Targeting an inducible SALL4-mediated cancer vulnerability with sequential therapy

Abstract

Oncofetal protein SALL4 is critical for cancer cell survival. Targeting SALL4, however, is only applicable in a fraction of cancer patients that are positive for this gene. To overcome this limitation, we propose to induce a cancer vulnerability by engineering a partial dependency upon SALL4. Following exogenous expression of SALL4, SALL4-negative cancer cells became partially dependent on SALL4. Treatment of SALL4-negative cells with the FDA-approved hypomethylating agent 5-Aza-2’-deoxycytidine (DAC) resulted in transient upregulation of SALL4. DAC pre-treatment sensitized SALL4 negative cancer cells to Entinostat, which negatively affected SALL4 expression through a microRNA, miRNA-205, both in culture and in vivo. Moreover, SALL4 was essential for the efficiency of sequential treatment of DAC and Entinostat. Overall, this proof-of-concept study provides a framework whereby the targeting pathways such as SALL4-centered therapy can be expanded, sensitizing cancer cells to treatment by transient target induction and engineering a dependency.

Graphical Abstract

Introduction

SALL4 is a zinc finger transcription factor that belongs to the spalt-like (SALL) gene family. SALL4 plays a significant role in the maintenance of self-renewal and pluripotency of embryonic stem cells(1,2). While SALL4 expression is silenced in most healthy adult tissues, it is re-expressed in various cancers and is regarded as an important biomarker for poor outcomes (3–6). Targeting SALL4 by RNA interference or peptides has been shown to be effective against a wide range of cancers, including lung cancer(7), endometrial cancer(8), gastric cancer(9), and liver cancer(10). However, the fact that SALL4 is only positive in a sub-set of cancer patients limits the application of SALL4-based cancer treatment(11–13), which is broadly true for all targeted therapies.

The function and downstream targets of SALL4 have been explored and identified(9,14,15). However, the regulation of SALL4 remains unexplored. Methylation of the SALL4 locus has been reported to be negatively associated with SALL4 expression (16). Hypomethylation agents (HMA), such as 5-aza-2’-deoxycytidine (DAC) and 5-azacytidine (5-Aza), are FDA-approved agents used to treat myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML)(17–19), functioning at least in part through inhibition of DNA methyltransferases (DNMTs)(20). The effects of HMA on SALL4 expression have not been carefully examined. SALL4 functions as a gene repressor by interacting with the Nucleosome Remodeling Deacetylase (NuRD) complex, in which histone deacetylases (HDACs) are critical components(21). We have previously shown that the gene signatures of the HDAC inhibitor Entinostat (ENT) and SALL4 were correlated using the connectivity map tool(7). We rationalized at that time that by blocking HDAC enzymatic activity, ENT functions as a SALL4 inhibitor by a mechanism akin to a SALL4 peptide blocker(7). Intriguingly, we also noticed that SALL4 protein level was significantly decreased upon ENT treatment, while the mechanism remained unclear. This prompted us to examine whether ENT can potentially function as a SALL4 inhibitory drug by modulating its expression possibly by upregulating a regulatory RNA.

MicroRNAs (miRNAs) are a class of small non-coding RNAs that act as post-transcriptional regulators by inducing mRNA degradation or translation repression of their targets through directly binding to the 3’untranslated region (3’UTR)(22–24). MiRNAs have been demonstrated to participate in various biological processes, such as self-renewal, cell proliferation, cell cycle, migration, and apoptosis(25–27). Dysregulation of miRNAs is frequently observed during the development and tumorigenesis(28–30). Several mechanisms have been identified contributing to the aberrant expression of miRNAs in cancers, in which epigenetic changes exhibit a significant role(31–33).

In this study, we demonstrate that SALL4 expression negatively correlated with its DNA methylation and can be upregulated by DAC treatment. Overexpressing SALL4 in negative cancer cells leads to a SALL4 partial dependency. Increased SALL4 expression can enhance cancer cells’ sensitivity to ENT, a drug that negatively regulates SALL4 expression post-transcriptionally through miRNA-205. We have further evaluated the potential of DAC plus ENT-based therapy to expand SALL4-targeted therapy in SALL4 negative cancers using cell culture and in vivo xenotransplantation models.

Materials and Methods

Cell culture and reagents

Lung cancer cell lines H661 and H1299, hepatocellular cancer cell lines SNU387 and SNU398, leukemic cell lines K562 and HL60 were purchased from the American Type Culture Collection with authentication and cultured in RPMI 1640 medium (Thermo Fisher Scientific, Waltham, USA) supplied with 10% fetal bovine serum and 1% antibiotics (100 U/ml penicillin and 100 mg/ml streptomycin) at 37°C in 5% CO2. 293T cells were cultured in DMEM medium (Thermo Fisher Scientific, Waltham, USA) with the same condition. All cell lines were used no more than five passages and tested with negative mycoplasma. ENT was provided by Jun Qi’s lab from Dana Farber Cancer Institute. 5-aza-2’-deoxycytidine was purchased from Sigma-Aldrich (Cat. A3656, Louis, USA).

Cell transfection

MiRNA negative controls, miR-205 mimics, and its inhibitor were purchased from Sigma-Aldrich (Louis, USA). shRNA against SALL4 was designed as previously described(1). MiRNAs or vectors were transfected in the amount of 2 μg for 6-well plate into indicated cells by Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific, Waltham, USA) according to the manufacturer’s instructions. After 48 hours, the cells were used for subsequent experiments.

Total RNA extraction and quantitative real-time PCR

Cells were dissolved in TRIzol reagent (Thermo Fisher Scientific, Waltham, USA) and total RNAs were obtained according to the manufacturer’s protocol and then quantified and synthesized into cDNA using an iScript cDNA Synthesis Kit (Bio-Rad, Hercules, USA) for SALL4 expression or using a High-capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, USA) for miRNA expression. Real-time PCR was performed using SYBR Green Super Mixes (Bio-Rad, Hercules, USA). GAPDH and U6 were used as endogenous controls for normalization. MiRNA-specific primers were purchased from Ribobio (Guangzhou, China). Relative levels of expression were normalized and calculated using 2−ΔΔCt method. Information on the primers is listed in supplemental file 1.

SALL4 expression in patient-derived samples

Patient samples were obtained with written informed consent in accordance with the Declaration of Helsinki and approval of the human research ethics committee of the Southeastern Sydney Local Health District (HREC ref# 17/295). Patient characteristics are available in Supplement Table 1. Bone marrow samples were collected prior to cycle 1 day 1 (C1D1) and following cycle 7 day 1 (C7D1) 6 cycles of subcutaneous administration of 5-azacytidine (Vidaza, Celgene), and mononuclear cells isolated using LymphoPrep (Axis-Shield). RNA was isolated with All-in-One DNA/RNA Mini-Preps Kit according to the manufacturer’s protocol (Bio Basic) and reverse transcribed using either a QuantiTect Revese Transcription Kit (Qiagen) or Maxima H Minus Reverse Transcriptase (Thermo Scientific) according to the manufacturer’s protocol. Quantitative PCR for SALL4 and GAPDH was performed using PowerUp SYBR Green Master Mix (Applied Biosystems) and an MX3000P thermocycler (Stratagene), and relative levels of expression were calculated using the 2-ΔΔCt method. Information on the primers is listed in supplemental file 1.

Western blot analysis

Cells were washed by cold PBS and treated with lysis buffer (150 mM NaCl, 50 mM Tris pH7.5, 1 mM EDTA, 0.5% Triton X-100, 0.1% deoxycholic acid sodium salt) on ice for 30 mins. Then cells were scraped and after centrifugation lysate supernatant was collected and stored at −80°C. The BCA Assay kit (Bio-Rad, USA) was used to test protein concentration. Protein samples were denatured and then separated on SDS-PAGE and transferred to PVDF membrane (Millipore, Burlington, USA). After blocking by non-fat milk for 1 hour, membranes were incubated overnight at 4°C with primary SALL4 antibody from Santa Cruz Biotechnology (Cat. EE-30, Dallas, USA), PARP antibody from Cell Signaling Technology (Cat. 9542S, MA, USA) and β-actin antibody from Sigma-Aldrich (Cat. A1978, St. Louis, USA) with 1:1000 dilution. Then after washing 3 times, membranes were incubated with anti-mouse HRP-conjugated secondary antibody from GE Healthcare (Cat. NA9311ML, Chicago, USA). The bands were analyzed by an Immobilon Western Chemiluminescent HRP Substrate (Millipore, Burlington, USA).

MiRNA sequencing

Total RNA was extracted from the H661 cells after DMSO or ENT treatment (duplicate samples for each treatment) for 8 hours using TRIzol (Invitrogen, Carlsbad, USA) and a miRNeasy mini kit (Qiagen, Hilden, Germany), following the manufacturer’s instructions. The RNA samples were sent to the Molecular Biology Core Facilities of Dana-Farber Cancer Institute for library preparation and Illumina MINIseq Next Generation Sequencing. Raw data could be accessed at GSE153589. miRNAs expression with at least 2-fold change was considered as significantly changed after ENT treatment.

RNA sequencing

Total RNA was extracted from H1299-SALL4 and H1299-GFP cells (duplicate samples for each treatment) using an RNeasy Micro Kit (Qiagen, Hilden, Germany) for RNA extraction, following the manufacturer’s instructions. The RNA samples were sent to the Molecular Biology Core Facilities of Dana-Farber Cancer Institute for library preparation and Illumina Next Generation Sequencing. Libraries were prepared using Qiagen Fast Select rRNA removal reagents followed by Kapa RNA Hyper Prep reagent kits from 200ng of purified total RNA according to manufacturers’ protocols. Raw gene counts were generated from STAR output. Raw data can be accessed at GSE180603. RNA-seq data was processed and analyzed for differential expression using CSI NGS portal(34). Expression with at least a 2-fold change between H1299-SALL4 and H1299-GFP cells was considered as significantly changed.

ChIP-qPCR

Briefly, 10 million cells after indicated treatment were crosslinked with 1% formaldehyde, then lysed in buffer containing 1% SDS, and spun at 20,000RPM to isolate the chromatin. Sonication was performed in lysis buffer containing 0.1% SDS. Protein A/G PLUS-Agarose (Santa Cruz Biotechnology, Dallas, USA) were conjugated with H3K27ac or IgG antibody (Cell Signaling Technology, Danvers, USA) and incubated with sonicated chromatin overnight at 4 degrees. Washes (0.1% SDS lysis buffer, LiCl wash buffer, and tris-EDTA) were performed and the chromatin-protein complex was reverse crosslinked by heating at 65 degrees with proteinase K (8ug/mL). The chromatin was incubated with RNaseA (8ug/mL) before it was extracted, purified with phenol chloroform (pH 8), and precipitated by ethanol. qPCR was performed using SYBR Green Super Mixes (Bio-Rad, Hercules, USA). Information on the primers is listed in supplemental file 1.

Cell immunofluorescence staining

H661 cells were washed twice with cold PBS and then fixed in 4% paraformaldehyde/PBS for 10 min. After washing with PBS three times, cells were incubated with 0.2% Triton X-100 for 10 min. Then the cells were incubated with primary antibody against SALL4 at 4°C overnight. Antibody was applied in PBS containing 1% bovine serum albumin, followed by incubation with secondary antibodies for 2 hours at 37°C. Images were obtained using a laser microscope (Olympus, Japan).

Cell viability and apoptosis assay

The Cell Counting Kit-8 (CCK8, Dojindo, Japan) was used to detect cell viability. Briefly, 1×10³ cells after the indicated treatment were seeded into 96-well plates and cultured for 5 days. Then 10 μL of CCK-8 solution was added to each well. After 3 hours of culturing, the absorbance at 450 nm was measured using a spectrophotometer. To determine cell apoptosis, the Apoptosis Detection Kit was used (BD Pharmingen, Bedford, USA), and cells were washed and resuspended in binding buffer, followed by staining with Annexin V and propidium iodide for 30 min prior to flow cytometry analysis.

Dual-luciferase reporter assay

The wild-type 3’UTR region of SALL4 mRNA or a mutant without the miR-205 binding site (Figure 4E) was amplified using PCR and cloned into the pGL3 vector (Promega, Madison, USA). HEK 293T cells were seeded into 24-well plates, then co-transfected with the indicated vectors and miR-205 mimics or the miR-negative control using Lipofectamine 2000 according to the manufacturer’s protocol. Finally, luciferase activities were measured using the dual-luciferase reporter gene assay kit (Promega, Madison, USA).

Figure 4. MiR-205 was involved in ENT-induced SALL4 inhibition.

(A) miRNA sequencing results of H661 cells after 8 hours treatment with 2.5 μM ENT or DMSO. (B) miR-205 was up-regulated by ENT and predicted to target SALL4 by the Targetscan database. (C) Real-time PCR of miR-205 expression in H661 cells after 2.5 μM ENT or DMSO treatment for 8 hours. U6 was used as a normalization control. (D) ChIP-qPCR of the miR-205 promoter region and non-specific control region after 2.5 μM ENT treatment. IgG was used as a normalization control. (E) Sequence of wild-type SALL4 3’UTR with miR-205 binding site or mutant one without binding site. (F) Luciferase reporter result in 293T cells after transfection with the indicated vectors. (G) mRNA and (H) protein levels of SALL4 in H661 cells after transfection with miR-205 or non-specific control. (I) Cell viability of H661 cells after indicated treatment for 5 days. (J) Flow cytometry of H661 cells after the indicated treatment. (K) Percent apoptotic and dead cells in (J). (L) Cell viability of ENT pre-treated H661 cells after transfection of miR-205 or non-specific control. (M) Flow cytometry of ENT pre-treated H661 cells after indicated treatment. (N) Percent apoptotic and dead cells in (M). n.s. means P>0.05, *P<0.05, **P<0.01, ***P<0.001, N=3.

Bisulfite treatment and sequencing

SALL4 Exon 1 region methylation status was assessed using bisulfite sequencing or pyrosequencing. In brief, 1 μg of genomic DNA extracted using the PureLink Genomic DNA Mini Kit (Invitrogen) was bisulfite-converted by using the Epimark Methylation kit (NEB). Each experiment included non-CpG cytosines as internal controls to detect incomplete bisulfite conversion of the input DNA. For manual methylation tests on H1299 cells, PCR products were gel purified (Qiagen) from the 1.5% TAE gel and cloned into the pMD20T vector (Takara Bio) for transformation. The cloned vectors were transformed into ECOS 101 DH5α cells and miniprep was performed to extract plasmids. Sequencing results were analyzed using BiQ analyser software. The number of clones for each sequenced condition was 10.

Pyrosequencing of the SALL4 gene in H661 cells was performed by EpigenDx, Inc. (Hopkinton, MA, United States). The PCR product was bound to Streptavidin Sepharose HP (GE Healthcare Life Sciences), after which the immobilized PCR products were purified, washed, denatured with a 0.2 μM NaOH solution, and rewashed using the Pyrosequencing Vacuum Prep Tool (Pyrosequencing, Qiagen), as per the manufacturer’s protocol. Pyrosequencing of the PCR products was performed using 0.5 μM of sequencing primer on the PSQ96 HS System (Pyrosequencing, Qiagen) according to the manufacturer’s instructions. The mean methylation level was calculated using methylation levels of all measured CpG sites within the targeted region of SALL4 gene.

Tumor cell implantation and treatment

All experimental procedures involving animals were conducted in accordance with the institutional guidelines set forth by the Children’s Hospital Boston (CHB animal protocol number 11–09-2022). Eight to ten-week-old non-obese diabetic scid gamma (NSG) mice were housed in a specific pathogen-free facility. One million H1299 cells subjected to different drug treatments were injected subcutaneously into the hindlimbs on both sides. Cells were resuspended in 0.1 mL of PBS and then mixed with 0.1 mL of Matrigel. Tumors were harvested and weighed after 30 days. For drug treatment, after 10 days of H1299 cells implantation, mice received an intraperitoneal injection of either vehicle or DAC (2.5 mg/kg) for 5 days. Then an intraperitoneal injection of either vehicle or ENT (2.5 mg/kg) was given for the following two weeks (5 consecutive injections per week). Tumor volume was calculated by using the formula: Tumor volume=length x width² / 2.

Statistical analysis

Statistical analysis was performed by GraphPad Prism 8 (GraphPad Software, San Diego, USA). All experiments were conducted for triplicate. All data were presented as mean ± SD. The data were analyzed by Student t test. P < 0.05 was considered statistically significant.

Data and materials availability

miRNA-sequencing data and RNA-sequencing data have been deposited in Gene Expression Omnibus (GSE153589 and GSE180603). All other data associated with this study are present in the paper or the additional information.

Results

A SALL4-partial dependency can be established in SALL4 negative cancer cells

Previously, we and others have shown that SALL4 is critical for cancer cell survival in SALL4 positive tumors(7,8,10). Using the Dependency Map (DepMap) (https://depmap.org/portal/) analytical tool(35,36), we observed that cancer cells with high SALL4 expression exhibited more dependency on SALL4 compared to those with low SALL4 expression (Supplement Figure 1A). Cancer cell lines with high SALL4 expression (232 cell lines, using H661 and SNU398 as references) were more dependent on SALL4 compared to those with low SALL4 expression (449 cell lines, using H1299 and SNU387 as references). However, it remains unknown whether SALL4-vulnerability can be established in SALL4 negative cancers. To answer this question, SALL4 was stably introduced into SALL4 negative H1299 lung cancer cells by retroviral transduction (designated H1299-SALL4). The level of expressed SALL4 was comparable to the SALL4 positive H661 lung cancer cells (Supplement Figure 1B&C). Intriguingly, treatment with shRNA against SALL4 led to significant inhibition of cell viability and induction of apoptosis in H1299-SALL4, but not in control H1299-GFP cells (Figure 1A–D). Loss of viability and induction of apoptosis in H1299-SALL4 following SALL4 repression mirrored the phenotype observed when knocking down SALL4 in H661 lung cancer cells, which express endogenous SALL4 (Figure 1E–H). This suggests that SALL4 negative cancer cells can become addicted to SALL4 once expressed. A similar SALL4-partial dependency was also observed in the SALL4 positive liver cancer cell line SNU398 (Supplement Figure 1D) and paired isogenic SALL4 negative liver cancer SNU387 cells with and without SALL4 expression (Supplement Figure 1E). The level of expressed SALL4 in SNU387 was comparable to the SALL4 positive SNU398 liver cancer cells (Supplement Figure 1 F&G). We have previously observed that exogenous expression of SALL4 in negative SNU387 cells liver cancer cells could enhance their sensitivity to oxidative phosphorylation inhibitors due to a shift in metabolism(37), suggesting SALL4 could reprogram cancer cells. RNA-seq results of H1299-SALL4 cells revealed alterations in various cellular processes compared to H1299-GFP cells, especially in metabolic processes (Supplement Figure 1H–J). Furthermore, inhibition of SALL4 in a xenograft mouse model of H1299-SALL4 also led to a highly significant reduction of tumor growth (Figure 1I–K).

Figure 1. Engineered SALL4-partial dependency in SALL4-negative cancer cells.

(A) Western blot of SALL4 protein in GFP control (H1299-GFP) or exogenous SALL4-expressing H1299 (H1299-SALL4) cells 48 hours after transfection of shRNA negative control (NC) or SALL4 shRNA (SALL4), β-actin was used as normalized control. (B) Cell viability of GFP control or SALL4-expressing H1299 cells after the indicated treatments. (C) Flow cytometry plots showing Annexin/PI staining of GFP control or SALL4-expressing H1299 cells after the indicated treatments. (D) Percent apoptotic (Annexin⁺) and dead cells (Annexin⁻ /PI⁺) from (C). (E) Western blot of SALL4 protein in H661 cells 48 hours after transfection of negative control sh-NC or SALL4 shRNA. β-actin was used as normalized control. (F) Cell viability of H661 cells after indicated treatment. (G) Flow cytometry plots showing Annexin/PI following the indicated treatment. (H) Percent apoptotic and dead cells from (G). (I) Representative images of tumors harvested from mice 30 days following sub-cutaneous injection of H1299 cells expressing the indicated vectors. (J) Tumor weight and (K) body weight from indicated groups (N=5). n.s. means P>0.05, *P<0.05, **P<0.01, ***P<0.001, N=3.

Taken together, these results demonstrated that expression of SALL4 in negative cancer cells can sensitize these cells to SALL4 targeting with inhibition of cell growth both in cell culture and in vivo.

Upregulation of SALL4 with decreased DNA methylation post HMA treatment

We next evaluated whether inducing SALL4 pharmacologically creates a vulnerability that could be therapeutically targeted. SALL4 expression is negatively correlated with DNA methylation at a CpG region of 5’UTR-Exon 1 (16). We, therefore, explored the relationship between SALL4 RNA expression levels and density of DNA methylation at 5’UTR-Exon 1 region in 386 cancer cell lines (CCLE database) and in 722 primary tumors (TCGA database, including lung adenocarcinoma and squamous cell carcinoma, colon adenocarcinoma, hepatocellular carcinoma, and gastric adenocarcinoma). As shown in Figures 2A and B, compared to the high methylation (>75% methylation) groups, the expression of SALL4 was significantly increased (4-fold in cell lines, and 2.4-fold in primary cancer patients) in the low methylation (<25% methylation) groups. 5-azacitidine has been approved by the FDA for the treatment of myelodysplasia, chronic myelomonocytic leukemia (CMML), and acute myeloid leukemia. Following six cycles of 5-azacitidine therapy, SALL4 expression was increased in bone marrow mononuclear cells in 7/13 patients, compared with pre-treatment levels (Figure 2C).

Figure 2. Upregulation of SALL4 by hypomethylating agents.

(A) SALL4 expression levels in cancer cell lines with either high- (>75%) and low- (<25%) CpG methylation at exon 1, data generated from CCLE database. (B) SALL4 expression levels in primary tumors from lung adenocarcinoma and squamous cell carcinoma, colon adenocarcinoma, hepatocellular carcinoma, and gastric adenocarcinoma with either high- (>75%) and low- (<25%) CpG methylation at exon 1, data generated from TCGA database. (C) Fold change of SALL4 expression measured by real-time PCR from patients’ bone marrow mononuclear cells after six cycles of 5-Aza treatment (C7D1/C1D1), GAPDH was used for a normalization control. (D) RNA-seq data of H1299 cells after treatment of 1 μM 5-Aza or DMSO, data generated from GSE5816. (E) mRNA and (F) protein levels of SALL4 in H1299 cells after 5 days treatment of DMSO or DAC. (G) Map of CpG islands tested within the first exon of SALL4 locus. (H) Methylation levels of CpG islands within the first exon of SALL4 genome in H1299 cells after indicated treatment. *P<0.05, ***P<0.001, N=3.

Upon data mining from a published gene expression database (38), we noticed that DAC treatment led to significantly increased SALL4 RNA expression levels in H1299 cells (Figure 2D, generated from GSE5816). We then confirmed that SALL4 mRNA and protein expression in H1299 cells was indeed significantly up-regulated upon DAC treatment (Figure 2E and F) along with decreased methylation level at 5’UTR-Exon 1 (Figure 2G and Supplemental Figure 2A and B). Similar results were also observed in the SALL4 negative leukemia cell lines K562 and HL60 (Supplement Figure 2C&D). These data suggested that there is a correlation between upregulation of SALL4 and decreased DNA methylation post HMA treatment. This contrasts with the results of H661 cells (high in endogenous SALL4 expression, low in SALL4 DNA methylation at 5’UTR-Exon 1 CpG region) after DAC treatment, in which no change in SALL4 mRNA (Supplement Figure 2E) or protein (Supplement Figure 2F) expression was observed.

Interestingly, using a Crispr-dCas9 based gene locus-specific demethylation system, we have observed that demethylation of this 5’UTR-Exon 1 region could also lead to upregulation of SALL4 in several SALL4 negative cancer cell lines, suggesting a critical role of DNA methylation in regulating SALL4 expression (39–41).

Entinostat (ENT) represses SALL4 expression post transcriptionally

There is no SALL4 specific drug in the clinic yet, drug developments on SALL4 peptide or siRNA approaches are still at the pre-clinical stage. We then evaluated existing drugs including at various clinical trial stages that could target or affect SALL4 in cancer. We have observed Entinostat (ENT) as a potential SALL4 inhibitory drug, however, its treatment triggered decreased SALL4 protein expression through an unknown mechanism(7). To elucidate the effects of ENT on SALL4, we first examined SALL4 protein expression levels at various time points post treatment. SALL4 protein expression in SALL4+ H661 cells was significantly decreased within 48 hours of ENT treatment at a concentration of 2.5 μM (Figure 3A), along with significantly decreased mRNA expression (Figure 3B and Supplement Figure 3A). Surprisingly, we noticed that SALL4 pre-mRNA level was increased after ENT treatment, in contrast to its mature mRNA (Figure 3C). Similar results were observed in SALL4 high SNU398 cells after ENT treatment (Supplement Figure 3B and C). In addition, we also examined the methylation status at exon-1 before and after DAC treatment. As shown in Supplement Figure 3D, compared to control DMSO treatment, there was no significant change in the average methylation levels after ENT treatment at different time points. Meanwhile, a significant up-regulation was observed in H3K27 acetylation (H3K27ac), which in general correlates with open chromatin and gene activation, at the SALL4 promoter region (Supplement Figure 3E). These results indicate that ENT-induced SALL4 repression was probably at a post transcriptional level.

Figure 3. ENT treatment decreased cell viability and induced cell apoptosis via inhibition of SALL4.

(A) Protein and (B) mRNA levels of SALL4 in H661 cells after the indicated treatment of ENT or DMSO control. (C) Pre-mRNA level of SALL4 in H661 cells after 48 hours treatment of ENT or DMSO control. (D) Cell viability of H661 cells after indicated treatment for 5 days. (E) Flow cytometry of H661 cells after indicated treatment. (F) Percent apoptotic and dead cells in (E). (G) Cell viability of ENT pre-treated H661 cells after transfection of negative control or SALL4 overexpression vector. (H) Flow cytometry of ENT pre-treated H661 cells after indicated treatment. (I) Percent apoptotic and dead cells in (H). *P<0.05, **P<0.01, ***P<0.001, N=3.

Furthermore, the IC₅₀ of lung cancer cell lines for ENT was negatively correlated with their endogenous SALL4 levels (Supplement Figure 3F). ENT treatment of H661 cells led to significantly decreased cell growth (Figure 3D) and increased apoptosis (Figure 3E and F). To confirm the ENT-mediated cellular effects were through SALL4 inhibition, we overexpressed SALL4 in H661 cells before ENT treatment. As shown in Figure 3G–I, SALL4 overexpression rescued the loss of viability and induction of apoptosis triggered by ENT treatment, compared to the control group. Taken together, our data suggested that ENT could suppress cell growth and induce apoptosis in cancer cells by post-transcriptional inhibition of SALL4.

ENT-induced SALL4 inhibition is mediated by microRNA miR-205

As a class of post-transcriptional regulators, miRNAs were considered as the potential mediator in ENT-induced SALL4 inhibition. MiRNA-sequencing was performed using H661 cells after 8 hours of ENT treatment, at a time point much earlier than when SALL4 RNA expression changes were observed (Figure 4A). Using a 2-fold change as a threshold, 38 miRNAs were significantly up-regulated by ENT treatment, and 32 miRNAs were significantly decreased (Supplemental file 2). After overlapping with the Targetscan database, miR-205 was identified as the only potential candidate that was significantly enhanced by ENT treatment and predicted to target SALL4 (Figure 4B). MiR-205 is highly conserved among species. In humans, it is located at the intronic region of the host gene on chromosome 1. Pathway analysis of predicted targets showed that miR-205 was potentially involved in tumor development, cell cycle, and DNA repair (Supplement Figure 4A). The expression level of miR-205 was confirmed to be significantly up-regulated following ENT treatment in H661 cells (Figure 4C). A similar induction of miR-205 following ENT treatment was also observed in SNU398 cells (Supplement Figure 4B). Moreover, as shown in Figure 4D, ENT significantly increased the H3K27ac level at the miR-205 promoter region, which correlated with increased expression of miR-205.

Next, we examined whether miR-205 could target SALL4 to repress its expression. As shown in Figures 4E and F, luciferase reporter assays demonstrated that miR-205 could directly target SALL4 wild-type mRNA and inhibit luciferase activity, while it had no significant effect on mutant constructs lacking a binding site for miR-205. Meanwhile, overexpression of miR-205 in H661 cells significantly repressed SALL4 expression at both mRNA and protein levels compared to control (Figure 4G and H). Immunofluorescent staining of H661 cells after transfection with GFP-tagged scramble or miR-205 plasmid also showed that SALL4 level was significantly lower in cells with miR-205 overexpression (Supplement Figure 4C). Furthermore, similar to ENT treatment, miR-205 overexpression in H661 cells also repressed cell number (Figure 4I) and induced apoptosis (Figure 4J and K), which was rescued by SALL4 overexpression (Figure 4L–N). Inhibition of miR-205 by its antisense oligo (miR-205AS) was sufficient to restore SALL4 expression in the presence of ENT treatment (Supplement Figure 4D and E) and prevented cells from ENT-induced growth repression and apoptosis (Supplement Figure 4F–H). Taken together, these data support the premise that miR-205 was responsible for ENT-mediated SALL4 inhibition.

DAC pre-treatment can prime SALL4 negative cells to become sensitive to ENT in culture

Based on the results shown above that SALL4 could be induced by DAC and targeted by ENT, we further tested whether DAC-primed SALL4 negative cancer cells could become sensitive to ENT. Five days of DAC treatment at a concentration of 1 or 2 μM was performed in both H661 and H1299 cells followed by the evaluation of drug sensitivity to ENT (Figure 5A). As shown in Figure 5B, there was no significant change of the IC₅₀ of ENT for SALL4 positive H661 cells. In contrast, upregulation of SALL4 in originally negative H1299 cells rendered them to be more sensitive to ENT, with a significant decrease of IC₅₀ from 11.77 μM to 1.42 μM (Figure 5C), similar to the IC₅₀ level of H661. Notably, ENT treatment also induced miR-205 expression in DAC-treated H1299 cells (Supplement Figure 5A), suggesting the same mechanism for SALL4 targeting. Moreover, similar results were also observed in SALL4 negative liver cancer SNU387 cell line (Supplement Figure 5B), whereby pretreatment of SNU387 can sensitize these cells to ENT with reduced IC₅₀, and leukemia K562 and HL-60 cell lines, whereby DAC priming followed by ENT treatment in these cells led to significantly decreased cell viability (Supplement Figure 5C and D). The SALL4 protein level was monitored during the sequential treatment and found to be upregulated by DAC treatment in SNU387 and H1299 cells, but not in H661 and SNU398 cells, downregulated the following ENT treatment (Supplement Figure 5E, F, G, and H).

Figure 5. DAC/ENT combination treatment targeted SALL4 negative cancer cells in culture.

(A) Schema of DAC and ENT combination treatment of H1299 or H661 cells. (B) Relative cell viability and IC₅₀ of DMSO or DAC pre-treated H661 cells following ENT therapy. (C) Relative cell viability and IC₅₀ of DMSO or DAC pre-treated H1299 cells following ENT therapy. (D) Schema of experiment design in (E&F). (E) Protein level of SALL4 in shNC or shSALL4 stable expressing H1299 cells after 5 days treatment of DMSO or DAC. (F) Percentage of IC₅₀ against ENT for indicated H1299 cells, IC₅₀ of DMSO treated group was used as normalization control. (G) Schema of experiment design in (H&I). (H) Protein level of SALL4 in H1299 cells after 7, 14, or 21 days of DMSO or DAC treatment. (I) Percentage of IC₅₀ against ENT for indicated time point of H1299 cells, IC₅₀ of DMSO treated group was used as normalization control. n.s. means P>0.05, *P<0.05, **P<0.01, ***P<0.001, N=3.

SALL4 is required for increased sensitivity to ENT treatment

To determine whether upregulation of SALL4 is required for the increased sensitivity to ENT after DAC treatment, we introduced SALL4shRNA into H1299 cells before DAC treatment. Intriguingly, pre-transduction of SALL4 shRNA in H1299 cells blocked the reactivation of SALL4 by DAC treatment and abolished cells’ vulnerability to ENT treatment compared to control groups (Figure 5D–F). This result support that upregulation of SALL4 is responsible for the enhanced drug sensitivity in the combination treatment strategy. To examine the dynamics of the upregulation of SALL4 by DAC, time point studies were carried out. We found that the induction of SALL4 expression by DAC only lasted for 2 weeks, and SALL4 expression was gone at 21 days post DAC treatment (Figure 5G and H). The same result was also identified in SNU387 and K562 cells as shown in Supplement Figure 6A&B. Intriguingly, increased sensitivity to ENT as assayed by drug IC₅₀ was only observed when SALL4 was induced post DAC treatment day 7 and day 14, not at day 21 whereby SALL4 induction was gone (Figure 5I). This result indicated that DAC treatment could only transiently reactivate SALL4 expression in its negative cancer cells, giving a window period for drug combination treatment.

Together, results from our studies on genetic knockdown using shRNA, time point pharmacological SALL4 induction by DAC, support that SALL4 is important for this sequential combination therapy (Figure 5D to I).

DAC pre-treatment can prime SALL4 negative cells to become sensitive to ENT in vivo

We next evaluated whether this sequential drug treatment strategy would work in vivo. H1299 cells were implanted into non-obese diabetic scid gamma (NSG) mice and treated with sequential injections of DAC and ENT both at a dose of 2.5 mg/kg according to published protocols (Figure 6A)(42–44). As shown in Figure 6B, tumors were monitored and measured on day 17, 24, and day 31, and were harvested on day 32 after various treatments. Following DAC and the sequential ENT injection, the sizes of tumors from mice that received both DAC and ENT treatment were significantly smaller than other groups on day 24 and day 31 (Figure 6B and C). At the end point of the study, the size and weight of tumors from mice that received DAC+ENT treatments were remarkably smaller than vehicle-, DAC-, and ENT- treated controls (Figure 6D and E). The body weights of all mice after different treatments exhibited no change (Figure 6F). Intriguingly, DAC alone would not increase the tumor progression compared to the control group (Figure 6B–F). Collectively, the results from cell culture and xenograft tumor models demonstrate that DAC treatment can prime SALL4 negative cancer cells to be targeted by ENT.

Figure 6. DAC/ENT combination treatment targeted SALL4 negative cancer cells in vivo.

(A) Schema of sequential drug treatment in a murine xenograft model. (B) Tumor growth measured at different time points. (C) Tumor volume measured at the end point from the indicated treatment groups. (D) Image of tumors harvested from mice with H1299 cells xenotransplantation after indicated treatment. (E) Tumor weights and (F) body weights from indicated groups. n.s. means P>0.05, *P<0.05, **P<0.01, ***P<0.001.

Discussion

As an important oncofetal gene, SALL4 is silenced in most healthy adult tissues but is reactivated in some tumors and contributes significantly to the survival of tumor cells. These features of SALL4 make it a promising target for cancer treatment. Studies have utilized RNA interference methods or peptides against SALL4 to effectively inhibit tumor growth in cell culture and xenotransplant models (8–10), and drugs that can be used to target SALL4 specifically in cancer patients are being developed. However, as in all targeted therapies, SALL4-centered treatment is limited to patients with tumors that are SALL4 positive, which represent only a sub-set of all cancer patients (11–13).

We have previously observed that exogenous expression of SALL4 in negative liver cancer cells could enhance their sensitivity to oxidative phosphorylation inhibitors due to a shift in metabolism(37), suggesting SALL4 could reprogram cancer cells. In our current studies, we further observed that SALL4 negative cancer cells can become SALL4 dependent. Knocking down SALL4 by shRNA in lung cancer H1299-SALL4 or liver cancer SNU387-SALL4 cells with exogenous SALL4 expression significantly inhibited cell viability and mirrored the phenotypes observed in paired H661 and SNU398 cancer cells with endogenous SALL4 expression. Xenotransplants also confirmed that knocking down SALL4 in isogenic H1299-SALL4 cells led to decreased tumor growth in vivo.

In our previous studies, we have shown that the HDAC inhibitor ENT functions as a SALL4 inhibitor based on drug and gene signatures observed using the connectivity map as an analytical tool(7). In our current study, we have identified a mechanism by which ENT can act as a SALL4 inhibitor drug. We report that ENT negatively regulates SALL4 expression post transcriptionally, at least in part, by epigenetically up-regulating miR-205 expression. Importantly, the impact of ENT on cell viability and tumor growth was dependent on miR-205 expression and SALL4 inhibition. We previously reported that HMA treatment could upregulate SALL4 in acute B cell lymphoblastic leukemia(16). Here we show that treatment with HMA results in hypomethylation of this region and increased SALL4 expression. Indeed, several SALL4 negative cancer cell lines display significantly increased sensitivity to ENT after DAC treatment. The degree of ENT sensitivity in these SALL4 negative cancer cells in which SALL4 re-expression had been pharmacologically engineered by prior HMA therapy was comparable to ENT sensitivity of SALL4 positive cancer cells. Additionally, we have demonstrated that SALL4 is required to sensitize cancer cells to ENT treatment (Figure 5D–I). As these observations are based on cell line studies, additional experiments will be required prior to recommending this as a clinical trial. SALL4 was shown to be transiently reactivated within 3 weeks after DAC treatment (Figure 5H and Supplement Figure 6A&B), and DAC alone treatment did not lead to increased tumor growth in vivo (Figure 6B–F), which eases the concern of long-term safety issues regarding of upregulation of SALL4 by HMA.

This therapeutic effect of DAC plus ENT for SALL4 non-expressing and otherwise ENT insensitive tumors was confirmed in xenograft murine models. This observation supports a prior report that ENT in combination with 5-AZA could repress tumor growth and ablate up to 75% of tumor mass in a murine lung cancer model(45). In clinical phase II studies, combination therapy with low dose 5-azacytidine and ENT were well tolerated in patients with refractory advanced non-small cell lung cancer(46), advanced breast cancer(47), and recurrent metastatic colorectal cancer(48), but limited response or improvement was observed. The limited efficacy in these clinical trials might be due to the lack of biomarkers to select patients. Since our study has demonstrated that only SALL4 negative and not SALL4 positive cancer cells become more vulnerable by combination treatment with DAC and ENT, we predict that it would be possible to improve the efficacy in clinical trials only if SALL4 negative patients enrolled, and we plan to investigate this in cancer patients receiving HMA and ENT in the future.

Altogether, our study has demonstrated that induction/upregulation of SALL4 in negative cancer cells can sensitize these cells to SALL4 targeting drug(s). More specifically, in this study we propose that hypomethylation treatment (HMA) can pharmacologically reactivate/upregulate SALL4 in cancer cells and prime them to a SALL4 inhibitory drug (ENT) treatment. Our studies have also highlighted a SALL4-mediated cancer vulnerability, which can be used as a rationale for sequential combination therapy of DAC and ENT to expand the scope of SALL4-centered cancer therapy.

Statement of Significance:

Our findings provide a therapeutic approach for patients harboring no suitable target by induction of a SALL4-mediated vulnerability.

List of abbreviations

DAC

5-Aza-2’-deoxycytidine

ENT

Entinostat

SALL

spalt-like

MDS

myelodysplastic syndrome

AML

acute myeloid leukemia

DNMTs

DNA methyltransferases

HMA

hypomethylation agents

NuRD

Nucleosome Remodeling Deacetylase

HDACs

histone deacetylases

miRNAs

microRNAs

3’UTR

3’untranslated region

DepMap

Dependency Map

CMML

chronic myelomonocytic leukemia

H3K27ac

H3K27 acetylation

NSG

non-obese diabetic scid gamma