TOPICAL TREATMENT OF BASAL CELL CARCINOMAS IN NEVOID BASAL CELL CARCINOMA SYNDROME WITH A SMOOTHENED INHIBITOR

Hans Skvara1, Frank Kalthoff2, Josef G. Meingassner2, Barbara Wolff-Winiski2, Heinrich Aschauer2, Joseph F. Kelleher3, XuWu4, Shifeng Pan4, Lesanka Mickel1, Christopher Schuster1, Georg Stary1, Ahmad Jalili1, Olivier J David5, Corinne Emotte5, Ana Monica Costa Antunes6, Kristine Rose7, Jeremy Decker3, Ilene Carlson8, Humphrey Gardner8, Anton Stuetz2, Arthur P. Bertolino6, Georg Stingl1 and Menno A. De Rie6,9

Basal cell carcinoma (BCC) is a distinctive manifestation in nevoid basal cell carcinoma syndrome (NBCCS) patients. Both inherited and acquired mutations of patched 1 (PTCH1), a tumor-suppressor gene controlling the activity of Smoothened (SMO), are the primary cause of the constitutive activation of the Hedgehog (HH) pathway, leading to the emergence of BCCs in NBCCS. LDE225, a distinct, selective antagonist of SMO, showed potent inhibition of basaloid tumor nest formation and mediated regression of preformed basaloid tumors in organ cultures of skin derived from Ptch1 heterozygous knockout mice. In a double-blind, randomized, vehiclecontrolled, intraindividual study, a total of 8 NBCCS patients presenting 27 BCCs were treated twice daily with 0.75% LDE225 cream or vehicle for 4 weeks. Application of 0.75% LDE225 cream was well tolerated and showed no skin irritation. Of 13 LDE225-treated BCCs, 3 showed a complete, 9 a partial, and only 1 no clinical response. Except for one partial response, the vehicle produced no clinical response in any of the 14 treated BCCs. Treatment with 0.75% LDE225 cream in NBCCS patients was very well tolerated and caused BCC regression, thus potentially offering an attractive therapeutic alternative to currently available therapies for this indication.

Journal of Investigative Dermatology advance online publication, 24 March 2011; doi:10.1038/jid.2011.48

1Division of Immunology, Department of Dermatology, Allergy and Infectious Diseases, Medical University of Vienna, Vienna, Austria; 2Novartis Institutes for BioMedical Research, Dermatology/ATI, Vienna, Austria; 3Novartis
Institutes for BioMedical Research, Cambridge, Massachusetts, USA; 4Genomics Institute of Novartis Foundation, San Diego, California, USA; 5Novartis Institutes for BioMedical Research, DMPK, Basel, Switzerland; 6Novartis Institutes for BioMedical Research, Translational Medicine, Basel, Switzerland; 7Novartis Pharmaceuticals Corporation, Oncology Translational Medicine, East Hanover, New Jersey, USA and 8Novartis Institutes for Biomedical Research, Oncology Biomarkers and Imaging, Cambridge, Massachusetts, USA Correspondence: Menno A. De Rie, Erasmus Medical Center, Department of Dermatology, Postbus 2040, 3000 CA Rotterdam, The Netherlands.  E-mail:menno_derie@hotmail.com 9Current address: Erasmus Medical Center, Department of Dermatology, Postbus 2040, 3000 CA Rotterdam, The Netherlands. Abbreviations: BCC, basal cell carcinoma; 3D, three-dimensional; HH, hedgehog; IC50, half-maximal inhibitory concentration; NBCCS, nevoid basal cell carcinoma syndrome; PTCH1, patched 1; SMO, smoothened
Received 4 September 2010; revised 24 November 2010; accepted 17 December 2010

INTRODUCTION

Nevoid basal cell carcinoma syndrome (NBCCS), also known as Gorlin–Goltz syndrome, Gorlin syndrome, or basal cell nevus syndrome, is a rare autosomal dominant disorder of high penetrance but variable phenotype expression (Wicking et al., 1997; Boutet et al., 2003). NBCCS has a minimum prevalence of 1:57,000 and is characterized by multiple developmental defects and cancer susceptibility, particularly to basal cell carcinomas (BCCs) and several other malignant or benign tumors including medulloblastoma as well as ovarian and cardiac fibroma (Shafei-Benaissa et al., 1998; Lo Muzio, 2008). The most frequent developmental abnormalities include dyskeratotic palmar and plantar pits, odontogenic keratocysts of the jaw, ectopic calcifications, spine and rib malformations, macrocephaly, and generalized overgrowth (Kimonis et al., 1997).

NBCCS is usually caused by mutations in the patched 1 (PTCH1) gene. This gene has been mapped to the long arm of chromosome 9 (q22.3) and encodes the transmembrane protein PTCH1, the primary receptor for ligands of the Hedgehog (HH) signaling pathway (Farndon et al., 1992; Epstein, 2008). HH signaling regulates a variety of processes during embryonic development and adult tissue homeostasis, and its uncontrolled activation is known to be involved in cancer formation (Jiang and Hui, 2008). Under physiological conditions this pathway is actively repressed by PTCH1, which inhibits Smoothened (SMO)—a seven-transmembrane protein and the key activator of downstream signals. Binding of HH ligands to PTCH1 blocks its repression toward SMO, thereby activating the HH signaling cascade. Consequently, Gli transcription factors get stabilized and translocate to the nucleus to induce gene transcription and thus control proliferation, survival, and differentiation of HH-responsive cells (Rohatgi and Scott, 2007; Ruiz i Altaba et al., 2007). Emergence of BCCs in NBCCS patients in general is because of somatic inactivation of the residual normal PTCH1 allele, resulting in constitutive activation of SMO and consequent uncontrolled proliferation and blockage of cellular differentiation (Laimer et al., 2008). Thus, inhibition of the HH pathway seems to be beneficial for patients suffering from this disease. 

Successful treatment of locally advanced or metastatic BCCs in humans has been proven recently after oral administration of a specific SMO inhibitor, referred to as GDC0449, which showed an overall response rate of 55% (Von Hoff et al., 2009). Additionally, oral administration of GDC0449 quite impressively mediated tumor regression in a patient with metastatic medulloblastoma and significantly reduced the BCC burden in a patient with NBCCS (Rudin et al., 2009; Goldberg et al., 2010). We have developed LDE225, a distinct and selective antagonist of human and murine SMO protein with single-digit nanomolar potency in SMO binding and cellular assays (Pan et al., 2010b), for the topical application in BCC. In this study, we will provide evidence that this SMO inhibitor, even when applied topically, can successfully interfere with the HH-dependent signaling pathway in mice and humans.

RESULTS

LDE225 induces regression of murine basaloid tumor nests 
Skin punches derived from Ptch1þ/_LacZ heterozygous mice were cultured in the presence of the small-molecule SMO agonist HH Ag-1.3 (Frank-Kamenetsky et al., 2002). Because Ptch1 is a target gene of HH signaling, the Ptch1-LacZ transgene allows for monitoring of the pathway activation by assaying b-D-galactosidase expression. LDE225 reduced X-gal staining of skin explants more profoundly than cyclopamine when added to SMO-agonist-induced cultures (Figure 1a). It has previously been shown that basaloid tumor nests form when whole-skin biopsies are cultured in the presence of recombinant Sonic HH protein or a smallmolecule SMO agonist (Williams et al., 2003). LDE225 efficiently blocked basaloid tumor nest formation in skin biopsies cultured in the presence of HH Ag-1.3 with the half-maximal inhibitory concentration (IC50) of p150 nM, whereas cyclopamine showed only weak antagonistic activity with an IC50 of B10 mM (data not shown). In a regression assay, basaloid tumor nests were allowed to form by stimulation with HH Ag-1.3 for 7 days. Subsequently, LDE225 or the reference SMO inhibitor cyclopamine was added and cultures maintained for another 8 days in the presence of HH Ag-1.3 agonist. Preformed basaloid tumor nests were induced to regress by the presence of LDE225 (but not cyclopamine) with an IC50 of p150 nM (Figure 1b and c). 

LDE225 inhibits HH gene expression and prevents hair growth in mice 
Activation of the HH signaling pathway plays an important role in hair follicle morphogenesis during the early anagen phase of the hair growth cycle (Oro and Higgins, 2003). Therefore, hair growth inhibition could serve as a pharmacodynamic read-out for SMO antagonists. Photographs taken on days 15 and 22 (e.g., 1 day and 8 days after cessation of treatment) have shown complete inhibition of anagen progression and hair growth in all LDE225-treated animals in contrast to vehicle controls (Figure 2). In the LDE225- treated group, skin pigmentation as first sign of anagen commenced on day 22, and hair growth progressed normally thereafter. Topical application of LD225 as 1% solution inhibited expression of Gli1, Gli2, and Sox9 transcripts almost completely, and reduced sFRP1 (secreted frizzledrelated protein 1) mRNA levels by 80%. Treatment with 0.3% LDE225 inhibited Gli1, Gli2, and Sox9 expression by B80, 50, and 40%, respectively, whereas sFRP1 was not affected (Supplementary Figure S1 online). 

Skin penetration/permeation studies ex vivo and in vivo 
Ex vivo, a 0.3% LDE225 solution in propylene-glycol penetrated well into the human (27±8 mgg–1) and pig (11±3 mgg–1) skin, whereas the permeation rate through human and pig skin was as low as 6 ng cm–2 h–1. In vivo, topical application of the 0.75% LDE225 cream on the dorsolateral trunk of pigs resulted in dermal concentrations of 1–1.5 mg g–1 at time points between 1 and 8 hours. Upon one application of LDE225 to 10% body surface of pigs, dermal concentrations of 1 mgg–1 and blood levels p0.15 ng ml–1 were observed after 24 hours (data not shown). 

Clinical and histopathological results after topical application of 0.75% LDE225 cream in NBCCS patients 
A total of 8 patients (male: 4, 50%) with NBCCS were enrolled into the study. Mean (range) age of all patients was 49.5 years (27–66), weight was 91.3 kg (69.4–110.7), and height was 180.9 cm (164–196). Figure 3 represents the flow of patients through the study. A total of 27 BCCs were treated twice daily with 0.75% LDE225 cream or vehicle for 4 weeks. Every patient received the vehicle and the active compound. In all, 14 lesions were treated with vehicle, which included 7 nodular and 7 superficial BCCs, and the active compound was applied to 13 lesions, comprising 8 nodular and 5 superficial BCCs. 

LDE225 was well tolerated and did not lead to skin irritation. Plasma LDE225 concentrations after 4 weeks were below detection level (0.06 ng ml–1) in 4/8 patients (highest plasma level detected was 0.11 ng ml–1). Mean LDE225 skin concentrations were 737 ng g–1 (BCC) and 605 ng g–1 (uninvolved skin). No clinically significant abnormalities in physical examination, vital signs, electrocardiographs, biochemistry, hematology, or urinalysis results were recorded during the study. Out of 13 (both nodular and superficial) BCCs treated with 0.75% LDE225 cream, 12 showed a clinical response (complete clinical response: 3; partial response: 9). Except in 1 out of 14 lesions, the vehicle produced no clinical response (Figure 4a). At the patient level, all 8 patients showed at least partial response for lesions treated with active compound, compared with 1 out of 8 for lesions treated with vehicle, a difference of 87.5%.

Journal_of_Investigative_Dermatology1

Figure 1. Regression of hedgehog (HH) Ag-1.3 preinduced basaloids in embryonic skin punches by LDE225 but not by cyclopamine. (a) Blue color due to X-gal staining indicates b-D-galactosidase induction from the patched 1 (Ptch1) promoter; scale bar¼1 mm. (b) Basaloid lesions represented as dark, roundish structures of variable sizes (all such lesions are marked with black arrows in the HH Ag-1.3 (1 mM) group) were induced by incubation of skin explants from embryonic Ptch1þ/_LacZ mice with HH Ag-1.3 (1 mM) for 7 days, and skin explants were maintained for another 8 days in the presence of HH Ag-1.3 and the indicated concentrations of LDE225 or cyclopamine; scale bar¼0.25 mm. (c) The number of lesions were counted (four slides per condition) and plotted against the concentration of LDE225 (light gray bars) or cyclopamine (gray bars) versus positive (hatched bar) or negative (black bar) controls.

The probability that the matched difference between active compound and vehicle is at least 50% was calculated to be 0.972. The clinical response parameters were defined as follows: complete response—there is no longer any visible evidence of a lesion consistent with BCC at this site; partial response—although a BCC still remains at this site, it has demonstrated a visible decrease in size compared with baseline; no response—the BCC has not demonstrated any visible decrease in size compared with baseline. The clinical evaluation was performed by an investigator who was blinded to treatment groups. Macroscopic and dermatoscopic images of one representative tumor at baseline (before treatment) and after 4 weeks of treatment with 0.75% LDE225 cream are shown in Figure 4b–e. Figure 4f and g show the three-dimensional (3D) image of the tumor at baseline (before treatment) and after 4 weeks of treatment with 0.75% LDE225 cream. Volumetric measurements of the BCCs on day 29 revealed a mean tumor volume reduction of 56% in the LDE225-treated lesions (Figure 4h), and was significantly different from the reduction in vehicle-treated tumors (P¼0.011). The mean surface area reduction on day 29 was 39%, which was significantly different from that seen with vehicle (P¼0.001; Figure 4i). 

In each patient, one vehicle-treated (n¼8) and one LDE225-treated BCCs (n¼8) were biopsied for histopathological analysis. Routine hematoxylin and eosin staining of day 29 skin biopsies of vehicle- and LDE225-treated tumors revealed that in all biopsied lesions tumor nests were still present. However, the number of proliferating Ki-67þ tumor cells was significantly lower in LDE225-treated BCCs compared with the vehicle-treated lesions. Markers for apoptosis (TUNEL, B-cell lymphoma 2 (Bcl-2), and cleaved caspase-3) or senescence (b-galactosidase) demonstrated no difference between vehicle- and LDE225-treated groups. Images for Ki-67 and data for Ki-67, TUNEL, and Bcl-2 are presented in Figure 5 (panels ; data for cleaved caspase-3 and b-galactosidase not shown). Gli1 (GLI family zinc-finger 1) staining revealed various degrees of reduction of nuclear reactivity in all the LDE225-treated BCCs compared with the vehicle-treated group (data not shown). An example illustrating a reduction in the number and intensity of Gli1-positive cells is shown in Supplementary Figure S2 online.

Figure 2. Efficacy of topical LDE225 to inhibit entry into anagen and hair growth. Mice were treated once daily with a 1% solution of LDE225 applied topically to depilated back skin. (a) On day 15, all vehicle-treated controls have entered anagen phase with visible hair re-growth, (b) whereas almost all LDE225-treated animals visibly remained in telogen phase. (c) Photos were also taken 8 days following treatment cessation, i.e., on day 22, when a complete hair coat was observed for all animals treated with vehicle, (d) whereas LDE225-treated mice just showed signs of anagen initiation with normal hair growth progression from day 22 onward. Shown are 4 representative animals of groups of 15.

Figure 2. Efficacy of topical LDE225 to inhibit entry into anagen and hair growth. Mice were treated once daily with a 1% solution of LDE225 applied topically to depilated back skin. (a) On day 15, all vehicle-treated controls have entered anagen phase with visible hair re-growth, (b) whereas almost all LDE225-treated animals visibly remained in telogen phase. (c) Photos were also taken 8 days following treatment cessation, i.e., on day 22, when a complete hair coat was observed for all animals treated with vehicle, (d) whereas LDE225-treated mice just showed signs of anagen initiation with normal hair growth progression from day 22 onward. Shown are 4 representative animals of groups of 15.

Downregulation of hedgehog target genes 
Taqman quantitative real-time reverse transcriptase-PCR was used to measure the expression levels of Gli1, Gli2, PTCH1, and PTCH2 genes in BCC biopsies from patients treated with 0.75% LDE225 cream and vehicle. The expression levels of Gli1, Gli2, and PTCH2 in biopsies from LDE225-treated BCCs were downregulated 2- to 16-fold in 6/8 patients and correlated with clinical response end points in all but one patient (Supplementary Figure S3 online).

PTCH mutation analysis 
Germline PTCH1 mutations were identified using material from the peripheral blood from six out of eight NBCCS patients (Supplementary Table S1 online). Two were missense mutations (found together in two patients—patients 5 and 7— mother and son), one was a nonsense mutation, one was a deletion of eight amino acids, and two were frameshift mutations leading to the truncation of the protein. One of the frameshift mutations was accompanied by the P1315L change, previously reported in NBCCS patients (Pastorino et al., 2005), and was present in two other patients from this study. All these variations were also observed in the vehicleand LDE225-treated tumors (germline mutations). Additionally, one of the patients with germline mutations showed an additional nonsense somatic mutation in exon 8 of PTCH1 in the vehicle-treated biopsy, whereas one of the patients with no germline mutation also showed a somatic nonsense mutation in the vehicle-treated tumor. Sequencing of the PTCH2 gene in DNA from peripheral blood revealed two previously reported missense variations (rs11573586 and rs11573590) in two of the six patients with PTCH1 mutations. No PTCH2 mutation was observed in patients tested negative for PTCH1 germline mutations.

Figure 3. The CONSORT flowchart. In total, eight patients with multiple basal cell carcinomas (BCCs) were randomized. In every patient, selected BCCs were randomly allocated to treatment with LDE225 or vehicle.

Figure 3. The CONSORT flowchart. In total, eight patients with multiple basal cell carcinomas (BCCs) were randomized. In every patient, selected BCCs were randomly allocated to treatment with LDE225 or vehicle.

Figure 4. Tumor response in nevoid basal cell carcinoma syndrome (NBCCS) patients after a 4-week treatment with 0.75% LDE225 cream. (a) The results of the clinical evaluation are displayed as no response (light gray), partial response (gray), and complete response (dark gray). Macroscopic and dermatoscopic images of a nodular ulcerated basal cell carcinoma (b, d) before (baseline) and (c, e) after treatment. Three-dimensional (3D) surface reconstruction of the same lesion (f) before and (g) after treatment, showing a 79% volume reduction compared with baseline. Mean percentage (%) change from baseline in (h) tumor volume (P¼0.011) and (i) tumor surface area (P¼0.001) for basal cell carcinomas (BCCs) treated with 0.75% LDE225 cream (black) and vehicle (light gray) was measured on days (d) 8, 15, 22, and 29 (within-patient two-sided t-test referring to day 29 only). Scale bar¼5mm for all micrographs.

Figure 4. Tumor response in nevoid basal cell carcinoma syndrome (NBCCS) patients after a 4-week treatment with 0.75% LDE225 cream. (a) The results of the clinical evaluation are displayed as no response (light gray), partial response (gray), and complete response (dark gray). Macroscopic and dermatoscopic images of a nodular ulcerated basal cell carcinoma (b, d) before (baseline) and (c, e) after treatment. Three-dimensional (3D) surface reconstruction of the same lesion (f) before and (g) after treatment, showing a 79% volume reduction compared with baseline. Mean percentage (%) change from baseline in (h) tumor volume (P¼0.011) and (i) tumor surface area (P¼0.001) for basal cell carcinomas (BCCs) treated with 0.75% LDE225 cream (black) and vehicle (light gray) was measured on days (d) 8, 15, 22, and 29 (within-patient two-sided t-test referring to day 29 only). Scale bar¼5mm for all micrographs.

Figure 5. Evaluation of an antiproliferative and/or proapoptotic effect of 0.75% LDE225 cream. Immunohistological analyses show representative slides stained for proliferating cells (Ki-67þ) after a 4-week twice-daily treatment with (a) 0.75% LDE225 cream compared with (b) vehicle-treated basal cell carcinomas (BCCs). (c) Box plots illustrate numbers of Ki-67þ cells detected in epidermis, BCC tumor nests, and dermis for vehicle (n¼8, white) and 0.75% LDE225 cream (n¼8, dark gray). The only significant differences were observed for Ki67 in the BCC tumors. (d) B-cell lymphoma 2 (BCL-2) and (e) TUNEL staining did not reveal any significant differences between LDE225 or vehicle groups in any compartment. Scale bar¼100 mm (a and b left panels); 25 mm (a and b right panels).

Figure 5. Evaluation of an antiproliferative and/or proapoptotic effect of 0.75% LDE225 cream. Immunohistological analyses show representative slides stained for proliferating cells (Ki-67þ) after a 4-week twice-daily treatment with (a) 0.75% LDE225 cream compared with (b) vehicle-treated basal cell carcinomas (BCCs). (c) Box plots illustrate numbers of Ki-67þ cells detected in epidermis, BCC tumor nests, and dermis for vehicle (n¼8, white) and 0.75% LDE225 cream (n¼8, dark gray). The only significant differences were observed for Ki67 in the BCC tumors. (d) B-cell lymphoma 2 (BCL-2) and (e) TUNEL staining did not reveal any significant differences between LDE225 or vehicle groups in any compartment. Scale bar¼100 mm (a and b left panels); 25 mm (a and b right panels).

DISCUSSION

LDE225 is a selective antagonist of SMO that inhibits HH-dependent signaling, with IC50s in the low nanomolar range based on cellular assays (Pan et al., 2010b). In this study, LDE225 efficiently mediated regression of basaloid tumor nests induced by a specific agonist in organ cultures of skin explants taken from embryonic Ptchþ/_LacZ heterozygous knockout mice with IC50 p150 nM. Topical application of LDE225 solutions to C57/BL6 mice was found to prevent hair growth and to inhibit HH target gene expression. Pharmacokinetic studies revealed that LDE225 penetrates well into porcine and human skin but permeates through skin only to a very low extent. These preclinical data suggested that LDE225 has a high potential to interfere with the HH pathway in BCCs after topical treatment, which encouraged us to conduct a short-term trial in patients with NBCCS. In the recent past, a handful of reports were published on HH inhibitors systemically applied to humans afflicted with a cancerous disease (Rudin et al., 2009; Von Hoff et al., 2009; Goldberg et al., 2010; Yang et al., 2010). Overall, these compounds showed a favorable, yet limited, response in terms of clearance. To our knowledge, it has never been demonstrated that a topical inhibitor of the HH pathway can also be beneficial, in this case for patients with NBCCS. This syndrome is hallmarked by a high penetrance of inactivating mutations of the PTCH1 gene (Wicking and McGlinn, 2001), leading to the development of multiple BCCs and, therefore, served perfectly for a therapeutic trial with a topical SMO inhibitor. In a previous trial, another SMO inhibitor termed Cur61414 failed to show efficacy when applied in a topical formulation to BCCs, possibly because of inadequate penetration into the skin (Tang et al., 2007). In our proofof- concept study, 12 out of 13 BCCs, both nodular and superficial, responded clinically after a 4-week twice-daily treatment period with topical LDE225. The average volume reduction of the LDE225-treated BCCs was 56%. A detailed inspection of the individual data revealed that some tumors (4 out of 13) responded rapidly in terms of volume shrinkage by 466% or surface area reduction by 450% in the first 2 to 3 weeks and reached a plateau thereafter. The rest of the responding tumors showed a slow but continuous reduction of tumor size over the entire 4-week treatment period. 

The number of Ki-67þ cells in the tumor was significantly reduced, when measured directly after treatment on day 29 when compared with vehicle, whereas the number of apoptotic cells revealed no difference in TUNEL and cleaved caspase-3 staining at this time point. Using a K5-tTA/TREGli2 transgenic BCC mouse model, certain investigators (Hutchin et al., 2005) demonstrated that after transgene inactivation of Gli2, a key transcription factor of the HH pathway, previously developed BCCs showed a volume reduction of over 90% after 3 weeks. In this experiment, distinct tumor cell apoptosis, measured by TUNEL staining, peaked after 3 days and returned to baseline levels on day 21. In contrast, tumor volume and Ki-67þ cells showed an almost linear decrease until day 21. This could explain the fact that our immunohistological results (only available from day 29 after treatment initiation) still revealed reduced proliferation, but did not show an increased apoptosis any longer. Although three BCCs were considered as complete clinical responders, none of the lesions showed evidence of histological clearance after 4 weeks of treatment. Currently, it remains unsolved whether topical or even systemic HH inhibition has the potential to cause complete tumor regression, or whether HH inhibition induces demise of only a portion of the tumor cells, keeping cancer stem cells in a dormant state as long as the activation of downstream transcription factors is suppressed. In fact, a recent study suggested the latter by showing that even prolonged inhibition of HH signaling is not sufficient to cause complete tumor clearance, indicating the persistence of a quiescent cell population in regressed BCC (Hutchin et al., 2005). Future trials with different dosing and prolonged treatment are required to answer these questions. 

In line with the clinical response, we demonstrated that the expression of Gli1, Gli2, PTCH1, and PTCH2 genes showed a 42-fold reduction in six out of eight NBCCS patients. In addition, Gli1 nuclear staining was reduced in all seven BCC skin samples that were tested. These findings confirm the efficacy of LDE225 in inhibiting the SMO pathway. For the future, the expression of Gli1, which can easily be demonstrated by immunohistology, might be of relevance to select sporadic BCCs for treatment with LDE225 cream. 

The PTCH mutation analysis confirmed previous reports from the literature regarding the different PTCH1 mutations that can be identified in NBCCS patients (Pan et al., 2010a). In two patients, there were also changes within exons coding for the PTCH2 gene. Although it is well documented that lossof- function mutations within PTCH1 are the cause for deregulated SMO activity, the relevance of PTCH2 for the suppression of SMO is not fully understood. Certain investigators (Rahnama et al., 2004) concluded that PTCH1 and PTCH2 have distinct roles and do not function similarly in terms of SMO suppression and HH pathway activation. This is in line with the observation that Ptch2-deficient animals were born alive, showed no obvious defects, and were not cancer prone, in contrast to Ptch1 deficiency (Lee et al., 2006). However, in animals with heterozygous deficiency of both Ptch1 and Ptch2, a higher incidence of tumors and a broader spectrum of tumor types were observed when compared with Ptch1 heterozygous knockout mice (Lee et al., 2006). Therefore, it could be assumed that PTCH2 modulates tumorigenesis associated with loss-of-function mutations of PTCH1. 

Although surgical excision is the standard treatment of BCCs, with recurrence rates between 1 and 3%, there exists no evidence-based recommendation of how to treat BCCs in NBCCS patients. These individuals require a dual approach, focusing on long-term preservation of healthy skin on the one hand and reversing the growth of invasive tumors on the other hand. Nonsurgical interventions include cryotherapy, photodynamic therapy, ablative laser therapy, and topical imiquimod 5% cream (van der Geer et al., 2009). One of the great advantages of the nonsurgical therapies is the preservation of the skin and the achievement of good cosmetic results. However, some of these treatment modalities have high recurrence rates (8–22%) and side effects such as erosions, ulcerations, pain, and also scarring, which have to be considered (Morton et al., 2008). The systemic approach of the SMO inhibitor GDC0449 (Genentech, San Francisco, CA) in a patient with NBCCS suffering from multiple BCCs resulted in the clinical disappearance of nearly all BCCs, but was also escorted by almost complete alopecia (Goldberg et al., 2010). In contrast, topically applied LDE225 showed an excellent safety profile with no local or systemic side effects, while being able to induce regression in all BCCs except one. Further exploration will demonstrate how effective 0.75% LDE225 cream can be in terms of a complete histological clearance when applied long term, and if so, whether persistent treatment is required to prevent the recurrence of BCCs at treated sites. Taken together, the topical route of administration of 0.75% LDE225 cream appears to be an attractive alternative to surgical intervention, oral SMO antagonists, and to other topical drugs available for BCC treatment, which frequently cause local or systemic side effects.

MATERIALS AND METHODS

Basaloid tumor formation and regression assay in vitro 
Ptchþ/_LacZ heterozygous knockout mice in which the bacterial LacZ gene is placed under the control of the promoter of the replaced Ptch1 allele were purchased from Jackson Laboratories (Bar Harbor, ME). Mouse embryos were collected and killed at late gestation (embryonic day 17.5), and the skins were excised. Ptchþ/_ mouse skin punch animal protocol was approved by the institutional animal care and use committee of the Genomics Institute of the Novartis Research Foundation. The punches were placed in a collagen-coated Transwell (Corning Costar, Sigma-Aldrich, St Louis, MO) with the epidermal side facing up and were incubated at the air–liquid interface in DMEM/F12 (3:1) culture medium with 5% fetal bovine serum further supplemented with EGF (20 ng ml–1; R&D Systems, Minneapolis, MN), insulin (5 mgml–1; Sigma, Sigma-Aldrich, St Louis, MO), and hydrocortisone (0.536 mgml–1; Sigma, Sigma-Aldrich). The presence of 1 mM of the SMO agonist HH Ag-1.3 (Frank-Kamenetsky et al., 2002) induced basaloid lesions within 7 days. SMO antagonists LDE225 or cyclopamine were added at different concentrations to address their activity to inhibit basaloid formation. In the regression assay, basaloid tumor lesions were induced first by HH Ag-1.3, and LDE225 (1.5 mM, 750 nM, and 150 nM) or cyclopamine (10, 5, and 1 mM) was added subsequently to cultures maintained in the presence of 1 mM HH Ag-1.3 for another 8 days. Pathway activation was assessed by assaying b-D-galactosidase activity using 5-bromo-4-chloro-3-indolyl b-D-galactoside (Xgal) staining of skin punches (1 mgml–1 of X-gal in phosphatebuffered saline) for 4 hours. Samples were then rinsed with phosphate-buffered saline and fixed with 4% paraformaldehyde and processed for histological analysis. Basaloid lesions were evaluated by microscopy of fixed hematoxylin and eosin-stained tissue sections and counted according to the hair follicle development staging system as described previously (Chuong, 1998; Paus et al., 1999). Basaloids morphologically similar to developing hair follicles of stages 1–4 were counted in each field (average field size: 1.5mm in length and 0.25mm in width) using four fields per culture condition. Mean basaloid counts±SEM were calculated and plotted against the compound concentration using EXCEL (Microsoft Corporation, Redmond, WA).

Inhibition of hair growth and HH target genes in C57/BL6 mice by topically applied LDE225 
Female C57/BL6 mice (aged 6 to 7 weeks) were depilated with a melted beeswax/resin mixture (depilation induces hair growth in telogen skin) when at telogen phase (pink skin color). Studies described in this report were performed according to a notification (y 9, O¨ sterreichisches Tierversuchsgesetz) approved by the local governmental authority (Landesregierung Wien, Magistratsabteilung 58, Vienna, Austria). Starting 24 hours after depilation, skin areas were treated with LDE225 dissolved in an experimental vehicle comprising propylene-glycol and ethanol (7þ3 volumes) once daily for 14 consecutive days using 12 mice per test article or vehicle. Treated skin areas were examined daily for pigment formation (indicating start of hair re-growth) and appearance of hair. Photographs were taken on days 15 and 22. Mice were followed up until recovery from treatment, visible as hair re-growth. In another experiment, LDE225 was applied as a 1 or 0.3% solution for 7 consecutive days, and skin biopsies were taken 4 hours after last drug application for quantitative gene expression analysis by reverse transcriptase-PCR. Biopsies were homogenized with a bead-based homogenizer for 30 seconds with 5.5ms–1 at room temperature. Total RNA was extracted by the Trizol method (Invitrogen, Carlsbad, CA), and RNA concentrations and purity were determined spectrophotometrically against a known RNA standard. RNA samples were transcribed into complementary DNA using the high-capacity complementary DNA ‘‘Archive Kit’’ followed by specific amplification of 20 ng triplicate samples using the Assay-on-Demand reagents for specific amplification and detection of mouse transcripts of the reference gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH; assay on demand Mm4352339E) and the specific HH target genes Gli1, Gli2, Sox-9, and sFRP-1 (assays on demand Mm01160468_g1, Mm01293117_m1, Mm00448840_m1, and Mm00489161_m1, respectively, all purchased from Applied Biosystems, ABI, Foster City, CA). Reverse transcriptase-PCR reactions were run on an ABI 7500 Fast Real-time PCR system machine. Threshold cycle (Ct) values obtained for GAPDH were subtracted from the Ct obtained for each of the target genes. Resulting DCt values were then converted to linear positive values using the formula 2(DCt) and multiplied by 1,000 to show relative expression levels in the graphs. 

Skin penetration/permeation studies 
For ex vivo studies, saturated solutions of LDE225 in propyleneglycol were applied epicutaneously to human cadaver skin or pig skin mounted in static Franz-type diffusion cells. Permeation (flux) through the skin was measured by sampling 100 ml of receptor fluid five times up to 48 hours as described previously (Schmook et al., 1993). Skin penetration was determined, at the end of a 48-hour exposure time, by drug concentration measurements in the skin (after removal of the stratum corneum). For in vivo studies, a 0.75% experimental formulation was topically applied to 4 cm2 skin of domestic pigs to determine dermal concentrations at selected time points and to 10% body surface area to evaluate systemic exposure at selected time points up to 24 hours after application by plasma level determinations. Samples were analyzed for LDE225 in the sub ng ml–1 range by reversed-phase HPLC with mass spectrometry detection in the MS/MS mode, using an 1100 Series LC System (Agilent, Santa Clara, CA) coupled to a LTQ mass spectrometer (Thermo Finnigan, San Jose, CA) as described previously (Billich et al., 2004).

Study in NBCCS patients 
Study CLDE225B2203 (EudraCT number: 2008 005506-40) was a 4-week, double-blind, randomized, vehicle-controlled study to evaluate the safety, local tolerability, and pharmacokinetics of twice-daily topical administration of 0.75% LDE225 cream and to explore the pharmacodynamics in skin BCCs in NBCCS patients. The study was conducted in Vienna (Department of Dermatology, Division of Immunology, Allergy and Infectious Diseases, Medical University of Vienna, Austria) after approval of the local ethics committee of the Medical University of Vienna and the Austrian health authority (Bundesministerium fu¨ r Gesundheit, Vienna, Austria). The study was conducted according to the principles embodied in the Declaration of Helsinki Principles, and all patients provided written informed consent before participating in the study. Patients aged between 18 and 75 years with multiple BCCs, fulfilling the diagnostic criteria for NBCCS (Kimonis et al., 1997), were enrolled. The subjects needed to present BCCs on two different body areas (scalp, arm, anterior or posterior trunk, and legs). Each of the two areas selected needed to have either one BCC 410mm in diameter or two BCCs 45mm in diameter. The investigator enrolled patients into the study and assigned a lesion number to each BCC to be treated. Each patient was assigned a randomization number by a central randomization office. The randomization number was linked to a randomization scheme generated by a validated system indicating which study treatment was to be applied to each lesion number for each patient. The BCC(s) in the respective area was (were) treated for 4 weeks twice daily with 0.75% LDE225 cream or vehicle. BCCs treated with vehicle and BCCs treated with 0.75% LDE225 cream were strictly located in different body areas. During the treatment phase, the patients were monitored as outpatients on a weekly basis. Blood pressure, pulse rate, and hematology/blood chemistry laboratory parameters were examined on day –1 and throughout the study period, including the follow-up at days 36 and 56. At the end of treatment on day 29, one skin biopsy was taken from an LDE225-treated BCC and one skin biopsy was taken from a vehicle-treated tumor in every patient (4-mm punch biopsy). Each specimen was divided into four parts for further analyses (paraffin sections, cryopreserved sections, and DNA and RNA isolation). Two additional 4-mm punch biopsies were taken from an LDE225-treated tumor and from perilesional LDE225-treated skin for pharmacokinetic assessments from every patient.

Imaging and volumetric measurements
Digital photography including dermatoscopic, macroscopic, and 3D images of the BCCs were performed on days _1, 8, 15, 22, and 29, and during follow-up on days 36 and 56. The volume of the respective lesion was measured with a special device, the ‘‘3D LIFEVIZ Micro system’’ (QuantifiCare, Sophia Antipolis, France). Thereby, a lens splitter is used to produce two images of the surface, captured at the same time, with viewing angle differences close to human vision. A stereovision algorithm is applied to reconstruct and quantitatively analyze the skin surface in 3D.

Histopathological biomarkers
Paraffin sections from skin biopsies of LDE225- or vehicle-treated BCCs were used for immunohistochemical analysis. To determine the antiproliferative and/or proapoptotic effects of LDE225, we performed Ki-67 (clone Ki-67, IgG1, DakoCytomation, Glostrup, Denmark), Bcl-2 (clone 124, IgG1, DakoCytomation) immunohistochemical staining, and TUNEL labeling (In Situ Cell Death Detection Kit, POD; Roche Molecular Biochemicals, Basel, Switzerland) according to the manufacturer’s instructions. Cryopreserved sections from vehicle- and LDE225-treated tumors were subjected to staining for the apoptosis marker cleaved caspase-3 (cleaved Caspase-3 (Asp175) (5A1E) rabbit monoclonal antibody) and the senescence marker b-galactosidase (both from Cell Signaling, Danvers, MA), according to the manufacturer’s instructions. Routine hematoxylin and eosin staining was performed on all formalin-fixed biopsy specimens (n¼16) and were evaluated by a histopathologist who was blinded to treatment groups. Immunohistochemical staining for Gli-1 was performed using the Ventana Discovery XT (Ventana Discovery Systems, Tucson, AZ) on skin biopsies of LDE225- and vehicle-treated BCCs taken from seven NBCCS patients after 4 weeks of twice-daily treatment.

Molecular biomarkers
Total RNA was extracted from BCC biopsies using the RNeasy mini-kit from Qiagen according to the manufacturer’s instructions (Qiagen, Hilden, Germany) and was reverse transcribed to complementary DNA using random hexamers and the High-Capacity complementary DNA archive kit from Applied Biosystems. 

Real-time PCR was performed using the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). Primers were obtained from Applied Biosystems (Supplementary Table S2 online). The expression results for the three control genes, UBC, DNAJB12, and C19ORF56, were used to control for intersample variability. Data are presented graphically as fold change. Fold change is calculated using the formula 2_DDCT, where CT is the threshold cycle (the fractional cycle number at which the amount of amplified target reaches a fixed threshold) and DCT is the difference in threshold cycles for the target gene and the average of the control gene Cts. DDCT is the difference in DCT for the vehicle- and LDE225-treated biopsies.

PTCH1 and PTCH2 sequencing analysis
Sequencing of the complete coding regions (exons and exon boundaries) of PTCH1 (NM_000264.3) and PTCH2 (NM_0037 38.8) was performed by direct sequencing. PCR was performed using unique primers carrying a universal M13 sequence tag (Supplementary Table S3a and S3b online). All amplifications except for exons 1 and 2 of each gene was performed with AmpliTaq Gold Master Mix (Applied Biosystems) on a GeneAmp 9700 PCR System (Applied Biosystems) according to the manufacturer’s instructions. All sequence variations were noted and corresponding coding changes characterized.

Statistical analysis
A total of eight patients were enrolled to ensure that six patients completed the study. Six patients would provide at least 80% probability that a 50% level of proof would be achieved that the matched difference between active compound and vehicle for at least partial response rates is at least 50%, if the true vehicle partial response rate was p10% and the true active partial response rate was 75%. 

Patients may have been treated for one or two lesions with each of LDE225 and the vehicle cream. For the assessment of clinical response, where two lesions were treated with a particular treatment for a patient, the overall patient/treatment response was determined. If both lesions had the same level of response (no, partial, complete) then this was the overall response for that treatment. Where the two lesions had different responses, the overall patient/treatment response was determined as complete response if both lesions had complete response and partial response otherwise. The probability that the matched difference between the at least partial response rates (at the patient level) for active compound and vehicle is at least 50% was calculated as a posterior probability by sampling from the joint Dirichlet distribution of the four possible outcomes (response for both treatments, response for active compound only, response for vehicle only, and no response for both treatments). For the analysis of tumor volume and size derived from 3D digital photography, the values were summed for each patient and treatment to give the total volume and surface area for LDE225- and vehicle-treated BCCs for each patient. The percentage change from baseline was then calculated and compared on day 29 between LDE225 and vehicle using a withinpatient two-sided t-test. No adjustment was made for multiple testing. Statistical analysis was performed using SAS version 9.2, SAS Institute, Cary, NC.

CONFLICT OF INTEREST

G Stingl has received consultation honoraria from Novartis. FK, JM, BW-W, HA, JK, XW, SP, OD, CE, KR, JD, IC, HG, AS, and AB are employees of Novartis. AMCA and MDeR were employees of Novartis during the study phase and manuscript preparation. HS, LM, CS, G Stary, AJ, AMCA, and MDeR have no conflict of interest to declare pertaining to this study.

ACKNOWLEDGMENTS

We thank the patients who kindly participated in this study. In addition, we thank Marion Dorsch for excellent technical help, Julie Jones for statistical advice, Jean-Philippe Thirion (Quantificare) for his support, and Constanze Jonak and Alice Pinc for their precise work as study pharmacists. We thank Dibyajyoti Mazumder (Novartis Healthcare) for providing editorial assistance.

SUPPLEMENTARY MATERIAL

Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid

REFERENCES

Billich A, Aschauer H, Aszodi A et al. (2004) Percutaneous absorption of drugs used in atopic eczema: pimecrolimus permeates less through skin than corticosteroids and tacrolimus. Int J Pharm 269:29–35
Boutet N, Bignon YJ, Drouin-Garraud V et al. (2003) Spectrum of PTCH1 mutations in French patients with Gorlin syndrome. J Invest Dermatol 121:478–81
Chuong C-M (1998) Molecular Basis of Epithelial Appendage Morphogenesis. R.G. Landes Company: Texas, 75–110 
Epstein EH (2008) Basal cell carcinomas: attack of the hedgehog. Nat Rev Cancer 8:743–54
Farndon PA, Del Mastro RG, Evans DG et al. (1992) Location of gene for Gorlin syndrome. Lancet 339:581–2
Frank-Kamenetsky M, Zhang XM, Bottega S et al. (2002) Small-molecule modulators of Hedgehog signaling: identification and characterization of Smoothened agonists and antagonists. J Biol 1:10
Goldberg LH, Firoz BF, Weiss GJ et al. (2010) Basal cell nevus syndrome: a brave new world. Arch Dermatol 146:17–9
Hutchin ME, Kariapper MS, Grachtchouk M et al. (2005) Sustained Hedgehog signaling is required for basal cell carcinoma proliferation and survival: conditional skin tumorigenesis recapitulates the hair growth cycle. Genes Dev 19:214–23
Jiang J, Hui CC (2008) Hedgehog signaling in development and cancer. Dev Cell 15:801–12
Kimonis VE, Goldstein AM, Pastakia B et al. (1997) Clinical manifestations in 105 persons with nevoid basal cell carcinoma syndrome. Am J Med Genet 69:299–308
Laimer M, Onder K, Schlager P et al. (2008) Nonsense-associated altered splicing of the Patched gene fails to suppress carcinogenesis in Gorlin syndrome. Br J Dermatol 159:222–7
Lee Y, Miller HL, Russell HR et al. (2006) Patched2 modulates tumorigenesis in patched1 heterozygous mice. Cancer Res 66:6964–71
Lo Muzio L (2008) Nevoid basal cell carcinoma syndrome (Gorlin syndrome). Orphanet J Rare Dis 3:32
Morton CA, McKenna KE, Rhodes LE (2008) Guidelines for topical photodynamic therapy: update. Br J Dermatol 159:1245–66
Oro AE, Higgins K (2003) Hair cycle regulation of Hedgehog signal reception. Dev Biol 255:238–48
Pan S, Dong Q, Sun LS et al. (2010a) Mechanisms of inactivation of PTCH1 gene in nevoid basal cell carcinoma syndrome: modification of the two-hit hypothesis. Clin Cancer Res 16:442–50
Pan S, Wum X, Jiang J et al. (2010b) Discovery of NVP-LDE225, a potent and selective smoothened antagonist. ACS Med Chem Lett 1:130–4
Pastorino L, Cusano R, Nasti S et al. (2005) Molecular characterization of Italian nevoid basal cell carcinoma syndrome patients. Hum Mutat 25:322–3
Paus R, Muller-Rover S, Van Der Veen C et al. (1999) A comprehensive guide for the recognition and classification of distinct stages of hair follicle morphogenesis. J Invest Dermatol 113:523–32
Rahnama F, Toftgard R, Zaphiropoulos PG (2004) Distinct roles of PTCH2 splice variants in Hedgehog signalling. Biochem J 378 (Part 2):325–34
Rohatgi R, Scott MP (2007) Patching the gaps in Hedgehog signalling. Nat Cell Biol 9:1005–9
Rudin CM, Hann CL, Laterra J et al. (2009) Treatment of medulloblastoma with hedgehog pathway inhibitor GDC-0449. N Engl J Med 361:1173–8
Ruiz i Altaba A, Mas C, Stecca B (2007) The Gli code: an information nexus regulating cell fate, stemness and cancer. Trends Cell Biol 17:438–47
Schmook FP, Stutz A, Reinhardt J (1993) Penetration of Sandimmune (cyclosporin A) in rat skin in vitro. Effects of penetration enhancers and solvents. Skin Pharmacol 6:116–24
Shafei-Benaissa E, Savage JR, Babin P et al. (1998) The naevoid basal-cell carcinoma syndrome (Gorlin syndrome) is a chromosomal instability syndrome. Mutat Res 397:287–92
Tang JY, So PL, Epstein EH Jr (2007) Novel Hedgehog pathway targets against basal cell carcinoma. Toxicol Appl Pharmacol 224:257–64
van der Geer S, Ostertag JU, Krekels GA (2009) Treatment of basal cell carcinomas in patients with nevoid basal cell carcinoma syndrome. J Eur Acad Dermatol Venereol 23:308–13
Von Hoff DD, LoRusso PM, Rudin CM et al. (2009) Inhibition of the hedgehog pathway in advanced basal-cell carcinoma. N Engl J Med 361:1164–72
Wicking C, Shanley S, Smyth I et al. (1997) Most germ-line mutations in the nevoid basal cell carcinoma syndrome lead to a premature termination of the PATCHED protein, and no genotype-phenotype correlations are evident. Am J Hum Genet 60:21–6
Wicking C, McGlinn E (2001) The role of hedgehog signalling in tumorigenesis. Cancer Lett 173:1–7
Williams JA, Guicherit OM, Zaharian BI et al. (2003) Identification of a small molecule inhibitor of the hedgehog signaling pathway: effects on basal cell carcinoma-like lesions. Proc Natl Acad Sci USA 100:4616–21
Yang L, Xie G, Fan Q et al. (2010) Activation of the hedgehog-signaling pathway in human cancer and the clinical implications. Oncogene 29:469–81

Skvara H, et.al., Topical treatment of basal cell carcinomas in nevoid basal cell carcinoma syndrome with a smoothened inhibitor. Journal of Investigative Dermatology, 2011.48.