SAMSN1 is a tumor suppressor gene in multiple myeloma.

Multiple myeloma (MM), a hematological malignancy characterized by the clonal growth of malignant plasma cells (PCs) in the bone marrow, is preceded by the benign asymptomatic condition, monoclonal gammopathy of undetermined significance (MGUS). Several genetic abnormalities have been identified as critical for the development of MM; however, a number of these abnormalities are also found in patients with MGUS, indicating that there are other, as yet unidentified, factors that contribute to the onset of MM disease. In this study, we identify a Samsn1 gene deletion in the 5TGM1/C57BL/KaLwRij murine model of myeloma. In addition, SAMSN1 expression is reduced in the malignant CD138+ PCs of patients with MM and this reduced expression correlates to total PC burden. We identify promoter methylation as a potential mechanism through which SAMSN1 expression is modulated in human myeloma cell lines. Notably, re-expression of Samsn1 in the 5TGM1 murine PC line resulted in complete inhibition of MM disease development in vivo and decreased proliferation in stromal cell-PC co-cultures in vitro. This is the first study to identify deletion of a key gene in the C57BL/KaLwRij mice that also displays reduced gene expression in patients with MM and is therefore likely to play an integral role in MM disease development.

Introduction

Multiple myeloma (MM) is an incurable hematological malignancy characterized by the clonal proliferation of malignant plasma cells (PCs) within the bone marrow (BM). MM is the second most common hematological malignancy after non-Hodgkin’s lymphoma, with approximately 20,000 newly diagnosed patients each year in the USA [1]. The main clinical manifestations of MM are the development of osteolytic bone lesions, bone pain, hypercalcemia, renal insufficiency, suppressed immunoglobulin production, and increased BM angiogenesis. Despite recent advances in treatment, MM remains almost universally fatal with a 10-year survival rate of approximately 17% [2].

MM encompasses a range of clinical variants ranging from monoclonal gammopathy of undetermined significance (MGUS) and smoldering/indolent MM to more aggressive disseminated forms of MM and PC leukemia. It is now widely accepted that most, if not all, MMs are preceded by a premalignant MGUS [3]. MGUS is defined as a benign proliferation of PCs and is clinically characterized by the presence of monoclonal protein (or “paraprotein”) of < 3 g/dl, clonal PCs constituting < 10% of the BM, and the absence of organ damage [4]. Patients with MGUS have a risk of developing overt MM at a rate of 1% per year; however, the time to progression varies greatly between patients [5].

Patients with MM can be stratified into various subgroups based on the presence of defined genetic abnormalities in their malignant PCs [6]. These genetic abnormalities include, but are not limited to, del(13) [7], del(16q) [8], del(17p) [9], gain of 1q21 [10], as well as translocations involving the immunoglobulin heavy chain locus, i.e., t(4;14)(p16.3;q32) and t(14:16)(q32;q23) [11], [12], all of which are associated with poor prognosis. While these abnormalities are believed to play a major causative role in disease development, studies suggest that many of them are already present in the clonal PCs of patients with MGUS [13], [14]. Therefore, it is likely that additional, as yet undefined, factors are required for the progression from asymptomatic MGUS to overt malignant MM. To this end, recent studies suggest that changes in gene expression, including up-regulation and/or down-regulation of key genes, which occurs through epigenetic mechanisms, may play a key role in the development of symptomatic MM [15].

In the present study, we used gene expression arrays to identify transcriptomic differences between the closely related C57BL/KaLwRij and C57BL/6 mouse strains that could account for the age-dependent predisposition of C57BL/KaLwRij mice to develop a MM-like disease [16], [17]. From these analyses, Samsn1 was identified as one of the most significantly downregulated genes in the C57BL/KaLwRij mouse strain. Further analysis revealed that the loss of Samsn1 expression was due to a homozygous gene deletion encompassing the entire Samsn1 coding region in the C57BL/KaLwRij mouse.

SAMSN1 (also known as HACS1, SLY2, and SASH2) encodes a member of the SLY family of cytoplasmic adaptor proteins and is predominantly expressed in the hematopoietic compartment, with lower levels of expression in heart, brain, placenta, and lung [18]. Studies in Samsn1 knockout mice, which are viable and fertile, indicate that it may act to moderate adaptive immune responses [19]. Samsn1 has also been shown to play a role in B cell activation and differentiation [20] and has been implicated as a tumor suppressor gene in lung cancer [21]. However, to date, the precise cellular role of SAMSN1 remains poorly understood.

In this study, we also show that SAMSN1 expression is low or absent in a proportion of human myeloma cell lines (HMCLs) and MM patient–derived PCs compared to PCs isolated from patients with MGUS and healthy controls. Furthermore, we show that methylation of the SAMSN1 promoter is a likely mechanism of reduced SAMSN1 expression in human MM cells. Importantly, overexpression of Samsn1 in the murine MM cell line, 5TGM1, completely inhibited MM disease development in vivo and reduced proliferation in stromal cell–PC co-cultures in vitro, suggesting that SAMSN1 may function as a tumor suppressor in MM.

Materials and Methods

Gene Expression Profiling by Microarray

Femora and tibiae from age- and sex-matched C57BL/6 and C57BL/KaLwRij mice were extracted, cleaned thoroughly, and snap-frozen in liquid nitrogen. Frozen bones were pulverized under liquid nitrogen in a mortar and pestle, and RNA was extracted using the high salt method [22]. Briefly, bone powder was homogenized in 4 M guanidinium thiocyanate containing 0.5% sarcosyl and 4.5 mM 2-mercaptoethanol by drawing repeatedly through a 19 G needle. Debris was removed by centrifugation and RNA was separated from proteins and DNA using acidified phenol/chloroform. Salt was removed by isopropanol precipitation and washing with 70% ethanol. The extracted RNA was further purified using a silica gel-based membrane column (Qiagen RNeasy kit; Qiagen, Valencia, CA). RNA was submitted to the Australian Genome Research Facility (Parkville, Australia) for labeling and array hybridization to the Illumina Mouse WG-6 v2.0 BeadChip. Data analysis was performed at the Australian Genome Research Facility. Four biologic replicates of each mouse strain were analyzed.

Real-Time Polymerase Chain Reaction

Total RNA was isolated using TRIzol reagent (Life Technologies, Carlsbad, CA) unless otherwise specified. RNA was reverse transcribed using Superscript III (Life Technologies) or Sensiscript (Qiagen) as per the manufacturers’ instructions. Real-time polymerase chain reaction (PCR) was conducted on a Rotor-Gene 3000 (Corbett Research, Mortlake, Australia) using the following primers: mouse/human Actb (F: 5′-TTGCTGACAGGATGCAGAAG-3′ and R: 5′-AAGGGTGTAAAACGCAGCTC-3′), mouse Samsn1 (F: 5′-TGCCTGCTCTCAGTTGTCTC-3′ and R: 5′-TCCGAAAACGGTCAAAATTC-3′), human B2M (F: 5′-AGGCTATCCAGCGTACTCCA-3′ and R: 5′-CGGCAGGCATACTCATCTTT-3′), and human SAMSN1 (F: 5′-TCCCTCAAAGCCAGTGACTC-3′ and R: 5′-GCCACAGAATGGTCCTGAAT-3′). Changes in gene expression were calculated relative to β-actin or β2-microglobulin using the ΔΔCt method [23].

Flow Cytometry

Approximately 500 μl of peripheral blood was obtained from mice by cardiac puncture and collected in microfuge tubes containing 50 μl of 0.5 M EDTA. Femora and tibiae were flushed and BM cells were collected. Red blood cells were removed by hypotonic lysis and leukocytes were stained with phycoerythrin (PE)-Cy7–conjugated rat anti-mouse B220 (eBioscience, San Diego, CA), fluorescein isothiocyanate–conjugated rat anti-mouse CD3 (eBioscience), PE-Cy5–conjugated rat anti-mouse CD11b (BioLegend, San Diego, CA), APC-conjugated rat anti-mouse Gr1 (eBioscience), and PE-conjugated rat anti-mouse NK1.1 (BD Biosciences, San Jose, CA). FluoroGold (Life Technologies) was used to exclude dead cells. Cells were sorted on a FACSAria II (BD Biosciences) into B cell (B220+), T cell (CD3+NK1.1−), monocyte (CD11bhiGr1lo), and granulocyte (CD11bhiGr1hi) populations for RNA extraction and real-time PCR. PCs were isolated from flushed long bones and identified using rat anti-mouse CD138 (R&D Systems, Minneapolis, MN) followed by PE-conjugated goat anti-rat IgG.

Deletion Mapping

Genomic DNA was isolated from mouse tissues using the DNeasy Blood and Tissue Kit (Qiagen) and PCRs were performed using AmpliTaq Gold Taq DNA Polymerase (Applied Biosystems, Foster City, CA), in accordance with the manufacturers’ recommendations. Primers and annealing temperatures are indicated in Supplementary Table S1. PCR products spanning the breakpoint were cloned into pGEMT-Easy vector (Promega, Madison, WI) before sequencing.

Patient Samples

BM trephines were collected, with informed consent, from patients with MM or MGUS and from hematologically normal controls. All MM samples were collected from patients at diagnosis with no prior therapy. This study was approved by the Royal Adelaide Hospital Human Research Ethics Committee.

CD138 + Magnetic-Activated Cell Sorting and RNA/DNA Isolation

CD138 + PCs were isolated from human BM samples from patients with MM (at diagnosis) using CD138 microbeads (Miltenyi Biotec, Auburn, CA) as per the manufacturer’s instructions. Briefly, cryopreserved human BM samples (approximately 107 cells/ml) were thawed into 10 ml of Dulbecco's modified Eagle's medium (high glucose) with 15% fetal calf serum (FCS) and DNase. Sample was centrifuged at 300g for 10 minutes and the supernatant was aspirated. Cell pellet was resuspended in magnetic-activated cell sorting (MACS) buffer (2mM EDTA and 0.5% deionized BSA in phosphate-buffered saline) and CD138 microbeads were added. Cells/beads were incubated on ice for 15 minutes, washed in 1 ml of MACS buffer, and centrifuged at 300g for 10 minutes. Cells were resuspended in MACS buffer and applied to a pre-rinsed MS column. The column was washed three times with MACS buffer, followed by elution in 1 ml. The purity of final elution was determined by FACS analysis using a CD138-PE antibody and samples were confirmed to be > 85% CD138 + following MACS. Total RNA and DNA were subsequently isolated using an All Prep DNA/RNA Micro Kit (Qiagen).

Cell Culture

Mouse 5TGM1 myeloma cells were maintained in Iscove’s modified Dulbecco’s medium (Sigma, St Louis, MO) with 20% FCS. HMCLs were maintained in RPMI-1640 medium (Sigma) with 10% FCS. BM stromal cells (BMSCs) were maintained in α-minimum essential medium (Sigma) with 10% FCS and 100 mM l-ascorbate-2-phosphate. All culture media were supplemented with 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, and 10 mM Hepes buffer (Life Technologies).

Methylation Analysis

Genomic DNA from myeloma cell lines was isolated using the DNeasy Blood and Tissue Kit (Qiagen). Two micrograms of DNA was bisulfite modified using the EpiTect Bisulfite Kit (Qiagen). Modified DNA was PCR amplified using EpiMark Hot Start Taq DNA Polymerase (New England Biolabs, Ipswich, MA) using the following primers: for downstream promoter: SAMSN1.30mer.F1 (5′-AGTTATGTTTTTATTTATATTTAGAATGGG-3′) and Downstream.01.R (5′-TCACCCAAACTAAAATACAATAACA-3′); for upstream promoter: SAMSN1.BIS.F (5′-TTGTTTTTATTTTGAGTTGTGTTTGT-3′) and SAMSN1.BIS.R (5′-ACTAAACTTCCTCCATTACTCTCTCTC-3′). PCR products were subcloned into pGemT-Easy vector (Promega) before sequencing. CpG methylation was assessed using QUMA (quma.cdb.riken.jp/).

Generation of Samsn1-Overexpressing Cell Lines

A luciferase-expressing 5TGM1 cell line (as described previously [24]) and an HMCL (H929)–overexpressing Samsn1 were generated by infection with a retroviral vector (pRUFimCH2 or pRUFiG2, respectively) harboring a full-length cDNA encoding murine or human Samsn1. pRUFiG2 was generated from pRUFneo [25]. Briefly, oligonucleotides encoding NotI and loxp sequences were cloned into the ClaI site at the 5′ end of the MC1Neo gene. An oligonucleotide encoding an loxp sequence and a new multiple cloning site (including BamHI, HpaI, EcoRI, BglII, SacII, SnaBI, NdeI, and XhoI) was cloned into the BamHI site of pRUFneo. The MC1Neo gene was excised and replaced with an internal ribosome entry site, green fluorescent protein (IRES-GFP) cassette from pMSCV-IRES-GFP to generate pRUFiG2. To generate pRUFimCH2, the IRES-GFP cassette was excised from pRUFiG2 with XhoI and NotI and replaced with the IRES-mCherry cassette from pcDNA3-IRES-mCherry.

The full-length murine Samsn1 cDNA was PCR amplified from a Mus musculus Samsn1 cDNA clone (clone 30077237; Open Biosystems, Huntsville, AL) and subcloned into pRUFimCH2 to generate pRUFimCH2-Samsn1. The full-length human SAMSN1 cDNA was PCR amplified from a Homo sapiens SAMSN1 cDNA clone (clone 4343284; Open Biosystems) and subcloned into pRUF-iG2 to generate pRUF-iG2-SAMSN1. Retroviral vectors were transfected into HEK-293T cells and viral particle-containing supernatant was used to infect 5TGM1-luc cells or H929 cells, as previously described [26]. Cell lines were sorted on a Beckman Coulter Epics Altra HyperSort, using Expo MultiComp Software version 1.2B (Beckman Coulter, Miami, FL) and pooled cell lines were established from the top 30% of mCherry or GFP-expressing cells. Resultant Samsn1-overexpressing cell lines (and empty vector controls) were used for subsequent in vitro and in vivo assays.

Western Blot

5TGM1 cells (2 × 107) were lysed in 1 ml of lysis buffer containing 1% NP-40, 20 mM Hepes, 150 mM NaCl, 10% glycerol, 2 mM Na3VO4, 10 mM Na4P2O7, 2 mM NaF, and Complete EDTA-free Protease Inhibitor Cocktail (Roche, Mannheim, Germany). One hundred micrograms of lysate was loaded on a 10% acrylamide gel and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Proteins were transferred to polyvinylidene difluoride membrane overnight and the membrane was subsequently incubated at room temperature in blocking buffer [Tris-buffered saline containing 0.1% Tween 20 and 2.5% ECL Blocking Agent (GE Healthcare, Little Chalfont, United Kingdom)] for 6 hours. The membrane was incubated overnight at 4°C with rabbit polyclonal anti-SAMSN1 antibody (Sigma) diluted 1:500 in blocking buffer, followed by alkaline phosphate-conjugated anti-rabbit IgG (Millipore, Billerica, MA) diluted 1:2500 in blocking buffer for 1 hour at room temperature. Proteins were visualized using ECL detection reagent (GE Healthcare) on a Typhoon FLA 7000 IP2 (GE Healthcare).

5-Aza-2′-Deoxycytidine Treatment in HMCL

HMCLs were seeded at 2 × 105 cells/ml and treated with 500 nM 5-aza-2′-deoxycytidine (Life Technologies) diluted in culture medium for 96 hours, replenishing treatment media daily. Total RNA was isolated from treated cells and untreated controls and specific gene expression was determined by real-time PCR.

Adhesion Assays

BMSCs were seeded at 8 × 103 cells per well in a 96-well plate and allowed to adhere overnight. Empty wells (plastic) were used as controls for adhesion. H929 or 5TGM1 cells (expressing an empty vector control or SAMSN1) were seeded at 2 × 104 cells per well in 100-μl volume and incubated for 10 minutes at 37˚C with 5% CO2. Cells were gently aspirated followed by three washes with 100 μl of HBSS with 5% FCS to remove non-adherent cells. One hundred microliters of standard culture medium was added to each well and four images taken per well at × 10 magnification. The number of GFP-positive cells per field of view was determined using FIJI analysis software (http://fiji.sc).

Proliferation Assays

5TGM1 cells were seeded at 2000 cells per well in triplicate in a 96-well plate. BrdU (Roche) was added to the cells and incubated for 24 hours at 37˚C with 5% CO2. Cells were fixed and stained as per the manufacturer’s protocol and absorbance was measured at 450 nm.

BMSCs were seeded at 5 × 104 cells per well in a 96-well plate (black plate with clear, flat bottom; Corning Life Science, Pittston, PA) in 100 μl of α-minimum essential medium with 10% FCS and allowed to adhere overnight. The medium was aspirated and 5TGM1-Samsn1 or 5TGM1-vector cells were seeded (in triplicate) at 5 × 103 cells per well in 100 μl of Iscove’s modified Dulbecco’s media + 20% FCS. Cells were incubated for 3 days at 37˚C with 5% CO2. Bioluminescence was determined using the Xenogen IVIS 100 Bioluminescence Imaging System (Caliper Life Sciences, Hopkinton, MA), following addition of 100 μl of 300 ng/ml luciferin per well and analysis using Living Image software (PerkinElmer, Waltham, MA). Absolute cell number per well was determined using a standard curve for bioluminescence.

Animals

C57BL/6 and C57BL/KaLwRij mice were bred and housed at the Institute of Medical and Veterinary Science Animal Care Facility. Animal studies were approved by the Institute of Medical and Veterinary Science/Adelaide Health Service and University of Adelaide Animal Ethics Committees. C57BL/KaLwRij mice (aged 6-8 weeks) received 5 × 105 luciferase-expressing 5TGM1 cells in 100 μl of sterile phosphate-buffered saline through the tail vein. At weekly intervals, mice were administered 150 mg/kg luciferin intraperitoneally and imaged using the Xenogen IVIS 100 Bioluminescence Imaging System until termination of the experiment at day 28. Tumor burden was quantitated using Living Image software.

μ-CT Analysis

Bone volume was evaluated using μ-CT (Skyscan 1174 X-ray Microtomograph; Bruker MicroCT, Kontich, Belgium). All tibiae were scanned at 48 kV/800 μA, with an isometric resolution of 6.49 μm/pixel using a 0.25-mm aluminium filter and two-frame averaging. Reconstruction of the original scan data was performed using NRecon. Analysis of bone volume fraction (BV/TV) was performed using CTAn. A total of 300 slices (1.947 mm) was analyzed for each tibia, commencing 85 slices (0.552 mm) distal to the growth plate. Digital segmentation of the bone from air/tissues was performed by adaptive (median-C) thresholding. The volume of interest were reconstructed in three dimensions using ANT software.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism version 5.02 for Windows (GraphPad Software, San Diego, CA; www.graphpad.com). Variance between patient groups was compared using analysis of variance (ANOVA) with Tukey’s multiple comparison test. Differences in tumor burden were compared between groups using a Mann-Whitney U test. In vitro assays were analyzed using t tests or ANOVA as appropriate. A P value of .05 was considered statistically significant.

Results

Samsn1 Expression Is Significantly Reduced in C57BL/KaLwRij Mice

To identify factors that may contribute to the development of MM, we compared the gene expression profiles from the closely related C57BL/6 and C57BL/KaLwRij mouse strains. While both the C57BL/6 and C57BL/KaLwRij strains have been demonstrated to develop monoclonal gammopathy at a similar rate (60-70% by 2 years old) [27], [28], the C57BL/KaLwRij mice are unique in their ability to spontaneously develop MM at a low frequency (0.5% of mice > 2 years old). Furthermore, the C57BL/KaLwRij strain allows the successful engraftment of exogenous murine myeloma PCs, while the C57BL/6 strain does not [16], [17], [29]. The C57BL/KaLwRij model is one of the most widely studied preclinical animal models of MM and the MM disease exhibited by these animals faithfully recapitulates the symptoms of the human disease, including osteolysis [16], [30], [31], [32]. In addition, these animals have previously been used to identify factors that play a role in the pathogenesis of human MM [33], [34].

Total RNA was extracted from the long bones (tibiae and femora) of age- and sex-matched C57BL/6 and C57BL/KaLwRij mice. Expression profiling of the long bones revealed a small number of genes (87) that are differentially expressed between the two strains by two-fold or more (Figure 1A). Of these, 22 were unnamed genes with no identified function and a further 7 had no human orthologue. Literature searches were conducted for the remaining 58 genes (Table 1) and these were subsequently prioritized for further investigation based on their potential relevance to MM. Statistically significant differences in expression were confirmed by quantitative PCR for 5 of the 20 most promising candidates (data not shown). Of particular interest was the expression of Samsn1, which was shown to be absent in the bone of C57BL/KaLwRij mice at both the mRNA and protein levels (Figure 1B). Samsn1 has been shown to play a role in regulating adaptive immune responses and the development of B cells [19], [20], which, coupled with our microarray data, highlighted Samsn1 as a promising candidate for further investigation into its potential role in the development of MM.

Figure 1

Global loss of Samsn1 expression is a feature of the 5TGM/C57BL/KaLwRij mouse model of myeloma. (A) The gene expression profiles of the long bones of C57BL/6 (n = 4) and C57BL/KaLwRij (n = 4) mice were compared using the Illumina Mouse WG-6 v2.0 BeadChip. Fold change in gene expression between the strains is plotted against the significance of the change. Each data point represents one gene. Samsn1 is circled. (B) The reduced expression of Samsn1 in C57BL/KaLwRij bones was confirmed in independent samples by real-time PCR (n = 4 per group) and Western blot. ****P < .0001, t test. (C) RNA was isolated from C57BL/KaLwRij (n = 3) and C57BL/6 (n = 3) derived tissues and Samsn1 mRNA expression was assessed by real-time PCR. (D) Cells were isolated from C57BL/KaLwRij (n = 3) and C57BL/6 (n = 3) BM (PCs and stromal cells) and peripheral blood (other cell subsets). Samsn1 mRNA expression was assessed by real-time PCR. (E) 5TGM1 cells express negligible Samsn1, as determined by real-time PCR. C57BL/KaLwRij and C57BL/6 mouse PCs are shown for comparison. ****P < .0001; ***P < .001; **P < .01; t test.

Table 1

Genes Differentially Expressed in the Long Bones of C57BL/KaLwRij and C57BL/6 Mice

Symbol	Definition	Fold change⁎	Direction
Gbp1	Guanylate binding protein 1	4.450412306	Increase
Ddit4	DNA damage–inducible transcript 4	4.240428329	Increase
Hist4h4	Histone cluster 4, H4	3.37437661	Increase
Tmod4	Tropomodulin 4	3.353338216	Increase
Angptl7	Angiopoietin-like 7	3.283043806	Increase
Mx2	Myxovirus (influenza virus) resistance 2	3.249176241	Increase
Lyplal1	Lysophospholipase-like 1	3.139500408	Increase
Ifit3	Interferon-induced protein with tetratricopeptide repeats 3	3.030790225	Increase
Pttg1	Pituitary tumor–transforming 1	2.94256646	Increase
Oas2	2′-5′ oligoadenylate synthetase 2	2.731815404	Increase
Oas1g	2′-5′ oligoadenylate synthetase 1G	2.704346478	Increase
Usp18	Ubiquitin specific peptidase 18	2.538251457	Increase
Ifi27l2a	Interferon, alpha-inducible protein 27	2.533970976	Increase
Gbp7	Guanylate binding protein 7	2.508261909	Increase
Pisd-ps3	Phosphatidylserine decarboxylase, pseudogene 3	2.436705673	Increase
Per1	Period homolog 1 (Drosophila)	2.397020089	Increase
Sct	Secretin	2.384120243	Increase
Irf7	Interferon regulatory factor 7	2.378730781	Increase
Pdk4	Pyruvate dehydrogenase kinase, isoenzyme 4	2.346674916	Increase
Rtp4	Receptor transporter protein 4	2.337282308	Increase
Oasl2	2′-5′ oligoadenylate synthetase-like 2	2.242105245	Increase
Msc	Musculin	2.210238416	Increase
Hist1h4b	Histone cluster 1, H4b	2.207251464	Increase
Igh-VJ558	PREDICTED: immunoglobulin heavy chain (J558 family)	2.193584071	Increase
Rbp7	Retinol binding protein 7, cellular	2.191502007	Increase
Pydc4	Pyrin domain containing 4	2.187228581	Increase
Npy	Neuropeptide Y	2.170017985	Increase
Erdr1	Erythroid differentiation regulator 1	2.158546243	Increase
Thsd4	Thrombospondin, type I, domain containing 4, transcript variant 1	2.133757168	Increase
Bfsp2	Beaded filament structural protein 2, phakinin	2.110095069	Increase
Paip1	Polyadenylate binding protein-interacting protein 1, transcript variant 1	2.098810627	Increase
Cyr61	Cysteine rich protein 61	2.036144571	Increase
Mgst1	Microsomal glutathione S-transferase 1	57.60499461	Decrease
Samsn1	SAM domain, SH3 domain and nuclear localization signals, 1	51.76062024	Decrease
Entpd4	Ectonucleoside triphosphate diphosphohydrolase 4	19.42186718	Decrease
Slc25a37	Solute carrier family 25, member 37	16.9916612	Decrease
Slc22a4	Solute carrier family 22 (organic cation transporter), member 4	14.15791584	Decrease
H2afj	H2A histone family, member J	11.74400026	Decrease
Sparc	Secreted acidic cysteine rich glycoprotein	9.001990237	Decrease
Bbs4	Bardet-Biedl syndrome 4 (human)	3.525644472	Decrease
Kif3a	Kinesin family member 3A	3.134149503	Decrease
Tipin	Timeless interacting protein	2.959086767	Decrease
Cdkn1b	Cyclin-dependent kinase inhibitor 1B	2.908346173	Decrease
Vsig4	V-set and immunoglobulin domain containing 4	2.825793418	Decrease
Hist1h3f	Histone cluster 1, H3f	2.804314199	Decrease
Sfrp4	Secreted frizzled-related protein 4	2.725447254	Decrease
Rps3a	Ribosomal protein S3a	2.493150544	Decrease
Pml	Promyelocytic leukemia (Pml), transcript variant 1	2.333694965	Decrease
Hist1h4i	Histone cluster 1, H4i	2.240988119	Decrease
Dcp1b	DCP1 decapping enzyme homolog b (S. cerevisiae)	2.233498058	Decrease
Aoah	Acyloxyacyl hydrolase	2.202379368	Decrease
Adpgk	ADP-dependent glucokinase	2.172734661	Decrease
Trim2	Tripartite motif protein 2	2.151279536	Decrease
Hdgf	Hepatoma-derived growth factor	2.128178262	Decrease
Hist1h4f	Histone cluster 1, H4f	2.085821933	Decrease
Plac9	Placenta specific 9	2.038744521	Decrease
Tbc1d9b	TBC1 domain family, member 9B	2.012406842	Decrease
Arhgap26	Rho GTPase activating protein 26	2.005116414	Decrease

Expression in C57BL/KaLwRij bone relative to expression in C57BL/6 bone.

Loss of SAMSN1 Is a Feature of the 5TGM1/C57BL/KaLwRij Mouse Model of MM

To identify whether Samsn1 may play a role in the development of MM in the mouse model, we first investigated Samsn1 mRNA expression in various tissues and found it to be consistently absent within the C57BL/KaLwRij mice compared to C57BL/6 controls (Figure 1C). Previous studies have shown that Samsn1 is expressed in cells of the hematopoietic compartment [18]. As such, hematopoietic cell subsets (B cells, T cells, monocytes, granulocytes, PCs, and stromal cells) were isolated by FACS and Samsn1 expression was evaluated by real-time PCR. As seen in Figure 1D, Samsn1 mRNA expression was absent across the entire range of hematopoietic cell subsets, including PCs, the effector cells of MM. Notably, this lack of Samsn1 expression was also evident in the well-characterized 5TGM1 cell line (Figure 1E), a finding consistent with its C57BL/KaLwRij origin. From these findings, we hypothesized that the absence of Samsn1 expression in these mice, and specifically within the malignant 5TGM1 PCs, may represent a crucial mechanism by which MM disease develops in this murine model.

The Samsn1 Gene Is Deleted in C57BL/KaLwRij Mice and 5TGM1 Murine PCs

To identify a mechanism for the global lack of Samsn1 expression in the C57BL/KaLwRij mice, we attempted to PCR amplify the two promoter regions of Samsn1 to analyze this region for CpG methylation (see Figure 2A for overview of the gene structure). Interestingly, using primers designed specifically to the downstream promoter region of the Samsn1 gene, we were unable to amplify the genomic DNA from tissues derived from the C57BL/KaLwRij mice (data not shown). Upon further analysis, we identified a large chromosomal deletion, encompassing the entire coding region of the Samsn1 gene in the C57BL/KaLwRij mice. Notably, the corresponding region in the C57BL/6 control strain was not deleted (Figure 2B, Supplementary Table S1, Supplementary Figure S1). PCR performed with primers BP.1F and DEL+55kb.F, which flank the deletion interval (Figure 2A), only yielded a product with KaLwRij-derived genomic DNA (Figure 2C). Sequencing of this PCR product revealed that the deletion is 179,971 bp in length and extends from midway through intron 2 to an intergenic region between Samsn1 and the next nearest downstream gene, Hspa13. The deleted region extends from 75,816,191 to 75,996,161 of mouse chromosome 16, and the three genes Hspa13, Rbm11, and Lipi that are located immediately downstream of Samsn1 were not deleted (data not shown; Supplementary Table S1). Notably, the 179,971 bp deletion was also shown to be present in the 5TGM1 malignant PC line (Figure 2, B and C). These data indicate that the C57BL/KaLwRij mice are genetically null for Samsn1. This model of MM therefore does not express Samsn1 in the microenvironment or in the malignant PCs and therefore provides an ideal model through which to investigate the role of PC-specific Samsn1 expression in MM disease development in vivo.

Figure 2

The Samsn1 gene is deleted in C57BL/KaLwRij mice. (A) Extent of genomic deletion of the Samsn1 gene in C57BL/KaLwRij mice. The two predominant protein coding isoforms from ensembl.org Mouse Genome Assembly are shown [Samsn1-001 (ENSMUST00000114240) and Samsn-003 (ENSMUST00000114239)]. Solid and open rectangles represent coding and non-coding exons, respectively. The large hatched horizontal rectangle at the bottom represents the genomic interval that is deleted, with the letters indicating the location of deleted (B–D) versus non-deleted (A and E) PCR primer binding sites. (B) PCR analysis of genomic DNA isolated from C57BL6 and C57BL/KaLwRij tissues/cell lines. Examples of PCRs that detect regions that are present (A and E) or deleted (B–D) in C57BL/KaLwRij tissues/cell lines are shown. The PCR primer names on the right-hand side correspond to those in Supplementary Table S1. NTC, no template (negative) control. (C) Breakpoint PCR with primer BP1F/DEL+55kb.F is only detectable in C57BL/KaLwRij-derived tissues or cell lines. The position of the breakpoint is indicated by the change from emboldened to normal text above the sequence chromatogram.

The Samsn1 gene is deleted in C57BL/KaLwRij mice. (A) Extent of genomic deletion of the Samsn1 gene in C57BL/KaLwRij mice. The two predominant protein coding isoforms from ensembl.org Mouse Genome Assembly are shown [Samsn1-001 (ENSMUST00000114240) and Samsn-003 (ENSMUST00000114239)]. Solid and open rectangles represent coding and non-coding exons, respectively. The large hatched horizontal rectangle at the bottom represents the genomic interval that is deleted, with the letters indicating the location of deleted (B–D) versus non-deleted (A and E) PCR primer binding sites. (B) PCR analysis of genomic DNA isolated from C57BL6 and C57BL/KaLwRij tissues/cell lines. Examples of PCRs that detect regions that are present (A and E) or deleted (B–D) in C57BL/KaLwRij tissues/cell lines are shown. The PCR primer names on the right-hand side correspond to those in Supplementary Table S1. NTC, no template (negative) control. (C) Breakpoint PCR with primer BP1F/DEL+55kb.F is only detectable in C57BL/KaLwRij-derived tissues or cell lines. The position of the breakpoint is indicated by the change from emboldened to normal text above the sequence chromatogram.

Samsn1 Expression Is Reduced in a Subset of Patients with MM

To ascertain whether SAMSN1 is also aberrantly expressed in patients with MM, RNA was extracted from total BM trephine biopsies recovered from patients with MM (n = 34) and MGUS (n = 9) at diagnosis and from hematologically normal, age-matched controls (n = 5). SAMSN1 expression was shown to be significantly reduced in the BMs of patients with MM when compared with those of patients with MGUS and age-matched controls (P = .0019, one-way ANOVA; Figure 3A). We next wanted to investigate the PC-specific expression of SAMSN1 in patients with MM. Therefore, we specifically isolated CD138 + PCs from BM samples of patients with MM using MACS techniques and were subsequently able to determine that patients presenting with higher BM PC burden at diagnosis had significantly lower PC-specific SAMSN1 expression (n = 10, r2 = 0.6147, P = .0043; Figure 3B). In support of our findings, in silico analyses of publicly available data sets from large-scale microarray studies (GEO Accession Nos GSE4581 and GSE5900; Myeloma Institute, University of Arkansas) also showed that SAMSN1 mRNA expression was significantly reduced in purified PCs from patients with MM (n = 414) compared with patients with MGUS (n = 44) and healthy age-matched controls (n = 22; P < .0001, one-way ANOVA; Figure 3C). Interrogation of these data sets revealed that approximately 25% of patients with MM express SAMSN1 below the normal range compared to only 7% of patients with MGUS. Low SAMSN1 expression was also observed in four of six HMCLs examined (Figure 3D). Together, these data suggest that there is a significant reduction in SAMSN1 expression in myeloma PC.

Figure 3

SAMSN1 expression is reduced in CD138 + PCs of patients with MM and HMCL. (A) SAMSN1 expression (as determined by real-time PCR) is significantly reduced in the BMs of patients with MM (n = 34) compared with patients with MGUS (n = 9) and healthy age-matched controls (n = 5; *P < .05, **P < .001, one-way ANOVA with Tukey’s multiple comparison test). (B) SAMSN1 expression in CD138 + MACS isolated PCs from patients with MM negatively correlates with BM PC burden (n = 10, r2 = 0.6147, P = .0043). (C) In silico analysis of published microarray data. CD138+ PCs were isolated by MACS from 414 patients with MM, 44 patients with MGUS, and 22 age-matched controls. RNA was extracted and analyzed using the Affymetrix U133Plus2.0 microarray platform (GEO Accession Nos GSE4581 and GSE5900). Expression of SAMSN1 is significantly reduced in PCs of patients with MM compared to those of patients with MGUS and normal controls. P < 0.0001, one-way ANOVA with Tukey’s multiple comparison test. (D) Total RNA was extracted from six HMCLs and reverse transcribed. The levels of SAMSN1 expression were assessed by real-time PCR.

The Samsn1 Promoter Is Methylated in HMCLs

Despite its reduced expression levels in lung cancer cell lines [21], methylation of the SAMSN1 promoter has not been previously investigated in the context of human disease. We hypothesize that, due to the significant reduction of SAMSN1 expression in patients with MM and HMCLs, SAMSN1 expression may be modulated by differential methylation of the promoter. In support of this, a recent study by Heller et al. listed SAMSN1 as a significantly upregulated gene in two myeloma cell lines following treatment with the DNA methyltransferase inhibitor 5-aza-2′-deoxycytidine [35]. Consistent with these findings, we showed an increase in mRNA expression of SAMSN1 in two HMCLs (U266 and H929) following treatment with 500 nM 5-aza-2′-deoxycytidine for 96 hours (Figure 4A). Interestingly, although displaying similar low basal SAMSN1 expression levels, the JIMI and LP-1 cell lines did not exhibit an increase in SAMSN1 expression following treatment with 5-aza-2′-deoxycytidine (Figure 4A).

Figure 4

The SAMSN1 gene is methylated in HMCLs. (A) H929, U266, LP-1, and JIMI cells were treated with 500 nM 5-aza-2′-deoxycytidine for 96 hours. Total RNA was isolated from treated and untreated controls. SAMSN1 expression was significantly increased in response to 5-aza-2′-deoxycytidine treatment in the H929 and U266 cell lines, as determined by real-time PCR. *P < .05, t test. (B) SAMSN1 promoter methylation in HMCLs. Patterns of CpG methylation revealed by sequencing of cloned PCR products for the two promoters of SAMSN1 are shown. Closed circles represent methylated CpGs and open circles represent unmethylated CpGs. The HMCLs are grouped according to SAMSN1 expression levels. The − 470/− 460 CpGs whose methylation status correlates with expression level are highlighted by the vertical hatched rectangle. The numbers above the CpGs are relative to the transcription start sites of the SAMSN1-001 (ENST00000285670) isoform for the upstream promoter and the SAMSN1-002 (ENST00000400566) and SAMSN1-002 (ENST00000400564) isoforms for the downstream promoter. SAMSN1 isoform data are from ensembl.org Human Genome Assembly GRCh37.p13.

To investigate this further, we examined the methylation pattern of the SAMSN1 promoter region in a panel of HMCLs using bisulfite-sequencing techniques. Similar to its mouse orthologue, the human SAMSN1 gene has two promoter regions associated with the full-length protein coding isoforms (see Figure 2A). As seen in Figure 4B (left panel), there is evidence of high levels of methylation of the CpG dinucleotides in the upstream SAMSN1 promoter region of all HMCLs screened. In contrast, the downstream SAMSN1 promoter exhibited differential methylation, particularly in the two central CpGs at − 470 and − 460 relative to the transcription start site (Figure 4B, right panel, hatched box). Importantly, the degree of methylation of the − 470 and − 460 CpGs of the downstream promoter corresponded with SAMSN1 expression levels in the HMCLs, with the lowest levels of combined methylation of these two CpGs seen in the RPMI-8226 (28%) and WL-2 (19%) cell lines, which show the greatest expression of SAMSN1 (see Figure 3D). In contrast, H929 (56%), U266 (100%), and LP-1 (100%) cells exhibit higher degrees of methylation at the − 470 and − 460 CpGs and do not express SAMSN1. Consistent with the lack of induction of SAMSN1 expression following treatment with 5-aza-2′-deoxycytidine, JIMI cells exhibit − 470/− 460 CpG methylation levels equivalent to that seen in the RPMI-8226 and WL-2 cell lines (5%). Interestingly, although displaying low basal SAMSN1 expression and a high degree of promoter methylation, the LP-1 cell line also did not respond to 5-aza-2′-deoxycytidine treatment. These data suggest a role for promoter methylation in the modulation of SAMSN1 expression and also highlight the existence of alternative, yet to be defined, mechanisms.

Overexpression of Samsn1 in 5TGM1 Cells Completely Inhibits MM Disease Development In Vivo

As detailed above, our studies show that Samsn1 expression is absent in the 5TGM1/C57BL/KaLwRij mouse model of myeloma, and this is consistent with reduced expression in a subset of patients with MM. We next aimed to determine whether restoration of Samsn1 expression in the 5TGM1 cells affects the development of MM in vivo. Luciferase-expressing 5TGM1 cells (which do not express Samsn1; see Figure 1E) were transduced with a Cherry-labeled Samsn1 expression construct (5TGM1-Samsn1) or vector control (5TGM1-vector), and Samsn1 expression was confirmed by real-time PCR (data not shown) and Western blot (Figure 5A). The expression of Samsn1 in 5TGM1-Samsn1 cells was equivalent to levels seen in PC isolated from C57BL/6 mice (data not shown). The use of luciferase-labeled cells allowed us to track the progression of MM disease spread and development in vivo using bioluminescence imaging techniques [24], [34].

Figure 5

Overexpression of Samsn1 completely inhibits MM tumor growth in vivo. (A) Western blot for Samsn1 in 5TGM1-vector and 5TGM1-Samsn1 cells. (B) No significant difference was seen in the proliferation of 5TGM1-vector and 5TGM1-Samsn1 cells as determined by BrdU incorporation. (C) C57BL/KaLwRij mice were injected intravenously with 5 × 105 5TGM1 cells. On day 28, mice received luciferin intraperitoneally and tumor growth was assessed using the Xenogen IVIS 100. Bioluminescence images of tumor growth in mice receiving control (upper panel) and Samsn1-expressing (lower panel) 5TGM1 cells (n = 5 per group) are shown. The in vivo growth of Samsn1-expressing 5TGM1 cells is significantly reduced compared to control cells as determined by quantification of photon flux. **P = .0079, Mann-Whitney U test. (D) Representative three-dimensional images of the tibial trabecular structure of tumor-bearing mice. Osteolysis was significantly lower in mice carrying 5TGM1-Samsn1 tumors as determined by bone volume analysis (*P = .0244, t test, n = 5). (E) Representative three-dimensional images of the tibial cortical structure of tumor-bearing mice. There was no significant difference in the number of cortical pits.

Re-expression of Samsn1 in 5TGM1 cells had no effect on their proliferative capacity in vitro, as determined by a BrdU incorporation assay (Figure 5B). However, re-introduction of Samsn1 to the 5TGM1 cells was found to significantly inhibit the development of MM disease in vivo, as seen by a significant decrease in bioluminescence of 5TGM1-Samsn1–inoculated mice compared to the 5TGM1-vector controls (P = .0079, Mann-Whitney U test; Figure 5C). In keeping with the absence of tumor in mice inoculated with 5TGM1-Samsn1 cells, the bone volume fraction in the tibiae of these animals was significantly greater than that of control tumor-bearing mice (32.18 ± 0.3815% compared with 30.38 ± 0.5256%, P = .0244, two-tailed t test; Figure 5D). While the number of resorption lacunae that traversed the cortices of the tibiae of both groups was similar (63.8 ± 6.402 compared with 71.9 ± 3.519, P = .2998; Figure 5E), the resorption lacunae in the 5TGM1-Samsn1 group tended to be smaller. These data indicate that restoring expression of Samsn1 in the murine 5TGM1 PCs can completely abolish the capacity of these cells to form intramedullary tumors in vivo and therefore prevents the osteolysis that is commonly observed in this tumor model in response to MM.

SAMSN1 Overexpression Reduces Cell Growth in the Presence of BMSCs In Vitro

To determine a mechanism by which Samsn1 expression may inhibit the development of MM, we investigated the adhesive properties of the vector- and Samsn1-expressing cells in vitro. 5TGM1-Samsn1 cells exhibited an increased capacity to adhere to a BMSC layer compared to the 5TGM1-vector control cells in a short-term in vitro adhesion assay (Figure 6A). This was confirmed in the H929 HMCL (Supplementary Figure S2). As adherence of cells of the hematopoietic lineage to stroma has previously been demonstrated to reduce proliferation [36], [37], [38], we next wanted to ascertain whether this increase in adhesion was associated with an altered proliferative capacity of the 5TGM1-Samsn1 cells. 5TGM1-Samsn1 and 5TGM1-vector cells were grown in the presence of BMSC and total cell number was quantified after a 72-hour incubation period. The total number of 5TGM1-Samsn1 cells was reduced compared to the vector control cells (Figure 6B), suggesting that the expression of Samsn1 in PCs may play a role in regulating cell adhesion and proliferation, thereby accounting for the lack of tumor growth observed in vivo.

Figure 6

Overexpression of Samsn1 increases adhesion and decreases cell growth on stroma in vitro. (A) 5TGM1-Samsn1 or -vector control cells were allowed to adhere to a monolayer of BMSCs for 10 minutes, followed by gentle washing to remove non-adherent cells. The number of GFP + adherent cells was determined using FIJI analysis software. Mean ± SEM, n = 4, ***P < .001, t test. (B) Equal numbers of 5TGM1-Samsn1 or -vector control cells were seeded on BMSC and the total number of cells after 72 hours was determined by bioluminescence imaging techniques. Mean ± SEM, n = 3, *P <. 05, paired t test.

Discussion

This is the first study to identify a key genetic variation, specifically the deletion of the Samsn1 gene, in the C57BL/KaLwRij mice, which may contribute to the propensity of these mice to spontaneously develop myeloma-like disease. Moreover, analysis of patient data showed that while expressed in PCs from healthy donors and patients with MGUS, SAMSN1 expression is significantly reduced in patients with MM suggesting a role for SAMSN1 in MM disease development and/or progression. Furthermore, we also showed that restoration of Samsn1 expression in MM PCs inhibited myeloma disease development in vivo, highlighting a potential tumor suppressor role for this protein in MM.

Although loss of heterozygosity at the chromosomal region 21q11-21, which includes the SAMSN1 gene, has previously been associated with lung cancer development [21], this is the first study in which SAMSN1 has been shown to play a tumor suppressor role in a hematological malignancy. Interestingly, this chromosomal region has also been identified as a region of frequent translocation events in hematological malignancies [39] and more recently was identified specifically as a region of chromosomal gain in a small number of patients with MM [40]. Furthermore, studies have previously shown an increase in SAMSN1 expression both in PC leukemia and MM [18], [41]. The contradiction between these studies and our own findings can likely be attributed to the large degree of genetic heterogeneity observed in patients with MM. Indeed, the study by Ni et al. identified a gain of 21q (specifically associated with a gain of SAMSN1) in only a very small number of patients on the background of a BCL1/JH t(11;14) (q13;q32) translocation, while we observe significantly reduced SAMSN1 expression in an unbiased sample of total BM trephines from patients with MM compared to healthy donors. Furthermore, our findings are supported by preliminary in silico analysis of the University of Arkansas gene expression data set with approximately 25% of MM patient–derived PC expressing SAMSN1 below the normal range. Further in silico analysis also suggests that reduction of SAMSN1 is not associated with any particular MM genetic subtype. Rather, patients with low SAMSN1 expression can be found in most subgroups. However, within each subgroup, these patients tend to have poorer disease-related survival compared to patients with SAMSN1 expression levels in the normal range (data not shown). This is also consistent with the correlation observed between increased BM PC burden and low CD138 + PC-specific SAMSN1 expression. It will be important in the future to determine what common genetic anomalies may be associated with reduced SAMSN1 expression in patients, e.g., activated nuclear factor kappa B pathway signaling, deregulated MYC, or common translocations, which may account for the enhanced severity of disease. It will also be necessary to dissect out whether these genetic signatures are the cause or consequence of reduced SAMSN1 expression.

Although Yamada et al. demonstrated frequent down-regulation of SAMSN1 in lung cancer cell lines, no mechanism for this down-regulation was identified [21]. A number of recent studies have shown global hypomethylation of the genome to be associated with the progression of MM disease from the non-malignant MGUS stage, with the majority of this hypomethylation occurring outside of defined CpG islands [15], [42], [43]. Gene-specific hypermethylation, associated with CpG islands within promoter regions, is also evident in MM progression. This hypermethylation has, to date, largely been associated with genes involved in developmental processes, cell cycle, and regulation of transcription [15], [43]. In this study, we present SAMSN1 as a novel gene exhibiting hypermethylation in cell lines derived from patients with MM. We have correlated differential methylation of specific CpGs within the downstream SAMSN1 promoter with gene expression levels. Furthermore, we demonstrated an increase in SAMSN1 expression in HMCLs following treatment with the DNA methyltransferase inhibitor 5-aza-2′-deoxycytidine. This is supported by a previous study showing a significant increase in SAMSN1 expression in an HMCL upon treatment with 5-aza-2′-deoxycytidine [35]. This finding provides evidence of SAMSN1 hypermethylation and subsequent down-regulation as a potential marker of MM disease.

C57BL/KaLwRij and C57BL/6 are closely related strains of mice. The C57BL/KaLwRij strain uniquely exhibits an inherent ability to develop MM disease as well as permitting the growth of the exogenous 5T series of mouse MM lines, while the C57BL/6 strain does not [16], [17], [29]. The 5TGM/C57BL/KaLwRij model provided an ideal opportunity to investigate the effect of PC-specific Samsn1 expression on the in vivo development of MM, as we have shown a large genomic deletion within chromosome 16, encompassing the entire Samsn1 coding region, resulting in an animal that is null for Samsn1 expression. As this deletion was shown to be specific to the C57BL/KaLwRij strain, we hypothesized that the deletion of Samsn1 was necessary but not sufficient for the unique development of MM in these animals. Despite being Samsn1 null, C57BL/KaLwRij mice only develop MM at a very low rate (approximately 0.5% greater than 2 years old) [16], suggesting that the loss of Samsn1 alone is not sufficient to drive the development of this disease. It is therefore likely that the 5TGM1 cells contain further genetic or epigenetic alterations that, in combination with the loss of Samsn1, account for the malignancy in this system.

It is evident that the PC-specific loss of Samsn1 expression is critical in the development of MM, as restoration of Samsn1 expression in the 5TGM1 cell line significantly inhibited disease development in vivo. In keeping with this finding, SAMSN1 expression is significantly reduced both in a subset of HMCLs and within CD138 + purified PCs from BMs of patients with MM. Together, these data suggest that the role of SAMSN1 in MM development is likely to be largely PC-specific.

Although the precise cellular function of SAMSN1 is largely unknown, it has been reported to be expressed primarily in the hematopoietic compartment [18] and to be up-regulated by B cell stimulators resulting in the activation and differentiation of B cells [20]. In addition, overexpression of Samsn1 in murine splenic B cells resulted in inhibition of proliferation, while mice with knock-out Samsn1 showed increased proliferation of naïve B cells [19], [20]. Collectively, these data support a role for SAMSN1 in B cell development and function. In addition, a role for SAMSN1 in actin cytoskeleton organization and B cell spreading has also been proposed [44]. Our preliminary findings suggest that expression of SAMSN1 in myeloma cell lines results in increased adhesion to BMSCs and subsequent decreased cell growth in short-term in vitro assays, accounting for the reduced tumor growth observed in vivo. This is in contrast to cells grown in the absence of BMSC wherein no change in proliferation is observed. These findings are supported by previous studies that have demonstrated a link between increased adhesion of hematopoietic progenitor cells to stroma and decreased cellular proliferation [36], [37]. In addition, specific binding of hematopoietic progenitor cells through the cell adhesion molecule PSGL-1 has been associated with suppressed cellular proliferation [45] and adhesion of lymphoma cells to BMSCs results in G1 arrest [38]. Taken together, these data implicate SAMSN1 as not only a modulator of B cell development and function [19], [20], [44] but also a protein that plays a role in controlling cell adhesion and proliferation of PCs within the BM microenvironment. However, it remains to be fully elucidated as to how loss of SAMSN1 contributes to MM development and this requires further investigation.

The structural nature of SAMSN1, specifically the presence of SH3 and SAM domains, places it in the class of intracellular adaptor molecules [46], [47]. Therefore, it is probable that key interactions and signaling pathways are modulated through SAMSN1 and the deregulation of these pathways contributes to the development of MM. Further investigation is required to identify a mechanism of action for SAMSN1 in suppressing MM tumor development, including potentially important interacting partners and downstream activated and/or repressed pathways.

Conclusions

In summary, we have identified Samsn1 as an MM tumor suppressor gene with reduced SAMSN1 expression observed both in the C57BL/KaLwRij mouse model of MM and a proportion of patients with MM. Further studies of the prognostic significance of reduced SAMSN1 expression in patients with MM are warranted. Identification of differential methylation of the SAMSN1 promoter suggests that demethylating agents may be clinically useful in the treatment of MM. Furthermore, investigation into the biologic outcomes of SAMSN1 molecular interactions may reveal novel therapeutic targets for the treatment of MM. Finally, further analysis of the C57BL/KaLwRij genome may provide key insights into genes that may be involved in the initiation and/or progression of MM disease.

The following are the supplementary data related to this article.

Supplementary Figures

Supplementary Table S1

PCR Primer Details