GPCR-like signaling mediated by smoothened contributes to acquired chemoresistance through activating Gli

Background Smoothened (Smo), which possesses a structural similarity with classic G-protein coupled receptors (GPCR), is the most important molecular target in Hedgehog (Hh) signaling system for developing anticancer drugs; however, whether Smo may transmit GPCR-like signaling to activate the canonical transcriptional factor Gli of Hh signaling system and consequently to be involved in the Gli-dependent biological events remains controversial. Results In this study, using the acquired chemoresistant cancer cell lines and their respective parental cells, we found that Smo may activate Gli through Gαi, Gβγ-JNK signaling axis, thereby promoting the Gli-dependent acquired chemoresistance. These observations were further complementarily strengthened by data obtained from chemosensitive cancer cells with artificially elevated Hh pathway activity. Conclusions Hence, our data demonstrate that GPCR-like signaling mediated by Smo contributes to the acquired chemoresistance through activating the canonical Hh transcriptional factor Gli; therefore improving our knowledge of the nature of the signal transduction of Smo and the molecular mechanisms responsible for the acquired chemoresistance maintained by Hh pathway. Moreover, our data that JNK after activated by Smo-Gβγ signaling axis may stimulate the Gli activity and consequently promotes acquired chemoresistance expose a promising and potential target for developing anti-cancer drugs aimed at Hh pathway and for combating the acquired resistance raised by using of anti-cancer drugs targeting Smo.


Introduction
The Hedgehog (Hh) signaling pathway plays critical roles for embryonic development and postnatal tissue homeostasis in organisms ranging from insects to mammals [1]. The activation Hh pathway is initiated by binding of the secreted Hh proteins, including Sonic, Indian and Desert Hedgehog (SHh, IHh and DHh, respectively) to the 12-transmembrane receptor Patched (Ptch), thereby liberating the Ptch-mediated inhibitory effect on Smoothened (Smo), a 7-pass transmembrane protein and a key component in the Hh signaling pathway. This subsequently causes the accumulation of Smo in the primary cilium and a series of consequent intracellular events, finally resulting in the activation of the canonical transcriptional factor Gli which consists of Gli1, Gli2 and Gli3 [2]. Analogous to other pathways active during embryonic development, inappropriate Hh pathway activity has been demonstrated to be critical for the initiation and progression of various kinds of tumors. Aberrant Hh pathway activity for tumors may occur either by mutations in key components of Hh pathway or by the production of Hh ligands in tumor cells in an autonomous and non-autonomous manner [3].
Given that the addiction of many types of tumors to aberrant Hh pathway activity, a variety of antagonists targeting Hh pathway have been developed for the treatment of cancers. Among them, the majority function as inhibitors of Hh pathway by targeting Smo, a critical component for canonical Hh pathway [4]. In this regard, dissecting the characteristics of signal transduction elicited from Smo is crucial and an area of intense investigation, as it may help us with the development of antagonists targeting Smo and its downstream effectors.
Smo, which possesses a structural similarity with classic G-protein coupled receptors (GPCR), has long been suggested to couple with heterotrimeric G proteins [5][6][7]. Indeed, it has been shown that Smo may interact with Gαi and subsequently acitivate the transcriptional activity of Gli (Ci in drosophila cells) in Drosophila Cl8 cells, Sf9 cells, and NIH3T3 cells, indicating the requirement of Gαi for the activation of Gli mediated by Smo [8,9]. However, this argument is challenged by the observations that pertussis toxin (PTX), which may ADPribosylate and consequently uncouple Gi from GPCR, fails to impact the Gli-dependent biological events such as chick neural tube patterning and some patterning events in zebrafish embryonic development [5,10]. On the other hand, after coupled to Gαi, Smo may exert a variety of biological activities independently of Gli, such as migration of murine embryonic fibroblasts, tubulogenesis of endothelial cells, and calcium spike activity of embryonic spinal cells [11][12][13]. Moreover, recent study indicate that Smo may contribute to the survival of diffuse large B-cell lymphoma cells by coupling to Gαi and Gα12 and subsequently activate NF-κB independently of Gli [14]. These studies suggest that the association between heterotrimeric Gαi proteins and Smo remains far from being fully understood, especially in the context of cancer biology. Meanwhile, in the case of canonical signal transduction of GPCR, ligand binding causes conformational changes in the structure of GPCRs, endowing them with abilities to function as a guanine nucleotide exchange factor (GEF). The exchange of GDP for GTP at the Gα subunit induces its dissociation from Gβγ dimmer [15]. To our knowledge, whether and how Gβγ, after dissociated from Gα subunit, may impact the Smo dependent Gli activity remains as well unclear.
Although great achievements have been made for the molecular-targeted anti-cancer drugs, traditional chemotherapy is still one of the most efficient approaches for treatments of cancers. Many studies have shown that Hh signaling pathway activity plays critical roles in maintaining the chemoresistant phenotype of acquired chemoresistant cancer cells [16][17][18][19][20][21][22][23]. In this study, utilizing well established acquired chemoresistant cancer cell lines and their respective parental ones, we provide a series of complementary evidences to show that Smo may promote acquired chemoresistance by activating Gli through Gαi and Gβγ−JNK signaling axis, therefore revealing that GPCR-like signaling elicited from Smo is involved in the canonical Hedgehog-Gli signaling pathway activation and the acquired chemoresistance.

Cell culture
The K562 human chronic myelogenous leukemia cell line, KB human epidermoid carcinoma cell line, NIH-3 T3 mouse embryo fibroblast cells, and HEK293T human epithelial kidney cells were purchased from the American Type Culture Collection and cultured according to the manufacturer's instructions. The Dox selected multidrug tolerant K562/A02 subline was obtained from the Institute of Hematology, Chinese Academy of Medical Sciences (Tianjin, China), which was routinely maintained in medium containing 200 ng/ml of Dox [24]. The VCR selected multidrug tolerant KB/VCR subline was obtained from Zhongshan University of Medical Sciences (Guangzhou, China) and was routinely maintained in medium containing 200 ng/ml of VCR [25]. Both resistant cells were authenticated by comparing their fold resistance with that of the parental cells and examining the expression levels of ABC transporters. All experiments using K562/A02, and KB/VCR cells were performed with cells growing in the absence of Dox or VCR for at least 5-7 days to avoid drug associated secondary effects.
Transient transfections were performed using Lipofectamine 2000 reagent (Invitrogen, Grand Island, NY) according to the manufacturer's instructions. The viral stocks were prepared and infections were performed according to previously reported [26].

Cell viability assay
The MTT assay was conducted as previously described to determine the sensitivity of cells to chemotherapeutic drugs [27].

Dual-luciferase reporter assay
Cells were seeded into 48-well plates. Twenty-four hours later, cells were cotransfected with luciferase expression constructs as indicated and Renilla luciferase using Lipofectamine 2000 (Invitrogen). Luciferase activities present in cellular lysates after indicated treatments were measured using a Dual-Luciferase reporter assay system from Promega (Madison, WI) according to the manufactu-rer's instructions and a luminometer (Molecular Device; Sunnyvale, CA). The firefly luciferase values were normalized to Renilla values.

Acquired chemoresistant cancer cells exhibit aberrant cell-autonomous Hh pathway activity
The requirement of Hh pathway activity for maintaining the acquired chemoresistance indicates that acquired chemoresistant cancer cells may harbor aberrant Hh pathway activity via cell autonomous manner [3]. Elaborate verification of this argument is a prerequisite for dissecting the nature of the signal transduction from Smo to Gli in acquired chemoresistant cancer cells. In this regard, we first examined the expressions of ligands of Hh pathway in acquired chemoresistant cancer cells compared to their respective parental ones. Using two well established acquired chemoresistant cancer cells K562/ A02, KB/VCR and their respective parental cells human chronic myelogenous leukemia cell line K562, human epidermoid carcinoma cell line KB, we found that the abundance of Hh ligands SHh, IHh and DHh were all obviously elevated when compared to their respective parental cancer cells as revealed by QT-PCR analysis ( Figure 1A), suggesting the possibility of cell-autonomous Hh pathway activity harbored by acquired chemoresistant cancer cells. Next, we set out to evaluate whether the elevated production of Hh ligands correlates with the aberrant Hh pathway activity in acquired chemoresistant cancer cells using Gli-luciferase assay to rule out the non-cell autonomous Gli activation [3]. We observed that the chemoresistant cancer cells harbored aberrant Hh pathway activity relative to respective parental cells ( Figure 1B). Meanwhile, treatment of acquired chemoresistant cancer cells with Robo and cyc, specific small molecular inhibitors targeting SHh and Smo, respectively [28,29], caused significant reductions of the aberrant Hh pathway activity in acquired chemoresistant cancer cells K562/A02 and KB/VCR, whereas both Robo and cyc did not affect Hh pathway activity in respective chemosensitive cells ( Figure 1B). Moreover, tomatidine, a steriodal alkaloid structurally similar to cyc and lacking activity against Hh pathway [30], exhibited no effect on the Hh pathway activity in both chemoresistant and respective chemosensitive cancer cells ( Figure 1B). These observations derived from Gli-luciferase reporter assay were faithfully recapitulated by QT-PCR analysis of the Gli1 ( Figure 1C), a transcriptional target of Hh pathway and served as readout of the Hh pathway activity. Considering that the chemoresistant cancer cells harbor aberrant cell-autonomous Hh pathway activity, it is conceivable that interference Hh pathway by targeting its ligands may sensitize acquired chemoresistant cancer cells to chemotherapy. Indeed, we found that exposure of chemoresistant cancer cells K562/A02, KB/VCR to Robo resulted in obviously sensitizing them to Dox and VCR, respectively ( Figure 1D), while Robo treatment failed to impact the sensitivity of K562 and KB cells to Dox and VCR (data not shown). Hence, these data together clearly demonstrate that chemoresistant cancer cells harbor cell-autonomous Hh pathway activity.

Gαi and Gβγ are involved in mediating the Hh pathway activity in chemoresistant cancer cells
Having established that acquired chemoresistant cancer cells harbor cell-autonomous Hh pathway activation and represent ideal models for dissecting the signal transduction nature of Hh pathway, we then asked whether and   [8,9], we first set out to examine whether interference with Gαi may repress the Gli activity in chemoresistant cancer cells. Exposure of chemoresistant cancer cells to PTX, which may uncouple Gαi from receptor activation by ADP ribosylating Gαi [31], obviously suppressed the transcriptional activity of Gli in K562/A02 cells and KB/ VCR cells as revealed by Gli-luciferase reporter assay (Figure 2A Figure 2C) and Gαt ( Figure 2D) obviously reduced the Gli-luciferase activity in response to SAG. These observations obtained by using KB/VCR cells were well recapitulated in NIH-3 T3 cells exposed to SAG ( Figure 2E-F). Moreover, we observed that transfection of Gβ1 and Gγ2 plasmids into NIH-3 T3 cells remarkably stimulated the Gli-lucidferase reporter activity ( Figure 2G); further demonstrating that Gαi and Gβγ upon dissociated from Gαi may transmit the signal from Smo to Gli. Taking these results together, we can conclude that Smo may couple to Gαi and both Gαi and Gβγ are involved in activating Gli mediated by Smo in chemoresistant cancer cells. Next, we examined whether interference with Gαi and Gβγ may circumvent the chemoresistance of acquired chemoresistant cancer cells. We found that treatment of KB/VCR cells ( Figure 2H) with PTX or ectopic expression of Gαt into K562/A02 cells ( Figure 2I) by lentivirus approach resulted in remarkably sensitizing the KB/VCR cells and K562/A02 cells to chemotherapeutic drugs VCR and Dox, respectively. Moreover, these reversals of acquired chemoresistance elicited by PTX and Gαt were accompanied by the repressions of Hh pathway activity in KB/VCR cells and K562/A02 cells as reflected by the reductions of transcript levels of Gli1 ( Figure 2J-K). Meanwhile, we observed that treatment of K562 and KB cells with PTX or transfection of Gαt did not increase the sensitivity of K562 and KB cells to chemotherapy (data not shown). Therefore, these data further demonstrate that Gαi and Gβγ are required for Smo-mediated Gli activation and consequently for Glidependent chemoresistance in acquired chemoresistant cancer cells.
Gαi and Gβγ are required for the chemoresistance promoted by reconstituted Hh pathway activity in KB cells Gβγ may promote Gli activity through JNK in chemoresistant cancer cells JNK is a well known downstream effector of Gβγ [35]. We then asked whether Gβγ, when released from Gαi after Smo activation, may activate Gli via JNK. To this end, we first tested whether inhibition of the Hh pathway may repress JNK activation in chemoresistant cancer cells. Exposure of K562/A02 cells with Robo ( Figure 4A) or treatment of K562/A02 cells and KB/ VCR cells with cyc ( Figure 4B) led to reductions of the phosphorylation of JNK, indicating that Hh signaling may activate JNK in chemoresistant cancer cells. This was further supported by the observation that SHh significantly provoked JNK activation in 293 T cells, reaching a maximum at 10 min ( Figure 4C). Next, we set out to determine the requirement of JNK activation for Gli activity in chemoresistant cancer cells. Inhibition of JNK in K562/A02 cells by JIP ( Figure 4D), a peptide inhibitor specifically targeting JNK [36,37], or by transfection K562/A02 and KB/VCR cells ( Figure 4E) with JNK1 (APF), a plasmid of dominant negative mutant of JNK [38], abundantly impaired the Gli activity in chemoresistant cancer cells as judged by Gli-luciferase reporter assay, therefore suggesting that JNK is required for maintaining the cell-autonomous Gli activation in acquired chemoresistant cancer cells. This argument was further confirmed by the results that JIP obviously abolished the Gli activity provoked by SAG in KB/VCR cells ( Figure 4F). Together, these findings suggest that JNK is involved in the cell-autonomous Gli activation in chemoresistant cancer cells.
We next set out to examine whether Gβγ may activate Gli though JNK in acquired chemoresistant cancer cells. We observed that treatment of chemoresistant cancer cells K562/A02 ( Figure 5A) and KB/VCR ( Figure 5B K562/A02 ( Figure 5C,D). Hence, these results suggest that Gβγ may mediate Gli activation elicited by Smo through JNK. This argument was further strengthened by the observation that JIP significantly inhibited the Gli-luciferase reporter activity provoked by transfection of Gβ1 and Gγ2 subunits into NIH-3T3 cells ( Figure 5E). Collectively, our findings demonstrate that after released from Gαi, Gβγ may activate Gli via JNK in chemoresistant cancer cells.

JNK activity is required for maintaining the chemoresistance maintained by Hh pathway
Having established that Gβγ may transduce the signaling from Smo to Gli via activating JNK in chemoresistant cancer cells, we next tested the biological relevance of Gli activation mediated by JNK to chemoresistance using acquired chemoresistant cancer cells. We saw that interference of JNK function with JIP or with forced expression of JNK dominant negative mutant JNK1a1 (APF) by lenti-virus approach significantly sensitized the K562/ A02 cells to Dox ( Figure 6A,C), concomitantly accompanying the reductions of the expressions of Gli1 at mRNA level ( Figure 6B,D). Given that JNK may promote chemoresistance through activating Gli, it is conceivable that artificial activation of JNK in chemosensitive cancer cells may result in Gli activation and subsequently chemoresistance. Indeed, taking advantage of the MKK7 and JNK1 fusion plasmid (MKK7-JNK1) engineered into a lenti-virus vector, which may activate JNK activity [39], we observed that artificial JNK activation rendered chemosensitive cancer cells K562 tolerant to Dox ( Figure 6E), simultaneously increasing the Gli activity as reflected by QT-PCR analysis of the Gli1 expression ( Figure 6F); while the negative control MKK7-JNK1(APF) for MKK7-JNK1 did not impact either the sensitivity of K562 cells to Dox ( Figure 6E) or the expression of Gli1 at mRNA level ( Figure 6F). Interestingly, GANT58, a small molecular antagonist specifically targeting Gli [40], restored the sensiti- vity of K562 cells with ectopic expression of MKK7-JNK1 to Dox ( Figure 6E). Taken together, these results complementarily demonstrate that JNK may activate Gli in chemoresistant cancer cells, thereby maintaining the chemoresistance phenotype.
We next investigated whether the JNK activation is required for the chemoresistance promoted by ectopic expression of SmoA1. To this end, we artificially activated the Hh pathway activity using  Figure 7C) and BCR-ABL tyrosine kinase inhibitors Imatinib ( Figure 7D), simultaneously accompanying increased Hh pathway activity as reflected by enhancement of Gli1 mRNA expression ( Figure 7E). Moreover, JNK1 (APF) restored the sensitivity of the K562 cells with artificial elevated Hh pathway to Dox, VP16 and Imatinib ( Figure 7C-D), concomitantly reducing the expression of Gli1 provoked by SmoA1 ( Figure 7E). Collectively, our findings further confirm that JNK is involved in the chemoresistance mediated by Hh pathway and that after dissociation from Gαi initiated by Smo activation, Gβγ may stimulate the Gli activity through JNK and subsequently promote chemoresistance.

Discussion
Hh signaling pathway has been shown to be critical for a variety of physiological and pathological conditions, such as embryonic patterning, maintenance of postnatal tissue homeostasis, as well as initiation and progression of cancers [41], whereas the molecular mechanisms responsible for its signaling transduction remain to be fully understood. Because of the structural homology with classical GPCRs, Smo has been suspected to have the ability to couple with heterotrimeric G proteins Gαi [5][6][7]. However, the contribution of Gαi to the Hh signaling transduction is quite controversial and unclear, especially in the cancer biology. Information available so far suggests that it is context-dependent and cell-type dependent for the ability of Smo coupling to Gαi and for the subsequent participation of Gli in the biological significance initiated by the interaction of Smo and Gαi [42,43]. In the present study, we provide complementary evidences to show that both Gαi and Gβγ are required for the Hh pathway activity and the subsequent acquired chemoresistance by activating its canonical transcriptional factor Gli, confirming the ability of Smo coupling to Gαi and the requirement of Gαi for the Glidependent biological significance in the context of acquired chemoresistance. Moreover, we found that Gβγ, after released from Gαi, are also be involved in the Gli activation and acquired chemoresistance through activating JNK. Indeed, by artificially increasing the Hh pathway activity in chemosensitive cancer cells, we determined that both Gαi and Gβγ-JNk signaling axis are required for the Gli activity and Gli-dependent acquired chemoresistance mediated by SmoA1. Our data that GPCR-like signaling mediated by Smo contributes to the acquired chemoresistance by activating Gli improve our interpretations of the underlying mechanisms for the acquired chemoresistance promoted by Hh pathway and help us with improving the chemotherapeutic efficiency by using Hh inhibitors [44]. Meanwhile, this study shed light on the understanding the nature of signaling transduction of Smo in cancer biology. Deregulated Hh signaling has been implicated in a wide range of cancers, such as medulloblastoma, basal cell carcinoma, glioblastoma, leukemia, breast cancer, pancreatic cancer, prostate cancer, lung caner, colon cancer, to name a few [45]. Aberrant Hh pathway activity may result from gain-of-function and loss-of-function mutations in key components in Hh pathway, such as PTCH, Smo, and Sufu [3]. Hh pathway may as well be activated in tumors by overexpression of Hh ligands functioning in a cell-autonomous or non-cell autonomous manner. However, the activation of Hh pathway in tumor cells via a cell-autonomous manner is challenged by many controversial observations and remains to be fully elucidated [3], for example, the inability of mutationally activated Smo expressed in pancreatic epithelial to initiate pancreatic cancer [46]. In this study, we used the well established acquired chemoresistant cancer cell lines as an experimental model system for investigating the contribution of heterotrimeric G proteins and their downstream effectors to Gli activation mediated by Smo. Our data clearly demonstrate that acquired chemoresistant cancer cells harbor aberrant Hh pathway activity in a cell-autonomous manner, thus increasing our knowledge about the mechanisms behind Hh activation in cancers. On the other hand, many studies have shown the loss of Hh pathway activity in cancer cells possessing elevated Hh pathway activity after cultured in vitro [47][48][49], arguing against the use of in vitro cultured cancer cell lines for many kinds of investigations related to Hh pathway in cancer biology, ranging from dissecting molecular mechanisms underlying Hh signaling transduction to preclinical evaluation of Hh inhibitors. In this regard, our data that acquired chemoresistant cancer cells harbor aberrant Hh pathway activity in a cell-autonomous manner identify acquired chemoresistant cancer cell lines as potential and useful in vitro experimental model systems for investigations related to Hh pathway in cancer biology.
How does the Smo-coupled Gαi signaling link the transcriptional factor Gli in chemoresistant cancer cells? In the case of classical GPCR signaling transduction, the exchange of GDP for GTP at Gαi subunit results in the activation of Gαi, thereby repressing the adenyl cyclase and subsequently decreasing the conversion of ATP to cAMP. Reduced cAMP level implies downregulation of the activity of PKA [15]. Considering that PKA is the key determinant for proteasome proteolysis of Gli by phophorylating it at multiple sites [50,51], we can envision that Gαi after activated by Smo signaling may protect Gli from proteasome degradation by inhibiting the activity of PKA in chemoresistant cancer cells in despite of required further verifications. On the other hand, in the case of classic GPCR signaling transduction, the Gβγ dimmer after releasing from Gαi may stimulate a couple of downstream effectors, such as PKC, PI3K and JNK [35]. Data from other labs indicate that dissociated Gβγ dimmer initiated by Smo signaling may potentially promote the activation of Gli via PKC and PI3K in chemoresistant cancer cells [52][53][54][55]. However, in the present study, we provide complementary evidences showing that Smo may as well promote the activation of Gli via Gβγ-JNK signaling axis. Hence, our data together with that from other labs suggest that Smo utilizes the G protein signaling to its full potential for activating the transcriptional factor Gli. JNK, a key member of the family of MAPKs, is also called stress activated protein kinase (SAPK) and can be activated by environmental and genototoxic stress and other extracellular stimulus [56]. JNK activation has also been linked to acquired chemoresistance by promotion of chemoresistance or by reversal of chemoresistance, relying on the duration and strength of the signaling [56,57]. Here, we show that JNK may function as a downstream effector of Gβγ for transmitting the signaling from Smo to Gli, thereby promoting the Glidependent acquired chemoresitance. Thus, this finding will help us with better understanding the role of JNK in acquired chemoresistance. Similar to ERK1/2, another critical member of MAPKs, JNK signaling is as well deregulated in many types of cancers [56,58]. However, the contribution of JNK in cancer development is complex and far from being fully elucidated, in other words, exhibiting context-specific and cell type-specific manner. JNK has been well known to confer the positive impact on proliferation and survival of cancer cells via its target AP1, a transcriptional factor composing Jun and Fos [59]. Of interest, data in our study imply that Gli represents a putative downstream target of JNK, thus facilitating our better interpretation of the molecular mechanisms responsible for promoting the development of cancers by JNK. Although inhibitors of membrane protein Smo have been approved for treatment of basal cell carcinoma, the early acquired resistance to such inhibitors proposes the need for additional downstream targets [60]. Hence, our data imply JNK as a new target for the treatment of the tumors with acquired resistance to Smo inhibitors. In this regard, how JNK promotes the activation of Gli is quite interesting, and is currently being investigated in our lab.

Conclusions
In this study, we demonstrate that GPCR-like signaling mediated by Smo contributes to the acquired chemoresistance through activating the canonical Hh transcriptional factor Gli. Our data improve our knowledge of the nature of the signal transduction of Smo and the molecular mechanisms responsible for the acquired chemoresistance maintained by Hh pathway. Moreover, the findings that JNK after activated by Smo-Gβγ signaling axis may stimulate the Gli activity and consequently promote acquired chemoresistance expose a promising and potential target for developing anti-cancer drugs aimed at Hh pathway and for combating the acquired resistance raised by using of anti-cancer drugs targeting Smo.