The LIN28B/let-7 axis is a novel therapeutic pathway in multiple myeloma.

MYC is a major oncogenic driver of multiple myeloma (MM) and yet almost no therapeutic agents exist that target MYC in MM. Here we report that the let-7 biogenesis inhibitor LIN28B correlates with MYC expression in MM and is associated with adverse outcome. We also demonstrate that the LIN28B/let-7 axis modulates the expression of MYC, itself a let-7 target. Further, perturbation of the axis regulates the proliferation of MM cells in vivo in a xenograft tumor model. RNA-sequencing and gene set enrichment analyses of CRISPR-engineered cells further suggest that the LIN28/let-7 axis regulates MYC and cell cycle pathways in MM. We provide proof of principle for therapeutic regulation of MYC through let-7 with an LNA-GapmeR (locked nucleic acid-GapmeR) containing a let-7b mimic in vivo, demonstrating that high levels of let-7 expression repress tumor growth by regulating MYC expression. These findings reveal a novel mechanism of therapeutic targeting of MYC through the LIN28B/let-7 axis in MM that may impact other MYC-dependent cancers as well.

Introduction

Multiple myeloma (MM), a tumor originating from plasma cells in the bone marrow (BM), has an annual incidence of 6.3 new cases per 100,000 individuals[1]. Despite the major advances in therapy for MM, it remains incurable and there are no targeted therapies for MM, in part due to the lack of therapies that target specific oncogenes involved in the pathogenesis of the disease. Genomic events such as chromosomal translocations, copy number variation, somatic mutations, and epigenetic modifications all contribute to gene deregulation of specific oncogenes or tumor suppressors during MM tumorigenesis[2]. Among those, MYC plays a central role in the progression of the disease. Approximately two thirds of newly diagnosed patients harbor MYC activation, which correlates with adverse clinical outcome[3]. MYC activation is commonly driven by translocation or copy number gain of chromosome 8q24, which contains the MYC locus[4, 5]. Despite the dominant role of MYC in MM, there are very few therapeutic options targeting MYC. Previous studies have attempted to target MYC by using a bromodomain inhibitor to target BET proteins, which regulate MYC[6, 7].

The let-7 miRNA was originally discovered in C. elegans as a regulator of developmental timing and cell proliferation[8]. Let-7 expression increases as cells become more differentiated. In humans, let-7 miRNAs comprise a family of 12 members distributed over 8 genomic loci[9] that are often repressed in cancer[10]. Let-7 miRNAs function as a tumor suppressor through regulation of key oncogenes, including MYC and RAS, by binding specific sites in the mRNA 3'-UTRs and inhibiting translation of these targets[11],[12]. Low expression of let-7 family members is associated with poor prognosis in several cancer types[13, 14].

In humans, let-7g and let-7i are located individually on chromosomes 3 and 12 respectively. The remaining let-7 family members are distributed among six miRNA clusters at genetically distinct loci. The let-7a2 and let-7c clusters are involved in hematopoietic stem and progenitor cell (HSPC) homeostasis by regulating the balance between TGFβ and Wnt signaling[15], whereas the let-7e cluster is highly expressed in HSPC and confers hematopoietic phenotypes[16]. However, the exact role of the various let-7 family members in mammalian development has not yet been fully elucidated[17,18], in large part because it is technically difficult to knock out multiple let-7 family members in the same individual cell. Moreover, these multiple let-7 family members are likely to have functionally similar roles.

LIN28B is an RNA-binding protein highly expressed in stem cells and developing tissues where it impairs the processing of let-7 precursors into mature, functional miRNAs[19]. Over-expression of LIN28B has been reported in several cancers[20] and is associated with advanced disease and poor outcome in ovarian[14], breast[21], colon[22], hepatocellular carcinoma[23, 24] and neuroblastoma[25, 26]. Transgenic LIN28B has been shown to induce multiple tumors types in mice including liver, Wilms, colon, and neuroblastoma, all of which solidify its oncogenic role[24, 25,27,28]. LIN28B has also been reported to act through a let-7-independent manner, especially via regulation of IGF2[24].

While the LIN28B/let-7 axis has been implicated in the regulation of MYC in different tumor types[20], its potential as a therapeutic target has not yet been explored, specifically in blood cancers. In this study, we define the mechanistic activity of the LIN28/let-7 axis in clonal plasma cells and establish a potential therapeutic role of this pathway in targeting MYC in MM, which could lead to significant therapeutic advances in MM and other cancers.

Methods

Cell and primary cells

The MM cell lines MM.1S and RPMI8226 were purchased from ATCC; KMS12BM and MOLP-8 were purchased from DSMZ and KMM-1 was purchased from JCRB Cell Bank. The MM.1S GFP+Luc+ cell line was generated by retroviral transduction with the pGC-GFP/Luc vector (gift of A. Kung, Dana-Farber Cancer Institute). Cells were authenticated by short tandem repeat DNA profiling. Primary samples were obtained from bone marrow aspiration from both MM patients and healthy controls. Plasma cells were isolated using CD138+ microbead selection (Miltenyi Biotec®, Auburn, CA). All patients were diagnosed with active MM at diagnosis or at relapse, based on criteria of the International Myeloma Working Group[29]. Informed consent was obtained from all patients and healthy volunteers in accordance with the Declaration of Helsinki protocol.

Lentivirus-mediated shRNA silencing

LIN28B shRNA in lentiviral plasmid (TRCN0000122191 and TRCN0000122599) and control shRNA (SHC216V) were purchased from Sigma-Aldrich. For viral production, 293T cells were transfected with lentiviral gag/pol, VSV-G, and the lentiviral plasmid, at a ratio of 1:0.4:1, using Lipofectamine 2000. Viral particles were harvested after 24hrs and 48hrs. Two milliliters of viral supernatant were used to infect 1,000,000 cells in the presence of Polybrene (8 ng/µL). Infected cells were selected on Puromycin antibiotic before subsequent analysis.

Lentivirus-mediated CRISPR silencing

LentiCRISPRv2 (Addgene plasmid #52961) and lentiCRISPR:EGFPsgRNA-1 (#51760) were gifts from Feng Zhang[30]. LIN28B sgRNA were designed using the MIT Optimized CRISPR design tool. Sequences of sgRNA were: 5’-CATCGACTGGAATATCCAA G-3’ for sgLIN28B#1 and 5’-CAGAGCAAACTATTCATGGA-3’ for sgLIN28B#2. Human U6 (hU6) primer 5’-GAGGGCCTATTTCCCATGATT-3' was used for validation by Sanger sequencing after cloning. Lentivirus production was processed as above for shRNA lentiviral production.

miRNA mimic transfection

Cell lines were transfected with hsa-let-7b mimic or with a control probe (mirVana miRNA mimic, Life Technology) at final concentration of 40 nM, using Lipofectamine 2000 according to manufacturer’s instructions (Invitrogen). Culture medium was changed to regular medium 24 hours after transfection and cells were used for functional assays at 48 hours. For the rescue experiment, MOLP-8 sgGFP or sgLIN28B#1 were transfected with a control probe or a mix of anti-let-7a, b, d, and g (mirVana anti-miRNA, Life Technology) at a final concentration of 40 nM, using X-tremeGENE 9 according to manufacturer’s instructions (Roche, Life Science).

Quantitative reverse transcription PCR

Mature miRNA and mRNA expression were analyzed by qRT-PCR using SYBR green dye on an Applied Biosystems AB7500 Real Time PCR System. All PCR reactions were run in triplicate. Ct values were normalized on RNU6B and 18S, respectively, and relative changes were calculated using the 2-ΔΔCt. The following primer sequences were used: LIN28B-F: 5’-GCCCCTTGGATATTCCAGTC-3’; LIN28B-R: 5’-TGACTCAAGGCCTTTGGAAG-3’; MYC-F: 5’-TCGGTCCTCGGATTCTCTGCTCT-3’; MYC-R: 5’-GCCTCCAGCAGAAGGTGATCCA-3’; KRAS-F: 5’-TGTGTCTCATATCAGGTTGACGA-3’; KRAS-R: 5’ -CAAGAGTCGAGTGTGGTCTCA-3’; CCND1-F: 5’-TCTACACCGACAACTCCATCCG-3’; CCND1-R: 5’-TCTGGCATTTTGGAGAGGAAGTG-3’; DICER1-F: 5’-CTCCTACCACTACAATACTATCACT-3’; DICER1-R: 5’-GGTCTTCATAAAGGTGCTTGGT-3’; E2F6-F: 5’-GCGGAGAGTGTATGACATCACC-3’; E2F6-R: 5’-GTCAGAAAGTTCCTCCTGTAGCT-3’; HMGA1-F: 5’-GAAGTGCCAACACCTAAGAGACC-3’; HMGA1-R: 5’-GGTTTCCTTCCTGGAGTTGTGG-3’; pri-let-7d-F: 5’-GCCAAGTAGAAGACCAGCAAG-3’; pri-let-7d-R: 5’-CAAGGAAACAGGTTATCGGTG-3’; pri-let-7g-F: 5’-GTTCCTCCAGCGCTCCGTT-3’; pri-let-7g-R: 5’-CCATTACCTGGTTTCCCAGAGA-3’. Sequences for full mature let-7 miRNA were used to design let-7 forward primers, in combination with universal 3’ miRNA reverse primer.

Immunoblotting

Whole-cell lysates were subjected to SDS-PAGE, and transferred to polyvinyldene fluoride (PVDF) membrane (Bio-Rad Laboratories). For immunoblotting we used antibodies against LIN28B (Cell signaling #4196S), c-MYC (Cell signaling #9402S) and GAPDH (Cell signaling #2118S).

Proliferation assay

Proliferation rates were measured by DNA synthesis, using [3H]-thymidine uptake (Perkin Elmer, Boston, MA) as described[31].

RNA-sequencing

RNA was extracted using Qiagen RNeasy Kit. Whole RNA (500ng) was subject to library preparation with NEBNext Ultra RNA Library prep for Illumina kit (BioLabs). A single unique index was assigned to each sample. Quality control of the libraries was evaluated by Bioanalyzer analysis with High Sensitivity chips (Agilent). Libraries were quantified by qPCR (Kapa assay) and multiplex before sequencing on a HiSeq 2000 (2×50bp paired-end reads) at the Biopolymers Facility of Harvard Medical School. Cutadapt was used to trim adapters; trimmed reads were aligned to Human reference genome (GRCh37) with tophat2; and read counts for each gene was calculated by HT-seq. RNA-seq data have been deposited to the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession numbers GSE71100.

In vivo studies

SCID/bg mice (n=5/group) used for xenograft experiments were injected subcutaneously with MOLP-8 cells that were infected with either a LIN28B shRNA, or a control shRNA. Tumor volume (measured by caliper) were calculated by the formula: length × width2 × 0.52[32]. To evaluate the effect of let-7 LNA-GapmeR (containing the sequence 5’-AGGTAGTAGGTTGTGT-3’), SCID/bg mice (n=5/group) were i.v. injected with 5×106 MM.1S GFP+Luc+ cells on day 0; followed by i.p. injections of let-7 (20 mg/kg) or control LNA-GapmeR 2 times a week starting on day 2. Tumor growth was evaluated by bioluminescence imaging (BLI) and mice were followed for survival.

Results

Deregulation of LIN28B/Let-7 axis in MM

We first sought to determine whether LIN28B is deregulated in MM, and therefore analyzed two independent publicly available gene expression profiling datasets containing plasma cells from patients with newly diagnosed MM and healthy control donors. These included GSE24080 and GSE2658, containing 22 normal donor plasma cell samples and 559 plasma cell samples from patients with newly diagnosed patients - both from UAMS - and GSE16558, with 5 normal donor plasma cells and 65 plasma cell samples from patients with newly diagnosed MM. We identified a significant overexpression of LIN28B in MM cells compared to normal plasma cells in both datasets (Fig. 1a). We next sought to determine the prognostic relevance of LIN28B in the survival of patients with MM. We performed Kaplan Meier analysis on a cohort of 542 patients treated with Total Therapy 2 (GSE2658), and observed that high expression of LIN28B was associated with significantly worse overall survival, (p=0.0075) (Fig. 1b).

Figure 1

LIN28B expression level in MM

LIN28B expression level in MM patients, compared to healthy individuals, by analysis of (a) Left panel: GSE16558 (5 normal donor plasma cells and 65 plasma cell samples from patients with newly diagnosed MM) and right panel: GSE24080 – containing 22 normal donor plasma cell samples - and GSE2658 – 559 plasma cell samples from patients with newly diagnosed patients - both from UAMS, which were normalized using GeneSpring. (b) Kaplan-Meier analysis of 542 patients with MM of the Total Therapy 2 cohort (GSE2658). Patients were classified as high vs. low expression of LIN28B based on the mean expression level. LIN28B was associated with significantly worse overall survival, (p=0.0075). (c) MOLP-8, KMS12BM, RPMI 8226 and KMM-1 cells infected with control shRNA, or 2 different LIN28B shRNAs, were studied for relative expression of LIN28B mRNA as determined by qRT-PCR.

Therefore, we explored whether targeting the LIN28B/let-7 axis would have a therapeutically relevant role in MM and silenced LIN28B using two lentiviral short hairpins (shRNA) constructs that target the mRNA coding sequence in several LIN28B-expressing MM cell lines: MOLP-8, KMS12BM, RPMI8226 and KMM-1. The shRNAs caused degradation of LIN28B mRNA in all cell lines (Fig. 1c). Moreover, pri-let-7 RNA was not modified as shown for pri-let-7g, consistent with the post-transcriptional regulation of let-7 by LIN28B (Supplemental Fig. 1a).

LIN28B regulates let-7 and MYC in MM

Consistent with the role of LIN28B in regulating let-7 expression, we observed de-repression of let-7 family members in cells with LIN28B knockdown compared to non-targeting control (Fig. 2a). We next analyzed the downstream targets of LIN28B/let-7 and found a decreased protein expression of MYC in LIN28B-silenced cells in several MM cell lines (Fig. 2b). Moreover, knockdown of LIN28B significantly impaired the proliferation of these cell lines tested (Fig. 2c). Consistent with these observations, we identified, through gene set enrichment analysis (GSEA), an enrichment of let-7 target genes in MM patient samples who displayed a high level of LIN28B expression in two independent datasets (GSE2658 and MMRC dataset), suggesting that LIN28B represses let-7 and indirectly enriches let-7 target expression (Fig. 3a).

Figure 2

The LIN28B/let-7/MYC axis in Multiple Myeloma

(a) relative expression of mature let-7 miRNA species as determined by quantitative PCR in 4 MM cell lines infected with control shRNA or 2 different LIN28B shRNAs. (b) Protein blot analysis for LIN28B and MYC expression in MOLP-8, KMS12BM, RPMI 8226 and KMM-1 cells infected with control shRNA or 2 different LIN28B shRNAs, and (c) proliferation assay by thymidine uptake over a 48hr time in the same cell lines. P values were obtained by two-tailed Student t test (*P < 0.05)

Figure 3

The LIN28B/let-7 axis in data sets and in vivo

(a) GSEA analysis showing an enrichment of let-7 target genes in MM patients who display a high level of LIN28B expression in two independent datasets (GSE2658 and MMRC dataset). (b) Tumor volume and (c) survival of SCID/bg mice (5 per group) injected subcutaneously with 5.106 MOLP-8 cells expressing pLKO control shRNA or pLKO.LIN28BshRNA; average survival time was 26 days versus 33 days, respectively, P = 0.0045. Bars indicate mean ± SD (n = 3).

In addition, among let-7 target genes, MYC expression significantly correlated with LIN28B expression level in two independent datasets (GSE16558 and GSE2658). As shown in Supplemental Fig. 1b, a strong correlation existed between LIN28B and MYC in MM tumor cells from patient samples in GSE16558 and in total therapy 2 GSE2658, p<0.028 and p<0.001, respectively.

To validate whether MM cells are dependent on LIN28B for growth, we examined tumor growth of MOLP-8 cell line with a LIN28B specific hairpin or non-targeting control in a xenograft mouse model using SCID/bg mice. Tumor growth was significantly lower in mice injected subcutaneously with LIN28B knockdown compared to scrambled control (Fig. 3b), resulting in a significant prolongation of survival in the experimental group, p = 0.0045 (Fig. 3c). In addition, MYC expression was significantly reduced in cells obtained ex-vivo from mice injected with LIN28B knockdown compared to scrambled control (Supplemental Fig. 1c).

Together, these data suggest a deregulation of LIN28B/let-7 axis in a proportion of patients with MM and can function as a novel therapeutic target of MYC regulation in MM.

MYC and cell cycle pathways are highly enriched in cells with high expression of LIN28B

To control for possible off-target effect of shRNAs and for incomplete LIN28B knockdown mitigating the observed phenotype, we next used CRISPR/Cas9 technology to knock-out (KO) LIN28B in the MOLP-8 cell line. Single guide RNAs (sgRNA) targeting exons 2 and 3 of LIN28B were used (Supplemental Fig. 2a). Significant decrease in LIN28B protein levels were observed, indicating high frequency LIN28B KO in the cell population. MYC protein level were similarly decreased (Fig. 4a). Moreover, LIN28B KO resulted in de-repression of let-7, which was consistent with the shRNA experiment (Fig. 4b). Moreover, LIN28B KO led to the reduced proliferation by thymidine uptake assay (Fig. 4c).

Figure 4

LIN28B KO with CRISPR/Cas9 leads to MYC regulation and Let-7 upregulation

MOLP-8 cell line was infected with a lentiCRISPR control (sgGFP) or 2 different sgLIN28B and studied for (a) Protein blot analysis, (b) relative expression of let-7 miRNAs by qRT-PCR and (c) proliferation assay by thymidine uptake. Bars indicate mean ± SD (n = 3). P values were obtained by two-tailed Student t test (*P < 0.01).

For further characterization of the LIN28B/let-7 axis in MM, we performed RNA-sequencing of the LIN28B CRISPR-KO cells and GFP sgRNA control cells in triplicate. LIN28B was confirmed to be the most significantly down-regulated gene in LIN28B-silenced cells (Fig. 5a), confirming efficient knock-out. We then queried the top 150 down regulated genes in LIN28B KO cells against the MSigDB ‘H’ hallmark, ‘c2’ canonical pathways and ‘c3’ transcription factor target gene sets. The 10 most enriched gene sets were present in E2F cell cycle pathway regulation (Fig. 5b). Of note, E2F2, a let-7 target gene, was one of the most down-regulated genes in LIN28B KO cells. Moreover, we found a significant enrichment of let-7 target genes in control compared to LIN28B KO cells (Fig. 6a). Using an unsupervised GSEA analysis of the whole data set against the MSigDB ‘H’ hallmark gene sets, we found that MYC pathway gene set was in the top 5 genes sets, which further suggests that the LIN28B/let-7 axis regulates MYC in MM. Moreover, significant enrichment of MYC pathway in control cells was also validated using several MYC pathway gene sets (Fig. 6b). These results support a model where LIN28B represses let-7, thereby enriching let-7 targets such as E2F2 and MYC in MM.

Figure 5

RNA sequencing and differential expression of genes downstream of LIN28B

RNA sequencing was performed with MOLP-8 lentiCRISPR control and sgLIN28B#2. (a) Scatter plot showing differential expression of genes ranked by Z score (metric of fold change and −log10 of the p value) of control cells against LIN28B-silenced cells. (b) Heat map of the top 150 down-regulated and up-regulated genes in LIN28B-silenced cells. Bar plot representing the p value of the top 10 gene sets enriched in the high-LIN28B signature.

Figure 6

RNA sequencing and gene set enrichment analysis

(a) Gene set enrichment analysis (GSEA) for the let-7 target gene set and (b) GSEA analysis for several MYC gene sets in control vs. LIN28B KO cells. (c) MOLP-8 cells were infected with either lentiCRISPR control (sgGFP) or sgLIN28B#1, and transfected with a control probe or a mix of anti-let-7 b and g. Cells were studied for let-7 expression level by qRT-PCR.

Given that LIN28B has been reported to act both in a let-7-dependent and independent manner[33], we next asked whether the functional role of LIN28B is mediated via let-7 in MM. We thus performed a rescue experiment by transfection of a let-7 inhibitor into LIN28B-KO MOLP-8 cells. We observed a significant decrease of let-7 family members in LIN28B KO cells plus the anti-let-7 (Fig. 6c). Let-7 inhibition also significantly rescued cell proliferation (Supplemental Fig. 2b), indicating that let-7 is likely the main target of direct regulation by LIN28B, in let-7 dependent manner in MM.

Deregulation of let-7 in MM

Given that we demonstrated that LIN28B regulates MM proliferation through let-7, we sought to define the direct role of let-7 in MM. We first assessed the expression level of let-7 family members in primary bone marrow MM CD138+ cells compared to healthy control CD138+ bone marrow plasma cells by qRT-PCR. Let-7 miRNAs were lower in plasma cells from six patients with relapsed MM and in four MM cell lines, as compared to plasma cells from three healthy donors (Fig. 7a). To determine whether let-7 directly regulates MYC in MM, we transfected a let-7b mimic into MM.1S cells (Supplemental Fig. 3a) and observed a reduction of cell proliferation (Fig. 7b) as well as a decrease level of MYC protein (Supplemental Fig. 3b). To validate these findings in patient samples, we assessed the correlation between let-7 and MYC expressions in publicly available gene expression profiling with matched miRNA array from MM patients (GSE16558). We found a significant inverse correlation between expression of let-7b and g and the expression level of MYC in a cohort of 60 patients (Fig. 7c). These data support the idea that LIN28B acts in a let-7-dependent manner in MM and suggest that low expression levels of let-7 in MM patients contributes to MYC dysregulation and tumor proliferation.

Figure 7

Let-7 is down-regulated in MM and regulates MYC

(a) qRT-PCR analysis of let-7 family members in primary plasma cells from healthy donors and MM patients, as well as in 4 MM cell lines. (b) proliferation assay using thymidine uptake and 72 hours after transfection of a let-7b mimic or a control probe in MM.1S cells. Bars indicate mean ± SD (n = 3). P values were obtained by two-tailed Student t test (*P < 0.05). (c) Correlation between let-7 and MYC expression in patients with MM. Scatter plot showing the correlation between let-7b and g, respectively, and MYC from GSE16558. A Pearson correlation coefficient and a two-tailed p value were computed for each of them.

Let-7 as a potential therapeutic target that regulates MYC in MM

We next sought to confirm the importance of let-7 in MM in vivo and to assess whether let-7 could serve as a therapeutic strategy to directly target MYC in MM. We therefore developed a let-7 Locked Nucleic Acid (LNA)-GapmeR mimic, which was designed based on the seed region of let-7 miRNAs to efficiently mark let-7 target genes for degradation by RNAse H. We first tested the ability of the let-7 LNA-GapmeR to decrease let-7 target genes in vitro. MM.1S cell line was cultured in presence of a control probe or three different concentrations of let-7 LNA GapmeR (10nM, 100nM and 1uM). By qRT-PCR we observed a consistent decrease of MYC, KRAS, CCND1, E2F6, DICER1 and HMGA1 expression levels in parallel to increased concentration of the GapmeR (Supplemental Fig. 4a). We next tested the let-7 LNA-GapmeR in vivo in a xenograft mouse model. SCID/bg mice were injected with 5×106 MM.1S GFP+Luc+cells intravenously, followed by intraperitoneal (i.p.) injections of let-7 LNA-GapmeR two times a week. The treatment was well tolerated and did not induce obvious toxicity or weight loss. The tumor growth was evaluated by BLI, and was significantly reduced in the group of mice that received let-7 LNA-GapmeR as compared to the control group, p=0.018 (Fig. 8a and Supplemental Fig. 4b). This was associated with consequent significant survival benefit, p=0.026 (Fig. 8b). Ex-vivo, MM cells were analyzed for MYC expression, confirming the repression of MYC in the let-7 LNA-GapmeR group (Fig. 8c). This experiment provides proof of principle that let-7 could represent a new therapeutic approach targeting MYC in MM.

Figure 8

Increased expression of let-7 in vivo decreases MM proliferation

(a) Mice were followed by bioluminescence intensity (BLI), after injection of 5 million of MM.1S GFP+Luc+cells. SCID/bg mice (5 per group) were injected i.p. 2 times a week with 20mg/kg of LNA control or LNA let-7 mimic. (b) Survival of the mice by Kaplan Meier analysis. Average survival time was 35 days in control group versus 42 days in LNA let-7 mimic group, P = 0.0026. (c) qRT-PCR analysis of MYC in MM.1S cells ex vivo. Bars indicate mean ± SD (n = 3). P values were obtained by two-tailed Student t test (*P < 0.05).

Discussion

In this study, we describe the role of the LIN28B/let-7 axis in MM and identify let-7 as a potential new therapeutic approach for targeting MYC. High expression of LIN28B in MM is associated with both adverse outcomes and MYC overexpression. LIN28B represses let-7 expression in MM cells, resulting in deregulation of MYC protein and cell proliferation in vitro and in vivo. A pathway enrichment analysis in LIN28B KO cells reveals the importance of the MYC and E2F cell cycle pathways within the LIN28B/let-7 axis. Moreover, LIN28B-induced proliferation and MYC deregulation is let-7 dependent. The tumor growth impairment in vivo by administration of a let-7b based LNA-GapmeR provides proof of principle for a new therapeutic option to target MYC in MM (Supplemental Fig. 5).

Let-7 miRNAs have been described as tumor suppressor in several cancers, by targeting major oncogenic pathways[9]. Copy number loss[34-36] or epigenetic silencing[37] of individual let-7 family members has been reported in some cancers. In MM, several let-7 genes are located in frequently deleted regions, such as let-7g at 3p21, let-7i at 12q14, or let-7a-2 at 11q24. These copy number aberrations might participate in deregulation of the LIN28B/let-7 axis in MM. Here, we have described a mechanism of regulation of let-7 miRNAs in MM involving LIN28B, which inhibits let-7 miRNAs, resulting in deregulation of the MYC and E2F cell cycle pathways. Although LIN28B has been reported in liver cancer to act through both let-7-dependent and let-7-independent manners[24], our findings suggest the predominance of a let-7-dependent mechanism in MM.

LIN28B is located in the 6q21 cytogenetic band, which is amplified in some cases of neuroblastoma, resulting in LIN28B overexpression[25]. In MM, previously published CGH array did not find amplification at 6q21 locus[38]. The increased expression of LIN28B might therefore result from epigenetic changes such as altered methylation or histone modification or deregulation of upstream processes. MiR-125b was reported to inhibit LIN28B in embryonic stem cells as well as in some case of cancers[25, 39, 40]. Interestingly, miR-125b is located in 11q23, which is frequently deleted in MM. Of note, let-7 miRNAs themselves have also been reported to regulate LIN28B expression, in a feedback loop manner[41]. Moreover, a recent report suggests that inactivation of DIS3 increases LIN28B expression in MM[42], by impairing its mRNA degradation. DIS3 is an exosome endoribonuclease involved in the turnover and degradation of mRNA in the cytoplasm. Strikingly, DIS3 is one of the most frequently mutated genes in MM[43, 44], further suggesting a central role for LIN28B/let-7 axis in MM.

Despite key roles for MYC in MM, there are very few therapeutic options targeting MYC. Previous studies attempted to target MYC by using a bromodomain inhibitor to target BET proteins, which regulate MYC[6, 7]. The clinical impact of these inhibitors is being elucidated in early phase clinical trials with some potentially promising results in hematological malignancies. Our findings provide proof of principle that therapeutic use of let-7 can allow the repression of multiple oncogenes concurrently in MM. We show that in vivo use of let-7 effectively regulates MYC, which is an essential regulator of tumor progression in MM and other cancers. Our findings indicate the importance of let-7 regulation in MM and suggest that let-7 may be an effective therapeutic option that can directly target MYC in MM.