1. Introduction

Multiple myeloma (MM) is a hematological neoplasm characterized by the uncontrolled aberrant clonal proliferation of plasma cells (PCs) in the bone marrow ^[1]. This disease is the second most common hematological malignancy, after non-Hodgkin lymphoma ^[2][3], affecting elderly patients with a median age of 65 years ^[4]. Despite the latest advances in treatment strategies, which have significantly increased patient survival, MM is still considered an incurable disease, with a median overall survival of 7 years.

MM is a very heterogeneous disease, which is reflected in the inter-individual differential diagnosis and survival of patients. Different studies have associated this variability with a wide range of genetic and epigenetic alterations present in MM patients ^[5][6], including distinct molecularly defined subtypes with different features ^[7]. Regarding the genetic variability, MM is divided into hyperdiploid (HRD) and non-HRD subtypes ^[7]. HRD MM is characterized by the trisomy of chromosomes 3, 5, 7, 9, 11, 15, 19 and 21 ^[6], whereas non-HRD MM is characterized by translocations of the immunoglobulin (Ig) alleles. The majority of these translocations affect chromosome 14, where the Ig H-chain is located ^[6][7]. However, Ig translocations can also affect the kappa and lambda light chains, the co-occurrence of which is common with HRD MM. Besides, some of these light-chain translocations are associated with a poor outcome for MM patients, as is the case for IgL-MYC translocations ^[8]. Some of the common heavy-chain translocations are also considered as high-risk prognostic factors, such as t(4;14) and t(14;16), which affect 

MMSET

MAF genes, respectively ^[6][7][9]. Epigenetic aberrations of the DNA methylation and histone modifications are also thought to play an important role in MM pathogenesis. The study of global DNA methylation of MM has led to the identification of a highly heterogeneous DNA methylation pattern, which results in extensive DNA hypomethylation of intergenic regions and DNA hypermethylation associated with intronic and enhancer regions ^[2][5]. In addition, the study of histone modifications in MM has revealed a de novo gain of active chromatin marks preferentially located in regulatory elements, such as enhancer and promoter regions, which arise from heterochromatic regions in normal B cells ^[10][11][12]. These results suggest the possibility that these epigenetically regulated non-coding genomic regions could lead to the transcription of non-coding RNA genes (ncRNAs) and, in particular, to the expression of long non-coding RNAs (lncRNAs), which may play a relevant role in the pathobiology and clinical outcome of MM ^[13]. Nowadays, studies about the role of certain lncRNAs in MM are emerging. However, more comprehensive analyses are required to better understand their function in this disease. In this review, we summarize the current knowledge regarding the role of lncRNAs in the development and outcome of MM and discuss the possibility of lncRNAs as targets for the development of novel RNA-based therapeutic strategies for MM patients.

2. Features of lncRNAs

Traditionally, cellular functions of DNA and proteins have overshadowed the roles of RNAs. In recent years, the development of high-throughput techniques, such as RNA sequencing (RNA-seq), has brought great advances in the understanding of the cell transcriptome. So far, it is known that, although only 1–2% of the human genome is translated into proteins, around 70–90% of it is transcribed into RNA, resulting in a huge amount of ncRNAs ^[14]. Among these ncRNA genes, lncRNAs are defined as those non-coding transcripts longer than 200 nt that do not encode proteins, with open reading frames (ORFs) smaller than 100 amino acids, and with a lack of or low coding potential. However, the latest RNA-seq studies have shown that some lncRNAs contain cryptic ORFs, which could encode for small ORFs or non-conserved peptides ^[15][16][17].

MALAT1

NEAT1 (nuclear paraspeckle assembly transcript 1) ^[18][19]. Although the size of lncRNAs varies between 200 nt and more than 1 MB (known as macro lncRNAs), 42% of lncRNAs only present two exons ^[19][20]. In contrast to mRNAs, which are located at the cytosol, lncRNAs can be located either in the nucleus or in the cytoplasm, where they can exert various functions. Thus, regarding the location where lncRNAs act and their transcription site, they are capable of acting as 

cis

trans transcripts ^[15][21]. 

Cis

trans lncRNAs act over distal genes ^[22][23][24]. Interestingly, lncRNAs are cell- and tissue-specific, and they may affect different biological processes, such as chromosome conformation, imprinting of genomic loci, or gene and protein regulation ^[15][25]. lncRNAs have the ability to regulate at DNA, RNA and protein levels, and their functions can be divided into four different groups depending on their molecular mechanisms ^[18][26]: (1) signal lncRNAs are regulatory molecules that can trigger the transcription of other genes by their presence. They can infer chromatin states, affect gene imprinting or mark certain spaces, times or stages for gene regulation, such as 

Air

PANDA (p21-associated ncRNA DNA damage activated) ^[26][27]. (2) Decoy lncRNAs are transcripts that bind to targets and prevent them from binding to their own targets, thus leading to the alteration of post-transcriptional control. This type of lncRNA can act as an miRNA sponge, binding to miRNAs thanks to their complementary sequence (

Figure 1

PTENP1 (phosphatase and tensin homolog pseudogene 1), for example, leads to tumor suppressor activity due to the decoy of different miRNAs ^[28][29][30]. (3) Guide lncRNAs can regulate gene expression through the recruitment and re-localization of ribonucleoprotein complexes at specific chromatin loci, such as 

MEG3

Table 1) ^[18][31][32]. (4) Scaffold lncRNAs can act as central platforms upon the assembly of different ribonucleoprotein complexes, affecting their molecular components (

Figure 1

HOTAIR

Table 1), promoting gene repression ^[18][26].

Figure 1.

 Mechanisms by which 

MALAT1

 acts in MM cells. (

A

) 

MALAT1

 acts as scaffold lncRNA, binding to PARP1 protein, which binds to a complex of DNA-repair enzymes consisting of LIG3 among others. Then, the protein–

MALAT1

 complex repairs the damaged DNA, triggering the proliferation of MM cells. However, when binding of 

MALAT1 and PARP1 does not occur, damaged DNA is not repaired, triggering MM cell death ^[33]. (

 and PARP1 does not occur, damaged DNA is not repaired, triggering MM cell death [70]. (

B

) 

MALAT1

 can also act as a miRNA sponge (decoy), binding to different miRNAs such as 

miR-1271-5p

, a tumor-suppressor miRNA that negatively regulates SOX13. Binding of 

MALAT1

 and 

miR-1271-5p

 triggers overexpression of SOX13 and proliferation of MM cells, whereas knockdown of 

MALAT1

 releases 

miR-1271-5p

, which binds and prevents translation of SOX13. MM = multiple myeloma.

Table 1.

 Summary of deregulated lncRNAs in MM. MM = multiple myeloma; KD = knockdown; UR = upregulation; up = upregulated; down = downregulated; NA = not available; PFS = progression-free survival; OS = overall survival.

Gene	Location	Gene Type	Expression in MM	Molecular Mechanism	Molecular Interaction in MM	Biological Effect after lncRNA KD	Biological Effect after lncRNA UR	Prognosis in MM	References

ANRIL	9p21.3	Antisense	Up	Decoy	Binds to miR-34a, miR-125a, miR-186 and miR- 411–3p	Decreases cellular proliferation and increases apoptosis	NA	High expression levels associated with worse PFS and OS	[34][35][33,34]
BM742401	18q11.2	LincRNA	Down	NA	NA	NA	Decreases cell migration	Methylated lncRNA associated with worse OS	[36][35]
Circ_0000190	1q42.12	Circular lncRNA	Down	NA	NA	NA	NA	High expression levels associated with better PFS and OS	[37][38][39][36,37,38]
CRNDE	16q12.2	LincRNA	Up	Decoy	Binds to miR-451	Decreases cellular proliferation, increases apoptosis and triggers cell cycle arrest	NA	High expression levels associated with worse OS	[40][41][39,40]
DARS-AS1	2q21.3	Antisense	Up under hypoxia	Decoy	Interacts with RBM39 and HIP-1α, suppressing mTOR pathway	Decreases cellular proliferation and increases apoptosis. Decreases tumorigenesis in vivo. Its upregulation reduces the sensitivity to bortezomib in vitro	NA	NA	[42][41]
ENSG00000249988	4p15.33	LincRNA	Up	NA	NA	NA	NA	High expression levels associated with worse PFS andbetter OS	[13]
ENSG00000254343	8q24.12	LincRNA	Up	NA	NA	NA	NA	High expression levels associated with worse PFS	[13]
FEZF1-AS1	7q31.32	Antisense	Up	Decoy	Binds to miR-610 and regulates AKT3	Decreases cellular proliferation, increases apoptosis and triggers cell cycle arrest	NA	NA	[43][42]
GAS5	1q25.1	Processed transcript	Down	NA	NA	NA	Decreases cellular proliferation	NA	[44][43]
H19	11p15.5	Processed transcript	Up	Decoy	Binds to miR-152-3p and miR-29b-3p	Decreases cellular proliferation, increases apoptosis and triggers cell cycle arrest	NA	High expression levels associated with worse PFS	[45][46][47][44,45,46]
HOTAIR	12q13.31	Antisense	Up	NA	Activates NF-κB pathway	Decreases cellular proliferation, triggers cell cycle arrest and decreases chemoresistance to dexamethasone	NA	NA	[21][48][49][50][21,47,48,49]
HOXB-AS1	17q21.32	Antisense	Up	Scaffold	Scaffold for ELAVL1. Interacts with FUT4-mediated Wnt/β-catenin pathway	Decreases cellular proliferation and increases apoptosis	NA	NA	[51][50]
IRAIN	15q26.3	Antisense	Down	Decoy	Binds to miR-125b and regulates IGF-1 signaling	NA	Increases apoptosis	NA	[52][53][51,52]
LINC00152	2p11.2	LincRNA	Up	Decoy	Binds to miR-497	Decreases cellular proliferation, increases apoptosis and triggers cell cycle arrest. Decreases tumorigenesis in vivo	NA	High expression levels associated with worse OS	[54][53]
LINC00461	5q14.3	LincRNA	Up	NA	NA	Decreases cellular proliferation and increases apoptosis	NA	High expression levels associated with worse OS	[55][56][57][54,55,56]
LINC00515	21q21.3	LincRNA	Up	Decoy	Binds to miR-140-5p	Increases apoptosis	NA	NA	[58][57]
LINC00665	19q13.12	LincRNA	Up	Decoy	Binds to miR-214-3p	Decreases cellular proliferation and increases apoptosis	NA	NA	[59][58]
LINC01234	12q24.13	LincRNA	Up	Decoy	Binds to miR-124-3p	Decreases cellular proliferation and increases apoptosis. Decreases cell proliferation and tumor growth in vivo	NA	High expression levels associated with worse OS	[60][59]
lnc-ANGPTL1-3	1q25.2	Antisense	Up	Decoy	Binds to miR-30a-3p	Increases the sensitivity to bortezomib	NA	High expression levels associated with worse OS	[61][60]
lnc-TCF7	5q31.1	NA	Up	NA	NA	NA	NA	High expression levels associated with worse PFS and OS	[62][61]
LUCAT1	5q14.3	LincRNA	Up	NA	Activates the TGF-β signaling pathway	Decreases cellular proliferation, increases apoptosis and triggers cell cycle arrest	NA	High expression levels associated with shorter five-year survival	[63][62]
MALAT1	11q13.1	LincRNA	Up	Decoy and Scaffold	Binds to miR-1271-5p, miR-181a-5p and miR-509- 5p. Scaffold for PARP1	Decreases cellular proliferation and increases apoptosis	NA	High expression levels associated with worse PFS and OS	[64][65][66][67][68][69][70][33][63,64,65,66,67,68,69,70]
MEG3	14q32.2	LincRNA	Down	Decoy	Binds to miR-181a	NA	Decreases cellular proliferation and increases apoptosis	High expression levels associated with better PFS and OS	[71][72][73][74][71,72,73,74]
MIAT	22q12.1	LincRNA	Up	Decoy	Binds to miR-29b	Sensitizes MM cells to bortezomib	NA	High expression levels associated with worse PFS and OS	[75][76][75,76]
NEAT1	11q13.1	LincRNA	Up	Decoy	Binds to miR-214 and miR-125a	Decreases cellular proliferation	NA	High expression levels associated with worse PFS and OS	[77][78][79][80][77,78,79,80]
OIP5-AS1	15q15.1	Processed transcript	Down	Decoy	Binds to miR-410 and miR-27a-3p	NA	Decreases cellular proliferation and increases apoptosis	NA	[81][82][81,82]
PCAT-1	8q24.21	LincRNA	Up	Decoy	Binds to miR-129	Increases apoptosis and sensitizes MM cells to bortezomib	NA	NA	[83][84][83,84]
PDIA3P	1q21.1	Pseudogene	Up	NA	NA	Decreases cellular proliferation. Increases the sensitivity to bortezomib	NA	High expression levels associated with worse OS	[85]
PDLIM1P4	3q12.1	Pseudogene	Up	NA	NA	NA	NA	High expression levels associated with worse PFS and OS	[13]
PRAL	17p13.1	NA	Down	Decoy	Binds to miR-210	NA	Decreases cellular proliferation and increases apoptosis. Increases the anti-tumor effect of bortezomib	High expression levels associated with better PFS and OS	[86][87][86,87]
PVT1	8q24.21	Processed transcript	Up	Decoy	Binds to miR-203a. It is inhibited by BRD4	Decreases cellular proliferation and increases apoptosis	NA	NA	[88][89][88,89]
SMILO	1q42.2	LincRNA	Up	NA	Regulates IFN pathway	Decreases cellular proliferation and increases apoptosis	NA	High expression levels associated with better OS	[13]
SNHG16	17q25.1	Processed transcript	Up	Decoy	Binds to miR-342-3p	Decreases cellular proliferation, increases apoptosis and triggers cell cycle arrest	NA	NA	[90]
SOX2OT	3q26.3	Sense overlapping	Up	Decoy	Binds to miR-144-3p	Decreases cellular proliferation, increases apoptosis and triggers cell cycle arrest. Decreases tumor growth in vivo	NA	NA	[91]
ST3GAL6-AS1	3q12.1	Antisense	Up	NA	NA	Decreases cellular proliferation, increases apoptosis and triggers cell cycle arrest	NA	High expression levels associated with worse PFS	[92][93][92,93]
TUG1	22q12.2	Antisense	Up	Decoy	Binds to miR-29b-3p and targets HDAC4	Decreases cellular proliferation and increases apoptosis	NA	NA	[44][94][43,94]
UCA1	19p13.12	Processed transcript	Up	Decoy	Binds to miR-1271-5p and miR-331-3p	Decreases cellular proliferation and increases apoptosis	NA	High expression levels associated with worse OS	[95][96][95,96]
XLOC_013703	20p11.21	NA	Down	NA	Involved in NF-κB signaling activation	NA	Decreases cellular proliferation and increases apoptosis	NA	[97]

3. Role of lncRNAs in the Pathobiology of MM

Diverse studies have pointed to the importance of lncRNAs in different biological processes, such as immune response, cell differentiation, gene expression modulation and chromatin reorganization ^[98][99]. Intriguingly, their deregulation also contributes to the development of carcinogenesis, metastasis and even anti-cancer treatment resistance ^[64]. The deregulation of the expression of lncRNAs can thus impact on relevant pathways involved in the pathogenesis and/or progression of certain human tumors, including MM ^[44][71]. We have recently demonstrated that 82% of the transcriptome, including coding genes and all types of polyA+ and non-polyA lncRNAs, in plasma cells from MM correspond to lncRNAs, compared to 18% of coding genes ^[13].

Diverse studies have pointed to the importance of lncRNAs in different biological processes, such as immune response, cell differentiation, gene expression modulation and chromatin reorganization [98,99]. Intriguingly, their deregulation also contributes to the development of carcinogenesis, metastasis and even anti-cancer treatment resistance [63]. The deregulation of the expression of lncRNAs can thus impact on relevant pathways involved in the pathogenesis and/or progression of certain human tumors, including MM [43,71]. We have recently demonstrated that 82% of the transcriptome, including coding genes and all types of polyA+ and non-polyA lncRNAs, in plasma cells from MM correspond to lncRNAs, compared to 18% of coding genes [13].

Some deregulated lncRNAs in MM also appear deregulated in the same way in other types of human cancer: for example, 

HOTAIR

 is upregulated in hepatocellular carcinoma (HCC), 

PDIA3P

 (protein disulfide isomerase family A member 3 pseudogene 1) in lung cancer, and 

LINC00461 in both HCC and lung cancer, all three of them being also upregulated in MM ^{[48][49][50][55][56][57][100]}. 

 in both HCC and lung cancer, all three of them being also upregulated in MM [47,48,49,54,55,56,100]. 

PRAL

 (P53 regulation associated lncRNA) is downregulated in HCC, lung cancer and MM, and 

GAS5 (growth arrest specific 5) in breast, prostate, renal cancer and MM ^{[44][86][87][101][102]}. However, there are lncRNAs that are deregulated in MM while they show the opposite direction of expression in other neoplasms. For instance, 

 (growth arrest specific 5) in breast, prostate, renal cancer and MM [43,86,87,101,102]. However, there are lncRNAs that are deregulated in MM while they show the opposite direction of expression in other neoplasms. For instance, 

MALAT1 is upregulated in MM, lung cancer, gallbladder cancer, colorectal carcinoma and HCC, whilst this lncRNA is downregulated in colorectal and glioma cancer ^{[65][103][104]}. 

 is upregulated in MM, lung cancer, gallbladder cancer, colorectal carcinoma and HCC, whilst this lncRNA is downregulated in colorectal and glioma cancer [64,103,104]. 

NEAT1

 is also upregulated in MM, lung cancer and HCC, but is downregulated in acute promyelocytic leukemia [105]. Finally, 

Circ_0000190 is downregulated in MM and gastric cancer, whereas it displays overexpression in lung cancer ^[37][38][39]. These results highlight the cell- and tissue- specificity of lncRNAs, showing that their deregulation—and thus, their potential function—needs to be addressed in each tumor. For example, 

 is downregulated in MM and gastric cancer, whereas it displays overexpression in lung cancer [36,37,38]. These results highlight the cell- and tissue- specificity of lncRNAs, showing that their deregulation—and thus, their potential function—needs to be addressed in each tumor. For example, 

MALAT1

 (one of the most widely studied lncRNAs [103]) and 

NEAT1

 are able to bind or interfere with different molecules and pathways depending on the tissue or disease (

Table 1) ^[77][106]. 

) [77,106]. 

MALAT1

 acts as an miRNA sponge binding to 

miR-1271-5p

 (

Figure 1

), 

miR-181a-5p

 and 

miR-509-5p

 in MM, to 

miR-195

 in HCC or to 

miR-206 and miR-363-3p in gallbladder cancer ^{[64][66][67][68][69]}. In the case of 

 and miR-363-3p in gallbladder cancer [63,65,66,67,68]. In the case of 

NEAT1

, it binds to 

miR-125a

 in MM and to 

miR-193a-3p in lung adenocarcinoma, among others ^[77][106]. Usually, the expression of lncRNAs and miRNAs is negatively correlated. Therefore, overexpression of one lncRNA could trigger the downregulation of miRNAs, whereas downregulation of one lncRNA could promote the overexpression of different miRNAs ^[34][77]. Likewise, there are other examples of lncRNAs which act as miRNA sponges in MM (

 in lung adenocarcinoma, among others [77,106]. Usually, the expression of lncRNAs and miRNAs is negatively correlated. Therefore, overexpression of one lncRNA could trigger the downregulation of miRNAs, whereas downregulation of one lncRNA could promote the overexpression of different miRNAs [33,77]. Likewise, there are other examples of lncRNAs which act as miRNA sponges in MM (

Table 1

). In MM, some of these lncRNAs, such as 

CRNDE

 (colorectal neoplasia differentially expressed) and 

IRAIN

 (IGF1R antisense imprinted non-protein coding RNA) are associated with the regulation of one single miRNA. However, an increasing number of studies in MM are showing that lncRNAs can regulate or can be regulated by more than one miRNA, such as 

H19

 (H19 imprinted maternally expressed transcript), 

UCA1

 (urothelial cancer associated 1) or 

OIP5-AS1

 (OIP5 antisense RNA 1). Remarkably, there are cases like 

TUG1

 (taurine up-regulated 1) and 

H19

 that are associated with the regulation of the same miRNA, 

miR-29b-3p

 (

Table 1) ^{[34][40][43][45][46][52][54][58][59][60][61][72][81][82][83][88][90][91][94][95][96]}. These results highlight the relevance of the miRNA sponge function of lncRNAs in MM.

) [33,39,42,44,45,51,53,57,58,59,60,72,81,82,83,88,90,91,94,95,96]. These results highlight the relevance of the miRNA sponge function of lncRNAs in MM.

Different studies have revealed how the knockdown or upregulation of certain lncRNAs is also associated with different biological and phenotypic effects in MM cells, such as the decrease in cell proliferation or viability, the decrease in cellular migration, the increase in cellular apoptosis and cell cycle arrest (

Table 1

).

Furthermore, various studies have demonstrated the in vivo biological effect of lncRNA knockdown in MM. For example, the inhibition of 

DARS-AS1

 (DARS antisense RNA 1) or 

LINC00152

 reduces the tumorigenesis of MM cells, whilst the knockdown of 

SOX2OT

 (SOX2 overlapping transcript) reduces tumor growth. Moreover, the knockdown of 

LINC01234

 increases 

miR-124-3p

 and suppresses 

GRB2

 expression, resulting in a decrease of cell proliferation and the inhibition of MM growth. These results demonstrate that lncRNAs play an important role in the pathobiology of MM (

Table 1) ^{[13][34][35][36][40][41][42][43][44][45][46][49][51][52][54][57][58][59][60][63][70][72][78][79][81][82][83][84][85][86][88][90][91][92][94][95][96][97][107]}.

) [13,33,34,35,39,40,41,42,43,44,45,48,50,51,53,56,57,58,59,62,69,72,78,79,81,82,83,84,85,86,88,90,91,92,94,95,96,97,107].