Low nephron endowment increases susceptibility to renal stress and chronic kidney disease

Abstract

Preterm birth results in low nephron endowment and increased risk of acute kidney injury (AKI) and chronic kidney disease (CKD). To understand the pathogenesis of AKI and CKD in preterm humans, we generated potentially novel mouse models with a 30%-70% reduction in nephron number by inhibiting or deleting Ret tyrosine kinase in the developing ureteric bud. These mice developed glomerular and tubular hypertrophy, followed by the transition to CKD, recapitulating the renal pathological changes seen in humans born preterm. We injected neonatal mice with gentamicin, a ubiquitous nephrotoxic exposure in preterm infants, and detected more severe proximal tubular injury in mice with low nephron number compared with controls with normal nephron number. Mice with low nephron number had reduced proliferative repair with more rapid development of CKD. Furthermore, mice had more profound inflammation with highly elevated levels of MCP-1 and CXCL10, produced in part by damaged proximal tubules. Our study directly links low nephron endowment with postnatal renal hypertrophy, which in this model is maladaptive and results in CKD. Underdeveloped kidneys are more susceptible to gentamicin-induced AKI, suggesting that AKI in the setting of low nephron number is more severe and further increases the risk of CKD in this vulnerable population.

Keywords: Chronic kidney disease; Nephrology.

Figure 1. Inhibition of Ret tyrosine kinase resulted in reduced N^glom.

Pregnant Ret^flox-V805A mice were injected daily with vehicle or a small-molecule Ret tyrosine kinase inhibitor, NA-PP1, beginning E16.5 daily for 3 days. Renal structure and N^glom in offspring were analyzed. (A) PAS staining of P1 kidneys exposed to 50 mg/kg of NA-PP1 shows that kidneys are smaller than vehicle-exposed controls and have no tubular dilatation or hydronephrosis; images were obtained with a Zeiss M2 Bio microscope with 10× eyepiece and mechanical stage adjusted to a total of 33× magnification. Trichrome staining of P1 kidneys indicates no evidence of fibrosis (magnification, 25×). (B) N^glom was reduced in mice with prenatal exposure to NA-PP1 (32.25 mg/kg, n = 13; 50 mg/kg, n = 33; or 62.5 mg/kg, n = 28) compared with vehicle-exposed controls (n = 17). *P < 0.0001, 1-way ANOVA followed by Tukey’s test for multiple comparisons; all groups significantly different from vehicle. (C) NA-PP1–exposed mice show decreased UB branching and truncated UB tips. Whole-mount P1 kidneys from vehicle- or NA-PP1–exposed (50 mg/kg) pups were immunolabeled with antibody to calbindin, followed by optical clearing before image acquisition with a laser confocal microscope and 3-D reconstruction (100× magnification). (D) Two-week-old kidneys of mice with prenatal exposure to NA-PP1 (50 mg/kg) contain proximal tubules, thick ascending limbs, distal convoluted tubules, and glomerular and peritubular capillaries. Kidneys were stained with proximal tubule marker LTA (white) and distal tubule marker TSC (red), shown in top panel. Thick ascending limb is labeled with THP (red) and collecting ducts with DBA (green) shown in middle panel. Bottom panel shows capillaries and small veins labeled with endomucin (red). Laminin immunostaining (green) highlights tubular basement membrane (400× magnification). (E) Kidney/body weight ratio in NA-PP1–exposed (50 mg/kg) adult mice is significantly lower than age-matched vehicle-exposed controls. *P < 0.01, Welch’s t test, vehicle n = 7; NA-PP1, n = 23.

Figure 2. Ret deletion results in decreased N^glom in a time-dependent manner.

Pregnant female mice carrying the Hoxb7-rtTA;tet-O-Cre;Ret^flox-V805A transgenes were crossed with male mice of the same genotype and exposed to doxycycline (Dox) in the drinking water beginning E15.5, E16.5, or E17.5 through delivery. Offspring carrying all transgenes are named Ret^{UB del}, while littermates carrying tet-O-Cre; Ret^flox-V805A without Hoxb7rtTA are controls. (A) UB-specific deletion of Ret during kidney development leads to significant reduction of N^glom with greater reduction on earlier exposure (71% at E15.5, n = 10; 60% at E16.5, n = 47; and 36% at E17.5, n = 10; compared with controls, n = 18). *P < 0.0001, 1-way ANOVA followed by Tukey’s test for multiple comparisons; all groups significantly differed from controls. (B) PAS staining of P1 kidneys from Ret^{UB del} pups exposed to Dox E16.5 shows no tubular dilatation or hydronephrosis (imaged at 33×). (C) Ret deletion (starting E16.5) resulted in decreased UB branching and truncated UB tips with areas of absent Six2-expressing cap mesenchyme. Whole-mount P1 kidneys were labeled with antibody to calbindin to identify UB and its derivatives (green) or antibody to Six2 to identify nephron progenitors in the cap mesenchyme (red), followed by optical clearing prior to confocal microscopy and 3-D image reconstruction (100× magnification). (D) PAS staining of adult Ret^{UB del} mouse kidneys (6 weeks old) shows thinner cortex with decreased glomerular number (40× magnification). (E) Adult kidney/body weight ratio in Ret^{UB del} mice (Dox E16.5) is significantly lower than age-matched littermate controls. *P < 0.001, Welch’s t test with n = 16 controls and n = 30 Ret^{UB del} mice.

Figure 3. Ret^{UB del} mice develop glomerular and tubular hypertrophy.

(A) Top panel, PAS-stained kidneys at 6 weeks of age show glomerular hypertrophy (imaged at 400×) in Ret^{UB del} mice. Bottom panel, quantification of glomerular surface area at 2 and 6 weeks of age. While at 2 weeks there was no difference between groups (mixed-effects regression; mice, n = 10; glomeruli, n = 1,741; P = 0.61), by 6 weeks, Ret^{UB del} mice developed glomerular enlargement, with mean glomerular surface area 1,332 μm² larger in Ret^{UB del} compared with controls (mixed-effects regression; mice, n = 10; glomeruli, n = 1,351; P = 0.002). (B) Top panel, PAS-stained kidneys at 6 weeks of age show tubular hypertrophy (imaged at 400×) in Ret^{UB del} mice. Bottom panel, quantification of proximal tubular diameter (measured in the S3 segments) at 2 and 6 weeks of age. While, at 2 weeks, there was no difference between groups (mixed-effects regression; mice, n = 10; tubules, n = 1,000; P = 0.317), by 6 weeks, Ret^{UB del} mice developed tubular enlargement, with mean tubular diameter 7.8 μm larger in Ret^{UB del} compared with controls (mixed-effects regression; mice, n = 10; tubules, n = 1,000; P = 0.007). (C) Correlation of glomerular and tubular size with tubular diameter plotted against glomerular surface area showed that mean tubular diameter and mean glomerular surface area are highly correlated; R = 0.917. (D) Ret^{UB del} mice at 6 weeks of age have robust endolysosomal structures. Lamp1 labels exuberant and abundant late endosomes and lysosomes in Ret^{UB del} mice. Immunolabeling of N⁺/K⁺-ATPase highlights basolateral cell membranes (400× magnification). Quantification of Lamp1 fluorescence integrated density corrected for number of nuclei in proximal tubules localized to the cortex shows a significant increase in Lamp1 expression at 6 weeks of age in Ret ^{UB del} mice. *P = 0.03 using Welch’s t test; n = 3 for each group.

Figure 4. Adult Ret^{UB del} mice develop a CKD phenotype.

(A) Top panel, trichrome staining shows focal subcapsular interstitial fibrosis and inflammation in Ret^{UB del} mouse kidneys at 6 weeks of age (400× magnification left 2 panels, 5.8× magnification right panel). Bottom panel left and middle, collagen I immunostaining shows increased collagen I expression in 6-week-old Ret^{UB del} mice compared with littermate control. Kidney images were contoured, tile scanned, and stitched into single mosaics (200× magnification) for collagen quantification. Bottom right, collagen I and CD45 staining show colocalization of areas of collagen I deposition and inflammatory cell infiltrates (200× magnification; inset 400× magnification) (B) Quantification of collagen I integrated density shows a significant increase in collagen I deposition at 6 weeks of age in Ret^{UB del} mice compared with controls (*P = 0.016, Welch’s t test, n = 7 controls, n = 5 Ret^{UB del} mice). (C) At 6 weeks of age sCr was similar in control and Ret^{UB del} mice (P = 0.13, Welch’s t test, n = 5 controls, n = 6 Ret^{UB del}) with increased creatinine evident by 12 weeks of age (*P = 0.02, Welch’s t test, n = 6 control, n = 5 Ret^{UB del}). By 12 weeks of age, there was increased urinary albumin excretion (*P = 0.02, Welch’s t test, n = 6 control, n = 9 Ret^{UB del}). (D) PAS staining of kidneys at 9 months of age reveals perihilar hyalinosis (arrow) and segmental accumulation of endocapillary foam cells (arrowheads) forming a cellular lesion of segmental sclerosis (400× magnification).

Figure 5. Neonatal Ret^{UB del} mice have more severe injury following gentamicin-induced AKI.

Newborn mice (control and Ret^{UB del}) were injected with saline or gentamicin once a day for 7 days from P3 to P9. Kidneys were collected for analysis 1 day after completing treatment (P10). (A) Top panel, PAS staining shows more tubular vacuolization (arrowhead) and loss of brush border in proximal tubules of Ret^{UB del} mouse kidneys (arrow). Bottom panel, more abundantly enlarged Lamp1-expressing lysosomes (red) in proximal tubules labeled with LTA (white) in Ret^{UB del} mice. Images were obtained at 630× magnification. Inset indicates swollen lysosomes (inset 1,260× magnification). (B) Cellular injury shown by electron microscopy. Top panels, lysosomes (arrowheads) containing myelin bodies (arrows) and cellular debris in proximal tubules of control and Ret^{UB del} mice. Middle left, swollen endosomes and lysosomes containing cellular debris in proximal tubules of control mice. Middle right, interstitial edema (asterisk) and a marginating polymorphonuclear leukocyte (arrow) adjacent to a severely injured proximal tubule in Ret^{UB del} kidney. Bottom, electron-dense cytoplasmic vacuoles, swollen endosomes, and lysosomes (arrowheads) in control and Ret^{UB del} mouse kidneys. Note the degenerating proximal tubular cell spilling cytoplasmic contents into the lumen in Ret^{UB del} mouse (arrow). (C) Ret^{UB del} mice have higher expression of kidney injury molecule 1 (Kim1) and more CD45⁺ leukocyte infiltration. The entire kidneys were scanned in tiles and stitched together (200× magnification, inset 400×). (D) Quantification of Kim1 expression and CD45⁺ cell infiltration by fluorescence integrated density using FIJI software. There is a significant increase in Kim1 expression (*P = 0.018, n = 5 per group) and CD45⁺ cell infiltration (*P = 0.019, n = 5 per group) in gentamicin-exposed Ret^{UB del} mice versus gentamicin-exposed controls using 1-way ANOVA followed by Tukey’s test for multiple comparisons. Gent, gentamicin.

Figure 6. Neonatal Ret^{UB del} mice have incomplete renal repair after gentamicin-induced AKI.

Newborn mice (control and Ret^{UB del}) were injected with saline or gentamicin once a day for 7 days from P3 to P9. Kidneys were analyzed 1 day (P10) or 1 month after completing treatment. (A) Expression of phospho-histone H3 (pH3, red) in the proximal tubules labeled with LTA (green) in mice 1 day after completion of injections (200× magnification) (B) While saline-injected control and Ret^{UB del} mice show no difference in the baseline expression of pH3⁺ proliferating cells in the renal cortex and outer medulla, gentamicin injury induces a significant increase in the expression of pH3 in control mice (1-way ANOVA followed by Tukey’s test for multiple comparisons, *P = 0.015, n = 4–5 per group) but not in Ret^{UB del} kidneys (P = 0.28, n = 5 per group). Comparison of gentamicin-injected control and Ret^{UB del} kidneys also shows a significantly lower number of pH3⁺ cells in the cortex and outer medulla of Ret^{UB del} mice compared with controls (**P = 0.0377, n = 5 per group). (C) At 1 month following completion of gentamicin injections, control mice have near complete tubular repair, whereas Ret^{UB del} kidneys have areas of tubular atrophy (arrow), interstitial fibrosis, chronic inflammation, and residual tubular vacuolization (arrowhead). (D) Control mice had no significant increase in injury score 1 month after gentamicin injection (1-way ANOVA followed by Tukey’s test for multiple comparisons, P = 0.77, n = 4–5 per group), whereas Ret^{UB del} mice had significantly elevated injury scores following gentamicin-induced AKI (*P < 0.01, n = 5 per group, 1-way ANOVA followed by Tukey’s test for multiple comparisons). The difference between injury scores in control and Ret^{UB del} mice 1 month after gentamicin-induced AKI were also significant (**P < 0.01, n = 5 per group, 1-way ANOVA followed by Tukey’s test for multiple comparisons). Gent, gentamicin.

Figure 7. Ret^{UB del} mice exhibit a unique inflammatory response to gentamicin-induced AKI.

Newborn mice (control and Ret^{UB del}) were injected with saline or gentamicin at 100 mg/kg once a day for 7 days at P3–P9. Kidneys were analyzed 1 day after the completion of injections (P10). (A) Representative images of cytokine arrays of kidney homogenates. Cytokines are identified in the chart below the array membrane. (B) Of 40 total cytokines tested, 11 cytokines that had detectable expression following saline or gentamicin injection are shown, and levels of Timp-1, MCP-1, CXCL10, and IL-1ra are significantly different between groups tested by 1-way ANOVA (*P < 0.01, n = 4 per group). (C) One-way ANOVA followed by Tukey’s test for multiple comparisons indicates significant differences in levels of Timp-1 and MCP-1 in saline versus gentamicin injected controls as well as in saline and gentamicin exposed Ret^{UB del} mice (*P < 0.05, **P < 0.0001, n = 4 per group). Note higher expression of Timp-1 and MCP-1 in gentamicin-injected Ret^{UB del} compared with gentamicin-injected control mice (*** P ≤ 0.003). CXCL10 was uniquely elevated in Ret^{UB del} mice after gentamicin injection (^†P < 0.0001). Gent, gentamicin.

Figure 8. Expression of CXCL10 and MCP-1 in gentamicin-injured proximal tubular cells.

(A) qPCR analysis shows significantly different mRNA expression levels for Timp-1, MCP-1, and CXCL10 (1-way ANOVA, *P < 0.05), with higher MCP-1 and CXCL10 mRNA levels in Ret^{UB del} mice compared with controls after gentamicin injection (1-way ANOVA followed by Tukey’s test for multiple comparisons, **P < 0.05, n = 4 per group). (B) RNAscope using probes to CXCL10 (white) and MCP-1 (red) was coupled with immunostaining of CD13 (green) to identify proximal tubules in kidneys 1 day after completing saline or gentamicin injections. In the saline-exposed groups, there was minimal expression of CXCL10 and MCP-1 (top panel). Gentamicin injection resulted in the increase of mRNA of MCP-1 (arrow) in the damaged proximal tubules of both control and Ret^{UB del} kidneys and in the increased expression of CXCL10 in Ret^{UB del} tubules (arrowhead) (630× magnification). Gent, gentamicin.