Targeting the MALAT1/PARP1/LIG3 complex induces DNA damage and apoptosis in multiple myeloma

Abstract

Metastasis-associated lung adenocarcinoma transcript 1(MALAT1) is a highly conserved long non-coding RNA (lncRNA). Overexpression of MALAT1 has been demonstrated to related to poor prognosis of multiple myeloma(MM) patients. Here, we demonstrated that MALAT1 plays important roles in MM DNA repair and cell death. We found bone marrow plasma cells from patients with monoclonal gammopathy of undetermined significance (MGUS) and MM express elevated MALAT1 and involve in alternative-non-homozygous end joining (A-NHEJ) pathway by binding to PARP1 and LIG3, two key components of the A-NHEJ protein complex. Degradation of the MALAT1 RNA by RNase H using antisense gapmer DNA oligos in MM cells stimulated poly-ADP-ribosylation of nuclear proteins, defected the DNA repair pathway, and further provoked apoptotic pathways. Anti-MALAT1 therapy combined with PARP1 inhibitor or proteasome inhibitor in MM cells showed a synergistic effect in vitro. Furthermore, using novel single wall carbon nanotube (SWCNT) conjugated with anti-MALAT1 oligos, we successfully knocked down MALAT1 RNA in cultured MM cell lines and xenograft murine models. Most importantly, anti-MALAT1 therapy induced DNA damage and cell apoptosis in vivo, indicating that MALAT1 could serve as a potential novel therapeutic target for MM treatment.

Introduction

Multiple myeloma (MM), a cancer of terminally differentiated plasma cells, is the second most frequently diagnosed hematologic cancer in the United States¹. MM is nearly always preceded from an age-related progressive pre-malignant condition termed monoclonal gammopathy of undetermined significance (MGUS). The finding of long non-coding RNA (lncRNA) transcripts from genomic regions is one of the most unexpected findings of the genomics era. lncRNAs are a group of RNA transcripts longer than 200nt that do not encode proteins but are involved in various forms of gene expression regulation². Rapidly accumulating evidences indicate that lncRNAs are involved in the initiation and progression of almost all kinds of cancer, including MM^3–5. These findings collectively support the possibility that systematic investigation of lncRNA function in tumorigenesis will yield novel insights into diagnosis and treatment of cancers.

Metastasis-associated lung adenocarcinoma transcript 1(MALAT1), also known as nuclear-enriched noncoding transcript 2(NEAT2), is a highly conserved nuclear lncRNA (~8.7kb in human)⁶. MALAT1 was initially found to play an important role in nuclear speckles and to interact with pre-mRNA splicing factors in Hela cells^7–10 through regulating a variety of biological process¹¹. It was originally identified in metastatic non-small cell lung cancer (NSCLC)¹² and over-expressed in different types of tumor such as hepatocellular carcinoma¹³, breast cancer¹⁴ and prostate cancer¹⁰. In MM, MALAT1 is reported to be the most highly expressed lncRNA and correlated with poor prognosis^{4, 15} and significantly unregulated in fatal course extramedullary MM compared with newly diagnosed MM patients¹⁶.

In our current study, we sought to determine the oncogenic role of MALAT1 and explore it as a possible therapeutic target for MM. We found that MALAT1 is highly expressed in MGUS, smoldering MM (SMM) and MM compared to normal plasma cells. We further identified the function of MALAT1 involving in alternative non-homologous end joining (A-NHEJ) pathway through binding with PARP1/ LIG3 complex, and regulated apoptosis via co-acting with PARP1. Finally, we developed a novel single-wall carbon nanotube (SWCNT)-conjugated anti-MALAT1 oligo, and used it in two MM xenograft murine models, and observed remarkable therapeutic outcomes.

Methods

Cell lines, plasmids and human MM tissues

Human MM cell lines including MM.1S, H929, RPMI8226 and HEK293T were obtained from ATCC. Bortezomib-resistant (RPMI8226/V10R) cell line is a kind gift of Dr. Robert Orlowski (The University of Texas M.D. Anderson Cancer Center, Houston, TX, USA). Melphalan-resistant (RPMI8226/LR5) and doxorubicin-resistant (RPMI8226/DOX40) cell lines are gift from Dr. William Dalton (Moffitt Cancer Center, Tampa, FL). Bone marrow (BM) plasma cells were isolated from 4 healthy donors (HDs) and 7 MM patients using CD138 magnetic beads (Miltenyi Biotec). Total RNA was isolated using Trizol reagent (Thermo Fisher Scientific). To establish A-NHEJ, NHEJ and homologous recombination (HR) DNA repair pathway reporter stable cell lines, pEJ2GFP-puro (#44025, Addgene)¹⁷, pimEJ5GFP (#44026, Addgene) ¹⁷ and pDRGFP (#26475, Addgene)¹⁸ vectors were transfected into HEK293T cells separately and selected with 2μg/mL puromycin. To construct the MALAT1 overexpression vector, full length of human MALAT1 cDNA was cloned into pCDH-MSCV-MCS-EF1-copGFP-T2A-Puro plasmid (System Biosciences), between NotI and SwaI sites. The packaging system was used according to the manufacturer’s protocol. MM.1S cells were infected by MALAT1 overexpression (V-MALAT1) or empty control virus (V-ctrl), and subjected to flow sorting by flow cytometry using green fluorescent protein copGFP as a marker.

Formalin-fixed paraffin-embedded (FFPE) BM blocks of 11 HDs and 9 MM patients were obtained from the myeloma tissue bank of the Cleveland Clinic Taussig Cancer Institute and the Norman Bethune International Peace Hospital. All participants signed informed consent forms. This study was approved by institutional review boards (IRB) of both Cleveland Clinic and Norman Bethune International Peace Hospital. MM tissue microarray (TMA) was purchased from US Biomax, lnc (T291b), which contained BM from 2 HDs and 4 MM patients.

RNA Antisense Purification (RAP)

Nuclear extracts were incubated with a 59bp biotin-labeled MALAT1 probe GTGCCTTTAGTGAGGGGTACCTGAAAAATCTTAAAAAAAGGCTTAGCGCCCACCTCACC/3Bio/ or sequence-scrambled probe TCAACCTTTACACCGATCTAGAATCGAATGCGTAGATTAGCCAGGTGCAAACCAAAAAT/3Bio/ and hybridized at 4°C for 2 hours. Hybridized material was captured with magnetic streptavidin beads (Thermo Fisher Scientific). Bound material was washed and eluted with RNaseH (New England Biolabs) as previously described¹¹. The proteins were separated by SDS-PAGE and stained using Coomassie blue. Specific bands were isolated for whole proteomic mass spectrometry (MS) analysis.

Ribonucleoprotein immunoprecipitation (RIP)

2×10⁷ H929 or MM.1S cells were rinsed with PBS and then irradiated with 150 mJ/cm² at 254 nm using a UV cross-linker. Cell pellets were resuspended in 100 μL cytoplasmic extract (CE) buffer (10 mM HEPES, 60 mM KCl, 1 mM EDTA, 0.075% NP40, 1 mM DTT, pH 7.6). The cells were incubated on ice for 3 min and then centrifuged at 250g for 5 min. The cell nuclei were washed wit 500 μL CE buffer without NP40 and then resuspended in lysis buffer (150mM NaCl, 50mM Tris-HCl pH 7.5, 0.5% Triton X-100 supplemented with protease inhibitor cocktail and RNase inhibitor). The lysate was sonicated for 5 min with 30sec on/off intervals on ice, and then centrifuge at 14,000g for 10 minutes. The cell lysate was further diluted (1:5) with NT2 buffer (50 mM Tris-HCl pH7.4, 150 mM NaCl, 1 mM MgCl₂, 0.05% NP-40 supplemented with fresh 200U RNaseOut, 400μM VRC, 1mM DTT, 20mM EDTA, and protease inhibitor cocktail).

Protein A/G magnetic beads (Pierce protein A/G magnetic beads; Thermo scientific) were washed with NT2 buffer 6 times and then pre-coated using 5% BSA NT2 buffer (1:5 v/v) at room temperature for 1 h. anti-PARP1 or anti-LIG3 antibody, 2μg, was added to 500 μL of the bead mixture and incubated at 4°C for 12 hours. The beads were washed in ice-cold NT2 buffer for 5 times and resuspended in 850 μL NT2 buffer.

The cell lysate was mixed with the antibody-coated beads, and an aliquot of the mixture was removed for total RNA and protein determination. The remaining lysate was incubated with beads at 4°C for 4 hours. After co-IP, the beads were washed as follows: twice with lysis buffer; thrice with the lysis buffer containing 900 mM NaCl and 1% NP-40; and twice more with lysis buffer. The beads were then transferred to a fresh tube and subjected to a final wash with the lysis buffer containing 0.05% NP-40. Following the washes, an aliquot of beads was removed from each sample and mixed with 2× LDS sample buffer for western blot analysis. Another aliquot of beads was used for RNA extraction.

Detailed description is provided in the Supplementary Materials and Methods.

Results

MALAT1 is the most highly expressed lncRNA in MGUS and MM

We first analyzed gene expression microarray datasets uploaded by 3 different groups, including Zhan dataset (GSE5900)¹⁹, Gutiérrez dataset (GSE16558)²⁰ and López-Corral dataset (GSE47552)²¹. Analysis of all 3 datasets showed that MALAT1 expression was higher in MGUS, SMM and MM compare with healthy donors (HDs, Fig. S1A).

We next used in situ hybridization (ISH, Fig. S1B) and qRT-PCR (Fig. S1C) to detect MALAT1 in clinical MM samples and cell lines, and verified that MALAT1 was highly expressed in BM CD138⁺ cells from MM patients compared with HDs, which was consistent with microarray data. Furthermore, two groups have reported that MALAT1 overexpression was significantly correlated to poor prognosis in MM patients, including shorter progression-free survival (PFS) and overall survival (OS)^{15, 22}.

MALAT1 over-expression accelerated proliferation and repressed apoptosis in MM

To explore the functions of over-expressed MALAT1 in MM, we infected V-MALAT1 or V-ctrl into MM.1S cells, and added puromycin for selection, then injected subcutaneously to the shoulders of SCID mice (Fig. 1A). Diameters of tumor were measured once a week, the growth of MM.1S-V-MALAT1 xenografts was significantly faster than controls (Fig. 1A). MALAT1 levels in MM.1S-V-MALAT1 xenografts were over-expressed confirmed by qRT-PCR (Fig. 1B). MM.1S-V-MALAT1 xenografts compared with the MM.1S-V-ctrl xenografts have higher proliferation and less apoptosis according to immunohistochemistry staining of Ki-67 and c-caspase3 (Fig. 1C).

Fig. 1. MALAT1 overexpression promoted the tumorigenesis of MM.

(A) 2×10⁶ MALAT1 overexpressed or control MM.1S cells were injected subcutaneously to the shoulder of SCID mice. The sizes of xenograft were measured once a week. Mice were sacrificed 30 days after injection, and xenografts were weighted. (B) MALAT1 level was determined by qRT-PCR. (C) The levels of Ki-67 and c-Caspase3 were detected by immunohistochemistry. (*p<0.05, **p<0.01, ***p<0.001)

MALAT1 binds with PARP1/LIG3 complex in MM

To investigate the co-factors binding to MALAT1 in MM cells, we used RNA antisense purification-mass spectrum (RAP-MS) to identify MALAT1 binding proteins (Fig. 2A). Biotin-labeled anti-MALAT1 DNA probe was used to pull-down MALAT1 in H929 cells, then MALAT1 pull-down sample was used to run a PAGE gel and subjected to MS analysis. (Fig. S2A–B). Using RAP-MS whole proteomic analysis, we identified 23 MALAT1 binding proteins (Table. S1). STRING database functional enrichment analysis revealed 10 of these proteins were related to DNA repair pathways (GO:0006281, false discovery rate 9.89e-08), including PARP1, LIG3, XRCC1, XRCC5, XRCC6, SUPT16H, NPM1, RFC1, SSRP1 and MPG (Fig. S2C). The notable proteins with strong signals, including PARP1, LIG3, and XRCC5 were further verified by western blot using MALAT1 pull-down protein lysate from H929, MM.1S and RPMI8226 cells, respectively (Fig. 2B). The co-localization between MALAT1 and PARP1 was further confirmed by immunofluorescence staining. As shown in Fig. 2C, more than 70% of the PARP1 signal was co-localized with MALAT1 signal in H929, MM.1S, and RPMI8226 cells.

Fig. 2. PARP1/LIG3 complex was identified as MALAT1 binding target by RAP-MS.

(A) Schematic diagram of the RAP-MS strategy used to identify MALAT1 binding proteins. (B) PARP1, LIG3 and XRCC5 were verified as MALAT1 binding proteins. (C) MALAT1 co-localized with PARP1 in H929, MM.1S and RPMI8226 cells (scale bar=1μM).

PARP1 was intensively investigated multiplefunctional protein which has been implicated in recognition of DNA single and double strand break (SSB and DSB)s during DNA repair and catalyzes PAR formation to induce cell apoptosis²³. According to our RAP-MS results, MALAT1 pulled down PARP1, as well as other DNA repair proteins, thus we hypothesized that MALAT1 acts as a scaffold, to form functional complexes through bundling PARP1 and other proteins, then exerted its function in DNA repair pathway(s). To validate this hypothesis, we firstly used ribonucleoprotein immunoprecipitation (RIP) strategy to further prove the binding between MALAT1 and PARP1 in myeloma cells. As shown in Fig. 3A, the RNA-protein complexes in myeloma cells were first cross-linked by UV, then the cell lysate was incubated with PARP1 antibody-coated magnetic beads. After washing, total RNA was extracted from the precipitate, then the MALAT1 level was determined by qRT-PCR. We found PARP1 antibody-coated beads specifically enriched PARP1 signal (Fig. 3B), and MALAT1 was also enriched by PARP1 antibody-conjugated beads exclusively (Fig. 3C), which demonstrated direct interaction between MALAT1 and PARP1. Although there no RNA binding domain on LIG3, poly (ADP-ribose) polymerase and DNA-Ligase Zn-finger (zf-PARP) regions are present that can bind PARP1²⁴. PARP1 and LIG3 are critical molecules involved in the A-NHEJ DNA repair pathway²⁵. Thus, we postulated that MALAT1 might play its role in A-NHEJ DNA repair by direct binding with PARP1 and indirect binding with LIG3.

Fig. 3. Verification of the binding between MALAT1 and PARP1 by RIP-PCR.

(A) Schematic diagram of the RIP-PCR assay. (B) The level of PARP1 before or after RIP was determined by immunoblotting. (C) MALAT1 pulled-down by PARP1 antibody-conjugated beads were determined by qRT-PCR (right panel), left panel is input (*p<0.05, **p<0.01, ***p<0.001).

MALAT1 inhibition induced DNA damage and apoptosis in MM

To demonstrate our postulation, we used two gapmer DNA antisense oligos targeting MALAT1 (anti-MALAT1-1/2) to knock-down MALAT1 expression and perform loss-of-function study of MALAT1 in MM cells. The gapmer DNA was flanked by blocks of 2′-OMe-modified RNAs, which would bind to MALAT1 RNA and induce cleavage of MALAT1 by RNase H²⁶. qRT-PCR analysis showed MALAT1 was efficiently knocked-down in H929(Fig. 4A), MM.1S (Fig. 4B) and RPMI8226 cells (Fig. 4C). The frequency of DNA break increased substantially as revealed by immunofluorescence staining and western blot for γH2A.X. Interestingly, we found that PAR signal increased after MALAT1 knock-down indicating that MALAT1 antagonist did not inhibit, but enhanced PARP1 catalytic activity by releasing PARP1 from MALAT1/PARP1 complex, which would induce cell apoptosis directly²⁷. Furthermore, defective DNA repair induced more cleavage of PARP1 and caspase3 (c-PARP1 and c-caspase3), which also contributed to cell apoptosis in MM cells. We also observed significantly increased apoptosis by flow cytometry analysis after anti-MALAT1 treatment (Fig. 4A–C).

Fig. 4. MALAT1 inhibition induced DNA damage and cell death in MM.

2′-OMe-modified anti-MALAT1 oligos or control oligos were transfected into H929(A), MM.1S(B) or RPMI8226(C) cells, respectively. At 48h after transfection, cells were collected and subjected to qRT-PCR, immunofluorescence staining for γH2A.X, immunoblotting of γH2A.X, PARP1, c-PARP1, caspase-3, and c-caspase3 and apoptosis assay. (*p<0.05, **p<0.01).

To determine whether MALAT1 specificly involved in A-NHEJ DNA repair pathway, we generated pEJ2GFP-puro (A-NHEJ reporter), pimEJ5GFP (NHEJ reporter) and pDRGFP (HR reporter) stable cell lines in HEK293T cells separately^{17, 18}. I-SceI was used to generate DNA damages at I-SceI sites on these plasmids. MALAT1 overexpression vector or anti-MALAT1 gapmer DNA were used to up-/down-regulate the expression of MALAT1. All vectors would produce GFP once the plasmid DNAs were repaired by correspondent functions, thus we could evaluate which DNA repair pathway involved by examining GFP positive ratio by flow cytometry. As shown in Fig. 5A, HEK293T-EJ2GFP with over-expressed MALAT1 had significant increase of GFP positive cells, whereas HEK293T-EJ2GFP with knocked-down MALAT1 had significant decrease of GFP positive cells. However, no significant difference was found in the HEK293T-imEJ5GFP and HEK293T-DRGFP cells with up-/down-regulation of MALAT1(Fig. S3), indicating that MALAT1 specificly involved in A-NHEJ pathway. The result were further verified by immuofluoresence staining of LIG3/PARP1 and LIG3/γH2A.X in H929 cells after we knocked down MALAT1. We found MALAT1 knock-down had no influence on LIG3/PARP1 co-localization (Fig. 5B, Fig. S4), but interrupted the co-localization between LIG3 and γH2A.X (Fig. 5C, Fig. S4). These results demonstrated that MALAT1 is crucial in A-NHEJ pathway through recruiting LIG3 to γH2A.X loci on DSB.

Fig. 5. MALAT1 coordinated with PAPR1 inhibitor through inhibiting A-NHEJ.

(A) In A-NHEJ reporter plasmid (pEJ2-GFP-puro) stable expression 293T cells, pCBA-SceI was transient transfected with MALAT1 overexpression/empty vector or anti-MALAT1/control gapmer. Then GFP positive cells were determined by flow cytometry. (B) The co-localization between LIG3 and PARP1 or γH2A.X were determined by immunofluorescence staining (scale bar=5μM). (C) H929, MM.1S and RPMI8226 cells transfected with anti-MALAT1 or control oligos were treated with ABT-888. Cells were collected for WB. MALAT1 were determined by qRT-PCR.(*p<0.05, **p<0.01)

To further understand how does MALAT1 knock down affect other components of A-NHEJ DNA repair pathway, NHEJ pathway, and HR pathway, we detected the expression levels of proteins involved in these pathways including CtIP, MRE11, RAD50, NBS1, p-ATM, p-ATR, XRCC5 and XRCC6. We found that RAD50, pATM, or pATR upregulated in all three MM cell lines, MRE11 upregulated in RPMI8226 cell line, after MALAT1 was knockdown, but not CtIP, NBS1, XRCC5 and XRCC6 protein levels (Fig. S5A). To determin if MALAT1 inhibition will affect MRE11-RAD50-NBS1 (MRN) complex or XRCC5 and XRCC6 complex formation, we did immunoflurecence staining of MRE11 and NBS1, XRCC5 /XRCC6 complex in MM cell lines. We found that MALAT1 knock down didn’t influence the MRN complex and XRCC5/XRCC6 complex formation (Fig. S5B–C), indicating MALAT1 is dispensable for the initial DSB recognition of either A-NHEJ or NHEJ or HR pathways.

PARP1 inhibitor cooperated with MALAT1 antagonist to induce DNA damage and apoptosis in MM

To determine whether inhibiting the dissociated PARP1 catalytic activity induces additional DNA damage to further increase cell death after anti-MALAT1 treatment in MM cells, we used the PARP1 inhibitor ABT888 to specially inhibit PARP1 activity in MALAT1 knocked-down MM cells. We observed increased PAR signal in H929 and MM.1S when treated with anti-MALAT1 only, but saw dose-dependent decreased PAR signal in the same cell lines treated synchronously with anti-MALAT1 and ABT888. Combination treatment significantly increased the level of γH2A.X, cleaved PARP1, and cleaved caspase-3(Fig. 5C).

MALAT1 inhibition potentiates the cytotoxic effects of bortezomib in MM

Bortezomib treatment could induce “BRCAness” in MM and impair HR pathway²⁸. Our results have demonstrated that repressing MALAT1 inhibited A-NHEJ activity, thus we postulated that MALAT1 antagonists might have synergistic effect with bortezomib by disabling both HR and A-NHEJ pathways, then provoked cell death by inducing severe DNA damages in MM. To verify our assumption, anti-MALAT1-1 or scrambled DNA oligos was transfected into H929, MM.1S and RPMI8226 cells and treated with various doses of bortezomib, then cells were collected for apoptosis assay. We found expression of BRCA1 and BRCA2 were dramatically down-regulated in all 3 cell lines received high-dose bortezomib treatment (Fig. 6A–C). Whereas γH2A.X signals and apoptosis ratio were increased by both high-/low-dose bortezomib treatment, and these effects were amplified by combining with anti-MALAT1 treatment. Meanwhile, anti-MALAT1 treated MM cells were more sensitive to bortezomib compared to untreated cells according to our cell viability assay, which showed the IC50 reduced from 4.9 nM to 3.6 nM in H929 cells, from 8.9 nM to 6.6 nM in MM.1S cells, and from 10.1 nM to 8.2 nM in sensitive to bortezomib compared with control cells (Fig. 6A–C). These results implied that bortezomib and anti-MALAT1 acted synergistically to induce MM cell death via promoting DNA damages.

Fig. 6. Synergistic effect of anti-MALAT1 and bortezomib in MM.

H929(A), MM.1S(B) and RPMI8226(C) cells were transfected with 1nM anti-MALAT1 or control oligos and treated with bortezomib. Cells were collected for immunoblotting, apoptosis assay and qRT-PCR. Cell viability was measured and IC50 was calculated before and after MALAT1 knockdown.( *p<0.05, **p<0.01, ***p<0.001)

To understand the role of MALAT1 in drug resistance in MM, we used bortezomib-, melphalan- and doxorubicin- resistance MM cell lines, RPMI8226/V10R, RPMI8226/LR5 and RPMI8226/DOX40 and their parental cell line RPMI8226 used as control. We found MALAT1 expression was significantly higher in these resistant MM cell lines compared with RPMI8226 cells (Fig. S6A). Furthermore, after MALAT1 level was knocked-down (Fig. S6B), an increased apoptotic cells numbers were observed in all three resistant cell lines (Fig. S6C). IC50 of RPMI8226/LR5 cells to melphalan decreased from 53.4 μM to 34.0 μM, RPMI8226/DOX40 cells to doxorubicin decreased from 2.46 μM to 1.48 μM and RPMI8226/V10R to bortezomib decreased from 193.8 nM to 143.9 nM, respectively (Fig. S7A–C). Those results indicated that anti-MALAT1 treatment resensitized resistant MM cells to their corresponding drugs again.

SWCNT-anti-MALAT1 oligo repressed MM proliferation and induced cell apoptosis in vivo

Our data indicated that MALAT1 antagonist was a robust tool to provoke DNA damage and apoptosis in MM. However, the efficient delivery of anti-MALAT1 oligos in vivo was the main obstacle that limits clinical application of this type of therapy. As a novel nanomaterial for drug delivery, SWCNT may deliver nucleic-acid drugs stablely and efficiently with good tolerability and minimal toxicity in vitro²⁹ and in vivo³⁰. To track the delivery visible, we conjugated SWCNT with Cy3-labeled-anti-MALAT1 oligos (SWCNT-anti-MALAT1-Cy3)(Fig. 7A)¹³, and then added it into culture medium of H929-GFP and MM.1S-GFP cells to validate delivery efficiency. As shown in Fig. 7B, SWCNT-anti-MALAT1-Cy3 was delivered into the nucleus of MM cells efficently and suppressed the endogenous MALAT1 level in both H929 and MM.1S cells significantly (Fig. 7C).

Fig. 7. SWCNT-anti-MALAT1 showed high delivery efficiency and minimal toxicity.

(A) Schematic diagram of SWCNT-anti-MALAT1-Cy3 gapmer oligos (Scale bars=100μM). (B) H929-GFP and MM.1S-GFP cells were co-cultured with SWCNT-anti-MALAT1-cy3 for 48h. (C) MALAT1 level was knocked-down successfully.(**p<0.01, ***p<0.001).

To further estimate the treatment potential of SWCNT-anti-MALAT1 in vivo, we subcutaneously injected MM.1S-Luc-GFP cells on the back of SCID mice to establish human MM xenograft murine model (Fig. 8A). At day 14 after tumor cell injection, SWCNT-anti-MALAT1 or SWCNT-ctrl oligos were injected directly into tumors and repeated at days 21, 24 and 28, respectively. We observed tumor burden with IVIS after luciferin injection, and found the luciferin signal was significant lower in SWCNT-anti-MALAT1 treated group compared with SWCNT-ctrl treated group. Then we measured MALAT1 level with RNA samples extracted from tumor xenografts and found that MALAT1 expression was significantly downregulated by SWCNT-anti-MALAT1 treatment, which indicated SWCNT delivered MALAT1 antisense oligo efficiently into MM cells in this human MM xenograft murine model. Western-blot results showed c-PARP1 increased in anti-MALAT1 treatment group. Immunohistochemistry results revealed decreased Ki-67 and increased c-caspase3 signals on SWCNT-anti-MALAT1 treated tumor sections.

Fig. 8. SWCNT-anti-MALAT1 treatment repressed myeloma growth in both xenograft and disseminated murine models.

(A) MM.1S-Luc-GFP cells were injected subcutaneously to SCID mice (5 mice each group). SWCNT-anti-MALAT1 or SWCNT-ctrl was injected into the tumors at the indicated days. Tumor growth was monitored by IVIS. Mice were sacrificed on day 30, tumor samples were subjected to qRT-PCR, WB and immunohistochemistry (Scale bars=100μM). (B) SCID mice (7 mice each group) were intravenously injected with MM.1S-Luc-GFP cells, then injected with SWCNT-anti-MALAT1 or SWCNT-ctrl once a week through tail veins. Kaplan-Meier analysis indicated SWCNT-anti-MALAT1 prolonged mouse lifespan significantly (P=0.04). (C) Proposed model of MALAT1 antagonist induces MM cell apoptosis.

MM is a systematic disease and malignant cells usually involve multiple organs in patient. To mimic this situation, we generated a disseminated MM murine model, and used SWCNT-anti-MALAT1 to test treatment effect on it (Fig. 8B). We firstly intravenously injected MM.1S-Luc-GFP cells into SCID mice through tail veins. At day 7 after tumor cell injection, SWCNT-anti-MALAT1 or SWCNT-ctrl oligos were injected through tail veins and repeated at days 14, 21, 28 and 35, respectively. We detected luciferin signal at day 35 and recorded their survival status, then we found SWCNT-anti-MALAT1 treatment not only reduced the tumor burden, but extended lifespan significantly (P=0.04) compare with SWCNT-ctrl treated group.

To further confirm the therapeutic effect of anti-MALAT1 in MM in vivo, we generated another murine xenograft model and disseminated model with H929-Luc-mCherry cells (Fig. S8). We found SWCNT-anti-MALAT1 inhibited H929 growth dramatically in both models. Meanwhile, SWCNT-anti-MALAT1 efficiently knocked-down MALAT1 expression, upregulated c-PARP1 and c-caspase3, and inhibited Ki-67 signal. In H929 disseminated model, SWCNT-anti-MALAT1 treatment extended mice lifespan significantly (P=0.009).

Discussion

This study is the first to elucidate the function of the lncRNA MALAT1 in MM. PARP1 and LIG3 are two key molecules required for the highly error-prone A-NHEJ³¹ DNA repair pathway. We demonstrated that MALAT1 is critical for PARP1/LIG3 complex to recognize DSBs γH2A.X loci on DNA, then activated A-NHEJ DNA repair in MM. LIG3 is upregulated in multiple myeloma³², chronic myeloid leukemia³³, and breast cancer³⁴. Strikingly, A-NHEJ is associated with frequent chromosome abnormalities such as deletions, translocations, inversions, and other complex rearrangements³⁵. Thus over-expression of MALAT1 in MM may enhance A-NHEJ DNA repair pathway to induce secondary chromosome changes³⁶, which may promote disease progression, but also induce drug resistance. Dissecting the mechanism of how MALAT1 directly or indirectly recruits LIG3 to γH2A.X loci, which represents DSBs, to favor A-NHEJ repair pathway will be our next focus by investigating the function of different domains on LIG3 and PARP1. Further gain-of-function studies of MALAT1 in normal or precursor cells and transgenic mice will be needed in the future to confirm our findings.

To evaluate MALAT1 as a possible therapeutic target, we pursued antisense inhibition and observed increased DNA damage and apoptosis in MM cells due to dissociation of the MALAT1/PARP1/LIG3 complex and deregulation of A-NHEJ pathway. We verified that knock-down MALAT1 in MM cells had synergistic effect with PARP1 inhibitor or bortezomib through inducing more cell apoptosis. Most of U.S. Food and Drug Administration (FDA)-approved PARP1 inhibitors are used to inhibit the catalytic activity of PARP1 and increase DNA damage in ovarian cancer with BRCA1/2 mutations, where the HR pathway is defected^37–39. Unrepaired DNA damage will induce cell apoptosis. MALAT1 antagonist acts its role through PARP1 but the underlying mechanism is different from PARP1 inhibitors. In contrast, Anti-MALAT1 treatment disrupts MALAT1/PARP1/LIG3 DNA repair complex, then dissociated free PARP1 will induce polyADP-ribosylation in the nucleus, which will promote cell apoptosis directly. Anti-MALAT1 treatment will also impaired the A-NHEJ DNA repair pathway, which will further induce cell apoptosis due to unrepaired DSB DNA. When combined bortezomib and anti-MALAT1 therapy in MM, bortezomib repressed HR through reducing BRCA1/2 expression, meanwhile anti-MALAT1 inhibited A-NHEJ activity, thus apoptosis accumulated dramatically through synchronous dysfunction of two DSB repair pathways. These results provide new therapeutic strategy for MM patients.

To today, FDA has approved several antisense oligonucleotide drugs, including nusinersen for spinal muscular atrophy⁴⁰, mipomersen for homozygous familial hypercholesterolemia⁴¹, fomivirsen for cytomegalovirus retinitis⁴², and eteplirsen for Duchenne muscular dystrophy⁴³. We used gapmer DNA flanked by 2′-OMe-modified MALAT1 antisense oligonucleotides to achieve remarkable inhibition effects, and applied SWCNT as delivery system to improve affinity, stabilize the oligos and against nuclease degradation during delivery. To the best of our knowledge, this is the first report to use functionalized SWCNTs to deliver anti-sense oligos targeting lncRNAs in tumor. Due to their surface chemistry properties for delivery and large cargo capability, SWCNT represents a novel and useful nanomaterial for drug delivery, which not only stabilize nucleic acid molecule from digestion of nucleases, but increase penetration of DNA/RNA dramatically without toxicity²⁹. In our study, SWCNT was functionalized covalently and then conjugated with anti-MALAT1, which allow anti-MALAT1 to be released with high concentration in MM cells and induced DNA damage and apoptosis effectively in both in vitro and in vivo experiments without bringing any toxicity in normal cells. Thus SWCNT-anti-MALAT1 is an ideal therapeutic method for the MM patients.

In conclusion, we have shown that MALAT1 exerted DNA protective and anti-apoptotic functions via binding to PARP1/LIG3 protein complexes, targeting MALAT1 induced DNA damage and apoptosis, therefore inhibited MM growth (Fig. 8C). Furthermore, we demonstrated that MALAT1 could be targeted via neutralization by antisense in vitro and in vivo, this treatment extended lifespan of MM-bearing mice significantly. Synergism of MALAT antisense with proteasome or PARP1 inhibitors further illustrated the potential therapeutic value of MALAT1 for MM patients.