade, in contrast to other patients for whom other agents or combination treatment strategies may be more appropriate. Mutational or neoantigen load, PD-ligand 1 (PD-L1) expression, and defects in mismatch repair have all been proposed as potential predictors of treatment response to PD-1 inhibition (4, 13–19). Data have also indicated that patients with CD8+ T-cell infiltrate within or in proximity to the tumor microenvironment may preferentially respond to immunotherapies (18, 20, 21). Thus, profiling the immune microenvironment of individual tumor types can provide a rational basis to guide the clinical evaluation of checkpoint blockade.

Given that immune checkpoint blockade therapy can potentially provide durable disease control, increasing the number of patients that respond to these agents is of critical importance (22). Targeted agents such as trastuzumab and cetuximab can have significant immunologic effects (23). In melanoma, specific oncogenic pathways (i.e., Wnt/β-catenin pathway) have been suggested to inversely correlate with T-cell infiltration and therefore serve as targets for immune potentiation (24); however, it is unknown whether this correlation is relevant in other tumor types. Additionally, traditional cytotoxic chemotherapy and radiation can also generate or potentiate immune responses under the proper circumstances, with some evidence for potential synergy with immune checkpoint blockade (25, 26).

Adenoid cystic carcinoma (ACC) is a malignant salivary gland tumor characterized by a long natural history with a propensity for indolent but relentless growth and dissemination (27). Translocations resulting in fusion of the vMYB gene on chromosome 6 to the NFIB gene on chromosome 9 have been observed in the majority of cases, and appear to contribute to pathogenesis (28, 29). The standard of care for localized disease includes maximal surgical resection with adjuvant radiotherapy with or without chemotherapy in most cases (30–32); however, approximately 50% of patients develop distant metastases, and up to one third die within 2 years of diagnosis (33). There are no proven or effective therapies in the setting of metastatic disease, and chemotherapy specifically provides very little, if any, benefit (34). Given the indolent growth that can, in some cases, progress over many years, local therapies such as palliative radiation are frequently used. Very little is known about endogenous immune responses directed against ACC at baseline, or following conventional treatments such as chemotherapy and/or radiation. Here, we report on our evaluation of ACC specimens for infiltrating immune cells and expression of immune checkpoint ligands PD-L1 and PD-L2. We also determined mRNA expression profiles of over 1,400 oncogenic and immune-related genes to identify genes that may be associated with a paucity of tumor-infiltrating immune cells. In addition, we assessed potential effects of chemoradiation on circulating immune mediators and antitumor humoral responses.

Patient information

Tumor samples were obtained from primary or metastatic lung lesions in 21 patients via biopsy or resection. Seven patients had tissue obtained from both metastatic and primary lesions, 2 patients had tissue that was obtained from more than one metastatic lung lesion, and 1 patient had tissue that was obtained both before and after chemoradiation for unresectable primary disease in the hard palate. All tissue was formalin fixed and paraffin embedded before storage. Additional clinical information, including the location of the primary or metastatic lesion, was abstracted from the clinical record. Peripheral blood and saliva samples were obtained longitudinally from the patient who received chemoradiation with concurrent weekly cisplatin (40 mg/m2) over the course of chemoradiation (at 1 week, 4 weeks, 6–7 weeks, and another 7 weeks after therapy). All samples were obtained/reviewed on Institutional Review Board (IRB)–approved or IRB-exempt protocols.

Immunohistochemistry

Antibodies to PD-L1 (clone 405.9A11), PD-1 (clone EH33), and PD-L2 (clone 366C.9E5) are mouse monoclonal antibodies obtained from Gordon Freeman (Dana-Farber Cancer Institute). 405.9A11 recognizes an epitope in the PD-L1 cytoplasmic domain, and reactivity confirms the expression of full-length PD-L1 protein. 366C.9E5 recognizes an epitope in the IgV domain of PD-L2. PD-1, PD-L2, and PD-L1 staining were performed as previously described (35–37). PD-L1 staining was observed in cytoplasm and membranes of tumor cells and was considered positive if ≥5% of tumor cells had membranous staining. As in previous studies, PD-1 was considered positive if ≥5% of the cells visualized displayed PD-1 staining (38–40). We also quantified the number of specimens with tumor-infiltrating immune cells with any degree of positive membranous PD-L1 staining. PD-L2 staining (cytoplasmic and membranous) on tumor cells was scored based on the percentage of cells with positive staining in conjunction with staining intensity. The intensity was scored as follows: 0, negative; 1, weak; 2, moderate; 3, strong. The cutoff of PD-L2 positivity in tumor cells on whole-tissue slides was ≥10% with 1+, 2+, and 3+ staining intensity. Slides indicating >100 distinct immune cells per 200-magnification high-power field were considered positive for significant intratumoral immune infiltration. Florescence in situ hybridization (FISH) was used to screen ACC tissues for the MYBNFIB translocation as previously described (41). In the patient with samples obtained before and after radiotherapy, the tissue was also stained for CD4+, CD8+, and FOXP3+ cells, with 3 representative sections counted under a ×200 high-powered field (HPF).

RNA extraction

Up to 3 unstained 5-μm sections of tissue were used to scrape the tumor region of interest identified from H&E annotated by a pathologist. All scrapings for a given case were pooled and the Agencourt Formapure kit for isolation of nucleic acids from formalin-fixed paraffin-embedded tissue (cat #A33342) was implemented on a Biomek FXP Laboratory Automation Workstation (Dual-arm system with multichannel pipette and span-8 pipetters; cat #A31844). The automated Formapure protocol is based upon Solid Phase Reversible Immobilization (SPRI) paramagnetic bead-based technology, which does not require vacuum filtration, centrifugation, or organic solvents such as phenol or xylene associated with traditional methods. RNA isolates were eluted in a 55-μL volume of RNase/DNase-free H2O.

RNA isolates were quantified using the Quant-iT RiboGreen assay (Life Technologies; cat# R11490). RNA (1 μL) was required for quantification. Concentration was measured as ng/μL. For RNA quantification, isolates were excited at 485 ± 10 nm, and the fluorescence emission intensity was measured at 530 ± 12 nm using a Victor X3 spectrophotometer (Perkin Elmer; cat# 2030-0030). Fluorescence intensity was plotted versus RNA concentration over the calibration range 0–100 ng/μL.

Oncogenic and immune mRNA profiling

All procedures related to mRNA quantification, including sample preparation, hybridization, detection, and scanning, were carried out as recommended by nanoString Technologies for both the PanCancer and PanCancer Immune Profiling codesets (Nanostring Technologies). Raw nanoString counts for each gene within each experiment were subjected to a technical normalization using the counts obtained for positive control probe sets prior to a biological normalization using housekeeping genes included in the CodeSet. Normalized data were log2-transformed and then used as input for differential expression clustering in response to selected covariates (immune infiltration, PD-L1 IHC status, etc.) and for heat maps. For the PanCancer Immune codeset, genes previously shown to be characteristic of various immune cell populations were used to measure the abundance of these populations. nSolver Software (Nanostring Technologies) was used for analysis.

Flow cytometry

We isolated peripheral blood mononuclear cells (PBMC) via centrifugation (1,500 × g, 20 minutes) and stored the PBMCs in freezing media (10% FBS RPMI + 10% DMSO) at −80°C on the same day as the blood draw. Flow cytometry was performed to quantify activated/cytotoxic T cells (CD3-PC7+, CD8-APC+, CD69-PE+), regulatory T cells (CD4-FITC+, CD25-PE+, CD127-APClow), and T cells expressing immune checkpoint receptors (CD8-APC+, LAG3-Per710+/TIM3-Bv421+/PD1–PE+) via established protocols. All antibodies were obtained from eBioscience except for CD8-APC (Miltenyi Biotec Inc.) and TIM3-Bv421 (Biolegend Inc.). FlowJo was used for analysis.

Seromics

Proteomic analysis was performed with isolated patient serum or saliva using ProtoArray Immune Response Biomarker Profiling (ThermoFisher Scientific) using the manufacturer's protocols to detect the presence of IgG antibodies directed against potential tumor antigens before and after chemoradiotherapy. As suggested by the manufacturer, candidate antibodies were considered to be significantly increased if signal intensity >3,000 relative fluorescence units (RFU), the signal-to-noise ratio was >1.5, and the fold change post-radiotherapy (RT) to pre-RT was >2. Additionally, antibodies with RFU >65,000, which approached the upper boundary of the dynamic range as reported by the manufacturer, were considered to have the greatest interaction with candidate antigens.

Statistical analysis

Two-sided P values < 0.05 were considered statistically significant. All statistical analyses were computed using GraphPad Prism (GraphPad Software Inc.), STATA 14.0 (StataCorp), or JMP Pro12 (SAS Institute Inc.), or as previously specified.

Immune microenvironment

Tissue from primary and metastatic ACC deposits were used to profile intratumoral infiltrating immune cells, PD-L1 expression on both tumor and immune cells, and PD-L2 expression on tumor cells, both cytoplasmic and membranous (Fig. 1). No patients had significant expression of PD-L1 on tumor cells. In contrast, PD-L2 was expressed on 60% of the primary (9/15) and 73% (8/11) of metastatic tumor samples. Forty-two percent of both primary and metastatic deposits contained many infiltrating immune cells (>100/HPF). Among the patients with significant immune infiltration, all but one had PD-L1 expression noted on the infiltrating tumor cells. PD-1 expression was also examined on tumor deposits and on infiltrating immune cells. As expected, PD-1 expression was absent on tumor cells (0/13 primary, 0/8 metastatic deposits screened). In contrast, 50% (7/14) of primary and 38% (3/8) of metastatic deposits showed PD-1 positivity in infiltrating immune cells. The expression levels of PD-1 accounted for ≤10% of tumor stroma in all cases.

Figure 1.

IHC indicating percentages of primary and metastatic tumor deposits with PD-L2 surface expression on tumor cells (left), and an example of a deposit with positive PD-L2 staining (right). Less than half of primary and metastatic tumor deposits showed substantial intratumoral immune infiltration as defined by greater than 100 distinct immune cells per HPF (middle). The majority of deposits with substantial infiltrating immune cells also expressed PD-L1 on immune cells (inset).

Figure 1.

IHC indicating percentages of primary and metastatic tumor deposits with PD-L2 surface expression on tumor cells (left), and an example of a deposit with positive PD-L2 staining (right). Less than half of primary and metastatic tumor deposits showed substantial intratumoral immune infiltration as defined by greater than 100 distinct immune cells per HPF (middle). The majority of deposits with substantial infiltrating immune cells also expressed PD-L1 on immune cells (inset).

Close modal

Immune mRNA profiling

In 12 cases for which enough archival tumor tissue was available, RNA was extracted from primary and metastatic lesions and assayed via the PanCancer Immune Profiling Panel (Nanostring) to assess expression levels of 770 immune-related genes. The resulting expression profiles (Fig. 2A) show distinct clustering of tumors with and without intratumoral immune infiltrate observed on IHC, consistent with increased expression of known immune-related genes in tumors with more infiltrating immune cells. In order to potentially discern which immune cell types were present within these tumors, samples were clustered by the presence or absence of immune cells (as determined by IHC) and combined with groups of mRNA expression markers that have been shown to correlate with distinct immune populations as per the manufacturer's protocol (Fig. 2B). When normalized and compared with tumors without significant immune cells on IHC, those with immune infiltrates showed 62% more B-cell–associated mRNA transcripts, 38% more regulatory T-cell transcripts, 30% more NK-associated transcripts, 25% more CD8+ T-cell transcripts, and 12% more activated CD4 T-cell transcripts. However, if the tumors were instead stratified by PD-L2 expression on IHC, all the tumors with higher PD-L2 expression indicated a consistent decrease in immune infiltrates (Fig. 2C). Specifically, changes included a 33% drop in mRNA transcripts identifying NK-cell abundance, a 16% drop in activated CD4 T-cell–associated transcripts, and a 17% drop in regulatory T cells.

Figure 2.

Immune gene signatures associated with tumor infiltration. A, heat map indicating differential mRNA expression profiles of immune-related genes in samples with or without immune infiltration as observed on IHC. B, clusters of RNA transcripts associated with immune subpopulations were compared among tumor deposits harboring a more pronounced immune infiltrate and those with few infiltrating immune cells. C, clusters of RNA transcripts associated with immune subpopulations were compared among tumor deposits with significant surface expression of PD-L2 by IHC and those without PD-L2 expression.

Figure 2.

Immune gene signatures associated with tumor infiltration. A, heat map indicating differential mRNA expression profiles of immune-related genes in samples with or without immune infiltration as observed on IHC. B, clusters of RNA transcripts associated with immune subpopulations were compared among tumor deposits harboring a more pronounced immune infiltrate and those with few infiltrating immune cells. C, clusters of RNA transcripts associated with immune subpopulations were compared among tumor deposits with significant surface expression of PD-L2 by IHC and those without PD-L2 expression.

Close modal

Oncogenic mRNA profiling

Extracted RNA was also assayed with the nCounter PanCancer Pathways Panel (Nanostring) to assess expression of an additional 700 essential genes representing all major cancer pathways. Using probes of exons upstream (MYB exons 1–2, 3–4, 5, 6; NFIB exons 2, 3–4, 5–6, 7) and downstream (MYB exons 15, 16; NFIB exon 12) of the putative translocation sites for MYB and NFIB, we were in parallel able to determine that 58% of primary tumors harbored a MYB–NFIB translocation. The presence of MYB translocations was detected by an imbalance of mRNA expression between upstream and downstream MYB exons; similar upstream and downstream expression indicated no translocation and a full-length gene expression. These results were consistent between 3 of 3 matched primary–metastatic pairs, and in 2 of 2 cases for which FISH results were also available. When examining the genes with the highest expression (top 10%), in at least half of all tumor samples, we identified overexpression of the PI3K pathway genes, as well as overexpression of genes in the cell cycle, MAPK, and Wnt pathways (Fig. 3). Specifically, our screen identified several genes known to be overexpressed in ACC mouse xenograph models, including MYB, JAG1, FGFR1, and NOTCH1 (42).

Figure 3.

mRNA expression imbalance among MYB and NFIB exons was used to determine rates of MYB–NFIB translocations. A, the majority of samples indicate a MYBNFIB fusion. B, mRNA transcripts belonging to pathways known to be implicated in ACC are overexpressed in the majority of tumor samples. Transc, transcripts.

Figure 3.

mRNA expression imbalance among MYB and NFIB exons was used to determine rates of MYB–NFIB translocations. A, the majority of samples indicate a MYBNFIB fusion. B, mRNA transcripts belonging to pathways known to be implicated in ACC are overexpressed in the majority of tumor samples. Transc, transcripts.

Close modal

We then stratified tumors by the presence of significant intratumoral immune infiltrate on histology and noted differential expression of these 700 genes between these two groups (Fig. 4A). Our screen identified several genes in the Wnt and PI3K pathways, with overexpression that potentially correlated with reduced immune infiltrate, although without correction for multiple testing (Fig. 4B; Supplementary Table S1). For example, increased expression of the genes for BAMBI (BMP and activin membrane-bound inhibitor, P = 0.02), β-catenin (P = 0.02), FGF17 (P = 0.004), and Bcl-2 (P = 0.01) all correlated with absent or near-absent immune infiltrates across the ACC samples. Conversely, our analysis also identified genes with expression that was correlated with immune infiltration, including Syk (P = 0.04), IL2RB (P = 0.02), and TGFβ (P = 0.02). Higher expression of these immune-related genes is likely related to the infiltrating immune cells themselves and is therefore expected to correlate with higher immune infiltrates on IHC. When the tumors were instead stratified as primary or metastatic, certain genes with protein products linked to cell signaling, including CACNA2D2 (P < 0.0001), BMP5 (P < 0.0001), FGFR4 (P = 0.0003), LAMC3 (P = 0.0001), and others, were more highly expressed in metastatic than primary tumors.

Figure 4.

Genes associated with immune cell infiltration of tumors. A, volcano plot indicates expression of genes that correlate with intratumoral immune invasion as determined by IHC. B, select transcripts associated with lack of immune infiltrate on IHC. Transcripts are shown in relation to the percentage of samples without substantial immune infiltrate in which overexpression was noted as compared with infiltrated samples.

Figure 4.

Genes associated with immune cell infiltration of tumors. A, volcano plot indicates expression of genes that correlate with intratumoral immune invasion as determined by IHC. B, select transcripts associated with lack of immune infiltrate on IHC. Transcripts are shown in relation to the percentage of samples without substantial immune infiltrate in which overexpression was noted as compared with infiltrated samples.

Close modal

Impact of chemoradiation

In one patient with ACC, tumor samples were obtained before and after chemoradiation, and peripheral blood samples were obtained over the course of therapy and at follow-up. Fractionated radiotherapy to the tumor appeared to increase the number of intratumoral CD8+ T cells and decrease the number of FOXP3+ regulatory T cells (Fig. 5A). The percentage of CD8+ effector T cells appeared to increase and regulatory T cells decreased in peripheral circulation, which then returned to baseline numbers following the completion of therapy (Fig. 5B). The number of CD8+ cells expressing LAG-3, TIM-3, or PD-1 also increased in the peripheral blood over the course of therapy and returned to baseline at the follow-up time point (Fig. 5C).

Figure 5.

Modulation of immune cell phenotypes and subsets by chemoradiation therapy. A, 7 weeks of chemoradiation appears to increase CD8+ T cells and decrease regulatory T cells within the tumor microenvironment. B, percentage of CD8+ cells and regulatory T cells in the peripheral circulation. Changes over the course of chemoradiation appear to mirror the effects seen within the tumor. C, percentage of CD8+ T cells expressing checkpoint receptors Lag-3, Tim-3, and PD-1 over the course of chemoradiation. Percentages appear to rise through the course of therapy and then return closer to baseline levels during follow-up (FU). D and E, proteomic arrays were used to quantify targeting of specific proteins by antibodies present in the serum before as compared with after chemoradiation. Statistically significant increases in signal were produced by antibodies targeting multiple protein targets after as compared with before radiation. F, increases in the number of antibody targets yielding the highest signals were also shown (>60,000 RFU) on the array.

Figure 5.

Modulation of immune cell phenotypes and subsets by chemoradiation therapy. A, 7 weeks of chemoradiation appears to increase CD8+ T cells and decrease regulatory T cells within the tumor microenvironment. B, percentage of CD8+ cells and regulatory T cells in the peripheral circulation. Changes over the course of chemoradiation appear to mirror the effects seen within the tumor. C, percentage of CD8+ T cells expressing checkpoint receptors Lag-3, Tim-3, and PD-1 over the course of chemoradiation. Percentages appear to rise through the course of therapy and then return closer to baseline levels during follow-up (FU). D and E, proteomic arrays were used to quantify targeting of specific proteins by antibodies present in the serum before as compared with after chemoradiation. Statistically significant increases in signal were produced by antibodies targeting multiple protein targets after as compared with before radiation. F, increases in the number of antibody targets yielding the highest signals were also shown (>60,000 RFU) on the array.

Close modal

Seromic analyses of this patient's serum and saliva examined the impact of treatment on humoral immunity. We identified the protein targets of antibodies for which signal increased substantially after treatment (Supplementary Table S2). Many more antibodies robustly increased after therapy than before, with antibodies in the blood that targeted 259 proteins robustly increased, as compared with decreases in antibodies targeting nine other protein targets (P < 0.0001, binomial test; Fig. 5D and E). Although the significance of these protein targets is unknown, antibodies directed against five of the protein targets also increased in the saliva following chemoradiation. In addition, the number of high-titer antibodies against self-proteins (defined as RFU > 60,000) detectable in serum was higher (38) after treatment than before (ref. 9; Fig. 5F).

ACC is a relatively rare tumor with a poor prognosis for which there were no previous data, to our knowledge, on immune infiltration, expression of checkpoint ligands, and the interplay between these factors, cell-signaling pathways, and conventional treatments. Here, we profiled a variety of primary and metastatic tumor deposits. We also conducted an in-depth analysis evaluating the effect of chemoradiation on an unresectable tumor to evaluate immune effects in the local tumor environment as well as more distantly.

We found multiple potential mechanisms of immune suppression that may be relevant in this tumor type, including PD-L2 expression in tumor cells (both cytoplasmic and membranous), most tumors containing a relatively sparse immune infiltrate, and PD-L1 expression on infiltrating lymphocytes when they are present. PD-1 expression was found in the immune cells from a significant percentage of tumor deposits. We also found that chemoradiation may affect the tumor microenvironment, leading to an increased number of infiltrating CD8+ T cells and a decreased number of infiltrating regulatory T cells. These effects present in the tumor microenvironment were mirrored in the peripheral blood. However, these changes were also observed in the context of increased levels of circulating CD8+ T cells expressing the checkpoint receptors Lag-3, Tim-3, and PD-1, thus potentially representing dysfunctional, or “exhausted” T-cell populations (43, 44). However, because these markers are also expressed on activated T cells, additional investigations of the functional activities of these cells are required to help us understand more fully their potential role in antitumor immunity. Chemoradiation also had a demonstrable effect on systemic humoral immunity, with increased targeting of specific proteins, as inferred from the presence of specific antibodies in the circulation and in serially collected saliva specimens. Radiotherapy may promote antitumor immunity via a humoral response, as has been observed in a previous study of patients with prostate cancer undergoing radiotherapy, as well as in a patient with melanoma irradiated for progressive disease in the setting of CTLA-4 blockade (25, 45). Although it is currently unclear whether the antibody responses we detect are generated in response to tumor neoantigens, the potential for monitoring radiation-induced changes in salivary antibodies is intriguing given the potential accessibility for serial monitoring.

In several samples in which enough archival tumor tissue was available after IHC, we used mRNA profiling to interrogate the immune microenvironment in more detail and also to identify candidate oncogenic pathways that may be linked to intratumoral immune infiltration. Distinct gene expression profiles suggested various immune populations present intratumorally. When compared with tumors without immune infiltration on IHC, tumor deposits with immune cells indicated significantly higher expression of B-cell–, regulatory T-cell–, CD8+ T-cell–, and NK-cell–associated mRNA transcripts. Conversely, when we stratified tumors by PD-L2 IHC status, positive staining for PD-L2 expression was associated with decreased mRNA expression indicating immune infiltration across all cell types, including NK cells and CD4+ and CD8+ T cells. We were unable to correlate with PD-L2 mRNA expression as it was not possible to disambiguate signal from tumor cells compared with immune cells.

ACC is characterized by MYB overexpression and the presence of MYB–NFIB translocations, which have been well described (27, 28, 41). Here, we used a relatively novel technique that employed the imbalance of signal obtained from upstream and downstream probes to assess translocation status in parallel with RNA profiling of an additional 700 oncogenic genes. The translocation results were consistent between matched primary and metastatic deposits, as well as with FISH data when available. mRNA profiling indicated the overexpression of several genes within the PI3K, MAPK, Wnt, and Notch pathways. Specifically, our screen detected several genes known to be overexpressed or mutated in ACC, including MYB, P53, and NOTCH1 (46). Many of the genes detected as overexpressed in our samples (i.e., FGFR1, JAG1, and BAMBI) have also been detected in salivary gland tumor xenograft data (University of Virginia ACC database). Our results also complement the findings of ACC whole-exome sequencing studies, which have identified recurrent mutations in the FGF/IGF/PI3K pathway, which had the most overexpressed genes in our study and may contribute to the aberrant overexpression (47, 48).

mRNA profiling performed in conjunction with immunohistochemical analysis allowed us to investigate pathways and individual genes that may be linked to the presence or absence of intratumoral immune infiltrates, which may, in turn, be linked to responsiveness to immune checkpoint blockade (24). When the gene expression was stratified by tumor infiltration on IHC, overexpression of several genes was correlated with poor immune infiltration. Members of the PI3K (i.e., FGF17, Bcl-2), and Wnt pathways (β-catenin, BAMBI) represented the majority of genes with expression that was correlated with poorer immune infiltrates. Although these hypothesis-generating results are not adjusted for multiple testing, they are consistent with a recent analysis showing activation of the WNT/β-catenin signaling pathway correlating with the absence of a T-cell gene expression signature in melanoma (24). Future studies are needed to better elucidate the mechanisms that potentially connect PI3K or Wnt pathway activation to poorer intratumoral immune infiltration and pathogenesis.

Although PD-L1 is expressed in multiple tumor types (49), comparatively less is known about the frequency and role of PD-L2 expression in modulating antitumor immune responses. Like PD-L1, PD-L2 also binds to the PD-1 receptor and decreases T-cell proliferation and cytokine production (50). PD-L2 can bind to the RGMb receptor on the surface of macrophages and other immune cells to promote immune tolerance, although the significance of this observation in cancer is unclear (51). PD-L2 expression has been noted in esophageal adenocarcinoma, breast cancer, and renal cell cancer (35, 52–54). The significance of PD-L2 expression on epithelial cells is less explored, compared with PD-L1 expression. Akin to PD-L1, PD-L2 is also implicated in dampening T-cell responsiveness and prompting immune tolerance. Stimulation via IFNγ enhanced PD-L2 expression, similar to what is observed for PD-L1 (55). However, given that we observed PD-L2 expression in the absence of PD-L1, our data suggest that IFN-induced expression may not be the dominant mechanism relevant to ACC. Therefore, other potential mechanisms, such as genomic changes or induction via Th2 signaling pathways, should be explored, especially given our finding that the most common ACC infiltrating lymphocytes were B cells, suggestive of a Th2 response. Regardless, because we identified PD-L1 expression on infiltrating immune cells in certain tumor deposits and PD-L2 expression on tumor cells in other deposits, PD-1 inhibitors may represent a better choice of therapy in patients with ACC as compared with PD-L1 inhibitors to stimulate an effective antitumor immune response.

Given the rarity of ACC and risks associated with tissue collection, we were limited in the amount of tumor tissue that could be evaluated for various immune markers and mRNA expression, despite retrieving samples collected over many years. However, even with relatively few patients, we could profile the immunologic microenvironment in both primary and matched metastatic ACC, study potential correlates to immune infiltration, and assess potential immunologic effects of chemoradiation. Our results suggest that PD-L2 expression may be common in ACC, and immune infiltration may be associated with upregulation of specific signaling pathways such as β-catenin/Wnt previously described in other tumor types. Chemoradiation could potentiate antitumor immune responses, as has been suggested in previous studies across a variety of tumor types (56). Collectively, these data suggest that PD-1 inhibition, potentially in combination with conventional treatments, could be a promising strategy for patients with ACC, a population that currently has few effective treatment options.

R.I. Haddad is a consultant/advisory board member for the Adenoid Cystic Carcinoma Research Foundation. G.J. Freeman has an ownership interest (including patents) in Novartis, Roche, Bristol-Myers Squibb, Merck, Amplimmune, EMD Serono, and Boehringer-Ingelheim and is a consultant/advisory board member for Novartis, Roche, Bristol-Myers Squibb, Seattle Genetics, Eli Lilly, Bethyl Labs, and Surface Oncology. F. Stephen Hodi reports receiving commercial research support from Bristol-Myers Squibb to his institution and is a consultant/advisory board member for Genentech, Synta, and Celldex. S.J. Rodig reports receiving commercial research support from Bristol-Myers Squibb, has received speakers bureau honoraria from Bristol-Myers Squibb, and is a consultant/advisory board member for Perkin Elmer. G. Dranoff reports receiving commercial research support from Bristol-Myers Squibb. No potential conflicts of interest were disclosed by the other authors.

Conception and design: V. Sridharan, F.S. Hodi, S.J. Rodig, J.D. Schoenfeld

Development of methodology: V. Sridharan, X. Liao, M. Severgnini, G.J. Freeman, S.J. Rodig, J.D. Schoenfeld

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): V. Sridharan, X. Liao, N.G. Chau, R.I. Haddad, M. Severgnini, F.S. Hodi, J.D. Schoenfeld

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): V. Sridharan, E. Gjini, X. Liao, R.I. Haddad, F.S. Hodi, S.J. Rodig, J.D. Schoenfeld

Writing, review, and/or revision of the manuscript: V. Sridharan, X. Liao, N.G. Chau, R.I. Haddad, P. Hammerman, A. El-Naggar, G.J. Freeman, F.S. Hodi, S.J. Rodig, G. Dranoff, J.D. Schoenfeld

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R.I. Haddad, A. El-Naggar, J.D. Schoenfeld

Study supervision: R.I. Haddad, S.J. Rodig, J.D. Schoenfeld

This work was supported by the Adenoid Cystic Cancer Research Foundation (ACCRF). We acknowledge assistance provided by the Center for Immuno-Oncology at the Dana-Farber Cancer Institute. We acknowledge assistance provided by the Center for Molecular Oncologic Pathology at the Dana-Farber Cancer Institute, specifically by Michaela Bowden, Meredith Smith, and Chensheng Zhou. We acknowledge assistance provided by Nanostring Inc., specifically Kit Fuhrman.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Brahmer
JR
,
Tykodi
SS
,
Chow
LQM
,
Hwu
W-J
,
Topalian
SL
,
Hwu
P
, et al
Safety and activity of anti-PD-L1 antibody in patients with advanced cancer
.
N Engl J Med
2012
;
366
:
2455
65
.
2.
Hodi
FS
,
O'Day
SJ
,
McDermott
DF
,
Weber
RW
,
Sosman
JA
,
Haanen
JB
, et al
Improved survival with ipilimumab in patients with metastatic melanoma
.
N Engl J Med
2010
;
363
:
711
23
.
3.
Robert
C
,
Thomas
L
,
Bondarenko
I
,
O'Day
S
,
Weber
J
,
Garbe
C
, et al
Ipilimumab plus dacarbazine for previously untreated metastatic melanoma
.
N Engl J Med
2011
;
364
:
2517
26
.
4.
Topalian
SL
,
Hodi
FS
,
Brahmer
JR
,
Gettinger
SN
,
Smith
DC
,
McDermott
DF
, et al
Safety, activity, and immune correlates of anti-PD-1 antibody in cancer
.
N Engl J Med
2012
;
366
:
2443
54
.
5.
Schoenfeld
JD
. 
Immunity in head and neck cancer
.
Cancer Immunol Res
2015
;
3
:
12
7
.
6.
Borghaei
H
,
Paz-Ares
L
,
Horn
L
,
Spigel
DR
,
Steins
M
,
Ready
NE
, et al
Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung cancer
.
N Engl J Med
2015
;
373
:
1627
39
.
7.
Motzer
RJ
,
Escudier
B
,
McDermott
DF
,
George
S
,
Hammers
HJ
,
Srinivas
S
, et al
Nivolumab versus everolimus in advanced renal-cell carcinoma
.
N Engl J Med
2015
;
373
:
1803
13
.
8.
Seiwert
TY
,
Burtness
B
,
Weiss
J
,
Gluck
I
,
Eder
JP
,
Pai
SI
, et al
A phase Ib study of MK-3475 in patients with human papillomavirus (HPV)-associated and non-HPV-associated head and neck (H/N) cancer
.
J Clin Oncol
2014
;
32
:
5s
(
suppl; abstr 6011
).
9.
Segal
NH
,
Antonia
SJ
,
Brahmer
JR
,
Maio
M
,
Blake-Haskins
A
,
Li
X
, et al
Preliminary data from a multi-arm expansion study of MEDI4736, an anti-PD-L1 antibody
.
J Clin Oncol
2014
;
32
:
5s
(
suppl; abstr 3002⁁
).
10.
Robert
C
,
Schachter
J
,
Long
GV
,
Arance
A
,
Grob
JJ
,
Mortier
L
, et al
Pembrolizumab versus ipilimumab in advanced melanoma
.
N Engl J Med
2015
;
372
:
2521
32
.
11.
Larkin
J
,
Hodi
FS
,
Wolchok
JD
. 
Combined nivolumab and ipilimumab or monotherapy in untreated melanoma
.
N Engl J Med
2015
;
373
:
1270
1
.
12.
Mellman
I
,
Coukos
G
,
Dranoff
G
. 
Cancer immunotherapy comes of age
.
Nature
2011
;
480
:
480
9
.
13.
Rizvi
NA
,
Hellmann
MD
,
Snyder
A
,
Kvistborg
P
,
Makarov
V
,
Havel
JJ
, et al
Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer
.
Science
2015
;
348
:
124
8
.
14.
Le
DT
,
Uram
JN
,
Wang
H
,
Bartlett
BR
,
Kemberling
H
,
Eyring
AD
, et al
PD-1 blockade in tumors with mismatch-repair deficiency
.
N Engl J Med
2015
;
372
:
2509
20
.
15.
Wang
L
,
Wang
H
,
Chen
H
,
Wang
WD
,
Chen
XQ
,
Geng
QR
, et al
Serum levels of soluble programmed death ligand 1 predict treatment response and progression free survival in multiple myeloma
.
Oncotarget
2015
;
6
:
41228
36
.
16.
Van Allen
EM
,
Miao
D
,
Schilling
B
,
Shukla
SA
,
Blank
C
,
Zimmer
L
, et al
Genomic correlates of response to CTLA-4 blockade in metastatic melanoma
.
Science
2015
;
350
:
207
11
.
17.
Meng
X
,
Huang
Z
,
Teng
F
,
Xing
L
,
Yu
J
. 
Predictive biomarkers in PD-1/PD-L1 checkpoint blockade immunotherapy
.
Cancer Treat Rev
2015
;
41
:
868
76
.
18.
Herbst
RS
,
Soria
JC
,
Kowanetz
M
,
Fine
GD
,
Hamid
O
,
Gordon
MS
, et al
Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients
.
Nature
2014
;
515
:
563
7
.
19.
Garon
EB
,
Rizvi
NA
,
Hui
R
,
Leighl
N
,
Balmanoukian
AS
,
Eder
JP
, et al
Pembrolizumab for the treatment of non-small-cell lung cancer
.
N Engl J Med
2015
;
372
:
2018
28
.
20.
Ji
RR
,
Chasalow
SD
,
Wang
L
,
Hamid
O
,
Schmidt
H
,
Cogswell
J
, et al
An immune-active tumor microenvironment favors clinical response to ipilimumab
.
Cancer Immunol Immunother
2012
;
61
:
1019
31
.
21.
Tumeh
PC
,
Harview
CL
,
Yearley
JH
,
Shintaku
IP
,
Taylor
EJ
,
Robert
L
, et al
PD-1 blockade induces responses by inhibiting adaptive immune resistance
.
Nature
2014
;
515
:
568
71
.
22.
Schadendorf
D
,
Hodi
FS
,
Robert
C
,
Weber
JS
,
Margolin
K
,
Hamid
O
, et al
Pooled analysis of long-term survival data from phase II and phase III trials of ipilimumab in unresectable or metastatic melanoma
.
J Clin Oncol
2015
;
33
:
1889
94
.
23.
Vanneman
M
,
Dranoff
G
. 
Combining immunotherapy and targeted therapies in cancer treatment
.
Nat Rev Cancer
2012
;
12
:
237
51
.
24.
Spranger
S
,
Bao
R
,
Gajewski
TF
. 
Melanoma-intrinsic beta-catenin signalling prevents anti-tumour immunity
.
Nature
2015
;
523
:
231
5
.
25.
Postow
MA
,
Callahan
MK
,
Barker
CA
,
Yamada
Y
,
Yuan
J
,
Kitano
S
, et al
Immunologic correlates of the abscopal effect in a patient with melanoma
.
N Engl J Med
2012
;
366
:
925
31
.
26.
Chandra
RA
,
Wilhite
TJ
,
Balboni
TA
,
Alexander
BM
,
Spektor
A
,
Ott
PA
, et al
A systematic evaluation of abscopal responses following radiotherapy in patients with metastatic melanoma treated with ipilimumab
.
Oncoimmunology
2015
;
4
:
e1046028
.
27.
Mitani
Y
,
Li
J
,
Rao
PH
,
Zhao
YJ
,
Bell
D
,
Lippman
SM
, et al
Comprehensive analysis of the MYB-NFIB gene fusion in salivary adenoid cystic carcinoma: incidence, variability, and clinicopathologic significance
.
Clin Cancer Res
2010
;
16
:
4722
31
.
28.
Mitani
Y
,
Rao
PH
,
Futreal
PA
,
Roberts
DB
,
Stephens
PJ
,
Zhao
YJ
, et al
Novel chromosomal rearrangements and break points at the t(6;9) in salivary adenoid cystic carcinoma: association with MYB-NFIB chimeric fusion, MYB expression, and clinical outcome
.
Clin Cancer Res
2011
;
17
:
7003
14
.
29.
Persson
M
,
Andren
Y
,
Mark
J
,
Horlings
HM
,
Persson
F
,
Stenman
G
. 
Recurrent fusion of MYB and NFIB transcription factor genes in carcinomas of the breast and head and neck
.
Proc Natl Acad Sci U S A
2009
;
106
:
18740
4
.
30.
Meyers
M
,
Granger
B
,
Herman
P
,
Janot
F
,
Garrel
R
,
Fakhry
N
, et al
Head and neck adenoid cystic carcinoma: a prospective multicenter REFCOR study of 95 cases
.
Eur Ann Otorhinolaryngol Head Neck Dis
2016
;
133
:
13
7
.
31.
Bjorndal
K
,
Krogdahl
A
,
Therkildsen
MH
,
Charabi
B
,
Kristensen
CA
,
Andersen
E
, et al
Salivary adenoid cystic carcinoma in Denmark 1990–2005: Outcome and independent prognostic factors including the benefit of radiotherapy. Results of the Danish Head and Neck Cancer Group (DAHANCA)
.
Oral Oncol
2015
;
51
:
1138
42
.
32.
Chae
YK
,
Chung
SY
,
Davis
AA
,
Carneiro
BA
,
Chandra
S
,
Kaplan
J
, et al
Adenoid cystic carcinoma: current therapy and potential therapeutic advances based on genomic profiling
.
Oncotarget
2015
;
6
:
37117
34
.
33.
Liu
J
,
Shao
C
,
Tan
ML
,
Mu
D
,
Ferris
RL
,
Ha
PK
. 
Molecular biology of adenoid cystic carcinoma
.
Head Neck
2012
;
34
:
1665
77
.
34.
Laurie
SA
,
Ho
AL
,
Fury
MG
,
Sherman
E
,
Pfister
DG
. 
Systemic therapy in the management of metastatic or locally recurrent adenoid cystic carcinoma of the salivary glands: a systematic review
.
Lancet Oncol
2011
;
12
:
815
24
.
35.
Derks
S
,
Nason
KS
,
Liao
X
,
Stachler
MD
,
Liu
KX
,
Liu
JB
, et al
Epithelial PD-L2 expression marks Barrett's esophagus and esophageal adenocarcinoma
.
Cancer Immunol Res
2015
;
3
:
1123
9
.
36.
Chen
BJ
,
Chapuy
B
,
Ouyang
J
,
Sun
HH
,
Roemer
MG
,
Xu
ML
, et al
PD-L1 expression is characteristic of a subset of aggressive B-cell lymphomas and virus-associated malignancies
.
Clin Cancer Res
2013
;
19
:
3462
73
.
37.
Shi
M
,
Roemer
MG
,
Chapuy
B
,
Liao
X
,
Sun
H
,
Pinkus
GS
, et al
Expression of programmed cell death 1 ligand 2 (PD-L2) is a distinguishing feature of primary mediastinal (thymic) large B-cell lymphoma and associated with PDCD1LG2 copy gain
.
Am J Surg Pathol
2014
;
38
:
1715
23
.
38.
Taube
JM
,
Klein
A
,
Brahmer
JR
,
Xu
H
,
Pan
X
,
Kim
JH
, et al
Association of PD-1, PD-1 ligands, and other features of the tumor immune microenvironment with response to anti-PD-1 therapy
.
Clin Cancer Res
2014
;
20
:
5064
74
.
39.
Schmidt
LH
,
Kummel
A
,
Gorlich
D
,
Mohr
M
,
Brockling
S
,
Mikesch
JH
, et al
PD-1 and PD-L1 expression in NSCLC indicate a favorable prognosis in defined subgroups
.
PLoS One
2015
;
10
:
e0136023
.
40.
D'Incecco
A
,
Andreozzi
M
,
Ludovini
V
,
Rossi
E
,
Capodanno
A
,
Landi
L
, et al
PD-1 and PD-L1 expression in molecularly selected non-small-cell lung cancer patients
.
Br J Cancer
2015
;
112
:
95
102
.
41.
Mitani
Y
,
Liu
B
,
Rao
P
,
Borra
V
,
Zafereo
M
,
Weber
RS
, et al
Novel MYBL1 gene rearrangements with recurrent MYBL1-NFIB fusions in salivary adenoid cystic carcinomas lacking t(6;9) translocations
.
Clin Cancer Res
2016
;
22
:
725
33
.
42.
Moskaluk
CA
,
Baras
AS
,
Mancuso
SA
,
Fan
H
,
Davidson
RJ
,
Dirks
DC
, et al
Development and characterization of xenograft model systems for adenoid cystic carcinoma
.
Lab Invest
2011
;
91
:
1480
90
.
43.
Wherry
EJ
,
Kurachi
M
. 
Molecular and cellular insights into T cell exhaustion
.
Nat Rev Immunol
2015
;
15
:
486
99
.
44.
Nirschl
CJ
,
Drake
CG
. 
Molecular pathways: coexpression of immune checkpoint molecules: signaling pathways and implications for cancer immunotherapy
.
Clin Cancer Res
2013
;
19
:
4917
24
.
45.
Nesslinger
NJ
,
Sahota
RA
,
Stone
B
,
Johnson
K
,
Chima
N
,
King
C
, et al
Standard treatments induce antigen-specific immune responses in prostate cancer
.
Clin Cancer Res
2007
;
13
:
1493
502
.
46.
Bell
D
,
Hanna
EY
,
Miele
L
,
Roberts
D
,
Weber
RS
,
El-Naggar
AK
. 
Expression and significance of notch signaling pathway in salivary adenoid cystic carcinoma
.
Ann Diagn Pathol
2014
;
18
:
10
3
.
47.
Ho
AS
,
Kannan
K
,
Roy
DM
,
Morris
LG
,
Ganly
I
,
Katabi
N
, et al
The mutational landscape of adenoid cystic carcinoma
.
Nat Genet
2013
;
45
:
791
8
.
48.
Stephens
PJ
,
Davies
HR
,
Mitani
Y
,
Van Loo
P
,
Shlien
A
,
Tarpey
PS
, et al
Whole exome sequencing of adenoid cystic carcinoma
.
J Clin Invest
2013
;
123
:
2965
8
.
49.
Teng
MW
,
Ngiow
SF
,
Ribas
A
,
Smyth
MJ
. 
Classifying cancers based on T-cell infiltration and PD-L1
.
Cancer Res
2015
;
75
:
2139
45
.
50.
Latchman
Y
,
Wood
CR
,
Chernova
T
,
Chaudhary
D
,
Borde
M
,
Chernova
I
, et al
PD-L2 is a second ligand for PD-1 and inhibits T cell activation
.
Nat Immunol
2001
;
2
:
261
8
.
51.
Xiao
Y
,
Yu
S
,
Zhu
B
,
Bedoret
D
,
Bu
X
,
Francisco
LM
, et al
RGMb is a novel binding partner for PD-L2 and its engagement with PD-L2 promotes respiratory tolerance
.
J Exp Med
2014
;
211
:
943
59
.
52.
Baptista
MZ
,
Sarian
LO
,
Derchain
SF
,
Pinto
GA
,
Vassallo
J
. 
Prognostic significance of PD-L1 and PD-L2 in breast cancer
.
Hum Pathol
2016
;
47
:
78
84
.
53.
Shin
SJ
,
Jeon
YK
,
Kim
PJ
,
Cho
YM
,
Koh
J
,
Chung
DH
, et al
Clinicopathologic analysis of PD-L1 and PD-L2 expression in renal cell carcinoma: association with oncogenic proteins status
.
Ann Surg Oncol
2016
;
23
:
694
702
.
54.
Barrett
MT
,
Anderson
KS
,
Lenkiewicz
E
,
Andreozzi
M
,
Cunliffe
HE
,
Klassen
CL
, et al
Genomic amplification of 9p24.1 targeting JAK2, PD-L1, and PD-L2 is enriched in high-risk triple negative breast cancer
.
Oncotarget
2015
;
6
:
26483
93
.
55.
Rozali
EN
,
Hato
SV
,
Robinson
BW
,
Lake
RA
,
Lesterhuis
WJ
. 
Programmed death ligand 2 in cancer-induced immune suppression
.
Clin Dev Immunol
2012
;
2012
:
656340
.
56.
Sridharan
V
,
Schoenfeld
JD
. 
Immune effects of targeted radiation therapy for cancer
.
Discov Med
2015
;
19
:
219
28
.