MALAT1 long non-coding RNA is overexpressed in multiple myeloma and may serve as a marker to predict disease progression

The study cohort included adult patients (aged 20 years and older) with multiple myeloma diagnosed at Kaohsiung Medical University Hospital from 2007 to 2012 who were free from other coexisting malignant diseases. The diagnosis of multiple myeloma was confirmed by bone marrow analysis which revealed a monoclonal plasma cell count over 10% by definition and related laboratory tests. The patients of extramedullary myeloma were not enrolled to this study. The diagnostic criteria, disease status and response to treatment were based on the criteria of the International Myeloma Working Group [17‐19]. Forty-five samples were collected from newly diagnosed patients (29 males, 16 females; median age 62.3 years, range 49 to 79 years) with different subtypes (IgG: 21, IgA: 13, light chain: 11) and clinical stages (Durie-Salmon stage 1: 1, stage 2: 6, stage 3: 38 or international staging system stage 1: 7, stage 2: 17, stage 3: 21). In addition, 61 samples were collected from patients after myeloma treatment, and 18 samples from patients who had experienced disease progression or relapse. The disease status of the post-treatment patients was mainly a complete response (CR) and very good partial response (VGPR) based on the criteria of International Myeloma Working Group. In addition, the percentage of plasma cells in the patients achieving VGPR or CR after treatment was less than 5%.

We also enrolled 20 healthy and genetically unrelated Taiwanese volunteers (healthy individuals) as the control group. These healthy individuals had undergone bone marrow analysis to investigate cytopenia that had been noted in blood tests, but whose bone marrow examinations revealed no abnormalities. All patients and healthy individuals signed informed consent forms after the study had been thoroughly explained.

The research protocol was created in accordance with the Declaration of Helsinki, and it was reviewed, approved and registered by the Ethics Committee of Kaohsiung Medical University Hospital (KMUHIRB-2012-01-08(II)).

Bone marrow mononuclear cells were isolated for this study. First, the bone marrow samples were collected in tubes containing ethylenediaminetetraacetic acid (EDTA), preserved at 4°C and processed within 4 hours of collection. The bone marrow samples were then centrifuged at 12,000 × g for 15 minutes, after which ammonium chloride lysis buffer (10 mM NH4Cl, 10 mM KHCO3, 0.1 mM EDTA) was used to clear the red blood cells and effectively isolate the fraction of mononuclear cells.

The isolated bone marrow samples were stored at -80°C until RNA extraction. Isolation of RNA from 200 μL of cell suspension was carried out using the TRIzol protocol (Invitrogen). The extracted RNA was then treated with DNase (Promega) and the concentration was determined by spectrophotometric OD260 measurement. The integrity of the RNA was examined by 1.2% RNA denaturing agarose gel electrophoresis.

Reverse transcription was performed to generate complementary DNA in a final volume of 20 μL, containing 2 μg RNA, 25X dNTP mix (100 mM), 10X random primer (0.5 μM), RNase inhibiter, reverse transcriptase, reverse transcriptase buffer (10X) and diethylpyrocarbonate (DEPC)-treated water. The procedure was performed according to the manufacturer’s protocol (Applied Biosystems).

Real-time quantitative RT-PCR was performed in a final volume of 10 μL containing 1 μL of RT product, 0.6 μL of primer (Roche), 1.2 μL of probe (Roche, cat. no. 04688945001), 2.2 μL of DEPC H₂O and 5 μL of qPCR Master Mix (2X) (KAPA Biosystem, KK 4600). Analysis of the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene was used as the internal control.

The primer sequences of MALAT1 were as follows: forward, 5’-GACCCTTCACCCCTCACC-3’; reverse, 5’-TTATGGATCATGCCCACAAG-3’, and the primer sequences of GAPDH were as follows: forward, 5’-AAAGTCCGCCATTTTGCCACT-3’; and reverse, 5’-CCAAATCGTTAGCGCTCCTT-3’.

Real-time quantitative RT-PCR was performed in a LightCycler 480 Real-Time PCR System (Roche). The PCR cycling program consisted of incubation for enzyme activation at 95°C for 10 minutes, followed by melting at 95°C for 10 seconds, annealing at 60°C for 30 seconds, and then extension at 72°C for 1 second, for a total of 50 cycles.

The expression levels of MALAT1 were normalized to the internal control GAPDH reference to obtain the relative threshold cycle (∆C_T). The relative expression levels were calculated by the comparative C_T (∆∆C_T) method, and relative expression folds (2^-ΔΔCT) were calculated.

All statistical analyses were based on two-sided hypothesis tests with a significance level of p < 0.05. The analyses were performed using SPSS version 17.0 (SPSS, Chicago, IL, USA).

The expression of MALAT1 was significantly higher in the patients at diagnosis compared with the patients post-treatment (mean ∆C_T: -5.54 ± 0.16 vs. -3.84 ± 0.09, 3.25-fold change; p < 0.001) or the healthy individuals (mean ∆C_T: -5.54 ± 0.16 vs. -3.95 ± 0.21, 3.01-fold change; p < 0.001) (Table 1). This suggests that MALAT1 may be deregulated and overexpressed in patients with multiple myeloma.

The association of MALAT1 expression pattern with disease status was further analyzed. The expression of MALAT1 was found to be significantly decreased in the post-treatment patients to a level that was similar to that of the healthy individuals (mean ∆C_T: -3.84 ± 0.09 vs. -3.95 ± 0.21, p = 0.614). In addition, in the patients in whom the disease had progressed or relapsed, the expression of MALAT1 was significantly increased compared with the post-treatment patients (mean ∆C_T: -4.92 ± 0.23 vs. -3.84 ± 0.09, 1.89-fold change; p < 0.001) (Table 1). For the patients who underwent multiple bone marrow examinations during treatment and follow-up, the expression of MALAT1 changed dynamically and was correlated with disease status (Figure 1).

The clinical relevance of MALAT1 was analyzed. The expressions of MALAT1 in the 45 newly diagnosed patients with different clinical characteristics were listed in Additional file 1: Table S1. We noticed that the expression of MALAT1 was not associated with the percentage of plasma cells in the bone marrow (r = -0.037, p = 0.808) (Additional file 2: Table S2). With regards to the association between MALAT1 expression and prognosis, the results showed that the initial higher MALAT1 expression level (Cut-off value: ∆C_T = -5.30) determined by ROC analysis was not associated with inferior prognosis including progression-free survival (PFS) (median PFS: 21.0 ± 9.9 vs. 15.0 ± 6.0 months, p = 0.390) or overall survival (OS) (median OS: Not reached; mean OS: 31.9 ± 4.3 vs. 37.8 ± 3.6 months, p = 0.172) (Additional file 3: Figure S1). However, we hypothesized the magnitude of the change (decrease) after myeloma-related therapy may have been related to the degree of treatment response and prognosis, because the expression of MALAT1 changed after treatment.

Among the 45 patients, 36 (including 21 men) with a mean age of 61.3 years received bone marrow examinations at least twice including at diagnosis and after treatment, and hence the expression of MALAT1 was available for further analysis. The median PFS of these 36 patients was 18 months (95% CI 8.95-25.35), and we used this median PFS as a cut-off value to divide the patients into two groups of early (PFS ≤18 months) or late (PFS >18 months) progression, with 18 patients in each group. We found that the only parameter which showed a significant difference between these two groups was the magnitude of MALAT1 change after treatment. The patients with early progression had a significantly smaller magnitude of MALAT1 change after treatment compared with the patients with late progression (mean difference of ∆C_T: 1.26 ± 1.06 vs. 2.09 ± 0.79; p = 0.011) (Table 2).

The previous results showed that the magnitude of MALAT1 change (as quantified by the difference in ∆C_T) was the only parameter associated with PFS. We then used ROC analysis and obtained a cut-off expression change value of 1.5 (post-treatment ∆C_T – pre-treatment ∆C_T; approximately a 2.8-fold change) with an estimated area under the ROC curve of 0.79 (p = 0.003). The proportion of patients with a lower MALAT1 change (difference in ∆C_T ≤1.5) was significantly higher in those who displayed early progression (n = 13, 72.2%) compared with those who displayed late progression (n = 4, 22.2%; p = 0.007).

In terms of an association between clinical characteristics and magnitude of MALAT1 change, the patients with greater changes (difference in ∆C_T >1.5) may have had a better treatment response (Additional file 4: Table S3). The PFS and OS rates were also analyzed using the cut-off value from ROC analysis. With a minimum follow-up period of 12 months (range: 12 to 48 months), the patients with a greater MALAT1 decrease (difference in ∆C_T >1.5) had a significantly prolonged median PFS (24 months, range 11-48 months) compared with the patients with a smaller MALAT1 change (11 months, range: 6-21 months; p = 0.001). There was no significant difference in OS between the two groups (median OS: Not reached; mean OS: 39.2 ± 3.6 months, range: 12-48 months vs. 32.8 ± 4.2 months, range: 12-48 months; p = 0.313) (Figure 2).

Cox regression analysis was used to identify the relative risk of early progression, which revealed that autologous hematopoietic stem cell transplantation (Auto-HSCT) and the magnitude of MALAT1 change were significantly associated with the prognosis. For all post-treatment patients (n = 36), those with a smaller MALAT1 change (difference in ∆C_T ≤1.5) had a significantly higher risk of early progression of disease (OR 4.89, 95% CI 1.73-13.86; p = 0.003), while auto-HSCT reduced the risk of early progression (OR 0.22, 95% CI 0.05-0.97; p = 0.046). For the post-treatment patients with a VGPR or CR (n = 33), a smaller MALAT1 change (difference in ∆C_T ≤1.5) remained the single factor predictive of early progression of multiple myeloma (OR 4.38, 95% CI 1.48-12.99; p = 0.008) (Table 3). Using the cut-off value to predict the patients who would show early progression (PFS ≤18 months), the estimated accuracy was 75% with a sensitivity of 72.2%, a specificity of 77.8%, a positive predictive value of 76.5%, and a negative predictive value of 73.7%.