Inactivation of Human SRBC, Located within the 11p15.5-p15.4 Tumor Suppressor Region, in Breast and Lung Cancers¹

Allelic loss at the chromosome segment 11p15.5 is frequently observed in a variety of adult solid tumors including breast (1, 2, 3), lung (4, 5), ovarian (6, 7), bladder (8), stomach (9), and adrenal cortical (10) cancers as well as some childhood tumors including Wilms’ tumor (11, 12, 13), rhabdomyosarcoma (14, 15), and hepatoblastoma (16). In breast cancer, the frequency of LOH³ for 11p15.5 detected in invasive ductal breast cancers is around 30–60% (1, 2). Two distinct regions on chromosome 11p15 that are subjected to LOH in breast cancer have been identified and refined by Karnik et al. (Fig. 1,A; Ref. 3). Region 1 extends from D11S1318 to D11S4088, overlapping with the previous identified LOH regions in breast cancer (2, 17), ovarian carcinoma (7), Wilms’ tumor (14), rhabdomyosarcoma (14, 15), and gastric adenocarcinoma (9). Region 2 is defined by D11S1338 and D11S1323, which overlaps with LOH regions described for breast cancer (18), NSCLC (5), and Wilms’ tumor (11). Region 1 also overlaps with a locus that contains several imprinted genes such as H19, IGF2, INS, TH, HASH2, KVLQT1, and p57Kip2 (Fig. 1,A), some of which are implicated in a variety of cancers (19). The breakpoints of chromosome translocation and inversion associated with malignant rhabdoid tumors and Beckwith-Wiedemann syndrome are also within this region (20). Region 2 is more centromeric (Fig. 1 A) and LOH in Region 2 is associated with clinical parameters of more aggressive breast tumors and a poorer prognostic indicators, such as aneuploidy, high S-phase fraction, and the presence of metastasis in regional lymph nodes (3). Similarly, 11p allele loss is also observed in lung cancer at a frequency ranging between 11 and 50% (4, 5). LOH at 11p in lung cancer is correlated with advanced T stage and nodal involvement in NSCLC (21). O’Brant and Bepler (22) have mapped two regions (LOH11A and LOH11B) on chromosome 11p15.5 that have frequent allele loss in lung cancer. LOH11A is centromeric between loci D11S1758 and D11S12, and LOH11B is telomeric between HRAS and D11S1363.

Nonrandom chromosome deletion and allele loss often marks the site of inactivation of TSGs residing in a particular chromosomal region. Thus, the finding of frequent LOH in breast and lung cancer strongly suggests the presence of one or several 11p TSGs. The observation that physical transfer of 11p chromosomal fragments into tumor cell lines was able to reverse the tumorigenic phenotype of cancer cells provided additional functional evidence for the existence of TSGs in this region (23, 24, 25). Because of the complex LOH pattern in this region and the lack of homozygous deletions to help target a positional cloning effort, it has been hard to identify candidate TSG(s) for this chromosome region. In this study, we described such a candidate, the hSRBC gene. The gene was isolated in a two-hybrid screen for proteins interacting with the product of the breast cancer susceptibility gene BRCA1. RH mapping localized hSRBC gene to chromosome region 11p15.5 between D11S1323 and D11S1338. Moreover, several mutations including frame shift and truncation mutations were identified in ovarian and lung cancer cell lines. Most significantly, the expression of hSRBC mRNA and protein was down-regulated in a large fraction of breast and lung cancer cell lines, and the down-regulation is associated with hypermethylation in the hSRBC promoter region. The evidence provided here suggests that inactivation of hSRBC may contribute to the pathogenesis of several types of cancer.

All of the breast and lung cancer cell lines were from the Hamon Center collection (UT Southwestern) generated by the authors and were cultured with RPMI 1640 containing 5% fetal bovine serum. Most are also available from American Type Culture Collection (Chantilly, VA). Ovarian cancer cell line 2008 was provided by Dr. C. Muller (UT Southwestern) and cultured in DMEM with 10% fetal bovine serum. All of the genomic DNAs and total RNAs of the cultured cell lines were prepared as described previously (26). HMECs, normal human bronchial/tracheal epithelial cells (cultured with retinoic acid), and normal human lung SAECs were purchased from Clonetics Corporation, cultured as recommended, and analyzed at passages 7–10. Formalin-fixed, paraffin-embedded primary lung and breast carcinoma specimens were retrieved from the surgical pathology archives of UT Southwestern-affiliated hospitals, microdissected using a laser capture microscope (27), and DNA extracted as described previously (28). Northern blotting and Western immunoblotting were performed as described previously (29). Twenty μg of total RNA prepared from the indicated cell lines were electrophoresed, transferred, and hybridized with a full open reading frame cDNA probe for hSRBC using standard procedures. The monoclonal antibody against hSRBC was generated by immunizing Balb/C mice with a glutathione S-transferase fusion protein containing amino acids 201–261 of hSRBC by standard fusion and screening procedures.

BAC clone 368 H20 was obtained from Research Genetics, maintained as recommended, and DNA isolated using a Midi plasmid kit (Qiagen). The PCR reaction was performed with Amplitaq Gold (Perkin-Elmer) in 1× buffer containing 1.5 mm MgCl₂, 5% DMSO, 200 μM dNTP, 0.2 mM primers, and 50 ng of BAC DNA in a 20-μl reaction. The primers for detecting hSRBC are: forward, 5′-GGAGCAGAGCGGTCAGGGATC-3′, and reverse, 5′-GGACCGTTTGAGGTCACTGAC-3′; and the cycle conditions for hSRBC-specific primers were 75°C-65°C touchdown [one cycle at 95°C (5 min); 10 touchdown cycles at 94°C (30 s), 75°C (30 s) with a decrease of 1°C each cycle, and at 72°C for 30 s; followed by 30 cycles of 94°C (30 s), 65°C (30 s), and 72°C (30 s), with a final extension at 72°C for 7 min]. The primers for D11S1323 are: forward, 5′-TCTAATTTCTTCTCCCACCC-3′, and reverse, 5′-TGGGAGTTTCTCTGTGCTAG-3′. The primers for D11S1338 are: forward, 5′-TCAGAAATCTGATGGAAAAGTC-3′, and reverse, 5′-TGCTACTTATTTGGAGTGTGAA-3′. The cycle conditions for D11S1323- and D11S1338-specific primers were 68°C-58°C touchdown [one cycle at 95°C (5 min); 10 touchdown cycles at 94°C (30 s), 68°C (30 s) with a decrease of 1°C each cycle and at 72°C for 30 s; followed by 30 cycles of 94°C (30 s), 58°C(30 s) and 72°C (30 s), with a final extension at 72°C for 7 min].

Mapping of the chromosome location of the hSRBC gene was performed using Stanford TNG RH panel (Research Genetics). The PCR reaction for hSRBC was performed as described above for the BAC DNA. The RH raw data from these tests are: 0100000100 000000R100 0000000000 1000110001 0001000000 1000010010 0001001001 0000000000 0100000010.

The primer pairs of hSRBC gene for PCR amplification from 100 ng of genomic DNA are: 5′-GGAGCAGAGCGGTCAGGGATC-3′ and 5′-GGACCGTTTGAGGTCACTGAC-3′ for exon 1; and 5′-GCTGTGTCCGTCACATGCAG-3′ and 5′-AGGCAGGCAACACCAGCCC-3′ for exon 2. The PCR reaction was performed as described above (75°C-65°C touchdown). The PCR products were purified with Qiagen PCR purification kit and sequenced with ABI 377 automatic DNA sequencer.

Genomic DNA (1 μg) was treated with sodium bisulfite as described previously (30). Forty ng of treated DNA was used as template in PCR amplification reaction in a volume of 25 μl. The primer pairs for PCR amplification are: forward, 5′-ATTTTTATTGGTGTGGGAGG-3′, and reverse, 5′-CCCTACAAACCCTCTCACTCT-3′. The PCR reaction was performed in 1× Hot Star Taq buffer containing 2.5 mm MgCl₂, 500 nm primers, 160 μM dNTP, and 1 unit of Hot Star Taq (Qiagen). The cycle conditions were: one cycle at 94°C (12 min); 2 touchdown cycles at 94°C (30 s), 61°C (45 s) with a decrease of 1°C each cycle, and at 72°C for 30 s; followed by 35 cycles of 94°C (30 s), 60°C(30 s) and 72°C (30 s), with a final extension at 72°C for 7 min. A second PCR was performed using 0.5 μl of the first reaction with the same conditions. The PCR products were purified with Qiagen PCR purification kit, and sequenced with an ABI 377 automatic DNA sequencer.

MCF7 cells grown to 30% confluency in 24-well plates were treated with 5′-azacytidine (Sigma Chemical Co.) at different concentrations for 2 days. Then, different concentrations of TSA were added, and cells were incubated for 2 more days before harvest. On harvest, cells were lysed directly in SDS-PAGE loading buffer, and the expression of hSRBC was determined by immunoblotting with the hSRBC monoclonal antibody.

In an effort to study the BRCA1-mediated tumor suppression pathway, a cDNA clone was isolated when a normal mammary epithelial cDNA library was screened by the yeast two-hybrid method, using a recombinant protein containing amino acids 1–304 of BRCA1 as the “bait” (31). This was performed by L. C. W. in R. B.’s laboratory at UT Southwestern. The cDNA contains an open reading frame of 261 amino acids (Accession no. AF339881). BLAST search revealed the cDNA encodes the human orthologue of the rat SRBC gene that was isolated in a far-Western approach using PKC-δ isoform (PKC-δ) as the probe (32). Thus, we refer to the human orthologue as hSRBC. hSRBC is also identical to the “unknown” NH₂-terminal cellular fusion protein in one form of activated c-Raf-1 (X06409; Ref. 33). hSRBC is 81% identical to its rat orthologue at the amino acid level, and, like rat SRBC, the amino acid sequence of hSRBC reveals a leucine zipper-like motif in its NH₂-terminal region. As was reported previously, SRBC is also homologous to several molecules in the database: both human and mouse SDPR factor (AF085481 and S67386, respectively; Ref. 34) and PTRF (AF036249; Ref. 35), and a putative leucine-zipper protein that appears to be the chicken orthologue of PTRF (D82079 and D26315).

A BLAST search of the NCBI expressed sequence tag database found many expressed sequence tags matched to the sequence of hSRBC gene, one of which (AA030008) has already been placed at chromosome 11p15.5-p15.4 on the GeneBridge 4 RH map of NCBI GeneMap 99.⁴ To confirm the chromosomal location of hSRBC, we performed RH mapping using the Stanford TNG RH panel (Research Genetics), which gives higher resolution than GB4 and G3 panel. The hSRBC gene was shown to be closely linked to SHGC-2070 (Genome Database locus D11S1323) at 11p15.5 on the SHGC G3 map,⁵ with a log of odds (LOD) score of 12.48 and the distance between hSRBC and D11S1323 of 13 centirays at 50,000 rads (cR_50,000; ∼26–39 kb). Further analysis using the RH raw data of telomeric marker SHGC-14206 (RHdb: RH95600) and centromeric marker SHGC-30684 (RHdb: RH96895) suggested that the hSRBC gene is located centromeric to D11S1323, thus residing between D11S1323 and D11S1338 (Fig. 1,A), right within the LOH Region 2 as described by Karnik et al. (3), and very close to the LOH region LOH11A in lung cancer as described by O’Briant and Bepler (22). Moreover, a specific DNA fragment could be amplified with hSRBC-, or D11S1323-specific primer pairs from the same BAC clone (368 H20 from Research Genetics; Fig. 1 B), confirming that they are less than 200 kb apart. This BAC clone was partially sequenced providing 7466 bp of genomic DNA (AF408198).

Comparison of the cDNA sequence of hSRBC with its corresponding genomic DNA sequence, which was obtained by sequence analysis of BAC clone 368 H20 (Research Genetics), revealed that the coding region of hSRBC gene is organized into two exons: exon 1 covers codons 1–128, and exon 2 contains codons 129–261. The two exons are separated by a small intron of 529 bp. Also PCR-single-strand-conformation-polymorphism analysis of genomic DNAs from 53 normal Caucasian individuals identified five common polymorphisms in the hSRBC coding sequence (Table 1). Because the hSRBC gene was mapped to a region that shows frequent LOH in multiple types of human cancers including breast and lung, we first examined LOH within the hSRBC coding region in lung cancer cell lines with single-strand conformation polymorphism analysis, using the polymorphisms identified in hSRBC coding sequence (data not shown). LOH was detected at a frequency of 46% in 15 informative pairs of lung cancer cell lines compared with their matched B lymphoblastoid cell lines (data not shown). In a prior separate study, LOH was 44% at D11S1338, a marker near hSRBC, in lung cancer cell lines (36).

Mutational analysis of hSRBC gene was then performed by PCR amplification followed by DNA sequencing using primer pairs covering the whole coding region and exon/intron junctions on 135 genomic DNAs derived from 35 breast cancer cell lines, 6 ovarian cancer cell lines, 60 lung cancer cell lines, 10 microdissected breast primary tumors, and 24 microdissceted primary lung tumors. Mutations were detected in four cancer cell lines (Table 1). Direct sequencing of PCR amplified genomic DNA prepared from the 2008 ovarian cancer cell line revealed a heterozygous 1-bp deletion at nucleotide position 89 (G18 frameshift to 23 stop), resulting in the loss of most of the coding region of hSRBC. The genomic DNA of the NSCLC cell line NCI-H358 contains a heterozygous nonsense mutation at nucleotide position 604, changing an arginine (CGA) to a termination (TGA) codon at residue 190. The SCLC cell line NCI-H510 harbors a heterozygous 21-bp deletion at nucleotide position 177, resulting in the removal of seven amino acids; SCLC cell line (NCI-H2081) contains a heterozygous missense mutation (C611G), resulting in an A192G amino acid substitution. This A192 is conserved among all of the homologous proteins. None of the four alterations are present in 60 normal control DNAs and 134 other cancer cell lines or tumor DNAs (total 388 chromosomes tested). However, we cannot formally exclude that they represent rare polymorphisms, because no matched normal samples were available to allow us to directly determine whether the alterations are tumor-acquired. All of the alterations were confirmed with multiple PCR amplification and sequence analysis using two batches of independent DNA preparations.

The hSRBC protein is expressed in both HMEC, and NHBE cell cultures as well as in a BRCA1 mutant breast cancer cell line HCC1937 (Fig. 2,A), as detected by immunoblot (37). In addition, hSRBC mRNA was also detected in NHBE cells and SAEC (Fig. 2,B). By contrast, loss of hSRBC protein expression was observed in 4 of 11 breast (Fig. 2C) and 26 of 32 lung cancer cell lines (Fig. 2,D). In contrast to the conventional two-hit model, in those tumor cell lines in which a hSRBC mutation was detected, the wild-type allele was also retained. However, immunoblotting with hSRBC-specific antibody failed to detect the wild-type protein products in three hSRBC mutant cell lines tested (Fig. 2 E), indicating that the wild-type allele was silent. These expression studies suggested that down-regulation of expression of hSRBC gene occurs commonly in sporadic breast and lung cancers.

Selective down-regulation of tumor suppressor gene expression is often associated with abnormal methylation of CpG dinucleotides in the promoter region (38). We submitted the 7466 bp of genomic DNA containing hSRBC to our software program PANORAMA, which indicated there were three CpG islands (at nt 1360–1662, 2816–3082, 3338–5636 of the AF408198 sequence; Ref. 39). To investigate the mechanism of hSRBC down-regulation in tumor cells, we examined the methylation status of CpG dinucleotides in a 210-bp fragment that was 240 bp upstream of the hSRBC translation initiation codon. This region contained a CCAAT box motif, a TATA box motif, and two Ap-1 consensus-binding sites (Fig. 3,A). Bisulfite genomic sequencing revealed that the methylation status of CpG dinucleotides was correlated with their hSRBC expression levels (Fig. 3,B). In two cell lines (HTB129 and NCI-H2052) that express hSRBC at normal level, eight of nine CpG dinucleotides in this region examined by bisulfite genomic sequencing were completely unmethylated. In 11 tumor cell lines in which expression of hSRBC was undetectable (HTB131, HTB132, UCI 101, 2008, MCF7, NCI-H187, NCI-H378, NCI-H1092, HTB19, NCI-H1299, and NCI-H1672), we found that 3 (NCI-H378, NCI-H1299, and NCI-H1672) were completely methylated and 8 were heavily methylated, with mCpG density ranging from 3.01 to 4.52 per 100 bp. In two cell lines (HTB130 and NCI-H1437) that have reduced hSRBC expression level, the CpG dinucleotides in the promoter region were also partially methylated, with mCpG densities varying from 2.11 to 2.26 per 100 bp. HMECs that express hSRBC (Fig. 2,A) were found to have all nine CpG dinucleotides unmethylated (not shown in Fig 3,B). In addition, the expression of hSRBC in MCF7 cells was partially restored by treatment with the histone deacetylase inhibitor TSA, whereas treatment with both demethylation reagent 5′-azacytidine and TSA additionally induced the expression of hSRBC in MCF7 cells (Fig. 3 C). Finally, we found four lung cancer lines (HCC78, NCI-H1993, H2052, and H2347) that expressed hSRBC and were also heterozygous for single nucleotide polymorphisms within the gene. We performed reverse transcription-PCR expression analysis on these and found that both alleles were expressed. Thus, at least in these cases, we can rule out imprinting of one of the alleles.

We examined the methylation status of the hSRBC promoter region in primary tumor tissues (Fig. 3 D). In five breast tumors examined, one was largely unmethylated (case 13 mCpG density, 0.90), one was partially methylated (case 67 mCpG density, 1.20), and three were heavily methylated (cases 2, 32, and 70 mCpG density, >3.0). In 14 NSCLCs examined, 3 were partially methylated (mCpG density, between 1.0 and 3.0), 9 were heavily methylated (mCpG density, >3.0), and 2 were completely methylated. Although the available hSRBC-specific antibody did not allow us to examine the expression of hSRBC in primary tumor tissues, the methylation study in primary tumor tissues indicated that 3 (60%) of 5 breast tumors and 11 (79%) of 14 lung tumors underwent hSRBC promoter methylation.

We have provided evidence suggesting that hSRBC is a candidate TSG at chromosome region 11p15.5-p15.4, predominantly undergoing epigenetic inactivation by promoter region methylation. Previously, LOH in this region was found to be associated with several types of adult and childhood tumors. Mutations of the hSRBC gene, including frame shift and truncation mutations, were detected in four cancer cell lines, which suggested that although rare, they do occur. Most importantly, we detected a high frequency (30 of 43) of loss of expression of hSRBC mRNA and protein in breast and lung cancer cell lines, with retained hSRBC expression in normal epithelial cultures from these organs. The loss of hSRBC expression is associated with hSRBC promoter region hypermethylation. This mechanism has been suggested as an alternative way to inactivate TSGs (40). Although we did not find hSRBC mutations in the limited numbers of primary tumors that we examined, hSRBC promoter region hypermethylation was observed in 60% of primary breast tumors and 79% of primary lung cancers, which suggested that inactivation of hSRBC gene function by epigenetic mechanisms is common in breast and lung cancers. In addition, the cancer lines with amino acid sequence alterations retained the wild-type allele yet failed to express hSRBC and exhibited promoter region methylation. Thus, promoter region methylation appears to be the dominant mode of inactivation of hSRBC in human lung and breast cancers. In addition, among the five polymorphisms detected in the hSRBC coding region, three (C59G, T509C, and C799T) result in amino acid changes and two (C59G and C799T) occur at a very low frequency. It is possible that such low-frequency, dramatic amino acid sequence alterations may confer interindividual variation in hSRBC functions and may result in different cancer risks.

SRBC belongs to a superfamily of proteins that thus far has three members: SDPR, SRBC, and PTRF. Mouse SDPR (mSDPR; S67386; 418 aa) was the first to be cloned in a differential hybridization screen, in which SDPR mRNA levels were found to be induced by serum starvation but not by contact inhibition in NIH3T3 cells (34). Later the human orthologue of mouse SDPR (hSDPR; AF085481; 425 aa) was isolated as a substrate and binding protein of PKC-α isoform (PKC-α) (41). Human SDPR resides on the caveolae of the plasma membrane and may help to recruit PKC-α to caveolae. Human SDPR is also identical to PS-68 (42), a previously characterized phosphatidylserine-binding protein purified from human platelets (43). The overall amino acid identity between human and mouse SDPR is 83%, with 92% identical in the first 307 amino acids. Rat SRBC (D85435; 263 aa), isolated as a PKC-δ substrate and interacting protein, shares several similarities with SDPR, including binding phosphatidylserine and the regulatory domain of PKC in the absence of Ca²⁺, and being phosphorylated by PKC in vitro (32). HSRBC described in this paper has an overall amino acid identity of 81% to the rat orthologue, with the first 190 amino acids 91% identical. The third member of the family, PTRF (AF036249; 392 aa), identified in a yeast two-hybrid screen with terminator TTF-1, induces dissociation of paused ternary rRNA transcription complexes (35). PTRF is also a very conserved protein. Mouse PTRF shows 72% identity to the chicken orthologue (D82079 and D26315). The homology among different family members is mostly localized in two regions, homologous region 1 at the NH₂ terminus (aa 20–139 of hSRBC), which includes the leucine-zipper motif that is completely conserved among three members (Fig. 4,A), and homologous region 2 in the middle (172–194 aa of hSRBC), which contains the putative PKC phosphorylation site (Fig. 4,B). Strikingly, in addition to the sequence similarity, the domain organization patterns among the three members are very similar (Fig. 4 C). Each of them has the homologous region 1 that contains the leucine-zipper motif at the NH₂ terminus, followed by an acidic region that has an aspartic/glutamic acid content above 40% (aa 143–173 of hSRBC). Immediately after the acidic region is a basic region (aa 174–197 of hSRBC) that contains about 38% basic residues and that overlaps and extends the second homologous region. Both SDPR and PTRF have a second acidic region (30% aspartic/glutamic acid) after the basic region. SRBC differs from SDPR and PTRF in that it has a proline-rich region (31% P) instead. The sequence and structural resemblance suggests that the three members may have similar functions that are yet to be discovered.

Although the biochemical function of SRBC is still unknown, several lines of evidence suggest that it may have a tumor suppressor function. The expression patterns of SDPR and SRBC under various growth conditions are very similar (32, 34). The mRNA levels of both SRBC and SDPR are induced on serum starvation and are down-regulated during G₀-G₁ transition, which suggests that they may be involved in cell cycle control. In addition, the induction of SDPR mRNA on serum withdrawal is abolished in transformed cells (34). Interestingly, the human SDPR gene was mapped to chromosome region 2q33, at which a high incidence of LOH has also been observed in lung cancer (44). It also overlaps the homozygous deletion region that has been detected in a SCLC (45). The expression of the rat SRBC was also induced in the livers of animals treated with dithiolethione, a chemopreventive agent, consistent with its anticancer role (46). Rat SRBC binds and is phosphorylated by PKC-δ, which itself is a potential tumor suppressor involved in the regulation of cell growth, differentiation, and apoptosis (47). We isolated hSRBC as a candidate BRCA1-interacting protein in a two-hybrid assay (31). The interaction between BRCA1 and hSRBC is currently under active investigation, but this potential interaction suggests the possibility that hSRBC may act in the BRCA1 pathway. Mutations in the BRCA1 gene are responsible for familial breast cases; more than 80% of families with both breast and ovarian cancers carry germ-line BRCA1 mutations, and the inheritance of a mutant BRCA1 accounts for 45% of families with breast cancer only (48). Mutations in BRCA2 gene are responsible for another 45% of cases of familial breast cancer (49). However, unlike other classic TSGs, no or very few mutations of either the BRCA1 or BRCA2 genes have been found in sporadic breast tumors, which occupy more than 90% of total breast cancer cases (50). Thus, the role of BRCA1 and BRCA2 in the pathogenesis of sporadic breast cancer is not clear. It is possible that lesions (both genetic and epigenetic) occur in other components of the BRCA1 pathway that may similarly compromise BRCA1-mediated tumor suppression functions. BRCA1 has been proposed to be involved in DNA repair processes, particularly double-strand DNA break repair (51) and transcription-coupled DNA repair (52). Thus, it will be of interest to see whether hSRBC also plays a role in transcription-coupled DNA repair. In conclusion, hSRBC is an attractive TSG candidate that undergoes frequent epigenetic inactivation in human lung and breast cancers. Thus, additional studies are indicated to elucidate the molecular details of the function of SRBC and its role in the biology of breast, lung, and other types of cancer.

We thank E. Forgacs, M. Yang, and T. C. Ayi for technical support and C. Muller for providing the 2008 cell line.