¹Department of Radiation Oncology, Duke University Medical Center, Durham, North Carolina, NC 27710, USA

²Department of Pathology, Vanderbilt University Medical Center, Nashville, Tennessee, TN 37232, USA

Correspondence to: R L Jirtle, Department of Radiation Oncology, Duke University Medical Center, Durham, North Carolina, NC 27710, USA

In addition to the intracellular sorting of lysosomal enzymes, the mannose 6-phosphate/insulin-like growth factor II receptor (M6P/IGF2R) plays a critical role in regulating the bioavailability of extracellular proteolytic enzymes and growth factors. It has also been shown to be mutated in a number of human cancers, and to suppress cancer cell growth. The purpose of this study was to determine if the M6P/IGF2R is mutated in lung cancer, a leading cause of cancer death worldwide. Archival pathology specimens were obtained on 22 patients with newly diagnosed, untreated squamous cell carcinoma of the lung. Two polymorphisms in the 3'-untranslated region of the M6P/IGF2R were used to screen lung tumors for loss of heterozygosity (LOH) by PCR amplification of DNA. Nineteen of 22 (86%) patients were informative (heterozygous), and 11/19 (58%) squamous cell carcinomas of the lung had LOH at the M6P/IGF2R locus. The remaining allele in 6/11 (55%) LOH patients contained mutations in either the mannose 6-phosphate or the IGF2 binding domain of the M6P/IGF2R. Thus, the M6P/IGF2R is mutated frequently in squamous cell carcinoma of the lung, providing further support for its function as a tumor suppressor. Oncogene (2000) 19, 1572-1578.

M6P/IGF2R; loss of heterozygosity; lung cancer; mutation; tumor suppressor

Introduction

Lung cancer is a major cause of mortality worldwide, and it is the leading cause of cancer death in the United States (Landis et al., 1998). There were 171 500 new cases of lung cancer diagnosed and 160 100 deaths in 1998, making lung cancer a major health problem in this country. The constitutive activation or overexpression of proto-oncogenes and the inactivation of tumor suppressor genes are critical in the multistep process of lung carcinogenesis. These oncogenic changes are not only present in lung tumors, but some can already be detected in the nonmalignant bronchial epithelium of current and former smokers (Wistuba et al., 1997). A number of proto-oncogenes have been shown to be altered in non-small cell lung cancer (NSCLC) including RAS, MYC, HER2/NEU, MYB, RAF and JUN (Fong et al., 1995). Lung cancer also demonstrates loss of imprinting (LOI) at the insulin-like growth factor II (IGF2) locus (Suzuki et al., 1994) resulting in the biallelic expression of a growth factor known to be oncogenic when overexpressed (Bates et al., 1995; Rogler et al., 1994). Tumor suppressor genes mutated in NSCLC include p53, Rb, p16INK4A and FHIT (Sekido et al., 1998). The presence of deletions at numerous other chromosomal locations in NSCLCs, however, suggests the presence of additional lung tumor suppressor genes.

The mannose 6-phosphate/insulin-like growth factor II receptor (M6P/IGF2R) is located at 6q26, a chromosomal location commonly deleted in NSCLCs (Bepler and Koehler, 1995; Petersen et al., 1997). It encodes for a multifunctional receptor required for the activation of the growth inhibitor, transforming growth factor 1 (TGF1), the degradation of the mitogen, IGF2, and the intracellular sorting of lysosomal enzymes (Dennis and Rifkin, 1991; Kornfeld, 1992; Jirtle, 1999a). Evidence to date suggests, however, that the M6P/IGF2R is not involved in cell signaling (Korner et al., 1995); this function is mediated primarily by the insulin-like growth factor I receptor (IGFIR) and the insulin receptor isoform A (Czech et al., 1989; Frasca et al., 1999). The M6P/IGF2R has also been shown to be mutated in a number of human cancers, including those that develop in the liver, breast and colon (Jirtle et al., 1999b), and to suppress cancer cell growth (Kang et al., 1999; O'Gorman et al., 1999; Souza et al., 1999). These findings are consistent with the M6P/IGF2R functioning normally as a tumor suppressor.

The purpose of this investigation was to determine if the M6P/IGF2R is a gene target for mutation in human lung cancer, a disease strongly associated with tobacco smoking. Herein, we report that the M6P/IGF2R is mutated in approximately 60% of squamous cell carcinomas of the lung, a mutation frequency comparable to that of p53 in lung cancer (Sekido et al., 1998).

Results and discussion

There are two mannose 6-phosphate receptors involved in the intracellular trafficking of lysosomal enzymes, the 46 kDa cation-dependent mannose 6-phosphate receptor (CDM6PR) and the 275 kDa cation-independent M6P/IGF2R (Kornfeld, 1992; Jirtle, 1999a). In contrast to the CDM6PR, the M6P/IGF2R also binds extracellular phosphomannosyl glycoproteins and IGF2 when located at the cell surface. Thus, in addition to the intracellular sorting of lysosomal enzymes, the M6P/IGF2R controls the extracellular levels of potent growth factors and proteolytic enzymes that govern cell death, proliferation and invasion (Kornfeld, 1992; Jirtle, 1999a).

The earliest experimental evidence for the involvement of the M6p/Igf2r in tumor formation came from rat liver carcinogenesis studies (Jirtle et al., 1994). A small subset of diethylnitrosamine initiated, phenobarbital promoted preneoplastic liver lesions failed to stain immunohistochemically for the M6p/Igf2r while the majority of hepatocellular carcinomas were either partially or uniformly negative for this receptor. This demonstrated that loss of M6p/Igf2r function occurs early in rat liver carcinogenesis, and suggested that preneoplastic lesions lacking this receptor preferentially progress to carcinomas.

M6P/IGF2R inactivation has also now been demonstrated in human tumors. LOH at the M6P/IGF2R locus occurs in 60% of human HCCs (De Souza et al., 1995a; De Souza et al., 1995b; Yamada et al., 1997) and 30% of breast cancers (Hankins et al., 1996), and missense mutations that reduce receptor function have been identified in the remaining allele (De Souza et al., 1997; Yamada et al., 1997; Byrd et al., 1999; Devi et al., 1999). The M6P/IGF2R also contains a poly-G region that is a common target for frame-shift mutations in colon, gastric, endometrial and liver tumors with mismatch repair deficiencies and microsatellite instability (MI) (Ouyang et al., 1997; Souza et al., 1996; Yamada et al., 1997). Moreover, the M6P/IGF2R is mutated in human gliomas that do not contain mutations in the TGF type II receptor or Bax genes (Leung et al., 1998).

Induction of exogenous wild-type M6P/IGF2R in human colorectal cancer cells with an inactivated allele significantly decreases growth rate and enhances apoptosis (Souza et al., 1999). In contrast, receptor down-regulation in choriocarcinoma cells by antisense M6P/IGF2R RNA increases both tumor cell growth rate and tumorigenicity (O'Gorman et al., 1999). The M6P/IGF2R also contains a retinoic acid (RA) bind site (Kang et al., 1997), and receptor overexpression in RA-resistant cancer cells lacking functional RA nuclear receptors confers susceptibility to RA-induced apoptosis (Kang et al., 1999). The results of these functional studies, coupled with frequent M6P/IGF2R mutation in a number of human cancers, demonstrate the importance of this genetic target in human carcinogenesis.

To determine if the M6P/IGF2R is mutated in squamous cell carcinoma of the lung, we screened 22 patients for LOH at this locus. Of these 22 patients, 19 (86%) were informative (heterozygous) for a polymorphic dinucleotide repeat sequence and a tetranucleotide insertion/deletion polymorphism located in the 3'-untranslated region of the receptor (Hol et al., 1992) (Figure 1). There were six female and 13 male patients in the informative group. The average age of these patients was 65»plus; 3 years, and they all had a history of smoking (Table 1). LOH at the M6P/IGF2R locus was observed in 11/19 (58%) informative patients. The average age of the patients in the tumor LOH group (66»plus;3 years) was not significantly different (P>0.1) from that in the tumor non-LOH group (63»plus;5 years). The tumor stage also did not vary significantly between these two tumor groups with most patients being either stages 3a or 3b (Table 1).

Although there is significant sequence homology between the extracellular region of CDM6PR and all 15 repeat domains in the M6P/IGF2R, only repeat domains 3 and 9 contain an M6P binding site (Dahms et al., 1987, 1993; Lobel et al., 1988). The minimal binding region for IGF2 in the human receptor is located in the N-terminal portion of repeat 11 (Dahms et al., 1994; Garmroudi et al., 1996; Schmidt et al., 1995), and repeat 13 further enhances the IGF2 binding affinity of the M6P/IGF2R (Devi et al., 1998). The mannose 6-phosphate binding sites and the IGF2 binding and enhancing regions of the M6P/IGF2R gene were screened for mutations in the tumors with LOH. Mutations were detected in these receptor binding regions in 6/11 (55%) tumors (Figure 2). The codon changes, amino acid changes and the predicted changes in secondary protein structure resulting from these mutations are presented in Table 2.

No mutations were identified in exons 8 to 11 indicating that the M6P binding site present in repeat 3 is not commonly mutated in squamous cell carcinomas of the lung. In contrast, 2/11 lung tumors were mutated in the M6P binding domain contained in repeat 9. A G:CC:G transversion was identified at a non CpG site in exon 27 that results in the substitution of Arg for Gly1296 (Figure 2a). This amino acid is conserved among human, bovine, rat, mouse and chicken indicating that Gly1296 is important for receptor function. The secondary structure of this repeat region has been modeled on the 3-D structure of the CDM6PR which forms a nine stranded flattened barrel (Roberts et al., 1998). Based upon this predicted secondary structure, Gly1296 of repeat 9 resides in the loop region between beta strands 4 and 5, the transition region between the two orthogonal sheets. Therefore, the substitution of Arg, a large charged amino acid, for Gly1296, a small neutral amino acid, would be expected to significantly alter the interaction between the two beta sheets thereby reducing receptor function.

The second mutation identified in repeat 9 is a G insertion into the M6P/IGF2R poly-G region of exon 28 (Figure 2b). This single base frameshift mutation results in the formation of the stop codon TAA for Lys1321, thereby inhibiting the production of a functional receptor as previously shown by immunohistochemistry (Yamada et al., 1997). The poly-G region in the M6P/IGF2R gene is a target of MI in replication/repair error-positive (RER+) endometrial, stomach and colorectal tumors (Ouyang et al., 1997; Souza et al., 1996), and it is also mutated in approximately 25% of hepatocellular carcinomas (Yamada et al., 1997). A deoxyguanine insertion into the poly-G region of the M6P/IGF2R indicates this region may also be a target of MI in squamous cell carcinomas of the lung. Our finding contradicts the recent reported absence of mutations in the poly-G region of the M6P/IGF2R in lung tumors (Gotoh et al., 1999). This most likely results from the inclusion of RER+ tumors with MI in our study population, even though its frequency appears to be low in lung cancer (Sekido et al., 1998).

Three different missense mutations were also detected in repeat domain 11 that contains the IGF2 binding site, but no mutations were found in the IGF2 binding enhancer site present in repeat domain 13 (exons 37-39) (Devi et al., 1998). A G:CC:G transversion was identified at a non CpG site in exon 33 that results in the substitution of Arg for Gly1564 (Figure 2c). This amino acid alteration occurs within the identified minimal IGF2 binding site (Dahms et al., 1994; Garmroudi et al., 1996; Schmidt et al., 1995), and is predicted from the 3-D structure of the CDM6PR to be in a putative loop region between beta strands 2 and 3 (Roberts et al., 1998). The substitution of Arg, a large charged amino acid, for Gly1564, a small neutral amino acid, would be expected to significantly alter receptor tertiary structure and function. Additionally, Gly1564 is conserved among mammalian species, but it is mutated in the chicken. This provides further evidence that Gly1564 plays a significant role in binding IGF2 since unlike the mammalian M6P/IGF2R homologues, the chicken receptor does not bind IGF2 (Zhou et al., 1995).

The second mutation detected in repeat domain 11 of the M6P/IGF2R is a G:CA:T transition at a non CpG site in exon 34 that results in the substitution of Thr for Ala1618 (Figure 2d). This amino acid change is predicted from the 3-D structure of the CDM6PR to reside in a putative loop region between beta strands 7 and 8 (Roberts et al., 1998). The substitution of Thr, an uncharged polar amino acid, for Ala1618, a nonpolar hydrophobic amino acid, would be expected to affect the receptor tertiary structure and function since the interactions between the two orthogonal beta sheets is principally hydrophobic in nature (Roberts et al., 1998). The hydrophobic interactions between the beta sheets makes the overall structure very tight and compact. An alteration in the receptor compactness may not only reduce IGF2 binding but also receptor half-life since it resides part of the time in the proteolytic environment of the lysosomes (Kornfeld, 1992). Ala1618 is also conserved among human, rat, mouse and chicken suggesting that this amino acid plays an important role in the function and/or stability of the receptor.

The third mutation detected in repeat domain 11 of the receptor is a G:CA:T transition at a CpG site in exon 34 that results in the substitution of Arg for Gly1619 (Figure 2e,f). This amino acid alteration is also predicted from the 3-D structure of the CDM6PR to reside in a putative loop region between beta strands 7 and 8 (Roberts et al., 1998). The substitution of Arg, a large charged amino acid, for Gly1619, a small neutral amino acid, would be expected to significantly alter receptor tertiary structure and function. Gly1619 is conserved among bovine, human, mouse, rat and chicken. Furthermore, this mutation was detected in two lung tumors highly suggesting that this amino acid is important in maintaining receptor function and/or stability.

The presence of mutations in both the M6P binding site in repeat 9 and the IGF2 binding site in repeat 11 suggests that both phosphomannosyl containing glycoproteins and IGF2 are involved in the etiology of lung cancer. Additionally, every M6P/IGF2R missense mutation detected in squamous cell carcinoma of the lung occurred within the predicted transition region between beta strands, and 80% of them involved Gly (Table 1). Thus, these loop regions appear to be critical to receptor ligand binding function and/or its stability. This postulate is further supported by our recent finding that substitution of a Val for Gly1449 in a putative loop region of repeat 10 markedly reduces receptor binding of both IGF2 and M6P containing ligands (Byrd et al., 1999; Devi et al., 1999).

We also screened M6P/IGF2R for the G:CT:A transversion in exon 31 that results in the substitution of a Val for Gly1449. Although this mutation is present in 13% of HCCs (De Souza et al., 1995b; Yamada et al., 1997), it was not detected in squamous cell carcinoma of the lung. This is particularly interesting since carcinogens found in smoke such as benzo[a]pyrene diol epoxide preferentially attack deoxyguanine, frequently resulting in a G:CT:A transversion (Greenblatt et al., 1994). Furthermore, none of the M6P/IGF2R missense mutations found in squamous cell carcinomas of the lung have been detected in either HCCs or breast cancer (De Souza et al., 1995b; Hankins et al., 1996; Yamada et al., 1997). Thus, the M6P/IGF2R mutational spectrum appears to be tumor specific and/or dependent upon the carcinogenic agent involved in the etiology of tumor formation.

M6P/IGF2R immunohistochemical staining was absent in the six squamous cell carcinomas of the lung in which both alleles of the gene were found to be mutated, but was present in tumors without LOH (Figure 3). The lack of immunohistochemical staining in tumors harboring not only frameshift but also missense mutations could result from structural changes in the receptor altering the primary epitopes recognized by the M6P/IGF2R antibody. Alternatively, receptors with reduced compactness because of missense mutations could have an enhanced susceptibility to proteolytic degradation when they reside in the lysosomes, thereby reducing the steady-state receptor level below that which is detectable. This latter postulate is supported by our recent finding that receptor stability in vitro is decreased by a missense mutation in repeat 10 of the M6P/IGF2R (Byrd et al., 1999; Devi et al., 1999). We further observed that the majority of tumors with LOH at the M6P/IGF2R locus (10/11) lack detectable receptor protein even though the remaining allele is not mutated in the receptor binding regions. This could result from the allele being mutated in gene regions not screened in this investigation. Gene expression could also be inhibited by epigenetic alterations, as occurs frequently with p16INK4A in lung cancer (Gazzeri et al., 1998). Additionally, one allele of the M6P/IGF2R could be transcriptionally silenced in lung tumors because of imprinting.

Genomic imprinting is an epigenetic form of gene regulation that results in the expression of only one parental allele, and a number of imprinted genes have been determined to be oncogenic targets (Falls et al., 1999; Jirtle et al., 1999). The M6p/Igf2r gene is imprinted in both mice (Barlow et al., 1991; Wang et al., 1994) and rats (Mills et al., 1998), whereas imprinting of the M6P/IGF2R is a polymorphic trait in humans with most adults expressing both alleles (Falls et al., 1999; Jirtle et al., 1999). Thus, imprinting of the M6P/IGF2R could result in allelic inactivation, a postulate supported by the discovery that 50% of patients with Wilms' tumor are imprinted at this locus (Xu et al., 1997). Unfortunately, it was not possible to determine if M6P/IGF2R was imprinted in these lung tumors because RNA could not be isolated from the formalin fixed, paraffin embedded tissues.

Although radiographic techniques have improved the detection of small lung tumors, 90% of lung cancer patients still die of their disease because of a propensity for lung cancer to metastasize early. For those patients in whom NSCLC is detected at an early stage, the 5 year survival following surgery exceeds 60% (Landis et al., 1998). The cure rates exceed 90% with surgical resection for patients with truly occult NSCLC (i.e., asymptomatic, fortuitously discovered lesions not visible on X-ray) (Ginsberg et al., 1997). Thus, while effective treatment exists for early stage NSCLC, to date there remains no widely accepted screening intervention for this disease (Rimer and Schildkraut, 1997). Since M6P/IGF2R mutation is an early event in liver (Yamada et al., 1997) and breast carcinogenesis (Chappell et al., 1997; Hankins et al., 1996), it may provide an important molecular marker for early neoplastic changes in the lung. The vast majority of lung cancers are smoking related, and the ability to detect early lesions in this high risk population could save over 100 000 lives per year in the United States alone (Landis et al., 1998).

In conclusion, M6P/IGF2R mutation is a frequent occurrence in squamous cell carcinoma of the lung, indicating that it normally functions as a lung tumor suppressor gene. Additional studies are required to determine if M6P/IGF2R mutation is an early event in lung carcinogenesis, and if gene loss alters patient prognosis.

Materials and methods

Patients

Archival paraffin embedded tissue sections from 22 randomly selected patients with histopathologically confirmed squamous cell carcinoma of the lung were obtained from the Department of Pathology, Duke University Medical Center. All patients were treated between 1992 and 1995, and had a history of smoking.

Tissue microdissection from paraffin embedded sections

Microdissection of tumor and normal tissue from 6 m histology sections was performed as previously described (De Souza et al., 1995a; Hankins et al., 1996). Normal stroma not immediately adjacent to cancer cells was used as the control tissue. The areas chosen for microdissection were identified by a board-certified pathologist (MK Washington) who also classified the lesions. Paraffin embedded sections were deparaffinized in xylene (2´5 min), placed in 100% ethanol (2´5 min) followed by 50% ethanol (2´5 min) and then rehydrated in H₂O prior to staining. The tissue sections were stained for 30 s with 2% (w/v) methylene blue, rinsed in H₂O and allowed to air dry. Tumor and normal tissue (>50 cells) were then dissected and put in 50 l of 1´ PCR buffer (10 mM Tris.HCl, pH 8.3 at 25°C, 50 mM KCl) containing 2 l of 20 mg/ml proteinase K (Roche Boehringer Mannheim, Indianapolis, IN, USA). The mixture was first incubated at 56°C for 60 min and then at 100°C for 10 min. 5-10 l of this mixture were used in the PCR analysis described below.

PCR analysis for loss of heterozygosity

Two polymorphisms are present in the 3'-untranslated region of the human M6P/IGF2R (Hol et al., 1992). There is a tetranucleotide insertion/deletion close to a dinucleotide repeat which together give observed heterozygosity of approximately 60%. We utilized these polymorphisms to determine, by PCR amplification of DNA, the frequency of LOH at the M6P/IGF2R locus in patients with squamous cell carcinomas of the lung. The forward and reverse PCR primers used were JJMIF (5'-TTGCCGGCTGGTGAATTCAA-3') and JJMIR (5'-CTCTTCAGGTTCTCATGATA-3'), respectively (De Souza et al., 1995a). The reaction conditions for PCR were as follows: 10 mM Tris.HCl, pH 8.3 at 25°C, 50 mM KCl, 1 mM MgCl₂, 200 M 4dNTP mix, 200 nM forward and reverse primers, 5 l microdissected template and 2.5 units Taq DNA polymerase (PE Biosystems, Foster City, CA, USA) in a total volume of 100 l. Hot-start PCR was carried out under the following conditions: 1 min denaturing at 94°C, 1 min annealing at 55°C and 1 min extension at 72°C for 30 cycles with an additional 10 min extension for cycle 30 on a DNA thermal cycler. The DNA generated by PCR was characterized by agarose gel electrophoresis and dideoxy sequencing.

Excess primers were removed using a Qiagen PCR purification kit (Qiagen, Valencia, CA, USA). 10 l of purified PCR products were digested with 10 units EcoRI, and then labeled with -³³P ATP (2000 Ci/mmol, 10 mCi/ml) using 2.5 units of Large Fragment DNA Polymerase I. Sequencing gel-loading buffer [95% (v/v) deionized formamide, 20 mM EDTA (pH 8.0), 0.05% (w/v) xylene cyanol FF, and 0.05% (w/v) bromophenol blue] was added to the labeled PCR products in a ratio of 1 : 1 prior to heat denaturation (80°C for 3 min) and electrophoresis was performed with a 6% denaturing polyacrylamide sequencing gel. The electrophoretic products were visualized by autoradiography. Due to the potential of contamination from normal stromal tissue, allele loss in informative patients was defined as a >50% decrease in the ratio of the two alleles in the tumor tissue versus that in the surrounding normal stromal tissue; this was quantified using a densitometer.

M6P/IGF2R mutation detection

In lung tumors with LOH at the M6P/IGF2R locus, the remaining allele was screened for mutations in the ligand binding regions by direct sequencing of PCR products. The regions screened for mutations were exons 8-11 (repeat 3), exons 27-29 (repeat 9), exon 31 (repeat 10), exons 33 and 34 (repeat 11) and exons 37-39 (repeat 13) (Killian and Jirtle, 1999). Because Taq DNA polymerase can introduce base errors during the PCR process, a number of precautions were taken to address this potential problem. Mutant DNA templates were amplified in two or more independent PCR reactions. The corresponding normal template, comprized of normal stromal tissue distant from the cancer cells, was always amplified in parallel with the tumor template, and the tumor mutations were confirmed by direct sequencing in both directions. Additionally, the genomic DNA was amplified using nested primers. The first round primers were removed using a Qiagen gel filtration kit (Clontech, Palo Alto, CA, USA), and 5 l of the elute were used in a second round amplification. Gel purified second round PCR products were directly sequenced using the AmpliCycle^TM kit (PE Biosystems, Foster City, CA, USA) and the second round primers. The exon specific forward and reverse PCR primers used have been previously published (http://www.geneimprint .com) (Killian and Jirtle, 1999).

M6P/IGF2R immunohistochemical staining

The technique used to immunohistochemically detect M6P/IGF2R protein in formalin fixed, paraffin embedded tissues sections was modified from Jirtle et al. (1994). Following tissue section deparaffinization and hyaluronidase digestion, the 6 m sections were blocked with protein solution (BioGenex, San Ramon, CA, USA) and exposed overnight at 4°C to the C-1 rabbit polyclonal antibody to M6P/IGF2R (Scott and Baxter, 1987); non-immune rabbit IgG was used as a control. The tissue sections were then processed according to the recommended procedures provided with the Multilink-Alkaline Phosphatase Detection Kit (BioGenex, San Ramon, CA, USA). The alkaline phosphatase-containing Fast Red Tablet Kit (BioGenex, San Ramon, CA, USA) was used to immunohistochemically stain the tissues, and levamisole was applied to block endogenous alkaline phosphatase activity. All the tissue sections were counterstained with hematoxylin. Aqueous mounting media was used to cover the immunohistochemically stained slides.

Acknowledgements

The authors would like to thank Hongtao Xing and Hong Huang for their assistance with data analysis, and David Pulford and Monica Bandera for M6P/IGF2R mutation confirmation. This research is supported by NIH grants CA25951 and ES08823, DOD Grant DAMD17-98-1-8305. Rohm & Haas Chemical Company, Inc., Sumitomo Chemical Company, Ltd. and Zeneca Pharmaceuticals, Ltd. For additional information on the M6P/IGF2R, visit the website: (http://www.geneimprint.com).

Barlow DP, Stöger R, Herrmann BG, Saito K and Schweifer N. (1991). Nature 349, 84-87. MEDLINE

Bates P, Fisher R, Ward A, Richardson L, Hill DJ and Graham CF. (1995). Br J Cancer 72, 1189-1193. MEDLINE

Bepler G and Koehler A. (1995). Cancer Genet Cytogenet 84, 39-45. MEDLINE

Byrd JC, Devi GR, De Souza AT, Jirtle RL and MacDonald RG. (1999). J Biol Chem 274, 24408-24416. MEDLINE

Chappell SA, Walsh T, Walker RA and Shaw JA. (1997). Br J Cancer 76, 1558-1561. MEDLINE

Czech MP, Lewis RE and Corvera S. (1989). Ciba Found Symp 145, 27-41. MEDLINE

Dahms NM, Lobel P, Breitmeyer J, Chirgwin JM and Kornfeld S. (1987). Cell 50, 181-192. MEDLINE

Dahms NM, Rose PA, Molkentin JD, Zhang Y and Brzycki MA. (1993). J Biol Chem 268, 5457-5463. MEDLINE

Dahms NM, Wick DA and Brzycki-Wessell MA. (1994). J Biol Chem 269, 3802-3809. MEDLINE

De Souza AT, Hankins GR, Washington MK, Fine RL, Orton TC and Jirtle RL. (1995a). Oncogene 10, 1725-1729. MEDLINE

De Souza AT, Hankins GR, Washington MK, Orton TC and Jirtle RL. (1995b). Nat Genet 11, 447-449. MEDLINE

De Souza AT, Yamada T, Mills JJ and Jirtle RL. (1997). FASEB J 11, 60-67. MEDLINE

Dennis PA and Rifkin DB. (1991). Proc Natl Acad Sci USA 88, 580-584. MEDLINE

Devi GR, Byrd JC, Slentz DH and MacDonald RG. (1998). Mol Endocrinol 12, 1661-1672. MEDLINE

Devi GR, De Souza AT, Byrd JC, Jirtle RL and MacDonald RG. (1999). Cancer Res 59, 4314-4319. MEDLINE

Falls JG, Pulford DJ, Wylie AA and Jirtle RL. (1999). Am J Pathol 154, 635-647. MEDLINE

Fong KM, Zimmerman PV and Smith PJ. (1995). Pathology 27, 295-301. MEDLINE

Frasca F, Pandini G, Scalia P, Sciacca L, Mineo R, Costantino A, Goldfine ID, Belfiore A and Vigneri R. (1999). Mol Cell Biol 19, 3278-3288. MEDLINE

Garmroudi F, Devi G, Slentz DH, Schaffer BS and MacDonald RG. (1996). Mol Endocrinol 10, 642-651. MEDLINE

Gazzeri S, Gouyer V, Vour'ch C, Brambilla C and Brambilla E. (1998). Oncogene 16, 497-504. MEDLINE

Ginsberg RJ, Vokes EE and Raben A. (1997). Principles and Practice of Oncology DeVita VT Jr, Hellman S and Rosenberg SA eds, . Lippincott-Raven Publishers New York, pp858-911.

Gotoh K, Yatabe Y, Sugiura T, Takagi K, Ogawa M, Takahashi T and Mitsudomi T. (1999). Carcinogenesis 20, 499-502. MEDLINE

Greenblatt MS, Bennett WP, Hollstein M and Harris CC. (1994). Cancer Res 54, 4855-4878. MEDLINE

Hankins GR, De Souza AT, Bentley RC, Patel MR, Marks JR, Iglehart JD and Jirtle RL. (1996). Oncogene 12, 2003-2009. MEDLINE

Hol FA, Geurds MP, Hamel BC and Mariman EC. (1992). Hum Mol Genet 1, 347. MEDLINE

Jirtle RL. (1999a). Encyclopedia of Molecular Biology Creidton, TE ed, . Wiley-Liss, Inc New York, New York, pp1441-1447.

Jirtle RL. (1999b). Exp Cell Res 248, 18-24.

Jirtle RL, Hankins GR, Reisenbichler H and Boyer IJ. (1994). Carcinogenesis 15, 1473-1478. MEDLINE

Killian JK and Jirtle RL. (1999). Mamm Genome 10, 74-77. MEDLINE

Kang JX, Li Y and Leaf A. (1997). Proc Natl Acad Sci USA 94, 13671-13676. MEDLINE

Kang JX, Bell J, Beard RL and Chandraratna RAS. (1999). Cell Growth Differ 10, 591-600. MEDLINE

Korner C, Nurnberg B, Uhde M and Braulke T. (1995). J Biol Chem 270, 287-295. MEDLINE

Kornfeld S. (1992). Annu Rev Biochem 61, 307-330. MEDLINE

Landis SH, Murray T, Bolden S and Wingo PA. (1998). CA Cancer J Clin 48, 6-29. MEDLINE

Leung SY, Chan TL, Chung LP, Chan AS, Fan YW, Hung KN, Kwong WK, Ho JW and Yuen ST. (1998). Am J Pathol 153, 1181-1188. MEDLINE

Lobel P, Dahms NM and Kornfeld S. (1988). J Biol Chem 263, 2563-2570. MEDLINE

Mills JJ, Falls JG, De Souza AT and Jirtle RL. (1998). Oncogene 16, 2797-2802. MEDLINE

Morgan DO, Edman JC, Standring DN, Fried VA, Smith MC, Roth RA and Rutter WJ. (1987). Nature 329, 301-307. MEDLINE

O'Gorman DB, Costello M, Weiss J, Firth SM and Scott CD. (1999). Cancer Res 59, 5692-5694. MEDLINE

Ouyang H, Shiwaku HO, Hagiwara H, Miura K, Abe T, Kato Y, Ohtani H, Shiiba K, Souza RF, Meltzer SJ and Horii A. (1997). Cancer Res 57, 1851-1854. MEDLINE

Petersen I, Bujard M, Petersen S, Wolf G, Goeze A, Schwendel A, Langreek H, Gellert K, Reichel M, Just K, du Manoir S, Cremer T, Dietel M and Ried T. (1997). Cancer Res 57, 2331-2335. MEDLINE

Rimer BK and Schildkraut J. (1997). Principles and Practice of Oncology DeVita Jr VT, Hellman S and Rosenberg SA, eds. Lippincott-Raven Publishers New York, pp619-631.

Roberts DL, Weix DJ, Dahms NM and Kim JJ. (1998). Cell 93, 639-648. MEDLINE

Rogler CE, Yang D, Rossetti L, Donohoe J, Alt E, Chang CJ, Rosenfeld R, Neely K and Hintz R. (1994). J Biol Chem 269, 13779-13784. MEDLINE

Schmidt B, Kiecke-Siemsen C, Waheed A, Braulke T and von Figura K. (1995). J Biol Chem 270, 14975-14982. MEDLINE

Scott CD and Baxter RC. (1987). Endocrinology 120, 1-9. MEDLINE

Sekido Y, Fong KM and Minna JD. (1998). Biochim Biophys Acta 1378, F21-F59. MEDLINE

Souza RF, Appel R, Yin J, Wang S, Smolinski KN, Abraham JM, Zou T-T, Shi Y-Q, Lei J, Cottrell J, Cymes K, Biden K, Simms L, Leggett B, Lynch PM, Frazier M, Powell SM, Harpaz N, Sugimura H, Young J and Meltzer SJ. (1996). Nat Genet 14, 255-257. MEDLINE

Souza RF, Wang S, Thakar M, Smolinski KN, Yin J, Zou T-T, Kong D, Abraham JM, Toretsky JA and Meltzer J. (1999). Oncogene 18, 4063-4068. MEDLINE

Suzuki H, Veda R and Takahashi T. (1994). Nat Genet 6, 332-333. MEDLINE

Wang ZQ, Fung MR, Barlow DP and Wagner EF. (1994). Nature 372, 464-467. MEDLINE

Wistuba II, Lam S, Behrens C, Virmani AK, Fong KM, LeRiche J, Samet JM, Srivastava S, Minna JD and Gazdar AF. (1997). J Natl Cancer Inst 89, 1366-1373. Article MEDLINE

Xu YQ, Grundy P and Polychronakos C. (1997). Oncogene 14, 1041-1046. MEDLINE

Yamada T, De Souza AT, Finkelstein S and Jirtle RL. (1997). Proc Natl Acad Sci USA 94, 10351-10355. Article MEDLINE

Zhou M, Ma Z and Sly WS. (1995). Proc Natl Acad Sci USA 92, 9762-9766. MEDLINE

Figure 1 Representative M6P/IGF2R informative patients (a,b) without LOH (a) and with LOH (b) in human squamous cell carcinomas of the lung. A non-informative lung cancer patient is also shown (c). Alleles are defined as A1 and A2 for the informative patients and A1/A2 for the non-informative patient. Arrowhead marks the lost allele; faint band due to contaminating normal stromal tissue and/or PCR artifact

Figure 2 M6P/IGF2R mutations detected in human squamous cell carcinomas of the lung with LOH. Five tumors (a, c-f) have missense point mutations (arrowheads) and 1 tumor (b) has a frameshift mutation caused by a single G insertion (arrowhead) (sequenced in reverse direction). Horizontal line in (c) identifies exon/intron boundary

Figure 3 Immunohistochemical staining for M6P/IGF2R in a human squamous cell carcinoma of the lung without LOH (a) and with LOH (b). (a) M6P/IGF2R was highly expressed (pink) in the squamous cell carcinoma of the lung (SC) without LOH (Table 1, Patient 12), and slightly expressed in the adjacent normal connective tissue (N). (b) M6P/IGF2R was undetectable in the squamous cell carcinoma of the lung (SC) with LOH and a Gly1296Arg mutation in the remaining allele (Tables 1 and 2, Patient 1); the adjacent normal connective tissue (N) was slightly stained. Control rabbit IgG staining was negative (data not shown). Tissues were counterstained with hematoxylin, X500

Table 1 Summary data for the M6P/IGF2R informative patients

Table 2 M6P/IGF2R mutations in lung cancer