Combined microarray analysis uncovers self-renewal related signaling in mouse embryonic stem cells

Gene-Environment Interactions: My Unique Journey

I am deeply honored to be invited to write this scientific autobiography. As a physician-scientist, pediatrician, molecular biologist, and geneticist, I have authored/coauthored more than 600 publications in the fields of clinical medicine, biochemistry, biophysics, pharmacology, drug metabolism, toxicology, molecular biology, cancer, standardized gene nomenclature, developmental toxicology and teratogenesis, mouse genetics, human genetics, and evolutionary genomics. Looking back, I think my career can be divided into four distinct research areas, which I summarize mostly chronologically in this article: (a) discovery and characterization of the AHR/CYP1 axis, (b) pharmacogenomics and genetic prediction of response to drugs and other environmental toxicants, (c) standardized drug-metabolizing gene nomenclature based on evolutionary divergence, and (d) discovery and characterization of the SLC39A8 gene encoding the ZIP8 metal cation influx transporter. Collectively, all four topics embrace gene-environment interactions, hence the title of my autobiography.

Divalent Metal Uptake and the Role of ZIP8 in Host Defense Against Pathogens

Manganese (Mn) and Zinc (Zn) are essential micronutrients whose concentration and location within cells are tightly regulated at the onset of infection. Two families of Zn transporters (ZIPs and ZnTs) are largely responsible for regulation of cytosolic Zn levels and to a certain extent, Mn levels, although much less is known regarding Mn. The capacity of pathogens to persevere also depends on access to micronutrients, yet a fundamental gap in knowledge remains regarding the importance of metal exchange at the host interface, often referred to as nutritional immunity. ZIP8, one of 14 ZIPs, is a pivotal importer of both Zn and Mn, yet much remains to be known. Dietary Zn deficiency is common and commonly occurring polymorphic variants of ZIP8 that decrease cellular metal uptake (Zn and Mn), are associated with increased susceptibility to infection. Strikingly, ZIP8 is the only Zn transporter that is highly induced following bacterial exposure in key immune cells involved with host defense against leading pathogens. We postulate that mobilization of Zn and Mn into key cells orchestrates the innate immune response through regulation of fundamental defense mechanisms that include phagocytosis, signal transduction, and production of soluble host defense factors including cytokines and chemokines. New evidence also suggests that host metal uptake may have long-term consequences by influencing the adaptive immune response. Given that activation of ZIP8 expression by pathogens has been shown to influence parenchymal, myeloid, and lymphoid cells, the impact applies to all mucosal surfaces and tissue compartments that are vulnerable to infection. We also predict that perturbations in metal homeostasis, either genetic- or dietary-induced, has the potential to impact bacterial communities in the host thereby adversely impacting microbiome composition. This review will focus on Zn and Mn transport via ZIP8, and how this vital metal transporter serves as a "go to" conductor of metal uptake that bolsters host defense against pathogens. We will also leverage past studies to underscore areas for future research to better understand the Zn-, Mn- and ZIP8-dependent host response to infection to foster new micronutrient-based intervention strategies to improve our ability to prevent or treat commonly occurring infectious disease.

SLC39A8 gene encoding a metal ion transporter: discovery and bench to bedside

SLC39A8 is an evolutionarily highly conserved gene that encodes the ZIP8 metal cation transporter in all vertebrates. SLC39A8 is ubiquitously expressed, including pluripotent embryonic stem cells; SLC39A8 expression occurs in every cell type examined. Uptake of ZIP8-mediated Mn2+, Zn2+, Fe2+, Se4+, and Co2+ represents endogenous functions-moving these cations into the cell. By way of mouse genetic differences, the phenotype of "subcutaneous cadmium-induced testicular necrosis" was assigned to the Cdm locus in the 1970s. This led to identification of the mouse Slc39a8 gene, its most closely related Slc39a14 gene, and creation of Slc39a8-overexpressing, Slc39a8(neo/neo) knockdown, and cell type-specific conditional knockout mouse lines; the Slc39a8(-/-) global knockout mouse is early-embryolethal. Slc39a8(neo/neo) hypomorphs die between gestational day 16.5 and postnatal day 1-exhibiting severe anemia, dysregulated hematopoiesis, hypoplastic spleen, dysorganogenesis, stunted growth, and hypomorphic limbs. Not surprisingly, genome-wide association studies subsequently revealed human SLC39A8-deficiency variants exhibiting striking pleiotropy-defects correlated with clinical disorders in virtually every organ, tissue, and cell-type: numerous developmental and congenital disorders, the immune system, cardiovascular system, kidney, lung, liver, coagulation system, central nervous system, musculoskeletal system, eye, and gastrointestinal tract. Traits with which SLC39A8-deficiency variants are currently associated include Mn2+-deficient hypoglycosylation; numerous birth defects; Leigh syndrome-like mitochondrial redox deficiency; decreased serum high-density lipoprotein-cholesterol levels; increased body mass index; greater risk of coronary artery disease, hypotension, cardiovascular death, allergy, ischemic stroke, schizophrenia, Parkinson disease, inflammatory bowel disease, Crohn disease, myopia, and adolescent idiopathic scoliosis; systemic lupus erythematosus with primary Sjögren syndrome; decreased height; and inadvertent participation in the inflammatory progression of osteoarthritis.

Autophagy and mTOR pathways in mouse embryonic stem cell, lung cancer and somatic fibroblast cell lines

Embryonic developmental stages and regulations have always been one of the most intriguing aspects of science. Since the cancer stem cell discovery, striking for cancer development and recurrence, embryonic stem cells and control mechanisms, as well as cancer cells and cancer stem cell control mechanisms become important research materials. It is necessary to reveal the similarities and differences between somatic and cancer cells which are formed of embryonic stem cells divisions and determinations. For this purpose, mouse embryonic stem cells (mESCs), mouse skin fibroblast cells (MSFs) and mouse lung squamous cancer cells (SqLCCs) were grown in vitro and the differences between these three cell lines signalling regulations of mechanistic target of rapamycin (mTOR) and autophagic pathways were demonstrated by immunofluorescence and real-time polymerase chain reaction. Expressional differences were clearly shown between embryonic, cancer and somatic cells that mESCs displayed higher expressional level of Atg10, Hdac1 and Cln3 which are related with autophagic regulation and Hsp4, Prkca, Rhoa and ribosomal S6 genes related with mTOR activity. LC3 and mTOR protein levels were lower in mESCs than MSFs. Thus, the mechanisms of embryonic stem cell regulation results in the formation of somatic tissues whereas that these cells may be the causative agents of cancer in any deterioration.

Learning Differential Module Networks Across Multiple Experimental Conditions

Module network inference is a statistical method to reconstruct gene regulatory networks, which uses probabilistic graphical models to learn modules of coregulated genes and their upstream regulatory programs from genome-wide gene expression and other omics data. Here, we review the basic theory of module network inference, present protocols for common gene regulatory network reconstruction scenarios based on the Lemon-Tree software, and show, using human gene expression data, how the software can also be applied to learn differential module networks across multiple experimental conditions.

In utero gene expression in the Slc39a8(neo/neo) knockdown mouse

Slc39a8 encodes ZIP8, a divalent cation/bicarbonate symporter expressed in pluripotent mouse embryonic stem cells, and therefore ubiquitous in adult tissues; ZIP8 influxes Zn²⁺, Mn²⁺ and Fe²⁺. Slc39a8(neo/neo) knockdown mice exhibit 10-15% of wild-type ZIP8 mRNA and protein levels, and show pleiotropic phenotype of stunted growth, neonatal lethality, multi-organ dysmorphogenesis, and dysregulated hematopoiesis manifested as severe anemia. Herein we performed RNA-seq analysis of gestational day (GD)13.5 yolk sac and placenta, and GD16.5 liver, kidney, lung, heart and cerebellum, comparing Slc39a8(neo/neo) with Slc39a8(+/+) wild-type. Meta-data analysis of differentially-expressed genes revealed 29 unique genes from all tissues - having enriched GO categories associated with hematopoiesis and hypoxia and KEGG categories of complement, response to infection, and coagulation cascade - consistent with dysregulated hematopoietic stem cell fate. Based on transcription factor (TF) profiles in the JASPAR database, and searching for TF-binding sites enriched by Pscan, we identified numerous genes encoding zinc-finger and other TFs associated with hematopoietic stem cell functions. We conclude that, in this mouse model, deficient ZIP8-mediated divalent cation transport affects zinc-finger (e.g. GATA proteins) and other TFs interacting with GATA proteins (e.g. TAL1), predominantly in yolk sac. These data strongly support the phenotype of dysmorphogenesis and anemia seen in Slc39a8(neo/neo) mice in utero.

Zinc- and bicarbonate-dependent ZIP8 transporter mediates selenite uptake

Selenite (HSeO₃⁻) is a monovalent anion of the essential trace element and micronutrient selenium (Se). In therapeutic concentrations, HSeO₃⁻ has been studied for treating certain cancers, serious inflammatory disorders, and septic shock. Little is known, however, about HSeO₃⁻ uptake into mammalian cells; until now, no mammalian HSeO₃⁻ uptake transporter has been identified. The ubiquitous mammalian ZIP8 divalent cation transporter (encoded by the SLC39A8 gene) is bicarbonate-dependent, moving endogenous substrates (Zn²⁺, Mn²⁺, Fe²⁺ or Co²⁺) and nonessential metals such as Cd²⁺ into the cell. Herein we studied HSeO₃⁻ uptake in: human and mouse cell cultures, shRNA-knockdown experiments, Xenopus oocytes, wild-type mice and two transgenic mouse lines having genetically altered ZIP8 expression, and mouse erythrocytes ex vivo. In mammalian cell culture, excess Zn²⁺ levels and/or ZIP8 over-expression can be associated with diminished viability in selenite-treated cells. Intraperitoneal HSeO₃⁻ causes the largest ZIP8-dependent increases in intracellular Se content in liver, followed by kidney, heart, lung and spleen. In every model system studied, HSeO₃⁻ uptake is tightly associated with ZIP8 protein levels and sufficient Zn²⁺ and HCO₃⁻ concentrations, suggesting that the ZIP8-mediated electroneutral complex transported contains three ions: Zn²⁺/(HCO₃⁻)(HSeO₃⁻). Transporters having three different ions in their transport complex are not without precedent. Although there might be other HSeO₃⁻ influx transporters as yet undiscovered, data herein suggest that mammalian ZIP8 plays a major role in HSeO₃⁻ uptake.

SLC39A8 Deficiency: A Disorder of Manganese Transport and Glycosylation

SLC39A8 is a membrane transporter responsible for manganese uptake into the cell. Via whole-exome sequencing, we studied a child that presented with cranial asymmetry, severe infantile spasms with hypsarrhythmia, and dysproportionate dwarfism. Analysis of transferrin glycosylation revealed severe dysglycosylation corresponding to a type II congenital disorder of glycosylation (CDG) and the blood manganese levels were below the detection limit. The variants c.112G>C (p.Gly38Arg) and c.1019T>A (p.Ile340Asn) were identified in SLC39A8. A second individual with the variants c.97G>A (p.Val33Met) and c.1004G>C (p.Ser335Thr) on the paternal allele and c.610G>T (p.Gly204Cys) on the maternal allele was identified among a group of unresolved case subjects with CDG. These data demonstrate that variants in SLC39A8 impair the function of manganese-dependent enzymes, most notably β-1,4-galactosyltransferase, a Golgi enzyme essential for biosynthesis of the carbohydrate part of glycoproteins. Impaired galactosylation leads to a severe disorder with deformed skull, severe seizures, short limbs, profound psychomotor retardation, and hearing loss. Oral galactose supplementation is a treatment option and results in complete normalization of glycosylation. SLC39A8 deficiency links a trace element deficiency with inherited glycosylation disorders.

Integrative multi-omics module network inference with Lemon-Tree

Module network inference is an established statistical method to reconstruct co-expression modules and their upstream regulatory programs from integrated multi-omics datasets measuring the activity levels of various cellular components across different individuals, experimental conditions or time points of a dynamic process. We have developed Lemon-Tree, an open-source, platform-independent, modular, extensible software package implementing state-of-the-art ensemble methods for module network inference. We benchmarked Lemon-Tree using large-scale tumor datasets and showed that Lemon-Tree algorithms compare favorably with state-of-the-art module network inference software. We also analyzed a large dataset of somatic copy-number alterations and gene expression levels measured in glioblastoma samples from The Cancer Genome Atlas and found that Lemon-Tree correctly identifies known glioblastoma oncogenes and tumor suppressors as master regulators in the inferred module network. Novel candidate driver genes predicted by Lemon-Tree were validated using tumor pathway and survival analyses. Lemon-Tree is available from http://lemon-tree.googlecode.com under the GNU General Public License version 2.0.

Oct4 promoter activity in stem cells obtained through somatic reprogramming

Multiple methods exist that can reprogram differentiated cells to a pluripotent state similar to that of embryonic stem cells (ESCs). These include somatic cell nuclear transfer (SCNT), fusion-mediated reprogramming (FMR) of somatic cells with ESCs, and the production of induced pluripotent stem cells (iPSCs). All of these methods yield cells in which the endogenous Oct4 gene is reactivated. We were interested in comparing the activity of the Oct4 promoter in three different classes of pluripotent cells, including normal ESCs, FMR cells (FMRCs), and iPSCs. We prepared cells of all three types that harbor a transgene composed of the mouse Oct4 promoter driving green fluorescent protein (Oct4-GFP). All cell derivations started with a characterized transgenic Oct4-GFP mouse, and from this we derived ESCs, FMRCs, and iPSCs with the Oct4-GFP transgene present in an identical genomic integration site in all three cell types. Using flow cytometry we assessed Oct4 promoter expression, cell cycle behavior, and differentiation kinetics. We found similar levels of GFP expression in all three cell types and no significant alterations in pluripotency or differentiation. Our results suggest that the pluripotent condition is a potent "local attractor" state, because it can be achieved through three vastly different avenues.

Cdk2ap2 is a novel regulator for self-renewal of murine embryonic stem cells

In this study we present data to support the role for Cdk2ap2 in regulating self-renewal of mouse embryonic stem cells (mESCs) under permissive conditions, and cell survival during differentiation of the mESCs into terminally differentiated cell types. To understand the function of Cdk2ap2 during early development, we generated mESCs with homozygous disruption of the endogenous Cdk2ap2 locus (Cdk2ap2(tr/tr)). The Cdk2ap2(tr/tr) mESCs, when grown in a complete growth medium containing leukemia inhibitory factor (LIF), showed an early differentiation phenotype characterized by flattened colonies and a distinct intercellular boundary. We also observed downregulation of Nanog and upregulation in markers of mesoderm and endoderm differentiation, including Brachyury (T), Afp, and S100a, when compared to Wt mESCs. Cdk2ap2(tr/tr) mESCs were able to form embryoid bodies (EBs); however, those EBs were unhealthy and had an increased level of apoptosis. Furthermore, Cdk2ap2(tr/tr) mESCs were unable to form teratomas in severe combined immunodeficiency (SCID) mice. Cdk2ap2 under normal conditions has a biphasic expression, suggesting regulatory roles in early-versus-late stem cell differentiation. These data begin to add to our understanding of how Cdk2ap2 may be involved in the regulation of self-renewal of stem cells during early embryogenesis.

ZIP8 zinc transporter: indispensable role for both multiple-organ organogenesis and hematopoiesis in utero

Previously this laboratory characterized Slc39a8-encoded ZIP8 as a Zn(2+)/(HCO(3)(-))(2) symporter; yet, the overall physiological importance of ZIP8 at the whole-organism level remains unclear. Herein we describe the phenotype of the hypomorphic Slc39a8(neo/neo) mouse which has retained the neomycin-resistance gene in intron 3, hence causing significantly decreased ZIP8 mRNA and protein levels in embryo, fetus, placenta, yolk sac, and several tissues of neonates. The Slc39a8(neo) allele is associated with diminished zinc and iron uptake in mouse fetal fibroblast and liver-derived cultures; consequently, Slc39a8(neo/neo) newborns exhibit diminished zinc and iron levels in several tissues. Slc39a8(neo/neo) homozygotes from gestational day(GD)-11.5 onward are pale, growth-stunted, and die between GD18.5 and 48 h postnatally. Defects include: severely hypoplastic spleen; hypoplasia of liver, kidney, lung, and lower limbs. Histologically, Slc39a8(neo/neo) neonates show decreased numbers of hematopoietic islands in yolk sac and liver. Low hemoglobin, hematocrit, red cell count, serum iron, and total iron-binding capacity confirmed severe anemia. Flow cytometry of fetal liver cells revealed the erythroid series strikingly affected in the hypomorph. Zinc-dependent 5-aminolevulinic acid dehydratase, required for heme synthesis, was not different between Slc39a8(+/+) and Slc39a8(neo/neo) offspring. To demonstrate further that the mouse phenotype is due to ZIP8 deficiency, we bred Slc39a8(+/neo) with BAC-transgenic BTZIP8-3 line (carrying three extra copies of the Slc39a8 allele); this cross generated viable Slc39a8(neo/neo)_BTZIP8-3(+/+) pups showing none of the above-mentioned congenital defects-proving Slc39a8(neo/neo) causes the described phenotype. Our study demonstrates that ZIP8-mediated zinc transport plays an unappreciated critical role during in utero and neonatal growth, organ morphogenesis, and hematopoiesis.

RNA and DNA microarrays

The development of microarray technology has revolutionized RNA and deoxyribonucleic acid (DNA) research. In contrast with traditional biological assays, microarrays allow the simultaneous measurement of tens of thousands of messenger RNA (mRNA) transcripts for gene expression or of genomic DNA fragments for copy number variation analysis. Over the past decade, genome-wide RNA or DNA microarray analysis has become an essential component of biology and biomedical research. The successful use of microarrays requires attention to unique issues of experimental design and execution. This chapter provides an overview of the methodology and applications of RNA and DNA microarrays in various areas of biological research.

Age-related synaptic loss of the medial olivocochlear efferent innervation

Age-related functional decline of the nervous system is consistently observed, though cellular and molecular events responsible for this decline remain largely unknown. One of the most prevalent age-related functional declines is age-related hearing loss (presbycusis), a major cause of which is the loss of outer hair cells (OHCs) and spiral ganglion neurons. Previous studies have also identified an age-related functional decline in the medial olivocochlear (MOC) efferent system prior to age-related loss of OHCs. The present study evaluated the hypothesis that this functional decline of the MOC efferent system is due to age-related synaptic loss of the efferent innervation of the OHCs. To this end, we used a recently-identified transgenic mouse line in which the expression of yellow fluorescent protein (YFP), under the control of neuron-specific elements from the thy1 gene, permits the visualization of the synaptic connections between MOC efferent fibers and OHCs. In this model, there was a dramatic synaptic loss between the MOC efferent fibers and the OHCs in older mice. However, age-related loss of efferent synapses was independent of OHC status. These data demonstrate for the first time that age-related loss of efferent synapses may contribute to the functional decline of the MOC efferent system and that this synaptic loss is not necessary for age-related loss of OHCs.

Src kinase up-regulates the ERK cascade through inactivation of protein phosphatase 2A following cerebral ischemia

BACKGROUND

The regulation of protein phosphorylation requires a balance in the activity of protein kinases and protein phosphatases. Our previous data indicates that Src can increase ERK activity through Raf kinase in response to ischemic stimuli. This study examined the molecular mechanisms by which Src activates ERK cascade through protein phosphatases following cerebral ischemia.

RESULTS

Ischemia-induced Src activation is followed by phosphorylation of PP2A at Tyr307 leading to its inhibition in the rat hippocampus. SU6656, a Src inhibitor, up-regulates PP2A activity, resulting in a significant decreased activity in ERK and its targets, CREB and ERalpha. In addition, the PP2A inhibitor, cantharidin, led to an up-regulation of ERK activity and was able to counteract Src inhibition during ischemia.

CONCLUSION

Src induces up-regulation of ERK activity and its target transcription factors, CREB and ERalpha, through attenuation of PP2A activity. Therefore, activation of ERK is the result of a crosstalk between two pathways, Raf-dependent positive regulators and PP2A-dependent negative regulators.

Discovery of ZIP transporters that participate in cadmium damage to testis and kidney

It has been known for decades that cadmium (Cd) must enter the cell to cause damage, but there was no mechanism to explain genetic differences in response to Cd toxicity until 2005. Starting with the mouse Cdm locus associated with differences in Cd-induced testicular necrosis between inbred strains, a 24.6-centiMorgan region on chromosome 3 was reduced ultimately to 880 kb; in this segment is the Slc39a8 gene encoding the ZIP8 Zn(2+)/HCO(3)(-) symporter. In endothelial cells of the testis vasculature, Cd-sensitive mice exhibit high ZIP8 expression, Cd-resistant mice exhibit very low expression. A 168.7-kb bacterial artificial chromosome (BAC) from a 129S6 (Cd-sensitive) BAC library containing the Slc39a8 gene was inserted into the Cd-resistant C57BL/6J genome: Cd treatment produced testicular necrosis in BAC-transgenic BTZIP8-3 mice but not in non-transgenic littermates, thereby proving that the Slc39a8 gene is indeed the Cdm locus. Cd-induced renal failure also occurred in these BTZIP8-3 mice. Immunohistochemistry showed highly expressed ZIP8 protein in the renal proximal tubular epithelial apical surface, suggesting that ZIP8 participates in Cd-induced renal failure. Slc39a14, most closely evolutionarily related to Slc39a8, encodes differentially-spliced products ZIP14A and ZIP14B that display properties similar to ZIP8. ZIP8 in alveolar cells brings environmental Cd into the organism and ZIP14 in intestinal enterocytes carries Cd into the organism and into the hepatocyte. We believe these two transporters function endogenously as Zn(2+)/HCO(3)(-) symporters important in combating inflammation and carrying out other physiological functions; Cd is able to displace the endogenous cation, enter the cell, and produce tissue damage and disease.

Module networks revisited: computational assessment and prioritization of model predictions

MOTIVATION

The solution of high-dimensional inference and prediction problems in computational biology is almost always a compromise between mathematical theory and practical constraints, such as limited computational resources. As time progresses, computational power increases but well-established inference methods often remain locked in their initial suboptimal solution.

RESULTS

We revisit the approach of Segal et al. to infer regulatory modules and their condition-specific regulators from gene expression data. In contrast to their direct optimization-based solution, we use a more representative centroid-like solution extracted from an ensemble of possible statistical models to explain the data. The ensemble method automatically selects a subset of most informative genes and builds a quantitatively better model for them. Genes which cluster together in the majority of models produce functionally more coherent modules. Regulators which are consistently assigned to a module are more often supported by literature, but a single model always contains many regulator assignments not supported by the ensemble. Reliably detecting condition-specific or combinatorial regulation is particularly hard in a single optimum but can be achieved using ensemble averaging.

AVAILABILITY

All software developed for this study is available from http://bioinformatics.psb.ugent.be/software.