Background
Diabetic nephropathy (DN) is an endemic complication of diabetes and the leading cause of end-stage renal failure. Clinical features of DN are progressive albuminuria, proteinuria, and an eventual reduction in the glomerular filtration rate [
1]. The complex progressive histopathological changes associated with DN include mesangial matrix expansion, thickening of basement membranes, glomerular and tubular hypertrophy, podocyte loss, and glomerulosclerosis and tubulointerstitial fibrosis [
2]. High glucose is a primary initiating factor of multiple molecular, metabolic, and hemodynamic changes resulting in kidney damage, including intrarenal tissue hypoxia [
3]. Tissue hypoxia activates multiple pathways, such as profibrotic growth factors, hemodynamic cytokines (angiotensin II), advanced glycation end products (AGE), and reactive oxygen species (ROS). Thus, both hyperglycemia and hypoxia are major determinators of the chronic complications associated with diabetes.
A master regulator of transcriptional responses to hypoxia is hypoxia inducible factor 1 (HIF-1). HIF-1 has been recently associated with the progression of chronic renal injuries including DN [
4‐
6]. HIF-1 consists of two subunits, HIF-1α, an O
2-labile subunit, and constitutively expressed HIF-1β [
7].
Hif1α
+/−
heterozygote mutants demonstrate impaired responses when challenged with hypoxic conditions after birth [
8,
9]. HIF-1 directly regulates the expression of more than 1000 human genes (for review, see [
10]). Although the expression of a subset of HIF-1 target genes is induced by hypoxia in most or all cell types, the majority of these genes are induced by hypoxia in a cell type–specific manner. In addition to hypoxia, the HIF-1α subunit activity is regulated by numerous other factors, including growth factors, cytokines, sirtuins, ROS, and intracellular metabolites, even under normoxic conditions [
11]. However, the mechanism of HIF-1α stabilization in a hyperglycemic environment is controversial. Hyperglycemia upregulates HIF-1α in the glomeruli of diabetic model mice regardless of the etiology of the diabetes [
5,
12]. The activation of HIF-1 in the diabetic kidney may be suboptimal despite profound renal hypoxia, as suggested by a large body of evidence showing that the diabetic milieu deregulates the HIF-1α pathway [
13‐
15]. It remains controversial whether the activation of HIF-1 signaling exerts a beneficial or harmful role in the progression of renal diseases, particularly DN. An indirect approach using YC-1 [3-(5′-hydroxymethyl-2′-furyl)-1-benzyl indazole], a HIF-1 inhibitor, reduced glomerular hypertrophy and AGE-tissue modifications in the type 1 diabetes mouse model [
6]. In contrast, an activation of HIF-1α by CoCl
2 reduced proteinuria and histological markers of kidney injury in an obese type 2 diabetes model [
16] and in STZ-induced DN in rats [
3].
To provide more insight into the functional role of HIF-1α pathways, we examine the relationship between diabetes-induced kidney injury and the partial deficiency of HIF-1α caused by the global deletion of the Hif1α functional allele with a specific focus on the early phase of diabetes-exposure. Together, our data suggest the potential roles of HIF-1α and Hif1α genetic variations in the manifestation of DN. Furthermore, our data point out the necessity of optimizing any possible pharmacological inhibition of HIF-1 in therapeutic applications of DN and diabetes-associated pathologies.
Discussion
Our partial deficiency
Hif1α model provides the first model that tests in vivo the function of HIF-1α in the development and progression of diabetes-induced renal damage. Previous work has only provided indirect evidence for the role of HIF-1α, using a HIF-1 inhibitor [
6] or HIF-1 activator [
16]. Our data extend previous findings that HIF-1α signaling is activated in the kidneys of experimental models with type I and type II diabetes and that it may be relevant to the development of DN [
4,
5,
12]. We examined the role of HIF-1α in the early stage of disease using the STZ-induced diabetic mouse model characterized by hyperglycemia (blood glucose levels > 13.9 mmol/L) and insulinopenia. We found that
Hif1α partial deficiency significantly accelerated the manifestation of pathological changes associated with the progression of DN. Changes in serum biochemical parameters associated with diabetic glomerular injury and progression of chronic kidney disease were more significant in diabetic
Hif1α
+/−
compared to diabetic
Wt mice. The combination of
Hif1α deficiency and diabetes resulted in an altered transcriptional expression profile of the renal cortex and decreased survival of podocytes.
Hypoxia represents an early and potentially initiating factor in the development and progression of chronic kidney diseases including DN [
4,
36]. HIF-1 mediates hypoxia-induced cellular responses through the regulation of genes involved in cell metabolism, glucose utilization, angiogenesis, oxidative stress, apoptosis, and proliferation. However, the activation of HIF-1 in the diabetic kidney may be suboptimal despite profound renal hypoxia, as suggested by a large body of evidence showing that the diabetic milieu deregulates the HIF-1α pathway [
13‐
15]. In recent years, HIF-1α genetic polymorphisms have emerged as potentially important determinants of disease severity and adverse outcomes [
37,
38]. Nonetheless, given the diversity of HIF-1 signaling, it remains controversial whether the activation of HIF-1 signaling exerts a beneficial or harmful role in the progression of renal diseases, particularly DN.
Persistent, chronic exposure to hypoxia is associated with structural tissue remodeling, such as renal fibrosis, inflammation, apoptosis and loss of microvasculature. HIF-1 signaling is an important protective physiological mechanism activated to counteract hypoxia and prevent renal damage (for review, see [
39]). For example, the global inactivation of the
Vhlh gene by the Cre-loxP system resulted in HIF-1α and HIF-2α stabilization and suppressed fibrogenesis in mice subjected to unilateral ureteral obstruction [
40]. Other studies using pharmacological approaches for systemic HIF-1 activation demonstrated improved proteinuria and histological parameters in experimental chronic kidney disease models [
41,
42]. In contrast, other studies have shown that sustained HIF-1 activation may have unfavorable effects. Genetic inactivation of the
Vhlh gene in tubular epithelial cells resulted in constitutive HIF-1α stabilization and accelerated renal fibrosis [
43]. Similarly, the genetic ablation of
Hif1α in the renal proximal tubule inhibited tubulointerstitial fibrosis in the in vivo model of unilateral ureteral obstruction [
44]. These data suggest that HIF-1α may play different roles in the progression of chronic kidney diseases depending on the mode of activation, cell-type specific action, and local versus global HIF-1α stabilization. Thus, these conflicting results reflect the complexity of the adaptive responses mediated by HIF-1.
Similar discrepancies have been reported regarding the role of HIF-1 in DN. An indirect approach using YC-1 [3-(5′-hydroxymethyl-2′-furyl)-1-benzyl indazole], a HIF-1 inhibitor, reduced glomerular hypertrophy and AGE in the type 1 diabetes mouse model [
6]. In contrast, an induction of HIF-1α by CoCl
2 reduced proteinuria and histological markers of kidney injury in an obese type 2 diabetes model [
16] and in STZ-induced DN in rats [
3]. In conjunction with these studies, our data demonstrate that a partial
Hif1α deficiency promotes the diabetes-induced kidney injury.
Hif1α partial deficiency was associated with a reduced expression of HIF-1-targeted genes
Pdk1, Ntn1, Ctgf, and
Fn1. Serum glucose levels were significantly increased in
Hif1α
+/−
mice compared to
Wt, implying systemic changes in glucose metabolism in association with
Hif1α partial deletion, which may contribute to the enhanced pathogenesis. HIF-1, by regulating the expression of glucose transporter GLUT1 and glycolytic enzymes, affects glucose homeostasis, including the regulation of glucose-stimulated insulin secretion (GSIS) from the pancreatic beta-cells [
45]. Targeted disruption of
Hif1α in pancreatic beta-cells resulted in glucose intolerance, impaired GSIS, and beta-cell dysfunction [
46]. Thus, the increased serum glucose levels in our diabetic
Hif1α
+/−
mice were in accordance with the changes in beta-cell function and impaired glucose homeostasis.
These changes were accompanied by glomerular damage, as indicated by a significant loss of podocytes and increased expression of podocin, a marker for podocyte damage, in the diabetic
Hif1α
+/−
renal cortex. These results suggest that HIF-1α functional impairment affected the survival of podocytes in the diabetes-exposed kidney. It is important to notice that systemic pharmacological approaches used in previous studies of DN [
3,
6,
16] may produce HIF-1-independent effects and may also affect other tissues resulting in different responses in diabetes-exposed kidneys.
In response to injury, mesangial cells transdifferentiate and synthesize different extracellular matrix proteins, which is an important pathological event during glomerulosclerosis and the progression of DN. The increased expression of transcription factor SOX9 has been associated with changes in mesangial cells and expansion of the mesangial area in the progression of DN [
28]. Additionally, the activation of SOX9 is critical for the early damage and repair response of injured renal tubule cells [
47]. This repair response in the chronically active form may represent an additional mechanism triggering long-term pathological responses resulting in kidney damage. Not only HIF-1 mediates
Sox9 expression, ERK1/2 signaling [
48] or BMP4 [
28] may also induce
Sox9 expression. Furthermore, advanced glycation end products (AGEs) have been shown to induce
Sox9 expression [
28]. Thus, we can postulate that increased
Sox9 expression in the diabetic
Hif1α
+/−
renal cortex may indicate a) an early transcriptional response to renal injury or/and b) regulatory compensatory response to
Hif1α deficiency and diabetic environment.
We found increased collagen accumulation in both diabetic
Hif1α
+/−
and
Wt mice. Correspondingly, the expression of markers of fibrosis and extracellular matrix accumulation,
Tgfβ1, fibronectin,
Ctgf, and α-SMA were increased in both diabetic
Hif1α
+/−
and
Wt mice. These results indicate that fibrosis in the diabetic kidney was not affected by the global reduction of
Hif1α, at least not in the early phase of diabetic exposure. In line with our observations are studies where the global
Hif1α deletion using the
Ubc-cre/ERT2 system did not affect collagen accumulation, although inflammation and renal injury were enhanced by
Hif1α deletion in the model of unilateral ureteral obstruction [
49].
VEGFA stimulates endothelial cell proliferation and has a key role in physiologic and pathologic angiogenesis in different tissues. In the kidney, VEGFA regulates glomerular permeability and maintenance of the glomerular tuft, and overall maintenance of kidney integrity [
50]. VEGFA is tightly regulated as shown by glomerular-selective overexpression or deletion of VEGFA resulting in severe and early renal pathologies [
33]. Renal diseases are frequently associated with impaired angiogenesis, capillary loss, and a reduction of VEGFA expression. In contrast, in diabetic nephropathy, renal VEGFA levels are elevated in experimental models as well as in diabetic patients [
51‐
53] The upregulation of VEGFA has been proposed as a contributing mechanism to renal dysfunction during the early phase of diabetes [
53,
54]. Inhibition of VEGFA at the onset of diabetes abolished the associated diabetes-glomerular hyperfiltration, glomerular hypertrophy, and urinary albumin excretion in the type I diabetes model [
53]. In our study, consistent with previously published data, VEGFA expression was significantly increased in the glomerulus of diabetic
Hif1α
+/−
compared to the diabetic
Wt, indicating a faster progression of renal dysfunction in diabetes (Fig.
4). The cause of the upregulation of VEGFA in the diabetic kidney remains speculative; however, multiple factors may be implicated [
53]. Renal dysfunction of diabetic
Hif1α
+/−
mice was further supported by the increased expression of
Adm in the diabetic
Hif1α
+/−
renal cortex. The upregulation of
Adm, which encodes a potent vasorelaxant peptide, is associated with glomerular hyperfiltration and dilatation of the glomerular capillaries in the acute phase of type 1 diabetes [
55]. Notably, serum albumin levels were significantly decreased in diabetic
Hif1α
+/−
mice (Fig.
1).
A limitation of our study is the global nature of the Hif1α deletion. We are unable to determine which cell type or which combinations of cell types are contributing to the increased susceptibility of Hif1α
+/−
mice to DN. The global deletion of Hif1α may affect other tissues and it may indirectly escalate pathological functional and structural changes in the kidney of Hif1α
+/−
mutants. At the same time, our model reproduces the conditions of a global inhibition of HIF-1 signaling, such as in pharmacological targeted-HIF-1 inhibition.