Lack of Liver X Receptors Leads to Cell Proliferation in a Model of Mouse Dorsal Prostate Epithelial Cell

Recent studies underline the implication of Liver X Receptors (LXRs) in several prostate diseases such as benign prostatic hyperplasia (BPH) and prostate cancer. In order to understand the molecular mechanisms involved, we derived epithelial cells from dorsal prostate (MPECs) of wild type (WT) or Lxrαβ−/− mice. In the WT MPECs, our results show that LXR activation reduces proliferation and correlates with the modification of the AKT-survival pathway. Moreover, LXRs regulate lipid homeostasis with the regulation of Abca1, Abcg1 and Idol, and, in a lesser extent, Srebp1, Fas and Acc. Conversely cells derived from Lxrαβ−/− mice show a higher basal phosphorylation and consequently activation of the survival/proliferation transduction pathways AKT and MAPK. Altogether, our data point out that the cell model we developed allows deciphering the molecular mechanisms inducing the cell cycle arrest. Besides, we show that activated LXRs regulate AKT and MAPK transduction pathways and demonstrate that LXRs could be good pharmacological targets in prostate disease such as cancer.


Introduction
Patient cohorts and human cancer cell lines give suitable biological samples to identify genetic defects correlated with disease but are still limited to investigate the molecular mechanisms involved. In vivo mouse models of prostate cancer represent a powerful tool to complement these human studies. A second advantage is that mice provide the possibility to derive long-term cell culture system genetically engineered. Hence, we have developed a useful cell culture system of prostatic epithelial cells (MPECs) from wild-type (WT) and transgenic mice with genetic ablation of both isoforms of the Liver X Receptor (Lxr). LXRa (NR1H3) and b (NR1H2) belong to the nuclear receptor superfamily and act as ligand-inducible transcription factors when heterodimerized with the Retinoid X Receptor (RXR, NR2B1 to B3). LXRs are involved in numerous physiological regulations such as cholesterol, fatty acids and glucose homeostasis, steroidogenesis and immunity [1,2]. Various studies have highlighted the therapeutic potential of LXR agonists in the treatment of prostate diseases such as benign prostatic hyperplasia (BPH) [3,4] and prostate cancer [5,6]. Indeed, LXR agonists like T0901317 slow down proliferation of various prostate cancer cell lines and decrease growth of prostate cancer cells in xenografted nude mice [7][8][9]. This lower proliferation of human prostatic cancer cell line LNCaP is characterized by a reduced amount of cells in S phase correlated with an accumulation of SKP2, a kinase involved in the degradation of cell cycle inhibitors. This cell cycle regulation was proposed to be connected with cholesterol homeostasis through LXR-regulated expression of the cholesterol transporter ABCA1 that supports cholesterol efflux [10]. Additionally, LXR activation by the synthetic ligand T0901317 results in a decrease of cell survival. The lowering of cellular cholesterol induced by the cholesterol transporter ABCG1 accumulation leads to a reduction in lipid raft size inactivating the AKT pathway and consequently limiting cell survival ability [9]. These previous studies demonstrated a putative combined anti-proliferative and pro-apoptotic effect of LXR activation in prostate cancer cell lines, suggesting a protective role of these nuclear receptors in this cancer via their effects on the cell cycle and apoptosis balance [11].
Even if prostate is anatomically different between human and mouse since the rodent prostate is divided into distinct lobes whereas three zones are found in human [12], transcriptional analyses linked the mouse dorsolateral lobe to the human peripheral zone which develops cancer [13].
Here, we present an original prostate epithelial cell model derived from the dorsal lobe of WT or Lxrab2/2 mice. This new cell models allowed deciphering the role of LXRs on cell cycle regulation.

Genotyping
Lxra and Lxrb loci with the neo cassette as well as the location of the primers are indicated on

Cell Culture and Treatment
Mouse embryonic fibroblasts (MEFs) were cultured in DMEM supplemented with L-glutamine (2 mM) and penicillin streptomycin (100 mg/ml). MPECs were trypsinized and seeded at 3610 5 cells in 10 mm culture dishes. After 16h cells were starved for 24h in a minimal medium DMEM-F12 (50:50, Invitrogen) supplemented with L-glutamine (2 mM) and penicillin streptomycin (100 mg/ml) and then with the same medium containing DMSO, T0901317, GW3965 or 22(R)-hydroxycholesterol for 48h. Cells were harvested and protein or RNA extraction, or flow cytometry analyses were performed.

Flow Cytometry Analysis
Cells were recovered and washed in PBS. Pellets were resuspended in RNAse A (500 mg/ml), propidium iodide (50 mg/ml) solution and kept 1h at 4uC in the dark. Cell suspensions were analyzed using a Beckman Coulter fluorescence-activated cell sorter. At least 15,000 events were measured for each sample.

Statistical Analysis
Values are expressed as means 6 SEM. Statistical comparisons were performed using a two-tailed Student's t test. A p, 0.05 was considered statistically significant.

Establishment of MPECs from Mouse Dorsal Prostate Lobe Explant
To generate MPEC culture, dorsal prostate lobes were dissected from wild-type (WT) and Lxrab2/2 mice (Fig. 1A). Prostate samples were minced and seeded onto ECM-coated plates in epithelial specific defined-medium (procedure details were recapitulated in Methods section). Cells migrated from the explant and culture was maintained until auto-immortalization of epithelial cells. Upon immortalization, cell genotypes were checked by PCR (Fig. 1B), and Lxra and Lxrb transcript accumulations were determined and compared to WT MEFs (Fig. 1C). It appears that Lxra slightly shows similar levels between MPECs and MEFs, while Lxrb is more accumulated in MPECs than in MEFs. As expected, both messengers were undetectable in Lxr2/2 MPECs.
Once cultured, MPECs formed islands and displayed typical cobblestone features of epithelial cells compared to spindle shape of MEFs ( Fig. 2A). In order to check for contamination of the culture by stromal fibroblasts, stromal and epithelial differentiation markers were investigated. Absence of both Tenascin and Snail expressions avoid potential contamination by mesenchymal cells   (Fig. 2B). Immunofluorescence and qPCR detection of specific markers E-cadherin and Cytokeratin 8 confirmed the epithelial origin of the culture (Fig. 2A,B), and as expected these markers were absent in MEFs. The epithelial characteristics of the MPECs were strengthened by accumulation of CK18 and PSCA proteins (Fig. 2C). We concluded that MPECs could be considered as a good tool to study the physiological role of LXRs in epithelial cells from dorsal prostate.

LXRs Control Cell Proliferation in MPECs
LXR agonists have been previously described to exert antiproliferative activities [16]. Thus, we investigated the effect of Lxrablation in MPECs. Treatments with various concentrations of 22R-hydroxycholesterol, a natural LXR-ligand, or GW3965, a synthetic LXR-ligand, on WT MPECs result in a dosedependent accumulation of cells in G0/G1 phases and conversely a decrease of the percentage of cells in S phase (Fig. 3A). These observations show that the effect of the ligands on cell cycle is also visible in non-cancerous cells and is rather an intrinsic property of prostate epithelial cells. The effect of T0901317 (synthetic LXR ligand) was analyzed on Lxrab2/2 MPECs and, as expected, had no effect on cell cycle (Fig. 3B) and proliferation (Fig. 3C). Interestingly, Lxrab2/2 MPECs exhibit a higher basal proliferation ( Fig. 3B and D). LXRs are thus regulators of cell cycle, which correlates with Cyclin D1 and Cdk2 protein accumulation (Fig. 3E). While GW3965 or T0901317 treatment decreases proliferation in WT MPECs, these ligands have no effect on Lxrab2/2 MPECs (Fig. 3B). It could thus be concluded that GW3965 or T0901317 cell cycle inhibition is strictly LXRdependent in MPECs, without potential interferences with other nuclear receptors. Interestingly, although T0901317 induction increased p21 cell-cycle inhibitor accumulation in WT MPECs that is consistent with proliferation decrease, Lxrab2/2 MPECs reached a high and constitutive level of p21 that contrasts with the high proliferative rate (Fig. 3E). Paradoxically, p21 targets Cdk2/ 4, CyclinD1 and CyclinE were increased in Lxrab2/2 MPECs indicating that driving forces leading to proliferation overcome p21 regulation. Concomitantly to p21 deregulation in Lxrab2/2 MPECs, p27, initially identified as a downstream target of LXR activation in LNCaP cells [7], exhibit a strong accumulation as well. As observed in prostate cancer PTEN-null mice model, we postulated that p21 and p27 accumulation level in Lxrab2/2 MPECs could result from a defensive mechanism to counterbalance the proliferation increase due to the combined lack of Lxra and Lxrb [17].

LXRs and Lipid Homeostasis in MPECs
As LXRs are transcription factors involved in cholesterol and fatty acids metabolism regulation, we have analyzed the effects of LXRs on these metabolic pathways. WT and Lxrab2/2 MPECs were treated with increasing amounts of T0901317 and expression of well-known LXR target genes was analyzed by qPCR. As expected, WT MPECs displayed increased accumulation of Abca1 and Abcg1, encoding cholesterol transporters, after LXR activation (Fig. 4A, left). A higher basal level of Abca1 was observed in Lxrab2/2 MPECs, as already described [18]. Conversely, no increased accumulation of Abca1 and Abcg1 was seen in Lxrab2/2 MPECs. Idol whose product controls the degradation of LDLreceptor (LDLR) is induced by T0901317 (Fig. 4A, right), indicating that the LXR target genes involved in cholesterol homeostasis are inducible in MPECs. Co-stimulation of LXR/ RXR heterodimer by T0901317 and 9-cis retinoic acid reproduces the canonical synergistic transcriptional activation of Abca1 expression (Fig. 4B), thus demonstrating that LXR-signaling pathway is fully functional in immortalized MPECs. Interestingly, LXR-target genes involved in fatty acid synthesis such as Srebp1c, Acc and Fas showed a modest but significant increase in their expression when LXRs are activated in WT MPECs (fold range induction 61.7 to x2.4) (Fig. 4C). These findings were correlated with western blot analyses in which WT MPECs displayed increased protein accumulation of ABCA1 and the non-cleaved form of SREBP1c in response to T0901317, conversely to what was observed in Lxrab2/2 MPECs (Fig. 4D). Altogether, expression profiles of genes involved in lipogenesis postulate for a tenuous storage of triglycerides in MPECs in response to LXR activation. Oil red-O staining confirm that point since a few number of cells exhibits lipid droplets filling (Fig. 4E) in WT MPECs compared to MEFs treated with T0901317 (Fig. 4F). These results show that in WT MPECs, LXRs can be efficiently activated and subsequently reproduces canonical LXR mediated lipid metabolism regulation.

LXRs are Connected to AKT and MAPK Transduction Pathways
Previous studies showed that LXR activation could impact transduction pathways such as PI3K/AKT signaling in human prostatic cancer cell lines [9]. To examine whether such regulation was still present in MPECs, we monitored various protein accumulations and/or phosphorylations involved in this transduction pathway. Both AKT phosphorylation of Serine 473 (pAKT S473) and Threonin 308 (pAKT T308) are sensitive to T0901317 or 22(R)-hydroxycholesterol treatment, as described in LNCaP cells (Fig. 5A,B). As expected this sensitivity was not seen in Lxr-deficient cells (Fig. 5A,B). However it has been puzzling to observe that AKT phosphorylation level was basally found substantially higher (Fig. 5A,B) and insensitive to IGF-I ( Fig. 5C and data not shown) in Lxrab2/2 MPECs. In order to investigate whether AKT phosphorylation modifications induced by LXR stimulation or ablation was relevant, we monitored phosphorylation levels of GSK3b and Bad, two known downstream targets of AKT. As expected, both proteins exhibited a decreased phosphorylation in WT MPECs treated with T0901317 or 22(R)hydroxycholesterol and a strong increased of basal phosphorylation in Lxrab2/2 MPECs (Fig. 5A). These results were paralleled by immunofluorescence detection of phosphorylated AKT. Indeed a similar membrane staining to those observed in LNCaP cells [9] could be seen. This signal was lost in WT MPECs incubated with T0901317, but remained persistent in Lxrab2/2 MPECs whatever the conditions. Abnormal transduction signaling in Lxrab2/2 MPECs is not restricted to AKT pathway as phosphorylation forms of MAPK (pp42/pp44) were also basally found highly accumulated in these cells (Fig. 5A,B). These observations on PI3K/AKT and MAPK transduction pathway deregulations suggest that they could sustain proliferation activity and act as mediators of the LXR effects. Indeed, incubation of WT and Lxrab2/2 MPECs in a medium supplemented with increasing amounts of pharmacological inhibitors of PI3-Kinase (LY-294002) or MEK (PD-98059) leads to a strong inhibition of cell growth. Consistent with the deregulated activity of AKT and MAPK in Lxrab2/2 compared to WT MPECs, Lxrab2/2 cells exhibit a higher sensitivity to both LY-294002 and PD-98059 treatments in MTT assays (Fig. 5D). Taken together, these results provide evidences that LXRs are important regulators of PI3K/ AKT and MAPK transduction pathways, two central crossroads that mediate cell cycle control in MPECs.

Discussion
This study characterizes MPECs derived from WT and Lxrab2/2 mice. Our results indicate that LXRs are necessary to restrain cell cycle in non-cancerous prostate cells. Moreover, various LXR agonists, e.g. T0901317, GW3965 and 22(R)hydroxycholesterol, clearly exert an anti-proliferative effect through LXRs, excluding the involvement of other nuclear receptors in the process we identified in MPECs. Molecular links between LXR target genes and proliferation process remains however to be elucidated. In the present study, we show that LXRs have a crucial role in the regulation of genes involved in cholesterol homeostasis and in a lesser extent in fatty acid synthesis in MPECs. Finally, we show that LXRs modulate AKT and MAPK phosphorylation accumulation, making them potential mediators of LXRs in the control of cell cycle.
LXRs were shown to modulate cell proliferation when they are activated by synthetic or natural ligands in prostate tumor cell lines [7]. Nevertheless, several studies questioned the specificity of the ligands used. Thus, some of these ligands have been proposed to interfere with distinct nuclear receptors [19]. In Lxrab2/2 MPECs, none of the tested LXR ligands (T0901317, GW3965) displayed modulation of proliferation, demonstrating that LXRs are involved in the control of cell cycle in this context.
The results raise the question of the link between LXRs and the cell cycle effectors. Fukuchi et al. suggested ABCA1 to be a key-regulator of the cell cycle in response to LXR-activation in LNCaP cells [10] and suspected that cholesterol efflux is of major importance in this context. Consistent with these previous observations, LXR-target genes involved in cholesterol homeostasis (e.g. Abca1, Abcg1 and Idol) were found up-regulated in WT MPECs in response to LXR ligands. These facts indicate that MPECs are responsive to cholesterol network homeostasis such as efflux/uptake activities. To date, various studies identified fatty acid metabolism as a crossing point in prostate tumor cell lines that is tightly associated to cell growth and proliferation [20][21][22]. Indeed, inhibition of acetyl CoA carboxylase (ACC) by Soraphen A efficiently blocks the proliferation of LNCaP cells, showing that tumor cell growth is dependent of fatty acid synthesis [23]. In MPECs, Srebp1, Fas and Acc display a significant but modest induction. Accordingly, the weak neutral lipid accumulation of MPECs (Fig. 5B) correlated the fact that induction of the fatty acid synthesis gene program is poorly sensitive to LXR regulation in these cells. Futhermore, the present work suggests that the capacity of LXRs to modulate cell cycle in MPECs is mostly dependent of their regulation of cholesterol homeostasis.
Unexpectedly, we observed a higher basal phosphorylation of AKT in MPECs lacking LXRs. It could be postulated that AKT signaling perturbation is a consequence of cholesterol metabolism alteration in Lxrab2/2 MPECs. This hypothesis is consistent with the previous study demonstrating that LXRs control AKT phosphorylation levels in a raft-dependent manner in LNCaP cells [9].
Altogether, we show that LXRs per se modulate cell cycle in noncancerous epithelial cell model. Lxrab2/2 MPECs provide thus a powerful tool to investigate intrinsic connections between oxysterols, cholesterol metabolism and cell proliferation.