Sirolimus Therapy for Angiomyolipoma in Tuberous Sclerosis and Sporadic Lymphangioleiomyomatosis: A Phase 2 Trial

Tuberous sclerosis is an autosomal dominant disorder caused by mutations in either TSC1 or TSC2 (1). It is characterized by development of tumors in many organs, including the kidneys and brain, and by a range of neuropsychiatric manifestations that include seizures, cognitive impairments, and autism (1). Renal angiomyolipomas affect up to 80% of patients causing morbidity and mortality because of hemorrhage and renal insufficiency. They do not regress spontaneously and current treatments include embolization and nephrectomy (2, 3). Lymphangioleiomyomatosis (LAM) is the main pulmonary manifestation and occurs almost exclusively in females. It can lead to respiratory failure as a consequence of proliferation of smooth muscle–like cells (LAM cells) and cystic degeneration and may require lung transplantation (4, 5). LAM also occurs as a rare sporadic disorder associated with acquired mutations of TSC2 in women without tuberous sclerosis, 40% of whom also have renal angiomyolipomas. Identical TSC2 mutations have been identified in their angiomyolipomas and pulmonary LAM cells indicating a clonal origin and a metastasis-like disease process (6). The neuropsychiatric manifestations of tuberous sclerosis are ranked by families and caregivers among the most significant problems (7). Forty percent to 50% of patients have intellectual disability, but even those with normal intellectual abilities may have specific neurocognitive problems, including deficits in executive function and memory (8, 9).

Mutation of TSC1 or TSC2 leads to overactivation of signaling via the mTOR complex 1 (mTORC1), a regulator of cell growth (10). Angiomyolipomas exhibit biallelic TSC1 or TSC2 mutation and mTORC1 pathway activation (2). In the central nervous system, mTORC1 regulates neuronal differentiation, synaptic plasticity, and the encoding of spatial and auditory memory (11, 12). Sirolimus (rapamycin), a potent mTORC1 inhibitor, reversed renal tumor growth, abnormal synaptic plasticity, and deficits in spatial recall memory in rodent models of tuberous sclerosis (13–15). Clinical studies indicate that sirolimus also reduces the size of kidney and brain tumors in patients with tuberous sclerosis (16–18). In a phase 1 to 2 trial in 20 patients with tuberous sclerosis or sporadic LAM, renal angiomyolipoma volumes were reduced by approximately 50% after 12 months of sirolimus treatment but the tumors increased in size when treatment was stopped (19). In an interim report of the current trial, we also showed size reduction of angiomyolipomas in patients who had been treated for up to 12 months (20). We have now evaluated the longer term efficacy and safety of sirolimus treatment for angiomyolipoma in adults with tuberous sclerosis or sporadic LAM. We also monitored aspects of neurocognitive function and, in those with LAM, assessed changes in lung function.

Patients with a definite diagnosis of tuberous sclerosis (21) or sporadic LAM (22) were eligible if they were 18 to 65 years of age, had at least 1 renal angiomyolipoma of 2 cm or more in longest diameter, and an estimated glomerular filtration rate (GFR; ref. 23) of at least 40 mL/min. The major exclusion criteria were angiomyolipoma embolization within 6 months or bleeding within 12 months, urinary protein excretion of more than 1 g/d, inability to walk 100 meters on the flat, requirement for continuous supplemental oxygen, or an intelligence quotient (IQ) of less than 70.

This trial was a prospective multicenter, phase 2 study (ClinicalTrials.gov number, NCT00490789) conducted at the University Hospital of Wales (United Kingdom), Nottingham City Hospital (United Kingdom), Royal Sussex Hospital (United Kingdom), and University Hospital, Zurich (Switzerland) from October 2005 through September 2009 in accordance with the United Kingdom Medicines for Human Use Regulations 2004, the Declaration of Helsinki, the Good Clinical Practice guidelines of the International Conference of Harmonization, and local regulatory requirements. The Thames Valley Multi-Centre Research Ethics Committee approved the protocol, and all patients gave written informed consent.

At baseline, angiomyolipomas were visualized by abdominal MRI without contrast media and measured. Up to 5 angiomyolipomas with longest diameters of 2.0 cm or more were selected in each kidney as target lesions and their longest diameters were summated for each patient. Angiomyolipomas with a longest diameter of less than 2.0 cm were recorded as nontarget lesions. Oral liquid sirolimus was initiated at a daily dose of 0.5 mg/m² body-surface-area, and the dosage was adjusted to achieve steady-state levels between 3 and 6 ng/mL. Follow-up visits were at 3 weeks and at 2, 4, 6, 9, 12, 18, and 24 months after initiation, and abdominal MRI was repeated at 2, 6, 12, and 24 months. If the longest diameter of all target lesions was not reduced by at least 10% at 2 months, the sirolimus dose was increased to achieve a steady-state level of 6 to 10 ng/mL.

At each visit, blood creatinine, lipids, liver enzymes, and hematologic parameters were measured, and protein and creatinine were measured in spot urine. GFR was estimated with the MDRD equation (23). Steady-state levels of sirolimus were determined by liquid chromatography–mass spectrometry from whole blood. Adverse events were classified according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 3.0.

Transaxial unenhanced abdominal MRI scans were done with 1.5 Tesla systems. Scans were acquired with T1-weighted fast spoiled gradient echo and T2-weighted fast spin echo protocols and reviewed by a single radiologist (T.D.). Cranial MRI was undertaken at baseline and, in patients with tuberous sclerosis, at 12 months by using 1.5 Tesla systems. Axial T2-weighted images, axial FLAIR images, and coronal 3D SPGR were obtained and reviewed by a single neuroradiologist (J.J.C). Computed tomography (CT) of the lungs was done at baseline and at 24 months in patients with LAM. Images were obtained during full inspiration by using a low dose protocol and reviewed by a single radiologist (K.P.).

Forced expiratory volume in 1 second (FEV₁), forced vital capacity (FVC), and diffusing capacity of the lung for carbon monoxide (DL_co) were measured at baseline and at 4, 6, 12, and 24 months in a lung function laboratory in all patients with LAM and expressed as percentage of predicted values (24) to allow for change over time.

Neurocognitive tests were undertaken at baseline, 4 and 12 months in patients with English as their first language. IQ was determined for eligibility by using the National Adult Reading Test (25). Immediate recall memory was tested by using the list learning, story, and complex figure tests from the Adult Memory and Information Processing Battery (26); immediate recognition memory by using the spatial recognition memory and pattern recognition memory subtests from the Cambridge Neuropsychological Test Automated Batteries (CANTAB; ref. 27), and executive functions by using the self-ordered spatial working memory task (SWM), the stockings of Cambridge (SoC) planning task, and the intra-dimensional extra-dimensional (IDED) set-shifting task, from CANTAB (27).

The primary outcome was the number of patients in whom renal angiomyolipomas responded to sirolimus therapy. Response was evaluated by the Response Evaluation Criteria in Solid Tumors (RECIST; ref. 28) that define a “complete response” as the disappearance of all lesions, a “partial response” as decrease of at least 30% in the sum of the longest diameters of target lesions without progression of nontarget lesions, “progressive disease” as a 20% increase in the sum of the longest diameters of target lesions compared with the smallest value recorded since treatment started or progression of nontarget lesions or appearance of 1 or more new lesions, and “stable disease” as neither response nor progression. In this trial, a response was defined as a complete or partial response that occurred at any time during the trial.

We aimed initially to recruit 15 patients with sporadic LAM and 15 with tuberous sclerosis but because of slow recruitment we modified our plan to recruiting at least 14 patients in total. The minimum sample size of 14 was based on Fleming's single stage design (29). p1 (the response rate below which the treatment would not be studied further) was set at 0.1 and p2 (the response rate above which the treatment would be studied further) at 0.4. We set α (the probability of concluding that the response rate is greater than p1 when that is false) at 0.05 and β (the probability of concluding that the response rate is less than p2 when that is false) at 0.1. The primary analysis was based on all patients.

The per protocol population included all patients in the study apart from those who received less than 4 months treatment at the higher sirolimus blood level of 6 to 10 ng/mL, unless a response had been achieved.

Annual rate of change in FEV₁, FVC, and DL_co was determined by linear regression.

For neurocognitive outcomes, we used published UK percentile bands for age and sex for each measure to derive summary scores (30) for each neurocognitive domain, (recall memory, recognition memory, and executive function).

Sixteen patients, 13 female and 3 male, were enrolled. Ten had tuberous sclerosis (including 3 with LAM) and 6 had sporadic LAM with no skin signs of tuberous sclerosis or manifestations of tuberous sclerosis on brain MRI scan. Four had undergone previous unilateral nephrectomy. Among the 12 patients with 2 kidneys the target angiomyolipomas (i.e., those of 2 cm or more in longest diameter) were bilateral in 6. During the trial, 1 patient with severe sporadic LAM died from a respiratory infection, 1 withdrew for lung transplantation, 1 for elective surgery for angiomyolipoma, 2 because of protocol deviations, and 1 because of likely sirolimus-related peripheral edema. Ten patients completed the 24-month trial (Fig. 1). At 2 months, 4 patients were dose escalated to steady-state blood levels of 6 to 10 ng/mL. Others were maintained at 3 to 6 ng/mL without further dose escalation.

The overall response rate by the RECIST criteria (28) was 50% (8 of 16) and in the per protocol group it was 80% (8 of 10). All were partial responses that occurred in 4 of 10 patients with tuberous sclerosis and 4 of 6 with sporadic LAM. At 24 months, a partial response was present in 4 of 10 patients (40%) remaining in the trial. Moreover, in every patient the sum of the longest diameters of target angiomyolipomas was reduced throughout the trial compared with baseline (Fig. 2 and Supplementary Table S1). Of 23 target angiomyolipomas evaluated at 24 months, 21 were smaller and 2 were unchanged (Table 1). Of all 48 target angiomyolipomas evaluated during the trial, 41 (85%) were smaller at the last measurement than at baseline, 2 were unchanged, and 5 were larger (including 4 from 1 patient, TSC4, and a 1-mm increase in the fifth; Table 1). Angiomyolipomas shrank most quickly early in the trial (Fig. 2). For angiomyolipomas measured at baseline and at 12 and 24 months, there was a mean reduction in longest diameter of 7.3 mm at 12 months but a further mean reduction of only 0.7 mm at 24 months (Table 1).

Lung function in patients with sporadic and tuberous sclerosis-associated LAM showed wide variation at baseline and fell slightly in most patients during the trial (Fig. 3 and Supplementary Table S2). For the 5 patients with measurements over 2 years the mean ± SD FEV₁ declined by 76 ± 52 mL/y, mean FVC by 55 ± 94 mL/y and mean DL_co by −0.49 ± 0.55 mL/min/mmHg/y.

Four patients (LAM1, LAM2, LAM4, and TSC1) had serial measurements of FEV₁ before enrollment for 16, 132, 42, and 106 months, respectively. Mean annual changes in FEV₁ for these patients before and during the trial were −172 and −94, −122 and −89, −90 and +10, and −49 and −69 mL/yr, respectively. There were no clinically relevant changes in lung CT appearances during the trial.

The mean ± SD IQ of patients with tuberous sclerosis was 107 ± 12 and of sporadic LAM patients 105 ± 15. Four patients with tuberous sclerosis were taking antiepileptic drugs during the trial, but none reported seizures during the trial or in the preceding year. Baseline cranial MRI scans in patients with tuberous sclerosis showed cortical tubers and subependymal nodules, but no subependymal giant cell astrocytomas and no changes were seen at 12 months.

Increased immediate recall memory scores were seen in 7 of 8 tested patients with tuberous sclerosis. By contrast, immediate recognition memory scores fell in 5 and none showed an increase. Executive function scores increased in 5 of 8 patients with tuberous sclerosis. Changes in neurocognitive scores were also observed in patients with sporadic LAM (Fig. 4 and Supplementary Table S3).

The most common adverse events were oral mucositis (6 of 16 patients), respiratory infections (5 patients) and proteinuria (5 patients; Table 2). Seven serious adverse events occurred of which 3 were possibly related to sirolimus. One patient with severe sporadic LAM died following a respiratory infection and 2 further patients with LAM were hospitalized, 1 with pharyngitis and 1 with a chest infection. The 4 remaining serious adverse events were categorized as not related or unlikely to be related to sirolimus: a fractured tibia and fibula, chest pain of unknown cause, urinary obstruction secondary to an ovarian cyst, and musculoskeletal back pain.

This trial determined the response of renal angiomyolipomas in patients with tuberous sclerosis or sporadic LAM to 2 years of sirolimus treatment. A tumor response by RECIST criteria (28) was observed in 8 of 16 patients overall and 8 of 10 in the per protocol group. Of 23 angiomyolipomas evaluated at 24 months, 21 were smaller and 2 were unchanged. Although all target angiomyolipomas shrank in most patients, one (TSC4) who had the most target lesions and who withdrew at 12 months, showed heterogeneity of tumor response, with 5 of 9 angiomyolipomas shrinking and 4 growing. One previous trial reported by Bissler and colleagues investigated sirolimus treatment for angiomyolipomas in patients with tuberous sclerosis or sporadic LAM and found a mean reduction in angiomyolipoma volume of 47% at 12 months (19). In the current trial, angiomyolipomas measured at 12 months showed a mean reduction in their longest diameters of 25% compared with baseline, equivalent to a volume reduction of 60% if a spherical shape is assumed. Although 12 of 16 patients in the current trial were maintained at trough blood levels of 3 to 6 ng/mL and the other 4 at 6 to 10 ng/mL, all but one of the patients in the trial by Bissler and colleagues were escalated to trough blood levels of 10 to 15 ng/mL. The current trial shows that sirolimus levels at the bottom end of the immunosuppressive range are effective in reducing angiomyolipoma size in tuberous sclerosis or sporadic LAM.

In the trial reported by Bissler and colleagues, angiomyolipoma volumes increased after sirolimus therapy was withdrawn at 12 months, returning from 53% to 86% of the baseline value by 24 months (19). In the current trial, the mean longest diameter of angiomyolipomas measured at both 12 and 24 months was 2.19 and 2.11 cm, respectively, indicating that tumor response was maintained by continuation of therapy, but little further shrinkage occurred during the second year of treatment. Long-term sirolimus therapy may be needed to prevent tumor regrowth in patients with tuberous sclerosis or LAM and this could be acceptable, as it is in organ transplant recipients.

In recent studies of LAM, the mean rate of decline in FEV₁ has ranged from 75 to 118 mL/y (31–33). During this trial, it was 76 mL/yr in the 5 patients who had measurements more than 2 years and 49 mL/y in 7 patients with measurements over 1 year or more. We did not observe a significant improvement in lung function during sirolimus therapy as was reported by Bissler and colleagues (19) but neither study was powered to evaluate lung function. A recently reported randomized control trial of sirolimus for the treatment of LAM demonstrated that sirolimus may prevent or reduce decline in lung function, consistent with our findings (34).

The reversal of spatial learning deficits by rapamycin (sirolimus) treatment in heterozygous Tsc2 mice (15) has suggested that mTORC1 inhibitors might also improve specific neurocognitive problems associated with tuberous sclerosis (35). The nonrandomized design and small patient numbers restricted the analysis of neurocognitive outcomes in this trial, and the patients with tuberous sclerosis were mildly affected with far fewer neurocognitive problems than more typical patients. Nevertheless, scores for recall memory that has been associated with mTOR activity (15) increased in 7 of 8 patients with tuberous sclerosis whereas those for recognition memory did not. Larger randomized control trials are warranted to determine whether treatment-related changes in neurocognition occur with mTORC1 inhibition in patients with or without tuberous sclerosis.

Adverse events were common, consistent with the known toxicities of sirolimus and mostly of low grade. However, 1 patient with severe LAM died of a respiratory infection during this study. Because of their immunosuppressive properties, the risks and benefits of mTORC1 inhibitors need specific and careful evaluation in this setting.

Although our trial involved a small number of patients and was noncontrolled and open label, it reports one of the first examples of therapeutic targeting of tumors in the context of both a Mendelian and a sporadic disorder by inhibition of a shared signaling pathway. The high response rate seen in the trial underscores the potential for effective targeted treatment when the setting is one of relative molecular homogeneity.

Recruitment to the trial proved difficult, reflecting the rarity of these disorders and our restrictive inclusion criteria, particularly for patients with tuberous sclerosis. Nonetheless, it provides evidence that renal angiomyolipomas in patients with tuberous sclerosis or sporadic LAM decrease in size in response to treatment with sirolimus and that this response is maintained by continuation of treatment for 2 years. Larger trials should now address whether in patients with large tumors and/or extensive renal involvement, mTORC1 inhibition leads to reduction of the clinically important complications of renal angiomyolipoma, notably hemorrhage and renal failure.

J.R. Sampson and A.L. Serra received fees to attend a Wyeth Medical Advisory Board meeting. All other authors declared no conflicts of interest.

J.R. Sampson and D.M. Davies had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors contributed to writing this report and vouch for its fidelity to the trial protocol and statistical analysis plan.

We thank A. Hunt, F. Dunstan, K. Hood, and S. Lewis for planning and statistical advice and colleagues at the study centers for pulmonary function testing. Wyeth provided sirolimus for this trial but did not provide financial support.

This work was supported by the Tuberous Sclerosis Association (to J.R. Sampson and P.J. de Vries), the James Tudor Foundation, and the Wales Gene Park.

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.