HTLV-1 bZIP factor HBZ promotes cell proliferation and genetic instability by activating oncomiRs

Viruses disrupt the host cell microRNA network to facilitate their replication. HTLV-1 replication relies on the clonal expansion of its host CD4 + and CD8 + T-cells, yet this virus causes adult T-cell leukemia/lymphoma (ATLL) that typically has a CD4 + phenotype. The viral oncoprotein Tax, which is rarely expressed in ATLL cells, has long been recognized for its involvement in tumor initiation by promoting cell proliferation, genetic instability, and miRNA dysregulation. Meanwhile, HBZ is expressed in both untransformed infected cells and ATLL cells, and is involved in sustaining cell proliferation and silencing virus expression. Here, we show that an HBZ/miRNA axis promotes cell proliferation and genetic instability as indicated by comet assays that showed increased numbers of DNA strand breaks. Expression profiling of miRNA revealed that infected CD4 + cells, but not CD8 + T-cells, overexpressed oncogenic miRNAs, including miR-17 and miR-21. HBZ activated these miRNAs via a posttranscriptional mechanism. These effects were alleviated by knocking down miR-21 or miR-17 and by ectopic expression of OBFC2A, a DNA damage factor that is down-regulated by miR-17 and miR-21 in HTLV-1 infected CD4+ T-cells. These findings extend the oncogenic potential of HBZ and suggest that viral expression might be involved in the remarkable genetic instability of ATLL cells. assess whether HBZ had a significant effect on genomic integrity. To this end, we used an alkaline comet assay to determine whether HBZ expression triggered DNA ATLL samples (30-33, 47-50). Our results here add to these findings, as the miRNA interferes with numerous transcription factors, it did not appear to stimulate miR-17 and miR-21 expression via transcriptional activation, as is the case for Tax and miR-146a expression via NF- κ B Rather, our results demonstrated that HBZ increased the expression of miR-17 and miR-21 without increasing the transcription of their pri-miRNA transcripts. This HBZ-dependent post-transcriptional deregulation represents a the HBZ DNA sequence is regularly spared from epigenetic silencing in Taken together with our results, this suggests that HBZ might be more influential than Tax in sustaining persistent clonal expansion and promoting early leukemogenesis through its dual cellular effect, namely proliferation and Whitney alleviates the effect of HBZ on cell growth. cells were transiently co-transfected with vectors encoding an OBFC2A mRNA unresponsive to miRNA-mediated modulation in the presence or absence of NCS. Cell proliferation was assayed using MTT


Introduction
4 HBZ/miRNA axis promotes genetic instability key role of HBZ in the continuous expansion of ATLL cells (19). HBZ is also involved in the non-malignant clonal expansion of infected cells, as its expression is correlated with proviral loads, inflammatory markers, and disease severity in TSP/HAM (20). In line with this activity, transgenic mice expressing HBZ develop symptoms similar to those observed in HTLV-1 carriers, including T-cell lymphomas and inflammatory lesions (21). These data indicate that, in addition to Tax, HBZ might play a significant role in early leukemogenesis through molecular mechanisms that remain to be elucidated. To date, the effect of HBZ on genetic instability has not been investigated, even though ATLL cells express HBZ and display a dynamic spectrum of both complex cytogenetic abnormalities and somatic mutations in vivo (13,(22)(23)(24).
MicroRNAs (miRNAs) are evolutionarily conserved, small (~21 nucleotides), noncoding RNAs that are encoded within the genomes of almost all eukaryotes from plants to mammals. Most organisms express hundreds of miRNAs that are integral to almost all known biological processes. In general, miRNAs post-transcriptionally regulate protein synthesis by base pairing to partially complementary sequences in the 3′ untranslated regions (UTRs) of target mRNAs (25)(26)(27)(28)(29). In human diseases, particularly cancer, epigenetic and genetic defects in miRNAs and their processing machinery are common hallmarks of disease. ATLL cells exhibit global repression of miRNA expression (30), including the loss of miR-31 expression that has been shown to activate NF-κB (30). At the molecular level, both Tax and the viral RNA binding protein Rex modulate miRNA expression in immortalized cell lines by altering NF-κB activation or the RNAi machinery (12,(31)(32)(33). To our knowledge, no study to date has assessed the effects of HBZ on the biogenesis and activity of miRNAs. Furthermore, whereas human lymphocyte subsets are known to possess specific miRNA signatures involved in Tcell differentiation and activation (34,35), little is known about the role of miRNA dysregulation in the clonal expansion of untransformed infected CD4 + and CD8 + T-cells in Research. on September 6, 2017. © 2014 American Association for Cancer cancerres.aacrjournals.org Downloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
In order to assess the effect of HTLV-1 infection on the miRNA expression profiles of host cells in vivo, we performed an integrated analysis of miRNA-and mRNA-expression profiles of cloned CD4 + and CD8 + T-cells derived from infected individuals without malignancy. When compared to their uninfected counterparts, each T-cell subset displayed specific changes in miRNA expression. Accordingly CD4 + T-cells, but not CD8 + non-immortalized T-cells, overexpressed known oncogenic miRNAs such as miR-17 and miR-21, which target the DNA damage effector OBFC2A/hSSB2 in these cells. Surprisingly, in CD4 + T-cells derived from infected individuals, the expression of these miRNAs strongly correlated with HBZ expression, but not Tax expression. In fact, miR-17 and miR-21 were found to be posttranscriptionally upregulated by HBZ, while HBZ/miRNA-mediated downregulation of OBFC2A expression triggered both cell proliferation and genomic instability. Together, our results reveal a close association of HBZ expression, miRNA dysregulation, genetic instability, and abnormal cell proliferation. Research.
on September 6, 2017. © 2014 American Association for Cancer cancerres.aacrjournals.org Downloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Materials/Subjects and Methods
Materials and methods are detailed in the Supplemental Information. Peripheral blood mononuclear cells (PBMC) were obtained from HTLV-1-infected individuals after obtaining written informed consent according to the principles expressed in the Declaration of Helsinki.
Pertinent clinical data accompanying the samples are given in Table 1. The study was approved by the Institutional Review Board of the Hospices Civils de Lyon, France. HeLa cells were obtained from the American Type Culture Collection.

Cell culture and microarray analysis
HeLa cells were cultured in Dulbecco's modified Eagle medium (DMEM), supplemented with 10% heat-inactivated FCS, L-glutamine, penicillin (5 U/mL) and streptomycin (5 µg/mL) at 37 °C in a humidified atmosphere with 5% CO 2 . PBMCs were cultured by limiting dilution cloning as previously described (6) and detailed in the Supplemental Information. During the growth period, cellular clones were not immortalized and required IL-2 and stimulation with phytohemagglutinin (PHA) every two weeks for continued growth. RNA expression profiling of 24 cloned T-cells (6 infected CD4+, 6 infected CD8+, 6 uninfected CD4+, and 6 uninfected CD8+ T cell clones) was carried out using an Agilent V3 miRNA microarray (Agilent, Santa Clara, CA) and a GeneChip Human Exon 1.0 ST array (Affymetrix, Santa Clara, CA) according to the manufacturer's instructions. The miRNA and mRNA array data have been submitted to GEO #GSE46345 and #GSE46518, respectively.

Quantitative RT-PCR and luciferase reporter assay
Quantitative RT-PCR assays are detailed in the Supplemental Information. Oligonucleotide sequences are available upon request. Reporter assays using the 3'UTR-OBFC2A target region fused to the 3'UTR of the luciferase gene in the pMIR-REPORT™ miRNA expression vector (Invitrogen, Carlsbad, CA) were used to test the suppressive capacity of miR-17 and
The minimum number of nucleotides for a miRNA response element (MRE) was set to 7 (pvalue = 0.05). The search for MREs was restricted to the 3'-UTR of mRNAs. Using Access software, in silico predicted miRNA targets were cross-matched with downregulated mRNAs (fold change ≥ 1.5, p-value = 0.05) of infected CD4 + T-cells compared with those of uninfected cells.

Comet Assay
Research.
on September 6, 2017. © 2014 American Association for Cancer cancerres.aacrjournals.org Downloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Transiently transfected HeLa cells were grown for 24h and exposed to 5 ng/mL of neocarzinostatin (NCS) or control medium without NCS for 3 h. After 24h of additional growth, cells were collected and resuspended in cold PBS at a density of 10 5 cells/mL and a comet assay was performed, according to the manufacturer's protocol (Trevigen).
Experiments were done in triplicates and DNA damage was quantified for 50 cells for each experimental condition by determining the tail moment, a function of both the tail length and intensity of DNA in the tail relative to the total DNA amount, using the software Comet Score (TriTek).

MTT proliferation assay
Cell proliferation rates were assessed by MTT assay (Cell Titer 96 ® Non-Radioactive Cell Proliferation Assay, Promega). Fifty thousand cells were plated and cultured for 24 h in 100 µL medium in 96-well plates. The MTT dye (Promega, 15 µL) was added to each well for the last hour of culture incubation. Supernatants were decanted and 100 µL of the Solubilization Solution/Stop Mix was added to each well, and cells were incubated for 1 hour at room temperature. The OD value was measured using an ELISA reader at 570 nm, with 650 nm as a reference. Experiments were done in triplicate and repeated at least three times. MTTpredicted cell proliferation was also confirmed by cell counting using Trypan blue dye (not shown).

Results miRNA signature of non-immortalized HTLV-1-positive T-cells
Human lymphocyte subsets possess specific miRNA signatures that can be modified upon viral infection (34,35). To address which changes in miRNA expression accompany cellassociated replication of HTLV-1, we carried out miRNA expression profiling of infected and uninfected CD4 + and CD8 + T-cells derived from naturally infected individuals without malignancy ( Table 1). Twenty-four T-cell clones were assayed for miRNA microarray analysis as described in the Methods section. They included 6 uninfected CD4+ T-cell clones, 6 infected CD4+ T cell clones, 6 uninfected CD8+ T-cell clones, and 6 infected CD8+ T-cell clones. The complete microarray data set has been deposited en toto into the GEO database and is available under record number GSE46345. HTLV-1-positive CD4 + T-cells expressed higher levels of the miRNAs miR-27b, miR-23b, and miR-34b compared to uninfected CD4 + T-cells (fold change > 1.5, p < 0.05). In contrast, infected CD8 + T-cells expressed higher levels of miR-34b, let-7a, and let-7f compared to their uninfected counterparts. Additional miRNA gene families such as mir-17, mir-21, mir-15, and let-7 were found to be differentially expressed among infected clones. The differential expression of these miRNAs was confirmed in T-cell clones used for microarray analysis using miRNA-specific real time quantitative RT-PCR (qRT-PCR) (not shown). We next assessed the significance of the microarray data by quantifying the expression level of some miRNAs using qRT-PCR analysis of 65 additional T-cell clones. Figure 1 shows that infected CD4 + clones expressed significantly higher levels of miR-17 (p = 0.0056, Mann-Whitney test), miR-21 (p = 0.011), miR-23b (p = 0.0063), and miR-27b (p = 0.023) than uninfected CD4 + clones. Infected CD4+ clones exhibited lower expression levels of miR-374b than their CD8+ counterparts, however this difference was not statistically significant between the infected and uninfected cell categories (Supplemental Figure 1). In contrast, qRT-PCR analysis revealed a wide range of expression level of other miRNAs within infected cells and failed to demonstrate any significant changes across the different subtypes (Supplemental Figure 1). Together, these results indicated that significant high levels of miR-17, miR-21, miR-23b, and miR-27b constitute a miRNA signature of non-immortalized HTLV-1-infected CD4+ cells but not CD8+ T cells in vivo.
Similarly, there was no correlation between the expression of Tax/Rex mRNA and that of the 6 analyzed miRNAs ( Figure 2B). To more confidently assess the role of HBZ in miRNA expression, we performed transient transfection assays of HeLa cells with increasing amounts of an HBZ-expressing vector and quantified the expression of endogenous miR-17, miR-21, miR-23b, and miR-27b by qRT-PCR. As shown in Figure 3A, HBZ was capable of upregulating the expression of these miRNAs in a concentration-dependent manner. In contrast, transfection with a vector expressing Tax/Rex did not significantly affect the expression of the 6 microRNAs (data not shown). Research.
on September 6, 2017. © 2014 American Association for Cancer cancerres.aacrjournals.org Downloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

HBZ stimulates the expression of miR-17 and miR-21 at a post-transcriptional level
The miR-21 and miR-17 are two of the best characterized oncogenic miRNAs and are involved in numerous malignancies (36)(37)(38)(39). The gene miR-21, along with its own promoter, is located in an intronic region of a coding gene, TMEM49. Meanwhile, miR-17 is encoded by the polycistronic miRNA cluster mir-17-92, also called oncomir-1 (36). The primary transcript (pri-miR) of this cluster contains six tandem stem-loop hairpin structures that encode 6 mature miRNAs: miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, and miR-92-1 (40). Our microarray data indicated that these miRNAs did not follow the same expression pattern upon infection, thereby suggesting that miR-17 up-regulation likely occurs at a posttranscriptional level. Having found that HBZ can stimulate miR-17 expression, we examined the effects of HBZ on the expression of the endogenous pri-miRNA that encodes miR-17 in HeLa cells ( Figure 3B). In contrast to mature miR-17, pri-miR-17-92 expression was not significantly altered by HBZ expression. This result confirmed that HBZ stimulates the biogenesis of miR-17 at a post-transcriptional level. Similar results were obtained for miR-21 ( Figure 3B). In line with this result, pri-miR-17 and pri-miR-21 expression was not significantly different between infected and uninfected CD4+ clones (Supplemental Figure   2). Collectively, these data suggest that HBZ can promote the post-transcriptional maturation of certain miRNAs.

HTLV-1-associated miRNAs from CD4 + T-cells regulate genes involved in genetic disorders
To identify the mRNA targets of miRNAs that are deregulated in HTLV-1 infected cells, we examined the entire genome expression profile for the 12 CD4 + clones used for miRNA analysis. The data are available online in the GEO database (GSE46518). Overall, 224 genes showed modified expression levels in infected cells as compared with their uninfected counterparts (fold change > 1.5, p < 0.01). In parallel, MiRWalk computational analysis was used to identify the putative target genes of miR-17, miR-21, miR-23b, and miR-27b, as detailed in the Methods section. Among 117 repressed genes, 63 (53%) were computationally predicted to be targeted by miR-17, miR-21, miR-23b, or miR-27b. Fortytwo, 37, 29, and 45 transcripts harbored at least one miRNA response element (MRE) for miR-17, miR-21, miR-23b, or miR-27b, respectively. Eighty percent ( Quantitative RT-PCR analysis confirmed the microarray data for 5 of these genes (Supplemental Figure 3).
Pathway analysis was carried out using Ingenuity Pathway Analysis (IPA; Ingenuity Systems). Of the 224 genes that were found to be dysregulated in HTLV-1-positive CD4 + Tcells, the majority are involved in genetic disorders (45%) and neurological diseases (33%) (Supplemental Figure 4). These two dominant pathways were over-represented, even when the analysis was restricted to the 63 genes targeted by miR-17, miR-21, miR-23b and miR-27b (41% and 23%, respectively). In contrast, the remaining pathways had varying representations between the two groups of genes.

HBZ promotes genomic instability in a miR-17-and miR-21dependent manner
We next wanted to assess whether HBZ had a significant effect on genomic integrity. To this end, we used an alkaline comet assay to determine whether HBZ expression triggered DNA putative MREs for miR-17, miR-21, miR-23b, and miR-27b. We next used qRTPCR analysis to confirm OBFC2A expression in T-cell clones derived from patients. As shown in Figure  This difference remained significant when the outlier in the control group was withdrawn from the analysis (p=0.03, Mann Whitney test). We then examined the effect of HBZ expression on endogenous OBFC2A gene expression in HeLa cells. As shown in Figure 5B and 6C, HBZ expression decreased OBFC2A expression at both the mRNA (p=0.0095, Mann-Whitney test) and protein (hSSB2) level.
To clarify the role of miRNAs in the regulation of OBFC2A by HBZ, we co-transfected HeLa cells with an HBZ-expressing vector alongside LNA-antimiR-17, LNA-antimiR-21, or a scrambled miRNA control. As shown in Figure 5C, the functional inactivation of miR-17 and miR-21 alleviated the repressive effect of HBZ on hSSB2 expression while ectopic overexpression of miR-17 and miR-21 resulted in a reduced level of endogenous OBFC2A/hSSB2 mRNAs (Supplemental Figure 6). Taken together, these results suggested that OBFC2A/hSSB2 might be a bona fide target of HBZ-induced miRNAs. To test this hypothesis, we constructed pMIR-REPORT constructs that express a luciferase RNA fused with either the wild-type sequence of the OBFC2A 3'-UTR (p3'-OBFC2A-WT) or a 3'UTR sequence mutated for the MREs of miR-17 (p3'-OBFC2A-mut17), miR-21 (p3'-OBFC2A-mut21), or both (p3'-OBFC2A-ALL; Figure 5D). A significant decrease in luciferase activity was observed in cells co-transfected with the p3'-OBFC2A-WT and miRNA mimics (either synthetic pre-miR-17 or pre-miR-21), but not in cells co-transfected with p3'-OBFC2A-mut17 and p3'-OBFC2A-mut21, suggesting that OBFC2A is a direct target of miR-17 and miR-21 ( Figure 5E). Consistent with this result, increased luciferase activity was detected in cells transfected with p3'-OBFC2A-WT along with LNA-antimiR-17 or LNA-antimiR-21 (Supplemental Figure 6). As shown in Figure 5F, ectopic HBZ expression significantly We then investigated the proliferation rate of HBZ cells exposed to a low dose of NCS, which was sufficient to cause a significant reduction in MTT activity in control cells ( Figure 6A).
Interestingly, compared to control cells, NCS-treated HBZ-expressing cells were unaffected in their proliferation capacity, suggesting that HBZ enabled cells to tolerate DNA damage. This To determine the role of hSSB2 in HBZ/miRNA-mediated cell proliferation in the presence of DNA damage, we co-transfected HeLa cells with the HBZ vector and an hSSB2-expressing vector. The latter construct lacked the OBFC2A-3'UTR, rendering it unresponsive to miRNAmediated modulation. Ectopic hSSB2 expression induced a mild decrease in MTT activity in untreated control cells, whereas it significantly reinforced the growth inhibition in NSCtreated cells ( Figure 6B). The largest difference was observed in HBZ-expressing cells wherein ectopic hSSB2 expression caused a decrease in the proliferative capacity of these cells by 44% and 33% in the presence and absence of NCS, respectively ( Figure 6B). We next investigated whether repression of the proliferation of HBZ-expressing cells by hSSB2 was correlated with a decrease in DNA damage levels. As shown in Figure 6C, hSSB2 overexpression significantly reduced the frequency of DNA strand breaks in NCS-exposed HBZ-expressing cells after 24 hours of recovery. A similar effect was observed in the absence of NCS exposure ( Figure 6C). In addition, H2AX (γH2AX) phosphorylation was also decreased in cells expressing ectopic hSSB2, which confirmed the decrease in DNA damage, as H2AX is a sensitive indicator of both DNA damage and DNA replication stress ( Figure 6D). Taken together, these results strongly suggest that HBZ/miRNA-mediated downregulation of hSSB2 contributes, at least in part, to abnormal cell proliferation and genetic instability. In vivo, T-cells persistently infected with HTLV-1 acquire numerous phenotypic changes with respect to cell morphology, apoptosis, proliferation, DNA content, and telomere homeostasis expression profile of non-immortalized infected T-cells derived from carriers without malignancies was significantly distinct from that of uninfected cells. Remarkably, infected CD4 + T-cells overexpressed known oncogenic miRNAs, such as miR-17 (36), miR-21 (37)(38)(39), miR-23b (51,52), and miR-27b (51,53,54) (Figure 1). Interestingly, we found that the predicted targets for the oncogenic miRNAs overexpressed in these HTLV-1 + CD4 + T-cells were significantly enriched for genes involved in genetic disorders, neurological diseases, cellular growth and proliferation, gene expression, and cellular function and maintenance ( Figure 2). Together, these differences in miRNA expression are consistent with the known differences in the effects of the virus on these T-cell subsets and fit well with the preleukemic phenotype of HTLV-1-positive non-immortalized CD4 + T-cells (5,6).
The pleiotropic effects of Tax were previously thought to govern HTLV-1-associated miRNA dysregulation. By targeting Drosha to the proteasome, Tax influences global miRNA biogenesis (49). Furthermore, Tax activates the transcription of miR-130b (33) and miR-146a (32), and downregulates miR-149 and miR-873, which both directly target the chromatin remodeling factors p300 and p/CAF that are known to play a critical role in HTLV-1 pathogenesis (47). Non-immortalized HTLV-1-positive clones express a wide range of Tax RNA levels, likely modifying numerous cellular pathways (5)(6)(7)(8). Here, we found no relationship between Tax and miRNA expression in these cells. Instead, we found that the level of HBZ expression strongly correlated with that of four oncomiRNAs that are overexpressed in CD4 + T-cells (Figure 2). Transient transfection assays in HeLa cells confirmed these relationships and characterized HBZ as a new player in miRNA homeostasis.
Interestingly, although HBZ interferes with numerous transcription factors, it did not appear to stimulate miR-17 and miR-21 expression via transcriptional activation, as is the case for Tax and miR-146a expression via NF-κB (31). Rather, our results demonstrated that HBZ novel mechanism that allows the activation of a restricted number of miRNAs belonging to a polycistronic microRNA cluster, such as the miR-17-92 cluster ((36) and present results). As such, future investigations will be important to address the molecular mechanisms underlying the post-transcriptional effect of HBZ.
Genetic instability is a hallmark of HTLV-1 infection that helps the infected cells to escape the strong anti-HTLV-1 cytotoxic T-lymphocyte (CTL) response. Genetic instability can also protect infected cells from clonal dominance and further transformation by promoting neoantigen formation (5,23,24). On the other hand, HTLV-1-triggered gene alterations have been proposed to mount a mutator phenotype that propitiates leukemogenesis (5,23,24).
So far, Tax has been recognized as the main source of HTLV-1-associated genetic instability.
Here, we found that HBZ promotes oncomiRNA expression as well as DNA strand breaks.
These effects relied, at least in part, on the downregulation of OBFC2A via the HBZdependent post-transcriptional activation of miR-17 and miR-21 ( Figures 5 and 6). OBFC2A encodes hSSB2, which intervenes with ATM signaling and promotes subsequent activation of DNA repair and cell cycle checkpoints (55,56). Importantly, our results also showed that HBZ expression does not promote growth arrest in DNA-damaged cells. On the contrary, HBZ-expressing cells continued to proliferate, even in the presence of NCS ( Figure 6). This phenotype was reversed by ectopic expression of OBFC2A, which decreased the proliferation rate and restored the DNA damage response in the same HBZ-expressing cells ( Figure 6).
Given this novel source of genetic instability, to what extent can it be exploited, if at all, by the infected cell? During non-malignant stages of infection when HBZ is expressed (43), we can propose that, similar to Tax, HBZ helps persistent clonal expansion in the face of host immune response. However, in sharp contrast to Tax, the HBZ DNA sequence is regularly spared from epigenetic silencing in vivo (43). Taken together with our results, this suggests that HBZ might be more influential than Tax in sustaining persistent clonal expansion and promoting early leukemogenesis through its dual cellular effect, namely proliferation and genetic instability. At the malignant stage, our results indicate that HBZ expression might be implicated in the plethora of genetic defects that characterize ATLL cells. These cells regularly display somatic mutations (23,24) and complex cytogenetic abnormalities with evidence of frequent clonal cytogenetic evolution during disease progression (22,(57)(58)(59)(60).
These acquired genetic abnormalities and their clonal evolution impact on disease aggressiveness, relapse risk, and resistance to treatment (22,(57)(58)(59)(60). Given that ATLL cells are Tax-negative but regularly express HBZ (43), the intraclonal genetic drift that characterizes ATLL cells was thought to result only from the tumor genotype, as is the case for virus-unrelated leukemias (61). However, our results indicating that HBZ triggers DNA strand breaks in proliferating cells suggest that HBZ expression in tumor cells might contribute to the dynamic spectrum of complex cytogenetic abnormalities and somatic mutations that characterize ATLL cells (22,(57)(58)(59)(60). However, these cells also have low levels of numerous miRNAs, including miR-17 (30), suggesting that the HBZ/miRNA axis might be impaired at the late tumor stage. These data argue for a role of the HBZ/miRNA axis, and subsequent genetic instability, in initiating rather that promoting HTLV-1-associated leukemogenesis. Interestingly, Figure 4C shows that HBZ remains capable of triggering DNA damage in the absence of functional miR-17 and miR-21, suggesting that additional mechanisms connect HBZ to genetic instability.
In conclusion, our results suggest that the preleukemic phenotypes of HTLV-1-positive CD4 + T-cells include an oncogenic miRNA profile that is promoted by HBZ. We show that HBZ inactivates OBFC2A via oncomiRNAs, and thereby promotes cell proliferation and genetic instability. Thus, the notion that Tax is the unique generator of genetic instability may require revision. In addition, as in the case of DNA tumor viruses (62), HTLV-1 expression in tumor cells might be an important factor in triggering clonal evolution. As a corollary, targeting HBZ might be a promising approach for preventing and treating ATLL.