Introduction

Renal cell carcinoma (RCC) represents the most common adult kidney cancer, accounting for >200,000 new cases and >100,000 deaths worldwide annually [1]. The incidence of RCC has been steadily increasing at a rate of 2 to 3 % per year across the past several decades [2, 3].

Frequent loss or inactivation of VHL E3 ubiquitin ligase and consequent overexpression of hypoxia-inducible factor (HIF) has been implicated in the pathophysiology of RCC [4–7]. In normal tissues, the product of the VHL gene is associated with ubiquitination and degradation of the HIF through an oxygen-sensing mechanism [8, 9]. In normoxia, HIF-1α and HIF-2α are hydroxylated by prolyl-hydroxylases in an iron-dependent manner. This post-translational modification allows the recognition of HIF by the VHL complex and leads to its degradation by the proteasome [10, 11]. However, in the absence of oxygen, or in the presence of a mutated VHL gene, HIF-1α and HIF-2α are stabilized and induce the expression of a panel of transcriptional target genes such as VEGF, PDGF, and GLUT1, supporting the metabolic shift that underlies RCC tumorigenicity [12]. Even in the presence of oxygen, clear cell kidney cancer cells with VHL gene mutation display a “pseudo-hypoxic” phenotype. Although the degradation of both HIF-1α and HIF-2α are regulated by VHL [6, 13], HIF-2α has been thought to be the dominant oncoprotein in VHL-deficient RCC cells [14–16]. It was also recently suggested that HIF-1α may function as a tumor suppressor gene in VHL-deficient clear cell kidney cancer [17].

PI-3K/AKT/mTOR pathway is the key convergent pathway, and PI-3K is a critical “signaling hub” downstream of multiple growth factor receptors (VEGFR, PDGFR, EGFR etc.). Recent literature revealed a major role of activated AKT in metastatic RCC, an effect mediated through PI-3K activation or loss of associated PTEN, a phosphatidylinositol (3–5)-triphosphate phosphatase that negatively regulates the PI-3K/AKT signaling pathway in renal cell carcinoma [18–22]. Cystic lesions of patients with VHL show hyperactivated PI-3K/AKT signaling [23]. Increased phospho-AKT levels were found in about 50 % of RCC tumor samples and most commonly in the clear cell subtype [24]. Combined mutations of VHL and PTEN lead to kidney cysts in mice [23]. PTEN-inactivating mutations [25, 26] or decreased PTEN expression has been identified in about 30 % of clear cell RCCs [27, 28]. Human renal cancer cell lines also show constitutive activation of AKT, and PI-3K/AKT inhibitor treatment induces apoptosis and inhibits cell growth in vitro and in xenografts [29]. Recent literature provides evidence that PI-3K/AKT signaling regulates HIF-1α stabilization [30–35]. Some reports have suggested that HIF-1α protein expression is dependent on mTOR in certain cellular contexts [36], and mTOR is a crucial component of a downstream PI-3K/AKT pathway. The importance of HIF-1α/HIF-2α in RCC progression and angiogenesis, and regulation of HIF-α by PI-3K/AKT/mTOR pathway, suggests that a pan-PI-3 kinase inhibitor should be therapeutic choice in RCC treatment.

SF1126 is a vascular-targeted pan-PI-3K inhibitor prodrug with antitumor and antiangiogenic activity [37]. SF1126 is a RGDS-conjugated LY294002 prodrug, which is designed to exhibit increased solubility and bind to specific integrins within the tumor compartment, resulting in enhanced delivery of the active compound to the tumor vasculature and tumor. This drug has completed phase I clinical trial in solid tumors and B-cell malignancies [38]. SF1126 showed good results on a clear cell renal cancer (RCC) patient who had stable disease (SD) for 84 weeks (21 cycles), suggesting the efficacy of this drug in the treatment of RCC patients. However, the mechanism by which SF1126 exerts its antitumor and antiangiogenic properties in RCC is unknown.

In the present study, using VHL-null (786-0) and VHL-WT (Caki) renal cell carcinoma cell line, we provide evidence that SF1126 blocks phosphorylation of AKT (in both 786-0 and Caki cell lines) and phosphorylation of ERK (in 786-0) in RCC cell lines. Interestingly, the inhibition of PI-3K/AKT signaling by SF1126 in vitro suppressed both constitutive and hypoxia-induced expression of HIF-2α in 786-0 and HIF-1α in Caki cells. Immunohistochemical analysis of CD31 staining demonstrated that SF1126-treated tumors had significantly decreased vascularization compared with control tumors. These findings revealed a putative antiangiogenic mechanism associated with SF1126 treatment which might contribute to its antitumor effect. The preclinical studies presented here have the potential to provide important information about SF1126 as we are using this drug in the treatment of RCC patients.

Methods

Synthesis of prodrugs: SF1126 and SF1326, cleavage and pharmacokinetic studies

The RGDS-conjugated and RADS-conjugated prodrugs, termed SF1126 and SF1326, respectively, were described previously [37]. Upon cleavage at neutral pH, SF1126 or SF1326 spontaneously cleave to release the SF1101 active chemical moiety to inhibit PI-3K activity. SF1101 is referred to as commercially available LY294002. Synthesis and characterization of SF1126 and SF1326 have been described before [37]. Briefly, the RGDS and RADS peptides were prepared in fully protected form on either Wang resin or 2-Cl-Trityl resin using standard Fmoc solid-phase peptide synthesis chemistry. The crude samples were purified on reverse-phase semi-prep HPLC (water/acetonitrile with 0.1 % trifluoroacetic acid added) to >95 % purity. A more efficient solution-phase synthesis route has been developed that is more amenable to large-scale GMP manufacturing of SF1126.

Tissue culture, cell lines and reagents

VHL-null (786-0) and VHL-WT (Caki) renal carcinoma cell lines were purchased from the American Tissue Culture Collection (Rockville, MD, USA). These cell lines were propagated in Dulbecco’s modified Eagle’s medium (Cellgro), 10 % FBS, 20 mM HEPES, 2 mM l-glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin (all from Invitrogen). Cells were routinely cultured in 95 % O2 and 5 % CO2 at 37 °C and made hypoxic by placing them in hypoxic chamber (Billups chamber) at 1 % O2, 5 % CO2, and 94.9 % nitrogen. Cells were pretreated with indicated compounds for 30 min prior to insulin-like growth factor (IGF) or hypoxia treatment. SF1126 was obtained from SignalRx Pharmaceuticals [38]. Compounds were diluted into medium to the appropriate concentration before adding to cells to ensure that all treated dishes received the same concentration, and DMSO was added to control or untreated dishes at a final dilution of 1:1,000. Antibodies specific for PTEN, AKT, phospho-S473-AKT, ERK and phospho-44/42 ERK were obtained from Cell Signaling Technology. Antibodies against HIF-1α and HIF-2α were from BD Transduction Laboratories and Novus Biological Inc., respectively. All other chemicals were purchased from Sigma unless otherwise stated.

Plasmid constructs and transient transfections

Transient transfections in VHL-null 786-0 cells were done using Lipofectamine plus (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Briefly, cells were split into 10-cm dishes so that 24 h later they were 70 % confluent. At this time, each dish was transfected with 10 µg plasmid using lipofectamine plus. After 48 h of transfection, cells were serum-starved for 4 h followed by hypoxia of 4 h and lysate preparation.

Biochemical analysis

For all Western blots, 2 × 106 cells were plated in 10-cm tissue culture dishes such that the density of the cells at the time of lysis was 70–80 % confluent. Cells were allowed to adhere overnight and next day were treated with 5 and 10 µM of SF1101, SF1126 (RGDS targeting moiety) and SF1326 (RADS-targeted prodrug derivative) for 30 min. Whole-cell lysates were prepared using RIPA buffer: 150 mM NaCl, 50 mM Tris, pH 7.4, 5 mM EDTA, 0.1 % SDS, 20 mM glycerophosphate, 10 mM NaF, 250 µM NaVO4, 1 mM phenylmethylsulfonyl fluoride and complete protease inhibitor (Roche Molecular Biochemicals). Clarified lysates were resolved in 10 % SDS-PAGE as described before [39, 40] and probed for phospho-AKT, AKT, phospho-ERK, ERK (Cell Signaling Technology Inc, Beverly, MA, USA) and actin (Sigma, St Louis, MO, USA).

For normoxic or hypoxic HIF-1α/HIF-2α stabilization studies, following treatment with 25 µM SF1126 or SF1101 or SF1326, cells were exposed to hypoxia or left under normoxia for 4 h. Lysis was performed immediately at the bench in ambient air. Extracts were resolved by 10 % SDS-PAGE, electrotransferred to a nitrocellulose membrane, and incubated with a mouse monoclonal antibody against HIF-1α (BD Transduction Laboratories) or rabbit polyclonal antibody against HIF-2α (Novus Biological Inc.).

HIF-1α transcription Activity

To determine the effects of SF1126 on HIF-1α function, Caki cells were plated in 6-well tissue culture plates at 70–80 % confluence for 18 h. Cells were then cotransfected with 1 µg/well of a reporter plasmid containing VEGF hypoxic response element promoter sequence (HRE) upstream of the firefly luciferase gene and 20 ng of pRL-TK vector as internal control plasmid that expresses a renilla luciferase gene (Promega). Cells were treated with SF1126 at 25 µM concentration for 30 min prior to placing cells into hypoxic chamber (1 % O2) for 24 h at 37 °C. The HRE response element was induced under hypoxic conditions (1 % oxygen for 24 h), resulting in an activation of HIF-1α transcriptional activity. The luciferase activity in the cell lysates was measured with a luminometer (Lumister galaxy, BMG, Winooski, VM, USA) using Dual Luciferase Assay System (Promega). The relative luciferase activity for each sample was calculated as a ratio of firefly luciferase activity divided by renilla luciferase activity. The level of renilla luciferase activity was used as an internal control to normalize the differences in transfection efficiencies among the samples.

Reverse transcription-PCR (RT-PCR) analyses

Total RNA was isolated from cell lines using the Qiagen RNAeasy kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions. cDNA was prepared from 1 µg RNA sample using iscript cDNA synthesis kit (Bio-Rad, Hercules, CA, USA). cDNA (2 µl) was amplified by RT-PCR reactions with 1× SYBR green supermix (Bio-Rad, Hercules, CA, USA) in 96-well plates on an CFX96 Real-time system (Bio-Rad, Hercules, CA, USA), using the program: 5 min at 95 °C, and then 40 cycles of 20 s at 95 °C, 1 min at 58 °C and 30 s at 72 °C. Specificity of the produced amplification product was confirmed by the examination of dissociation reaction plots. Relative expression levels were normalized to Gapdh expression according to the formula <2(Ct gene of interest−Ct Gapdh) [41].

Animal studies

Athymic female mice (CD-1 nu/nu, 20–25 g) were used for in vivo tumor growth inhibition studies. Mice were obtained from the NIH/NCI repository and housed in on a 12-h light/dark cycle with food and water ad libitum under pathogen-free conditions, according to the guidelines of the Association for the Assessment and Accreditation for Laboratory Animal Care, International. All of the in vivo studies were carried out under approved Emory University experimental animal care and use protocols. Five million 786-0 or Caki cells in 100 µl PBS were injected subcutaneously into the right flank of each mouse. Tumor growth was monitored twice per week for external measurements using vernier calipers. Tumor volume was calculated using the formula V = (A × B 2)/2 where A and B represent length and width of the tumor, respectively.

For SF1126 experiments, treatment was initiated when tumors reached a tumor volume of 80–100 mm3. Mice were divided randomly into two groups receiving vehicle (acidified sterile water diluents for SF1126) or SF1126. SF1126 (100 µl volume) was injected subcutaneously three times weekly (Monday, Wednesday and Friday) for 3 weeks at a dose of 25 mg/kg in the left flank of the mice. No toxicities were noted in mice treated with SF1126 or vehicle.

CD31 immunohistochemistry

At the end of the efficacy studies, tumors were harvested and placed in OCT blocks for frozen section analysis or fixed in 10 % buffered formalin and/or processed into paraffin. Sections of tumor tissue at 4 µm thickness were stained with rat antimouse CD31 antibody for the detection of the murine tumor microvasculature [42].

Integrin-induced cell migration

Cell migration assays were performed on polycarbonate membranes using Transwell migration chambers (pore size 8 µm; Coster Corp., Cambridge, MA, USA). The underside of the membrane to which cells migrate was coated with 10 µg/ml vitronectin in PBS for 1 h at 37 °C. Surfaces were subsequently blocked with heat-denatured BSA. Transwells were placed into the lower chamber containing 600 µl of serum-free medium. RCC cells were treated with SF1126 at 20 µM concentration for 30 min prior to adding cells (2 × 105/well) to the top of the migration chamber (uncoated side) and allowed to migrate to the coated side of the chamber for 4 h at 37 °C. Haptotaxis assay was quantitated as described previously. The haptotaxis response was further confirmed by demonstrating the complete abrogation of migration by coating both sides of the membrane with vitronectin.

Actin dynamics

VHL-null, 786-0 were seeded on vitronectin-coated cover slips in 6-well plates. Cells were treated with 25 µM SF1126 for 30 min and were then processed for Phalloidin 555 staining. Nuclei were stained with DAPI. Stained cells were photomicrographed for actin polymerization using confocal microscopy. Cells were imaged using a Zeiss (Thornwood, NY, USA) LSM 510 Meta confocal microscope with a 63x (1.4-numerical-aperture) or 100x (1.4-numerical-aperture) Plan-Apochromat oil objective. All images were acquired using Zeiss LSM 510 software and processed using Adobe Photoshop 7.0 as described elsewhere [43].

Integrin-induced Rac1 activation

VHL-null, 786-0 were stimulated in 10 cm of nontissue culture-coated Petri dishes coated with 10 µg/ml vitronectin in the presence or absence of SF1126 (10 µM) for 30 min. Following adhesion, cells were chilled with ice-cold HANKs’ balanced salt solution. Cell lysates were prepared in 25 mM HEPES, pH 7.5, 150 nM NaCl, 1 % Igepal CA-630, 10 mM MgCl2, 1 mM EDTA, 10 % glycerol, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 25 mM NaF and 1 mM sodium orthovanadate. The binding reaction was initiated by adding 10 µl of Pak-1 agarose (GST fusion protein, corresponding to the p21 binding CRIB domain, PBD, residues 67–150, of human PAK-1, expressed in E. coli and bound to glutathione agarose) to each sample and incubated for 1 h at 4 °C with gentle shaking and processed as described [44, 45]. The membrane was probed with monoclonal Rac1 antibody (1:1,000, Upstate Biotechnology). The amount of total Rac1 protein in each lysate was quantified as an additional loading control.

Statistical analysis

The Student’s t test was used to evaluate differences observed between experimental groups and to compare tumor volume differences between SF1126 and vehicle-treated controls.

Results

The pan-PI-3 kinase inhibitor, LY294002 has been shown in a number of in vivo models to possess antitumor activity [46, 47] but suffers from poor pharmacokinetics [48], poor water solubility, and undesirable toxicity. The chemical, pharmacological and toxicological data of LY294002 suggested that it will not be a viable drug candidate [49]. Considering the published literature using this compound and its desirable inhibition properties, our lab has developed water soluble, clinically viable, tumor-targeted prodrug form of LY294002/SF1101 that have better pharmacokinetic and toxicological profile and hence have better anticancer efficacy [37]. The RGDS and RADS-conjugated prodrugs, termed SF1126 and SF1326, respectively, were developed in solid-phase organic synthesis procedure and currently RGDS-conjugated SF1126 has completed phase I clinical trials for solid tumors and B-cell malignancies [38]. We have previously shown that the inhibitory effects of this drug (RGDS-conjugated SF1126) in in vitro cell-based assays were blocked (at least 50 %) by the prepulsing with RGDS peptide by virtue of occupying the RGDS binding sites of the αvβ3/αvβ5 and α5β1 integrins in different cell lines (U937, U87MG, U251MG) and primary primate endothelial progenitor cells [37]. In the present study, we investigated the effect of this drug on VHL-null and WT renal cell carcinoma cell lines.

SF1126 reduces level of p-AKT and p-ERK in RCC cell lines

Frequent loss or inactivation of VHL is the major cause in RCC malignancy; however, current literature also revealed a major role of activated AKT in metastatic RCC, an effect mediated through the loss of tumor suppressor protein, PTEN [18, 20, 22]. The capacity of RGDS-conjugated pan-PI-3K inhibitor, SF1126 to suppress the phosphorylation of AKT (phospho-Ser473-AKT) at 5 and 10 µM concentration in RCC cell line (786-0) was significant as determined by Western blot analysis (Fig. 1a). Similarly, treatment with SF1101/LY294002 or SF1326 (at 5 and 10 µM concentration) abrogated phospho-Ser473-AKT levels in RCC cell line (786-0) (Fig. 1a). To validate the potency of SF1126 in VHL-WT cell lines, we used Caki cell line. SF1126 significantly blocked phosphorylation of AKT in Caki cell line in concentration-dependent manner (Fig. 1b). Importantly, RGDS alone has no effect on phosphorylation of AKT (lane 2, Fig. 1b).

Fig. 1
figure 1

SF1126 inhibits activation of AKT and ERK in renal carcinoma cell lines. a Subconfluent renal carcinoma cell line (VHL-null, 786-0) was incubated with SF1126 or SF1101 or SF1326 at 5 and 10 µM concentrations for 30 min. Cell lysates were subjected to Western blot analysis for pAKT, pERK, total AKT and total ERK1/2. β-actin served as loading control N.T. shows no treatment. b VHL-WT renal carcinoma cell line (Caki) was incubated with SF1101 or SF1126 at 5 and 10 µM concentrations for 30 min or RGDS alone at 50 µM concentration for 15 min. Cell lysates were subjected to Western blot analysis for the expression of pAKT and pERK. Data show that RGDS alone at 50 µM concentration has no effect either on AKT phosphorylation or ERK 1/2 phosphorylation

Full size image

We next examined the ability of SF1126 to block activation of MAPK pathway by measuring ERK1/2 phosphorylation in VHL-null renal carcinoma cell line (786-0). RAF–MEK–ERK is one of the key pathways (other than PI-3K/AKT/mTOR pathway) downstream of growth factor receptors for proliferation/growth. It has been reported that RAS–RAF–MEK–ERK and PI-3K/AKT/mTOR pathways interact with each other to regulate growth/survival in normal cells and in some cases of tumorigenesis [50]. Our results demonstrate that treatment of 786-0 cells with SF1126 or SF1101/LY294002 or SF1326 attenuated the activation of ERK1/2 (Fig. 1a) at lower concentration (5 µM). However, treating the cells with higher concentration (10 µM) of SF1126 or SF1101 resulted in increased p-ERK levels (Fig. 1a). The results provide evidence for the activation of MAPK pathway in RCC cells treated with PI-3K inhibitors. In support of this notion, the existing literature suggest that MAPK pathway (Ras–Raf–MEK–ERK) is involved in cross talk with the PI-3K pathway, such that inhibition of the PI-3K pathway can detrimentally turn on signaling through the MAPK pathway [51, 52]. Contrary to 786-0 cells, Caki cells treated with SF1126 or LY294002/SF1101 or SF1326 showed concentration-dependent inhibition of ERK1/2 phosphorylation (Fig. 1b). These data suggest that SF1126 controls renal carcinoma cell survival and proliferation via the regulation of AKT–mTOR and MEK–ERK pathways.

SF1126 reduces expression of HIF-1α/HIF-2α and its transcriptional activity in RCC cell lines

Previous reports have suggested that PTEN and PI-3K pathway regulate HIF-α transcription factor, a control point for the hypoxic induction of VEGF [31, 32]. So next we investigate the effect of SF1126 on normoxic/hypoxic HIF-1α/HIF-2α stabilization and on HIF-1α/HIF-2α’s transcription activity in RCC cell lines. Figure 2a demonstrates that there is no HIF-1α expression in VHL-null RCC cell line (786-0). These results are consistent with previous reports which suggest that loss of HIF-1α protein expression was a common event in RCC cell lines [10, 53]. Previous reports also suggest that RCC cells lacking HIF-1α expression maintains VEGF through HIF-2α [10]. Results presented in Fig. 2b suggest that HIF-2α is the predominant isoform of HIF-α in 786-0 cell line and SF1126 blocks normoxic and hypoxic accumulation of HIF-2α, as well as inhibits HIF-2α transcriptional activity (Fig. 2c). Interestingly, inhibition of HIF-2α accumulation following 25 µM of SF1126 treatment was markedly higher than the equimolar amount of SF1101 or false-targeted SF1326 (Fig. 2b). SF1126 significantly blocked transcription of VEGF in 786-0 cells (Fig. 2c). In order to confirm these results, we determined the effect of SF1126 on the hypoxia-induced HIF-1α accumulation in the VHL-WT Caki cell line. Similar to 786-0, SF1126 at 25 µM concentration significantly and more potently than equimolar concentration of SF1101 blocked hypoxia-induced HIF-1α accumulation in VHL-WT Caki cell line (Fig. 2d). Treatment with SF1126 was specific for the regulated α subunit of HIF-1 and HIF-2, but had no effect on the protein levels of HIF-1β (data not shown). We next examined whether SF1126 has any effect on HIF’s transcriptional activity. For this, Caki cells were transiently transfected with a construct containing Luciferase gene under the control of the hypoxia response element (HRE) from the VEGF promoter [54]. The firefly luciferase and Renilla luciferase activities were measured in the lysate using Dual Luciferase Assay System (Promega). Results show that under hypoxia, promoter activities of HRE-luciferase constructs were increased in tumor cell line (Fig. 2e). Interestingly, increased levels of hypoxia-induced HIF transcription was inhibited in the presence of pan-PI-3K inhibitor SF1126. These results were also validated by decreased VEGF mRNA levels (downstream target for hypoxic HIF transcription) in SF1126-treated Caki cells (Fig. 2f). Collectively, these data suggest that antiangiogenic activity of SF1126 correlates with a block in the HIF-α-VEGF signaling in RCC cells.

Fig. 2
figure 2

Effects of pan-PI-3K inhibitors on HIF-1α/HIF-2α expression and its transcriptional activity. a, b VHL-null RCC cells were treated with 25 µM SF1126 or parental compound SF1101 or false-targeted compound SF1326 or no treatment for 30 min at 37 °C and were exposed to 20 % (normoxic) or 1 % (hypoxic) O2 for 4 h. Immunoblot assays of HIF-1α (a) and HIF-2α expression was performed. As expected there is no HIF-1α protein expression in normoxic or hypoxic conditions. Data show that inhibition of HIF-2α expression following 25 µM of SF1126 treatment was markedly higher than equimolar amount of SF1101 or SF1326 in both normoxic and hypoxic conditions. These data providing evidence that a functional VHL protein is not required for the inhibition of HIF-2α by SF1126. N.T. shows no treatment, +C indicates positive control (Lysates derived from bone marrow derived macrophages isolated form C57BL/6 mice) for HIF-1α and HIF-2α. c VHL-null 786-0 cells treated with 25 µM SF1126 were exposed to normoxia or hypoxia for 24 h, followed by RNA isolation and Real-time PCR for VEGF. d To further substantiate the above observation, VHL-WT RCC cells (Caki) were treated with 25 µM SF1126 or parental compound SF1101 or no treatment for 30 min at 37 °C and were exposed to 1 % (hypoxic) O2 for 4 h, followed by lysate preparation and immunoblot analysis of HIF-1α. e Downregulation of HIF-1α promoter activity (for VEGF) in hypoxic Caki cells in the presence of SF1126. The promoter activity of VEGF-HRE (V6R) luciferase gene construct was determined in VHL-WT Caki cell lines under conditions of normoxia versus cellular hypoxia. The relative V6-HRE-luciferase activity was expressed as a ratio of firefly luciferase activity divided by renilla luciferase activity (relative luciferase unit, RLU). Normoxic samples served as control for hypoxia. The bars in the figure represent the mean value and SD of the relative V6-HRE-luciferase activity from three independent experiments. *P < 0.05 (t test). f Caki cells treated with 25 µM SF1126 were exposed to normoxia or hypoxia for 24 h, followed by RNA isolation and Real-time PCR for VEGF. Graphs present mean ± SD, *P < 0.05, **P < 0.01 and ***P < 0.001 versus control (c, f) in hypoxic conditions as determined using pairwise two-sided Student’s t test

Full size image

SF1126 regulates MDM2-induced hypoxic degradation of HIF-1α/HIF-2α via 26S proteasome

To obtain better understanding of the mechanism involved in HIF-α inhibition by SF1126, RCC cells were treated with SF1126 in the presence or absence of the proteosome inhibitor MG132. Data show that MG132 restored HIF-1α level in SF1126-treated VHL-WT Caki cells and HIF-2α levels in 786-0 cells (Fig. 3a, b) implying that SF1126-induced proteosomal-mediated degradation of HIF-1α/HIF-2α during hypoxia. The observation that HIF-α can be degraded under hypoxic conditions in PI-3K-dependent manner suggested the involvement of a potential E3 ligase in proteasome-dependent degradation. In a recently published work from our laboratory, we reported that MDM2, whose transport between the cytoplasm and nucleus is regulated by PI-3K/AKT, is the E3 ligase controlling HIF-degradation [34]. In order to test whether this hypothesis is true in VHL-null 786-0 cells, phosphorylation site mutants of MDM2 which regulate subcellular localization of this protein were employed. S166A-MDM2 plasmid with a serine-to-alanine substitution cannot be phosphorylated by AKT; thus, it remains in the cytoplasm [55], whereas S166E-MDM2 construct where serine 166 is mutated to glutamic acid mimics a phosphorylated MDM2 and is predominantly localized in the nucleus [55, 56]. In this manner, we were able to test the effect of cytoplasmic versus a nuclear MDM2 on the hypoxic stability of HIF-1α under conditions of PI-3K inhibition. In these experiments, VHL-null 786-0 cells were transduced with S166A-MDM2 or S166E-MDM2 plasmid, serum-starved for 4 h followed by treatment with or without MG132 and hypoxia for 4 h. S166A MDM2 which is known to be localized to the cytoplasmic compartment degrades HIF-1α in a PI-3K-dependent manner (Fig. 3c). In contrast, the S166E mutant of MDM2 which is obligately localized to nuclear compartment (Fig. 3c) does not induce the hypoxic degradation of HIF-1α (Fig. 3c). Interestingly, we note that blocking the proteasome with MG132 reverse the effect caused by S166A MDM2 plasmid (Fig. 3c). Taken together, we can conclude that transfection of an obligate cytoplasmic MDM2 (S166A) mutant, results in the degradation of HIF-1α in a PI-3K-independent manner (Fig. 3c), confirming a model which predicts that MDM2 in the cytoplasm mediates the degradation of HIF-1α under hypoxia.

Fig. 3
figure 3

SF1126 regulate MDM2-induced hypoxic degradation of HIF-1α/HIF-2α via the 26S proteasome pathway. a, b VHL-null 786-0 (a) and VHL-WT Caki (b) cell lines were treated with 25 µM SF1126 in the presence of proteosome inhibitor 10 µM MG132 for 30 min before giving hypoxia (1 % O2) for 4 h, followed by cell lysate preparation. Equal amounts of protein from each cell lysate were resolved by SDS-PAGE, transferred, and immunoblotted with antibodies against HIF-1α, HIF-2α and β-actin (as loading control). Data show that SF1126-induced degradation of HIF-1α was restored following the treatment of proteosome inhibitor, MG132. c VHL-null 786-0 cells were transfected with 10 µg of S166A or S166E MDM2 plasmids. 48 h after transfection cells were serum-starved, followed by pulse with 10 µM MG132 and hypoxia for 4 h and whole-cell extract preparation. These lysates were run on SDS-PAGE, transferred on nitrocellulose membrane and probed for HIF-2α and FLAG tag

Full size image

Antitumor activity of SF1126 in human xenograft

SF1126 significantly demonstrated robust antitumor efficacy in xenograft tumor models and tumor-induced angiogenesis in vivo using different human carcinoma cell lines [37]. Figure 4a, c illustrates the antitumor efficacy of RGDS-conjugated SF1126 in RCC tumor xenograft models (786-0 and Caki cells, respectively). Mice-bearing 100 mm3 tumors were treated subcutaneously with SF1126 at 25 mg/kg, every alternate day (Monday/Wednesday/Friday) for 3 weeks. SF1126 produced tumor growth inhibition with no evidence of toxicity, as measured by weight loss relative to control animals or drug-related lethality. Furthermore, we evaluated the effects of SF1126 on the capacity of RCC tumor cells to recruit a blood supply in vivo during three times per week treatments with SF1126 at 25 mg/kg. A quantitation of microvessel density (MVD) in control versus SF1126-treated tumors demonstrated a significant decrease in MVD in SF1126-treated tumors (Fig. 4b, d), suggesting that pan-PI-3K inhibition could also impair tumor growth through effects on tumor vasculature.

Fig. 4
figure 4

SF1126 demonstrates antitumor efficacy in human xenograft model and blocks angiogenesis. a, c VHL-null renal carcinoma cells (786-0) (a) and VHL-WT Caki (c) cells were implanted sc in the right flank of athymic mice as described in “Methods” section. Treatment was initiated on the day shown when all mice had tumors ranging in size from 80 to 100 mm3. SF1126 was given sc at 25 mg/kg for every alternate day (Monday/Wednesday/Friday) for 3 weeks. There was no lethality and no increase in weight loss in treated group relative to the corresponding control group. Growth rate of tumors in vivo was significantly different when compared to vehicle-treated mice (n = 7) Graphs present mean ± SD. *P < 0.05, **P < 0.01 versus vehicle as determined using pair wise two-sided Student’s t test. b, d PI-3K inhibitor, SF1126 blocks angiogenesis. Left MVD based on CD31 immunohistochemical staining of tumor from vehicle-treated mice or animals treated with SF1126 (25 mg/kg for every alternate day for 3 weeks) concomitant with tumor implantation in 786-0 (b) and Caki cells (d). Right bar diagrams of the quantitation of CD31-positive microvessels with in the tumor tissue in 786-0 (b) and Caki cells (d) at ×40 and ×20, respectively. We note a significant inhibitor effect of SF1126 on tumor-induced angiogenesis. Bars represent SD of mean (n = 7, P < 0.05)

Full size image

SF1126 inhibits integrin- mediated RCC (786-0) cell migration

Renal cell carcinoma (RCC) is a highly metastatic tumor. Metastatic tumors spread to different organs and are the primary cause of death in cancer patients. Integrin-mediated migration is a necessary step for invasive and metastatic property of any tumor. Several studies show that a major role of small GTPase, Rac1, in the regulation of actin polymerization hence controls integrin-directed migration [57]. The ability to block the migratory and invasive capacity of tumor cells offers a new avenue to treat patients with malignant disease. It has also been reported by other laboratories and unpublished observation from our laboratory that PI-3K activity is required for Rac-induced cell motility [58, 59]. Figure 5a, b clearly shows that integrin-mediated migratory property of VHL-null-786-0 cells and dynamic reorganization of actin cytoskeleton were significantly abrogated while treating with pan-PI-3K inhibitor, SF1126. Untreated cells display features consistent with polarization and movement, including strong F-actin staining in the extended lamellipodia and filopodia. In contrast, the SF1126-treated cells display contracted/condensed and disorganized actin structures, and lack polarized orientations in these structures. To determine whether the inhibition of PI-3K impairs Rac1 activation, we analyzed the activities of Rac1 by GTP binding in SF1126-treated cells in response to integrin engagement. Consistent with abrogation of cell migration (Fig. 5a) and dynamic reorganization of actin filament (Fig. 5b) following SF1126 treatment, treatment of 786-0 cells with SF1126-inhibited Rac1 activation in response to vitronectin stimulation (Fig. 5c). From these data, we can suggest that PI-3K regulates integrin- mediated dynamic reorganization of actin cytoskeleton and activation of small G-protein Rac1 required for a motile phenotype of RCC.

Fig. 5
figure 5

Effect of pan-PI-3K inhibitors on integrin-mediated clear renal carcinoma cell migration. a Transwells with polycarbonate membranes (8 µm pore size) were coated with 10 µg/ml vitronectin for 1 h at 37 °C. Following the treatment with SF1126 (20 µm for 30 min), cells (2 × 105) in 100 µl serum-free media were added to the upper chamber of the well. Left the photomicrograph of 786-0 cells which have migrated to the lower chamber as quantified by crystal violet staining. The migration was quantified by counting the migrated cells from 10 to 12 randomly selected fields. Right represents the bar diagram showing quantification of no. of cells migrated per field. Data represents mean ± SD representative of 5–6 independent experiments, P < 0.002. Results show that SF1126 blocked integrin-mediated migration in renal carcinoma cells. b Effect of SF1126 on the state of actin polymerization in 786-0 cells on vitronectin. Cells were plated on vitronectin and treated with SF1126 (20 µM) for 30 min. Fixed cells were processed for confocal microscopy (see details in “Methods” section). The figure shows dramatic change in the cytoskeletal organization of actin in SF1126 treated cells. c The conversion of GDP-Rac1 to activated GTP-Rac1 in SF1126 (10 µM) treated 786-0 cell lines was measured following stimulation with vitronectin at the indicated time. Levels of GTP-bound Rac1 were measured by pull-down experiments using GST fusion protein representing the GTP-Rac1 binding CRIB domain of PAK kinase. It is noted that SF1126 markedly blocked Rac1-GTP activity following vitronectin stimulation. From these data, we can suggest that Pan-PI-3K inhibitor, SF116 inhibits integrin-mediated dynamic reorganization of actin cytoskeleton, activation of small G-protein Rac1 leading to motile phenotype of RCC

Full size image

A schematic model for the interaction between PI-3K/AKT/mTOR pathway and RAS–RAF–MEK pathway is presented in Fig. 6 which clearly shows how these pathways are blocked by SF1126, a pan-PI-3K inhibitor.

Fig. 6
figure 6

A schematic model for the interaction between PI-3K/AKT/mTOR pathway and RAS–RAF–MEK pathway. Activation of these pathways leads to phosphorylation of S6 Kinase and ERK. S6 Kinase and ERK are critical components of general protein translation machinery important for cell survival/proliferation. HIF-α, the potential proangiogenic factor is also regulated by mTOR. The pathway is negatively regulated by PTEN and pan-PI-3K inhibitor SF1126

Full size image

Discussion

There is a significant unmet need for the development of effective therapy that can stabilize or slow the progress of renal cell carcinoma. Although, FDA-approved two antiangiogenic agents, sorafinib (Nexavar) and sunitinib (Sutent), have shown some clinical benefits for treating advanced RCC [60–63], there are also reports that patients treated with these drugs develop resistance and experience disease progression. SF1126 is a vascular-targeted pan-PI-3K inhibitor prodrug with antitumor and antiangiogenic activity [37]. We recently published that SF1126 can block lymphoproliferative response in mouse lymphoma model [64]. This drug has completed phase I clinical trial in solid tumors and B-cell malignancies [38]. In the present study, we have described the efficacy of RGDS-conjugated pan-PI-3K inhibitor, SF1126, in renal tumor progression and tumor angiogenesis.

The major identified problems in renal cell carcinoma are, (1) in most of the cases VHL is inactivated as a result HIF-α accumulation and its high transcription activity (level of VEGF, PDGF etc.) which leads to highly angiogenic tumor (2) PTEN level is very low or negligible, so AKT activation is high and proliferation rate is high and (3) renal cell carcinoma is highly metastatic, so these tumor cells easily and rapidly invade to the other organs. We clearly need a drug or combination of drugs that can control all of the above.

SF1126 used in the present study, markedly inhibited AKT phosphorylation, an indicator of AKT–mTOR cell survival pathway activity in VHL-null (786-0) or VHL-WT (Caki) RCC cell lines. In various malignant tumor cell lines, where activation of PI-3K–AKT/mTOR pathways were observed, a disruption of AKT activation has been shown to inhibit cell proliferation/survival and an increase apoptosis. A major role of activated AKT in metastatic RCC has been reported, an effect mediated through the loss of PTEN in renal cell carcinoma [18–20, 22]. Our data show that high basal level of activated AKT in RCC cell lines was markedly inhibited by the treatment of SF1126 (Fig. 1a, b). SF1126 also blocked ERK1/2 phosphorylation in VHL-null RCC cell line (Fig. 1a). One possible explanation of this, PI-3K activates Rac-GTPase, which once activated can induce numerous downstream biochemical pathways including p21-activated kinase (PAK) and PAK activates RAF kinase. MEK once activated by RAS and RAF can phosphorylate and activate ERK1/2 and provides signal for proliferation/survival [65, 66]. However, SF1126 treatment increased phosphorylation of ERK at higher concentrations. It is well known that the MAPK pathway (Ras–Raf–MEK–ERK) is involved in cross talk with the PI-3K pathway such that inhibition of the PI-3K pathway can turn on signaling through the MAPK pathway [51, 52]. Likewise, inhibition of the MAPK pathway can detrimentally turn on the PI-3K pathway [67]. These studies suggest that targeting both PI-3K and Raf/MEK/ERK signaling axes in the RCC might provide long-term tumor control or eradication of the disease.

The most common molecular abnormality in renal cell carcinoma is loss of VHL leading to high steady-state levels of HIF-α expression and its target genes e.g., VEGF, PDGF, TGF-α, tissue factor, HGF, EPO etc. A large body of clinical evidence suggests that intra-tumoral hypoxia and high level of HIF-α correlate with elevated aggressive behavior of cancer cells and their resistance to therapy, leading to poor prognosis [68, 69]. Recently, Shinojima et al. [53] analyzed a panel of RCC cell lines for the expression levels of HIF-1α and HIF-2α and found that the loss of HIF-1α protein expression was a common event in RCC cell lines. Our data also show that there is no HIF-1α protein expression in VHL-null RCC cell line (786-0) (Fig. 2a). Loss of HIF-1α protein expression may be due to truncated HIF-1α mRNA transcripts and also with post-transcriptional silencing [53]. It has been suggested by number of investigators, that HIF-2α plays a major role in renal tumor growth and angiogenesis [15, 16, 70, 71]. HIF-2α, which has been reported to be regulated similarly to HIF-1α [72, 73] was also downregulated by SF1126 (Fig. 2b). These data clearly demonstrate that HIF-1α/HIF-2α accumulation and transcriptional activity were markedly reduced following the treatment of SF1126 (Fig. 2b–f) in RCC cell lines. Furthermore, from proteosomal inhibitor (MG132) studies, we can suggest that SF1126-induced proteosomal degradation of HIF-1α during hypoxia in Caki cells and HIF-2α in 786-0 cells (Fig. 3a, b). We recently published that HIF-1α can be degraded under hypoxic conditions via the 26S proteasome and that MDM2, a E3 ligase is necessary for the hypoxic degradation of HIF-1α [34]. Similar to this report, we observed that MDM2 is mediating the degradation of HIF-2α in VHL-null 786-0 cells in proteasome and PI-3K-dependent manner (Fig. 3c).

Previous report from our laboratory established that LY294002/SF1101 blocks tumor progression in vivo [47]. Our previous studies on SF1126 demonstrate that this drug has potent antiangiogenic and antitumor activity in different xenograft models [37]. In addition to inhibiting AKT and ERK1/2 activation, we show here that SF1126 exhibits potent inhibition of HIF-1α/HIF-2α accumulation/activity which may play a major role in angiogenesis. SF1126 demonstrated a single-agent antitumor activity (Figs. 3c, 4a) against a preclinical xenograft model using human malignant renal cancer cell lines (786-0, and Caki). SF1126 acts predominantly to prevent growth of the tumors in this preclinical study. Since treatment with SF1126 (25 mg/kg in alternate days for 3 weeks) is well tolerated, it is possible that increasing the duration of therapy or dose would have produce tumor regression. This pan-PI-3K inhibitor, SF1126 appears to generate its effects via a dual pathway affecting both angiogenesis through inhibitory effects on endothelial cells and pericytes (significantly lowering MVD in tumor tissues obtaining from SF1126-treated mice compared with control, Fig. 4b, d) and tumor cells directly through inactivation of AKT and ERK1/2.

Malignant carcinomas (such as RCC) are highly motile cells and have dramatic alteration in their actin-cytoskeletal organization that facilitates their invasive and migratory behavior. During migration cell projects lamellipodia (organized changes in the cytoskeletal actin filaments) and attach to the extracellular matrix (ECM) protein (integrin-mediated adhesion to the substratum) and simultaneous break existing ECM contacts at the trailing edge. This allows cells to move forward. Our data demonstrate that SF1126 markedly blocked integrin-mediated renal carcinoma cell migration and deregulated actin filaments (Fig. 4a, b). Several reports support that blocking PI-3K activation using pharmacological inhibitors effectively blocked actin polymerization and cell migration [74]. It has been suggested by different laboratories that Rho family of small GTPases-including Rho, CDC42 and Rac can regulate actin dynamics hence cell movement [57]). Furthermore, multiple lines of evidence demonstrate that several Rac regulated cellular functions depend on PI-3K activity [75] and products of PI-3K activate Vav, a guanine nucleotide exchange factor which promote the conversion of inactive Rac-GDP to active Rac-GTP [76]. In line with these reports our data show that SF1126 significantly blocked Rac1 activity following αvβ3/αvβ5 engagement (Fig. 4c). Taken together, these results support the likelihood that SF1126 affects tumor growth by blocking angiogenesis, tumor cell proliferation/survival and metastasis.

In our recent phase I clinical study on RCC patients, we found that SF1126 treatment (continued at 840 mg/m2 dose for 84 weeks) prolonged clinical benefit in patient suffering from metastatic RCC resistant to mTORC1 therapy [38]. The results provided in the present study strongly validate the in vivo efficacy of this drug in clinical trials and provides a mechanism of the action of drug. In conclusion, our results demonstrate that SF1126 has potent antitumor and antiangiogenic activity in vitro and in vivo in RCC cell lines which lead us to propose that it administration in patients with a highly activated PI-3K/AKT/mTOR pathway should benefit most from this therapy.