Introduction
TS is a common autosomal-dominant disorder characterized by the development of tumors of the brain, kidney, skin and lung. The disorder is characterized by mutations or deletions in one of two large genes, hamartin (
TSC1) and tuberin (
TSC2). In addition, neuro-developmental complications of TS, including seizures and autism, lead to significant morbidity. While no strict genotype-phenotype relationship has been established, the disease is more severe in patients with tsc2 mutations. Loss of heterozygosity (LOH), which deletes the unaffected allele, is required for the development of some neoplastic features of TS, including renal angiomyolipomas, lymphangiomyomatosis, and skin lesions, while LOH is not observed in brain tubers which often cause epileptic foci [
1]. Apart from deletions, there are other mechanisms causing inactivation of tsc1 and tsc2.
The signaling pathways implicated in TS are complex. A hotspot for mutations in tsc2 involves regions implicated in controlling rheb, although multiple other signaling pathways have also been linked to TS-related neoplasia, including mTORC1, notch, p42/44 MAP kinase, NFkB, and Akt [
2‐
11]. Tuberin (
TSC2) was weak or absent in angiomyolipomas, but present in healthy kidney, whereas, phosphorylated p70 S6 kinase and pS6 were present only in angiomyolipomas. Activation of a mammalian target of rapamycin metabolic pathway in tuberous sclerosis lesions, which contributes to their growth [
2]. The upregulation of mTORC1 observed in neoplasms of patients with TS has served as a rationale for recent use of mTOR inhibitor rapamycin (sirolimus) to treat TS. Rapamycin treatment was shown to result in partial regression of kidney, lung and skin lesions but not complete disappearance of tumors. In addition, tumor re-growth was observed upon cessation of therapy, consistent with the known cytostatic rather than cytotoxic effects of rapamycin. The use of mTOR inhibitors is becoming increasingly accepted, especially for the treatment in TSC [
2,
12,
13]. These clinical findings suggest that additional signaling pathways are active in TS-related tumors [
14].
We have previously demonstrated that platelet-derived growth factor β receptor (PDGFRβ) is present and active in human and murine TS lesions [
15‐
17]. Other groups have demonstrated an inverse relationship between mTOR activation and PDGFRβ levels in TS-derived cells [
18]. Therefore, we reasoned that mTOR blockade might be compensated for by PDGF activation in vitro and
in vivo, and combined blockade might be more efficacious than rapamycin monotherapy. To test this idea, we used the FDA-approved drug imatinib (gleevec) to treat mouse tumors formed by implanted tsc2ang1 cells, isolated from a cutaneous sarcoma that arose in a tsc2 heterozygous mouse and a well-validated model of TS-related neoplasia. Imatinib functions as a specific inhibitor of a number of tyrosine kinase enzymes leading to downregulation of MMP-2 [
19]. Currently, imatinib is approved for chronic myelogenous leukemia carring the bcr-abl translocation, as well as hypereosinophilic syndrome and eosinophilic leukemia with the FIP1L1-PDGFRα fusion kinase. Solid tumors for which imatinib is approved include c-kit positive gastrointestinal stromal tumor and inoperable dermatofibrosarcoma protuberans (
http://www.gleevec.com). We demonstrate that addition of imatinib to rapamycin decreases the levels and phosphorylation of PDGFRβ. Consistent with this, combination treatment resulted in a greatly decreased phosphorylation of Akt, and Akt is a major determinant of tumorigenesis
in vivo. Finally, we show that the combination of rapamycin and imatinib has a greater antitumor effect compared to vehicle alone than either rapamycin or imatinib compared to vehicle
in vivo. These findings provide a rationale for combination therapy with rapamycin and imatinib in TS.
Discussion
TS is a multisystem disorder characterized by benign or malignant neoplasia, as well as autism and seizures. No highly effective and long lasting therapy exists for TS neoplasia. Renal lesions such as angiomyolipomas may cause massive bleeding and compromise renal function, often requiring a kidney transplant. Lymphangiomyomatosis, a neoplastic complication most commonly seen in young women, can only be cured by lung transplantation. Deforming skin lesions cause significant psychological distress [
21]. Finally, brain tumors such as subependymal giant cell astrocytomas, can cause morbidity and mortality, and multiple tubers results in refractory seizures and exacerbation of mental retardation. Thus, urgent therapy is required for TS. Here, we demonstrate in a validated preclinical model a nearly complete tumor inhibition with a combination of two FDA-approved drugs (rapamycin and imatinib) targeting two distinct signaling pathways (mTORC1 and PDGFRβ, respectively) implicated in TS-associated neoplasia.
The lack of total blockade by rapamycin is consistent with the incomplete clinical response of tumors to systemic rapamycin, implying that either additional pathways are already present in these tumors, or are induced by rapamycin monotherapy. High among the candidates are the PDGFRβ mediated signal pathways, especially since mTORC1 activation and PDGFRβ signaling have been shown to have an inverse relationship [
18].
Treatment with combination of rapamycin and imatinib led to downregulation of PDGFRβ in tsc2ang1 cells, in a c-cbl-independent manner (data not shown). In vitro proliferation assays demonstrated an additive effect of rapamycin and imatinib on tsc2ang1 cells.
Rapamycin has been used as monotherapy in patients with TS, resulting in benefit only as long as the patients are exposed to the drug [
12]. Our findings that rapamycin therapy alone does not address PDGFβ signaling, as well as modest efficacy against Akt activation are potential reasons for the failure of rapamycin as monotherapy in solid tumors. In highly malignant tumors, mTORC1 inhibition has shown to cause a paradoxical activation of Akt, in part through activation of mTORC2, which is an Akt kinase [
22,
23]. We demonstrate that the combination of rapamycin and imatinib blocks Akt activation compared with monotherapy, and likely accounts for the decrease in tumor volume. Our findings demonstrate the following. Combination therapy of TS relevant cells with rapamycin and imatinib results in downregulation of PDGFRβ and Akt signaling. This decrease is independent of c-cbl-mediated degradation. Finally, the combination of rapamycin and imatinib has at least additive activity against the tsc2ang1 model of TS
in vivo. Given that the combination of these two approved drugs is well tolerated
in vivo, and that rapamycin alone does not cause long-lasting remissions, this study provides a rationale for the combination of these two drugs in humans.
Materials and methods
Generation of murine model of TS
Tsc2ang1 (ATCC CRL 2620) is a murine cell line derived from a cutaneous sarcoma that arose in the extremity of a mouse heterozygous for tsc2; these mice develop cutaneous sarcomas at a frequency of 10 to 15%. The cells were cultured in complete DMEM medium supplemented with 10% FBS (Sigma Aldrich, St Louis, MO).
In vitro proliferation assay
10,000 tsc2ang1 cells per well were plated in 24-well dishes in triplicate. The next day, fresh medium containing the compounds or vehicle controls was added. Cells were incubated at 37° C for 24 h, and cell number was determined using a Coulter Counter (Hialeah, FL).
Western blot analysis
Lysates of Tsc2ang1 cells treated with vehicle or indicated drugs were prepared in NP-40 lysis buffer (1% NP-40, 150 mmol/L NaCl, 10% glycerol, 20 mmol/L HEPES, 1 mmol/L phenylmethylsulfonyl fluoride, 2.5 mmol/L EDTA, 100 μmol/L Na3VO4, and 1% aprotinin). Protein concentration in cleared lysates was determined using an Eppendorf BioPhotometer. Samples were resolved using SDS-PAGE (National Diagnostics) and transferred to nitrocellulose membranes. The membranes were blocked with 5% nonfat dry milk in 10 mmol/L Tris/0.1% Tween 20/100 mmol/L NaCl and incubated with primary antibodies followed by horseradish peroxidase–conjugated secondary antibody. The immunoreactive bands were visualized by enhanced chemiluminescence (Amersham Biosciences). The antibodies used were: Phospho-PDGFR- β antibody(Tyr 1021) (Cell signaling Laboratories); Akt antibody (9272) (Cell signaling Laboratories), pAKT (4058) antibody (Cell signaling Laboratories) monoclonal anti-GAPDH antibody (Santa Cruz L-18 S-48167) was used as a loading control.
The presence of MMPs in conditioned media samples was determined using MMP ELISA Quantikine Kits (R&D Systems, Inc.). Specimens, standards and reagents were prepared according to manufacturer's instructions. Protein concentration was determined via the Bradford method using bovine serum albumin as the standard as described previously [
24].
In vivo tumor growth
To test the
in vivo activity of compounds that inhibit tsc2ang1 growth in vitro
, groups of four nude mice per compound (or control). We injected 1 million Tsc2ang1 cells s.c. into 4 nude mice in each group. I.p. treatment with imatinib, rapamycin and imatinib and rapamycin were conducted for 30 days. Rapamycin and imatinib were obtained from LC Laboratories (Woburn, MA). Beginning 2 days later, the mice received daily i.p. injections of vehicle (control), rapamycin (12 mg/kg/day), imatinib (120 mg/kg/day) or imatinib plus rapamycin. The compounds were suspended in 0.1 ml of ethanol and 0.9 ml of Intralipid solution (Fresenius Kabi, Uppsala, Sweden) [
25]. No local or systemic toxicity was observed in any of the animals. Injections were given over a period of 4 weeks, after which the mice were sacrificed due to overwhelming tumor burden in the control group. Tumor volume was calculated using the equation (w
2 xL)0.52, where w(width) represents the shortest diameter of the tumor.
Statistics
One way ANOVA, and non parametric test were preformed for the tumor volume statistics. We did parametric analysis when the conditions for ANOVA were satisfied and in case where conditions are not satisfied and if the variables are not normally distributed, we conducted corresponding non parametric test and opted to present results for Wilcoxon test.
Acknowledgements
JLA was supported by the grant RO1 AR47901and P30 AR42687 Emory Skin Disease Research Core Center Grant from the National Institutes of Health, a Veterans Administration Hospital Merit Award, as well as funds from the Rabinowitch-Davis Foundation for Melanoma Research and the Betty Minsk Foundation for Melanoma Research. JLA was also funded by Robert Margolis Liposarcoma Research Fund. HB was supported by NIH grants CA87986, CA105489, CA 116552 and CA99163 and MAM and ASC were supported by NIH grant P01 CA045548.
Competing interest
United States Patent Application 20070078142 Inventor: Jack L. Arbiser, patent filed by Novartis Treatment of tuberous sclerosis associated neoplasms.
Authors’ contribution
BG,LW, ASC,MYB, and MGA performed experimental studies. SC and EV performed statistical analysis. HB, MAM and JLA designed experiments and wrote the manuscript.