The Role of Relapse Specific over-Expression of S100A8/A9 B Leukemic Cells in Promoting Chemoresistance

Background: We previously discovered a novel superenhancer (SE) adjacent to S100A8/9 at relapse in a majority of patient samples indicating a role in clonal evolution. In agreement with these findings, a metanalysis of gene expression studies demonstrated that S100A8 is the top overexpressed transcript at relapse in childhood ALL. S100A8 is a member of the S100 calcium-binding protein family and some studies suggest a role in mediating intrinsic drug resistance. However, secreted S100A8/9 proteins play a major role in inflammation and it is possible that their role in clonal evolution occurs by altering the tumor microenvironment (TME).

Design/Method

Isogenic B-ALL cell lines (697 and RS4;11) were created to overexpress an empty vector control (EV), S100A8 or S100A9 alone, or S100A8+S100A9 together. Protein expression and secretion were measured using an ELISA for human S100A8/9 (Abcam). Proliferation over a 7-day period was measured using the Beckman Coulter Vi-CELL counter. Chemoresistance was measured using Promega's CellTiter-Glo® assay following exposure to chemotherapeutic agents used in the treatment of ALL including mercaptopurine, prednisolone, methotrexate, vincristine, cytarabine, and etoposide for 72-120 hours. Colony forming potential was assessed after 14 days in MethoCult (STEMCELL Technologies).

To determine that expression originated from leukemia cells one sample pair (SGZ) known to overexpress S100A8 at relapse using bulk RNA expression, was compared to 3 paired patient samples without overexpression using single cell RNAseq (scRNAseq). CD45+CD19+ (blasts) and CD45+CD19- (microenvironment) were sorted and then pooled to increase the fraction of cells from the microenvironment before scRNAseq using 10X Genomics. CellPhoneDB analysis was used to predict ligand-receptor interactions between different cell types.

Results

Western blot demonstrated stable expression of S100A8 and S100A9 proteins when co-expressed but protein was not detected when either was alone. Overexpression of S100A8+S100A9 was confirmed via ELISA using both cell lysate and condition media suggesting secretion. No differences in proliferation, colony formation, or chemoresistance were observed between cells overexpressing S100A8/9 compared to the empty vector control line, suggesting no cell intrinsic effect of S100A8/9 overexpression

ssRNAseq analysis validated upregulation of S100A8/9 specifically in leukemic blasts (cluster 12, L2FC of 5.6 and 5.3 (p<.0001) for S100A8 and S100A9, respectively). In contrast, two myeloid monocytic-like populations showed minimal to no change in expression level from diagnosis to relapse (L2FC 1.13 and 0.18 for S100A8 and S100A9, respectively (p<.01). Pathway analysis of differentially regulated genes in SGZ (L2FC cutoff of 1 and p<.001, Enrichr, Panther 2016) revealed known relapse pathways, including Ras, p53, and p38 MAPK, but also those known to impact the TME including Inflammation mediated by chemokine and cytokine signaling, Toll receptor signaling, and TFG-beta signaling.

Interestingly, we also observed an increase in, and higher percentage of non-classical monocytes within the non-leukemic cell population at relapse in SGZ (>20%) compared to minimal (<5%) in the other three patients. In SGZ, we found 455 and 573 significantly upregulated (p<.05, abs L2FC>.32) and 686 and 362 significantly downregulated genes (p <.05, absolute L2FC <.32) in classical monocytes and non-classical monocytes, respectively. In contrast, we found less than 20 differential genes within these clusters in any of the other patient pairs, suggesting a profound impact of S100A8/A9 expression on the transcriptome of these myeloid populations. Using CellPhone DB we studied the cell-cell communication mediated by ligand-receptor unique to SGZ by comparing the blast and the two monocytic populations. A number of unique interactions were seen at relapse, including CCL7-CCL3/CCR1, CCL3/IDE and CCL5/CCR5 known to induce the migration of leukocytes and myeloid cells. Other interactions involved in tumorigenesis and inflammation upregulated at relapse were seen: NRP2/VEGFA, LGALS9/SLC1A5.

Conclusion

Our results indicate that S100A8/A9 overexpression in B-ALL cells at relapse has no cell intrinsic impact but triggers remodeling of the TME by inducing an increase in myeloid subpopulations and unique cell to cell interactions.

Raetz:BMS: Other: Data and Safety Monitoring Board; Pfizer: Research Funding.