Article Text

Download PDFPDF

496 Immune cell profiling across solid tumor types by mass cytometry reveals tumor enrichment of PD-1+/LAG-3+ CD8 memory T cells that exhibit tumor-reactive yet dysfunctional features
  1. Bradley Garman,
  2. Laurence Menard,
  3. Can Jiang,
  4. Sherif Daouti,
  5. Priyanka Mehta,
  6. Miye Jacques,
  7. Mohan Bolisetty,
  8. Christos Hatzis,
  9. Nataly Manjarrez Orduno,
  10. Michaela Bowden,
  11. Justin David,
  12. Justin David and
  13. Justin David
  1. Bristol Myers Squibb, Lawrenceville, NJ, USA


Background Characterization of human immune responses by profiling immune cells from patients is critical for the successful development of immuno-oncology agents and is useful to understand mechanism-of-action, identify pharmacodynamic/response biomarkers, and guide patient selection strategies. Extensive immune cell heterogeneity necessitates comprehensive high parameter immunophenotyping to yield these actionable insights.

Methods Cytometry by time-of-flight (CyTOF) was performed on homogenates from commercially procured tumors (n=28) and matched PBMCs (n=7) from patients with various solid tumors (colon (n=10), endometrial (n=9), kidney (n=4), liver (n=2), skin (n=1), lung (n=1), and gastro-intestinal (n=1)). Two antibody panels, recognizing a total of 18 lineage and 31 target proteins, were used to profile marker expression among the major lymphocyte and myeloid lineages. Data were analyzed using manual gating and non-linear dimensionality reduction (tSNE and UMAP), and expression was measured by frequency (% gate) and arcsinh-transformed median ion counts. Pairwise Wilcoxon Rank Sum tests were performed on arcsinh-transformed median ion counts to determine statistically significant differences in marker expression, and P values were adjusted using Benjamini-Hochberg correction (p<0.05 considered statistically significant). Cell subpopulation percentages were compared using unpaired two-sided T-tests. Sample populations with less than 150 events were excluded from analysis. The initial analysis focused on CD8 T cells as a primary mediator of antitumor immunity.

Results Matched samples revealed enrichment of effector memory (EM) and central memory (CM) CD8 T cells in tumors compared to PBMCs, as expected. EM cells represented on average 63.36% of the CD8 T cells in tumors vs 30.31% in PBMCs (p=0.0067), and CM cells 12.11% vs 5.58% respectively (p=0.1558). Non-linear dimensionality reduction mapping of these CD8 EM and CM cell subtypes among tumors displayed an activated but potentially dysfunctional phenotype, characterized by substantially higher expression of multiple coinhibitory receptors (PD-1, LAG-3, TIM-3, TIGIT) and activation markers (HLA-DR, ICOS) compared to PBMCs. Among these cells, a PD-1+/LAG-3+ subset, observed in 17/28 TIL samples, expressed TIM-3, TIGIT, HLA-DR, and ICOS at significantly higher levels compared to other PD1/LAG3 expression subsets. Interestingly, CD137 (4-1BB), a marker of potentially tumor-reactive cells, is expressed predominantly in PD-1+ memory CD8 T cells, with the most intense expression levels observed in the PD-1+/LAG-3+ subset.

Conclusions The present results provide insight into the relative (co)expression of potentially targetable immunological pathways, and suggest a biological basis for informing approaches to combination checkpoint inhibition therapy.

Acknowledgements We thank Paul Fischer for his contributions in acquiring the CyTOF data and performing initial data QC and analysis.

Ethics Approval This study was approved by Bristol Myers Squibb’s Global Data Repository (Biological Assessment of Risk (BAR) number LVL_2020_12339). Samples were provided by Discovery Life Sciences (CA), MT Group (CA), Avaden BioSciences (WA), or BioOptions (CA). All patients gave written informed consent at the time of sample collection according to the IRB protocols of each provider.

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.