Germ Line BAX Alterations Are Infrequent in Li-Fraumeni Syndrome

The Li-Fraumeni syndrome (LFS) defines a rare familial aggregation of cancers (1), which characteristically includes brain tumors, leukemia, osteosarcomas, soft tissue sarcomas, adrenocortical carcinomas, and breast cancer (1, 2). LFS is inherited as an autosomal dominant disorder with high penetrance; 90% of carriers develop cancer by age 70 years (3). Additionally, LFS patients frequently develop multiple primary tumors (4). More than 60% of patients with LFS harbor germ line mutations of the TP53 tumor suppressor gene (5, 6). LFS-like (LFS-L) families have similar cancer histories but do not meet the strict definition of LFS (cancer under age 45 years or the presence of a sarcoma). Typically, the LFS-L families exhibit an incidence of germ line TP53 alteration of only 20% (6). The lack of germ line TP53 mutations in a significant proportion of LFS patients suggests that other genetic alterations may be important in determining their susceptibility to cancer. Recent studies have essentially excluded a significant role in the genesis of cancer in LFS families for CDKN2A, PTEN, BCL10, and TP63 (7-11). The role of CHEK2 in LFS is still not clear, although current work suggests that it is altered only in very rare families (12-14). Deletions, intronic mutations, and promoter region defects of TP53 have rarely been reported (15-17). Furthermore, the clinical description of LFS-L families suggests that variations in phenotype may complicate refinement of the genotype (6). This study proposes that germ line mutations in genes regulated by p53, such as BAX, may lead to the cancers in the LFS and LFS-L families without germ line TP53 mutations. Furthermore, aberrant expression of these proteins might be involved in the transformation of LFS-derived cell lines. BAX (BCL2-associated x protein), a proapoptotic protein that antagonizes the survival function of BCL2, is also involved in regulating mitochondrial membrane permeability. In normal cells, BAX is a cytosolic monomer (18). When activated, BAX integrates into the outer mitochondrial membrane and oligomerizes, allowing the release of cytochrome c, which binds APAF1 and subsequently activates the initiator caspase-9 (19, 20). p53 positively regulates BAX (21). Mutations of BAX have only rarely been reported as somatic alterations in human malignancy (22, 23). A BAX frameshift mutation resulting from a repeat of seven or nine guanine residues instead of the normal eight residues has been identified in 21 of 41 colon adenocarcinomas of the multiple mutator phenotype (23). Two missense mutations, a G67R alteration found in a T-cell acute lymphoblastic leukemia line and a G108V alteration found in a Burkitt lymphoma cell line, have also been reported (24). However, their potential role in the development of cancer in the presence or absence of germ line TP53 mutations in LFS has not been explored previously. We postulated that germ line alteration of the p53 downstream gene BAX could be responsible for the cancer-prone phenotype of LFS families harboring wild-type (WT) TP53.

One fibroblast and 11 lymphoblastoid cell lines derived from LFS patients or unaffected family members were the primary source of RNA. Lymphoblastoid cell lines, as well as primary lymphocytes, were the source of genomic DNA for a total of 37 LFS and 36 LFS-L patients.

DNA was extracted from peripheral blood leukocytes from LFS or LFS-L patients or from cell lines of LFS patients. Single-strand conformational polymorphism was preformed on this DNA using a protocol and primers from Chou et al. (22). Samples that were shifted were gel purified and cloned into a t-tailed pBluescript vector, transfected into competent Escherichia coli, and sequenced using primers specific to the vector and the standard dideoxy termination method as described in the Sequenase version 2.0 DNA sequencing kit (Amersham, Piscataway, NJ).

Allelic frequencies of the BAX polymorphism 14 bp 3′ of exon 3 were calculated from the genotype frequencies determined by single-strand conformational polymorphism. χ² analysis was done to determine if the observed allele frequencies were significantly different from those reported by Centre d'Etude de Polymorphism Humain (CEPH). The significance level of each test was taken to be 0.05.

Thirty-seven LFS and 36 LFS-L individuals were screened for BAX alterations. Among the LFS samples, 17 carry WT TP53 and 20 are mutant. Among the LFS-L samples, 29 have WT TP53 and 7 are mutant (Table 1). Band shifts were detected in exons 1, 5, and 6 in a total of 3, 1, and 12 samples, respectively; however, bidirectional direct sequencing of these samples showed no mutations. Single-strand conformational polymorphism analysis of exons 2 and 3 showed numerous band shifts resulting from a known C/T polymorphism 14 bp 3′ of exon 3. This polymorphism occurs at a frequency of 32% of alleles in 31 CEPH individuals (22). Subcloning and sequencing of the samples and control DNA confirmed the presence of this polymorphism. The known BAX frameshift mutation reported in colon cancers of the microsatellite mutator phenotype was not observed, nor were any other alterations (23).

χ² analysis of the observed allele frequencies of the BAX polymorphism indicated no significant difference between the 37 LFS and the 36 LFS-L samples analyzed (χ² = 0.446 and 0.266, respectively; Table 1A). A χ² of 3.84 is required for significance at 0.05. When segregated into WT TP53 (n = 17) or mutant TP53 (n = 20) groups and compared with the CEPH sample, the LFS samples were not significantly different (χ² = 0.170 and 2.190, respectively; Table 1B). Similar analysis of LFS-L samples segregated into WT TP53 (n = 29) or mutant TP53 (n = 7) groups, and compared with the CEPH sample, LFS-L WT TP53 and mutant TP53 samples were not significantly different (χ² = 0.015 and 2.019, respectively; Table 1C).

To date, several genes have been examined in an attempt to explain the inherited familial cancer predisposition in TP53 WT Li-Fraumeni families. To add to the current body of work, this article examines the status of BAX in individuals with LFS or LFS-L with and without germ line TP53 mutations. This p53 downstream gene is partially responsible for the p53 regulation of apoptosis.

Our results indicate an absence of BAX mutations in LFS or LFS-L samples with either WT or mutant TP53, suggesting that germ line alterations of BAX are not responsible for the cancer-prone phenotype of LFS or LFS-L. It is possible that the observed polymorphism has a role in altering BAX RNA stability or is involved in the alternatively spliced transcript BAXδ, in which exon 3 is excluded from the mRNA and protein product resulting in the removal of the BH3 domain (25). The significance of the BAXδ transcript is unknown, although BH3-only proteins are critical to inducing apoptosis (26). The BH3 domain along with BH1 and BH2 form a hydrophobic pocket that is critical for BAX function (27). BAXδ may lack such a pocket and fail to be activated in response to an apoptotic signal, although this has not been shown to date. To determine if the incidence of the polymorphism correlates with the predisposition of LFS, χ² analysis of the observed frequency of the BAX polymorphism, C/T 14 bp 3′ of exon 3, was done. The results indicate that the frequency of the polymorphism among LFS and LFS-L samples is not significantly different from the frequency of a CEPH control population (Table 1). Based on these results, it seems that the molecular defect responsible for the inherited cancer syndrome in LFS families lacking TP53 mutations is not an alteration of BAX. Similar studies of other genes such as CDKN2A, PTEN, BCL10, CHK2, and TP63 have generally failed to explain the familial cancer predisposition in these families (7-14).

Several novel downstream target genes of p53 including NOXA, p53AIP1, and PIDD have been identified, but it is not yet known if these proteins are involved in oncogenesis. Perhaps proteins that functionally interact with the p53 protein such as p33 and JMY will provide new and interesting candidates to explain these cancers. Proteins upstream of p53 such as ATM, ATR, and Chk2 phosphorylate p53, regulating its function. Recently, germ line alterations of CHEK2 have been implicated in one LFS TP53 WT family (12); however, these findings have not been further substantiated in families meeting the strict definition of LFS (13, 14). Additionally, PIN1, an enzyme capable of altering protein conformation, has recently been shown to activate p53 in response to genotoxic agents and may represent a reasonable target to account for poor p53 function. Perhaps there are several genes each responsible for a handful of cancer families, with TP53 responsible for the largest proportion. Such a finding would be confirmation of the hypothesis that newly discovered tumor suppressors are unlikely to be as widely involved in cancer as TP53. Rather, they are likely to occur infrequently and in special cases where the gene is occasionally involved as a tumor suppressor in a cell type–specific or tissue-specific fashion (28).

We thank Dr. Linda Penn (Ontario Cancer Institute, Toronto, Ontario, Canada) for providing Rat1 cells with and without a human BCL2 transgene, Gillian Bromfield for providing a p21^WAF1/CIP1 positive control, and Susan Quesnel for technical assistance with cell lines and Western blotting.