Loss of PTH1R signaling in osteoblasts increases migration of 4T1 breast cancer cells in vitro. To determine whether migration of 4T1 breast cancer cells toward osteoblasts is affected by PTH1R signaling, primary calvarial osteoblasts were isolated from neonatal mice with a conditional ablation of either PTH1R (PTH1ROsxKO mice) or the Gs α subunit (Gsα) (GsαOsxKO mice) in osteoprogenitors, and cocultured with 4T1 cells in a transwell migration assay system. We plated 4T1 cells in the transwell inserts in serum-free medium, and they were allowed to migrate toward calvarial osteoblasts plated in the bottom chamber. Migration of 4T1 breast cancer cells was increased in the presence of primary calvarial osteoblasts from PTH1R control (PTH1Rfl/fl) or Gsα control (Gsαfl/fl) mice, as compared with wells with serum-containing medium only (Figure 1, A–D). Migration of 4T1 breast cancer cells was further increased toward primary calvarial osteoblasts lacking either PTH1R or Gsα, suggesting that PTH1R signaling in osteoblasts negatively regulates secreted promigratory factors acting on breast cancer cells.

Loss of PTH1R signaling in calvarial cells enhances migration of 4T1 cells in vitro and alters expression of target genes involved in the breast–bone vicious cycle. To study the effects of PTH1R knockdown in osteoblasts (OB), primary osteoblasts were isolated from neonatal calvariae of PTH1ROsxKO and GsαOsxKO mice, and migration assays were set up against 4T1 cells in transwell chambers. (A) Representative images and (B) numbers of 4T1 breast cancer cells that migrated toward osteoblasts from PTH1ROsxKO mice vs control (PTH1Rfl/fl) mice. All values represent the mean ± SEM of 6 individual experiments conducted in triplicate. Statistical significance was evaluated using 2-way ANOVA with Tukey’s test as the post hoc analysis. Scale bar: 200 μm. (C) Representative images and (D) numbers of 4T1 breast cancer cells that migrated toward osteoblasts from GsαOsxKO vs control (Gsαfl/fl) mice. All values represent the mean ± SEM of 6 individual experiments conducted in triplicate. Statistical significance was evaluated using 2-way ANOVA with Tukey’s test as the post hoc analysis. Scale bar: 200 μm. (E) Target gene expression in bone from PTH1ROsxKO mice (n = 3) compared with PTH1Rfl/fl (n = 3) using RNA-Seq analysis as described in Methods. The red dotted line represents expression in PTH1Rfl/fl mice set to 1. (F) Gene expression of Gnas (the gene encoding Gsα) and (G) Bglap (the gene encoding osteocalcin) in forelimbs of control and GsαOsxKO mice that were administered doxycycline in utero until weaning. All values represent the mean ± SEM. Statistical significance was evaluated using 1-way ANOVA with Tukey’s test as the post hoc analysis. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

We recently performed RNA-Seq on osteoprogenitors from bones of PTH1Rfl/fl and PTH1ROsxKO mice (29). We examined the expression pattern of several genes implicated in the breast–bone vicious cycle in the bone microenvironment and found that conditional ablation of PTH1R in osteoblasts is associated with increased expression of Pthlh, Cxcl12, Tgfb2, Tgfb3, Tnf, Flt1 (encoding VEGFR1), and Vcam1 (Figure 1E). Osteoblasts from PTH1ROsxKO mice had significantly lower levels of Pth1r mRNA. Gene expression analysis revealed that the expression of Gnas, which encodes Gsα, was significantly lower in the bones of GsαOsxKO mice (Figure 1F). Gnas mRNA is still present at a low level because of ubiquitous expression of Gsα within non–Cre-targeted cells. We have previously demonstrated that conditional ablation of Gsα in osteoprogenitors disrupts PTH1R signaling (24). Consistent with this, the expression of PTH target osteocalcin (encoded by Bglap) is significantly decreased in GsαOsxKO bone (Figure 1G). These changes suggest that the lack of PTH1R or disruption of its signaling pathways significantly alters the bone microenvironment in both the PTH1ROsxKO and GsαOsxKO mice.

Ablation of Gsα, a downstream mediator of PTH1R signaling in bone, abolishes the inhibitory effects of PTH treatment on breast cancer bone metastasis. To study the role of PTH1R signaling in breast cancer bone metastasis, we attempted to generate PTH1ROsxKO mice that survive to adulthood. In both PTH1ROsxKO and GsαOsxKO mice, constitutive expression of Osx-driven Cre recombinase throughout embryonic development results in early death (30–32). However, the osterix promoter–driven GFP:Cre fusion protein is regulated by a Tet-Off tetracycline transactivator (33), and administration of doxycycline in drinking water to pregnant females from mating until delivery can delay Cre recombinase expression until after birth (24). We backcrossed PTH1ROsxKO mice to a Balb/c background to provide a syngeneic background for the 4T1 tumor injections. Unfortunately, PTH1ROsxKO mice on the Balb/c background could not be rescued by doxycycline-mediated delay of PTH1R deletion until postnatal life. We have previously shown that in the GsαOsxKO mice, the anabolic actions of PTH on bone are mediated by Gs (24). Hence, we have used GsαOsxKO mice backcrossed onto a Balb/c background for our studies.

We injected 4T1 breast cancer cells into the mammary fat pads of 10-week-old control (Gsαfl/fl) and GsαOsxKO mice. PTH treatment was initiated 24 hours later (Figure 2A), and mice were monitored for the next 4 weeks. GsαOsxKO mice were significantly smaller than their control littermates throughout the study period, and PTH treatment did not affect the body weights of either control or GsαOsxKO mice (Figure 2B). Palpable tumors were detected in all groups by week 1, and tumor volumes were similar in all 4 groups, reaching approximately 800 to 1000 mm3 (Figure 2C) by the end of the treatment period. PTH treatment did not affect the growth of the tumors in GsαOsxKO mice or Gsαfl/fl littermate controls. Mice were euthanized 4 weeks after tumor implantation, and individual organs were dissected and underwent bioluminescence imaging (BLI) to detect presence of metastases (Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.157390DS1). We confirmed that bioluminescence correlated with metastasis by histological analysis of paraffin-embedded sections from GsαOsxKO mice (Supplemental Figure 2). Metastases to the lungs, liver, and spleen were similar in both the PBS-treated control and GsαOsxKO mice, and intermittent PTH had no additional effect (Table 1 and Figure 2, D and E). As expected, PTH treatment significantly reduced bone metastases to the hind limbs in control mice (n = 3 of 12) compared with the PBS-treated control mice (n = 8 of 14). However, in GsαOsxKO mice, PTH administration had no effect on bone metastases, suggesting that PTH1R signaling is required in osteoblasts for PTH to decrease bone metastasis.

Ablation of Gsα, a downstream target of PTH1R signaling in bone, abolishes the beneficial effects of PTH treatment on skeletal metastasis with breast cancer. (A) Experimental design. Mice were 10 weeks old at the time of 4T1 cell injection. PBS or intermittent PTH (80 mg/kg) was administered Monday through Friday. (B) Weekly BW measurements and (C) weekly tumor volume measurements in control (Gsαfl/fl) and GsαOsxKO mice treated with PBS or PTH. (D) Representative bioluminescent images of metastases to hind limbs in Gsαfl/fl and GsαOsxKO mice treated with PBS or PTH (80 mg/kg/d, Monday through Friday) (E) Quantification of BLI in lungs, liver, spleen, and hind limbs with metastases. All values represent the mean ± SEM of 10 mice for each group. Statistical significance was evaluated using 2-way ANOVA with Tukey’s test as the post hoc analysis. *P < 0.05 when compared with respective control. Bg, background.

Summary of number of metastases per the total number of control (Gsαfl/fl) and GsαOsxKO mice treated with PBS or intermittent PTH

Decreased expression of Pth1r is associated with progression and metastasis of breast cancer in both mice and humans. PTH1R is also expressed by breast cancer cells (20, 22). To examine the relationship between PTH1R expression and breast cancer progression, we analyzed publicly available microarray data sets in the National Center for Biotechnology Information (NCBI) database. A total of 10 data sets using murine models of breast cancer, human cell lines, or patient samples were analyzed for PTH1R expression. All 3 murine studies analyzed showed significant decreases in Pth1r expression in breast cancer samples compared with normal breast tissue (Figure 3A). In the first study (34) , using 8 different genetically engineered murine models of spontaneous breast cancer, Pth1r was significantly lower in the breast tumor tissue compared with normal mammary gland from pregnant mice on mixed background (Figure 3A). In needle aspirates of bone metastases (35), Pth1r expression was decreased compared with the 4T1-derived primary tumor. In another murine study (36), involving transgenic mice that synthesize the large T antigen under the control of the whey acidic protein promoter–TNP8 mice resulting in multifocal ductal carcinoma in situ and, ultimately, invasive ductal breast carcinoma (IDC), Pth1r levels were significantly lower in IDC than in breast tissue from WT nonlactating mice (Figure 3A).

Microarray analyses of publicly available data sets for the expression of Pth1r in breast cancer. Publicly analyzed microarray data sets from (A) murine breast cancer studies, (B) human breast cancer cell lines, and (C and D) studies of human patients with breast cancer were analyzed for the expression of Pth1r in murine and human breast cancer tissues using the GEO2R (an online data analysis tool used to analyze GEO data sets under the same experimental conditions) as described in Methods. Values are represented as log2 fold change in comparison to their respective controls, as indicated. *P < 0.05, ***P < 0.001, ****P < 0.0001 when compared to their respective controls. Mets, metastasis.

Results from the 7 analyzed human studies were more varied (Figure 3, B–D), with 4 studies (37–39) reporting reduced levels of PTH1R expression and 3 studies (40–42) reporting no change. PTH1R expression in the bone variant MDA-MB-231 human breast cancer cell line, characterized by highly metastatic clones that preferentially migrate to the bone, did not differ from the parental cell line in 1 study. In a second study using the same MDA-MB-231 human breast cancer cell line, PTH1R expression was significantly decreased in cells from bone metastases compared with naive bone (Figure 3B). In 2 of the 3 studies involving patients with IDC, PTH1R expression was significantly lower in cancerous tissue compared with adjacent, normal, healthy breast stroma (Figure 3C). In 2 studies of patients with bone metastases, PTH1R expression in 1 study was significantly lower in bone with metastases compared with normal bone, but PTH1R expression did not differ between bone metastases and the primary tumor in the other study (Figure 3D). In no instance was an increase in Pth1r/PTH1R expression associated with progression of disease. More importantly, we did not find any cases in which reduced PTH1R expression was associated with a better clinical outcome. Together, these analyses of microarray data sets reveal reduced PTH1R expression is often associated with breast cancer progression and bone metastasis.

Knockdown of Pth1r in 4T1 cells diminishes responsiveness to intermittent PTH treatment in vivo. To examine the role of PTH1R in breast cancer and bone metastasis, we knocked down Pth1r expression in 4T1 cells (Pth1rKD-4T1) by shRNA. Pth1r mRNA levels were decreased 86% in Pth1rKD-4T1 cells compared with the LMP vector control cells (Cont-4T1), but Pthlh mRNA levels were unchanged (Figure 4A). Basal cAMP levels did not differ between Pth1rKD-4T1 and Cont-4T1 cells. Although a robust increase in cAMP was elicited with a single dose of dibutyryl cAMP (500 μM) in both the Cont-4T1 and Pth1rKD-4T1 cells, PTH treatment (single dose, 10 nM) increased cAMP production in Cont-4T1 but not Pth1rKD-4T1 4T1 cells. PTH had no effect on the proliferation of Cont-4T1 or Pth1rKD-4T1 cells as measured by DNA content (Figure 4B). Migration of Pth1rKD-4T1 breast cancer cells toward calvarial osteoblasts treated with PBS was unchanged in transwell assays compared with the Cont-4T1 cells (Figure 4, C and D). Although PTH treatment reduced the migration of the Cont-4T1 cells toward the calvarial cells, it had no effect on Pth1rKD-4T1 cells. Collectively, the data demonstrate that knockdown of Pth1r rendered the Pth1rKD-4T1 cells resistant to the effects of PTH in vitro.

Pth1r knockdown in 4T1 cells attenuates the inhibitory effect of PTH on migration toward osteoblasts. (A) Relative mRNA expression (n = 9) (left panel) and cAMP assay (n = 8) (right panel). *P < 0.05 when compared to PBS-treated controls. ***P < 0.001 when compared to Cont-4T1. (B) DNA content in Cont-4T1 and Pth1rKD-4T1 cells (n = 9). *P < 0.05. (C and D) Representative images and numbers of Cont-4T1 and Pth1rKD-4T1 breast cancer cells that migrated towards calvarial osteoblasts treated with PTH or PBS in transwell assays (n = 12). All values represent mean ± SEM of experiments conducted in triplicate. Statistical significance was evaluated using 2-way ANOVA with Tukey’s test as the post-hoc analysis. ***P < 0.001, ****P < 0.0001. DiBcAMP, dibutyryl cAMP. Scale bar: 200 μm.

To determine whether PTH1R signaling in breast cancer cells is required for bone metastases, intramammary orthotopic tumors of Cont-4T1 or Pth1rKD-4T1 cells were established in 10-week-old female Balb/c mice. Treatment with intermittent PTH (80 μg/d/kg BW administered Monday through Friday) or vehicle (PBS) (Supplemental Figure 3A) was started 24 hours after tumor injection and continued for 4 weeks. BWs were monitored weekly, and no changes were evident with intermittent PTH treatment in mice with Cont-4T1 or Pth1rKD-4T1 tumors compared with their respective PBS-treated control groups (Supplemental Figure 3, B and C). Tumor volumes were equivalent in mice bearing either Cont-4T1 or the Pth1rKD-4T1 tumors (~800 to 900 mm3) at the end of 4 weeks by BLI and digital caliper measurements (Supplemental Figure 3, D–F). PTH treatment did not affect growth of either the Cont-4T1 or Pth1rKD-4T1 tumors when compared with PBS treatment. Four weeks after tumor implantation, mice were euthanized, and individual organs underwent BLI to determine the presence of metastasis (Figure 5A and Supplemental Figure 4). Metastases to liver, spleen, and lungs were similar in both number and BLI intensity in all groups of mice (Figure 5B) and did not change with PTH treatment. Bone metastases to hind limbs were also comparable in mice bearing Cont-4T1 or Pth1rKD-4T1 tumors with PBS treatment. However, PTH treatment significantly reduced the incidence of bone metastasis in mice bearing Cont-4T1 tumor but not in mice bearing Pth1rKD-4T1 tumors (4/20 vs 11/20) (Table 2 and Figure 5B). These results confirm that knockdown of Pth1r in 4T1 breast cancer cells renders them nonresponsive to the effects of PTH treatment.

Treatment with intermittent PTH does not reduce skeletal metastasis in orthotopic models of tumors lacking Pth1r. (A) Representative bioluminescent images of metastases to hind limbs in mice bearing either Cont-4T1 or Pth1rKD-4T1 tumors treated with PBS or intermittent PTH. (B) Quantitation of BLI in lungs, liver, spleen, and hind limbs with metastases. All values represent the mean ± SEM of 20 measurements for each group. Statistical significance was evaluated using 2-way ANOVA with Tukey’s test as the post hoc analysis.

Summary of number of metastases per the total number of mice bearing either Cont-4T1 or Pth1rKD-4T1 tumors treated with PBS or intermittent PTH

Gene changes with PTH treatment in the tumor–bone microenvironment is abolished with PTH1R knockdown in 4T1 cells. Next, we examined the effects of PTH1R knockdown in breast cancer cells on the tumor–bone microenvironment. RNA was isolated from tumors and long bones of tumor-bearing mice (Cont-4T1 or Pth1rKD-4T1) treated with either vehicle or PTH for 4 weeks. Analyses of key target genes involved in the tumor–bone vicious cycle showed similar patterns of gene expression in the vehicle-treated Cont-4T1 and Pth1rKD-4T1 tumors except for Il6, Ackr3 (encoding CXCR7), and Cyp19a1, which were reduced in the Pth1rKD-4T1 cells (Figure 6A). Intermittent PTH treatment reduced the expression of several of these key genes (Pthlh, Il6, Cyp19a1, Ackr3, and Tnf) in the Cont-4T1 tumors. These changes, however, were not seen in the Pth1rKD-4T1 tumors treated with intermittent PTH. Cdkn1A (encoding the protein p21), a known marker for cell cycle arrest, remained unchanged. It is interesting to note that PTHrP (encoded by Pthlh), whose actions are also mediated through the PTH1R, is significantly upregulated in the Pth1rKD-4T1 cells with intermittent PTH treatment. Target gene expression analyses of bones from mice bearing Cont-4T1 tumors showed reciprocal changes in the expression of several genes involved in the tumor–bone vicious cycle (Figure 6B). Significant decreases in the expression of Tgfb, Mmp13, Flt1, Vcam1, and Cxcl12 were seen in response to PTH treatment. Additionally, significant reduction in Tnfsf11 (encoding RANKL) expression accompanied by an increase in Tnfrsf11b (encoding OPG) expression resulted in a significant decrease in the RANKL to OPG ratio. Basal expression of various target genes was similar in the mice bearing either the Cont-4T1 or Pth1rKD-4T1 tumors with PBS treatment. PTH treatment, however, did not induce any changes in the expression of various target genes studied except for Vcam1 in the mice bearing the Pth1rKD-4T1 tumors. These results suggest that the tumor Pth1r expression is critical for the actions of PTH in reducing bone metastasis. Knockdown of Pth1r in tumors rendered both the tumors and the bone nonresponsive to the actions of intermittent PTH.

Target gene expression in primary tumors and bone of mice injected with Cont-4T1 and Pth1rKD-4T1 cells. (A and B)Target gene expression analysis in primary tumors and hind limbs of mice bearing Cont-4T1 and Pth1rKD-4T1 tumors treated with either PBS or intermittent PTH for 4 weeks, using gene-specific primers as described in Methods. All values represent the mean ± SEM of 10 mice for each group. Statistical significance was evaluated using 2-way ANOVA with Tukey’s test as the post hoc analysis. *P < 0.05, **P < 0.01, ***P < 0.001, when compared with PBS-treated Cont-4T1 mice. +P < 0.05, ++P < 0.01 when compared with the Pth1rKD-4T1 PBS group.

PTHrP is a key mediator of the actions of PTH on breast cancer cell migration. PTHrP (encoded by Pthlh) is a key mediator of the bone–breast vicious cycle, and treatment with intermittent PTH significantly decreased Pthlh expression in Cont-4T1 primary tumors while increasing the expression of Pthlh in Pth1rKD-4T1 primary tumors (Figure 6A). To determine whether suppression of PTHrP is required for inhibition of 4T1 breast cancer cell migration by PTH treatment, we knocked down Pthlh in 4T1 cells (PthlhKD-4T1 cells) using siRNA. Migration toward MC3T3-E1 osteoblastic cells in a transwell system was significantly reduced in PBS-treated PthlhKD-4T1 cells compared with PBS-treated control Cont-4T1 cells. Treatment with intermittent PTH further decreased migration (Figure 7, A–C). Conversely, overexpression of Pthlh in 4T1 cells increased migration toward MC3T3-E1 cells and rendered them resistant to the effects of intermittent PTH treatment (Figure 7, D–F). Overall, these results demonstrate that PTHrP is an important mediator of breast cancer cell migration toward the bone and suggest that inhibition of migration by PTH is mediated, at least in part, by suppression of PTHrP expression.

Pthlh expression is a key mediator of the actions of PTH on breast cancer cell migration. (A) Relative Pthlh mRNA levels in 4T1 cells with Pthlh knockdown (PthlhKD-4T1) using siRNA. (B) Representative images and (C) numbers of Cont-4T1 and Pth1rKD-4T1 breast cancer cells that migrated toward MC3T3 osteoblasts treated with PBS or PTH in transwell assays. Scale bar: 200 μm. (D) Relative Pthlh mRNA levels in PthlhOE-4T1 cells. (E) Representative images and (F) numbers of Cont-4T1 and PthlhOE-4T1 breast cancer cells that migrated toward MC3T3 osteoblasts treated with PBS or PTH in transwell assays. All values represent the mean ± SEM of at least 6 individual experiments conducted in triplicate. Scale bar: 200 μm. Statistical significance was evaluated using 2-way ANOVA with Tukey’s test as the post hoc analysis. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 when compared with the PBS-treated control. ++P < 0.01 when compared with the PTH-treated control.