Prognosis Value of RBBP8 Expression in Plasma Cell Myeloma
Total Page:16
File Type:pdf, Size:1020Kb
Cancer Gene Therapy https://doi.org/10.1038/s41417-018-0069-3 ARTICLE Prognosis value of RBBP8 expression in plasma cell myeloma 1 2 3 4 5 1 1 1 Weilong Zhang ● Ying Song ● Xue He ● Xiaoni Liu ● Ye Zhang ● Zuozhen Yang ● Ping Yang ● Jing Wang ● 1 4 3 4 1 Kai Hu ● Weiyou Liu ● Xiuru Zhang ● Xiaoliang Yuan ● Hongmei Jing Received: 30 July 2018 / Revised: 30 October 2018 / Accepted: 2 November 2018 © The Author(s) 2019. This article is published with open access Abstract Plasma cell myeloma (PCM) secretes monoclonal immunoglobulin (Ig) by clonal plasma cells of abnormal proliferation in the bone marrow. As PCM is incurable, it is necessary to find new biomarkers to predict the prognosis and recurrence of PCM. The relationship between cancer and RBBP8 has not been fully studied. The role of RBBP8 in tumorigenesis remains inconsistent. We described the expression of RBBP8 in the gene expression profile of 1930 PCM samples (1878 PCM patients) from seven independent data sets. We analyzed the relationship between RBBP8 and survival prognosis, recurrence, and treatment response in patients with PCM, and the biological significance of RBBP8 in PCM. The gene expression level of RBBP8 was significantly related to the International staging system (ISS) grade of PCM (P = 0.0012). RBBP8 expression in different molecular subtypes was different (P < 2.2e-16). High RBBP8 expression is associated with 1234567890();,: 1234567890();,: poor survival in PCM (P < 0.0001). High expression of RBBP8 indicates that PCM patients are more likely to relapse (P = 0.0078). The biological significance of RBBP8 in PCM is related to the cell cycle (P < 0.05). High RBBP8 expression predicts poorer survival and more likely relapse in PCM. RBBP8 plays an important role in the cell cycle of PCM. RBBP8 can be considered an independent prognostic factor for PCM. RBBP8 can be used as a potential biomarker for assessing the prognosis of PCM patients. Introduction These authors contributed equally: Weilong Zhang, Ying Song, Xue Plasma cell myeloma (PCM) secretes monoclonal immu- He noglobulin (Ig) by clonal plasma cells of abnormal pro- liferation in the bone marrow. Clinical features of PCM Supplementary information The online version of this article (https:// doi.org/10.1038/s41417-018-0069-3) contains supplementary typically present with bone damage, hypercalcemia, renal material, which is available to authorized users. impairment, and anemia. PCM accounts for ~ 10% of cancers in the blood system [1]. The average median * Xiuru Zhang survival of the disease is 3–4 years. Costa LJ summarized [email protected] that the early mortality was 4–25% of newly diagnosed * Xiaoliang Yuan [email protected] PCM patients in a randomized phase 3 clinical trial in the * Hongmei Jing past decade [2]. However, he found that the early mor- [email protected] tality rate is more higher than what has been reported in clinical trials [2]. At present, the treatment of novel 1 Department of Hematology, Lymphoma Research Center, Peking myeloma drugs of immunomodulatory agents and pro- University Third Hospital, Beijing 100191, China teasome inhibitors improve survival rates [3, 4]. In addi- 2 Gannan Medical University, Ganzhou 341000, China tion, combinations of high dose-chemotherapy with 3 fi autologous stem cell transplantation (ASCT) resulted in Department of Pathology, Beijing Tiantan Hospital Af liated with – Capital Medical University, No. 6 Tiantan Xili, Beijing 100050, better overall survival (OS) [5 7]. However, there is still a China high recurrence rate in PCM. Therefore, it is necessary to fi 4 Department of Respiratory Medicine, The First Affiliated Hospital nd new biomarkers to predict the prognosis and recur- of Gannan Medical University, Ganzhou 341000, China rence of PCM. fi 5 Melbourne School of Population and Global Health, The According to the genetic classi cation, PCM can be University of Melbourne, Victoria 3010, Australia classified based on the translocation and cyclin D (TC) W. Zhang et al. and the University of Arkansas for Medical Sciences Materials and methods (UAMS) system. The TC classification distinguishes eight subgroups by the deactivation of primary immunoglobulin Data source and gene expression analysis H translocations and transcriptional activation of cyclin Dgene[8]. The UAMS molecular classification were Probe set measurements for all arrays were calculated using classified into seven subtypes by different gene expression the RMA (robust multiarray averaging) method. Logarith- profiles, including MMSET [t(4;14)], MAF [t(14;16)/t mic conversion of relative RNA expression values was (14;20)], CD1/2 [t(11;14), and t(6;14)], HY (hyper diploid performed using log2. According to the RBBP8 gene cluster), PR (proliferation), and a cluster mainly char- expression level, RBBP8-high group, and RBBP8-low acterized by a low percentage of bone disease (LB) [9]. group using survminer package with maximally selected BasedontheUAMSclassification in 2010, PCM is rank statistics arithmetic. Only genes with foldchanges reclassified as CD1, CD2, CTA, HY, MF, MS, myeloid, (log2) > 0.8 or < −0.8 were considered different expressed NFKB, and PR [10]. According to treatment response genes. P values < 0.05 were defined to be statistically sig- with bortezomib and dexamethasone (Dex), patients were nificant. We obtained seven independent data set totally divided into CR (complete response) group, PR (partial 1930 PCM samples (1878 PCM patients) from the seven response) group, MR (minimum response) group, NC (no independent Gene Expression Omnibus (GEO) data sets. change) group, and PD (progressive disease) group [11]. This study was in accordance with the Helsinki Declaration. Similarly, according to the treatment response with after GSE24080 of 559 patients were obtained from the GEO induction chemotherapy (pre-1st) and after ASCT, database. The gene expression was detected by Affymetrix patients were classified into CR group, VGPR (very good Human Genome U133 Plus 2.0 Array [25]. We analyzed partial response) group, PR group, NR (stable disease) the relationship between RBBP8 expression and Interna- group, and Prog (no response, progressive disease) group tional staging system (ISS) stage, 1q21 amplification, [12]. molecular subtype, and survival. The human RBBP8 (retinoblastoma-binding protein 8), GSE19784 of 311 patients were obtained from the GEO also known as CTIP (CTBP (C-terminal-binding protein)- database. The gene expression was detected by using interacting protein), is a protein coding gene. This protein Affymetrix GeneChip U133 plus 2.0 arrays [10]. Hier- interacts with other factors and participates in a variety of archical clustering identified 10 distinct subgroups. We nuclear pathways. Overexpression of RBBP8 in tumors is analyzed the relationship between RBBP8 expression and mainly related to cyclin D1 transcription [13]. CTIP/RBBP8 molecular subtype (9 subgroups). gene accelerates tumorigenesis through transcriptional GSE9782 of 477 patients were obtained from the GEO activity [14]. CTIP/RBBP8 was described as a key check- database. Gene expression profiling was detected using the point of G1 phase by initiation S-phase and DNA replication Affymetrix 133 A/B microarray [11]. We analyzed RBBP8 [15]. CTIP/RBBP8 mediates DNA double-strand breaks expression in different therapeutic response with bortezo- repair in the cell cycle through homologous recombination mib or dexamethasone (Dex). [16, 17]. CTIP/RBBP8 interacts with proliferating cell GSE83503 of 585 patients were obtained from the GEO nuclear antigen (PCNA) at specific localization and acti- database. The gene expression array was Affymetrix vates DNA damage checkpoints leading to DNA damage, Human Exon 1.0 ST Array [26]. We analyzed the rela- suppressing DNA replication at S and G2 phases [18]. tionship between RBBP8 and PCM recurrence. However, the relationship between cancer and RBBP8 has GSE82307 of 66 samples (33 patients) were obtained not been fully studied. from the GEO database. Samples were tested by Affymetrix Investigation found RBBP8 genes was associated with Human Genome U133 Plus 2.0 Array [27]. We analyzed sporadic brain arteriovenous malformations [19]. Advanced the relationship between RBBP8 expression in presentation invasive bladder cancer was associated with the deletion of (baseline) and recurrence. nuclear RBBP8 protein [20]. Deletion of the RBBP8 gene GSE19554 of 38 samples (19 patients) were obtained was associated with significantly worse prognosis in ovar- from the GEO database. Samples were tested by Affymetrix ian cancer [21]. The poor prognosis of breast cancer was Human Genome U133 Plus 2.0 Array [28]. We analyzed related to the low or no expression of RBBP8 [22, 23]. the relationship between RBBP8 expression at diagnosis RBBP8 is also overexpressed in certain tumors [24]. (baseline) and after induction chemotherapy (pre-1st). However, despite these associations with cancer, the GSE39754 of 136 patients were obtained from the GEO expression level of RBBP8 has not been reported in PCM. database. The gene expression was detected by Affymetrix We analyzed the association of RBBP8 expression with Human Exon 1.0 ST Array [12]. We analyzed RBBP8 PCM prognosis, relapse, and event-free survival (EFS) or expression in different therapeutic response with the pre-1st OS. and after ASCT. Prognosis value of RBBP8 expression in plasma cell myeloma Wilcoxon test. For all statistical methods, the P < 0.05 was considered to indicate statistical significance. Results Expression level of RBBP8 in different molecular subtypes of PCM The amplification of 1q21 was related to the expression level of RBBP8 (Fig. 1a, P = 0.0013, Kruskal–Wallis test). Compared with the two copies of 1q21 PCM samples, RBBP8 was significantly increased in the 1q21 amplifica- tion samples ( ≥ 4 copies) (Fig. 1a, P = 0.00055, Wilcoxon test). The expression of RBBP8 in different molecular subtypes was different (Fig. 1b, P < 2.2e-16, one-way ana- lysis of variance analysis test). The expressions of RBBP8 in CD1 and PR subtypes were increasing, compared with the mean of all subtypes (Fig. 1b, CD1, P ≤ 0.05; PR, P ≤ 0.0001, unpaired t test, two sided). Although MMSET subtype showed the lower RBBP8 expressions (Fig.