Chapter 3 │ Page 92 COMBOlow vs CDDPlow in Cal27 (p = 0.0441), it was striking to see that NTP-CDDP combinations persistently decreased the levels detected for IFN-β (Figure 10A and Suppl. Figure 2A-B). Progressing in the assessment of the pro-inflammatory interleukins IL-6 and IL-8, the data showed the same cell line-dependent variation in expression profiles. Indeed, SCC22B spheroids exhibited consistently low absolute levels of both IL-6 and IL-8, similar to what was previously observed for IFN-α (Suppl. Figure 2C-D). While IL-8 expression was higher in Cal27 and SCC61 compared to SCC22B, no significant treatment e ects were observed for this cytokine. IL-6, however, was predominantly upregulated by CDDP monotherapy, with peaks reaching up to 11.6-fold increases compared to untreated controls (p ≤ 0.0496) (Figure 10A and Suppl. Figure 2C). The expression of immune suppressive and tumor-promoting cytokine IL-1β, IL-10 and TNF-α was generally limited and showed some decreases in the NTP combination treatments compared to CDDP alone, although the di erences did mostly not reach statistical significance (expect for TNF-α in Cal 27, CDDPlow vs COMBOlow, p = 0.0035). This trend was particularly noticeable for IL-10 and TNF-α in Cal27 cells and IL-1β in both Cal27 and SCC61. TGF-β, notorious for inhibiting host immunosurveillance, was not detected in this experimental set-up (Suppl. Figure 2). Another aspect of immunogenicity of cancer cell killing is the production of various chemokines, especially important in immune cell attractance and migration [15, 4850]. We studied four HNSCC-relevant chemokines, including CXCL9, CXCL10, CCL4 and CCL5, and could conclude that NTP-CDDP treatment e ects were also limited in this context (Figure 10, Suppl. Figure 2I-L). In general, expression levels were diminished in treated spheroids compared to untreated controls, reaching significance for CXCL10 and CCL5 in the Cal27 and SCC61 cell type. Next, we aimed to gain a global and more profound understanding of the biological significance of this expression data (Figure 10B). By identifying clusters among the cytokine and chemokine profiles, we aimed to uncover the underlying biological
RkJQdWJsaXNoZXIy MTk4NDMw