Angiotensin II Promotes Development of the Renal Microcirculation through AT₁ Receptors

Abstract

Pharmacologic or genetic deletion of components of the renin-angiotensin system leads to postnatal kidney injury, but the roles of these components in kidney development are unknown. To test the hypothesis that angiotensin II supports angiogenesis during postnatal kidney development, we quantified CD31⁺ postglomerular microvessels, performed quantitative PCR analysis of vascular growth factor expression, and measured renal blood flow by magnetic resonance. Treating rats with the angiotensin II type 1 receptor antagonist candesartan for 2 weeks after birth reduced the total length, volume, and surface area of capillaries in both the cortex and the medulla and inhibited the organization of vasa recta bundles. In addition, angiotensin II type 1 antagonism inhibited the transcription of angiogenic growth factors vascular endothelial growth factor, angiopoietin-1, angiopoietin-2, and the angiopoietin receptor Tie-2 in cortex and medulla. Similarly, Agtr1a^−/−;Agtr1b^−/− mouse kidneys had decreased angiopoietin-1, angiopoietin-2, and Tie-2 mRNAs at postnatal day 14. To test whether increased urinary flow leads to microvascular injury, we induced postnatal polyuria with either lithium or adrenalectomy, but these did not alter vascular endothelial growth factor expression or vasa recta organization. Compared with vehicle-treated rats, renal blood flow was significantly (approximately 20%) lower in candesartan-treated rats even 14 days after candesartan withdrawal. Taken together, these data demonstrate that angiotensin II promotes postnatal expansion of postglomerular capillaries and organization of vasa recta bundles, which are necessary for development of normal renal blood flow.

In rodents, significant parts of the kidneys develop after birth. Nephrogenesis proceeds in the subcapsular region until postnatal day 7 (P7) to 8 followed by growth and functional maturation of the kidney medulla. The kidney medulla proliferates through mitotic activity in loops of Henle that peaks around P14 and is completed in the fourth week.¹ This is reflected by a marked surge in urine-concentrating ability. Pharmacologic inhibition or targeted gene deletion of components of the renin-angiotensin system (RAS) leads to similar postnatal phenotypes with diminished kidney growth, fewer glomeruli, and medullary injury with decreased ability to concentrate the urine.^2–8 The mechanism for the postnatal injury of impaired RAS activity on kidney development is not known. Angiotensin II (AngII) promotes urine propagation,⁹ and inhibition of AngII signaling leads to pressure-related damage to the kidney epithelium,¹⁰ which supports that medullary injury is secondary to increased urinary output.¹¹ Other studies showed increased accumulation of extracellular matrix,^3,12 increased apoptosis, decreased cell proliferation,¹³ and impaired expression of growth factors in kidney after RAS inhibition.¹⁴ A prerequisite for tissue growth is sufficient supply with oxygen and nutrients provided by capillaries. Impaired AngII signaling leads to a reduced number of glomeruli and to pathologic alterations of the preglomerular vasculature as shown by Tufro-McReddie et al. ^15,16 AngII promotes angiogenesis in vitro^17,18 and stimulates vascular growth factor expression in diabetic kidney.¹⁹ The possibility that AngII is causally related to capillary expansion and patterning in the postnatal phase with marked growth of kidney medulla has not yet been explored. We hypothesized that postnatal development of the kidney depends on AngII-stimulated angiogenesis and that medullary injury after RAS inhibition is caused by impaired development of the microvasculature. The hypothesis was addressed in vivo by application of unbiased quantitative stereologic methods. Postglomerular microvascular length, volume, and surface area were determined in rat kidneys at P14 after treatment with AngII type 1 (AT₁) inhibitor during the first 2 weeks of postnatal life. Mechanisms were addressed by analysis of vascular growth factor expression and localization in postnatal kidney tissue. The consequence for renal blood flow (RBF) in adolescence was measured by magnetic resonance imaging (MRI).

Results

Effect of AT₁ Receptor Blockade on Postglomerular Microvessel Quantity and Kidney Size in the Postnatal Period

Total length of cortical peritubular capillaries was 288 ± 20 m in control rat kidneys at P14 (n = 10; Figure 1A). In the candesartan-treated rats, the length of postglomerular microvessels in the cortex was reduced significantly (211 ± 18 m; P < 0.01; n = 10; Figure 1A). Total surface area of peritubular cortical capillaries was reduced significantly in the AT₁ blocker group (Figure 1A). Total volume of cortical peritubular capillaries was not different between the two experimental groups (Figure 1A). Immunolabeling of kidney sections for the endothelial cell marker CD31 showed a dense capillary network in the cortical labyrinth, where virtually all basolateral membranes of tubular cells were directly apposed to endothelial cells. There was distinct labeling of glomerular capillaries (Figure 1C). In response to the AT₁ inhibitor, the cortical labyrinth was more cell dense and tubules displayed irregular shapes with dilation, atrophy, and closed luminas. The capillary network was less widespread. The major part of basolateral epithelial cell boundaries were not in contact with CD31⁺ cells (Figure 1D). In medulla of control rats, microvessel length was 235 ± 16 m (n = 10; Figure 1B). In the AT₁ blocker group, total length was reduced significantly to 133 ± 6 m (P < 0.0001; n = 10; Figure 1B). Total capillary volume and total capillary surface area were significantly reduced in kidney medulla by AT₁ inhibition (Figure 1B). Cross-sections of the renal outer medulla from control rats at P14 displayed multiple CD31⁺ regularly distributed vasa recta bundles with approximately 30 microvessels in each and interbundle capillary profiles (Figure 1E). AT₁ inhibitor–treated kidney medulla showed no regularly organized vascular bundles and less patent tubular profiles (Figure 1F). In cortex and medulla, the radius (largest distance in tissue from a capillary) and area of diffusion were increased significantly in the candesartan-treated animals (Table 1). There were no differences in the calculated diameter of capillaries between the two experimental groups: In cortex 8.6 ± 0.2 (control) versus 8.9 ± 0.3 μm (AT₁ blocker) and medulla 8.8 ± 0.4 (control) versus 9.8 ± 0.4 μm (candesartan). Kidney/body weight ratios were not different between AT₁ inhibitor–treated rats and controls (data not shown). The inner medulla volume was significantly reduced (approximately 50%) after candesartan treatment (Table 2). No significant changes were seen in the total volume of cortex, outer stripe of outer medulla, or inner stripe of outer medulla (Table 2). There was no significant difference between total volume of kidney medulla in the two groups (Table 2). Plasma renin concentration (control 19.6 × 10⁻⁵ GU/ml ± 1.5; AT₁ blocker 7372.6 × 10⁻⁵ GU/ml ± 958; n = 7 in each group; P < 0.001) and renin mRNA level (control 11.5 ± 1.2 renin/18S ratio; AT₁ blocker 308 ± 30.1 renin/18S ratio; n = 9 in each group; P < 0.0001) increased significantly compared with the control group.

Figure 1.

(A) Quantitative stereologic estimations of kidney cortical peritubular capillaries in control rats (C; n = 10) and in candesartan-treated animals (Can; 1 mg/kg per d, P1 to P13; n = 10). Total length (in meters; top) and surface area (in cm²; middle) of microvessels were significantly decreased after AT₁ receptor inhibition. Total volume in mm³ (bottom) was not affected. Data are means ± SEM. *P < 0.05. (B) Quantitative stereologic estimations of kidney medullary CD31⁺ microvessels including vasa recta in control rats (C; n = 10) and in candesartan-treated animals (Can; 1 mg/kg per d, P1 to P13; n = 10). Total length in meters, surface area in cm², and volume in mm³ were significantly decreased in response to AT₁ receptor inhibition. Data are means ± SEM, *P < 0.05. (C and D) Immunohistochemical staining of kidney cortex for the endothelial cell marker CD31. (C) In control rats, cortex displayed a dense capillary network in which virtually all tubular cells were directly apposed to endothelial cells and distinct labeling of glomerular capillaries. (D) In candesartan-treated rats, the capillary network was less widespread with increased tubular-endothelial distance(n = 10 in each group). Bar = 50 μm. (E and F) Immunohistochemical staining of cross-sections of renal outer medulla with an antibody specific for the endothelial cell marker CD31. (E) Sections from control rat pups displayed distinct vasa recta bundles and capillaries. (F) Candesartan-treated rats showed disrupted vasa recta architecture and abnormally thickened capillaries (n = 10 animals in each group). Bar = 50 μm.

Table 1.

Stereologic analysis of total area and radius of diffusion in cortex and medulla after candesartan treatment from P1 to P13

Parameter	Control (n = 10)	Candesartan (n = 10)
Radius of diffusion in cortex (μm)	12.7 ± 0.3	14.3 ± 0.4a
Area of diffusion in cortex (μm2)	505 ± 24	649 ± 41a
Radius of diffusion in medulla (μm)	10.5 ± 0.4	13.4 ± 0.4b
Area of diffusion in medulla (μm2)	351 ± 23	566 ± 32b

Data are means ± SEM.

^aP < 0.01.

^bP < 0.0001.

Table 2.

Stereologic analysis of total volume of the various kidney zones after candesartan treatment from P1 to P13

Parameter	Control (n = 10)	Candesartan (n = 10)
Whole kidney (mm3)	253 ± 12	236 ± 18
Cortex (mm3)	142 ± 5	135 ± 12
OSOM (mm3)	27 ± 4	28 ± 4
ISOM (mm3)	55 ± 4	63 ± 4
IM (mm3)	25 ± 2	12 ± 2a
Medulla total (mm3)	108 ± 8	102 ± 7

Data are means ± SEM. IM, inner medulla; ISOM, inner stripe of outer medulla; OSOM, outer stripe of outer medulla.

^aP < 0.0001.

Effect of AT₁ Receptor Antagonist Treatment and Targeted Deletion of AT₁ Receptors on Expression of Angiogenic Factors in the Postnatal Kidney

In renal cortex, vascular endothelial growth factor (VEGF), angiopoietin-1, and angiopoietin-2 mRNA levels were significantly decreased in the candesartan-treated animals compared with the control group (Figure 2A). Messenger RNA levels of the receptors Flk-1, Flt-1, and Tie-2 did not change (Figure 2A). In kidney medulla, VEGF, angiopoietin-1, angiopoietin-2, Flk-1, Flt-1, and Tie-2 mRNAs were significantly reduced by candesartan (Figure 2B). In whole kidney tissue from Agtr1a^−/−Agtr1b^−/− mice at P14, there was a significantly lower level of angiopoietin-2, angiopoietin-1, Tie-2, and Flk-1 mRNAs (Figure 2C). Renin mRNA was significantly increased compared with wild-type mice (28.8 ± 1.9 renin/18S versus 221 ± 24.7 renin/18S; n = 8 in each group; P < 0.0001; data not shown). At P14, VEGF mRNA was detected in microdissected outer medullary collecting duct (OMCD), medullary thick ascending limb of loop of Henle (mTAL), but not in thin descending limb of Henle's loop (Figure 2D, Table 3). Aquaporin 2 (AQP-2) was amplified in OMCD only (Figure 2D, Table 3). Amplification of sodium-potassium 2 chloride co-transporter and AQP-1 confirmed segment identity (Figure 2D, Table 3). Western blotting for VEGF confirmed a significant lower protein level in medulla after AT₁ inhibition (Figure 2E). AT_1A receptor was detected in OMCD and mTAL but not thin descending limb of Henle's loop (Figure 2F, Table 3). AT_1B was not detected (data not shown). At P14, VEGF immunoreactivity was associated with collecting ducts, primarily in cortex and outer medulla (Figure 2G). There were fewer VEGF-positive segments in candesartan-treated kidneys (Figure 2G, right). Omission of VEGF antibody and preabsorption of VEGF antibody with the immunizing peptide abolished immunostaining (data not shown). In vitro, incubation of immature glomeruli from control rats at P14 with AngII increased significantly VEGF mRNA level (control 52.8 ± 4.8; AngII 10⁻⁷ mol/L 84 ± 2.6; AngII 10⁻⁶ mol/L 82.2 ± 4.3; all values VEGF/18S ratio, P < 0.001).

Figure 2.

(A) Vascular growth factor and growth factor receptor mRNA levels in rat kidney cortex measured by quantitative PCR in control rats (n = 9) and in candesartan-treated rats at P14 (1 mg/kg per d; n = 9). AT₁ receptor blockade in the postnatal period resulted in significantly decreased mRNA levels of VEGF, angiopoietin-1, and angiopoietin-2. Growth factor receptor expression was not affected. Messenger RNA level is shown relative to 18S rRNA level and is expressed as the percentage change from the expression level in control animals. Columns show means ± SEM. *P < 0.05. (B) Vascular growth factor and growth factor receptor mRNA levels in rat kidney medulla measured by quantitative PCR in control rats (n = 9) and in candesartan-treated rats at P14 (1 mg/kg per d; n = 9). AT₁ receptor inhibition in the postnatal period resulted in significantly suppressed levels of VEGF, angiopoietin-1, and angiopoietin-2 mRNAs and, in addition, in impaired growth factor receptor mRNAs (Flt, Flk, and Tie-2). Messenger RNA level is shown relative to 18S rRNA level and is expressed as the percentage change from the expression level in control animals. Columns show means ± SEM. *P < 0.05; **P < 0.01. (C) Vascular growth factor and growth factor receptor mRNA levels in AT_1A/1B^−/− double-knockout mouse kidneys (n = 9) and wild-type littermate kidneys at P14 (n = 9). AT_1A/1B^−/− mice displayed suppressed angiopoietin-1, angiopoietin-2, and Tie-2 mRNAs, whereas VEGF mRNA level was unaltered. The mRNA level of the products of interest are presented relative to 18S rRNA levels and are shown as the percentage change from the expression level in wild-type littermates. Columns show means ± SEM.*P < 0.05; **P < 0.01. (D) PCR analysis of microdissected outer medullary nephron segments from control rats (P14). VEGF mRNA was observed in OMCDs and in mTAL but not in descending thin limb of loop of Henle (DTL). AQP-2 served as a positive control for OMCD, sodium-potassium 2 chloride co-transporter (NKCC2) as positive control for mTAL, and AQP-1 as positive control for DTL. Negative controls were omission of reverse transcriptase (−RT) and addition of water instead of cDNA in the amplification. Size marker is ΦX174DNA/HaeIII fragments. (E) Western blotting experiment for VEGF protein abundance in kidney medulla homogenates from control (C; n = 9) and candesartan-treated rats (Can; 1 mg/kg per d; n = 8). *P < 0.05. (F) Amplification products for AT_1A receptor mRNA was observed in OMCD and mTAL. Negative controls were omission of reverse transcriptase (−RT) and addition of water instead of cDNA in the amplification. Size marker is ΦX174DNA/HaeIII fragments. (G) Immunohistochemical labeling of rat kidney sections from P14 for VEGF. In control rats, distinct labeling of outer medullary tubular structures compatible with OMCD and mTAL segments was seen. In outer medulla from candesartan-treated rat pups (P1 to P13), fewer VEGF-positive tubular segments were present and the tissue appeared more irregular with fewer tubular profiles (n = 10 in each experimental group). Bar = 50 μm.

Table 3.

Summary of expression studies by PCR analysis of microdissected structures obtained in rat kidney in the third postnatal week

Product	Glomeruli	Outer Medullary Vasa Recta	Descending Limb of Henle's Loop	Thick Ascending Limb of Henle's Loop	OMCDs
AT1A	+	+	−	+	+
VEGF	+	NE	−	+	+
CD31	+	+	NE	NE	NE
AQP2	NE	NE	−	−	+
NKCC2	NE	NE	−	+	−
AQP1	NE	NE	+	−	−

NE, not examined; NKCC2, sodium-potassium 2 chloride co-transporter.

Effect of an AT₁ Receptor Inhibitor on Renal Interstitial Cells

In control kidneys, α-smooth muscle actin (α-SMA) was associated with preglomerular arteries and arterioles and vasa recta bundles (Figure 3, A and B). In the AT₁ blocker group, labeling for α-SMA was associated with spindle-shaped cells that populated the interstitium in outer medulla, in cortical medullary rays, and in the subcapsular region (Figure 3, D and E). α-SMA protein abundance was elevated in cortex and medulla from candesartan-treated rats at P14 (Figure 3G). Interstitial cells were negative for E-cadherin (Figure 3, C and F). E-cadherin was associated with basolateral membranes. There were no significant differences in E-cadherin protein abundance between control and AT₁ inhibition in cortex and medulla (data not shown). The renal stem cell marker CD133 was associated with endothelial cells in vasa recta in controls and after AT₁ inhibition (Figure 3, H and I). Outer medullary interstitial cells were negative for CD133 (Figure 3I). A distinct subset of interstitial cells in the papilla were CD133⁺ in candesartan-treated rats (Figure 3J).

Figure 3.

(A and B) Immunohistochemical staining of kidney sections from control rats at P14 (n = 8) with an antibody specific for α-SMA. α-SMA was restricted to preglomerular arteries and arterioles (A) and medullary vasa recta bundles (B). Bar = 500 μm in A and 50 μm in B (outer medulla). (D and E) Immunohistochemical staining of kidney sections from candesartan-treated rat pups (1 mg/kg per d, P1 to P13; n = 8) for α-SMA. Immunolabeling was associated with preglomerular vessels. α-SMA–positive spindle-shaped cells populated predominantly the outer medullary interstitium and cortical medullary rays (E). Bar = 500 μm in D and 50 μm in E. (C and F) Immunohistochemical staining of outer medulla from control and candesartan-treated rats (1 mg/kg per d, P1 to -P13) with an antibody specific for E-cadherin. Distinct E-cadherin labeling was associated with renal tubules, most clearly collecting ducts, in both control and AT₁ inhibitor–treated animals. Cells populating the outer medullary interstitium in candesartan-treated animals were negative for E-cadherin (n = 8). Bar = 50 μm. (G) Western blotting experiments for α-SMA protein abundance in kidney tissue homogenates from control (C; n = 9) and candesartan-treated rats (Can; 1 mg/kg per d; n = 8). A band of 42 kD was detected for α-SMA. AT₁ receptor inhibition resulted in increased α-SMA protein abundance in both cortex and medulla compared with control animals. Columns show means ± SE. *P < 0.05. (H through J) Immunohistochemical staining of outer medulla (H and I) and inner medulla (J) from control (H) and candesartan-treated rats (I and J; 1 mg/kg per d, P1 to P13) with an antibody specific for the stem cell marker CD133. Distinct CD133 labeling was associated with endothelial cells and confirmed irregular vasa recta bundle organization after candesartan treatment (H versus I). (I) Interstitial cells in outer medulla were negative for CD133. (J) CD133⁺ interstitial cells appeared in candesartan-treated rat kidney inner medulla. Bars = 50 μm in H and I and 200 μm in J.

Effect of Postnatal Polyuria on VEGF Expression and Medullary Microvessel Organization

VEGF mRNA level was not altered significantly at P20 in kidneys from rats adrenalectomized (ADX) at P10 compared with sham-operated controls or by lithium (Li) treatment from P0 through P14 compared with controls (Figure 4, A and D). CD31 labeling showed vasa recta bundles in kidneys from both ADX rats and Li-treated rats (Figure 4, B and E). Immunohistochemical labeling of kidney sections from ADX rats and Li-treated rats for VEGF showed immunoreactive protein associated with collecting ducts similar to control kidneys in both series (Figure 4, C and F).

Figure 4.

(A) The bar graph shows the effect of Li treatment from P0 to P14 on VEGF mRNA levels in rat kidney cortex and medulla as measured by quantitative PCR. C, control (n = 10); Li, Li-treated rats (50 mmol/kg chow; n = 5). There was no significant effect of Li treatment on mRNA level of VEGF. Messenger RNA level is shown relative to TATA-box binding protein (TBP) RNA level. Columns show means ± SEM. (B) Immunohistochemical labeling of kidney section from Li-treated rat (P0 to P14) for the endothelial marker CD31. A descending vasa recta bundle at the cortical outer medulla junction is seen. Bar = 50 μm. In each experimental group, n = 5. (C) Immunohistochemical labeling of rat kidney sections from control rats (n = 5) and Li-treated rats (n = 5) for VEGF (P0 to P14, 50 mmol/kg chow). There was no difference in the labeling pattern; the micrograph shows immunoreactive VEGF protein associated with outer medullary tubular structures compatible with OMCD and mTAL segments. Bar = 50 μm. (D) The graph shows the effect of ADX at P10 on VEGF mRNA levels in rat kidney at P20 as measured by quantitative PCR (sham-operated control n = 9, ADX-treated rats n = 6). There was no significant effect of ADX on VEGF mRNA level. Messenger RNA level is shown relative to 18S rRNA level. Columns show means ± SEM. (E) The micrograph shows immunohistochemical labeling of kidney section for VEGF. The tissue was from rats adrenalectomized at P10 and analyzed at P20 (n = 4 to 5). A vasa recta bundle is not different from sham-operated rat kidney at the cortical outer medulla junction is seen. Bar = 50 μm. (F) Effect of ADX at P10 on VEGF immunoreactivity in rat kidney. Sections were from P20 and show tissue from sham-operated control rats and adrenalectomized rats (n = 4 to 5). There was no difference in the labeling pattern; immunoreactive protein was associated with outer medullary tubular structures compatible with OMCD and mTAL segments. Bar = 50 μm.

Effect of an AT₁ Receptor Antagonist on Vascular Proliferation

There was no significant difference in proliferating cell nuclear antigen (PCNA) abundance in response to candesartan in kidney cortex and medulla tissue by Western blotting (Figure 5A). PCNA mRNA level decreased significantly between P14 and adult rat in both cortex (approximately 32 times) and medulla (approximately eight times; Figure 5B). Vascular bundles were microdissected from rat pups treated with candesartan or vehicle for 3 days: P11 to P13, P13 to P15, or P17 to P19. PCNA mRNA level was significantly lower after candesartan treatment from P11 to P13 but not at P16 and P20, compared with vehicle control (P < 0.05; Figure 5C). The microdissected outer medullary vascular bundles expressed CD31 and AT_1A receptors (Figure 5D, Table 3).

Figure 5.

(A) Western blotting experiments for the proliferation marker PCNA in homogenates from control rat kidney cortex and medulla (n = 5) and candesartan-treated (P1 to P13, 1 mg/kg per d) rat kidney cortex (n = 4) and medulla (n = 4). Expected size is 42 kD. No changes in PCNA protein abundance was seen between the two experimental groups. (B) Effect of rat age on PCNA mRNA level in kidney cortex and medulla. Kidney tissue was from P14 and adult rats (>2 months). PCNA mRNA expression was significantly decreased in both cortex and medulla from adult rats compared with P14. Columns show means ± SEM (SE not visible; n = 3 in each group). (C) PCNA mRNA level in vasa recta bundles microdissected from rats treated with vehicle or candesartan (1 mg/kg per d) during one of the following time periods: P11 to P13 (n = 4 control, n = 4 candesartan), P13 to P15 (n = 4 control, n = 4 candesartan), or P17 to P19 (n = 4 control, n = 4 candesartan). AT₁ receptor blockade at P11 to P13 but not at later stages resulted in significantly decreased PCNA mRNA in vasa recta bundles compared with the control group. The results are normalized to TBP mRNA level. Data are means ± SEM. *P < 0.05. Micrograph (right) shows a microdissected vasa recta bundle from rat outer medulla. (D) PCR analysis of microdissected vasa recta bundles for the endothelial marker CD31 (left; Can, candesartan; C, control; P, postnatal day) and AT_1A receptor. CD31 and AT_1A were readily amplified from all samples of microdissected vasa recta bundles. Negative controls were omission of reverse transcriptase (−RT) and cDNA (−cDNA). Size marker is ΦX174DNA/HaeIII fragments. (E) Schematic outline of proposed mechanisms. AngII through the AT₁ receptor in collecting duct and loop of Henle epithelium supports expression of VEGF (and other vascular growth factors) that are released to the renal interstitium. VEGF and angiopoietins occupy cognate receptors end stimulate endothelial proliferation and vascular patterning in the cortex and the medulla.

Effect of Postnatal AT₁ Receptor Blockade on Kidney Morphology and RBF in Young Rats

In rats treated with candesartan from P1 through P13 and left untreated from P14 through P30, the kidneys exhibited injury of the medulla with a small papilla and thickened vessels (Figure 6, A through D). There was accumulation of α-SMA in the outer medulla interstitium at P30 (Figure 6B). Total RBF as determined by dynamic MRI at P30 was reduced significantly in the candesartan group (Figure 6, E through H). Total kidney volume at P30 was significantly smaller in the candesartan group (576 ± 20 mm³; n = 7) compared with control kidneys (667 ± 20 mm³; n = 6).

Figure 6.

(A and B) α-SMA immunostaining of kidney sections from control rat (A) and a rat treated with candesartan (1 mg/kg per d, P1 to P13; B). Rats were not treated from P14 until P30. The renal papilla displayed significant injury at P30 (B versus A). (B) α-SMA–positive cells accumulated in the interstitium, especially in the outer medulla (n = 4). Bar = 200 μm. (C and D) At P30, labeling for α-SMA yielded positive signals from preglomerular arteries and arterioles. (D) Candesartan treatment (P1 to P14) resulted in thickened arterioles with hypertrophied media layer compared with control kidneys. Bar = 50 μm. (E and F) Representative images from magnetic resonance visualization show the relative intrarenal blood flow in rats treated with vehicle (E) or candesartan (1 mg/kg per d, P0 to P13; F). Red arrows indicate the site (aorta) where the arterial input function was obtained necessary in the kinetic analysis using a two-compartment model, describing the intravascular and extravascular spaces, respectively. (G) Control high-resolution image obtained by MRI for anatomic localization of kidneys and aorta in rat (P30) treated with vehicle from P1 to P13. (H) Effect of candesartan treatment from P1 to P13 followed by withdrawal from P14 to P30 on total RBF as measured by MRI at P30. Columns show means ± SE (n = 8 in each group). *P < 0.05.

Discussion

This study used unbiased quantitative stereologic methods to show that impaired AngII AT₁ receptor signaling through the first 2 postnatal weeks leads to a significant reduction in the absolute length, volume, and surface area of renal postglomerular microvessels, a greater intercapillary distance, and a disturbed microvessel organization, most prominent in kidney medulla. AT_1A receptor and vascular growth factors were detected in outer medullary tubular epithelium. In response to AT₁ receptor blockade, vascular growth factor mRNAs and cell proliferation in vasa recta bundles were suppressed. After a period of AT₁ inhibitor withdrawal, the renal injury was exacerbated and total RBF was reduced in young rats. In two separate experimental rat models with postnatal polyuria and intact AngII signaling, VEGF expression and vascular bundle organization were not affected. The data are compatible with a specific vascular growth–promoting effect of AngII-AT₁ signaling in the early postnatal period.

The calculated diameter of cortical and medullary microvessels yielded values of 8 to 9 μm, which provides a valuable internal control for the validity of the primary stereologic estimates of capillary dimensions. In accordance with previous observations, the inner medulla was the most severely affected kidney region,^2,6,7 and the medullary microvasculature exhibited a larger sensitivity to AT₁ inhibition compared with cortex. Little is known about vascular proliferation in the renal medulla. Elongation of vasa recta and assembly into bundles are obligate postnatal processes in rodents because immature bundles are first observed postnatally at approximately day 14 and developed fully at the time of weaning in rat (P21).²⁰ This study shows a significant role of AngII-AT₁ receptors to govern these patterning processes in the medulla. The coexpression of AT₁ receptors and VEGF in developing collecting duct and mTAL epithelium and the suppression of VEGF by AT₁ inhibition support that the epithelium relays AngII signals to promote vascular endothelial growth. AngII may also stimulate vasa recta pericyte proliferation directly through AT₁ receptors.²¹ The ureteric bud stimulates growth of the mesenchyme and differentiation of nephrons at early developmental stages and at later stages, collecting duct clusters are believed to constitute the “organizing unit” around which vasa recta and probably also capillaries group.²² Cultured renal epithelial cells secrete VEGF and stimulate migration of endothelial cells and capillary formation.²³ Infusion of AngII stimulates VEGF expression in adult rat kidney through AT₁ receptors.²⁴ VEGF and its receptors are crucial for vessel formation.^25–31 The observed downregulation of Flt and Flk mRNAs after AT₁ inhibition is in line with suppression of VEGF, because it is observed typically after diminished ligand binding.³² Suppression of angiopoietin-1 and -2 and their receptors by AT₁ inhibition may also contribute to disturbed vessel expansion.³² Targeted deletion of AT₁ receptors in mice yielded essentially similar data on renal vascular growth factor level as observed after AT₁ antagonist treatment in rat, which supports a specific effect of the antagonist on AT₁ receptors. The observed injury is less likely to be caused by blood pressure or blood flow changes by AT₁ inhibition, because administration of hydralazine¹⁶ or a calcium channel antagonist, nifedipine,³³ in the postnatal period does not mimic the effect of AT₁ inhibition.

Ureter obstruction is associated with decreased tubular VEGF expression and regression of peritubular capillaries,^10,34 and AngII promotes pelvic peristalsis.^9,35 We observed no changes in VEGF mRNA level, VEGF immunolocalization, and vascular bundle organization in two experimental rat models with postnatal polyuria and intact AngII signaling. The microvascular injury is therefore not caused by excessive urine flow. Taken together, the data support that AngII promotes capillary endothelial expansion and outer medullary vascular bundle assembly during postnatal kidney development and that AngII-mediated stimulation of vascular growth factor release from the epithelium is likely to be involved. The postnatal renal developmental stages in rats and mice reflect late intrauterine stages in humans. At gestational week 35, human fetal kidney tissue from a mother treated with losartan (50 mg/kg per d) displayed pathologic changes similar to our observations in rat kidney: Poorly developed and short vasa recta, abundant mesenchymal tissue, shortening of loops of Henle, fewer developed collecting ducts, and vessel media thickening with hyperplasia of juxtaglomerular granular cells.⁸ This is in agreement with the notion that similar AngII-AT₁ receptor signaling mechanisms contribute to intrauterine development of the human kidney and to postnatal development of rat kidney.

In adults, AngII promotes fibrosis and end-stage kidney failure³⁶; however, in the developing kidney, AT₁ receptor inhibition promoted accumulation of α-SMA–positive cells in the renal interstitium, particularly in the outer medulla and cortical medullary rays, that was further aggravated 17 d after candesartan withdrawal. The cells were not positive for the stem cell marker CD133 and are most likely injury-related fibroblasts. The observation of injury predominantly along collecting ducts in outer medulla and cortex further supports a primary role for these tubular segments in the normal growth and differentiation response to AngII.

Inhibition of RAS in the postnatal period leads to hypertension in adulthood.^15,37 Decreased medullary blood flow has been associated with increased systemic BP.^38,39 Impaired vascular development in the kidney was associated with a lower basal blood flow at P30. Altered regulation of medullary hemodynamics could be causally involved in the development of hypertension. The injured medullary microvasculature would be expected to compromise countercurrent exchange and contribute also to reported decreases in sodium and urea concentrations in the inner medullary interstitium.³³ In summary, this study presents the first quantitative in vivo data that establish a role of AngII for capillary development in the immature kidney coupled to disturbed RBF in adolescence. The data show that impaired AngII AT₁ receptor signaling in the postnatal period leads to a significant reduction in postglomerular renal capillary length, surface area, and volume, especially in the renal medulla, to a reduced level of vascular growth factor expression, especially VEGF and angiopoietins, to a reduced proliferative index of developing vasa recta, to impaired vasa recta organization, and to a reduction in basal RBF flow in adolescence. We conclude that kidney injury after inhibition of the RAS during development may be caused by impaired AngII-mediated postnatal angiogenesis.

Concise Methods

Animals and Experimental Design

Animals were housed at the Biomedical Laboratory, University of Southern Denmark. All procedures conformed to the Danish national guidelines for the care and handling of animals and the published guidelines from the National Institutes of Health. Female Sprague-Dawley rats were used for breeding. Within 24 hours after birth, litters were reduced to 10 pups to ensure equal feeding. Dams and pups were kept on a 12:12-hour light:dark cycle and had free access to standard pathogen-free rat chow (Altromin-1310; 2 g/kg Na⁺, 5 g/kg Cl⁻) and tap water.

Series 1

Newborn rat pups were treated with subcutaneous injections of candesartan (1 mg/kg body wt per d) from P1 to P13. On P14, pups were anesthetized with intraperitoneal injections of pentobarbital sodium (Mebumal 50 mg/kg), and kidneys were fixed by systemic perfusion through the left cardiac ventricle with 4% phosphate-buffered paraformaldehyde for 4 minutes. Kidneys were removed, weighed separately, and further fixed in 4% paraformaldehyde for 24 hours before processing. Right kidney from each animal was used for stereologic analysis.

Series 2

Newborn rat pups were treated with candesartan as described in series 1. On P14, pups were killed by decapitation. Mixed trunk blood was collected in EDTA-coated tubes and centrifuged at 3000 × g for 10 minutes. Plasma was removed and stored at −20°C. Kidney tissue was quickly removed and kept on ice. Cortex and medulla were separated using a stereomicroscope. Tissue was snap-frozen in liquid nitrogen and stored at −80°C.

Series 3

Rat pups were treated with candesartan (1 mg/kg body wt per d) for 3 days in one of the following time periods: P11 to P13, P13 to P15, or P17 to P19. Pups were killed and kidneys used for microdissections of vasa recta bundles on P14, P16, or P20.

Series 4

Agtr1a −/−Agtr1b −/− double-knockout mice were bred and housed at the animal facility of the Durham Veterans Affairs Medical Center according to National Institutes of Health guidelines with the approval of Institutional Animal Care and Use Committees of the Duke University and Durham VA Medical Centers. Two-week-old pups were killed, and kidneys removed and snap-frozen. Genotypes of mouse pups were determined from tail biopsies as described previously.⁶

Series 5

A group of rat pups treated with candesartan from P1 through P13 were used for measurements of RBF on P30. Animals were anesthetized by inhalation of O₂ and isoflurane in a ratio of 3:1.5%, and intravenous access was made through the tail vein. RBF was analyzed by dynamic MRI.

Series 6

Rat pups were adrenalectomized at P10 and were killed at P20. All procedures were as described previously.⁴⁰ This procedure leads to disappearance of aldosterone and impaired somatic growth and urine-concentrating ability and elevates plasma renin concentration approximately 30 times while the renal papilla is intact.⁴⁰

Series 7

Litters were reduced to a gender-matched size of eight, and dams were given pellets with 50 mmol/kg Li from P0 through P14.⁴¹ Pups were killed at P14, and kidneys were rapidly frozen or fixed in formaldehyde. This protocol yields plasma Li concentration of approximately 1 mmol/L and elevated urea at day 20 and results in loss of urine-concentrating ability and polyuria in young adult rats.⁴¹

Stereology

Stereologic measurements were done according to Gundersen et al.,⁴² Nyengaard et al.,⁴³ and Skyum et al.⁴⁴ Slices were cut 1 mm thick at right angle to the long axis of the fixed kidneys using a device with parallel razor blades. Every second slice was embedded in paraffin. Tissue slices were cut 4 μm thick, stained with periodic acid-Schiff, and used for estimation of volume of kidney zones. The rest of the kidney slices were processed for quantitative stereologic measurements of total capillary length, capillary surface area, and capillary volume. For these measurements, isotropic orientation of structures of interest is necessary. This was obtained using a modification of the previously published isector method.⁴⁵ Kidney slices were cut into approximately 1 × 1-mm large pieces using a scalpel blade and mixed. One sixth of the pieces were systematically randomly chosen and embedded in paraffin. Tissue sections were cut 4 μm thick and stained with the endothelial cell marker CD31 for correct identification of kidney capillaries. Stereologic quantifications were done using a light microscope with a motorized stage table connected to a video camera and a computer using the software Cast 2000 (Olympus). Volume fractions of kidney zones were done at ×4 magnification using point counting, and measurements of postglomerular capillaries were done at ×40 magnification. Grid density was determined from pilot studies. Preglomerular arteries, glomeruli, efferent arterioles, and veins were not included in the measurements.

Estimations of postglomerular microvessel length [L_(cap)] were made on the basis of the assumption that the structures of interest are much longer than their diameter and that the density of the kidney tissue is 10¹² μm³/g.⁴⁴ L_(cap) = 2 × Q_A(cap) × V_v × weight of the kidney × density, where Q_A(Cap) is the number of capillary profiles per area and V_v is the volume fraction of the region of interest (ROI; cortex or medulla). Surface area of capillaries was calculated by S_(cap) = 2 × I_(cap) × V_v × weight of the kidney × density, where I_(cap) is the number of profile intersections per area. Volume of capillaries was determined by V_(cap) = V_v(cap) × V_v × weight of the kidney × density, where V_v(cap) is the volume fraction of the capillaries. Capillary diameter was calculated through d = 2 × √V_(cap)/[π × L_(cap)]. The estimations are made from the assumption that the capillaries are cylindrical. Area and radius of diffusion were estimated by A = 1/L_v and radius of diffusion, r = √1/(L_v × π), where L_v is the relative length of the capillaries (μm/μm³ tissue). The estimations are made from the assumption that the capillaries are cylindrical and that the area and radius of diffusion are the same in all directions.

Microdissection of Nephron Segments and Vasa Recta Bundles

Kidney tissue was cut in thin coronal slices through the tip of the papilla and enzymatically treated with Collagenase A (1 mg/ml; Roche) in DMEM + 0.1% BSA for 25 minutes at 37°C during modest shaking. Microdissections were done at 4°C using a stereomicroscope. Nephron segments and vasa recta bundles were identified because of localization and appearance. Nephron segments were dissected at P14 from control rat kidneys. Vasa recta bundles were microdissected from rat pups treated with candesartan or vehicle between P11 and P13 (n = 4), P13 and P15 (n = 4), or P17 and P19 (n = 4). Length of segments was measured using a scale built in the ocular of the microscope. Segments were transferred to 4 M guanidinium thiocyanate with β-mercaptomethanol.

Glomeruli

Glomeruli were isolated from control rats at P14. Kidneys were collected and immediately transferred to ice-cold PSS (115 mM NaCl, 25 mM NaHCO₃, 1.2 mM MgSO₄, 2.5 mM K₂HPO₄, 1.3 mM CaCl₂, 5.5 mM glucose, and 100 mM HEPES) supplemented with 0.1% BSA. Cortex was cut into thin slices and pressed through a 100-μm mesh followed by sieving through a graded series of mesh screens (150 to 53 μm). Glomeruli captured in the 53-μm mesh were resuspended in PSS + 0.1% BSA and centrifuged for 5 minutes at 1000 × g. Glomeruli were resuspended in RPMI supplemented with 0.5% BSA, penicillin (1000 U/ml), and streptomycin (0.1 mg/ml). Glomeruli were collected using a microscope and transferred to 12-well plates in a density of 100 glomeruli per well. Glomeruli were incubated for 16 to 18 hours at 37°C and 5% CO₂ before addition of AngII (Sigma). Glomeruli were harvested for RNA isolation after 6 hours of incubation with active substances.

Messenger RNA Analysis

Messenger RNA was isolated from kidney tissue using the RNeasy Protect Mini Kit (Qiagen). Messenger RNA was isolated from nephron segments, vasa recta bundles, and glomeruli using phenol-chloroform extraction. cDNA synthesis, PCR analysis, and quantitative PCR experiments were done as described previously.⁴⁶ For quantitative PCR experiments, a standard curve was constructed by plotting threshold cycle (C_t values) against serial dilutions of purified PCR product. Specificity of the product was confirmed by postrun melting point analysis and by gel electrophoresis. Primer sequences used throughout this study are shown in Supplementary Table S1.

Western Blotting

Tissue samples were homogenized and prepared as described previously in detail.⁴⁶ Samples were run on 4 to 12% Bis-Tris gels (Invitrogen), and protein was transferred to polyvinylidene difluoride membranes (Immobilon-P Transfer membrane; Millipore). Membranes were blocked with 5% nonfat dry milk in TBST (137 mM NaCl, 20 mM Tris [pH 7.6], and 0.05% Tween 20) and incubated overnight at 4°C with primary antibodies diluted in 5% nonfat dry milk in TBST. Primary antibodies were VEGF (sc-152, dilution 1:500; Santa Cruz Biotechnology), PCNA (FL-261, dilution 1:2000; Santa Cruz Biotechnology), α-SMA (ab5694, dilution 1:5000; Abcam), and E-cadherin (#610181, dilution 1:20,000; BD Transduction Laboratories). The antigen-antibody complex was visualized by horseradish peroxidase–conjugated secondary antibodies (P0447 and P0448, diluted 1:2000, Dako) using the enhanced chemiluminescence system (Amersham Pharmacia Biotech).

Immunohistochemistry

Tissue sections were deparaffinized and rehydrated as described previously.⁴⁷ Antigen retrieval was carried out using pepsin digestion (0.4% pepsin diluted in 0.1 M HCl) for 30 minutes at 37°C or boiling for 20 minutes in TEG buffer (10 mM Tris and 0.5 mM EGTA [pH 9.0]). Primary antibodies used were VEGF (sc-152, 1:500; Santa Cruz Biotechnology), CD31 sc-1506, 1:400; (Santa Cruz Biotechnology), α-SMA (ab7817, dilution 1:50; Abcam), E-cadherin (#610181, dilution 1:1000; BD Transduction Laboratories), and CD133 (Miltenyi Biotec). CD133 was visualized with The Catalyzed Signal Amplification II kit (Dako). CD31 was visualized using the Envision⁺ System (Dako). All other antibodies were visualized using horseradish peroxidase–conjugated secondary antibodies (P0447, P0448, and P0449 diluted 1:1000; Dako).

A negative control with omission of the primary antibody was always run in parallel. Whenever the peptide used for immunization was available, negative controls were included using peptide-absorbed primary antibody.

Plasma Hormone Analysis

Plasma renin concentration was analyzed as described previously.⁴⁸

Magnetic Resonance Imaging

MRI was performed using a Philips 1.5 T clinical MRI system (Philips Medical Systems). Animals were placed in the supine position on a 4-cm surface coil for data acquisition. Control high-resolution images were acquired for anatomic localization of kidneys and aorta using a gradient echo sequence with the following parameters: Recovery time/echo time/flip angle 250 ms/4.0 ms/45°, slice thickness 2 mm, field-of-view 13 × 13 cm, and the acquisition and reconstruction matrix 256 × 256. A dose of 0.1 mmol/kg gadodiamide (Omniscan; GE Health Care) was administered as a rapid bolus in <1 second. The gadolinium-enhanced series was acquired in the coronal plane using a spoiled gradient-echo sequence with a linear k-space ordering. A series of 500 images were obtained using the following sequence parameters: Recovery time/echo time/excitation flip angle 10 ms/2 ms /40°, allowing images to be acquired with an interval of 0.85 seconds. The slice thickness was 2 mm, field-of-view was 110 × 70 mm, and the acquisition matrix was 0 filled to 256 × 256 before Fourier transformation. The sequence and inversion time were chosen for optimum enhancement in the rat kidney. Postprocessing was accomplished using the Mistar software (Apollo Imaging Technology). Dynamic data were evaluated on a raw signal intensity basis using the assumption that a change in signal is linearly related to a change in the concentration of gadodiamide. ROIs were manually drawn over left and right kidneys and abdominal aorta on high-resolution anatomic localized images. The kidney ROIs included the parenchyma only. These ROIs were used on the perfusion-weighted data. Because the presence of contrast agent induced a signal increase in the dynamic image, a threshold of S(0) ± 2 × SD of mean was used to determine the bolus arrival time, where S(0) represents the signal intensity baseline before injection. Time signal intensity curves were generated for each ROI. RBF was measured according to Auman et al.⁴⁹

Statistical Analysis

Values are presented as means ± SEM. Comparison of means between two groups was made using a standard unpaired t test. Comparison of means between more than two groups was made using one-way ANOVA with post hoc unpaired t test with Bonferroni correction. P < 0.05 was considered statistically significant. Statistical analysis was done using the GraphPad InStat 3.0 software.

Disclosures

None.