|Home | About | Journals | Submit | Contact Us | Français|
Chromosome 13 deletions (del), detected by metaphase cytogenetics, predict poor outcome in multiple myeloma (MM), but the gene(s) responsible have not been conclusively identified. We sought to identify tumor suppressor genes on chromosome 13 using a novel array comparative genomic hybridization (aCGH) strategy.
We identified DNA copy number losses on chromosome 13 using genomic DNA isolated from CD138 enriched bone marrow cells (tumor) from twenty patients with MM, monoclonal gammopathy of undetermined significance (MGUS) or amyloidosis. We used matched skin biopsy (germline) genomic DNA to control for copy number polymorphisms and a novel aCGH array dedicated to chromosome 13 to map somatic DNA gains and losses at ultra-high resolution (>385,000 probes; median probe spacing 60bp). We analyzed microarray expression data from an additional 262 patient samples both with and without del.
Two distinct minimally deleted regions at 13q14 and 13q13 were defined that affected the RB1 and NBEA genes, respectively. RB1 is a canonical tumor suppressor previously implicated in MM. NBEA is implicated in membrane trafficking in neurons, PKA binding, and has no known role in cancer. Non-coding RNAs on chromosome 13 were not affected by interstitial deletions. Both the RB1 and NBEA genes were deleted in 40% of cases (8/20; 5 patients with del detected by traditional methods and three patients with interstitial deletions detected by aCGH). Forty-one additional MM patient samples were used for complete exonic sequencing of RB1, but no somatic mutations were found. Along with RB1, NBEA gene expression was significantly reduced in cases with del.
The NBEA gene at 13q13, and its expression, are frequently disrupted in MM. Additional studies are warranted to evaluate the role of NBEA as a novel candidate tumor suppressor gene.
Numeric or structural chromosomal abnormalities are detected in nearly all patients with plasma cell dyscrasias, including primary amyloidosis, monoclonal gammopathy of undetermined significance (MGUS) and multiple myeloma (MM) . Chromosome 13 deletions, most frequently monosomy 13 (del) are detected in 10–20% of MM cases by routine cytogenetics or metaphase fluorescent in situ hybridization (FISH) and are a significant predictor of shortened survival ,, and  Previous efforts to map somatically acquired DNA copy number losses on chromosome 13 in MM have identified 13q14-q21 and 13q34 as commonly affected regions , , , , , [10,] and . These studies used 10Mb resolution CGH and 1Mb resolution FISH techniques. Whole genome array CGH (0.73Mb resolution) combined with gene expression analysis identified CUL4A (13q34) as a potentially relevant gene located within a 0.77Mb deletion on chromosome 13  and . Single nucleotide polymorphism analysis (10Kb resolution) revealed a 1.9Mb minimally deleted region (MDR) spanning 13q13.3 to q21.3 . The relatively low resolution of these studies has precluded consistent identification of biologically relevant genes targeted by del. Additional studies using higher resolution techniques therefore, are needed.
The widespread contribution of DNA copy number polymorphisms to variability among human genomes has recently been appreciated  and . In previous CGH studies that have used pooled genomic DNA as controls, false positive identification of DNA gains and losses due to DNA copy number polymorphisms could not be excluded. Here, we definitively identified somatic changes by using genomic DNA isolated from matched skin biopsy specimens from our patients to control for DNA copy number polymorphisms. Furthermore, we used a novel, chromosome 13-dedicated CGH array to identify genes affected by DNA copy number loss with unprecedented resolution. Our novel approach allowed the mapping of extremely small deletions. We found a 3.49Kb MDR in 13q14 that mapped to exon 20 of RB1 encoding the highly conserved ‘pocket domain’ responsible for binding E2F transcription factors , , and . A second MDR at 13q13 was definitively mapped to NBEA, a gene encoding a PKA binding site and predicted to regulate membrane transport in neurons with a role in cancer that has yet to be defined.
Bone marrow and skin biopsy samples were obtained from patients with plasma cell dyscrasias following informed consent. The study was approved by the Washington University Institutional Review Board and the Siteman Cancer Center Patient Review Monitoring Committee. Clinical data including routine cytogenetics and M-FISH were obtained anonymously using Unique Patient Numbers (UPN). I-FISH was performed on marrow sections using probes: LSI 13/RB1, and CEP7 (Vysis Inc., Downers Grove, IL, USA). For each hybridization, a minimum of 100 non-overlapping nuclei were analyzed. For molecular analyses, CD138+ bone marrow mononuclear cells were isolated by Ficoll gradient (Stem Cell Technologies, Vancouver, BC) followed by separation using CD138 microbeads and an AutoMACS Cell Separator (Miltenyi Biotec, Aubern, CA). Fluorescence activated cell-sorting analysis using a PE-CD138 antibody (Miltenyi Biotec, Aubern, CA) confirmed >97% purity. Genomic DNA was isolated with Qiagen Miniprep Kits (Valencia, CA).
The first 20 patient sampes with ≥500,000 CD138+ cells were selected for aCGH, which required 1.5μg DNA. CD138+ (tumor) DNA was labeled with Cy3 and skin (germline) DNA was labeled with Cy5. The custom array contained 385,272 oligonucleotide probes. Nimblegen built the array, performed probe design, and sample hybridization to the custom array (www.Nimblegen.com). Sequence source for the probe design was HG17/UCSC (http://genome.ucsc.edu/).
Circular Binary Segment (CBS) Analysis: CBS  was performed using Signal Map Software (Nimblegen, Madison, WI). Data was analyzed using a non-overlapping window, which averaged the signal intensity from each probe over a 600bp region. Since probes were spaced approximately every 60bp apart, each window averaged 10 probes. This approach was used to condense the data and provided clean segment breaks. Systematic criteria set to eliminate false positives included: ≥3 data points involved (representing ~30 probes; 180bp) and log2 ratio <−0.25. Magnified plots were generated with Graphpad/Prism 4, Version 4.02 (Graphpad Software Inc, San Diego, CA).
Process Control analysis was performed on unaveraged data set (no windows used to condense data). Data was normalized using qspline . To eliminate outliers, the raw data for the skin (reference) samples from each patient was analyzed. Probes with signal intensity >3X SD above the mean were discarded (range: 8,000–17,000, averaged 10,000 per patient (2–4.4% of total)). Process control employs techniques using a Shewhart control chart , a graphical and analytical tool used in industry for quality control purposes. It is applied to aCGH analysis to determine which probe intensities are different enough from mean variability to be considered meaningful. Probe intensity ratios were considered “significant” if they satisfied: eight probes in a row on one side of the overall mean. They also had to pass either A) two of three probes in a row beyond two units of overall SD, or B) four of five points in a row beyond one unit SD [23,24]. If an eight-probe region (representing ~480bp) passed the criteria, the first and last of the eight probes were mapped. Genes were mapped by aligning probes of interest to the Human Build 36.2 genome. Whole chromosome plots using this same data set were generated using the program [R] (Figure 1A).
High-throughput sequence analysis of RB1 was performed by the Genome Sequencing Center at Washington University (WUSTL) as described . Detailed protocols are available on WUSM GSC website (http://genome.wustl.edu/platforms.cgi?id=7).
PCR validation of RB1 micro-deletion was performed on genomic DNA isolated from CD138+ selected bone marrow (tumor) and skin biopsy (germline) patient samples. The independent control DNA was kindly provided by Rhonda Reis, Division of Oncology, WUSTL. Primers: RBValFWD3: CCATTGCCCACAGTCAGAAA RBValREV3:GGTAGGGGAATAGGGGGTGA. Products were cloned into and sequenced from TOPO2.1 vector (Invitrogen, Carlsbad, CA).
Assays were performed on original patient genomic DNA with Taqman Universal PCR Master Mix. Primer concentration: 900nM; probe concentration: 2.5mM, 10ng template. Reactions were run on 7300 Real Time PCR System, and analyzed using 7300 System Software (Applied Biosystems, Foster City, CA). RBRTFwd1:5′GAATTAGAACATATCATCTGGACCCTTT3′
The ΔΔCt method was used since equal efficiency of primer/probes was shown.
Two datasets were analyzed. The Mayo Clinic dataset included 162 samples (101 MM, 24 SMM, 22 MGUS, and 15 normal PC’s; GEO GSE6477). Chromosome 13 status was determined by FISH. The MMRC dataset (http://www.themmrc.org) included 100 samples. Chromosome 13 status was determined by aCGH. Expression values were derived against a PM/MM difference background using Robust Multichip Average (RMA, ). Present/Absent probes were called using Affymetrix Microarray Suite version 5. Only probes detected in at least one sample were used in subsequent comparisons. In pooled Chromosome 13 Deletion vs. No Deletion comparisons, Significance Analysis of Microarrays (SAM, ) was used to detect differentially expressed genes based on a q-value of less than 5%. SAM was run with 100 permutations for correction of False Discovery Rate. These genes were clustered and visualized in DChip  (http://www.dchip.org). aCGH data was first smoothed with region=2, outlier scale =4, smoothing SD=2 and trimming proportion of 0.025. CBS was then run with default parameters (alpha=0.01, window.size= NULL, with 10000 permutations).
RNA was isolated using Trizol and cDNA generated using SuperScript First Strand Synthesis Kit (Invitrogen, Carlsbad, CA) per manufacturer’s directions.
hGAPDHFwd:GAAGGTGAAGGTCGGAGTC hGAPDHRev:GAAGATGGTGATGGGATTTC GAPDHProbe:5′56FAMGGCTGAGAACGGGAAGCTTGTAMSp3′ Human brain sample provided by Bob Schmidt, Pathology, WUSTL. Samples were run in triplicate and performed twice. Error bars are SD of two experiments.
LP-1, KMS-11, OPM-2 and UTMC2 lines were provided by W. Michael Kuehl, Genetics Branch, NIH) and maintained in RPMI with 1% Penicillin/Streptomycin (both Cambrex Bioscience, Walkersville, MD), 10% fetal bovine serum (HyClone, Logan, Utah). RPMI-8226, U266, and H929 cells were obtained from and grown per ATCC recommendations. Lysates were prepared as described . Antibodies: total RB1: IF8 (Santa Cruz Biotechnologies, Santa Cruz, CA); anti-phospho-RB1(Serine Ser807/811; Cell Signaling, Beverly, MA); Actin (Sigma, St. Louis, MO). NBEA antibody kindly provided my Manfred Kilimann, Institute for Anatomy, Bochum, Germany and used as described .
Deletions affecting chromosome 13 occur at similar frequencies in a variety of plasma cell dyscrasias , so patients with the diagnosis of MM, MGUS, or amyloidosis, regardless of chromosome 13 status, were selected for array comparative genomic hybridization (aCGH). Twenty patients were selected solely on the basis bone marrow plasma cell yield after CD138 enrichment (Table 1 and S1). We excluded low-yield samples to avoid the need for whole genome amplification (WGA), which can introduce bias or mutation (personal communication Matthew Walter, 2008). Genomic DNA was isolated from CD138-enriched bone marrow samples (tumor) as well as from patient-matched skin biopsy samples (germline) controls. Patient-matched skin biopsy samples were an important internal control for copy number polymorphisms known to occur in healthy populations  and . To identify DNA copy number alterations across chromosome 13 with the greatest possible resolution, we performed comparative genomic hybridization using a custom CGH array (Nimblegen Inc, Madison, WI) dedicated to chromosome 13. The custom array had 385,272 probes spanning the entire length of chromosome 13 with median probe spacing of 60 base pairs.
Array CGH data were plotted linearly along chromosome 13 using log2 tumor: germline signal intensity ratios (Figure 1). By eye, some regions of copy number change were obvious, but to systematically identify regions of DNA copy number loss across chromosome 13, we performed two independent, unsupervised analyses of the data. To facilitate identification of minimally deleted regions (MDRs) across patient samples, we first used a CBS algorithm using stringent criteria to identify interstitial deletions . By CBS analysis, eight of the 20 patient samples (40%) harbored at least one region of interstitial DNA copy number loss with a mean deletion size of 596Kb (range: 1.2Kb to 16Mb, Table 2). Among the eight patients with DNA copy number loss, the mean number of deletions was five (range: 1 to 13). The finding of a greater number of chromosome 13 deletions than previously reported using lower resolution techniques  suggested that our strategy could be useful for finding novel regions on chromosome 13 contributing to plasma cell diseases.
We compared chromosome 13 status determined by aCGH to analyses of chromosome 13 using standard techniques including metaphase cytogenetics, metaphase fluorescent in situ hybridization (M-FISH) and interphase fluorescent in situ hybridization (I-FISH; Figure 1). Because the aCGH raw data were normalized to balance fluorochrome intensity, monosomy 13 (i.e. non-interstitial deletions) was undetectable via aCGH analysis. We were therefore forced to rely on clinical cytogenetic data for detection on monosomy 13 (Table 1, Figure 1). By cytogenetics, five of the 20 patient samples (25%) had monosomy 13 (Table 1, Figure 1). Additionally, two patients with monosomy 13 (22848 and 92896) also had aCGH-detected DNA copy number losses suggesting homozygous deletion at those loci (Table 2). A 1200bp deletion in patient sample 22848 did not map to known genes or micro RNA at 13q31, while patient sample 92896 harbored two deletions affecting KATNAL1 and DNAJC3 genes.
Notably, cytogenetic and FISH analysis failed to detect chromosome 13 DNA copy number loss in 25% of cases (5/20) that were positive by aCGH (Figure 1). This data demonstrates that high-resolution array CGH has the ability to detect chromosome 13 deletions undetected by standard FISH and cytogenetics. This result also highlights the utility of unbiased analysis of the entire chromosome to identify novel regions on chromosome 13 whose copy number changes could direct the study of genes relevant to MGUS and MM pathogenesis.
To identify chromosome 13 genes whose loss could contribute to MM pathogenesis, we mapped all known genes that fell within the regions of copy number loss identified by CBS analysis. We found 28 of 43 (65%) deleted segments mapped to at least one gene, rather than non-coding DNA (Table 2). None of the regions identified in our study mapped to known non-coding RNA loci. Specifically, the two micro RNA clusters on chromosome 13 known to contribute to CLL (miR 15–16 at 13q14 and MiR 17–92 at 13q31.3) were not affected in our samples.
To independently identify genes with copy number loss on chromosome 13, we performed a separate, unsupervised analysis of the data by using an independent Process Control algorithm  shown to reliably call aCGH probe signals that deviate significantly from baseline (Figure S1). This second analysis identified 216 probes that mapped to 42 genes (Table S2). Twenty of the 42 genes (48%) identified by Process Control were also identified by the CBS analysis, underscoring the robustness of the aCGH data set (Table 2).
The region most affected in our patient group encompassed 13q12 to 13q14.3, (25 to 50Mb), Figure 1). CBS analysis of the log2 plots from three of five patient samples with interstitial deletions (58762, 64511, and 95295) revealed two distinct MDRs within 13q12-14.3 (Figure 1 shaded bars, Figure 2–3). Patient sample 95295 harbored two DNA copy number losses that were extremely small (106Kb and 1200bp, respectively) and defined the MDRs at 13q14.2 and 13q13 (Figure 1–3, Table 2). Within 13q14.2, only the RB1 gene was affected in all three patient samples (Figure 1, ,2).2). Strikingly, the 13q14.2 MDR mapped to exon 20 of RB1, encoding the ‘pocket domain’ of RB1 critical to its tumor suppressor function .
Inspection of the log2 plots from the same three patient samples revealed each harbored a second and distinct interstitial deletion at 13q13, 13Mb centromeric to the RB1 locus (Figure 1, ,3).3). This second 13q13 MDR mapped to a single gene not previously implicated in myeloma biology: neurobeachin (NBEA, BCL8B, Figure 3). Every patient sample in our set that harbored a deletion affecting RB1 (three with interstitial deletions and five with monosomy 13) simultaneously harbored copy number losses affecting the novel myeloma associated gene NBEA (Table 2, Figure 1–3).
Since the segment of DNA copy number decrease within RB1 in patient 95295 was small (3.49Kb) and contributed significantly to the mapping of the 13q14 MDR, we first performed PCR spanning the microdeletion on the same tumor and skin genomic DNA used in the aCGH analysis (Figure 2). Amplification of a truncated band and sequence analysis of the PCR product confirmed this deletion tumor-associated (Figure 2). This result confirms the micro-deletion affecting RB1 in patient sample 95295.
To quantify and confirm the DNA copy loss across all three patient samples with interstitial RB1 deletions (58762, 64511 and 95295), real time PCR was performed on CD138 purified tumor genomic DNA (Figure 2). Consistent with the qualitative PCR, patient sample 95295 had virtually no signal using a primer-probe set at this locus (fold copy number: 0.02). Patient samples 58762 and 64511 had a fold copy number of 0.86 and 0.62, respectively, consistent with loss of one copy of RB1. These results are concordant with the aCGH log2 ratios for this region (average log2 ratio of probes that span microdeletion: 95295: −0.977; 58762: −0.518; 64511: −0.754).
A similar analysis was used to quantify and confirm the interstitial NBEA deletions in patient samples 58762, 64511, and 95295. Consistent with the aCGH data, patient sample 95295 revealed homozygous deletion (fold copy number: 0.14). Patient samples 58762 and 64511 revealed heterozygous loss of NBEA (fold copy number: 1.22 and 0.9, respectively; Figure 3). These data confirm non-contiguous interstitial deletions on chromosome 13, affecting simultaneously the NBEA and RB1 genes in three of 20 patient samples (15%).
Our data suggested RB1 is a target of deletions in MM, yet in most patient samples (7/8 in our set) the other copy is retained. Limited sequence analysis in myeloma failed to show mutations in RB1 exons 20–24  (mutation hotspots in retinoblastoma) , but other domains of RB1 have not been re-sequenced in MM. We performed sequencing of all 27 RB1 exons and surrounding intronic sequences in 41 MM/MGUS patient samples (including 16 of our 20 patient set; Table S3). We found no non-synonymous sequence changes affecting the coding or promoter sequences (bp −474 to −182) of RB1, suggesting that, in contrast to retinoblastoma tumors, most myeloma tumors retain at least one wild-type RB1 allele.
We detected eleven intronic SNPs (Table S3). Since myeloma is twice as prevalent in African American populations compared to Caucasians (www.seer.cancer.gov), race matched minor allele frequencies (MAFs) from our patient samples were compared to published MAFs in the Hap Map database for the nine of eleven SNPs with available data (Table S4). Two RB1 SNPs (rs198580 and rs198617) were significantly more common in our Caucasian patients (P < 0.001 and P < 0.018, respectively), suggesting a possible role in MM pathogenesis. No significant differences were found between subgroups for the other seven SNPs.
We identified six novel SNPs, not reported in the NCBI or HapMap databases (Table S5). These were validated by repeat sequencing and were detected in both tumor and matched skin genomic DNA, demonstrating these to be germline sequence variants. Since identified SNPs were located near exon boundaries (range: 10–171bp from boundary), we considered the possibility that RB1 SNPs might play a role in MM pathogenesis by affecting RNA splicing. Examination of RB1 cDNA isolated from seven patients with reported or novel SNPs revealed only RB1 transcripts of expected size (not shown). Together, our re-sequencing analysis demonstrated no somatic mutations in retained RB1 alleles.
To determine whether retained RB1 alleles were expressed, we performed Western Blot analysis on a panel of MM cell lines. RB1 protein was detected in all cell lines that retained at least one RB1 allele. U266, shown to have undergone rare biallelic loss of RB1 , expressed no RB1 protein as expected, as did UTMC2 cells (Figure 2). LP-1 and KMS-11 cells, which retain one copy of chromosome 13 , expressed lower levels of RB1 protein than OPM-2 and RPMI-8226 cells, which retain two copies of RB1 . These data suggest RB1 protein levels are related to RB1 genomic copy number.
Given the lack of RB1 mutations identified in this study, we hypothesized RB1 protein would be inactivated by phosphorylation in MM. All MM lines that retained at least one copy of RB1, expressed phosphorylated (Ser807/811) RB1 protein (Figure 2) consistent with a previous analysis  and  whose levels also appeared to correlate with RB1 copy number. These data show retained RB1 alleles are expressed and raise the possibility that RB1 haploinsufficiency contributes to MM/MGUS pathogenesis.
We sought to validate NBEA as a deletion target by characterizing its expression in MM cells. We anticipated that patient samples with del would have lower NBEA expression than patient samples with without del. We analyzed two large microarray data sets (Mayo GSE 6477) and MMRC (http://www.themmrc.org; total n=262) for expression changes based on chromosome 13 status. In both datasets, NBEA transcript levels were significantly decreased in patient samples with del (Figure 4, S2; Table 3, S6, S7).
We developed a quantitative real-time reverse transcriptase PCR (Q-RT-PCR) assay to validate NBEA transcript expression, and assayed a panel of MM cell lines (MMCL). Some MMCL expressed low levels of NBEA transcript, as anticipated, but surprisingly, several MMCL expressed NBEA at high levels (Figure 3C). We found UTMC2 cells expressed NBEA at levels three times than higher than in a human brain sample, where NBEA is normally most highly expressed , , and . OPM2 cells had levels 30% of brain while U266 had levels 18% of brain. RPMI-8226 and LP1 had low/undetectable transcripts (Figure 3C). We examined NBEA protein levels in these MMCL by Western Blotting of whole cell lysates. Consistent with the Q-RT-PCR data, we found NBEA protein expression varied significantly between cell lines (Figure 3D). The UTMC2, OPM2, and H929 cell lines had the highest NBEA protein levels, while RPMI 8226, U266 and LP1 had low to undetectable NBEA protein.
Finally, we measured NBEA transcripts and protein levels in a set of CD138-enriched primary MM bone marrow samples (n=14) using Q RT PCR and Western blotting. We found NBEA transcript expression varied significantly across samples and, consistent with our MMCL data, some MM patients, even with del, harbored high NBEA transcript levels (Figure 3E). Because of the large number of CD138 cells needed for Western analysis we were forced to analyze a separate cohort of MM patient samples by Western blot (only sample 14216 had both RNA and protein data, and expression was low by both analyses). Consistent with the RNA data, Western blotting using NBEA-specific antibodies demonstrated that NBEA protein was strikingly dysregulated in patient MM cells (Figure 3F).
We used a novel ultra high-resolution aCGH strategy to map somatic chromosome 13 deletions with unprecedented resolution in 20 patients with MM, MGUS or amyloidosis. We used a custom CGH array dedicated solely to chromosome 13 (60bp median probe spacing) and genomic DNA from patient-matched germline (skin) biopsy samples as controls to eliminate signal noise due to DNA copy number polymorphisms  and . We avoided noise introduced by whole genome amplification strategies by using non-amplified genomic DNA from cases with high yields of CD138+ bone marrow mononuclear cells. Patients with low bone marrow tumor burden may therefore have been underrepresented in this study. However, analysis by standard techniques of FISH and cytogenetics detected chromosome 13 deletions at expected frequencies  and  (Table 1) suggesting our patients were generally representative of other MM, MGUS and amyloidosis cohorts. Although our detection of del by I-FISH was lower than other reports, our analysis was performed on non-enriched paraffin embedded bone marrow samples.
We found two regions of recurrent DNA copy number loss that were non-overlapping and mapped to two genes: RB1, the canonical tumor suppressor at 13q14.2, and NBEA at 13q13, a gene whose role in cancer is less clear. Two independent, unsupervised analyses (CBS and Process Control) generated gene lists affected in our patient set that largely overlapped (Table 2 and S2) demonstrating the high quality of our aCGH data. Both lists included the RB1 and NBEA genes. Visual inspection of log2 plots at these loci in high-resolution and PCR confirmed the identification of bona fide deletion events (Figure 2–3 and data not shown). Extremely small deletions (3.49Kb and 106Kb) in a single patient (95295) significantly narrowed the MDRs we identified. In sample 95295, NBEA and RB1 were the only two genes on chromosome 13 affected by DNA copy number loss. Re-sequencing analysis of the RB1 gene within this patient sample revealed only homozygous SNPs (Table S4) demonstrating isodisomy/gene conversion across all or part of chromosome 13. These data strongly suggest that chromosome 13 DNA copy number decreases in this patient (i.e. RB1 and/or NBEA loci) were selected for during disease development and likely contribute to MM biology.
Homozygous deletions of RB1 are rare in MM , and we considered the possibility that 95295 might be an outlier. This patient harbored the t(4;14) translocation, and had rapidly progressive disease characterized by treatment resistance (not shown). If the 95295 sample is removed from our analysis, however, our conclusions remain substantially unchanged. Two distinct MDRs are still defined by the remaining interstitial deletions and identify a small number of candidate genes. At 13q13, NBEA remains the sole gene affected. Without 95295, the MDR at the gene-rich 13q14.2 locus expands to include: SUCLA2, NUDT15, MED4, ITM2B, RB1, P2RY5 and RCBTB2. Unlike in retinoblastoma, the retained RB1 allele is not affected by mutations in MM, so we cannot formally exclude the possibility that additional 13q14.2 genes contribute to myeloma biology.
Our expression analysis identified fourteen chromosome 13 genes with decreased expression in patient samples harboring monosomy 13, including RB1. Detection of decreased RB1 levels in del samples has been inconsistent in prior studies , [13,  and . We found decreased levels in del samples consistent with prior reports  and . The inconsistencies are probably due to variability of RB1 expression, which we have also observed (data not shown) or differences in criteria used in gene expression analyses. We also identified Ring finger 6 (RNF6) previously shown to have tumor suppressor function  and Integral membrane protein 2B (ITM2B) identified to undergo single copy number loss and promoter methylation in bladder cancer , both identified in a prior report to be decreased in del . Although a gene of interest due to prior reports of MM involvement, we did not identify CUL4A in the aCGH or expression analysis  and . This does not exclude a potential role for CUL4A in MM biology, as MM is a disease with diverse clinical presentations and complex genetics. Further analysis is required to determine the role of these genes, if any, to myeloma biology.
Our genetic data suggest that NBEA may be a novel plasma cell dyscrasia tumor suppressor gene. NBEA shares sequence homology to the Drosophila melanogaster protein kinase A anchoring protein rugose and the mammalian gene LRBA. NBEA encodes a PKA binding domain and is a predicted PKA anchoring protein (AKAP)  and . The C-terminal region of NBEA encodes a BEACH domain that is situated next to a WD40 domain, suggesting a role in protein-protein interactions. Crystal structure analysis revealed a structurally conserved PH domain intimately associated with the BEACH domain of NBEA . NBEA is a large gene (730Kb) encompassing the FRA13A fragile site . The breakpoints we observed were centromeric to the most fragile FRA13A breakpoint region in NBEA, suggesting that the NBEA deletion events we observed were not “bystander mutations.”
A functional role for NBEA was suggested by the dysregulation of its expression in MM patient samples. DNA microarray data from a large number of patient samples (n=262) demonstrated that compared to patients with normal chromosome 13, there was a decrease in NBEA expression in delpatients. These data suggest that NBEA may be a novel tumor suppressor gene in MM. On the other hand, our Q-RT-PCR and Western blot analysis performed on a separate cohort of MM samples revealed that some patients, even with del, harbored very high NBEA expression (Figure 3). A prior study showed increased NBEA expression with advanced disease stage in primary plasma cell dyscrasias . Furthermore, NBEA was one of a small set of genes whose expression was increased in del patients in another study (supplemental data of ). Inactivating mutations in the p53 tumor suppressor gene is often associated with high p53 expression  so these data may all be consistent with a role for NBEA as a tumor suppressor. Sequencing of NBEA genes in MM patient samples will be required to confirm this hypothesis.
NBEA shares 62% sequence identity at the amino acid level to LRBA, a homolog also implicated in cancer cell growth . Intriguingly, knock-down of LRBA in cancer cell lines decreased the growth of cells in culture, and these authors proposed that LRBA functions as an oncogene by facilitating EGFR . Additional studies are required to elucidate the contribution of NBEA to MM disease biology.
Figure S1. Significant chromosome 13 probes representing DNA copy number changes identified by Process Control.
A. Probes representing deleted and amplified regions in any of the 20 patient samples identified by Process Control Analysis. Data are plotted in a linear fashion across chromosome 13 beginning with the p arm to the left and extending throughout the long q arm to the right. 8844 probes marked regions of DNA copy number decrease and 6706 probes marked regions of DNA copy number increase. Arrowhead indicates region of copy number loss in 13q14. B. Process Control probes as in A detected in two or more patients. In two or more patient samples, 216 probes indicated DNA copy loss and 274 probes indicated DNA copy number gains. We found 69 of the 216 probes with DNA copy number decrease mapped to 42 genes (Table S2, Table 2).
Figure S2. Gene expression profile of patient samples with or without chromosome 13 deletion
Microarray analysis revealed genes differentially expressed in patient samples with or without monosomy 13. MMRC dataset is shown.
The authors thank the patients who donated samples. We thank Patrick Cahan, John DiPersio, Julie Fortier, Megan Janke, Tim Ley, Luke Starnes, Matt Walter, Zhifu Xiang and Katherine Weilbaecher for critical review of the manuscript. We thank AnnaLynn Molitoris and the staff of the Siteman Cancer Center Tissue Procurement Core for technical assistance. We thank Rafael Fonseca and Kenneth Anderson for helpful discussions. This work was supported by the National Institute of Health grant: CA116168 and the Washington University Division of Oncology.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.