Search

Background: Novel drug discoveries have shifted the treatment paradigms of most hematological malignancies including multiple myeloma (MM), but minimal residual disease and drug resistance underlie relapses in MM. Although many genetic and epigenetic alterations regulate MM progression, MM cells are not autonomous. Dynamic interactions between MM cells and cells of the bone marrow (BM) microenvironment have been reported by our group and others. MM plasma cells (PCs) depend on interactions with bone marrow stromal cells (BMSCs) for their survival and growth, but little is known about the specific genetic events taking place in the MM BM microenvironment. Methods: Here we report a detailed analysis of the genetic and epigenetic events that are characteristic of MM BMSC as compared to HD-BMSC interacting with BM PCs. To evaluate genetic and epigenetic landscapes, RNA was extracted from bulk sorted populations of 16 MM-BMSC, 3 HD-BMSC, and 10 autologous MM cells. We prepared libraries for 32 samples using the NEBNext Ultra II Stranded Poly A kit, and then sequenced on the NextSeq 500, PE150. Sequencing data were analyzed using a custom computational and statistical pipeline at the Department of Biostatistics, School of Public Health and Partek software. Results: Unsupervised clustering showed that MM-BMSC samples clustered as a distinct and completely separate cluster from HD-BMSC and autologous MM cells. Gene level analyses of these three groups identified 990 genes differentially expressed (upregulated or downregulated, P< 0.005). Sequential filtration analyses of the differentially expressed genes in MM-BMSC identified significant deregulation of : transcripts in the Jak-STAT signaling pathway (JAK3, PIM1, IL6, CSF2R, AKT1/2, BCL2L1, CDKN1A and range of IL transcripts); genes encoding extracellular matrix interacting proteins (CD36, CD49, LAMA3, CD44, CD47); and various plasma membrane proteins that define different subpopulations of hematopoietic cells. These genes were deregulated in >24% of MM-BMSC samples analyzed as compared to HD-BMSC samples. These transcripts were downregulated in autologous tumor cells. Next, we interrogated the epigenomic landscape and identified the splicing signature of MM-BMSC as compared to HD-BMSC, and autologous MM cells. Comparison of the splicing patterns (exon skipping, intron retention, novel splice acceptor and/or donor activation) of these three distinct groups showed that a total of 2,100 genes were differentially expressed and 566 were alternatively spliced among the three groups (P < 0.01). These analyses identified a limited number of the transcripts with ~3% significantly spliced in MM-BMSC compared to HD-BMSC. However, comparing MM-BMSC splicing events to MM cells splicing events, we identified >30% of genes which were alternatively spliced in MM cells but not in MM-BMSC. Further, gene enrichment and pathway analyses identified a selective set of transcripts that were alternatively and differentially spliced in MM-BMSC including genes involved in MAPK and Ras signaling pathways, homologous recombination, mismatch repair, and adherens junction. Conclusions: Taken together, our studies identified marked differences between important stromal elements in MM- and HD-BM. We identified genes that were specifically upregulated/suppressed in MM-BMSC compared to MM-cells and HD-BMSC. Within MM BMSCs, we identified several splicing events on genes of signaling pathways implicated in development and progression of MM. Furthermore, altered splicing events identified on these transcripts represent potential new immunotherapeutic targets. Disclosures Chu: Gilead: Honoraria; Celgene: Honoraria; Teva: Consultancy; Amgen Inc.: Honoraria; AstraZeneca: Honoraria. Anderson:C4 Therapeutics: Other: Scientific founder ; Gilead Sciences: Other: Advisory Board; OncoPep: Other: Scientific founder ; Sanofi-Aventis: Other: Advisory Board; Janssen: Other: Advisory Board.

Background: Receptor for hyaluronan-mediated motility (RHAMM) or CD168 has been a promising target for MM immunotherapy because it is overexpressed in MM cells. RHAMM has been tested as a target for an anti-RHAMM peptide vaccination approach in MM and other hematological malignancies. Although RHAMM peptide-induced immune response in patients, clinical outcomes were mixed, that can be a result of equal expression levels of RHAMM in different subpopulation of bone marrow (BM) cells in MM patients and healthy donors (HD). To enhance current RHAMM-peptide and future immunotherapeutic approaches, we investigated the cause of altered RHAMM mRNA splicing in MM patients. mRNA splicing has the potential to produce numerous mis-spliced genes, creating novel disease markers; some resulting proteins are likely to contain neoantigens selectively expressed on MM tumor cells. Methods/results: Splicing alterations can be caused by single nucleotide variations (SNVs) that affect splicing regulatory elements (SREs), or by deregulated splicing factors (SFs). We evaluated the incidence of SNVs located in the vicinity of the RHAMM. We identified a total of 57 SNVs: 72% SNVs are in the intronic region, and 28% are in the RHAMM coding region. We used the "HEXplorer" tool and predicted that four SNVs have the potential to contribute to aberrant RHAMM splicing in MM either by altering SF binding to SREs or by impacting splice site selection. Predicted SNVs were evaluated using an in vivo splicing assay to identify SNV-clusters causing aberrant RHAMM splicing. We have observed progressive overexpression of core SF PTBP1/2 (polypyrimidine track-binding protein) in MM patients and associated with disease progression. Since SNVs on the RHAMM modulate canonical SF binding sites, we tested the effects of PTBP1/2 deregulation on RHAMM splicing. We expressed PTBP1/2 in H929 cells, and then evaluated the RHAMM splicing pattern in transfected cells at a single cell (SC) level. SC analyses showed that overexpression of PTBP1/2 increased (2.5-fold) the RHAMM-V3:FL ratio in MM cells. SC analyses also identified overexpression of the RHAMM-V3 splice variant in 18% of H929 SCs expressing PTBP1, and in 37% of cells expressing PTBP2, confirmed at the single cell (SC) level. In BM-infiltrating myeloid cells, analyses showed 50% of myeloid cells express the RHAMM-V3 variant alone, and 79% of plasma cells (PCs) express this variant in combination with RHAMM-FL. Moreover, the RHAMM-V3/FL ratio in PCs is elevated (2.6-fold), further confirming a correlation between the RHAMM variant ratio and the clinical outcome. Next, we determined RHAMM-V3/FL ratios in BM stromal cells from 16 MM patients: MM-BMSC samples exclusively express the RHAMM-V3 in combination with RHAMM-FL and the RHAMM-V3:FL is 1.8 fold. In BMSC samples derived from healthy donors (HD), we detected relatively low-level expression of RHAMM-FL as compared to expression levels of RHAMM-FL in MM patients, while RHAMM-V3 transcripts were undetectable. SC analysis of RHAMM FL and splice variant transcripts in MM BMSC and HD-BMSC agreed with the analyses done on the MM HD-BMSC bulk population. We did not detect any MM BMSC cells expressing RHAMM-V3 alone and the RHAMM V3/FL ratio was 1.6-fold, which is lower than that in MM-PCs. MM-BMSC screening also identified a new splice variant of RHAMM, that was absent in MM PCs or in MM myeloid cells. Conclusions: Our study suggests that aberrant RHAMM splicing in MM can result from SNPs/SNVs affecting SRE due to the upregulation of PTBP1/2. Our study is the first to show that the RHAMM-V3 variant is associated with PTBP2 overexpression. The identification of cell type-specific RHAMM splicing events identifies novel targets for improved immunotherapy in MM. Disclosures Chu: Celgene: Honoraria; AstraZeneca: Honoraria; Gilead: Honoraria; Teva: Consultancy; Amgen Inc.: Honoraria. Anderson:C4 Therapeutics: Other: Scientific founder ; OncoPep: Other: Scientific founder ; Gilead Sciences: Other: Advisory Board; Janssen: Other: Advisory Board; Sanofi-Aventis: Other: Advisory Board.

Genome-wide transcriptome profiling detected an increased splicing alterations in MM and AML. While these malignancies are derived from different cell linages, their tumor cells acquire similar aberrant splicing (AbSp), mostly intron retentions. To delineate AbSp mechanism in MM/AML, we focused on PTBPs (1/2/3) that play a critical role in intron excision. We have previously reported deregulated expression of splicing factors (SFs) in MM/AML patient and healthy donor (HD) bone marrow (BM). As MM progressed, PTBP1/2 progressively increased, and PTBP3 gradually decreased (ASH 2017). In AML, PTBP2/3 upregulation and PTBP1 downregulation were detected in patient samples in which increased intron retentions were identified by genome-wide splicing analysis (CCR 2015). Here, these findings were validated by TaqMan assays for PTBPs in 48 MM and 325 AML patient samples, 16 MM/AML cell lines, and in plasma cells (PCs) and CD34+cells from 14 HDBM. Results were consistent with differential expressions of PTBPs in MM/AML previously analyzed. Upregulation of PTBP1/2 and PTBP2/3 proteins were detected in MM and AML cell lines, respectively. PTBP1/2 upregulation was pronounced when MM cell lines were cocultured with BM stromal cells (MMBMSC) derived from MM patients' BM. Importantly, we detected increased proliferation and decreased apoptosis in MM/AML cell lines overexpressing PTBPs. These effects were most evident after coculturing MM cell lines with MMBMSC as compared to HDBMSC. To evaluate PTBP effects on AbSp, we knocked down and/or overexpressed PTBPs in MM/AML cell lines and assessed PDL1 splicing in MM, and NOTCH2 and FLT3 splicing in AML. PDL1 is spliced in ~ 30% of 90 MM patients; while NOTCH2/FLT3/CD13 are spliced in 78%/50% of 387 AML patients, respectively. After PTBP1/2 knockdown in MM cells we detected 4- to 11-fold downregulation of PDL1 splice variants, with proportional upregulation of wild-type PDL1 levels; concurrently, MM cell proliferation was decreased and apoptosis increased. To evaluate MM specific splicing alterations in the context of the MM BM microenvironment, PDL1 splicing, and PTBP1/2 mRNA/protein expressions, were monitored in MM cell lines cocultured with MMBMSC or HDBMSC. We observed time-dependent PDL1 splice variant upregulation, and higher levels of PTBP1/2 in MM cells. We also noted time-dependent PDL1 variant expression switching in association with PTBP1/2 deregulation in the BM microenvironment. RNA-seq analysis and western blotting showed that MMBMSC culture with tumor cells increased intron retention, and altered SF expressions, including PTBPs in BMSC. We next monitored PTBP effects on splicing in AML cells by evaluating NOTCH2 and FLT3 splicing in an TF1, an AML cell line that overexpresses PTBP2/3. RT-PCR analysis showed association between PTBP3 overexpression and NOTCH2 and FLT3 AbSp in TF1 cells. For further validation, we developed an ex vivo splicing assay composed of an FLT3/CD13 splicing cassette with a GFP reporter, which allows for evaluation of splicing events by flow cytometry (FACS) and microscopy. In this assay, cells overexpressing PTBP2/3 caused FLT3/CD13 minigene splicing similar to that detected in AML patients. Also, by RT-PCR, we showed that overexpression of PTBPs caused intron retention, that was confirmed by cloning and sequencing of PCR products and consistent with the FACS analysis and microscopy. Finally, we have tested effects of PTBPs using in vivo AML models (BMT and xenograft). In the BMT model, animal median survival was 48 days post-BMT for MLL-AF9/PTBP3 and 56 days in the control group. In the xenograft model, animal median survival was 66, 94, & 106 days after injection of TF1-PTBP3, TF1-PTBP2, and TF1 cells in mice, respectively (P>0.0001). These studies suggest that PTBP3 overexpression in partnership with the MLL-AF9 promotes occurrence of AML in mice. Tumor RNA samples harvested from these animals were subjected to RNA-seq analysis, which showed increased AbSp association with PTBP3 overexpression. Our studies indicate that deregulated PTBP1/2 expression in MM and of PTBP2/3 in AML drive time-dependent AbSp (intron retention) and splice variant switching, which in MM is induced by culture with BMSC; and conversely, that analogous changes are induced in BMSC cultured with tumor cells. They define role of deregulated expression of the PTBPs in MM/AML pathogenesis, and suggest novel targets for therapeutic intervention. Disclosures Stone: Pfizer: Consultancy; Jazz: Consultancy; Fujifilm: Consultancy; Merck: Consultancy; Cornerstone: Consultancy; Celgene: Consultancy, Other: Data and Safety Monitoring Board, Steering Committee; Novartis: Consultancy, Research Funding; Sumitomo: Consultancy; Ono: Consultancy; Orsenix: Consultancy; Otsuka: Consultancy; AbbVie: Consultancy; Agios: Consultancy, Research Funding; Amgen: Consultancy; Argenx: Other: Data and Safety Monitoring Board; Arog: Consultancy, Research Funding; Astellas: Consultancy. Griffin:Astellas Pharma: Consultancy; Novartis Pharma: Other: Grant, Patents & Royalties: Royalties ; Analysis Group: Consultancy; Sun Pharmaceuticals: Consultancy; RXi Pharmaceuticals: Consultancy; Lilly Pharmaceuticals: Other: Grant; Myeloproliferative Neoplasia Foundation: Other: Grant . Anderson:C4 Therapeutics: Equity Ownership, Other: Scientific founder; Celgene: Consultancy; OncoPep: Equity Ownership, Other: Scientific founder; Bristol Myers Squibb: Consultancy; Gilead: Membership on an entity's Board of Directors or advisory committees; Millennium Takeda: Consultancy.

Although the in vitro expansion of the multiple myeloma (MM) clone has been unsuccessful, in a novel three-dimensional (3-D) culture model of reconstructed bone marrow (BM, n = 48) and mobilized blood autografts (n = 14) presented here, the entire MM clone proliferates and undergoes up to 17-fold expansion of malignant cells harboring the clonotypic IgH VDJ and characteristic chromosomal rearrangements. In this system, MM clone expands in a reconstructed microenvironment that is ideally suited for testing specificity of anti-MM therapeutics. In the 3-D model, melphalan and bortezomib had distinct targets, with melphalan targeting the hematopoietic, but not stromal com-partment. Bortezomib targeted only CD138+CD56+ MM plasma cells. The localization of nonproliferating cells to the reconstructed endosteum, in contact with N-cadherin–positive stroma, suggested the presence of MM-cancer stem cells. These drug-resistant CD20+ cells were enriched more than 10-fold by melphalan treatment, exhibited self-renewal, and generated clonotypic B and plasma cell progeny in colony forming unit assays. This is the first molecularly verified demonstration of proliferation in vitro by ex vivo MM cells. The 3-D culture provides a novel biologically relevant preclinical model for evaluating therapeutic vulnerabilities of all compartments of the MM clone, including presumptive drug-resistant MM stem cells.

Multiple myeloma (MM) is a cancer of plasma cells (PCs) expressing immunoglobulin heavy chain (IgH) postswitch isotypes. The discovery of earlier stage cells related to postswitch PCs, called preswitch clonotypic IgM (cIgM) cells led to the hypothesis that cIgM cells may be MM progenitors, replenishing the tumor throughout malignancy. cIgM cells may do this by undergoing class switch recombination (CSR), a process detectable in postswitch PCs as multiple IgH switch junctions associated with a single clonotypic IgH V/D/J. We addressed this with a specific clonotypic-switch polymerase chain reaction (PCR), informative for 32 of 41 cases. Here we made 2 significant discoveries: (1) in all cases, we detected only a single clonotypic switch fragment that persists over time (1-7.6 years), and (2) we detected ongoing mutation upstream of the switch junction in 5 of 6 patients, often targeting the intronic enhancer, a key control region in IgH expression. The presence of a single, unchanging clonotypic switch junction suggests that cIgM cells are not MM-PC progenitors; rather, postswitch PCs arise from a single cIgM cell, and MM-PC progenitors reside in the postswitch population. Furthermore, mutations revealed here provide a new marker to identify MM-PC progenitors and aggressive clones that evolve throughout malignancy.

Analysis of clonotypic isotype class switching (CSR) in Waldenström macroglobulinemia (WM) and IgM monoclonal gammopathy of undetermined significance (MGUS) reveals a normal initial phase of B-cell activation as determined by constitutive and inducible expression of activation-induced cytidine deaminase (AID). Switch μ (Sμ) analysis shows that large deletions are not common in WM or IgM MGUS. In CD40L/IL-4-stimulated WM cultures from 2 patients, we observed clonotypic IgG exhibiting intraclonal homogeneity associated with multiple hybrid Sμ/Sγ junctions. This suggests CSR had occurred within WM cells. Nevertheless, the estimated IgG/IgM-cell frequency was relatively low (1/1600 cells). Thus, for the majority of WM B cells, CSR does not occur even when stimulated in vitro, suggesting that the WM cell is constitutively unable to or being prevented from carrying out CSR. In contrast to WM, the majority of IgM MGUS clones exhibit intraclonal heterogeneity of IgH VDJ. Furthermore, most IgM MGUS accumulate more mutations in the upstream Sμ region than do WM, making them unlikely WM progenitors. These observations suggest that switch sequence analysis may identify the subset of patients with IgM MGUS who are at risk of progression to WM.

In this study, we show that the hyaluronan synthase 1 (HAS1) gene undergoes aberrant intronic splicing in multiple myeloma (MM). In addition to HAS1 full length (HAS1FL), we identify 3 novel splice variants of HAS1, HAS1Va, HAS1Vb, and HAS1Vc, detected in patients with MM or monoclonal gammopathy of undetermined significance (MGUS). HAS1Vb and HAS1Vc undergo intronic splicing with creation of a premature stop codon. MM cells expressing one or more HAS1 variants synthesize extracellular and/or intracellular hyaluronan (HA). Expression of the HAS1Vb splice variant was significantly correlated with reduced survival (P = .001). Together, alternative HAS1 gene splicing, the correlations between HAS1 splicing and HA synthesis, and the correlations between HAS1 splicing and reduced survival of MM patients support the hypothesis that the family of HAS1 protein plays a significant role in disease progression. Further, expression of HAS1Vb, in conjunction with HAS1FL and/or other HAS1 variants, may lead to accumulation of intracellular HA molecules and an impact on receptor for HA-mediated motility (RHAMM)-mediated mitotic abnormalities in MM. This study highlights the potential importance of HAS1 and its alternative splicing in pathophysiology of MGUS and MM. (Blood. 2005;105: 4836-4844)

Background Multiple myeloma (MM) remains incurable despite treatment advances. While passive immunotherapy such as anti-CD38 antibodies is highly effective, active immunotherapy may provide long-lasting remissions by virtue of triggering memory. A phase 1 nivolumab study, an antibody targeting programmed death-1 (PD1), was unable to yield any responses in multiply relapsed MM patients. Conversely, preliminary trial data of lenalidomide combined with pembrolizumab, a different anti-PD1 antibody, found significantly higher response rates. These two differing outcomes reflect our limited understanding of checkpoint inhibition and immunotherapy in MM. There is a paucity of preclinical models to guide therapeutic studies. Cell lines and xenografted murine models are incapable of exploring active immunotherapy due to a lack of microenvironment and endogenous immune cell signals. Furthermore, malignant cells responsive to drugs in 2-dimensional (2D) cultures are known to display a more resistance in 3D. We have previously demonstrated that B-cell malignancies can be accurately studied using a 3D culture system of patient bone marrow mononuclear cells (BMCs) and can better inform translational trials. Herein we describe an ex vivo, 3D tissue culture model of patient-derived MM samples to more accurately test therapeutics including checkpoint inhibition using ipilimumab, a monoclonal antibody targeting cytotoxic T-lymphocyte antigen 4 (CTLA) which is crucial in co-stimulatory signaling of effector T-cells. Methods A 3D extracellular matrix was created using matrigel in 12-well plates. BMCs were isolated from marrow aspirates of 5 MM patients at time of diagnosis and individually cultured. Each patient sample was tested for sensitivity against increasing concentrations of ipilimumab (1X, 3X, and 10X clinical doses) added into supportive medium. Plates were monitored visually by microscopy followed by harvest on day 21 using enzymatic degradation. Unique clonotypic heavy chain immunoglobulin rearrangement (IgH VDJ) from each sample was sequenced, validated and used for semi-quantitative PCR. Semi-QT PCR with clone-specific primers estimated malignant cell survival after harvest. Flow cytometry was used to define cell populations present in culture and to correlate with clonotypic PCR data. T-cell mediated activity was examined by reverse transcription of trizol-extracted, T-cell RNA after harvest. Results All samples were successfully cultured, followed for 21 days and harvested. Flow cytometry confirmed presence of T-cell subsets, B-cells, NK cells and dendritic cells before and after culture in 3D. Minimal depletion of clonotypic cells was observed at 3x clinical levels of drug. At 10x simulated clinical therapeutic levels, 3 MM samples demonstrated >90% death of clonotypic MM cells while the other 2 demonstrated 62% and 72% death, respectively, compared to untreated control cultures. The extent to which the drug diffuses into the matrigel is as yet unknown. Flow cytometry of harvested cells suggest that T-cells demonstrate a modest shift toward CD4 and CD8 effector cells. Preliminary mechanistic data from one MM sample using trizol-extracted RNA and reverse transcriptase PCR harvested at 21 days from 3D culture suggests that anti-malignant, cytotoxic T-cell effect may be driven by granzyme B expression. Expanded data from the remaining samples will be presented. Conclusions We demonstrate that an ex vivo 3D tissue culture model of MM is both feasible and informative in studying immunotherapy. By culturing unselected BMCs which include stromal cells, immune cells and malignant populations, the 3D culture more closely mimics the tumor microenvironment with both the patient's immune system present as well as stromal supportive signals. In this study, we show that in the presence of active immune effector cells, ipilimumab has activity against patient-derived MM cells. The data suggests the importance of targeted cytotoxic T-cell activation as a primary mechanism of action. We have previously studied standard MM therapeutics such as cytotoxic chemotherapy, immunomodulatory drugs, and proteasome inhibitors in the same way. Consequently, this model is well positioned to study other immunotherapies such as other checkpoint inhibitors, cellular therapy, and combinations. Further testing with therapeutics targeting PD1/PDL1, and adenosine receptors are underway. Disclosures Venner: Takeda: Honoraria; Celgene: Honoraria, Research Funding; J+J: Research Funding; Janssen: Honoraria; Amgen: Honoraria. Belch:Celgene: Honoraria; Janssen: Honoraria; Amgen: Honoraria; Takeda: Honoraria.

The myelomagenic capacity of clonotypic myeloma cells in G-CSF mobilized blood was tested by xenotransplant. Intracardiac (IC) injection of NOD SCID mice with peripheral cells from 5 patients who had aggressive myeloma led to lytic bone lesions, human Ig in the serum, human plasma cells, and a high frequency of clonotypic cells in the murine bone marrow (BM). Human B and plasma cells were detected in BM, spleen, and blood. Injection of ex vivo multiple myeloma cells directly into the murine sternal BM (intraosseus injection [IO]) leads to lytic bone lesions, BM plasma cells, and a high frequency of clonotypic cells in the femoral BM. This shows that myeloma has spread from the primary injection site to distant BM locations. By using a cellular limiting dilution PCR assay to quantify clonotypic B lineage cells, we confirmed that peripheral myeloma cells homed to the murine BM after IC and IO injection. The myeloma progenitor undergoes self-renewal in murine BM, as demonstrated by the transfer of human myeloma to a secondary recipient mouse. For 6 of 7 patients, G-CSF mobilized cells from patients who have minimal disease, taken at the time of mobilization or after cryopreservation, included myeloma progenitors as identified by engraftment of clonotypic cells and/or lytic bone disease in mice. This indicates that myeloma progenitors are mobilized into the blood by cyclophosphamide/G-CSF. Their ability to generate myeloma in a xenotransplant model implies that such progenitors are also myelomagenic when reinfused into patients, and suggests the need for an effective strategy to purge them before transplant.

Although the mechanism(s) underlying mobilization of hematopoietic progenitor cells (HPCs) is unknown, detachment from the bone marrow (BM) microenvironment and motility are likely to play a role. This work analyzes the motile behavior of HPCs and the receptors involved. CD34+45lo/medScatterlo/med HPCs from granulocyte colony-stimulating factor (G-CSF)–mobilized blood and mobilized BM were compared with steady-state BM for their ability to bind hyaluronan (HA), their expression of the HA receptors RHAMM and CD44, and their motogenic behavior. Although RHAMM and CD44 are expressed by mobilized blood HPCs, function blocking monoclonal antibodies (MoAbs) identified RHAMM as a major HA binding receptor, with a less consistent participation by CD44. Permeabilization of mobilized blood HPCs showed a pool of intracellular (ic) RHAMM and a smaller pool of icCD44. In contrast, steady-state BM HPCs have significantly larger pools of icRHAMM and icCD44. Also, in contrast to mobilized blood HPCs, for steady-state BM HPCs, MoAbs to RHAMM and CD44 act as agonists to upregulate HA binding. The comparison between mobilized and steady-state BM HPCs suggests that G-CSF mobilization is associated with depletion of intracellular stores of HA receptors and modulates HA receptor usage. To confirm that mobilization alters the HA receptor distribution and usage by HPCs, samples of BM were collected at the peak of G-CSF mobilization in parallel with mobilized blood samples. HA receptor distribution of mobilized BM HPCs was closely matched with mobilized blood HPCs and different from steady-state BM HPCs. Mobilized BM HPCs had lower pools of icHA receptors, similar to those of mobilized blood HPCs. Treatment of mobilized BM HPCs with anti-RHAMM MoAb decreased HA binding, in contrast to steady-state BM HPCs. Thus, G-CSF mobilization may stimulate an autocrine stimulatory loop for HPCs in which HA interacts with basal levels of RHAMM and/or CD44 to stimulate receptor recycling. Consistent with this, treatment of HPCs with azide, nystatin, or cytochalasin B increased HA binding, implicating an energy-dependent process involving lipid rafts and the cytoskeleton. Of the sorted HPCs, 66% were adherent and 27% were motile on fibronectin plus HA. HPC adherence was inhibited by MoAbs to β1 integrin and CD44, but not to RHAMM, whereas HPC motility was inhibited by MoAb to RHAMM and β1 integrin, but not to CD44. This finding suggests that RHAMM and CD44 play reciprocal roles in adhesion and motility by HPCs. The G-CSF–associated alterations in RHAMM distribution and the RHAMM-dependent motility of HPCs suggest a potential role for HA and RHAMM in trafficking of HPCs and the possible use of HA as a mobilizing agent in vivo.

Although the mechanism(s) underlying mobilization of hematopoietic progenitor cells (HPCs) is unknown, detachment from the bone marrow (BM) microenvironment and motility are likely to play a role. This work analyzes the motile behavior of HPCs and the receptors involved. CD34+45lo/medScatterlo/med HPCs from granulocyte colony-stimulating factor (G-CSF)–mobilized blood and mobilized BM were compared with steady-state BM for their ability to bind hyaluronan (HA), their expression of the HA receptors RHAMM and CD44, and their motogenic behavior. Although RHAMM and CD44 are expressed by mobilized blood HPCs, function blocking monoclonal antibodies (MoAbs) identified RHAMM as a major HA binding receptor, with a less consistent participation by CD44. Permeabilization of mobilized blood HPCs showed a pool of intracellular (ic) RHAMM and a smaller pool of icCD44. In contrast, steady-state BM HPCs have significantly larger pools of icRHAMM and icCD44. Also, in contrast to mobilized blood HPCs, for steady-state BM HPCs, MoAbs to RHAMM and CD44 act as agonists to upregulate HA binding. The comparison between mobilized and steady-state BM HPCs suggests that G-CSF mobilization is associated with depletion of intracellular stores of HA receptors and modulates HA receptor usage. To confirm that mobilization alters the HA receptor distribution and usage by HPCs, samples of BM were collected at the peak of G-CSF mobilization in parallel with mobilized blood samples. HA receptor distribution of mobilized BM HPCs was closely matched with mobilized blood HPCs and different from steady-state BM HPCs. Mobilized BM HPCs had lower pools of icHA receptors, similar to those of mobilized blood HPCs. Treatment of mobilized BM HPCs with anti-RHAMM MoAb decreased HA binding, in contrast to steady-state BM HPCs. Thus, G-CSF mobilization may stimulate an autocrine stimulatory loop for HPCs in which HA interacts with basal levels of RHAMM and/or CD44 to stimulate receptor recycling. Consistent with this, treatment of HPCs with azide, nystatin, or cytochalasin B increased HA binding, implicating an energy-dependent process involving lipid rafts and the cytoskeleton. Of the sorted HPCs, 66% were adherent and 27% were motile on fibronectin plus HA. HPC adherence was inhibited by MoAbs to β1 integrin and CD44, but not to RHAMM, whereas HPC motility was inhibited by MoAb to RHAMM and β1 integrin, but not to CD44. This finding suggests that RHAMM and CD44 play reciprocal roles in adhesion and motility by HPCs. The G-CSF–associated alterations in RHAMM distribution and the RHAMM-dependent motility of HPCs suggest a potential role for HA and RHAMM in trafficking of HPCs and the possible use of HA as a mobilizing agent in vivo.

The receptor for hyaluronan (HA)-mediated motility (RHAMM) controls motility by malignant cells in myeloma and is abnormally expressed on the surface of most malignant B and plasma cells in blood or bone marrow (BM) of patients with multiple myeloma (MM). RHAMM cDNA was cloned and sequenced from the malignant B and plasma cells comprising the myeloma B lineage hierarchy. Three distinct RHAMM gene products, RHAMMFL, RHAMM−48, and RHAMM−147, were cloned from MM B and plasma cells. RHAMMFL was 99% homologous to the published sequence of RHAMM. RHAMM−48 and RHAMM−147 variants align with RHAMMFL, but are characterized by sequence deletions of 48 bp (16 amino acids [aa]) and 147 bp (49 aa), respectively. The relative frequency of these RHAMM transcripts in MM plasma cells was determined by cloning of reverse-transcriptase polymerase chain reaction (RT-PCR) products amplified from MM plasma cells. Of 115 randomly picked clones, 49% were RHAMMFL, 47% were RHAMM−48, and 4% were RHAMM−147. All of the detected RHAMM variants contain exon 4, which is alternatively spliced in murine RHAMM, and had only a single copy of the exon 8 repeat sequence detected in murine RHAMM. RT-PCR analysis of sorted blood or BM cells from 22 MM patients showed that overexpression of RHAMM variants is characteristic of MM B cells and BM plasma cells in all patients tested. RHAMM also appeared to be overexpressed in B lymphoma and B-chronic lymphocytic leukemia (CLL) cells. In B cells from normal donors, RHAMMFL was only weakly detectable in resting B cells from five of eight normal donors or in chronically activated B cells from three patients with Crohn’s disease. RHAMM−48 was detectable in B cells from one of eight normal donors, but was undetectable in B cells of three donors with Crohn’s disease. RHAMM−147 was undetectable in normal and Crohn’s disease B cells. In situ RT-PCR was used to determine the number of individual cells with aggregate RHAMM transcripts. For six patients, 29% of BM plasma cells and 12% of MM B cells had detectable RHAMM transcripts, while for five normal donors, only 1.2% of B cells expressed RHAMM transcripts. This work suggests that RHAMMFL, RHAMM−48, and RHAMM−147 splice variants are overexpressed in MM and other B lymphocyte malignancies relative to resting or in vivo–activated B cells, raising the possibility that RHAMM and its variants may contribute to the malignant process in B-cell malignancies such as lymphoma, CLL, and MM.

Monoclonal antibodies (MoAbs) that selectively identify Muc-1 core protein (MoAbs DF3-P, VU-4H5) determinants were used to identify the Muc-1 glycoform present on 7 multiple myeloma (MM) cell lines, 5 MM patient plasma cells, 12 MM patient B cells, as well as 32 non-MM cell lines and normal hematopoietic cells. Flow cytometry studies demonstrated that all MM cell lines, MM patient plasma cells, and MM patient B cells expressed Muc-1 core protein epitopes. Circulating B cells from 4 normal donors also expressed Muc-1 core protein. In contrast, Muc-1 core protein was absent on 28 of 32 non-MM neoplastic cell lines, 17 of which expressed Muc-1. Splenic and tonsillar B cells, CD34+ stem cells, resting T cells, and bone marrow plasma cells obtained from normal donors both lacked Muc-1 glycoforms. We next studied the effects of estrogen, progesterone, and glucocorticoid receptor agonists and antagonists on Muc-1 expression, because consensus sequences for the response elements of these steroids are present on the Muc-1 gene promoter. These studies showed that dexamethasone (Dex) induced Muc-1 expression on MM cell lines, as determined by both flow cytometry and Western blot analyses. Dex also induced upregulation of Muc-1 on prostate and ovarian cancer cell lines. Time and dose-response studies demonstrated that Dex induced maximal cell surface Muc-1 expression by 24 hours at concentrations of 10−8 mol/L. Dex induced Muc-1 upregulation could be blocked with a 10-fold excess of the glucocorticoid receptor antagonist RU486, confirming that Dex was acting via the glucocorticoid receptor. No changes in Muc-1 expression were observed on MM cells treated with estrogen and progesterone receptor agonists and antagonists or with RU486. These studies provide the framework for targeting Muc-1 core protein in vaccination and serotherapy trials in MM. In addition, the finding that Muc-1 expression on MM cells can be augmented by Dex at pharmacologically achievable levels suggests their potential utility in enhancing treatments targeting Muc-1 in MM.

In multiple myeloma (MM), the VDJ rearrangement of the immunoglobulin heavy chain expressed by MM plasma cells provides a unique clonotypic marker. Although clonotypic MM cells have been found in the circulation, their number has been controversial. Our objective was to provide direct evidence, using single-cell assays, for the frequency of clonotypic cells in blood of 18 MM patients, and to confirm their identity as B cells. The clonotypic Ig heavy-chain (IgH) VDJ was determined from single plasma cells using consensus reverse transcriptase-polymerase chain reaction (RT-PCR), subcloning, and sequencing. For all patients, using patient-specific primers, clonotypic transcripts were amplified from 10 or more individual plasma cells. Using in situ RT-PCR, for all patients greater than 80% of plasma cells were found to be clonotypic. Three separate methods, RT-PCR, single-cell RT-PCR, and in situ RT-PCR, were used to analyze clonotypic cells in peripheral blood mononuclear cells (PBMC) from MM patients. Sequencing of the IgH transcripts expressed by individual cells obtained by limiting dilution of freshly isolated PBMC from a MM patient showed that all B cells expressed an identical CDR3. This intraclonal homogeneity indicates an escape from antigenic-selection, characteristic of malignant B cells. For this patient, the frequency of clonotypic PBMC, about 25%, was comparable to the number of PBMC B cells (34%). Because the PBMC included less than 1% plasma cells, virtually all clonotypic PBMC must be B cells. Using single-cell RT-PCR, clonotypic IgH transcripts were identified in individual sorted B cells from blood. To accurately quantify the number of clonotypic B cells, sorted B cells derived from 18 MM patients (36 samples) and 18 healthy donors (53 samples) were analyzed using in situ RT-PCR with patient-specific primers. Clonotypic transcripts were not detectable among normal B cells. For the 18 MM patients, a mean of 66% ± 4% (SE) of blood B cells were clonotypic (range, 9% to 95%), with mean absolute number of 0.15 ± .02 × 109/L blood. Over time in individual patients, conventional chemotherapy transiently decreased circulating clonotypic B cells. Their numbers were increased in granulocyte colony-stimulating factor (G-CSF)– mobilized blood of one patient. However, clonotypic B cells of a one patient became undetectable after allogeneic transplant, correlating with complete remission. Although contributions to MM spread and progression is likely, their malignant status and impact has yet to be clarified. Their high frequency in the blood, and their resistence to conventional chemotherapy suggests that the number of circulating clonotypic cells should be clinically monitored, and that therapeutic targeting of these B cells may benefit myeloma patients. © 1998 by The American Society of Hematology.

Although chemotherapy effectively reduces the plasma cell burden in multiple myeloma (MM), the disease recurs. MM includes circulating and bone marrow (BM) localized components. A large majority of circulating CD11b+ MM B cells (81%) express an IgH VDJ rearrangement identical to that of autologous BM plasma cells. Unlike plasma cells, these monoclonal circulating B cells exhibit dye and drug transport activity before and throughout chemotherapy. Drug resistance was measured as the ability to export the fluorescent dye Rhodamine123 (Rh123) or the drug adriamycin, using flow cytometry. The role of P-glycoprotein 170 (P-gp), the multidrug transporter, was defined by cyclosporin A (CsA)-sensitive dye export. Only 8% to 11% of BM-localized plasma cells exported dye with the majority retaining dye, identified as bright staining. Circulating leukemic plasma cells were also unable to export dye and remained Rh123bright. However, 53% of circulating clonotypic MM B cells exhibited CsA-sensitive dye export. BM plasma cells taken before or after initiation of first line chemotherapy were equally unable to export dye. Thus in myeloma, differentiation to the plasma cell stage is accompanied by a loss of P-gp function, although P-gp phenotypic expression is retained. In contrast, for monoclonal gammopathy of undetermined significance (MGUS), 54% of BM-localized plasma cells exported dye, comparable to the 53% of circulating MGUS B cells that also exported dye, suggesting that the apparent defect in P-gp function is unique to myeloma plasma cells. Virtually all BM plasma cells in MM retained the drug adriamycin, consistent with their initial drug sensitivity in vivo, in contrast to circulating MM B cells, or to T cells in BM or blood. Thus, circulating B cells appear to be the predominant drug resistant component of the MM B-lineage hierarchy. This report suggests that successful therapeutic strategies will be those that target circulating B cells. Chemosensitization methods involving inhibition of P-gp are likely to improve depletion of these cells by compromising their ability to exclude drug. This work suggests that circulating clonotypic B cells should be monitored in clinical trials to confirm their depletion and the overall efficacy of novel treatment strategies.

Peripheral blood mononuclear cells (PBMC) from patients with multiple myeloma (MM) are here shown to include 23% ± 2% of CD34+ cells, the majority of which coexpress CD19, as identified by a panel of 17 anti-CD34 antibodies. The expression of CD34 mRNA by sorted CD34+ PBMC from MM was confirmed by in situ reverse transcriptase-polymerase chain reaction (RT-PCR) with CD34-specific primers. The majority of CD34+ MM PBMC were CD19+ cells that expressed mRNA for CD19 and for rearranged IgH as identified with consensus IgH VDJ primers, as well as having cytoplasmic Ig, definitively identifying them as B cells, in absolute numbers of 0.06 to 0.69 × 109/L of blood. CD34 is largely absent from normal B cells. To determine the clonal relationship of CD34+ B cells to autologous MM plasma cells, IgH VDJ DNA rearrangements of sorted CD34+ MM blood B cells were amplified by nested PCR using consensus primers followed by Southern blotting with allele-specific oligonucleotides for 7 MM patients, and clonotypic IgH mRNA expression was assessed for 4 MM patients using quantitative patient-specific in situ RT-PCR. For 9 of 11 myeloma patients tested, CD34+ blood B cells included IgH gene rearrangements or expressed IgH mRNA identical to that of autologous bone marrow plasma cells. For 4 of 4 MM patients, 74% to 94% of individual sorted CD34+19+ B cells expressed clonotypic IgH mRNA, as detected by in situ RT-PCR with patient-specific primers. Clonotypic IgH VDJ sequences were absent from B cells of unrelated MM patients and of normal donors. Clonotypic CD34+ B cells were detected before, during, and after treatment, and during relapse. Our results indicate a clonal relationship between CD34+ MM B cells and malignant plasma cells. We speculate that CD34 may play an important role in the biology of myeloma by facilitating extravasation from blood and thus spread of myeloma through the skeletal system.

Peripheral blood mononuclear cells (PBMC) from patients with multiple myeloma (MM) are here shown to include 23% ± 2% of CD34+ cells, the majority of which coexpress CD19, as identified by a panel of 17 anti-CD34 antibodies. The expression of CD34 mRNA by sorted CD34+ PBMC from MM was confirmed by in situ reverse transcriptase-polymerase chain reaction (RT-PCR) with CD34-specific primers. The majority of CD34+ MM PBMC were CD19+ cells that expressed mRNA for CD19 and for rearranged IgH as identified with consensus IgH VDJ primers, as well as having cytoplasmic Ig, definitively identifying them as B cells, in absolute numbers of 0.06 to 0.69 × 109/L of blood. CD34 is largely absent from normal B cells. To determine the clonal relationship of CD34+ B cells to autologous MM plasma cells, IgH VDJ DNA rearrangements of sorted CD34+ MM blood B cells were amplified by nested PCR using consensus primers followed by Southern blotting with allele-specific oligonucleotides for 7 MM patients, and clonotypic IgH mRNA expression was assessed for 4 MM patients using quantitative patient-specific in situ RT-PCR. For 9 of 11 myeloma patients tested, CD34+ blood B cells included IgH gene rearrangements or expressed IgH mRNA identical to that of autologous bone marrow plasma cells. For 4 of 4 MM patients, 74% to 94% of individual sorted CD34+19+ B cells expressed clonotypic IgH mRNA, as detected by in situ RT-PCR with patient-specific primers. Clonotypic IgH VDJ sequences were absent from B cells of unrelated MM patients and of normal donors. Clonotypic CD34+ B cells were detected before, during, and after treatment, and during relapse. Our results indicate a clonal relationship between CD34+ MM B cells and malignant plasma cells. We speculate that CD34 may play an important role in the biology of myeloma by facilitating extravasation from blood and thus spread of myeloma through the skeletal system.

We investigated the ability of blood B cells, bone marrow (BM) plasma cells, and terminal leukemic plasma cells (T-PCL) from patients with multiple myeloma (MM) to migrate on extracellular matrix proteins. Hyaluronan (HA), but not collagen type I, collagen type IV, or laminin, promoted migration of MM blood B cells, as determined by time-lapse video microscopy. Between 13% and 20% of MM blood B cells migrated on HA with an average velocity of 19 micron/min, and greater than 75% of MM blood B cells exhibited vigorous cell movement and plasma membrane deformation, as did circulating T-PCL and extraskeletal plasma cells from patients with MM. In contrast, plasma cells obtained from BM of patients with MM lacked motility on all substrates tested and did not exhibit cell membrane protrusions or cellular deformation. MM blood B cells and MM plasma cells from all sources examined expressed the HA- binding receptors receptor for HA-mediated motility (RHAMM) and CD44. On circulating MM B cells, both RHAMM and CD44 participated in HA- binding, indicating their expression ex vivo in an activated conformation. In contrast, for the majority of BM plasma cells in the majority of patients with MM, expression of RHAMM or CD44 was not accompanied by HA binding. A minority of patients did have HA-binding BM plasma cells, involving both RHAMM and CD44, as evidenced by partial blocking with monoclonal antibodies (MoAbs) to RHAMM or to CD44. Despite HA binding by both RHAMM and CD44, migration of MM blood B cells on HA was inhibited by anti-RHAMM but not by anti-CD44 MoAbs, indicating that RHAMM but not CD44 mediates motility on HA. Thus, circulating B and plasma cells in MM exhibit RHAMM- and HA-dependent motile behavior indicative of migratory potential, while BM plasma cells are sessile. We speculate that a subset(s) of circulating B or plasma cells mediates malignant spread in myeloma.

Multiple myeloma (MM) is characterized by a plasma cell infiltrate of the bone marrow (BM). However, late-stage monotypic B cells have been detected in the blood. This work analyzes the effects of clinical treatment on late stage CD19+ B cells present in 752 blood samples from 152 MM patients. MM patients have 2 to 8 times as many circulating CD19+ cells as do normal donors. Analysis of the Ig heavy chain (IgH) gene rearrangements using polymerase chain reaction indicates that the CD19+ population includes cells sharing the same clonotypic CDR3 region as is detected in the BM plasma cells, for patients analyzed during chemotherapy or in relapse. They are also monotypic as defined by their cytoplasmic or surface expression of Ig kappa or lambda light chain. The light chain restriction is the same as that of the BM plasma cells. Individual patients observed over 1- to 2-year periods exhibit considerable variation in the number of B cells present in blood; this number does not correlate with the concentration of serum monoclonal Ig. The monoclonal blood CD19+ cells are not eliminated by any of the chemotherapy regimens analyzed and remain at high levels during transient remissions. Patients in the progressive phase of disease or in relapse have significantly higher numbers of B cells than do patients in transient remission or untreated patients. During periods when the quantity of blood B cells approaches normal, phenotypically their quality is highly abnormal, with physical and phenotypic heterogeneity. Most B cells express CD45R0, a high density of CD38, and CD56 characteristic of late-stage B or pre-plasma cells. CD38hi blood B cells had a cyclical presence. We conclude that monoclonal B cells in the blood of myeloma patient populations include drug-resistant reservoirs of clonotypic cells that may underlie relapse.

Introduction: With the widespread adoption of novel agents (NA) in all lines of therapy patients are being exposed to both proteosome inhibitors (PIs) and immunomodulatory drugs (IMiDs) early in their treatment course. This has lead to marked improvements in survival in the frontline setting. Data is limited with respect to patient outcomes after exposure to both these drug classes in the real world setting. Here we present our experience examining outcomes after each line of therapy whereby bortezomib-based induction followed by lenalidomide-based therapy at first relapse has become the standard of care. We further explored the outcomes of patients who were exposed to both active classes of drugs within their first 2 lines of therapy. Patients and methods: This series includes patients seen through the provincial Alberta Myeloma and Dysproteinemia Program in Canada. Only patients treated between 2005-2013 were included to allow at least 2 years of follow-up beyond first line therapy. Only those treated with a NA-containing regimen as part of their first line treatment were examined. The cohort was split based on eligibility for autologous stem cell transplant (ASCT). Double exposed patients were those who had been treated with, but were not necessarily refractory to both an IMiD and PI within the first 2 lines of treatment. Outcomes were measured after first, second and third line therapy. Survival outcomes were measured in months (m). OS was measured from the start of each line of therapy until death or last follow-up. PFS was from the start of each line of therapy to relapse, death or last follow-up. Response was measured as per the most recent International Myeloma Working Group criteria. Near complete response (nCR) was used when the monoclonal protein disappeared on protein electrophoresis but was not confirmed by immunofixation. Results: Two hundred forty eight patients had received upfront therapy (non-ASCT = 113 and ASCT = 135). One hundred twenty seven had received second line therapy (non-ASCT = 62 and ASCT = 65). Sixty-four had received third line therapy (non-ASCT = 31 and ASCT = 33). The median OS and PFS after each line of therapy are shown in table 1. After first line therapy the OS (p < 0.001) and PFS (p< 0.001) were significantly better in the ASCT cohort. There were no significant differences in survival outcomes based on transplant eligibility in subsequent lines of therapy (figure 1A and B). The overall response rate to third line therapy was 45% (VGPR = 14% and nCR = 7%) for non-ASCT patients and 52% (VGPR = 15% and nCR = 6%) for ASCT patients. Fifty-five percent of non-ASCT patients failed to respond during third line therapy (34% with progressive (PD) and 21% with stable disease (SD)). Forty-eight percent of ASCT patients failed to respond (PD = 27% and SD = 21%). Forty-seven patients were double exposed within the first 2 lines of therapy (non-ASCT = 26 and ASCT = 21). In this cohort, the OS and PFS after double exposure (i.e. third line therapy) was 15m and 5m respectively with no significant difference based transplant eligibility (figure 1C and D). The response rate to third line therapy was 46% (VGPR = 17% and nCR = 8%) for ASCT patients and 43% (VGPR = 14% and nCR = 5%) for non-ASCT patients. Fifty-five percent failed to respond (PD = 38% and SD = 17%) in the non-ASCT group. Fifty-seven percent failed to respond (PD = 38% and SD = 19%) in the ASCT group. Summary: The introduction of NAs earlier in the management of patients with myeloma has improved OS. This is driven by improvements in PFS to frontline therapy and after first relapse. However, with current therapeutic approaches patients will be exposed to both IMiDs and PIs much earlier in their disease. In many jurisdictions, the limited treatment options in third line and beyond, especially in double exposed patients, poses a significant therapeutic challenge. Durable responses are limited in this setting with most patients relapsing after only 6 months. In addition, approximately a third of patients have overtly progressive disease. Interestingly, front-line ASCT eligibility had no impact on outcome with subsequent relapses, emphasizing the fact that ASCT only improves the outcome for the line in which it is employed. Further study regarding resistant mechanism and clonal evolution after exposure to both IMiDs and PIs will be important in developing rationally designed therapeutic regimens for this population. Disclosures Venner: J&J: Honoraria, Research Funding; Amgen: Honoraria; Celgene: Honoraria, Research Funding. Off Label Use: Some patients in this series will have received frontline lenalidomide which is not yet an approved indication for this drug in Canada.. Bahlis:Johnson & Johnson: Speakers Bureau; Johnson & Johnson: Consultancy; Amgen: Consultancy; Celgene: Consultancy, Honoraria, Research Funding, Speakers Bureau; Johnson & Johnson: Research Funding. Neri:Celgene: Research Funding. Sandhu:Amgen: Consultancy, Honoraria; Novartis: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Celgene: Consultancy, Honoraria. Duggan:Jansen: Honoraria; Celgene: Honoraria. Belch:Janssen-Cilag: Consultancy. Jimenez-Zepeda:Celgene: Honoraria; J&J: Honoraria; Amgen: Honoraria.

Multiple myeloma is basically an incurable cancer. Most patients respond initially to chemotherapy with reduction in bone marrow (BM) plasma cells and monoclonal Ig levels, but the disease nearly always recurs and becomes refractory to therapy. The objective of this study was to characterize the expression of the multidrug transport pump, P- glycoprotein 170 (P-gp), in myeloma. The great majority of B cells from peripheral blood mononuclear cells (PBMCs) in myeloma express P-gp, detected by the monoclonal antibody MRK-16. P-gp+ blood B cells exhibit extensive DNA hyperdiploidy, suggesting replicative abnormality characteristic of malignant growth. We speculate these represent a stem cell population in myeloma. The proportion of B cells expressing P-gp was comparable among untreated myeloma patients and those treated with chemotherapy, biologic response modifiers, or off treatment. Among BM cells, P-gp was absent or low in untreated myeloma patients but was expressed at high levels on BM cells from patients previously treated with chemotherapy. For untreated patients the majority of B/plasma cells expressing P-gp are located in PBMCs, not the BM cells. Flow cytometric analysis of rhodamine 123 dye efflux indicated a functional P-gp that was efficiently blocked by verapamil or cyclosporin A (CsA). Both the CD11bhi CD19+ B cells and the T cells in myeloma PBMCs had active CsA-inhibited dye efflux, but monocytes lacked the ability to efflux dye. Nearly all CD38hi plasma cells from myeloma BM cells retained dye, indicating their lack of a functional transport pump. Thus, PBMC B cells may be the predominant set of drug-resistant tumor cells. Myeloma PBMC B cells were cultured with Adriamycin with or without CsA and drug toxicity evaluated by the induction of apoptosis, using flow cytometry to quantitate DNA disruption. No apoptosis was detectable at 0.01 microgram/mL adriamycin, the in vivo steady-state level, with or without CsA. With 0.1 microgram adriamycin, no apoptosis was detectable in the absence of CsA, but with CsA, 66% of B cells initiated DNA disruption, whereas most T cells were spared. This work suggests that currently used drug dosages are too low to effect P-gp+ B- cell death, even in the presence of CsA. We suggest that blood B cells comprise a highly drug-resistant subset of the myeloma B lineage that escapes conventional chemotherapy and may underlie the almost uniform fatal relapse in myeloma patients.

The peripheral blood lymphocytes from 42 patients with multiple myeloma (MM) and 13 patients with monoclonal gammopathy of undetermined significance (MGUS) were studied by three-color immunofluorescence (IF) using antibodies directed to a broad range of B-cell markers (CD19, CD20, CD21, CD24), CALLA (CD10), PCA-1 (a plasma cell marker), and to the high and low molecular weight isoforms of the leukocyte common antigen, CD45RA (p205/220) and CD45RO (p 180). CD45RA is expressed on pre-B and B cells, and a transition from CD45RA to CD45RO defines differentiation towards plasma cells. Peripheral blood mononuclear cells (PBMC) from patients with myeloma included a large subset of B- lineage cells (mean of 39% to 45%) that were CALLA+ and PCA-1+ in all patients studied, including newly diagnosed patients and patients undergoing chemotherapy. Southern blot analysis indicated the presence of monoclonal Ig rearrangements in PBMC and a substantial reduction in the germ-line bands consistent with the presence of a large monoclonal B-cell subset. Avoidance of purification methods involving depletion of adherent cells was essential for detection of the abnormal B cells. Phenotypically, this abnormal B-cell population corresponded to late B or early pre-plasma cells (20% to 80% of PBMC), as defined by the concomitant expression of low densities of CD19 and CD20, moderate densities of CALLA and PCA-1, and strong expression of CD45RO on all B cells, with weakly coexpressed CD45RA on a small proportion. Heterogeneity in the expression of CD45RA and CD45RO within the abnormal B-cell population from any given patient suggested multiple differentiation stages. Abnormal B cells similar to those in MM were also detected in MGUS, although as a lower proportion of PBMC (26%). Abnormal B cells from patients with MGUS expressed predominantly the CD45RO isoform, but had a lower proportion of CALLA+ and PCA-1+ cells than were found on B cells from MM. This work indicates that the large subset of circulating monoclonal B lymphocytes from myeloma patients are at a late stage in B-cell differentiation, continuously progressing towards the plasma cell stage.

To characterize genetic contributions toward aberrant splicing of the hyaluronan synthase 1 (HAS1) gene in multiple myeloma (MM) and Waldenstrom macroglobulinemia (WM), we sequenced 3616 bp in HAS1 exons and introns involved in aberrant splicing, from 17 patients. We identified a total of 197 HAS1 genetic variations (GVs), a range of 3 to 24 GVs/patient, including 87 somatic GVs acquired in splicing regions of HAS1. Nearly all newly identified inherited and somatic GVs in MM and/or WM were absent from B chronic lymphocytic leukemia, nonmalignant disease, and healthy donors. Somatic HAS1 GVs recurred in all hematopoietic cells tested, including normal CD34+ hematopoietic progenitor cells and T cells, or as tumor-specific GVs restricted to malignant B and plasma cells. An in vitro splicing assay confirmed that HAS1 GVs direct aberrant HAS1 intronic splicing. Recurrent somatic GVs may be enriched by strong mutational selection leading to MM and/or WM.

The aurora kinases facilitate transit from G2 through cytokinesis and, thus, are targets in cancer therapy. Multiple myeloma (MM) is a malignancy characterized by genetic instability, suggesting a disruption of checkpoints that arrest cells at G2M when injury to the mitotic machinery occurs. Since deficient checkpoints would prevent cell cycle arrest and may render cells susceptible to apoptosis in mitosis and since aurora kinases are intermediaries in checkpoint pathways, we tested antimyeloma effects of 2 agents that inhibit aurora kinases. Both inhibited growth of MM lines and primary myeloma samples at nanomolar concentrations while having less of an effect on proliferating lymphocytes and hematopoietic cells. MM cells were not protected by IL-6 or activating mutations of Ras. Antimyeloma effects included induction of tetraploidy followed by apoptosis. Apoptosis correlated with inhibition of aurora activity as shown by reduction of histone 3B phosphorylation. Ectopic expression of aurora A protected MM cells against aurora inhibitors but had no effect on apoptosis induced by bortezomib. As expression of RHAMM in MM contributes to genetic instability, we tested effects of RHAMM. RHAMM overexpression enhanced sensitivity to apoptosis and RHAMM silencing decreased sensitivity. These results suggest potential for aurora kinase inhibitors in MM especially in patients in whom RHAMM is overexpressed.

Multiple myeloma (MM) plasma cells (PCs) express receptor for hyaluronan-mediated motility (RHAMM), a hyaluronan-binding, cytoskeleton and centrosome protein. The most abundant RHAMM isoforms in MM are full-length RHAMM (RHAMMFL) and the splice variant RHAMM-exon4. We separately examined the significance of RHAMM expression, and isoform balance, in 2 groups of MM patients. In oligonucleotide microarray experiments (n = 210, Arkansas), increasing RHAMM mRNA expression in MM PCs is strongly associated with osteolytic bone lesions (P < .001), and event-free (P = .05) and overall (P = .04) survival. Semiquantitative determination of RHAMM isoform expression (Alberta, Canada) used capillary electrophoretic detection and measurement of RHAMM-exon4/RHAMMFL reverse-transcriptase-polymerase chain reaction (RT-PCR) products. RHAMM isoforms are rarely expressed concurrently in single MM PCs; the pattern of isoform expression, at the single-cell level, is approximated in larger numbers of cells by the RHAMM-exon4/RHAMMFL ratio. Absolute RHAMM expression and the RHAMM-exon4/RHAMMFL ratio are only partially correlated in MM PCs; in cell lines, absolute RHAMM expression is elevated in mitosis, while RHAMM ratios remain stable. Temporal examination of MM patients' peripheral blood reveals that the RHAMM-exon4/RHAMMFL ratio increases with disease burden. The RHAMM-exon4/RHAMMFL ratio in diagnostic bone marrow samples (n = 101, Alberta) is an independent prognostic factor. Thus, expression and splicing of RHAMM are important molecular determinants of disease severity in MM. (Blood. 2004;104:1151-1158)

Analysis of clonotypic immunoglobulin M (IgM) from 15 patients with Waldenstrom macroglobulinemia (WM) showed a strong preferential use of the VH3/JH4 gene families. Identification of the WM IgM V/D/J was validated using single-cell analysis, confirming its presence in most B cells. Despite the extensive hypermutated VH genes in 13 of 15 patients, statistical analysis of framework/complementary-determining region (FR/CDR) mutation patterns suggests that they might have escaped antigenic selection. Neither intraclonal diversity nor isotype switching was detectable. Membranous and secreted forms of clonotypic IgM transcripts were present in bone marrow and blood. Single-cell analysis showed that clonotypic B cells coexpress CD20, surface IgM (sIgM), and sIgD but that they lack CD138. Most B cells lacked memory marker CD27 despite their hypermutated variable regions otherwise suggestive of memory status. At diagnosis, circulating B cells in WM are largely clonotypic. However, when monoclonal IgM levels are decreased, clonotypic frequencies are substantially reduced despite elevated CD20+ cells, shown to be polyclonal by DNA sequencing and CDR3 fragment analysis. Thus, WM includes the expansion of circulating, polyclonal B cells. Overall, this work suggests that WM may originate from a largely VH3-restricted, somatically mutated, predominantly CD27-IgM+IgD+ population that cannot undergo class switching, suggestive of B cells that might have bypassed the germinal center. (Blood. 2004;104:2134-2142)

This study analyzed the frequency and clinical significance of t(4;14)(p16;q32) in multiple myeloma (MM) among 208 patients with MM and 52 patients with monoclonal gammopathy of undetermined significance (MGUS); diagnosed between 1994 and 2001. Patients with the translocation were identified using reverse transcription–polymerase chain reaction (RT-PCR) to detect hybrid immunoglobulin heavy chain (IgH)–MMSET transcripts from the der(4) chromosome. We found 31 (14.9%) t(4;14)+ MM patients and 1 (1.9%) t(4;14)+ MGUS patient. IgH-MMSET hybrid transcripts were detected in bone marrow (BM) and blood. Breakpoint analysis revealed that 67.7% of t(4;14)+ patients expressed hybrid transcripts potentially encoding full-length MMSET, whereas the remainder lacked one or more amino terminal exons. Expression of fibroblast growth factor receptor 3 (FGFR3), presumptively dysregulated on der(14), was detected by RT-PCR in only 23 of 31 (74%) patients with t(4;14)+ MM. Patients lacking FGFR3 expression also lacked detectable der(14) products. Longitudinal analysis of 53 MM patients with multiple BM and blood samples showed that, over time, BM from t(4;14)+ patients remained positive and that t(4;14)− patients did not acquire the translocation. IgH-MMSET hybrid transcripts and FGFR3 transcripts disappeared from blood during response to therapy. No correlation was observed between the occurrence of t(4;14) and known prognostic indicators. However, we find the t(4;14) translocation predicts for poor survival (P = .006; median, 644 days vs 1288 days; hazard ratio [HR], 2.0), even in FGFR3 nonexpressors (P = .003). The presence of t(4;14) is also predictive of poor response to first-line chemotherapy (P = .05). These results indicate a significant clinical impact of the t(4;14) translocation in MM that is independent of FGFR3 expression.

Multiple myeloma (MM) is identified by unique immunoglobulin heavy chain (IgH) variable diversity joining region gene rearrangements, termed clonotypic, and an M protein termed the “clinical” isotype. Transcripts encoding clonotypic pre and postswitch IgH isotypes were identified in MM peripheral blood mononuclear cells (PBMCs), bone marrow (BM), and mobilized blood. For 29 patients, 38 BM, 17 mobilized blood, and 334 sequential PBMC samples were analyzed at diagnosis, before and after transplantation for 2 to 107 months. The clinical clonotypic isotype was readily detectable and persisted throughout treatment. Eighty-two percent of BM and 38% of PBMC samples also expressed nonclinical clonotypic isotypes. Clonotypic immunoglobulin M (IgM) was detectable in 68% of BM and 25% of PBMC samples. Nonclinical clonotypic isotypes were detected in 41% of mobilized blood samples, but clonotypic IgM was detected in only 12%. Patients with persistent clonotypic IgM expression had adverse prognostic features at diagnosis (lower hemoglobin, higher β2-microglobulin) and higher numbers of BM plasma cells compared with patients with infrequent/absent clonotypic IgM. Patients with persistent clonotypic IgM expression had significantly poorer survival than patients with infrequent IgM expression (P

Soluble MUC1 (sMUC1) levels are elevated in many MUC1+cancers. We and others have shown that MUC1 is expressed on multiple myeloma (MM) plasma cells and B cells. In this study, we measured sMUC1 levels in bone marrow (BM) plasma from 71 MM patients and 21 healthy donors (HDs), and in peripheral blood (PB) plasma from 42 MM patients and 13 HDs using an immunoassay that detects the CA27.29 epitope of MUC1. sMUC1 levels were found to be significantly greater (mean 31.76 U/mL, range 5.69 to 142.48 U/mL) in MM patient BM plasma versus HD BM plasma (mean 9.68 U/mL, range 0.65 to 39.83 U/mL) (P

Abstract 3930 Risk associations with genetic changes, especially SNPs, can be most meaningfully interpreted when the functional relevance of the involved gene(s) is known. Further, using targeted sequencing to detect SNPs provides for unequivocal allele “calling” in each patient, as well as identification of any linked low penetrance mutations that might influence risk. HAS1 is overexpressed and aberrantly spliced in malignant cells from multiple myeloma (MM) and Waldenstrom macroglobulimenia (WM), but not in healthy donors (HD); HAS1Vb correlated with reduced survival in a cohort of MM patients1. In transfectants, HAS1 variants are oncogenic in vivo and/or in vitro2. As shown here, a set of three intron 3 SNPs in linkage disequilbrium have significantly different genotype and allele frequencies and robust hazard ratios in people with B lineage malignancies as compared to age matched controls and to a set of unlinked exon 3 HAS1 SNPs. In all patient and control groups, all five SNPs met Hardy-Weinberg equilibrium. We sequenced an 850bp segment of the HAS1 gene (exon 3, intron 3) from PBMC of 307 Caucasian individuals, including 86 MM, 70 monoclonal gammopathy of undetermined significance (MGUS), 25 WM, 40 B chronic lymphocytic leukemia (CLL), 15 affected and 21 unaffected members of a monoclonal gammopathy prone Icelandic kindred, and 60 age-matched HD. Using direct or subclone sequencing, we evaluated the frequencies of NCBI designated minor alleles for the 5 SNPs. Bioinformatics and in vitro mutagenesis experiments confirmed that intron 3 plays a central role in clinically relevant splice site selection and aberrant intronic HAS1 splicing3. The linked intron 3 HAS1 SNPs (rs11084110, rs11084109 and rs11669079) in patient groups have significantly different host genotype and allele frequencies from those of HD. The frequencies of the unlinked exon 3 HAS1 polymorphisms (rs61736495, rs11084111) on the same 850bp amplicon were not significantly different between patients and HD, providing internal controls for sequencing artefacts. Associations between risk of B cell malignancy and HAS1 SNPs were evaluated using the logistic regression model. Compared to age matched controls, the HAS1 intron 3 SNPs were significantly associated with MM and MGUS for genotype frequencies (p.01 to.05) and allele frequencies (p.01 to.0007); the association was even stronger for CLL and WM (genotype frequencies p 1. Blood 2005 105:4836 2. JBC 2009 285:18840 3. Blood 2008 512:5111 Disclosures: Belch: Celgene: Research Funding; Onyx: Research Funding.

Abstract 3931 Clinically important aberrant splicing of the hyaluronan synthase 1 gene (HAS1) occurs in malignant B cells from multiple myeloma (MM) and Waldenstrom macroglobulimenia (WM) patients but is undetectable in B cells from healthy donors (HD)1. HAS1 splice variants are generated by aberrant alternative splicing (AS) within exons 3, 4 and 5, involving exon 4 skipping (Va), partial intron 4 retention (Vc, Vd) or a combination of both (Vb). Aberrant splicing of HAS1Vb has previously shown to correlate with reduced survival in a cohort of MM patients1. In patients, frequent mutations have been identified in introns 3 and 4, suggesting their roles in AS1. To identify genetic regions involved in splice site selection leading to HAS1Vb expression, we performed in vitro sequence manipulation of introns 3 and 4. We established a mammalian expression system to analyse HAS1 splicing by fusing a minigene extending from exon 3 to exon 5 with the upstream cDNA sequence. In an unaltered sequence of the HAS1 minigene, splicing of HAS1 full length predominated, with trace amount of variants, including the novel HAS1Vd, identified for the first time in transfectants. HAS1Vd shared an alternative acceptor site with HAS1Vb (retained 59 bp of downstream intron 4) but unlike HAS1Vb, retained exon 4. Analysis of peripheral blood mononuclear cells (PBMC) revealed that 9% of healthy donors (n=102) and MM patients (n=93) expressed HAS1Vd. About 2% coexpressed HAS1Vb and Vd. In this cohort of patients, HAS1Vb was found in 20% of unfractionated MM PBMC compared to 5% in HD PBMC, supporting our previous finding that HAS1Vb was restricted to sorted MM B cells1and suggesting that in HD PBMC, non-B cells are responsible for this splicing. While MM PBMCs expressed HAS1Vb>Vd, HD PBMCs and transfectants expressed HAS1Vd>Vb, indicating that splicing directed by the minigene construct was substantively different from that occurring in MM patients. In transfectants, partial deletion of HAS1 intron 4 increased HAS1Vd expression but did not affect HAS1Vb despite the fact that both variants use the same 3' splice site in intron 4. Proper joining of exons 3, 4 and 5 required a minimum of 172 bp of downstream intron 4 and acceptor site selection may be regulated by sequence between 172 and 84 bp upstream of exon 5. Thus, changes in intron 4 alone were insufficient to promote the splicing pattern observed in patients (i.e., elevated HAS1Vb). We then employed site directed mutagenesis to alter multiple G-rich regions in downstream intron 3, a strategy which enhanced exon 4 skipping as shown by increased HAS1Va expression, but did not lead to HAS1Vb splicing. Minimal manipulation to promote exon 4 skipping included base changes within the 2 tandem G-repeats (8 bp) located 73 bp upstream of exon 4, where splicing enhancer/silencer sequences were also identified. Thus, changes in intron 3 also affected the HAS1 splicing profile but like deletions in intron 4, were insufficient on their own to promote HAS1Vb expression. We next developed expression constructs that combined deletion in intron 4 with mutations in intron 3. We are able to demonstrate that when both introns were co-modified, the splicing pattern shifted towards increased HAS1Vb expression, similar to that observed in malignant cells from MM patients. Our previous work showed that deletions or mutations in the 3' end of intron 4 are frequent in MM; in silico analysis predicts splicing to form HAS1Vb2. To determine whether intron 3 genetic changes occur in patients, we sequenced intron 3 from genomic DNA of 50 MM PBMC. In 22/50 patients, 18 recurrent mutations unique to MM were identified in intron 3; many were also recurrent in our previous study2. Individual mutations recurred in 2–7 patients. Among these, 17/18 recurrent mutations increased the G-C content of intron 3 and 6/18 created or disrupted G runs in intron 3. This supports the idea that in MM patients, cumulative variations in introns 3 and 4 alter splice site selection, operationally resulting in loss of HAS1Vd and enhanced expression of the clinically relevant HAS1Vb variant. We speculate that individuals who accumulate genetic variations in introns 3 and 4 of HAS1 are predisposed to aberrant splicing of HAS1 which may contribute to development of malignancy in MM and WM. Disclosures: Belch: Celgene: Research Funding; Onyx: Research Funding.

Abstract 4045 Multiple Myeloma (MM) is characterized by accumulations of chromosomal abnormalities throughout the course of disease. Recurrent IgH translocations such as t(4;14) or deletion (del)13 occur during early stages of disease or perhaps as originating events. Other abnormalities appear to be acquired late in the disease process, after development of frank MM, including p53 deletion (Ch17) and amplification of CKS-1B (Ch1). We have previously identified drug resistant clonotypic B lymphocytes in MM that have malignant characteristics and can xenograft MM to immunodeficient mice. Using an automatic scanning system we scanned bone marrow (BM) cytospin slides stained with May-Grünwald Giemsa to identify lymphocytes and to determine whether or not they share the same chromosomal abnormalities that are found in autologous plasma cells (PC). For 200 MM patients, we performed interphase FISH using commercial probes, to detect deletion of chromosome 13, any translocations in 14q32, t(4;14)(p16;q32),t(11;14)(q13;q32), deletion of p53 and a custom probe to detect amplification of CKS-1B. For patients harboring a given chromosomal abnormality in their PC, we found the identical abnormalities in BM lymphocytes for 35% of patients having del13, 30% having t(4;14) and 23% having t(11;14). Likewise we detected amplification of locus 4p16, 11q13, 14q32, trisomy 1 and trisomy 17 in 45%, 15%, 30%, 37.5% and 44% respectively in the lymphocytes of patients demonstrating those same abnormalities in the PC. In a more limited series of experiments using immuno-FISH, we observed that BM lymphocytes with chromosomal abnormalities expressed CD20. In contrast, amplification of CKS-1B or p53deletion were undetectable in lymphocytes from patients having these abnormalities in their PC. We are currently analyzing the impact of abnormalities present in BM lymphocytes with regards to the clinical parameters and survival. An update of these results will be given at the meeting. Given the compelling evidence that B cells comprise a clinically important compartment of the MM clone, this work suggests that fundamental genetic events shared by all compartments of the MM clone include recurrent IgH translocations and Ch13 deletions. However, further clonal expansion occurring during disease progression and after relapse may be restricted to the PC population, characterized by acquisition of p53 deletions and CSK1B amplification. Furthermore in overt MM, clonal expansion may simultaneously occur on at least two levels, firstly that of the B cells which persistently give rise to PC, and secondly that of PC themselves as they continue clonal expansion in the absence of contributions from the B cells. Abnormalities such as del13 or IgH translocations characterize early events in the disease while p53del or amplification of CKS-1B may occur as progression events restricted to PC. Disclosures: No relevant conflicts of interest to declare.

Abstract 2981 Typically MM involves a single monoclonal protein, predicting a single CDR3 peak. We have shown that a sizeable number of patients do not fit this pattern. RT-PCR and genomic PCR (gPCR) were performed using universal primers that bind to FR3/JHc or FR3/CH1 to identify the monoclonal CDR3. In a cohort of 49 MM patients, 10 (20%) exhibited two CDR3 peaks, one of the clinically identified MM clone and another “second” clone. With one exception, a biclonal gammopathy, these second clones were not detected serologically by serum protein electrophoresis or immunofixation. Molecular biclonality is categorized into 3 groups based on the isotype of second clones: group I (1 IgD and 4 IgG MM) where second clone has a pre-switch isotype; group II (3 IgG MM) where the second clone has a post-switch isotype and group III (2 IgA MM) where the second clone exhibits ongoing class switch recombination. All MM clones undergo somatic hypermutation (2.44-15.03%, median = 8.55%). For second clones, 2/10 are unmutated (0-1.05%, below 2% cut off) and 8/10 are mutated (2.44-13.54%, median = 5.90%). The VH3 gene family was utilized in 6/10 of MM clones and 5/10 of second clones. For 10/10 patients, the clinical MM clones are unrelated to second clones as shown by different combinatorial VDJ, different length of CDR3 and different VDJ junction. Biclonal distribution was compared by grading of CDR3 peak heights generated by gPCR. For 9/10 patients, the MM clone is most frequent in BM, 3 of which also predominated in blood. For 6/10 patients, second clones predominated in blood. Q-PCR of mononuclear cells (MNC) showed that second clones ranged from 0.06–24% (median = 1%, n=7) in BM and 0.1–31% (median = 1.45%, n=6) in blood. Chromosomal abnormalities in plasma cells (PC) were seen in 8/8 patients analysed; among these, 4 had t(11;14). Other genetic abnormalities included deletion 13q14, amplification 1q21, trisomy and tetrasomy of chromosome 1. For 4/5 patients where second clones could be distinguished from the MM clone by IgH immuno-FISH, chromosomal abnormalities were detected in the second clone, which in some cases were different from those in the MM clone. Longitudinal analysis was studied in 3 patients. For case 1, a biclonal gammopathy, the IgGλ clone exhibited deletion 13q14 and amplification 1q21, and was dominant in a sternal biopsy (IgGλ/IgGκ = 3.6%/ Disclosures: No relevant conflicts of interest to declare.

Abstract 783 Malignant plasma cells from patients with multiple myeloma (MM) display recurrent chromosome translocations involving the immunoglobulin heavy chain locus (IGH) and CCND1 [t(11;14)] or FGFR [t(4;14)]. It is poorly understood why recombination events between chromosomes recur at specific breakpoints in the human genome, but spatial proximity of translocation-prone gene loci (TPGL) is influential. Utilizing 3D fluorescent in situ hybridization and 3D analysis techniques, we have measured the spatial and radial positioning of translocation-prone and control loci in the nuclei of non-malignant cells to determine a propensity to translocate. Purified CD34+ progenitors and CD19+ B cells were analyzed; none had translocations. Utilizing these cell subsets (n=900 cells) from MM patients having recurrent IGH translocations in autologous plasma cells, and comparable subsets (n=300 cells) from healthy donors, we show that spatial genome organization contributes to the formation of recurrent translocations. We show that IGH, CCND1, and FGFR3 are preferentially positioned in close proximity relative to each other in presumptively normal CD34+ progenitors and CD19+ B-cells from these MM patients, and that the clinical frequency of IGH translocations correlates with relative positioning (p=0.002 for CD34+ cells; p=0.017 for CD19+ cells), and with radial positioning (p=0.05 for CD34+ cells; p=0.043 for CD19+ cells). We also show that non-malignant CD19+ B-cells from MM patients display cancer-specific radial positioning of TPGL in the nucleus (p≤0.008), suggesting a predisposition to translocate. In addition to preferred positions within the nucleus, genes also localize to favoured locations within their respective chromosome territory (CT). Genes have been shown to position outside of their respective chromosome territory. Although proximity of potential translocation partners is necessary for translocation events to occur, others have shown that translocation events between adjacent chromosomes take place at CT boundaries, or in the space of intermingling between adjacent CTs. Active transcription ‘factories' are also present within the area of intermingling. We observed that at least one of the two alleles of CCND1 and FGFR3, is positioned outside of its chromosome territory in 59% of CD19+ cells from MM patients, whereas a control locus TGFBR2, is positioned outside its chromosome territory in only 25% of the same cells. We find myeloma-specific positioning of TPGL in a subset of B-cells that lack translocations and are predominantly polyclonal, defining them as non-malignant. It seems likely that the original parent B-cell that gave rise to MM also harboured spatially proximal TPGL prior to the formation of physically translocated loci. Together, these observations support the likelihood that translocations occur only in specific sites within the nucleus. Recent reports indicate that ongoing transcription is necessary for functional activity of IGH recombination enzymes, and that these same enzymes have off-target effects on proximal genes. IGH and its translocation partners FGFR3 and CCND1 may come together briefly to a transcription ‘factory' outside of their respective CTs and be acted upon by IGH recombination enzymes. We speculate that cancer-specific positioning of active TPGL in B-cells from MM patients promotes the formation of clinically important IGH translocations, perhaps through co-localization of TPGL to nuclear transcription ‘factories' outside of CTs. The strong correlation between locus positioning and the clinical frequency of recurrent IGH translocations involving these loci is consistent with the idea that spatial proximity in the nucleus is an important contributor to the high frequency of recurrent translocations. CCND1 and FGFR3 are positioned outside their CT and in close proximity to IGH as compared to TGFBR2 and c-MAF, both of which are more peripheral in the nucleus and position out of their respective CTs at lower frequency. This is consistent with the fact that c-MAF/IGH translocations are relatively infrequent and TGFBR2/IGH translocations have not been reported. Our results suggest that the formation of recurrent IGH translocations in MM is determined not only by spatial proximity of gene loci and radial positioning in the cell nucleus, but also by positioning of loci outside of their chromosomal territories. Disclosures: No relevant conflicts of interest to declare.

Abstract 2842 Poster Board II-818 In addition to full length (FL) transcripts, clinically significant HAS 1 splice variants (Va, Vb and Vc) have been previously identified in multiple myeloma (MM) and Waldenstrom's macroglobulinemia. Increased HAS1Vb expression correlates with poor survival in a cohort of MM patients. Here, we show that directed mutation of HAS1 intron 3 alters HAS1 splicing and generates a pattern of HAS1 variant expression that mimics patterns detected in MM patients. This suggests that hypermutation of HAS1 and consequent expression of HAS1 splice variants may contribute to oncogenesis in MM. HAS1FL comprises of 5 exons (2089 bp); Va skips exon 4 (133 bp); Vb skips exon 4 and partially retains 59 bp of intron upstream of exon 5 (+59); Vc has all 5 exons and partially retains 26 bp of intron downstream of exon 4. In MM, frequent intronic mutations have been observed in introns 3 and 4, suggesting possible contributions to HAS1 alternative splicing. We have utilized a mammalian expression system to analyse HAS1 splicing by fusing a minigene extending from exon 3 to exon 5 (g345) with the upstream cDNA sequences. HAS1 expression is determined by transfection and RT-PCR using appropriated primer sets. This study focuses on identification of intronic mutations that may affect HAS1 splicing. We target mutations on (A/U)GGG motif because of its high abundance in HAS1 intron 3. The (A/U)GGG repeat was also shown to enhance the splicing of alternative intron in chicken β-tropomyosin (Sirand-Pugnet, P, et al, NAR, 1995, 23, 3501) and intronic G runs could work in a combinatorial way to control the selection of the proper 3' splice site in human thrombopoietin (Marcucci, R, et al, NAR, 2006, 35, 132). A 580 bp long human HAS1 intron 3 is GC-rich and comprises of 28 (A/U)GGG motifs (sequentially identified as G1, G2.., G28). HeLa cells transfected with an unmutated intron 3 construct mainly produce FL with a small amount of HAS1Va, a profile that is similar to CD40L/IL-4 activated normal B cells. Site directed mutagenesis of all 28 (A/U)GGG motifs (G1-28) abolished FL expression, but not HAS1Va, suggesting that these sequence alterations are highly unfavorable for constitutive splicing. It may be due to the loss of essential cis-acting element(s) and/or undesirable conformational changes that prevent spliceosome formation. Mutagenesis of G1-18 is shown to eliminate constitutive expression by increasing the usage of multiple alternative donor sites. Mutagenesis of G19-28 produces more HAS1Va than FL, presumably due to increased exon 4 skipping events. An increased Va/FL ratio could also be achieved by mutagenesis of G25-28 or G27-28, suggesting that this subregion is important for pathway selection. Mutagenesis was also studied in del1 construct, a unique derivative of HAS1 minigene that partially deletes intron 4. Similar to g345, del1 produces FL and HAS1Va as well as promotes expression of novel HAS1Vd, an isoform that includes +59 bp (like Vb) and exon 4. Alteration of HAS1 splicing profile caused by mutagenesis shown in g345 series is also observed in del1 series. Additionally, there is a shift from Vd to Vb expression in all constructs analysed (del1/G1-28, del1/G1-18, del1/G19-28, del1/G25-28 and del1/G27-28), a pattern of aberrant splicing that found in MM patients. Thus, in del1, increased exon 4 skipping events promote both Va and Vb expression. Sequencing of HAS1 intron 3 in a cohort of 50 MM patients indicates that recurrent mutations are found in the G repeat regions and that new repeats are generated by recurrent MM-specific HAS1 mutations. This suggests that mutation of the HAS1 construct mimics HAS1 mutation events that occur in MM patients themselves, and contributes to the clinically significant aberrant HAS1 splicing we have reported in MM (Adamia et al. Blood, 2008, 112, 5111; Blood, 2005, 105, 4836). Overall, critical mutations that could alter HAS1 expression and the ratio of HAS1 variants to FL were identified in intron 3. In intron 4, critical mutations that increase the usage of alternative splice site (+59) remain to be studied. We speculate that cumulative mutations within these two intronic sequences could bring the two events together to promote HAS1Vb splicing. While trans-acting elements are likely to regulate RNA splicing and its pathway, our studies clearly suggest that intronic mutations play an important role in the aberrant splicing of human HAS1, with probable contributions to disease progression in MM. Disclosures: No relevant conflicts of interest to declare.

Abstract 2841 Poster Board II-817 Introduction: Monoclonal gammopathy of undetermined significance (MGUS) is the most common plasma cell disorder. A small proportion of MGUS transform to multiple myeloma (MM) or other B lineage malignancies. Hyaluronan synthase I (HAS1) is overexpressed in many human cancers, including MM. Here we show that intron 3, a region involved in aberrant splicing of HAS1, is heavily mutated in 34 patients with MGUS. Aberrant splicing of HAS1 correlated with poor outcome in a cohort of MM patients. HAS1 is GC-rich and as such is likely to be subject to mutation events. We recently reported a series of nearly 200 HAS1 mutations in 17 MM and Waldenstrom's macroglobulinemia (WM) patients, that were absent from healthy donors (Adamia et al, Blood, 2008, 112:5111). Forty-nine of these HAS1 mutations were recurrent, defined as being present in more than one patient. None of them were present in sequenced regions of HAS1 from a small series of MGUS patients. Mutated HAS1 DNA directs aberrant splicing of HAS1 and in vitro mutagenesis of HAS1 intron 3 alters HAS1 splicing patterns (see abstract by Kriangkum et al.), mimicking the aberrant splicing seen in MM patients. To sequence HAS1 from a larger cohort, we developed a streamlined sequencing strategy that allowed rapid sequencing of exon 3 and intron 3, regions of the HAS1 gene rich in mutations. Our objective was to identify HAS1 mutations in MGUS as compared to a larger cohort of MM and healthy donors. Many reports have shown familial clusters of MGUS and MM and of other B cell malignancies involving MGUS, supporting the idea that genetic factors, and possibly genetic predispositions, contribute to the development of disease. Patients and Methods: We sequenced HAS1 regions from 34 MGUS patients, 50 MM patients, and 45 healthy donors. In all cases, we sequenced an 850 bp region covering exon 3 and intron 3 of the HAS1 gene. For each patient, we sequenced eight HAS1 region subclones, in both directions. To predict impact of recurrent HAS1 mutations on the splicing process, we performed in silico analysis using online bioinformatic tools. Results: To date we have identified, 117 HAS1 mutations in HAS1 gene in MGUS patients including 5 frequently detected NCBI single nucleotide polymorphisms (SNPs). Although the same NCBI SNPs were also identified in healthy donors, increased homozygosity was detected in MGUS patients compared to healthy donors. Overall, among the novel HAS1 mutations, 32 (27%) were shared between MGUS and MM but absent in healthy donors, defining them as recurrent. Seven of the 32 mutations (23%) shared between MGUS and MM were recurrent in 2-3 individual MGUS patients; three of the recurrent MGUS mutations were missense mutations present in the coding region of exon 3, changing the protein sequence. Our initial bioinformatic analysis showed that the two of the shared recurrent mutations were found in the SR protein binding sites of the pre-mRNA and may contribute to aberrant splicing of the primary HAS1 transcript. We predict that our ongoing analysis HAS1 mutations in MGUS patients will identify a subpopulation(s) at high risk of aberrant splicing and thus at increased risk of transformation to overt MM. Conclusion: HAS1 appears to be hypermutated, perhaps due to its GC-rich nature. HAS1 mutations were identified in patients with a clinical diagnosis of MGUS (range of 1-14 mutations per MGUS). Thirty-two HAS1 mutations were recurrent in MGUS and MM but were absent from healthy donors. Some of these mutations are predicted to have functional significance for protein sequence and pre-mRNA splicing, possibly leading to increased risk for or predisposition to disease. Although genetic background and environmental context are likely to be important factors, our report of novel mutations in HAS1 of MM and WM patients, confirmed now in MGUS, suggests that mutations in HAS1 may play a key role in oncogenesis. Further genetic analysis of patient populations and the biochemical characterization of the mutated proteins will provide considerable insight into the function of HAS1 during transformation from MGUS to overt MM. Disclosures: No relevant conflicts of interest to declare.

Abstract 2822 Poster Board II-798 Background: Oct-2 and Bob1 are B-lymphoid lineage specific transcription factors which drive the expression of genes involved in B-lymphocyte and plasma cell development and differentiation, including the immunoglobulin heavy chain (IgH) genes and many others. It is also thought that these transcription factors drive the overexpression of oncogenes involved in IgH translocations through upregulation of the IgH enhancers. We sought to evaluate the expression of Oct-2 and Bob1 in multiple myeloma, a malignancy of the B cell/plasmacyte lineage, in order to determine their potential biological significance as determinants of disease behaviour/prognosis and as potential therapeutic targets. Methods: A tissue microarray containing triplicate samples from 220 myeloma patient samples was generated. Samples were collected at the time of diagnosis between 1993 and 2008. To measure Oct-2 and Bob-1 protein expression in the CD138+ plasma cells of each sample, fluorescent immunohistochemistry was used to stain the samples and the HistoRx PM-2000 platformTM was used to generate an AQUA score for each patient. Oct-2 AQUA scores were available for 169 patient samples and Bob-1 AQUA scores were available for 169 patient samples. The analysis was repeated using conventional immunohistochemistry and visual scoring of staining intensity. Results: The Oct-2 and Bob-1 AQUA scores were continuously distributed over the patient cohort; thus, subsequent analyses were done by dividing the patient cohort by the median Oct-2 or Bob-1 AQUA score and comparing the group with high AQUA scores to the group with low AQUA scores. The patient group with high Oct-2 AQUA scores had an OS of 52.8 months compared to 73.1 months in the group with low Oct-2 AQUA scores (hazard ratio [HR]=1.74, 95% confidence interval [CI]=1.11-2.73, p=0.0164). The patient group with high Bob-1 AQUA scores had an OS of 70.5 months compared with 38.3 months in the group with low Bob-1 AQUA scores (HR=0.46, 95% CI=0.29-0.71, p=0.0008). In a Cox regression proportional hazards model including autologous stem cell transplant (ASCT) status, serum albumin, serum β2-microglobulin and Oct-2 or Bob-1 AQUA scores, only ASCT status was independently predictive of OS. The patient group with low Oct-2 AQUA scores had a significantly higher proportion of ASCT recipients than the group with high Oct-2 AQUA scores. Conversely, the patient group with high Bob-1 AQUA scores had a significantly higher proportion of ASCT recipients than the group with low Bob-1 AQUA scores. Results were similar but not statistically significant when analyzing the TMA with conventional immunohistochemistry and visual scoring of staining intensity. Discussion: In this retrospective study high Oct-2 expression was found to be associated with decreased OS, while high Bob-1 protein expression was associated with increased OS. These results suggest that Oct-2 is a negative prognostic factor in myeloma, while Bob-1 is a favorable prognostic factor, and demonstrates the applicability of the HistoRx platform to myeloma bone marrow TMA analysis. Since this was a small, exploratory study these results must be interpreted with caution and must be confirmed in a larger, prospective analysis. Disclosures: No relevant conflicts of interest to declare.

Abstract 1813 Poster Board I-839 Malignant clones in multiple myeloma (MM) express unique clonotypic VDJ sequences associated with the clinical isotype. It has been widely assumed that MM patients have only one clone. However, our analysis reveals that a high frequency of MM patients have two clones. Analysis of complementary determining region 3 (CDR3) in bone marrow (BM) by DNA fragment analysis using florescent labeled consensus primers that bind to the framework region 3 and the joining region (FR3/JHc) or the framework region 3 and the constant region (FR3/CH1) related to the clinical isotype typically yields a single CDR3 peak, corresponding to the clonotypic VDJ gene segments. Signature sequences identified from CDR2 and CDR3 of clonotypic VDJ provides a molecular tool to monitor the malignant clone within a given patient. Study of CDR3 in BM and peripheral blood (PB) provides valuable information about the IgH VDJ repertoire and Ig expression profile, especially in cases where two clonal sequences are molecularly identified or when validation by single cell analysis did not show high frequency for a presumptive clonotypic IgH VDJ. Here, we analysed CDR3 expression profile in 23 MM patients. The results indicated a high incidence of MM having a second clone (6/23 patients, 26%). Reverse transcription-polymerase chain reaction (RT-PCR) of FR3/JHc and FR3/CH1 reveals a CDR3 peak representing the major MM clone in all 23 BM analysed. Among 23 patients, 6 (4 IgG, 2 IgA) exhibit a second monoclonal peak which can be detected both in PB and BM. Fluorescent in situ hybridization was unable to detect an increased incidence of IgH translocations within these 6 patients. Characterization of VDJ sequences indicated that the second clone mainly utilizes VH3 gene family (5/6), undergoes somatic hypermutation (5/6 at 2% cutoff) and does not derive from MM clone (6/6) as shown by different combinatorial and junctional diversity. Repertoire analysis determined by FR3/JHc amplification of genomic DNA excluded biallelic expression of the MM clone. In patient MM1, the second clone (μ) predominates in PB while the MM clone (γ) predominates in BM. In MM2, the second clone has a very short CDR3 (18 nt), predominates in blood compartment and has undergone isotype switching (μ, γ, á). In MM3, the M protein is of IgG type, consistent with transcripts in an FR3/Cγ RT-PCR peak in BM, but repertoire analysis of genomic DNA showed that the majority of B lineage cells in bone marrow and blood have a rearranged μ isotype. In MM4, having a homogeneous MM (á) clone, the second clone (á) exhibited intraclonal diversity, as defined by sequencing of VDJ subclones. In MM5, clonal transcripts of both dominant γ clone and the second clone (á) persisted in blood following initial chemotherapy. In MM6, the MM clone (γ) predominates over the second clone (á) both in BM and PB. Overall, our studies indicate a high incidence of biclonal expansion in MM. The unrelated VDJ origin of sister clones detected in MM patients exhibiting biclonality suggests that they may have arisen from different transformation event, which may be promoted when cumulative genetic abnormalities reach a critical threshold. Although clinical manifestations in a patient may reflect a take-over by the most aggressive clone, our studies suggest that transformation may be an ongoing event that occurs at a heretofore underestimated rate in MM B lineage cells. Disclosures No relevant conflicts of interest to declare.

Abstract 603 Background: Acquired drug resistance often limits treatment efficacy in multiple myeloma, and may in part be caused by selection for cells bearing mutations conferring drug resistance that were either present in an extant subset of myeloma cells prior to treatment, or were acquired during therapy. We are applying whole transcriptome shotgun sequencing (WTSS) to myeloma BM at diagnosis and drug-resistant relapse, in order to identify gene expression and mutational signatures of drug resistance. We have an archive of frozen myeloma total bone marrow mononuclear cell fractions, as well as a smaller archive of CD138+ enriched specimens. The enriched specimens can be expected to contain purer populations of malignant cells, but may under-represent the heterogeneity of the disease. We are therefore interested in a comparison of the results of WTSS on these two different types of specimens. Methods: Bone marrow from a myeloma patient who responded to front line drug therapy and who subsequently experienced drug resistant disease relapse was identified from our biorepository. RNA was isolated from the total bone marrow mononuclear cell population and from a CD138+ enriched population at diagnosis and at relapse, yielding four samples for analysis (pre-treatment unprocessed, pre-treatment enriched, relapse unprocessed, relapse enriched). Poly-adenylated RNA from each of the four samples was converted to cDNA using a random primed approach, and paired-end sequences collected using Illumina instruments. Results: We sequenced the transcriptomes of the four samples by generating a total of 694,076,590 36- and 50-mer sequences (30,681,959,896 base-pairs). Of these, 325,998,000 (14,063,396,516 base-pairs) were mapped to reference genome or transcriptome databases using MAQ. Each base in 3,094 transcripts was represented at a minimum of 6 fold redundant coverage. 68.57 % (223,555,599) of the reads matched perfectly to reference sequences. Reads containing single nucleotide mismatches to the reference sequence were evaluated using a Bayesian mixture model to estimate the presence of single nucleotide variants (SNVs) which yielded 89,944 high confidence predictions divided amongst the four samples. Of these, 74,704 (83.1%) were identical to known SNVs present in dbSNP and other sequence datasets, and 13,073 represented non-coding or synonymous variants. The remaining 2,167 mismatches most likely represent undiscovered variants also present in the patient's normal DNA or somatic mutations. These two possibilities will be distinguished by sequencing the candidate variants using DNA purified from the patient's non-malignant cells. For candidate sequence variant positions well represented at the position and gene level in each of the two sample pairs we were comparing (pre- and post-treatment unprocessed, and pre- and post-treatment enriched) we found 85 unique to the pre-treatment enriched sample, 145 unique to the post-treatment enriched sample, 250 unique to the pre-treatment unprocessed sample and 51 unique tunique o the post-treatment unprocessed sample. Some of the genes affected by these candidate mutations include genes that have previously been observed in cancer, such as PIM2, SHC1 and SCRIB. Gene expression levels were estimated using the previously published RPKM (reads per kilobase of exon model per million mapped reads) metric. For the unprocessed samples, 108 genes were found to be up-regulated pre-treatment and 95 post-treatment. The enriched samples presented 80 genes with higher expression before treatment and 54 after treatment. Because we are using a paired end sequencing approach, the potential exists to identify candidate expressed gene fusions. Initial searches for candidate gene fusion events have resulted in the detection of a fusion involving IgH and MMSET, corresponding to a t(4;14) translocation which had previously been detected with standard methods. Discussion: It is possible to sequence and compare the expressed myeloma genome of patients at diagnosis and at relapse with WTSS, in order to gain insight into the evolution of myeloma under drug selection and its role in the development of treatment resistance. With the study of a larger cohort of patients and subsequent functional characterization of the candidate sequence variants, we hope to identify recurrent mechanisms of myeloma drug resistance which might be amenable to therapeutic intervention. Disclosures: No relevant conflicts of interest to declare.

Despite improvement in therapeutic regimes which have extended survival of patients with multiple myeloma, the search for novel targets to further advance treatment options continues. It is critical to the success of targeted cancer therapy that the rational selection of drugs be derived from the molecular and cellular biology of protein targets. Furthermore, studies have indicated that focused combination drug therapy could produce remarkable improvement in patient outcomes. Aurora kinases A/B/C (AURK), components of the centrosome and important for cell cycle function, are promising new targets in the treatment of numerous types of cancer. We have previously validated AURKs as a therapeutic target in pre-clinical models of multiple myeloma. In this study our aim is to better define the biology of AURKs in both normal and malignant cells, with the goal of identifying rational strategies for improving the therapeutic potential of Aurora kinase inhibitors (AKIs) as anti-cancer agents. We use a systematic two-pronged approach to map post-translational modifications and novel protein-protein interactions that define the interactome of AURKs. We also examine whether Aurora kinase interactions and posttranslational modifications are significantly different in cancer cells compared to normal cells. Increased understanding of Aurora kinase biology will help develop mechanisms to intensify the anti-cancer effects of AKIs. SUMOylation is a recently discovered posttranslational modification of numerous proteins that has been linked to many cancer types. A small ubiquitin-like modifier (SUMO) group is conjugated at the canonical ubiquitin site of a protein to alter its function, stability or sub-cellular localization. We believe that Aurora kinases may be differentially SUMOylated in cancer cells compared to normal cells, and that SUMOylation of Aurora kinases in cancer cells may alter their function and their susceptibility to AKI therapy. Here we have determined putative SUMOylation sites of all three AURKs in silico and also have evidence of AURK SUMOylation using an in vitro assay and also in human HEK293 cells. Furthermore, the gene encoding the central SUMO conjugating enzyme - UBC9 - is expressed at higher levels in CD138+ malignant plasma cells compared to other CD138−bone marrow cells in myeloma patients. As part of our strategy, we have also used Tandem Affinity Purification (TAP) technology to identify novel protein-protein interactions of Aurora kinases. In the initial stage, we have established stable expressing clones in a human cancer cell line and a normal line that express TAP-tagged AURKA/B or C. In pilot experiments we have successfully isolated pure recombinant AURKA along with interacting proteins using the TAP method. Using mass spectrometry (MS/MS) protein ID, we have identified putative protein interaction partners with AURKA which have implications in general cancer biology and centrosomal structure. In the future, these interactions will be validated using conventional biochemical methods before being tested in myeloma cell lines.

BACKGROUND: The centrosome is the cellular organelle that nucleates the mitotic spindle. Polo-like kinase 1 (Plk1), a centrosome-associated serine/threonine kinase, serves as a key regulator of multiple steps in mitosis. Plk1 is overexpressed in a broad spectrum of tumor types, and its expression often correlates with poor patient prognosis. We have previously observed centrosome amplification in myeloma, and have validated other components of the myeloma centrosome and mitotic apparatus as therapeutic targets. The present study explored the expression of Plk1 in myeloma and the effect of BI2536, a potent and selective inhibitor of Plk1, on myeloma cells in the pre-clinical setting. METHODS: Plk1 expression was assayed in a panel of myeloma cell lines (MMCL) and AutoMACS-purified CD138+ patient bone marrow plasma cells (BMPC). The anti-myeloma effects of BI2536, alone or in combination with conventional agents, were assayed on myeloma cells with proliferation (MTS) and apoptosis (Annexin V/propidium iodide) assays. The phenotype of treated cells was examined with DNA content analysis and immunofluorescence microscopy. The efficacy of BI2536 monotherapy was evaluated in NOD/SCID mice bearing RPMI8226 xenografts. RESULTS: Plk1 is ubiquitously expressed in myeloma to varying degrees in both MMCL and BMPC. BI2536 inhibited the proliferation of MMCL (RPMI 8226, U266, LP-1 and KMS-11) and patient BMPC at nanomolar concentrations. The addition of BI2536 was able to overcome resistance to dexamethasone. Bortezomib in combination with BI2536 had significantly increased anti-myeloma effects compared to the use of either agent alone. BI2536-treated MMCL accumulated 4N DNA content prior to undergoing apoptosis. The phenotype of BI2536-treated cells is consistent with inhibition of Plk1, showing prometaphase arrest and monopolar mitotic spindles in a dose-dependent fashion. BI2536 induces regression of human myeloma xenografts in NOD/SCID mice. Taken together, BI2536 is a promising new agent for the treatment of multiple myeloma. This work provides further evidence that Plk1 and the amplified myeloma centrosome are targets for therapy.

Multiple myeloma (MM) is a clonal B-cell malignancy with an accumulation of plasma cells in the bone marrow (BM). Newer treatments continue to increase remission occurrence and decrease side effects, however, relapse invariably occurs. Current diagnostic and monitoring criteria such as bone marrow plasmacytosis (BMPC) and serum M-protein levels (Mpr) fail to predict the length of remission for a given patient. The clonal IgH VDJ gene rearrangement provides a molecular method of measuring the MM disease burden. This unique biomarker can identify and quantify the number of malignant cells through analysis of genomic DNA with quantitative PCR. Here we use SYBR Green real-time quantitative PCR (rqPCR) to measure the level of clonal cells, termed VDJ%, in 91 previously untreated and 51 relapsed pre-treatment BM aspirates from MM patients. Kaplan-Meier analysis showed that VDJ%s lower than the median were associated with longer event free survivals (EFSs) (p=0.0077, HR=0.57). These patients were treated with a variety of different regimens (including VAD, ASCT, Melphalan, Prednisone, Thalidomide, Dexamethasone, Velcade or Revlimid). A modest correlation between the VDJ% and the BMPC was observed (p=0.0003, r2=0.099) but, the BMPC was not able to stratify these patients into groups with significantly different EFSs (p=0.1153, HR=0.6521) as was the VDJ%. This result confirms the BMPC’s poor association with outcome. The Mpr was not significantly correlated with VDJ% (p=2312, r2=0.012). The IMWG response subcategories correlated well with EFS where the poorer the response the shorter the EFS. However, there was no significant correlation between the pre-treatment VDJ% and the response subcategories. In addition to the pre-treatment BM samples, the tumor burden in 30 remission BM samples was quantified. To date, neither the remission VDJ% nor the reduction in VDJ% relative to the pre-treatment VDJ% was associated with EFS in a Kaplan-Meier analysis. The remission VDJ% had a significant yet modest correlation with the remission BMPC (p=0.014, r2=0.27) and nadir Mpr (p=0.0001, r2=0.61). Additionally, the IMWG response subcategory correlated with the proportionate reduction of the remission BM (p=0.036). Taken together these results indicate that VDJ% in pre-treatment BMs, but not the matched BMPC or Mpr, is a significant measure of outcome. The percent of clonal cells as measured in BM aspirates with rqPCR could be developed as a novel tool for predicting outcome by placing patients into a high risk or low risk category for EFS. Patients with a higher than median VDJ% have a shorter EFS than patients with a VDJ% below the median. RqPCR enumerates all cells containing the clonal sequence and thus may include relevant cells types not counted in the BMPC or reflected in the protein electrophoresis and not proportionally represented in remission BMs. The predictive value of pre-treatment VDJ% occurred despite varying types of therapy and tumor response, suggesting that these factors might provide independent prognostic information.

Bone marrow (BM), a site of hematopoiesis, is a multicellular tissue with a complex architecture. Multiple myeloma (MM) is an incurable plasma cell malignancy where even patients in remission succumb to an inevitable relapse. While considerable progress has been made towards understanding and treating MM, to date, there is no culture system which can recapitulate the complex interactions within the BM microenvironment. Current failure to grow the MM clone within the context of human microenvironment hampers progress into the understanding of the biology of MM and design of biologically relevant therapies. Here we present an in vitro three-dimensional (3-D) tissue culture model which recapitulates the human BM microenvironment allowing for the growth and expansion of the MM clone. Cells from the BM aspirates are grown in a fibronectin, laminin and collagen rich ECM designed to reconstruct in vitro endosteum and central marrow, mimicking the in vivo microenvironment of the BM. Proliferation and redistribution of cells within reconstructed ECM results in stratification of the culture, mimicking the in vivo condition where cells occupy individual niches. Cellular composition of the culture is maintained in accordance with the proliferation properties of the BM where osteoblasts, osteoclasts, adipocytes and stromal cells differentiate along with the full complement of the hematopoietic cells. BM cultures from normal donors are well-organized with osteoclasts and hematopoietic cells occupying distinct positions in the ECM. In contrast, reconstructed BM from MM patients is disorganized in 3-D where osteoclasts intermingle with the hematopoietic compartment. The MM malignant clone is expanded in 3-D cultures as measured by real-time quantitative PCR (rqPCR) for genomic clonotypic VDJ sequences. Malignant B and plasma cells proliferate in these cultures and FISH analysis reveals that their progeny harbor chromosomal abnormalities identical to those that mark the malignant clone prior to culture. Preclinical testing of emerging therapeutics targeted for multiple myeloma is hindered by the failure of the current models to sustain growth of the myeloma clone. In the 3-D culture, myeloma clone expands within its native environment providing an ideal preclinical model where conventional (Melphalan) and novel (Velcade) therapeutics efficiently and selectively kill their target cells. In the 3-D BM culture model, non-proliferating, label retaining cells (LRC) concentrate at a putative endosteum-marrow junction, where hematopoietic stem cells have been shown to localize in vivo, suggesting that the drug-resistant myeloma stem cells localize to the endosteal niche. In a colony-forming assay, drug-resistant LRC purified from the 3-D cultures form clonal colonies composed of malignant cells with patient specific clonotypic VDJ sequences. Recapitulation of the BM architecture in vitro is a first step towards the identification and therapeutic targeting of the elusive myeloma stem cell.

High dose melphalan is an essential component in the treatment of patients with advanced multiple myeloma. However few data are available regarding genetic polymorphisms possibly correlated with patient outcome or toxicity in this setting. To identify polymorphisms predictive of survival and/or toxicity we performed a retrospective single nucleotide polymorphism (SNP) analysis using APEX technology in 169 patients having received high dose melphalan for multiple myeloma. All patients had received one or two courses of high dose melphalan with stem cell reinfusion. DNA was obtained from peripheral blood mononuclear cells and analyzed for 230 SNPs in genes selected for their role in drug metabolism, DNA repair, cell cycling or apoptosis. There were 108 males and 61 females and the median age was 55 years. The average number of VAD courses received before high dose melphalan was 4 (3–6). 58 patients (34%) received two high dose melphalan courses. A complete/very good response was observed in 43 of 103 evaluable patients (42%). Forty patients among 107 evaluable (37%) experienced high grade (III–IV) mucositis after high dose melphalan. The median freedom from progression was 45 months and the overall survival was 91 months. SNPs in SULT1A1, MDR, ERCC5, GSTP1, NQO1, CYP1A2, CCND1, PARP, RAD51, DRD4, P53 or associated proteins were found to be correlated with response to VAD therapy. SNPs in GRPR, ERCC5, GSTM3, GSTP1, XRCC1, ERCC4, CYP1B1, BARD1, MDR1, P53BP1, TERT, DRD4 were found to be correlated with response to high dose melphalan. Polymorphisms in CYP2E1, CYP2C9, CYP1A1, CYP1B1, TMPT, OGG1, PCNA, DRD2, XPC were correlated with disease progression while polymorphisms in ERCC1, DRD4, CYP2D6, CYP2C9, CYP1A1, CYP1A2, CYP1B1, ALDH, BARD1, BRCA1, MYH, TP53BP1 were correlated with death. Polymorphisms in XRCC1, SNP103, CYP2C9, CYP2C19, BRCA1, BRCA2, NBS1, CDKN1A were correlated with severe mucositis. These results suggest that SNPs of genes involved in drug metabolism or repair could be used to distinguish patients subgroups with different toxicity/efficacy profiles after high dose melphalan.

Hyaluronan synthase 1 splice variants (HAS1Va, Vb and Vc) have been previously identified in multiple myeloma (MM) and Waldentsrom’s macroglobulinemia (WM) in addition to HAS1 full length (FL) transcript, and overexpression of HAS1Vb correlates with poor survival in MM (Adamia, et al, Blood2005; 105(12): 4836–44). Unlike HAS1FL that has 5 exons, the aberrantly spliced HAS1Vb transcript excludes exon 4, but retains the last 59 bp of intron 4. Sequence analysis of HAS1 intron 4 (2.1 kb) in MM and WM, but not in B-CLL, reveals frequent point mutations and other genetic variations in genomic HAS1 from bone marrow and blood (Adamia et al, submitted). Intronic mutations have been shown to regulate and modulate RNA splicing. Although it is likely that the mutations identified in HAS1 contribute to aberrant splicing events, an in vitro splicing assay was established to confirm that genomic HAS1 from patient DNA directed the aberrant splicing of HAS1 pre-mRNA to generate HAS1Vb. A 3.7 kb long HAS1 minigene extending from exon 3 to exon 5 was amplified from a WM sample in which HAS1Vb was expressed. Sequence comparison of this minigene to the germline indicated two point mutations in intron 4 (G >A at positions 13 and A>G at position 1837 downstream of exon 4), a T insertion within the T’s stretch, 1708 bp downstream of exon 4 and an insertion of 2 extra TTTA repeats, 1861 bp downstream of exon 4. There were no mutations identified in intron 3 of this patient. The minigene was fused to the upstream HAS1 cDNA sequence in a mammalian expression vector, pcDNA3, and transfected into HeLa cells, which do not otherwise express either HAS1 or HAS1 splice isoforms. RNA splicing as determined by RT-PCR indicated that both constitutive and aberrant splicing of HAS1, in particular, HAS1Vb, occurred in transfected cells, indicating a correct assembly of the construct and the presence in HeLa cells of the required spliceosome components. It also suggests that HAS1 minigene sequences used in this analysis provided cis-acting elements required for HAS1Vb production and that the aberrant HAS1 splicing is not restricted by B cell specific trans-acting elements. Site-directed mutagenesis of intron 4 is currently under investigation to determine the significance of each point mutation/insertion in the HAS1 clone studied. In combination with sequence information of HAS1 intron 4 compiled from sequencing data from a large number of MM and WM patients generated in this laboratory, future site-directed mutagenesis and deletion analysis will target on intron 4 subregions where mutations are shown to cluster. Overall, our studies will unveil the essential cis-acting elements that regulate the aberrant splicing of HAS1Vb. Our work to date confirms that intronic mutations in HAS1 direct aberrant HAS1 pre-mRNA splicing to exclude exon 4 and include a fragment of intron 4, thereby generating an intronic splice variant previously shown to correlate with poor survival in MM. This suggests that in MM and WM but not B-CLL, inherited and somatic variations in genomic HAS1 may contribute to early events in the oncogenic processes giving rise to these malignancies.

In MM and WM, we identified aberrant HAS1 splice variants that were absent from normal donors (HD) and B-CLL. Here we sequenced multiple subclones from multiple cell subsets to show that aberrant HAS1 splicing results from cryptic splice sites activation. Aberrant splicing defects are the consequences of genetic variations (GVs) detected in the sequence of classical splicing elements as well as within exons and introns. To investigate HAS1 splicing in MM and WM patients, we sequenced the HAS1 gene segments involved in abnormal splicing events. HAS1 from buccal epithelial cells (BEC) represented the host genotype, and hematopoietic progenitors (HP), T, B and plasma cells (PC) as the normal and malignant components of the hematopoietic lineage in MM/WM. 197 GV were found in 16 MM or WM patients, but none in 9B-CLL, MGUS or HD. We found 60 germline (defined as present in BEC and hematopoietic cells) and 137 somatic GV (defined as GV found in HP, T, B and/or PC, but absent from BEC). These somatic GV include 97 tumor-specific GV found in MM and/or WM B and PC and 40 hematopoietic origin GV identified in HP, T, B and PC, but not in BEC. Some GV were recurrent, detected in more than one patient. Recurrent GV (24 in MM and 22 in WM) included both germline and somatic GVs, 6 tumor-specific, 6 hematopoietic and 14 germline origin GV, as well as 20 NCBI-SNPs. The distribution of GV indicated that some of the recurrent germline and somatic GV are restricted to MM, some are restricted to WM and some are shared by both MM and WM. None were found in B-CLL, MGUS or HD. The patterns of germline GV observed in MM and WM suggests that MM and WM patients, but not B-CLL, inherit recurrent germline GV that are necessary but not sufficient for progression to malignancy. Acquisition of recurrent, somatic HAS1 GV in HP further increases the risk of developing MM or WM. Transformation may become inevitable when tumor-specific recurrent GV are acquired. Interestingly, we detected increased homozygosity for the mutated allele of some germline GV in all cell types (PC, B, T, HPs and BECs) from MM and/or WM patients. These GV were detected in 75-90% of subclones analyzed. Mutational analysis of minigene constructs demonstrated a distribution of GV as clusters in the vicinity of HAS1splicing elements, allowing us to classify them as “splicing mutations”. This is supported by an in vitro splicing assay, which confirmed that a combination of germline and somatic GVs leads to aberrant HAS1 splicing (see abstract by Kriangkum et al., ASH 2007). Our study demonstrates that the impact of inherited and acquired GV on HAS1 gene splicing is manifested only in the context of accompanying tumor-specific HAS1 GV, that in combination give rise to the clinically significant aberrant splicing of HAS1. It also indicates that MM and WM are closely related at the genomic level, with the same recurrent somatic mutations independently arising in both diseases. Accumulation of somatically acquired HAS1 mutations in HP, in the context of inherited predispositions to MM and WM may represent a very early stage of pre-malignant development. Similar to leukemias, initiation events that contribute to MM and WM pathogenesis may arise from mutations which first accumulate during the non-malignant or pre-malignant stages of hematopoietic differentiation in progenitor cells.

BACKGROUND: We have previously found that in multiple myeloma there is amplification of the centrosome, the organelle that nucleates the mitotic spindle. We have reported that RHAMM is a component of the centrosome which interacts with TPX2, the protein that targets Aurora A kinase to the mitotic spindle. We have also shown that RHAMM expression and alternative splicing in myeloma correlate with centrosome amplification, aggressive disease and poor survival. These findings led us to speculate that Aurora A and possibly other Aurora kinases are potential therapeutic targets in myeloma. METHODS: We examined the expression of Aurora A, B and C kinases in 5 myeloma cell lines and autoMACS-purified CD138+ myeloma bone marrow plasma cells from 20 patients. We assessed the anti-proliferative and pro-apoptotic effects of Aurora A knockdown in myeloma cell lines with RNA interference. We investigated the anti-myeloma activity of two potent, selective Aurora kinase inhibitors, VE-465 (Merck/Vertex) and AZD1152 (AstraZeneca), in 5 myeloma cell lines, in CD138+ bone marrow plasma cells from 2 myeloma patients, and in a NOD/SCID murine xenograft model. RESULTS: Aurora A, B and C kinases are ubiquitously expressed in both myeloma cell lines and myeloma bone marrow plasma cells. Expression levels vary among patients. Aurora A and B are expressed in myeloma plasma cells at levels comparable to that seen in the CD138- cells from the same marrow sample, and comparable to the levels seen in normal marrow from control individuals. Aurora C, while expressed at low levels, is consistently ectopically overexpressed in myeloma plasma cells relative to coexisting CD138- cells and normal marrow. In myeloma cell lines, Aurora A knockdown with RNA interference induces apoptosis and cell killing. In all five myeloma cell lines tested, and in myeloma bone marrow plasma cells from two patients, both VE-465 and AZD1152 induce apoptosis and myeloma cell killing at nanomolar concentrations, to varying degrees (20–80% reduction in cell viability). VE-465 is known to inhibit all three Aurora kinases with comparable specificity, while AZD1152 is known to inhibit Aurora B and C more selectively than Aurora A. Despite these differences in activity, both compounds have comparable pre-clinical efficacy against myeloma. Myeloma cell lines treated with either agent demonstrate a phenotype consistent with target inhibition. Both drugs show additive effects on killing of cell lines and primary myeloma cells when combined with dexamethasone, even in dexamethasone-resistant cells. Anti-myeloma activity was seen with single agent Aurora kinase inhibition in the murine model, at well tolerated doses. CONCLUSIONS: Aurora kinases are potential therapeutic targets in myeloma. Aurora kinase inhibitors comprise an emerging class of anti-cancer drug therapy that deserves further evaluation for myeloma patients.

Interphase fluorescence in situ hybridization (FISH) is widely used as a diagnostic tool for known genetic abnormalities due to its sensitivity for detection of cryptic aberrations, such as t(4;14)(p16;q32) in multiple myeloma (MM). In many cancers, chromosomal abnormalities are prognostic indicators that also predict response to therapy. Tests to determine the type and extent of these abnormalities are increasingly essential for more informed diagnosis and choice of treatment strategies. However the cost and complexity of the current FISH protocols, and the variability arising from differences in technical approach and subjective evaluation of hybridization patterns has compromised its widespread utilization. To create a standardized platform that will be accurate, robust, cost-effective and easy to use in any clinical setting, we have developed a microfluidic platform that enables simultaneous assessment of 10 chromosomal abnormalities or 10 patients, on a single chip. Microfluidic chips are hybrid polymer/glass microsystems with miniaturized networks of wells and channels, incorporating valves, heaters and fluidic control. The 10 channel microfluidic chip used here is the size of a microscope slide. Each channel requires only 1/10 the amount of probe used in conventional FISH, thus substantially reducing the cost per test. All the probes tested gave comparable results to conventional testing. Three cell lines and three ex-vivo PBMC samples from MM patients were tested against four different chromosomal probe sets, to detect translocation (4:14), any 14q32 translocation, deletion of chromosome 13 or deletion of p53. We used a mixture of patient sample and cell line to test the robustness of our technology and were able to successfully distinguish abnormal patterns with percentages that were comparable to FISH on microscope slides. On-chip FISH was highly reliable with consistent results in multiple test runs. To automate the process of reading slide, a computer vision algorithm was developed to provide a quantitative and objective measure of staining patterns, and to eventually eliminate the requirement for human intervention. This strategy uses artificial intelligence to distinguish probe from background staining, to identify and quantify the number of cells with different chromosomal patterns. This visual processing algorithm has been validated against human interpretation and provides a sensitive and unbiased method to distinguish signal and noise within stained cells. Although reliable and reproducible hybridization occurred in as little as four hours, to further reduce the time required for FISH testing, methods to enhance the hybridization were examined. These included chip designs that implemented mechanical or electrokinetic pumping. Both methods improved the hybridization and are currently being optimized. On-chip FISH appears to be versatile, fast and inexpensive, making fully automated FISH testing a possibility. Compared to conventional methods, these first iterations of on-chip FISH provide a 10-fold higher throughput and a 10 fold reduction in the cost of testing. On-chip FISH technology holds promise for sophisticated and cost-effective screening of cancer patients at every clinic visit in any health care setting, thus facilitating the delivery of personalized cancer care targeted to the genetic characteristics of each individual. Funded by CIHR, NSERC and Western Economic Diversification.

In multiple myeloma (MM) and Waldenstrom’s macroglobulinemia (WM), we identified three alternatively and/or aberrantly spliced HAS1 transcripts—HAS1Va, HAS1Vb and HAS1Vc. Statistical analysis of samples taken from 172 untreated MM patients showed that expression of HAS1Vb, an intronic splice variant, strongly correlates with poor survival (P=0.005). We investigated the molecular basis of intron retention during HAS1 splicing in MM and WM patients. We speculated that aberrant HAS1 splicing and the associated reduced survival of MM patients, resulted from an accumulation of mutations in aberrantly spliced regions of HAS1. Exons and introns 3 and 4 of the HAS1 gene were sequenced, because they are hotspots for splicing aberrations. Sequencing of HAS1 was performed for a total 11 patients with WM and MM and 2 healthy donors. HAS1 gene templates for sequencing were isolated from a multiple sorted cell subpopulations, including malignant B and plasma cells (PC), non-malignant T cells and buccal epithelial cells (BECs), as well as hematopoietic progenitor cells (HPCs) from mobilized blood of MM patients. We detected sets of inherited and acquired genetic variations in HAS1 that were recurrent within 5–11 of the MM and WM patients analyzed, but absent from healthy donors. We also identified genetic variations that were unique to individual patients. Those HAS1 mutations found in all cell types tested, including BECs and from the hematopoietic cells (B, PC, T and HPCS) were classified as germline mutations. Those mutations found in hematopoietic cells but absent from BECs were classified as hematopoietic origin which acquired during the lifetime of the individual. Mutations identified only in malignant MM and WM B cells and PCs (absent from T cells, HPCs and BECs) were classified as acquired tumor specific mutations. Recurrent HAS1 mutations were found among both inherited and acquired sets of mutations. Some recurrent HAS1 mutations were common to both MM and WM. The high frequency of inherited HAS1 mutations suggests that they confer predisposition for developing MM or WM. Our sequencing analysis suggests that in MM and WM, sequential accumulation of genetic variations occurs as hematopoietic cells differentiate. Our data also suggest that hematopoietic origin mutations are necessary but by no means sufficient to drive HAS1 gene splicing. Effects of hematopoietic origin mutations on HAS1 splicing are manifested in malignant MM cells in context of additional tumor specific mutations, which are acquired by circulating B cells and passed to their plasma cell progeny. This suggests that mutations which lead to aberrant splicing of HAS1 pre-mRNA undergo mutational selection events, and leave a mutational “trace” throughout the hematopoietic cell lineage, including tumor cells. Existence of same mutational events detected in HAS1 gene from MM and WM supports the speculation that the precursors of both diseases may undergo a series of shared genetic events, diverging only when tumor specific mutations accumulate in distinct subsets of B lineage cells. In silico comparison of splicesomal assembly between wild type and mutated HAS1 gene gave a pattern that precisely predicts partial retention of intron and aberrant splicing of the HAS1 gene.

Malignant B lineage cells in Waldenstrom’s macroglobulinemia (WM) express a unique clonotypic VDJ associated with IgM. Studies of WM patients revealed a frequent incidence of biclonal B cells (16%) as defined by the presence of two distinct clonotypic VDJ sequences. This is the first report to estimate the frequency of biclonality in WM. Four WM cases are reported: two with distinct IgM clones (WM1-19 and WM1-09), one with distinct IgM/IgA clones (WM1-18), and one with related but diversified IgM/IgG clones (WM10). In 2 cases (WM1-19 and WM10), the two clonotypic signatures were found to be abundant in bone marrow (BM) but less frequent in blood, reminiscent of monoclonal WM cases. In WM1-19, single cell analysis showed that only one partner VDJ was expressed per cell, excluding the possibility of aberrant biallelic rearrangements. The distribution ratio between the two tumor clones was 2:1, suggestive of their mutual role in the clinical manifestation. In the other 2 cases (WM1-09 and WM1-18), partner clones were shown by CDR3 fragment analysis to be anatomically distinct, with one BM clonotypic signature and one in blood. In these 2 patients, the BM clone was hypermutated (6.2% and 3.8%) while the blood clone was germline or minimally mutated (0% and 1.0%). Partner clones lacked intraclonal diversity and Ig class switching, characteristic of malignant WM clones, suggesting relatively frequent transformation events throughout B lineage differentiation. The biological events leading to the appearance of clones in two different anatomic sites and the clinical implications remain to be understood. CDR3 fragment analysis in the longitudinal studies of WM1-09 and WM1-18 suggested that the repertoire of blood B cells may recover some level of diversity after successive cycles of treatment (WM1-18), while the persistence of the monoclonal peak (WM1-09) likely reflects a tumor that does not respond to treatment. In WM10, biclonal IgM/IgG B cell clones shared common VDJ gene rearrangement. The IgG clone displayed a higher mutation rate than did the IgM clone (9.7% vs. 6.1%). Increased somatic hypermutation in IgG consists mostly of replacement mutations that are clustered within the CDR regions, strongly supporting a contention that the IgG clone has undergone affinity maturation. There are 11 point mutations in the IgM that are not present in the IgG, suggesting that the two clones are distinctly different and belong to different branches of the genealogical tree. These distinct mutational signatures suggest that biclonal IgM/IgG clones are unlikely derived from clonal evolution. Overall, our results suggest that for the four biclonal WM evaluated here, the partner B cell clones appear to have undergone separate transformation events in the development of biclonality. The extent to which each partner clone contributes to disease progression and death, whether separately or in synergy, is as yet unknown.