Smac-Mimetic–Induced Epithelial Cell Death Reduces the Growth of Renal Cysts

Abstract

Past efforts to pharmacologically disrupt the development and growth of renal cystic lesions focused primarily on normalizing the activity of a specific signaling molecule, but the effects of stimulating apoptosis in the proliferating epithelial cells have not been well studied. Although benign, ADPKD renal cysts created by the sustained proliferation of epithelial cells resemble tumors, and malignant cell death can be achieved by cotreatment with TNF-α and a mimetic of second mitochondria-derived activator of caspase (Smac). Notably, TNF-α accumulates to high levels in ADPKD cyst fluid. Here, we report that an Smac-mimetic selectively induces TNF-α–dependent cystic renal epithelial cell death, leading to the removal of cystic epithelial cells from renal tissues and delaying cyst formation. In vitro, a Smac-mimetic (GT13072) induced the degradation of cIAP1 that is required but not sufficient for cell death. Cotreatment with TNF-α augmented the formation and activation of the RIPK1-dependent death complex and the degradation and cleavage of FLIP, an inhibitor of caspase-8, in renal cystic epithelial cells. This approach produced death specifically in Pkd1 mutant epithelial cells, with no effect on normal renal epithelial cells. Moreover, treatment with the Smac-mimetic slowed cyst and kidney enlargement and preserved renal function in two genetic strains of mice with Pkd1 mutations. Thus, our mechanistic data characterize an apoptotic pathway, activated by the selective synergy of an Smac-mimetic and TNF-α in renal cyst fluid, that attenuates cyst development, providing an innovative translational platform for the rational development of novel therapeutics for ADPKD.

Autosomal dominant polycystic kidney disease (ADPKD) is caused by mutations in one of two genes: polycystic kidney disease 1 (PKD1) and PKD2.¹ The gene product of PKD1, polycystin-1 (PC1), either alone or in complex with the gene product of PKD2 polycystin-2 (PC2), regulates a wide variety of cellular functions, including proliferation, apoptosis, fluid secretion, adhesion, and morphogenesis,² features common in all hereditary renal cystic diseases.³ Epithelial cells lining renal cysts resemble benign neoplasms, in which cell proliferation forces sustained cyst expansion throughout the lifespan of patients.^4,5 In the past, efforts have focused on targeting specific pathways to normalize a cystic epithelial cell function, thus preventing cyst formation.⁶ Recent studies showing apoptosis of malignant cells treated with a second mitochondria-derived activator of caspase (Smac) -mimetic plus TNF-α^7,8 suggested that amplifying a pathway that induces cell death exclusively in cystic epithelia, while sparing wild-type cells, might possibly reduce cyst growth and secondary destruction of parenchyma.

TNF-α is a constant feature of cyst fluids sampled from the kidneys of ADPKD patients.⁹ TNF-α binds to receptor I (TNFR1) to initiate the formation of a multimeric signaling complex that regulates cell survival and cell death. The TNF-α/TNFR1 complex also includes the TNF-α receptor-associated protein with death domain (TRADD), TNF-α receptor-associated protein 2, receptor-associated protein kinase 1 (RIPK1), and cellular inhibitor of apoptosis protein 1 (cIAP1) and cIAP2. This large complex then recruits the IκB kinase composite, leading to the activation of NF-κB.^10–12 NF-κB activation prevents cell death by leading to dependent gene transcription, including additional cytokines and antiapoptotic proteins, such as cellular FLICE (FADD-like IL-1β-converting enzyme)-inhibitory protein (FLIP) (a protease-dead caspase-8 homolog that competes for caspase-8 binding to Fas-associated protein with death domain [FADD]).^13–16 For this reason, the TNFR1-associated complex is referred to as the prosurvival complex I.^17–19 A prodeath complex (complex II) is also formed after internalization of the TNFR1 receptor and consists of RIPK1, FADD, and caspase-8.²⁰ The activity of complex II can be inhibited by endogenous FLIP,²¹ which competes for caspase-8 binding to FADD.

TNF-α together with Smac-mimetic induces cancer cell death.^22,23 Smac-mimetics are cell-permeable synthetic compounds designed to mimic the N-terminal 4 amino acids of Smac, a mitochondrial protein that binds to and antagonizes inhibitors of apoptosis proteins (IAPs), including cIAP1, cIAP2, and X-linked inhibitor of apoptosis protein.^22,23 Several IAP antagonists have been developed that mimic the interactions of the Smac amino-terminal peptide with IAP proteins. These antagonists possess proapoptotic activity both in vitro and in vivo.^24–28 Smac-mimetics induce the degradation of cIAP1 and/or cIAP2 in complex I, resulting in the formation of the prodeath complex II, which promotes cancer cell death. Moreover and key to the current study, this process has an absolute requirement for TNF-α.^8,29 The extent and mechanisms to which an Smac-mimetic can induce TNF-α–dependent cell death of nontransformed cystic renal epithelial cells and slow cyst expansion are addressed in this study.

Results

TNF-α Induces Its Own Transcription in Pkd1 Mutant Cystic Renal Epithelial Cells

TNF-α is constantly present at measurable levels in ADPKD cyst fluids,⁹ although the mechanisms underlying TNF-α accumulation are unknown. The expression of TNF-α is regulated through its receptor-mediated activation of NF-κB.²⁹ Quantitative RT-PCR showed that TNF-α mRNA was increased in Pkd1 null mouse embryonic kidney (MEK) cells (Figure 1A) and postnatal Pkd1 homozygous PN24 cells (Figure 1B) as well as the kidneys from Pkd1^flox/flox:Ksp-Cre and Pkd1^nl/nl mice (Supplemental Figure 1) compared with Pkd1 wild-type MEK cells, Pkd1 heterozygous PH2 cells, and Pkd1 wild-type kidneys, respectively. TNF-α mRNA was further increased in response to external TNF-α stimulation in Pkd1 null MEK cells and PN24 cells (Figure 1, A and B). This response is mediated through canonical NF-κB signaling, because adding an NF-κB inhibitor, SN50, prevented the increase in TNF-α mRNA in Pkd1 mutant renal epithelial cells treated with TNF-α (Figure 1A). TNF-α induces its own transcription in Pkd1 mutant renal epithelial cells, suggesting that TNF-α in cyst fluid may induce its own transcription by cyst-lining epithelial cells, thereby magnifying its levels in cyst fluid.

Figure 1.

TNF-α exerts a prosurvival effect on the cystic epithelium through NF-κB activation. (A and B) TNF-α induced its own transcription through NF-κB in (A) embryonic and (B) postnatal renal epithelial cells null for Pkd1 as assayed by quantitative RT-PCR. Each sample was run in triplicate in every experiment, and each experiment was repeated three times. The expression level of TNF-α was normalized to the expression level of actin (P<0.05). (C) The components of complexes I and II were upregulated in cystic renal epithelial cells. The expression of cIAP1, RIPK1, and TRADD in complex I and the expression of caspase 8 and FADD in prodeath complex II were increased in Pkd1 (null) MEK cells compared with wild-type MEK cells. The expression of the upregulated proteins in Pkd1 null MEK cells was quantified from three independent immunoblots and presented as the relative protein expression level standardized to actin (P<0.05). (D) Western blot analysis of the expression of I-κB from whole-cell lysates of Pkd1 wild-type and Pkd1 null MEK cells treated with TNF-α. (E) Western blot analysis of the expression of cIAP1 and FLIP from whole-cell lysates of Pkd1 wild-type and Pkd1 null MEK cells treated with TNF-α. h, hour.

TNF-α Exerts a Prosurvival Effect on Pkd1 Mutant Cystic Epithelium through NF-κB Activation

TNF-α binds with its receptor I, TNFR1, to initiate the formation of a prosurvival complex. This complex activates NF-κB and NF-κB–mediated expression of FLIP, an inhibitor of caspase 8, to regulate cell survive and death.²⁰ TNFRI expression is increased in Pkd1 null renal epithelial cells and tissues.⁹ In the current study, cIAP1, RIPK1, and TRADD, components of complex I, as well as caspase-8 and FADD, components of prodeath complex II, were upregulated in Pkd1 null MEK cells compared with Pkd1 wild-type MEK cells, even in the absence of TNF-α (Figure 1C). In addition, the long form of FLIP was also elevated in Pkd1 null MEK cells (Figure 1C). The expressions of the components of complexes I and II as well as FLIP were also increased in postnatal Pkd1 homozygous renal epithelial (PN24) cells and tissues compared with postnatal Pkd1 heterozygous renal epithelial (PH2) cells and Pkd1 wild-type kidneys (Supplemental Figure 2). The increased expression of the components in complexes I and II in cystic renal epithelial cells and tissues may facilitate the formation of prosurvival complex I.

The formation of prosurvival complex I prevents cell death through activation of NF-κB.²⁰ Activation of NF-κB by TNF-α is characterized by the degradation of I-κBα, an inhibitor of NF-κB.⁷ We found that TNF-α treatment induced a rapid but transient degradation of I-κBα in Pkd1 null MEK cells (Figure 1D), whereas TNF-α induced minimal degradation of I-κBα in Pkd1 wild-type MEK cells (Figure 1D). Activation of NF-κB elevated the expression of FLIP,¹⁵ which inhibits caspase-8 and prevents caspase-8/RIPK1/FADD complex II-induced apoptosis.²⁰ Here, we found that TNF-α treatment increased the expression of FLIP and unexpectedly, increased the expression of cIAP1 in Pkd1 null MEK cells (Figure 1E). However, all of the other components of complexes I and II failed to respond to TNF-α treatment in these cells (Supplemental Figure 3). The significant I-κBα degradation and the increased expression of cIAP1 and FLIP in Pkd1 null MEK cells treated with TNF-α compared with the minimal degradation of I-κBα and the expression of cIAP1 and FLIP in Pkd1 wild-type MEK cells treated with TNF-α suggest that activation of NF-κB by TNF-α might be a key step in protecting Pkd1 null MEK cells from apoptosis. A similar pathway is active in cancer cells.¹⁶ In support of this notion, we found that TNF-α did not induce apoptosis in either Pkd1 wild-type or null MEK cells, even at high concentration (200 ng/ml) (Supplemental Figure 4A); however, treatment with TNF-α plus an NF-κB inhibitor, SN50, significantly decreased the viability of Pkd1 null MEK cells (Supplemental Figure 4B). These results suggest that cyst fluid TNF-α may protect cystic renal epithelial cells from apoptosis through NF-κB activation.

Smac-Mimetic Induces TNF-α–Dependent Death in Pkd1 Null But Not Wild-Type Renal Epithelial Cells

TNF-α administered with an Smac-mimetic induces cancer cell death.^22,23 Here, we found that TNF-α combined with the Smac-mimetic GT13072, an analog of TetraLogic’s lead Smac-mimetic drug birinapant (formerly TL32711), which is in phase 2 clinical trials and being developed for both solid tumors and hematologic malignancies as a single agent,³⁰ significantly induced cell death in Pkd1 null MEK cells (Figure 2A) (P<0.001) and postnatal Pkd1 homozygous PN24 cells (Figure 2B) (P<0.001). Importantly, Smac-mimetic plus TNF-α induced cell death in epithelial cells obtained from primary cultures of mural epithelial cells from human ADPKD cysts but not cells from normal human kidney (Figure 2C) (P<0.001). Furthermore, Pkd1 null MEK cells treated with Smac-mimetic alone or TNF-α alone remained viable (methylene blue staining³¹) after 5 days, whereas Pkd1 null MEK cells treated with the Smac-mimetic plus TNF-α died (Figure 2D). Consistent with the proposed antagonism of caspase inhibition by Smac-mimetics, Pkd1 null MEK cells treated with the Smac-mimetic and TNF-α plus z-Vad, a pan caspase inhibitor, remained viable (Figure 2D).

Figure 2.

Smac-mimetic induced TNF-α–dependent cystic epithelial cell death. (A) Apoptosis was significantly increased in Pkd1 null MEK cells treated with Smac-mimetic (SM) plus TNF-α compared with Pkd1 wild-type (WT) MEK cells. Pkd1 WT and null MEK cells were treated with 1 µM Smac-mimetic plus TNF-α (50 ng/ml) for 24 hours. The apoptotic cells were analyzed by FACS (P<0.001). (B) Smac-mimetic plus TNF-α induced apoptosis in Pkd1 null PN24 cells but not Pkd1 heterozygous postnatal PH2 cells (P<0.001). (C) Smac-mimetic plus TNF-α induced apoptosis in primary human ADPKD cells but not primary normal human kidney (NHK) cells (P<0.001). (D) The effect of TNF-α (50 ng/ml) and/or Smac-mimetic (1 or 5 μM) on MEK cell viability. We found that Pkd1 null MEK cells treated with Smac-mimetic alone or TNF-α alone remained viable after 5 days, whereas cells treated with Smac-mimetic plus TNF-α died. By contrast, similar treatment had no effect on WT MEK cells. It was interesting to find that Pkd1 null MEK cells treated with Smac-mimetic and TNF-α plus z-Vad, a pan caspase inhibitor, also survived, suggesting that z-Vad could rescue Pkd1 null MEK cells from death after treatment with Smac-mimetic plus TNF-α and the involvement of caspases in this process.

Smac-Mimetic Reduces Cyst Growth in Pkd1 Animal Models

Given the robust cystic epithelial cell death induced by Smac-mimetic in cell culture, we examined the effect of Smac-mimetic on cyst formation and growth in vivo using animal models bearing Pkd1 mutations. We hypothesized that the presence of TNF-α in cyst fluid⁹ plus the addition of an Smac-mimetic would result in a reduction of cyst growth in situ. First, we determined if Smac-mimetic reduced cyst growth in a Pkd1-floxed mouse model combined with a kidney-specific Ksp-cadherin driving Cre expression, which is activated mid-gestation in Pkd1^flox/flox:Ksp-Cre progeny. Cyst progression is aggressive in the kidneys of Pkd1^flox/flox:Ksp-Cre mice.³² Smac-mimetic (7.5 mg/kg) or DMSO (control) was injected intraperitoneally to lactating mothers at postnatal day 3 (PN3) and PN5, and kidneys were removed and analyzed at PN7. We confirmed that Smac-mimetic was effectively delivered to the pups through lactation, because a concentration that was well tolerated by the dams (30 mg/kg) was toxic to neonates. We found that administration of Smac-mimetic delayed cyst growth in PN7 kidneys (n=10) in Pkd1^flox/flox:Ksp-Cre mice compared with pups injected with DMSO (n=10; P<0.001) (Figure 3, A and B). Treatment with Smac-mimetic decreased the kidney weight/body weight ratio and blood urea nitrogen (BUN) levels in Pkd1^flox/flox:Ksp-Cre mice (P<0.05) compared with DMSO controls (Figure 3, C and D). Apoptotic cells in epithelia lining cysts were strikingly increased in Pkd1 knockout PN7 kidneys treated with Smac-mimetic compared with DMSO treatment (P<0.01) (Figure 3E and Supplemental Figure 5), whereas apoptotic cells were not detected in Pkd1 wild-type PN7 kidneys treated with Smac-mimetic or DMSO (Supplemental Figure 5). Smac-mimetic did not affect cystic renal epithelial cell proliferation and did not induce fibrosis in PN7 kidneys from Pkd1^flox/flox:Ksp-Cre mice compared with the kidneys from age-matched Pkd1^flox/flox:Ksp-Cre mice with DMSO treatment (Figure 3, F and G). These results suggest that Smac-mimetic delayed cyst growth in this mouse model by inducing the death of Pkd1 null epithelial cells lining the cysts. Smac-mimetic–treated Pkd1^flox/flox:Ksp-Cre mice had a significantly longer mean survival compared with those animals treated with vehicle (P<0.01) (Figure 3H).

Figure 3.

Smac-mimetic delayed cyst formation in Pkd1 conditional knockout postnatal mouse kidneys. (A and B) Smac-mimetic reduced cyst formation in Pkd1^flox/flox:Ksp-Cre neonate kidneys at PN7. (A) Histologic examination revealed smaller kidneys with more preserved parenchyma in Pkd1^flox/flox:Ksp-Cre PN7 Smac-treated (10 pairs) than DMSO-treated (10 pairs) animals. (Scale bar, 2 mm.) (B) The cystic index was also reduced by treatment with Smac-mimetic (P<0.001). Shown is mean ± SEM of all sections quantified for each condition. At least five evenly spaced sections from each kidney were analyzed to determine the cystic index. (C) The ratio of kidney weight/body weight (KW/BW) was decreased in PN7 Pkd1^flox/flox:Ksp-Cre mice treated with Smac-mimetic compared with treatment with DMSO (control; P<0.01). (D) BUN levels in Pkd1^flox/flox:Ksp-Cre mice treated with Smac-mimetic were lower than in DMSO-treated controls (mean ± SEM; P<0.05). (E) The terminal deoxynucleotidyl transferase-mediated digoxigenin-deoxyuridine nick-end labeling (TUNEL) assay showed that Smac-mimetic significantly increased cystic epithelial cell death in Pkd1 conditional knockout mouse kidneys at PN7 compared with age-matched DMSO-treated controls. (Scale bar, 20 µm.) Cyst-lining epithelial cells positive for TUNEL staining expressed as a percentage of all cyst-lining cells (identified by 4',6-diamidino-2-phenylindole nuclear staining) from at least three different animals (mean ± SEM; P<0.01). (F) Smac-mimetic did not affect cyst epithelial cell proliferation in PN7 kidneys from Pkd1^flox/flox:Ksp-Cre mice compared with the kidneys from age-matched Pkd1^flox/flox:Ksp-Cre mice with DMSO treatment; 1000 nuclei per mouse kidney section were counted. Shown is mean ± SEM (P>0.05). (G) Smac-mimetic treatment did not induce fibrosis in PN7 kidneys from Pkd1^flox/flox:Ksp-Cre mice. Tissues stained with Masson trichrome. (H) Survival was increased in Pkd1^flox/flox:Ksp-Cre mice treated with Smac-mimetic compared with Pkd1^flox/flox:Ksp-Cre mice treated with DMSO (control; P<0.01). PCNA, proliferating cell nuclear antigen.

We also examined the effect of Smac-mimetic on cyst growth in hypomorphic Pkd1^nl/nl mice, which have a slower progression of disease.³³ Smac-mimetic (7.5 mg/kg) or DMSO (control) was injected intraperitoneally to pups from PN9 to PN25 every other day, and kidneys were removed and analyzed on PN28. The administration of Smac-mimetic delayed cyst progression (Figure 4, A and B) and induced apoptosis of epithelial cells lining cysts (P<0.01) (Figure 4C and Supplemental Figure 6) in PN28 kidneys (n=6) of Pkd1^nl/nl mice compared with age-matched DMSO-injected Pkd1^nl/nl mice (n=6). In contrast, apoptotic cells were not detected in Pkd1 wild-type PN28 kidneys treated with Smac-mimetic or DMSO (Supplemental Figure 6). Treatment with Smac-mimetic also decreased the kidney weight/body weight ratio and the levels of BUN in Pkd1^nl/nl mice (Figure 4, D and E), indicating that Smac-mimetic preserved normal parenchyma and improved renal function. Smac-mimetic administration did not affect cystic renal epithelial cell proliferation and did not induce additional fibrosis in PN28 kidneys from Pkd1 homozygous Pkd1^nl/nl mice compared with kidneys from age-matched Pkd1 homozygous Pkd1^nl/nl mice treated with DMSO (Figure 4, F and G).

Figure 4.

Smac-mimetic reduced cyst growth in Pkd1 hypomorphic homozygous Pkd1^nl/nl mice. (A) Histologic examination of Pkd1^nl/nl PN28 kidneys treated with Smac-mimetic showed reduced renal size and more compact residual parenchyma compared with DMSO control. (Scale bar, 2 mm.) (B) Cystic index was decreased in PN28 kidneys from Pkd1^nl/nl mice treated with Smac-mimetic compared with kidneys from the age-matched Pkd1^nl/nl mice treated with DMSO. Shown is mean ± SEM of at least five evenly spaced sections from each kidney quantified for each condition (P<0.001). (C) Smac-mimetic significantly increased apoptosis of cyst renal epithelial cells in PN28 Pkd1^nl/nl kidneys compared with age-matched DMSO-treated kidneys (TUNEL assay). (Scale bar, 20 μm.) Shown is mean of at least three mice ±SEM (P<0.01). (D) The ratio of KW/BW was decreased in PN28 Pkd1^nl/nl mice treated with Smac-mimetic compared with treatment with DMSO (control; P<0.01). (E) BUN levels were lower in PN28 Pkd1^nl/nl mice treated with Smac-mimetic than mice treated with DMSO (mean ± SEM; P<0.01). (F) Smac-mimetic did not affect cyst epithelial cell proliferation in PN28 kidneys from Pkd1 homozygous Pkd1^nl/nl mice compared with kidneys from age-matched Pkd1 homozygous Pkd1^nl/nl mice with DMSO treatment; 1000 nuclei per mouse kidney section were stained, and only strongly stained nuclei were considered PCNA-positive (mean ± SEM; P>0.05). (G) Smac-mimetic treatment did not induce additional fibrosis in PN28 kidneys from Pkd1^nl/nl mice. Tissues stained with Masson trichrome. Shown is mean ± SEM (P>0.05).

Smac-Mimetic Targets cIAP1 But Not cIAP2 in Renal Epithelial Cells

To investigate further the mechanisms of Smac-mimetic–induced specific cell death in Pkd1 null renal epithelial cells, we examined the potential target of the Smac-mimetic. Smac-mimetic–induced degradation of endogenous cIAP1 occurred within 30 minutes but had no effect on cIAP2 in Pkd1 null MEK and PN24 cells (Figure 5A and Supplemental Figure 7). In addition, Smac-mimetic did not alter the expression of other components of complexes I and II in cystic renal epithelial cells (Supplemental Figure 7). Of note, Smac-mimetic induced the degradation of cIAP1 in Pkd1 wild-type MEK cells and Pkd1 heterozygous postnatal PH2 cells, raising the possibility that cIAP1 degradation is necessary but not sufficient to induce cell death (Figure 5, A and B). cIAP1 is an ubiquitin E3 ligase that is capable of mediating autoubiquitination and ubiquitination of several of their binding partners, such as RIPK1.³⁴ In support of this premise, we found that a proteasome inhibitor, carbobenzoxy-Leu-Leu-leucinal (MG132), efficiently blocked the Smac-mimetic–dependent decrease of cIAP1 (Figure 5, C and D), whereas z-Vad had no effect on the degradation of cIAP1 (Figure 5C). Thus, Smac-mimetics stimulate proteasomal degradation of cIAP1 independent of caspase activation. This proteasome-mediated degradation of cIAP1 was observed in both embryonic and postnatal renal epithelial cells independent of TNF-α signaling (no TNF-α was administered in the above experiments) (Figure 5, A and B and Supplemental Figure 7).

Figure 5.

Smac-mimetic induced cIAP1 degradation in renal epithelial cells. (A) Smac-mimetic induced cIAP1 degradation in Pkd1 WT and null MEK cells. Western blot analysis of cIAP1 and cIAP2 expression in whole-cell lysates of Pkd1 WT and null MEK cells treated with Smac-mimetic at indicated time points. (B) Western blot analysis of cIAP1 and cIAP2 expression from whole-cell lysates of Pkd1 WT MEK cells and postnatal Pkd1 heterozygous PH2 cells treated with Smac-mimetic at indicated time points. (C) Western blot analysis of cIAP1 expression in Pkd1 WT and null MEK cells treated with Smac-mimetic and/or z-Vad or MG132. (D) Western blot analysis of the expression of cIAP1 and cIAP2 from whole-cell lysates of Pkd1 null MEK cells and postnatal Pkd1 homozygous PN24 cells treated with Smac-mimetic plus MG132 at indicated time points. MG132 efficiently blocked Smac-mimetic–dependent degradation of cIAP1 in Pkd1 null MEK cells and PN24 cells. min, minute.

Smac-Mimetic and/or TNF-α Only Induce the Formation of TNFR1-Associated Complex and Caspase 8/RIPK1/FADD Complex in Cystic But Not Normal Renal Epithelial Cells

To compare the effect of Smac-mimetic and/or TNF-α on the formation of TNFR1-associated complex and caspase-8/RIPK1/FADD complex in renal epithelial cells, we performed coimmunoprecipitations with anti-TNFR1 and anti-caspase 8 antibodies. Coimmunoprecipitation of TNFR1 showed that TNF-α alone resulted in modest recruitment of both cIAP1 and RIPK1 to the receptor in Pkd1 null MEK cells, whereas Smac-mimetic alone showed similar effects on RIPK1 and induced the degradation of cIAP1 (Figure 6A). The addition of both Smac-mimetic and TNF-α resulted in additional RIPK1 recruitment to the TNFR1 complex, despite the loss of cIAP1 (Figure 6A). In contrast, treatment with TNF-α alone, Smac-mimetic alone, or Smac-mimetic plus TNF-α failed to induce detectable association of RIPK1 and/or cIAP1 with TNFR1 in Pkd1 wild-type MEK cells (Figure 6A). This result may be because of lower endogenous levels of TNFR1, cIAP1, and RIPK1 and the transient nature of this complex in these cells. We found that recruitment of RIPK1, which was initially enhanced in Pkd1 null MEK cells on TNF-α plus Smac-mimetic cotreatment, was subsequently released from the receptor over a 6- to 12-hour period (Figure 6B). This finding was consistent with observations in cancer cells; TNF-α induction caused rapid recruitment of RIPK1 to the receptor and was polyubiquitinated, and in the presence of Smac-mimetic, RIPK1 was deubiquitinated and dissociated from the receptor faster than in TNF-α–only treated cells.⁷ Similar to the result observed in sensitive tumor cells, we found that Smac-mimetic plus TNF-α treatment significantly increased formation of the caspase-8/RIPK1/FADD complex compared with TNF-α or Smac-mimetic treatment alone in Pkd1 null MEK cells (Figure 6C). The addition of z-Vad (a pan caspase inhibitor) further enhanced the formation of the caspase-8/RIPK1/FADD complex (Figure 6C), probably by preventing the complex from falling apart after caspase-8 activation. The addition of MG132 (a proteasome inhibitor) together with the Smac-mimetic, TNF-α, and z-Vad decreased the level of the caspase-8/RIPK1/FADD complex formation compared with treatment of TNF-α or Smac-mimetic alone (Figure 6C). In contrast, treatment with TNF-α alone, Smac-mimetic alone, Smac-mimetic plus TNF-α, or Smac-mimetic plus TNF-α and z-Vad failed to induce detectable association of RIPK1 and/or FADD with caspase 8 in Pkd1 wild-type MEK cells (Figure 6C). These data suggest that the Smac-mimetic induced degradation of cIAP1, and the TNF-α–mediated increased RIPK1 binding to TNFR1 and subsequent release from it are necessary for increasing the formation of caspase-8/RIPK1/FADD complex in cyst epithelial cells (Figure 6C). To explore the need for cIAP1 elimination and TNF-α for RIPK1-dependent caspase-8 complex formation in Pkd1 null MEK cells, we knocked down cIAP1 proteins with siRNA oligos. The results showed that the formation of the RIPK1/FADD/caspase-8 complex was increased in cIAP1 knockdown cells in response to TNF-α alone (Figure 6D). In addition, silencing of RIPK1 with siRNA abolished the interaction between caspase-8 and FADD in Pkd1 null MEK cells treated with both Smac-mimetic and TNF-α (Figure 6E), suggesting that RIPK1 is a required component of the FADD/caspase-8 interaction. The caspase-8/RIPK1/FADD complex apparently activated downstream caspase-3 in Pkd1 null renal epithelial cells but not Pkd1 wild-type renal epithelial cells treated with Smac-mimetic plus TNF-α (Figure 6F). Caspase-3 activation was further confirmed by the appearance of cleaved poly(ADP-ribose) polymerase, a substrate of caspase-3 (Figure 6F).

Figure 6.

Smac-mimetic and TNF-α cotreatment induced RIPK1-dependent TNFR1-associated complex and caspase 8/RIPK1/FADD complex (death complex) formation in Pkd1 cystic renal epithelial cells only. (A) The formation of TNFR1-associated complex was increased in cystic renal epithelial cells on treatment with TNF-α and/or Smac-mimetic for 6 hours. In contrast, TNFR1-associated complex was not formed in Pkd1 WT renal epithelial cells on treatment with TNF-α and/or Smac-mimetic. (B) Recruitment of RIPK1 was initially enhanced but subsequently released from the TNF-α receptor 1 in Pkd1 null MEK cells on TNF-α + Smac-mimetic cotreatment. (C) The formation of prodeath complex was induced in cystic renal epithelial cells on treatment with TNF-α and/or Smac-mimetic. In contrast, caspase 8-associated prodeath complex was not formed in Pkd1 WT renal epithelial cells on treatment with TNF-α and/or Smac-mimetic. (D) Knockdown of cIAP1 with siRNA increased the formation of the RIPK1/FADD/caspase-8 complex in response to TNF-α alone. (E) Silencing of RIPK1 with siRNA abolished the interaction between caspase-8 and FADD in Pkd1 null MEK cells treated with both Smac-mimetic and TNF-α. (F) Treatment with TNF-α plus Smac-mimetic increased the activation of caspase 3 in Pkd1 null MEK cells but not Pkd1 WT MEK cells. Caspase 3 activation was further confirmed by the appearance of cleaved poly(ADP-ribose) polymerase (PARP), a substrate of caspase-3. CK, control; IB, immunoblot; IP, immunoprecipitation; S, Smac-mimetic; T, TNF-α; Z, z-Vad.

Degradation and Cleavage of FLIP Were Induced in Cyst Epithelial Cells Treated with Smac-Mimetic Plus TNF-α But Not Cells Treated with Smac-Mimetic Alone

FLIP has been reported to inhibit caspase-8– and caspase-8/RIPK1/FADD complex-induced apoptosis.²⁰ We found that treatment with TNF-α plus Smac-mimetic induced not only the degradation but also the cleavage of FLIP in Pkd1 null MEK cells (Figure 7A and Supplemental Figure 8A). This result is in contrast with cancer cells, where TNF-α plus Smac-mimetic induced only the cleavage of FLIP but did not affect the levels of FLIP.⁷ Of interest, treatment with Smac-mimetic alone did not induce the degradation and cleavage of FLIP in either Pkd1 wild-type or null MEK cells (Figure 7B and Supplemental Figure 8B). Thus, our results suggest that both cleavage and degradation of FLIP are required for Smac-mimetic and TNF-α to induce death in cyst epithelial cells. This finding may explain why Smac-mimetic induced the degradation of cIAP1 and caused a minor formation of the caspase-8/RIPK1/FADD complex (complex II) in Pkd1 mutant MEK cells but could not induce death in these cells. Thus, the increased formation of the caspase-8/RIPK1/FADD complex and the cleavage and degradation of FLIP may be key factors underlying the specificity of Smac-mimetic plus TNF-α to destroy epithelial cells lining renal cysts.

Figure 7.

Smac-mimetic induced TNF-α–dependent degradation and cleavage of FLIP in Pkd1 null MEK cells but not Pkd1 WT renal epithelial cells. Western blot analysis of cIAP1 and/or FLIP expression in whole-cell lysates of Pkd1 WT and Pkd1 null MEK cells treated with (A) Smac-mimetic plus TNF-α or (B) Smac-mimetic alone at indicated time points.

Discussion

It is generally agreed that ADPKD is a neoplastic condition in which the renal cysts resemble unusual tumors that fill with fluid as they progressively enlarge. Although partially dedifferentiated, the mural cells are not malignantly transformed. The respective roles of cell proliferation and apoptosis in PKDs have not been clearly resolved. In the current study, we investigated the role of apoptosis in ADPKD and delineated the mechanisms by which TNF-α signaling regulates the survival and death of epithelial cells lining renal cysts. We showed that, similar to malignant cells, Pkd1 null cyst cells are highly susceptible to apoptotic destruction induced by the combined administration of a Smac-mimetic and TNF-α (Figure 2, A and B). The similarity of the response to TNF-α and Smac-mimetic in human cyst epithelial cell primary cultures (Figure 2C) reinforces the likelihood that the exuberant apoptotic response is also provoked in patient cysts in situ, which provides the opportunity to translate fundamental understanding of cell survival and death pathways into therapeutic interventions for ADPKD.

We have developed a working model that elucidates how Smac-mimetic and TNF-α may lead to apoptosis only in epithelial cells lining the cysts of patients with ADPKD (Figure 8A, upper panel). There is abundant evidence to indicate that, in ADPKD (in which all renal cells carry a germ-line mutation of PKD1), a second hit or biochemical change focused on the normal allele lowers PC-1 levels below a critical threshold.³⁵ Reduced PC-1 then leads to aberrant cell proliferation, which can be accelerated by cAMP agonists and growth factors.³⁶ Progressive cell division causes the wall of the nascent cyst to extend into the interstitium, eventually reaching a size at which the cyst separates from the parent tubule and becomes an isolated sac. Sustained cyst expansion continues aided by the transepithelial secretion of sodium chloride and water into the lumen. After the cyst separates from the tubule and the wall seals, TNF-α accumulates within the cyst cavity to relatively high levels and can interact with Smac-mimetic to promote apoptosis specifically limited to the mural cells within the cysts.⁹ One can appreciate that, depending on the respective amounts of TNF-α and Smac-mimetic reaching the cyst, the number of mural cells may not change or may even decrease as the battle between cell proliferation and apoptosis determines how many viable mural cells remain.

Figure 8.

Model depicting the effect of Smac-mimetic on cyst enlargement. (A) Cysts arise primarily in collecting ducts when cell polycystin-1 levels fall to a critical threshold. A new phenotype arises that proliferates in response to cAMP and other growth factors, and it generates cytokines and chemokines, including TNF-α, which has been found to accumulate in cyst fluid in an isolated sac formed after the cyst separates from the tubule in ADPKD patients (upper panel). Although the cyst remains attached to the tubule, some of the TNF-α synthesized in the mutant cells may escape into the urine, but some of the TNF-α may retain at cyst developing sites through binding to its receptor, TNFR1 (lower panel). TNF-α is trapped within the cysts or with TNFR1 to partner with Smac-mimetic and promote apoptosis of mural cells. Removal of apoptotic mural cells would reduce cyst surface area and decrease the rate of cyst growth. Extensive apoptosis could cause cysts to shrink. (B) Schematic of extrinsic apoptotic pathway involving TNF-α regulation of renal epithelial cell survival and death in PKD1. Loss of PC-1 results in the upregulation of the components in complexes I and II as well as FLIP. In PKD1, TNF-α is autotranscribed in epithelial cells lining cysts. In the absence of Smac-mimetic, TNF-α activates apoptosis signaling (here referred to as complex I), working through IκB kinase (IKK) and NF-κB to protect PKD1 null cells from premature death and leading to sustained cyst growth. The administration of Smac-mimetic promotes disassociation of apoptosis signaling components from the TNFRI receptor, formation of complex II, and cleavage and degradation of FLIP to relieve inhibition of complex II, leading to cell apoptosis and a reduced rate of cyst growth.

It is unclear whether the kidneys of Pkd1^flox/flox:Ksp-Cre and Pkd1^nl/nl mice at the ages examined have isolated cysts that are detached from the nephrons or whether TNF-α is retained or accumulated within cysts that remain attached to the tubule. Because the lumens of dilated cystic tubules are open to the flow of urine in the tubules, some TNF-α produced within the cyst-lining epithelial cells may flow downstream; however, the rate of fluid flow in the dilated cystic tubules may be very slow, and the fluid containing TNF-α may pool within the dilated cavities. In addition, TNFR1 is upregulated in Pkd1 mutant renal epithelial cells,⁹ suggesting that TNFR1 may bind and retain TNF-α within the microenvironment of cyst (Figure 8A, lower panel). In this scenario, immune cells, which infiltrate the kidney in response to renal lesions and infections,³⁷ may be an initial source of TNF-α. Immune cell-derived TNF-α further stimulates the expression of TNF-α in renal epithelial cells through TNF-α–mediated NF-κB activation (Figure 1A). Then, before isolated cysts form, TNFR1, aberrantly expressed by cyst-lining epithelial cells, binds and retains TNF-α within the cyst fluid at the site of cyst development. Although it is unclear how much TNF-α is required to overcome the survival threshold and induce apoptosis in response to Smac-mimetic treatment, the findings that Smac-mimetic treatment delays cyst formation in Pkd1 mutant mouse kidneys without addition of TNF-α (Figures 3 and 4) suggest that TNF-α in the cystic microenvironment may be sufficient to synergize the proapoptotic effect of Smac-mimetic in vivo.

New features discovered in the current study and relevant to the molecular model described in Figure 8B are listed. (1) Pkd1 mutation results in the upregulation of the components in complexes I and II as well as FLIP. (2) During initial cyst development, immune cells secrete TNF-α, which stimulates its own transcription in mural epithelial cells through the upregulated TNF receptor I and the upregulated components in complex I to activate NF-κB (Figure 1, A and B). Aberrant expression of TNFR1 binds and retains TNF-α during cyst development, leading to elevated levels of TNF-α in early dilated cystic nephrons and the accumulation of TNF-α within isolated cysts detached from the parent nephron segment. Similar autocrine behavior has been observed in cultured human cyst epithelial cells in respect to monocyte chemotactic protein-1 and periostin.^38,39 (3) We discovered that TNF-α within the cyst fluid has the potential to regulate mural epithelial cell survival through activation of NF-κB and its target FLIP (Figures 1, 2D, and 7 and Supplemental Figure 8). (4) We found that a Smac-mimetic in specific partnership with TNF-α can provoke the death of Pkd1 null renal epithelial cells in vitro (Figure 2). (5) We observed that the administration of a Smac-mimetic to two different strains of mice with homozygous mutations of Pkd1 reduced disease progression, presumably by partnering with TNF-α produced within in situ cysts (Figures 3 and 4). (6) We observed that promoting apoptosis of cyst epithelial cells in Pkd1 null mice by administering Smac-mimetic did not alter the rate of cell proliferation (Figures 3 and 4) but was associated with reduced kidney size, cyst area, and BUN and increased survival (Figures 3 and 4).

The data presented here show that Smac-mimetic, only combined with TNF-α, will induce apoptosis exclusively in cystic renal epithelia under the influence of mutated Pkd1. Four mechanisms support the selective effect in Pkd1 mutant renal cysts of the Smac-mimetic, including (1) the upregulation of TNFR1 and associated cIAP1, RIPK1, and TRADD (Figure 1C and Supplemental Figure 2), which were adjusted by c-Myc expression in Pkd1 mutant cystic renal epithelial cells (Supplemental Figure 9), and the upregulation of caspase-8 and FADD, which might be adjusted by the expression of histone deacetylases or histone methyltransferases and augment apoptotic signaling in response to treatment with Smac-mimetic plus TNF-α in Pkd1 mutant but not wild-type cells (Figure 1C and Supplemental Figure 2). In particular, on Smac-mimetic plus TNF-α treatment, (2) the upregulation of the components of TNFR1-associated complex increases the release of RIPK1 from this complex in Pkd1 null epithelia (Figure 6B). (3) Then, the released RIPK1 together with the upregulated caspase-8 and FADD augment the formation and activation of the RIPK1-dependent caspase-8/RIPK1/FADD death complex in Pkd1 null epithelia only but not Pkd1 wild-type renal epithelial cells (Figure 6C). (4) The degradation and cleavage of FLIP are induced in Pkd1 null but not Pkd1 wild-type renal epithelial cells (Figure 7 and Supplemental Figure 8). Please note that the heterogeneity of the cell population in the kidneys from DMSO- and Smac-mimetic–treated mice limited us to confirm the formation of complexes I and II in vivo, because we did not get positive immunoprecipitation results with anti-TNFR1 antibody to pull down RIPK1 and cIAP1 or anti-caspase 8 antibody to pull down RIPK1 and FADD in Pkd1 wild-type and knockout kidneys, respectively.

The weight of our new evidence leads us to think that, under certain conditions, enhanced apoptosis may preserve renal structure by eliminating mural cells from cysts that otherwise would expand endlessly. These findings contradict previous studies, in which reductions in apoptosis were associated with decreases in cyst and kidney expansion as well as improvement in renal function.^40,41 As emphasized in the work by Goilav,⁴² whether apoptosis promotes or retards cyst growth has been muddled by confounding factors, such as animal versus human models, early versus late disease, apoptosis in tubules and interstitial cells versus apoptosis in cysts, and different primary end points. The nature of the mutation responsible for renal cystic disease may be relevant as well. Studies of recessive Pkd models have revealed that increased apoptosis is likely mediated by B cell lymphoma (Bcl2) family members.⁴⁰ Although increased apoptosis is thought to play a role in recessive animal models of PKD (Bcl2 mouse, Pck rat, and congenital polycystic kidney mouse),⁴⁰ the study by Hughes et al.⁴³ showed that loss of Pkd1 and loss of Bcl-2 elicit cyst formation through distinct mechanisms. Also, the study by Hughes et al.⁴³ showed that ablation of one or both alleles of the proapoptotic gene Bim prevents cyst formation in mice deficient in the prosurvival protein Bcl-2, whereas loss of Bim had no effect on cyst development in Pkd1 homozygous mutant mice. Also, loss of Bcl-2 alleles did not significantly influence the Pkd1 mutant phenotype. These studies suggest that Bcl-2–mediated apoptosis is not involved in cystogenesis in mice with Pkd1 deficiency.

Increased apoptosis was not observed in Pkd1^flox/−:Ksp-Cre⁴⁴ or Pkd2-WS25 mouse kidneys in association with cystic disease.⁴⁵ Moreover, the overall number of apoptotic nuclei in these kidneys was very low and not significantly different between cystic and normal kidneys.^44,45 These studies in models deficient in PKD1 suggest that apoptosis may not contribute to cyst progression in polycystin-related ADPKD. The results of the current study are consistent with the data from these orthologous ADPKD models, because apoptotic cells were rarely seen in the kidneys from Pkd1 conditional knockout and Pkd1 hypomorphic Pkd1^nl/nl mice (Figures 3E and 4C and Supplemental Figures 5 and 6). Our findings in cells and intact kidneys of polycystic mice with Pkd1 mutations strengthen the view that apoptosis may not contribute to cyst progression in polycystin-mediated PKD. The increased apoptosis observed in human kidneys in the late stages of ADPKD⁴⁶ may reflect the well known secondary fibrotic process and the development of acquired cysts observed in all renal diseases that progress to the end stage.⁴⁷

Use of the drugs targeting specific pathways to normalize a cystic epithelial cell function, thus preventing cyst formation, also induces cystic epithelial cell apoptosis, such as mammalian target of rapamycin and Sirtuin 1 inhibitors, etc.^48–50 It might be argued that the effects on apoptosis induced with these drugs might be more of a readout or side effect, but nonetheless, induced apoptosis was noted in the treatment of ADPKD animal models, which did not worsen but were delayed in the growth of renal cyst. These studies also support that apoptosis may not contribute to cyst progression in polycystin-mediated PKD. Thus, a key point is whether and how we can specifically increase cystic epithelial cell apoptosis to delay cyst growth. Our study presents for the first time that Smac-mimetic induces TNF-α–dependent cystic renal epithelial cell death only but has no effects on healthy renal cells, leading to the removal of cystic epithelial cells from renal tissue and thus, delaying cyst formation. In addition, reducing apoptosis threshold by Smac-mimetic may also sensitize cystic epithelial cells to other agents, such as mammalian target of rapamycin and histone deacetylase inhibitors,^48,50,51 which induce cystic epithelial cell death to delay cyst growth. However, because the toxicity of Smac-mimetic (GT13072) is unknown, the application of GT13072 to ADPKD patients is still waiting for the ongoing clinical trial results.

In summary, the current study helps to clarify the role of apoptosis in the regulation of cyst size, and in a larger sense, it may open a new approach to target renal cysts and prevent their endless expansion by the administration of Smac-mimetics. This study encourages a paradigm shift from current efforts that focus on normalizing cell function in cystic epithelial cells to directly targeting these cells for removal.

Concise Methods

Cell Culture and Reagents

Dolichos biflorus agglutinin-positive Pkd1 wild-type and Pkd1 null MEK cells were maintained as previously described.⁹ Postnatal Pkd1 heterozygous PH2 and homozygous PN24 cells, which were derived from the proximal tubules, were maintained as previously described.⁴⁴ Primary human ADPKD and normal human kidney cells were cultured as described.⁹ TNF-α was purchased from Sigma. z-Vad and MG132 were purchased from EMD Chemicals. Smac-mimetic (GT13072) was kindly provided by TetraLogic Pharmaceuticals (Malvern, PA).

Western Blot and Immunoprecipitation

Immunoprecipitation was performed on whole-cell lysates as described in the work by Wang et al.⁷ The antibodies used for Western blotting included anti-cIAP1, anti-FLIP, and anti-FADD (Enzo Diagnostics, NY); anti-casp8, anti-X-linked inhibitor of apoptosis, anti–poly(ADP-ribose) polymerase, and anticleaved caspase 3 antibodies (Cell Signaling Technologies, MA); anti-cIAP2, anti-TNFR1, anti–TNF-α receptor-associated protein 2, anti-TRADD, and anticaspase 8 antibodies (Santa Cruz, CA); anti-RIPK1 antibody (BD Biosciences, CA); and antiactin antibody (Sigma, MO). Donkey anti-rabbit, donkey anti-mouse, and donkey anti-rat IgG-horseradish peroxidase (Santa Cruz, CA) were used as secondary antibodies.

RNA Interference

The RNA oligonucleotides targeted for mouse cIAP1 and mouse RIPK1 were purchased from Thermo Dharmacon and transfected with the DharmaFECT siRNA transfection reagent (Dharmacon Technologies, CO). After transfection for 48 hours, cells were harvested and analyzed by Western blotting.

Apoptosis Assays

Terminal deoxynucleotidyl transferase-mediated digoxigenin-deoxyuridine nick-end labeling assay was performed according to the manufacturer’s protocols (In Situ Death Detection Kit; Roche Diagnostics, Germany). Apoptosis was also measured by FACS with the FITC Annexin-V Apoptoois Detection Kit (BD Pharmingen) according to the manufacturer’s instruction. The apoptotic cells were marked as Annexin V⁺PI⁻ cells. Kidney sections from three animals in each group were scored in a blinded fashion for terminal deoxynucleotidyl transferase-mediated digoxigenin-deoxyuridine nick-end labeling-positive cystic lining cells and total cystic lining cells (identified by 4′,6-diamidino-2-phenylindole staining), and the results were expressed as percentage of apoptotic cells.

Immunohistochemistry

Kidneys were fixed with 4% paraformaldehyde (pH 7.4). For proliferating cell nuclear antigen staining, monoclonal mouse antiproliferating cell nuclear antibody (1:1000; Cell Signaling Technologies, Danvers, MA), a biotinylated secondary antibody (1:100; Sigma, St. Louis, MO), and a 3,3′-diaminabenzidine tetrahydrochloride substrate system were used. Images were analyzed with a NIKON ECLIPSE 80i Microscope.

Quantitative RT-PCR

Cells were treated with or without SN50 (50 µg/ml) for 1 hour and then treated with TNF-α (10 ng/ml) for 3 hours. Total RNA was extracted using the RNeasy Plus Mini Kit (Qiagen, MD). cDNA was synthesized using the Iscript cDNA Synthesis Kit (BioRad, CA). TNF-α mRNA expression profile was analyzed by real-time PCR using iTaq SYBER Green Supermix with ROX (BioRad, CA) and carried in an icycler iQTM Real-time PCR Detection System. Each sample was run in triplicate in every experiment, and each experiment was repeated three times. The expression level of TNF-α was normalized to the expression level of actin.

Mouse Strains and Treatment

For analyzing postnatal kidneys treated with Smac-mimetic, the lactating Pkd1^flox/+:Ksp-Cre female mice were injected intraperitoneally with Smac-mimetic (7.5 mg/kg) or DMSO (control) at PN3 and PN5, and postnatal kidneys of Pkd1^flox/flox:Ksp-Cre were harvested and analyzed at PN7. In addition, the hypomorphic Pkd1 homozygous Pkd1^nl/nl pups were injected intraperitoneally with Smac-mimetic (7.5 mg/kg) or DMSO (control) from PN9 to PN25 every other day, and kidneys were harvested and analyzed at PN28.

Statistical Analyses

Data are presented as mean ± SEM. An unpaired two-tailed t test was used to determine the significant differences. A P value less than 0.05 is considered significant.

Disclosures

None.