A sodium-glucose cotransporter 2 inhibitor attenuates renal capillary injury and fibrosis by a vascular endothelial growth factor–dependent pathway after renal injury in mice

basic research

www.kidney-international.org

A sodium-glucose transporter 2 inhibitor attenuates renal capillary injury and fibrosis by a vascular endothelial growth factor–dependent pathway after renal injury in mice Yifan Zhang1,2,6, Daisuke Nakano1,6, Yu Guan1,6, Hirofumi Hitomi1, Akiyoshi Uemura3, Tsutomu Masaki4, Hideki Kobara4, Takeshi Sugaya5 and Akira Nishiyama1 1 Department of Pharmacology, Kagawa University Medical School, Kagawa, Japan; 2Department of No. 2 Orthopedics, Shijiazhuang City No. 1 Hospital, Shijiazhuang, Hebei, China; 3Department of Retinal Vascular Biology, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan; 4Department of Gastroenterology & Neurology, Kagawa University, Kagawa, Japan; and 5Department of Internal Medicine, St. Marianna University School of Medicine, Kanagawa, Japan

Multiple large clinical trials have shown that sodiumglucose cotransporter (SGLT) 2 inhibitors reduce the risk of renal events. However, the mechanism responsible for this outcome remains unknown. Here we investigated the effects of the SGLT2 inhibitor luseogliflozin on the development of renal fibrosis after renal ischemia/ reperfusion injury in non-diabetic mice. Luseogliflozin significantly suppressed development of renal fibrosis, prevented peritubular capillary congestion/hemorrhage, attenuated CD31-positive cell loss, suppressed hypoxia, and increased vascular endothelial growth factor (VEGF)-A expression in the kidney after ischemia/reperfusion injury. Luseogliflozin failed to induce the above-mentioned protection in animals co-treated with sunitinib, a VEGF receptor inhibitor. Additionally, luseogliflozin reduced glucose uptake and increased VEGF-A expression in the kidneys of glucose transporter 2 (GLUT2)-downregulated mice following ischemia/reperfusion and in GLUT2-knockdown cells compared with those in normal controls. Withdrawal of glucose from cultured medium, to halt glucose uptake, remarkably increased VEGF-A expression and reversed the luseogliflozin-induced increase in VEGF-A expression in the proximal tubular cells. Thus, luseogliflozin prevented endothelial rarefaction and subsequent renal fibrosis after renal ischemia/reperfusion injury through a VEGF-dependent pathway induced by the dysfunction of proximal tubular glucose uptake in tubules with injuryinduced GLUT2 downregulation. Kidney International (2018) j.kint.2018.05.002

-, -–-;

https://doi.org/10.1016/

KEYWORDS: glucose uptake; renal fibrosis; sodium glucose co-transporter 2; vascular endothelial growth factor

Correspondence: Daisuke Nakano, Department of Pharmacology, Faculty of Medicine, Kagawa University, 1750-1 Ikenobe, Miki, Kita, Kagawa 761-0793, Japan. E-mail: [email protected] 6

These authors contributed equally to this work.

Received 15 September 2017; revised 25 April 2018; accepted 3 May 2018 Kidney International (2018) -, -–-

Copyright ª 2018, International Society of Nephrology. Published by Elsevier Inc. All rights reserved.

S

odium-glucose cotransporter 2 (SGLT2), which is expressed in the convoluted segments of renal proximal tubules, transports sodium ion (Naþ) and glucose at a 1:1 stoichiometry from the tubule lumen into the cells, and it mediates approximately 90% of glucose reabsorption under physiological conditions.1 Regardless of its extensive role in glucose transport, little attention was initially paid to the pathophysiological role of SGLT2 in the proximal tubules and surrounding renal environment because patients with familial renal glucosuria due to congenital SGLT2 mutation did not show apparent renal function changes.2 However, the sudden inhibition of glucose transport via an SGLT2 inhibitor may observably influence renal pathophysiology due to its acute nature, in contrast with that of a congenital SGLT2 mutation. Notably, the EMPA-REG OUTCOME trial reported that empagliflozin, an SGLT2 inhibitor, reduced the risk of major adverse cardiovascular events3 as well as the risks of serum creatinine doubling and renal replacement therapy initiation4 in patients with type 2 diabetes at high risk for cardiovascular events. Similar renal protection was observed recently by the CANVAS Program for canagliflozin, another SGLT2 inhibitor.5 Acute kidney injury (AKI) is a common clinical syndrome defined as a sudden onset of reduced renal function,6 which could occur during and/or after heart failure, and the AKI incidence has steadily increased over the last decade. Furthermore, AKI survivors have an increased risk of developing chronic kidney disease (CKD),7 heart failure, and myocardial infarction.8 Numerous studies have revealed the multiple pathways, such as autophagy,9 endothelial rarefaction,10 cell cycle arrest, and cell senescence,11,12 involved in the mechanism of AKI and subsequent development of CKD. We recently reported that an acute SGLT2 knockdown in renal proximal tubular cells prevented high-glucose–induced tubular cell senescence.13 Furthermore, the occurrence of AKI tended to be reduced, or at least not increased, in SGLT2 1

RESULTS Effect of luseogliflozin on post-IR renal function and morphology

This study employed nondiabetic animals because AKI severity and recovery could be affected by blood glucose levels. Throughout the experimental period, the blood glucose level did not significantly change following luseogliflozin treatment in either the sham-operated or IR groups (Supplementary Figure S1A). Luseogliflozin increased the urine volume significantly in the sham-operated groups at days 1, 3, and 7 (Supplementary Figure S1B), but the similar trend in the IR groups did not reach statistical significance. Additionally, luseogliflozin markedly increased urinary glucose excretion levels in both the sham-operated and IR groups, although the response was greater in the shamoperated groups (Supplementary Figure S1C). IR induced an increase in blood urea nitrogen (BUN) and a decline in creatinine clearance, which both reflect the severity of renal malfunction, at day 1 after reperfusion (n ¼ 5–8) (Figure 1a and b). The BUN and creatinine clearance levels were relatively recovered in the following days (days 3 and 7), but these levels remained different from those in the sham-operated groups. Luseogliflozin did not affect the creatinine clearance and BUN levels at day 1 or the recovery of BUN and creatinine clearance from IR-induced injury compared with vehicle treatment (Figure 1a and b). There was no statistically significant difference in the AKI-induced histological damage between the luseogliflozin-treated and vehicle-treated groups at days 1 and 3 (Figure 1c); however, treatment with luseogliflozin attenuated the histological damage at day 7. 2

a

Sham

BUN (mg/l)

800

Ischemia-reperfusion

n.s.

1000

*

*

n.s.

600

*

*

n.s.

n.s. n.s.

400 200 0 Vehicle Luseo Vehicle Luseo Vehicle Luseo Day 1

b

Day 7

Day 3

200

n.s.

n.s. CCr (μl/min)

inhibitor-treated subjects in both the EMPA-REG OUTCOME4 and CANVAS Program,5 and a recent study using mice reported the protective effects of dapagliflozin against ischemia-reperfusion (IR) injury,14 despite its presumed risk of dehydration. These findings led us to hypothesize that SGLT2 inhibition might prevent AKI after IR injury in mice by accelerating tubular recovery and might prevent renal interstitial fibrosis development, a common outcome of CKD. Therefore, we assessed the ability of luseogliflozin treatment initiated after IR in nondiabetic mice to prevent AKI or to accelerate recovery from AKI, and, furthermore, observed its effects on renal fibrosis development. We also investigated the potential mechanism for this effect. Proximal tubules express 2 glucose transporters, SGLT2 and glucose transporter 2 (GLUT2). Because GLUT2 is a facilitated diffusion transporter, inhibition of SGLT2 alone should not influence the net glucose uptake in proximal tubules in vivo; GLUT2 can take up glucose from the interstitial spaces. Importantly, IR injury changes proximal tubule polarity and influences protein expression levels.15 Thus, we also hypothesized that AKI disrupted the cell membrane GLUT2 expression and, thus, that luseogliflozin induced a depletion of glucose uptake. This loss of glucose uptake could change the phenotype of proximal tubules and the long-term outcome of kidney injury.

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

n.s. n.s.

n.s.

150 100 50

*

* n.o.

n.o. 0

Vehicle Luseo Vehicle Luseo Vehicle Luseo Day 1 n.s.

c

5

*

Day 7

Day 3

*

n.s.

*

4 HE score

basic research

*

3

* 2

*#

1 0 Vehicle Luseo Vehicle Luseo Vehicle Luseo Day 1

Day 3

Day 7

Figure 1 | Time course of changes in blood urea nitrogen (BUN), creatinine clearance, and histological damage score in the kidney after ischemia-reperfusion. The (a) blood urea nitrogen levels, (b) creatinine clearance, and (c) histological damage kidney scores were measured on days 1, 3, and 7 in luseogliflozin (luseo)- or vehicle-treated groups of mice that underwent a sham-operation or ischemiareperfusion (n ¼ 5–8). n.o., not observed; n.s., not significant. *P < 0.05 versus sham-operated group, #P < 0.05 versus vehicle-treated group.

AKI is characterized by an increase in tubular cell death, proliferation,11 and autophagy.9 Luseogliflozin did not affect the numbers of propidium iodide-positive dead cells at day 1, of LC3-positive puncta in GFP-LC3#53 reporter mice at day 3, or of bromodeoxyuridine- and Ki67-positive cells at day 3 (n ¼ 3–4) (Supplementary Figure S2), supporting the observation that luseogliflozin treatment, when started at 6 hours after reperfusion, did not alleviate AKI. Kidney International (2018) -, -–-

basic research

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

Luseogliflozin treatment attenuates post-IR renal interstitial fibrosis

Renal interstitial fibrosis occurs in the chronic phase after IRinduced AKI.16 Thus, we further analyzed renal interstitial fibrosis following IR injury. Schematics illustrating the experimental protocols are shown in Figures 2a and 3a. There were more Sirius red stain-positive areas in the vehicle-treated group (n ¼ 7) (Figure 2b and c) compared with the luseogliflozin-treated group (n ¼ 7) (Figure 2d and e) and the contralateral kidneys at week 1 after IR (Figure 2f and g). The

a

relative differences in Sirius red stain-positive areas between the groups described above are shown in Figure 2h (total cross section area) and Supplementary Figure S3 (cortex and medulla). The TGF-b mRNA level was increased in the vehicle-treated IR group, and luseogliflozin suppressed this increased TGF-b expression (n ¼ 8 and 9, respectively) (Figure 2i). The 1-week treatment with luseogliflozin after IR attenuated the development of renal interstitial fibrosis at week 4 post-IR (n ¼ 6 vs. n ¼ 5 in the vehicle group) (Figure 3).

Reperfusion

Luseogliflozin or vehicle,p.o., daily

Ischemia

30 min

1 week histopathology and RNA isolation

6 hrs

Ischemic (b, c, d and e) contralateral (f and g)

b

Vehicle, cortex

d

Luseo, cortex

c

Vehicle, outer medulla

e

Luseo, outer medulla

10

i

Sirius red–positive area

8

*

6

#

4 2 0

Vehicle Luseo Vehicle Luseo Contralateral

Ischemic

Expression level (2-δδCT)

Relative difference (vs. vehicle, contralateral)

h

f

1.5

g

Vehicle, cortex

Luseo, cortex

TGF-β mRNA

*

1.0

#

0.5 0.0

Vehicle Luseo Vehicle Luseo Contralateral

Ischemic

Figure 2 | Assessment of renal fibrosis development following luseogliflozin treatment at 1 week after renal ischemia-reperfusion injury. (a) Schematic of the experiment schedule. The left kidney was subjected to unilateral renal ischemia for 30 minutes. Luseogliflozin was administered 6 hours after reperfusion and then daily until day 7. Kidney samples were collected at day 7. Representative images of Sirius red–stained sections from the ischemic kidney of vehicle-treated mice (b, cortex; c, outer medulla) or luseogliflozin (luseo)-treated mice (d, cortex; e, outer medulla) at 1 week after reperfusion. Representative images of Sirius red–stained sections from the contralateral kidneys of (f) vehicle- or (g) luseo-treated mice. (h) The relative difference in Sirius-red–positive area in the kidney at 1 week after reperfusion (n ¼ 7). (i) The transforming growth factor-b (TGF-b) levels in the kidneys of the vehicle- or luseo-treated groups at 1 week after reperfusion (n ¼ 8–9). Bar ¼ 100 mm. *P < 0.05 versus contralateral control, #P < 0.05 versus vehicle-treated group. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org. Kidney International (2018) -, -–-

3

basic research

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

Reperfusion

a

Luseogliflozin or vehicle

Ischemia

30 min

6 hrs

1 week

b

Vehicle #1, Sirius red

d

Luseo: Sirius red

c

Vehicle #1, Azan

e

Luseo, Azan

g

15

Sirius red

12

*

9 6

#

3 0

Contralateral

Ischemic

Relative difference (vs. vehicle, contralateral)

f Relative difference (vs. vehicle, contralateral)

4 weeks histopathology

25

Azan

20 15

*

10 5

#

0

Contralateral

Ischemic

Figure 3 | Assessment of renal fibrosis development following luseogliflozin treatment at 4 weeks after renal ischemia-reperfusion injury. (a) Schematic of the experiment schedule. The left kidney was subjected to unilateral renal ischemia for 30 minutes. Luseogliflozin was administered 6 hours after reperfusion and then daily until day 7. Kidney samples were collected at 4 weeks after reperfusion. Neither luseogliflozin nor vehicle was administered during the period of 1 to 4 weeks after reperfusion. Representative images of (b,d) Sirius-red–stained or (c,e) Azan-stained renal outer medulla from (b,c) vehicle-treated or (d,e) luseogliflozin (luseo)-treated mice at 4 weeks after reperfusion. The relative differences in (f) Sirius red or (g) Azan staining-positive area (n ¼ 5–6). Bar ¼ 100 mm. *P < 0.05 versus contralateral control, #P < 0.05 versus vehicle-treated group. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

Renal function was analyzed after uninephrectomy of the contralateral kidney at 4 weeks post-unilateral IR (Figure 4a). Luseogliflozin treatment tended to lower the plasma creatinine level and urine volume, and it significantly preserved creatinine clearance compared with vehicle treatment (Figure 4b). Luseogliflozin ameliorated renal endothelial rarefaction post-IR injury

We next focused more closely on the difference in the post-IR morphological damage between vehicle- and luseogliflozintreated groups at week 1 shown in Figure 1 (uninephrectomized 4

IR model). There was little difference in the scores for the brush border loss, tubular dilatation, and cast formation between luseogliflozin- and vehicle-treated groups, whereas luseogliflozin treatment markedly attenuated the IR-induced congestion and hemorrhage, which was reproducible in a non-nephrectomized IR model at week 1 (n ¼ 6) (Figure 5a and Supplementary Figure S4A and B). To further examine the damage in the renal capillary, we performed immunohistochemistry for CD31, an endothelial marker (n ¼ 8) (Figure 5b and c). IR significantly reduced the CD31-positive area in the ischemic kidney of vehicletreated mice compared with the contralateral kidney at week 1, Kidney International (2018) -, -–-

basic research

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

a

Reperfusion

Urine collection for 24 hrs

Left kidney ischemia 30 min

Luseogliflozin or vehicle 6 hrs

1 week

4 weeks UNX

6 hrs 30 hrs after UNX after UNX

UNX: uninephrectomy of right kidney

b Plasma creatinine (mg/dl) 1.5

Urine volume (ml) 6

Creatinine clearance (μl/min) 200

*

150

4 1.0

100 2

0.5

50

0 Vehicle

Luseo

0 Vehicle

Luseo

Vehicle

Luseo

Figure 4 | Assessment of renal function following luseogliflozin treatment at 4 weeks after renal ischemia-reperfusion injury. (a) Schematic of the experiment schedule. The left kidney was subjected to unilateral renal ischemia for 30 minutes. Luseogliflozin was administered 6 hours after reperfusion and then daily until day 7. Neither luseogliflozin nor vehicle was administered during the period of 1 to 4 weeks after reperfusion. At 4 weeks after reperfusion, the mice underwent nephrectomy of the contralateral right kidney. Urine samples were collected from 6 to 30 hours after uninephrectomy. (b) Plasma creatinine level (left), 24-hour urine volume (center), and creatinine clearance (right) after uninephrectomy at 4 weeks after unilateral ischemia-reperfusion. *P < 0.05 versus vehicle-treated group.

and luseogliflozin treatment ameliorated this decline in the CD31-positive area (Figure 5b and c). Tissue hypoxia was examined by pimonidazole staining (n ¼ 4).17 IR significantly augmented the level of pimonidazole adducts in the ischemic kidney of vehicle-treated mice compared with the contralateral kidney at week 1, and luseogliflozin treatment ameliorated this increase in pimonidazole adducts (Figure 5d). The area positive for NG2, an activated pericyte marker,18 was decreased by luseogliflozin treatment (n ¼ 5) (Supplementary Figure S5). The above findings suggest that luseogliflozin ameliorated IR-induced renal capillary rarefaction and the detachment of pericytes from endothelial cells, both of which have been implicated in renal interstitial fibrosis development.19 We next hypothesized that luseogliflozin maintained the renal capillary network by accelerating post-IR neoangiogenesis via increasing the vascular endothelial growth factor (VEGF)-A level. In situ hybridization revealed that the proximal tubules expressed VEGF-A mRNA (Supplementary Figure S6). Luseogliflozin treatment increased the VEGF-A mRNA expression in the kidney at week 1 compared with vehicle treatment (n ¼ 5) (Figure 5e). To determine whether VEGF-A plays a critical role in the luseogliflozin-induced beneficial effects, we performed experiments using sunitinib, an inhibitor of the tyrosine kinase in the VEGF receptor. Luseogliflozin treatment failed to improve the congestion/ hemorrhage score (images in Supplementary Figure S7) or the levels of renal interstitial fibrosis, transforming growth factor-b (TGF-b) mRNA expression, NG2 staining, or CD31 staining in the VEGF receptor inhibitor–treated mice at week 1 after IR (n ¼ 5–12) (Figure 6). The above results indicate that there is an interaction between the proximal tubules and peritubular capillaries because SGLT2 is expressed in only the early segments of Kidney International (2018) -, -–-

proximal tubules, not the capillary endothelium (Supplementary Figure S8). IR reduced the number of SGLT2-positive tubules regardless of luseogliflozin treatment, whereas SGLT2 expression was still maintained at the apical membrane following IR. Therefore, we next focused on the relationship between proximal tubular glucose uptake and the VEGF-A mRNA level. An in vivo analysis of the uptake of a fluorescent glucose analogue, NBDG, in mouse kidneys (Supplementary Figure S9) revealed that both IR in vehicletreated mice and luseogliflozin treatment in sham-operated mice mildly suppressed glucose uptake in the proximal tubules and that luseogliflozin treatment post-IR dramatically reduced the glucose uptake in the proximal tubules (Figure 7). We speculated that GLUT2, another glucose transporter, contributes to the difference in glucose uptake by the luseogliflozin-treated proximal tubules between the IR and control kidneys. This idea is supported by the GLUT2 downregulation that accompanied the marked reduction of glucose uptake in the luseogliflozin-treated IR kidneys (Figure 8). We then determined whether the glucose uptake level affects the VEGF-A mRNA level in mProx24 cells, a cultured proximal tubular cell line that express SGLT2 (Supplementary Figure S10). Under normoxic conditions, glucose concentrations of 5 to 25 mmol/l strongly suppressed the VEGF-A mRNA level in the proximal tubular cells compared with that in the glucose-free condition (Figure 9a). The effect was greater under conditions with higher glucose concentrations (17.5 and 25 mmol/l), indicating that the presence of glucose is itself a strong suppressor of VEGF-A induction. In experiments blocking both SGLT2 and GLUT2, we blocked GLUT2 by using a GLUT2 inhibitor, cytochalasin B, or by using cells transfected a siRNA for 5

basic research

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

a Congestion/hemorrhage (score)

5 4 3

*

2

#

1

Contralateral

b

CD31 IHC

Lu se o

Ve hi cle

Lu se o

Ve hi cle

0

Ischemic

c Area positive for CD31 (%/field) 20 15

#

10

*

5

Contralateral

Vehicle, ischemic Luseo, ischemic

Lu se o

#

3.0 2.0

0.0

Contralateral

Lu se o

*

1.0

Ve hi cle

Luseo, contralateral

VEGF-A mRNA

Lu se o

e

Pimonidazole IHC

Vehicle, contralateral

Ischemic

Vehicle, ischemic Luseo, ischemic

Ve hi cle

d

Luseo, contralateral

Expression level (2-δδCT)

Vehicle, contralateral

Ve hi cle

Lu se o

Ve hi cle

0

Ischemic

Figure 5 | Effect of luseogliflozin treatment on endothelial rarefaction after renal ischemia-reperfusion injury. Vehicle- or luseogliflozin (luseo)-treated mice were analyzed at day 7 after ischemia-reperfusion injury. (a) A representative image (left) and the corresponding congestion-hemorrhage scores (right) of the renal outer medulla. Yellow arrows indicate congestion and hemorrhage (n ¼ 6). Bar ¼ 50 mm. (b) Representative images and (c) corresponding quantification of the area positive for anti-CD31 antibody in the contralateral and ischemic kidneys (n ¼ 8). Bar ¼ 300 mm. Bottom images show enlargements of the boxed areas in the top images. (d) Representative images of the pimonidazole staining in the renal cortex (n ¼ 4). (e) Quantification of the renal vascular endothelial growth factor-A (VEGF-A) mRNA expression in the contralateral and ischemic kidneys (n ¼ 9–10). Bar ¼ 100 mm. *P < 0.05 versus contralateral control. #P < 0.05 versus vehicletreated group. To optimize viewing of this image, please see the online version of this article at www.kidney-international.org.

6

Kidney International (2018) -, -–-

basic research

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

Congestion/hemorrhage

a

n.s.

5

Score

4

*

*

3

*

2 1 0 Vehicle

Vehicle

Vehicle

b

Sirius red

Sunitinib

c

n.s.

*

3

*

1.5 1.0

n.s.

*

*

0.5 0.0

0 Vehicle

Vehicle

Vehicle

Luseo

Sunitinib

Vehicle

(x103 pixel/field)

*

*

6

Expression level (2-δδCT)

9

Vehicle

Area positive for NG2 n.s.

e 24

Luseo

Sunitinib

Vehicle

Area positive for CD31 n.s.

20

8

*

*

6

*

4

(%/field)

Relative difference (vs. vehicle, contralateral)

12

10

TGF-β mRNA 2.0

15

d

Luseo

16 12

4

0

0 Vehicle Vehicle

Vehicle

Luseo

Sunitinib

*

8

2

*

*

Vehicle Vehicle

Vehicle

Luseo

Sunitinib

Figure 6 | Effect of treatment with sunitinib. Vehicle- or luseogliflozin (luseo)-treated mice were co-treated with sunitinib, a vascular endothelial growth factor (VEGF) receptor inhibitor, during days 3 to 7 after ischemia-reperfusion (n ¼ 5–12). The (a) congestion-hemorrhage scores, (b) Sirius red–positive areas, (c) renal transforming growth factor-b (TGF-b) mRNA expression levels, (d) NG2-positive areas, and (e) CD31positive areas of these mice. Open symbols: contralateral kidney; closed symbols: ischemia-reperfusion kidney. n.s., not significant. *P < 0.05 versus vehicle control.

GLUT2, which reduced the GLUT2 mRNA level to 59%  4% (n ¼ 3) compared with a scrambled control. Treatment with luseogliflozin in cytochalasin B- or siRNA-treated cells reduced the glucose uptake level (Supplementary Figure S11) and the increased the VEGF-A expression level compared with luseogliflozin treatment in otherwise untreated cells (Figure 9b and c). Next, we evaluated the effects of glucose with and without luseogliflozin under cobalt-induced hypoxia and cobaltwithdrawal-induced reoxygenation (n ¼ 6–12) (Figure 9d). Luseogliflozin did not alter VEGF-A mRNA levels in cells that did not undergo hypoxia-reoxygenation. Compared with the

Kidney International (2018) -, -–-

normoxic control in the proximal tubular cells, the VEGF-A mRNA levels trended higher following hypoxiareoxygenation, but this increase did not reach statistical significance. Luseogliflozin significantly increased the VEGF-A mRNA levels after hypoxia-reoxygenation at all investigated glucose concentrations. Furthermore, luseogliflozin showed no effect on the VEGF-A mRNA level when glucose-free medium was used (Supplementary Figure S12), suggesting that glucose uptake is indispensable for the induction of VEGF-A by luseogliflozin. These data indicate that the proximal tubular cells could be one source of the increased VEGF-A in luseogliflozin-treated kidneys after IR.

7

basic research

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

Sham: vehicle

Sham: luseo NBDG uptake (Δ intensity vs. baseline)

3000 * # 2000

*

1000

†

IR: Vehicle

IR: Luseo

Sham

Lu se o

Ve hi cl e

Lu se o

Ve hi cl e

0

Ischemia/ reperfusion

Figure 7 | Glucose uptake in the kidney. Representative images (left) and quantification (right) of the uptake of a fluorescent glucose analogue, NBDG, in mouse proximal tubules at day 3 after ischemia-reperfusion (IR) (n ¼ 7). Bar ¼ 50 mm. *P < 0.05 versus vehicle, #P < 0.05 versus sham, †P < 0.05 versus interaction. To optimize viewing of this image, please see the online version of this article at www.kidneyinternational.org.

DISCUSSION

AKI induced by renal IR is characterized by severe tubular damage, such as tubular cell sloughing and cast formation,20 and by progressive renal interstitial fibrotic changes with capillary rarefaction of peritubular capillaries.10 Here, we showed that treatment with luseogliflozin started after IR attenuated the endothelial rarefaction, renal hypoxia, and development of renal interstitial fibrosis. Our results also suggest that the luseogliflozin treatment–induced reduced glucose uptake in the GLUT2-downregulated tubules increased VEGF-A expression, and the increased VEGF-A level in the proximal tubules can, at least partially, explain the mechanism for this effect. The in vivo glucose uptake test conducted by multiphoton imaging allows visualization of changes in glucose uptake in the tubules of superficial nephrons, and the brush border membrane of the tubules have been reported to express SGLT2.21,22 Luseogliflozin treatment in vivo did not strongly influence the glucose uptake in these tubules; this may be due to the action of GLUT2, a bidirectional glucose transporter23 expressed at the basolateral membrane of SGLT2-expressing tubules.22 GLUT2 can transport glucose from the interstitial space to the intracellular space after the inhibition of transapical glucose transport through SGLT2, to equilibrate intraand extra-cellular glucose levels. Indeed, we found that the inhibition of both GLUT2 and SGLT2 reduced glucose uptake more efficiently than SGLT2 inhibition alone in proximal tubular cells. Thus, we propose that the inhibition of SGLT2 in AKI results in the remarkable reduction of glucose uptake in the proximal tubules with reduced GLUT2 expression. Furthermore, several previous reports in multiple cell types have demonstrated that both hypoxia and hypoglycemia are factors capable of inducing VEGF-A.24–27 Therefore, we presume that this marked reduction in the cellular glucose uptake resulting from the combination of luseogliflozin and IR mimicked the in vitro study environment with glucose-free medium and became a strong inducer of VEGF-A (Figure 10). 8

The cell origin of VEGF in the present study has not been definitively confirmed; however, cultured proximal tubular cells expressed more VEGF-A under luseogliflozin treatment compared with the vehicle control in response to hypoxiareoxygenation stress. Quaggin’s group28 reported that renal tubule-derived VEGF-A maintained the perivascular capillary architecture under physiological conditions. Their study reported that mice in which VEGF-A was inductively deleted in Pax8-expressing cells, which are mainly in renal tubules, showed a decreased density of peritubular capillaries, indicating that the continuous secretion of VEGF-A from tubules is essential to preserve the capillary network. Thus, one possibility is that the treatment with luseogliflozin in the present study restored the tubular cell VEGF-A levels and preserved the capillary network. Indeed, our results indicate that luseogliflozin increased the VEGF-A mRNA level by regulating glucose uptake into the proximal tubular cells after hypoxia-reoxygenation. Alternatively, SGLT2 inhibition may increase the glucose load on the SGLT1-expressing straight segments of proximal tubules, resulting in an increase in ATP consumption in the SGLT1-expressing cells. This might further activate hypoxiainducible factor 1a and subsequent VEGF-A production from the tubules, thereby ameliorating the endothelial rarefaction of peritubular capillary beds after IR. Notably, changes in the extracellular glucose level from 5 to 25 mmol/l did not enhance the responses to VEGF-A expression, although we were able to detect SGLT1 in mProx24 cells. Another possibility is that SGLT2-inhibited proximal tubules secrete factors that accelerate VEGF-A production in the surrounding cells, additional to the VEGF-A from the tubules, and thus contribute to the angiogenesis of peritubular capillaries as a repair mechanism. Because there was no difference in the levels of BUN and creatinine clearance between the vehicle- and luseogliflozintreated groups in the acute phase, it seems likely that the Kidney International (2018) -, -–-

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

a

GLUT2 on membrane

β-actin Luseo

+

+

−

−

+

+

Sham

−

−

Ischemia/ reperfusion

b Ratio to β-actin

1.5

1.0

# #

0.5

0.0

le ic eh V

o se u L

Sham

le ic eh V

o se u L

Ischemia/ reperfusion

Figure 8 | Glucose transporter 2 (GLUT2) expression in the membranous fraction of the kidney. Western blots show the GLUT2 protein expression in the ischemia-reperfusion kidneys (at day 3). Representative (a) Western blots and (b) quantitative data are shown (n ¼ 4). Note: there are multiple bands for GLUT2 due to glycosylation. The bands that are indicated by an arrow, which are detectable mainly in the cell membrane and are potentially glycosylated, were analyzed, and the resulting data are shown in Figure 8b. #P < 0.05 versus sham.

post-IR luseogliflozin treatment did not exaggerate dehydration. The effects of luseogliflozin on renal function and tubular damage appear to be minimal; there was no effect on renal function, cast formation, tubular dilatation, autophagy, or the number of dead cells in the first 3 days after IR. Furthermore, luseogliflozin is unlikely to accelerate the regeneration of tubular cells after IR based on the observed numbers of Ki67positive cells. Rather, it might improve renal capillary perfusion by preventing the capillary network loss in AKI after IR. A recent study reported that 2-day treatment with dapagliflozin, another SGLT2 inhibitor, prior to IR protected kidneys against the IR-induced injury at 24 hours after reperfusion.14 The authors proposed that the preconditioning effect was induced through a mechanism by which dapagliflozin increased hypoxia-inducible factor 1. Here, the luseogliflozin treatment did not alleviate kidney injury in the acute phase, and it did not significantly increase the mRNA level of VEGF-A, a target of hypoxia-inducible factor 1, in the contralateral kidneys or in proximal tubular cells cultured under normoxia. Both drugs have been reported to be absorbed within a half hour after an oral administration and Kidney International (2018) -, -–-

basic research

to induce glycosuria at least for 18 hours, although the glycosuria duration induced by dapagliflozin is longer than that induced by luseogliflozin (>24 hours).29 This longer glycosuria duration could be key for the ability to induce a preconditioning effect. The number of NG2-positive cells was increased after IR, which is consistent with the findings of previous studies.30 Activated pericytes change their phenotype from periendothelial cells to myofibroblasts that secrete collagen and promote the development of fibrosis.31 The pericyte-detached endothelium seems to be destabilized, causing endothelial rarefaction and cell cycle arrest, and endothelium-free pericytes change their phenotype to a mesenchymal phenotype.32 Thus, the loss of the pericyte-endothelium interaction, which may also be regulated by VEGF-A,33 triggers phenotypic changes in both cell types and worsens renal interstitial fibrosis. A limitation of the current study is that we used healthy young subjects, who commonly show resistance to AKI stress and possibly to AKI-induced CKD stress. Further studies need to investigate the effects of SGLT2 inhibition on the development of fibrosis in subjects with known AKI risk factors, such as older mice. In conclusion, luseogliflozin treatment, when started after the onset of AKI, protected the renal capillary network from IR injury and, thereby, inhibited the development of renal interstitial fibrosis after IR injury in mice. The luseogliflozininduced reduction of glucose uptake and the following expression of VEGF-A play a critical role in this pathway. Treatment with SGLT2 inhibitors is usually prohibited in patients undergoing major cardiovascular surgery because of the risk of dehydration and the high susceptibility to ureteral infection. However, this study provides evidence that the initiation of treatment with luseogliflozin after an AKI event might be beneficial for protecting the kidney against capillary rarefaction and progressive interstitial fibrosis. MATERIALS AND METHODS A detailed description of the materials and methods can be found in the Supplementary Materials and Methods. Animals Six-week-old male C57BL/6J mice were purchased from Clea Japan (Tokyo, Japan). GFP-LC3#53 reporter mice with green fluorescent protein (GFP) expression under the control of the LC3-promoter were obtained from Riken BioResource Center (Kyoto, Japan).34 All animal experimental procedures were performed according to the guidelines for the care and use of animals established by Kagawa University. Renal IR protocol Two different IR models were used in this study. To observe treatment effects on AKI (data from this experiment are shown in Figure 1), unilateral renal IR injury of the left kidney was performed 11 days after nephrectomy of the right kidney under isoflurane anesthesia (1.0%–1.5%), with mouse body temperature maintained at 37  C using a heat pad. The renal artery and vein were occluded by a vascular clamp for 30 minutes to induce renal ischemia. To observe 9

basic research

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

0.5 0.0 0

*

#

*

#

5 17.5 25 mmol/l Glucose concentration

d

2

*

1 0

o se Lu

2.5

VEGF-A mRNA * †

2.0 1.5 1.0 0.5 0.0

VEGF-A mRNA 4

Expression level (2-δδCT)

* †

Lu se o

1.0

3

c

Ve hi cl e Lu se o

Expression level (2-δδCT)

1.5

*

VEGF-A mRNA

Expression level (2-δδCT)

b

VEGF-A mRNA

Ve hi cl e Lu se o

Expression level (2-δδCT)

a

Hypoxia/ reoxygenation

Normoxia †

3

†

†

2

#

1

#

5 mmol/l glucose

Ve hi cl e Lu se o Ve hi cl e Lu se o

Ve hi cl e Lu se o Ve hi cl e Lu se o

Ve hi cl e Lu se o Ve hi cl e Lu se o

0

17.5 mmol/l glucose

25 mmol/l glucose

Figure 9 | Effect of glucose on vascular endothelial growth factor-A (VEGF-A) mRNA expression. (a) VEGF-A mRNA expression in mProx24 cells cultured with media containing different concentrations of glucose (n ¼ 9–12). Note: 17.5 mmol/l is the glucose concentration in the standard medium for mProx24 cells. *P < 0.05 versus 0 mmol/l glucose, #P < 0.05 versus 5 mmol/l glucose. (b) Effects of cytochalasin B (GLUT2i) treatment on VEGF-A mRNA in luseogliflozin (luseo)-treated mProx24 cells (n ¼ 6). *P < 0.05 versus vehicle, †P < 0.05 versus luseo. (c) Effects of GLUT2 siRNA on VEGF-A mRNA in luseo-treated mProx24 cells (n ¼ 6). *P < 0.05 versus vehicle, †P < 0.05 versus luseo. (d) Effects of luseo on VEGF-A mRNA in mProx24 cells under CoCl2-induced hypoxia (for 20 hours) and CoCl2-free reoxygenation (for 6 hours) in media with various glucose concentrations (n ¼ 10–18). #P < 0.05 versus vehicle þ normoxia in 5 mmol/l glucose, †P < 0.05 versus vehicle control in hypoxia-reoxygenation.

renal fibrosis after IR (data from these experiments are shown in Figures 2–4), we used a unilateral renal IR injury model without nephrectomy.35 We employed this model to analyze the long-term outcome because the other IR models induce severe AKI, which is accompanied by unacceptably high mortality, to induce evident fibrosis.11 Mice were randomly separated into groups, and IR was performed as described above. The SGLT2 inhibitor luseogliflozin or its vehicle, carboxymethyl cellulose (0.1%), was administered 6 hours after reperfusion (p.o. 30 mg/kg/d) and then daily until day 7. The treatment was initiated only after IR because our preliminary study showed that treatment administered both before and after IR tended to exaggerate dehydration in the mice. The animals were killed at day 1, 3, or 7 or at 4 weeks post-reperfusion, and plasma and kidney samples were collected for further analysis. The data from the nonischemic contralateral right kidneys in the vehicle-treated mice were used as a control for normalization in the non-nephrectomized model. A separate set of animals received sunitinib, a tyrosine kinase inhibitor that inhibits the VEGF receptor, at a dosage of 50 mg/kg/d, or its vehicle during days 3 to 736 after a 30-minute IR. Luseogliflozin or its vehicle was administered 6 hours after reperfusion and then daily until day 7. The animals were killed at day 7 after reperfusion, and kidney samples were collected for further analysis. 10

Statistical analysis Statistical significance was assessed using a 1-way or 2-way analysis of variance followed by a Tukey’s multiple-comparison test to evaluate the differences between groups. A Mann-Whitney U test was performed to compare the means in experiments with 2 individual groups. All statistical analyses were performed using GraphPad Prism 6 (GraphPad Software Inc., La Jolla, CA), with values of P < 0.05 considered to be statistically significant. Data are presented as means þ SEM. DISCLOSURE

All the authors declared no competing interests. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript, except for providing information on the pharmacokinetics of luseogliflozin. ACKNOWLEDGMENTS

This study was sponsored by Taisho Pharmaceutical Co., Ltd., the manufacturer of luseogliflozin. We thank Ms. Megumi Yamamoto for her technical assistance in sectioning the paraffin blocks and performing hematoxylin and eosin, Sirius red, and Azan staining. We also thank Katie Oakley, PhD, from Edanz Group (www.edanzediting. com/ac) for editing a draft of this manuscript. Kidney International (2018) -, -–-

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

: SGLT2 Lumen

: GLUT2

Proximal tubular cells

Interstitial glucose

glucose Uninjured and untreated

glucose Glucose level equilibrated

glucose

glucose glucose glucose

Uninjured + luseogliflozin Glucose level equilibrated

glucose

glucose

glucose

Tubular injury + luseogliflozin

glucose Glucose level reduced

VEGF-A

glucose

Figure 10 | Working hypothesis for the sodium-glucose cotransporter 2 (SGLT2) inhibitor-induced depletion of glucose uptake in proximal tubules. Under normal (uninjured) conditions, glucose is taken up through SGLT2 and transferred to the interstitial space via glucose transporter 2 (GLUT2). Although SGLT2 inhibition may transiently reduce the intracellular glucose level under normal conditions, GLUT2, an equilibration transporter, transfers glucose from the interstitial side to the intracellular side, so the intracellular glucose level will return to a level similar to that in the interstitial space. In contrast, the GLUT2 level is reduced in injured proximal tubular cells and is insufficient to equilibrate the intracellular glucose level following treatment by SGLT2 inhibitors. Thus, under these conditions, treatment with SGLT2 inhibitors induces the reduction of glucose uptake and stimulates the expression of vascular endothelial growth factor-A (VEGF-A). The VEGF-A may act to maintain neoangiogenesis and prevent endothelial rarefaction. SUPPLEMENTARY MATERIAL Supplementary Materials and Methods Figure S1. Changes in blood glucose, urine volume, and urinary glucose over time. Changes in (A) blood glucose, (B) urine volume, and (C) urinary glucose excretion by the kidney over the week after ischemia-reperfusion (IR) in luseogliflozin (luseo)- or vehicle-treated sham-operated or IR groups of mice. Figure S2. Assessment of acute kidney injury following treatment with luseogliflozin after renal ischemia-reperfusion injury. (A) The numbers of dead cells in the contralateral and ischemic kidneys of vehicle- or luseogliflozin (luseo)-treated mice, as analyzed by propidium-iodide (PI) staining. (B) The numbers of LC3-GFP puncta, which correspond to the accumulation of autophagosomes, in vehicle- or luseo-treated GFP-LC3#53 reporter mice. Quantitation of staining with bromodeoxyuridine, which was taken up into the S-phase cells as a (C) tyrosine analogue or (D) Ki67 in vehicle- or luseo-treated mice. Figure S3. Assessment of renal fibrosis development following luseogliflozin treatment after renal ischemia-reperfusion injury. The relative difference in Sirius red stain-positive areas in either the (A) cortical or (B) medullary area of sections from the ischemic or contralateral kidney in vehicle- or luseogliflozin (luseo)-treated mice Kidney International (2018) -, -–-

basic research

at 1 week after reperfusion. *P < 0.05 versus contralateral control. #P < 0.05 versus vehicle-treated group. Figure S4. Assessment of the effect of luseogliflozin treatment on injury to the renal capillary network at day 7 after ischemiareperfusion. (A) Congestion and/or hemorrhage scores in the cortical (left) and medullary (right) areas of ischemic or contralateral kidneys in vehicle- or luseogliflozin (luseo)-treated mice at day 7 after reperfusion. (B) The cast formation (left) and loss of brush border membrane (right) in kidney sections from mice treated as described above. #P < 0.05 versus vehicle-treated group. Figure S5. Numbers of NG2-positive cells in luseogliflozin- and vehicle-treated mice. Above: representative images of NG2 staining (green) in the cortex (top) and the medulla (bottom) of ischemic or contralateral kidneys in vehicle- or luseogliflozin (luseo)-treated mice. Arrows indicate the cast in the tubular lumen, which can be easily identified by the shape of its auto-fluorescence. Red: isolectinB4Alexa594, which represents the capillary endothelium. Below: quantification of the NG2-positive areas in these groups. *P < 0.05 versus contralateral control. #P < 0.05 versus vehicle-treated group. Bar ¼ 50 mm. Figure S6. In situ hybridization for VEGF-A. A representative image of VEGF-A (purple) in situ hybridization in a kidney section from a luseogliflozin-treated IR mouse. The staining was performed on a 300mm thick section. Yellow arrows indicate VEGF-A-positive proximal tubules, and the white arrow indicates a glomerulus. Bar ¼ 100 mm. Figure S7. Effect of sunitinib treatment on congestion and hemorrhage. Vehicle- or luseogliflozin (luseo)-treated mice were co-treated with sunitinib at day 7 after ischemia-reperfusion (n ¼ 5–12). Representative images of hematoxylin and eosin (HE)stained kidney sections of these mice are shown. The corresponding semiquantitative results are shown in Figure 6a. Yellow arrows indicate congestion and hemorrhage. Bar ¼ 50 mm. Figure S8. Immunohistochemistry for SGLT2. Representative images showing immunoreactivity for SGLT2 (brown) at the apical membrane of tubules in the renal cortical region. Note: there is nonspecific staining on the cast in the image of the ischemic kidney from vehicletreated mice (top right-hand side of the image). Bar ¼ 100 mm. Figure S9. In vivo glucose uptake test results. A fluorescent glucose analogue, NBDG, was injected i.v., and the changes in fluorescent intensity (green) were analyzed. The faint green dots visible at baseline are due to tubule autofluorescence, and autofluorescence was not counted. The bottom image represents tubule autofluorescence excited by a 720-nm laser; the blue fluorescence mainly derived from cytosolic NADH overlapped with NBDG-derived fluorescence. L, tubular lumen. Figure S10. Expression of SGLT1 and SGLT2 in mProx24 cells. Polymerase chain reaction (PCR) was performed to assess the SGLT1 and SGLT2 expression levels in mProx1 cells. Mouse kidney tissue was used as a positive control, and mouse heart, brain, and liver were used as negative controls. The PCR products for SGLT2 (170 bp) and SGLT1 (128 bp) were electrophoresed in 2% agarose gels in Trisborate EDTA buffer and then stained with ethidium bromide. Figure S11. In vitro glucose uptake test in GLUT-inhibited cells. GLUT2 was inhibited in cultured proximal tubular cells by using a GLUT2 inhibitor, cytochalasin B (left), or transfection with a GLUT2specific siRNA (right). These cells were subsequently treated with luseogliflozin or vehicle, and in vitro glucose uptake tests were conducted by using a fluorescence glucose analogue, NBDG. *P < 0.05 versus control or scramble, yP < 0.05 versus luseogliflozin. Figure S12. VEGF-A mRNA expression level after hypoxiareoxygenation in cultured mouse proximal tubular cells. The VEGF-A mRNA expression level of proximal tubular (mProx24) cells, cultured in the absence of glucose, from vehicle- or luseogliflozin (luseo)treated cells after hypoxia-reoxygenation. 11

basic research

Supplementary material is linked to the online version of the paper at www.kidney-international.org. REFERENCES 1. Ferrannini E, Solini A. SGLT2 inhibition in diabetes mellitus: rationale and clinical prospects. Nat Rev Endocrinol. 2012;8:495–502. 2. Santer R, Calado J. Familial renal glucosuria and SGLT2: from a mendelian trait to a therapeutic target. Clin J Am Soc Nephrol. 2010;5:133–141. 3. Zinman B, Wanner C, Lachin JM, et al. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. 2015;373: 2117–2128. 4. Wanner C, Inzucchi SE, Lachin JM, et al. Empagliflozin and progression of kidney disease in type 2 diabetes. N Engl J Med. 2016;375:323–334. 5. Neal B, Perkovic V, Mahaffey KW, et al. Canagliflozin and cardiovascular and renal events in type 2 diabetes. N Engl J Med. 2017;377:644–657. 6. Nakano D, Doi K, Kitamura H, et al. Reduction of tubular flow rate as a mechanism of oliguria in the early phase of endotoxemia revealed by intravital imaging. J Am Soc Nephrol. 2015;26:3035–3044. 7. Munshi R, Hsu C, Himmelfarb J. Advances in understanding ischemic acute kidney injury. BMC Med. 2011;9:11. 8. Gammelager H, Christiansen CF, Johansen MB, et al. Three-year risk of cardiovascular disease among intensive care patients with acute kidney injury: a population-based cohort study. Crit Care. 2014;18:492. 9. Takahashi A, Kimura T, Takabatake Y, et al. Autophagy guards against cisplatin-induced acute kidney injury. Am J Pathol. 2012;180:517–525. 10. Basile DP, Friedrich JL, Spahic J, et al. Impaired endothelial proliferation and mesenchymal transition contribute to vascular rarefaction following acute kidney injury. Am J Physiol Renal Physiol. 2011;300:F721–F733. 11. Nishioka S, Nakano D, Kitada K, et al. The cyclin-dependent kinase inhibitor p21 is essential for the beneficial effects of renal ischemic preconditioning on renal ischemia/reperfusion injury in mice. Kidney Int. 2014;85:871–879. 12. Braun H, Schmidt BM, Raiss M, et al. Cellular senescence limits regenerative capacity and allograft survival. J Am Soc Nephrol. 2012;23: 1467–1473. 13. Kitada K, Nakano D, Ohsaki H, et al. Hyperglycemia causes cellular senescence via a SGLT2- and p21-dependent pathway in proximal tubules in the early stage of diabetic nephropathy. J Diabetes Complications. 2014;28:604–611. 14. Chang YK, Choi H, Jeong JY, et al. Dapagliflozin, SGLT2 inhibitor, attenuates renal ischemia-reperfusion injury. PLoS One. 2016;11: e0158810. 15. Wald FA, Figueroa Y, Oriolo AS, et al. Membrane repolarization is delayed in proximal tubules after ischemia-reperfusion: possible role of microtubule-organizing centers. Am J Physiol Renal Physiol. 2003;285: F230–F240. 16. Basile DP, Bonventre JV, Mehta R, et al. Progression after AKI: understanding maladaptive repair processes to predict and identify therapeutic treatments. J Am Soc Nephrol. 2016;27:687–697. 17. Aguilera KY, Brekken RA. Hypoxia studies with pimonidazole in vivo. Bio Protoc. 2014;4. 18. You WK, Yotsumoto F, Sakimura K, et al. NG2 proteoglycan promotes tumor vascularization via integrin-dependent effects on pericyte function. Angiogenesis. 2014;17:61–76.

12

Y Zhang et al.: SGLT2 inhibition attenuates renal capillary injury

19. Gomez IG, Roach AM, Nakagawa N, et al. TWEAK-Fn14 signaling activates myofibroblasts to drive progression of fibrotic kidney disease. J Am Soc Nephrol. 2016;27:3639–3652. 20. Nakano D, Nishiyama A. Multiphoton imaging of kidney pathophysiology. J Pharmacol Sci. 2016;132:1–5. 21. Sabolic I, Vrhovac I, Eror DB, et al. Expression of Naþ-D-glucose cotransporter SGLT2 in rodents is kidney-specific and exhibits sex and species differences. Am J Physiol Cell Physiol. 2012;302:C1174–C1188. 22. Kanai Y, Lee WS, You G, et al. The human kidney low affinity Naþ/ glucose cotransporter SGLT2. Delineation of the major renal reabsorptive mechanism for D-glucose. J Clin Invest. 1994;93:397–404. 23. Thorens B, Mueckler M. Glucose transporters in the 21st century. Am J Physiol Endocrinol Metab. 2010;298:E141–E145. 24. Shweiki D, Neeman M, Itin A, et al. Induction of vascular endothelial growth factor expression by hypoxia and by glucose deficiency in multicell spheroids: implications for tumor angiogenesis. Proc Natl Acad Sci U S A. 1995;92:768–772. 25. Stein I, Neeman M, Shweiki D, et al. Stabilization of vascular endothelial growth factor mRNA by hypoxia and hypoglycemia and coregulation with other ischemia-induced genes. Mol Cell Biol. 1995;15:5363–5368. 26. Sone H, Kawakami Y, Okuda Y, et al. Vascular endothelial growth factor is induced by long-term high glucose concentration and up-regulated by acute glucose deprivation in cultured bovine retinal pigmented epithelial cells. Biochem Biophys Res Commun. 1996;221:193–198. 27. Carmeliet P, Dor Y, Herbert JM, et al. Role of HIF-1alpha in hypoxiamediated apoptosis, cell proliferation and tumour angiogenesis. Nature. 1998;394:485–490. 28. Dimke H, Sparks MA, Thomson BR, et al. Tubulovascular cross-talk by vascular endothelial growth factor a maintains peritubular microvasculature in kidney. J Am Soc Nephrol. 2015;26:1027–1038. 29. Tahara A, Takasu T, Yokono M, et al. Characterization and comparison of sodium-glucose cotransporter 2 inhibitors in pharmacokinetics, pharmacodynamics, and pharmacologic effects. J Pharmacol Sci. 2016;130:159–169. 30. Birbrair A, Zhang T, Files DC, et al. Type-1 pericytes accumulate after tissue injury and produce collagen in an organ-dependent manner. Stem Cell Res Ther. 2014;5:122. 31. Duffield JS. Cellular and molecular mechanisms in kidney fibrosis. J Clin Invest. 2014;124:2299–2306. 32. Grgic I, Krautzberger AM, Hofmeister A, et al. Translational profiles of medullary myofibroblasts during kidney fibrosis. J Am Soc Nephrol. 2014;25:1979–1990. 33. Caporali A, Meloni M, Nailor A, et al. p75(NTR)-dependent activation of NF-kappaB regulates microRNA-503 transcription and pericyteendothelial crosstalk in diabetes after limb ischaemia. Nat Commun. 2015;6:8024. 34. Mizushima N, Yamamoto A, Matsui M, et al. In vivo analysis of autophagy in response to nutrient starvation using transgenic mice expressing a fluorescent autophagosome marker. Mol Biol Cell. 2004;15:1101–1111. 35. Kusaba T, Lalli M, Kramann R, et al. Differentiated kidney epithelial cells repair injured proximal tubule. Proc Natl Acad Sci U S A. 2014;111: 1527–1532. 36. Matsumoto S, Batra S, Saito K, et al. Antiangiogenic agent sunitinib transiently increases tumor oxygenation and suppresses cycling hypoxia. Cancer Res. 2011;71:6350–6359.

Kidney International (2018) -, -–-