HCC spatial transcriptomic profiling reveals significant and potentially targetable cancer-endothelial interactions.

BACKGROUND: HCC is a highly vascular tumor, and many effective drug regimens target the tumor blood vessels. Prior bulk HCC subtyping data used bulk transcriptomes, which contained a mixture of parenchymal and stromal contributions. METHODS: We utilized computational deconvolution and cell-cell interaction analyses to cell type-specific (tumor-enriched and vessel-enriched) spatial transcriptomic data collected from 41 resected HCC tissue specimens. RESULTS: We report that the prior Hoshida bulk transcriptional subtyping schema is driven largely by an endothelial fraction, show an alternative tumor-specific schema has potential prognostic value, and use spatially paired ligand-receptor analyses to identify known and novel (LGALS9 tumor-HAVCR2 vessel) signaling relationships that drive HCC biology in a subtype-specific and potentially targetable manner. CONCLUSIONS: Our study leverages spatial gene expression profiling technologies to dissect HCC heterogeneity and identify heterogeneous signaling relationships between cancer cells and their endothelial cells. Future validation and expansion of these findings may validate novel cancer-endothelial cell interactions and related drug targets.

Spatially resolved analysis of pancreatic cancer identifies therapy-associated remodeling of the tumor microenvironment.

In combination with cell-intrinsic properties, interactions in the tumor microenvironment modulate therapeutic response. We leveraged single-cell spatial transcriptomics to dissect the remodeling of multicellular neighborhoods and cell-cell interactions in human pancreatic cancer associated with neoadjuvant chemotherapy and radiotherapy. We developed spatially constrained optimal transport interaction analysis (SCOTIA), an optimal transport model with a cost function that includes both spatial distance and ligand-receptor gene expression. Our results uncovered a marked change in ligand-receptor interactions between cancer-associated fibroblasts and malignant cells in response to treatment, which was supported by orthogonal datasets, including an ex vivo tumoroid coculture system. We identified enrichment in interleukin-6 family signaling that functionally confers resistance to chemotherapy. Overall, this study demonstrates that characterization of the tumor microenvironment using single-cell spatial transcriptomics allows for the identification of molecular interactions that may play a role in the emergence of therapeutic resistance and offers a spatially based analysis framework that can be broadly applied to other contexts.

Tumor cell-based liquid biopsy using high-throughput microfluidic enrichment of entire leukapheresis product.

Circulating Tumor Cells (CTCs), interrogated by sampling blood from patients with cancer, contain multiple analytes, including intact RNA, high molecular weight DNA, proteins, and metabolic markers. However, the clinical utility of tumor cell-based liquid biopsy has been limited since CTCs are very rare, and current technologies cannot process the blood volumes required to isolate a sufficient number of tumor cells for in-depth assays. We previously described a high-throughput microfluidic prototype utilizing high-flow channels and amplification of cell sorting forces through magnetic lenses. Here, we apply this technology to analyze patient-derived leukapheresis products, interrogating a mean blood volume of 5.83 liters from patients with metastatic cancer, with a median of 2,799 CTCs purified per patient. Isolation of many CTCs from individual patients enables characterization of their morphological and molecular heterogeneity, including cell and nuclear size and RNA expression. It also allows robust detection of gene copy number variation, a definitive cancer marker with potential diagnostic applications. High-volume microfluidic enrichment of CTCs constitutes a new dimension in liquid biopsies.

Transfer Learning Reveals Cancer-Associated Fibroblasts Are Associated with Epithelial-Mesenchymal Transition and Inflammation in Cancer Cells in Pancreatic Ductal Adenocarcinoma.

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive malignancy characterized by an immunosuppressive tumor microenvironment enriched with cancer-associated fibroblasts (CAF). This study used a convergence approach to identify tumor cell and CAF interactions through the integration of single-cell data from human tumors with human organoid coculture experiments. Analysis of a comprehensive atlas of PDAC single-cell RNA sequencing data indicated that CAF density is associated with increased inflammation and epithelial-mesenchymal transition (EMT) in epithelial cells. Transfer learning using transcriptional data from patient-derived organoid and CAF cocultures provided in silico validation of CAF induction of inflammatory and EMT epithelial cell states. Further experimental validation in cocultures demonstrated integrin beta 1 (ITGB1) and vascular endothelial factor A (VEGFA) interactions with neuropilin-1 mediating CAF-epithelial cell cross-talk. Together, this study introduces transfer learning from human single-cell data to organoid coculture analyses for experimental validation of discoveries of cell-cell cross-talk and identifies fibroblast-mediated regulation of EMT and inflammation. SIGNIFICANCE: Adaptation of transfer learning to relate human single-cell RNA sequencing data to organoid-CAF cocultures facilitates discovery of human pancreatic cancer intercellular interactions and uncovers cross-talk between CAFs and tumor cells through VEGFA and ITGB1.

Mesothelin CAR T Cells Secreting Anti-FAP/Anti-CD3 Molecules Efficiently Target Pancreatic Adenocarcinoma and its Stroma.

PURPOSE: Targeting solid tumors with chimeric antigen receptor (CAR) T cells remains challenging due to heterogenous target antigen expression, antigen escape, and the immunosuppressive tumor microenvironment (TME). Pancreatic cancer is characterized by a thick stroma generated by cancer-associated fibroblasts (CAF), which may contribute to the limited efficacy of mesothelin-directed CAR T cells in early-phase clinical trials. To provide a more favorable TME for CAR T cells to target pancreatic ductal adenocarcinoma (PDAC), we generated T cells with an antimesothelin CAR and a secreted T-cell-engaging molecule (TEAM) that targets CAF through fibroblast activation protein (FAP) and engages T cells through CD3 (termed mesoFAP CAR-TEAM cells). EXPERIMENTAL DESIGN: Using a suite of in vitro, in vivo, and ex vivo patient-derived models containing cancer cells and CAF, we examined the ability of mesoFAP CAR-TEAM cells to target PDAC cells and CAF within the TME. We developed and used patient-derived ex vivo models, including patient-derived organoids with patient-matched CAF and patient-derived organotypic tumor spheroids. RESULTS: We demonstrated specific and significant binding of the TEAM to its respective antigens (CD3 and FAP) when released from mesothelin-targeting CAR T cells, leading to T-cell activation and cytotoxicity of the target cell. MesoFAP CAR-TEAM cells were superior in eliminating PDAC and CAF compared with T cells engineered to target either antigen alone in our ex vivo patient-derived models and in mouse models of PDAC with primary or metastatic liver tumors. CONCLUSIONS: CAR-TEAM cells enable modification of tumor stroma, leading to increased elimination of PDAC tumors. This approach represents a promising treatment option for pancreatic cancer.

Interplay between B7-H3 and HLA class I in the clinical course of pancreatic ductal adenocarcinoma.

Human leukocyte antigen (HLA) class I defects are associated with cancer progression. However, their prognostic significance is controversial and may be modulated by immune checkpoints. Here, we investigated whether the checkpoint B7-H3 modulates the relationship between HLA class I and pancreatic ductal adenocarcinoma (PDAC) prognosis. PDAC tumors were analyzed for the expression of B7-H3, HLA class I, HLA class II molecules, and for the presence of tumor-infiltrating immune cells. We observed defective HLA class I and HLA class II expressions in 75% and 59% of PDAC samples, respectively. HLA class I and B7-H3 expression were positively related at mRNA and protein level, potentially because of shared regulation by RELA, a sub-unit of NF-kB. High B7-H3 expression and low CD8+ T cell density were indicators of poor survival, while HLA class I was not. Defective HLA class I expression was associated with unfavorable survival only in patients with low B7-H3 expression. Favorable survival was observed only when HLA class I expression was high and B7-H3 expression low. Our results provide the rationale for targeting B7-H3 in patients with PDAC tumors displaying high HLA class I levels.

Dysregulated Repeat Element Viral-like Immune Response in Hepatocellular Carcinoma.

Purpose: Dysregulation of viral-like repeat RNAs are a common feature across many malignancies that are linked with immunological response, but the characterization of these in hepatocellular carcinoma (HCC) is understudied. In this study, we performed RNA in situ hybridization (RNA-ISH) of different repeat RNAs, immunohistochemistry (IHC) for immune cell subpopulations, and spatial transcriptomics to understand the relationship of HCC repeat expression, immune response, and clinical outcomes. Experimental Design: RNA-ISH for LINE1, HERV-K, HERV-H, and HSATII repeats and IHC for T-cell, Treg, B-cell, macrophage, and immune checkpoint markers were performed on 43 resected HCC specimens. Spatial transcriptomics on tumor and vessel regions of interest was performed on 28 specimens from the same cohort. Results: High HERV-K and high LINE1 expression were both associated with worse overall survival. There was a positive correlation between LINE1 expression and FOXP3 T-regulatory cells (r = 0.51 p < 0.001) as well as expression of the TIM3 immune checkpoint (r = 0.34, p = 0.03). Spatial transcriptomic profiling of HERV-K high and LINE-1 high tumors identified elevated expression of multiple genes previously associated with epithelial mesenchymal transition, cellular proliferation, and worse overall prognosis in HCC including SSX1, MAGEC2, and SPINK1. Conclusion: Repeat RNAs may serve as useful prognostic biomarkers in HCC and may also serve as novel therapeutic targets. Additional study is needed to understand the mechanisms by which repeat RNAs impact HCC tumorigenesis.

Proceedings of the inaugural Dark Genome Symposium: November 2022.

In November 2022 the first Dark Genome Symposium was held in Boston, USA. The meeting was hosted by Rome Therapeutics and Enara Bio, two biotechnology companies working on translating our growing understanding of this vast genetic landscape into therapies for human disease. The spirit and ambition of the meeting was one of shared knowledge, looking to strengthen the network of researchers engaged in the field. The meeting opened with a welcome from Rosana Kapeller and Kevin Pojasek followed by a first session of field defining talks from key academics in the space. A series of panels, bringing together academia and industry views, were then convened covering a wide range of pertinent topics. Finally, Richard Young and David Ting gave their views on the future direction and promise for patient impact inherent in the growing understanding of the Dark Genome.

The landscape of human SVA retrotransposons.

SINE-VNTR-Alu (SVA) retrotransposons are evolutionarily young and still-active transposable elements (TEs) in the human genome. Several pathogenic SVA insertions have been identified that directly mutate host genes to cause neurodegenerative and other types of diseases. However, due to their sequence heterogeneity and complex structures as well as limitations in sequencing techniques and analysis, SVA insertions have been less well studied compared to other mobile element insertions. Here, we identified polymorphic SVA insertions from 3646 whole-genome sequencing (WGS) samples of >150 diverse populations and constructed a polymorphic SVA insertion reference catalog. Using 20 long-read samples, we also assembled reference and polymorphic SVA sequences and characterized the internal hexamer/variable-number-tandem-repeat (VNTR) expansions as well as differing SVA activity for SVA subfamilies and human populations. In addition, we developed a module to annotate both reference and polymorphic SVA copies. By characterizing the landscape of both reference and polymorphic SVA retrotransposons, our study enables more accurate genotyping of these elements and facilitate the discovery of pathogenic SVA insertions.

Spatial transcriptomics reveals distinct tissue niches linked with steroid responsiveness in acute gastrointestinal GVHD.

Severe acute graft-versus-host disease (aGVHD) is associated with significant mortality and morbidity, especially in steroid-resistant (SR) cases. Spatial transcriptomic technology can elucidate tissue-based interactions in vivo and possibly identify predictors of treatment response. Tissue sections from 32 treatment-naïve patients with biopsy-confirmed lower gastrointestinal (GI) aGVHD were obtained. The GeoMx digital spatial profiler was used to capture transcriptome profiles of >18 000 genes from different foci of immune infiltrates, colonic epithelium, and vascular endothelium. Each tissue compartment sampled showed 2 distinct clusters that were analyzed for differential expression and spatially resolved correlation of gene signatures. Classic cell-mediated immunity signatures, normal differentiated epithelial cells, and inflamed vasculature dominated foci sampled from steroid-sensitive cases. In contrast, a neutrophil predominant noncanonical inflammation with regenerative epithelial cells and some indication of angiogenic endothelial response was overrepresented in areas from SR cases. Evaluation of potential prognostic biomarkers identified ubiquitin specific peptidase 17-like (USP17L) family of genes as being differentially expressed in immune cells from patients with worsened survival. In summary, we demonstrate distinct tissue niches with unique gene expression signatures within lower GI tissue from patients with aGVHD and provide evidence of a potential prognostic biomarker.

Therapy-associated remodeling of pancreatic cancer revealed by single-cell spatial transcriptomics and optimal transport analysis.

In combination with cell intrinsic properties, interactions in the tumor microenvironment modulate therapeutic response. We leveraged high-plex single-cell spatial transcriptomics to dissect the remodeling of multicellular neighborhoods and cell-cell interactions in human pancreatic cancer associated with specific malignant subtypes and neoadjuvant chemotherapy/radiotherapy. We developed Spatially Constrained Optimal Transport Interaction Analysis (SCOTIA), an optimal transport model with a cost function that includes both spatial distance and ligand-receptor gene expression. Our results uncovered a marked change in ligand-receptor interactions between cancer-associated fibroblasts and malignant cells in response to treatment, which was supported by orthogonal datasets, including an ex vivo tumoroid co-culture system. Overall, this study demonstrates that characterization of the tumor microenvironment using high-plex single-cell spatial transcriptomics allows for identification of molecular interactions that may play a role in the emergence of chemoresistance and establishes a translational spatial biology paradigm that can be broadly applied to other malignancies, diseases, and treatments.

Cancer cells co-evolve with retrotransposons to mitigate viral mimicry.

Overexpression of repetitive elements is an emerging hallmark of human cancers 1 . Diverse repeats can mimic viruses by replicating within the cancer genome through retrotransposition, or presenting pathogen-associated molecular patterns (PAMPs) to the pattern recognition receptors (PRRs) of the innate immune system 2-5 . Yet, how specific repeats affect tumor evolution and shape the tumor immune microenvironment (TME) in a pro- or anti-tumorigenic manner remains poorly defined. Here, we integrate whole genome and total transcriptome data from a unique autopsy cohort of multiregional samples collected in pancreatic ductal adenocarcinoma (PDAC) patients, into a comprehensive evolutionary analysis. We find that more recently evolved S hort I nterspersed N uclear E lements (SINE), a family of retrotransposable repeats, are more likely to form immunostimulatory double-strand RNAs (dsRNAs). Consequently, younger SINEs are strongly co-regulated with RIG-I like receptor associated type-I interferon genes but anti-correlated with pro-tumorigenic macrophage infiltration. We discover that immunostimulatory SINE expression in tumors is regulated by either L ong I nterspersed N uclear E lements 1 (LINE1/L1) mobility or ADAR1 activity in a TP53 mutation dependent manner. Moreover, L1 retrotransposition activity tracks with tumor evolution and is associated with TP53 mutation status. Altogether, our results suggest pancreatic tumors actively evolve to modulate immunogenic SINE stress and induce pro-tumorigenic inflammation. Our integrative, evolutionary analysis therefore illustrates, for the first time, how dark matter genomic repeats enable tumors to co-evolve with the TME by actively regulating viral mimicry to their selective advantage.

Clinical, pathological, genetics and intratumoural immune milieu of micropapillary carcinoma of the colon.

AIM: Micropapillary carcinoma (MPC) is a recognised WHO variant of colonic carcinoma (CC), although little is known about its prognosis, immune microenvironment and molecular alterations. We investigated its clinical, pathological and immunological characteristics. METHODS: We assessed 903 consecutive CCs and used the WHO definition to identify MPC. We recorded serrated and mucinous differentiation and mismatch repair (MMR) status. We performed immunohistochemistry and quantification on tissue microarrays for HLA class I/II proteins, beta-2-microglobulin (B2MG), CD8, CD163, LAG3, PD-L1, FoxP3, PD-L1and BRAF V600E. RESULTS: We classified 8.6% (N=78) of CC as MPC. Relative to non-MPC, MPC was more often high grade (p=0.03) and showed serrated morphology (p<0.01); however, we found no association with extramural venous invasion (p=0.41) and American Joint Committee on Cancer stage (p=0.95). MPCs showed lower numbers of CD8 positive lymphocytes (p<0.01), lower tumour cell B2MG expression (p=0.04) and lower tumour cell PD-L1 expression (p<0.01). There was no difference in HLA class I/II, LAG3, FOXP3, CD163 and PD-L1 positive histiocytes. There was no association with MMR status or BRAF V600E relative to non-MPC. MPC was not associated with decreased disease-specific survival (p=0.36). CONCLUSION: MPCs are associated with high-grade differentiation and a less active immune microenvironment than non-MPC. MPC is not associated with inferior disease-specific survival.

SULT1A1-dependent sulfonation of alkylators is a lineage-dependent vulnerability of liver cancers.

Adult liver malignancies, including intrahepatic cholangiocarcinoma and hepatocellular carcinoma, are the second leading cause of cancer-related deaths worldwide. Most individuals are treated with either combination chemotherapy or immunotherapy, respectively, without specific biomarkers for selection. Here using high-throughput screens, proteomics and in vitro resistance models, we identify the small molecule YC-1 as selectively active against a defined subset of cell lines derived from both liver cancer types. We demonstrate that selectivity is determined by expression of the liver-resident cytosolic sulfotransferase enzyme SULT1A1, which sulfonates YC-1. Sulfonation stimulates covalent binding of YC-1 to lysine residues in protein targets, enriching for RNA-binding factors. Computational analysis defined a wider group of structurally related SULT1A1-activated small molecules with distinct target profiles, which together constitute an untapped small-molecule class. These studies provide a foundation for preclinical development of these agents and point to the broader potential of exploiting SULT1A1 activity for selective targeting strategies.

TGF-ß in the microenvironment induces a physiologically occurring immune-suppressive senescent state.

TGF-ß induces senescence in embryonic tissues. Whether TGF-ß in the hypoxic tumor microenvironment (TME) induces senescence in cancer and how the ensuing senescence-associated secretory phenotype (SASP) remodels the cellular TME to influence immune checkpoint inhibitor (ICI) responses are unknown. We show that TGF-ß induces a deeper senescent state under hypoxia than under normoxia; deep senescence correlates with the degree of E2F suppression and is marked by multinucleation, reduced reentry into proliferation, and a distinct 14-gene SASP. Suppressing TGF-ß signaling in tumors in an immunocompetent mouse lung cancer model abrogates endogenous senescent cells and suppresses the 14-gene SASP and immune infiltration. Untreated human lung cancers with a high 14-gene SASP display immunosuppressive immune infiltration. In a lung cancer clinical trial of ICIs, elevated 14-gene SASP is associated with increased senescence, TGF-ß and hypoxia signaling, and poor progression-free survival. Thus, TME-induced senescence may represent a naturally occurring state in cancer, contributing to an immune-suppressive phenotype associated with immune therapy resistance.

Ultrasensitive detection of circulating LINE-1 ORF1p as a specific multi-cancer biomarker.

Improved biomarkers are needed for early cancer detection, risk stratification, treatment selection, and monitoring treatment response. While proteins can be useful blood-based biomarkers, many have limited sensitivity or specificity for these applications. Long INterspersed Element-1 (LINE-1, L1) open reading frame 1 protein (ORF1p) is a transposable element protein overexpressed in carcinomas and high-risk precursors during carcinogenesis with negligible detectable expression in corresponding normal tissues, suggesting ORF1p could be a highly specific cancer biomarker. To explore the potential of ORF1p as a blood-based biomarker, we engineered ultrasensitive digital immunoassays that detect mid-attomolar (10-17 M) ORF1p concentrations in patient plasma samples across multiple cancers with high specificity. Plasma ORF1p shows promise for early detection of ovarian cancer, improves diagnostic performance in a multi-analyte panel, and provides early therapeutic response monitoring in gastric and esophageal cancers. Together, these observations nominate ORF1p as a multi-cancer biomarker with potential utility for disease detection and monitoring.

Quantitative p53 immunostaining aids in the detection of prevalent dysplasia.

AIMS: The lack of accepted scoring criteria has precluded the use of p53 in routine practice. We evaluate the utility of automated quantitative p53 analysis in risk stratifying Barrett's oesophagus (BE) patients using non-dysplastic BE (NDBE) biopsies in a multicentric cohort of BE progressor (P) and non-progressor (NP) patients. METHODS: NDBE biopsies prior to the diagnosis of advanced neoplasia from 75 BE-P, and index and last surveillance biopsies from 148 BE-NP were stained for p53, and scored digitally as 1+, 2+ and 3+. A secondary cohort of 30 BE-P was evaluated. RESULTS: Compared with BE-NP, BE-P was predominantly men (p=0.001), ≥55 years of age (p=0.008), with longer BE segments (71% vs 33%; p<0.001). The mean number of 3+p53 positive cells and 3+ positive glands were significantly more in BE-P versus BE-NP NDBE biopsies (175 vs 9.7, p<0.001; 9.8 vs 0.1; p<0.001, respectively). At a cut-off of ≥10 p53 (3+) positive cells, the sensitivity and specificity of the assay to identify BE-P were 39% and 93%. On multivariate analysis, scoring p53 in NDBE biopsies, age, gender and length of BE were significantly associated with neoplastic progression. 54% of patients classified as prevalent dysplasia showed an abnormal p53 immunohistochemical stain. These findings were validated in the secondary cohort. CONCLUSIONS: Automated p53 analysis in NDBE biopsies serves as a promising tool for assessing BE neoplastic progression and risk stratification. Our study highlights the practical applicability of p53 assay to routine surveillance practice and its ability to detect prevalent dysplasia.

Immunological and Clinico-Molecular Features of Tumor Border Configuration in Colorectal Cancer.

INTRODUCTION: Infiltrating tumor border configuration (ITBC) portends a poor prognosis compared with pushing tumor border configuration (PTBC) in colorectal cancer. The tumor and its surrounding immune microenvironment of tumor border configuration is not well-characterized. We aim to elucidate the differences in expression of molecular markers between the 2 groups using tissue microarray (TMA). STUDY DESIGN: Immunohistochemistry was performed on TMAs of surgical pathology specimens obtained from colorectal cancer patients consecutively operated at our institution from 2004 to 2015. TMAs were stained for immune cells (CD8, FOXP3, LAG3, PU1, CD163, and PDL1); HLA II, beta 2 microglobulin, and HC10 on tumor cells; BRAFV600E mutation; and DNA mismatch repair proteins (MMR) status. Patients who received neoadjuvant therapy were excluded. RESULTS: There were 646 tumors with ITBC and 310 tumors with PTBC. There was a significantly lower expression (p < 0.05) of immune components, namely CD8, FOXP3, LAG3, PU1, PDL1 immune cells, and Beta-2 Microglobulin on tumor cells in the tumors with ITBC compared with PTBC, except CD163 immune cells, and HC10 and HLAII on tumor cells. Tumors with ITBC were less likely to be associated with BRAFV600E mutations and deficient MMR proteins (p < 0.001). On analyzing MMR-proficient tumors separately, we could not find any difference in the expression of any molecular marker (including BRAF), except a lower expression of PDL1 immune cells in tumors with ITBC (p < 0.001). CONCLUSIONS: Colorectal tumors with ITBC are associated with a generalized low immune microenvironment and low rates of BRAFV600E mutation compared with tumors with PTBC. However, the molecular expression of tumor border configuration seems confounded by the MMR molecular signature. MMR-proficient colorectal tumors with ITBC are associated with a lower expression of only PDL1 immune cells among all immune markers examined.

Targeting breast and pancreatic cancer metastasis using a dual-cadherin antibody.

The successful application of antibody-based therapeutics in either primary or metastatic cancer depends upon the selection of rare cell surface epitopes that distinguish cancer cells from surrounding normal epithelial cells. By contrast, as circulating tumor cells (CTCs) transit through the bloodstream, they are surrounded by hematopoietic cells with dramatically distinct cell surface proteins, greatly expanding the number of targetable epitopes. Here, we show that an antibody (23C6) against cadherin proteins effectively suppresses blood-borne metastasis in mouse isogenic and xenograft models of triple negative breast and pancreatic cancers. The 23C6 antibody is remarkable in that it recognizes both the epithelial E-cadherin (CDH1) and mesenchymal OB-cadherin (CDH11), thus overcoming considerable heterogeneity across tumor cells. Despite its efficacy against single cells in circulation, the antibody does not suppress primary tumor formation, nor does it elicit detectable toxicity in normal epithelial organs, where cadherins may be engaged within intercellular junctions and hence inaccessible for antibody binding. Antibody-mediated suppression of metastasis is comparable in matched immunocompetent and immunodeficient mouse models. Together, these studies raise the possibility of antibody targeting CTCs within the vasculature, thereby suppressing blood-borne metastasis.

FGFR mRNA Expression in Cholangiocarcinoma and Its Correlation with FGFR2 Fusion Status and Immune Signatures.

PURPOSE: Selective FGFR inhibitors are effective against cholangiocarcinomas that harbor gene alterations in FGFR2. Clinical trials suggest that expression of wild-type FGFR mRNA can predict sensitivity to FGFR inhibitors, but this biomarker has not been well characterized in cholangiocarcinoma. This study explores the prevalence of FGFR mRNA overexpression in cholangiocarcinoma, its role in predicting sensitivity to FGFR inhibitors, and its association with immune markers. EXPERIMENTAL DESIGN: Tissue microarrays of intrahepatic (ICC) and extrahepatic cholangiocarcinomas (ECC) resected between 2004 and 2015 were used to evaluate FGFR1-4 mRNA expression levels by RNA in situ hybridization (ISH). Expression levels of FGFR2 mRNA were correlated with FGFR2 fusion status and with patient outcomes. Immune markers expression was assessed by IHC and CSF1 and CSF1 receptor expression were examined by RNA ISH. RESULTS: Among 94 patients with resected cholangiocarcinoma, the majority had ICC (77%). FGFR2 fusions were identified in 23% of ICCs and 5% of ECCs. High levels of FGFR mRNA in FGFR2 fusion-negative ICC/ECC were seen for: FGFR1 (ICC/ECC: 15%/0%), FGFR2 (ICC/ECC: 57%/0%), FGFR3 (ICC/ECC: 53%/18%), and FGFR4 (ICC/ECC: 32%/0%). Overall, 62% of fusion-negative cholangiocarcinomas showed high levels of FGFR mRNA. In patients with advanced FGFR2 fusion-positive ICC, high levels of FGFR2 mRNA did not correlate with clinical benefit. FGFR2 fusion-positive tumors showed a paucity of PD-L1 on tumor cells. CONCLUSIONS: FGFR mRNA overexpression occurs frequently in cholangiocarcinoma in the absence of genetic alterations in FGFR. This study identifies a molecular subpopulation in cholangiocarcinoma for which further investigation of FGFR inhibitors is merited outside currently approved indications.