Polarization of THP-1 cells into alternatively activated M2 macrophages is associated with reorganization of the microtubule cytoskeleton
Initially, we differentiated THP-1 monocytes into M0 macrophages using 100 nM PMA. The differentiated macrophages were polarized to the alternatively activated M2 phenotype by incubation with IL-4 and IL-13 (20 ng/ml each) [36]. Immunofluorescence analysis confirmed strong expression of CD68 (a classical pan macrophage marker) on the surface of the differentiated cells compared to THP-1 monocytes (Supplementary data, Fig. S1a). Furthermore, the quantification of immunofluorescence images showed that CD68 was expressed in 79.66 ± 3.21% and 81.3 ± 3.78% of M0 and M2 macrophages, respectively, compared with 46.3 ± 5.67% of monocytes (Supplementary data, Fig. S1b). Quantitative RT‒PCR confirmed that CD68 transcripts were significantly upregulated in M0 and M2 macrophages compared with monocytes (Supplementary data, Fig. S1c). Next, we determined the expression of a panel of cytokines and growth factors associated with the M2 polarized phenotype, including the surface markers CD206 (mannose receptor C-type 1) and CD163 (a scavenger receptor). Immunofluorescence analysis showed increased expression of CD206 in M2 macrophages (78.6 ± 4.04%) compared to that in M0 macrophages (29 ± 2%) (Supplementary data, Fig. S1d, e). Furthermore qRT‒PCR results showed an increase in the expression of both CD206 and CD163 in M2-polarized macrophages in comparison to M0 macrophages (Supplementary data, Fig. S1 f, g). Flow cytometry also confirmed increased expression of CD163 in M2 macrophages compared with that in M0 macrophages (Supplementary data, Fig. S1h). M2 TAMs modulate the tumor microenvironment for tumor growth and metastasis by secreting various tumorigenic factors, including proteases, cytokines, and chemokines, such as matrix metalloproteases (MMPs), VEGF-A, IL-10, and TGF-β [4, 37]. As expected, qRT‒PCR analysis showed an increase in the cytokines CCL18, CCL22, TGF-β, MMP9, MMP2, fibronectin, IL-10, and arginase in M2 macrophages compared to M0 macrophages (Supplementary data, Fig. S1i). Gelatin zymography showed an increase in the activity of secreted MMP9 in M2 macrophages compared to M0 macrophages (Supplementary data, Fig. S1j). These results suggest that THP-1cells were polarized into M2 macrophages, which exhibit intrinsic traits of tumor-associated macrophages.
Reportedly, LPS-mediated classical activation of macrophages is associated with an increase in acetylated microtubules in the cytoplasm [37]. Hence, we analyzed the status of microtubules in M0 and M2 macrophages. Our immunofluorescence data demonstrated that M0 and M2 macrophages exhibit a well-developed and spread network of microtubule filaments (Fig. 1a), with a microtubule network more elaborately organized in M2 macrophages. However, monocytes, which are mainly spherical, exhibit peripheral α-tubulin staining with essentially no microtubule network. We also examined the acetylation level of microtubules (Lys 40 of α-tubulin) as a marker of microtubule stability [38]. Interestingly, we confirmed that acetylation of α-tubulin at lysine 40 residue was significantly higher in M2 macrophages than in resting M0 macrophages (Fig. 1b, c).
2-Methoxyestradiol (2ME2) suppresses M2 polarization of macrophages
To elucidate the effect of 2ME2 on M2 macrophages, we first evaluated the effect of 2ME2 on cell viability. The results of the MTT assay showed that 2ME2 inhibited the viability of M2 TAMs in a concentration-dependent manner. The cell viability was not discernibly affected up to a concentration of 20 µM 2ME2, while cell viability was significantly reduced at higher concentrations (50 and 100 µM) (Fig. 2a). Additionally, annexin-V/PI staining confirmed that 2ME2 treatment did not induce significant cell death of M2 TAMs at low concentrations (5 and 20 µM) (Fig. 2b). Subsequent experiments were carried out with 5 and 20 µM 2ME2 to determine its cell death-independent effects.
Since 2ME2 is a well-known microtubule-depolymerizing agent, we analyzed the status of microtubules in 2ME2-treated TAMs. Our data revealed that 5 µM 2ME2 partially depolymerized the microtubule network, whereas 20 µM 2ME2 led to a complete disruption of the microtubule network, concomitantly reducing microtubule acetylation (Fig. 2c). Similarly, Western blotting results also showed a decrease in microtubule acetylation after treatment with 5 and 20 µM 2ME2 (Fig. 2d, e). We also examined the effect of 2ME2 on HIF-1α in M2 macrophages [39]. As shown, Western blotting results showed a decrease in the expression of HIF-1α in 2ME2-treated M2 macrophages (Fig. 2f). Taken together, our results indicated that 2ME2 is taken up by THP1 cells and that even non-cytotoxic concentrations of 2ME2 affected its well-known cellular targets (α-tubulin and HIF-1α) in M2 TAMs.
Next, we investigated the effect of 2ME2 on the polarization of M2 macrophages. As shown by qPCR analysis, the expression of the M2 surface markers CD206, CD209, and CD163 decreased after treatment with 2ME2 (Fig. 3a-c). In line with the gene expression data, flow cytometry analysis also demonstrated a decrease in the expression of the M2 macrophage surface marker CD163 in 2ME2-treated macrophages (Fig. 3d). The data showed that 2ME2 notably reduced the expression of typical macrophage surface markers associated with the M2 phenotype.
We then hypothesized that 2ME2-mediated suppression of M2 polarization may reduce the abundance of pro-tumoral factors in the conditioned medium (CM) from M2 TAMs. Our results also showed a decrease in the mRNA expression of anti-inflammatory factors, including CCL18, TGF-β, MMP9, interleukin-10 (IL-10), fibronectin (FNT1), arginase and CXCL12 (Fig. 3e). Furthermore, Western blotting revealed reduced expression of MMP9 in 2ME2-treated M2 macrophages (Fig. 3f). In accordance, gelatin zymography of the CM also showed a concentration-dependent decrease in MMP9 activity in 2ME2-treated M2 macrophages (Fig. 3g). Moreover, 2ME2 treatment (5 and 20 µM) also decreased the expression of VEGF-A, a key angiogenic molecule secreted by TAMs (Fig. 3h, i).
2-Methoxyestradiol impairs the protumor functions of alternatively activated M2 macrophages in breast cancer cells
We next hypothesized that, in addition to M2 macrophage polarization, 2ME2 could impair the pro-tumoral functions of M2 macrophages. To examine this, breast cancer cells were incubated with drug-free conditioned medium (CM) from 2ME2-treated M2 macrophages. THP-1 cells were polarized to the M2 phenotype in the presence of either vehicle or 2ME2 (5 and 20 µM) for 24 h, 2ME2-containing medium was removed, and the cells were cultured in FBS-free RPMI medium for 6 h. As shown in Fig. 4a, b, M2 macrophage CM resulted in increased proliferation of MCF-7 cells, which was markedly reduced upon incubation with 2ME2-treated CM. We then determined the effect of CM from M2 macrophages on the migration of breast cancer cells. The percent wound closure at 24 h in scratch-wounded MDA-MB-231 cells decreased by 31.29 ± 1.18% (p < 0.001) and 40.03 ± 3.74% (p < 0.005) when cells were grown in CM from M2 macrophages treated with 5 µM and 20 µM 2ME2, respectively, compared to CM from vehicle-treated M2 macrophages (Fig. 4c, d), indicating a decrease in cell migration. The Transwell assay also showed that 2ME2 reduced the migration-inducing ability of M2 macrophages (Fig. 4e, f). Furthermore, gelatin zymography results showed that conditioned medium from control M2 macrophages stimulated the secretion of MMP9 from MCF-7 cells. A drastic reduction in secreted MMP9 was observed in MCF-7 cells induced by TAMs polarized with 2ME2 in comparison with control M2 macrophages (Fig. 4g). Overall, these observations showed that treatment with 2ME2 impaired the pro-tumoral functions of M2 macrophages.
2ME2 decreased M2 macrophages and inhibited breast tumor metastasis in vivo
To confirm the effects of 2ME2 on tumor-associated macrophages and breast cancer metastasis in vivo, we developed a 4T1 syngeneic breast tumor model in immunocompetent BALB/c mice and administered vehicle or 75 or 125 mg/kg 2ME2 to these mice. The tumors were excised after 25 days. Immunohistochemistry analysis showed reduced expression of the M2 macrophage marker CD163 in tumors isolated from 2ME2-treated mice (Fig. 5a). Concomitantly, in comparison with vehicle-treated mice, 2ME2-treated mouse tumors showed a decrease in the mRNA expression of CD163 and cytokines and growth factors, including VEGF-A, CCL22, FNT1 and IL-10(Fig. 5b, c).
Next, we analyzed the effect of 2ME2 on tumor metastasis. Our results showed that 2ME2-treated mice showed reduced tumor volume and tumor weight in comparison with vehicle-treated mice (Fig. 5d-f). However, irrespective of the treatment provided, no significant reduction in mouse body weight was observed (Fig. 5g). Interestingly, our results showed fewer metastatic lung nodules after 2ME2 treatment (Fig. 5h, i) which was confirmed by H&E staining of the lungs of 2ME2 and vehicle-treated mice (Fig. 5j). Together, these results suggest that, consistent with our in vitro results, 2ME2 inhibits M2 macrophage polarization and suppresses cancer metastasis to the lungs in vivo.
2-Methoxyestradiol inhibits STAT3 in alternatively activated M2 macrophages
To further investigate the mechanism by which 2ME2 dampens the polarization and pro-tumoral functions of M2 macrophages, we determined its effect on STAT3 signaling, which is implicated in the polarization and functions of M2 macrophages. Compared to that in M0 macrophages, the level of p-STAT3 (pY705) was found to be nearly two-fold higher in M2 TAMs (Fig. 6a, b). Immunofluorescence imaging further showed that M0 macrophages had a uniform punctate distribution of STAT3 in the cytoplasm, while in M2 macrophages, activated STAT3 accumulated in the nucleus (Fig. 6c, d). After analyzing the kinetics of activated STAT3 nuclear translocation, our results revealed that almost all M2 cells exhibited accumulation of pSTAT3 in the nucleus within 2 h of IL-4/13 polarization (data not shown). Western blotting showed that 2ME2 decreased pSTAT3 in M2 TAMs (Fig. 6e, f). Furthermore, immunofluorescence images showed diffuse localization of STAT3 primarily in the cytoplasm of M2 macrophages polarized with 2ME2. The inhibitory effect of 2ME2 on the phosphorylation and nuclear transfer of pSTAT3 was diminished in the presence of colivelin, a STAT3 activator (0.5 µM) (Fig. 6g, h) [40]. Western blotting also showed the inhibition of the nuclear translocation of STAT3 after 2ME2 treatment, which was partly reversed by colivelin (Fig. 6i-k). The data show that 2ME2 inhibits the phosphorylation and nuclear translocation of STAT3 in M2 TAMs, suggesting that the functional impairment of TAMs could be mediated by the inhibitory effect of 2ME2 on STAT3 signaling.
To confirm that STAT3 is essential for M2 functions and that its pharmacological inhibition can impair their functions and target M2 TAMs, we polarized TAMs in the presence of WP1066, a specific inhibitor of STAT3 phosphorylation in clinical trials, and examined the polarization and pro-tumoral functions of M2 TAMs. Using the MTT assay, we first determined the concentrations of WP1066 with less cytotoxicity (data not shown). We then examined whether inhibition of STAT3 activation could indeed inhibit the polarization and protumor functions of M2 macrophages. Gene expression analysis by qPCR revealed significantly decreased expression of CD206 and CD163 after treatment with sub-cytotoxic concentrations of WP1066 (Supplementary data, Fig. S2a, b). Flow cytometry analysis also showed a decrease in the CD163 + cell population after treatment with WP1066, suggesting that STAT3 inhibition affected M2 polarization (Supplementary data, Fig. S2c). Similarly, we found reduced expression of M2 macrophage-secreted cytokines, including FNT-1, MMP9, VEGF-A, IL-10 and TGF-β, in WP1066-treated macrophages (Supplementary data, Fig. S2d). Gelatin zymography also showed decreased gelatinolytic activity of MMP9 in conditioned medium from WP1066-treated M2 macrophages (Supplementary data, Fig. S2e). We found that WP1066-mediated inhibition of STAT3 also reduced the migration-augmenting effects of M2 macrophages in breast cancer cells (Supplementary data, Fig. S2f-h). Overall, these results show that STAT3 activity is required for M2 polarization and pro-tumoral functions and that 2ME2 dampens the functions of STAT3 in M2 macrophages and phenocopies the effects of STAT3 inhibition.