SALL4, the missing link between stem cells, development and cancer

Abstract

There is a growing body of evidence supporting that cancer cells share many similarities with embryonic stem cells (ESCs). For example, aggressive cancers and ESCs share a common gene expression signature that includes hundreds of genes. Since ESC genes are not present in most adult tissues, they could be ideal candidate targets for cancer-specific diagnosis and treatment. This is an exciting cancer-targeting model. The major hurdle to test this model is to identify the key factors/pathway(s) within ESCs that are responsible for the cancer phenotype. SALL4 is one of few genes that can establish this link. The first publication of SALL4 is on its mutation in a human inherited disorder with multiple developmental defects. Since then, over 300 papers have been published on various aspects of this gene in stem cells, development, and cancers. This review aims to summarize our current knowledge of SALL4, including a SALL4-based approach to classify and target cancers. Many questions about this important gene still remain unanswered, specifically, on how this gene regulates cell fates at a molecular level. Understanding SALL4’s molecular functions will allow development of specific targeted approaches in the future.

1. Introduction

SALL4, a member of the spalt-like (SALL) gene family (SALL1 to SALL4), was originally cloned based on its DNA sequence homology to the homeotic gene in Drosophila, spalt (sal) ^1,2,3. SALL4 plays an essential role in maintaining the pluripotent and self-renewal properties of embryonic stem cells (ESCs). After birth, SALL4 expression is down-regulated and absent in most adult tissues. However, SALL4 is re-expressed in various cancers. This review summarizes our current knowledge about the structure and expression patterns of SALL4, its role in stem and cancer cells, and the existing understanding of its molecular functions.

1.1. SALL4 Gene and Protein Structure

In humans, SALL4 is located in 20q13.2, consisting of four coding exons and 3,162 bp of coding sequence (SALL4A; AY172738)^4,5. GenBank sequences AY170621 (SALL4B mRNA, 1851bp) and AY170622 (SALL4C mRNA, 831bp) suggest that two alternative splicing products exist in addition to the full length SALL4A mRNA. The B mRNA consists of exon 1, 1,020 bp of 5’ end of exon 2, exons 3 and 4, whereas in the C mRNA, exon 2 is spliced out ⁶ (Figure 1). Another SALL4 transcript containing different exons 1a and 1b instead of the original exon 1 identified by 5’ RACE has also been reported⁷.

Figure 1. Human SALL4 gene structure and isoforms.

Alternative splicing generates 3 forms of SALL4 mRNA. SALL4A (encoding 1053 amino acids) has all 4 exons and contains 4 zinc finger domains (ZF1–4, indicated by blue bars). SALL4B (encoding 617 a.a.) is generated via an alternative splice donor site that results in the deletion of large portion of exon 2 and a protein that contains only ZF1 and ZF4. SALL4C (encoding 277 a.a.) does not have exon 2 and encodes a version that only has ZF1. A conserved N-terminal 12 amino acid (N-12aa) domain of SALL4 (indicated by red bars) is sufficient for recruiting the nucleosome remodeling and histone deacetylase (NuRD) complex and mediating transcriptional repression. A glutamine (Q)-rich region (indicated by yellow bars) that is highly conserved in all invertebrate and vertebrate Spalt family members, was shown to be necessary in protein interactions between Spalt family proteins. A conserved nuclear localization signal (NLS) has been reported and is indicated by green bars.

Like the protein encoded by sal, SALL4 belongs to a group of transcription factors characterized by multiple cys(2)his(2) (C2H2)-type zinc finger domains distributed over the entire protein. Human SALL4 contains a single C2H2 zinc finger near the N-terminus, two C2H2 zinc finger clusters in the middle portion, and one at the C-terminus of the protein⁸ (Figure 1). SALL4A has all four zinc finger clusters while SALL4B lacks the zinc finger clusters ZF2 and ZF3. The most C-terminal ZF4 cluster of Sall4 is both necessary and sufficient for its localization to the heterochromatin⁹. In addition to zinc fingers, SALL4 also contains a glutamine (Q)-rich region that is highly conserved in all invertebrate and vertebrate SALL family members. This domain was shown to be necessary for interactions between Spalt family proteins¹⁰. Furthermore, SALL4 A and B isoforms are able to form homodimers and heterodimers¹¹, possibly through their Q-rich regions. SALL4 is a nuclear protein and its subcellular localization is mediated through at least one conserved nuclear localization signal (NLS) at amino acids (AA) 64–67. A single mutation that changes lysine 64 into arginine (K64 into R64) is sufficient to disrupt its subcellular distribution and compromises its function in vivo¹².

Post-translational modifications of SALL4 such as phosphorylation¹³, ubiquitination¹⁴ and sumoylation¹⁵ have been reported. The functions of these post-translational modifications are largely unknown. One study reported that lysine residues 156, 316, 374, and 401 were essential for SALL4B sumoylation, which seemed to affect the protein stability, its interaction with Oct4, and its transactivation function¹⁵. The role of ubiquitination of SALL4 is still unknown.

1.2. SALL4 Expression in Development

In murine development, Sall4 protein expression is first observed at the two-cell stage due to maternal contribution, and later in some cells of the 8- to 16-cell-stage mouse embryo after zygotic transcription has initiated¹⁶. In late blastocysts, the Sall4 RNA and Sall4 protein become enriched in the inner cell mass (ICM) and the trophectoderm. Finally, by E11.5, Sall4 expression is observed in the midbrain, the rostral edge of the forebrain, maxillary arch, genital tubercle, limb buds, tail, and left ventricular myocardium^17,18. In adult mice, Sall4 expression is mostly restricted to germ cells, wherein it is highly expressed in undifferentiated spermatogonia and oocytes in primordial, primary, and secondary follicles^19,20,21. Similarly, expression of SALL4 in adult human tissue is restricted to the testis and ovary⁴ (Figure 2). One exception to this expression pattern is human CD34+ hematopoietic stem/progenitor cells (HSPCs)²².

Figure 2. SALL4 expression during development.

Sall4 protein expression is first observed at the two-cell stage, and later in some cells of the 8- to 16-cell-stages of the mouse embryo. In late blastocysts, Sall4 protein becomes enriched in the inner cell mass (ICM) and the trophectoderm. By E11.5, Sall4 expression is observed in the midbrain, genital tubercle, limb and tail buds^17,18. In adult mice, Sall4 expression is mostly restricted to the germ cells, wherein it is highly expressed in undifferentiated spermatogonia and oocytes in primordial, primary, and secondary follicles^19,20,21. The red circles represent SALL4 expression.

SALL4 has isoform-specific gene expression pattern in the testis, ESCs, and fetal liver cells. Sall4a is expressed in postnatal day 7 (PND7) testis, while Sall4b is expressed from PND0 onward ²³. Sall4a is more abundant than Sall4b in undifferentiated mouse ESCs, and neither isoform is found upon induction of ESC differentiation¹¹. Furthermore, the A isoform, and not B, is detected in fetal livers, while several human hepatocellular carcinoma (HCC) cell lines express both SALL4A and SALL4B²⁴.

1.3. SALL4 Gene Regulation

Very little is known about the expression regulation of SALL4 and its isoforms. Several studies have focused on the promoter region of this gene^7,25,26. While one study reported on SALL4’s intronic enhancer²⁷, its distal regulatory element(s) remains uninvestigated. In breast cancer, SALL4 was proposed to be a downstream target gene of STAT3²⁵. During intestinal metaplasia of the gastric epithelial cells, SALL4 was identified as a direct target of Caudal-related homeobox 1 (CDX1) protein²⁸. In murine ESC, several reports have demonstrated that Sall4 protein participates in an interconnected autoregulatory circuit with Oct4, Sox2, and Nanog, wherein each of the four factors can regulate its own expression as well as that of others^29,30,31. Within this circuit, one study reported that Sall4 can negatively self-regulate and antagonize Oct4’s activation function to balance its own expression level²⁶. Further, a TCF/LEF consensus sequence was reported in the SALL4 promoter region and SALL4 expression could be activated by LEF1 or TCF4E, indicating that SALL4 is a target of the canonical WNT signaling⁷. Posttranscriptional regulation of SALL4 has also been observed. Specifically in glioma, one study reported that miR-107 can bind the 3’UTR of SALL4 mRNA and modulate its expression. In a metastatic breast cancer model³², SALL4 was found to be a target of miR-33b. In murine ESC, miR-294 and let-7c were reported to have opposing effects on Sall4 expression³³.

Epigenetic regulation of SALL4 has also been observed. Specifically, the DNA methylation status of SALL4 seems to correlate with its expression. In fact, SALL4 is hypomethylated and highly expressed in induced pluripotent stem (iPS) cells and ESCs, comparing to differentiated cells (Figure 3)^34,35. While one study focused on the methylation status of SALL4 promoter region³⁶, the major CpG island of SALL4 is located at the Exon 1/Intron 1 region as reported by other SALL4 methylation studies^37,38. Future experiments are thus needed to explore the effects of upstream regulator(s) on SALL4 DNA methylation status and how this epigenetic status affects cell fate.

Figure 3. SALL4 CpG island.

SALL4 DNA methylation changes at the single CpG level during reprogramming of K562 cells to leukemia-derived pluripotent stem cells (LiPS) cells. Results from K562 and two separate LiPS lines are shown³⁵.

2. Function of SALL4 in Stem Cells and Development

Sall4 has been shown to be involved in the proper development of the ICM and Sall4-null mice do not survive beyond embryonic day E6.5^16,9,30. Although Sall4-null blastocysts have no defects in lineage commitment of the ICM or trophectoderm in vivo, Sall4-null ICM cells show defected proliferation in vitro⁹. Down-regulation of Sall4 in ESCs via shRNA leads to reduced Pou5f1 expression and increased Cdx2 expression, which results in an expansion of trophectoderm cells³⁰. When cultured with mouse embryonic fibroblast (MEF) feeder cells, one study reported that Sall4 is essential for efficient proliferation and stabilization, but not for pluripotency, of ESCs by repressing trophectoderm gene expression³⁹. Another study, however, showed that Sall4 is essential for the pluripotency of ESCs, especially under feeder-free conditions³⁰. Additional reports suggested that Sall4 is important for ES cell fate^{26, 27,30, 31, 40, 41, 42}. Sall4 was also reported to play a positive role in the generation of iPS cells from fibroblasts⁴³. SALL4 expression needs to be reactivated during the reprogramming process by which MEF cells are converted to iPS cells⁴⁴. Furthermore, SALL4 was used to enhance iPS cell reprogramming efficiency in a cell fusion approach⁴⁵, and inclusion of Sall4 in the Oct4, Sox2, and Klf4 (OSK) reprogramming system leads a more consistent and stable induction of iPS cells with a higher efficiency⁴⁶. Interestingly, ectopic expression of the combination of Sall4, Nanog, Esrrb, and Lin28 (SNEL), or the combination of Sall4, Sall1, Utf1, Nanog and Myc in MEFs can reprogram them into iPSCs. These results indicate that Sall4, in combination with other factors, can substitute for OSKM factors^47,48. Furthermore, SALL4 can interact with the Nucleosome Remodeling and Histone Deacetylation (NuRD) complex^39,49. MBD3, an essential structural component of the NuRD complex, has been shown to be a key regulator in the reprogramming process ^50,51,52,53. Though it has been suggested that the SALL4/NuRD interaction may not be required in iPSC reprogramming in a cell fusion model⁴⁵, it remains to be seen whether the same conclusion can be drawn from other approaches in generating iPSCs.

Mutations in the SALL4 gene have been shown to cause autosomal dominant diseases such as Okihiro/Duane-radial ray syndrome (DRRS; MIM 607323)^4,54, acro-renal-ocular syndrome (AROS; MIM 102490), Instituto Venezolano de Investigaciones Cientificas syndrome (IVIC)⁵⁵, and are suspected to cause thalidomide embryopathy^56,57. The phenotype of Okihiro/Duane-radial ray syndrome has overlap with several other syndromes such as Holt Oram syndrome (HOS; MIM 142900)⁵⁸, which is caused by TBX5 mutations and Townes-Brocks syndrome, which is caused by SALL1 mutations (TBS; MIM 107480)⁵⁹.

Okihiro syndrome refers to the familial occurrence of radial-sided hand malformations in association with Duane, a congenital eye movement disorder. The range of associated features among the described families is broad, including anal stenosis, atrial septal defect, hearing impairment, pigmentary disturbance, renal abnormalities, external ear malformations, and facial asymmetry in the affected individuals. Significant intrafamilial variability is a feature of the reported pedigrees⁴. In addition to these developmental defects, SALL4 was reported as a potential candidate gene that mutated in Han Chinese women with premature ovarian failures⁶⁰.

Homozygous mutations within the SALL4 gene have not been reported in humans, possibly because this would result in early embryonic lethality. All reported SALL4 mutations are heterozygous and are located in exons 2 and 3, with no mutations found in exons 1 or 4. These are nonsense mutations, short duplications, or deletions. Two cases of missense SALL4 mutations around the coding region for the ZF4 cluster were also reported^61,62.

In mice, Sall4 has been reported to be involved in neural tube closure, limb and heart development^39,63,9,18. Heterozygous mice carrying a targeted Sall4 allele with deletion of exons 2–3 that contain all ZF domains exhibit embryonic exencephaly, anorectal malformations, and heart defects⁹. However, in this particular study, limb defects were not described. In a different report, heterozygous mice with a targeted SALL4 allele that lacks exons 2–4 exhibit deafness, lower anogenital tract abnormalities, anencephaly, Hirschsprung’s disease, renal hypoplasia, and skeletal defects⁶³. However, heart anomalies were not described in this model. Mice with a Sall4 gene trap allele with an insertion in exon 2 exhibit variable embryonic phenotypes ranging from ventricular septum defect (VSD) to mild limb abnormalities¹⁸.

3. SALL4 in cancers

Analyses of its expression and epigenetic status have shown that SALL4 is de-regulated and aberrantly expressed in various cancers (Review in⁶⁴) such as leukemia⁶⁵, germ cell tumors^66,67, hepatocellular carcinoma (HCC)^24,68, gastric cancer^20,69–73, colorectal carcinoma^74–76, esophageal squamous cell carcinoma^77,78, breast cancer^79–81, endometrial cancer ^82,83, lung cancer^84–86 and glioma⁸⁷. Functionally, SALL4 is important for the survival^{38,80,82,88–93}, drug resistance^82,94 and metastasis^69,82 of many of these cancer cells (Figure 4). The main challenge in this area is the need for standardization of the SALL4-based diagnostic methods. While some reports focus on the RNA expression level of SALL4, others focus on the protein expression level. Frequently, there is a discrepancy between these two detection methods. Even for SALL4 protein immunohistochemistry (IHC) studies, various antibodies and different diagnostic criteria were used. Both percentage of positive cells and intensity of staining, as well as the staining pattern, such as nuclear versus cytoplasmic, should be considered. Quality control of SALL4 antibodies should be evaluated before their use in IHC experiments. Cell type-specific genes, in addition to SALL4, should be examined in various SALL4-positive cancers. Standardization of SALL4-based diagnosis is important for future development of SALL4-specific approaches to classify and target cancers. Below, we have detailed selected studies of SALL4’s mis-regulation in cancer.

Figure 4. SALL4 expression in cancer.

SALL4 is de-regulated and aberrantly expressed in various cancers.

In hematological malignancies, high level of SALL4 expression is correlated with high-risk myelodysplastic syndromes (MDS) patients with poor survival^95,96. Consistent with this observation, aberrant hypomethylation of SALL4 gene in MDS patients has also been reported^37,97. Increased SALL4 expression was detected in patient samples from blastic stage of chronic myeloid leukemia (CML), compared to the chronic phase, CML patients who have achieved complete remission (CR) after chemotherapy groups⁹⁰, or those who have tyrosine kinase inhibitor (TKI) resistance⁹⁸. SALL4 is constitutively expressed in most human primary acute myeloid leukemia (AML) and myeloid leukemia cell lines⁶⁵. Further supporting this, transgenic mice over-expressing SALL4B develop hematopoietic disorders, including MDS-like symptoms and AML⁶⁵. SALL4 is also abnormally expressed in B cell acute lymphocytic leukemia (B-ALL) with hypomethylation at the previously described CpG island^38,99. SALL4 expression is the highest in B-ALL patients with TEL-AML1 translocation, which is the most common genetic abnormality in pediatric B-ALL. In contrast, SALL4 is negative in T cell acute lymphocytic leukemia (T-ALL), and most lymphomas⁹⁹ with the exception ALK-positive anaplastic large cell lymphoma (ALK+ ALCL)¹⁰⁰.

In solid tumors, SALL4 was first studied in germ cell tumors, where it is a sensitive diagnostic marker for primary central nerve system (CNS) germ cell tumors⁶⁶, testicular germ cell tumors⁶⁷, primary extragonadal germ cell tumors^101,102, mediastinal seminoma^103,104, ovarian malignant germ cell tumor¹⁰⁵, and metastatic germ cell tumors¹⁰⁶. In yolk sac tumors and primary extragonadal germ cell tumors, one study reported that the circulating tumor cells are detectable by anti-SALL4 antibody in peripheral blood samples, making this protein a more sensitive marker than α-fetoprotein and glypican-3¹⁰⁷. In addition, SALL4 is expressed in a majority of malignant rhabdoid tumors (MRTs) ^{20,108,109,110,111} but rarely expressed in epithelioid sarcomas. Therefore, SALL4 immunoexpression may be a useful diagnostic tool to distinguish epitheloid sarcoma from malignant rhabdoid tumor.

For tumors in the gastrointestinal & digestive system, SALL4 is well studied in HCC. SALL4 is expressed in human fetal livers and absent in adult livers^24,68, but it is re-expressed in a subgroup of HCC patients. SALL4-positive HCC cells have a gene expression pattern that is similar to that of fetal hepatic progenitor cells. These HCCs tend to be more aggressive and associated with a poor prognosis⁶⁸. Compared to SALL4-negative cases, HCC patients with SALL4 expression show higher frequency of tumor cell invasion, intrahepatic metastasis, and are positive for cytokeratin 19 and Epithelial Cell Adhesion Molecule (EpCAM), as well as other cancer stem cell markers¹¹². A positive correlation between SALL4 and EpCAM expression have been reported in other HCC studies ^{91,113,114,115}. Liver cancer patients with high serum level of SALL4 have poor prognosis as evidenced by both tumor recurrence and overall survival rate¹¹⁶.

4. Molecular dissection of SALL4 functions in cells

To understand the mechanism(s) underlying SALL4 function(s) in various cells, efforts have been made to search for both SALL4 protein interaction partners and its downstream targets. The PTEN/AKT pathway is the best-understood example that integrates our current understanding of these two aspects of SALL4’s molecular functions. Based on chromatin immunoprecipitation followed by microarray (ChIP-chip) data in normal human CD34+ cells, leukemic NB4 cells, and human ESCs, PTEN was identified to be a key SALL4 downstream target gene^49,92. SALL4 directly interacts with the NuRD complex, and recruits it to the promoter region of PTEN⁴⁹. A peptide that competes with SALL4 in its interaction with the NuRD complex can reverse its effect on PTEN repression, which leads to the activation of the AKT pathway^88,93. Here we have summarized some of the known interacting partners of SALL4 and its downstream target genes.

4.1. Protein-partners of SALL4

It has been shown that SALL4 can interact with many proteins, which can define its molecular function. Specifically, SALL4 can bind the NuRD complex, LSD1, or DNMTs, thus functions as a transcriptional repressor. SALL4 can also form a protein complex with transcriptional regulators such as Mixed Lineage Leukemia (MLL) or β-catenin to activate genes. Furthermore, Sall4 interacts with Oct4, Nanog, or Sox2 to maintain “stemness” of ESCs. There are further reports of SALL4 interaction with TBX5 in cardiac tissues¹⁸, with PLZF in spermatogonial progenitors in postnatal testes¹¹⁷, and with Rad50 to stabilize the Mre11–Rad50–Nbs1 complex during the DNA damage repair process¹¹⁸. We have detailed some of these interactions below.

A conserved N-terminal 12 amino acid (N-12aa) domain of SALL4 that is also present in other SALL family members can recruit the NuRD complex and mediate transcriptional repression^119,120. The NuRD complex comprises many different subunits, the most frequently associated ones are histone deacetylases HDAC1/2, ATP-dependent remodeling enzymes CHD3/4, histone chaperones RbAp46/48, CpG-binding proteins MBD2/3, the GATAD2a (p66a) and/or GATAD2b (p66b), and sequence-specific DNA-binding proteins MTA1/2/3^121–123. SALL4’s recruitment of the NuRD complex was validated by studies of a SALL4 target gene, PTEN, which encodes an important tumor repressor that regulates the AKT signaling pathway and cell growth. SALL4 co-occupies the PTEN promoter region with a NuRD component, HDAC2, to repress its gene expression⁴⁹. Treatment of SALL4-expressing malignant cells with a synthesized N-12aa peptide to compete for SALL4’s interaction with NuRD leads to cell death due to PTEN de-repression, which can be rescued by co-treatment of cells with a PTEN inhibitor^88,93.

SALL4’s interaction with NuRD may mediate its binding to the NuRD-associated LSD1 protein, a histone demethylase that specifically converts histone H3K4me2 to H3K4me1 or unmethylated H3K4, thus functioning as a transcriptional co-repressor ^124,125. Both WT SALL4A and SALL4B proteins can co-immunoprecipitate (IP) with LSD1¹²⁶. Based on serial truncations of SALL4 in co-IP and demethylation experiments, the region of SALL4 containing the Q-rich domain is essential for its binding to LSD1. Both gain- and loss-of-function approaches revealed that SALL4 controls the histone binding levels of LSD1, which is accompanied by changes in the methylation status of H3K4 at promoter regions of genes that they co-regulate¹²⁶. In addition to its function in modulating the covalent modification of histones, there is also evidence that SALL4 may directly interact with DNA methyltransferases (DNMTs) to mediate gene repression³⁶. Furthermore, the C-terminal portion of SALL4 can bind Cyclin D1, which together with SALL4, can associate with heterochromatic region and modulate gene repression¹²⁷.

Through interacting with other proteins, SALL4 can also activate gene expression. For example, SALL4 recruits the MLL complex, which contains histone methyltransferase activity for K4 and K79 on the histone 3 tail. In the context of AML, this interaction results in HOXA9 up-regulation and subsequently more aggressive leukemogenesis⁸⁹. Members of the MLL complex, including RbBp5 and Menin, were also found in SALL4 pull-down experiments⁸⁹. In addition, MLL complex and SALL4 are known to mediate signaling from the Wnt/β-catenin pathway in AML development. Wnt is a highly conserved pathway that plays an important role in development and cancer¹²⁸. In amphibian neural development, Sall4 is known to be activated by Wnt signaling to specify posterior neural fate¹²⁹. Furthermore, both SALL4A and SALL4B isoforms can bind β-catenin protein and synergistically activate the Wnt/β-catenin signaling pathway⁶⁵. The interaction domain between SALL4 and β-catenin was not identified. However, based on binding data between SALL1 and β-catenin, the C-terminal portion of SALL4 is likely involved in this interaction¹³⁰. Elucidating the relationship among SALL4, MLL, and Wnt signaling can lead to novel targeting opportunities in AML treatment.

Many of the interactions described above between SALL4 and its co-regulators were observed in the context of leukemogenesis or cancer development. Due to its key roles in ESCs, SALL4’s interactions with transcription factors crucial for pluripotency have also been explored. Two independent studies, using unbiased approaches of affinity purification of Oct4 followed by mass spectrometry, identified Sall4 to be one of its binding partners in mouse ESCs^131,132. Both approaches identified components of the NuRD complex, raising the possibility that Sall4 may bridge the interaction between Oct4 and this epigenetic repressor complex. Since Oct4, Sox2, and Nanog form a central pluripotency transcription factor network in ESCs, Sall4’s interaction with these proteins was also investigated. Through affinity precipitation using oligonucleotides containing OCT-SOX elements, Sall4 was found to be specifically associated with Oct4/Sox2/DNA complex in mouse ESCs. Though it is unclear which part of Sall4 is responsible for binding to Sox2 or whether the interaction is direct, this association is found at the promoter of a subset of pluripotent genes⁴². Another pluripotency factor, Nanog, also seem to be present in the same protein complex with Sall4 in ESCs, with the N-terminal portion of Sall4 important for this interaction²⁷. As is the case in regards to Sox2, it is unclear whether the interaction between Nanog and Sall4 is direct. Moreover, a recent study suggested that SALL4 can recruit polycomb complex members such as Bmi-1 to the Sox gene locus, thus facilitating changes in higher-order chromatin structure¹³³. These studies further support that Sall4 is part of the important transcription regulatory complex that is responsible for maintaining the pluripotency of mouse ESCs.

SALL4 may play a role in the DNA damage response. Sall4 is required for efficient activation of ATM-dependent responses to DNA double stranded breaks (DSBs) damage in ESCs. Sall4 was observed to rapidly mobilize to the sites of DSBs after DNA damage. Identified by mass-spec and confirmed by co-IP, Sall4 interacts with Rad50 and stabilizes the Mre11–Rad50–Nbs1 complex for the efficient recruitment and activation of ATM. Further, Sall4 interacts with Baf60a, a member of the SWI/SNF (switch/sucrose nonfermentable) ATP-dependent chromatin-remodeling complex, which may be responsible for recruiting Sall4 to the site of DNA DSB¹¹⁸.

4.2 Downstream targets and pathways

The DNA binding ability of SALL4 remains to be resolved. Genome-wide analyses such as ChIP followed by next-generation sequencing (seq) or ChIP-chip have suggested several potential binding motifs for SALL4. The common SALL4 DNA binding motif suggested by several groups^11,134,135 is “ATTTGCAT”, which appears similar to the consensus motif utilized by multiple factors such as Nanog, Oct4, and Sox2 ^135–139. A study on differential DNA binding potential of SALL4 isoforms¹¹ suggested that another motif, “TTGTCTACTTGGTA“, exists for Sall4A, and was predominantly found in differentiation and patterning genes. A different study reported Sall4 ChIP-chip data that revealed an enrichment of the sequence “TCGCCATA” in Sall4 peaks with the highest enrichment³¹. Further studies are needed to verify which motif is important using known SALL4 downstream target genes.

Many critical downstream targets of SALL4 have been identified. SALL4 can bind its own promoter region and exerts a strong self-repressive auto-regulation, which in turn acts as a “gate keeper” for the SALL4/OCT4 positive feedback loop²⁶. SALL4 can also regulate other ESC genes such as Nanog, Pou5f1, and SOX2. In addition, in normal human hematopoietic and leukemic cells, SALL4 can up-regulate the expression of BMI-1^95,98,140. HOXA9^22,89, and down-regulate PTEN expression⁴⁹. Being a survival factor for cancer cells, SALL4 can affect apoptosis-inducing genes (e.g., TNF, TP53, PTEN, CARD9, CARD11, CYCS, LTA) and apoptosis-inhibiting genes (e.g., BMI-1, BCL2, XIAP, DAD1, TEGT)⁹². MYC has been shown to be a SALL4 target both in cancer (endometrial cancer) and ESCs^29,82,92. SALL4 is enriched in side population (SP) cells and can regulate the expression of ATP-binding cassette multidrug transporter, ABCA3 and ABCB1⁹⁴. Genome-wide analysis revealed that SALL4 regulates the Sonic Hedgehog (SHH) pathway²⁹. Sall4 and Gli3 cooperate for proper development of the anterior-proximal skeletal elements and also function upstream of Sonic hedgehog (Shh)-dependent posterior skeletal element development¹⁴¹. Finally, SALL4 can regulate the expression of cell surface molecules such as EpCAM ^91,113, epithelial-mesenchymal transition (EMT) related factors (e.g. SNAI1¹¹³, CXCR4, TWIST1⁹¹, E-Cadherin^69,83,142, VIM, and ZEB1¹⁴³) in cancers (Figure 5).

Figure 5. Summary of regulation and downstream targets of SALL4.

Many critical interacting partners, upstream regulators, as well as downstream targets of SALL4 are summarized here. The red arrow indicates activation and the blue bar indicates repression.

5. Conclusion remarks

The fate of a specific cell type is governed by genetic and epigenetic factors, and can be redirected or reprogrammed through modifications of the key factors involved. Identifying and understanding the key players in cell fate determination will help us direct a cell type-specific gene expression profile, which is ultimately responsible for cellular identity and function. The knowledge thus gained can help us direct cellular differentiation for the purpose of tissue regeneration and repair, or control malignant cell growth. The zinc-finger protein SALL4 has a unique role in connecting ESCs, development, and cancer. SALL4 affects its downstream target gene expression by interacting with various epigenetic modulators and sequence-specific transcription factors, thus making it a key cell fate regulator. Additional studies are needed to shed light on the molecular mechanism(s) of SALL4 functions in various cell types in order to open avenues for new therapeutic approaches to target cancers as well as generate stem-cell related applications for tissue regeneration.

Highlights.

This review summarizes the gene structure, expression and function(s) of SALL4.

SALL4 plays an essential role in embryonic stem cells.

Mutations in the SALL4 gene causes Okihiro/Duane-radial ray syndrome.

SALL4 is de-regulated and aberrantly expressed in various cancers.

SALL4 can be a cancer-specific target.