Allelic loss of chromosome 8p21–22 occurs frequently in cancer, including lung and head and neck squamous cell cancer. The tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) receptors, including proapoptotic DR4 and KILLER/DR5, are located on 8p21–22. TRAIL receptors are candidate tumor suppressor genes, because their inactivation would be expected to result in deficient apoptotic signaling. To investigate the involvement of DR4 in human cancer, we have determined the genomic structure of DR4 and screened 31 lung cancer cell lines [14 small cell lung cancer and 17 non-small cell lung cancer (NSCLC)], many with deletions at 8p21–22, and 21 primary NSCLC samples for mutations in DR4. We found two missense alterations in the ectodomain of DR4. One, at nucleotide 626, changes a cytosine to a guanine (C626G) and results in a substitution of an arginine for threonine. The other, at nucleotide 422, changes a guanine to adenine (G422A) and results in a substitution of a histidine for arginine. Using genomic DNA sequencing and RFLP analysis, we show that these two alterations cosegregated in 96% of all of the samples (n = 243) evaluated (tumor and normal). The frequency of being homozygous for both altered alleles was 35% in the lung cancer cell lines but only 13% in age- and race-matched controls, which was a significant increase (χ2 = 5.2, P = 0.023). The frequency of homozygosity for both alleles was also significantly increased in the primary NSCLC samples (χ2 = 9.2, P = 0.002) as compared with the age- and race-matched controls. To determine whether the altered alleles are specific for lung cancer, we evaluated 19 head and neck squamous cell cancer and 25 gastric adenocarcinoma samples. Forty-seven % of the former and 44% of the latter were homozygous for both the C626G and G422A alterations, and this was significantly elevated relative to age- and race-matched controls (χ2 = 8.6, P = 0.003 and χ2 = 8.2, P = 0.004). These alterations result in amino acid changes in or near the ligand-binding domain of DR4 and, based on the crystal structure of DR5 and its homology with DR4, have the potential to affect TRAIL binding to DR4. Our results suggest that the altered DR4 alleles may be associated with, and should be investigated additionally as potential markers for, predisposition to common malignancies.

Lung cancer is the leading cause of cancer-related mortality in both men and women in the United States. In 1996, there were 177,000 new cases reported in the United States. The overall 5-year survival is 13%. The WHO classifies lung cancer into two groups, SCLC3 and NSCLC, which includes squamous cell carcinoma, adenocarcinoma, and large cell carcinoma (1).

Multiple genetic lesions have been described in lung cancer (2). Mutations of both dominant oncogenes and of tumor suppressor genes have been noted. Allelic loss (LOH) at chromosomal regions 3p and 9p is noted in both invasive lung cancer and preneoplastic lesions (2, 3, 4). Allelic loss of chromosome 8p21–22 is a frequent event in various cancers including lung, prostate, colon, hepatocellular carcinoma, and HNSCC (5, 6, 7, 8, 9). Deletions of 8p21–23 have been detected in high frequency as an early event in both SCLC and NSCLC cell lines and primary tumors (10).

The family of TRAIL receptors, including the proapoptotic DR4 and KILLER/DR5, as well as the decoy receptors TRID and TRUNDD, are all located on human chromosome 8p21–22 (11, 12, 13, 14). DR4 and KILLER/DR5 have significant homology with the other members of the tumor necrosis factor receptor family, which are characterized by a cysteine-rich, extracellular ligand-binding domain (14). Signaling through DR4 and KILLER/DR5 is dependent on the presence of a cytoplasmic DD in these receptors (15). Binding of TRAIL to DR4 and KILLER/DR5 initiates a signaling pathway that triggers a caspase cascade, which results in apoptosis (15, 16). TRAIL receptors are excellent candidate tumor suppressor genes, because their inactivation would be expected to result in deficient apoptotic signaling.

Mutations in the DD of KILLER/DR5 have been noted in both HNSCC (17) and lung cancer (18). In the former case, the mutation resulted in the loss of the proapoptotic function of the receptor. To date, no reports of DR4 mutations in human cancer have been identified. Previous work has identified a polymorphism in the DD of DR4 in both a bladder and ovarian cell line and in 20% of normal individuals (19). This polymorphism was an A-to-G transition at nucleotide 1322 (A1322G), resulting in the conversion of amino acid lysine (codon 441) to arginine (K441R). To additionally investigate the involvement of DR4 in human cancer, we screened 31 lung cancer cell lines (14 SCLC and 17 NSCLC), many with deletions at 8p21–23, and 21 primary NSCLC specimens for mutations in DR4. We first characterized DR4 by defining its exon-intron boundaries and generating a detailed physical map of the coding region. We then evaluated for DR4 mutations by direct sequencing of genomic DNA derived from tumor and normal samples. The results reveal two alterations in the ectodomain of DR4 that occur together in a homozygous fashion at an increased frequency in lung cancer cell lines (P = 0.023) and primary NSCLC (P = 0.002), HNSCC (P = 0.003), and gastric adenocarcinoma (P = 0.004) samples relative to age- and race-matched controls. These results suggest that the altered DR4 alleles may be associated with a predisposition to common malignancies.

Cell Line Specimens.

Thirty-one paired lung cancer cell lines (14 SCLC and 17 NSCLC) and their corresponding B-lymphoblastoid cell lines were evaluated. The cell lines used have been described previously (10). Complete demographic data were available on only 30 of the cell lines. This group was 63% male and 37% female, 90% Caucasian and 10% African-American, and had a median age of 55 years (range, 34–80 years).

Tumor and Matched Normal Specimens.

Twenty-one paired primary NSCLC samples and their corresponding matched DNA from adjacent, normal lung tissue were analyzed. DNA was isolated as described previously (20). This group was 100% Caucasian with a median age of 66 years (range, 37–87 years) and consisted of 65% males and 35% females (demographic data were available on all but 1 of the specimens). Nineteen primary HNSCC samples were evaluated. DNA was derived from freshly frozen tissue as described previously (17). Complete demographic data were available for all but 2 of the specimens. The median age of this group was 62 years (range, 41–86 years) with 76% males, 24% females, 88% Caucasians, and 12% African-Americans. No matched, normal DNA was available for these samples. Twenty-five surgically resected, gastric adenocarcinoma specimens were collected in standard RPMI 1640 and stored on ice for xenografting. Normal tissue from resected specimens or peripheral blood samples was also procured. Xenografting was performed as described previously (21). Briefly, small pieces of tumor tissue were soaked in Matrigel (Collaborative Biomed Research) and then implanted s.c. into the flanks of immunodeficient mice (nu/nu mice from Harlan or severe combined immunodeficient mice from Charles River) for xenografting growth. Passage tumors were harvested when their diameter reached ∼1 cm, and genomic DNA was then extracted by a standard organic method as described previously (22). Corresponding normal DNA from each patient was also extracted in a similar fashion. Complete demographic data were available on only 16 of these specimens. This group was 100% Caucasian with a median age of 65 years (range, 41–83 years) and composed of 62% males and 38% females.

Normal Controls.

Three groups of normal volunteers were evaluated. Group 1 consisted of 10 individuals (70% male, 30% female; 50% Caucasian, 50% Asian; median age, 36 years; age range, 25–43 years) and has been described previously (19). Group 2 consisted of 15 individuals. DNA was isolated from peripheral blood lymphocytes (20) after centrifugation over a ficoll gradient. This group had a median age of 34 years (range, 26–57 years) and was 60% male, 40% female, 60% Asian, and 40% Caucasian. Group 3 consisted of 45 individuals frequency matched for age and race to the combined lung cancer cell line, primary NSCLC, and HNSCC groups. For each individual group comparison, there were no statistically significant differences in race between the cases and the controls (P > 0.5 in each case). DNA was obtained and processed from the group 3 individuals as described previously (23).

Exon-Intron Boundaries.

Exon-intron boundaries were determined by direct DNA sequencing of subclones derived from EcoRI digestion of a P1 clone containing the DR4 gene (Genome Systems clone 18847). Template DNA was isolated by CsCl banding. Sequencing was performed using the T7 Sequenase Version 2.0 DNA Sequencing kit (United States Biochemical Corp.) as per the manufacturer’s instructions. Primers were designed from the known cDNA sequence. After sequencing, samples were heated to 75°C for 2 min and loaded on a 6% denaturing sequencing gel. Primers used in sequencing to determine exon-intron boundaries were as follows: DR4-7, 5′-AAGCCAGCCCTCGGCTCCGGGT-3′; DR4-3R, 5′-GGACACAACTCTCCCAAAG-3′; DR4-8, 5′-CAGGTCGTACCTAGCTCAGC-3′; DR4-2, 5′-CATCTGATTTACAAGCTGTACA-3′; DR4-23, 5′-AACGTCCTGGAGCCTGTAACCG-3′; DR4-20, 5′-TGTGCTGCACTTCCGGCACATC-3′; DR4-24, 5′-AGAGAAGTCCCTGCACCACG-3′; DR4-5R, 5′-AACCAAAATCACCCATATATTAT-3′; DR4-21, 5′-ACAACAGACAATCAGCACAG-3′; DR4-17, 5′-GCAATGGACATAATATATGG-3′; DR4-15S, 5′-AACACTGCCCTCTAAGAAGG-3′; DR4-16, 5′-AGCAGACACTGTGCCTCCCC-3′; DR4-19, 5′-TGGCGCTTGGGTCTCCTACG-3′; DR4-10, 5′-TCTCAGTGGGGTCAGCACCA-3′; DR4-18, 5′-GACCGGCAGAAGCTGAAGGG-3′; and DR4DD-2R, 5′-CTTCTCTTTTGCATGTCTCTCTTCC-3′.

Genomic Map of the Human DR4 Coding Region.

Most intron sizes were determined by PCR amplification of each intron using the DR4 P1 clone as a template. Primers were designed from the known cDNA sequence as well as intronic sequence derived during the course of sequencing for exon-intron boundaries. The primer pairs for each intron were as follows: intron 3, DR4-23 and DR4-20; intron 4, DR4-24 and DR4-I4R (5′-TTCTTTGTGGACACACTCG-3′); intron 6, DR4-17 and DR4-15 (5′-ACTTGGGGTCCCCTCCACAA-3′); intron 7, DR4-22 (5′-TGTGGAGGGGACCCCAAGTG-3′) and DR4-13 (5′-CCCTCGTAGGAGACCCAAGC-3′); intron 8, DR4-19 and DR4-10; and intron 9, DR4-18 and DR4DD-2R. PCR amplification was carried out in 50-μl reactions consisting of 1× buffer [16.6 mm (NH4)2SO4, 67 mm Tris-HCl (pH 8.8), 6.7 mm MgCl2, and 10 mm β-mercaptoethanol], 0.35 mm each deoxynucleoside triphosphate (dATP, dTTP, dCTP, and dGTP), 150 ng of each primer for a given intron as above, 5 units of Taq DNA polymerase (Perkin-Elmer Biosystems), and 100 ng of DR4 P1 clone template DNA. PCR conditions consisted of an initial denaturation step of 94°C for 3 min followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 62°C–68°C (depending on the primer pair) for 1 min, extension at 72°C for 2 min, and a final elongation step of 72°C for 5 min. The PCR products were electrophoresed on a 1% agarose gel, and band sizes were determined. Intron 5 size was determined by direct sequencing of the P1 clone as described above using primer DR4-21. Intron 2 size was determined using nested PCR and direct sequencing. Initial PCR was performed using primers DR4-I2F (5′-CAATTGGCACACAGCAATGG-3′) and DR4-I2RS (5′-GGGAGGCAAGCAAACAAATTG-3′). One μl of the 50-μl PCR product was used in a second PCR reaction with primers DR4-3 (5′-CTTTGGGAGAGTTGTGTCC-3′) and DR4-I2R (5′-AGGACGTTCTGATCTATGAG-3′). PCR conditions for both reactions were as follows: initial denaturation step of 94°C for 3 min followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 58°C for 45 s, extension at 72°C for 2 min, and a final elongation step of 72°C for 5 min. After electrophoresis on a 2% agarose gel to determine band size, the band was purified using a QIAquick Gel Extraction kit (Qiagen) as per the manufacturer’s instructions. The product was sequenced using the nested set of primers (DR4-3 and DR4-I2R) to confirm both intron-exon boundaries. Sequencing was performed using the SequiTherm Cycle Sequencing kit (Epicentre Technologies). After performing the sequencing reaction, samples were heated to 80°C for 5 min, loaded on a 6% denaturing polyacrylamide gel, and processed as described previously (19).

Genomic Sequencing.

Sequence analysis of the DD in exon 10 as well as the cysteine-rich, extracellular ligand-binding domain coded for by exons 3 and 4 of DR4 was performed using all of the 31 lung cancer cell lines and their matched B-lymphoblastoid cell lines. Each region was amplified from genomic DNA by PCR. Primers pairs were as follows: DD, DR4-11 (5′-CTCTGATGCTGTTCTTTGAC-3′) and DR4-12 (5′-TCACTCCAAGGACACGGCAG-3′); exon 3, DR4-I2 (5′-ATCCTCTGGGAACTCTGTGG-3′) and DR4-E3R (5′-TACCACTCCCACCTCACTGC-3′); and exon 4, DR4-E4F (5′-AAGGTCAAGGGACACGTCAGG-3′) and DR4-E4R (5′-GCTTCTGTGGTTTCTTTGAGG-3′). PCR amplification was performed in 50-μl reactions consisting of 1× PFU buffer (Stratagene), 0.35 mm each deoxynucleoside triphosphate (dATP, dTTP, dCTP, and dGTP), 150 ng of each primer for a given exon as listed above, 2.5 units PFU Turbo DNA polymerase (Stratagene), and 50 ng of template DNA. PCR conditions were as follows: initial denaturation step of 94°C for 3 min followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 58°C for 45 s, extension at 72°C for 1 min, and a final elongation step of 72°C for 5 min. After electrophoresis on a 2% agarose gel, the bands were purified using a QIAquick Gel Extraction kit (Qiagen) as per the manufacturer’s instructions. Sequencing was performed using the forward primer and the SequiTherm Cycle Sequencing kit (Epicentre Technologies) as described above.

RFLP Analysis of C626G and G422A Alleles.

To detect the C626 and 626G alleles, exon 4 was amplified as described above. Thirty-nine μl of the PCR product was then incubated in a 50-μl reaction with 20 units DraIII, 0.1 mg/ml BSA, and 1× NEB buffer #3 (New England BioLabs, Inc.) at 37°C for 90 min. The products were then separated on a 2% agarose gel. To detect the G422 and 422A alleles, exon 3 was amplified as described above. Thirty-eight μl of the PCR product was then incubated in a 50-μl reaction with 8 units of FokI, 0.1 mg/ml BSA, and 1× NEB buffer #4 (New England BioLabs, Inc.) at 37°C for 90 min. The products were then separated on a 2% agarose gel.

Statistical Analysis.

Statistical analysis was performed using the STATA software program. Unless otherwise specified, comparisons were made between each tumor group and the age- and race-matched control group.

Genomic Structure of the Human DR4 Gene.

To examine the status of DR4 in human cancer, we first determined its genomic structure by direct sequencing of subclones derived from EcoRI digestion of a P1 clone containing the DR4 gene and PCR amplification of the introns (Fig. 1). There are 10 exons and 9 introns in the coding region. Exon sizes vary from 32 to 320 bp, and intron sizes vary from 86 to 4700 bp. Exact exon-intron boundaries were determined by consensus splice site rules with all of the introns beginning with the dinucleotide GT and ending with AG. Table 1 shows the exon-intron boundaries and the exon sizes. Of note, the genomic structure is strikingly similar to the genomic structure of DR5, both with 10 exons, 9 introns, and similar intron and exon sizes (17).

A1322G Polymorphism Found in the DD of Lung Cancer Cell Lines.

The A1322G polymorphism has been identified previously in the DD of DR4 in both a bladder and ovarian cancer cell line and in 20% of normal volunteers (19). This DR4 polymorphism is believed to make the receptor less sensitive to TRAIL. Evaluation of the 31 lung cancer cell lines revealed the A1322G alteration in 2 of 31 lung cancer cell lines (data not shown). Analysis of matched normal DNA revealed that both of the tumor cell lines with the alteration were heterozygous for the alteration in their matched specimens. Of note, both of these tumor specimens have been described previously to have areas of 8p21 deletion (10). In addition, two other normal specimens demonstrated the A1322G alteration, whereas their corresponding tumor samples did not. No other mutations were identified in the DD of DR4.

Identification of a C626G Allele in Exon 4 of TRAIL Receptor DR4 in Lung Cancer.

Binding of TRAIL to its proapoptotic receptors, KILLER/DR5 and DR4, induces apoptosis (15). The ligand-binding motif of the receptors is a cysteine-rich domain, which is characteristic of the extracellular domain of the tumor necrosis factor receptor family. The principle elements of this domain are encoded by exons 3 and 4 of DR4. We performed sequence analysis of exons 3 and 4 in 31 lung cancer cell lines. Cycle sequencing of exon 4 revealed that 21 (68%) of the lung cancer cell lines had a missense alteration at nucleotide 626 changing a cytosine to a guanine (C626G), resulting in a substitution of an arginine for threonine at codon 209 (T209R) in the ectodomain of DR4 (Fig. 2). Eleven (35%) of the lung cancer cell lines exhibited only the altered allele. Analysis of matched normal DNA from the same patients revealed that 7 were homozygous and 4 were heterozygous for the alteration. Of the 10 tumors heterozygous for the C626G allele, 9 of their matched normals were also heterozygous, whereas 1 displayed only the common allele (change subsequently acquired in the tumors). Of the 10 tumors without the C626G change, 5 of the matched normals were found to be heterozygous for the alteration (Table 2). These 5 samples have all been previously noted to have deletions at 8p21 (10). Of note, there was no difference in the frequency of the C626G allele between the SCLC and NSCLC cell lines.

The C626G alteration eliminated a unique DraIII restriction site in exon 4. PCR amplification of exon 4, as described above, resulted in a 220-bp DNA product. Subsequent incubation with DraIII endonuclease yielded two smaller fragments (164 and 56 bp) if the allele was wild type but yielded only the uncut 220-bp fragment if the allele was homozygously altered. This allowed for rapid identification of specimens that were wild type, heterozygous, and homozygous at the C626G allele (Fig. 3 A).

Initial evaluation of two groups of normal healthy volunteers revealed that 3 of 25 (12%) were homozygous for the C626G change (Table 2). The frequency of being homozygous for C626G (35%) was significantly increased in patients with lung cancer as compared with the normal volunteers (χ2 = 4.1, P = 0.044). Because it is possible that racial differences between cases and normal volunteers could influence the results, we repeated the analysis with a race-matched control group. Evaluation of these 45 normal control patients matched for age and race (group 3; Table 2) revealed that 10 (22%) controls were homozygous for the altered allele. Although the frequency of homozygosity for C626G was also elevated in lung cancer cell lines relative to matched controls, this was no longer statistically significant.

Subsequent evaluation of 21 primary NSCLC tumors revealed 10 (48%) specimens homozygous for the C626G alteration, which was significantly increased relative to the age- and race-matched controls (χ2 = 4.4, P = 0.037; Table 3). The matched normal DNA for all of the 10 specimens was also homozygous for the C626G change. Of the other 11 tumor samples, 5 were heterozygous and 6 had only the wild-type allele (Table 2).

Identification of a C626G Allele in Exon 4 of TRAIL Receptor DR4 in HNSCC and Gastric Adenocarcinoma.

8p21 LOH has been described in HNSCC (5) and other malignancies as well as lung cancer. In addition, mutations in KILLER/DR5 have been identified in HNSCC, and DR4 is known to be homozygously deleted in FaDu cells, a human nasopharyngeal squamous cancer cell line (24). 8p LOH has also been described in gastric adenocarcinoma (25), although the exact chromosomal region has not yet been localized. To determine whether the increased C626G allele frequency is specific for lung cancer, we evaluated 19 HNSCC and 25 gastric adenocarcinoma samples. Nine of 19 (47%) of the HNSCC samples were homozygous, and 7 of 19 (37%) were heterozygous for the C626G alteration (Tables 2 and 3). The frequency of individuals homozygous for C626G was elevated relative to age- and race-matched controls (χ2 = 4.0, P = 0.044; Table 3). Of the gastric adenocarcinoma specimens, 11 of 25 (44%) were homozygous for the C626G alteration. Although the frequency of C626G homozygosity in the gastric adenocarcinoma samples was increased relative to the matched (group 3) controls, this did not reach statistical significance (χ2 = 3.6, P = 0.057).

Identification of a G422A Allele in Exon 3 of TRAIL Receptor DR4.

Sequence analysis of exon 3 in 31 lung cancer cell lines revealed that 11 (35%) were homozygous for a missense alteration at nucleotide 422 changing a guanine to adenine (G422A), resulting in a substitution of a histidine for arginine at codon 141 (R141H) in the ectodomain of DR4 (Tables 4 and 5). Analysis of matched, normal DNA was similar to that of the C626G allele, because the C626G and G422A allele status was identical in all of the 31 lung cancer cell lines.

PCR amplification of exon 3 of DR4 yielded a 230-bp DNA fragment. The G422A alteration created a new and unique FokI restriction site in exon 3, which on incubation with FokI endonuclease resulted in 160- and 70-bp fragments. This allowed for rapid determination of G422A status in subsequent samples (Fig. 3 B).

Subsequent evaluation of the 21 primary NSCLC, 19 HNSCC, and 25 gastric adenocarcinoma samples revealed that 48, 47, and 48%, respectively, were homozygous for the G422A alteration, whereas only 13% of the 45 age- and race-matched control samples were homozygous (Table 4). In comparison with the matched controls, the frequency of G422A homozygosity was significantly elevated in all of the three primary tumor groups (χ2 = 9.2, P = 0.002 for NSCLC; χ2 = 8.6, P = 0.003 for HNSCC; and χ2 = 10.1, P = 0.001 for gastric adenocarcinoma) and in the lung cancer cell lines (χ2 = 5.2, P = 0.023; Table 5).

Evaluation of both the C626G and G422A alleles revealed that they segregated together in 96% of all of the samples (n = 243) evaluated (tumor and normal). In only 9 cases (2 gastric adenocarcinoma samples and 7 matched controls) was the C626G and G422A allele status different. All of the lung cancer cell lines and the primary NSCLC and HNSCC samples that expressed the G422A homozygosity also expressed the C626G homozygosity. All but 1 of the gastric adenocarcinoma specimens expressing G422A homozygosity also expressed the C626G homozygosity. The frequency of being homozygous for both alleles was significantly elevated in lung cancer cell lines as well as primary NSCLC, HNSCC, and gastric adenocarcinoma samples relative to matched controls (group 3; Table 6).

The TRAIL receptor genes are all located on human chromosome 8p21–22 (11, 12, 13, 14). Previous studies have found allelic loss at chromosome 8p21 in SCLC, NSCLC, and HNSCC (5, 10). Such LOH studies suggest that one or more tumor suppressor genes might be located in this region. The proapoptotic KILLER/DR5 and DR4 are candidate tumor suppressor genes because deficient apoptosis would be expected to contribute to a transformed phenotype and tumor expansion. Mutations in the DD of KILLER/DR5 have been noted in both HNSCC (17) and lung cancer (18). In addition, DR4 is deleted in a homozygous fashion in the FaDu nasopharyngeal cancer cell line (24). It is therefore possible that point mutations in DR4 might be found in both lung cancer and HNSCC cells.

We have identified a missense alteration at nucleotide 626 of DR4 that changes a cytosine to a guanine and results in a substitution of an arginine for threonine at codon 209 (T209R). This altered allele, although present in a homozygous state in only 22% of age- and race-matched controls, was noted at a frequency of 35, 48, 47, and 44% in lung cancer cell lines and primary NSCLC, primary HNSCC, and gastric adenocarcinoma samples, respectively. We have also identified a missense alteration at nucleotide 422 of DR4 that changes a guanine to an adenine and results in a substitution of histidine for arginine at codon 141 (R141H). This altered allele was found at a frequency of 35, 48, 47, and 48% in lung cancer cell lines and primary NSCLC, primary HNSCC, and gastric adenocarcinoma samples, respectively, but was present homozygously in only 13% of matched controls. DR5 mutations described previously occurred at a frequency of <11% in tumor samples (17, 18).

It is quite possible that some of the “homozygotes” identified in our tumor samples may actually be hemizygous (have only one allele present). Our assays did not allow us to unambiguously distinguish these two possibilities. 8p21 deletions are frequent events in both lung cancer and HNSCC. Although the 8p21 status of our primary tumors is unknown, the majority of the lung cancer cell lines studied here have large areas of 8p21 deletion (10). In fact, 8 of the 11 lung cancer cell lines “homozygous” for the C626G and G422A alleles have known 8p21 deletions, whereas 3 do not exhibit allelic losses in this area (10). Functionally, the presence of either one or two copies of an altered allele in the absence of any wild-type alleles would be expected to be equivalent. In addition, there were 7 tumor samples (4 lung cancer cell lines and 3 gastric adenocarcinomas) homozygous and 7 tumor samples (5 lung cancer cell lines and 2 gastric adenocarcinomas) wild type for both alterations, the matched normal DNA of which was heterozygous. This result is likely explained by LOH, because all 9 of these lung cancer cell lines have known 8p21 deletions, and it is quite possible that the gastric adenocarcinoma specimens do as well. One possible interpretation of these results is that the C626G and G422A alleles may predispose to cancer under certain conditions, but once cancer develops they may not be required for tumor maintenance.

The significantly increased frequency of homozygosity for both alleles and the G422A allele individually in all of the four tumor groups relative to matched controls, as well as the significantly increased frequency of C626G homozygosity in primary NSCLC and HNSCC, suggests that these alleles are more than silent polymorphisms. In addition, the C626G and G422A alleles were not present homozygously in the germ-line of any matched normal sample, the tumor of which displayed only the wild-type allele. In addition, although both alterations appear to be germ line in nature, analysis of matched normal DNA reveals that the alteration occurred somatically in at least 1 case.

Although it is accepted that tumorigenesis is the result of accumulation of many genetic alterations, the frequency of the C626G and G422A alleles in tumor samples and their occurrence as both a somatic and germ-line alteration suggest that, at a minimum, the altered alleles may be a marker for predisposition to lung cancer, HNSCC, and gastric adenocarcinoma. Genetic alterations may change the susceptibility of a cell for cancer. Environmental influences may interact with genetic predispositions to contribute to tumorigenesis. Smoking is a well-defined risk factor for lung cancer (1). Perhaps one or both of these alterations confer an increased risk of development of lung cancer on heavy smokers as compared with light smokers or nonsmokers. This is a hypothesis that may be further investigated in a larger case control study.

Recently, the crystal structure of the TRAIL/DR5 complex has been described (26). Two main interface regions were identified, the 50’s loop and 90’s loop. DR4 has 69% sequence identity with DR5 in its cysteine-rich, ligand-binding motif (Fig. 4). The C626G alteration changes a threonine to an arginine (charge change) immediately 3′ to one of these two main interface regions between TRAIL and its proapoptotic receptors. The G422A alteration causes an amino acid change (arginine to histidine) just 5′ to the ligand-binding domain of DR4. This suggests that one or both of these alterations may alter TRAIL binding to DR4 and, perhaps, result in deficient apoptotic signaling.

In summary, we have identified two alterations in the ectodomain of DR4 that occur in increased frequency in lung cancer, HNSCC, and gastric adenocarcinoma samples compared with matched controls. Because other cancers have been identified with 8p21 LOH, it is quite possible that similar or other alterations in the DR4 gene have yet to be identified and may contribute to tumorigenesis or be a marker for predisposition to malignancy.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

        
1

Supported in part by a Hematology-Oncology Training Grant to the Children’s Hospital of Philadelphia (to M. J. F.). W. S. E-D. is an Assistant Investigator of the Howard Hughes Medical Institute.

                
3

The abbreviations used are: SCLC, small cell lung cancer; NSCLC, non-small cell lung cancer; HNSCC, head and neck squamous cell cancer; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand; DD, death domain; LOH, loss of heterozygosity.

Fig. 1.

Genomic structure of the coding region of the human DR4 gene. Exons are shown as black rectangles with the exon (e) number and its size (in kb) indicated above the rectangles. The introns are shown as open rectangles, within which are their respective sizes. The translation initiation codon (ATG), translation stop codon (TGA), transmembrane domain (TM), death domain (DD), and the positions and the number of the extracellular Cys (C) residues are indicated above the bar. Gray arrows, location of nucleotides 422, 626, and 1322.

Fig. 1.

Genomic structure of the coding region of the human DR4 gene. Exons are shown as black rectangles with the exon (e) number and its size (in kb) indicated above the rectangles. The introns are shown as open rectangles, within which are their respective sizes. The translation initiation codon (ATG), translation stop codon (TGA), transmembrane domain (TM), death domain (DD), and the positions and the number of the extracellular Cys (C) residues are indicated above the bar. Gray arrows, location of nucleotides 422, 626, and 1322.

Close modal
Fig. 2.

C626G polymorphism found in DR4 exon 4. Representative sequence analysis of DR4 exon 4 showing the C-to-G transversion at nucleotide 626 in both heterozygous and homozygous samples (A) and in a panel of tumors and normal population (B) is shown. Sequencing of the PCR-amplified exon 4 from genomic DNA was performed as described in the text. In B, samples from each termination mix were loaded together for easy comparisons.

Fig. 2.

C626G polymorphism found in DR4 exon 4. Representative sequence analysis of DR4 exon 4 showing the C-to-G transversion at nucleotide 626 in both heterozygous and homozygous samples (A) and in a panel of tumors and normal population (B) is shown. Sequencing of the PCR-amplified exon 4 from genomic DNA was performed as described in the text. In B, samples from each termination mix were loaded together for easy comparisons.

Close modal
Fig. 3.

RFLP analysis of C626G and G422A. In A, exon 4 was amplified and incubated with DraIII endonuclease as described in the text. Wild-type (WT) specimens have a DraIII DNA restriction site and yield two bands (164 and 56 bp). Specimens homozygous (HOM) for C626G no longer possess a DraIII site and yield only one band (220 bp). Three bands are present in heterozygous (HET) samples. A representative sample of lung cancer cell lines analyzed is shown. In B, exon 3 was amplified and incubated with FokI endonuclease as described in the text. WT specimens reveal only one band (230 bp). Specimens homozygous for G422A have a new FokI DNA restriction site and therefore yield two bands (160 and 70 bp). HET samples are diagnosed by the presence of three bands. A representative sample of lung cancer cell lines analyzed is depicted.

Fig. 3.

RFLP analysis of C626G and G422A. In A, exon 4 was amplified and incubated with DraIII endonuclease as described in the text. Wild-type (WT) specimens have a DraIII DNA restriction site and yield two bands (164 and 56 bp). Specimens homozygous (HOM) for C626G no longer possess a DraIII site and yield only one band (220 bp). Three bands are present in heterozygous (HET) samples. A representative sample of lung cancer cell lines analyzed is shown. In B, exon 3 was amplified and incubated with FokI endonuclease as described in the text. WT specimens reveal only one band (230 bp). Specimens homozygous for G422A have a new FokI DNA restriction site and therefore yield two bands (160 and 70 bp). HET samples are diagnosed by the presence of three bands. A representative sample of lung cancer cell lines analyzed is depicted.

Close modal
Fig. 4.

Homology of the cysteine-rich domain (CRD) of DR5 (amino acids 43–130) and DR4 (amino acids 147–234). Alignment map showing that DR4 has 69% identity with DR5 in its ligand-binding motif. Alignment was carried out using the MacVector 6.5 ClustalW Alignment program (Oxford Molecular Group). The two main interface regions, the 50’s loop (50s Loop) and the 90’s loop (90s Loop), between TRAIL and its receptors are indicated. Gray arrows, location of the amino acid (threonine) affected by the C-to-G alteration at nucleotide 626 (T209R) and the amino acid (arginine) affected by the G-to-A alteration at nucleotide 422 (R141H).

Fig. 4.

Homology of the cysteine-rich domain (CRD) of DR5 (amino acids 43–130) and DR4 (amino acids 147–234). Alignment map showing that DR4 has 69% identity with DR5 in its ligand-binding motif. Alignment was carried out using the MacVector 6.5 ClustalW Alignment program (Oxford Molecular Group). The two main interface regions, the 50’s loop (50s Loop) and the 90’s loop (90s Loop), between TRAIL and its receptors are indicated. Gray arrows, location of the amino acid (threonine) affected by the C-to-G alteration at nucleotide 626 (T209R) and the amino acid (arginine) affected by the G-to-A alteration at nucleotide 422 (R141H).

Close modal
Table 1

Exon-intron boundaries of the human DR4 TRAIL receptor gene

Exon-intron boundaries of the human DR4 TRAIL receptor gene
Exon-intron boundaries of the human DR4 TRAIL receptor gene
Table 2

C626G allele status in patients with cancer and matched controls

C626G allele status in patients with cancer and matched controls
C626G allele status in patients with cancer and matched controls
Table 3

Increased frequency of C626G homozygosity in primary NSCLC and HNSCC specimens relative to matched controls

Homozygousχ2P                  aHeterozygousWild TypeTotal
Lung cancer cell lines 11 (35%) 1.6 0.204 10 (32%) 10 (32%) 31 
Primary NSCLC 10 (48%) 4.4 0.037 5 (24%) 6 (29%) 21 
HNSCC 9 (47%) 4.0 0.044 7 (37%) 3 (16%) 19 
Gastric adenocarcinoma 11 (44%) 3.6 0.057 6 (24%) 8 (32%) 25 
Matched controls 10 (22%)   20 (44%) 15 (33%) 45 
Homozygousχ2P                  aHeterozygousWild TypeTotal
Lung cancer cell lines 11 (35%) 1.6 0.204 10 (32%) 10 (32%) 31 
Primary NSCLC 10 (48%) 4.4 0.037 5 (24%) 6 (29%) 21 
HNSCC 9 (47%) 4.0 0.044 7 (37%) 3 (16%) 19 
Gastric adenocarcinoma 11 (44%) 3.6 0.057 6 (24%) 8 (32%) 25 
Matched controls 10 (22%)   20 (44%) 15 (33%) 45 
a

Comparisons in each case are between cancer sample and matched controls.

Table 4

G422A allele status in patients with cancer and matched controls

G422A allele status in patients with cancer and matched controls
G422A allele status in patients with cancer and matched controls
Table 5

Increased frequency of G422A homozygosity in lung cancer cell lines, primary NSCLC, HNSCC, and gastric adenocarcinoma specimens relative to matched controls

Homozygousχ2P                  aHeterozygousWild TypeTotal
Lung cancer cell lines 11 (35%) 5.2 0.023 10 (32%) 10 (32%) 31 
Primary NSCLC 10 (48%) 9.2 0.002 5 (24%) 6 (29%) 21 
HNSCC 9 (47%) 8.6 0.003 7 (37%) 3 (16%) 19 
Gastric adenocarcinoma 12 (48%) 10.1 0.001 5 (20%) 8 (32%) 25 
Matched controls 6 (13%)   27 (60%) 12 (27%) 45 
Homozygousχ2P                  aHeterozygousWild TypeTotal
Lung cancer cell lines 11 (35%) 5.2 0.023 10 (32%) 10 (32%) 31 
Primary NSCLC 10 (48%) 9.2 0.002 5 (24%) 6 (29%) 21 
HNSCC 9 (47%) 8.6 0.003 7 (37%) 3 (16%) 19 
Gastric adenocarcinoma 12 (48%) 10.1 0.001 5 (20%) 8 (32%) 25 
Matched controls 6 (13%)   27 (60%) 12 (27%) 45 
a

Comparisons in each case are between cancer sample and matched controls.

Table 6

Increased frequency of homozygosity for both the C626G and G422A alleles in lung cancer cell lines, primary NSCLC, HNSCC, and gastric adenocarcinoma specimens relative to matched controls

Homozygousχ2P                  aAltered alleleWild TypeTotal
Lung cancer cell lines 11 (35%) 5.2 0.023 21 (68%) 10 (32%) 31 
Primary NSCLC 10 (48%) 9.2 0.002 15 (71%) 6 (29%) 21 
HNSCC 9 (47%) 8.6 0.003 16 (84%) 3 (16%) 19 
Gastric adenocarcinoma 11 (44%) 8.2 0.004 17 (68%) 8 (32%) 25 
Matched controls 6 (13%)   30 (67%) 12 (27%) 45 
Homozygousχ2P                  aAltered alleleWild TypeTotal
Lung cancer cell lines 11 (35%) 5.2 0.023 21 (68%) 10 (32%) 31 
Primary NSCLC 10 (48%) 9.2 0.002 15 (71%) 6 (29%) 21 
HNSCC 9 (47%) 8.6 0.003 16 (84%) 3 (16%) 19 
Gastric adenocarcinoma 11 (44%) 8.2 0.004 17 (68%) 8 (32%) 25 
Matched controls 6 (13%)   30 (67%) 12 (27%) 45 
a

Comparisons in each case are between cancer sample and matched controls.

1
Ginsberg R. J., Vokes E. E., Raben A. Non-small cell lung cancer Ed. 5 DeVita V. T. Hellman S. Rosenberg S. A. eds. .
Cancer: Principles and Practice of Oncology
,
:
858
-911, J. B. Lippincott Co. Philadelphia  
1997
.
2
Minna J. D., Sekido Y., Fong K., Gazdar A. F. Molecular biology of lung cancer Ed. 5 DeVita V. T. Hellman S. Rosenberg S. A. eds. .
Cancer: Principles and Practice of Oncology
,
:
849
-857, J. B. Lippincott Co. Philadelphia  
1997
.
3
Hung J., Kishimoto Y., Sugio K., Virmani A., McIntire D. D., Gazdar A. F. Allele-specific chromosome 3p deletions occur at an early stage in the pathogenesis of lung carcinomas.
JAMA
,
273
:
558
-563,  
1995
.
4
Kishimoto Y., Sugio K., Mitsudomi T., Oyama T., Virmani A., McIntire D. D., Gazdar A. F. Allele-specific loss of chromosome 9p in preneoplastic lesions accompanying non-small cell lung cancers.
J. Natl. Cancer Inst.
,
87
:
1224
-1229,  
1995
.
5
El-Naggar A. D., Coombes M. M., Batsakis J. G., Hong W. K., Goepfert H., Kagan J. Localization of chromosome 8p regions involved in early tumorigenesis of oral and laryngeal squamous carcinoma.
Oncogene
,
16
:
2983
-2987,  
1998
.
6
Fujiwara Y., Ohata H., Emi M., Okui K., Koyama K., Tsuchiya E., Nakajima T., Monden M., Mori T., Kurimasa A., Oshimura M., Nakamura Y. A 3-Mb physical map of chromosome region 8p21.3–p22, including a 600-kb region commonly deleted in human hepatocellular carcinoma, colorectal cancer, and non-small cell lung cancer.
Genes Chromosomes Cancer
,
10
:
7
-14,  
1994
.
7
Kagan J., Stein J., Babaian R. J., Joe Y-S., Pisters L. L., Glassman A. B., von Eschenbach A. C., Troncoso P. Homozygous deletions at 8p22 and 8p21 in prostate cancer implicate these regions as the sites for candidate tumor suppressor genes.
Oncogene
,
11
:
2121
-2126,  
1995
.
8
Yaremko M. L., Wasylyshyn M. L., Paulus K. L., Michelassi F., Westbrook C. A. Deletion mapping reveals two regions of chromosome 8 allele loss in colorectal carcinomas.
Genes Chromosomes Cancer
,
10
:
1
-6,  
1994
.
9
MacGrogan D., Levy A., Bostwick D., Wagner M., Wells D., Bookstein R. Loss of chromosome arm 8p loci in prostate cancer: mapping by quantitative allelic imbalance.
Genes Chromosomes Cancer
,
10
:
151
-159,  
1994
.
10
Wistuba I. I., Behrens C., Virmani A. K., Milchgrub S., Syed S., Lam S., Mackay B., Minna J. D., Gazdar A. F. Allelic losses at chromosome 8p21–23 are early and frequent events in the pathogenesis of lung cancer.
Cancer Res.
,
59
:
1973
-1979,  
1999
.
11
Marsters S. A., Sheridan J. P., Pitti R. M., Huang A., Skubatch M., Baldwin D., Yuan J., Gurney A., Goddard A. D., Godowski P., Ashkenazi A. A novel receptor for Apo2L/TRAIL contains a truncated death domain.
Curr. Biol.
,
7
:
1003
-1006,  
1997
.
12
Degli-Esposti M. A., Smolak P. J., Walczak H., Waugh J., Huang C-P., DuBose R. F., Goodwin R. G., Smith C. A. Cloning and characterization of TRAIL-R3, a novel member of the emerging TRAIL receptor family.
J. Exp. Med.
,
186
:
1165
-1170,  
1997
.
13
Degli-Esposti M. A., Dougall W. C., Smolak P. J., Waugh J., Smith C. A., Goodwin R. G. The novel receptor TRAIL-R4 induces NF-κB and protects against TRAIL-mediated apoptosis, yet retains an incomplete death domain.
Immunity
,
7
:
813
-820,  
1997
.
14
Wu G. S., Burns T. F., McDonald E. R., III, Jiang W., Meng R., Krantz I. D., Kao G., Gan D-D., Zhou J-Y., Muschel R., Hamilton S. R., Spinner N. B., Markowitz S., Wu G., El-Deiry W. S. KILLER/DR5 is a DNA damage-inducible p53-regulated death receptor gene.
Nat. Genet.
,
17
:
141
-143,  
1997
.
15
Ashkenazi A., Dixit V. M. Death receptors: signaling and modulation.
Science (Wash. DC)
,
281
:
1305
-1308,  
1998
.
16
Bodmer J-L., Holler N., Reynard S., Vinciguerra P., Schneider P., Juo P., Blenis J., Tschopp J. TRAIL receptor-2 signals apoptosis through FADD and caspase-8.
Nat. Cell Biol.
,
2
:
241
-243,  
2000
.
17
Pai S. I., Wu G. S., Ozoren N., Wu L., Jen J., Sidransky D., El-Deiry W. S. Rare loss-of-function mutation of a death receptor gene in head and neck cancer.
Cancer Res.
,
58
:
3513
-3518,  
1998
.
18
Lee S. H., Shin M. S., Kim H. S., Lee H. K., Park W. S., Kim S. Y., Lee J. H., Han S. Y., Park J. Y., Oh R. R., Jang J. J., Han J. Y., Lee J. Y., Yoo N. J. Alterations of the DR5/TRAIL receptor 2 gene in non-small cell lung cancer.
Cancer Res.
,
59
:
5683
-5686,  
1999
.
19
Kim K., Fisher M. J., Xu S-Q., El-Deiry W. S. Molecular determinants of response to TRAIL in killing of normal and cancer cells.
Clin. Cancer Res.
,
6
:
335
-346,  
2000
.
20
Herrmann B. G., Frischauf A. M. Isolation of genomic DNA.
Methods Enzymol.
,
150
:
182
-183,  
1987
.
21
Hahn S. A., Seymour A. B., Hoque A. T., Schutte M., da Costa L. T., Redston M. S., Caldas C., Weinstein C. L., Fischer A., Yeo C. J., Hruban R. H., Kern S. E. Allelotype of pancreatic adenocarcinoma using a xenograft model.
Cancer Res.
,
55
:
4670
-4675,  
1995
.
22
Goelz S., Hamilton S., Vogelstein B. Purification of DNA from formaldehyde-fixed and paraffin-embedded human tissue.
Biochem. Biophys. Res. Commun.
,
130
:
118
-126,  
1985
.
23
Walker A. H., White D. L., Jaffe J. M., Kanetsky P. A., Rebbeck T. R. Collection of genomic DNA by buccal swabs for PCR-based biomarker assays.
Environ. Health Perspect.
,
107
:
3
-7,  
1999
.
24
Ozoren N., Fisher M. J., Kim K., Liu C. X., Genin A., Shifman Y., Dicker D. T., Spinner N. B., Lisitsyn N. A., El-Deiry W. S. Homozygous deletion of the death receptor DR4 gene in a nasopharyngeal cancer cell line is associated with TRAIL resistance.
Int. J. Oncol.
,
16
:
917
-925,  
2000
.
25
Yustein A. S., Harper J. C., Petroni G. R., Cummings O. W., Moshalik C. A., Powell S. M. Allelotype of gastric adenocarcinoma.
Cancer Res.
,
59
:
1437
-1441,  
1999
.
26
Hymowitz S. G., Christinger H. W., Fuh G., Ultsch M., O’Connell M., Kelley R. F., Ashkenazi A., de Vos A. M. Triggering cell death: the crystal structure of Apo2L/TRAIL in a complex with death receptor 5.
Mol. Cell
,
4
:
563
-571,  
1999
.