Role of GALNT2 on Insulin Sensitivity, Lipid Metabolism and Fat Homeostasis

O-linked glycosylation, the greatest form of post-translational modifications, plays a key role in regulating the majority of physiological processes. It is, therefore, not surprising that abnormal O-linked glycosylation has been related to several human diseases. Recently, GALNT2, which encodes the GalNAc-transferase 2 involved in the first step of O-linked glycosylation, has attracted great attention as a possible player in many highly prevalent human metabolic diseases, including atherogenic dyslipidemia, type 2 diabetes and obesity, all clustered on the common ground of insulin resistance. Data available both in human and animal models point to GALNT2 as a molecule that shapes the risk of the aforementioned abnormalities affecting diverse protein functions, which eventually cause clinically distinct phenotypes (a typical example of pleiotropism). Pathways linking GALNT2 to dyslipidemia and insulin resistance have been partly identified, while those for type 2 diabetes and obesity are yet to be understood. Here, we will provide a brief overview on the present knowledge on GALNT2 function and dysfunction and propose novel insights on the complex pathogenesis of the aforementioned metabolic diseases, which all impose a heavy burden for patients, their families and the entire society.


Introduction
O-linked glycosylation, the greatest diverse form of post-translational modifications, affects the folding, stability, processing and trafficking of most secreted proteins [1], thus playing a key role in regulating many physiological processes. It is, therefore, not surprising that abnormal O-glycosylation has been related to several human diseases [2]. GALNT2 encodes the polypeptide N-acetyl-galactosaminyl-transferase 2 (ppGalNAc-T2, also referred to as GalNAc-T2) [3,4], which transfers N-acetyl galactosamine to the serine/threonine hydroxyl group of specific acceptor peptide motifs [5][6][7] in the first step of O-glycosylation. In the last few years, GALNT2 has attracted great attention as a possible player in many human metabolic abnormalities that all share the common ground of insulin resistance, including atherogenic dyslipidemia [8][9][10][11][12][13][14], type 2 diabetes [15,16] and obesity [17,18]. After very briefly introducing insulin resistance, its role in several highly prevalent human diseases and how abnormal insulin signaling may cause it, we will offer to the readers an overview on what is known about GALNT2 function and dysfunction, with the aim of providing novel insights on the complex pathogenesis of the aforementioned metabolic diseases. which act as "docking molecules" for Src-Homology 2 domain-containing proteins, eventually enabling downstream signal transmission [38][39][40]. The effects of insulin on glucose metabolism are mediated by IRS1, which activates phosphatidyl inositol-3-kinase, which then produces phosphatidyl-inositol-3,4,5-triphosphate (PIP3) [38][39][40]. PIP3, in turn, activates the serine/threonine protein kinase beta Akt2, the central mediator of insulin's effects on glucose and lipid metabolism, including glucose uptake and the synthesis of glycogen and fatty acids [38][39][40]. Several genes encode molecules that inhibit this pathway by various mechanisms, including a direct interaction with the insulin receptor (ectonucleotide pyrophosphatase/phosphodiesterase 1, ENPP1) [41][42][43][44][45][46] or Akt2 (tribbles pseudokinase 3, TRIB3) [47], tyrosine dephosphorylation of the insulin receptor (protein tyrosine phosphatase non-receptor type 1, PTPN1), serine phosphorylation of IRS1 (tumor necrosis factor, TNF and interleukin 6, IL6) and 5 dephosphorylation of PIP3 (inositol polyphosphate phosphatase-like 1, INPPL1; also known as SH2-containing inositol phosphatase 2, SHIP2) [38][39][40]48,49]. In some individuals, insulin resistance and related traits are likely to be caused by abnormalities in this complex network that is modulated by both stimulatory and inhibitory molecules [38,39,48,49]. In addition, some studies point to insulin resistance at the level of the pancreatic β cells as an important determinant of impaired insulin secretion [50,51], the best predictor of future loss of glucose homeostasis [52,53]. Overall, abnormal insulin signaling, through a combined effect on both insulin resistance and deficient insulin secretion, becomes a key factor in the pathogenesis of type 2 diabetes, atherogenic dyslipidemia, metabolic syndrome and all related metabolic and cardiovascular abnormalities. Although our knowledge has improved over the last three decades [40], several gaps on the precise molecular events leading to the failure of insulin signaling still remain to be understood. Some novel cues linking insulin signaling, insulin resistance, atherogenic dyslipidemia and adipose tissue dysfunction will be provided in the following part of this article.

GALNT2 Gene, Protein Function and Targets
GALNT2, located on chromosome 1q42. 13 [4], belongs to a large family of homologous genes coding in humans for 20 different glycosyltransferases all involved in Oglycosylation [54]. As indicated by intron/exon positioning analyses [54], GALNT2 likely shares a common ancestor with the other GALNT genes [54] and exhibits more than 98% sequence similarity among species [3]. Of note, phylogenetic analyses suggest that the enzymatic function of ppGalNAc-T2 has also been highly conserved throughout evolution [55,56].
ppGalNAc-T2 participates in the initiation step of O-linked glycosylation [57,58] by covalently linking N-acetyl-D-galactosamine (GalNAc) to the hydroxyl group of serine or threonine on specific peptide acceptor motifs [5][6][7]. It is of note that in contrast to the other isoenzymes of the same family, which show distinct non-redundant substrate specificity [59][60][61], ppGalNAc-T2 serves a broad spectrum of target peptides [57], thus modifying a large number of substrates and contributing enormously to widen protein diversity [57]. In addition, GALNT2 is expressed in nearly all cell types and tissues [62] and contributes to the function of several proteins in an organ-specific manner [9,54,[63][64][65][66][67][68][69][70]. Based on these premises, it is not unexpected that GALNT2/ppGalNAc-T2 is considered by far the major contributor of the whole O-glycoproteome [57].

GALNT2 and Insulin Sensitivity
Evidence on the role of GALNT2 as a mediator of insulin sensitivity comes from our studies on human liver cells (HepG2) as well as on mouse fibroblasts (3T3L1) during their differentiation into mature adipocytes [83,84]. The reduction in GALNT2 expression in HepG2 cells impairs insulin signaling and action, as indicated by the reduced ability to stimulate insulin receptor, IRS-1 and protein kinase beta Akt2 phosphorylation [83] and to down-regulate the gluconeogenetic enzyme phosphoenolpyruvate carboxykinase (PEPCK) [83]. Coherently, data in pre-adipocytes show that GALNT2 over-expression significantly improves insulin signaling and, in turn, stimulates adipocyte maturation and enlargement [84]. Unfortunately, the mechanism through which GALNT2 affects insulin sensitivity is not yet known. In this regard, it is of note that in both HepG2 and 3T3L1 cells, the expression of GALNT2 is inversely correlated with that of ENPP1, a well-established negative modulator of insulin signaling [41][42][43][44][45][46]. This correlation makes it possible to hypothesize that ENPP1 down-regulation is likely to mediate, at least partly, the insulinsensitizing effect of GALNT2. Whether ENPP1 is a primary ppGalNAc-T2 target or its modulation is secondary to upstream O-glycosylation events is not yet known. Taken as a whole, our data on two different insulin responsive cells do suggest that GALNT2 is a positive modulator of insulin sensitivity.
A role of GALNT2 in modulating insulin sensitivity is also suggested by preliminary genetic studies indicating that GALNT2 variability is associated with HOMA-insulin resistance index in women with polycystic ovary syndrome [85].
Extreme conditions of GALNT2 loss of function (LOF), ranging from humans (homozygous individuals carrying the GALNT2 p.Phe104Ser or p.Gln289 * mutations) to genetically modified nonhuman primates and mice, affect HDL-C consistently across species [9,86]. Notably, GALNT2-LOF also increased TG levels in monkeys and mice but not in humans [9], thus proposing that GALNT2 modulates lipid metabolism in a species-specific manner.
Additionally, fine tuning of GALNT2 affects lipid metabolism. In detail, data from genome-wide studies highlight that the common rs4846914 SNP at the GALNT2 locus, which is associated with lower GALNT2 expression in human liver [9], is also associated with increased TG and lower HDL-C levels [14], a combination known as atherogenic dyslipidemia, which is a marker of insulin resistance and an established risk factor for coronary artery disease (CAD) [81,87]. Along the same line, data in peripheral white blood cells (PWBC) from individuals with a wide range of metabolic clinical conditions show that GALNT2 expression levels are correlated directly with HDL-C and inversely with TG serum levels [10]. Of note, GALNT2 mRNA levels remain significantly associated with TG but no longer with HDL-C in conditional analyses including both lipid fractions, thus suggesting the association with HDL-C is mainly mediated by the very much expected strong and inverse correlation between HDL-C and TG levels [88]. Finally, the GALNT2 promoter is hyper-methylated, a proxy of gene down-regulation, in patients with CAD and low HDL-C levels [19,89].
Overall, the available data depict a complex scenario compatible with the hypothesis that GALNT2 exerts pleiotropic effects [88] on different serum lipid fractions by glycosylating alternative substrates, including ANGPTL3, ApoC-III and PLTP [88].

GALNT2, Type 2 Diabetes and Hyperglycemia
Studies in human PWBC show that GALNT2 levels are reduced in obese patients with type 2 diabetes as compared to obese patients with no diabetes and even more to non-obese control individuals [15]. These human data resemble those reporting GALNT2 downregulation in liver of Goto-Kakizaki diabetic rats as compared to their normoglycemic counterparts [16]. The biology underlying these associations (e.g., is GALNT2 downregulation causing hyperglycemia or, vice versa, is hyperglycemia causing GALNT2 downregulation?) is clearly indicated by data in human cultured monocytes, where a high glucose concentration causes GALNT2 down-regulation [15]. These data together with those on the direct relationship between GALNT2 expression and insulin signaling [83,84] make it possible to hypothesize that hyperglycemia-induced GALNT2 down-regulation is part of the mechanism underlying hyperglycemia-induced insulin resistance [90].

GALNT2 and Obesity
Given the well-established role of insulin sensitivity in promoting fat mass expansion, it is not surprising that GALNT2 over-expression in pre-adipocytes, by improving insulin signaling, stimulates adipocyte maturation [84] and leads to enlarged mature adipocytes characterized by small and clustered LD [91]. However, it cannot be excluded that GALNT2 also affects adipogenesis independently of insulin signaling, possibly mediating coordinated changes in key genes affecting adipocyte maturation (such as Adig, Retn, Nr1h3, Ucp1, Slc2a4, Lipe, Fgf2 and Tcs22d3) [84]. Of note, GALNT2 expression is increased in subcutaneous adipose tissue from obese individuals [17], thus making it possible to hypothesize a cause-effect relationship, with GALNT2 expression changes contributing to the increase in adiposity. This hypothesis is further supported by models of GALNT2-LOF mutations in both cattle [18] and rodents [86] which, in addition to severe congenital abnormalities, show reduced body weight as compared to their wild-type counterparts [18,86].

Conclusions
Taken altogether, the data that we have reported here are quite consistent with the contribution of GALNT2 to several highly prevalent metabolic abnormalities sharing the common ground of insulin resistance, namely atherogenic dyslipidemia, type 2 diabetes and obesity. Unfortunately, they leave open the question on which mechanisms underlie some of the effects of GALNT2 on the aforementioned diseases. Pathways through which GALNT2 impairs lipid metabolism and insulin resistance have been partly identified, while those connecting GALNT2 to hyperglycemia/type 2 diabetes and obesity are only scarcely understood. With this caveat in mind, it is possible to envision a scenario in which GALNT2 operates through various mechanisms (Figure 1). The simplest, and probably too simple, picture is with GALNT2 changes inducing insulin resistance, which is pathogenic for virtually all the metabolic disorders mentioned above. Conversely, in a more complex and probably more realistic scenario, alternative pathogenic pathways, uniquely affecting different metabolic abnormalities, can be envisioned. This second hypothesis is facilitated by the knowledge that different degrees of GALNT2 changes affect differently specific protein functions in an organ-specific manner, thus causing clinically distinct phenotypes (pleiotropism) [88]. A clear paradigm of this pleiotropic effect comes from data showing that while mild changes in ppGalNAc-T2 activity have been associated only with dyslipidemia [9,10,13,14], very rare and severe LOF homozygous mutations, other than heavily impairing lipid metabolism [9,86,92], cause the multisystem GALNT2congenital disorder of glycosylation (CDG2T), which severely affects the central nervous system, muscle function, immunity control, several aspects of the endocrine system, and coagulation [18,86]. clinically distinct phenotypes (pleiotropism) [88]. A clear paradigm of this pleiotropic effect comes from data showing that while mild changes in ppGalNAc-T2 activity have been associated only with dyslipidemia [9,10,13,14], very rare and severe LOF homozygous mutations, other than heavily impairing lipid metabolism [9,86,92], cause the multisystem GALNT2-congenital disorder of glycosylation (CDG2T), which severely affects the central nervous system, muscle function, immunity control, several aspects of the endocrine system, and coagulation [18,86]. Figure 1. Changes in GALNT2 expression and/or function and insulin resistance, atherogenic dyslipidemia and obesity. Notes: GALNT2 promotes insulin action by down-regulating ectonucleotide pyrophosphatase/phosphodiesterase 1 (ENPP1) [83], a negative modulator of insulin signaling [41][42][43][44][45][46] (orange arrow). Whether ENPP1 is a primary ppGalNAc-T2 target or its downregulation is secondary to upstream O-glycosylation events is unknown (dashed box). The role of GALNT2 down-regulation in atherogenic dyslipidemia (low HDL-Cholesterol and/or high triglycerides) is best understood. This effect is mediated by the O-glycosylation of three established targets of ppGalNAc-T2 [9] (the product of GALNT2) (blue arrow). Although no data are currently available, it cannot be excluded that some of the deleterious effects of GALNT2 down-regulation on dyslipidemia are mediated by the concurrent deleterious effect on insulin signaling which, in turn, causes insulin resistance and low HDL-cholesterol/high-triglyceride levels [19]. Experimental data suggest that the effect of GALNT2 on adipocyte maturation, adipogenesis and eventually obesity is at least partly due to its positive modulation of insulin signaling [84]. Whether GALNT2 also affects adipogenesis through other mechanisms, possibly mediating coordinated changes in key genes affecting adipocyte maturation [84], is an alternative possibility that deserves further studies to be addressed (purple dashed arrow). Finally, a high glucose concentration decreases GALNT2 expression and this may play a role in hyperglycemia-induced insulin resistance (glucose toxicity) [15]. The mechanisms through which this down-regulation occurs are not known (green arrow, dashed box).
Overall, we must admit that, unfortunately, no single scenario is available within a solid framework for the many effects exerted by GALNT2 on the cluster of metabolic Figure 1. Changes in GALNT2 expression and/or function and insulin resistance, atherogenic dyslipidemia and obesity. Notes: GALNT2 promotes insulin action by down-regulating ectonucleotide pyrophosphatase/phosphodiesterase 1 (ENPP1) [83], a negative modulator of insulin signaling [41][42][43][44][45][46] (orange arrow). Whether ENPP1 is a primary ppGalNAc-T2 target or its down-regulation is secondary to upstream O-glycosylation events is unknown (dashed box). The role of GALNT2 down-regulation in atherogenic dyslipidemia (low HDL-Cholesterol and/or high triglycerides) is best understood. This effect is mediated by the O-glycosylation of three established targets of ppGalNAc-T2 [9] (the product of GALNT2) (blue arrow). Although no data are currently available, it cannot be excluded that some of the deleterious effects of GALNT2 down-regulation on dyslipidemia are mediated by the concurrent deleterious effect on insulin signaling which, in turn, causes insulin resistance and low HDL-cholesterol/high-triglyceride levels [19]. Experimental data suggest that the effect of GALNT2 on adipocyte maturation, adipogenesis and eventually obesity is at least partly due to its positive modulation of insulin signaling [84]. Whether GALNT2 also affects adipogenesis through other mechanisms, possibly mediating coordinated changes in key genes affecting adipocyte maturation [84], is an alternative possibility that deserves further studies to be addressed (purple dashed arrow). Finally, a high glucose concentration decreases GALNT2 expression and this may play a role in hyperglycemia-induced insulin resistance (glucose toxicity) [15]. The mechanisms through which this down-regulation occurs are not known (green arrow, dashed box).
Overall, we must admit that, unfortunately, no single scenario is available within a solid framework for the many effects exerted by GALNT2 on the cluster of metabolic abnormalities, all sharing the common ground of insulin resistance. Several data indicate that they may actually represent different effects with different mechanisms (e.g., pleiotropism), although it is possible that, for at least some of them, a pathogenic link mediated by abnormal insulin signaling and insulin resistance does actually exist.
A further lack of knowledge concerns the possibility of modulating GALNT2 with the drugs currently in use for other purposes (i.e., drug repositioning). To the best of our knowledge, there are no in vivo or in vitro studies that have specifically investigated the effect of any drug on GALNT2 expression or ppGalNAc-T2 enzyme activity. Given the potential role of GALNT2 on the aforementioned severe and highly prevalent disorders, studies addressing the potential role of drugs already available today on GALNT2 would certainly be useful and are, therefore, timely.
In all, several studies both in humans and animal models point to GALNT2 as a contributing molecule to highly prevalent metabolic abnormalities, including atherogenic dyslipidemia, type 2 diabetes and obesity, which all impose a heavy burden for patients, their families and society as a whole. Conversely, our knowledge of the intimate mechanisms through which GALNT2 exerts these effects is still poor. A full understanding of GALNT2 function and dysfunction is, therefore, mandatory and will certainly improve our comprehension of the aforementioned metabolic diseases. Such knowledge gain is a prerequisite for discovering new pathogenic targets to be addressed with new therapies, so as to hopefully ameliorate the strategies currently available to treat patients with related insulin-resistance abnormalities.

Acknowledgments:
We would like to thank Marco Giorgio Baroni (University of L'Aquila, Department of Clinical Medicine), for inviting us to write this review.

Conflicts of Interest:
The authors declare no conflict of interest.