Three to five millilitres of bone marrow samples was collected from leftover specimens from the Human Genetic Laboratory, Department of Pathology, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand during February 2017 to October 2018. Bone marrow samples of 15 newly diagnosed MM patients were divided into two groups. Five samples were allocated to the high cytogenetic risk MM group, and the other 10 samples were assigned to the standard cytogenetic risk MM group. Among those samples, four of five samples in high-risk group were subdivided into t(4;14) and non-t(4;14) subgroups in order to perform the functional enrichment analysis comparison. Group classification was determined according to the criteria set forth in the International Myeloma Working Group consensus multiple myeloma with high-risk cytogenetics 2016. The clinical and laboratory features including serum β₂-microglobulin (β₂M), serum creatinine, serum hemoglobin, follow-up time and clinical response were collected from medical records. The clinical and laboratory features of 15 MM patients were shown in Table 1.

Plasma cells (PCs) were isolated from bone marrow samples by immunomagnetic bead selection with monoclonal mouse antihuman CD138 antibodies using a magnetic column (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. PC purity was confirmed by MACs technique as greater than 95% CD138 + cells and morphology by Wright-Giemsa staining. RNA was extracted using a mirVana™ miRNA Isolation Kit (Life Technologies, Carlsbad, CA, United States), and its concentration was determined using a NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA, United States) according to the manufacturer’s instructions.

Gene expression profiling was performed on an Affymetrix GeneChip Human Clariom S Pico Assay (Affymetrix, Santa Clara, CA, United States). All CEL files were analyzed using Transcriptome Analysis Console (TAC) 4.0 software (Thermo Fisher Scientific). Expression analysis to determine genes that significantly differentially express was performed using analysis of variance (ANOVA). Significant differential expression was defined as gene level fold change less than −2 or more than 2, and a p-value < 0.05. A Probeset (Gene/Exon) was considered expressed if ≥ 50% of samples had a detection above background (DABG) value below the DABG threshold (DABG <0.05). Functional enrichment analysis was performed using 3 types of software. including Gene Set Enrichment Analysis (GSEA), FUNRICH, and Gene Ontology (GO) Panther pathway. For GSEA, a collection of annotated gene sets was assembled and uploaded into Molecular Signatures Database version 7.0 (updated August 2019), which was used as a gene sets database. The chip platform used in this study was Clariom S Human HT.r1.chip. The weighted enrichment statistic was used to analyze the expression profile of each sample. The metric for ranking genes was set as ratio of classes. Genes from expression profile analysis were evaluated using FunRich software, and there were 941 of 960 genes mapped in the FUNRICH database. Hypergenometric p-value test was performed against all 20,515 genes in the FunRich database. All of the 960 genes identified from gene expression profiling were analyzed in the GO Panther pathway. Over-representation test was performed using version 14.1 PANTHER annotation platform. All genes in the database of Homo sapiens were selected as the reference list. Fisher's exact test was used to compare data from the PANTHER pathways annotation dataset. Protein-protein interaction (PPI) network was created using STRING database version 1 [23]. We used search tool for multiple proteins by name to analyze the PPI network.

MM cell lines, including KMM1 and KMS11, were kindly provided by Professor Seiji Okada of the Center of AIDS Research, Kumamoto University, Kumamoto, Japan. All cell lines were cultured in Roswell Park Memorial Institute (RPMI)-1640 medium supplemented with 15% fetal bovine serum (FBS) (Thermo Fisher Scientific) and 100 U/ml penicillin-streptomycin at 37°C in a humidified atmosphere containing 5% CO₂. All cell lines were treated with UCHL1 inhibitor LDN-5744 (Sigma-Aldrich Corporation, St. Louis, MO, United States) at concentrations of 10, 20, and 40 ng/μl for 24 h and DMSO (AMRESCO) as control. These doses were selected based on previous IC50 studies that showed a specific inhibition of the proteasome activity for UCH-L1 [24]. Cells were then counted using trypan blue staining. The cells obtained from both treated and untreated conditions at about 1 × 10⁶ cells were analyzed for DNA index by propidium iodide (PI) staining. The cells were centrifuged at 450 g for 3 min and then the supernatant was discarded. After that, 1 µl of PI/RNase staining solution (BD Biosciences, Franklin Lakes, NJ, United States) and 1 µl of Triton X-100 solution (Fluka Analytical, Buchs, Switzerland) were added and incubated for 20 min. The cells were then analyzed using a Cytomics FC 500 flow cytometer (Beckman Coulter, Brea, CA, United States). Independent-samples t test was used as statistical testing of cell counting by trypan blue staining and DNA index.

Total RNA was reverse transcribed into cDNA using a SuperScript VILOTM cDNA Synthesis Kit (Thermo Fisher Scientific). Real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) was performed using Express SYBR GreenER™ qPCR SuperMix (Life Technologies) and normalized to RNA level by GAPDH. The primer sequences were, as follows: UCHL1—sense 5′-CCC​AGC​ATG​AGA​ACT​TCA​GG-3′ and anti-sense 5′-CAC​AGG​AAT​TCC​CAA​TGG​TC-3′; CCND2—sense 5′-GCG​GAG​AAG​CTG​TGC​ATT​TA-3′ and anti-sense 5′-CTG​CCA​GGT​TCC​ACT​TCA​AC-3′; CDK4—sense 5′-GTG​GAA​ACT​CTG​AAG​CCG​AC-3′, and anti-sense 5′-AAG​TCA​GCA​TTT​CCA​GCA​GC-3′; mTOR—sense 5′-ACC​TCA​CAA​GAC​ATC​GCT​GA-3′ and anti-sense 5′-CTC​TCT​CAC​CCA​GCA​GAA​CA-3′; CTNNB1—sense 5′-AAG​GTA​GAG​TGA​TGA​AAG​TTG​TT-3′ and anti-sense 5′-CAC​CAT​GTC​CTC​TGT​CTA​TTC-3′; and, GAPDH—sense 5′-CCT​GTT​CGA​CAG​TCA​GCC​G-3′ and anti-sense 5′-CGA​CCA​AAT​CCG​TTG​ACT​CC-3′. All primers were synthesized by Macrogen (Seoul, Korea). Independent-samples t test was used as statistical testing of gene expression comparing of standard-risk and high-risk MM subgroups.

Cell lines were lyzed in radioimmunoprecipitation assay (RIPA) buffer (Cell Signaling Technology, Danvers, MA, United States) at 4°C for 5 min. The cell lysate was then centrifuged at 14,000g for 5 min and the supernatant was collected. Proteins were mixed with 1:1 Laemmli Sample Buffer (Bio-Rad Laboratories, Hercules, CA, United States), and then incubated at 95°C for 5 min. A slab of 10% SDS-polyacrylamide gel was prepared using a TGX FastCast™ Kit (Bio-Rad Laboratories). Proteins were separated by gel electrophoresis at 100 V for 1 h. After that, proteins were transferred to a membrane in transfer buffer (Bio-Rad Laboratories) at 100 V for 2 h, and then the membrane was blocked with blocking buffer that contained 5% non-fat dry milk in Tris-buffered saline with Tween 20 (TBST). The membrane was incubated with primary antibody UCHL1 (Cell Signaling Technology) and GAPDH (Abcam, Cambridge, United Kingdom) as an internal reference at room temperature for overnight and then incubated with secondary antibody anti-rabbit IgG, HRP-linked antibody at room temperature for 1 h (Cell Signaling Technology). The intensity of western blot results were performed by GeneTools software.

We included 15 newly diagnosed MM patients in this study. Risk classification of patients was based on the IMWG consensus of risk stratification in MM [25]. There were 10 patients categorized as standard-risk, and five patients were allocated to the high-risk group. The clinical and laboratory features of the 15 included MM patients are presented in Table 1. Cytogenetic analysis by combination standard karyotyping and fluorescence in situ hybridization (FISH) revealed that t(4;14) and aberration on chromosome 13 were the most commonly observed cytogenetic abnormality in this study with a positivity rate among all cases of 47%. Four out of five cases in the high-risk group harbored a t(4;14). In addition, copy number gain of CCDN1 was positive in 27% of all cases in this study while previous report about CCDN1 copy number variation in Thai population was 15% [26]. Other genetic abnormalities observed in this study included copy number variation of CCND1 (27%), FGFR3 (20%), TP53 (20%), IGH (7%), CCND3 (7%), and D17Z1 (7%). The overall genetic aberrations observed in this study are shown in Figure 1. Patient age ranged from 51 to 87 years, with a median of 65 years. The average values of β₂-microglobulin, serum creatinine, and hemoglobin concentration are shown in Table 1. High-risk MM patients had a significantly higher β₂-microglobulin value compared to standard-risk MM (p < 0.05). In addition, the median follow-up time of samples in the high-risk group was lower than the median follow-up time of all samples (30 months). Serum creatinine and hemoglobin concentrations were not significantly different between the high-risk group and standard-risk groups (Table 1).

To identify genes and pathways that were differentially expressed in the high-risk group compared with the standard-risk group, we performed gene expression profiling using Clariom™ S Array GeneChip from Affymetrix, and we applied ANOVA to filter the differentially expressed genes using TAC software. As expected, the expression profiles of high-risk and standard-risk MM exhibited a high degree of difference in gene transcription levels. The hierarchical plot that is shown in Figure 2A clearly demonstrates the differences in cytogenetic risk between groups. The top 20 upregulated and downregulated genes compared between the high-risk and standard-risk groups are shown in Figure 2A. Interestingly, the highest differentially overexpressed gene is UCHL1 and we found CCND2, one of its downstream target, also overexpressed in the top 20 upregulated genes. Details specific to the top 50 upregulated and downregulated genes are shown in Supplementary Tables S1 and S2, respectively. There were 960 genes differentially expressed between the high-risk and standard-risk groups with a fold change greater than the 2.0 cutoff (p < 0.05) (Figure 2B). Among those 960 signatures, there were 610 genes (63.54%) found to be overexpressed (fold change >2.0), as shown by Venn diagram comparing the high-risk group with the standard-risk group in Figure 2C. Interestingly, several upregulated genes in high-risk MM were commonly involved in cell cycle control, the Wnt signaling pathway, and the ubiquitin proteasome pathway. In contrast, there were 350 genes (36.46%) found to be significantly downregulated (fold change <2.0, p < 0.05), as shown by Venn diagram comparing the high-risk group with the standard-risk group in Figure 2D. Collectively, these findings further highlight the difference in the transcriptional signatures of these two cytogenetic risk subgroups of MM. Additionally, dysregulation of a key component of the proteasome pathway, UCHL1 might contribute to high-risk transcriptional signature by disrupting several components of cell cycle control, including CCND2.

To explore the certain biological process or molecular function of differentially expressed genes, we performed three different classes of pathway analysis, including Gene Set Enrichment Analysis (GSEA), FUNRICH, and Gene Ontology (GO) Panther pathway enrichment analysis, to analyze our data. GSEA revealed that 16 gene sets were significantly enriched at a nominal p-value < 5% and a false discovery rate (FDR) < 25%, as shown in Supplementary Table S3. These included hallmark of mTOR signaling, G2M cell cycle checkpoint, E2F targets, MYC targets, and protein secretion pathways (Figure 3). Similar to GSEA, FUNRICH pathway analysis showed several pathways associated with tumorigenesis and tumor progression to be enriched. Those included Wnt signaling, cell cycle control, G2M cell cycle checkpoint, ubiquitin-dependent degradation of cyclin D, mTORC1 signaling, PI3K-AKT-mTOR signaling, and DNA repair pathways (Figure 4). Consistent to GSEA and FUNRICH, GO enrichment analysis further confirmed cell cycle control and the ubiquitin proteasome pathway to be enriched in the high-risk MM group (Figure 5). Taken together, our data indicate the disruption of several molecular signaling pathways in genes differentially expressed among t(4;14) MM and non-t(4;14) MM. Additionally, these data further highlight the alteration of the ubiquitin proteasome pathway, and they suggest that UCHL1 influences tumorigenesis and tumor progression among t(4;14) MM.

Gene set enrichment and pathway analysis data both demonstrated that the ubiquitin proteasome pathway via UCHL1 and CCND2-dependent cell cycle control pathways are copersistent and differentially expressed in t(4;14) high cytogenetic risk MM compared to the other non-t(4;14) MM group. It was earlier reported that UCHL1 regulates cell cycle control and enhances cell proliferation by activation of several cyclin-dependent kinase proteins [21]. Perhaps more importantly, previous study found that UCHL1 promotes progression of colorectal cancer cell lines via the activation of CCND2, which is a downstream target of the Wnt β-catenin/TCF pathway [18]. Furthermore, UCHL1 activates the Akt pathways to promote the progression of osteosarcoma [27] and breast cancer [28]. These findings prompted us to further investigate whether key components of those pathways were differentially expressed in our tested samples. By RT-qPCR, UCHL1 and its downstream target CCDN2 were significantly differentially overexpressed in the high cytogenetic risk group (p < 0.05) (Figure 6). Our data further highlight the copersistence of upregulation of UCHL1 and CCND2 in t(4;14) high cytogenetic risk MM, which may be involved in disease progression in this MM subgroup.

To gain further insights into the mechanism of UCHL1 that is responsible for the progression of t(4;14) cytogenetic high-risk MM, we performed in vitro pharmacological targeting of UCHL1 in high-risk with t(4;14) (KMS11) and non-t(4;14) (KMM1) cell lines using UCHL1 inhibitor (LDN-57444). LDN-57444 has been used to inhibit UCHL1 hydrolase activity leading to cell death via apoptosis by decreasing the activity of ubiquitin proteasome [29]. Both KMS11 and KMM1 were treated with 20 ng/μl of LDN-57444 for 24 h. Cell proliferation results established by using tryphan blue staining showed significantly decreased cell numbers in KMM1 and KMS11 cell lines (p-value < 0.05) as shown in Supplementary Figure S1. The efficacy of LDN-57444 to target UCHL1 was confirmed Western blotting. As expected, targeting of UCHL1 was able to suppress the protein level and proliferation of both KMS11 and KMM1 cell lines in vitro. Interestingly, KMS11 cells with phenotypically overexpressed UCHL1 protein were strongly suppressed by LDN-57444 compared to KMM1 cells with a low level of UCHL1 protein (Figure 7A). The intensity of western blot results of treated and untreated conditions in KMM1 were 6.85 × 10⁴ and 1.62 × 10⁵ respectively, whereas in KMS11 were 3.8 × 10⁵ and 3.44 × 10⁶, respectively. Additionally, cell cycle arrest in G1 was observed in both KMS11 and KMM1 cell lines after treatment with UCHL1 inhibitor (Figure 7B). Statistical analysis demonstrated that both cell lines had significantly increased proportion in G1 phase after treatment (Figures 7C,D). These results suggest that both cell lines were sensitized by UCHL1 inhibitor resulting in increased cell cycle proportion in G1 phase even though in KMS11 cells with phenotypic overexpression of UCHL1. We further investigated the crosstalk between UCHL1 and its downstream targets, including CCND2, CDK4, mTOR, and β-catenin. Using STRING database to predict protein-protein interaction (PPI). We found that UCHL1 is able to interact with CCND2 via CDKN1B and β-catenin (Supplementary Figure S2A). Another PPI network was generated using UCHL1 and the top 100 upregulated genes identified in the expression profile analysis in this study. The result showed that UCHL1 can interact with the molecular network involved with CCND2 by indirect interaction (Supplementary Figure S2B). Moreover, functional enrichment in the predicted protein network was mostly observed in proteins that influence cell cycle regulatory pathways, which is similar to our findings from functional enrichment analysis (Supplementary Tables S3–S5). Taken together, our data further highlight the potential molecular mechanism of UCHL1 to manipulate the cell cycle in MM cell lines.