Rapamycin Inhibits Anal Carcinogenesis in Two Preclinical Animal Models

Anal cancer is a disease of increasing incidence in the general population (1), and much more so among HIV-infected men who have sex with men, particularly in the era of highly effective anti-HIV therapies, which have prolonged the life of HIV-infected persons (2). Anal cancer treatment has essentially remained static over the past 2 decades, and is often associated with a high degree of morbidity. Better clinical treatments are clearly needed for anal cancer patients, especially those with more advanced stages of disease, for whom the 5-year survival rates are dismally low (1).

Like cervical cancer, the vast majority of anal cancer is etiologically associated with high-risk HPVs. As in cervical cancer, HPV16 is the most common genotype found in anal cancer, being present in 66% of these cancers (3). Of the HPV-associated cancers, however, anal cancer is one of the least well studied owing to the absence of laboratory model systems with which to pursue experiments. For example, there are no HPV-positive anal cancer cell lines yet reported in the literature. For this reason, we established 2 new preclinical animal models for human anal cancer, providing us experimental platforms for better understanding the role of HPV in anal cancer and identifying novel approaches for preventing and/or treating this debilitating disease. Our first animal model for HPV-associated anal cancer was recently described (4) and is based in the use of HPV16 transgenic mice that have been used previously to develop mouse models for HPV-associated cervical (5–16) and head/neck (17–19) cancers. In this mouse model, expression of HPV16 E6 and E7 oncogenes in the stratified epithelium of the anus synergized with the topically applied carcinogen, DMBA, to cause formation of a progressive neoplastic disease culminating in anal carcinoma. Biomarker expression (p16 and MCM7) paralleled that observed in human anal neoplastic disease (4). A second mouse model that we have developed comprises HPV16-positive human anal cancer xenografts passaged subcutaneously in immunodeficient (scid or nude) mice. This model is first described in this study. Using these 2 mouse models, we set out in this study to identify novel strategies for preventing and/or treating HPV-associated anal carcinomas. Because anal cancers in these mice arise on the exposed surfaces of the animals they can be easily monitored longitudinally, facilitating these studies. We focused our initial drug studies on rapamycin.

Rapamycin was originally isolated and identified as an antifungal agent (20), then discovered to have immunosuppressive activity (21). The molecular targets of rapamycin (TOR) were defined and the molecular pathway inhibited by rapamycin, the so-called mTOR pathway, characterized (for review, see ref. 22). Rapamycin inhibits proliferation of mammalian cells (23–25). Furthermore, the mTOR pathway that is targeted by rapamycin is induced in many cancers (26, 27) including squamous cell carcinomas of the cervix (28) and the head and neck region (29–31), both sites of HPV-associated neoplasia. Preclinical studies showed that many cell lines derived from such cancers or cancers arising in mice are also induced for the mTOR pathway and inhibited in their growth by rapamycin (for review, see ref.32) including in the case of squamous cell carcinomas of the head and neck (29, 33–35). This has led to the clinical trials evaluating the efficacy of rapamycin or like drugs that inhibit the mTOR pathway in the treatment of human cancer (for review, see ref. 36).

In this study, we determined that the mTOR pathway was active in our 2 preclinical animal models for human anal cancer. This led us to evaluate whether rapamycin could prevent the onset of and/or treat anal cancers in these preclinical models. In both models rapamycin was reproducibly found to reduce significantly or stop altogether the growth of preexisting anal tumors. Its effectiveness in preventing the onset of anal cancers was not significant; however, it did significantly prevent overall onset of tumors, which includes both benign and malignant lesions. These results provide the first preclinical evidence for the effectiveness of rapamycin in treating human anal cancer.

Generation of K14E6 and K14E7 mice has been previously described (37, 38). These mice were maintained on the inbred FVB/N genetic background. E6/E7 transgenic mice were generated by crossing K14E6 females with K14E7 males. Immunodeficient mice used for xenograft development included male and female SCID and Nude (FOXN1 mutant) mice obtained from Taconic and Harlan, respectively. All mice were kept in American Association for Accreditation of Laboratory Animal Care-approved McArdle Laboratory Cancer Center Animal Care Facility and studies with them were carried out in accordance to an approved animal protocol.

Anal carcinogenesis was achieved as previously described by weekly (20-week-treatment period) topical treatment of DMBA (dimethylbenz[a]anthracene) to the anus at a dose of 0.12 μmoles (4). The protocol was modified in that DMBA was dissolved in 60% acetone/40% DMSO (dimethylsulfoxide) versus 100% DMSO; this modification led to increased absorption of the carcinogen and increased tumor incidence compared with the original study. Mice were monitored weekly for appearance of overt tumors and change in size. Overt tumor size was measured at week 20 and the time of sacrifice (28th week) for all tumors. Tumor size was measured using calipers measuring the length and width of each tumor in mm.

All K14E6/E7 mice began rapamycin treatment 120 days prior to the calculated sacrifice date based on DMBA protocol of 20 treatments followed by an 8-week hiatus that correlated to the 11th week of the DMBA treatment regimen. No mouse had an overt tumor present at the start of rapamycin treatment. Rapamycin treatment was achieved with 7.5 mg 60-day slow release pellets (Innovative Research) implanted subcutaneously resulting in a dose of 5 mg/kg/d based upon the average mouse weight of 25 g. Six of 27 tumor-bearing mice were sacrificed early due to tumor size resulting in preventative rapamycin treatment ranging from 56 to 107 days. Matched K14E6/E7 control mice receiving no treatment also had no tumors at the 11th week of DMBA protocol. Mice were assigned to control or preventative arms at week 0 when DMBA treatments began.

All K14E6/E7 mice began rapamycin treatment 60 days prior to the calculated sacrifice date based on DMBA protocol of 20 treatments followed by an 8 week hiatus, which correlated to the 20th (final) week of DMBA treatment regimen. All mice had an overt tumor present at the start of rapamycin treatment. Rapamycin treatment was achieved with 7.5 mg 60-day slow release pellets (Innovative Research) implanted subcutaneously resulting in a dose of 5 mg/kg/d based upon the average mouse weight of 25 g. Five of 12 tumor-bearing mice were sacrificed early due to tumor size resulting in therapeutic rapamycin treatment ranging from 19 to 54 days. K14E6/E7 mice receiving no treatment that had also developed tumors at or before the 20th week of DMBA protocol were used as control mice (note this is a subset of the control mice for preventative treatment study). Mice were assigned to control or therapeutic arms at week 0 when DMBA treatments began.

After IRB approval, squamous cell anal carcinoma tissue was obtained by biopsy of the primary lesion from a HIV-seropositive male without previous clinical therapy (radiation, chemotherapy, or surgery). Within 3 hours of tissue acquisition, the tissue was minced in sterile saline and injected into the subcutaneous plane of immunodeficient mice at the shoulder and hip areas as passage zero (P0). All tumor passages were monitored for tumor incidence and timing of first and last tumor onset (Supplementary Table S1). Tumor passage was optimized at P4 by injecting 100 μL minced tumor pieces in sterile saline mixed 1:1 with reduced growth factor Matrigel (BD Matrigel Basement Membrane Matrix, Growth Factor Reduced (GFR) Catalog no. 354230).

P0 tissue at the time of harvest from immunodeficient mice was snap frozen. DNA and RNA from this tissue were isolated using Promega Wizard genomic DNA kit and Quigen RNAeasy minikit, respectively. To detect HCV, RT-PCR was conducted using 1 μg template RNA and Invitrogen Superscript III One Step RT-PCR kit. PCR conditions were 1) 50°C for 30 minutes, 2) 94°C for 4 minutes, 3) 94°C for 15 seconds, 4) 55°C for 30 seconds, 5) 68°C for 1 minute, 6) repeat steps 3 to 5 for 40 cycles, and 7) final extension at 68°C for 5 minutes. HCV primers used were 5′primer (5′AAGCGTCTAGCCATGGCG35′) and 35′primer (5′CACTCGCAAGCACCCTATCA35′) at 0.2 μmol/L for a approximetly400-bp product. To detect HIV, anal xenograft template DNA was subjected to nested PCR. Both PCR rounds (50 μL) had final concentrations of 1x PCR buffer, 0.2 mmol/L dNTPs, 1.5 mmol/L MgCl₂, and 0.02U Invitrogen Platinum Taq. In the first round of PCR, external primers were used with 100 ng template DNA and were denatured at 94°C for 2 minutes followed by 35 cycles of 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 3 minutes 30 seconds with final extension at 72°C for 10 minutes. In the second round of PCR, 5 μL of the external primer PCR product and primers to the HIV gag region were used under the same conditions except the extension step was shortened to 30 seconds from 3 minutes 30 seconds. External HIV primers used were 5′primer (5′GCGRCTGGTGAGTACGCC3′) and 35′primer (5′CACYAGCCATTGCTCTCC3′) at 0.2 μmol/L for a 3,566-bp product. Gag region HIV primers used were 5′ primer (5′GATGACAGCATGTCAGGG35′) and 35′ primer (5′RGGAAGGCCAGATYTTCC35′) at 0.2 μmol/L for a 283-bp product. To detect HPV, 100 ng anal xenograft DNA was subjected PCR using JumpstartTaq with the same final concentrations as HIV detection and were denatured at 94°C for 4 minutes followed by 40 cycles of 94°C for 15 seconds, 55°C for 30 seconds, and 68°C for 1 minute with final extension at 68°C for 5 minutes. HPV primers used were MY09 5′ primer (5′CGTCCMARRGGAWACTGATC35′) and MY11 35′primer (5′GCMCAGGGWCATAAYAATGG35′) at 0.2 μmol/L for a 448-bp product. Final products for PCR reactions were run on a 2.5% agarose gel electrophoresis at 100 V and stained with ethidium bromide. The HPV band was isolated by UV guidance, extracted using Quiagen gel extraction spin column protocol, and sequenced.

Total genomic DNA from P4 snap frozen tissue was isolated as previously described (39) such that high and low molecular weight DNAs were collected. DNA was incubated with NcoI or BamHI, at 37°C overnight and then run on a 0.8% agarose gel. Gels were denatured, neutralized, and DNA transferred to charged nylon membrane overnight. After UV cross-linking, the membrane was blocked and probed with radioactively labeled HPV16 probe.

Tumors that developed in P4 and P5 xenografts were stratified into treatment and control arms based on tumor size such that each group had tumors with a similar range in size. Rapamycin treatment was achieved in P4 xenografts with 7.5 mg 60-day slow release pellets (Innovative Research) implanted subcutaneously resulting in a dose of 4.17 mg/kg/d based upon the average mouse weight of 30 g. Rapamycin treatment was done on P5 xenografts by intraperitoneal (IP) injection of rapamycin dissolved in aqueous 5.2% polyethylene glycol and 5.2% Tween 80 for a dose of 5 mg/kg/d for 5 days followed by 2 days without treatment. Control P5 mice were injected with vehicle only. Tumor size was measured twice weekly using calipers measuring the length and width of each tumor in mm. Both P4 and P5 mice were treated for a total of 4 weeks.

The anal canal tissue was fixed in 10% formalin, paraffin embedded, and serially sectioned at 5 μm. Every 30th section was stained with H&E and histopathologically analyzed for normal/hyperplasia, papilloma, atypia, or carcinoma. Anal xenograft tumors were fixed and sectioned in the same method. The 5th section was H&E stained and confirmed to be squamous cell cancer on histopathology.

Analyses were carried out as previously described (40). Primary antibodies were applied overnight at 4°C at the following concentrations in 2.5% horse serum for human xenograft tissues (1:200 p16^INK4a, Neomarker; 1:200 K10 Sigma 8.6; 1:200 anti-MCM7, Neomarkers; 1:1,000 K14, Covance; 1:100 pERK1/2, Cell Signaling). For detecting markers of mTOR pathway in both mouse and human tissues the following concentrations in 2.5% horse serum were used: 1:100 pS6 Ser235/236 (Cell Signaling); 1:50 pAkt Ser473 (Cell Signaling).

Two-sided Wilcoxon rank-sum test was used to determine significant differences in tumor size, onset, and growth rate. Two-sided Chi-square test was used to determine differences in rates of disease in preventative and therapeutic treatment of rapamycin groups.

As described in the introduction, there is evidence, based upon biomarker expression studies, that the mTOR pathway is activated in human squamous cell carcinomas including those arising in the cervix and head/neck region wherein HPVs can be an etiologic factor. Also, Gutkind et al. have found the same pathway activated in squamous cell carcinomas arising in the oral cavity of mice. We previously generated a mouse anal cancer model in which HPV16 transgenic mice expressing the E6 and E7 oncogenes in stratified epithelium and treated topically with DMBA developed a progressive neoplastic anal disease culminating in squamous cell carcinoma (4). In this study, we asked if the mTOR pathway is active in the squamous cell carcinomas arising in this mouse model for anal cancer.

Nontransgenic and HPV16 transgenic (K14E6/E7) mice were treated topically with DMBA as described previously (4) but with a modification to the vehicle used to deliver DMBA (changed from 100% DMSO to 60%DMSO/40% acetone to improve adsorption of the carcinogen). This change led to an increased incidence of anal carcinomas both in the nontransgenic and HPV16 transgenic mice from 0% and 23%, respectively, observed in the original study (4) to 10% and 85%, respectively, in this study. This allowed us to compare the activation of the mTOR pathway in both HPV-negative and HPV-positive anal carcinomas. Both phospho-AKT (pAKT) and phopho-S6 (pS6), 2 biomarkers commonly used for measuring activation of the mTOR pathway, scored uniformly positively in the squamous cell carcinomas of the mouse anus arising in the HPV16 transgenic mice (Fig. 1A). In contrast, the squamous cell carcinomas arising in the nontransgenic mice stained poorly for pAKT and nonuniformly for pS6. In normal anal epithelium, both pAKT and pS6 were largely restricted in their detection to the more terminally differentiated compartment (Fig. 1A). These biomarkers results are consistent with the mTORC1 pathway being activated in these cancers, though at least one of these markers, pAKT, is also a marker for mTORC2 pathway. These data provided us cause to assess whether an activated mTORC1 pathway is required for the development and/or maintenance of the cancer state in this mouse model for anal carcinoma, by monitoring the effects of rapamycin in this mouse model. Because the mTOR pathway seems more uniformly activated in the anal cancers arising in the HPV16 transgenic mice (Fig. 1A), and because a majority of anal cancers in humans are HPV-positive, we focused our further analysis on anal carcinogenesis in the context of the HPV16 transgenic mice.

To assess whether inhibition of the mTOR pathway prevents tumors from arising in our mouse model for HPV-associated anal cancer, we gave K14E6/E7 mice the drug rapamycin, using continuous release pellets that delivered the drug at 5 mg/kg/d, beginning at the 11th week of the 20-week topical DMBA treatment (Fig. 1B: preventative). The 11th week of DMBA-treatment was chosen because no overt tumors had yet arisen at that time in the DMBA treatment. After 20 weeks of DMBA treatment, mice were held an additional 8 weeks, at which time tumors were harvested for histopathologic analysis. Rapamycin was given to the mice throughout the remaining time course of the experiment, that is, from week 11 to the week 28 endpoint. Time of onset of overt tumors, tumor growth rates, tumor size, and tumor grade were compared between the cohort of mice given rapamycin to a control cohort not given the drug (Table 1). Rapamycin treated K14E6/E7 mice had significantly decreased overt tumor incidence (overt tumors include papillomas, atypia, and carcinomas) compared with control K14E6/E7 mice (73% vs. 100%, P = 0.01); though, on histologic analysis, the difference in carcinoma incidence (70% in rapamycin-treated vs. 85% in control) was not significant (P = 0.22) (Table 1). Although the time of tumor onset was only slightly delayed in the rapamycin-treated cohort, the tumors that did develop in the rapamycin-treated cohort were smaller both at the 20th week of DMBA-treatment and at the 28-week endpoint. These decreases in tumor size translated to a nearly 3-fold reduction in tumor growth rate in the rapamycin-treated mice (Table 1, Fig. 1C), from on average 1.5 mm/week in the control group down to 0.6 mm/week in the rapamycin-treated group. This 2.5-fold decrease in tumor growth rate approached the 95% confidence limit of significance (P = 0.08).

To assess the effect of rapamycin on already established tumors, K14E6/E7 mice treated with DMBA for 20 weeks that had all developed tumors by that time point, were treated with rapamycin, starting at the last week of DMBA treatment, for the remaining 8 weeks, until the 28-week endpoint of the study (Fig. 1B: therapeutic), at which time tumors were harvested. These mice were compared with a cohort of DMBA-treated mice not exposed to rapamycin, who also bore tumors at the 20-week-time point. There was no significant difference in the histologic grade of tumor in the 2 groups of mice at the 28-week endpoint (Table 2). However, tumor size at endpoint was significantly lower in the rapamycin-treated K14E6/E7 mice (Table 2) translating to an approximate 3-fold decreased tumor growth rate compared with the control mice (Table 2, Fig. 1D), from on average 1.7 mm/week in the control mice down to 0.5 mm/week in the rapamycin-treated mice. This 3.4-fold difference was significant (P = 0.01) and was similar to what was observed in the mice treated with rapamycin in the prevention trial (Table 1, Fig. 1C). Together, the results of these prevention and therapeutic trials in DMBA-treated K14E6/E7 mice indicate that rapamycin could reduce growth of mouse anal tumors.

Given the data from our mouse anal cancer study, we wanted to learn whether the mTOR pathway is activated in human anal cancers. With IRB approval, we identified 10 patients who were treated for anal cancers at the University of Wisconsin Hospitals for which the presence of squamous cell carcinoma could be confirmed histopathologically in paraffin-embedded biopsies taken prior to chemoradiation. These were subjected to phospho-S6 specific immunohistochemistry. All the cancers showed staining for pS6 throughout the malignant lesions (shown in Fig. 2A–C are 3 representative cancers). In contrast, in nearby normal anal epithelium, cells staining positively for phospho-S6 were limited to the more superficial portion of the spinous layer with some weak, occasional staining in the basal/parabasal compartment (Fig. 2D). These findings provided evidence that in human anal cancers, the mTOR pathway may be activated.

To address the potential effects of rapamycin on human anal cancer, we established anal cancer xenografts from human squamous cell anal cancer tissue. A biopsy from a patient presenting with anal cancer was acquired under IRB approval from the University of Wisconsin Hospital, transferred to McArdle Laboratory for Cancer Research on campus and injected subcutaneously into the flanks of SCID mice. The cancer grew as tumor xenografts at multiple sites and could be passaged multiple times (to date it has been passaged 6 times). The efficiency of passaging (i.e., time of initial onset of tumors and efficiency of tumor growth) was positively affected by mixing the tumor with Matrigel and it could grow in both SCID and Nude mouse strains (Supplementary Table S1). Histologically, the tumor xenograft retained over passaging the histopathologic characteristics of being a poorly differentiated squamous carcinoma (Fig. 3A). It displayed strong positive staining for MCM7, and p16 (Fig. 3B). It stained positively for Keratin 14 but not Keratin 10, consistent with its poorly differentiated characteristics (data not shown). To assess whether the mTOR pathway was activated, the xenograft was scored for pAKT and pS6 and found to be positive for both (Fig. 3B). The P0 xenograft was initially determined to be HPV16-positive (data not shown), based upon sequence analysis of a PCR amplimer generated in a PCR reaction on the xenograft-derived DNA using the MY9/11 primer pair (see methods). This primer pair is routinely used to detect common anogenital HPV genotypes in clinical samples. Subsequent passages of the xenograft retained HPV positive status by PCR. Southern analysis of P4 xenograft derived total genomic DNA confirmed that the cancer is HPV16 positive, with a hybridization pattern consistent with the HPV16 DNA being integrated (Fig. 3C). The P0 xenograft was also screened by PCR and found to be negative for HIV and HCV (data not shown). The patient was previously screened to be positive for both of these pathogens.

To assess the effect of inhibiting the mTOR pathway on growth of human anal cancers, mice- bearing tumors that developed from P4 anal cancer xenografts were placed into control or rapamycin-treatment groups such that the average tumor size at the outset of rapamycin treatment was not significantly difference between the 2 groups. In the initial experiment mice were given rapamycin using slow release pellets, providing a dose of 4.17 mg/kg/d. Tumor size was measured using calipers weekly for 4 weeks. The tumors in the mice treated with rapamycin had lower tumor growth rates (avg. = 0.24 ±0.3 mm/week) compared with that in the control group (avg. = 0.74 mm/week) (Table 3, Fig. 3D: left graph). This difference was highly significant (P < 0.01).

Prior studies used in mice with oral cancers (34) had used a slightly higher dose of rapamycin 5 mg/kg/d delivered IP. We therefore repeated the experiment in passage 5 tumor xenografts using this higher dose and alternative delivery method. In this experiment, the tumors were more completely inhibited in their growth (Table 3, Fig. 3D: right graph), showing on average a slight negative growth (−0.02 mm/week) during the 4-week treatment period compared with the tumors in the paired untreated group (0.74 mm/week). This complete block in tumor growth was also highly significant (P < 0.01). These results show that an HPV16-positive human anal cancer, carried as a xenograft in mice, is inhibited significantly in its growth by rapamycin.

Prior studies have shown that the MAPK pathway can be activated by rapamycin treatment of tumors (41). As such activation of the MAPK pathway provides a useful readout for the rapamycin activity. To learn if the MAPK pathway was activated in the human xenografts treated with rapamycin, we monitored levels of phosphorylated ERK (pERK) by immunoblot analysis. We found a strong induction of pERK in the human anal cancer xenografts treated with rapamycin (Fig. 4).

In this study, we show that both mouse and human anal tumors are activated in the mTOR pathway and are significantly reduced in their growth rates upon administration of rapamycin. These results provide preclinical data to support the use of mTOR inhibitors in the therapeutic treatment of human anal cancers. Our prevention trial in the context of the mouse anal cancer model did not show efficacy at the level of preventing frank cancer, though overall tumor incidence and growth was reduced by early onset treatment with rapamycin prior to onset of overt tumors.

The utility of mTOR inhibitors in the clinic has been found to be maximal when these inhibitors are combined with other anticancer treatments. This may reflect the fact that, although mTOR inhibitors can induce cell death, they primarily act as cytostatic agents (42, 43). In addition, rapamycin has been observed to induce MAPK in some cancers, and this can negatively effect its therapeutic effect as evidenced by the fact that drugs that inhibit MAPK can synergize with rapamycin to cause further reduction in tumor growth/persistence (41). Consistent with these prior observations, we observed induction of activated ERK (Fig. 4). Rapamycin also has been found to act as a radiation-sensitizer in many cancers (44–47). These studies with other cancer types and our own observations (Fig. 4) point to the value of pursuing further studies using our preclinical animal models for anal cancer to identify combinatorial therapeutic regimens including the use of mTOR inhibitors that could have greater effect in treating human anal cancer with reduced morbidity associated with standard of care treatments available today.

For our therapeutic trials on the human anal cancer xenograft, we used 2 modes of rapamycin delivery, one using a slow release pellet that, based upon the weight of the recipient animals, should have delivered approximately 4.2 mg/kg/d or rapamycin, the other mode was 5x weekly IP injections at a dose of 5 mg/kg/treatment. Although both modes of delivery were effective in decreasing the growth rates of the xenografts, clearly the latter was more effective. Whether this reflects differences in maximum and/or trough levels of drug is unclear; however, the latter clearly has importance in other mouse models (48).

In the context of these studies, we developed a new human, HPV-positive, anal cancer xenograft model. This animal model provides a powerful laboratory experimental system in which to assess the role of HPV and cellular genes/pathways in this poorly studied human cancer. As such it complements the utility of our previously described mouse anal cancer model for HPV-associated human anal cancer. That both the mouse anal cancer and human anal cancer mouse models respond similarly to rapamycin provides further validation of the former model.

No potential conflicts of interest were disclosed.

We thank Dawn Dudley of Dr. David O'Connor's Lab and Israr Ansari of Dr. Rob Striker's lab, both at UW-Madison, for help with PCR screening for HIV and HCV, respectively.

Grant Support

This study was supported by NIH grants (R01 DE017315, U01 CA141583) funded to P.F. Lambert. M.K. Stelzer was supported by a NIH training grant (NIH/NCI T32 CA090217) for clinical fellows.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.