
	Version PDF

Induction of the differentiation and apoptosis of tumor cells in vivo with efficiency and selectivity

European Journal of Dermatology. Volume 14, Number 2, 96-102, March-April 2004, Investigative report

Summary

Author(s) : Sinan TAŞ, Oktay AVCI , Yasemin Sokak No. 6, Narlidere, Izmir 35320, Turkey Manavkuyu Mahallesi 244 Sokak, Demirler Sitesi 5‐1\12, Bornova, Izmir, Turkey .

Summary : Hedgehog\\smoothened signaling is active in a variety of tumors and is also involved in the maintenance of normal stem cells in vivo. We evaluated the possibility of preferential affection of tumor versus normal cells following inhibition of this signaling. We applied a cream preparation of cyclopamine (an inhibitor of the hedgehog\\smoothened signaling) onto skin tumors in patients who were scheduled for the excision of these tumors (four basal cell carcinomas and a trichoepithelioma in four unrelated patients). All of the cyclopamine‐treated tumors regressed rapidly. Histological and immunohistochemical analyses showed inhibition of the proliferation and highly efficient induction of the differentiation and apoptosis of tumor cells by a non‐genotoxic mechanism. No adverse effects were noted and normal skin tissue and putative stem cells that were exposed to cyclopamine together with tumors were well preserved under the conditions we describe. Our findings show selective and highly efficient induction of the differentiation and apoptosis of tumor cells in vivo by transient inhibition of the hedgehog\\ smoothened signaling and provide a rational cancer therapy.

Keywords : apoptosis, basal cell carcinoma, cancer therapy, cyclopamine, differentiation, hedgehog\\smoothened signaling, stem cells

Pictures

ARTICLE

Auteur(s) : Sinan TAŞ ^a, Oktay AVCI^b

^a Yasemin Sokak No. 6, Narlidere, Izmir 35320, Turkey ^b Manavkuyu Mahallesi 244 Sokak, Demirler Sitesi 5-1/12, Bornova, Izmir, Turkey

Article accepted on 19/01/2004

The hedgehog family of proteins, first identified in drosophila, have been found to act as inducers of cellular differentiation and tissue patterning during development [1]. Patched encodes for a transmembrane protein that serves as a receptor for hedgehog proteins [1]. When not liganded by hedgehog, patched inhibits intracellular signal transduction by another transmembrane protein, the smoothened [1]. Binding of hedgehog to the patched relieves its inhibitory action on smoothened [1]. Hedgehog/smoothened signaling overactivity (resulting from loss-of-function mutations of patched and/or mutations of other elements of the signaling pathway) are found in all basal cell carcinomas (BCCs) as well as in a variety of other tumors [1-6]. Hedgehog/smoothened signaling is also employed by a number of normal cell types in adults and also in the maintenance of stem cells [7, 8].
Cyclopamine is a steroidal alkaloid identified first as a teratogenic compound of the Veratrum plants causing holoprosencephaly in the lambs of the sheep grazing on these plants [9]. Holoprosencephaly was found in later studies to arise from inhibition of the differentiation of hedgehog target cells in developing brain by cyclopamine [10,11]. Inhibition of the hedgehog/ smoothened signaling by cyclopamine has been reported to cause inhibition of cellular differentiation in other systems as well, including the differentiation of bone marrow cells to erythroid cells [12] and the differentiation of urogenital sinus to prostate [13]. After testing a treatment schedule of topical cyclopamine on normal skin without adverse effects, we were interested in studying the influence of a similar treatment on the BCCs and other tumors displaying hedgehog/smoothened signaling overactivity.

Methods

Patients and Tumors

Patients with facial tumors, who were scheduled for excision of these tumors, were recruited. The tumors had not received any treatment previously. The study was explained and written informed consent was obtained from each patient in accordance with the Helsinki Declaration. Patient 1 was a 85 year old man with Gorlin’s syndrome. Tumor 1 (T1) was on his cheek and measured 4 × 3.5 mm on the surface, T2 was on the temporal region (4 × 5 mm), T3 was on the nasolabial region (4 × 4.5 mm) and T4 was frontal (4 × 3.5 mm). Following the findings with the first patient, the men consulting one of us (OA) with facial tumors and who had a clinical diagnosis of BCC were included in the study when they volunteered. A pre-treatment punch biopsy was obtained from an edge of the tumor when its size permitted. Patient 2, 68 years old, had a large (10 × 11.5 mm) ulcerated tumor (T5) on his upper nasal region. Patient 3, 59 years old, had a pigmented tumor (T6) on the zygomatic region measuring 8 × 10 mm on the surface. Patient 4, 82 years old, had a flesh-colored tumor (T7) on his cheek that measured 4 × 4 mm.

Cyclopamine and Placebo Applications

We dissolved cyclopamine (a kind gift of W. Gaffield and also purchased from Toronto Research Chemicals, Inc, North York, Canada) in ethanol and mixed with a base cream [14] to a final concentration of 18 mM. The base cream mixed with ethanol similarly served as placebo. In patient 1, T3 and T4 received the cyclopamine cream and T1 and T2 the placebo. One of us (OA) applied the creams to the assigned BCCs without knowing which cream was placebo (obvious clinical regressions of the cyclopamine-applied tumors, however, revealed the cream identities by day 2). Each BCC received ~ 10 µl of cream on each occasion with the aid of a steel spatula. We applied the creams directly on top of each BCC (with ~ 1 mm wide spread to the surrounding skin) four times per day starting ~ 9.00 a.m. with ~ 3 1/2 hours in between. Night-time applications were avoided in view of the possible loss of cream to linens during sleep. Similar testing of the cyclopamine cream previously on normal skin on ourselves (on the dorsum of the hand and on the deltoid region) had revealed no detectable adverse effect (over 5 weeks of observation at the time of treatment of the first patient; over 35 months at this writing). Cyclopamine and other Veratrum alkaloids had been tested earlier on rodents and teratogenicity was observed only with amounts that are several times greater than we used here [15, 16]. However no previous human study or application of cyclopamine on skin existed. In the absence of information on the skin penetration and metabolism of cyclopamine, we extrapolated from such data on similar steroidal molecules and relied on empirical monitoring of biological responses. With further understanding of the skin penetration, metabolism and pharmacokinetics of cyclopamine in future studies, improvements may be expected in the concentration, frequency and mode of administration of cyclopamine. We kept the cyclopamine cream closed in eppendorf tubes in the freezing compartment of a refrigerator ( ~ – 5 °C) during the course of applications. Unused tubes with cream were stored at – 20 °C and did not appear to lose potency for at least 4 months.
The relatively large sizes of the tumors 5 and 6 (in patients 2 and 3) allowed us to obtain a punch biopsy from an edge of each prior to treatment. In patient 2, we applied ~ 20 µl cyclopamine cream onto the lower half of the tumor uninterruptedly every third hour and none to the remainder (Fig. 1 e; arrow points to the non-applied region). Thus the tumor cells in the uppermost part are expected to receive the lowest concentration of cyclopamine (by possible diffusion from the directly applied region), if any. In patient 3, ~ 20 µl cyclopamine cream was applied onto the tumor (T6) uninterruptedly every fourth hour, except for a nodule in the periphery of the tumor that was left untreated (Fig. 1g, arrow). In patient 4, ~ 25 µl cyclopamine cream was applied onto the tumor (T7) uninterruptedly every third hour. Cyclopamine applied tumors regressed rapidly in all cases (see below) and we discontinued use of cyclopamine before full disappearance of tumors so as to leave behind tumor material for investigations. Except for T6, the placebo- and cyclopamine-applied tumors (T1-5 and T7) were excised 3 to 4 hours after the last application together with a margin of ~ 5 mm surrounding skin. In case of T6, after discontinuing treatment, we left the residual tumor (Fig. 1h) in place for 6 days and excised on the 6^th day of non-treated follow-up. The ~ 5 mm excision margins were from the pre-treatment tumor margins.

Histopathological and Immunohistochemical Investigations

Excised tissues were fixed immediately in neutral buffered formalin (4 % formaldehyde), embedded in paraffin blocks and serial sections were subjected to hematoxylin-eosin (H&E) and immunohistochemical staining. For immunohistochemical labeling, all primary antibodies were mouse monoclonal antibodies against human antigens. Antibodies against epithelial antigen (Ber-EP4), cytokeratin 15 (C8/144B), cytokeratin 19 (RCK108), Ki-67 (M7187) and p53 (DO-7) were from DAKO (Glostrup, Denmark). For p53, we used DO-7 on tumors 1-4 and additionally used a “cocktail” of the antibodies DO-7 and BP53-12 (NeoMarkers, Lab Vision Corp, Fremont, CA, U.S.A.) on tumors 5-7. Anti-CD44 (F10-44-2; reacts with the CD44 standard) was from Novocastra Labs Ltd (Newcastle upon Tyne, U.K.). Immunohistochemical staining with all antibodies employed peroxidase activity in the detection step and all reaction conditions, including for epitope retrieval, were as recommended by the manufacturer. Previously characterized tissue sections placed on the same slides as the test tissue sections served as control material for immunohistochemical staining. Normal tissue components present in the same tissue sections as tumors provided further control material. Tissue sections were also subjected to conventional periodic acid-Schiff (PAS), diastase-PAS and Alcian Blue (pH 2.5) staining.
We used the morphological criteria of apoptosis [17] for quantifying apoptotic activity. For this purpose we counted all cells in representative high-power (1000X) microscopic fields (number of fields indicated) and determined the proportion of cells that displayed at least two of the following morphological signs: a) nuclear fragmentation, b) homogenous darkly staining compacted nucleus, c) strongly eosinophilic cytoplasm, d) cytoplasmic retraction.

Results

Tumors exposed to cyclopamine regressed rapidly. In patient 1, despite discontinuation of cyclopamine applications during night time, the height of cyclopamine-treated tumors (T3 and T4) decreased visibly on day 2 and several parts of the BCCs disappeared on the 5^th and 6^th days (Fig. 1a vs b and c vs d). In addition cyclopamine treated BCCs started to loose translucency after day 2 (cf, Fig. 1). Placebo-treated tumors, on the other hand, showed no change in size or appearance (not shown). Tumor 5 (in patient 2), which received cyclopamine to its lower half, decreased markedly in this region by 24^th hour (Fig. 1f) while the uppermost part (Fig. 1e, arrow) changed the least (Fig. 1f). The height of tumor 6 (in patient 3) was decreased on day 2 and tumor nodules were much smaller or had disappeared on day 3, except for the nodule onto which we did not apply cyclopamine (Fig. 1h). Tumor 7 (in patient 4) displayed the fastest regression in this series and became nearly undetectable by the 24^th hour (Fig. 1j vs i). Another noteworthy finding in patient 4 was the decreased size and pigmentation of a mole (a benign melanocytic tumor) located adjacent to the treated tumor on the 24^th hour of treatment (Fig. 1j vs i).
The pigmented nature of the BCC in patient 3 facilitated its follow-up and, instead of excising soon after discontinuance of cyclopamine treatment, we left the residual tumor in place for a study of possible late effects. In the absence of treatment no clear further regression was observed and the skin area corresponding to the original tumor was excised on the sixth day of nontreated follow-up.
Histological examinations revealed similar findings with all residual tumors that were excised while still under the influence of cyclopamine (T3, T4, T5, T7). Regions on serial tissue sections corresponding to the tumor nodules that had become undetectable during treatment showed large cystic structures (Fig. 2a, b) displaying either little material inside (Fig. 2a) or material showing featureless staining (Fig. 2b, the large cyst on the right). These cystic structures were devoid of a lining epithelium (Fig. 2a, b) and did not exist in the tissue sections of placebo-treated BCCs. Regions corresponding to the skin areas that contained visible residual tumor showed variously sized cysts forming among the tumor cells (Fig. 2c-e). Higher magnification inspection of these residual tumor areas revealed numerous cells displaying typical apoptotic morphology and the generation of the cysts as a result of massive apoptosis (Fig. 2d, e; note the imminent joining together of the three cysts on Fig. 2d upon removal of the apoptotic septal cells). The placebo-treated tumors (T1, T2) and non-treated tumors (pre-treatment punch biopsy material of T5 and T6), on the other hand, showed typical BCC histology and few or no apoptotic cells (Fig. 2f, Table I; see also Fig. 3l). Tumor 7 was found to be a trichoepithelioma upon microscopic examination (Fig. 2g). Residue of this tumor also showed numerous apoptotic cells and cystic spaces that were forming by the apoptotic removal of tumor cells (Fig. 2g, Table I).

Table I. Inhibition of the proliferation and induction of the differentiation and apoptosis of tumor cells following treatment with cyclopamine

	Percentage Of The Residual Tumor Cells Displaying
Tumor, Histology and Region	EA(Ber-EP4)	CD44	p53	Ki-67	Signs of Apoptosis
T3, BCC	0 ± 0	N.D.	7 ± 6*	N.D	23 ± 9
T4, BCC	0 ± 0	N.D.	9 ± 8*	N.D.	16 ± 8
T5, BCC, Treated Half	0 ± 0	100 ± 0	8 ± 6*	0 ± 0	14 ± 4
T7, TE, Interior	0 ± 0	100 ± 0#	6 ± 3*	0 ± 0	36 ± 14
T7, TE, Periphery	0 ± 0	100 ± 0	96 ± 2	2 ± 1	0 ± 0
C. Tumor treated with cyclopamine for 3 days, followed-up without treatment for 6 days and excised on the 6^th day of non-treated follow-up
	Percentage Of The Residual Tumor Cells Displaying
Tumor, Histology and Region	EA(Ber-EP4)	CD44	p53	Ki-67	Signs of Apoptosis
T6, BCC, Residual Nodules	2 ± 1	99 ± 1	1 ± 1	1 ± 1	3 ± 1
Adjacent To Epidermis
T6, BCC, Residual Nests Deep	64 ± 5	12 ± 5	4 ± 3	12 ± 6	1 ± 1
In Dermis, Invasion Border

Means ± standard deviations from at least 16 (BCCs) or 6 (trichoepithelioma) high power fields of the tissue sections of each tumor are shown.
* Weak intensity labeling.
# Excluding the keratin pearls and the surrounding layer of cells, which were non-labeled.
§ Cells that are more than four cells away from the outermost periphery.

The cyclopamine-treated but not the untreated or placebo-treated BCCs were consistently retracted from stroma (Fig. 2h vs i; see also Fig. 2c) and had an Alcian blue staining material in the retraction spaces (Fig. 2h, arrow) as well as in the cystic spaces forming in these tumors (Fig. 2h). While the molecular nature of this Alcian blue staining material remains to be determined, we note that heparan sulfate proteoglycans react with Alcian blue and are employed in hedgehog/smoothened signaling [18]. Similar material has been described in BCCs in association with degenerative changes [19]. Retraction from stroma has been associated with the cessation of proliferative activity in BCCs [20].
Loss of translucency of the cyclopamine-receiving BCCs suggested the possibility of cyclopamine-induced tumor cell differentiation. We tested this possibility with a number of immunohistochemical markers. The monoclonal antibody Ber-EP4 labels the outer root sheath of hair follicles where the putative hair follicle/epiderm stem cells reside and also labels the BCCs and trichoepitheliomas [21,22]. Squamous cell carcinomas and the more differentiated suprabasal cells of normal epidermis, on the other hand, are not labeled with this antibody [21,22]. In placebo-treated and untreated BCCs we found Ber-EP4 labeling of all peripheral cells and nearly all of the interior cells (Fig. 3a, Table I). In contrast the tumors that had been treated with cyclopamine were completely devoid of Ber-EP4 labeling (Fig. 3b, c, Table I). The hair follicles and normal epidermis found on the sections of cyclopamine-treated tumors showed a normal pattern of Ber-EP4 labeling (i.e. labeling of the outer root sheath cells in hair follicles and labeling of some of the basal but none of the upper layer cells in epidermis; notice in Fig. 3c that while the hair follicle is Ber-EP4 labeled, the nearby trichoepithelioma cells are non-labeled).
Expression of the adhesion molecule CD44, reported to be weak in BCCs and strong in squamous cell carcinomas, particularly in the more differentiated regions and types, increases markedly upon differentiation of epidermal basal cells to spinous cells [23,24]. We found weak, patchy and low frequency CD44 labelling in untreated BCCs (Fig. 3d, Table I). Cyclopamine-treated BCCs, on the other hand, exhibited very strong CD44 labeling of essentially all residual cells (Fig. 3e, Table I). Thus CD44 labelling confirmed the induction of differentiation of tumor cells by cyclopamine and indicated that the loss of Ber-EP4 labelling of BCCs cannot be due to a nonspecific (e.g. degenerative) loss of stainability. Additional immunohistochemical and morphological signs of tumor cell differentiation were also evident in cyclopamine-treated tumors (see below).
Immunohistochemical detection of p53 using monoclonal antibodies that bind both the wild type and most mutant forms revealed strong nuclear labeling in the untreated T5 and placebo-treated T1 and T2 (Fig. 3f, Table I). In comparison, T5 after cyclopamine treatment and T3 and T4, which were cyclopamine-treated, showed markedly decreased frequency and intensity of labeling (Fig. 3g, Table I). Tumor 6 had little immunodetectable p53 before or after treatment (Table I). The cyclopamine-treated trichoepithelioma showed strong p53 labeling of the cells in its periphery while the more differentiated cells towards the interior displaying relatively larger cytoplasm had weak or no labelling (Fig. 3h, Table I). In particular the cells in and immediately around the keratin pearls (an overt sign of differentiation) were uniformly non-labeled (Fig. 3h).
Tumor cell proliferation, measured by the frequency of Ki-67 expressing tumor cells, was inhibited following exposure to cyclopamine (Fig. 3i vs j, Table I).
The residual BCC that was excised after six days of non-treated follow-up (T6) showed a relative paucity of apoptotic cells (Table I) in accordance with the rapid removal of such cells from tissues in vivo. Nevertheless there was a greater frequency of apoptotic cells in the residual T6 than in the pre-treatment tumor (Table I). Residual T6 also showed increased frequency of regions with keratin pearls and cells with enlarged eosinophilic cytoplasm (Fig. 3k; Fig. 3l shows the pre-treatment T6). These morphological signs of tumor cell differentiation were again accompanied by immunohistochemical signs (Table I). Interestingly, residual T6 contained Ber-EP4 labeled cells that were located in the tumor regions expected to have received relatively lower concentrations of cyclopamine (e.g. deeper into dermis) (Table I). Indeed the tumor nodule onto which we had not applied cyclopamine (but could have received relatively lower concentrations by diffusion from the nearby application area) showed a clear gradient pattern of BerEp4 labeling (Fig. 3m). Relative frequencies of the CD44 and Ki-67 labeled and non-labeled cells through the residual T6 were also in accordance with the Ber-EP4 findings (Table I) and showed that the tumor cells that had differentiated beyond a critical step during treatment did not revert in the absence of treatment (the Ber-EP4(–), CD44(+) residual cells). The Ki-67(+) cells located predominantly in the regions of the residual T6 that had Ber-EP4(+) cells and weak or no CD44 labeling (Table 1) testify, on the other hand, to the proliferation of undifferentiated cells in the absence of cyclopamine.

Discussion

Contrary to the earlier reports that inhibition of hedgehog/smoothened signaling by cyclopamine prevents cellular differentiation [9-13], we found induction of the differentiation of BCC cells by cyclopamine. Analysis of gene expression profiles of BCC cells has revealed the epithelial stem cell like features of these tumor cells [25]. Monoclonal antibody Ber-EP4 normally labels the outer root sheath of hair follicles where the putative hair follicle/epiderm stem cells reside [21] and all BCCs have been found positive for the epithelial antigen bound by Ber-EP4 [21,22]. Our finding that treatment of BCCs with cyclopamine causes loss of Ber-EP4 labeling in them is consistent with a requirement for hedgehog/smoothened signaling in stem cell maintenance. However normal stem cells are also maintained by hedgehog/smoothened signaling [8] and we found that the putative hair follicle/epiderm stem cells that were exposed to cyclopamine together with tumor cells, continued to be labeled by Ber-Ep4. In addition, in further analyses of the same cyclopamine-treated tissues we found normal labeling of the hair follicle outer root sheath with antibodies to cytokeratins 15 and 19 (not shown). These antibodies have also been reported to label the putative hair follicle/epiderm stem cells [26]. How might such selectivity on tumor cells be achieved ? We hypothesize that the hedgehog/smoothened pathway in the normal epidermal basal cells and hair follicles, which, unlike the situation in tumors, is responsive to the environmental signals, was inhibited transiently by cyclopamine but this inhibition was later on overcome because of the continuing environmental cues maintaining these cells. The same transient inhibition in the tumor cells lacking a pathway capable of responding to the environmental signals (because of e.g. lacking a normal patched) might have sufficed to trigger a differentiation program leading to or beyond the step detected by Ber-EP4 and anti-CD44.
Another effect of cyclopamine was the causation of tumor cell apoptosis. As treatment of transformed cells with cyclopamine that caused inhibition of proliferation was reported to have no effect on the viability of cells [27], causation of tumor cell apoptosis by cyclopamine may also be surprising. Our findings on the relative p53 contents of tumor cell nuclei before and after treatment with cyclopamine are against a genotoxic mode of action of cyclopamine in the causation of apoptosis. The antibodies we used for immunostaining can detect not only the increases of p53 due to missense mutations but also the increase of normal p53 that occurs in nuclei prior to the genotoxicity- induced apoptosis [28]. However we found not an increase but rather a decrease of p53 in the nuclei of cyclopamine-exposed tumor cells (Fig. 3f vs g, Table I). In this respect p53 expression has been found to decrease in epidermal cells upon withdrawal from proliferation and commitment to differentiation [29]. We also found greatly decreased p53 in the more differentiated interior cells of trichoepithelioma in comparison to the peripheral cells (Fig. 3h, Table I). Decrease of p53 labelling in the cyclopamine-treated BCCs may therefore also be related to the differentiation that was caused by cyclopamine (Table I). In any case massive apoptotic activity in the cyclopamine-treated tumors despite markedly decreased p53 expression would mean that cyclopamine caused apoptosis of these tumor cells by a non-genotoxic mechanism.
Tumor cells in the BCCs and trichoepithelioma, like in many other tumors, proliferate in the absence of stimulating normal signals and despite inhibitory environmental circumstances, because of the genetic changes they had acquired. Thus inhibition of the hedgehog/ smoothened signaling by cyclopamine in these tumor cells, which are incapable of processing environmental hedgehog signals even when they are present, would have left them void of trophic stimulation and susceptible to apoptosis. Under the same conditions, normal tissue cells, including the putative stem cells, were well preserved. We propose that this selective advantage of the normal tissue cells over tumor cells may apply to tumors in general when subjected to the same type of intervention. The non-normal forcing of proliferation in tumors may result in accumulation of defects in these mutation-prone cells in their ability to respond to the normal tissue signals and, when the signaling pathways they rely upon are altered, they may become more susceptible to apoptosis than the normal cells.
Causations of the inhibition of proliferation and of the differentiation and apoptosis of tumor cells, all with the preservation of normal tissues, make the treatment described attractive for BCCs and other tumors that utilize the hedgehog/smoothened signaling pathway. While we did not observe adverse effects clinically and by microscopy, the known teratogenicity of cyclopamine cautions its use in women of child-bearing age until sufficient relevant data is available. Suitability of skin tumors to topical treatment facilitated our study as we were able to avoid systemic exposure. However cyclopamine is not a DNA-damaging molecule and our finding that normal tissues could be preserved under the conditions that caused differentiation and apoptosis of tumor cells is encouraging further evaluation of cyclopamine on internal tumors which utilize the hedgehog/smoothened pathway for survival and proliferation. n

Acknowledgements. We are grateful to W. Gaffield for the gift of cyclopamine. This work was carried out by the private funds of authors and by use of medical and laboratory facilities outside the university where the first author served on the faculty during parts of the work. We thank N. Özdemir and I. Kuzu for the laboratory facilities. Findings reported herein have been the subject matter of patent application by us (Tas S and Avci O, 02 July 2001, PCT/TR 01/00027); we declare otherwise no conflict of interest.

References

1. Kalderon D. Transducing the hedgehog signal. Cell 2000; 103: 371-4.

2. Gailani MR, Stahle-Backdahl M, Leffell DJ, Glynn M, Zaphiropoulos PG, Pressman C et al. The role of human homologue of Drosophila patched in sporadic basal cell carcinomas. Nat Genet 1996; 14: 78-81.

3. Unden AB, Zaphiropoulos PG, Bruce K, Toftgard R & Stahle-Backdahl M. Human patched (PTCH) mRNA is overexpressed consistently in tumor cells of both familial and sporadic basal cell carcinoma. Cancer Res 1997; 57: 2336-40.

4. Reifenberger J, Wolter M, Weber RG, Megahed M, Ruzicka T, Lichter P, et al. Missense mutations in SMOH in sporadic basal cell carcinomas of the skin and primitive neuroectodermal tumors of the central nervous system. Cancer Res 1998;58:1798-803.

5. Vorechovsky I, Unden AB, Sandstedt B, Toftgard R, Stahle-Backdahl M. Trichoepitheliomas contain somatic mutations in the overexpressed PTCH gene: support for a gatekeeper mechanism in skin tumorigenesis. Cancer Res 1997; 57: 4677-81.

6. Michimukai E, Kitumura N, Zhang Y, Wang H, Hiraishi Y, Sumi K et al. Mutations in the human homologue of the Drosophila segment polarity gene patched in oral squamous cell carcinoma cell lines. In Vitro Cell Dev Biol Anim 2001; 37: 459-64.

7. Sato N, Leopold PL, Crystal RG. Induction of the hair growth phase in postnatal mice by localized transient expression of Sonic hedgehog. J Clin Invest 1999; 104: 855-64.

8. Zhang Y, Kalderon D. Hedgehog acts as a somatic stem cell factor in the Drosophila ovary. Nature 2001; 410: 599-604.

9. Keeler RF. Teratogenic compounds of Veratrum californicum (Durand). VI. The structure of cyclopamine. Phytochemistry 1969; 8: 223-5.

10. Incardona JP, Gaffield W, Kapur RP, Roelink H. The teratogenic veratrum alkaloid cyclopamine inhibits sonic hedgehog signal transduction. Development 1998; 125: 3553-62.

11. Cooper MK, Porter JA, Young KE, Beachy PA. Teratogen mediated inhibition of target tissue to Shh signalling. Science 1998; 280: 1603-7.

12. Detmer K, Walker AN, Jenkins TM, Steele TA, Dannawi H. Erythroid differentiation in vitro is blocked by cyclopamine, an inhibitor of hedgehog signaling. Blood Cells Mol Dis 2000; 26: 360-72.

13. Berman DM, Chen JK, Beachy PA. Inhibition of prostate morphogenesis by the Sonic hedgehog pathway inhibitor cyclopamine. J Urol 2000; 163: 204.

14. Tas S, Avci O. Use of cyclopamine in the treatment of basal cell carcinoma and other tumors. Patent file PCT/TR 01/00027, 02 July 2001 (available at www.wipo.org; correction of the publication error of WIPO in PCT Gazette No. 24/2003, Section II ).

15. Keeler RF. Teratogenic effects of cyclopamine and jervine in rats, mice and hamsters. Proc Soc Exp Biol Med 1975; 149: 302-6.

16. Omnell ML, Sim FR, Keeler RF, Harne LC, Brown KS. Expression of Veratrum alkaloid teratogenicity in the mouse. Teratology 1990; 42: 105-19.

17. DeJong JS, vanDiest PJ, Baak JPA. Number of apoptotic cells as a prognostic marker in invasive breast cancer. Br J Cancer 2000; 82: 368-73.

18. Gallet A, Rodriguez R, Ruel L, Therond PP. Cholesterol modification of hedgehog is required for trafficking and movement, revealing an asymmetric cellular response to hedgehog. Dev Cell 2003; 4: 191-204.

19. Oliver GF, Winkelmann RK. Clear cell basal cell carcinoma: histopathological, histochemical and electron microscopic findings. J Cutan Pathol 1998; 15: 404-8.

20. Yoshimura K, Egewa K, Ono T. Proliferating patterns of basal cell carcinoma-immunohistological study by means of anti-BrdU monoclonal antibody. Nippon Hifuka Gakkai Zasshi 1990; 100: 575-80.

21. Jimenez FJ, Burchette JL, Grichnik JM, Hitchcock MG. Ber-Ep4 reactivity in normal skin and cutaneous neoplasms. Mod Pathol 1995; 8: 854-8.

22. Beer TW, Shepherd P, Theaker JM. BerEp4 and epithelial membrane antigen aid distinction of basal cell, squamous cell and basosquamous carcinomas of the skin. Histopathology 2000; 37: 218-23.

23. Seelentag WKF, Gunthert U, Saremaslani P, Futo E, Pfaltz M, Heitz PU et al. CD44 standard and variant isoform expression in human epidermal skin tumors is not correlated with tumor aggressiveness but down-regulated during proliferation and tumor de-differentiation. Int J Cancer 1996; 69: 218-24.

24. Kooy AJ, Tank B, de Jong AA, Vuzevski VD, vander Kwast TH, van Joost T. Expression of E-cadherin, alpha-& beta-catenin and CD44V6 and subcellular localization of E-cadherin and CD44V6 in normal epidermis and basal cell carcinoma. Hum Pathol 1999; 30: 1328-35.

25. Bonifas JM, Pennypacker S, Chuang PT, McMahon AP, Williams M, Rosenthal A, et al. Activation of expression of hedgehog target genes in basal cell carcinomas. J Invest Dermatol 2001; 116: 739-42.

26. Lyle S, Christofidou-Solomidou M, Liu Y, Elder DE, Albelda S, Cotsarelis G. The C8/144B monoclonal antibody stains the hair follicle bulge cells. J Cell Sci 1998; 111(Pt21): 3179-88.

27. Taipale J, Chen JK, Cooper MK, Wang B, Mann RK, Milenkovic L, et al. Effects of oncogenic mutations in smoothened and patched can be reversed by cyclopamine. Nature 2000; 406: 1005-9.

28. Lakin ND, Jackson SP. Regulation of p53 in response to DNA damage. Oncogene 1999; 18: 7644-55.

29. Dazard JE, Piette J, Basset-Seguin N, Blanchard JM, Gandarillas A. Switch from p53 to MDM2 as differentiating human keratinocytes lose their proliferative potential and increase in cellular size. Oncogene 2000; 19: 3693-705.

A. Non-treated and placebo-treated tumors
	Percentage Of The Tumor Cells Displaying
Tumor and Histology	EA(Ber-EP4)	CD44	p53	Ki-67	Signs of Apoptosis
T1, BCC	97 ± 7	N.D.	71 ± 18	N.D.	0.1 ± 0.2
T2, BCC	96 ± 5	N.D.	50 ± 21	N.D.	0.1 ± 0.3
T5, BCC	100 ± 0	11 ± 5	98 ± 1	30 ± 13	0 ± 0
T6, BCC	100 ± 0	3 ± 4*	4 ± 2	39 ± 16	0.2 ± 0.2
B. Tumors that were exposed to cyclopamine and excised before the complete disappearence of tumor, while under the influence of cyclopamine