1. Introduction
BET inhibitors, an emerging class of agents that target epigenetic dysregulated readers, BRD2, BRD3, BRD4 and BRDT, have recently been studied with a focus on the autophagic machinery. For example, Sakamaki et al.
[1] showed that treatment with the largely BET4 inhibitor, JQ1, induced autophagy in KP-4 cells, as evidenced by increased LC3 lipidation and puncta formation. These results were mirrored in vivo using mice models, in which JQ1 treatment caused the conversion of LC3I to LC3II, as well as increasing LC3 puncta formation in mouse small intestinal tissue
[1]. The BET inhibitor/degrader, ARV-825, was shown to reduce the levels of BRD2, BRD3 and BRD4 together with inducing autophagy, as shown by increased LC3II levels in KP-4 cells. The BET inhibitors JQ1 and I-BET151, as well as OTX015, upregulated the autophagic genes MAP1LC3B, SQSTM1, ATG2A and ATG16L2, as shown by RT-qPCR
[1].
2. Bladder Cancer
Li et al.
[2] studied the effect of the BET inhibitor JQ1 in the bladder carcinoma cell lines T24, 5637 and UMUC-3. Using an MTT assay, the counting of viable cells and clonogenic survival assays, JQ1 was shown to suppress the proliferation of the three cell lines in a dose-dependent manner. JQ1 treatment increased the number of autophagosomes and autolysosomes based on the GFP-RFP-LC3 fluorescence assay, consistent with autophagy induction
[3]. Autophagy induction by JQ1 was confirmed via the accumulation of LC3II and p-ULK1
[4], as well as p62/SQSTM1 degradation and the appearance of autophagic vacuoles by transmission electron microscopy (TEM), all of which was suggestive of a role for autophagy in the anti-proliferative effects of JQ1. To establish whether autophagy was, in fact, mediating drug action in these experimental models, the pharmacologic autophagy inhibitors 3-MA, BAF-A1 and NH
4Cl were found to interfere with the antitumor activity of JQ1. However, JQ1 could still partially inhibit cellular proliferation, suggesting that mechanistic aspects other than autophagy could also play a role in suppressing the proliferative ability of bladder cancer cells. These findings relating to autophagy were validated by demonstrating that the genetic inhibition of autophagy, using ATG5-directed siRNA, eliminated JQ1′s anti-proliferative actions, consistent with a cytostatic role of autophagy in this model. In this context, a cytostatic function of autophagy was described in early work from researcher's group
[5][6], although there have been few subsequent studies in the literature relating to this component of autophagic action. These results were further mirrored in vivo using a xenograft tumor model implanted with the T24 bladder cancer cell line. Here, JQ1 significantly inhibited tumor growth and induced autophagic flux, as indicated by elevated levels of LC3II and p-ULK1, as well as p62/SQSTM1 degradation.
In additional, quite rigorous mechanistic studies, JQ1 treatment was demonstrated to downregulate p-mTOR (consistent with the promotion of autophagy), accompanied by the upregulation of p-LKB1, p-AMPKα and its substrate p-ACC. Using AMPKα-directed siRNA, the JQ1 inhibitory effect was eliminated, accompanied by a reduction in autophagy, again shown by monitoring LC3II levels as well as the GFP-RFP-LC3 fluorescence assay. Furthermore, JQ1 could be shown to increase the recruitment of LKB1 and its interaction with AMPKα by co-immunoprecipitation assays, suggesting that the autophagy induced by JQ1 is dependent on the LKB1/AMPK/mTOR signaling pathway.
3. Ovarian Cancer
Luan et al.
[7] investigated the potential contribution of autophagy to JQ1 resistance in ovarian cancer using the A2780, HO-8910, SKOV-3 and HEY cell lines. JQ1 was shown to inhibit the proliferation of the four cell lines, with A2780 and HO-8910 cells showing less sensitivity than SKOV-3 and HEY cells. In these studies, JQ1 was shown to promote concentration-dependent apoptosis (using an annexin V assay). Consistent with the differential sensitivity observed, the SKOV-3 and HEY cells demonstrated a higher apoptosis response as compared to the A2780 and HO-8910 cell lines, in a dose-dependent manner. BRD4 and c-Myc levels were downregulated in the four cell lines upon JQ1 treatment, as would have been expected considering that these are established downstream targets of JQ1 and other BET inhibitors
[8][9].
Luan et al.
[7] further studied whether JQ1 induced autophagy in these preclinical models. In the less sensitive A2780 and HO-8910 cell lines, acridine orange staining, the accumulation of LC3 II, ATG5 and Beclin1 and p62/SQSTM1 degradation were indicative of autophagic flux induced by JQ1
[10]. However, this was not the case with the more sensitive SKOV-3 and HEY cell lines, which showed minimal changes in the autophagic markers with JQ1. These results suggested that JQ1-induced autophagy may have contributed to the reduced sensitivity to JQ1 in the A2780 and HO-8910 cells. The combination of JQ1 with the pharmacologic autophagy inhibitors, 3-MA and CQ, sensitized the A2780 and HO-8910 cell lines and increased the extent of apoptosis, as shown by annexin V staining and cleaved-PARP levels. Taken together, this series of experiments suggested a cytoprotective role of autophagy in two of the four cell lines. This cytoprotective role of autophagy was further confirmed in vivo using xenograft tumor models implanted subcutaneously with A2780 cells; specifically, JQ1 in combination with CQ showed higher antitumor activity than each drug alone in the mice models.
Luan et al.
[7] further studied the relation between JQ1-induced autophagy and the Akt/mTOR pathway. JQ1 treatment in the less sensitive A2780 and HO-8910 cell lines caused a reduction in the phosphorylation levels of Akt, and, consistent with the findings of Li et al.
[2], mTOR or p70S6K. However, this was not the case in the JQ1-sensitive SKOV-3 and HEY cell lines, where JQ1 did not affect the levels of Akt, mTOR and p70S6K.
4. Breast Cancer
Ali et al.
[11] studied the possible targeting of the BRD4/RAC 1 axis using the MCF-7, MDA-MB-231, JIMT1 and SKBR3 breast tumor cells lines, also using JQ1. BRD4 inhibition, using JQ1 in combination with the RAC1 inhibitor NSC23766, suppressed cellular growth, clonogenic survival, cell migration and mammary stem cell expansion. The anti-proliferative effect of the combination was confirmed in vivo, using the MDA-MB-231-based xenograft model, where JQ1 in combination with NSC23766 showed significant combined anti-proliferative effects. Here, JQ1 in combination with NSC23766 induced senescence in MCF-7, MDA-MB-231 and JIMT1 cells, but not in SKBR3 cells. The combination treatment resulted in the greater accumulation of LC3-II than each drug alone, suggesting that the combination of JQ1 and NSC23766 may induce (either cytostatic or cytotoxic) autophagy. However, these studies did not further explore the potential involvement of autophagy in JQ1 action, in contrast to the studies described above in the ovarian cancer and bladder cancer cell work.
5. Glioblastoma
Colardo et al.
[12] studied the relationship between the BET family and autophagy in glioblastoma (GBM) using U87MG (U87), GL15 and the patient-derived GH2 cell lines. These investigators reported that the expression levels of BRD2 and BRD4 were high in these three cell lines, as well as in GBM patient samples. JQ1 significantly suppressed the proliferation of both the U87 and GH2 cell lines. Furthermore, upon combining JQ1 with temozolomide, the standard-of-care of therapy for GBM
[13], a greater anti-proliferative response was evident compared to each drug alone, together with increased apoptosis. JQ1 treatment also promoted marked morphological changes in both cell lines, such as increasing the number of GBM cells with cytoplasmic extensions and elongation, suggesting the induction of a differentiation process. The induction of differentiation was further confirmed by the accumulation of synaptophysin and β3-tubulin, early markers of neuronal differentiation
[14][15].
With respect to autophagy, the protein levels of autophagy-related markers ULK1, ATG5, ATG7, Beclin1 and p62/SQSTM1 were examined by Western blotting in U87 and GH2 cells
[12]. Here, JQ1 caused the transient upregulation of ULK1, an upstream promoter of autophagy, but without promoting significant differences in ATG5, ATG7 and Beclin1 protein levels. Furthermore, a transient elevation in p62/SQSTM1 levels at 24 h after JQ1 treatment was observed, followed by a significant reduction, which was not prevented by CQ, suggesting that there was an impairment in p62/SQSTM1 expression instead of p62/SQSTM1 degradation. Despite these apparent inconsistencies, JQ1 induction of autophagy was confirmed in U87 and GH2 cells by immunofluorescence analysis of endogenous LC3 with the accumulation of LC3 dots
[12].
The relationship between GBM cell differentiation and autophagy was investigated using Beclin1-directed shRNA in GL15 cells
[12]. Upon combining genetic autophagy inhibition with JQ1, β3-tubulin and synaptophysin levels were lower than in the shRNA control samples. These results were confirmed with the pharmacological inhibition of autophagy using CQ, where the JQ1-induced accumulation of β3-tubulin and synaptophysin was suppressed upon CQ treatment, confirming the importance of the autophagic flux in JQ1-induced GBM differentiation
[12].
The role of BET family members in regulating stem cell differentiation, through the involvement of various signal transduction pathways, has been studied previously
[16][17]. BRD4 was shown to be occupied at the Notch1 promoter site, thus controlling the Notch1 signaling pathway that is involved in regulating the self-renewal capacity of glioma stem cells and their tumorigenicity
[18]. Interestingly, Li et al.
[19] reported that BET inhibition caused an increase in the number of neurons. Furthermore, gene expression profiling analysis demonstrated that BET bromodomain inhibition induced a transcriptional program enhancing the directed differentiation of neural progenitor cells into neurons, while suppressing cell cycle progression and gliogenesis. Regarding autophagy, it is well known that autophagy plays an important role during embryonic development and differentiation, maintaining cellular homeostasis as well as the stemness characteristics of self-renewing cells
[20][21][22]. Therefore, Colardo et al.
[12] proposed that dysregulated BET protein expression negatively regulates autophagy in GBM cells, maintaining stemness and contributing to tumor aggressiveness.
These data raise the question of whether a form of autophagy, either cytoprotective, cytotoxic or cytostatic, plays a prominent role in the differentiation process. It would appear to be the cytostatic form of autophagy, consistent with growth arrest in differentiated cells.
6. Pancreatic Cancer
Xu et al.
[23] studied the combination of JQ1 and arsenic trioxide (ATO) in pancreatic cancer. Using 11 pancreatic cancer cell lines, these investigators observed that ATO effectively induced ER stress, an unfolded protein response (UPR)
[13] and eventually apoptosis in some cell lines, including B×PC-3 and MIAPaCa-2 cells, independently of K-ras or p53 status. Furthermore, upon analyzing ATO-treated B×PC-3 microarray data, autophagy-related genes were upregulated, including GABARAPL1, GABARAPL2, ULK1 and ATG12, which was further validated using RT-PCR analysis. Autophagy activation was further confirmed by Western blotting, indicating the upregulation of LC3-II, ATG5, ATG7 and Beclin1 protein levels; however, p62/SQSTM1 protein as well as mRNA levels were elevated, which would appear to be a contradictory outcome indicative of autophagy inhibition. Therefore, using a GFP assay as well as microarray data, these authors showed that ATO promoted immature autophagosome accumulation via activating autophagosome formation but also inhibiting the lysosomal functions and consequently degradation. Treatment with the autophagy inhibitor CQ further reduced the viability of B×PC-3 and MIAPaCa-2 cells, suggesting that a cytotoxic form of autophagy resulted from the accumulation of autophagosomes and impaired lysosomal function, leading to cell death
[13]. Autophagy induction was also shown to result from ATO-induced ER stress, in which TEFB and TFE3 may play important roles.
Nuclear factor (erythroid-derived 2)-like 2 (NRF2) is a basic leucine zipper transcription factor within the cap “n” collar family
[24]. NRF2 regulates the activity of many enzymes and transporters that are involved in fatty acid synthesis and oxidation, xenobiotic detoxification and transportation, as well as conjugation reactions
[25][26]. Furthermore, NRF2 regulates the activity of other transcription factors, including AhR, PPAR γ, CEBPα and RXRα
[26][27][28]. Recently, attention has been directed to the possible relationships between NRF2 and cancer
[29]. Wu et al.
[26] suggested that NRF2 is a double-edged sword. NRF2 signaling pathways appear to be responsible for protection against chemical-induced oxidative damage, maintaining redox homeostasis and exerting anti-inflammatory as well as antineoplastic activity. However, NRF2’s persistent activation has been associated with metabolic reprogramming, apoptosis suppression and increasing the self-renewal abilities of cancer stem cells, as well as chemotherapeutic resistance
[26].
Interestingly, Xu et al.
[23] showed that ATO treatment altered the NRF2 expression profile, causing a decrease in the sensitive cells and an increase in the insensitive cells. Moreover, upon NRF2 knockdown, ATO treatment resulted in reduced cell viability and the induction of an ER stress response, as well as apoptosis in insensitive cells. Similar results were generated in mice models implanted with NRF2 knockdown cells, in which ATO treatment resulted in a significant reduction in tumor size
[23]. To investigate whether a possible relationship exists between autophagy and NRF2 in pancreatic cancer, NRF2 was depleted in the ATO-insensitive PANC-1 cell line. Here, autophagic genes were significantly elevated in the NRF2 knockdown cells upon ATO treatment. Autophagy induction was confirmed by the accumulation of LC3II, as well as by TEM-detected autophagosome formation
[23]. Moreover, lysosomal-related genes/proteins were suppressed in the NRF2 knockdown cells upon ATO treatment, suggesting the impairment of lysosome-related activity.
JQ1 sensitized the ATO-resistant cell lines, including PANC-1, YAPC, AsPC-1, HAPF-II, CFPAC-1 and HUP-T4 cells. JQ1 treatment led to the significant downregulation of NRF2 levels, as assessed by an immunoblotting assay, without affecting NRF2 mRNA, suggesting that NRF2 is regulated by BET proteins via indirect mechanisms, including translation or protein stability control
[23]. Furthermore, the combination of JQ1 and ATO produced a greater reduction in tumor size than each drug alone in mice models implanted with pancreatic cancer cells. JQ1 in combination with ATO generated a further increase in the levels of LC3 II, ATG-5, ATG-7 and Beclin1, together with p62/SQSTM1 accumulation, than each drug alone
[23]. Interestingly, these findings were accompanied by a reduction in the lysosomal protein CTSB, suggesting the impairment of lysosome function. The impairment of lysosomal function was further confirmed using immunofluorescent assays. Upon combining ATO and JQ1 together with CQ, the viability of the ATO/JQ1-treated population was significantly reduced, suggesting a possible cytotoxic role of autophagy due to the accumulation of autophagosomes, leading to cell death
[13]. Therefore, they proposed that JQ1 exerts an effect similar to NRF2 knockdown in sensitizing pancreatic cancer cells to ATO-induced autophagosome accumulation, ER stress/UPR and apoptosis
[23].
7. Leukemia
Jang et al.
[30] studied the relationship between autophagy and JQ1 resistance in leukemia stem cells (LSC), using the KG1, KG1a and Kasumi-1 cell lines, which are enriched in the LSC phenotype. JQ1 inhibited the proliferation of the Kasumi-1 cell line, with a lesser anti-proliferative response noted with KG1 and KG1a cells. Moreover, the apoptotic response to JQ1 varied, with the highest apoptotic population evident in Kasumi-1 cells, with minimal apoptosis in the KG1 and KG1a cell lines. These results were mirrored in AML patient samples, with variability in the apoptotic response to JQ1.
Mechanistically, JQ1 was able to suppress c-Myc expression in both JQ1-sensitive and insensitive cell lines
[30]. JQ1 induced autophagy in the KG1 and KG1a cell lines but not in Kasumi-1 cells, as evidenced by the conversion of LC3I to LC3II, as well as by p62/SQSTM1 degradation. The autophagy induction was further confirmed using TEM as well as GFP-LC3 puncta, showing a significant increase in the number of LC3 puncta in the KG1 and KG1a cell lines but not in Kasumi-1 cells
[30]. Autophagy induction was also confirmed in JQ1-resistant patient samples, as shown by increased LC3I/LC3II conversion and an increase in the number of GFP-LC3 puncta, suggesting a possible relationship between autophagy and JQ1 resistance, i.e., that autophagy was demonstrating a cytoprotective function. This assumption was confirmed with pharmacologic autophagy inhibition using bafilomycin A1, 3-MA or hydroxychloroquine, and genetically using Beclin1-directed siRNA
[30]. Autophagy inhibition sensitized the resistant cell lines and patient samples to JQ1, as confirmed by increased apoptosis and cleaved caspase-3 and cleaved-PARP accumulation, suggesting a cytoprotective role of autophagy in the resistant cells.
On the molecular level, Beclin1 levels were observed to be increased in parallel with LC3II in the resistant cell lines, which was not the case in the JQ1-sensitive Kasumi-1 cells
[30]. Furthermore, JQ1 increased the phosphorylation of AMPK (Thr172), ULK1 (Ser555), mTOR (Ser2448) and p70S6K in the resistant cells (in contrast to both Li et al.
[2] and Luan et al.
[7], confirming that autophagy is tumor/cell type-specific).
Wang et al.
[31] reported that the overexpression of c-Myc by tumor cells is needed to maximize glycolysis and oxidative phosphorylation in order to support the high level of ATP consumption required by rapid, proliferation-associated anabolism in tumor cells
[32][33]. Therefore, c-Myc inhibition is accompanied by metabolic de-regulation, mitochondrial atrophy, neutral lipid accumulation, cell cycle arrest, ATP depletion and an effort to replenish ATP via the upregulation of AMPK. Therefore, Jang et al.
[30] proposed that JQ1 suppresses c-Myc levels, which, in turn, depletes ATP stores and induces the phosphorylation of AMPK
[31]. The AMPK-mediated phosphorylation of ULK1 induces autophagy, which can be responsible for JQ1 resistance, independently of the mTOR pathway. The role of AMPK in JQ1 resistance was confirmed via AMPKα-targeted siRNA, as well as by the use of the AMPK pharmacological inhibitor, compound C. AMPK inhibition increased the extent of the apoptotic response in JQ1-treated resistant cell lines and patient samples, consistent with autophagy expressing a cytoprotective function.
On the other hand, JQ1 increased AMPK phosphorylation without p-ULK1 accumulation or autophagy induction, in Kasumi-1 cells and JQ1-sensitive patient samples, with a reduction in both mTOR and p70S6K phosphorylation. It was therefore proposed that the failure of JQ1 to promote ULK1 phosphorylation was the basis for the absence of autophagy induction. However, JQ1 increased the levels of apoptotic protein markers, including cleaved caspase-3, cleaved caspase-9 and PARP. These data raise the question as to why ULK-1 is induced in the JQ1-resistant cell line, leading to the induction of cytoprotective autophagy, and not the JQ1-sensitive cell lines, despite AMPK activation and mTOR inhibition.
This entry is adapted from the peer-reviewed paper 10.3390/ijms241612669