DEK protein level is a biomarker of CD138positive normal and malignant plasma cells

Overexpression of DEK oncogene is associated with increased proliferation of carcinoma cells and it is observed in several solid tumors due to the amplification of the 6p22.3 chromosomal region where DEK locates. Although the same chromosomal amplification occurs in multiple myeloma (MM), a plasma cell neoplasm, whether the expression and the copy number of the DEK gene are affected in MM remains elusive. We show that despite the increased copy number in CD138positive MM cells (4 out of 41 MM samples), DEK mRNA expression was down-regulated compared with that in CD138negative bone marrow (BM) cells of the same patients (P<0.0001). DEK protein was not detectable by immunohistochemistry (IHC) in CD138positive normal plasma cells or in malignant plasma cells of MM patients (n = 56) whereas it was widely expressed in normal and neoplastic B-cells. Stable knockdown or overexpression of DEK in CD138positive MM cell lines did not affect the proliferation and viability of the cells profoundly in the presence or absence of chemotherapeutic agent melphalan whereas knockdown of DEK moderately but significantly increased the expression level of CD138 (p<0.01). Decreased DEK expression in plasma cells suggests a potential role of this gene in plasma cell development and lack of detectable DEK protein by IHC could be used as a biomarker for normal and malignant plasma cells.


Introduction
Multiple myeloma (MM) is a malignancy characterized by invasion of the bone marrow (BM) and bones with abnormal plasma cells that are expanded clonally [1,2]. Cytogenetically, aberrations in MM can be divided into those carrying balanced translocations typically involving PLOS  the immunoglobulin heavy chain gene and those carrying numerical changes. The latter often involve trisomies but may comprise recurrent deletions or gains of subchromosomal material as well, including gains of 6p22.3-p21. 3, found in about 16% of MM patients [3,4].
The DEK oncogene, located on 6p22.3, was initially identified in acute myeloid leukemia as a partner of the DEK-CAN fusion gene [5]. It encodes a nuclear protein [6], which is highly expressed in proliferating cells, and it participates in several cellular processes, including chromatin modeling and inhibition of senescence [7,8]. DEK expression is upregulated, most commonly in association with amplification of the genetic locus, in several solid tumors including breast cancer [9,10], melanoma [11], bladder cancer [12], and retinoblastoma [13]. Consistently, DEK overexpression transforms epithelial cells and promotes cancer in mouse models, whereas DEK knockdown induces cell death in tumor cells but not in differentiated epithelial cells [14]. Although DEK has been shown to contribute to myeloid differentiation of hematopoietic stem/precursor cells and cell lines [11,15,16], it remains to be determined whether its expression affects the biology and function of normal and neoplastic plasma cells, especially in the context of 6p amplification.
Here we determined the expression level and copy number of the DEK gene in MM cells. To this end, we used formalin fixed paraffin embedded (FFPE) BM samples as well as CD138 positive (malignant plasma cells) and CD138 negative cells isolated from fresh or frozen BM samples of MM patients and analyzed the copy number and expression level of the DEK gene using qPCR and RT-qPCR, respectively. IHC analysis with antibodies against DEK and CD138 was performed on the FFPE samples of MM and monoclonal gammapathies of uncertain significance (MGUS) patients, the latter of whom carry a risk of progression to symptomatic MM of approximately 1% per year. Additional IHC analysis was also performed on the FFPE samples of Burkitt lymphoma (BL), mantle zone lymphoma (MZL) and diffuse large B cell lymphoma (DLBCL) patients. Finally, we stably knocked-down or overexpressed DEK in MM cell lines to determine if change in DEK expression influences the expression level of CD138 and the growth of MM cells in the presence or absence of the chemotherapy agent melphalan.

Patient samples
FFPE BM tissues of patient samples were obtained from Vanderbilt University (MM (n = 26), MGUS (n = 12), and control BM (n = 9), BL (n = 3), MZL (n = 7) and DLBCL (n = 12)) and Istanbul University, Istanbul Medical Faculty, Department of Pathology (MM (n = 30), control BM (n = 9)). CD138 positive and CD138 negative cells were isolated from 41 fresh/frozen BM samples of MM patients (Vanderbilt University), 12 of which were obtained concurrently with the FFPE samples listed above. All samples were obtained at diagnosis. The stage of disease was determined by Durie-Slamon criteria [17]. The study was approved by the Institutional Review Boards of Vanderbilt University and Istanbul University and informed consent was obtained from patients in accordance with the Declaration of Helsinki.

RNA isolation and RT-qPCR
Total RNA isolation was performed by using the Ambion RecoverAll™ Total Nucleic Acid Isolation Kit (Life Technologies, Grand Island, NY) and cDNA synthesis was performed using the High Capacity cDNA Reverse Transcriptase Kit (Life Technologies, Grand Island, NY) following the manufacturer's instructions. TaqMan, Applied Biosystems primer probes for ABL1 (Hs01104728_m1) and/or human ACTB (Hs01060665_g1) were used to normalize the expression of human DEK (Hs00180127_m1) or CD138 (Hs00896423_m1). Relative DEK expression was calculated using three pooled healthy BM samples as a calibrator based on the 2 [-ΔΔCT] method or by using a standard curve method. Samples with a cycle time (Ct) value 40 were deemed positive for the expression of the analyzed gene. All reactions were performed in triplicate.
Determination of the DEK gene copy number DNA was isolated using the QIAamp DNA Blood Mini Kit (Qiagen, Valencia, CA) and qPCR was performed in quadruplicate using 20 ng DNA and primer probe sets for RNase P and DEK (Applied Biosystems, 4403326 and Hs04904663_cn, respectively) following the manufacturer's instructions. qPCR values of the DEK gene were normalized against those of RNase P (2 copies per genome) and a relative copy number of DEK was calculated using control BM DNA (three pooled samples) as a calibrator based on the 2 [-ΔΔCT] method. Normalized qPCR values of control, carrying two gene copies, were scored as "1", while values of patient samples !1.5 fold or 0.5 fold were considered to represent an increased or decreased gene copy number, respectively.
Cell lines, growth curve and cell cycle analysis U266 (TIB-196) or RPMI-8226 (CCL-155) cells were obtained from ATCC (Manassas, VA) and maintained in RPMI-1640 containing glutamine (Gibco, ThermoFisher Scientific, 52400025, MA, USA), 15% fetal bovine serum (Gibco, 10270106), and 1% penicillin/streptomycin (Gibco, 15140122). For cell cycle and growth curve analyses, FACS-sorted GFP positive cells were seeded (2x10 5 cells/ml) into 6-well plates in 3ml medium in the presence of 10 or 20 μM melphalan or vehicle (acid alcohol), in duplicate daily, and the total cell number was counted using a Countess cell counter (Invitrogen, ThermoFisher Scientific, C10227). Viability was calculated for each analyzed time point (viable cell number/total cell number x 100) and the percentage of viable cells was represented relative to the viability at "time zero". Cell cycle analysis was performed using propidium iodide staining (PI) and FACS.

Statistical analysis
Statistical analyses were performed using GraphPad Prism, version 5.0 for Windows (Graph-Pad Software; www.graphpad.com). P values lower than 0.05 were considered significant.

Total BM samples of MM patients show decreased DEK mRNA expression
To determine the expression level of DEK mRNA in MM patients, we first analyzed the archival FFPE BM samples from MM patients and healthy controls using RT-qPCR. Using two different house-keeping genes (ACTB and ABL1) as normalizer [20,21] we found that although the expression level of DEK was similar between the control BM (n = 8) and patient samples with stage-I (n = 1) and stage-II (n = 7) MM (Fig 1A and 1B), there was no detectable DEK mRNA expression in 10 out of 21 patients with stage-III MM (Table 1). Moreover, when the patients with stage-III MM lacking DEK expression were excluded from the data, expression levels of DEK in the rest of the stage-III patient samples was significantly lower than those of healthy controls (P = 0.0006 (for DEK/ABL1) and P = 0.0096 (for DEK/ACTB)) and patients  with stage-I/stage-II MM (P = 0.0035 for DEK/ABL1 and P = 0.0036 for DEK/ACTB)) ( Fig 1A  and 1B). These data suggested that DEK expression is down regulated in the BM of stage-III MM patients.

DEK mRNA expression is reduced in CD138 positive plasma cells in MM
We hypothesized that the observed decrease in DEK expression in BM samples of MM patients reflects the level of DEK in malignant plasma cells, since the CD138 positive neoplastic plasma cell (aka MM cells) burden in these marrows was increased (greater than 10% of the total cellularity). To test this hypothesis, CD138 positive and CD138 negative cells were isolated (Fig 1C) from the fresh or frozen BM samples of MM patients (n = 41) and examined using RT-qPCR. Consistent with the FFPE RT-qPCR results, we found an overall statistically significant decreased DEK expression in the CD138 positive cells compared to both CD138 negative cells of the same patients (P<0.0001) and to the control total BM cells (P = 0.0009) (Fig 1D).  (Fig 1E). Together, these data indicated that DEK mRNA expression was significantly reduced in CD138 positive MM cells.

Copy number changes of the DEK gene in MM cells
Given that gains of the chromosome region 6p22.3 is observed in 16% of MM patients [4], we hypothesized initially that the DEK gene would be amplified and overexpressed in MM patients who carry 6p22.3 amplification, as observed in the other solid tumors [12,13]. To our surprise, none of the patient samples in our cohort showed DEK overexpression in CD138 positive MM cells (Fig 1E) suggesting that either DEK was not amplifid or its amplification did not cause overexpression in the CD138 positive MM cells. To distinguish between these possibilities, we analyzed the copy number variation (CNV) of the DEK gene (DEK-CNV) in the CD138 positive cells of these 41 samples using a qPCR assay. We found that the DEK allele was amplified in the CD138 positive MM cells in 4 of 41 samples (!1.5 fold, 3 copies or more) compared with the control DNA obtained from total BM samples of healthy donors (3 pooled DNAs) (Fig 2A and Table 2). Moreover, DEK amplification was specific to the CD138 positive cells in these samples while the CD138 negative cells of the same patients contained two copies of the DEK gene ( Fig 2B and Table 2). Interestingly, amplification of the DEK gene in the CD138 positive MM cells did not increase the expression level of DEK mRNA. Compared to paired CD138 negative cells from the same patients, DEK was down-regulated in these CD138 positive MM cells by 15, 1.6, 14, and 3.3-fold, respectively ( Fig 1E and Table 2, patient numbers 12, 13, 18 and 25 marked by asterices). These results suggest that DEK mRNA expression is decreased in terminally differentiated plasma cells, regardless of the copy number of this gene.

Lack of detectable DEK expression by IHC in normal and malignant CD138 positive plasma cells
To determine the expression level of DEK protein in MM cells, we analyzed the FFPE BM samples of MM patients (n = 56) (including the FFPE samples of the 12 out of 41 fresh or frozen samples used in RT-qPCR analysis and 30 BM samples used in FFPE-RT-qPCR analysis) by using either a single or a double IHC staining with antibodies against DEK and CD138. We were not able to perform Western Blot analysis due to limitations in the number of cells that were separated based on CD138 expression. Therefore, we preferred IHC analysis since it is one of the most convenient routine test performed by pathologists for the archived FFPE samples, which provide information about cellular localization and expression level of proteins as well as morphology of the cells. We found a moderate to high level of DEK expression in myeloid and erythroid cells, whereas there was no detectable DEK protein in CD138 positive plasma cells in BM samples of controls (n = 8), MGUS or MM patients (Fig 3A). Similarly, IHC analysis of a proliferative lymph node, used as a positive control, showed DEK expression in lymphocytes, particularly in the germinal center cells, but not in the CD138 positive plasma cells (Fig 3B, left and right panels). DEK expression was also undetectable in CD138 positive MM cells of the FFPE samples, which were available from 3 MM patients which showed DEK amplification (Fig 3C). Finally, analysis of patients with B cell malignancies, which included BL, MZL and DLBCL, showed similar staining pattern with DEK and CD138 antibodies (Fig  4). These results suggest that the level of DEK expression in mature plasma cells was below the detection limit of the IHC assay and the lack of detectable DEK protein might be an additional useful negative marker for the detection of CD138 positive normal and malignant plasma cells.

Stable knockdown of DEK in MM cell lines moderately increases CD138 expression without a profound effect on the proliferation and viability
Next we interrogated the biological effects of altered DEK expression, given that DEK was down regulated in primary CD138 positive MM cells (Fig 1D and 1E). We stably suppressed DEK expression in the MM cell lines RPMI-8226 and U266 using sh-RNA lentiviral constructs targeting DEK mRNA. RT-qPCR ( Fig 5A) and Western blot analysis (Fig 5B) of transduced and FACS-sorted GFP positive cells confirmed the knockdown with two different shDEK lentiviruses in both cell lines. Growth curve analysis of RPMI-shDEK and U266-shDEK cells in the presence or absence of melphalan, one of the chemotherapeutic agents used in the treatment of MM patients, did not show a significant difference compared with control sh-Negative cells (Fig 6A). Melphalan treatment of the RPMI-shDEK and U266-shDEK cells resulted in a similar level of cell death ( Fig 6B) and arrest in the S-G2M phase of the cell cycle in all cases ( Fig  6C and 6D). Similarly, overexpression of DEK in both cell lines (FACS-sorted RPMI--DEK-GFP or U266-DEK-GFP) did not change the growth profile or melphalan response of the cells, compared with that of FACS-sorted RPMI or U266 cells transduced with control GFP-only virus (data not shown). Given that DEK expression was lower in the primary CD138 positive cells, next we tested the expression level of CD138 in MM cell lines using RT-qPCR and found a moderate (1.8 fold) but significant (P<0.01) increase in the expression level of CD138 in RPMI-shDEK cells compared to control RPMI-sh-Negative cells (Fig 7A). Flow cytometry ( Fig 7B) and immunocytochemical analyses (Fig 7C and 7D) of the same cells revealed a mild increase in CD138 protein on the surface of the RPMI-shDEK cells without a profound effect on the adhesion to fibronectin (data not shown), a process partly mediated by CD138. We did not observe a change in the expression level of CD138 when we overexpressed DEK in the parental RPMI-8226 cells (data not shown), suggesting a possible indirect association between decreased DEK and increased CD138 expression in RPMI-8226 cells.

Discussion
Here, we report that DEK mRNA and protein expression is decreased in normal plasma cells and MM cells that express CD138 regardless of the level of amplification of the DEK gene. This finding is distinct from the concerted findings of DEK amplification and overexpression in epithelial cancers. Due to limitations in our resources, we performed qPCR analysis, a commonly used and accepted technique for large scale analysis of CNV in various diseases [13,22], to detect the correlation between the CNV and mRNA expression of DEK. Despite the relatively small number of MM cases that we have analyzed, we were able to detect a copy number gain in DEK gene in 10% of the samples (4/41), which was comparable to the study of Walker et al. showing a gain in the chromosome 6p22.3-p21.31 region, where DEK locates, in 16% of MM samples (19/114) using a high-resolution single nucleotide polymorphism mapping array [4]. High level of DEK expression is associated with increased cellular proliferation in the epithelial cells [11,13,23,24] and B-lymphoid cells [25]. Consistently, our IHC analysis of a proliferative lymph node indicated high level of DEK expression in the mantle zone and germinal center B cells (Fig 3B), suggesting that DEK expression is high in earlier B cell ontogeny, but is downregulated in post-germinal center plasma cells. Therefore amplification of the DEK gene may result in different expressional outcomes in different types of B cells and their progeny, depending on the cell-specific regulation of DEK expression that might be mediated via transcriptional, post transcriptional or translational mechanisms. Supporting our hypothesis, expression analysis of amplified genes in gliomas showed that not all amplified genes in these tumors are overexpressed and the repressed expression patterns of the genes in original (normal) tissue are maintained in the tumor tissue despite the amplification of the genes [26]. Similarly, we showed that different B-cell malignancies including BL, MZL and DLBCL show high level of DEK expression in malignant B cells (Fig 4), which was similar to their normal counterpart ( Fig 3B).  Reduced DEK expression has distinct effects in different cell types. In epithelial cells, decreased DEK expression induces senescence and reduces tumor formation [7,14], whereas it increases the number of myeloid progenitor cells in mice and stimulates myeloid colony formation in vitro [11]. In our study, stable knockdown or the overexpression of DEK in MM cell lines did not affect the proliferation or viability of the cells (Fig 6A and 6B). Similar to the literature [27], melphalan treatment of the MM cell lines induced cell death and cell cycle arrest, a process which was not affected by knockdown (Fig 6C and 6D) or overexpression of DEK (data not shown). Interestingly, knockdown of DEK in RPMI-8226 cells, which are already 98% positive for CD138 [28], resulted in a mild but significant increase in the expression level of this gene (Fig 7). All together, our results suggest a potential association between reduced DEK expression and level of CD138 expression on the plasma cells, which we aim in the future to further investigate using the primary plasma cell progenitors.
In conclusion, we have found a surprising down-regulation of DEK expression specifically in CD138 positive plasma cells, even in the setting of copy number gains of the DEK gene associated with neoplasia. Our findings suggest that high levels of DEK expression might be required for the proper proliferation of primary B cells whereas its downregulation might contribute to the development of terminally differentiated plasma cells. This hypothesis will be the subject of future research aiming to understand the role of DEK in normal and malignant plasma cell development.