Document resource
Background Tertiary lymphoid structures (TLS) are ectopic lymphoid aggregates found in several cancer types because of chronic inflammation. TLS are often linked to enhanced immunotherapy response and increased overall survival in most solid tumors. However, cues that drive differences in TLS formation and function are not clearly understood. Here, we spatially interrogated TLS in treatment naïve non-small cell lung cancer (NSCLC) patients to delineate distinct functional states.Methods We combined multispectral imaging and spatial transcriptomics (GeoMx® whole transcriptomic atlas) to comprehensively profile the tumor microenvironment (TME) in 70 NSCLC patients. We deployed a machine learning approach to identify variance in TLS composition and transcriptional states, and compared our new TLS insights to traditional pathological assessment of TLS. We used independent patient cohorts (NSCLC, ovarian, and brain metastases) to validate our findings in multiple tumor indices and correlate to survival outcomes.Results Our machine learning identified TLS (herein referred to as TLS type I state) with a distinct transcriptional signature which also correlated with improved prognosis. Importantly, we observed a trajectory of chemokine expression that heightened in the TLS type I state, indicative of TLS development in the TME. Furthermore, the TLS type I state had enhanced B cell receptor signaling pathways that coincided with high expression of CD37 and CD52. CD37 is a tetraspanin protein expressed by B cells and regulates B cell signaling and cell adhesion. CD52 is a cell surface glycoprotein and plays a critical role as a target in chronic lymphocytic leukemia treatment. In a separate cohort of 34 patients, we validated the expression of CD37 and CD52 by multispectral imaging. CD37 and CD52 were highly abundant in some TLS while lacking in others. Of note, expression of CD37 or CD52 in TLS correlated with enhanced TLS activity, measured by Ki67 expression within the TLS. Patients with a high density of CD37+ or CD52+ TLS or both had a higher probability of survival.Conclusions Our work utilizes spatial omics to delineate transcriptional differences that govern TLS function. We describe key pathways that are associated with TLS functional states, highlighting transcriptional changes in TLS state. Finally, for clinical relevance, we identified and validated CD37 and CD52 as important biomarkers for functional TLS in NSCLC. Our findings shed light on key developmental changes governing TLS function that could be used to modulate TLS in solid tumors for improved immunotherapy response.Ethics Approval This study was approved by the University of Pittsburgh institutional Ethics Review Board, (STUDY21020081).