Hexokinase Domain Containing Protein-1 in Metabolic Diseases and Beyond

Abstract

Glucose phosphorylation by hexokinases traps glucose in cells and facilitates its usage in metabolic processes dependent on cellular needs. Hexokinase domain containing protein-1 (HKDC1) is a recently discovered protein with wide expression containing hexokinase activity, first noted through a genome-wide association study to be linked with gestational glucose homeostasis during pregnancy. Since then, HDKC1 has been observed to be expressed in many human tissues. Moreover, studies have shown that HKDC1 plays a role in glucose homeostasis by which it may affect the progression of many pathophysiological conditions such as gestational diabetes mellitus, NASH and cancer. Here, we review the key studies contributing to our current understanding of the roles of HKDC1 in human pathophysiological conditions and potential therapeutic interventions.

Discovery and functional assessment of HKDC1

Upon cell entry, hexokinases (HKs) (see Glossary) phosphorylate hexose sugars, primarily glucose, trapping these sugars in the cell and facilitating their usage in metabolic processes dependent on cellular needs [1–3]. Four HKs have been characterized and extensively reviewed – HK1–3 have high enzymatic activity and mediate the phosphorylation of glucose for the purpose of cellular utilization, whereas glucokinase (GK) functions as a glucose sensor due to its higher K_m that is near physiological blood glucose levels [4–7]. The HKs are found in a wide range of tissues, but their patterns of expression differ [8–11]. In the mid-2000s, in an effort to better understand the evolutionary relationships of the HKs and to explain the evolution of GK and the diversification of the HKs, phylogenetic analyses were performed, which uncovered a novel HK-like gene called HKDC1 [12]. HKDC1 lies on chromosome 10 adjacent to HK1 and encodes a 100 kDa protein product sharing approximately 70% sequence homology to HK1 [12,13]. Review of the human Expressed Sequence Tags (EST) database suggested that HKDC1 was widely expressed and given its homology to HK1, along with the significance of the HKs for glucose metabolism, this preliminary study proposed that HKDC1 is, in fact, another HK.

HKDC1 is differentially expressed in many tissues

Studies utilizing a human EST database [12] and tissue samples from the Genotype-Tissue Expression Consortium [13] have found HKDC1 in a wide range of human and mouse tissues. Confirming this mRNA data, Khan et al. [14] performed immunohistochemical analyses on human surgical specimens and demonstrated that HKDC1 protein expression is similarly widely expressed. A review of the key findings from HKDC1 mRNA and protein expression work is summarized in Tables 1 and 2, respectively.

Table 1:

Organs and tissues with high levels of HKDC1 gene expression

Species	Source/Method	Organ/Tissue	Reference
Homo sapiens	Human EST database*	Colon	[12]
		Esophagus
		Eye
		Pharynx
		Thymus
Homo sapiens	Genotype-Tissue Expression (GTEx) project**	Brain Colon Kidney	[13]
		Liver
		Lung
		Nerve
		Pancreas
		Pituitary
		Small intestine
		Testes
		Thyroid
Mus musculus	RT-PCR	Brain	[13]
		Colon
		Kidney
		Lung
		Small intestine

^*Database website was retired in July 2019. The data utilized in the work of Irwin et al can be found at: https://ftp.ncbi.nlm.nih.gov/repository/UniGene/

^**Website: https://www.gtexportal.org

Table 2:

Immunohistochemical analyses of HKDC1 protein expression

System	Organ/Tissue	Comments	Reference
Gastrointestinal	Colon	Strong HKDC1 expression restricted to the epithelial cells of the villi	[13]
	Liver	Diffuse, low level expression in hepatocytes
Genitourinary	Kidney	HKDC1 expression predominately in the renal tubules	[13]
Cardiovascular	Heart	Weak intensity* staining in cardiomyocytes	[14]
Endocrine	Thyroid	Moderate to strong intensity staining in the follicular epithelium	[14]
Gastrointestinal	Esophagus	Moderate intensity staining in the suprabasal squamous epithelium	[14]
		Weak intensity staining in basal layer
	Stomach	Strong intensity staining in parietal and chief cells	[14]
		Moderate intensity staining in the superficial foveolar epithelium
	Small intestine	Strong intensity staining in the villous brush border	[14]
	Colon	Strong intensity staining in the surface epithelium	[14]
		Weak intensity staining in the crypt base epithelium
	Liver	Weak intensity staining in the interlobular bile duct	[14]
	Pancreas**	Strong intensity staining in the larger and interlobular pancreatic ducts, acinar centrocytes	[14]
Genitourinary	Kidney	Mostly weak intensity staining in the renal tubule and glomerulus	[14]
	Prostate	Moderate intensity staining in basal cells	[14]
		Weak to moderate intensity staining in stromal cells
		Weak intensity staining in the gland epithelium
	Testes	Weak intensity staining in spermatocytes	[14]
Integumentary	Skin	Moderate intensity staining in the sweat gland duct	[14]
		Weak intensity staining in the epidermis, hair follicle, and sweat gland
Lymphatic	Spleen	Weak to moderate intensity staining in white pulp	[14]
		Weak intensity staining in red pulp
Pulmonary	Lung	Strong intensity staining in alveolar macrophages	[14]
		Moderate intensity staining in the bronchial epithelium
		Weak intensity staining in alveolar lining cells

^*HKDC1 immunostaining assessed by estimating the intensity of cytoplasmic staining and assigning it to one of four categories: no staining, weak-positive, moderate-positive, or strong positive staining

^**In the pancreas, the most intense HKDC1 staining was seen in exocrine tissue; pancreatic islet cells of the endocrine system exhibited weak intensity staining

HKDC1 contains hexokinase activity

Considering the significant sequence homology to HK1, it was assumed that HKDC1 functions similarly to the known HKs. Assessment of the amino acid sequence of HKDC1 in comparison with other HKs demonstrated the presence of conserved amino acid residues in the N- and C-terminal domains that are predicted to be necessary for glucose and ATP-binding [12]. The presence of these residues suggests that HKDC1 may function as a hexokinase. To test this hypothesis, Guo et al. [15] transduced INS-1 rat pancreatic β-cells with adenovirus containing HKDC1 from a human cytomegalovirus promoter, then assessed hexokinase activity from cell lysates utilizing different concentrations of glucose. At baseline, INS-1 cells contain minimal detectable low K_m hexokinase activity. While the expression of the other HKs were unaffected by expression of HKDC1, the authors found that adenovirus-HKDC1 treated INS-1 cells resulted in significantly elevated HKDC1 protein and enhanced hexokinase activity across the range of glucose concentrations utilized. Furthermore, the shift in dose-response curve suggested that HKDC1 has a lower K_m than GK. Together, these findings suggested that HKDC1 contributes to cellular HK activity through either direct HK activity or modulation of the activity of the other HKs. To determine more specifically if HKDC1 has intrinsic HK activity, the authors purified HKDC1 and utilizing a hexokinase assay in vitro found that HKDC1 contained 20% of the specific activity of HK1 under the same conditions. Taken together, these studies indicated that HKDC1 is a widely expressed fifth HK, fueling interest in understanding the function of HKDC1 in metabolism.

Genetic Variations of HKDC1 in Human Disease

Although the role of HKDC1 as a canonical hexokinase was initially debated, it was believed that its potential for physiological impact was likely, since HKDC1 is both highly conserved between vertebrate species and is expressed in a multitude of human tissues [12]. Since then, multiple large-scale genetic studies have uncovered significant associations between genetic variations in HKDC1 and disease phenotypes.

The first breakthrough in determining HKDC1 functionality occurred when several gene variants in HKDC1 were identified to be in association with maternal glucose metabolism during a large-scale genome-wide association study (GWAS) investigating the genetic architecture underlying glycemic traits during pregnancy [16]. Specifically, the Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) Study sought to examine the genetic variation associated with glycemic traits during pregnancy in women of various ethnic backgrounds [17,18]. Utilizing a subset of participants in the main study, Hayes et al. [16] examined DNA samples and phenotypic data from four different ancestry populations, and cohort-specific and meta-analyses of genome-wide single nucleotide polymorphism (SNP) data were conducted to identify common genetic variants associated with glycemic traits, including maternal fasting plasma glucose and C-peptide levels, and one- and two-hour plasma glucose levels measured during an oral glucose tolerance test. The locus with the strongest association in the GWAS was 2-hour plasma glucose (2HPG) with region 10q22.1 – just upstream of the first intron of HKDC1.

The genetic locus of HKDC1 is comprised of a 3,653 base pair region upstream of HK1 [http://useast.ensembl.org/Homo_sapiens/Gene/Summary?db=core;g=ENSG00000156510;r=10:69220332-69267552ⁱ]. The GWAS from the HAPO subset revealed a 400-kb association locus with 2HPG spanning several genes, including HKDC1. The SNP with the strongest association was rs4746822, which reached genome-wide significance in a meta-analysis combining the four ancestry groups (Table 3). Other 2HPG-associated variants near HKDC1 were identified in regions of open chromatin and histones H3K27ac and H3K27me, indicating the presence of active regulatory elements [19]. These results linked HKDC1, for the first time, to having potential physiological function during pregnancy [20].

Table 3:

Genetic variants of HKDC1 and their associations

							Refs
Hayes et al.	rs5030937	gravid 2HPG	1.02E-22	Hispanic (n=817), Afro-Caribbean (n=1075), Northern European (n=4393*),	7,463 pregnant women	2-h plasma glucose collected after OGTT at ~28 weeks gestation	[16]
	rs4746822	gravid 2HPG	6.30E-16	Thai (n=1,178)
Guo et al.	rs4746822	gravid 2HPG	4.41E-13	Africa, Americas, Asia, Europe	4,437 pregnant women from Hayes et al. cohort	2-h plasma glucose collected after OGTT at ~28 weeks gestation	[15]
	rs10762264		3.75E-12	(n unspecified)
	rs2394529		1.88E-12
	rs9645501		8.61E-06
Kanthimathi et al.	rs4746822 rs10762264	GDM	0.004 0.002	South Indian	500 pregnant women with GDM, 510 non-GDM	Plasma glucose values exceed: fasting ≥5.1 mmol/L, 1-h ≥ 10.0 mmol/L and 2-h ≥ 8.5 mmol/L, following OGTT during pregnancy	[22]
	rs2394529		0.020
	rs9645501		0.059
Tan et al.	rs4746822	gravid 2HPG	0.016	Han Chinese	334 pregnant women, 367 controls	Plasma glucose values exceed: fasting ≥5.1 mmol/L, 1-h ≥ 10.0 mmol/L and 2-h ≥ 8.5 mmol/L, following OGTT during 24– 28 weeks gestation	[23]
Karolak et al.	rs202105269	Keratoconus	N/A**	Ecuadorian	8 family members	Based on visual acuity testing, intraocular pressure assessment, biomicroscopic evaluation, and fundus examination with dilation.	[24]
Zhang et al.	rs3740598	Retinitis pigmentosa	N/A**	Unknown	Two unrelated probands	Loss of vision, fundus photographic and angiographic changes	[25]
	rs3740600
Neale et al.	rs902619	ADHD	3.73E-05	Predominantly European/American	1,150 participants	DSM-IV criteria for ADHD	[27]

^*Includes external replication analysis

^**WES analysis, no P value

Utilizing the findings from their GWAS study in addition to genetic databases and several computational models, Guo et al. [15] identified four common genetic variants at the HKDC1 locus (Table 3). Because none of these variants were found in the coding region of the gene, the authors hypothesized that genetic variants alter HKDC1 expression through multiple regulatory elements. The four variants with the strongest association to maternal 2-hour glucose levels, namely, rs4746822, rs10762264, rs2394529, and rs9645501, lie in four separate enhancers that have a cooperative effect of reducing HKDC1 expression. These data support a model in which multiple genetically-linked regulatory variants coordinate to reduce HKDC1 expression in women with higher gestational glucose levels.

Since these discoveries, further studies have sought to verify these findings in additional ancestry groups. Because higher gestational 2-hour plasma glucose levels can indicate an inadequate response to a state of increased insulin resistance, one study aimed to evaluate the association of the four major regulatory variants of HKDC1 with gestational diabetes mellitus (GDM), a maternal disease linked to adverse fetal outcomes and an increased risk of type 2 diabetes mellitus (T2DM) in mothers [21]. This replication study utilized a cohort of pregnant women within a South Indian population – an ancestry not included in the HAPO cohort – and found a strong association with GDM in variants rs4746822 and rs10762264, as well as a modest association in variant rs2394529 [22] (Table 3). Additionally, the authors found that presence of the risk variant conferred 1.95, 1.71, and 1.64 times higher risk for GDM, respectively. Similarly, a GWAS seeking to replicate the results of the HAPO study in a Han Chinese population found a significant association between HKDC1 gene variant rs4746822 and GDM [23] (Table 3).

HKDC1 has appeared in several genetic studies linked to ocular diseases (Table 3). Karolak et al. [24] utilized both whole-exome sequencing (WES) and Sanger sequencing to identify coding sequence variants in an individual genome and sequenced DNA samples from an Ecuadorian family with keratoconus. They identified a new variant in HKDC1’s coding region resulting in a glycine-to-serine substitution which modeling studies suggested may affect the structure and function of HKDC1. Utilizing WES, Zhang et al. [25] identified a specific HKDC1 missense variant in patients with retinitis pigmentosa. To further assess the functional role of HKDC1 in this disease, the authors created an HKDC1 knockout mouse model using CRISPR/Cas9 technology and found that loss of HKDC1 led to mislocalization of rhodopsin to the inner segments and cell bodies of rods in some retina regions [26]. Additionally, in vitro studies revealed the missense mutation causes 30% reduction in HK activity. The authors speculated HKDC1 may be important for glucose sensing in the human retina given its similarity to HK1.

Recently, HKDC1 was mentioned in a case-controlled GWAS of attention deficit/hyperactivity disorder (ADHD) [27] (Table 3). The study found no significant genome-wide associations, but genome-wide SNP analyses suggested a handful of genes, including HKDC1, may have a role in influencing ADHD risk. Similarly, a study investigating genes involved in fat deposition identified HKDC1 as being upregulated in fast-growing compared to slow-growing chickens [28]. Although a potential mechanism linking HKDC1 to lipid metabolism was not proposed, the HAPO GWAS study also found a nominal association between HKDC1 variant rs4746822 and number of skinfolds in neonates—a common way to quantify body fat in neonates and young infants [29]. Because increases in neonatal adiposity have been described in infants exposed to uncontrolled GDM in utero, this study indirectly links HKDC1 to a glucometabolic role [30,31]. Taken altogether, genetic studies have begun to reveal the role of HKDC1 in metabolism and human disease.

As part of the Accelerating Medicines Partnership Type 2 Diabetes Knowledge Portal project (AMP-T2DKP), a human genetic database from T2DM patients, a GWAS utilizing a multi-parent advanced generation inter-cross (MAGIC) plus population of T2DM indicated a genetic link between at least 45 HKDC1 variants and hemoglobin A1c [https://t2d.hugeamp.org/region.html?chr=10&end=71077308&phenotype=HBA1C&start=70930088ⁱⁱ]. These variants differ from those affecting 2-hour glucose levels during pregnancy, with some predicted to affect intron sequences or transcription factor binding. In conjunction with prior work associating HKDC1 with 2-hour plasma glucose levels after an oral glucose tolerance test (OGTT) during pregnancy, as well as with GDM, these new studies provide additional evidence for roles of distinct HKDC1 variants in overall glucose metabolism and T2DM. To explore this, studies have utilized transgenic mouse models to vary expression of HKDC1 both globally and in an organ-specific pattern.

HKDC1 is involved in glucose metabolism

HKDC1 global knockdown mouse studies

Ludvik et al. [13] created and analyzed a mouse model in which HKDC1 was knocked down in all HKDC1-expressing tissues (Box 1). In response to an OGTT at 28 weeks of age, heterozygous HKDC1^+/− mice exhibited a significantly increased glucose excursion compared to wild type mice. This finding was not specific to glucose uptake via the intestine, as an increased glucose excursion was also seen during an intraperitoneal glucose tolerance test. These findings were not due to changes in insulin secretion, insulin sensitivity or gluconeogenesis, nor an overt phenotypic difference such as differences in weight, random blood glucose, or fasting glucose or insulin levels. The effect of HKDC1 knockdown was also analyzed in 8- to 12-week-old pregnant mice at day 15 of gestation, which is the peak of insulin resistance during mouse pregnancy [32]. Similar to 28-week-old mice, pregnant HKDC1^+/− mice exhibited an increased glucose excursion compared to pregnant wild type mice during an OGTT. Interestingly, non-pregnant females and male HKDC1^+/− mice at 8- to 12-weeks of age did not demonstrate a difference in glucose tolerance compared to age-matched controls. These results suggested an effect of HKDC1 on glucose metabolism under times of metabolic stress, such as in pregnancy or aging.

Box 1: Creation of HKDC1 global knockdown mice.

As the first study to investigate the role of HKDC1 in glucose metabolism by genetic mouse models, Ludvik et al [13] utilized a gene trap system to globally ablate HKDC1 expression through embryonic stem cells obtained from the Knockout Mouse Repository (www.komp.org) and generated by the Wellcome Trust Sanger Institute. In these mice, a LacZ-Neo cassette was inserted into the HKDC1 gene after exon 2, producing a truncated nonfunctional gene product. Interestingly, utilizing an HKDC1^+/− x HKDC1^+/− breeding approach, no homozygous (HKDC1^−/−) offspring were observed, whereas the remainder of the litters occurred at the expected 2-to-1 Mendelian ratio of heterozygous to wild type mice. Through this breeding schematic, heterozygous HKDC1^+/− mice exhibited approximately 50% reduction in total HKDC1 expression and were utilized for analysis. Though this finding suggests that HKDC1 may have a role in embryonic development, Zhang et al. [25] created viable HKDC1 KO mice using CRISPR/Cas9 technology, hence the significance of the global lethality in one model and not another is unclear. Prior reports have reported embryonic lethality in HK2 and GK transgenic mice [58,59].

To explore why HKDC1^+/− mice exhibit impaired glucose tolerance in the absence of alterations in baseline phenotype, insulin action or gluconeogenesis, further studies were performed [13]. An assessment of the livers of HKDC1^+/− mice indicated that these mice had overall lower levels of triglycerides, and glycogen levels trended lower in male mice. Furthermore, utilizing an oral bolus of [U-¹³C₆]glucose to more directly assess whole-body glucose utilization during an OGTT, the overall quantity of radiolabeled glucose found in tissues with high HKDC1 expression, such as the kidney and small intestine, was significantly lower in HKDC1^+/− mice compared to wild type, and trended to lower levels in other tissues. These data suggest that HKDC1 has roles in hepatic energy storage, and the reduced radiolabeled glucose uptake by tissues correlates with the impaired glucose tolerance seen in aged and pregnant HKDC1^+/− mice. These studies conducted with HKDC1 global knockdown mice suggest that HKDC1 is important for whole-body glucose utilization and hepatic energy storage, particularly under conditions of enhanced metabolic stress including pregnancy and with aging (Figure 1). These data support the GWAS-derived observation that HKDC1 is associated with maternal 2-hour glucose levels after an OGTT at 28 weeks’ gestational age in humans.

Figure 1: Effects of modulation of HKDC1 protein expression on glucose homeostasis.

C57B1/6J mice were created such that they expressed 50% of wild-type HKDC1 protein in all tissues (left side, green boxes; see Box 1 for protocol) or selectively overexpressed or did not express hepatic HKDC1 (right side, orange boxes; see Box 2 for protocols). Mice displaying whole-body HKDC1 knockdown were assessed after an overnight fast using an OGTT consisting of a bolus of 2g/kg body weight glucose. During pregnancy and in aged mice, HKDC1 knockdown exhibit impaired blood glucose excursions during the OGTT and overall display reduced tissue glucose uptake and overall glucose tolerance. These findings were only found in pregnancy and with age, which are conditions of greater metabolic stress, and were not found in younger non-pregnant mice. Mice displaying liver-specific modulation of HKDC1 expression were assessed during pregnancy after an overnight fast using an intraperitoneal GTT consisting of a bolus of 2g/kg body weight glucose. Similar to whole-body HKDC1 knockdown, mice devoid of hepatic HKDC1 expression exhibited impaired tissue uptake of glucose and overall glucose tolerance after receiving the intraperitoneal glucose load, and also increased hepatic gluconeogenesis. On the contrary, mice overexpressing hepatic HKDC1 displayed the opposite findings, with improved tissue glucose uptake, glucose tolerance and insulin sensitivity, and indications of a shift in metabolism from glucose reliance to fat oxidation and ketone utilization during fasting due to the presence of increased ketone bodies.

Modulation of liver-specific HKDC1 expression

Khan et al. [33] recently showed in an adult-onset hepatocyte-specific HKDC1 gain/loss mouse model (Box 2) that during the later stages of pregnancy in mice, modulation of hepatic HKDC1 expression alters systemic nutrient balance. While pregnancy-induced insulin resistance was evident in these mice, HKDC1 overexpression reversed this phenomenon. In contrast, pregnant mice with hepatocyte-specific HKDC1 knocked down exhibited impaired insulin tolerance compared to controls. As one of the probable causes of enhanced insulin sensitivity with HKDC1 overexpression, it was shown that there was enhanced insulin-induced phosphorylation of protein kinase B (Akt) compared to wild type mice. Moreover, congruent with enhanced insulin sensitivity in pregnant mice overexpressing hepatic HKDC1, there was significantly less glucose production during a pyruvate challenge and, consistent with this, hepatic HKDC1 knockdown mice had more pronounced hepatic gluconeogenesis. Complimenting this observation, genes involved in de novo gluconeogenesis showed a negative correlation with HKDC1 expression. Furthermore, this study showed that, in the fasting state, mice overexpressing HKDC1 had significantly higher plasma nonesterified fatty acids and ketones, the latter suggesting enhanced β-oxidation. This study demonstrates that hepatic HKDC1 influences maternal whole body nutrient balance by modulating insulin sensitivity, glucose tolerance, hepatic gluconeogenesis, and ketone production (Figure 1).

Box 2: Creation of mice lacking or overexpressing liver-specific HKDC1.

Khan et al. [33] created adenovirus containing recombinant human HKDC1 in preparation for introduction into adult mice. To assess the effect of overexpression of hepatic HKDC1 on glucose metabolism in pregnant mice, 10-day pregnant female C57B1/6J mice were injected into the lateral tail vein with saline containing adenoviral HKDC1 constructs, creating an adult-onset hepatocyte-specific human HKDC1 overexpression mice. Pregnant female mice injected with adenovirus containing GFP served as controls. These mice were analyzed at 17–18 days of pregnancy.

To create adult female mice specifically lacking hepatic HKDC1, HKDC1-floxed mice (HKDC1^fl/fl mice, www.komp.org) were injected in the lateral tail vein with saline containing adeno-associated virus serotype 8 bearing a thyroid-binding globulin promoter driven Cre recombinase. Female mice injected with a Null vector served as controls. These female C57B1/6J mice lacking hepatic HKDC1 were created prior to pregnancy. Following mating and attainment of pregnancy, these female mice lacking hepatic HKDC1 were analyzed at day 17–18 of pregnancy.

In a further study, Pusec et al. [34] characterized the role of hepatic HKDC1 in metabolism and provided insights into a possible role of HKDC1 in the pathogenesis of nonalcoholic fatty liver disease (NAFLD). In this report, it was shown that purified HKDC1 protein has very low hexokinase activity, similar to prior work [15], and that acute in vivo overexpression of human HKDC1 in mice mildly improves glucose tolerance. Although HKDC1 has low hexokinase activity, it appears that it might be modulating glucose metabolism by its localization to the mitochondria. Observing that HK1 and HK2 contain N-terminal mitochondrial localization sequences, the authors here assessed HKDC1’s cellular localization utilizing full-length protein and a truncated version lacking the N-terminal twenty residues. Using immunofluorescence, full-length HKDC1 localized to the mitochondria, whereas the N-terminal truncated version was found in various cytosolic compartments. Correlating with this finding, hepatocyte-specific overexpression of HKDC1 significantly reduced glycolytic capacity, maximal respiration, glucose oxidation, and diminished mitochondrial membrane potential, suggesting that HKDC1 may induce mitochondrial dysfunction in hepatocytes.

Examining how these findings may relate to disease, the authors utilized human liver biopsy samples and various mouse models to determine the correlation of hepatic HKDC1 mRNA and protein levels liver fat [34]. Interestingly, HKDC1 expression positively correlated with not just the degree of liver steatosis but steatohepatitis and fibrosis in both humans and mice, with the greatest levels of HKDC1 expression found in liver samples that had progressed to nonalcoholic steatohepatitis (NASH). Hence, hepatic HKDC1-induced mitochondrial interactions and/or dysfunction may have roles in the progression of NAFLD.

HKDC1 is involved in the adaptive integrated stress response

In response to cellular stress, eukaryotic cells adapt through phosphorylation of eukaryotic translation initiation factor 2 alpha (eIF2α), leading to translation of select transcripts that allow the cell to endure stress [35–40]. One of these transcripts is Activating Transcription Factor 4 (ATF4) – a leucine zipper transcription factor, that induces genes affecting a wide spectrum of cellular processes in order to best endure stressors [36, 41–43]. In a study aiming to identify novel transcription targets of ATF4, Evstafieva et al. [44] found HKDC1 to be highly responsive to ATF4 overexpression. In response to inhibition of the mitochondrial respiration chain and endoplasmic reticulum (ER) stress, HKDC1 is upregulated, however in the presence of the same cellular stress combined with inhibition of ATF4 by RNA interference, HKDC1 gene expression is attenuated. These findings suggest a role for HKDC1 in the cellular integrated stress response.

Role of HKDC1 in cancer

A link between the HKDC1 and cancer (Figure 2a) is suggested by studies that have observed changes in HKDC1 expression or splicing occurring in cancer, as compared to a normal physiological state. Li et al. [45] utilized RNA-seq data from The Cancer Genome Atlas to identify genetically changed genes in univariate survival analysis in patients with squamous cell lung carcinoma (SQCLC) [46]. Through analysis of RNA-seq data from 550 SQCLC patients, 7,222 genetically changed genes were identified of which HKDC1 was one of 14 feature genes with more than 100 frequencies that was associated with a poorer prognosis. Bong et al. [47] examined microRNAs (miRNAs), which are small non-coding double-stranded RNAs that bind to and downregulate target gene expression [48], amongst patients with and without multiple myeloma (MM) to identify those associated with multiple myeloma (MM) disease progression. They found five miRNAs in MM patients with significantly higher expression compared to healthy subjects that were associated with reduced HKDC1 expression. This finding suggests a regulatory role of specific miRNAs in curbing HKDC1 expression. Lian et al. [49] examined alternative splicing events in HKDC1 in colorectal cancer (CRC), a leading cause of mortality in developed countries [50]. Through RNA-seq analysis of paired CRC and normal tissues, significant splicing differences were found in nine genes in CRC and confirmed by The Cancer Genome Atlas data amongst CRC tumors and normal controls. Utilizing DEXSeq, aimed at finding changes in relative exon usage [51], the authors identified alternative regulation of the first exon in HKDC1, where HKDC1 E1a-E3a was upregulated in CRC, suggesting a potential functional impact due to a predicted alteration in the HKDC1 protein sequence. Furthermore, in an assessment of the clinical significance of splicing events to overall survival in CRC, patients with CRC were found to have higher expression of HKDC1 E1-E2 junction, which when combined with increased expression of COL6A3 E5-E6 junction, are associated with better survival in CRC.

Figure 2: Synopsis of HKDC1 roles in human cancers.

The diagram represents a summary of key studies identifying and assessing alterations in the HKDC1 gene (a) and HKDC1 protein expression (b) found in corresponding human cancer tissues and cell lines. In the studies noted, findings were determined by comparing data obtained from cancerous human tissue samples or cancer cell lines to matching normal human tissue and cell lines. A key finding amongst the data is the finding that greater levels of HKDC1 in cancerous tissue correlates with increased proliferation, migration, and invasion, leading to a poorer overall prognosis and reduced survival. Abbreviations: AICAR, 5-aminoimidazole-4-carboxamide ribonucleotide (analog of adenosine monophosphate that stimulates AMPK activity); CRC, colorectal carcinoma; HCC, hepatocellular carcinoma; LUAD, lung adenocarcinoma; miRNA, microRNA; ROS, reactive oxygen species; siRNA, small interfering RNA; VDAC1, voltage dependent anion channel 1. Cell lines: SW480, human colon adenocarcinoma cell line, grade 3–4; SW620, metastatic human colon adenocarcinoma cell line obtained from lymph node; A549, non-small cell lung carcinoma cell line derived from adenocarcinomic lung tissue; H1299, metastatic non-small cell lung carcinoma cell line obtained from lymph node; MCF7, human breast cancer cell line.

In another study focusing on CRC, Fuhr et al. [52] examined the connection between HKDC1 and circadian clock genes in SW480 CRC cells. They found a 13-hr phase-shift in HKDC1 expression between SW480 cells and their metastatic counterpart, SW620, which was in line with a phase shift in aryl hydrocarbon receptor nuclear translocator-like protein-1 (BMAL1; core clock gene). Silencing BMAL1 leads to upregulation of HKDC1 expression in SW480 cells and vice versa, which was lost in SW620 cells. These results suggest an interplay between HKDC1 and the circadian clock which is disrupted in metastatic cells.

Studies attempting to define mechanistic roles for HKDC1 in cancer have also begun (Figure 2b). In the liver, Zhang et al. [53] showed that HKDC1 expression is upregulated in human hepatocellular carcinoma (HCC) tissue and is positively correlated with aggressive phenotype and poor prognosis. Furthermore, genetic knockdown of HKDC1 in BEL-7402 and HepG2 human hepatocellular carcinoma cell lines significantly suppressed cellular proliferation and migration, likely due to concurrent downregulation of β-catenin and c-Myc expression. These findings indicate that therapeutic knockdown of HKDC1 in HCC may confer benefit by inactivating pro-migratory and proliferation signaling pathways. Similarly, Wang et al. [54] found higher levels of HKDC1 mRNA and protein levels in lung adenocarcinoma cell lines compared to normal lung epithelial cells. Furthermore, the level of HKDC1 protein expression directly correlated with histologic differentiation, tumor size, pN stage, poor prognosis and lower survival. Correlating with these findings, lung adenocarcinoma cells lines stably created to overexpress HKDC1 exhibited higher glucose consumption and lactate production, and showed greater proliferation, migration and invasion compared to normal lung epithelial cells. Furthermore, overexpression of HKDC1 led to reduced Thr172 phosphorylation of AMPKa and increased phosphorylation of mTOR at Ser2448 and p70S6 – findings which were reversed when HKDC1 expression was silenced. To verify an oncogenic role of HKDC1, rescue experiments using AICAR (activator of AMPK signaling) and rapamycin (inhibitor of mTORC signaling) within lung adenocarcinoma cells overexpressing HKDC1 showed attenuation of proliferation, migration, invasion, glycolysis, and epithelial-mesenchymal transition. Collectively, these results demonstrate that HKDC1 affects proliferation of lung adenocarcinoma, likely through regulation of AMPK/mTOR signaling.

With the observation that HKDC1 is significantly increased in breast cancer cells and tissues, Chen et al. [55] utilized MCF7 breast adenocarcinoma cells and found that, similarly to Pusec et al. [34], HKDC1 is located in the mitochondrial outer membrane and binds with voltage dependent anion channel 1 (VDAC1). This leads to increased mitochondrial membrane potential and glucose uptake, thus coupling glucose metabolism to mitochondrial ATP synthesis and the promotion of cell survival and proliferation. In line with this finding, knockdown of HKDC1 increased ROS formation, caspase-3 activity, and apoptosis.

Could the HKDC1-VDAC1 interaction serve as a potential therapeutic target point? In a recent study, Chen et al. [56] examined whether interruption of the HKDC1-VDAC1 interaction could affect the proliferation of extranodal nasal-type natural killer/T-cell lymphoma (ENKTL). Similar to prior work, HKDC1 expression was highly upregulated in ENKTL cells and silencing HKDC1 suppressed tumor growth. Furthermore, delivery of the C-terminal eight residues of HKDC1 using transferrin (Tf) receptor internalization sequence (Tf-D-HKC8) resulted in dissociation of HKDC1 from VDAC1, increased apoptosis, increased oxidative stress and DNA damage, reduced EBV replication, and smaller tumors. Of note, Pusec et al. [34] first showed that the N-terminal region of HKDC1 was critical for VDAC interaction, however the potential for existence of secondary interactions at other sites along HKDC1 that may stabilize the HKDC1-VDAC binding is suggested by these findings involving the C-terminus of HKDC1. This study by Chen et al. is the first to show that inhibition of the HKDC1-VDAC1 interaction can improve survival in cancer and the first to suggest that targeting this interaction may serve therapeutic benefit.

Concluding remarks and future perspectives

HKDC1 phosphorylates glucose and is widely expressed in tissues, yet its precise role in human metabolism remains to be completely understood. We have determined that genetic variants of HKDC1 are associated with increased glucose excursions after an OGTT during pregnancy. In correlation with this finding, genetic variants of HKDC1 are related with HgbA1c, suggesting a potential role in diabetes. Lower HKDC1 expression levels were shown to increase glucose excursions after a glucose load during pregnancy and during aging, suggesting a role of HKDC1 in maintaining glucose homeostasis during metabolic stress. More studies are needed to assess the function of HKDC1 in different tissue types, as well as to pinpoint why alterations in HKDC1 expression affect glucose homeostasis only under certain conditions (see Outstanding Questions).

Outstanding questions.

What is/are the function(s) of HKDC1 within each tissue type where it is expressed?

What is/are the precise biochemical mechanism(s) by which HKDC1 contributes to whole-body glucose utilization and homeostasis? How does HKDC1 expression levels affect glucose metabolism mainly under conditions of metabolic stress?

How do genetic variants of HKDC1 contribute to the development of diabetes?

What are the precise biochemical mechanisms by which HKDC1 promotes cell survival, proliferation, and metastasis in cancer?

Can targeting the HKDC1-VDAC interaction prevent the progression of multiple human cancers and are there other targets of HKDC1 activity that could serve as critical points for the development of potential therapeutics?

Could HKDC1 eventually be a target for new therapeutics aimed at improving survival in certain human cancers? Li et al. [57] utilized anticancer activity enrichment analysis to search for genes having a high likelihood (>60%) of being clinical targets for lung cancer treatment. Of 50 potential target genes, the only one in the top twenty for which no therapeutics had been designed was HKDC1. As discussed here, multiple studies suggest roles for HKDC1 in cancer progression. HKDC1 mRNA and protein are highly expressed in cancerous tissue compared to normal, matched tissue. In the presence of high HKDC1 levels, cancer cells show increased proliferation, migration, invasiveness and overall survival, whereas lowering HKDC1 mRNA levels in cancer cells leads to the opposite. Furthermore, inhibiting the HKDC1-VDAC1 interaction at the mitochondrial membrane increases apoptosis, oxidative stress and DNA damage, leading to cell death. These findings suggest that targeting HKDC1 could provide therapeutic benefits in cancer. More studies are needed to fully assess the precise interactions of HKDC1 with signal transduction pathways and to determine if inhibiting the interaction between HKDC1 and VDAC1 produces similar results in other cancers. Furthermore, the plausibility of targeted delivery will be critical in understanding the utility of targeting HKDC1 interactions in improving clinical disease and outcomes.

In conclusion, HKDC1 is a widely expressed protein with HK activity that plays important roles in glucose homeostasis during pregnancy and aging, and has been linked to tumor progression and poorer prognosis in human cancer. HKDC1 has emerged as a critical component of both normal physiological processes as well as human disease states, such as diabetes and cancer. Therefore, an understanding of the precise mechanisms of action of HKDC1 are likely to provide important benefits for human health.

Highlights.

Hexokinase domain containing protein-1 (HKDC1) is a widely expressed novel hexokinase.

Recent studies aimed at understanding the functionality of HKDC1 have identified roles for HKDC1 that extend beyond glucose sensing and phosphorylation.

Genome wide association studies have identified genetic variants of HKDC1 that are associated with 2-hour blood glucose levels during an oral glucose tolerance tests performed in pregnancy, as well as with hemoglobin A1c, suggesting a link between HKDC1 and diabetes.

Mouse models indicate that HKDC1 is important for glucose utilization and homeostasis during times of metabolic stress.

Beyond glucose metabolism, recent studies suggest roles for HKDC1 in the survival, progression, and metastases of several human cancers, conferring a poorer prognosis.

Glossary

Genome wide association study (GWAS)

study that scans markers across complete sets of DNA, or genomes, from different individuals with the aim of identifying genetic variations associated with a particular disease.

Gestational diabetes mellitus (GDM)

glucose intolerance discovered for the first time during pregnancy and not prior to pregnancy.

Hexokinase

an enzyme that transfers an inorganic phosphate from ATP to six-carbon sugars (hexoses) including glucose. In the case of glucose, phosphorylation to form glucose-6-phosphate traps glucose within the cell, allowing it to be metabolized according to current cellular needs.

Integrated stress response

conserved pathway activated by cells in response to stress aimed at restoring cellular homeostasis.

miRNA

short for “microRNA”; a noncoding single-stranded RNA molecule that binds to a complementary sequence within mRNA, downregulating and modulating the expression of a particular gene.

Oral glucose tolerance test (OGTT)

medical test in which an oral glucose load is administered and blood is analyzed at various time points to assess the body’s ability to clear glucose. A test used to measure the body’s response to sugar.

Single nucleotide polymorphism (SNP)

a change in a single nucleotide in the genome occurring within a population that may be the result of a nucleotide substitution, deletion, or insertion, and can occur within coding, noncoding, or intergenic regions. Analysis of SNPs in a population can be used to assess our susceptibilities to various diseases.

Voltage dependent anion channel (VDAC)

porin ion channel located in the outer mitochondrial membrane that facilitates ion transfer between the cytosol and mitochondria.

Whole-exome sequencing (WES)

technique utilized for determining the nucleotide sequences of the protein coding regions of an individual’s genome.