The role of PAK1 in the sensitivity of kidney epithelial cells to ischemia-like conditions

ABSTRACT

Kidney ischemia, characterized by insufficient supply of oxygen and nutrients to renal epithelial cells, is the main cause of acute kidney injury and an important contributor to mortality world-wide. Earlier research implicated a G-protein coupled receptor (NK1R) in the death of kidney epithelial cells in ischemia-like conditions. P21-associated kinase 1 (PAK1) is involved in signalling by several G-proteins. We explored the consequences of PAK1 inhibition for cell survival under the conditions of reduced glucose and oxygen. Inhibition of PAK1 by RNA interference, expression of a dominant-negative mutant or treatment with small molecule inhibitors greatly reduced the death of cultured kidney epithelial cells. Similar protection was achieved by treating the cells with inhibitors of MEK1, in agreement with the prior reports on PAK1-MEK1 connection. Concomitant inhibition of NK1R and PAK1 offered no better protection than inhibition of NK1R alone, consistent with the two proteins being members of the same pathway. Furthermore, NK1R, PAK and MEK inhibitors reduced the induction of TRAIL in ischemia-like conditions. Considering the emerging role of TRAIL in ischemia-mediated cell death, this phenomenon may contribute to the protective effects of these small molecules. Our findings support further exploration of PAK and MEK inhibitors as possible agents to avert ischemic kidney injury.

KEYWORDS:

• Acute kidney injury
• MAP kinase cascade
• tachykinin receptor 1

Introduction

Acute renal ischemia (ARI) is a major cause of acute kidney injury (AKI), which is characterized by a sharp decline in kidney filtration and associated with increasing incidence, high levels of morbidity and mortality, inadequate treatment options, and a huge financial burden worldwide [1,2]. Indeed, ischemic AKI is one of the main contributors to human mortality and morbidity, claiming 300,000 lives every year in the US alone [3]. One in five hospital patients worldwide experiences AKI, and the rates are higher in lower-income countries, where 13.3 million cases of AKI occur annually [4]. Renal ischemia results from impaired delivery of oxygen and nutrients to kidney cells. Diabetes, old age, heart failure and open-heart surgery are long-known risk factors for AKI [5]. Other prominent risk factors include liver failure, hypertension, diabetes mellitus, dehydration, sepsis and others [6] .

Initial death of kidney epithelial cells is followed by inflammation at time of reperfusion, which further exaggerates the condition and contributes to the ensuing fibrosis [7]. Patients who have survived an episode of AKI frequently develop chronic kidney disease, which, in turn, is a high risk factor for new AKI episodes [3]. This creates a life-threatening cycle of organ destruction, which often necessitates organ transplant. ARI is imminent during kidney transplantation, and can significantly delay graft function in transplant patients [8].

Despite extensive research efforts, the clinical practice still awaits a pharmacological intervention, which would reverse or avert the AKI-induced damage in at-risk population. Current options are merely supportive, while available pharmacological interventions give, at best, ambiguous results [1,9]. For example, pharmacological blockage of Renin‐Angiotensin‐Aldosterone system was considered a major path towards prevention of AKI, but, paradoxically, appears to increase the risk instead [9].

It has been argued that cell death in ischemia-like conditions involves maladaptive biochemical responses and active induction of cell death in kidney epithelium [7,10,11]. Indeed, previous work has shown that apoptosis is a major form of the cell death responsible for the kidney failure associated with AKI [12]. Thus, one may hope that correctly aimed interference with either biochemical or apoptotic circuitry may avert the initial death of affected epithelium and, hence, avoid the ensuing organ damage.

Importantly, while there is a tremendous progress towards understanding normal and pathological responses to hypoxia per se, such as VHL-dependent [13] and – independent [14,15] activation of Hypoxia Inducible Factor 1 (HIF1) and the ensuing switch toward glycolysis [13]. However, the protective role of such mechanisms is contingent upon abundance of glucose, which is problematic under ischemia. On the other hand, HIF1 is also known to activate pro-apoptotic pathways [16,17]. Similarly, the stress of glucose starvation in normoxic conditions can be attenuated by enhanced oxidation of fatty acids. However, this depends on abundance of oxygen, which is a problem for an ischemic tissue. Indeed, partial products of fatty acid oxidation are toxic and contribute to kidney injury [18]. Thus, it is likely that the challenges presented by concomitant lack of glucose and oxygen are principally different from what pertains to individual shortage of either of these resources, and the relevant resistance mechanisms need to be investigated accordingly.

Previously, we used an unbiased genetic screen to identify the genes whose inhibition can protect human renal epithelial cells under ischemia-like conditions [10]. Among the identified genes was TACR1, which codes for a G protein-coupled receptor, neurokinin 1 receptor (NK1R). NK1R serves as a receptor for multiple tachykinins, which are encoded by several distinct genes [19]. While interference with one of those genes, TAC1, has yielded protection in ischemia-like conditions [10], the gene encodes multiple bioactive peptides [19], and the precise nature of the relevant secreted molecule remains to be established. Importantly, genetic or pharmacologic inhibition of NK1R synthesis or function has shown protective effects in several pre-clinical models [10], but the corresponding signaling pathway, which might contain additional druggable targets, was not investigated in the original study.

P21-activated kinase 1 (PAK1) is a well-known mediator of signaling by G protein-coupled receptors [20,21]. Under some circumstances, a G protein-PAK1 interaction can participate in the transduction of apoptotic signals [22–24]. These considerations prompted us to investigate whether interference with PAK1 signaling affects the survival of human renal epithelium cells in conditions of low glucose and oxygen supply.

Results

Interference with PAK1, MEK1 signaling confers ischemia protection

To determine whether PAK1 signaling plays a role in ischemia-induced renal damage we engineered non-transformed immortal renal proximal tubule cell line HKC-8 [25] to express either an shRNA targeting PAK1 or a dominant negative mutant form of PAK1 (PAK1-K299R). We subjected those cells to conditions of low glucose and oxygen and compared cell survival to that of the control cells. Figure 1(a) shows that, when compared to HKC-8 cells expressing control vectors, HKC-8 cells expressing either an shRNA against PAK1 or the dominant negative PAK1 mutant displayed about 7-fold and 10-fold increases in survival, respectively. We employed quantitative PCR to verify the efficacy of the shRNA targeting PAK1, and our results revealed that PAK1 expression was knocked down by over 70% (Figure 1(b)).

Figure 1. Interference with PAK1 protects kidney epithelial cells in ischemia-like conditions. (a) HKC-8 cells expressing a PAK1 shRNA or a dominant negative PAK1 were subjected to glucose and oxygen deprivation for 48 hours. Following treatment, cell numbers for each culture were assessed relative to respective normoxic controls and values are presented relative to those from control cells expressing non-targeting shRNA. The means and standard deviations of three independent experiments are shown. (b) PAK1 mRNA levels in HKC-8 cells transduced with PAK1-specific shRNA were measured by qRT-PCR, normalized to mean GAPDH transcript levels, and reported relative to that in cells transduced with non-targeting shRNA. The means and standard deviations of three independent experiments are shown.

The biological effects of shRNAs that knock down a particular gene are often predictive of the biological effects of small molecule inhibitors targeting that gene’s protein product. Therefore, we tested whether PAK1 inhibition by IPA3 [26] or PF-3758309 [27] could protect HKC-8 cells from ischemic stress. Furthermore, PAK1 is a well-known modulator of MAP kinase signaling cascade, and MEK1, along with RAF proteins, is a reported direct target of PAK1 phosphorylation [20]. Interestingly, in the context of ischemic injury to neurons, MEK1 is a well-established mediator of cell death [28]. Therefore, we also examined whether inhibition of MEK1 by AZD6244 (aka Selumetinib) [29] or U0126 [30] could protect renal epithelial cells under ischemia-like conditions. The results from these experiments demonstrate that chemical inhibition of either PAK1 or MEK1 potently protects HKC-8 cells from ischemia-induced death (Figure 2).

Figure 2. Chemical inhibition of PAK1 or MEK1 protects kidney epithelial cells in ischemia-like conditions. HKC-8 cells were treated with a PAK1 inhibitor (10µM IPA-3, or15nM PF-3758309) or a MEK1 inhibitor (60nM AZD or 10µm U0126), and subjected to glucose and oxygen deprivation for 48 hours. Cell numbers were then assessed relative to respective normoxic controls and values are presented relative to those from control cells not treated with PAK1 or MEK1 inhibitor. The means and standard deviations of three independent experiments are shown.

PAK1 and NK1R appear to function in the same pathway of ischemia response

We have previously described that in our in vitro models of ischemia the protective effect of Aprepitant reaches its maximum at ~5–10 microM of the drug [10]. Further dose escalation fails to increase the fraction of surviving cells, but an additional boost in survival could be achieved by combined inhibition of NK1R and certain other proteins [10]. This is consistent with the hypothesis that multiple pathways contribute to the loss of cells under ischemia-like conditions, and even a complete inhibition of each individual pathway yields only a partial protection. We decided to investigate the effects of combined NK1R and PAK1 inhibition. As expected, each of the inhibitors by itself achieved a significant, but incomplete protection. Interestingly, the combination of the two failed to surpass the maximal effect of Aprepitant alone (Figure 3), in contrast with RHOB and C2ORF42 shRNAs [10] and a fatty acid oxidation inhibitor Trimetazidine (Supplementary Figure 1). This observation is consistent with the notion that PAK1 and NK1R are parts of the same molecular pathway of ischemia response.

Figure 3. A PAK inhibitor fails to increase the protective effect of Aprepitant. HKC-8 cells were treated with TACR1 inhibitor (8µM Aprepitant), PAK1 inhibitor (15nM PF-3758309), or a combination of the two treatments and subjected to glucose and oxygen deprivation for 48 hours. Cell numbers were then assessed relative to respective normoxic controls and values are presented relative to those from control cells not chemically treated The means and standard deviations of three independent experiments are shown.

NK1R, PAK1 and MEK regulate expression of TRAIL mRNA

Previous reports implicated pro-apoptotic cytokine TNF-Related Apoptosis Inducing Ligand (TRAIL) in the death of epithelial cells under ischemia-like conditions [31]. TRAIL, also known as Apo-2 ligand (APO2L) or TNF Superfamily Member 10 (TNFSF10), binds and activates Death Receptors 4 and 5 (DR4 and DR5, aka TNFRSF10A and TNFRSF10B), which, in turn, initiates the so-called extrinsic pathway of apoptosis. TRAIL-DR5 interaction, in particular, has been linked to acute kidney injury [32]. TRAIL-mediated killing may occur in an autocrine or paracrine manner: the cell producing TRAIL under stress conditions sets off apoptosis within itself or its neighbors, provided that the appropriate receptor is present. In our system, TRAIL has a clear role in cell death: depletion of TRAIL potently protects the cells in ischemia-like conditions (Figure 4).

Figure 4. Depletion of TRAIL (TNFSF10) mRNA protects kidney epithelial cells in ischemia-like conditions. (a) HKC-8 cells expressing shRNA targeting TNFSF10 were subjected to glucose and oxygen deprivation for 48 hours. Following treatment, cell numbers for each culture were assessed relative to respective normoxic controls and values are presented relative to those from control cells expressing non-targeting shRNA. The means and standard deviations of three independent experiments are shown. (b) TNFSF10 mRNA levels in HKC-8 cells transduced with TNFSF10-specific shRNA were measured by qRT-PCR, normalized to mean GAPDH transcript levels, and reported relative to that in cells transduced with non-targeting shRNA. The means and standard deviations of three independent experiments are shown.

Figure 5. Inhibition of NK1R (TACR1), PAK1, or MEK1 prevents ischemia-induced TRAIL induction. (a) HKC-8 cells were transduced to express shRNA targeting TACR1 mRNA or treated with NK1R inhibitor (8µM L-733060), and cultured under normoxia or subjected to glucose and oxygen deprivation for 48 hours. TNFSF10 mRNA levels were then measured by qRT-PCR, normalized to mean GAPDH transcript levels, and reported relative to those from untreated parental cells. The means and standard deviations of three independent experiments are shown. (b) HKC-8 cells were treated with PAK1 inhibitor (15nM PF-3758309) or MEK1 inhibitor (60nM AZD6244), and cultured under normoxia or subjected to glucose and oxygen deprivation for 48 hours. TNFSF10 mRNA levels were then measured by qRT-PCR, normalized to mean GAPDH transcript levels, and reported relative to those from untreated cells. The means and standard deviations of three independent experiments are shown.

We set forth to investigate whether the NK1R pathway affects production of TRAIL. As shown in Figure 5, TNFSF10 transcript was upregulated in ischemia-like conditions. However, suppression of NK1R using either an shRNA or a chemical inhibitor L-733060 reduced TRAIL mRNA levels (Figure 5(a)). Furthermore, the level of TRAIL mRNA was also attenuated by inhibitors of PAK1 and MEK (Figure 5(b)). Interestingly, L-733060 also diminished the stress-induced induction of TNFRSF10B mRNA (Supplementary Figure 2), suggesting that the levels of both TRAIL and its receptor are under the control of the NK1R -dependent pathway in these cells.

Discussion

Protection of at-risk population from acute kidney injury has attracted considerable attention, but the practical outcome to date has been rather limited. A possibility to pre-condition kidney to ischemia by transient ischemic treatment of remote organs is considered as a non-pharmacological measure [33]. However, clinical results for this technique vary considerably between studies [34–36], perhaps reflecting the need for further optimization of the procedure, as well as the differences in the study populations [37]. It has been argued that the method is less effective in the elderly [38], who represent the major group of risk for AKI [5,39]. In this context, a pharmacological nephroprotective approach remains a desirable goal, which, unfortunately, is yet to be fully realized. Every year, multiple new agents are reported to elicit some nephroprotection in pre-clinical models (e.g. [40–47]), but the proof of their efficacy in human trials is still lacking, and, hence, the unmet need remains.

It has been long recognized that the hopes of developing an effective intervention to prevent AKI rest, for the large part, on gaining a more complete understanding of the mechanisms of cellular injury [48]. In our prior work, we identified a number of genes, whose products play a role in this process [10]. While NK1R presents an attractive druggable point of intervention, the molecular mechanisms connecting this G-protein coupled receptor to other mediators of ischemia response remained elusive.

NK1R belongs to the family of GPCRs that signal through heterotrimeric G-proteins to a variety of intracellular targets, including ERKs [49]. These signaling cascades involve small GTPases [49], whose connection to ERKs is known to be mediated by PAK1 [20]. In addition, at least some heterotrimeric G-proteins are direct activators of PAK1 [50]. However, the specific G-proteins, which work downstream of NK1R in this pathway, are yet to be established. In the meantime, it remains formally possible that NK1R and PAK1 are separated by multiple biochemical steps, and more complex scenarios of NK1R-PAK1 relationship cannot be ruled out. Indeed, it has been reported that, at least in some systems, NK1R activation can lead to the release of EGFR ligands [51] (e.g. heparin-binding epidermal growth factor [52]), and EGFR signaling eventually activates of ERKs [51]. Importantly, PAK1 is a long-known mediator of EGFR signaling [53].

Overall, PAK1 appears as a likely mediator of TACR1 activity. Indeed, a p21-activated kinase has been previously implicated in a morphological change (“membrane blebbing”) that substance P elicits in an astrocytoma cell line [54], although the identity of the particular PAK isoform has not been elucidated in that study. Pharmacological targeting of PAK1 is an area of active research [27], which is stimulated, in part, by numerous findings that PAK1 regulates survival and drug resistance in cancer (e.g. in melanoma [55,56]). While the role of PAK1 as a promoter of cell growth and survival has received a lot of attention, there are evidences that in certain contexts PAK1 participates in pro-apoptotic signaling [22–24]. Our data presented herein indicate that PAK1 may play a role in the sensitivity of kidney epithelial cells to ischemia-like conditions and is likely to belong to the same signaling pathway as NK1R. Importantly, this role of PAK1 in our experiments has been validated using two different genetic inhibitors (an shRNA and a dominant-negative mutant) and two different chemical inhibitors. While the specificity of each of the individual tools may be questioned [20,57], their concordant behavior strongly argues in favor of an on-target effect.

There are evidences that ERK activity contributes to kidney injury, at least, in some circumstances. For example, ERK inhibition elicits nephroprotection in the context of endotoxin-induced injury [58] and protects kidney epithelial cells upon hypothermia [59] and oxidant exposure [60]. CDK5, whose inhibition was reported to protect porcine kidney cells under hypoxia, is also a regulator of the MAP kinase pathway [61]. The role of PAK1 in regulating the MAP kinase cascade in various cells and in response to various stimuli is well established [20,62]. We may speculate that the NK1R-PAK1-MEK-ERK axis might also play a role in some other NK1R-induced responses, which are reportedly mediated by MAPKs, and ERKs in particular [63–65]. Interestingly, while those reports come from the studies of skin [63] and neurons [64], they too describe a potentially pro-death activity of substance P and its receptor. Importantly, NK1R, PAK1 and the MAP kinase cascade are “druggable”, with multiple inhibitors either in clinical practice or at various stages of pre-clinical development, albeit not explicitly in the context of nephroprotection. Our findings support further investigation of such substances as protectors of kidney and, possibly, other organs. Considering that aging population is at a higher risk of AKI, it is noteworthy that MEK inhibition is being suggested as a general “geroprotective” strategy [66]. Encouragingly, addition of a MEK inhibitor to a BRAF inhibitor not only increased the anti-cancer activity of the latter, but also dramatically decreased the incidence of kidney injury among the treated population [67].

Our observations indicate that other pharmacological agents may cooperate with NK1R inhibitors in protecting kidney cells in ischemia-like conditions (Supplementary Figure 1). TMZ is a promising candidate nephroprotectant [68,69], and it may be warranted to explore a TMZ-Aprepitant combination as a way to enhance the protective effects of these drugs. Importantly, the lack of cooperation between maximally-effective doses of NK1R and PAK1 inhibitors does not necessarily rule out their concomitant use. For example, targeting the same pathway with multiple inhibitors may be warranted if the doses required for either molecule to fully suppress its respective target in a clinical setting are associated with unacceptable toxicities. It is also likely that effective prevention of AKI in high-risk patients would involve long-term treatment, and it would be desirable to alternate the drugs to minimize side-effects of prolonged exposure to ether one of them.

Cytokine TRAIL is a candidate mediator of cell death in ischemic epithelium. In particular, prevention of TRAIL binding to its receptor DR5 has been shown to reduce the extent of kidney injury, at least, in some models [31,32]. Our data is consistent with this role of TRAIL (Figure 4) and indicate that expression of both DR5 and TRAIL are sensitive to NK1R signaling. This suggests a plausible mechanism of NK1R involvement in ischemic cell death, although it does not exclude a contribution of other NK1R-dependent processes. The exact molecular mechanisms that link NK1R to TRAIL and DR5 expression in kidney cells are yet to be deciphered.

Finally, it is important to note that the danger of ischemia to the function of internal organs is not limited to kidneys. Furthermore, the shortage of oxygen and nutrients represents one of the key hurdles in the growth of most tumor types. Additional research would be warranted to elucidate whether the findings described herein pertain to other histological and pathophysiological contexts.

Materials and methods

HKC-8 cells were cultured and treated as previously described [10]. Cells were free of mycoplasma contamination, as tested using MycoAlert Mycoplasma Detection Kit (Lonza, Walkersville, MD, USA). No contaminating replication-competent retroviruses [70] were detected in the cells. The RNA interference reagents were obtained from the Genomics Shared Resource of the Roswell Park comprehensive Cancer Center. The viral constructs were used as reported elsewhere [55]. Trimetazidine, Aprepitant, U0126, IPA3 and AZD6244 were obtained from Selleck Chemicals (Houston, TX, USA). L-733060 was from Tocris (Bristol, UK), PF3758309 was purchased from ChemieTek (Indianapolis, IN, USA). Cell numbers were compared using the methylene blue staining and extraction method as described [55].

The transcripts were detected using quantitative real-time RT-PCR with GAPDH as an endogenous control. The nucleotide sequences were: PAK1 5′-GTGAAGGCTGTGTCTGAGACTC-3′ and 5′-GGAAGTGGTTCAATCACAGACCG-3′ TNFSF10 5′-TGGCAACTCCGTCAGCTCGTTA-3′ and 5′-AGCTGCTACTCTCTGAGGACCT-3′; TNFRSF10B 5′-AGCACTCACTGGAATGACCTCC-3′ and 5′-GTGCCTTCTTCGCACTGACACA-3′; GAPDH 5′-GTCTCCTCTGACTTCAACGCG-3′ and 5′-ACCACCCTGTTGCTGTAGCAA. RNA was isolated using the RNeasy kit (Qiagen, Germantown, MD, USA). Complementary DNA was synthesized using SuperScript III (Life Technologies). PCR was performed using an ABI Prism 7900 Sequence Detection System (Waltham, MA, USA) and IQ SYBR Green SuperMix (Bio-Rad, Hercules, CA, USA). The thermal cycling conditions comprised 2 min at 50 °C, 10 min at 90 °C, and 1 min at 60 °C for 40 cycles. Data were analyzed using RQ Manager 1.2.1 (ABI). The cycle threshold data was converted to change-fold in expression by the “delta delta Ct” method [71].

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary Material

Supplemental data for this article can be accessed here.

Additional information

Funding

This study was supported in part by the grants from the Roswell Park Alliance Foundation, Roswell Park Cancer Institute and the National Cancer Institute [CA205178] to ESK. The participation of MM was made possible by a grant from the National Cancer Institute [CA181003]. The use of the shared resources at the Roswell Park Comprehensive Cancer Center is supported through a grant from the National Cancer Institute [CA016056].