1. NLRP3 agonist improves radiation-induced primary tumor control.
We first explored the efficacy of combining the NLRP3 agonist with different doses of XRT on tumor growth in vivo. We established two bilateral tumor models, 344SQ-P and 344SQ-R, and delivered unilateral XRT to primary tumors, intratumor NLRP3 agonist, and intraperitoneal α-PD1 injections, as shown in Figure 1a. 1 Gy XRT, either with or without NLRP3/α-PD1, was unable to extend survival (all P>0.05 vs. XRT, supplementary figure S1a). Both 12 Gy and 5 Gy XRT + NLRP3 ± α-PD1 prolonged survival in both the 344SQ-P (all P<0.01 vs. XRT, Figure 1b, supplementary figure S1b) and 344SQ-R groups (all P<0.01 vs. XRT, Figure 1c, supplementary figure S1c). Consistent with the survival data, no enhanced tumor control was achieved for 1 Gy XRT, either with or without NLRP3/α-PD1 (all P>0.05 vs. XRT, supplementary figure S1d). Addition of the NLRP3 agonist significantly improved tumor control of 12 Gy and 5 Gy XRT in both 344SQ-P (P<0.0001 vs. XRT, Figure 1d; P=0.0015 vs. XRT, supplementary figure S1e) and 344SQ-R models (all P<0.0001 vs. XRT, Figure 1e, supplementary figure S1f). Moreover, the 12 Gy + NLRP3 agonist achieved better primary tumor control than the 5 Gy treatment regimen in 344SQ-P model. The addition of α-PD1 did significantly improve performance of the NLRP3 agonist in combination with 5 Gy XRT in both tumor models (P=0.017 and P<0.0001, respectively, supplementary figure S1e and S1f); however, α-PD1 did not boost the strength of 12 Gy XRT + NLRP3 combination in regards to primary tumors (all P>0.05, Figure 1d and 1e). Therefore, the combination of XRT + NLPR3 agonist halted the growth rate of primary tumors in a radiological dose-dependent manner.
2. NLRP3 agonist increases inflammasome-related immune priming.
The NLRP3 inflammasome is a key component of the pro-inflammatory response of macrophages [13]. NLRP3 is also expressed in tumor cells and is related to immune resistance through PD-L1/NLRP3 inflammasome signaling [14]. It has been well documented that radiation can trigger NLRP3 inflammasome activation via multiple mechanisms [15]. Hence, we assessed the effect of XRT on NLRP3 expression at the mRNA level in macrophages and tumor cells separately in vitro. XRT enhanced NLRP3 expression in peritoneal macrophages in a dose-dependent manner (Figure 2a). In the tumor cells, Nlrp3 was upregulated 24h after XRT, with expression peaking following a dose of 6 Gy at this time point. This increase was not significant in 344SQ-P cells (Figure 2b), but it was in 344SQ-R cells (P<0.05, Figure 2c).
To explore the mechanism by which the combination of XRT and the NLRP3 agonist improved primary tumor control, we isolated RNA from the primary tumors of both lines and analyzed the differential regulation of 770 immune related genes using the nCounter PanCancer Immune Profiling Panel from NanoString. Since the XRT regimen of 12 Gy × 3 fractions elicited the best response, we used this dose moving forward. The NLRP3 agonist significantly increased the expression of Nlrp3 and the inflammasome cytokines Il1b and Il18 relative to untreated controls or mice treated with XRT in 344SQ-P tumors (P<0.05, Figure 2d). The addition of α-PD1 to XRT + NLRP3 further increased the expression of Il1b (P=0.010) and Il18 (P=0.014) relative to XRT + NLRP3 group. Unlike in the 344SQ-P tumors, XRT alone was able to significantly elevate the expression of these genes, and neither addition of the NLRP3 agonist alone nor the agonist in concert with α-PD1 was able to further significantly increase this (Figure 2e). Baseline expression of Nlrp3 and Il18 was significantly higher in 344SQ-R compared to 344SQ-P (both P<0.05, supplementary table S1), so the lack of upregulation in 344SQ-P cells treated with XRT alone was not due to Nlrp3 levels already being elevated. Thus, 344SQ-R cells had higher initial expression of Nlrp3, and this expression, as well as the expression of the inflammasome-related cytokines Il1b and Il18, were further increased following agonism of NLRP3.
Most measured immune pathways were upregulated in the XRT + NLRP3 ± α-PD1 groups, as compared to control or XRT alone, and this was true for both tumor lines (supplementary figure 2a and 2b). In particular, the pathways related to innate immunity, antigen processing, macrophage function, interleukins, adaptive immunity, and T cell function were drastically enhanced in the treatments containing XRT + NLRP3 in both 344SQ-P (Figure 2f) and 344SQ-R (Figure 2g) models. Predictably, the addition of α-PD1 to XRT + NLRP3 reflected enhanced effect on gene expression in the 344SQ-P model vs. 344SQ-R.
3. NLRP3 agonist promotes abscopal responses.
The benefit of the NLRP3 agonist was not only limited to the primary tumors, but also impacted the growth of secondary, unirradiated tumors. As with the primary tumors, both 12 Gy and 5 Gy XRT in combination with NLRP3 agonist significantly depressed the growth of secondary tumors in the 344SQ-P model (Figure 3a, supplementary figure S3a) and 344SQ-R (Figure 3b, supplementary figure S3b) models (all P<0.05 vs. XRT). Surprisingly, this abscopal response was dramatically enhanced by the addition of α-PD1 to XRT + NLRP3 in the 344SQ-R (P=0.001, vs. XRT + NLRP3, Figure 3b), but the same was not seen in the 344SQ-P (P=0.321, vs. XRT + NLRP3, Figure 3a). These treatments also strongly reduced lung metastases in 344SQ-P and 344SQ-R models. Both 12 Gy and 5 Gy XRT regimens in concert with the NLRP3 agonist and α-PD1 strongly reduced lung metastases in 344SQ-P and 344SQ-R models (Figure 3c and 3d, supplementary figure S3c).
In addition to secondary tumor shrinkage and reduced lung metastases, multiple pro-inflammatory cytokines were elevated in the serum of mice treated with XRT+NLRP3 or triple therapy, as assessed by multiplex ELISA. IL-1b, IL-4, IL-12, IL-17, IFN-γ and GM-CSF were significantly increased in the combination of 12Gy XRT and NLRP3 agonist compared to XRT alone in 344SQ-P model (all P<0.05, Figure 3e). Although none of these cytokines were significantly increased in XRT + NLRP3 as to XRT alone in 344SQ-R model, IFN-γ and GM-CSF were drastically increased by the addition of α-PD1 to XRT + NLRP3 (all P<0.05, Figure 3f).