Increasing reactive oxygen species as a therapeutic approach to treat hereditary leiomyomatosis and renal cell carcinoma

Abstract

Hereditary leiomyomatosis renal cell carcinoma (HLRCC)-associated renal tumors are aggressive and tend to metastasize early. There are currently no effective forms of therapy for patients with advanced HLRCC-associated kidney cancer. We have previously shown that HLRCC cells express a high level of reactive oxygen species (ROS). In the present study we investigated the cytotoxic-effects of increasing ROS level using bortezomib in combination with cisplatin on HLRCC cells in vitro and in an in vivo xenograft model. The cytotoxic effect of several ROS inducers on FH-deficient cells was assessed by synthetic lethality. ROS inducers had a pronounced impact on the viability of FH-deficient cells. Because of its high potency, the proteasome inhibitor bortezomib was further investigated. Bortezomib induced apoptosis in vitro in HLRCC cells and inhibited HLRCC tumor growth in vivo. Bortezomib-associated cytotoxicity was highly correlated with cellular ROS level: combining bortezomib with other ROS inducers enhanced cytotoxicity, while combining bortezomib with a ROS scavenger inhibited its cytotoxic effect. Finally, HLRCC murine xenografts were treated with bortezomib and cisplatin, another ROS inducer. This regimen induced HLRCC tumor regression in vivo. These findings suggest that increasing ROS level in HLRCC above a certain threshold can induce HLRCC-tumor cell death. Increasing tumor ROS with bortezomib in combination with cisplatin represents a novel targeted therapeutic approach to treat advanced HLRCC-associated renal tumors.

Background

Hereditary leiomyomatosis and renal cell carcinoma (HLRCC) is an autosomal dominant hereditary familial cancer syndrome, characterized by germline mutation of the mitochondrial tricarboxylic acid (TCA) cycle enzyme, fumarate hydratase (FH).¹ Renal tumors associated with HLRCC are very aggressive and tend to metastasize early.² We have previously characterized a human FH (-/-) HLRCC kidney cancer cell line UOK262, which cannot use oxidative phosphorylation as a primary source of energy.^3,4 Instead, because of their impaired TCA cycle, these cells depend on glycolysis to provide energy.³ These kidney cancer cells, which are characterized by impaired oxidative phosphorylation and dependence on glucose, are a clear example of the “Warburg effect” in cancer.

We previously reported that the glycolytic switch induced by the loss of FH activates the NADPH oxidase enzyme complex and produces high levels of reactive oxygen species (ROS).³ ROS are implicated in several normal and/or pathological processes, including immunity, aging, tumorigenesis and metastasis.^5–8 At low or moderate concentration, ROS may have a beneficial effect on cellular physiology; for example, phagocytic cells produce ROS to promote defense against microorganisms.^9,10 However, higher concentrations of ROS may cause DNA and protein damage which can lead to apoptosis.^11,12

In the present report, we have identified a novel therapeutic approach to target selectively HLRCC tumor cells. We sought to take advantage of the naturally elevated ROS level in UOK262 cells.³ While many tumor cells generate ROS, they adapt by upregulating ROS scavengers such as glutathione or superoxide dismutase.^13,14 However, if the capacity of the scavenger system is exceeded, apoptosis is likely to result.^15,16 Targeting the oxidative stress process has been explored in a number of cancer models.^17,18 Reducing ROS level in tumor cells may impact oncogenic hallmarks such as genetic instability, high proliferation and cell motility and antioxidants have shown promising activity in several tumor models.^19,20 Alternatively, in some cases further upregulation of ROS with pro-oxidants may induce apoptosis. For example, the proteasome inhibitor bortezomib, a pro-oxidant, has shown promising antitumor activity in several murine models of human neoplasms including prostate cancer and multiple myeloma.^21,22 Bortezomib is currently used as a second line therapy for multiple myeloma.²³

In the present study, we evaluated whether increased or decreased ROS might affect UOK262 cell viability. Our data showed that bortezomib induced apoptosis in vitro and in vivo, and that this response correlated with elevated cellular ROS level. Furthermore, combination of bortezomib with cisplatin enhanced the anti-tumor properties of the proteasome inhibitor in vitro as well as in vivo. Our data thus provide a novel potential therapeutic approach for treatment of HLRCC-associated kidney cancer.

Results

ROS inducers cause death of FH-deficient cells.

In order to selectively evaluate the impact of ROS on HLRCC-tumor cell viability, we performed a synthetic lethal screen using five cell lines previously described: the HLRCC cell line UOK262, UOK262 cells stably transfected with either a control vector (UOK262CV) or a vector expressing functional FH (UOK262/FH), as well as the normal epithelial kidney cell lines HK2 stably transfected with either a control shRNA (HK2) or a shRNA against FH (HK2/D10).³ We treated the cells for 48 hours with several ROS inducers (menadione, bortezomib and cisplatin) over a wide range of concentrations.³ As shown in Table 1, the FH-deficient cell lines (UOK262, UOK262CV and HK2/D10) were more sensitive to ROS inducers than the matched cell lines expressing wild type FH activity (HK2 and UOK262/FH). Of the agents tested, bortezomib demonstrated a remarkable activity (IC₅₀ = 3.8–7.6 nM). These results suggest that induction of ROS may have therapeutic potential in HLRCC and identified bortezomib as a potential drug candidate for the treatment of this disease.

Table 1.

Cytotoxicity of ROS-inducing agents correlates inversely with fumarate hydratase activity status

Compounds	HK2 IC50 (µM ± STDEV)	HK2/D10 IC50 (µM ± STDEV)	UOK262/FH IC50 (µM ± STDEV)	UOK262CV IC50 (µM ± STDEV)	UOK262 IC50 (µM ± STDEV)
Menadione	11.19 ± 2.45	4.74 ± 1.94	23.9 ± 4.78	10.44 ± 4.47	4.6 ± 3.37
Bortezomib	50.13 × 10−3 ± 8.1 × 10−3	4.79 × 10−3 ± 2.84 × 10−3	17.2 × 10−3 ± 2.7 × 10−3	7.6 × 10−3 ± 1.6 × 10−3	3.8 × 10−3 ± 1.6 × 10−3
Cisplatin	50.62 ± 3.98	33.75 ± 5.74	14.5 ± 2.25	6.5 ± 4.4	11.7 ± 2.25

Cells were treated for 48 hours with a range of concentrations of the different drugs. Cytotoxicity assays where performed to determine the IC₅₀ value for each drug. Cytotoxicity is inversely correlated to fumarate hydratasae (FH) status. HK2 and HK2/D10 are HK2 cells stably transfected with, respectively, control shRNA or shRNA against FH. UOK262CV and UOK262/FH are UOK262 cells stably transfected with, respectively, a control vector or a vector expressing functional FH. UOK262 cells are the parental, untransfected cell line. N, 6–8; STDEV, standard deviation.

Proteasome inhibition induces apoptosis in HLRCC cells in vitro and in vivo.

We then investigated whether cell death induced by bortezomib was due to apoptosis. Caspase-3 is a critical executioner of apoptosis, and its activation requires proteolytic processing of its inactive zymogen, generating cleaved fragments.²⁴ As shown in Figure 1A and B, bortezomib (5 nM) caused a marked increase in cleaved-caspase 3 in UOK262 cells, and its effect on cell viability was reversed by pre-treating the cells for 1 h with the pan-caspase inhibitor Z-VAD (10 µM; Fig. 1A and B). To assess these results in vivo, we developed an HLRCC xenograft model. Twice weekly treatment of tumor-bearing mice with bortezomib (1 mg/kg) induced up to 70% tumor growth inhibition (Fig. 1C). TdT staining of excised tumor showed a significant increase of apoptotic cells in the bortezomib-treated-tumors compared to their controls (Fig. 1D), however, proliferation of tumor cells, measured by Ki67 index, was not affected by this treatment (data not shown).

Figure 1.

The proteasome inhibitor bortezomib induces apoptosis of HLRCC cells in vitro and in vivo. (A) The apoptosis marker cleaved caspase 3 was visualized by western blot. (B) Cell viability of UOK262 treated with the pan-caspase inhibitor Z-VAD (10 uM) 1 h prior to bortezomib (5 nM, 16 h). (C) Scid/beige mice were treated with bortezomib (1 mg/kg, twice weekly) for 3 weeks. Tumor volumes were measured with a caliper prior to each treatment. Arrows show when the mice were treated. (D) TdT staining was performed on five-micron slides from formalin-fixed, paraffin embedded HLRCC xenograft samples. Z-VAD: Z-VAD-FMK; * and ^#p < 0.05.

Bortezomib's pro-apoptotic effect is dependent on ROS level.

We confirmed, as previously shown,³ that the basal ROS level in FH-deficient cells is significantly higher than in their paired control cells (Fig. 2A, left 5 bars). When measured the effect of bortezomib on ROS, ROS level was increased following bortezomib treatment in vitro (Fig. 2A and B). This effect was inhibited by treatment with the ROS scavenger NAC. To further explore whether ROS may play a role in the sensitivity of UOK262 to bortezomib, we determined UOK262 cell viability after only 24 hours exposure to menadione (6 µM), bortezomib (5 nM) and NAC (5 mM) (Fig. 2B). The combination of the two ROS inducers bortezomib and menadione proved to be exceptionally cytotoxic to these cells. Moreover, while NAC did not affect UOK262 viability, it reversed the effect of menadione and bortezomib. These data suggest that bortezomib's pro-apoptotic effect is due to its pro-oxidant properties in FH-deficient cells.

Figure 2.

Bortezomib-induced toxicity correlates with ROS level in FH-deficient cells. (A) ROS level was measured with H2DCFDA after treatment for 4 h with bortezomib (bort., 5 nM), menadione (Mena., 6 uM), and/or n-acetyl-cysteine (NAC, 5 mM). (B) Viability of UOK262 cells after bortezomib (bort., 5 nM) treatment alone or in combination with ROS modulators (Mena: menadione, 6 uM; NAC, 5 mM); * and ^#p < 0.05.

Cisplatin amplifies bortezomib in vitro cytotoxicity and induces tumor regression when combined with bortezomib in vivo.

The previous data suggest that combining another ROS inducer with bortezomib may enhance tumor cytotoxicity. To test a more clinically usable agent for this purpose we chose cisplatin, a pro-oxidant that has been combined with bortezomib in several clinical studies.^25,26 Treatment of FH-deficient cells with cisplatin (2.5 µM) increased the impact of bortezomib (5 nM) on ROS and consequently on cell viability (Fig. 3A and B). Those effects were inhibited when combined with the ROS scavenger NAC. At the concentrations used, neither bortezomib nor cisplatin, whether administered as single agents or in combination, were cytotoxic for HRCE normal renal epithelial cells. Given this tumor specificity in vitro, we wished to assess the activity of this combination in vivo. We treated skid/beige mice bearing HLRCC xenograft tumors once weekly with either cisplatin (2.5 mg/kg) or bortezomib (0.1 mg/kg per week, 1/20^th of the dose per week displaying single agent activity in our model, see Fig. 1C), or with the combination of both drugs. After four weeks, tumors regressed by approximately 40% in the animals treated with bortezomib and cisplatin, while these agents individually had no significant effect on tumor growth at the doses and schedule used (Fig. 4A). At the end of the experiment, tumors were harvested and paraffin-embedded for immunohistochemistry. The Ki67 index was not affected by this treatment (data not shown). However, TdT staining showed that the bortezomib/cisplatin combination significantly increased the number of apoptotic cells compared to untreated specimens.

Figure 3.

Bortezomib's effect is enhanced in vitro by the ROS inducer cisplatin. (A) ROS level was measured with H2DCFDA after treatment for 4 h with bortezomib (bort., 5 nM), cisplatin (Cispl., 6uM), and n-acetyl-cysteine (NAC, 5 mM) in UOK262 cells. (B) Cell viability in UOK262 and HRCE cells 24 h after treatment with bortezomib (bort., 5 nM), cisplatin (Cispl., 6 uM), and n-acetyl-cysteine (NAC, 5 mM), alone or in combination. * and ^#p < 0.05.

Figure 4.

Combining cisplatin and bortezomib induces HLRCC tumour regression in vivo. (A) Scid/beige mice were treated i.p. once weekly with bortezomib (0.1 mg/kg) and cisplatin (2.5 mg/kg), alone or in combination. Arrows show when the mice where treated. (B) Five-micron slides from formalin-fixed, paraffin embedded samples were utilized for TdT staining. ^# and *p < 0.05.

Discussion

Although recent reports have described signaling pathways underlying the tumorigenesis of HLRCC kidney cancer in vitro and in vivo, few pre-clinical studies have evaluated possible therapeutic strategies based on these signaling derangements.²⁷ Because HLRCC tumor cells exhibit increased aerobic glycolysis, TCA dysfunction and high basal ROS level,^3,4 we chose to exploit their disrupted redox balance as a targeting strategy. The present study demonstrates that HLRCC kidney cancer cells are sensitive to pro-oxidants, and in particular to the combination of the proteasome inhibitor bortezomib with cisplatin.

Somewhat surprisingly, our data suggest that, at least in FH-deficient cancers, optimal bortezomib cytotoxicity requires a certain amount of ROS. To our knowledge, such a dependence of bortezomib activity on oxidative stress has not been previously described. Our data suggest that an elevated intracellular ROS level may sensitize cells to bortezomib and predict drug efficacy in vivo. A corollary of these observations suggests that the proapoptotic anti-tumor activity of bortezomib might be enhanced by combination with a second pro-oxidant agent. We confirmed this hypothesis for HLRCC both in vitro (bortezomib + menadione) and in vivo (bortezomib + cisplatin). Importantly, the bortezomib/cisplatin combination required less bortezomib than was necessary to obtain single agent activity and proved to be non-toxic to normal renal epithelial cells while highly cytotoxic to UOK262. In vivo, this combination led to tumor regression. These results support the hypothesis that proteasome inhibitors might have preferential activity in tumors expressing a high basal ROS level.

Combination of bortezomib and cisplatin has been evaluated in clinic trials with other cancers. A Phase II study used this regimen as a first line treatment for malignant mesothelioma (www.clinicaltrials.gov), and a Phase I study used the combination to supplement radiotherapy for advanced head and neck cancers.²⁸ However, neither of these clinical trials considered the combination as a pro-oxidant therapy and so did not examine tumor ROS levels before or after treatment. It would be of interest to determine whether clinical response to this drug combination depends on a pro-oxidant effect similar to the one we have described.

In conclusion, our data demonstrate the high sensitivity of HLRCC-associated kidney cancer cells to pro-oxidants, and they highlight the possible efficacy of combining bortezomib and cisplatin for treatment of HLRCC-associated kidney cancers. Further, our findings support the hypothesis that tumors with elevated ROS may be particularly sensitive to proteasome inhibition.

Materials and Methods

Cell lines.

Early passages (between 6 and 12) of the HLRCC cell line UOK262, established in the Urologic Oncology Branch (NCI, Bethesda, MD USA), were used.⁴ UOK262CV, UOK262FH, HK2CV (HK2) and HK2/D10 were generously provided by Dr. Sunil Sudarshan (UTSA, San Antonio, TX).³ Renal cortical epithelial cells (HRCE) were obtained from Clonetics (San Diego, CA). All the cells were cultured in Dulbecco's modified Eagle's medium High Glucose supplemented with 10% fetal calf serum, 5 mM non essential amino-acids and 5 mM HEPES.

Drugs.

Cisplatin, menadione, N-actetyl-L-cysteine (NAC), and N-Benzyloxycarbonyl-Val-Ala-Asp(O-Me) fluoromethyl ketone (Z-VAD-FMK or Z-VAD) were purchased from Sigma-Aldrich (St. Louis, MO). Bortezomib was provided by Millenium Pharmaceuticals, Cambridge, MA.

Cell viability.

Cell viability was measured using the cytotoxic assay kit purchased from Promega Biosciences, Inc., (San Luis Obispo, CA), following the manufacturer's protocol. Briefly, 5,000 cells were plated in black clear-view 96-well plates (PerkinElmer, Waltham, MA) and treated as described. A fluorogenic cell-permeable protease substrate was added to the wells and enzyme-dependent fluorescence was measured after 30 minutes in a spectrophotometer (Victor, PerkinElmer).

Reactive oxygen species measurements.

Cells were plated in Black clearview 96-well plates (PerkinElmer) and treated as described in Results or in the figure legends. Chloromethyldichlorodihydrofluorescein diacetate (CM-H2DCFDA, Invitrogen, Carlsbad, CA) was reconstituted in DMSO and then diluted in PBS. Plates were loaded with 10 µM dye for 1 hour at 37°C before being returned to pre-warmed media without phenol red. After 60 minutes, oxidation of CM-H2DCFDA in 2, 7-dichlorodihydrofluorescein diacetate (DCF) was measured in a spectrophotometer. Data were normalized to the total protein concentration of each well, as determined using the BCA reagent (Pierce Biotechnology, Inc., Rockford, IL).

Immunoblot analysis.

Equivalent amounts of whole cell protein lysates were separated on 7.5% SDS-PAGE gels (Bio-Rad) and transferred on PVDF membranes. Hsp70 was visualized using a mouse polyclonal antibody (Assay Designs, Inc., Ann Arbor, MI). Cleaved caspase 3 was visualized using a rabbit monoclonal antibody (Cell Signaling Technology, Inc., Danvers, MA). Horseradish peroxidase-linked secondary antibodies (Sigma-Aldrich) were used with the ECL protein detection system (Pierce). Experiments were performed in triplicate and equal loading of gels was confirmed by probing for β-actin (Cell Signaling Technology).

Immunohistochemistry.

Five-micron slides from formalin-fixed, paraffin embedded (FFPE) samples were utilized for immunohistochemical analysis. The indirect HRP-labeled antibody method was employed.²⁹ Briefly, sections were placed on positively charged glass slides and deparaffinized in xylene. After rehydration in a series of graded alcohols, heat-induced antigen retrieval in a solution of either citrate buffer (pH 6.0) or Tris-EDTA (pH 8.0) was carried out, before to incubate the samples with 3% hydrogen peroxide to inactivate endogenous peroxidases. Slides were washed 3 times with TBS containing 1% Tween-20 and incubated with mouse anti-Ki67 antibody (1 µgml⁻¹, Dako, Carpenteria, CA) for 1 hr at room temperature. Isotype-matched antibodies raised in the same species were substituted to the primary antibodies as negative controls. Sections were subsequently incubated with HRP-polymer-labeled secondary antibodies for 30 min at room temperature and reactions were developed with 0.025% 3, 3′-diaminobenzidine solution, lightly counterstained with Mayer's hematoxylin and permanently mounted for observation. At least two different sections per sample were analyzed.

Apoptotic cells were visualized using ApopTag^® Peroxidase In Situ Apoptosis Detection Kit (Millipore, Billerica, MA) following the manufacturer's protocol.

The percentage of positive nuclei over the total number of tumor cells was obtained from five to ten high power fields. Subcellular localization of immunostaining was also evaluated.

HLRCC xenograft studies.

Animal experiments were performed in accordance with the guidelines of the Animal Care and Use Committee of the National Institutes of Health. We established a xenograft transplantation model (UOK262XD) using an intraoperatory tumor specimen from the same HLRCC patient from which the UOK262 cell line was derived. The specimen was implanted directly into SCID/beige mice and maintained by passaging from mouse to mouse in the NIH animal care facility. We used 7 week old female skid/beige mice for both experiments. Pharmacologic treatments started 2 to 3 weeks after tumor implantation (average volume = 150 mm³). Mice were randomly separated into groups with comparable tumor volumes. Animals were either treated twice weekly by intraperitoneal (i.p.) injection with either bortezomib or vehicle (PBS/DMSO); or they were treated once weekly by i.p. with either bortezomib, cisplatin, bortezomib + cisplatin or vehicle (PBS/DMSO).

Statistics.

Unless specified, all values are expressed as mean ± Standard Error. All experiments have been performed at least in triplicate, with exception of the animal studies. Values were compared using the Student-Newman-Keul's test. p < 0.05 was considered significant.

Abbreviations

HLRCC

hereditary leiomyomatosis and renal cell carcinoma

FH

fumarate hydratase

ROS

reactive oxygen species