The patched polymorphism Pro1315Leu (C3944T) may modulate the association between use of oral contraceptives and breast cancer risk



The gene coding for the human homologue of the Drosophila segment polarity gene patched (PTCH1) is mutated in several common human tumors. In mice, haplodeficiency at the Ptch1 locus results in severe histologic defects in mammary ductal structure. We found no mutations within the coding region of PTCH1 in 17 human primary breast carcinomas. However, the biallelic Pro1315Leu (C3944T) polymorphism of PTCH1 was significantly associated with breast cancer in 41 Bavarian patients compared to 85 healthy controls. We investigated whether this variant influences susceptibility for breast cancer in 611 breast cancer patients diagnosed by age 50 years and 1,057 controls matched by age and study region in Germany and in 1,093 breast cancer patients from the United Kingdom. Allele and genotype frequencies were not different between cases and controls. However, multivariate logistic regression analysis revealed an effect modification of oral contraceptive use (OC) on breast cancer risk by Leu-carrier status. Compared to women who have Pro/Pro and never used OC, Pro/Pro OC users had an increased odds ratio for breast cancer of 1.7. The odds ratio was also 1.7 for Leu-carriers who never used OC, but this was attenuated among Leu-carriers who ever used OC by 20%. The gene-environmental interaction was confirmed in case-only analysis of the German and British studies, yielding an interaction odds ratio of 0.7 for premenopausal women (p = 0.06). Longer duration of pill use was associated with a significantly greater risk reduction (p for trend = 0.015). Our novel observation of a differential effect of OC use on breast cancer risk by PTCH1 1315Leu-carrier status suggests the interesting possibility of the Sonic hedgehog/Patched (SHH/PTCH1) signaling pathway being involved in hormone-induced development of breast carcinoma. © 2002 Wiley-Liss, Inc.

The tumor suppressor gene PTCH1 is a downstream receptor in the hedgehog family of cell signaling proteins and plays an essential role in many aspects of cell growth and cell differentiation. Germline mutations in PTCH1 have been detected in patients with nevoid basal cell carcinoma syndrome (NBCCS), in which patients are predisposed to developmental abnormalities and a variety of neoplasms including basal cell carcinoma and medulloblastoma.1, 2 In addition, somatic mutations in PTCH1 have been identified in sporadic cases of basal cell carcinomas and of medulloblastoma and in a variety of other tumors (see review).3, 4 Interestingly, nonsense mutations in PTCH1 have been reported in 2 of 7 breast carcinomas,5 suggesting an involvement of the SHH/PTCH1 signaling pathway in development of this kind of tumor. Although a similar study in a larger sample of 45 breast carcinomas did not reveal any inactivating PTCH1 mutations,6 the possible involvement of this gene in the formation of breast carcinoma has also been demonstrated in an animal model for Ptch1 haplodeficiency.7 All postpubescent virgin mice heterozygous for Ptch1 developed ductal hyperplasia and dysplasia of the mammary glands, which reverted during late pregnancy and lactation but returned upon involution and gland remodeling. Treatment with estradiol and progesterone enhanced the mutant histologic phenotype. These data suggest a role for the Ptch1 signaling network in mammary growth and differentiation.

To obtain further insights into the involvement of PTCH1 in breast carcinoma formation, we analyzed a series of breast carcinomas for PTCH1 mutations. In the process we observed that the biallelic Pro1315Leu (C3944T) polymorphism in PTCH1 was more common in Bavarian patients with breast tumors than in healthy controls. After using the case-control comparison to generate a hypothesis for an association, we employed extended epidemiologic studies to confirm the initial observations.


Tumor samples

Seven ductal and 10 lobular breast carcinomas were selected from the archives of the Institute of Pathology of the Technical University of Munich, Bavaria, Germany. In all 17 cases, parts of the tumor tissue had been frozen immediately after surgery and stored in liquid nitrogen until further use. Formalin-fixed paraffin blocks were available for another 5 ductal and 19 lobular breast carcinomas, resulting in an overrepresentation of lobular carcinoma cases (total 29). DNA was extracted from microdissected cells of interest.

Patients and controls

Bavaria, Germany.

The Bavarian sample consisted of the above-mentioned 41 patients with breast carcinoma and 85 healthy female inhabitants of Bavaria aged 25–74 years used as controls. The women were identified through the population-based KORA (Kooperative Gesundheitsforschung in der Region Augsburg) registry. Informed and written consent was obtained from all women. Our study was approved by the ethical committee of the Bayerische Landesärztekammer.

Baden-Württemberg, Germany.

The population-based case-control study consisted of German-speaking women with no former history of breast cancer and residing in 2 study regions covering a population of about 2.2 million in the State of Baden-Württemberg in Germany.8 Cases were patients up to the age of 50 years at the time of diagnosis of incident in situ or invasive breast cancer and complete ascertainment was achieved by surveying all hospitals serving the population. Two controls per case matched by exact age and study region were selected from random lists of residents supplied by the population registries and contacted by letter. Overall, 706 (70% of eligible) cases and 1,381 (61% of eligible) population controls participated by completing the study questionnaire for assessing demographic, anthropometric and other known or putative risk factors and providing a blood sample. Our study was reviewed and approved by the ethics committee of the University of Heidelberg. Informed written consent was obtained from each patient.

Blood samples were obtained from 95% of cases and 82% of controls. Restricted to women who have at least 1 parent of German nationality, comprising 91% of cases and 96% of controls, and for whom genotyping was successful, this analysis is based on 611 cases and 1,057 controls.

East Anglian, United Kingdom.

Cases (n = 1,093) were drawn from the Anglian Breast Cancer Study. This is an ongoing population-based study of breast cancer cases ascertained through the East Anglian Cancer Registry.9 All patients diagnosed below age 55 years since 1991 and still alive in 1996 (prevalent cases, median age 48 years), together with those below 65 years diagnosed from 1996–1999 (incident cases, median age 52), were eligible to take part. All study participants completed an epidemiologic questionnaire and provided a blood sample for DNA extraction. Participation rate was 70%. Of the 1,093 DNA samples, 161 [14.7%] could not be amplified by PCR. Ethical approval was obtained from the Anglia and Oxford Multicentre Research Committee in the UK. Informed written consent was obtained from each patient. These samples were used for the case-only analysis.

Statistical analysis

The observed genotype distributions were used to calculate the allelic frequencies. The χ2 test was used to assess deviations in the distribution of the genotype frequencies in the controls from those expected under Hardy-Weinberg equilibrium (HWE). Allele and genotype frequencies in cases and controls were compared by χ2 tests. We assessed the association between the risk of breast cancer and the PTCH1 genotypes using conditional logistic regression to obtain maximum likelihood estimates for the odds ratio and 95% confidence intervals (PROC PHREG, SAS 6.12). Relevant risk factors were accounted for in multivariate analysis. First order interactions between risk factors (OC use, breastfeeding, parity, menopausal status) and genotypes of the PTCH1 polymorphism were estimated under the standard multiplicative model.

Case-only analysis was also used to measure gene-environmental interaction, i.e., when the effect of genotype on disease risk depends on the level of exposure to an environmental factor or vice versa.10 Cases with and without the susceptibility genotype are compared with respect to the prevalence of the environmental exposure. The case-only study has been shown to provide more power to detect gene-environmental interaction and better precision for effect estimation due to elimination of control group variability.11 This approach however holds only under the assumption that genotype and environmental exposure are uncorrelated. Therefore, we first checked independence by χ2 tests among the controls. For a case-only analysis, the Pro1315Leu (C3449T) PTCH1 polymorphism was used as a dependent variable since this has to be binary and its dependence on other risk factors was assessed using unconditional logistic regression (PROC LOGISTIC, SAS 6.12). In the joint analysis, study center was incorporated in the regression model for adjustment. For the analysis by menopausal status, women who have had a hysterectomy but not oophorectomy, menopause was assumed to occur at the median age of menopause for the other cases.

DNA extraction

DNA was extracted from microdissected tumor or from normal breast tissue. Samples were sectioned at 10 μm, mounted on glass slides and neoplastic (consisting of a pure population of 90–95% of tumor cells) or normal tissue was microdissected using the Palm Laser-MicroBeam System (P.A.L.M., Bernried, Germany). After selecting the cells of interest, adjacent cells were photolysed by the microbeam. At least 1,000 selected cells (either tumor or normal glandular cells) were picked from the slides using conventional sterile needles and transferred into reaction tubes containing 200 μl STE buffer (20 mM Tris/HCl pH 8.0, 10 mM NaCl, 10 mM EDTA, 0.5% w/v sodium dodecyl sulfate and 1 mg/ml proteinase K) for isolation of genomic DNA. Following protein digestion (55°C, overnight) DNA was extracted from STE buffer by phenol-chloroform extraction according to standard procedures.

PCR and sequencing analysis

All 23 PTCH1 exons and the flanking exon/intron junctions were amplified by PCR from genomic DNA of the 17 (in liquid nitrogen frozen) breast carcinoma. The primer pairs used to analyze exons 3–11, 13, 14 and 16–22 were as described in Xie et al.5 Exons 1 and 2 as well as exon 12 were amplified using primers as described in Hahn et al.1 Primers for exon 15 were the same as primer pair 14 in Hahn et al.1 Exon 23 was amplified in 2 overlapping PCR products. The first fragment was amplified with primer 23F (5′ CCC TTC TAA CCC ACC CTC ACC CTC) and 23R.2 (5′ AAA CAG GCC GTG GTC AGT CTC ACC). The second fragment was amplified with primer 23F.3 (5′ CCG CGC AGA GAC GCT TTT GAA ATT) and 23R (5′ GAC ACA TCA GCC TTG CTC). Each PCR-product was sequenced in both directions using the same primers and the Big Dye Terminator Cycle Sequencing kit (Applied Biosystems, Weiterstadt, Germany). Sequencing analysis was performed on an automatic DNA sequencer (Applied Biosystems, Weiterstadt, Germany).

Constitutional DNA isolated from normal tissue of the described 17 cases and from additional 24 formalin-fixed breast carcinoma cases (see Samples) was investigated for the C3944T polymorphism by amplification and sequencing of exon 23 using the primer pair 23F and 23R.2.

Allelic discrimination by TaqMan

The C3944T polymorphism in exon 23 of PTCH1 in the German and British samples was genotyped using a TaqMan assay. The primer pair used for allelic discrimination assay was ex23.TaqF: 5′-GGG ACC CCC CCA GAG AA-3′ and ex23.TaqR: 5′-CAG AAT GCC CTT CAG TAG AAA TTT C-3′. To detect the polymorphism at position 3944 (C or T) of the human PTCH1 gene the flourigenic probes ex23.P1: FAM-TTG TGG CCA CCC CCC TAC AGA C-TAMRA and ex23.P2: VIC-CTT GTG GCC ACC CCT CTA CAG ACC-TAMRA were employed. PCR was carried out with the TaqMan Universal PCR Master Mix (PE, Applied Biosystems) using 50 ng of genomic DNA, 200 nM of each probe and 300 nM forward primer and reverse primer in a 30 μl final reaction mixture. After 2 min incubation at 50°C, AmpliTaq Gold was activated by incubation for 10 min at 95°C. Each of the 40 PCR cycles consisted of 15 sec denaturation at 95°C and hybridization of probe and primers for 1 min at 61°C. DNA samples with known genotypes were used as controls.

Sequencing of 700 randomly selected samples using the exon 23-specific primers 23F and 23R.2 revealed 5 samples that showed a different genotype in comparison to TaqMan analysis. For these samples, the sequencing results were used for statistical analysis.


Mutation screen of the PTCH1 coding region in 17 human breast carcinomas

To determine whether PTCH1 was inactivated by mutations in breast carcinoma, we amplified and sequenced the entire coding region of the PTCH1 gene in 17 tumors. Apart from nucleotide polymorphisms previously found in PTCH1,12 the mutation screening did not detect any new nucleotide changes in the protein-coding region of PTCH1 in any of the cases examined. These results suggest that primary sporadic breast carcinomas do not contain frequent inactivating mutations in the coding region of PTCH1. However, 13 of the 17 samples (76%) examined were heterozygous for the nucleotide polymorphism C3944T within exon 23, which leads to a Pro/Leu aa-exchange at position 1315 of the PTCH1 protein.

Association of the biallelic Pro1315Leu (C3944T) variant of PTCH1 with breast cancer risk

We sequenced exon 23 in additional 24 Bavarian women with breast carcinoma and found the biallelic Pro1315Leu variant of PTCH1 in 36 (88%) patients with breast carcinoma, including the 17 patients previously screened for mutations. This variant was present in 47% of the 85 controls (Table I), therefore, the genotype distribution was significantly different between the 2 groups (p < 0.001).

Table I. PTCH1 Pro1315Leu (C3944T) Polymorphism (as Per Genbank Entry U59464): Allele Frequencies and Odds Ratios for Breast Cancer in German Breast Cancer Patients and in Population Controls
 Total nFrequency of the PTCH1 C3944T genotypesFrequency of risk allele T (95% CI)p-valuea
CC (Pro/Pro) n (%)CT (Pro/Leu) n (%)TT (Leu/Leu) n (%)
  • 1

    Odds ratio and 95% confidence interval; univariate OR did not differ from OR adjusted for number of full-term pregnancies (0, 1–2, ≥3), age at menarche (≤12, ≥13), duration of breastfeeding (continuous), menopausal status, first-degree family history of breast cancer and alcohol consumption (1–18, ≥19 g/day).

  • 2p-value for Chi-squared test for difference in genotype distribution between cases and controls.

 Controls8545 (52.9)36 (42.4)4 (4.7)0.26 (0.19–0.32) 
 Cases415 (12.2)35 (85.4)1 (2.4)0.45 (0.34–0.56)0.001
Baden-Württemberg (DKFZ study)      
 Controls1,057452 (42.8)470 (44.5)135 (12.8)0.35 (0.33–0.37) 
 Cases611267 (43.7)273 (44.7)71 (11.6)0.34 (0.31–0.37)0.65
OR1 1.00.970.870.95 
 (95% CI)  (0.78–1.20)(0.63–1.21)(0.77–1.21) 

The association of the biallelic Pro1315Leu polymorphism of PTCH1 with breast carcinoma was then evaluated in the population-based case-control study of breast cancer by age of 50 years in Baden-Württemberg (hence DKFZ study). There was no significant difference in allele frequency or genotype distribution between the 611 breast cancer cases and 1,057 age-matched population controls (Table I). The allele frequencies did not deviate from the expected Hardy-Weinberg distribution. However, the frequency of the T-allele in the Baden-Württemberg population was significantly different from the Bavarian controls (p = 0.04). To detect possible effects of the polymorphism in specific subgroups of patients, we examined the distribution of genotypes in the breast cancer cases according to histologic type, grading and estrogen and progesterone receptor status and we detected no appreciable differences between any of the subgroups (data not shown).

Ovarian hormones contribute to the development of ductal hyperplasia and dysplasia of the mammary glands in mice haplodeficient in the Ptch1 gene. Therefore, we explored the effect of the Pro1315Leu polymorphism on disease risk depending upon the exposure to exogenous hormones. Whereas parity reduced risk of breast cancer, neither ever use of OC nor duration of OC use was associated with breast cancer risk in the entire study population (data not shown). We did not observe any interaction between parity and the Pro1315Leu variant. However, multivariate analysis revealed a differential effect of carrier status of 1315Leu on pill use and breast cancer risk (Table II). Compared to women who have Pro/Pro and never used OC, women who have Pro/Pro genotype and ever used OC had an increased risk for breast cancer (OR 1.7; 95% confidence interval [CI] 1.1–2.7). A risk elevation of similar magnitude was associated with being a carrier of 1315Leu and never OC user (OR 1.7; CI 1.0–2.8) and this was attenuated among those 1315Leu-carriers who ever used oral contraceptives (OR 1.4; CI 0.6–1.2), i.e., a risk reduction by 20%. The interaction odds ratio between Leu-carrier status and ever use of OC was 0.51 and statistically significant (p = 0.017). Furthermore, we observed an influence of the duration of OC use whereby more than 10 years of oral contraceptives use was associated with a greater risk reduction among 1315Leu-carriers (Table II).

Table II. Odds Ratios for Breast Cancer Associated with Oral Contraceptive (OC) Use According to PTCH1 Pro1315Leu (C3944T) Polymorphism in the DKFZ Study
VariablePTCH1 noncarrier (Pro/Pro)PTCH1 Leu-carrier (Pro/Leu or Leu/Leu)
Cases (n = 267)Controls (n = 452)Odds ratio1 (95% CI2)Cases (n = 344)Controls (n = 605)Odds ratio1 (95% CI2)Odds ratio13 (95% CI2)
  • 1

    Adjusted for number of full-term pregnancies (0, 1–2, ≥3), age at menarche (≤12, ≥13), duration of breastfeeding (continuous), menopausal status, first-degree family history of breast cancer and alcohol consumption (1–18, ≥19 g/day); reference group is PTCH1 noncarrier of T-allele without OC use.

  • 2

    CI, confidence interval.

  • 3

    Odds ratio calculated with reference group of PTCH1 1315Leu-carrier without OC use.

  • 4

    Includes 2 persons with missing data on ever/never OC use.

OC use       
 Never35891.06410741.7 (1.0–2.8)1.0
 Ever2323631.7 (1.1–2.6)2804981.4 (0.9–2.2)0.8 (0.6–1.2)
Duration of OC use       
 Never35891.06410741.8 (1.1–2.9)1.0
 1–10 years1232091.6 (1.0–2.6)1572871.4 (0.9–2.1)0.9 (0.6–1.3)
 11+ years1011421.9 (1.1–3.0)1052001.2 (0.8–1.9)0.8 (0.5–1.1)
 Missing duration812 1811  

Case-only analysis of Pro1315Leu (C3944T) status, use of oral contraceptives and breast cancer risk

After showing that OC use and PTCH1 genotype are independently distributed in the German population controls, we used a case-only analysis of the DKFZ study to confirm the results of the case-control analysis, suggesting an interaction between 1315Leu-carrier status and the use of oral contraceptives. The interaction odds ratio of OC use by 1315Leu-carrier status was 0.6 (CI 0.4–1.0), being stronger for premenopausal women (OR 0.5; CI 0.3–0.9) and very similar to the interaction odds ratio calculated in the case-control analysis (Table III).

Table III. Case Only Analysis of Oral Contraceptive Use and PTCH1 Pro1315Leu Polymorphism in Breast Cancer Patients from Germany (DKFZ Study) and the United Kingdom
OC usePremenopausalPostmenopausal
Leu-carrierNon carrierOR1 (95%CI2)Leu-carrierNon carrierOR1 (95%CI2)
  • 1

    Adjusted for age in 2-year intervals, study center, number of full-term pregnancies (0, 1–2, ≥3), age at menarche (≤12, ≥13), first-degree family history of breast cancer, alcohol consumption (1–18, ≥19 g/day).

  • 2

    CI, confidence interval.

  • 3

    Includes persons with missing data on ever/never OC use (premenopausal: 5 carrier, 1 noncarrier; postmenopausal: 7 carrier, 7 noncarrier).

 Ever use2692220.5 (0.3–0.9)1110
  1–10 years use1481180.6 (0.3–1.0)95
  11+ years use103960.5 (0.3–0.9)25
  Duration missing188 00 
United Kingdom      
 Ever use2031670.9 (0.6–1.5)157981.1 (0.7–1.7)
  1–10 years use1461120.9 (0.6–1.6)124711.2 (0.8–1.9)
  11+ years use55540.8 (0.4–1.4)33270.7 (0.4–1.5)
  Duration missing21 00 
Germany + United Kingdom      
 Ever use4723890.7 (0.5–1.0)1681081.1 (0.7–1.6)
  1–10 years use2942300.7 (0.5–1.1)133761.2 (0.8–1.8)
  11+ years use1581500.6 (0.4–0.9)35320.7 (0.4–1.3)
  Duration missing209 00 

An analogous analysis of the British patients did not reveal significant interaction between OC use and T-carrier status although the odds ratios were below 1 for premenopausal women and there was a tendency to further reduction with longer duration of OC use (Table III).

The joint case-only analysis of the DKFZ study and British study by menopausal status yielded an interaction odds ratio of 0.7 for premenopausal women with borderline significance (p = 0.06), such that the risk of developing breast carcinoma among 1315Leu-carriers was reduced by 30% with ever use of oral contraceptives. There was a significant greater risk reduction associated with longer duration of pill use (p for trend = 0.015). This gene-environmental interaction was not apparent among postmenopausal women, although the interaction odds ratio was decreased for more than 10 years of OC use (Table III).


Nonsense mutations in the SHH/PTCH1 signaling pathway have previously been reported in a small series of 7 breast carcinomas.5 We were not able to detect any mutations within the protein-coding region of PTCH1 in 17 breast carcinomas by direct sequencing. Our findings corroborate results of a recent study, which detected no mutations in PTCH1 in 45 breast carcinomas using SSCP analysis.6 In both studies, the regulatory elements and the introns of PTCH1 remained unscreened and some mutations may have been overlooked. However, one does not expect mutational hotspots in these elements because mutations are distributed evenly within the PTCH gene.4 The data thus suggest that it is very unlikely that PTCH1 is inactivated in a substantial fraction of breast carcinoma.

Based on a comparison of 41 breast cancer patients from Bavaria with 85 healthy controls, we made the interesting initial observation that the biallelic variant C3944T of the PTCH1 gene was significantly associated with breast carcinoma. We attempted to confirm this observation using 2 population-based case-control studies. However, we did not find any association of the biallelic variant Pro1315Leu (C3944T) with risk for breast cancer in the DKFZ study of breast cancer by age 50 years conducted in Baden-Württemberg. Of the initial series of Bavarian breast cancer patients, only 25% were diagnosed by the age of 50 years. On the other hand, all of the patients of the DKFZ study and almost 40% of the patients from the British study were diagnosed by 50 years. The allele frequency (T-allele) of 0.36 (95% CI 0.34–0.38) in the British patient is very similar to that of 0.34 (95% CI 0.31–0.37) in patients of Baden-Württemberg and not statistically significantly different from the frequency of 0.45 (95% CI 0.34–0.56) in the series of Bavarian patients. We did not observe differences in allele frequency by histologic type, grading and estrogen and progesterone receptor status.

However, a significant difference in the allele frequency was found between the 2 controls series, 0.35 (95% CI 0.33–0.37) in Baden-Württemberg and 0.26 (95% CI 0.19–0.32) in Bavaria. It seems unlikely that the allele frequency estimate based on 1,057 controls does not truly reflect the population frequency unless age was a modifier. The initial case-control difference was probably a spurious finding as a result of small study size and convenient sampling, which is not validated by the subsequent, better designed studies, as has been often observed.13 However, we cannot exclude a real difference in C3944T allele frequency between the 2 populations in Germany and the possibility that the gene effect is stronger in some subpopulations than in others.

Although we did not find any association of the biallelic variant Pro1315Leu (C3944T) with risk for breast cancer in the DKFZ population-based case-control study, we observed a significant interaction between 1315Leu-carrier status and use of oral contraceptives. This was confirmed in case-only analysis of the DKFZ study. The joint analysis including German and British patients yielded interaction odds ratios of borderline significance in premenopausal women, although the interaction was much weaker for the British study. However, there was no evidence of heterogeneity between the results of the 2 studies. This is the first report of a possible differential effect of OC use on breast cancer risk by 1315Leu-carrier status in the human PTCH1 protein, predominantly in premenopausal women.

The observation that Leu-carrier status at position 1315 of PTCH1 might lower the incidence of breast carcinoma in individuals who use oral contraceptives is highly interesting. There is compelling evidence from epidemiologic and experimental data that implicate endogenous estrogen levels as critical determinants of breast cancer risk.14, 15 Exogenous hormones used for contraception have thus been suggested for a long time to be related to breast cancer risk, although recent results from pooled analysis indicate only weak effects.16 It has recently been shown that PTCH1 is involved in response to steroid hormones such as estrogens and progesterone and also to the precursor of both of them, which is cholesterol.17 Furthermore, haploinsufficiency at the murine Ptch1 locus in heterozygous postpubescent virgin animals was found to result in severe histologic defects in the mammary gland reminiscent of the histology of human ductal carcinoma in situ (DCIS).7 Such changes suggest a loss of contact inhibition commonly associated with uncontrolled neoplastic cell division and have remarkable similarity to outgrowths of transplants of hormone-dependent tumors that arise during pregnancy and lactation.18 While hormone-dependent tumors worsen during pregnancy, Ptch1-induced dysplasia reverts to wild type during late pregnancy and lactation but returns upon involution and gland remodeling. The histologic defects can be mimicked by treatment of heterozygous mice with both estrogen and progesterone hormones. This suggests that murine Ptch1 plays a role in the response to ovarian hormones.7 Another connection between the Ptch1 signaling and ovarian hormones is the expression of one of the Ptch1 ligands Indian hedgehog (Ihh), which is induced by progesterone in the murine uterus.19 The authors proposed that this increase results in increased proliferation of the uterine stroma.

Interaction between Ptch1 and steroids is also suggested by the sterol-sensing domain of Ptch1, which is essential for its activity.20 The plant steroidal alkaloid, cyclopamine, can induce cyclopia and other birth defects by blocking the SHH/PTCH1 signaling pathway.21, 22, 23 Altogether the data show that sterols including hormones have an enormous impact on the SHH/PTCH1 pathway and may regulate its signaling function.

The C3944T polymorphism results in an amino acid exchange in the C-terminus of PTCH1. This C-terminus was recently identified as an important regulatory region and is essential for proper SHH/PTCH1 signaling.24 Therefore, an amino acid exchange in the regulatory C-terminal end of PTCH1 in combination with sterols may alter its function and may result in a change of PTCH1 activity and thus modulate the effect of exogenous hormones on tumor development. The observation that the combination of 2 separate risk factors for developing breast carcinoma, namely 1315Leu-carrier status and OC intake, attenuates the risk seems initially conflicting. However, it may be that 1315Leu-carrier status in PTCH1 normally results in a higher activity of the HH/PTCH-signaling pathway, which together with other factors may lead to enhancement of cellular proliferation. Since PTCH1 is involved in the response to steroid hormones,17 it may be possible that the combination of 1315Leu-carrier status and OC interact in such a way that the proliferative function of the pathway is reduced and the risk for developing breast carcinoma is attenuated.

In conclusion, our observation that Leu-carrier status at position 1315 in the human PTCH1 protein may modify the effect of oral contraceptives on breast cancer risk suggests the interesting possibility of the involvement of the SHH/PTCH1 signaling pathway in hormone-induced development of breast carcinoma and thus warrants further investigation.


We thank Ms. U. Eilber for competent data management and Dr. W. Koch, Deutsches Herzzentrum München, Munich, for technical advice. We are grateful to Dr. L. Wojnowski for comments on the manuscript. This work was supported by a BioFuture-Grant of the German Ministry for Education and Research BMBF (to U.S. and H.H.) and the epidemiologic case-control study was supported by the Deutsche Krebshilfe (Project number 70492).