The Role of an Inflammatory Phenotype in Driving Clonal Evolution in B-Acute Lymphoblastic Leukemia

Introduction: Although outcomes for children with B-acute lymphoblastic leukemia (B-ALL) have improved dramatically, survival for the 10-20% of children who relapse remains dismal. Our laboratory focuses on discovering genetic and epigenetic alterations leading to relapse. We discovered a novel superenhancer adjacent to S100A8 and S100A9 shared by a number of patients at relapse, implicating a role in driving resistance (Saint Fleur-Lominy et al., Cancer Res, 2020). S100A8/A9 are members of the S100 calcium-binding family and have been shown to mediate intrinsic drug resistance in cancer. However our previous work using engineered cell lines showed that S100A8/A9 expression alone did not influence growth, drug sensitivity or clonal potential. We reasoned that other co-regulated genes might be driving the clonal advantage. Single cell RNAseq data on 4 diagnosis/relapse pairs, 1 of which has increased S100A8/A9 at relapse, suggests that interleukin-6 receptor (IL6R) is upregulated with S100A8/A9. Pathway analysis on differential genes unique to the patient with upregulation of S100A8/A9 revealed inflammatory pathways, including IL6/JAK/STAT3. Our hypothesis is that S100A8/A9-expressing blasts interact with the tumor immune microenvironment (TIME) via the IL6/IL6R axis to create an inflammatory response that ultimately leads to resistance.

Methods: Pearson correlation coefficient was calculated to measure the linear correlation between the log10 fold change in expression of IL6R and S100A8/A9 from relapse to diagnosis using our previously published microarray datasets (Hogan et al., Blood, 2011). We also examined co-expression on bulk RNAseq data of 176 paired diagnosis/relapse samples from children enrolled on COG AALL1331. To determine the impact of IL6R expression on drug resistance, a panel of isogenic B-ALL cell lines (697, RS4;11, Nalm6 and RCH) was generated expressing empty vector (EV), S100A8/A9 (89), with and without IL6R and used to assess various phenotypic changes indicative of a clonal advantage. Flow cytometry was used to confirm surface expression of IL6R. Western blot was performed to assess downstream IL6R signaling with probes for total STAT3, phosphorylated STAT3 (pSTAT3) and actin as loading control. Cell viability in response to drugs used in B-ALL therapy was assessed by Promega’s CellTiter-Glo®. Apoptosis was measured using Annexin V and DAPI staining.

Results: Our bulk microarray expression data from primary patient pairs showed significant positive correlation between S100A8/A9 and IL6R (Pearson r=0.34, p=0.006 for S100A8 and IL6R; r=0.47, p=0.0001 for S100A9 and IL6R). We validated co-regulation of IL6R and S100A8 (corr 0.46, padj=0.0014) and S100A9 (corr 0.52, padj<1e-6) in an unbiased way using bulk RNAseq data from 176 paired samples. A total of 671 genes were found significantly co-regulated with both S100A8 and S100A9, including other genes involved in inflammation mediated resistance in AML (e.g., TLR4, ITGAM, CLEC7A). While we did not observe significant differences in chemosensitivity in the S100A8/A9 overexpressing cell lines, we did find that co-expression with IL6R provides a clonal advantage to cell lines under glucocorticoid exposure, a hallmark of disease relapse in 3 out of 4 cell lines tested. The IL6R mediated resistance was only observed when cells were exposed to IL6, either by culturing with HS-5 conditioned media (CM) or recombinant IL6 (rIL6). We confirmed activation of downstream signaling (pSTAT3) via the IL6-IL6R axis following growth in 0.01% IL6 CM or 100pg/ml rIL6 for 24 hours. Inhibition of pSTAT3 activation was observed when cells were pre-treated with 100ng/ml tocilizumab (toci) for 4 hours. Pre-treatment with toci led to partial restoration of drug sensitivity to prednisolone in the EV_IL6R and 89_IL6R cell lines (p<0.05). Likewise inhibition of pSTAT3 activation, as well as re-sensitization, were observed when cells were pre-treated with 100nM ruxolitinib for 4 hours.

Conclusions: Our data suggests that tumor escape is mediated by the acquisition of an inflammatory phenotype that includes increased IL6R expression on blasts activated by IL6 secretion from the surrounding TIME. S100A8/A9 secretion is known to remodel the TIME and we are testing whether this remodeling also leads to a protective niche in B-ALL. Inhibition of STAT signaling in this subgroup of relapsed patients may be a promising avenue to pursue.