Stress-induced Norepinephrine Downregulates CCL2 in Macrophages to Suppress Tumor Growth in a Model of Malignant Melanoma

Abstract

Psychological stressors have been implicated in the progression of various tumor types. We investigated a role for stress in tumor immune cell chemotaxis in the B16F10 mouse model of malignant melanoma. We exposed female mice to 6-hour periods of restraint stress (RST) for 7 days, then implanted B16F10 malignant melanoma tumor cells and continued the RST paradigm for 14 additional days. We determined serum corticosterone and liver catecholamine concentrations in these mice. To evaluate the tumor microenvironment, we performed IHC and examined cytokine expression profiles using ELISA-based analysis of tumor homogenates. We found that tumors in mice subjected to RST grew significantly slower, had reduced tumor C-C motif ligand 2 (CCL2), and contained fewer F4/80-positive macrophages than tumors from unstressed mice. We observed a concomitant increase in norepinephrine among the RST mice. An in vitro assay confirmed that norepinephrine downregulates CCL2 production in both mouse and human macrophages, and that pretreatment with the pan-β-adrenergic receptor inhibitor nadolol rescues this activity. Furthermore, RST had no effect on tumor growth in transgenic CCL2-deficient mice. This study suggests that stress reduces malignant melanoma by reducing recruitment of tumor-promoting macrophages by CCL2.

Figure 1.. Stress is protective against tumor growth in a murine model of melanoma.

A) Female C57BL/6J mice were subjected to 6 hours of HCC or RST for 7 days. B16F10 tumor cells were implanted and HCC or RST continued for 14 more days. Tumors were palpable after 3 days and measured daily. B) Before RST (Baseline) and during RST (Days 7, 14, and 21), blood was collected to determine serum corticosterone. Mice subjected to RST had significantly higher levels of serum corticosterone compared to mice that were not stressed (HCC) as evaluated by ANOVA analysis for group effect (p=0.039). n=5 mice for both RST and HCC groups. Error bars represent ±SEM. C) Subsequent to HCC/RST, the livers were removed, snap frozen, weighed, homogenized in homogenization buffer and centrifuged. Final concentrations of norepinephrine were determined by standardizing to total liver weight. HPLC/EC was performed following standard procedure. Livers from the RST group had significantly more norepinephrine than the HCC mice (p=0.041). n=10 for both HCC and RST groups. Error bars represent ±SEM. D) Female C57BL/6J mice were subjected to RST for seven days then subcutaneously injected with 2.5×10⁵ B16F10 tumor cells. Stress was continued in the RST group until the tumors reached 2cm² in their largest diameter, until the mouse died, or for an additional 14 days. For statistical analysis, the data points from were converted to growth slopes by using the natural log (Ln) of tumor growth and the resulting lines fitted using Waldman Test parameters from random-slope regression comparing HCC to RST groups (inset) and doubling times of tumor growth for both groups were calculated (HCC=3.7 days; RST=6.5 days) (p=0.001). Upon sacrifice, the tumors were resected and measured with calipers (right). Tumors from the HCC group were significantly larger than tumors from the RST group (p=0.029). n=30 for both HCC and RST groups, experiment repeated 3 independent times. Error bars represent ±SD.

Figure 2.. Tumors from mice subjected to stress have reduced macrophage infiltration and CCL2.

A) Tumors were removed upon sacrifice and immunostained using an antibody specific for F4/80 antigen on macrophages. A representative tumor from the HCC group (top left and bottom left) reveals more F4/80(+)-stained cells than in a tumor from the RST group (top right and bottom) at the same magnification. B) Quantification of F4/80(+)-stained cells as percent of brown-stained cells per high powered field (HPF). T-test analysis suggests a statistical difference of p=0.017 for macrophage infiltration between the two groups. HCC=6 tumors and RST=8 tumors. C) Tumors from each group were flash frozen and homogenized in Trizol for total RNA isolation. Real-time PCR analysis for murine Ccl2 mRNA was performed. Tumors from mice subjected to RST had significantly less Ccl2 mRNA than tumors from mice left unstressed (HCC) (p<0.05). n=6 mice per group and error bars represent ±SEM. D) Tumors from each group were weighed, homogenized, centrifuged and supernatant collected and subjected to Bioplex analysis. After standardization using tissue weight, tumors from the RST group had significantly less CCL2 compared to tumors from the HCC group (p=0.031). n=6 for both HCC and RST (left). Error bars represent ±SEM. CCL2 protein directly correlate with F4/80+ cells/HPF (p<0.001, right).

Figure 3.. Reduced macrophage population in stress mice results in less Texas Red-dextran+ tumor blood vessels.

Prior to sacrifice, C57BL/6J mice bearing B16F10 tumors that were subjected to HCC or RST for 21 days were injected with Texas Red-dextran tumor blood vessels. Images representative of n=6 for both HCC and RST groups (left). Fluorescent microscopy and quantification of red pixels per high powered field of the entire tumor surface suggest there are less tumor vessels in the tumors of the RST mice than the HCC mice (p=0.036) (right). n=6 for both HCC and RST groups. Error bars represent ±SEM.

Figure 4.. Norepinephrine inhibits CCL2 production from macrophages, in vitro.

A) Macrophages from non-stressed female C57BL/6J mice were derived in serum-containing media. B16F10 tumor cells were co-cultured with the macrophages or cultured alone. Macrophages alone, macrophages co-cultured with B16F10 tumor cells, or the B16F10 tumor cells alone were left untreated (Macs; Macs+F10; F10), treated with 50 μM norepinephrine (Macs+NE; Macs+F10+NE; F10+NE), or pre-treated for 30 minutes with 150 μM nadolol prior to treatment with norepinephrine (Macs+Nad+NE; Macs+F10+nad+NE; F10+Nad+NE) and allowed to incubate for 24 hours at 37°C and 5% CO₂. The cell-free supernatants were collected and subjected to murine CCL2 ELISA. The macrophages alone and the macrophages co-cultured with B16F10 cells produced more CCL2 compared to cells incubated with norepinephrine (p=0.036 and p=0.017, respectively). Pre-incubation with nadolol rescued CCL2 expression from both macrophages treated with norepinephrine and the macrophage/B16F10 co-culture treated with norepinephrine (p=0.05 and p=0.038, respectively.) n=8 for CCL2 in vitro from murine macrophages. Error bars represent ±SEM. B) Human macrophages were differentiated over 5 days from peripheral blood monocytes. Macrophages alone (Macs), macrophages treated with 50 μM norepinephrine (Macs+NE 50 μM), or pre-treated for 30 minutes with 300 μM nadolol prior to treatment with norepinephrine (Macs+Nad 300 μM+NE 50 μM) and allowed to incubate for 24 hours at 37°C and 5% CO₂. Cell-free supernatants were collected and subjected to human CCL2 ELISA. Macrophages alone produced more CCL2 compared to those cells incubated with norepinephrine (p=0.0058). Pre-incubation with nadolol rescued CCL2 expression from macrophages treated with norepinephrine (p=0.0075). n=4 per group. Error bars represent ±SEM. C) Human macrophages prepared as described above were treated with complete media or complete media plus 50 μM norepinephrine (NE) or 5 nM NE for 24 hours. Both NE concentrations significantly suppressed NE production compared to control (p=0.0027 and p=0.0047). n=4 per group. Error bars represent ±SEM.

Figure 5.. CCL2-deficiency abrogates protection of stress on tumor growth.

A) Age-matched, female C57BL/6J CCL2−/− or wild type CCL2+/+ mice were subjected to HCC or RST for seven days then implanted with 2.5×10⁵ B16F10 tumor cells. The HCC/RST paradigm persisted for 14 more days. Repeated measures ANOVA demonstrates a significant difference in tumor growth between the wild type mice subjected to HCC and RST (p=0.001) while there was no difference in tumor growth in the CCL2−/− mice (p=0.764). n=7 for CCL2+/+ HCC; n=6 for CCL2+/+ RST; n=6 for CCL2−/− RST and n=8 for CCL2−/− HCC. Error bars represent ±SD. B) Tumors from the CCL2+/+ and CCL2−/− mice were resected upon sacrifice and homogenized into a single cell suspension. The cells were immunostained with an antibody targeting macrophage F4/80 and differences in F4/80-positivity between RST and HCC groups were quantified using flow cytometry. RST significantly suppressed macrophage influx into the tumors from the CCL2+/+ mice (p<0.05) while RST had no effect on macrophage recruitment in the CCL2-deficient mice (p=NS). n=7 for CCL2+/+ HCC; n=6 for CCL2+/+ RST; n=6 for CCL2−/− RST and n=8 for CCL2−/− HCC. Error bars represent ±SEM.

Figure 6.. Selective β-AdR agonists reduce macrophage CCL2, in vitro.

We derived murine bone marrow into macrophages from C57BL/6J mice and cultured in serum-media (Media alone); 50 μM norepinephrine (NE) alone or pre-treated for 30 minutes with 150 μM nadolol (NAD); 50 μM epinephrine (EPI) alone or pre-treated for 30 minutes with 150 μM nadolol. Less CCL2 was produced by macrophages treated with 50 μM NE and 50 μM EPI compared to the cells treated with media alone (n=4 for epinephrine and n=4 for nadolol+epinephrine; n=8 for norepinephrine and n=8 for nadolol+norepinephrine) (p=0.007 and p=0.001, respectively). Pre-treatment with 150 μM nadolol rescued suppression of CCL2 to levels similar to cells in media alone. Also, murine macrophages were cultured in media alone with 1, 10, or 100 μM of the β₁-AdR agonist dobutamine; 1, 10, or 100 μM of the β₂-AdR agonist, formoterol; 1, 10, or 100 μM of the α₁-AdR agonist methoxamine; or 1, 10, or 100 μM of the α₂-AdR agonist clonidine. The β-AdR agonists dobutamine and formoterol reduced CCL2 production in the supernatants in a dose-dependent manner compared to vehicle (1 μM, p=0.005; 10 μM and 100 μM, p=<0.001 for both dobutamine and formoterol). The α-AdR agonists methoxamine and clonidine had no effect on the down regulation of CCL2 at any concentration tested. n=8 for dobutamine, n=8 for formoterol, n=4 for methoxamine, and n=4 for clonidine. Error bars represent ±SEM.