1.06 Suppression of CXCL10/CXCR3 Switches Polymetastatic Phenotype to Oligo- in a Melanoma Mouse Model

S. C. Wightman1, A. Uppal1, G. Oshima1, X. Huang1, S. Ganai2, N. N. Khodarev1, M. C. Posner1, R. R. Weichselbaum1  1University Of Chicago,Chicago, ILLINOIS, USA 2Southern Illinois University,Carbondale, ILLINOIS, USA

Introduction:   Oligometastasis is a state of limited metastasis with the potential for long-term disease free control by surgery or radiotherapy. We designed syngeneic mouse melanoma models with oligo-  and polymetastatic pulmonary disease. We found CXCL10, an interferon inducible chemokine that acts on the receptor CXCR3, consistently elevated in polymetastatic tumor clones both in vivo and in vitro.

Methods:   Two stable CXCL10 KDs (knockdowns) and two stable CXCR3 KDs of both oligo- and polymetastatic B16F1 clones (P2M5B and P2M3C respectively) were generated by lentiviral transfection and suppression confirmed by Western blot.  Lung metastases were counted 2.5-3.5 weeks after tail-vein injection of the KD clones and non-targeting controls (NTCs). Flank primary tumors were generated via injection of the KD clones or NTCs and measured serially over 19 days. 

Results:  The P2M3C CXCL10 KD #1 and KD #2 had decreased metastases, 27.7+/-21.1 and 26.7+/-16.9, compared to P2M3C-NTC with 89.7+/-14.3 metastases (p<0.001).  Similarly, the P2M5B CXCL10 KD #1 and KD #2 had decreased metastases, 9.4+/-6.2 and 2.8+/-1.6, compared to P2M5B-NTC with 25.4+/-11.0 metastases (p<0.01).  For CXCR3, the P2M3C CXCR3 KD #1 and KD #2 had decreased metastases, 86.1+/-32.3 and 77.3+/-19.8, compared to P2M3C-NTC with 134.4+/-39 metastases (p = 0.03 and 0.004 respectively).  The CXCR3 P2M5B KD #2 had decreased metastases to 1.7+/-0.8, compared to P2M5B-NTC with 31.3+/-25.5 metastases (p=0.02).  The CXCR3 P2M5B KD #1, while with a decreased number of metastases, was not statistically significant with 14.7+/-7.5 (p = 0.12). After experiments were done with KDs, primary tumors were grown in the flanks of mice to identify differences in the CXCL10 KDs outside of tumor metastases.  At day 19, volume in both of the P2M3C CXCL10 KDs were decreased with KD #1 having an average volume of 0.31+/-0.29 cm3 and KD #2 having an average volume of 0.15+/-0.15 cm3 compared to the P2M3C-NTC with a volume of 4.94+/-1.63 cm3 (p<0.001 for both).  The P2M5B CXCL10 KDs also both decreased with KD #1 having an average volume of 2.10+/-0.73 cm3 and KD #2 0.60+/-0.07 cm3 compared to the P2M5B-NTC with a volume of 7.81+/-1.80 cm3 (p = 0.03 and 0.001 respectively).

Conclusion:  Our experiments stress the vital role of the CXCL10/CXCR3 axis in metastases generation and primary tumor growth in our murine melanoma model.  As noted above, knocking down the CXCL10/CXCR3 axis is able to change the phenotype from polymetastatic to oligometastatic effecting the ability for metastatic tumor colonization. In primary tumors, decreasing the CXCL10 levels decrease tumor growth.  This potentially offers a gateway for therapeutic intervention in patients with melanoma.  These findings stress the importance of CXCL10 in tumor biology for both metastases and primary tumors in murine melanoma.  Further experiments are needed to dissect mechanisms through which CXCL10 regulates pulmonary metastases.