Estrogen activates pyruvate kinase M2 and increases the growth of TSC2-deficient cells

Lymphangioleiomyomatosis (LAM) is a devastating lung disease caused by inactivating gene mutations in either TSC1 or TSC2 that result in hyperactivation of the mechanistic target of rapamycin complex 1 (mTORC1). As LAM occurs predominantly in women during their reproductive age and is exacerbated by pregnancy, the female hormonal environment, and in particular estrogen, is implicated in LAM pathogenesis and progression. However, detailed underlying molecular mechanisms are not well understood. In this study, utilizing human pulmonary LAM specimens and cell culture models of TSC2-deficient LAM patient-derived and rat uterine leiomyoma-derived cells, we tested the hypothesis that estrogen promotes the growth of mTORC1-hyperactive cells through pyruvate kinase M2 (PKM2). Estrogen increased the phosphorylation of PKM2 at Ser37 and induced the nuclear translocation of phospho-PKM2. The estrogen receptor antagonist Faslodex reversed these effects. Restoration of TSC2 inhibited the phosphorylation of PKM2 in an mTORC1 inhibitor-insensitive manner. Finally, accumulation of phosphorylated PKM2 was evident in pulmonary nodule from LAM patients. Together, our data suggest that female predominance of LAM might be at least in part attributed to estrogen stimulation of PKM2-mediated cellular metabolic alterations. Targeting metabolic regulators of PKM2 might have therapeutic benefits for women with LAM and other female-specific neoplasms.


Introduction
Lymphangioleiomyomatosis (LAM) is a disease that develops almost exclusively in females of reproductive age and predominantly involves the lungs. Although the genetic basis is known, specifically mutations in either tuberous sclerosis 1 (TSC1) or the tuberous sclerosis 2 (TSC2) genes, the pathophysiology is poorly understood. It is hypothesized that smooth-muscle-like

Estrogen promotes the growth of TSC2-deficient cells in a glucosedependent manner in vitro
To elucidate pro-survival mechanisms regulated by estrogen in LAM cells [17,19,24], we used the well-established, estrogen-responsive Tsc2-deficient rat ELT3 cell line, which was engineered to express an empty vector (TSC2-) or TSC2 add-back (TSC2+) [21]. Consistent with our findings [17] and those of others [15,19], estrogen modestly promoted the growth of ELT3 (TSC2-) cells by 12%, 25%, and to 35% in a time-dependent manner from day 3 to day 5, respectively (p < 0.01; Fig 1A). However, estrogen treatment had no measurable effect on the growth of ELT3 cells expressing TSC2 (TSC2+) (Fig 1B). We also tested estrogen responsiveness in LAM patient-derived TSC2-deficient 621-101 cells that exhibit constitutively active mTORC1, and in TSC2-expressing 621-103 cells. Consistently, crystal violet assay showed that estrogen significantly increased cell number by approximately 33% (p < 0.01) but only in glucose-rich medium (Fig 1C). In contrast, and consistent with the results shown in Fig 1B, estrogen treatment did not affect the growth of TSC2-addback 621-103 (TSC2+) cells regardless of glucose concentrations (Fig 1D). These data indicate that estrogen selectively promotes the growth of TSC2-cells in a glucose-dependent manner.

Estrogen regulates PKM2 phosphorylation in TSC2-deficient cells in vitro
Although PKM2-stimulated glycolysis contributes to the development of tumors caused by hyperactive mTORC1 [10], the impact of PKM2 on HIF-1α transcription, and ultimately cell proliferation, requires phosphorylation of PKM2 at Ser37 by extracellular signal-regulated kinase (ERK1/2). Phosphorylated-PKM2 translocates from the cytoplasm to the nucleus and regulates the expression of genes involved in glycolysis [12,13]. We speculated that estrogen induces the phosphorylation of PKM2 at Ser37 through activating the ERK1/2 pathway. Estrogen treatment for 2 hours markedly increased the levels of phospho-PMK2 [Ser37], but had no effect on the protein levels of PKM2, relative to the vehicle control, in 621-101 cells (Fig 2A). Quantitative densitometry of three biological replicates showed the significant increase of 2.8-fold for phospho-PMK2, whereas no significant difference for PKM2 (Fig 2B). Importantly, E 2 -induced PKM2 phosphorylation was concomitant with robust phosphorylation of ERK1/2 in 621-101 cells (Fig 2A). Collectively, these results indicate that estrogen-induced PKM2 phosphorylation is unequivocal and, moreover, is associated with ERK1/2 activation in 621-101 cells.
Next, we assessed the effect of TSC2 expression on PKM2 phosphorylation. We found that the basal level of phospho-PKM2 [Ser37] was significantly lower in TSC2-corrected 621-103 cells relative to that in TSC2-null 621-101 cells (Fig 2A and 2B). Moreover, the level of E 2induced PKM2 phosphorylation was substantially higher in 621-101 (TSC2-) relative to that in 621-103 (TSC2+) cells. Our data indicate that TSC2 negatively regulates PKM2 phosphorylation and suppresses E 2 -induced PKM2 phosphorylation in LAM-derived cells.
To further investigate the effect of estrogen inhibition on activation of PKM2, we treated 621-101 TSC2-null cells with Faslodex, a pure estrogen receptor antagonist, for 24 hours. Faslodex treatment markedly decreased E 2 -induced PKM2 phosphorylation at Ser37, although the protein levels of PKM2 were not significantly affected (Fig 2C). To examine the effect of glucose and estrogen on PKM2 phosphorylation, we cultured 621-101 cells in glucose-rich or glucose-free conditions for 24 hours, and then treated cells with 10 nM E 2 for 2 hours. Under glucose-rich conditions (Glc 17.5 mM), E 2 stimulation largely increased the level of phospho-PKM2 [Ser37] relative to vehicle control (Fig 2C, left panel). Moreover, treatment with faslodex, an estrogen receptor alpha (ERα) antagonist, completely prevented E 2 -induced PKM2 phosphorylation (Fig 2C, left panel). Furthermore, E 2 stimulation did not increase the levels of phospho-PKM2 in 621-101 cells under glucose-free (Glc 0 nM) conditions (Fig 2C, right  panel), further supporting the important impact of glucose and estrogen on PKM2 phosphorylation in TSC2-null LAM patient-derived cells.

Estrogen induces nuclear localization of phosphorylated PKM2
Next, we examined the influence of estrogen on subcellular localization of phospho-PKM2 [Ser37] in TSC2-null 621-101 cells. The phospho-PKM2 [Ser37] was detected clearly as shRNA-PKM2 (#1 and #2) 621-101 cells were treated with E 2 (10 nM) or vehicle over a range of 4 days. Cell growth was measured by crystal violet assay; data were normalized to vehicle control at day 0 (n = 8). (G, I) shRNA-PKM2 (#1 and #2) 621-101 cells were cultured in glucose-rich (Glc 17.5 mM) or glucose-deprived (Glc 0 mM), and then treated with 10 nM E 2 or vehicle for 24 hours (n = 8). Cell growth was measured using crystal violet staining; data were normalized to the vehicle treatment and glucose-rich group. Data are represented as mean ± SEM, �� p<0.01, ns: not significant, two-sided Student's t-test. https://doi.org/10.1371/journal.pone.0228894.g001 Role of PKM2 in the growth of mTORC1 hyperactive cells fluorescent puncta in the nucleus after 30 min estrogen treatment, relative to vehicle control (Fig 3A, I-II). Phospho-PKM2 [Ser37] puncta in nuclei were still found after estrogen treatment for 24 hours (Fig 3A, III), indicating its stability in nuclei in response to estrogen. Importantly, the estrogen-induced nuclear localization of phospho-PKM2 [Ser37] was abrogated or reduced upon treatment with the potent and selective inhibitor of the estrogen receptor Faslodex (Fig 3A, IV), or MAPK inhibitor PD98059 (Fig 3A, V). These results suggest that the phosphorylation of PKM2 at Ser37 and its nuclear translocation are estrogen-dependent in 621-101 cells. Next, we treated TSC2-null 621-101 cells with 10 nM E 2 , E 2 plus 10 μM Faslodex, E 2 plus PD98059, or vehicle for 30 minutes or 24 hours, harvested cells, and performed subcellular fractionation. Immunoblot analysis showed that E 2 treatment for 30 minutes and 24 hours apparently induced the nuclear localization of phospho-PKM2 [Ser37] (Fig 3C). Concomitantly, E 2 treatment for 30 minutes and 24 hours decreased cytoplasmic localization of phospho-PKM2 [Ser37], consistent with findings of immunofluorescent staining. Importantly, Faslodex or PD98059 treatment markedly reduced E 2 -induced nuclear localization of phospho-PKM2. Together, our data using two independent methods demonstrate that E 2 promotes nuclear translocation of phospho-PKM2 [Ser37] in part via MAPK pathway in TSC2-null cells.

TSC2 negatively regulates PKM2 expression in rapamycin-insensitive manner
To determine how PKM2 expression is regulated, we first investigated the effect of TSC2 gene expression on the protein level and phosphorylation of PKM2. We found that TSC2-reexpression markedly decreased PKM2 protein levels by 59% and PKM2 phosphorylation [Ser37] by 85% in 621-101 cells (Fig 4A and 4B). However, to address a related authentication issue pertaining the potential clonal variation in 621-101 isogenic lines, we transiently transfected 621-101 cells with wild-type TSC2 (pcDNA3.1+TSC2) and empty vector pcDNA3.1+. Immunoblot analysis showed that TSC2 overexpression decreased S6 phosphorylation, (Fig 4C), as expected. Importantly, TSC2 overexpression markedly reduced levels of PKM2 phosphorylation by 64% and PKM2 protein by 76% relative to vector control, respectively (Fig 4C). Together, our data indicate that TSC2 negatively regulates PKM2 phosphorylation.
To further investigate whether the specific regulation of PKM2 in TSC2-deficient cells depends on mTORC1 activity, the mTORC1 inhibitor rapamycin was used to treat 621-101 cells lacking TSC2, at indicated time points, followed by immunoblotting. Rapamycin treatment did not affect the protein levels or Ser37 phosphorylation of PKM2, whereas phosphorylation of S6 [Ser235/236] was strongly suppressed (Fig 4D, left panel), as expected. To determine whether glucose availability affects the sensitivity of PKM2 phosphorylation to rapamycin, we cultured 621-101 cells in glucose-free (Glc 0 mM) or E 2 -free conditions. In addition, cells were treated with 10 nM rapamycin for 24 and 48 hours. Rapamycin treatment potently reduced S6 phosphorylation, as expected. However, rapamycin did not affect PKM2 phosphorylation (Fig 4D, middle and right panel). Together, our data indicate that PKM2 phosphorylation is insensitive to rapamycin treatment in either glucose-rich or glucose-free conditions.

Accumulation of phospho-PKM2 is evident in pulmonary LAM nodules from TSC/LAM patients
To determine the clinical relevance of phospho-PKM2, we assessed the abundance of phospho-PKM2 [Ser37] using immunohistochemical staining in two cases of pulmonary LAM lungs. Phospho-PKM2 [Ser37] accumulation was prominent in pulmonary LAM nodule cells that were positive for smooth muscle actin (SMA) and phospho-S6 [Ser235/236] (Fig 6A). Low levels of phospho-PKM2 were observed in SMA-positive bronchial smooth muscle cells in normal lung (Fig 6B). In contrast, immunofluorescent confocal microscopy showed that pulmonary LAM lesion cells accumulated both nuclear and cytoplasmic phospho-PKM2 in the two LAM lungs (Fig 6C). These data indicate that phosphorylation of PKM2 is likely associated with specific tumor growth in LAM lesions.

Expression of ERα and ERβ is evident in TSC2-null cells
Studies have shown that LAM patient-derived 621-101 cells and rat uterine leiomyomaderived ELT3 cells express ERα and respond to estrogen stimulation [20,23,25,26]. To examine the status of the expression of ERα and ERβ expression in these cell models, we measured their transcript levels using quantitative real-time RT-PCR. The relative transcript level of ERα (ESR1) was significantly higher in 621-101 cells (CT = 30.7) relative to lung adenocarcinoma A549 cells (CT = 36.4) (p < 0.001; Fig 7A), although TSC2-reexpression (621-103 cells) did not affect ESR1 expression. However, the transcript level of ESR1 was much lower in 621-101 cells than that in breast cancer MCF-7 cells, similar to previous findings [23]. Importantly, the  (Fig 7A). Moreover, the transcript level of ERβ (Esr2) (CT = 36.6) was much lower than ERα (Esr1) (CT = 32.0) in Tsc2-null ELT3-V3 cells, although TSC2-reexpression did not affect the expression of Esr1 or Esr2 in ELT3-T3 cells (Fig 7B). Role of PKM2 in the growth of mTORC1 hyperactive cells

Discussion
LAM is a devastating lung disease affecting young women that is characterized by metastasis of smooth muscle cells to the lungs and emphysema-like destruction of the lung parenchyma [27][28][29][30]. LAM is associated with TSC2 mutations [31] resulting in activation of the mechanistic target of rapamycin complex 1 (mTORC1) [32], A seminal clinical trial showed that the mTORC1 inhibitor sirolimus stabilizes lung function and improves symptoms in LAM patients while drug exposure continues [33], but long-term benefit and toxicity are unknown, and some patients do not respond. This limitation of sirolimus has engendered multiple efforts to attempt to develop better, or at least complementary, treatment approaches. Despite many advances in our understanding of the importance of mTORC1-dependent and -independent signaling pathways that are central to LAM pathogenesis, the underlying basis of the sexual dimorphism for symptomatic LAM is essentially unknown. In this study, we show evidence that this may at least in part be mediated by estrogen regulation of PKM2. We show that estrogen treatment increases the phosphorylation of PKM2 at Serine 37 and induces the nuclear translocation of phospho-PKM2; treatment with the estrogen receptor antagonist Faslodex Role of PKM2 in the growth of mTORC1 hyperactive cells blocks these actions. Morevover, accumulation of phosphorylated PKM2 was increased in pulmonary nodule cells from TSC/LAM patients.
Faslodex disrupts ligand binding, receptor dimerization and nuclear translocation, and degradation of both ERα and ERβ [34,35]. Studies including ours have shown the expression of both ERα and ERβ in LAM cells [26,36], suggesting the interplay between ERα and ERβ in LAM progression. Opposing effects of estrogen receptor subtypes ERα and ERβ have been implicated in breast cancer cells [37]. ERα plays a pro-proliferative role and ERβ can exert an anti-proliferative action by controlling cell-cycle regulators in most cell types, depending on the differential expression of ER subtypes [38,39]. In our cell-based models, the transcript levels of ERα were higher than that of ERβ in ELT3-V3 cells, but comparable in 621-101 and (Fig  7). Our previous studies have demonstrated that Faslodex inhibits the estrogen-promoted lung metastasis of ELT3 cells in vivo [40], although the specific role of ERα or ERβ was not addressed. In the present study, we observed that Faslodex treatment decreased estrogeninduced phosphorylation of PKM2 and nuclear translocation of phospho-PKM2 (Fig 3A, IV). We speculate that both ERα and ERβ mediate estrogen actions on PKM2 phosphorylation and proliferation of TSC2-null cells. Because agonists and antagonists to both ERα and ERβ are available, it would be of paricular interest to assess the specific impact of ER subtypes on celluar functions, which have not been explored in LAM. We postulate that ERβ-specific agonist would decrease, and ERβ-specific antagonists would increase proliferation and PKM2 activation in LAM cells, which could be addressed in future studies.
Alterations in cellular energy metabolism are a hallmark of cancer [41], and metabolic reprogramming is particularly critical for the survival of Tsc2-deficient tumor cells [14,[42][43][44]. TSC2 deficiency leads to increased transcription of glucose metabolism genes [9] and the cells with mTORC1 activation express high levels of the glycolytic protein PKM2 [10]. The glucose dependent survival of TSC2-deficient cells has been reported [43], suggesting that glucose metabolism is essential for the growth of TSC2-deficient cells. However, the mTORC1 inhibitor, rapamycin (or sirolimus) reduces lactate production but does not affect cellular ATP levels in Tsc2 -/-MEF cells [44]. It has been reported that TSC2-deficient cells exhibit autophagydependent alteration of glucose metabolism rewiring to pentose phosphate pathway [42]. Together, these reports highlight connections between cellular metabolic alterations and glucose utilization that likely impact the survival of TSC2-deficient cells.
Importantly, our study shows that phosphorylated PKM2 is evident in the nucleus of LAM patient-derived cells in vitro and in pulmonary LAM nodule cells in vivo. Phosphorylated PKM2 [S37] and its nuclear translocation promote the Warburg effect and tumorigenesis [12]. Monomeric PKM2 translocates into the nucleus, where it functions as a transcriptional coactivator of β-catenin and upregulates the expression of c-Myc and cyclin D1 [48], thereby promoting the Warburg effect and cell cycle progression, respectively [13]. These findings will warrant future investigation of the important role of PKM2 in estrogen-driven LAM progression.
LAM can lead to respiratory failure and death [49][50][51]. The Multicenter International LAM Efficacy of Sirolimus Trial (MILES Trial) demonstrated that the mTORC1 inhibitor sirolimus (rapamycin) stabilizes lung function and improves the symptoms in women with LAM. However, lung function decline resumed upon drug cessation [33]. Therefore, despite advances in the clinical care of women with LAM, there remains a critical need for improved therapeutic options. PKM2 expression can only be suppressed by TSC2 reconstitution and is not significantly affected by the mTORC1 inhibitor rapamycin, or Raptor depletion using shRNA, suggesting that TSC2 deficiency upstream of the mTORC1 pathway is the leading cause of PKM2 upregulation. Interestingly, we also showed that knockdown of mTORC2 component Rictor does not affect the protein of levels or phosphorylation of PKM2, indicative of mTORC2-indepent regulation of PKM2 expression and activation in TSC2-deficient cells. In this study, our data suggest that PKM2 upregulation is likely a direct consequence of TSC2 loss.
Collectively, our data suggest that inhibiting estrogen-dependent cellular metabolic pathways could block the pro-survival effects of estrogen on LAM cells without need to ablate the entire hormonal signaling axis. Hormonal ablation therapy for breast cancer patients decreases circulating hormone levels and increases the development of osteoporosis and bone fracture [52,53]. Thus, in the long term, it is possible that, compared to hormonal ablation, metabolically-focused strategies in LAM could have preferable side effect profiles with regard to bone health and biochemical parameters including serum calcium, serum phosphorus and bone specific isoform of alkaline phosphatase [54]. Recent studies have demonstrated therapeutic potentials of targeting dysregulated cellular metabolic pathways including glucose metabolism and autophagy addication using hydroxychloroquine or resveratrol in LAM [25,[55][56][57][58]. However, additional pre-clinical and clinical studies will be needed to test this concept.