Abstract
Background: PIN1, a peptidyl-prolyl isomerase, is overexpressed in many tumor types. Previous studies in pancreatic cancer have shown an association between higher PIN1 expression with an immunosuppressive tumor microenvironment (TME) and poor clinical outcomes. Inhibition of PIN1 in animal models renders pancreatic cancer eradicable by immunotherapy. However, studies of PIN1 overexpression in biliary tract cancers (BTCs) are limited. We thus investigated the relationship between PIN1 mRNA expression in BTCs and the TME, common somatic alterations, overall survival, and response to therapy in a large real-world dataset.

Methods: We used Tempus Lens to identify patients (pts) with a primary diagnosis of BTC who had Tempus xT DNA and xR RNA testing from a multimodal database (n=4479). RNA-seq data were normalized to correct for assay/batch effects, quantified as transcripts per million (TPM) and reported as log2(TPM+1). Pts were classified as PIN1 high (PIN1-H, n=2240) or low (PIN1-L, n=2239) based on median PIN1 mRNA expression. Immune cell proportions and cytolytic, cytotoxic, and interferon-γ immune scores were estimated from RNA expression. Chi-squared/Fisher’s exact or Wilcoxon rank sum tests were used to assess statistical significance. Real-world overall survival (rwOS) was defined as time from sample collection to death or loss to follow up. Median rwOS was calculated using the Kaplan-Meier estimator and significance using log-rank test.

Results: PIN1-H cases had a median PIN1 mRNA expression of 3.54 (range 3.31-5.78) while PIN1-L cases had a median expression of 3.06 (range 1.82-3.31). The TME of the PIN1-H group showed significantly higher enrichment in M1 macrophages (p<0.001) as well as cytolytic (p<0.001), cytotoxic (p=0.012), and interferon-γ (p=0.048) signatures compared to the PIN1-L group. Somatic alterations differed with PIN1-H status associated with lower rates of TP53 (40 vs. 48%) and ERBB2 alterations (2.4 vs 5.3%) and higher rates of IDH1 alterations (12 vs 7.1%) (all p<0.001). Median rwOS was significantly longer for PIN1-H vs. PIN1-L (11.3 vs. 8.5 mo), driven primarily by late-stage disease. There was no statistical difference when restricting to pts who underwent surgical resection. The significant difference favoring PIN1-H persisted when stratifying by TP53 status, as well as restricting to cases with no actionable alterations. There were no significant differences between the two groups’ response to first-line systemic treatment.

Conclusions: In this real-world genomic analysis of BTCs, we found that PIN1-H was associated with a more active immune microenvironment and better prognosis, even after accounting for differences in key genetic alterations. This contrasts with studies in other GI malignancies, emphasizing complexities of PIN1 activity. We note that our dataset had greater representation of stage IV disease compared to many analyses in the literature and focused on mRNA expression rather than immunohistochemistry. Further studies are needed to delineate the nuances of PIN1 activity in BTCs.

VIEW THE PUBLICATION

VIEW THE POSTER