Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Jun 13.
Published in final edited form as: Cancer Cell. 2016 Jun 13;29(6):874–888. doi: 10.1016/j.ccell.2016.04.016

ΔNp63/DGCR8-Dependent microRNAs Mediate Therapeutic Efficacy of HDAC Inhibitors in Cancer

Marco Napoli 1,2, Avinashnarayan Venkatanarayan 1,2,3, Payal Raulji 1,2, Brooke A Meyers 1,2, William Norton 4, Lingegowda S Mangala 5,6, Anil K Sood 5,6, Cristian Rodriguez-Aguayo 6,7, Gabriel Lopez-Berestein 6,7, Harina Vin 8, Madeleine Duvic 8, Michael B Tetzlaff 8,9, Jonathan L Curry 8,9, Alain H Rook 10, Hussein A Abbas 1,2, Cristian Coarfa 11, Preethi H Gunaratne 12, Kenneth Y Tsai 2,3,8, Elsa R Flores 1,2,3,6,*
PMCID: PMC4908836  NIHMSID: NIHMS785717  PMID: 27300436

SUMMARY

ΔNp63 is an oncogenic member of the p53 family and acts to inhibit the tumor suppressive activities of the p53 family. By performing a chemical library screen, we identified HDAC inhibitors (HDACi) as agents reducing ΔNp63 protein stability through the E3 ubiquitin ligase, Fbw7. ΔNp63 inhibition decreases the levels of its transcriptional target, DGCR8, and the maturation of let-7d and miR-128, which we found critical for HDACi function in vitro and in vivo. Our work identified Fbw7 as a predictive marker for HDACi response in squamous cell carcinomas and lymphomas and unveiled let-7d and miR-128 as specific targets to bypass tumor resistance to HDACi treatment.

eTOC Blurb

graphic file with name nihms785717u1.jpg

Napoli et al. show that HDAC inhibitors (HDACi) target the ΔNp63/DGCR8 axis to inhibit ΔNp63-dependent tumors, especially those lacking functional p53. ΔNp63 inhibition decreases DGCR8 expression and maturation of let-7d and miR-128. They alsoidentify Fbw7 level as a predictive marker for HDACi response.

INTRODUCTION

The p53 family member and p63 isoform, ΔNp63, is a transcription factor essential for terminal differentiation of stratified epithelia, such as the epidermis (Chakravarti et al., 2014). We found that ΔNp63 controls terminal differentiation through transcriptional regulation of DGCR8 (Chakravarti et al., 2014), a crucial component of the miRNA biogenesis pathway. In this manner, ΔNp63 controls the maturation of a specific group of miRNAs necessary for the induction of terminal differentiation of the skin (Chakravarti et al., 2014).

ΔNp63 is highly expressed in many tumors (Karni-Schmidt et al., 2011; Marchini et al., 2008; Orzol et al., 2014) suggesting its role as an oncogene. We have demonstrated the role of ΔNp63 as an oncogene in vivo using ΔNp63 conditional knockout mice (Venkatanarayan et al., 2015). By intercrossing these mice with p53−/− mice, which develop thymic lymphomas with high penetrance, we found that deletion of ΔNp63 in p53−/− tumors caused their complete regression (Venkatanarayan et al., 2015). Transcriptional profiling analysis comparing p53−/− thymic lymphomas to those lacking p53 and ΔNp63 revealed a transcriptional program of mRNAs and miRNAs critical for tumor regression. Based on these data, we deemed that the inhibition of the ΔNp63/DGCR8 axis may be crucial to inhibit ΔNp63 oncogenic functions and may provide additional therapeutic options to target p53-mutant or deficient tumors relying on this oncogenic axis.

RESULTS

Identification of ΔNp63/DGCR8 Inhibitors

In order to pharmacologically inhibit ΔNp63 function, we screened a drug library composed of 855 pharmacologically active compounds, including 415 FDA approved drugs, in MCF-10A cells. To assess the efficacy of these compounds in affecting the protein levels of ΔNp63 and DGCR8, cells were evaluated by immunofluorescence and quantified with the Celigo cell imaging cytometer (Nexcelom Bioscience). Only 5 out of the 855 tested compounds delivered at a concentration of 1 μM significantly reduced the expression of both ΔNp63 and DGCR8 by over 75% compared to the DMSO treated cells. Intriguingly, all 5 of the compounds were hydroxamate based pan-HDACi : AR-42, ITF2357, JNJ-26481585, panobinostat, and SB939 (Figure 1A). To verify the effect of these HDACi at lower doses, MCF-10A cells were treated at 30 nM concentration for 48 h, and western blot analysis of ΔNp63 and DGCR8 was performed. At this dosage, 3 HDACi (AR-42, JNJ-26481585, and panobinostat) were still effective in reducing ΔNp63 and DGCR8 (Figure 1B). This outcome was observed in a panel of transformed and non-transformed murine and human cell lines (Figures 1C, 1D and S1A–S1D) showing that HDACi can decrease ΔNp63 and DGCR8 protein levels in a broad range of cells. To assess the effect of the HDACi on ΔNp63 levels, a time course followed by western blot analysis of ΔNp63 was performed in a cutaneous squamous cell carcinoma (SCC) cell line, Colo16, treated with JNJ-26481585 (Figure 1E). We found a reduction in ΔNp63 protein levels as early as 6 hours after treatment.

Figure 1. Identification of ΔNp63/DGCR8 Inhibitors.

Figure 1

Open in a new tab

(A) Scatter graph of the ΔNp63 (◆) and DGCR8 (○) immunofluorescence signals in MCF-10A cells treated for 48 hr with 1μM of 855 compounds and compared with the DMSO condition. (B) Representative western blot (WB) analysis of MCF-10A cells treated for 48 hr with 30 nM of the indicated HDAC inhibitors (HDACi). Asterisk indicates a non-specific band recognized by the DGCR8 antibody. (C–D) Representative WB analyses of the specified murine (C) and human (D) cell lines treated for 48 hr with 30 nM of JNJ-26481585. (E) Representative WB analysis of Colo16 cells treated with 30 nM of JNJ-26481585 for the indicated time (hours). See also Figure S1.

HDACi Affect ΔNp63 Protein Stability in an Fbw7/HDM2-Dependent Manner

To understand whether HDACi affect ΔNp63 and DGCR8 either at the protein or at the transcriptional level, we analyzed mRNA levels of ΔNp63 and DGCR8 in two cutaneous SCC cell lines, Colo16 and SBR12, and in normal human epidermal keratinocytes (NHEKs). While ΔNp63 mRNA levels were unperturbed by the treatment, the mRNA levels of its direct target gene DGCR8 were significantly reduced in all 3 cell lines (Figure 2A). Thus, we hypothesized that these compounds may reduce ΔNp63 protein stability and in turn affect DGCR8 expression. This hypothesis was tested by adding a proteasome inhibitor to the 3 cell lines treated with the HDACi. This experiment resulted in the rescue of ΔNp63 protein levels as well as DGCR8 mRNA levels (Figures 2B and 2C). Similar results were obtained in H1299 and HeLa cells, suggesting that these HDACi share a common mechanism of action in a broad range of cell lines (Figure S2A–C).

Figure 2. HDACi Affect ΔNp63 Protein Stability in an Fbw7/HDM2-Dependent Manner.

Figure 2

Open in a new tab

(A) Quantitative real time PCR (qRT-PCR) in the indicated cell lines treated for 48 hr either with 30 nM of JNJ-26481585 or with vehicle (DMSO). Data are mean ± SD, n = 3, * versus DMSO, p < 0.005, two-tailed t-test. (B) Representative WB analysis of the indicated cell lines treated with DMSO (−) or JNJ-26481585 for 42 hr and with (+) or without (−) 5 μM of the proteasome inhibitor MG132 for additional 6 hr. (C) qRT-PCR in the same conditions as in B. Data are mean ± SD, n = 3, * versus DMSO, # versus JNJ-26481585 only, p < 0.005, two-tailed t-test. (D) qRT-PCR in Colo16 cells treated with DMSO or JNJ-26481585 and transfected either with the indicated siRNAs. Data are mean ± SD, n = 3, * versus respective siC, p < 0.005, two-tailed t-test. (E) Representative WB analysis of Colo16 cells treated with DMSO or JNJ-26481585 and transfected with the indicated siRNAs. (F) qRT-PCR in the same conditions as in E. Data are mean ± SD, n = 3, * versus siC DMSO, # versus siC JNJ-26481585, p < 0.005, two-tailed t-test. (G) Table summarizing the clinical data of the mycosis fungoides (cutaneous T cell lymphoma) biopsies treated with HDACi and analyzed by IHC for Fbw7 expression. Average score in HDACi-responder samples (green) = 4.72; Average score in HDACi-non-responder samples (red) = 2.72; p = 0.000219538, two-tailed t-test. (H) IHC for Fbw7 in biopsies of mycosis fungoides. Positive nuclei are brown. Hematoxylin (purple) was used as a counterstain. Scale bars represent 200 μm (panels) and 50 μm (magnified inserts). See also Figure S2.

Next, we investigated the mechanism through which these HDACi decrease ΔNp63 protein stability, and asked whether E3 ubiquitin ligases could be required for this process. Therefore, we focused on E3 ubiquitin ligases reported to target ΔNp63: Fbw7 (Galli et al., 2010; Restelli et al., 2014), Itch (Rossi et al., 2006a; Rossi et al., 2006b), HDM2 (Galli et al., 2010), Pirh2 (Jung et al., 2013; Yan et al., 2013), and WWP1 (Li et al., 2008). Colo16 cells were transfected with siRNAs against each of these factors individually and treated with the HDACi, JNJ-26481585, or the vehicle alone, DMSO. mRNA levels of each E3 ubiquitin ligase was measured by qRT-PCR subsequent to treatment with the corresponding siRNAs. In each case, a significant reduction in mRNA expression was achieved for each E3 ubiquitin ligase and this effect was not perturbed by the subsequent treatment with JNJ-26481585 (Figure 2D). We found that ΔNp63 protein downregulation by JNJ-26481585 was rescued by FBW7 depletion and, in part by HDM2 (Figure 2E). A similar rescue by the depletion of FBW7 and HDM2 was also observed in H1299 cells treated with AR-42 (Figure S2D), suggesting that the various HDACi reduce ΔNp63 protein levels in an Fbw7/HDM2-dependent fashion. As a consequence of the restoration of ΔNp63 levels by FBW7 depletion, the levels of its target gene DGCR8 concomitantly increased up to its physiological steady state conditions (Figure 2F).

Based on these data, we asked whether FBW7 expression is predictive of HDACi response. Indeed, we found that cutaneous squamous cell carcinoma (cSCC) and head and neck squamous cell carcinoma (HNSCC) cell lines and patient derived cutaneous T cell lymphoma (cTCL) cell lines with previously characterized responsiveness to HDACi (Zhang et al., 2005) all showed that higher FBW7 levels corresponded to higher sensitivity to HDACi (Figure S2E–G). Again, protein levels of Fbw7 were inversely correlated with levels of ΔNp63 and DGCR8 (Figure S2F & G). We further investigated the potential of Fbw7 as a predictive marker for the therapeutic response of HDACi by assessing Fbw7 protein levels in a set of 22 biopsies of mycosis fungoides, the most common form of cTCL that has gained FDA approval of HDACi as a single anti-cancer agent (West and Johnstone, 2014). These 22 biopsies were derived from 15 patients and reflect 11 tumors that responded to HDACi therapy and 11 tumors that were non-responsive to HDACi (Figure 2G). Importantly, we found that Fbw7 protein levels were higher and more frequent in biopsies of HDACi-responding tumors (Figure 2H), thus indicating that high Fbw7 is predictive of HDACi response. Overall, these data show that HDACi impair the stability of ΔNp63 protein through Fbw7 and consequently reduce DGCR8 mRNA expression. Moreover, Fbw7 can be used as a predictive marker for HDACi response as illustrated in patient derived cell lines and patient biopsies.

HDACi Impair DGCR8-Dependent miRNA Biogenesis

We have recently reported that ΔNp63 is essential for the maintenance of p53−/− thymic lymphomas (Venkatanarayan et al., 2015), thus understanding mechanisms for the inhibition of ΔNp63 is essential to targeting tumors with defects in the p53 pathway. We found that the reduced ΔNp63 mRNA levels were associated with a decrease in DGCR8 mRNA levels in tumors deficient for both p53 and ΔNp63 compared to tumors deficient for p53 only (Figures S3A and S3B), suggesting that the transcriptional regulation of DGCR8 by ΔNp63 that we have previously demonstrated in keratinocytes (Chakravarti et al., 2014) is also present in tumors. Because levels of DGCR8 are low in ΔNp63/p53 double deficient tumors, we analyzed the expression of miRNAs using next generation miRNA-sequencing and found that 16 miRNAs were significantly downregulated in ΔNp63/p53 double deficient (ΔNp63Δ/Δ;p53−/−) thymic lymphomas compared to those from p53−/− mice (ΔNp63fl/fl; p53−/−) (Figure 3A). To determine which of the 16 miRNAs are most relevant for the observed tumor regression, we performed functional pair analysis (Creighton et al., 2008) using data from our miRNA-seq and RNA-seq analysis of ΔNp63Δ/Δ;p53−/− thymic lymphomas to identify downregulated miRNAs whose putative mRNA targets were found to be upregulated in ΔNp63Δ/Δ;p53−/− thymic lymphomas (Venkatanarayan et al., 2015). Using this analysis, we identified let-7d and miR-128 (Figure 3B) and validated their reduced levels in ΔNp63/p53 double deficient thymic lymphomas by qRT-PCR (Figure 3C).

Figure 3. HDACi Impair DGCR8-Dependent miRNA Biogenesis.

Figure 3

Open in a new tab

(A) Heat map of miRNAs downregulated in ΔNp63Δ/Δ;p53−/− thymic lymphomas compared with ΔNp63fl/fl;p53−/− thymic lymphomas. Low miRNA expression is indicated in green and high expression in red. (B) Network showing the results of miRNA-mRNA functional pair analysis between the downregulated miRNAs listed in A and the mRNAs upregulated in ΔNp63Δ/Δ;p53−/− vs. p53−/− thymic lymphomas. A diamond indicates the miRNA and a circle marks the regulated mRNA. The lines represent the regulation from the miRNA to one of its putative target mRNAs. The legend depicts the colors used to indicate the direction and magnitude of expression of miRNAs and mRNAs. (C) qRT-PCR in ΔNp63Δ/Δ;p53−/− and ΔNp63fl/fl;p53−/− thymic lymphomas. Data are mean ± SD, n = 3, * versus ΔNp63fl/fl;p53−/− tumors, p < 0.005, two-tailed t-test. (D–E) qRT-PCR for the DGCR8-dependent (D) and -independent (E) miRNAs in ΔNp63fl/fl;p53−/− thymic lymphoma cells treated for 48 hr either with 30 nM of the specified HDACi or with DMSO as a control. Data are mean ± SD, n = 3, * versus DMSO, p < 0.005, two-tailed t-test. (F–G) qRT-PCR for the DGCR8-dependent (F) and -independent (G) miRNAs in the indicated cell lines treated for 48 hr either with JNJ-26481585 or with DMSO as a control. Data are mean ± SD, n = 3, * versus DMSO, p < 0.005, two-tailed t-test. (H) qRT-PCR for the primary miRNA (pri-miRNA) transcripts of the 2 DGCR8-dependent miRNAs in cell lines indicated on the x-axis and treated with JNJ-26481585 or with DMSO. (I) Ratio between the mature and the pri-miRNA levels of the 2 DGCR8-dependent miRNAs in the specified cell lines treated with JNJ-26481585 or DMSO. Data are mean ± SD, n = 3, * versus DMSO, p < 0.005, two-tailed t-test. See also Figure S3.

Since AR-42, JNJ-26481585, and panobinostat were able to decrease ΔNp63 and DGCR8 protein levels in a broad number of cell lines, including p53−/− thymic lymphoma cells, we asked whether these HDACi could also affect the levels of let-7d and miR-128. A significant downregulation of let-7d and miR-128 was found in all HDACi treatment conditions (Figure 3D). To verify whether the effect of HDACi on these miRNAs was dependent on DGCR8, we tested the expression of 6 miRNAs whose maturation was previously shown to be DGCR8-independent (miR-122, miR-320, miR-484, miR-668, miR-720, and miR-877) (Babiarz et al., 2011; Babiarz et al., 2008; Davis et al., 2012; Yi et al., 2009). In HDACi treated thymic lymphoma cells, the expression of these DGCR8-independent microRNAs was not affected (Figure 3E), indicating an important difference between DGCR8-dependent and -independent miRNAs subsequent to treatment with HDACi.

Next, we tested if the difference in expression of DGCR8-dependent and -independent miRNAs was conserved in human cancer cell lines. In all the human cancer cell lines tested, treatment with HDACi decreased the levels of the 2 DGCR8-dependent miRNAs (Figures 3F, S3C and S3E), while the levels of the 6 DGCR8-independent ones were unchanged (Figures 3G, S3D and S3F). A notable exception was represented by cancer cells with low Fbw7 levels, like HH-NR cells (cTCL cell lines previously reported to be non-responsive to HDACi) (Zhang et al., 2005), where ΔNp63 and DGCR8 levels were not affected by HDACi (see Figure S2G) and consequently the levels of the 2 DGCR8-dependent miRNAs were not perturbed (Figure S3C).

Intriguingly, in two normal cell lines, NHEKs (primary human epidermal keratinocytes) and HMECs (primary human mammary epithelial cells), the basal expression of the 2 DGCR8-dependent miRNAs were lower compared to their expression in cancer cells, and more importantly remained unchanged subsequent to treatment with the HDACi (Figure 3F and S3E). To determine whether the miRNA biogenesis pathway in normal cells is responsive to HDACi, we analyzed the levels of 3 other miRNAs (miR-31, miR-220a and miR-200c), that we previously found to be regulated by ΔNp63 and DGCR8 in NHEKs (Chakravarti et al., 2014). We found a significant downregulation in the levels of miR-31, miR-220a and miR-200c by the HDACi in NHEKs (Figure S3G), indicating a different response of let-7d and miR-128 to HDACi in normal cells compared to cancer cells. This difference may be due to the very low basal expression levels of let-7d and miR-128 in normal cells or to the involvement of other factors, such as specific RNA-binding proteins.

Given that HDACi may affect gene expression (West and Johnstone, 2014), we asked whether these compounds also altered the expression of the 2 corresponding primary transcripts of let-7d and miR-128. We found that the expression of the primary transcripts was not affected by the HDACi in any of the tested cell lines and, as seen in this case of the mature miRNAs, their expression levels were lower in NHEKs than in the cancer cell lines (Figure 3H). By measuring the ratio between the mature miRNAs and their pri-miRNAs, we observed a reduction upon treatment with JNJ-26481585 in both Colo16 and SRB12 cells, while no change was present in the NHEKs (Figure 3I). Taken together, these data indicate that HDACi exploit Fbw7 to target ΔNp63, thus diminishing the expression of miRNAs whose maturation is DGCR8-dependent.

Inhibition of the ΔNp63/DGCR8 Axis Reduces Cancer Cell Viability

Because deletion of ΔNp63 in p53−/− thymic lymphomas caused tumor regression by reducing cell proliferation and inducing cell death (Venkatanarayan et al., 2015), we asked if reduction of ΔNp63 through the HDACi achieves the same biological effects in these tumors. To address this hypothesis, p53−/− thymic lymphoma cells were treated with the HDACi (AR-42, JNJ-26481585, or panobinostat) and cell proliferation was assessed by EdU incorporation. We found a decreased number of EdU positive cells subsequent to all treatment conditions compared to the vehicle (DMSO) alone indicating a lower proliferation rate in treated cells (Figure S4A). This lower cell proliferation rate was combined with an increased percentage of apoptotic cells, as quantified by Annexin V staining (Figure S4B).

Based on these results, we asked whether these HDACi could also impair cell viability of human cancer cell lines. In all human cancer cell lines assayed, we found a 3 to 4-fold decrease in the number of EdU positive cells (Figures 4A, 4B, S4C and S4D), and a more than 10-fold increase in the number of Annexin V positive cells (Figures 4C, 4D, S4E and S4F) compared to the DMSO treated conditions. The amplitude of the biological effects was dependent on the Fbw7-mediated downregulation of ΔNp63 and DGCR8 levels. Indeed, cells with lower levels of Fbw7, that have no perturbation of ΔNp63 and DGCR8 levels (see Figure S2G), did not show any change in either cell proliferation or cell death after treatment with HDACi (Figures S4G and S4H). In addition, no difference in cell proliferation or apoptosis was observed in the normal cells (NHEKs) treated with HDACi or placebo, suggesting that inhibition of ΔNp63 and DGCR8 can be specifically detrimental to cancer cells (Figures 4A–D).

Figure 4. Inhibition of the ΔNp63/DGCR8 Axis Reduces Cancer Cell Viability.

Figure 4

Open in a new tab

(A–B) Immunofluorescence (IF) and quantification for EdU (green) incorporation in the indicated cell lines treated for 48 hr either with 30 nM of the specified HDACi or with DMSO. Scale bars represent 20 μm. Data are mean ± SD, n = 3, * versus DMSO, p < 0.005, two-tailed t-test. (C–D) IF and quantification for Annexin V (green) incorporation in cells treated as in A. Hoechst 33342 (blue) was used as a counterstain. Scale bars represent 20 μm. Data are mean ± SD, n = 3, * versus DMSO, p < 0.005, two-tailed t-test. (E) In vivo tumor growth curves of Colo16 cells infected with pLJM1-EGFP-DGCR8 or empty vector and s.c. injected in both flanks of athymic nu/nu mice. Mice with palpable tumors were treated twice a week with i.p. injections of either JNJ-26481585 or vehicle. Data are mean ± SD of the tumor volumes, n = 6, * versus vehicle p < 0.01, two-tailed t-test. (F) Images of the Colo16-derived tumors collected at the end-point of the experiment described in E. See also Figure S4.

In order to assess the effects of HDACi in vivo, we generated a xenograft mouse model using Colo16 cells injected subcutaneously (s.c.) in both right and left flanks of athymic nude (nu/nu) mice. When tumors were palpable, the mice were treated twice a week with intraperitoneal (i.p.) injections of JNJ-26481585 (40 mg/kg). Tumor volume measurements showed a significant decrease of the tumor growth in the drug treated mice compared to the vehicle alone (Figure 4E), indeed, after two weeks of treatment, all treated tumors were significantly smaller than the vehicle treated ones (Figure 4F). We found the in vivo effects of JNJ-26481585 towards the component of the ΔNp63/DGCR8/miRNAs pathway to be reminiscent of the in vitro ones. Indeed, upon the administration of JNJ-26481585, ΔNp63 and DGCR8 protein expression were diminished (Figure S4I), the levels of DGCR8-dependent miRNAs (let-7d and miR-128) were decreased, while those of the DGCR8-independent miRNAs were not perturbed (Figures S4J and S4K). Overall, these findings indicate that HDACi inhibit cancer cell growth both in vitro and in vivo, and this effect is associated with an inhibition of the ΔNp63/DGCR8-dependent miRNA biogenesis pathway.

The Effects of HDACi on Cancer Cell Viability Are Mediated by DGCR8 through let-7d and miR-128

To verify whether the reduced cancer cell viability due to these compounds was dependent on the decrease in ΔNp63 and DGCR8 levels as well as in the DGCR8-dependent miRNA biogenesis, a Myc-tagged version of ΔNp63 was over-expressed in Colo16 and in H1299 cells. Compared to the empty vector infected cells, ΔNp63 protein levels were higher in the JNJ-26481585 treated conditions (Figures 5A and data not shown). These increased levels were paralleled by an upregulation of DGCR8 mRNA expression as well as of the 2 DGCR8-dependent miRNAs (let-7d and miR-128) (Figures 5B, 5C and data not shown). Additionally, ΔNp63 over-expression was able to rescue the cell viability defects by augmenting cell proliferation and reducing cell apoptosis in JNJ-26481585 treated cancer cell lines (Figures 5D–F and data not shown). Similar effects were achieved by the overexpression of an EGFP-tagged version of DGCR8 in Colo16 cells (Figures S5A–F). Notably, DGCR8 overexpression in the JNJ-26481585 treated conditions was able to rescue the expression levels of the 2 DGCR8-dependent miRNAs (let-7d and miR-128) (Figure S5B), while no significant effects were detected in the expression levels of the 6 DGCR8-independent miRNAs (miR-122, miR-320, miR-484, miR-668, miR-720, and miR-877) (Figure S5C). We further validated the dependence of HDACi efficacy on the downregulation of DGCR8. To do this, Colo16 cells overexpressing DGCR8 were injected subcutaneously into nude mice (xenograft model of cSCC) (Park et al., 2005) and then treated with JNJ-26481585. DGCR8 overexpression resulted in increased levels of the 2 DGCR8-dependent miRNAs (let-7d and miR-128) and continued cSCC growth regardless of the treatment with JNJ-26481585 (see Figures 4E, 4F and S4G–I). Therefore, these data indicate that HDACi affect the miRNA biogenesis pathway and, ultimately, cancer cell viability in a ΔNp63/DGCR8-dependent manner.

Figure 5. The Effects of HDACi on Cancer Cell Viability Are Mediated by DGCR8 through let-7d and miR-128.

Figure 5

Open in a new tab

(A) Representative WB analysis of Colo16 cells infected with pLPC-Myc-ΔNp63α or empty vector and treated for 48 hr either with 30 nM of JNJ-26481585 (+) or with DMSO (−). (B–C) qRT-PCR in the same conditions as in A. Data are mean ± SD, n = 3, * versus DMSO, # versus JNJ-26481585 empty, p < 0.005, two-tailed t-test. (D–F) IF and quantification for EdU and Annexin V positive cells (green) in the same condition as in A. Hoechst 33342 (blue) was used as a counterstain. Scale bars represent 20 μm. Data are mean ± SD, n = 3, * versus DMSO, # versus JNJ-26481585 empty, p < 0.005, two-tailed t-test. (G) Representative WB analysis of Colo16 cells treated for 48 hr either with 30 nM of JNJ-26481585 or with DMSO or transfected with the indicated siRNAs or miRNA inhibitors. (H–J) IF and quantification for EdU (green) and Annexin V (green) positive Colo16 cells transfected with the indicated siRNAs or miRNA inhibitors. Hoechst 33342 (blue) was used as a counterstain in Annexin assays. Scale bars represent 20 μm. Data are mean ± SD, n = 3, * versus DMSO, # versus siC, § versus non targeting (NT) miRNA inhibitor, p < 0.005, two-tailed t-test. (K–L) Bar graphs showing the percentage of EdU (K) and Annexin V (L) positive cells transfected with the indicated miRNA inhibitors. Data are mean ± SD, n = 3, * versus non targeting (NT) miRNA inhibitor, p < 0.005, two-tailed t-test. (M–O) IF and quantification for EdU (green) and Annexin V (green) positive Colo16 cells treated for 48 hr with 30 nM of JNJ-26481585 and transfected with the indicated miRNA mimics. DMSO treatment was used as a control. Hoechst 33342 (blue) was used as a counterstain. Scale bars represent 20 μm. Data are mean ± SD, n = 3, * versus DMSO, # versus negative control mimic, p < 0.005, two-tailed t-test. See also Figure S5.

Given that loss of ΔNp63 may impair cell proliferation and cause cell death (Venkatanarayan et al., 2015), we asked whether depletion of DGCR8 or inhibition of let-7d or miR-128 could produce the same biological effects. We therefore compared cancer cells treated with: JNJ-26481585, siΔNp63, siDGCR8, or inhibitors of let-7d or miR-128. Downregulation of ΔNp63 or DGCR8 using siRNAs resulted in similar effects on cell proliferation and apoptosis as treatment with JNJ-26481585 (Figures 5G). To ask whether JNJ-26481585 functions through let-7d or miR-128, we treated cancer cell lines with inhibitors of let-7d and miR-128 and found that inhibition of both let-7d and miR-128 were needed to achieve the same effect as treatment with JNJ-26481585 (Figures 5G–J), suggesting that this HDACi functions through let-7d and miR-128. Interestingly, the combination of let-7d and miR-128 inhibitors reduced cell proliferation and increased cell death in HDACi-resistant cells to a similar extent to what occurs in HDACi-sensitive cells (Figures 5K and 5L), thus suggesting the therapeutic potential of these miRNA inhibitors to treat HDACi resistant tumors.

Since the combined inhibition of let-7d and miR-128 were critical and mimicked the analyzed biological effects of HDACi, we asked whether their over-expression in the drug treated cells would rescue the cell viability defects. To address this question, Colo16 and SRB12 cells were transfected with either each miRNA mimic or with a negative control. In line with the results from the above-described experiment using the miRNA inhibitors, over-expression of either let-7d or miR-128 in JNJ-26481585 cancer cell lines was able to partially rescue the cell viability defects and, notably, their combined over-expression completely rescued cell viability in both cell lines (Figures 5M–O and data not shown), suggesting again that let-7d and miR-128 are crucial for the function of JNJ-26481585. Taken together, these findings demonstrate that the biological effects of HDACi on these cancer cells are mediated by the ΔNp63/DGCR8/miRNAs axis, and more specifically through let-7d and miR-128.

let-7d and miR-128 Regulate Cancer Cell Viability through p21

To investigate how these let-7d and miR-128 were affecting cancer cell viability in response to HDACi, we evaluated the involvement of p21, since 1) p21 is a target of both miRNAs (Li et al., 2013; Liu et al., 2015), 2) its expression was reported to be induced by HDACi (Ocker and Schneider-Stock, 2007), and 3) it is an important factor through which HDACi achieve their biological effects (Bose et al., 2014). A time course performed in Colo16 cells treated with JNJ-26481585 indicated a rapid and progressive increase in P21 (also known as CDKN1A) mRNA levels (Figure 6A) with a concomitant decrease in ΔNp63 protein levels (Figure 1E). To assess whether p21 was required for the reduced proliferation and increased cell death in HDACi treated cells, Colo16 cells were transfected with an siRNA for P21 (sip21) or a scrambled control (siC), prior to the treatment with JNJ-26481585 (Figure 6B). Depletion of P21 was able to partially recover the cell viability defects in JNJ-26481585 treated cells (Figures 6C–E). To determine the relevance of p21 as a mediator of let-7d and miR-128 biological activities, P21 depleted Colo16 cells were transfected with inhibitors of these 2 miRNAs or with a non-targeting inhibitor (NT) as a control. In line with previous evidence (Li et al., 2013; Liu et al., 2015), the inhibition of let-7d and miR-128 in siC transfected cells increased p21 levels compared to the siC NT transfected cells (Figure 6F). Interestingly, the reduced cell proliferation and increased cell death caused by the inhibition of let-7d and miR-128 was partially rescued by depletion of P21 (Figures 6G–I). Taken together, these experiments indicate that p21 is a crucial mediator of the biological activities of HDACi on cancer cells subsequent to the inhibition of the ΔNp63/DGCR8/miRNAs axis.

Figure 6. let-7d and miR-128 Regulate Cancer Cell Viability through p21.

Figure 6

Open in a new tab

(A) qRT-PCR for CDKN1A (also known as P21) in Colo16 cells treated with 30 nM of JNJ-26481585 for the indicated hours. Data are mean ± SD, n = 3, * versus DMSO, p < 0.005, two-tailed t-test. (B) Representative WB analysis of Colo16 cells transfected with sip21 or siC (control) and treated for 48 h either with 30 nM of JNJ-26481585 or with DMSO using the indicated antibodies. HSP90 was used as a loading control. (C) Representative images of IF for EdU (green) and Annexin V (green) positive cells in the same conditions described in panel B. Hoechst 33342 (blue) was used as a counterstain. Scale bars represent 20 μm. (D-E) Bar graphs showing the percentage of EdU (D) and Annexin V (E) positive Colo16 cells represented in panel C. Data are mean ± SD, n = 3, * versus siC DMSO, # versus siC JNJ-26481585, p < 0.005, two-tailed t-test. (F) Representative WB analysis of Colo16 cells transfected with the indicated siRNAs and miRNA inhibitors using the indicated antibodies. HSP90 was used as a loading control. (G) Representative images of IF for EdU (green) and Annexin V (green) positive cells in the same conditions described in panel F. Hoechst 33342 (blue) was used as a counterstain. Scale bars represent 20 μm. (H–I) Bar graphs showing the percentage of EdU (H) and Annexin V (I) positive Colo16 cells represented in panel G. Data are mean ± SD, n = 3, * versus siC NT, # versus siC let7-d & miR-128 inhibitors, p < 0.005, two-tailed t-test.

Inhibition of the ΔNp63/DGCR8/miRNAs Axis Induces Tumor Regression in Genetically Engineered and Xenograft Mouse Models

Because we found that direct inhibition of DGCR8 as well as that of let-7d and miR-128 can reduce cancer cell viability in vitro, we wanted to determine their efficacy in vivo by utilizing the p53−/− thymic lymphoma model, where deletion of ΔNp63 was proven to cause tumor regression (Venkatanarayan et al., 2015). To address this, DOPC-incorporated siRNAs and miRNA inhibitors were used for intratumoral injection in 10 week-old p53−/− mice with thymic lymphomas with volumes of 8.0 to 12.0 mm3 at the time of injection and monitored weekly by magnetic resonance imaging (MRI). The delivery of the siRNAs was confirmed by IVIS Lumina imaging (Figure S6A), and their efficacy against the respective targets lasted up to 1 week, as assessed by Western blot analysis (Figure S6B). As seen in vitro, depletion of ΔNp63 and DGCR8 was achieved within 48 hr after the injection (Figure S6C), as well as inhibition of let-7d and miR-128, as assessed by the increased mRNA levels of some of their downstream targets (Figure S6D). Accordingly, the injected tumors showed decreased cell proliferation and increased cell death rates, as assessed by immunohistochemistry (IHC) for PCNA and cleaved caspase 3, respectively (Figures S6E and S6F). In line with these results, weekly injections of siRNAs for ΔNp63 or DGCR8 or miRNA inhibitors for let-7d or miR-128 caused a significant regression of the tumors within 3 weeks of treatment compared to the respective controls (Figures 7A–D) indicating the efficacy of this microRNA therapy as an anti-tumor agent. Next, we assessed whether let-7d and miR-128 inhibitors showed anti-tumor effects in a cSCC xenograft mouse model. cSCC generated in athymic nude (nu/nu) mice with Colo-16 cells were treated with weekly i.p. injections of either let-7d or miR-128 inhibitor or both in combination. Compared to the control, each miRNA inhibitor was able to significantly reduce tumor growth and, more importantly, when used in combination they caused rapid and complete tumor regression (Figures 7E and 7F). Overall, these results highlight the therapeutic potential of inhibiting the ΔNp63/DGCR8/miRNA axis and specifically indicate the broad function of HDACi in several tumor types through inhibition of let-7d and miR-128, which may be safe therapeutic targets in p53 deficient and mutant tumors.

Figure 7. Inhibition of the ΔNp63/DGCR8/miRNAs Axis Induces Tumor Regression in vivo.

Figure 7

Open in a new tab

(A) Representative magnetic resonance imaging (MRI) of p53−/− thymic lymphomas treated with weekly intratumoral injections of 5 μg of the indicated 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC) coated siRNAs. Tumor volume (mm3) shown within each panel. Tumors indicated by the dashed yellow line. (B) Quantification of the indicated thymic lymphomas shown in A. Data are mean ± SD of the relative tumor volumes, n ≥ 5, * versus siC, p < 0.01, two-tailed t-test. (C) Representative MRI of p53−/− thymic lymphomas treated with weekly intratumoral injections of 5 μg of the indicated DOPC-coated miRNA inhibitors. Tumor volume (mm3) shown within each panel. Tumors indicated by the dashed yellow line. (D) Quantification of the indicated thymic lymphomas shown in C. Data are mean ± SD of the tumor volumes, n ≥ 5, * versus NT inhibitor, p < 0.01, two-tailed t-test. (E) In vivo tumor growth curves of xenograft mouse models composed of Colo16 cells s.c. injected in both flanks of athymic nu/nu mice. Mice with palpable tumors were treated twice a week with i.p. injections of 5 μg of the indicated miRNA inhibitors. Data are mean ± SD of the tumor volumes, n = 6, * versus NT, p < 0.01, two-tailed t-test. (F) Images of the Colo16-derived tumors collected at the end-point of the experiment described in E. See also Figure S6.

DISCUSSION

The TP63 gene belongs to the p53 family of transcription factors and, due to the presence of two alternative promoters, drives the expression of two different groups of isoforms: TAp63, which contains an N-terminal acidic domain resembling the p53 transcriptional activation domain, and ΔNp63, which is devoid of that domain but it is still transcriptionally competent (Su et al., 2013). While TAp63 acts as a potent tumor and metastasis suppressor (Adorno et al., 2009; Girardini et al., 2011; Muller et al., 2009; Napoli and Flores, 2013; Su et al., 2010), increasing in vitro and in vivo evidence has shown ΔNp63 to be overexpressed in several human tumors, including basal-like bladder carcinomas (Karni-Schmidt et al., 2011), ovarian epithelial cancers (Marchini et al., 2008), as well as squamous cell carcinomas of head and neck, lung, skin and cervix (Orzol et al., 2014) consistent with its role as an oncogene. ΔNp63 elevated levels may be associated with its ability to directly bind to both p53 and TAp63 and inhibit their functions (Flores et al., 2002; Yang et al., 1998) and with its oncogenic activity. We have recently demonstrated in vivo oncogenic roles for ΔNp63 through deletion of ΔNp63 in p53−/− thymic lymphomas resulting in rapid tumor regression through activation of the tumor suppressive activities of TAp63 (Venkatanarayan et al., 2015). Here, we identified HDACi as compounds that downregulate ΔNp63 through Fbw7 to induce a tumor suppressive program in a broad range of cancer types. Importantly, we found Fbw7 as a predictive marker for HDACi response in patient derived cTCL biopsies and cell lines. We also show that inhibition of ΔNp63 through HDACi administration caused a reduction in the DGCR8-dependent maturation of let-7d and miR-128. HDACi-driven induction of tumor regression through activation of cell cycle arrest and apoptosis was dependent on inhibition of both let-7d and miR-128 and largely but not entirely relied on the consequent induction of P21. We further showed that combined inhibition of both let-7d and miR-128 delivered in xenograft and genetically engineered mouse models (GEMMs) with alterations in p53 resulted in rapid tumor regression demonstrating that HDACi and specific suppression of let-7d and miR-128 are potentially effective therapies for p53 deficient and mutant tumors.

The effect of ΔNp63 deletion in p53-altered tumors clearly indicates the exploitability of targeting ΔNp63 to therapeutically treat p53 mutant and deficient tumors. Therefore, in contrast to the several mechanisms used to reactivate mutant p53 proteins in human tumors that have unfortunately failed so far in clinical trials (Muller and Vousden, 2014), the inhibition of ΔNp63, which is rarely mutated in human cancers (Inoue and Fry, 2014), could result in a more efficacious strategy. Given the difficulties in directly targeting a transcription factor compared to “druggable” factors such as kinases (Yeh et al., 2013), we decided to screen more than 850 pharmacologically active compounds in an effort to repurpose existing drugs for their ability to mimic ΔNp63 deletion and downregulation of its downstream transcriptional target, DGCR8. We identified 5 compounds that effectively reduced expression of both ΔNp63 and DGCR8, and they are all HDACi. A growing body of evidence mainly relying on cell-based studies has led to interest of these compounds in cancer therapy. Indeed, HDACi may produce anti-tumor effects by promoting different biological processes, such as G1/S phase cell cycle arrest, intrinsic and extrinsic apoptosis, senescence, and differentiation (West and Johnstone, 2014). However, in addition to exerting these activities on tumor cells, HDACi have multiple adverse effects including leukocyte toxicity (Sweet et al., 2012), suppression of TREG cell functions (Shen et al., 2012), and inhibition of inflammatory cytokine production (Villagra et al., 2010). Because of these immunosuppressive consequences, the clinical use of the HDACi has been largely restricted to cutaneous and peripheral T cell lymphomas (West and Johnstone, 2014). The efficacy of HDACi has partially been linked to their ability to affect the activities of the p53 family members (Ramsey et al., 2011; Yamaguchi et al., 2009; Zhang et al., 2013); however, our report shows that they can also act through the ΔNp63/DGCR8 axis and through two specific microRNAs. We found that this effect is mediated by the E3 ubiquitin ligase Fbw7, which was previously demonstrated to degrade ΔNp63 in an HDM2-dependent manner (Galli et al., 2010). Fbw7 is a tumor suppressor able to cause the proteasome-driven degradation of several oncogenes (including cyclin E, c-myc and Notch), and because of its role it can be found mutated or inactivated in human cancers (Davis et al., 2014). Interestingly, Fbw7 was already described to collaborate with HDACi to promote the ubiquitination of other proteins, such as topoisomerase IIα (Chen et al., 2011), thus suggesting that this cooperation may be common in the regulation of different Fbw7 substrates. Importantly, we identified a crucial role of Fbw7 in determining the sensitivity to HDACi in patient derived cutaneous T cell lymphoma (cTCL) cell lines, cSCC, and HNSCC human cell lines. These data are further corroborated by our analysis of 22 biopsies of mycosis fungoides, the most common form of cTCL, where we found that high Fbw7 expression is a predictive marker of response to HDACi.

As a result of the cooperation between Fbw7 and HDACi, we found a reduction in the ΔNp63/DGCR8-dependent miRNA maturation process. This is in line with the observation that HDACi decrease miRNA expression levels in hematopoietic cell lines, although in that system the effect was due to diminished DGCR8 activity rather than an effect on its levels (Wada et al., 2012). Among the ΔNp63/DGCR8-dependent miRNAs affected by these HDACi, we identified two miRNAs that were critical for the activity of these compounds, let-7d and miR-128. Intriguingly, the expression levels of their primary transcripts in cancer cells were higher compared to those detected in normal cells (i.e. NHEKs), and the mature miRNA levels were significantly decreased in the treated conditions. On the contrary, the lack of effect of HDACi on let-7d and miR-128 levels in NHEKs could be due to the low expression of these miRNAs or to the contribution of other components, including RNA-binding proteins, that regulate miRNA in a timed and spatially controlled manner (van Kouwenhove et al., 2011). These molecular differences were mirrored by the observed biological responses (i.e. increased apoptosis and reduced cell proliferation) in cancer cell lines in contrast to primary normal cells. Importantly, we found that JNJ-26481585 functions through the inhibition of let-7d and miR-128, since forced re-expression of these miRNAs in cancer cells treated with JNJ-26481585 led to loss of the biological activity of this drug suggesting that this HDACi functions primarily through inhibition of these two miRNAs.

Both let-7d and miR-128 were reported to support tumor formation, even though this outcome is tissue-specific (Nuovo et al., 2012; Yu et al., 2015). Let-7d is very peculiar compared with the rest its family, which is composed by tumor suppressive miRNAs and proven to regulate different target genes (Kolenda et al., 2014): as a consequence, for example, let-7d upregulation is associated with breast cancer progression (Volinia et al., 2012). Mir-128 targets the tumor suppressor PTEN, thereby sustaining the proliferation of osteosarcoma cells (Shen et al., 2014). These properties are in agreement with our findings that inhibiting the activities of these two miRNAs through in vivo delivery of miRNA inhibitors effectively induces tumor regression in both GEMM models of p53−/− thymic lymphomas and xenograft mouse models derived from Colo16 cSCC cells bearing mutant p53.

The efficacy of RNA-based therapeutics in mouse models has led to the development of several delivery strategies that are currently under clinical evaluation (Yin et al., 2014). The system we chose to deliver let-7d and miR-128 inhibitors (i.e. neutral lipid based nanoliposomes) was demonstrated to increase the stability of the coated RNA molecules and their uptake by tumors compared to other delivery methods (Halder et al., 2006). More importantly, systemic administration of DOPC-based particles in mice is not toxic to normal tissues or the immune system, thus paving the way for the validation of these particles in clinical trials (Ozcan et al., 2015).

Our data support the treatment of cutaneous T cell lymphomas and squamous cell carcinomas with HDACi or the inhibition of specific miRNAs, as indicated by our data from patient biopsies and xenograft mouse models. However, our findings may also be useful for the treatment of other cancers based on the following reasons. First, they suggest that the hydroxamate class of pan-HDACi may have a large clinical spectrum of activity that extends to p53-mutant or deficient cancers that are dependent upon ΔNp63/DGCR8 function. Second, we identified a specific mechanism by which these anti-tumor effects are achieved, namely the inhibition of DGCR8-dependent maturation of let-7d and miR-128, in lymphomas and carcinomas. Third, we identified Fbw7 as a predictive marker of HDACi response, thus emphasizing the importance of considering this factor to select a larger yet more specific group of cancer subtypes that may benefit from HDACi treatment. Finally, we show that the combined inhibition of let-7d and miR-128 in the place of HDACi may represent a more effective and potentially less toxic alternative approach to effectively target p53 deficient and mutant tumors addicted to the ΔNp63/DGCR8/miRNA axis. Our data detail strong preclinical rationale for further clinical investigation of HDACi currently in clinical trials and inhibition of let-7d and miR-128 for cancer therapy.

EXPERIMENTAL PROCEDURES

Additional methods can be found in the Supplemental Experimental Procedures.

Drug Library Screening

MCF-10A cells were plated at a density of 1 × 104 cells in 4 replicates in 96-well dishes, and treated with 1 μM of each of the 855 bioactive chemical compounds (Selleck Chemicals). After 48 hr, immunofluorescence was performed as described previously (Chakravarti et al., 2014), using antibodies for ΔNp63 (1:250) (BioLegend) and DGCR8 (1:250) (ab36865, Abcam). The respective signals were quantified through a high-throughput immunofluorescence plate reader and accompanying software (Celigo), and compared to the DMSO treated condition.

In vivo Tumor Monitoring and Drug Treatment of Xenograft Mouse Models

4 x 105 Colo16 cells in 100 μl of PBS were subcutaneously injected in both flanks of six week-old female athymic nu/nu mice (xenograft model of cSCC) (Park et al., 2005). Once tumors became palpable, mice were treated as described. To test the in vivo efficacy of JNJ-26481585, mice were treated twice a week with intraperitoneal injections of JNJ-26481585 (40 mg/kg). The drug was prepared in distilled water solution containing 10% hydroxypropyl-beta cyclodextrin, 0.8% HCl (0.1 N), 0.9 % NaOH (0.1 N), and 3.4% mannitol. This vehicle solution was injected as a control. For the in vivo analysis of the miRNA inhibitors, mice were treated weekly with i.p. injections of 5 μg of 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC) incorporated miRNA inhibitors (let-7d or miR-128) prepared as described previously (Mangala et al., 2009). Tumor growth was monitored by caliper measurements. Tumor volume was calculated using the formula: tumor volume (mm3) = D × d2/2, where D and d are the longest and the shortest diameters of the tumor, respectively. The cohorts of mice were sacrificed when the biggest tumors of the respective controls were 1,000 mm3. All procedures were approved by the IACUC at the University of Texas MD Anderson Cancer Center.

Magnetic Resonance Imaging

MRI imaging was performed on mice with suspected thymic lymphomas at 10 weeks of age when the tumors were established and the volumes range from 8.0 mm3 to 12.0 mm3. To reduce the variation between different groups of mice, cohorts of 5 to 7 mice with similar tumor volumes were established and tumor regression was monitored by MRI. All mice were scanned once a week for a period of 4 weeks on a 7-T, 30-cm bore BioSpec MRI system (Bruker Biospin Corp., Billerica, MA).

Intra-Thymic Injection

Mice were anesthetized using isoflurane and 2% oxygen and placed on a custom surgical bed. An incision was performed to expose the sternum. Using a 28.5G and insulin syringe, 5 μg of DOPC-coated siRNAs or miRNA inhibitors in 30 μl of PBS were surgically administered by intra-thymic injection through the 2nd and 3rd sternum. The incision was sealed using wound clips and mice were allowed to recover. Mice were monitored for tumor regression by MRI.

Immunohistochemistry of human biopsies of mycosis fungoides

Procedures for the acquisition following informed patient consent and processing of human biopsies were approved by the IRB at the University of Texas MD Anderson Cancer Center and at the University of Pennsylvania. Formalin-fixed, paraffin embedded sections were de-waxed in xylene and re-hydrated using decreasing concentrations of ethanol. Antigens were unmasked in Tris buffer unmasking solution (10 mM Tris base, 1 mM EDTA, 0.05 % Tween-20, pH 9.0) followed by incubation with blocking solution, and an 18 hour incubation at 4 °C with Fbw7 antibody (1:200) (Bethyl). Visualization was performed using the DAB peroxidase substrate kit (SK4100, Vector Laboratories) and counter-stained with hematoxylin (H-3401, Vector Laboratories). The slides were mounted using Histomount (008030, Life Technologies). Images were acquired using a Zeiss Axio microscope and analyzed with ProgRes Capture Pro 4.5 software. Fbw7 staining was scored and quantified in a blinded fashion based on its intensity (from 1 = low to 3 = high) and the proportion of positive tissue (from 1 ≤ 25 % to 3 > 75 %) by a board certified dermatopathologist.

Supplementary Material

supplement

Significance.

HDACi based therapies are currently limited due to a paucity of predictive markers of therapeutic efficacy and their toxicity to cancer patients. Here, we show that HDACi function to inhibit the ΔNp63/DGCR8 axis to effectively treat tumors dependent upon oncogenic ΔNp63, especially therapeutically challenging ones devoid of a functional p53. We identified Fbw7 as a predictive marker for HDACi response. Importantly, our work suggests an expanded yet specific group of cancers that may be effectively treated by HDACi, and advances let-7d and miR-128 as potential specific targets for therapeutic intervention.

Highlights.

  1. HDAC inhibitors (HDACi) reduce ΔNp63 stability in an Fbw7-dependent manner

  2. HDACi cause tumor regression by impairing ΔNp63/DGCR8-dependent miRNA maturation

  3. Fbw7 levels are predictive of tumor response to HDACi

  4. Resistance to HDACi can be bypassed by direct inhibition of let-7d and miR-128

Acknowledgments

We thank F. Muller for kindly sharing the drug library, and J.T. Pietz for art production. This work was supported by grants to E.R.F. from NCI (R01CA160394 and R35CA197452), CPRIT (RP140271 and RP150094), NCI-Cancer Center Core Grant (CA-16672) (University of Texas M.D. Anderson Cancer Center), a development award from the Lymphoma SPORE (P50CA136411), the Hildegardo E. and Olga M. Flores Foundation, and the Mel Klein Foundation. This work was also funded in part by the RGK Foundation. E.R.F. is a scholar of the Leukemia and Lymphoma Society, the Rita Allen Foundation and the V Foundation for Cancer Research. M.N. is a CPRIT-TRIUMPH Scholar and was supported by a Research Training Award from the Cancer Prevention and Research Institute of Texas (RP140106). H.A.A. is an Odyssey Fellow at MD Anderson Cancer Center.

Footnotes

ACCESSION NUMBER

The accession number in the Gene Expression Omnibus public database for the miRNA sequencing experiment is GSE77889.

AUTHORS CONTRIBUTIONS

M.N. and E.R.F. conceived the study, designed experiments and analyzed data. M.N., A.V., P.R., B.A.M., W.N., L.S.M., A.K.S., C.R.-A., G.L.-B., H.V., M.D., M.B.T., J.L.C., A.H.R., H.A.A. and K.Y.T. designed and performed experiments. C.C. and P.H.G. performed bioinformatic analyses. E.R.F. and M.N. wrote the paper. All authors discussed the paper and commented on the manuscript.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Adorno M, Cordenonsi M, Montagner M, Dupont S, Wong C, Hann B, Solari A, Bobisse S, Rondina MB, Guzzardo V, et al. A Mutant-p53/Smad complex opposes p63 to empower TGFbeta-induced metastasis. Cell. 2009;137:87–98. doi: 10.1016/j.cell.2009.01.039. [DOI] [PubMed] [Google Scholar]
  2. Babiarz JE, Hsu R, Melton C, Thomas M, Ullian EM, Blelloch R. A role for noncanonical microRNAs in the mammalian brain revealed by phenotypic differences in Dgcr8 versus Dicer1 knockouts and small RNA sequencing. Rna. 2011;17:1489–1501. doi: 10.1261/rna.2442211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Babiarz JE, Ruby JG, Wang Y, Bartel DP, Blelloch R. Mouse ES cells express endogenous shRNAs, siRNAs, and other Microprocessor-independent, Dicer-dependent small RNAs. Genes & development. 2008;22:2773–2785. doi: 10.1101/gad.1705308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bose P, Dai Y, Grant S. Histone deacetylase inhibitor (HDACI) mechanisms of action: emerging insights. Pharmacology & therapeutics. 2014;143:323–336. doi: 10.1016/j.pharmthera.2014.04.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chakravarti D, Su X, Cho MS, Bui NH, Coarfa C, Venkatanarayan A, Benham AL, Flores Gonzalez RE, Alana J, Xiao W, et al. Induced multipotency in adult keratinocytes through down-regulation of DeltaNp63 or DGCR8. Proceedings of the National Academy of Sciences of the United States of America. 2014;111:E572–581. doi: 10.1073/pnas.1319743111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chen MC, Chen CH, Chuang HC, Kulp SK, Teng CM, Chen CS. Novel mechanism by which histone deacetylase inhibitors facilitate topoisomerase IIalpha degradation in hepatocellular carcinoma cells. Hepatology. 2011;53:148–159. doi: 10.1002/hep.23964. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Creighton CJ, Nagaraja AK, Hanash SM, Matzuk MM, Gunaratne PH. A bioinformatics tool for linking gene expression profiling results with public databases of microRNA target predictions. Rna. 2008;14:2290–2296. doi: 10.1261/rna.1188208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Davis MP, Abreu-Goodger C, van Dongen S, Lu D, Tate PH, Bartonicek N, Kutter C, Liu P, Skarnes WC, Enright AJ, Dunham I. Large-scale identification of microRNA targets in murine Dgcr8-deficient embryonic stem cell lines. PloS one. 2012;7:e41762. doi: 10.1371/journal.pone.0041762. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Davis RJ, Welcker M, Clurman BE. Tumor suppression by the Fbw7 ubiquitin ligase: mechanisms and opportunities. Cancer cell. 2014;26:455–464. doi: 10.1016/j.ccell.2014.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Flores ER, Tsai KY, Crowley D, Sengupta S, Yang A, McKeon F, Jacks T. p63 and p73 are required for p53-dependent apoptosis in response to DNA damage. Nature. 2002;416:560–564. doi: 10.1038/416560a. [DOI] [PubMed] [Google Scholar]
  11. Galli F, Rossi M, D’Alessandra Y, De Simone M, Lopardo T, Haupt Y, Alsheich-Bartok O, Anzi S, Shaulian E, Calabro V, et al. MDM2 and Fbw7 cooperate to induce p63 protein degradation following DNA damage and cell differentiation. Journal of cell science. 2010;123:2423–2433. doi: 10.1242/jcs.061010. [DOI] [PubMed] [Google Scholar]
  12. Girardini JE, Napoli M, Piazza S, Rustighi A, Marotta C, Radaelli E, Capaci V, Jordan L, Quinlan P, Thompson A, et al. A Pin1/mutant p53 axis promotes aggressiveness in breast cancer. Cancer cell. 2011;20:79–91. doi: 10.1016/j.ccr.2011.06.004. [DOI] [PubMed] [Google Scholar]
  13. Halder J, Kamat AA, Landen CN, Jr, Han LY, Lutgendorf SK, Lin YG, Merritt WM, Jennings NB, Chavez-Reyes A, Coleman RL, et al. Focal adhesion kinase targeting using in vivo short interfering RNA delivery in neutral liposomes for ovarian carcinoma therapy. Clinical cancer research : an official journal of the American Association for Cancer Research. 2006;12:4916–4924. doi: 10.1158/1078-0432.CCR-06-0021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Inoue K, Fry EA. Alterations of p63 and p73 in human cancers. Sub-cellular biochemistry. 2014;85:17–40. doi: 10.1007/978-94-017-9211-0_2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Jung YS, Qian Y, Yan W, Chen X. Pirh2 E3 ubiquitin ligase modulates keratinocyte differentiation through p63. The Journal of investigative dermatology. 2013;133:1178–1187. doi: 10.1038/jid.2012.466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Karni-Schmidt O, Castillo-Martin M, Shen TH, Gladoun N, Domingo-Domenech J, Sanchez-Carbayo M, Li Y, Lowe S, Prives C, Cordon-Cardo C. Distinct expression profiles of p63 variants during urothelial development and bladder cancer progression. The American journal of pathology. 2011;178:1350–1360. doi: 10.1016/j.ajpath.2010.11.061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kolenda T, Przybyla W, Teresiak A, Mackiewicz A, Lamperska KM. The mystery of let-7d - a small RNA with great power. Contemporary oncology. 2014;18:293–301. doi: 10.5114/wo.2014.44467. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Li M, Fu W, Wo L, Shu X, Liu F, Li C. miR-128 and its target genes in tumorigenesis and metastasis. Experimental cell research. 2013;319:3059–3064. doi: 10.1016/j.yexcr.2013.07.031. [DOI] [PubMed] [Google Scholar]
  19. Li Y, Zhou Z, Chen C. WW domain-containing E3 ubiquitin protein ligase 1 targets p63 transcription factor for ubiquitin-mediated proteasomal degradation and regulates apoptosis. Cell death and differentiation. 2008;15:1941–1951. doi: 10.1038/cdd.2008.134. [DOI] [PubMed] [Google Scholar]
  20. Liu B, Liu M, Wang J, Zhang X, Wang X, Wang P, Wang H, Li W, Wang Y. DICER-dependent biogenesis of let-7 miRNAs affects human cell response to DNA damage via targeting p21/p27. Nucleic acids research. 2015;43:1626–1636. doi: 10.1093/nar/gku1368. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
  21. Mangala LS, Han HD, Lopez-Berestein G, Sood AK. Liposomal siRNA for ovarian cancer. Methods in molecular biology. 2009;555:29–42. doi: 10.1007/978-1-60327-295-7_3. [DOI] [PubMed] [Google Scholar]
  22. Marchini S, Marabese M, Marrazzo E, Mariani P, Cattaneo D, Fossati R, Compagnoni A, Fruscio R, Lissoni AA, Broggini M. DeltaNp63 expression is associated with poor survival in ovarian cancer. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2008;19:501–507. doi: 10.1093/annonc/mdm519. [DOI] [PubMed] [Google Scholar]
  23. Muller PA, Caswell PT, Doyle B, Iwanicki MP, Tan EH, Karim S, Lukashchuk N, Gillespie DA, Ludwig RL, Gosselin P, et al. Mutant p53 drives invasion by promoting integrin recycling. Cell. 2009;139:1327–1341. doi: 10.1016/j.cell.2009.11.026. [DOI] [PubMed] [Google Scholar]
  24. Muller PA, Vousden KH. Mutant p53 in cancer: new functions and therapeutic opportunities. Cancer cell. 2014;25:304–317. doi: 10.1016/j.ccr.2014.01.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Napoli M, Flores ER. The family that eats together stays together: new p53 family transcriptional targets in autophagy. Genes & development. 2013;27:971–974. doi: 10.1101/gad.219147.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Nuovo GJ, Garofalo M, Valeri N, Roulstone V, Volinia S, Cohn DE, Phelps M, Harrington KJ, Vile R, Melcher A, et al. Reovirus-associated reduction of microRNA-let-7d is related to the increased apoptotic death of cancer cells in clinical samples. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2012;25:1333–1344. doi: 10.1038/modpathol.2012.95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Ocker M, Schneider-Stock R. Histone deacetylase inhibitors: signalling towards p21cip1/waf1. The international journal of biochemistry & cell biology. 2007;39:1367–1374. doi: 10.1016/j.biocel.2007.03.001. [DOI] [PubMed] [Google Scholar]
  28. Orzol P, Holcakova J, Nekulova M, Nenutil R, Vojtesek B, Coates PJ. The diverse oncogenic and tumour suppressor roles of p63 and p73 in cancer: a review by cancer site. Histology and histopathology. 2014 doi: 10.14670/HH-30.503. [DOI] [PubMed] [Google Scholar]
  29. Ozcan G, Ozpolat B, Coleman RL, Sood AK, Lopez-Berestein G. Preclinical and clinical development of siRNA-based therapeutics. Advanced drug delivery reviews. 2015 doi: 10.1016/j.addr.2015.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Park YW, Younes MN, Jasser SA, Yigitbasi OG, Zhou G, Bucana CD, Bekele BN, Myers JN. AEE788, a dual tyrosine kinase receptor inhibitor, induces endothelial cell apoptosis in human cutaneous squamous cell carcinoma xenografts in nude mice. Clinical cancer research : an official journal of the American Association for Cancer Research. 2005;11:1963–1973. doi: 10.1158/1078-0432.CCR-04-1665. [DOI] [PubMed] [Google Scholar]
  31. Ramsey MR, He L, Forster N, Ory B, Ellisen LW. Physical association of HDAC1 and HDAC2 with p63 mediates transcriptional repression and tumor maintenance in squamous cell carcinoma. Cancer research. 2011;71:4373–4379. doi: 10.1158/0008-5472.CAN-11-0046. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Restelli M, Lopardo T, Lo Iacono N, Garaffo G, Conte D, Rustighi A, Napoli M, Del Sal G, Perez-Morga D, Costanzo A, et al. DLX5, FGF8 and the Pin1 isomerase control DeltaNp63alpha protein stability during limb development: a regulatory loop at the basis of the SHFM and EEC congenital malformations. Human molecular genetics. 2014;23:3830–3842. doi: 10.1093/hmg/ddu096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Rossi M, Aqeilan RI, Neale M, Candi E, Salomoni P, Knight RA, Croce CM, Melino G. The E3 ubiquitin ligase Itch controls the protein stability of p63. Proceedings of the National Academy of Sciences of the United States of America. 2006a;103:12753–12758. doi: 10.1073/pnas.0603449103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Rossi M, De Simone M, Pollice A, Santoro R, La Mantia G, Guerrini L, Calabro V. Itch/AIP4 associates with and promotes p63 protein degradation. Cell cycle. 2006b;5:1816–1822. doi: 10.4161/cc.5.16.2861. [DOI] [PubMed] [Google Scholar]
  35. Shen L, Chen XD, Zhang YH. MicroRNA-128 promotes proliferation in osteosarcoma cells by downregulating PTEN. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine. 2014;35:2069–2074. doi: 10.1007/s13277-013-1274-1. [DOI] [PubMed] [Google Scholar]
  36. Shen L, Ciesielski M, Ramakrishnan S, Miles KM, Ellis L, Sotomayor P, Shrikant P, Fenstermaker R, Pili R. Class I histone deacetylase inhibitor entinostat suppresses regulatory T cells and enhances immunotherapies in renal and prostate cancer models. PloS one. 2012;7:e30815. doi: 10.1371/journal.pone.0030815. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Su X, Chakravarti D, Cho MS, Liu L, Gi YJ, Lin YL, Leung ML, El-Naggar A, Creighton CJ, Suraokar MB, et al. TAp63 suppresses metastasis through coordinate regulation of Dicer and miRNAs. Nature. 2010;467:986–990. doi: 10.1038/nature09459. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Su X, Chakravarti D, Flores ER. p63 steps into the limelight: crucial roles in the suppression of tumorigenesis and metastasis. Nature reviews Cancer. 2013;13:136–143. doi: 10.1038/nrc3446. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sweet MJ, Shakespear MR, Kamal NA, Fairlie DP. HDAC inhibitors: modulating leukocyte differentiation, survival, proliferation and inflammation. Immunology and cell biology. 2012;90:14–22. doi: 10.1038/icb.2011.88. [DOI] [PubMed] [Google Scholar]
  40. van Kouwenhove M, Kedde M, Agami R. MicroRNA regulation by RNA-binding proteins and its implications for cancer. Nature reviews Cancer. 2011;11:644–656. doi: 10.1038/nrc3107. [DOI] [PubMed] [Google Scholar]
  41. Venkatanarayan A, Raulji P, Norton W, Chakravarti D, Coarfa C, Su X, Sandur SK, Ramirez MS, Lee J, Kingsley CV, et al. IAPP-driven metabolic reprogramming induces regression of p53-deficient tumours in vivo. Nature. 2015;517:626–630. doi: 10.1038/nature13910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Villagra A, Sotomayor EM, Seto E. Histone deacetylases and the immunological network: implications in cancer and inflammation. Oncogene. 2010;29:157–173. doi: 10.1038/onc.2009.334. [DOI] [PubMed] [Google Scholar]
  43. Volinia S, Galasso M, Sana ME, Wise TF, Palatini J, Huebner K, Croce CM. Breast cancer signatures for invasiveness and prognosis defined by deep sequencing of microRNA. Proceedings of the National Academy of Sciences of the United States of America. 2012;109:3024–3029. doi: 10.1073/pnas.1200010109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wada T, Kikuchi J, Furukawa Y. Histone deacetylase 1 enhances microRNA processing via deacetylation of DGCR8. EMBO reports. 2012;13:142–149. doi: 10.1038/embor.2011.247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. West AC, Johnstone RW. New and emerging HDAC inhibitors for cancer treatment. The Journal of clinical investigation. 2014;124:30–39. doi: 10.1172/JCI69738. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Yamaguchi H, Woods NT, Piluso LG, Lee HH, Chen J, Bhalla KN, Monteiro A, Liu X, Hung MC, Wang HG. p53 acetylation is crucial for its transcription-independent proapoptotic functions. The Journal of biological chemistry. 2009;284:11171–11183. doi: 10.1074/jbc.M809268200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Yan W, Chen X, Zhang Y, Zhang J, Jung YS, Chen X. Arsenic suppresses cell survival via Pirh2-mediated proteasomal degradation of DeltaNp63 protein. The Journal of biological chemistry. 2013;288:2907–2913. doi: 10.1074/jbc.M112.428607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch V, Andrews NC, Caput D, McKeon F. p63, a p53 homolog at 3q27-29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Molecular cell. 1998;2:305–316. doi: 10.1016/s1097-2765(00)80275-0. [DOI] [PubMed] [Google Scholar]
  49. Yeh JE, Toniolo PA, Frank DA. Targeting transcription factors: promising new strategies for cancer therapy. Current opinion in oncology. 2013;25:652–658. doi: 10.1097/01.cco.0000432528.88101.1a. [DOI] [PubMed] [Google Scholar]
  50. Yi R, Pasolli HA, Landthaler M, Hafner M, Ojo T, Sheridan R, Sander C, O’Carroll D, Stoffel M, Tuschl T, Fuchs E. DGCR8-dependent microRNA biogenesis is essential for skin development. Proceedings of the National Academy of Sciences of the United States of America. 2009;106:498–502. doi: 10.1073/pnas.0810766105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Yin H, Kanasty RL, Eltoukhy AA, Vegas AJ, Dorkin JR, Anderson DG. Non-viral vectors for gene-based therapy. Nature reviews Genetics. 2014;15:541–555. doi: 10.1038/nrg3763. [DOI] [PubMed] [Google Scholar]
  52. Yu D, Green B, Marrone A, Guo Y, Kadlubar S, Lin D, Fuscoe J, Pogribny I, Ning B. Suppression of CYP2C9 by MicroRNA hsa-miR-128-3p in Human Liver Cells and Association with Hepatocellular Carcinoma. Scientific reports. 2015;5:8534. doi: 10.1038/srep08534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Zhang C, Richon V, Ni X, Talpur R, Duvic M. Selective induction of apoptosis by histone deacetylase inhibitor SAHA in cutaneous T-cell lymphoma cells: relevance to mechanism of therapeutic action. The Journal of investigative dermatology. 2005;125:1045–1052. doi: 10.1111/j.0022-202X.2005.23925.x. [DOI] [PubMed] [Google Scholar]
  54. Zhang J, Xu E, Chen X. TAp73 protein stability is controlled by histone deacetylase 1 via regulation of Hsp90 chaperone function. The Journal of biological chemistry. 2013;288:7727–7737. doi: 10.1074/jbc.M112.429522. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS785717-supplement.pdf (32.7MB, pdf)

RESOURCES