Naringenin prevents TGF-β1 secretion from breast cancer and suppresses pulmonary metastasis by inhibiting PKC activation

The murine breast cancer cell line 4T1 was purchased from American Type Culture Collection (Manassas, VA, USA). 4T1 cells, the vector control (4T1/RFP), and TGF-β1-overexpressing 4T1-Luc2 cells (4T1/TGF-β1) were cultured in RPMI 1640 medium. 1D11 antibody was from eBioscience Tech (San Diego, CA, USA). Naringenin was purchased from Shanxi Huike Botanical Development Co. (Xi'an, China).

Human growth hormone signal sequence was synthesized and fused with the full-length mouse Tgf-β1 gene using PCR. The hybrid gene of human growth hormone signal sequence and mouse Tgf-β1 was then ligated into pSin-EF2-Oct4-Pur plasmid (Addgene, Cambridge, MA, USA) to replace Oct4 with SpeI and EcoRI cloning sites at the 5′ and 3′ termini, respectively. RFP from the pmRFP-C1 plasmid (Clontech, Mountain View, CA, USA) was then cloned into the pSin-EF2-TGFβ1-Pur plasmid to replace puromycin. The generated Tgf-β1 overexpression vectors were then enveloped in 293T cells. The medium containing the packaged virus was used to infect 4T1-Luc2 breast cancer cells (PerkinElmer, Waltham, MA, USA) to generate 4T1/TGF-β1 transformants. The control transformants, 4T1/RFP cells, were generated using the vector without Tgf-β1 gene, following the same procedures. 4T1/RFP and 4T1/TGF-β1 transformants were then sorted by flow cytometry with excitation/emission of 578/603 nm.

Four-week-old female Balb/c mice were purchased from Weitonglihua Tech. (Beijing, China) and housed in the Animal Care Facility of the Institute of Biophysics, Chinese Academy of Sciences, China. All animal protocols used for this study were approved by the Institutional Animal Care and Use Committee. The fourth mammary fat pads of Balb/c mice were injected with 2 × 10⁴ 4T1/RFP or 4T1/TGF-β1 cells. Beginning on the same day, the mice were administered 200 mg/kg naringenin once daily for 30 days (suspension in 1 % sodium carboxyl methyl cellulose (CMCNa)) or 5 mg/kg 1D11 antibody (dilution in phosphate-buffered saline buffer) twice a week for 3 weeks. The primary tumor and lung metastases were imaged by bioluminescence using the IVIS Spectrum In Vivo Imaging System (Xenogen, Caliper Life Science, PerkinElmer, Hopkinton, MA, USA ) as described previously [28]. Briefly, tumor-bearing mice were given intraperitoneal injections with 150 mg/kg luciferin and the lung areas were imaged. To avoid the bioluminescence from the primary tumor, primary tumors were wrapped with light-proof bags. After 4 weeks of primary tumor growth, mice were sacrificed after intraperitoneal injection of luciferin for 15 minutes and the lungs were collected for imaging or weighing to determine the amount of metastases. The weight of tumor burden in the lung was calculated by subtracting the mean weight of normal lungs (0.15 g) from the weight of the lungs with metastatic tumors; primary tumors were also isolated and weighted. After treatment, the lungs and spleens of each mouse at day 14 and day 28 were dissected for quantitative PCR (qPCR), T-cell immunity assays, and immunohistochemistry. The levels of activated TGF-β1 in the primary tumors, spleens, lungs, and serum were detected by enzyme-linked immunosorbent assay (ELISA). The bioluminescence imaging, qPCR [29], T-cell immunity assays, immunohistochemistry, and TGF-β1 ELISA were performed as described previously [26] (details in Additional file 1). The survival endpoint of mice was defined as the day of the mouse’s death.

In vivo assays of T-cell activation were performed as described previously [26]. Briefly, isolated lymphocytes from the spleen and lung were surface-stained with fluorochrome-conjugated anti-CD4, anti-CD44, and anti-CD62L antibodies. To assay myeloid-derived suppressor cells (MDSC), isolated splenocytes or purified T cells from the lungs were stained with fluorochrome-conjugated anti-Gr-1 and anti-CD11b antibodies. To analyze the production of interferon gamma (IFNγ) in CD8⁺ T cells, isolated splenocytes were stimulated with 25 ng/ml phorbol 12-myristate 13-acetate (Sigma, St. Louis, MO, USA) and 500 ng/ml ionomycin (Sigma) for 5 hours. After further incubation with Brefeldin A Solution (eBioscience) for 3 hours, the stimulated cells were surface-stained with anti-CD8 antibodies, fixed and permeabilized, and then stained with anti-IFNγ antibody.

For Treg assays, the isolated splenocytes or purified T cells from lungs were separated using CD4 MACS beads (Miltenyi Biotec, Cologne, Germany). The separated CD4⁺ T cells were cultured in the supernatants with anti-CD3 antibody (3 μg/ml) and anti-CD28 antibody (1 μg/ml) for 72 hours. The cells were surface-stained with anti-CD4 and anti-CD25 antibodies, fixed and permeabilized, and then stained with anti-Foxp3 antibody. The stained cells were analyzed by flow cytometry using a FACSCalibur system with CellQuest software (BD Biosciences, San Jose, CA, USA).

T cells were isolated from the spleen of normal (control group) or 4T1/TGF-β1 tumor-bearing Balb/c mice with naringenin or 1 % CMCNa treatment. After red blood cells were lysed by ACK hypotonic lysis solution (Sigma), T cells were purified using a pan T-cell isolation kit (Miltenyi Biotech) in accordance with the manufacturer’s protocol. Balb/c nude mice were then intravenously injected with 5 × 10⁵ of these purified T cells mixed with 5 × 10³ 4T1/TGF-β1 tumor cells (100:1 ratio of T cells to 4T1/TGF-β1 tumor cells). T cells with tumor cells were transferred once a week for 3 weeks. The transfer of the mixed cells to nude mice was performed as described previously [30]. The bioluminescence of lung metastasis in nude mice on day 28 was imaged using the IVIS system as described previously [28] (details in Additional file 1).

Cellular localization of TGF-β1 in 4T1 cells with different treatments was determined by immunofluorescence staining with Alexa 633-labeled secondary IgG antibody after anti-TGF-β1 antibody binding. The TGN was stained with Alexa 488-labeled secondary IgG antibody after anti-TGN46 antibody binding. The nuclei were stained with Hoechst 33342. Stained cells were observed using confocal microscopy as described [31] previously (details in Additional file 1).

All in vivo experiments included at least five mice per group. Results are presented as median ± interquartile (IQ) range or median ± standard error (SE). Analysis of results containing two groups was carried out using the Student t test, assuming unequal variance. Two-way analysis of variance was performed to determine the variations of bioluminescent photons between the control and TGF-β1 overexpression groups. The survival curves were analyzed using log-rank (Mantel–Cox test) analysis. P <0.05 was considered statistically significant.

To investigate the role of TGF-β1 overexpression in 4T1 cells in pulmonary metastasis, we constructed a lentiviral vector that expressed a hybrid construct containing human growth hormone signal sequence and murine Tgf-β1, which enabled the transduced cells to secrete mature TGF-β1 (4T1/TGF-β1). Compared with the null-vector control cells (4T1/RFP), 4T1/TGF-β1 cells showed increased expression of Tgf-β1 mRNA (2332-fold increase; P <0.001), as well as increased cellular and secreted TGF-β1 (P = 0.05 and P = 0.01, respectively) (Fig. 1a, b). The 4T1/TGF-β1 cells also showed increased migration and invasion compared with 4T1/RFP cells in vitro (Fig. 1c). However, cellular invasion of both 4T1/RFP and 4T1/TGF-β1 was inhibited by addition of TGF-β1 neutralizing antibody (1D11) to the medium (Fig. 1c). Furthermore, we quantitated TGF-β1 secretion in several human breast cancer cell lines, and found that secreted TGF-β1 levels were positively related to invasiveness of the cancer cells (Additional file 2: Figure S1).

We next investigated whether Tgf-β1 overexpression led to extensive pulmonary metastases. 4T1/TGF-β1 cells and 4T1/RFP cells were injected into the fourth breast pads of mice, and pulmonary metastases were determined at day 28 after injection. The bioluminescence images showed that all of the mice bearing 4T1/TGF-β1 tumors had pulmonary metastases with brighter bioluminescence than the mice bearing 4T1/RFP tumors, in which only one mouse had detectable bioluminescence before the dissected lungs were exposed (for the in vivo lung imaging, P = 0.023; for the dissected lung imaging, P = 0.006) (Fig. 1d, e). The histological sections of lungs from mice bearing 4T1/TGF-β1 tumors showed larger metastatic lesions than those from mice bearing 4T1/RFP tumors (Fig. 1f). The tumor burden weight in removed lungs from mice bearing 4T1/TGF-β1 tumors was greater than that of lungs from mice bearing 4T1/RFP tumors (P <0.001) (Fig. 1h), while the tumor weight at the primary site did not show a significant difference between the mice bearing 4T1/RFP cells and the mice bearing 4T1/TGF-β1 cells (P = 0.43) (Fig. 1g and Additional file 3: Figure S2). The extensive pulmonary metastases in the 4T1/TGF-β1 tumor-bearing mice could lead to a decreased life span. Indeed, the survival analysis showed that all of the 4T1/TGF-β1 tumor-bearing mice without any treatment died at day 45 after tumor cell injection, while the 4T1/RFP tumor-bearing mice without treatment displayed an increased life span (P = 0.007) (Fig. 2c).

As shown in Fig. 2, naringenin administration had an inhibitory effect on pulmonary metastasis at day 34 (similar inhibitory effects on pulmonary metastasis at day 24 shown in Additional file 3: Figure S2A) and prolonged survival to levels comparable with 1D11 antibody administration in 4T1/TGF-β1 tumor-bearing mice, but better than 1D11 in 4T1/RFP tumor-bearing mice (for survival in 4T1/RFP: naringenin vs control, P = 0.032 and 1D11 vs control, P = 0.19; for survival in 4T1/TGF-β1: naringenin vs control, P = 0.012 and 1D11 vs control, P = 0.021) (Fig. 2a–c). In contrast to the untreated control, naringenin and 1D11 treatments decreased the incidence of pulmonary metastasis in tumor-bearing mice on day 24 and day 34 (for 4T1/TGF-β1 tumor-bearing mice on day 24: 57 % (4/7) in control group, 14 % (1/7) in naringenin group, and 29 % (2/7) in 1D11 group; for 4T1/RFP tumor-bearing mice on day 34: 100 % (7/7) in control group, 29 % (2/7) in naringenin group, and 43 % (3/7) in 1D11 group) (Additional file 4: Table S1). The growth of the primary tumor was not significantly affected by naringenin treatment or by 1D11 treatment compared with the untreated control mice (Additional file 3: Figure S2B, C). The reduction of pulmonary metastasis by naringenin and 1D11 treatments could result from lower levels of activated TGF-β1 circulating in the serum of tumor-bearing mice. Therefore, we determined the levels of the activated TGF-β1 in the serum and in the lung on day 34. Compared with normal mice, the activated TGF-β1 concentration in lung tissues and circulating serum was significantly elevated both in 4T1/RFP and 4T1/TGF-β1 tumor-bearing mice. As expected, naringenin and 1D11 treatments reduced active TGF-β1 levels both in the lung and serum (Fig. 2d).

To ascertain the role of tumor-secreted TGF-β1 in promoting pulmonary metastasis, we determined levels of the activated form of TGF-β1 in the tumor, spleen, lung, and serum on days 14 and 28 after tumor cell implantation. Activated TGF-β1 was increased in all samples, including tumor, lung, and serum from tumor-bearing mice compared with tumor-free mice (P <0.05). Naringenin treatment significantly reduced activated TGF-β1 levels in the tissues and serum in tumor-bearing mice (P <0.05) (Additional file 5: Figure S3A, B). Tumor-derived TGF-β1 has been showed to induce the expression of Foxp3, which is a firm link between TGF-β1 levels and Tregs [32, 33]. In agreement with these studies, an increase of Foxp3 expression was observed in the lungs of tumor-bearing mice, particularly in mice bearing 4T1/TGF-β1 tumors. Naringenin administration reduced Foxp3 expression in the lungs of tumor-bearing mice (Additional file 5: Figure S3D). The immunohistochemical images of the lungs on day 14 and day 28 after tumor injection showed a histologic disruption of lung structure and an increase of TGF-β1 staining. Naringenin treatment decreased TGF-β1 staining and restored normal lung architecture (Additional file 5: Figure S3C).

The elevated TGF-β1 levels were accompanied by increased Foxp3 expression in the lungs of 4T1/TGF-β1 tumor-bearing mice, which aroused our interest in the relationship between TGF-β1 production and the Treg population. Because previous studies have revealed that tumor-derived TGF-β1 mediates the conversion of CD4⁺Foxp3⁺ Tregs [34], we measured the percentage of CD4⁺CD25⁺Foxp3⁺ Tregs and CD103⁺CD4⁺Foxp3⁺ cells in lymphocytes from the spleens and lungs of tumor-bearing mice. The expression of CD103 in lymphoid CD4⁺Foxp3⁺ cells was reported to recruit Tregs into tumor sites [35]. Overproduction of TGF-β1 in 4T1/TGF-β1 tumor-bearing mice was associated with a higher percentage of CD4⁺CD25⁺Foxp3⁺ cells in the lungs and CD103⁺CD4⁺Foxp3⁺ cells in the spleens, compared with 4T1/RFP tumor-bearing mice and tumor-free mice. Both naringenin and 1D11 treatments significantly reduced the percentage of CD4⁺CD25⁺Foxp3⁺ and CD103⁺CD4⁺Foxp3⁺ cells in the lungs and spleens, respectively, of 4T1/TGF-β1 tumor-bearing mice (P <0.05) (Fig. 3a, b).

Several studies have indicated that TGF-β1 can induce an immunosuppressive microenvironment that promotes metastasis both in animal models and human cancer patients [36, 37]. The results from clinical trials indicate that inhibitors of TGF-β signaling may enhance anti-tumor immune responses in patients [38], causing us to wonder whether naringenin could also exhibit anti-tumor immune responses in 4T1/TGF-β1 tumor-bearing mice by reducing TGF-β1 production. As expected, tumor-derived TGF-β1 established a systemic immunosuppressive environment that was represented by an increase of the CD4⁺Gr11b⁺ T-cell population and a decrease of the activated subpopulation of CD4⁺CD44⁺CD62L^– T cells in lungs and spleens and CD8⁺IFNγ⁺ T cells in spleens (Fig. 4a–c). In addition, the mRNA levels of effector molecules such as IFNγ and granzyme-B in the lungs were reduced after tumor injection (P <0.001) (Fig. 4d). Naringenin treatment effectively reversed the tumor-mediated immunosuppressive environment; its effect was comparable with that of 1D11 treatment (Fig. 4a–d).

To further ascertain whether the enhanced T-cell activity after naringenin treatment facilitates its inhibitory effect on breast cancer pulmonary metastasis, we transferred the T cells from 4T1/TGF-β1 tumor-bearing mice into nude mice along with 4T1/TGF-β1 tumor cells. Compared with nude mice with T cells transferred from normal mice (Control) and from the mice without naringenin treatment (CMCNa), the nude mice that received T cells from mice with naringenin treatment showed a decreased incidence of pulmonary metastasis (100 % (6/6) vs 83 % (5/6) vs 50 % (3/6) in dissected lungs; and 100 % (6/6) vs 67 % (4/6) vs 33 % (2/6) in whole bodies) (Fig. 4e–g). In addition, one of the nude mice transferred with T cells from mice without naringenin treatment showed expanded whole body metastases besides lung metastases (Fig. 4f).

Overexpression of Tgf-β1 in 4T1 cells promoted pulmonary metastasis, and naringenin inhibited pulmonary metastasis by decreasing TGF-β1 levels. We then asked whether naringenin inhibited the secretion of TGF-β1 from 4T1/TGF-β1 cells. Compared with 4T1/RFP cells, the overexpression of Tgf-β1 in 4T1/TGF-β1 cells led to a significant increase of secreted TGF-β1 (2.3-fold difference, P = 0.02) and promoted invasion of the cells (Fig. 5a, b). Knockdown of Tgf-β1 by TGF-β1 specific small interfering RNA (siRNA) and neutralization of TGF-β1 by 1D11 antibody in 4T1/RFP cells showed an inhibitory effect on the invasion and TGF-β1 secretion of tumor cells (siRNA vs control, 0.46-fold difference, P = 0.01; 1D11 vs control, 0.31-fold difference, P = 0.003) (Fig. 5a, b and Additional file 6: Figure S4A, B). Naringenin treatment resulted in decreased TGF-β1 secretion (in 4T1/RFP cells, 0.64-fold difference, P = 0.04; in 4T1/TGF-β1 cells, 0.69-fold difference, P = 0.05) and cell invasion in both 4T1/RFP and 4T1/TGF-β1 cells. The combination of naringenin and TGF-β1 specific siRNA showed no difference in TGF-β1 secretion (P = 0.18) and tumor cell invasion (Additional file 6: Figure S4A, B) from each agent alone. These in vitro results provided complementary data to confirm that naringenin suppressed the invasion of breast cancer cells by reducing autocrine TGF-β1 signaling from the tumor cells.

TGF-β1 can convert lymphatic T cells to Tregs. We next investigated whether naringenin could suppress the conversion of Tregs in vitro. In agreement with the in vivo data, the supernatants from cells treated with naringenin or 1D11 showed a suppression of Treg conversion activity (Fig. 5c and Additional file 6: Figure S4C). TGF-β1 siRNA also inhibited the conversion of Tregs in 4T1/RFP cells (0.66-fold difference, P = 0.04) (Additional file 6: Figure S4C). Compared with that of control, the supernatant of 4T1/RFP cells treated with TGF-β1 siRNA or 4T1/TGF-β1 cells treated with naringenin, which contained a lower concentration of TGF-β1, showed a reduced capacity to induce the expression of MMP2 and MMP9 mRNA in 4T1 cells (Additional file 6: Figure S4D, E). Interestingly, although both naringenin and 1D11 reduced TGF-β1 secretion, 1D11 did not change the intracellular TGF-β1 concentration while naringenin resulted in an intracellular accumulation of TGF-β1 (Fig. 5b, d). The levels of secreted and intracellular TGF-β1 in different human breast cancer cell lines and  293T and MDCK cells treated with naringenin for 48 hours were also detected. The results showed that naringenin significantly decreased TGF-β1 secretion and increased intracellular TGF-β1 concentrations in multiple human breast cancer cells, including MCF7, MDA-MB-231, MDA-MB-436, and MDA-MB-468 cells, but not in 293T and MDCK cells (Additional file 7: Figure S5).

1D11 treatment had no effect on the intracellular TGF-β1 concentration in 4T1 cells, suggesting that naringenin may have a different mechanism of action from siRNA or 1D11 in reducing the secretion of TGF-β1. Thus, we investigated the mechanisms underlying the inhibitory effect of naringenin on TGF-β1 secretion, which involve a series of controlled processes, including transcription, post transcription, and post translation. We surprisingly found that the transcription of Tgf-β1 was not influenced by naringenin treatment (Additional file 8: Figure S6B). The post-transcriptional regulation of Tgf-β1 was also investigated using actinomycin D, which blocks mRNA synthesis, in the presence or absence of naringenin. Naringenin did not affect the decay of Tgf-β1 mRNA in the presence of actinomycin D (P = 0.65) (shown in Additional file 6: Figure S4C).

Next, we tested whether naringenin regulates the translation of TGF-β1 using cycloheximide (CHX) to block protein synthesis in 4T1 cells. CHX treatment induced a decrease of secreted TGF-β1 and intracellular TGF-β1 (P <0.05) (Additional file 9: Figure S7). Naringenin treatment showed that the time-dependent increase of intracellular TGF-β1 was in parallel to a gradual decrease of secreted TGF-β1 into the media (naringenin vs control, from 24 to 48 hours detected, P <0.05) (Fig. 6a and Additional file 8: Figure S6). The effects of naringenin on released TGF-β1 and intracellular TGF-β1 were still observed in the presence of CHX (P = 0.04 for intracellular TGF-β1; P = 0.05 for released TGF-β1 after 48 hours of treatment) (Fig. 6b). Flow cytometry analysis of intracellular TGF-β1 further confirmed that naringenin induced an accumulation of intracellular TGF-β1 in the presence or absence of CHX (in the absence of CHX, 1.4-fold difference, P = 0.02; in the presence of CHX, 1.7-fold difference, P = 0.04) (Fig. 6c).

TGF-β proteins, synthesized as precursor proteins, are cleaved in the ER and then transported through the TGN to the cell surface for release or to the lysosome for degradation. We therefore hypothesized that naringenin might reduce lysosomal degradation of TGF-β1, resulting in the observed cellular accumulation. To address this hypothesis, we utilized a specific lysosome inhibitor, bafilomycin A1, which prevents protein degradation by raising lysosomal pH. Interestingly, bafilomycin A1 treatment significantly increased the intracellular accumulation of TGF-β1 while also increasing extracellular TGF-β1 release (intracellular, twofold difference, P = 0.01; extracellular, 1.6-fold difference, P = 0.02), indicating that more TGF-β1 can be released by cells after blocking lysosomal degradation (Fig. 6d, e). In contrast to bafilomycin A1 treatment, naringenin treatment decreased the extracellular release of TGF-β1 even in the presence of bafilomycin A1 (0.57-fold difference, P = 0.02) (Fig. 6d, e), suggesting that the lysosomal degradation process does not involve in the effects of naringenin.

Previous studies have demonstrated that the protein kinase C (PKC) family of kinases is involved in the transport and sorting of many proteins from the TGN to cell membranes and organelles [39], prompting us to address whether PKC is involved in the trafficking of TGF-β1. To test this hypothesis, we first investigated the expression of 11 members of the PKC family in 4T1/RFP cells and 4T1/TGF-β1 cells (Additional file 10: Table S2). The results showed that Tgf-β1 expression positively correlated with PRKCE (3.1-fold difference, P = 0.008) and PRKCZ mRNAs (1.5-fold difference, P = 0.04), suggesting that PKC-ε and PKC-ζ proteins might be involved in the transport of TGF-β1. Silencing PKC-ε and PKC-ζ in 4T1 cells using the respective siRNAs showed a decrease of Tgf-β1 mRNA expression (PKC-ε silencing, 0.57-fold difference, P = 0.04; PKC-ζ silencing, 0.63-fold difference, P = 0.05) (Fig. 7a). These results suggest that TGF-β1 and PKC-ε or PKC-ζ may work in a positive feedback loop. To further investigate the relationship between TGF-β1 and PKC-ε or PKC-ζ, we determined the production of TGF-β1 protein in 4T1 cells after silencing PKC-ε or PKC-ζ; meanwhile, the phosphorylation status of PKC-ε and PKC-ζ was also detected. Compared with untreated cells, PKC-ε and PKC-ζ siRNA-treated cells showed reduced expressions of the targeted genes, both the total protein levels and the phosphorylation levels (Fig. 7b); correspondingly, the amount of TGF-β1 released from the cells was significantly reduced (PKC-ε silencing, 0.62-fold difference, P = 0.008; PKC-ζ silencing, 0.76-fold difference, P = 0.038), but resulted in an accumulation of cellular TGF-β1 (PKC-ε silencing, 1.57-fold difference, P = 0.015; PKC-ζ silencing, 1.54-fold difference, P = 0.04) (Fig. 7c). Silencing PKC-ε or PKC-ζ with siRNA had similar results to naringenin treatment, which decreased secretion of TGF-β1 (0.64-fold difference, P = 0.006) while increasing intracellular levels of TGF-β1 (1.54-fold difference, P = 0.012) (Fig. 7c). However, naringenin treatment had no effect on the mRNA and protein expression of PKC-ε or PKC-ζ, but did inhibit their phosphorylation (Fig. 7a, b). The combination of naringenin and siRNAs targeting PKC-ε or PKC-ζ exhibited a synergistic effect on decreasing TGF-β1 secretion (Fig. 7c). Taken together, our results suggest that both PKC-ε and PKC-ζ regulate TGF-β1 trafficking, and that naringenin inhibits the activation of PKC-ε and PKC-ζ, which prevents TGF-β1 secretion.

Golgi-bound PKC has been shown to have a crucial impact on vesicle formation at the TGN [39]. Therefore, we sought to address whether the naringenin-mediated cellular accumulation of TGF-β1 is due to blocking the transport of TGF-β1 from the TGN to the plasma membrane. We used phorbol 12-myristate 13-acetate (PMA) to activate PKC in 4T1 cells, and observed the distribution of the TGN and TGF-β1 by immunofluorescence. Compared with untreated cells, PMA-treated cells had an increased density of TGN in the vicinity of the plasma membrane with a more dispersed cytoplasmic distribution and lower fluorescence intensity of TGF-β1, indicating that more TGF-β1 was released from the cells (Fig. 7d). Indeed, we found that PMA treatment resulted in a 0.57-fold decrease of TGF-β1 concentration in the cells and a 1.77-fold increase of TGF-β1 concentration in the culture media (Fig. 7e, f). The addition of calphostin C, a specific inhibitor of PKC, resulted in an increased density of the TGN and TGF-β1-containing vesicles around the nucleus with brighter fluorescence (Fig. 7d), along with more cellular accumulation (2.5-fold difference, P = 0.01) and less release of TGF-β1 (0.42-fold difference, P = 0.02) (Fig. 7e, f). Cells treated with naringenin showed similar responses to the addition of calphostin C (Fig. 7d–f), suggesting that naringenin is a potential inhibitor of PKC. Furthermore, addition of naringenin to PMA-treated cells partially reversed the effects of PMA, but naringenin had no effect on calphostin C-treated cells (Fig. 7d–f). Taken together, naringenin reduced TGF-β1 secretion by blocking its cellular transport through inhibition of PKC rather than promoting its degradation and/or inhibiting its expression.

All animal protocols used for this study were approved by the Institutional Animal Care and Use Committee.