Concomitant attenuation of HMGCR expression and activity enhances the growth inhibitory effect of atorvastatin on TGF-β-treated epithelial cancer cells

Epithelial-mesenchymal transition (EMT) in primary tumor cells is a key prerequisite for metastasis initiation. Statins, cholesterol-lowering drugs, can delay metastasis formation in vivo and attenuate the growth and proliferation of tumor cells in vitro. The latter effect is stronger in tumor cells with a mesenchymal-like phenotype than in those with an epithelial one. However, the effect of statins on epithelial cancer cells treated with EMT-inducing growth factors such as transforming growth factor-β (TGF-β) remains unclear. Here, we examined the effect of atorvastatin on two epithelial cancer cell lines following TGF-β treatment. Atorvastatin-induced growth inhibition was stronger in TGF-β-treated cells than in cells not thusly treated. Moreover, treatment of cells with atorvastatin prior to TGF-β treatment enhanced this effect, which was further potentiated by the simultaneous reduction in the expression of the statin target enzyme, 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR). Dual pharmacological targeting of HMGCR can thus strongly inhibit the growth and proliferation of epithelial cancer cells treated with TGF-β and may also improve statin therapy-mediated attenuation of metastasis formation in vivo.

www.nature.com/scientificreports/ cells, mesenchymal cells, and adipocytes 14 . Additionally, elevated blood levels of TGF-β in cancer patients are associated with poor prognosis 15,16 . Metastases lead to approximately 90% of cancer-related deaths 17 , and EMT is considered an important first step in metastasis initiation. We previously reported that mesenchymal-like tumor cells displayed increased sensitivity to atorvastatin-induced growth delay 4,18 . Atorvastatin treatment also reduced metastasis formation in the lung and liver in two independent mouse models of spontaneous breast cancer metastasis without affecting growth in their respective primary tumors 19 . These data collectively suggest that cancer cells that initiated, undergo or have completed EMT are more susceptible to statins' inhibitory effects than those that do or have not, both in vitro and in vivo.
In this study, we examined whether TGF-β treatment of epithelial-like cancer cells and the concomitant downregulation of HMGCR expression counteracted the cells' resistance to atorvastatin. We also tested if atorvastatin's anticancer effects depended on when the statin treatment commenced. We found that TGF-β-treated epithelial cancer cell lines were more sensitive to atorvastatin. This sensitivity was more pronounced when statin addition preceded TGF-β treatment. We also demonstrated that such sensitivity was even further enhanced by the concomitant downregulation of HMGCR expression in cancer cells.

TGF-β1 induces markers of EMT initiation in epithelial cancer cell lines. To examine whether
inducing a mesenchymal cell phenotype in epithelial cancer cell lines increased their susceptibility to atorvastatin's growth-inhibitory effect, we first tested their response to TGF-β1. The experimental design is shown in Fig. 1A. Briefly, lung cancer-derived NCI-H322M cells were serum-starved for 24 h, then incubated with 1, 5, or 10 ng/mL of TGF-β1 for 72 h, as described previously [20][21][22] . Data are normalized to 18S rRNA level in each sample and are expressed as values relative to that of the internal control. RNA isolated from cells cultured without TGF-β1 is used as control. The measurement values for each group are compared using the Bonferroni-Dunn post-hoc test. Mean ± SD, n = 3, *p < 0.05, **p < 0.01; Comparison with TGF-β1 (−) control (0 ng/mL). (E) Protein levels of E-cadherin, N-cadherin, and vimentin are determined by western blot analyses. GAPDH expression is used as a loading control. Representative images from three independent experiments are shown. (F) Immunofluorescence staining images of EMT markers. E-cadherin (green), N-cadherin (red), and cell nuclei (blue, Hoechst staining) are shown. N-cadherin expression is only observed in TGF-β1 induced cells.

TGF-β1 treatment promotes atorvastatin sensitivity in epithelial cancer cell lines.
To further test the sensitivity of NCI-H322M cells to statin treatment we first treated cells with 10 ng/mL TGF-β1. Next, different atorvastatin concentrations were added to the cell culture medium ( Fig. 2A, TGF-β (+) group). The proliferation of cells untreated with TGF-β1 (TGF-β (−) group) was only slightly attenuated by 10 µM atorvastatin, but was significantly attenuated at a concentration of 30 µM (p < 0.01) (Fig. 2B). In contrast, in the TGF-β (+) cells, the growth of cells significantly decreased at both 10 µM and 30 µM atorvastatin concentrations (p < 0.05, and p < 0.01, respectively) (Fig. 2B, right). Similar effects were observed in TGF-β1-treated ovarian cancer-derived OVCAR3 cells (Supplementary Fig. S2). In both cell lines, induction with TGF-β1 alone led to a small decline in cell numbers ( Fig. 2B and Supplementary Fig. S2B). Thus, TGF-β1 treatment improved the growth inhibitory effects of atorvastatin, to some degree, in epithelial-like cancer cells.
Atorvastatin pretreatment of epithelial cancer cells increases their sensitivity to statin post TGF-β treatment. Next, we tested whether atorvastatin pretreatment improved the growth-inhibitory effect on TGF-β1-treated epithelial cancer cells. After serum starvation, NCI-H322M (Fig. 3A) and OVCAR3 ( Supplementary Fig. S3A) cells were incubated with and without 1 µM or 5 µM atorvastatin in serum-starved medium for 24 h prior to TGF-β1 treatment for 6 days (TGF-β (+) group). TGF-β1-untreated and atorvastatintreated cells served as the control (TGF-β (−)) group. The difference in cell number between the TGF-β (+) www.nature.com/scientificreports/ and TGF-β (−) groups at 0 µM atorvastatin concentration (Fig. 3C) was statistically not significant at the end of the experiment. However, at an increased concentration of atorvastatin, the cell numbers more significantly decreased in the TGF-β (+) group than those of the TGF-β (−) group (p < 0.01) receiving the same atorvastatin dose (Fig. 3C). Similar effects were observed in TGF-β1-treated OVCAR3 cells ( Supplementary Fig. S3B).
In previous studies, TGF-β1 treatment was shown to increase cell volume by activating the mammalian target of rapamycin (mTOR) pathway, which in turn was inhibited by rapamycin treatment 23 . Therefore, we also tested cell volume changes. The cell volume of NCI-H322M cells in the TGF-β (+) group grew to approximately 1.3 × that of cells in the TGF-β (−) group, regardless of atorvastatin treatment (p < 0.01) (Fig. 3B,D). This suggests that atorvastatin pretreatment did not affect cell growth via mTOR signaling.
At the end of the experiment (Fig. 3A), we also tested how EMT markers' expression changed in atorvastatin's presence, both before and after TGF-β1 treatment ( Fig. 4 and Supplementary Fig. S4, S5). E-cadherin expression remained essentially unchanged at both the mRNA (Fig. 4A) and protein levels ( Fig. 4F and Supplementary Fig. S4A) in both the presence and absence of TGF-β1 at a low (1 µM) atorvastatin dose. However, treatment with 5 μM atorvastatin slightly increased TGF-β1-induced downregulation of E-cadherin expression (Fig. 4A,F). Moreover, an increased atorvastatin concentration strongly inhibited TGF-β1-induced upregulation of N-cadherin expression, both at the mRNA and protein levels (Fig. 4B,F and Supplementary Fig. S4B). Treatment with 5 μM atorvastatin significantly accelerated TGF-β1-induced vimentin mRNA expression ( Fig. 4C) but no vimentin protein expression was detected ( Fig. 4F and Supplementary Fig. S4C). In the TGF-β (−) group, there was no significant difference in the expression of the cell cycle regulator, c-Myc, at both the mRNA and  4E). In contrast, after the addition of 5 μM atorvastatin, there was a decline in the induction of HMGCR gene expression in the TGF-β1-treated cells (Fig. 4E). These findings paralleled protein level changes; atorvastatin treatment strongly induced HMGCR protein expression (Fig. 4F), yet in TGF-β1-induced cells exposed to 5 μM atorvastatin this effect was weaker (Fig. 4F, Supplementary Fig. S4E).

Attenuated HMGCR expression promotes sensitivity to atorvastatin in TGF-β-treated epithelial cancer cells. Downregulation of HMGCR activity contributes to statins' anticancer effect by inhibiting
post-translational modification (prenylation) of small G-proteins (e.g., Ras, Rho, Rac, and Cdc42) 3 . We previously showed that siRNA-mediated attenuation of HMGCR expression enhanced the growth and proliferationinhibitory effect of atorvastatin on relatively statin-resistant NCI-H322M epithelial lung cancer cells 24 . Therefore, we next tested whether downregulating the expression of HMGCR further enhanced statin's inhibitory effect in atorvastatin-pretreated, TGF-β1-treated NCI-H322M cells. The experimental design is shown in Fig. 5A. Briefly, NCI-H322M cells were treated with 1 μM or 5 μM atorvastatin for 24 h. EMT was then induced with 10 ng/mL TGF-β1 for 6 days, alongside HMGCR siRNA treatment. Downregulation of HMGCR expression, observed at both the mRNA (Fig. 5B) and protein levels ( Fig. 5C and Supplementary Fig. S6), enhanced atorvastatin's inhibitory effects at both the 1 and 5 μM concentrations ( Fig. 5D; Supplementary Fig. S7; Supplementary Table S1). Thus, TGF-β1-treated NCI-H322M cells with reduced HMGCR expression were more sensitive to atorvastatin's inhibitory effects than those with uninhibited HMGCR expression.

Discussion
EMT is a stepwise cellular transdifferentiation process from epithelial to mesenchymal state 25 . EMT plays an important role in several physiological and pathological processes such as wound healing and metastasis formation 12,13,26 . EMT-induced initiation of the metastatic cascade is intimately related to the tumor microenvironment's state 26 . For example, senescent fibroblasts often acquire a proinflammatory state characterized by their www.nature.com/scientificreports/ secretion of extracellular proteases, growth factors, and cytokines 27 . Some of these factors, such as brain-derived neurotrophic factor (BDNF) and TGF-β1, can promote EMT-like phenotypes, in both normal cells [28][29][30] and in tumor cells bearing mutations in oncogenes and tumor suppressor genes 14,31 . TGF-β1 initiates EMT by activating Smad2/3. This is followed by increased expression of transcription factors such as ZEB1, which ordinarily suppresses the expression of E-cadherin 12,13,32 . Intracellular signals involved in the expression of E-cadherin are thereby altered, leading to the expression of mesenchymal markers like N-cadherin and vimentin 33 .  www.nature.com/scientificreports/ We previously reported that the epithelial lung cancer cell line NCI-H322M had high E-cadherin expression and very low or no vimentin expression 4 . Because statins exert anticancer effects most potently on mesenchymallike cancer cells 4 , we hypothesized that cancer cells that acquire at least some mesenchymal properties become more sensitive to statins. However, atorvastatin treatment after TGF-β1 treatment only slightly affected the proliferation of NCI-H322M and OVCAR3 cells. In contrast, atorvastatin pretreatment significantly attenuated cell proliferation in the TGF-β1-treated group in a dose-dependent manner. Expression levels of c-Myc are minimal in quiescent cells in vitro; however, once the cells are exposed to mitogenic stimuli, c-Myc mRNA and protein levels rapidly increase, and the cells enter the G1 phase of the cell cycle 34,35 . A previous study has reported that statins upregulate miR-33b expression and adversely affect c-Myc expression and function in cancer cells 36 38 similarly reported that atorvastatin treatment effectively negated the TGF-β1-stimulated (1 ng/mL) downregulation of E-cadherin and upregulation of vimentin in A549 lung cancer cells. In our study, atorvastatin pretreatment significantly suppressed the TGF-β1 treatment-induced increase in N-cadherin expression at both the mRNA and protein levels. This effect was inversely proportional to atorvastatin concentration. Loss or reduction in the expression of E-cadherin and upregulation of N-cadherin (the so-called "cadherin switch") are common features of full-or partial EMT. Functional E-cadherin expression inhibits cell migration, whereas aberrant N-cadherin expression enhances cell migration and invasion in cancer cells regardless of E-cadherin expression 39 . Furthermore, activated TGF-β signaling pathway can induce the expression of cell-surface glycoprotein CD146, a member of the immunoglobulin superfamily. A CD146/ERK cascade can enhance N-cadherin expression during TGFβ-induced EMT 40 . However, statins and farnesyl transferase inhibitors are known to significantly reduce ERK phosphorylation in non-small cell lung cancer cells 41 . It is thus likely that the reduction in N-cadherin expression this study observed occurred due to the inhibition of the CD146/ERK cascade, although the detailed mechanism remains unclear. Jiang et al. (2018) 42 demonstrated that statins both slightly upregulated HMGCR mRNA levels and increased HMGCR protein levels more than tenfold, mainly by preventing ubiquitination and degradation of HMGCR. Indeed, statin-induced upregulation of HMGCR expression appeared to be a common homeostatic reaction in cells 43 . In this study, we found that inducing a partial mesenchymal state did not affect HMGCR (mRNA) expression. However, HMGCR protein levels strongly increased in response to statin treatment, although they were slightly lower in cells after TGF-β1 treatment than without it (Fig. 4F, Supplementary Fig. S4E). This effect was countered by siRNA knockdown of the HMGCR gene and protein expression (Fig. 5B,C). In turn, this led to a much stronger attenuation of cell growth and proliferation than atorvastatin treatment alone (Fig. 5D).
Altogether, our study demonstrates that inducing phenotypic state switch in epithelial cancer cells and simultaneously downregulating HMGCR expression improves atorvastatin-induced attenuation of cell proliferation. We also found a consistent correlation between atorvastatin treatment-mediated attenuation of TGF-β1-inducedand mesenchymal-like cell proliferation in vitro 4,18 and metastasis formation in vivo 19 . The data of Ishikawa et al. (2018) 24 and Göbel et al. (2019) 44 , combined with our current results, moreover demonstrates that attenuating HMGCR expression further enhances atorvastatin's effect in vitro. Indeed, dual treatment with statin and the HMGCR degrader Cmpd81 synergistically decreased LDL-cholesterol levels (and reduced atherosclerotic plaque formation) in a murine model 42 . Therefore, we can hypothesize that the dual inhibition of HMGCR by, on the one hand inhibiting HMGCR activity by statins and, on the other, attenuating HMGCR expression by siRNA or Cmpd81 would also delay metastasis formation to a greater extent than atorvastatin treatment alone 19 (Fig. 6).
For TGF-β induction, 5 × 10 4 cells/mL were seeded in 12-well plates. After attachment, the cells were incubated with serum-starved medium (0.5% FBS) for 24 h to remove residual TGF-β1 from the media. Cells were then incubated with 1, 5, or 10 ng/mL of recombinant human TGF-β1 in serum-starved medium (Peprotech, Rocky Hill, NJ, USA) for 72 h, as described previously [20][21][22] . Changes in gene and protein expression levels of epithelial (E-cadherin) and mesenchymal (N-cadherin and vimentin) cell markers were analyzed. For some experiments (Fig. 3A), the culture medium was changed to fresh serum-starved medium supplemented with 10 ng/mL TGF-β1 and 0-5 µM atorvastatin. To maintain optimal culture conditions, serum-starved medium containing TGF-β1 and atorvastatin was replaced every 3 days. The cell number was counted after 6 days of incubation. Total RNA and protein were extracted from triplicate samples accordingly.
In siRNA experiments (see below), after 24 h of serum starvation, NCI-H322M cells were treated with atorvastatin at concentrations of 1 μM or 5 μM for 24 h. Cells were then treated with 10 ng/mL TGF-β1 for 6 days in the www.nature.com/scientificreports/ presence of 1 μM or 5 μM atorvastatin and HMGCR siRNA (Fig. 5A). Scrambled siRNA was used as a negative control for RNAi experiments. Cell numbers were counted on days 3 and 6 after TGF-β1 addition.

Downregulation of HMGCR expression by siRNA. Predesigned siRNA oligonucleotides specific to
HMGCR (NM_000859, siRNA ID#s142, targeted exon 12, siRNA location: 1698) were obtained from Thermo Fisher Scientific. Silencer negative control siRNA (#4390843, Thermo Fisher Scientific) was used as a scrambled siRNA. This sequence, provided by the manufacturer, showed no significant homology to any gene. Reverse transfections were performed in 12-well dishes (5 × 10 4 cells/mL) according to the manufacturer's instructions using Lipofectamine RNAiMax (Thermo Fisher Scientific), Opti-MEM (Thermo Fisher Scientific), and siRNAs (final concentration, 10 nM) for the respective targets. Atorvastatin-treated NCI-H322M cells were harvested on days 3 and 6 after transfection to analyze cell viability. Transfection efficiency was assessed by quantitative reverse transcription polymerase chain reaction (RT-PCR) and western blotting. The PCR primer set, which included siRNA target sites, was as follows: sense primer 5′-CCC AGC CTA CAA GTT GGA AA-3′ and anti-sense primer 5′-AAC AAG CTC CCA TCA CCA AG-3′ (PCR product: 152 bp).
Real-time PCR. An RNeasy mini kit (Qiagen, Hilden, Germany) was used to extract total RNA from the cells. cDNA was synthesized from 1 μg of total RNA using ReverTra Ace qPCR RT with gDNA Remover kit (Toyobo, Osaka, Japan). The primer sets used for PCR are shown in Supplementary Table S2. PCR was performed using LightCycler FastStart DNA MasterPLUS SYBR Green I mix and a LightCycler rapid thermal cycler system (Roche Diagnostics, Lewes, UK).
Western blotting. Cells in a petri dish were washed twice with cold PBS, followed by incubation with CelLytic M solution (Sigma-Aldrich) on ice for 5 min. Subsequently, the cells were scraped and collected into 1.5-mL microfuge tubes and homogenized by passing through a 27-gauge syringe needle. The cell lysates were centrifuged at 16,000 × g at 4 °C for 15 min, and the supernatants were transferred to new tubes. Protein concentrations were measured by the bicinchoninic acid (BCA) method using the BCA Protein Assay Kit (Thermo Fisher Scientific). Proteins were incubated at 90 °C for 3 min with Laemmli sample buffer (Bio-Rad, Hercules, CA, USA) prior to electrophoresis. NuPAGE® 4%-12% Bis-Tris gel (Thermo Fisher Scientific) was used for electrophoresis, and 10 μg of protein lysates were loaded per lane. After electrophoresis, the proteins were transferred onto a nitrocellulose membrane using iBlot® Gel Transfer Stacks Nitrocellulose (Thermo Fisher Scientific) and an iBlot® Gel Transfer Device (Thermo Fisher Scientific). After blocking the nitrocellulose membrane with 5% (w/v) skim milk (Morinaga Milk Industry, Tokyo, Japan), the membranes were incubated with primary antibodies according to the manufacturer's instructions. Anti-E-cadherin rabbit monoclonal antibody (  Cell number and cell volume measurement. NCI-H322M and OVCAR3 cells cultured in 12-well plates were washed with PBS and detached with 0.05% trypsin/EDTA (Fujifilm Wako Pure Chemical). Trypsin was inactivated by adding complete medium. The cells were counted using a Scepter handheld automated cell counter (Millipore, Billerica, MA, USA). The ratio of the number of cells in each experimental group to the average number in the control group was considered as the actual cell number and survival rate of the control group was defined as 100%.
Statistical analyses. Statistical analyses were performed using Excel Statistics 2016 for Windows (version 3.21; SSRI, Tokyo, Japan). The data were compared using Student' s two-tailed t-test and one-way or two-way analysis of variance, followed by Bonferroni-Dunn post-hoc tests, with a significance level of p < 0.05. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.