October 22, 2003

Media Contact: Joanna Downer
410-614-5105
Email: jdowner1@jhmi.edu

EMBRYONIC PATHWAY CRITICAL TO GROWTH OF DIGESTIVE TRACT TUMORS

A common signal critical for normal embryo development in many species also contributes to cancers of the esophagus, stomach and pancreas in people, according to Johns Hopkins researchers. Upwards of 50,000 cancer deaths a year may now be partly attributable to this pathway's activity, say the researchers. Their report appears in the Oct. 23 issue of Nature.

The signal, called Hedgehog, tells cells when and where to grow during embryonic development and is turned on in primitive cells, or stem cells, in adult tissues to trigger tissue repair. Researchers at Hopkins and elsewhere have already linked Hedgehog and its signaling pathway to a non-fatal skin cancer (basal cell carcinoma), a deadly lung cancer and the most common childhood brain cancer (medulloblastoma).

"Blocking this signal may one day help treat cancers for which there are currently few or no mechanism-based therapies," says senior author Philip Beachy, Ph.D., professor of molecular biology and genetics in Hopkins' Institute for Basic Biomedical Sciences and a Howard Hughes Medical Institute investigator. "For right now, the biggest question is whether it will pan out in people."

In experiments with cancer cell lines and tumor samples from patients, the scientists found that Hedgehog's signal is required for the cancers' growth. Moreover, a three-week course of a plant-derived chemical called cyclopamine, known to block Hedgehog, killed these cancers when grown in mice, causing no apparent harm to the animals.

"In mice, blocking the Hedgehog signal made the implanted tumors disappear," says the study's first author, David Berman, M.D., Ph.D., assistant professor of pathology at Hopkins. "It's been about three and a half months since we stopped the cyclopamine, and still the tumors haven't returned."

Unfortunately, cyclopamine is unlikely to be useful for patients because there just isn't enough of it, so the search is on to find Hedgehog blockers that could be made in quantities necessary for human studies, say the researchers.

The researchers checked for Hedgehog activation in cell lines and fresh samples of digestive tract tumors because the gut comes from the same part of the embryo -- the endoderm -- as the lung, says Anirban Maitra, MBBS, assistant professor of pathology. Earlier this year, a team from Hopkins linked Hedgehog activation to small cell lung cancer, providing reason to anticipate Hedgehog's involvement in a variety of other cancers, notes Berman.

"Because of Hedgehog's important roles in these tissues during development, we hypothesized that 'reactivation' of the pathway occurs in adult life during cancer development in these organs," adds Maitra, whose procedure for obtaining fresh samples from surgically removed tumors provided the opportunity to analyze cancers unaltered by years of laboratory growth. "Our studies prove this hypothesis to be true."

The pathway's link to another batch of cancers support the idea that cancer may arise -- in part -- from abnormal growth of stem cells inside mature organs.

The scientists speculate that primitive cells in the lining of the digestive tract may turn on the normal Hedgehog pathway to repair tissue damaged by long-term exposure to an environmental toxin or irritant, such as excess stomach acid chronically rising into the esophagus.

If the damaging environmental irritant is also carcinogenic, such as tobacco smoke, the chances go up that these long-lived primitive cells eventually may collect the right genetic mutations to trigger cancer development, suggests Beachy.

Authors on the study linking Hedgehog to digestive tract tumors are Berman, Maitra, Beachy, Sunil Karhadkar, Rocio Montes de Oca, Meg Gerstenblith, Antony Parker and James Eshleman of the Johns Hopkins School of Medicine; Kimberly Briggs and Neil Watkins of the Johns Hopkins Kimmel Cancer Center; and Yutaka Shimada of Kyoto University, Japan. The project was funded by the family of Margaret Lee and by grants from the National Institutes of Health.

A related paper, focusing on pancreatic cancer and written by researchers at the University of California at San Francisco and Harvard Medical School, appears in the same issue of the journal.

Under a licensing agreement between Curis Inc. and The Johns Hopkins University, Beachy and the University hold equity in Curis and are entitled to a share of royalties from sales of the products described in this article. Beachy also receives payment and equity for service as a consultant to Curis Inc. The terms of this arrangement are being managed by The Johns Hopkins University in accordance with its conflict of interest policies.

--JHMI--

On the Web: http://www.nature.com/nature

Widespread requirement for Hedgehog ligand stimulation in growth of digestive tract tumours

DAVID M. BERMAN1,2,3 (Genentech Consultant),*, SUNIL S. KARHADKAR1,2,*, ANIRBAN MAITRA2,*, ROCIO MONTES DE OCA1, MEG R. GERSTENBLITH1, KIMBERLY BRIGGS3, ANTONY R. PARKER2, YUTAKA SHIMADA4, JAMES R. ESHLEMAN2, D. NEIL WATKINS3 & PHILIP A. BEACHY1 (Curis & Genentech Consultant)

1 Department of Molecular Biology and Genetics and Howard Hughes Medical Institute, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA
2 Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA
3 Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA
4 Department of Surgery, Kyoto University, Kyoto 606-8507, Japan
* These authors contributed equally to this work

Correspondence and requests for materials should be addressed to P.A.B. (pbeachy@jhmi.edu).

Activation of the Hedgehog (Hh) signalling pathway by sporadic mutations or in familial conditions such as Gorlin's syndrome is associated with tumorigenesis in skin, the cerebellum and skeletal muscle1, 2. Here we show that a wide range of digestive tract tumours, including most of those originating in the oesophagus, stomach, biliary tract and pancreas, but not in the colon, display increased Hh pathway activity, which is suppressible by cyclopamine, a Hh pathway antagonist. Cyclopamine also suppresses cell growth in vitro and causes durable regression of xenograft tumours in vivo. Unlike in Gorlin's syndrome tumours, pathway activity and cell growth in these digestive tract tumours are driven by endogenous expression of Hh ligands, as indicated by the presence of Sonic hedgehog and Indian hedgehog transcripts, by the pathway- and growth-inhibitory activity of a Hh-neutralizing antibody, and by the dramatic growth-stimulatory activity of exogenously added Hh ligand. Our results identify a group of common lethal malignancies in which Hh pathway activity, essential for tumour growth, is activated not by mutation but by ligand expression.

The Hh signalling pathway specifies patterns of cell growth and differentiation in a wide variety of embryonic tissues. Mutational activation of the Hh pathway, whether sporadic or in Gorlin's syndrome, is associated with tumorigenesis in a small subset of these tissues, predominantly skin, the cerebellum and skeletal muscle1, 2. Mutations that activate the Hh pathway include those that impair the ability of the transporter-like Hh receptor3 Patched (PTCH, the target of Gorlin's syndrome mutations) to restrain Smoothened (SMO)-mediated activation of transcriptional targets through the Gli family of latent transcription factors1, 2, 4, 5. Paradoxically, Hh pathway activity is associated with increased expression of PTCH, which is a transcriptional target of the pathway but is unable to restrain SMO when bound by Hh protein. Pathway activation, whether triggered by Hh binding or by PTCH mutation, requires SMO, a seven-transmembrane protein that binds to and is inactivated by the pathway antagonist cyclopamine6.

The recent finding that Hh pathway activity is important for growth of a significant proportion of small-cell lung cancers7, a tumour type not associated with Gorlin's syndrome, suggested that other, non-Gorlin's tumours might require Hh pathway activity for growth. We investigate here the role of pathway activity in tumours derived from the gut, a tissue with prominent and diverse roles for Hh signalling in developmental patterning, and in mature tissue homeostasis. A role in homeostasis is suggested by the expression of Hh ligands and target genes in postnatal gut epithelium and mesenchyme8-11 (Fig. 1a).

Figure 1 Hh pathway activity in normal and neoplastic gut cells and tissues. Full legend

High resolution image and legend (168k) See link at end to obtain pictures

We began our examination of gut-derived tumours by testing for expression of Sonic hedgehog (SHH) and Indian hedgehog (IHH), which encode members of the Hh ligand family that are expressed in early endoderm and throughout gut development10, 12. We detected SHH and IHH messenger RNA in 37 of 38 (97%) cell lines from oesophageal, stomach, biliary tract, pancreatic and colon carcinomas (Fig. 1b). The Hh target genes PTCH and GLI were used as indicators of Hh pathway activity and were co-expressed in most cell lines from oesophageal (4/6), stomach (6/6), pancreatic (5/6) and biliary tract (5/9) tumours. By contrast, PTCH was not expressed in colon tumour cell lines (0/11), although a few of these cell lines expressed GLI in the absence of PTCH. The Hh pathway, however, is not active in these colon-tumour-derived cell lines, as indicated by our reporter assays (see below).

Expression of PTCH and GLI within cells from several types of digestive tract tumours that also express Hh ligands suggests the autonomous operation of an active signalling process. Autonomous pathway activity was confirmed by the high level of luciferase activity produced by an exogenously introduced Hh-inducible GLI–luciferase reporter13 in all cell lines producing detectable PTCH mRNA (Fig. 2a). Furthermore, Hh pathway activity in these cell lines was inhibited in a dose-dependent manner by the Hh-pathway-specific antagonist cyclopamine, but not by tomatidine, an inactive but structurally related compound14 (Fig. 2a). These results suggest that high levels of Hh pathway activity are a common feature of digestive tract tumours and prompted us to further investigate the role of the pathway in tumour growth. We found that cyclopamine treatment reduced the growth of tumour cell lines from the oesophagus, stomach, biliary tract and pancreas by 75–95% compared with tomatidine controls (Fig. 2b). Strikingly, significant growth inhibition was observed only in tumour lines expressing PTCH mRNA, confirming that the effects of cyclopamine treatment are pathway specific rather than generally cytotoxic.

Figure 2 Cyclopamine suppression of Hh pathway activity and growth in digestive tract tumour cell lines correlates with expression of PTCH mRNA. Full legend

High resolution image and legend (92k)

Because the properties of cell lines adapted to long-term growth in vitro can differ from those of tumours growing in vivo, we also examined pathway activation in freshly resected stomach and pancreatic tumours by measuring endogenous PTCH mRNA levels. For each specimen, RNA for quantitative polymerase chain reaction with reverse transcription (RT–PCR) analysis was isolated from 10 consecutive 10 µm cryosections after histological analysis of the immediately flanking sections to determine tumour content. We found that PTCH mRNA levels were 23–271 times (mean = 129; n = 9) higher in stomach tumours and 69–5,044 times (mean = 448; n = 15) higher in pancreatic tumours than in adjacent normal tissue (Fig. 3a).

Figure 3 Hh pathway activity and requirement for growth of tumour cells in vivo. Full legend

High resolution image and legend (60k)

To examine the role of Hh pathway activity in growth, we analysed pancreatic carcinomas that had been passaged once as xenografts in nude mice then cultured and immediately assayed in vitro. Of six such xenografts, four expressed PTCH mRNA (data not shown); two of these were a matched pair of primary and metastatic tumours from a single patient. All four of these PTCH-expressing primary xenografts expressed the GLI–luciferase reporter in a cyclopamine-sensitive manner (Fig. 3b). Cyclopamine treatment of these PTCH mRNA-expressing xenografts also resulted in decreased viable cell mass (Fig. 3c), demonstrating more extreme cell-killing effects of Hh pathway blockade than those observed in established tumour cell lines (Fig. 2b). By contrast, single-passage xenografts lacking PTCH mRNA grew equally well in control and cyclopamine-containing media (Fig. 3c), again confirming that cyclopamine effects are pathway specific rather than generally cytotoxic.

To examine the effects of cyclopamine treatment in vivo, subcutaneous xenografts from HUCCT1 cells, a metastatic cholangiocarcinoma cell line, were established in athymic mice. After the tumours had grown to an average size of 180 mm3, mice bearing these tumours were injected daily with cyclopamine or vehicle, and control tumours continued to grow until animals were euthanized at 22 days with an average tumour volume exceeding 800 mm3 (Fig. 3d, e). Tumours in cyclopamine-treated animals, by contrast, regressed completely by 12 days. Treatment continued for a further 10 days, followed by an observation period of 98 days. Remarkably, all treated tumours were grossly and histologically undetectable at the end of the 3 month observation period, indicating a complete and durable tumoricidal effect of blocking the Hh pathway with cyclopamine (Fig. 3d, e). As previously reported for shorter treatment courses7, 15, all mice survived cyclopamine treatment and the observation period with no obvious adverse effects.

These findings establish that the Hh pathway is widely activated in gut-derived tumours, and further demonstrate a role for pathway activity in tumour cell growth in vitro and in vivo. Yet Gorlin's syndrome is not associated with a higher incidence of gut-derived tumours, and PTCH mutations in these tumours have not been reported, which suggests a distinct mechanism for Hh pathway activation that does not involve mutation of pathway components. Given that SHH and IHH mRNA are expressed in nearly all gut-derived tumours examined, we investigated the role of Hh ligand binding in pathway activity. We measured Hh-inducible reporter activity in cells from tumours of the oesophagus, stomach, pancreas and biliary tract treated with 5E1 monoclonal antibody16, which binds to Shh and Ihh ligands17 and blocks signalling by disrupting ligand binding to Ptch18. Autonomous activation of the transfected reporter was not affected by the control antibody, but was markedly reduced by incubation with 5E1 at 0.1 or 10 µg ml-1 (Fig. 4a and Supplementary Fig. 1a). By contrast, reporter activity was augmented by around eightfold by the addition of purified Shh ligand to a concentration of 25 nM. Addition of 5E1 in combination with Shh ligand reduced reporter activity to a level intermediate between those produced with either reagent alone (Fig. 4a and Supplementary Fig. 1a), which indicates a mutual antagonism between 5E1 antibody and Hh ligand in activating the pathway.

Figure 4 Ligand dependence of Hh pathway activity and growth in digestive tract tumours. Full legend

High resolution image and legend (85k)

Reporter activity in cells from single-passage pancreatic cancer xenografts was also antagonized by 5E1 (Fig. 4b). In addition, treatment with the 5E1 antibody markedly reduced viable cell mass (Fig. 4c). This cell-killing effect and the reporter effect were observed exclusively in cells from tumours that expressed endogenous PTCH mRNA. We further investigated the relationship between ligand concentration and growth by adding 5E1 antibody to cells from a single-passage pancreatic tumour xenograft at a level that was just sufficient to block growth. We then added Shh protein and found that growth correlated positively with increasing concentrations (Fig. 4d). Rates of growth from this experiment plotted as a function of Shh concentration (Fig. 4e) indicate that ligand-induced pathway activation is rate limiting and that unperturbed growth of these cells is sub-maximal. Assays of tumour cell lines derived from the oesophagus, stomach, pancreas and biliary tract yielded similar results (Supplementary Fig. 1a–c), confirming a widespread requirement for Hh ligand in the growth of these tumours.

The Hh ligand and 5E1 antibody are mutually antagonistic in their effects on reporter activity and produce opposite effects on the growth of cells from these gut-derived tumours (Fig. 4a–e and Supplementary Fig. 1a–c). Thus, pathway activation and cell growth must be dependent on Hh ligand. By contrast, addition of neither Hh ligand nor 5E1 blocking antibody significantly affected the growth of cells from a single-passage pancreatic tumour xenograft that did not express PTCH mRNA (Fig. 4f), which demonstrates the specificity of antibody and ligand effects. We also observed no significant ligand- or antibody-induced change in growth of medulloblastoma cells derived from a mouse model of Gorlin's syndrome (Fig. 4f) in which the Hh pathway is activated through loss of Ptch function15, 19. In contrast to antibody-resistant xenograft cells, medulloblastoma-derived cells require pathway activity for growth and can be killed by cyclopamine treatment15 (Fig. 4f).

Ligand-independent mutational activation of the Hh pathway has been linked to the formation of tumours, such as medulloblastoma, associated with Gorlin's syndrome. Despite a widespread activation of and dependence on the Hh pathway for medulloblastoma growth15, only a fraction of sporadic tumours can be assigned to pathway-activating mutations, suggesting that other mechanisms of pathway activation may be at play. Here we establish such a mechanism by showing that pathway activation and growth of cells from a group of commonly lethal gut-derived malignancies is ligand dependent. Small-cell lung cancer, also arising from endodermally derived epithelium and associated with Hh ligand expression, has recently been linked to transient reactivation of the Hh pathway within the airway epithelium, where it regulates progenitor cell fates during injury repair7. A similar role for Hh signalling in renewal of mature digestive tract epithelium is suggested by expression of the Hh pathway targets Ptch and Gli9, 10 (Fig. 1a) and by the requirement for Hh signalling for proliferation of gut progenitor cells10. It is not known whether renewal of injured gut epithelium is associated with transient Hh pathway reactivation. However, increased rates of oesophageal, gastric and pancreatic carcinomas occur in association with acid injury in Barrett's oesophagus, in Helicobacter pylori infection, and with exposure to alcohol, cigarette smoke and certain dietary components20-22. Exposure to such factors probably causes injury to the gut epithelium, eliciting a chronic state of injury repair and a consequent increase in proliferative stem or progenitor cells that may arise through ligand-dependent reactivation of the Hh pathway. Many of these agents are also mutagenic, thus potentially enhancing tumour formation by subjecting an enlarged pool of stem or stem-like target cells to potentially oncogenic mutations. Our results identify a group of common and frequently lethal gut-derived tumours, readily diagnosed by their expression of endogenous pathway targets such as PTCH, which may respond to antagonist- or antibody-mediated pathway blockade, even at advanced stages of metastatic disease.

Methods
Detection of beta-galactosidase expression Ptch–lacZ mice were killed at 4–6 weeks of age. Staining was performed as described previously7.

Tumour cells and tissues Origins and sources of our cells and tissues are described in the Supplementary Table. First-passage pancreatic cancer xenografts were derived from freshly harvested pancreaticoduodenectomy specimens as described previously15. In previous experiments, approximately 65% of specimens yielded xenografts (data not shown), so we inferred that these xenografts represent pancreatic tumours in the general population. The diagnosis of frozen samples from gastric and pancreatic adenocarcinoma resections and adjacent normal stomach and pancreas was microscopically confirmed by two pathologists (D.M.B. and A.M.), and RNA was prepared as described previously15.

RT–PCR Templates were prepared and amplified as described elsewhere15. For all primer pairs, specificity was confirmed by sequencing of PCR products. For quantitative RT–PCR, 10% of the first-strand reaction was amplified using IQ-SYBR Green Supermix, an i-cyclerIQ real-time detection system (Bio-Rad) and specific oligonucleotide primers for PTCH or PGK (phosphoglycerate kinase). Amplification was performed at 95 °C for 5 min followed by 40 cycles of 10, 15 and 30 s at 95 °C, 55 °C and 75 °C, respectively. Bio-Rad software was used to calculate threshold cycle (CT) values for PTCH and for the housekeeping gene PGK. For each sample, PTCH expression was derived from the ratio of PTCH to PGK levels using the formula 2-DeltaCT, where DeltaCT = CT–PTCH - CT–PGK. PTCH levels in tumours were presented as a ratio to levels detected in adjacent normal tissue (Fig. 3a).

Hh-responsive reporter assays Hh-responsive firefly luciferase and control SV40 Renilla luciferase reporter assays were performed on subconfluent triplicate cultures as described previously23. Two days after transfection, culture medium was replaced for a 2-day culture period with assay medium: RPMI-1640 (Bio-Whittaker) supplemented with 0.5% (established cell lines) or 20% (first-passage xenografts) FBS and containing combinations of 5E1 anti-Hh monoclonal antibody, recombinant doubly lipid-modified Sonic hedgehog (ShhNp) peptide13, cyclopamine purified from Veratrum extract or tomatidine (ICN Pharmaceuticals) at the concentrations indicated in the main text. Lysates were prepared and analysed as described elsewhere13.

Proliferation assays Cells were cultured in triplicate in 96-well plates in assay media to which 5E1 monoclonal antibody, ShhNp and/or cyclopamine were added at 0 h at concentrations indicated in the main text. Viable cell mass was determined by optical density measurements at 490 nm (OD490) at 2 and 4 days using the CellTiter96 (Promega) colorimetric assay. Relative growth was calculated as OD (day 4) - OD (day 2)/OD (day 2).

Xenograft treatment In accordance with approved Johns Hopkins University Animal Care and Use Committee protocols, HUCCT1 tumours (n = 18) were grown in athymic (nude) mice and treated with cyclopamine (50 mg kg-1 d-1, subcutaneous injection) or control vehicle as described previously15.

Supplementary information accompanies this paper.

Received 28 May 2003;accepted 25 July 2003

1. Bale, A. E. & Yu, K. P. The hedgehog pathway and basal cell carcinomas. Hum. Mol. Genet. 10, 757−762 (2001) | Article | PubMed | ISI | ChemPort |
2. Wechsler-Reya, R. & Scott, M. P. The developmental biology of brain tumors. Annu. Rev. Neurosci. 24, 385−428 (2001) | Article | PubMed | ISI | ChemPort |
3. Taipale, J., Cooper, M. K., Maiti, T. & Beachy, P. A. Patched acts catalytically to suppress the activity of Smoothened. Nature 418, 892−897 (2002) | Article | PubMed | ISI | ChemPort |
4. Taipale, J. & Beachy, P. A. The Hedgehog and Wnt signalling pathways in cancer. Nature 411, 349−354 (2001) | Article | PubMed | ISI | ChemPort |
5. Ingham, P. W. & McMahon, A. P. Hedgehog signaling in animal development: paradigms and principles. Genes Dev. 15, 3059−3087 (2001) | Article | PubMed | ISI | ChemPort |
6. Chen, J. K., Taipale, J., Cooper, M. K. & Beachy, P. A. Inhibition of Hedgehog signaling by direct binding of cyclopamine to Smoothened. Genes Dev. 16, 2743−2748 (2002) | Article | PubMed | ISI | ChemPort |
7. Watkins, D. N. et al. Hedgehog signalling within airway epithelial progenitors and in small-cell lung cancer. Nature 422, 313−317 (2003) | Article | PubMed | ISI | ChemPort |
8. Roberts, D. J., Smith, D. M., Goff, D. J. & Tabin, C. J. Epithelial−mesenchymal signaling during the regionalization of the chick gut. Development 125, 2791−2801 (1998) | PubMed | ISI | ChemPort |
9. van den Brink, G. R. et al. Sonic hedgehog regulates gastric gland morphogenesis in man and mouse. Gastroenterology 121, 317−328 (2001) | PubMed | ISI | ChemPort |
10. Ramalho-Santos, M., Melton, D. A. & McMahon, A. P. Hedgehog signals regulate multiple aspects of gastrointestinal development. Development 127, 2763−2772 (2000) | PubMed | ISI | ChemPort |
11. Hebrok, M. Hedgehog signaling in pancreas development. Mech. Dev. 120, 45−57 (2003) | Article | PubMed | ISI | ChemPort |
12. Bitgood, M. J. & McMahon, A. P. Hedgehog and Bmp genes are coexpressed at many diverse sites of cell−cell interaction in the mouse embryo. Dev. Biol. 172, 126−138 (1995) | Article | PubMed | ISI | ChemPort |
13. Taipale, J. et al. Effects of oncogenic mutations in Smoothened and Patched can be reversed by cyclopamine. Nature 406, 1005−1009 (2000) | Article | PubMed | ISI | ChemPort |
14. Cooper, M. K., Porter, J. A., Young, K. E. & Beachy, P. A. Plant-derived and synthetic teratogens inhibit the ability of target tissues to respond to Sonic hedgehog signaling. Science 280, 1603−1607 (1998) | Article | PubMed | ISI | ChemPort |
15. Berman, D. M. et al. Medulloblastoma growth inhibition by hedgehog pathway blockade. Science 297, 1559−1561 (2002) | Article | PubMed | ISI | ChemPort |
16. Ericson, J., Morton, S., Kawakami, A., Roelink, H. & Jessell, T. M. Two critical periods of Sonic Hedgehog signaling required for the specification of motor neuron identity. Cell 87, 661−673 (1996) | Article | PubMed | ISI | ChemPort |
17. Wang, L. C. et al. Conditional disruption of hedgehog signaling pathway defines its critical role in hair development and regeneration. J. Invest. Dermatol. 114, 901−908 (2000) | Article | PubMed | ISI | ChemPort |
18. Fuse, N. et al. Sonic hedgehog protein signals not as a hydrolytic enzyme but as an apparent ligand for patched. Proc. Natl Acad. Sci. USA 96, 10992−10999 (1999) | Article | PubMed | ChemPort |
19. Goodrich, L. V., Milenkovic, L., Higgins, K. M. & Scott, M. P. Altered neural cell fates and medulloblastoma in mouse patched mutants. Science 277, 1109−1113 (1997) | Article | PubMed | ISI | ChemPort |
20. Chen, X. & Yang, C. S. Esophageal adenocarcinoma: a review and perspectives on the mechanism of carcinogenesis and chemoprevention. Carcinogenesis 22, 1119−1129 (2001) | Article | PubMed | ISI | ChemPort |
21. Peek, R. M. Jr Helicobacter pylori strain-specific modulation of gastric mucosal cellular turnover: implications for carcinogenesis. J. Gastroenterol. 37 (Suppl. 13), 10−16 (2002) | Article | PubMed | ISI | ChemPort |
22. Lowenfels, A. B. & Maisonneuve, P. Epidemiologic and etiologic factors of pancreatic cancer. Hematol. Oncol. Clin. North Am. 16, 1−16 (2002) | Article | PubMed | ISI |
23. Chen, J. K., Taipale, J., Young, K. E., Maiti, T. & Beachy, P. A. Small molecule modulation of Smoothened activity. Proc. Natl Acad. Sci. USA 99, 14071−14076 (2002) | Article | PubMed | ChemPort |

Acknowledgements. We thank E. Traband and K. Young for technical assistance, E. Montgomery, K. Miyazaki and J. Harmon for cell lines, J. Chen for help with cyclopamine purification and P. Fussell for help with figures. This work was supported by the family of Margaret Lee and grants from the National Institutes of Health (D.M.B., A.M., J.R.E. and P.A.B.). P.A.B. is an investigator and M.R.G. a medical fellow of the Howard Hughes Medical Institute.

Competing interests statement. The authors declare competing financial interests.
full article
PDF version of article