Renal olfactory receptor 1393 contributes to the progression of type 2 diabetes in a diet-induced obesity model

Abstract

Olfactory receptors are G protein-coupled receptors that serve to detect odorants in the nose. Additionally, these receptors are expressed in other tissues, where they have functions outside the canonical smell response. Olfactory receptor 1393 (Olfr1393) was recently identified as a novel regulator of Na⁺-glucose cotransporter 1 (Sglt1) localization in the renal proximal tubule. Glucose reabsorption in the proximal tubule (via Sglt1 and Sglt2) has emerged as an important contributor to the development of diabetes. Inhibition of Sglt2 is accepted as a viable therapeutic treatment option for patients with type 2 diabetes and has been shown to delay development of diabetic kidney disease. We hypothesized that Olfr1393 may contribute to the progression of type 2 diabetes, particularly the development of hyperfiltration, which has been linked to increased Na⁺ reabsorption in the proximal tubule via the Sglts. To test this hypothesis, Olfr1393 wild-type (WT) and knockout (KO) mice were challenged with a high-fat diet to induce early-stage type 2 diabetes. After 16 wk on the high-fat diet, fasting blood glucose values were increased and glucose tolerance was impaired in the male WT mice. Both of these effects were significantly blunted in the male KO mice. In addition, male and female WT mice developed diabetes-induced hyperfiltration, which was attenuated in the Olfr1393 KO mice and corresponded with a reduction in luminal expression of Sglt2. Collectively, these data indicate that renal Olfr1393 can contribute to the progression of type 2 diabetes, likely as a regulator of Na⁺-glucose cotransport in the proximal tubule.

INTRODUCTION

Olfactory receptors (ORs) are seven-transmembrane domain G protein-coupled receptors that act as chemical detectors in the nose to sense smell (2, 3). It is now appreciated that ORs are also expressed “extranasally,” where they contribute to important physiological processes (8, 13, 28, 40). Recently, we determined that one of these receptors, Olfr1393, is found in all three segments (S1, S2, and S3) of the renal proximal tubule. This receptor responds to a subset of small molecules containing preconstrained carbon rings and was found to contribute to renal glucose reabsorption (37). Olfr1393 knockout (KO) mice fed a normal-chow diet are euglycemic, yet they present with glycosuria, indicating a defect in glucose handling. Furthermore, KO mice have improved glucose tolerance compared with their wild-type (WT) littermates (37).

Under normal conditions, glucose is freely filtered by the kidney and is completely recovered in the proximal tubule via two Na⁺-glucose cotransporters, Sglt1 and Sglt2 (24, 25, 30, 38, 44). In the kidney, Sglt2 handles the bulk of glucose reabsorption (~90%) in the convoluted proximal tubule, while the remaining glucose (~10%) is reabsorbed in the straight proximal tubule by Sglt1. Together, these two transporters ensure that no glucose is wasted in the urine. Studies in Sglt1 and/or Sglt2 KO mice have found that these mice present with euglycemic glycosuria and improved glucose tolerance, phenotypes that are remarkably similar to that of the Olfr1393 KO mice (20, 30, 44). Indeed, Olfr1393 KO mice have reduced levels of Sglt1 on the luminal membrane of the renal proximal tubule, suggesting that Olfr1393 may be a novel regulator of Sglt1 (37).

Recently, the Sglts have emerged as novel therapeutic targets for patients with diabetes (5, 24, 49, 55). The mechanism of these drugs is simple: inhibition of Na⁺-glucose transport impairs glucose reabsorption, leading to glycosuria. This ultimately results in lowered blood glucose levels, improved cardiovascular outcomes, and an attenuation of renal hyperfiltration. While patients with end-stage renal disease present with a decreased glomerular filtration rate (GFR), at the outset of diabetes patients often present with hyperfiltration, and those who do are biased toward the development of diabetic nephropathy and end-stage renal disease (5, 26, 46). This phenomenon is best explained by the tubular hypothesis of GFR (15, 27): briefly, as blood glucose levels rise, the Sglts act to increase glucose reabsorption in the proximal tubule. Because this is a Na⁺-coupled process, proximal tubule Na⁺ reabsorption is also increased, leading to a decrease in distal Na⁺ delivery to the macula densa. Because this decrease in distal Na⁺ delivery is “misinterpreted” to mean that the GFR is too low, GFR is increased (hyperfiltration) via the tubuloglomerular feedback mechanism. In this case, hyperfiltration is due to vasodilation of the afferent arterioles (5, 26, 46). Diabetic mouse models replicate the glomerular hyperfiltration observed in patients with diabetes, and inhibitors of Sglt2 have shown promise in the normalization of GFR (43, 45, 49).

Given that Olfr1393 KO mice present with glycosuria, improved glucose tolerance, and decreased amounts of Sglt1 on the lumen of the proximal tubule, we hypothesized that these mice might be protected from some of the effects of diabetes, including the development of diabetes-induced hyperfiltration. To examine this hypothesis, we challenged Olfr1393 WT and KO mice with a high-fat diet (HFD) for 16 wk. In keeping with the idea that Olfr1393 is a novel regulator of proximal tubule glucose reabsorption, we found that Olfr1393 KO mice are protected from the worsening of glucose tolerance seen in male WT mice on a HFD, and that HFD-fed Olfr1393 KO mice also exhibit an attenuated hyperfiltration response.

MATERIALS AND METHODS

Mouse model and HFD.

Olfr1393 WT and KO mice (males and females) were generated and described previously (37). Mice were maintained on a normal-chow diet (Teklad Global Rodent Diet T.2018SX.15, Envigo) for the first 7 wk and then switched to a HFD (diet D12492, Research Diets) that consisted of 60% kcal from fat, 20% kcal from carbohydrates, and 20% kcal from protein for 16 wk. To examine any changes that may occur at earlier stages of HFD feeding, a separate cohort was maintained on the HFD for 6 wk. Mice were maintained on a 14:10-h light-dark cycle, with food and water provided ad libitum. All studies performed on the WT and KO mice were age-matched and included both males and females; data are displayed as an average of all mice included in the study, unless otherwise noted. To track weight gain, HFD-fed animals were weighed at the beginning of each week (plotted in Fig. 1). At the conclusion of the study (after 16 or 6 wk on diet), the mice were euthanized by CO₂ asphyxiation. Kidneys were collected, weighed for calculation of the kidney weight-to-body weight ratio, and processed for histological analysis, immunohistochemistry, and triglyceride measurements (see below for detailed methods). These values are listed in Table 1. The liver was also collected for triglyceride measurements. All animal experiments were approved and were performed in accordance with the policies and procedures of the Johns Hopkins University and Georgetown University Institutional Animal Care and Use Committees.

Fig. 1.

HFD-fed Olfr1393 KO mice have similar weight gain. Male and female Olfr1393 wild-type (WT) and KO mice were fed a HFD for 16 wk, and weights were tracked weekly. Values are means ± SE; n = 4 male and 3 female WT and 6 male and 7 female KO mice.

Table 1.

Body and kidney weight, fasting blood glucose, and kidney and liver triglycerides in HFD-fed Olfr1393 WT and KO mice

			Male		Female
	WT	KO	WT	KO	WT	KO
Body weight, g§	41.9 ± 3.2 (n = 7)	37.3 ± 2.5 (n = 10)	47.0 ± 1.8† (n = 4)	41.5 ± 2.2† (n = 6)	35.2 ± 5.3 (n = 3)	33.8 ± 4.2 (n = 4)
Kidney weight/body weight	0.009 ± 0.0008 (n = 7)	0.010 ± 0.0006 (n = 10)	0.008 ± 0.0009 (n = 4)	0.009 ± 0.0002 (n = 6)	0.009 ± 0.0015 (n = 3)	0.010 ± 0.0011 (n = 4)
Fasting blood glucose, mg/dl	164.1 ± 23.1 (n = 7)	127.0 ± 7.8 (n = 12)	207.8 ± 15.6‡ (n = 4)	140.6 ± 14.2* (n = 5)	106.0 ± 17.5 (n = 3)	117.3 ± 7.6 (n = 7)
Kidney triglycerides, mg/g protein	130.3 ± 44.7 (n = 6)	76.1 ± 14.3 (n = 10)	77.0 ± 37.7 (n = 3)	49.8 ± 12.5 (n = 3)	183.6 ± 75.6 (n = 3)	87.3 ± 18.6 (n = 7)
Liver triglycerides, mg/g protein	219.4 ± 38.3 (n = 7)	137.4 ± 30.3 (n = 13)	254.0 ± 48.9 (n = 4)	165.7 ± 55.3 (n = 6)	173.2 ± 60.2 (n = 3)	113.2 ± 31.8 (n = 7)

Values are means ± SE; n, no. of mice. Olfr1393, olfactory receptor 1393; WT, wild-type; KO, knockout; HFD, high-fat diet.

^†P = 0.018 (no differences among genotypes).

^‡P ≤ 0.001 (male WT vs. female WT).

^*P = 0.02 (male WT vs. male KO). §Body weight measured at time of euthanasia.

Glomerular filtration rate.

GFR was determined in conscious, unrestrained mice by transcutaneous measurement of FITC-sinistrin (Mannheim Pharma & Diagnostics, Mannheim, Germany) after 0, 11.5, and 16 wk on the HFD, as previously described (36, 37). All GFR measurements were performed between 8AM and 11AM. Briefly, on the day before the procedure, light anesthesia was induced using isoflurane, and hair on a small area on the back of each mouse was removed using a depilatory (Nair). On the following day, the transcutaneous measurement device was attached to the hairless region of the back and a baseline (1 min) measurement was obtained; then the mice were injected with FITC-sinistrin (Fresenius Kabi, Graz, Austria) retroorbitally. Mice weighing <30 g were injected with 15 mg of FITC-sinistrin per 100 g body weight, while those >30 g body weight were injected with 12.75 mg of FITC-sinistrin per 100 g body weight to ensure that the measuring device could detect the peak levels of circulating sinistrin. Mice were returned to their cage and had free access to food and water for 1 h; thereafter, the measuring device was removed and data were downloaded. GFR was determined by the decay rate of FITC-sinistrin using a three-compartment model that analyzes the entire FITC-sinistrin decay curve (7). Because we measured GFR in the same animal at different time points and the body weight of the animals changed over time, we chose to calculate GFR as microliters per minute, as described previously (16). To do so, we calculated the rate constant using the calculated half time (t_1/2) of the curve [ln(2)/t_1/2] and approximated the extracellular fluid volume (ECVS) (14,616.8/100 × body weight) for each animal. The complete calculation is as follows: GFR (µl/min) = calculated ECVS × rate constant. Detailed calculations for this method are available elsewhere (36).

Glucose and insulin tolerance tests.

After 16 wk on the HFD, mice were subjected to glucose and insulin tolerance tests (GTT and ITT). For the GTTs, mice were fasted overnight for 16 h and injected with glucose (1 g/kg body wt ip; Sigma). For the ITTs, mice were fasted for 2 h and injected with recombinant human insulin (0.7 U/kg body wt; GIBCO). All GTTs were performed between 9AM and 11AM; the ITTs were performed between 10AM and 12PM. After injection, blood was collected from a tail nick, and glucose values were measured with a glucometer (Accu-Chek Nano, Roche) at 0, 15, 30, 60, 90, and 120 min postinjection. Fasting blood glucose values were obtained from time 0 of the GTT.

Kidney and liver triglyceride measurements.

To measure total liver and kidney triglycerides, 75 mg of frozen tissue were homogenized in 0.5 ml of lysis buffer containing 150 mM NaCl, 0.25 M sucrose, 50 mM HEPES, and 0.1% IGEPAL, pH 7.0. Lysates were cleared at 700 g for 15 min, and 10 µl of lysate were used with 150 µl of triglyceride liquid stable reagent (Thermo Fisher). After the sample was incubated at 37°C for 15 min, absorption at 540 nm (Abs540) was measured. To calculate the triglycerides in mg/dl, Abs540 was compared with a standard curve of known triglyceride values (triglyceride standard, Pointe Scientific). All measurements were performed in triplicate. To normalize tissue triglycerides, protein lysate concentrations were calculated using the Bio-Rad protein assay according to the manufacturer’s protocol. Tissue triglycerides were calculated as milligrams of triglycerides per gram of protein.

Histology analysis.

At the conclusion of the study, kidneys were obtained from the WT and KO mice, drop-fixed in 10% buffered formalin, embedded in paraffin, and processed for hematoxylin-and-eosin staining by AML Laboratory (St. Augustine, FL). Light images were obtained on a fluorescence microscope (model BZ-X700, Keyence).

Immunohistochemistry.

At the conclusion of the study, kidneys were obtained from the WT and KO mice, drop-fixed in 10% buffered formalin, and embedded in optimum cutting temperature compound. Cryosections (8 µm) were stained for Sglt1 and Sglt2, as described previously (23). Briefly, sections were subjected to antigen retrieval in a citrate buffer, pH 6.0, and then permeabilized using Triton X-100 (0.5% for 15 min and 2% for 30 min). The sections were blocked in 1% BSA and then incubated with polyclonal Sglt1 (1:25 dilution) or Sglt2 (1:500 dilution) (44) antibody overnight at 4°C. The sections were washed with PBS, and staining was detected with Alexa Fluor 594 secondary antibodies. Proximal tubules were labeled with Lotus tetragonolobus lectin (LTL; Vector Laboratories, Burlingame, CA) and postfixed with 4% paraformaldehyde. Confocal images were taken with a confocal microscope (Leica SP8) at ×63, and the z stacks were compressed with Leica LAS software. To assess differences in localization of Sglt1 and Sglt2, luminal quantification was performed in ImageJ (National Institutes of Health, Bethesda, MD), as previously described (37). Briefly, random cortical fields from each kidney section were visualized, and the average pixel intensity was determined for 1) the luminal region and 2) the whole tubule for each LTL⁺/Sglt⁺ tubule within the field of view. The ratio of the luminal to total Sglt stain was determined for each tubule. In addition, to assess differences in luminal expression of Sglt1 and Sglt2, the average pixel intensity of the luminal membrane (as determined above) was normalized to the total area of each tubule. The localization and expression ratios from all tubules analyzed per kidney (n = 12–22 tubules) were averaged (for Sglt1, n = 6 WT and 7 KO; for Sglt2, n = 6 WT and 6 KO).

Statistics.

Values are means ± SE. Student’s t-test was used to compare two values (i.e., WT vs. KO for each time point of the GTT and ITT); significance was set at P < 0.05. When multiple comparisons were required (i.e., sex and genotype differences for GTT area under curve, ITT area under curve, fasting blood glucose levels, body weight, and kidney weight-to-body weight ratio), a two-way ANOVA was performed followed by Tukey’s post test; significance was set at P < 0.05. For the GFR data, where sex, genotype, and different time points were compared, a three-way ANOVA followed by Tukey’s post test was used; significance was set at P < 0.05. ANOVA tests were performed in SigmaPlot.

RESULTS

Inhibition of renal glucose reabsorption has emerged as a therapeutic strategy for patients with diabetes. We previously reported that Olfr1393 is required for complete glucose reabsorption; thus we hypothesized that it might contribute to the progression of diabetes. To examine the effects of Olfr1393 in the context of diabetes, Olfr1393 WT and KO mice were challenged with a HFD for 16 wk to promote diet-induced obesity. This rodent model has been shown to recapitulate the progression of early-stage type 2 diabetes, especially in males, and is known to induce significant weight gain, hyperglycemia, and hyperfiltration after only 9 wk on the diet (4, 16, 50).

Changes in body weight in male and female WT and KO mice.

Olfr1393 WT and KO mice exhibited steady weight gain over the course of the study (Fig. 1), and after 16 wk the males weighed 48.8 ± 1.9 (SE) g (WT) and 43.4 ± 2.3 g (KO), whereas the females weighed 35.8 ± 6.2 g (WT) and 34.1 ± 4.1 g (KO). In both males and females, there was no significant difference in weight gain between WT and KO mice, although the KO males consistently trended lower than the WT males in the later weeks of the study (weeks 9–16; Fig. 1). Sex differences in response to HFD feeding have been reported, and it is well documented that female mice show an overall diminished response (18). This likely accounts for the differences in weight gain between WT males and females (Fig. 1). Despite the difference in weight gain and the known sex differences in response to a HFD, we did not detect sex differences for the majority of the parameters that were measured in this study (determined by ANOVA). Thus, for clarity, we present the results of our data in both combined form (males and females) and separated by sex.

Glucose and insulin tolerance in HFD-fed WT and KO mice.

To assess the development of diabetes, WT and KO mice were subjected to a GTT. After 16 wk on the HFD, WT mice presented with impaired glucose tolerance, as indicated by their inability to return to baseline blood glucose levels 2 h after the glucose challenge (Fig. 2A). In contrast, KO mice performed significantly better than their WT littermates: their peak blood glucose level was lower, and their blood glucose values recovered at a faster rate (Fig. 2A). This finding is consistent with our previous finding that KO mice (fed a standard diet) can handle a glucose challenge better than their WT littermates (37), and it appears that the HFD further amplifies the previously reported KO phenotype.

Fig. 2.

Glucose tolerance is improved in high-fat diet (HFD)-fed olfactory receptor 1393 (Olfr1393) knockout (KO) mice. A–C: Olfr1393 KO (black lines) and wild-type (WT, gray lines) mice were subjected to a glucose tolerance test (GTT) after 16 wk on the HFD. Olfr1393 KO mice presented with a significant improvement compared with their WT littermates. This improvement in glucose tolerance was driven by changes in the male mice (B). Male WT mice presented with impaired glucose tolerance, which was significantly attenuated in the KO mice, whereas female WT and KO mice performed similarly (C). *P < 0.05 (by Student’s t-test). D: area under the curve (AUC) for combined male and female data (all), as well as male and female curves. KO AUCs were significantly lower in combined (all) and male groups; no difference was observed in females. AUC was also significantly lower for female WT than male WT mice. *P < 0.05 (by 2-way ANOVA with Tukey’s post test). Values are means ± SE; n = 7 (4 male and 3 female) WT and 12 (5 male and 7 female) KO mice.

Notably, however, this genotype difference was driven by changes in the male mice. Glucose intolerance was observed in male WT mice and significantly attenuated in male KO mice (Fig. 2B); however, there were no differences in the GTT response between the female WT and KO mice (Fig. 2C). The area under the curve (AUC) for the GTTs reconfirmed the observed sex differences (Fig. 2D). AUC was significantly decreased in KO mice compared with their WT littermates, indicating improved glucose tolerance, and this was driven by the differences between the male WT and KO mice. In addition, fasting blood glucose values were significantly lower in male KO than WT mice, as noted at time 0 of the GTT (Table 1).

In addition to the GTT, male and female mice were subjected to an ITT (Fig. 3). Consistent with our previous data from mice fed a normal-chow diet, both Olfr1393 WT and KO mice remained insulin tolerant after 16 wk on a HFD (Fig. 3A), and there were no differences between genotypes in the males (Fig. 3B) or females (Fig. 3C). Because AUC was significantly lower in the female than male mice, the female mice were considered to be more insulin-tolerant (Fig. 3D). This is consistent with the sex differences that we noted for the GTT.

Fig. 3.

HFD-fed olfactory receptor 1393 (Olfr1393) wild-type (WT) and knockout (KO) mice do not develop insulin intolerance. A–C: Olfr1393 KO (black lines) and WT (gray lines) mice were subjected to an insulin tolerance test (ITT) after 16 wk on the HFD. WT and KO mice performed similarly and did not present with insulin intolerance. When ITT data were analyzed by sex and plotted for males and females (B and C), no change was detected for either sex. D: area under the curve (AUC) for combined male and female data (all), as well as male and female curves. No genotype differences were detected, although AUC was significantly higher for male WT than female WT mice, and for male KO versus female KO mice. *P < 0.05 (by 2-way ANOVA with Tukey’s post test). Values are means ± SE; n = 7 (4 male and 3 female) WT and 13 (6 male and 7 female) KO mice.

Hyperfiltration in HFD-fed WT and KO mice.

To track the development of hyperfiltration in our mice over time, we measured GFR by plotting the transcutaneous decay of FITC-sinistrin, as previously described (36, 37). This method allows for real-time GFR measurements in conscious, unrestrained mice. Since this is not a terminal procedure, it also allows us to track changes in GFR over time within the same animal. Because previous studies determined that diabetes-induced hyperfiltration can be detected between weeks 9 and 12 of a HFD (16, 50), we measured GFR at 0, 11.5, and 16 wk. Consistent with our earlier findings (37), there was no difference in GFR between WT and KO mice at baseline (week 0; Fig. 4). However, in WT mice, HFD consumption led to a gradual increase in GFR, with significant hyperfiltration after 16 wk (WT: P < 0.001, baseline vs. 16 wk). While Olfr1393 KO mice also reached a state of hyperfiltration by 16 wk on the diet (KO: P < 0.001, baseline vs. 16 wk), hyperfiltration was significantly blunted compared with the WT mice (Fig. 4A; P = 0.037, WT vs. KO at 16 wk). There were no sex differences in the development of hyperfiltration in WT mice or in the attenuation of the phenotype in the KO mice; hyperfiltration was clearly observed in both male (Fig. 4B; P < 0.002, baseline vs. 16 wk) and female (Fig. 4C; P < 0.04, baseline vs. 16 wk) WT mice, and the protection was evident in both male and female KO mice.

Fig. 4.

Attenuated hyperfiltration response in high-fat diet (HFD)-fed olfactory receptor 1393 (Olfr1393) knockout (KO) mice. Glomerular filtration rate (GFR) was measured in Olfr1393 wild-type (WT) and KO mice at 0, 11.5, and 16 wk on the HFD. A: after 16 wk, Olfr1393 WT mice presented with hyperfiltration, and this response was significantly attenuated in KO mice. B and C: when the GFR response was analyzed by sex and plotted for males and females, no sex differences were detected; male and female WT and KO mice developed hyperfiltration after 16 wk on the HFD, but this response was attenuated in the KO mice. *P < 0.05, WT vs. KO at 16 wk; #P < 0.05, WT at 16 vs. 0 wk; +P < 0.05, KO at 16 vs. 0 wk (by 3-way ANOVA with Tukey’s post test). Values are means ± SE; n = 7 (4 male and 3 female) WT and 13 (6 male and 7 female) KO mice.

Sglt localization in HFD-fed WT and KO mice.

We previously reported decreased Sglt1 expression on the lumen of the renal proximal tubule in Olfr1393 KO mice fed a normal-chow diet (37). To examine expression and localization of Sglt1 and Sglt2 in HFD-fed mice, kidneys from WT and KO mice were harvested at the conclusion of the study and stained for these transporters (Fig. 5, A and B). Contrary to the response of the mice fed a normal-chow diet (37), there was no detectable difference in luminal localization of Sglt1 (Fig. 5C), nor was localization of Sglt2 altered (Fig. 5D) between WT and KO mice when normalized to total tubule fluorescence intensity. While the Sglt antibodies have great fidelity in staining of the brush border membrane, they can yield a nonspecific background stain in other regions of the cell (1, 44). To avoid including contributions of nonspecific staining in our analysis, we also quantified luminal expression by normalizing luminal fluorescence intensity to the total area of the tubule. Using this alternate analysis, we did not detect differences in Sglt1 expression in male or female KO mice (Fig. 5E). However, we did note a significant decrease in Sglt2 luminal expression of ~30% in the KO mice; this trend was noted in the male and female mice (Fig. 5F).

Fig. 5.

Expression and localization of Na⁺-glucose cotransporters 1 and 2 (Sglt1 and Sglt2) after 16 wk on the high-fat diet (HFD). A and B: after 16 wk on the HFD, kidneys from wild-type (WT) and knockout (KO) mice were excised, fixed, and stained for Sglt1 and Sglt2. Scale bar = 50 µm. C–F: luminal localization (C and D) and total luminal expression (E and F) were quantified as described in materials and methods. No differences in Sglt1 localization or expression were detected between WT and KO mice for either sex, although analysis of male and female data together shows a ~30% reduction in Sglt2 luminal expression. *P < 0.05 (by 2-way ANOVA with Tukey’s post test). Values are means ± SE; n = 6 (3 male and 3 female) Sglt1 WT, 7 (4 male and 3 female) Sglt1 KO, 6 (3 male and 3 female) Sglt2 WT, and 6 (3 male and 3 female) Sglt2 KO.

The genotype difference in Sglt1 luminal localization we previously reported in mice fed a normal chow diet was not observed after 16 wk on the HFD (Fig. 5). Because Sglt1 and Sglt2 are known to be upregulated under conditions of HFD feeding (38, 42, 47), we wondered if the genotype difference in Sglt1 localization had been muted by this upregulation. To examine this, we fed a separate cohort of WT and KO mice the HFD for only 6 wk. Previous studies reported that, by 6 wk, high-fat feeding leads to hyperglycemia but not overt glucose intolerance (16, 34). As such, we reasoned that compensation of transporter expression may not yet have occurred at this time point. Nevertheless, we did not detect major changes in Sglt1 or Sglt2 expression or localization between WT and KO mice at 6 wk (Fig. 6); the only genotype difference was a decrease in the ratio of Sglt2 luminal to total expression in male KO vs. male WT mice.

Fig. 6.

Expression and localization of Na⁺-glucose cotransporters 1 and 2 (Sglt1 and Sglt2) after 6 wk on the high-fat diet (HFD). A and B: after 6 wk on the HFD, kidneys from wild-type (WT) and knockout (KO) mice were excised, fixed, and stained for Sglt1 and Sglt2. Scale bar = 50 µm. C–F: luminal localization (C and D) and total luminal expression (E and F) were quantified as described in materials and methods. No differences were detected in Sglt1 localization (C), although total expression of Sglt1 was significantly higher in female KO than male KO mice (E). Luminal localization of Sglt2 was higher in male WT than male KO and female WT mice (D), although no differences in Sglt2 luminal expression were detected (F). *P < 0.05 (by 2-way ANOVA with Tukey’s post test). Values are means ± SE; n = 7 (3 male and 4 female) Sglt1 WT, 7 (4 male and 3 female) Sglt1 KO, 6 (3 male and 3 female) Sglt2 WT, and 7 (4 male and 3 female) Sglt2 KO.

Histology and triglycerides in HFD-fed WT and KO mice.

At the conclusion of the study, kidneys were obtained for histological analysis. In the male mice, we observed a number of microvacuoles within the tubules, which have been described previously in HFD-fed mice (54); no other gross morphological differences were detected in hematoxylin-eosin-stained kidneys from WT or KO mice (Fig. 7), nor were there differences in the kidney weight to body weight ratio (Table 1). We also did not detect significant differences in liver and kidney triglycerides, although the KO mice trended slightly lower in both of these measurements (Table 1).

Fig. 7.

Histological analysis of kidneys from high-fat diet (HFD)-fed olfactory receptor 1393 (Olfr1393) wild-type (WT) and KO mice. After 16 wk on the HFD, kidneys were excised, fixed, and stained with hemotoxylin and eosin. No gross morphological changes were detected.

DISCUSSION

Olfr1393 contributes to the development of type 2 diabetes.

Previously, we identified renal Olfr1393 as a novel regulator of Sglt1 in the proximal tubule. Here, we report that Olfr1393 contributes to the progression of type 2 diabetes in a diet-induced obesity model, possibly due to alterations in Sglt2 luminal expression. After 16 wk on a HFD, we found improved glucose tolerance and decreased fasting blood glucose levels in male Olfr1393 KO mice. Additionally, we found that both male and female Olfr1393 KO mice developed an attenuated hyperfiltration response and decreased luminal expression of Sglt2. Collectively, this study has identified a novel contributor to type 2 diabetes in the diet-induced obesity model.

Our studies were designed to directly compare HFD-fed Olfr1393 WT and KO mice, and our data from WT mice are well aligned with other reported HFD values in the literature. It is well documented that C57BL/6 male mice weight gain and develop impaired glucose tolerance when challenged with a HFD. After 16 wk on the diet, it has been shown that male mice gain ~25 g of body weight (53). On the other hand, female mice gain weight at a slower rate: ~13 g after 11 wk (11) and ~17 g after 20 wk (9) on the diet. These values are comparable to our WT male and female mice: ~28 g in male WT and ~17 g in female WT mice. For intraperitoneal GTTs, peak blood glucose values in male mice have been observed to reach ~400 mg/dl, which also matches our findings (peak blood glucose level in male WT mice = 519.3 ± 30.2 mg/dl) (11). Finally, studies from our laboratory and others showed hyperfiltration in HFD-fed male mice as early as 9 wk (300–400 μl/min) (16, 50). This is consistent with our observation of hyperfiltration in our WT mice after 16 wk on the diet (400 μl/min). At the conclusion of our studies, our HFD mice were ~6 mo of age. It has been shown that GFR typically decreases during the aging process in humans and rodents (31, 35, 41), although mice in the equivalent age range used in the present study maintain a consistent GFR (~240 μl/min at 10 wk compared with 280 μl/min in 6-mo-old male mice) (14). Thus it is quite clear that HFD feeding induced hyperfiltration and early-stage type 2 diabetes in Olfr1393 WT male mice and that the absence of Olfr1393 in males and females is sufficient to blunt this glomerular hyperfiltration.

Hyperfiltration mechanism.

Diabetes-induced hyperfiltration is thought to arise due to increased glucose filtration in the glomerulus and subsequent increased reabsorption of Na⁺ and glucose in the proximal tubule via Sglt1 and Sglt2. This, in turn, alters the tubuloglomerular feedback mechanism due to decreased Na⁺ delivery to the macula densa (24, 25, 46, 49). Clinical and experimental trials have shown that Sglt2 inhibition, which decreases glucose reabsorption in the proximal tubule, can restore GFR and significantly blunt diabetes-induced hyperfiltration by increasing the distal delivery of Na⁺ (24, 25, 46, 49). Based on the decrease in Sglt2 luminal expression alongside our previous data identifying Olfr1393 as a regulator of Sglt1, we propose that Olfr1393 KO mice have a reduced capacity to handle glucose in the proximal tubule, which leads to a significantly attenuated rise in GFR on the HFD. In one recent study using a Sglt2 inhibitor in diabetic Akita mice, Vallon and colleagues observed a reduction in GFR from ~400 to 250 μl/min with treatment (43). These values are similar to those of our HFD-fed WT and KO mice and lend credence to our hypothesis (Fig. 4).

While we previously observed a significant decrease in luminal expression of Sglt1 in Olfr1393 KO mice fed a normal-chow diet (37), this finding was not consistent with findings in HFD-fed KO mice (Figs. 5 and 6). Several studies have suggested that Sglt1 expression increases in conditions of hyperglycemia, as in the case of a HFD feeding (38, 42, 47). As such, it is possible that HFD-fed Olfr1393 WT and KO mice upregulate luminal expression of Sglt1, and thus an altered localization cannot be detected using the antibodies and quantification technique used in the present study. Additionally, although we did not detect changes in Sglt1 localization after 6 or 16 wk on the HFD, we cannot rule out the possibility that such changes may be apparent at other intermediate time points. After 16 wk on the HFD, however, we did note a decrease in Sglt2 luminal expression in Olfr1393 KO mice (Fig. 5), which may partially explain the attenuation of diabetes-induced hyperfiltration in these mice (Fig. 4). It is also possible that there are differences in Sglt1 (and/or Sglt2) activity in the KO mice that are independent of changes in localization; clearly, future studies are required to fully elucidate the effect of Olfr1393 signaling on Sglt1 in the kidney.

Sex differences.

Our studies were performed using cohorts that included male and female mice, and, as expected, there were clear differences between the male and female WT mice in response to the HFD. Overall, the female WT mice were much less susceptible to the diet: body weight and fasting blood glucose values were significantly reduced in the female cohort compared with the male WT mice. The females were also more tolerant of both glucose and insulin based on the AUCs for the GTTs and ITTs (Figs. 2 and 3). This finding is consistent with previous studies showing sex differences in response to a HFD (18). Nevertheless, despite a blunted diabetic phenotype in the WT female mice, our WT female mice did develop glomerular hyperfiltration after 16 wk on the diet. Notably, there were no sex differences at any time point with respect to GFR. Recently, a thorough analysis of the differences between male and female proximal tubule transporters determined that female rats rely more heavily on Sglt2 than on other Na⁺ transporters, including the Na⁺/H⁺ exchanger NHE3, for Na⁺ reabsorption (21). Inasmuch as diabetes-induced hyperfiltration is thought to be due to increases in proximal tubule Na⁺ reabsorption, this reliance on Sglt2 in females has the potential to lead to hyperfiltration, even in the absence of overt glucose intolerance, which may explain the discrepancy between the female glucose tolerance and GFR responses in the present study. While there were no genotype differences in glucose tolerance in HFD-fed female mice, both WT and KO female mice appear to have developed mild glucose intolerance compared with previously published data from mice fed a normal-chow diet (37). Thus the HFD-induced increase in the filtered glucose load in female (and male) mice must be sufficient to drive glomerular hyperfiltration. This same finding was previously described in a db/db diabetic mouse model: using radiolabeled creatinine to estimate GFR in female diabetic mice, Gartner observed a marked increase in hyperfiltration that did not correlate with the severity of hyperglycemia (10).

We previously observed sex differences between Olfr1393 WT and KO mice fed a normal-chow diet (37). In particular, the glycosuria phenotype was more prominent in the female than the male mice, suggesting that Olfr1393 has a more robust role in the female mice. It was previously noted by others that Sglt1 and Sglt2 protein expression is greater in female than male rats (1, 32, 33). It has also been shown that estrogen can play a renoprotective role in female mice, limiting the complications of diabetes (31). Although Olfr1393 is not activated by any known sex hormone (37), it is possible that the sex differences in the HFD-fed mice may be due to a combination of circulating hormone levels and a strong role for Olfr1393 in female mice. This would explain the sex differences in GTT data (estrogen plays a protective role against the development of type 2 diabetes, protecting WT and KO female mice) and the apparent lack of a sex difference in GFR (Olfr1393 plays a stronger role in female mice). Clearly, a more thorough investigation of the role of Olfr1393 in these sex differences is needed.

Sglt1 broad tissue expression.

While Sglt2 expression appears to be confined to the kidney tubule, Sglt1 expression is more varied. In particular, Sglt1 is also known to be expressed in the small intestine, liver, and pancreas, all of which can contribute to the development of diabetes and glucose tolerance (23). We previously reported that Olfr1393 also has a broad distribution, in both the liver and small intestine, but not in the pancreas (37). Although glomerular hyperfiltration is a kidney-specific phenotype, the Olfr1393 KO mice used in this study are whole animal KO mice; thus it is possible that some of the metabolic phenotypes observed particularly in our male mice (i.e., glucose tolerance) may be explained by an extrarenal role for this receptor.

Therapeutic potential.

The finding that Olfr1393 contributes to the progression of diabetes-induced hyperfiltration is potentially a clinically important one. While diet-induced obesity is an excellent model that mimics the progression of type 2 diabetes, including the development of hyperfiltration, in the future it will be important to examine the influence of Olfr1393 on other diabetic models as well. Additional models of type 2 diabetes, including the leptin signaling-deficient ob/ob and db/db mice, are of particular interest. Additionally, it would be interesting to examine the effects of Olfr1393 in the context of type 1 diabetes with use of the streptozotocin-induced diabetic model or the Akita mouse to more fully understand the therapeutic potential of this OR. In particular, it would be of great interest to use a model in which female control mice do exhibit impaired glucose tolerance to determine if the absence of Olfr1393 is beneficial in that setting.

Olfr1393 is a member of the G protein-coupled receptor superfamily, which is the largest family of “druggable” proteins (48). It remains to be determined how, exactly, Olfr1393 is regulating the Sglts and whether this regulation is specific to the kidney. However, we hypothesize that this regulation likely involves cAMP. ORs typically increase cAMP levels upon activation, and cAMP has also been shown to increase Sglt1 and Sglt2 protein expression and/or localization (12, 17, 19, 29, 38, 52). Thus, in WT animals, Olfr1393 may function to ensure that Sglt expression and function are maximized. In the absence of Olf1393, a decrease in Sglt function would result in a mild glycosuria and a concurrent increase in NaCl delivery to the macula densa, thereby attenuating the hyperfiltration associated with diabetes. Similarly, one can imagine that decreasing Olfr1393 activity through other mechanisms (i.e., via a chemical antagonist) could also lead to an attenuated hyperfiltration response. From future studies, we hope to gain a better understanding of how Olfr1393 functions in health and disease and to explore potential ways to manipulate this receptor to alter physiological outcomes.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Disease Grant K01-DK-106400 awarded to B. D. Shepard.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

B.D.S. and J.L.P. conceived and designed research; B.D.S. performed experiments; B.D.S. and J.L.P. analyzed data; B.D.S. and J.L.P. interpreted results of experiments; B.D.S. prepared figures; B.D.S. drafted manuscript; B.D.S. and J.L.P. edited and revised manuscript; B.D.S., H.K., and J.L.P. approved final version of manuscript.