Spatial immunophenotyping of PD1 and PDL1 in tumor cells and microenvironment of nodal follicular T helper cell lymphoma Free

Background: Nodal T follicular helper cell lymphomas (nTFHLs) represent a group of mature T cell lymphomas with heterogenous microenvironment and the phenotypic features and gene expression signature of TFH cells^{1-3, 6}. Although Immunohistochemistry (IHC) remains a valuable tool for targeted treatment regimens like PD1 (programmed cell death 1) antibodies for PDL1expressing tumors, it lacks spatial multimarker precision and quantification.^4-5 To date, only few studies evaluated simultaneous quantification of markers in follicular T helper cell lymphoma and immune cell populations by Multiplex IF ^{7-9, 12,.}

PD 1/PDL-1 Axis activates Intracellular ERK Signaling in Tumor Cells to Mediate Poor Prognosis in T-cell Lymphoma ^10-11. Recently, Geptanolimab (GB226), an anti-PD-1 antibody, demonstrated an overall RR of 40.4% in relapsed or refractory PTCL. A subgroup analysis showed better response and survival in patients with PD-L1 ≥ 50% ¹³. A combination of checkpoint inhibitors with other agents could be a promising option to enhance anti-tumor activity in T cell lymphoma¹⁴ . Spatial proximity of immune cells and tumor cells has been evaluated in solid tumors and is associated with response to check point inhibitors in our previous study¹⁵ . No study has evaluated the spatial proximity analysis of PDL1 in the immunomicroenviornement of follicular T helper cell lymphomas.

We used multiplex immunofluorescence (mIF) in FTH lymphomas to identify detailed spatial analysis of immune markers, offering insight into checkpoint pathways (PD-1/PD-L1).

Methods and materials:

Retrospective cohorts were retrieved from samples archived by the Department of Pathology, UCDH. A case of Angioimmunoblastic T Cell Lymphoma was selected with distinct immunophenotype lacking surface CD3 on flow cytometry. The Tumor rich area was annotated based on the morphology, and immunohistochemistry of lymphoma (CD3dim+, CD7-, CD4+.CD8-). FFPE tissue sections (4 µm) were stained with Leica BondRX autostainer (Leica) following pre-optimized immune panel (CD3, CD8, CD68, PD-1, PD-L1, panCK) ¹⁶. Stained slides were scanned with Vectra/Polaris multispectral imaging system (Akoya Biosciences) and imaging analysis was performed with QuPath¹⁷ . The tissue images from our tissue sample and human tonsil were used to build artificial intelligence (AI)- based cell classifier with cell specific markers. Regions of interest (ROI) were selected on IHC slide and equivalent area was annotated on QuPath to acquire cell segmentation data from the ROI. Data were extracted to CSV and interrogated by Python/R to quantify marker co-expression and spatial proximity relationships.

Results: High expression of PD-L1 >50% was noted on the lymph node with strong and diffuse staining for PD-1 on tumor cells. Images of single antibodies were merged to evaluate expression of multiple antibodies on the tumor cells and their microenvironment. The percentage of PD-L1+ cells was calculated on T cells and CD8+ T cells with and without PD-1 as well as on histiocytes. The percentage of other PD-L1+ cells with expression of PD-1 was also measured. We found that 71.07% of the histiocytes expressed PD-L1. 49.9% of T cells in tumor microenvironment (CD3+/PD1+) expressed PD-L1 including subset of CD8+ T cells. 19.58% of T-cells co-express PD-1 and PD-L1 and lack CD8 expression, suggesting a CD4⁺ subset. Next we applied stringent criteria to CD3+/CD8- to select CD3dim and intermediate to large nuclei to define lymphoma cells (CD3dim large). To define reactive T cells (CD3high small) we included high CD3 and small nuclei. Proximity analysis showed that CD3dim large cells are closer to CD68 than to CD3 high small cells. This was similar regardless of size and if only CD3 dim cells were included in the analysis. We also found that PDL1 expressions were higher on cells with CD3dim than CD3 high. (83.3% vs. 45.1%).

Conclusion: Proximity analysis revealed CD3 dim and large cells, selected by stringent criteria as lymphoma cells, were closer to macrophages. We also found that the population of CD8-/CD3 dim, defined as tumor population, expressed higher PDL1 as compared to CD3 high, representing reactive T cells. Our study suggests that Lymphoma cells may express PDL1. mIF enabled quantification and spatial analysis of cellular components within lymphoma. The case study provided insight into the expression of PD-L1 in the lymphoma cells and its microenvironment.