Fight-or-flight: murine unilateral ureteral obstruction causes extensive proximal tubular degeneration, collecting duct dilatation, and minimal fibrosis

Abstract

Unilateral ureteral obstruction (UUO) is the most widely used animal model of progressive renal disease. Although renal interstitial fibrosis is commonly used as an end point, recent studies reveal that obstructive injury to the glomerulotubular junction leads to the formation of atubular glomeruli. To quantitate the effects of UUO on the remainder of the nephron, renal tubular and interstitial responses were characterized in mice 7 and 14 days after UUO or sham operation under anesthesia. Fractional proximal tubular mass, cell proliferation, and cell death were measured by morphometry. Superoxide formation was identified by nitro blue tetrazolium, and oxidant injury was localized by 4-hydroxynonenol and 8-hydroxydeoxyguanosine. Fractional areas of renal vasculature, interstitial collagen, α-smooth muscle actin, and fibronectin were also measured. After 14 days of UUO, the obstructed kidney loses 19% of parenchymal mass, with a 65% reduction in proximal tubular mass. Superoxide formation is localized to proximal tubules, which undergo oxidant injury, apoptosis, necrosis, and autophagy, with widespread mitochondrial loss, resulting in tubular collapse. In contrast, mitosis and apoptosis increase in dilated collecting ducts, which remain patent through epithelial cell remodeling. Relative vascular volume fraction does not change, and interstitial matrix components do not exceed 15% of total volume fraction of the obstructed kidney. These unique proximal and distal nephron cellular responses reflect differential “fight-or-flight” responses to obstructive injury and provide earlier indexes of renal injury than do interstitial compartment responses. Therapies to prevent or retard progression of renal disease should include targeting proximal tubule injury as well as interstitial fibrosis.

chronic renal disease, regardless of etiology, leads ultimately to nephron loss and renal fibrosis. The model of murine unilateral ureteral obstruction (UUO) has been widely employed to study the underlying mechanisms of progressive chronic kidney disease, most focusing on interstitial fibrosis as the final common pathway (43). Extensive formation of atubular glomeruli in this model was revealed by loss of Lotus tetragonolobus lectin binding from cells forming Bowman's capsule as well as from cells of the glomerulotubular junction (14). Through a process of epithelial cell phenotypic transition and remodeling, the urinary pole of Bowman's capsule is sealed off from the atrophic proximal tubular segment (14). To determine the temporal evolution of the lesions in the entire nephron, the present study was undertaken to examine the segmental renal tubular responses following 7 and 14 days of UUO. The results reveal segment-specific responses to UUO that contribute to an enhanced understanding of marked renal tubular alterations that overshadow interstitial responses.

MATERIALS AND METHODS

Experimental animals and surgical procedures.

Male mice of the C57BL/6 strain were subjected to complete UUO or sham operation at 6 wk of age. All surgery was performed using sterile technique in accordance with an animal protocol approved by the University of Virginia Animal Care and Use Committee. All animals were anesthetized with isoflurane plus oxygen, and the left ureter was exposed through a flank incision. In animals undergoing UUO, the ureter was ligated with 8-0 nylon; in sham-operated mice, the ureter was left undisturbed. A total of 32 animals were used for the study.

Tissue collection and processing.

Animals were examined 7 days (n = 10 UUO + 5 sham) or 14 days (n = 12 UUO + 5 sham) after surgery. All animals were injected with pentobarbital sodium-phentoin sodium (Euthasol) solution (Virbac, Fort Worth, TX), and kidneys and ureters were exposed through an abdominal incision. Renal pelvic diameter and ureteral diameter proximal to the obstruction were measured in situ; then kidneys were removed and fixed by immersion in 10% phosphate-buffered formalin. In some cases, kidneys were perfused with 1.5% glutaraldehyde in a solution of 3% dextrose, 3% dextran (43,500 avg mol wt), and 50 mM CaCl₂. Formalin-fixed kidneys were washed in phosphate buffer, dehydrated through a graded series of ethanols and xylene, embedded in paraffin, and sagittally sectioned at 4 μm. Glutaraldehyde-perfused kidneys were cut into 50-μm coronal sections and processed for plastic embedment, as described previously (13). Plastic “semithin” (0.25-μm) sections of areas of interest were cut and stained with alkaline toluidine blue.

Staining.

Fragmented DNA was detected using Apoptag (Chemicon, Temecula, CA) with diaminobenzidine (DAB) development (Biogenex, San Ramon, CA) and methylene blue counterstaining. This method is based on the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labeling (TUNEL) reaction (15). For immunohistochemistry, sections were pretreated to quench endogenous peroxidase (H₂O₂ in methanol) and to neutralize endogenous biotin [avidin-biotin (ABC) blocking kit, Vector Laboratories, Burlingame, CA]. Localized production of superoxide was demonstrated in kidneys of 14-day UUO mice perfused with nitro blue tetrazolium (NBT), as previously described (14). Oxidative damage was localized by immunohistochemical staining with 4-hydroxynonenal antibody (ab485606, Abcam, Cambridge, MA) at 1:2,000 dilution or 8- hydroxy-2′-deoxyguanosine antibody (ab48508, Abcam) at 1:100 dilution. Mitotic cells were detected with phosphorylated (Ser¹⁰) histone H3 (Cell Signaling Technology, Beverly, MA) at a 1:200 primary antibody dilution. Interstitial cell α-smooth muscle actin (α-SMA) content was localized by immunohistochemical staining using antibody A-2547 (Sigma-Aldrich, St. Louis, MO) at a dilution of 1:800 and fibronectin antibody (ab6328, Abcam) at a dilution of 1:200. L. tetragonolobus lectin (Vector Laboratories) binds to proximal tubule epithelial cells in mouse and human kidney, an affinity that develops in utero (20, 21, 34). Paraffin sections of formalin-fixed kidney were treated by this staining procedure, which incorporated proteinase K enzymatic digestion before exposure to biotinylated Lotus lectin (1:50 dilution) and development by the ABC-DAB regimen.

Picrosirius red staining was used to identify collagen deposition (Polysciences, Warrington, PA). Morphometric analysis of histological sections stained with picrosirius red reveals a very high correlation with tissue hydroxyproline content (R = 0.89, P < 0.0001) and binds to types I, III, and IV collagen (22). Comparing serial sections of kidney tissue from mice subjected to UUO, we found that picrosirius red staining is superior to trichrome staining for quantitation of collagen by digital morphometry (5).

Quantitation.

Seven animals were used for documentation and stereology of specifically stained sections obtained 7 and 14 days after surgery. Parenchymal thickness was estimated as previously described (37); briefly, three radial measurements (both poles and opposite the hilum) were made on median sagittal sections (Fig. 1). TUNEL staining reveals damaged DNA in cells, which can result from apoptosis, necrosis, or autolysis (18). Since DNA fragmentation is not absolutely specific for apoptosis or necrosis (35), TUNEL staining was scored for net apoptosis counts by manual counting of clearly identifiable apoptotic nuclei (condensed chromatin and nuclear blebbing) in 10 fields at ×400 magnification in a median sagittal section of each kidney, with care taken to ensure that these fields sampled all regions of the kidney (Fig. 1B). Quantitation of mitosis was carried out in a similar fashion by manual counting of clearly identified phosphorylated histone-positive mitotic nuclei.

Fig. 1.

A: representative sagittal sections of kidneys from mice subjected to unilateral ureteral obstruction (UUO) or sham operation. Progressive loss of kidney mass occurs through parenchymal thinning of obstructed kidney. Solid lines superimposed on 14-day UUO kidney profile demonstrate procedure whereby average parenchymal thicknesses were established, with dotted line demarcating papilla. B: positioning of 10 microscopic fields for stereological assessment of parameters including mitosis, apoptosis, α-smooth muscle actin (α-SMA), and fibronectin. (The same sampling approach is utilized for contralateral and sham-operated kidneys.) C: sampling pattern utilized for measurement of proximal tubule contribution (also see Fig. 3).

Image analysis (ImagePro Plus 5.1, Media Cybernetics, Silver Spring, MD) was used to quantitate α-SMA, collagen, and fibronectin distribution, with cell staining with DAB reaction product expressed as a percent area value [volume fraction (V_v); Fig. 1B]. In the case of collagen, 20 microscopic fields of picrosirius red-stained sections were measured for each adult mouse kidney, alternating between cortical and medullary zones, thus providing a sampling procedure for the entire wall.

Vascular V_v [V_V(Vasc)] in platelet-endothelial cell adhesion molecule 1 (PECAM-1)-stained sections was measured by point counting, which allowed accurate location and identification of all vascular-specific components, including arteriolar walls and vessel lumina, that are not PECAM-positive. An eyepiece-mounted graticule was used to count 10 fields each of cortex and medulla (with glomerular capillary count as a subset of total cortical measurements). V_V(Vasc) for each parameter was calculated as the fraction of points falling on a blood vessel wall or lumen.

Renal cortical volume fraction of proximal tubules was measured using a stereological approach similar to that used for other stains: 10 fields were photographed at a total ×400 magnification but were restricted to the subcapsular cortex (Fig. 1C; see Fig. 3B).

Fig. 3.

Microscopic and morphometric evaluations of contributions of proximal tubule complement in sham-operated vs. obstructed and contralateral kidneys at 7 and 14 days after operation. A: Lotus lectin staining clearly delineates proximal tubules. B: 2 images of the same cortical field as they appear with image analysis software before (top) and after (bottom) digital contrasting of proximal tubules (shown in red) to measure their volume fraction within the field. C: proximal tubule contribution. Solid bars, obstructed kidney; open bars, sham and contralateral kidney. Values are means ± SE. *P < 0.05 vs. contralateral kidney. Considerable loss of proximal tubular mass has occurred by 7 days; by 14 days, many of the proximal tubular remnants have collapsed (also see Fig. 2). Contralateral kidneys' proximal tubules retain normal-appearing morphology at both stages, and there is no change in proximal tubular fraction in contralateral compared with sham-operated kidneys. Scale bar, 100 μm.

Statistical analysis.

SigmaStat version 3.0 (Aspire Software International, Ashburn, VA) was employed for statistical analysis. Parameters for contralateral vs. obstructed kidneys for each experimental category were compared using Student's t-test for paired observations. Comparisons between UUO and sham-operated groups were made using Kruskal-Wallis one-way ANOVA on ranks with pair-wise multiple comparisons by Dunn's method. Comparisons between 7- and 14-day groups were made using two-way ANOVA with pair-wise multiple comparisons by the Holm-Sidak method. Statistical significance was defined as P < 0.05.

RESULTS

Chronic UUO causes renal parenchymal loss.

As shown in Table 1, there was no significant change of body weight between 7 and 14 days following UUO or sham operation. Ureteral and pelvic diameter increased fourfold after 7 days of ipsilateral UUO, but there was no additional increase between 7 and 14 days of UUO. However, during this interval, weight of the obstructed kidney decreased by 26% and parenchymal thickness decreased by 33%, while weight of the contralateral kidney did not change (Fig. 1).

Table 1.

Characteristics of mice

	7 Days	14 Days
Body wt, g
UUO	24 ± 1.5	23 ± 0.4
Sham	22 ± 0.5	24 ± 0.6
Ureteral diameter, mm
UUO	1.83 ± 0.13‡	2.04 ± 0.26‡
Sham	0.45 ± 0.03	0.40 ± 0.00
Pelvic diameter, mm
UUO	5.00 ± 0.30‡	5.63 ± 0.30‡
Sham	1.15 ± 0.10	1.24 ± 0.11
Kidney wt, mg
UUO	168 ± 13	125 ± 9*†‡
Contralateral	178 ± 12	171 ± 5
Sham	143 ± 15	153 ± 7
Renal parenchymal thickness, mm
UUO	3.0 ± 0.3	2.0 ± 0.2*†
Contralateral	3.9 ± 0.4	4.6 ± 0.2

Values are means ± SE of 7 unilateral ureteral-obstructed (UUO) and 4-5 sham-operated mice.

^*P < 0.05 vs. contralateral kidney.

^†P < 0.05 vs. 7 days.

^‡P < 0.05 vs. sham.

Proximal tubular degeneration.

In sham-operated or contralateral kidneys (Fig. 2, A–C), the tall epithelial cells of proximal tubules and Bowman's capsule exhibited a consistent columnar morphology with abundant mitochondria (Fig. 2A). Perfusion with NBT resulted in deposition of blue diformazan crystals at the bases of these cells (Fig. 2B). The cells shared an affinity as well for L. tetragonolobus lectin (Fig. 2C). As previously shown (14), the glomerulotubular junctions in obstructed mouse kidneys underwent rapid deterioration. Downstream from the glomeruli, the proximal tubules fragmented into isolated collapsed segments containing numerous autophagic bodies (Fig. 2D). TUNEL positivity denotes apoptosis and necrosis (Fig. 2E) in such proximal tubular segments, which, despite their collapse and atrophy, retained their lectin affinity (Fig. 2F). By contrast, mitotic profiles were rare in proximal tubules. After NBT perfusion, diformazan crystals were concentrated within the epithelial cells of collapsed proximal tubules (Fig. 2, G and H), and markers of oxidative damage stained the tubule fragments positively, with cytoplasmic staining for 4-hydroxynonenal (Fig. 2G) and nuclear positivity for 8-hydroxy-2′-deoxyguanosine (Fig. 2H). The result of these multiple events was a 65% decrease in the cortical proximal tubule contribution (Fig. 3), which accounts for the majority of renal parenchymal loss following ureteral ligation.

Fig. 2.

Proximal tubular damage resulting from ureteral obstruction. A: sham-operated kidney. Semithin plastic section of a cortical field consists largely of proximal tubules, which are composed of columnar epithelial cells containing numerous rod-shaped mitochondria. Structurally similar cells make up the capsule of an adjacent glomerulus (G). B: contralateral kidney from 14-day UUO mouse perfused with nitro blue tetrazolium (NBT) to demonstrate sites of superoxide production in the form of blue diformazan crystals (arrowheads) concentrated at basal surfaces of proximal tubules (PT) and glomerular capsules (G). Note lack of labeling of cortical collecting duct (CCD). C: NBT-perfused, sham-operated kidney. Field is similar to B but stained with Lotus tetragonolobus lectin to confirm identity of diformazan-decorated profiles as belonging to proximal tubules and similar epithelial cells of the capsule of the glomerulus (G). D–H: obstructed kidney from 14-day UUO mice. D: semithin plastic section through the remaining cortex showing fragmented segments of collapsed proximal tubules. Arrows, apoptotic figures; ∗, autophagic bodies. E and F: serial consecutive sections subjected to terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labeling (TUNEL) procedure (E) and Lotus lectin staining (F). Two proximal tubule profiles, one collapsed and the other relatively normal, show strong lectin positivity; only degenerating tubule shows evidence of apoptosis (arrows). G: 4-hydroxynonenal immunostaining of a collapsed proximal tubule from a kidney perfused with NBT. Nuclei (Nu) are unstained. Note dense cytoplasmic immunostaining (denoting oxidant injury), along with blue diformazan crystals, which result from interaction of NBT with sites of superoxide activity. Inset: detail of diaminobenzidine chromogen (brown) and diformazan (blue). Micrograph was deliberately intensified by selection of the blue channel to distinguish between the 2 forms of precipitate. H: similar to G, but with 8-hydroxy-2′-deoxyguanosine immunostaining, which demonstrates oxidative damage to nuclei. Note opacified nuclei and abundant diformazan crystals, as well as a darkly stained apoptotic figure (arrow), within epithelial cells of fragmented proximal tubular segment. Scale bars, 50 μm (bar in C applies to B and C, bar in F applies to E and F, and bar in H applies to G and H). Magnification of inset in G is doubled.

Dilatation and remodeling of collecting ducts.

Dilatation of collecting ducts was evident following 7 days of UUO in the obstructed kidney (Fig. 4A), with a decrease in dilatation at 14 days (Fig. 4B). After 7 days of obstruction, collecting ducts (Fig. 4, C and E) were composed of flattened epithelial cells, in which mitotic and apoptotic nuclei were plentiful (Fig. 4, G and I). By 14 days, along with the lessening of collecting duct dilatation (Fig. 4B), epithelial cells became crowded and cuboidal (Fig. 4, D and F). By 14 days, mitosis was rare (Fig. 4H), but apoptosis was still prominent (Fig. 4J).

Fig. 4.

Temporal effects of UUO on collecting ducts in obstructed kidneys. A, C, E, G, and I: 7-day kidney. B, D, F, H, and J: 14-day kidney. A and B: survey views of median sagittal sections stained with picrosirius red. Dilated collecting ducts are evident after 7 days of obstruction, but their incidence decreases by 14 days as parenchyma becomes thinner. C–F: “semithin” (0.25-μm) plastic sections of glutaraldehyde-perfused obstructed kidneys. In C and D, red-rimmed dots are centered over epithelial cell nuclei (higher-magnification details of tubule walls are shown in E and F, respectively), illustrating attenuated cell profiles with widely separated nuclei at 7 days of UUO (C and E), in contrast to reduction in tubule dilatation and crowding of uniformly cuboidal epithelial cells after 14 days of UUO (D and F). G and H: phosphorylated histone-stained sections showing extensive mitotic activity in collecting ducts after 7 days (G) but a sharp decline by 14 days (H). I and J: TUNEL staining for apoptosis, which is found routinely in collecting ducts at 7 and 14 days of obstruction. Scale bars: 100 μm (B, applies to A and B); 50 μm (D, applies to C and D); 10 μm (F, applies to E and F); 50 μm (J, applies to G–J). K and L: morphometric measurements of mitosis and apoptosis, respectively. Solid bars, obstructed kidney; open bars, sham and contralateral kidney. Values are means ± SE. *P < 0.05 vs. contralateral kidney. *P < 0.05 vs. obstructed kidney. †P < 0.05 vs. 7 days.

The incidence of mitosis decreased significantly between 7 and 14 days (Fig. 4K), with the majority of cell division in the collecting ducts. While this contrasts with the progressive increase in renal apoptotic activity from 7 to 14 days (Fig. 4L), the relative proportion of apoptotic cells was greater than that of mitotic cells by severalfold at 14 days (Fig. 4, K and L). Thus, as a consequence of chronic UUO, in contrast to the collapse and destruction of proximal tubules, collecting ducts underwent dilatation, with marked remodeling and preservation of tubular integrity.

Effects of UUO on the interstitium.

In contralateral kidneys, collagen and α-SMA were largely restricted to the walls of arterioles. In the obstructed kidneys, accumulation of collagen was scattered (Fig. 5, A and D), whereas α-SMA staining was more evenly distributed throughout the interstitium (Fig. 5, B and E). The interstitial matrix protein fibronectin was sparsely distributed among the tubules in the contralateral kidney but was abundant in the obstructed kidney's interstitial spaces (Fig. 5, C and F). Between 7 and 14 days of UUO, interstitial collagen and fibronectin increased, while interstitial α-SMA did not change (Fig. 5, G–I).

Fig. 5.

Effects of UUO on interstitium of obstructed kidney. A–F: distribution of collagen (A and D), α-SMA (B and E), and fibronectin (C and F) staining in serial consecutive sections of kidneys following 14 days of UUO. In obstructed kidney of adult mouse, aside from the capsule, only scattered concentrations of collagen fibrils, identifiable by picrosirius red staining, are seen (A, examples at arrows in D), whereas α-SMA staining (B and E, brown coloration) is more generally distributed and cell-based, with fibronectin appearing throughout the interstitium (C and F). Scale bar: 100 μm (C, applies to A–C; F, applies to D–F). G–I: renal interstitial responses to UUO. G: collagen (picrosirius red staining). H: α-SMA. I: fibronectin. Solid bars, obstructed kidney; open bars, contralateral kidney. Values are means ± SE. *P < 0.05 vs. contralateral kidney. †P < 0.05 vs. 7 days.

Effects of UUO on vasculature.

Endothelial cell PECAM staining was strong in the obstructed kidney (Fig. 6, A, B, and F) but weaker in the contralateral kidney (Fig. 6, C, D, G, and H). At 7 and 14 days, V_v(Vasc) was higher in the cortex in the obstructed than contralateral kidney; V_v(Vasc) was higher in the medulla in the obstructed kidney only at 7 days (Fig. 6, E and I).

Fig. 6.

Renal microvascular distribution in adult kidneys after 7 days (A–E) and 14 days (F–I) of UUO. After 14 days, platelet-endothelial cell adhesion molecule (PECAM) staining is more intense in obstructed (F) than contralateral (G and H) kidney. Stereological evaluation (E and I) shows vascular volume fraction [V_v(Vasc)] values that are greater in cortex and medulla of obstructed (solid bars) than contralateral (open bars) kidney after 7 days and in cortex after 14 days (I). Scale bars, 100 μm (D, applies to A–D; H, applies to F–H). *P < 0.05 vs. contralateral kidney. †P < 0.05 vs. 7 days.

DISCUSSION

Over the past 30 years, surgical UUO in a variety of experimental animals has become established as a reproducible animal model of progressive chronic kidney disease, with renal interstitial collagen accumulation serving as a primary outcome (5). The majority of reports based on this model utilize mice, since genetically engineered mutants are most readily available in this species (1). Because of the numerous inferences being drawn from these studies regarding the pathogenesis of progressive chronic kidney disease, a better understanding of the renal cellular response to UUO in mice is needed. Findings in the present study complement and extend the recent discovery in the obstructed kidney of widespread phenotypic transition of cells of the glomerulotubular junction resulting in the formation of atubular glomeruli (14).

Chronic UUO induces renal parenchymal atrophy through proximal tubule oxidant injury, cell death, and tubular collapse.

Since it has a rich supply of mitochondria, the proximal tubule is particularly susceptible to hypoxic and oxidant injury resulting from vasoconstriction and ischemia of the obstructed kidney (9, 39). As shown in the present study, the corollary to the formation of atubular glomeruli is a major reduction in proximal tubular mass (Figs. 3 and 7). This is the result of a combination of necrosis, apoptosis, and autophagy, leading to the formation of tubule fragments. This is associated with nuclear and cytoplasmic tubular oxidative damage. There is little mitotic compensation by the proximal tubule, the epithelial cells of which become flattened, losing their apical microvilli and most mitochondria. This is accompanied by a shift in the localization of superoxide primarily from the peritubular surfaces of healthy epithelial cells to the collapsed lumina of degenerating tubules (Fig. 2).

Fig. 7.

Effects of 14 days of UUO on fractional distribution of renal parenchymal compartments. Parenchymal mass is estimated by kidney weight: note 18% loss of parenchyma by obstructed compared with sham-operated kidney. Fractional contributions to parenchyma of obstructed kidney have been reduced in proportion to 18% decrease in kidney weight compared with sham-operated kidney. The most dramatic effect of UUO is reduction of proximal tubular mass, which decreases from 65% to 23% after 14 days of obstruction. Fractional contribution of collagen increases from <1% to 3% in obstructed kidney, while fractional contribution of fibronectin increases from 2% to 11% and that of α-smooth muscle actin increases from <1% to 3%. Fractional vascular contribution changes little, from 8% to 9%, and remaining parenchyma (including dilated distal nephrons) increases from 24% to 32%.

Differential, segment-specific alterations in nephrons that result from ureteral ligation have been described in the neonatal mouse kidney (3). Proximal tubules furthermore are traditionally divided into S1, S2, and S3 segments, on the basis of their position, structure, and cytochemistry. It now appears that a further subdivision can be made within the initial segment of the murine proximal tubule, namely, for the tall epithelial cells that form the urinary pole portion of the glomerular capsule and the immediately contiguous glomerulotubular junction. These cells have been found to be identical in ultrastructure to the epithelial cells located downstream from the glomerulus (10) and share histochemical and immunohistochemical properties (e.g., periodic acid-Schiff-positive brush border and apical Lotus positivity) with them. When challenged by ureteral obstruction, however, the epithelial cells of Bowman's capsule exhibit only minimal evidence of cell death, instead undergoing substantial remodeling from columnar to squamous configuration, sealing off the urinary pole and resulting in the formation of atubular glomeruli (14). In contrast, the remaining proximal tubule portions exhibit multiple forms of cell death (apoptosis, necrosis, and autophagy) while initially retaining a degree of cell polarity, as shown by retention of lectin affinity (Fig. 2, E and F), even though they lose continuity through their fragmentation into isolated segments (Fig. 2D).

The functional implications of tubular loss resulting from clinical chronic kidney disease were described 45 years ago in a study demonstrating significant correlations between tubular atrophy and glomerular filtration rate or renal concentrating ability (32). In a recent review of mouse models of chronic kidney disease, Eddy et al. (12) highlight the statement that “renal parenchymal loss, not fibrosis severity per se, is the critical outcome.” Although previous studies used measurements of E-cadherin or tubular diameter as indexes of tubular mass (12), our approach in the present study allowed tracking the quantitative response of individual tubular segments in addition to the response of the vascular and interstitial compartments. It is becoming increasingly apparent that there is a continuum between acute kidney injury and progression in chronic kidney disease (41). The model of UUO begins as an acute mechanical insult, leading to rapid cellular responses that affect tubular segments differentially. As in models of ischemic acute kidney injury (19, 31), the mitochondria-rich proximal tubules appear to be the most susceptible to obstructive injury, as demonstrated by localized oxidant damage and rapid cell death.

Chronic UUO causes collecting duct dilatation, apoptosis, and remodeling with preservation of distal tubular integrity.

Dilated collecting ducts contained widespread mitotic and apoptotic epithelial cells following 7 days of UUO, with mitoses decreasing by 14 days. A similar sequence has been reported for the rat kidney following UUO: mitosis increases within the 1st wk and then declines, while apoptosis continues to increase through 3 wk (38). As a consequence of UUO, the entire nephron is subjected to increased hydrostatic pressure gradients (4). Relative dilatation of nephron segments initially depends on regional tubular compliance: distal tubules undergo significant dilatation, while the less compliant proximal segments show less distension (3, 8). There is a close correlation between distal nephron luminal area and tubular apoptosis in neonatal and adult mice subjected to UUO (3, 24). Mechanical stretching of cultured monolayers of mouse tubular cells stimulates apoptosis in proportion to the degree of axial strain (3). As shown in the present study, reduction of collecting duct dilatation at 14 days was accompanied by reversion of the epithelial cells to a cuboidal configuration, resulting in remodeling of this nephron segment (Fig. 4). Thus, collecting duct dilatation reflects not only the greater compliance of this tubular segment, but also a transient epithelial cell phenotypic transition and proliferation that favors tubular survival over degeneration. Injured distal tubular cells are less susceptible to oxidant injury than are proximal tubules (6). Generation by distal nephron cells of endogenous survival factors, such as superoxide dismutase and Bcl-2 (an antiapoptotic protein), is likely to contribute to this response (16, 25).

Expression of interstitial compartment α-SMA and collagen by the obstructed kidney fails to parallel obstructive cellular injury.

α-SMA in the contralateral kidney was mainly located in arteriolar smooth muscle cells, with extravascular α-SMA-containing cells (“myofibroblasts”) appearing only in the interstitium of the obstructed kidney, often in regions lacking collagen accumulations. There is thus a lack of correlation between α-SMA and collagen deposition, which suggests that α-SMA ought not serve as a surrogate for interstitial fibrosis in models of UUO. Of the three proteins measured as markers of interstitial expansion (collagen, α-SMA, and fibronectin), fibronectin is a definitive interstitial matrix protein, and its substantial increase with UUO indicates that it may be a more reliable index for the evaluation of renal fibrosis development.

Obstructive nephropathy is typically described as being characterized by the development of renal interstitial fibrosis. However, in the present study, renal α-SMA and collagen accumulation in the obstructed kidney did not exceed 5% of the fractional renal parenchymal area (Fig. 5). Extracellular fibronectin, which is necessary for fibroblasts to form a collagen matrix (40) and can also be synthesized by injured blood vessels (7), doubled from 7% after 7 days to 13% after 14 days of UUO (Fig. 5I). However, there was no further decrement in fractional proximal tubular mass between 7 and 14 days of UUO (Fig. 3). Although deposition of interstitial collagen has been viewed to be injurious to the kidney (11), an alternative school of thought contends that fibrosis is, instead, the by-product of a reparative process (23). Thus, decreasing renal function may be more closely linked to progressive proximal tubular cell loss than toxic effects of interstitial collagen deposition (41). The present study underscores a notable discordance: although there is a significant increase in interstitial collagen and fibronectin in the obstructed kidney, there is a far greater renal parenchymal loss (Fig. 7). The rapidity of proximal tubular cell death in the obstructed kidney, uncompensated by cell proliferation or regeneration, thus appears to be the primary determinant of renal parenchymal loss. Additionally, a steadily mounting body of work has failed to support evidence of epithelial-to-mesenchymal transition as a source of activated interstitial fibroblasts in models of renal disease, including UUO (27). These investigators conclude, instead, that interstitial cell accumulation is best explained by the proliferation of the existing population of resident interstitial fibroblasts (26, 30).

Response to UUO by the renal vasculature.

The measured V_V(Vasc) showed fractional increases in the obstructed kidneys relative to the contralateral kidneys (Fig. 6). Zhang et al. (44) reported superior values of V_V(Vasc) in urokinase receptor (uPAR) knockout (uPAR^−/−) relative to equivalent uPAR^+/+ animals under sham-operated conditions as well as after 3 and 14 days of UUO. Rouschop et al. (33) as well demonstrated a higher V_V(Vasc) in the cortex of CD44^−/− mice than their CD44^+/+ counterparts after 7 and 14 days of UUO. Mechanisms involving heightened angiogenesis or prevention of apoptosis in the knockout mice are proposed to account for these measured differences in V_V(Vasc) (33, 44). However, the present study suggests an alternate explanation: increases in relative vascular contribution to the obstructed kidney could be attributed, in large part, to parenchymal loss, which could obscure microvascular loss. Nonetheless, vascular profiles become tortuous (unpublished observations), and the intensity of their PECAM staining is greater in the obstructed kidney (Fig. 6). Basement membrane thickening of endothelium has been noted in the glomeruli of mouse kidneys after protracted periods (3–17 wk) of obstruction, reflecting microvascular injury (2). Renal blood flow is markedly reduced by obstruction (36), which likely contributes to changes in the vasculature of the obstructed kidney. Recent reports of phenotypic transition of endothelial cells and pericytes in the obstructed kidney indicate their potential contribution to the progression of interstitial injury (29, 42). Further investigation of the vasculature in the mouse UUO model is called for, with emphasis on establishing not only the quantity, but also the morphology and functionality, of the blood vessels in the obstructed kidney.

Conclusions and speculation.

In response to UUO, proximal tubules undergo progressive atrophy and collapse as a result of necrosis and apoptosis. The induction of autophagy fails to prevent progressive destruction of the proximal tubular mass (28). The response to UUO by the proximal tubule can be viewed as “failed differentiation” of epithelial cells (41). In contrast, the distal nephron appears to possess adaptive mechanisms for coping with obstructive injury: collecting ducts undergo cellular remodeling and maintain their patency. Cellular remodeling also plays a role in the sealing of the urinary pole of Bowman's capsule in the formation of atubular glomeruli that remain perfused and continue to produce renin (14). The rapidity of proximal tubular damage in the obstructed kidney appears to be a primary determinant of renal parenchymal loss, followed by progressive increase in extracellular matrix.

These segment-specific nephron responses to obstructive injury may be considered in the context of a cellular “fight-or-flight” response to stress (17). Selective oxidant injury to the obstructed proximal tubule causes widespread DNA damage and nuclear and cytoplasmic disruption (Fig. 2). This represents a “flight” reaction to a toxic microenvironment that results in necrosis and apoptosis and is not successfully controlled by autophagy (a fight response inadequate to counter the stimulus of prolonged complete UUO). In contrast, the collecting duct initially responds to UUO with a burst of mitosis and remodeling of the dilated epithelium: a fight response made possible by fewer metabolic demands and greater antioxidant defenses than exist in the proximal tubule (25). Although glomeruli initially remain viable by remodeling of Bowman's capsule (a fight response), they are nonfunctional, because filtrate cannot escape. Notably, in response to release of temporary partial UUO in the neonatal mouse, formation of atubular glomeruli is arrested, and nephron remodeling is virtually complete (37). Thus, in response to a less severe stimulus (transient partial, rather than persistent complete, UUO), the fight-or-flight cellular responses are adaptive. Application of this paradigm in the interpretation of the model of murine UUO may prove useful in focusing attention on segmental tubular responses to abort the entire cascade of events that eventually lead to interstitial fibrosis.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-45179 and a grant from the University of Virginia Children's Hospital.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

M.S.F. and R.L.C. are responsible for conception and design of the research; M.S.F., B.A.T., J.J.M., K.A.G., and C.I.G. performed the experiments; M.S.F., B.A.T., J.J.M., K.A.G., C.I.G., and R.L.C. analyzed the data; M.S.F., B.A.T., J.J.M., K.A.G., C.I.G., and R.L.C. interpreted the results of the experiments; M.S.F. prepared the figures; M.S.F. drafted the manuscript; M.S.F. and R.L.C. edited and revised the manuscript; M.S.F., B.A.T., J.J.M., K.A.G., C.I.G., and R.L.C. approved the final version of the manuscript.