Abstract
In a comparison of gene expression profile in unsorted bone marrow (BM) samples from patients with multiple myeloma (MM), acute leukemia, and diffuse large B-cell lymphoma infiltrating the BM, the leading myeloma distinguishing gene was GPRC5D. This gene was highly expressed in BM samples from the 10 MM cases examined as opposed to minimal expression in samples from the eight cases with other hematological malignancies. Moreover, following antimyeloma treatment the expression of GPRC5D decreased several folds. The strong and selective expression of GPRC5D in MM cells makes this gene and its encoded surface protein as promising markers for monitoring the tumor load and hopefully also as targets for antimyeloma antibodies.
Gene expression datasets
Introduction
Multiple myeloma (MM) cells are easily recognized by morphology and expression of certain antigens such as CD138. Still, no single marker has been proved to be useful for monitoring the tumor load directly, although multi-parameter flow cytometry applying a panel of antibodies is gaining interest.Citation1 Although SLAMF7 (CS1) is strongly expressed on the surface of malignant and normal plasma cells, its role as a tumor marker has not yet been tested; however, recently its selective expression has been utilized therapeutically.Citation2 Here, we present our gene expression profile (GEP) data which exhibited another myeloma expressed gene with a high preponderance to the tumor cells.
Methods
Sample collections
Bone marrow (BM) aspiration was collected into a heparinized syringe and two drops (∼100 µl) were transferred immediately into an Eppendorf tube and flash frozen with liquid nitrogen. Subsequent samples were frozen after longer incubation in the syringe pool (left in room temperature). Alternatively, two drops were transferred into tubes containing 1.2 ml RNAlater (Ambion, Inc., Austin, TX, USA).
Patients
Newly diagnosed patients with clinical and laboratory manifestations typical for MM with a high tumor load (severe cytopenias, leukemic phase, extreme paraprotein levels, multiple lytic lesions, hypercalcemia) entered the study. Also included were patients with acute leukemia presenting with severe cytopenias (grade 4) and no mature leukocytes seen in the peripheral blood. The samples from the leukemia cases were analyzed only if cytospine smears from the aspirates confirmed complete ‘replacement’ (∼100%) by tumor cells. Another case had leukemic relapse of diffuse large B-cell lymphoma (DLBCL). Patients provided written informed consent. The study was approved by Asaf Herofhe Institute, Zerifin, Israel review boards.
RNA extraction
Frozen samples were lyzed by adding 300 µl lysis buffer to tubes. Samples stored in RNAlater were fractionated and lyzed by 10× volumes (v/v) lysis buffer. RNA was isolated by MagNA Pure Compact RNA Isolation procedure using the MagNA Pure Compact instrument (Roche, Applied Science, Indianapolis, IN, USA). The integrity of RNA was examined by an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA).
Gene expression profile
Biotin-labeled cRNA was generated from 200 ng total RNA, hybridized onto GeneChip Human Gene1.0 ST Array (Affymetrix, Santa Clara, CA, USA), and the data were processed with the Affymetrix GeneChip Scanner 3000 and Affymetrix Expression Console. Normalization was done by the RMA method and fold change results were calculated relative to the reference sample of each case (which was fixed without delay). The dataset is accessible from the NCBI Gene Expression Omnibus GSE36036 and GSE39184:
Results
Our series included an overall group of 10 MM cases, each having several unsorted BM samples which were fixed at variable time points following aspiration and analyzed by microarray ().Citation3 GEP was also analyzed from the BM samples of six acute myeloid leukemia (AML) cases, one case with mixed lineage leukemia (AML/T-ALL) and another case with leukemic relapse of DLBCL (). A comparison of the GEPs analyzed from the BM samples of the 10 MM cases with the BM samples obtained from the patients with the various leukemia subtypes identified GPRC5D as the leading myeloma distinguishing gene in our series. More specifically, the average expression signal (log2) of GPRC5D in the BM samples from the eight MM cases with ≥70% plasma cells in the aspirates was 11.42 (10.96–11.88; CI, 0.05) as compared with 8.98 (8.63–9.32; CI, 0.05) in the samples from the two cases with a lower proportion of plasma cells, and only 5.06 (4.86–5.26; CI, 0.05) in the BM samples from the AML, AML/ALL, and DLBCL cases (). Furthermore, in one MM patient (case #8) who was re-examined after completing induction therapy (with bortezomib and dexamethasone) the average signal intensity of GPRC5D dropped from 10.6 prior to treatment to 7.6 during morphological complete remission ().
Table 1. Patient characteristics and the expression signal intensities
Table 2. Average expression signals
Discussion
To uncover the genes that are differently expressed in the myelomatous BM in comparison with BM tissues without MM, we compared GEP in BM samples from these patient groups. The results pointed towards GPRC5D as the most promising candidate gene for distinguishing MM cells from either normal or abnormal BM cell populations other than MM including the common leukemia and lymphoma subtypes. This conclusion is consistent with the high GPRC5D expression recorded from sorted (‘pure’) MM-cell populations in large series like GDS1284 where the average log2 signal intensity value of GPRC5D was 12.1 (11.9–12.2; CI, 0.05) in contrast to the minimal expression of this gene in lymphoblastic and myeloid leukemia datasets such as GDS2493, where the average expression of GPRC5D was 6.88 (6.76–6.98, CI, 0.05) versus 12.07 (11.87–12.45) for GAPDH or GDS3312 in which the average signal intensity of GPRC5D was 6.76 (6.72–6.81) versus 13.70 (13.60–13.80) for GAPDH. Similar differences were seen in unsorted BM samples from 48 cases with MM using reverse transcriptase-polymerase chain reaction (RT-PCR), where GPRC5D expression was highly variable (median, 288; quartiles, 17–928) compared to only low expression in normal tissues (median, 1; quartiles, 1–23). In the latter study however the authors concluded that the intensity of GPRC5D expression can be used as an independent prognostic factor in MM because higher expression was associated with reduced overall survival.Citation4 Although microarray and RT-PCR are incomparable, the 3-log reduction in GPRC5D expression observed in our series following antimyeloma treatment (case #8) and the lower GPRC5D signals obtained when the proportions of myeloma plasma cells in the marrow were lower () together with the consistently high GPRC5D expression with a small variability in the signal intensities seen in series which examined sorted MM-cell populations (like GDS1284) argue against the real existence of a large inherent variability in the expression of GPRC5D among MM-cell populations from different individuals, and suggest instead that the inverse correlation between GPRC5D expression and the overall survival seen in the Atamaniuk et al. series resulted from differences in the tumor burden. As such, measurement of GPRC5D expression in unsorted BM samples can be utilized for quantifying and monitoring the tumor load in MM. Furthermore, the still higher expression of GPRC5D in the aspiration samples from case #8 during morphological complete remission (combined with drop in her paraprotein levels from 6.41 g/dl to undetectable levels) compared with the expression of this gene in the non-myeloma samples raises the possibility that GPRC5D is also expressed in residual less defined clonotypic cells,Citation5 the persistence of which could explain the rapid and full blown relapse evolved within few weeks after the remission was documented. Finally, the high and selective expression of GPRC5D in MM cells as evident from our and other series and the surface localization of the encoded protein put GPRC5D in line with SLAMF7 in relation to the potential to serve as a specific target for antimyeloma antibodies. In summary, our data suggest that GPRC5D can be an ideal tumor marker for quantifying and monitoring the tumor load in myeloma objectively. The encoded surface protein can be an attractive target for antimyeloma antibodies.
Author's contribution
YC created and designed research, performed research, collected and analyzed data, and wrote the paper; all authors assisted in the recruitment of the patients, technical assist, advices, and gave final approval of the manuscript for publication.
Acknowledgements
The authors thank Dr Relly Forer, Dr Inna Vulin, and Dr Dina Volodarsky from Dyn Diagnostics Ltd for professional microarray and bioinformatic service and to Dr Yizhar Hardan for his kind support.
References
- Paiva B, Martinez-Lopez J, Vidriales MB, Mateos MV, Montalban MA, Fernandez-Redondo E, et al. Comparison of immunofixation, serum free light chain, and immunophenotyping for response evaluation and prognostication in multiple myeloma. J Clin Oncol. 2011;29:1627–33.
- Benson DM, Byrd JC. CS1-directed monoclonal antibody therapy for multiple myeloma. J Clin Oncol. 2012;30:2013–5.
- Cohen Y, Garach-Jehoshua O, Bar-Haim A, Kornberg A. Niche modulated and niche modulating genes in multiple myeloma. Blood Cancer J. 2012;2:e97.
- Atamaniuk J, Gleiss A, Porpaczy E, Kainz B, Grunt TW, Raderer M, et al. Overexpression of G protein-coupled receptor 5D in the bone marrow is associated with poor prognosis in patients with multiple myeloma. Eur J Clin Invest. 2012;42:953–60.
- Pilarski LM, Baigorri E, Mant MJ, Pilarski PM, Adamson P, Zola H, et al. Multiple myeloma includes phenotypically defined subsets of clonotypic CD20+ B cells that persist during treatment with rituximab. Clin Med Oncol. 2008;2:275–87.