Germline CDKN2A mutation implicated in predisposition to multiple myeloma

Following informed consent, cells from a bone marrow (BM) aspiration were acquired from patient No. II:1. Plasma cells were purified from the unsorted BM using the B-B4 monoclonal antibody (mAb) (α-syndecan/CD138; Serotec, Oxford, England) and immunomagnetic beads (DYNABEADS M450 rat antimouse immunoglobulin G₁[IgG₁]; Dynal, Oslo, Norway).14 

We prepared DNA from peripheral blood following standard protocols. We amplified CDKN2A exon 1 sequences from all available members of the family as well as the presort and postsort BM DNA obtained from patient No. II:1 by using the forward primer (p16E1-FF) 5′GAAAGAGGAGGGGCTGGCTGGTC and the reverse primer (p16E1-RR) 5′GCGCTACCTGATTCCAATTCCCCTGC. The reaction products were separated by polyacrylamide gel electrophoresis (PAGE) on an 8% ethidium-stained polyacrylamide gel and photographed under ultraviolet light.

Family A is of northern European descent and comprises 9 members, 4 of whom developed melanoma (including one case of multiple primary disease; Figure 1). In addition to melanoma, one sibling (patient No. II:1) developed multiple myeloma at the age of 53, and a second developed a “brain tumor” (patient No. II:2–it is unknown whether this was a primary tumor or metastatic lesion). We sequenced the genomic DNA of the index case, patient No. II:1, and found that she harbors a germline CDKN2A mutation in exon 1, designated +24 (not shown). This mutation results from a duplication of the first 24 nucleotides of the CDKN2A coding sequence and has been previously associated with melanoma predisposition by multiple investigators worldwide.15 

The +24 mutant and wild type CDKN2A exon 1 polymerase chain reaction (PCR) products can be resolved easily by PAGE. We obtained peripheral blood lymphocyte (PBL) genomic DNA from all available members of family A and used PAGE to analyze each member of the family (Figure 1). Interestingly, every member of this family possessing the +24 mutation has developed cancer; in contrast, the 3 noncarrier family members are cancer free. While the association of the germlineCDKN2A mutation and melanoma in this family was unsurprising, its role in the development of myeloma in patient No. II:1 was unclear. We hypothesized that the +24 mutation had predisposed patient No. II:1 to this disease.

If loss of CDKN2A expression played a role in the development of myeloma in patient No. II:1, the malignant cells should demonstrate loss or dysfunction of the wild type but not the mutantCDKN2A allele. To confirm this prediction, we performed LOH analysis on this patient's plasma cells. Following informed consent, we acquired an aliquot of a BM aspiration (performed as part of the routine clinical care of this patient). Examination of this BM sample revealed that it comprised both normal hematopoietic elements and 30% or less malignant plasma cells. To ensure that nonmalignant constituents of the BM did not interfere with the LOH analysis, we enriched the sample for the malignant clone using magnetic bead selection with a plasma cell-specific B-B4 mAb (α-syndecan/CD138, Serotec), which resulted in a final sample containing more than 90% plasma cells. We isolated genomic DNA from both the sorted and unsorted BM samples as well as from the PBLs from patient No. II:1, amplified exon 1 of CDKN2A from each sample, and resolved the PCR products on a PAGE gel.

Germline DNA amplified from the myeloma patient (No. II:1) demonstrated equal ratios of the wild type and mutant exon 1 alleles ofCDKN2A (Figure 1). Similarly, DNA from unsorted BM cells from this individual appears heterozygous as well because the minority of cells (30% or less) is malignant. However, DNA isolated from sorted BM, which is highly enriched with malignant plasma cells (greater than 90%), comprises mainly the mutant allele, as indicated by excess of the corresponding PCR product. The small amount of wild type PCR product present in this lane arises from residual nonmalignant cells in the sorted BM sample. We conclude that the wild type CDKN2Aallele is lost in the malignant cells and propose that this loss played a role in the development of myeloma in patient No. II:1 of family A.

Our hypothesis that abrogation of CDKN2A function plays a role in myeloma initiation or progression in patient No. II:1 is reinforced by 2 previous observations. First, germline CDKN2A mutations predispose mice to pristane-induced plasmacytoma. Second, loss ofCDKN2A function occurs commonly in human myeloma, albeit at a later stage of the disease.16,17 It is possible that the loss of CDKN2A in this patient may be a nonspecific event that occurred in the malignant cells due to an unstable genome. However, this explanation seems unlikely due to the important role played byCDKN2A in cell cycle control and the almost universal loss of control of the restriction point in the malignancy.

In light of the genetic complexity of both melanoma and myeloma susceptibility, it is not likely that most CDKN2Amutation-bearing individuals are equally predisposed to both diseases. Comparing an analogy between humans and BALB/c mice, which carry at least 4 Pctrl, germline CDKN2A mutations in humans may not be predisposed to myeloma unless other modifier genes that accelerate tumor development are present.

We thank the families and individuals who attended the Toronto Familial Melanoma Clinic and participated in the study. We thank Rakesh Nayer for technical assistance.