Targeted Deletion of Cxcl1 in MSCs Regulates Osteogenesis and Suppresses Bone-Metastatic Prostate Cancer

Abstract

Bone metastasis continues to be the greatest challenge in treating patients with prostate cancer despite ongoing research. In bone, prostate cancer tumors hijack normal bone remodeling processes to drive cancer progression. However, it is unclear how these interactions drive bone-metastatic prostate cancer growth in the bone environment. To understand the mechanisms associated with bone-metastatic prostate cancer regulation of mesenchymal stem cells (MSC), we previously identified that bone-metastatic prostate cancer induces MSC expression of the pro-inflammatory chemokine CXCL8 and its mouse functional homologue Cxcl1. To date, there has been little to no information about the role of CXCL1/8 in MSC biology and its impact in the tumor-bone environment. Using genetic deletion of Cxcl1, we discovered a novel role for Cxcl1/8 in regulating MSC osteoblast differentiation, such that targeted deletion of Cxcl1 enhanced MSC osteoblastogenesis. Despite the osteogenic nature of prostate cancer, co-injection of Cxcl1 knockout (KO) MSCs with bone-metastatic prostate cancer in bone significantly suppressed tumor growth compared with co-injection with scrambled control (non-targeting) MSCs, even in the presence of three times more prostate cancer to MSCs. Furthermore, bulk RNA sequencing revealed immune response pathways, both in Cxcl1-KO MSCs and bone-metastatic prostate cancer tumors containing Cxcl1-KO MSCs. In support of this, Cxcl1-KO MSCs reduced immature neutrophils in the bone environment, while increasing monocytes. These findings demonstrate the importance of MSC-derived Cxcl1 in the bone microenvironment and highlight the importance of Cxcl1 in bone-metastatic prostate cancer progression.

Implications: MSC-derived Cxcl1 regulates prostate cancer progression in bone.

Fig. 1. Cxcl1 KO Increases MSC Osteoblast Differentiation.

a) Schematic showing generation of Crispr-Cas9 Cxcl1 KO MSCs; b) Cxcl1 measured via ELISA for Scr. Ctrl. and Cxcl1 KO MSCs; graph shows protein expression normalized to total protein; ANOVA statistical analysis was performed. c) Alizarin red staining of mineralization assay of Cxcl1 KO and Scr. Ctrl. MSCs +/− osteogenic supplement (OS); representative images of staining (top), Graph shows quantification of staining (left); ANOVA statistical analysis was performed. d) Alkaline phosphatase activity of Cxcl1 KO sgRNA 2 and Scr. Ctrl. MSCs measured via conversion of pNPP to colored pNP; ANOVA statistical analysis was performed of each condition at 40 minutes. e) RealTime qPCR of osteoblast genes Osx, Ocn, and Col1α1 expression in Cxcl1 KO sgRNA 2 and Scr. Ctrl. MSCs +/− OS; graphs represent relative fold change of gene expression normalized to r36B4. ANOVA statistical analysis was performed. f) RealTime qPCR measurement of CXCL8, RUNX2, and OPN gene expression following siRNA targeting of CXLC8 or non-targeted Scr. Ctrl. siRNA in human MSCs; graphs represent relative fold change of gene expression normalized to r18S. ANOVA statistical analysis was performed.

Fig. 2. Cxcl1 KO impacts CXCR1 expression and alters the MSC Transcriptome.

a) RealTime qPCR of Cxcr1 in Scr. Ctrl. MSCs and Cxcl1 KO MSCs (sgRNA2). Graph represents relative fold change of gene expression normalized to r36B4. Two sample t-test statistical analysis was performed. b) Left, Cxcr1 gene expression in parental MSCs following siRNA targeting or non-targeted siRNA; graph represents relative fold change of gene expression normalized to r36B4; two sample t-test statistical analysis was performed. Right, Alkaline phosphatase activity of Cxcr1 KD and Scr. Ctrl. MSCs +/− osteogenic supplement treatment measured via conversion of pNPP to colored pNP. Two sample t-test statistical analysis was performed at 120-minute time point; c) Gene Ontology (GO) analysis of upregulated (left) and downregulated (right) pathways in Cxcl1 KOsgRNA 2 MSCs compared to Scr. Ctrl. MSCs; d) c) Gene set enrichment analysis (GSEA) plots: Regulation of Runx2 Signaling and Activity, Collagen Degradation, Assembly of Collagen Fibrils and Other Multimeric Structures; e) Validation of Cxcl1-dependent Changes to the MSC Transcriptome: a) Realtime qPCR validation of RNA sequencing identified changes in Fbln1, Postn, Gas6, Parm1 gene expression in Cxcl1 KO MSCs (top) and Cxcl1 KD MSCs (bottom). ANOVA statistical analysis was performed.

Fig. 3. Cxcl1 KO MSCs Suppress RM1 Growth In Vitro and In Vivo.

a) RM1-luciferase cell proliferation measured via luciferase expression following overnight co-culture with an equal number of Cxcl1 KO sgRNA 2 or Scr. Ctrl. MSCs; ANOVA statistical analysis was performed; b) RM1 proliferation in response to Cxcl1 KO sgRNA 2 or Scr. Ctrl. MSC conditioned media; graph shows cell count using Trypan Blue exclusion assay; ANOVA statistical analysis was performed at Day 3; c) SA-AXL3 cell proliferation in response to Cxcl1 KO sgRNA 2 or Scr. Ctrl. MSC conditioned media; graph shows cell count using Trypan Blue exclusion assay; ANOVA statistical analysis was performed at Day 3 and each condition compared to each. ***p<0.0001 and comparing Scr. Ctrl. CM to SA AXL cells in serum-free media unless noted. **p<0.001; d) RealTime qPCR gene expression of RM1 Cxcr1 and Cxcr2; graph represents relative fold change of gene expression normalized to Hprt and Scr. Ctrl. MSCs; ANOVA statistical analysis was performed; e) Bioluminescence imaging of RM1-luciferase growth in an intratibial tumor model in C57BL/6 male mice co-injected Cxcl1 KO sgRNA 2 or Scr. Ctrl. MSCs (n=15/group) (left), graph represents quantitation of bioluminescence imaging (right); two sample t-test statistical analysis was performed at Day 9, 11, and 14; f) X-ray of tumor-bearing tibias; (left) representative images, (right) quantification of osteolytic pit area normalized to total bone marrow cavity area; g) Trichrome stain of tumor-bearing tibia: left, representative images (scale bar 500 um); right, quantification of staining normalized to marrow cavity area (top right); two samples t-test statistical analysis was performed.

Fig. 4. Cxcl1 KO MSCs suppress BM-PCa Growth and promote osteogenesis in vivo at 3:1 (PCa: MSCs) Ratio.

a) Bioluminescence imaging of RM1-luciferase growth in an intratibial tumor model in immunocompetent mice with 1 x 10⁴ Cxcl1 KO sgRNA 2 or Scr. Ctrl. MSCs and 1 x 10⁴ RM1 cells (left) or 5 x 10³ MSCs and 1.5 x 10⁴ RM1 cells (right); tumor burden (top); representative bioluminescence imaging of mice (bottom); two sample t-test statistical analyses were performed at Day 5, 7, and 9; b) Trichrome staining of tumor-bearing tibia sections from 3:1 (PCa: MSCs) mice: left, representative images quantification of staining normalized to marrow cavity area (scale bar = 500 μm); right, quantification of staining normalized to marrow cavity area; two sample t-test statistical analysis was performed; c) Fluorescent IHC of a-SMA in tumor-bearing tibia from both 1:1 and 3:1 (PCa: MSC) tumors (n=6 Scr. Ctrl (3 each 10K/15K); n=7 KO (3 of 10K, 4 of 15K)): left, representative images from 3:1 ratio, SMA-positive MSCs (gold), nuclei stained with DAPI (blue) (scale bar = 100 μm); right, graph shows quantitation of SMA pixel intensity; ANOVA statistical analysis was performed.

Fig. 5. Impact of Cxcl1 KO MSCs on C42B growth in bone.

Luciferase-expressing C42B were injected intratibially into male NSG mice at a 1:1 ratio of C42B with Scr. Ctrl. MSCs or Cxcl1 KO MSCs (3x10⁵) a) Representative bioluminescence images; b) graph shows quantitation of tumor growth rate per group (each mouse normalized to its luminescence (total flux) at first imaging, Day 7). Left, growth rate up to Day 35; right, growth rate for full study (49 days); c) MicroCT imaging of C42B tumor bones; left, representative images of C42B tumors containing Scr. Ctrl. MSCs (top) or Cxcl1 KO MSCs (bottom); right, graph shows quantitation of bone volume/tissue volume (BV/TV) in tumor vs. sham (saline-injected) bones; ANOVA statistical analysis was performed; d) microCT quantitation of trabecular (tr.) thickness, number, and separation. Two sample t-test statistical analysis was performed; e) Scr. Ctrl. or Cxcl1 KO MSCs (3x10⁵) or saline (sham) were injected into tibia of male NSG mice (n=5/group). After 19 days, bone marrow was flushed and immune populations assessed via flow cytometry. Graphs show flow cytometry of total immune cells (% live CD45.2 cells) (left), myeloid cells (middle), and neutrophils (right); f) Left, histogram shows delineation of immature and mature neutrophils based on expression level of Ly6G; Right, graph shows percentage of live Ly6G-low (Ly6Glo) expresser neutrophils per group; g) Left, density plot of monocyte populations; right, graph shows percentage of monocytes in live CD11b+ cells per each group. ANOVA statistical analysis was performed.