INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE AVIAN KIDNEY features a number of unique morphological and physiological characteristics. For example, there is considerable heterogeneity in nephron structure, ranging from very short, cortical nephrons, which lack loops of Henle (loopless nephrons), to deeper, transitional nephrons and, finally, to looped "mammalian-type" nephrons. The latter, representing a minority population of nephrons, have a parallel arrangement of descending and ascending loop segments and collecting ducts organized into bundles called medullary cones (4). Another feature that differs from the mammalian kidney is the presence of a renal portal blood supply, whereby cortical regions of the kidney are perfused by a mixture of postglomerular arterial blood and venous blood entering the kidney via afferent portal blood vessels from the hindlimbs (31).

The physiological significance of these features is not clear but may be related to the marked secretory capacity of avian proximal tubules. Urate and phosphate, as well as other organic anions and cations, can be excreted by the kidney in quantities that exceed their filtered loads by as much as severalfold (10, 17, 32). In the case of urate, a high-capacity system transports this relatively insoluble compound from peritubular capillaries to the tubule lumen, where it then forms a colloidal suspension. Recent studies on isolated, perfused avian proximal tubules have added to our understanding of the cellular transport mechanisms for urate, but many details remain unresolved (6). Similarly, the transport properties of the phosphate secretory pathway are poorly understood. In birds, parathyroid hormone is known to inhibit a reabsorptive flux, as in mammals, and to stimulate a secretory flux, thus augmenting the excretory capacity of the kidney (32). Recent experiments with chicken renal brush-border membrane vesicles have identified a potentially unique K⁺- and voltage-sensitive phosphate carrier that may play a role in the secretory transport pathway (2).

Avian renal transport processes have been studied by a variety of experimental approaches, including whole animal clearance (21, 32), micropuncture (16, 17), in vitro microperfusion (6), and membrane vesicle preparations (21, 22). Whole kidney cell cultures have also been used, primarily to study biochemical properties of the kidney cells (13). However, a well-defined primary cell culture model for the avian proximal tubule has not been developed. Any such model, to be useful in the study of avian transport systems, must be shown to have characteristics consistent with those of vertebrate proximal tubules. We report here some of the morphological and physiological properties of chick proximal tubule cells grown on permeable filter supports.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals and reagents. White Leghorn chicks were hatched in incubators and fed chick starter mash and water. Seven to 10 chicks were killed at 3-5 days of age for each cell culture preparation. Collagenase (type A), Dispase II, and collagen (type I) were purchased from Boehringer Mannheim (Indianapolis, IN). All other cell culture reagents were purchased from Sigma Chemical (St. Louis, MO). Growth medium was serum-free DMEM-Ham's F-12 (1:1) supplemented with 5 µg/ml insulin, 5 µg/ml transferrin, 5 ng/ml selenite, 2 µg/ml hydrocortisone, 20 µM ethanolamine, and penicillin-streptomycin at 100 U/ml and 100 µg/ml, respectively. In later experiments, antibiotics were removed from the growth media and used only for the initial preparative steps. Cells were grown on 10-mm permeable filter inserts with 0.02-µm pore size (Nunc, Naperville, IL) placed in 24-well culture dishes.

Tubule isolation and cell culture. The techniques originally described by Vinay et al. (28) for isolation of rat tubule segments were modified for this study. Chicks were killed by cervical dislocation, and kidneys were removed aseptically by cutting away the synsacrum and teasing out large pieces of tissue. The kidney pieces were washed in ice-cold Hanks' balanced salt solution (HBSS), cleaned of visible blood vessels, ducts, and connective tissue, and minced with a scalpel. Tissue was kept on ice until all kidneys were harvested. Minced tissue (~1-2 g) was then incubated for 30 min in 10 ml of growth medium containing 1.0 mg/ml collagenase A and 0.6 U/ml Dispase II. The incubation was carried out at 37°C in the presence of 5% CO₂-95% air with gentle agitation.

Ultrastructure. Monolayers on permeable filters were fixed in 2% glutaraldehyde in Tyrode-cacodylate buffer at 7-10 days after seeding. For scanning electron microscopy, the entire filter was progressively dehydrated in an ethanol series, then critical-point dried. For transmission electron microscopy, monolayers were fixed as described above, then postfixed in osmium for 2.5 h at 23°C. The cells were then dehydrated and embedded in British araldite. The filter side of the block was attached to a Teflon-coated glass slide and dried at 60°C for 72 h. The cell side of the block was then glued to a flat substrate with cyanoacrylate glue, allowing the cell monolayer to be separated from the filter. Sections were cut with a diamond knife, counterstained in uranyl acetate and lead citrate, and viewed with a Philips 201 electron microscope.

Enzyme analyses. Glucose-6-phosphatase, a proximal cell enzyme marker, and hexokinase (HK), a soluble distal tubule marker, were assayed in the initial kidney homogenate (after enzymatic and mechanical dissociation) in the material from PT Percoll bands, and in F1 band tissue, which, as described by Vinay et al. (28), consisted of a heterogenous mixture of glomeruli, thin distal tubule fragments, and proximal tubule fragments. Tissue from each fraction was washed several times in HBSS, resuspended in 1 ml of sucrose-EDTA buffer (250 mM sucrose, 1 mM EDTA), and homogenized with a Dounce-type glass tissue grinder, then subjected to ultrasonic disruption. Samples were stored at 80°C until assays were performed. Glucose-6-phosphatase was assayed at 37°C by a method described by Baginsky et al. (1), with results expressed as micromoles of P_i produced per minute per milligram of protein. HK was assayed at 21°C by the method of Joshi and Jagannathan (15), as modified by Scholer and Edelman (24). Results are expressed as nanomoles of NADP reduced per minute per milligram of protein. Total protein was determined by the bicinchoninic acid assay kit of Pierce Chemical (Rockford, IL). In addition to these markers, the proximal brush-border membrane enzyme -glutamyltranspeptidase (-GT) was measured in an in situ assay designed to assess the cellular polarity of enzyme distribution in these monolayers. The assay was based on protocols described by K. Amsler (Technical Bulletin 404, Becton-Dickinson, Lincoln Park, NJ) with use of the enzyme assay of Naftalin et al. (19). Separate monolayers were exposed to a substrate solution containing -glutamyl-p-nitroanilide on the mucosal or serosal side of the insert. After a 20-min incubation at 21°C, the reaction was stopped and assays were performed on solutions from the mucosal (insert cup) and serosal (well of 24-well culture dish) sides. The concentration of the colored product, p-nitroaniline, was measured at 405 nM and compared with absorbances of known standards. The side not exposed to substrate served as a control for product leakage across the monolayer. This assay, therefore, demonstrates the relative enzyme activity on apical vs. basolateral membranes of the cultured monolayers. Results are expressed as nanomoles of p-nitroaniline produced per minute per monolayer.

Electrophysiology. The electrophysiological characteristics of these cultures were evaluated at 5-10 days after seeding. Insert cups with intact filters and monolayers were mounted, with use of a special adaptor, in modified Ussing flux chambers (MRA, Naples, FL). This approach thus avoids edge damage to the monolayers. A transport buffer containing 135 mM NaCl, 4 mM KCl, 1.3 mM CaCl₂, 1 mM MgSO₄, 5 mM HEPES, and 25 mM NaHCO₃, with or without 5 mM glucose, was circulated on both sides and gassed with 5% CO₂-95% air (pH 7.4). The monolayers were short circuited with an automatic voltage clamp (model DVC 1000, World Precision Instruments) with correction for fluid resistance compensation. Short-circuit current (I_sc) was continuously measured and displayed on a strip chart recorder, with intermittent measurements of open-circuit potential (PD). Tissue resistance (R) was determined by measuring current deflections in response to brief (1 s) changes in clamping voltage. After a stabilization period, reagents were added to the mucosal and/or serosal reservoirs, and the changes in I_sc from the previous, stable baseline were determined. Current responses to mucosal addition of the following reagents were determined: -methyl-D-glucose (-MG), a specific substrate for the Na⁺ gradient-driven glucose cotransporter (2 mM); phenylalanine (2 mM); phloridzin, a specific inhibitor of the Na⁺-glucose cotransporter (1 mM); and amiloride, an inhibitor of epithelial Na⁺ channels (10⁵ M). Addition of phloridzin to the basolateral side was also tested in a separate series of experiments. Transepithelial potentials were expressed with reference to the mucosal side; positive I_sc indicates mucosal-to-serosal flow of current.

Urate transport. Intact filter cup monolayers in 24-well plates were washed and preincubated for 15-20 min in transport buffer. The bathing solutions (0.5 ml both sides) were then replaced with transport buffer containing 15 mg/dl uric acid. The plates were then placed on a rocking platform in a 37°C incubator. After a 5-min mixing period, initial aliquots were taken from the apical (cup) and basolateral (well) sides of the monolayers (time 0). Incubation was then continued for an additional 120 min, and then a second set of samples was removed. Net urate transport was assessed from the change in urate concentration over the 2-h period in the apical vs. basolateral solutions. Urate was determined by the method of Morgenstern et al. (18).

Statistics. Values are means ± SE. Significant differences were determined by Student's t-test, paired or unpaired as appropriate.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

The tissue isolated from the PT band of the Percoll gradient was highly homogenous, consisting almost entirely of short, intact tubule fragments (Fig. 1A). These fragments had large tubule diameters and a characteristic yellow appearance of the cytoplasm, typical of proximal tubule cells. In contrast, whole kidney homogenate and F1 band tissue showed a high degree of heterogeneity, with a mixture of single cells, proximal fragments, and thin, distal segments.

View larger version (123K): [in this window] [in a new window]   Fig. 1.   Chick proximal tubule fragments (A) and 24-h cultures (B). A: sample of proximal tubule band material taken immediately after enzymatic digestion and Percoll gradient separation of chick kidney tissue. Phase contrast; scale bar, 100 µm. B: 24-h cultures of proximal tubule fragments. Phase contrast; scale bar, 50 µm.

Enzyme assays performed on the PT band (Table 1) showed a significant increase in specific activity of glucose-6-phosphatase, a proximal tubule marker (P < 0.05), and a significant decrease in specific activity of HK, a distal tubule marker (P < 0.05), compared with initial whole kidney homogenates. In contrast, glucose-6-phosphatase specific activity in the F1 fraction was significantly reduced (P < 0.05), whereas HK activity was unchanged. These data indicate an enrichment of the PT band in proximal tubule cells.

Tubule segments from the PT band adhered to the collagen-coated filters and formed flattened, proliferating colonies within 24-48 h in culture (Fig. 1B). Except for occasional loss of cultures due to microbial contamination, confluent, functional monolayers were obtained from every proximal tubule preparation. Monolayers typically reached confluence within 5-7 days, as determined by semiquantitative electrical resistance measurements (EVOM). Inulin restriction assays showed a fourfold lower inulin diffusion rate on 7- to 10-day cultures than on cell-free filters over a 3-h period: apical-to-basolateral inulin concentration ratios were 0.14 ± 0.02 and 0.56 ± 0.03 (SE) for monolayers and cell-free filters, respectively (n = 12). This difference was statistically significant (P < 0.001).

Electron microscopy of the confluent cultures revealed cuboidal cells with numerous mitochondria distributed throughout the cell and short but numerous apical microvilli (Fig. 2). At higher magnification, well-developed tight junctions and lateral infoldings were also evident, as were apical vacuoles. Subapical, electron-dense material, possibly components of the terminal web, can also be seen across the cell diameter. Examination of the apical surface with scanning electron microscopy (Fig. 3) showed a well-developed microvillus brush border as well as central cilia extending from many of the cells.

View larger version (146K): [in this window] [in a new window]   Fig. 2.   Transmission electron micrograph of proximal tubule cells in culture. Confluent monolayers (7-10 days) on permeable filter inserts were fixed in 2% glutaraldehyde in Tyrode-cacodylate buffer, postfixed in osmium, and embedded in British araldite. Filter was separated from monolayer before sectioning. Note cuboidal shape of cells and numerous microvilli and mitochondria. Scale bar, 5 µm.

View larger version (186K): [in this window] [in a new window]   Fig. 3.   Scanning electron micrograph of 7-day proximal tubule cell culture. Confluent monolayers were fixed in 2% glutaraldehyde, dehydrated, and critical-point dried. Apical surface of monolayer is shown. Note numerous microvilli and central cilia extending from individual cells. Scale bar, 2 µm.

An enzyme assay for -GT was performed on confluent, filter-grown monolayers to assess proximal tubule-like expression and functional polarity of these cultures. The results, shown in Fig. 4, indicate a threefold higher enzyme activity on apical than on basolateral sides of the monolayer (P < 0.001), indicative of some combination of preferential apical localization of -GT and/or apical surface area amplification.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   -Glutamyltranspeptidase (-GT) activity on apical and basolateral sides of proximal tubule cell monolayers. Confluent cell cultures were assayed for -GT activity in situ by measuring product formation in apical (insert cup) or basolateral (well) media. Separate monolayers were used for measurements on the 2 sides. Substrate solution contained 5 mM L-glutamyl-p-nitroanilide and 100 mM glycylglycine. Incubations were performed for 20 min at 21°C; n = 9 pairs from 4 proximal tubule preparations. * Significantly different from apical (P < 0.001).

Electrophysiological measurements were obtained for confluent monolayers under short-circuited conditions in the presence and absence of 5 mM glucose on both sides. The stable, initial values for PD, R, and I_sc are shown in Table 2. PD and R were not significantly different between the two conditions. I_sc, in contrast, was doubled in the presence of glucose (P < 0.01).

The presence of rheogenic transporters on these monolayers was further evaluated by mucosal addition of a nonmetabolizable hexose and an amino acid to inserts bathed in the glucose-free transport buffer (Fig. 5). Mucosal addition of 2 mM -MG increased I_sc by ~3 µA/cm², and 2 mM phenylalanine further increased current by an additional 2 µA/cm². Prior serosal addition of 2 mM glucose had negligible effects on I_sc. Phloridzin (1 mM), a specific inhibitor of the Na⁺-glucose luminal cotransporter (SGLT), inhibited I_sc by 3.5 µA/cm², i.e., slightly more than the net increase elicited by mucosal addition of -MG. In a separate series of experiments, basolateral addition of 1 mM phloridzin also inhibited I_sc by 1.07 ± 0.34 µA/cm² (n = 7 monolayers) or 47% of the -MG-stimulated current in this series. Subsequent addition of 1 mM phloridzin to the apical side inhibited the remaining fraction of -MG-stimulated current in every case. Phloridzin had no effect on I_sc in the absence of apical hexose. Finally, mucosal addition of 10⁵ M amiloride inhibited a significant portion of the remaining current after phloridzin (Fig. 5).

View larger version (36K): [in this window] [in a new window]   Fig. 5.   Changes in short-circuit current (Isc) of proximal tubule monolayers in response to various agonists and antagonists. Monolayers were placed into Ussing chambers, bathed on both sides with a glucose-free transport buffer, and voltage clamped. After stabilization of Isc, 2 mM -methyl-D-glucose (-MG), 2 mM phenylalanine (Phe), 1 mM phloridzin (Phlor), and 105 M amiloride (Amil) were sequentially added to mucosal bathing solution. Positive values indicate stimulation of Isc; negative values indicate inhibition. Between each addition 5-10 min were allowed. Each change was significantly different from immediately preceding stable baseline (P < 0.001, n = 13 monolayers from 6 different PT preparations).

Net urate transport was assayed in 13 monolayers from three culture preparations. Initial (time 0) urate concentrations were identical on the two sides and averaged 15.34 ± 1.82 mg/dl. Over a 2-h incubation period, these monolayers generated a concentration difference of 16.47 ± 1.43 mg/dl on the apical side vs. 13.93 ± 1.40 mg/dl on the basolateral side. The mean paired difference (apical  basolateral) of 2.43 ± 0.71 mg/dl was statistically significant (P < 0.005, paired t-test).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study was undertaken to develop suitable methods for producing a transporting, primary cell culture model of the avian proximal tubule and to validate the model by various criteria. The approach described by Vinay et al. (28) for isolation of rat proximal tubule segments has been widely adapted for primary cell culture of proximal cells from a number of mammalian species (3, 14). In the present study, modifications to this procedure, including reduced collagenase concentration, addition of Dispase, and a trituration step, resulted in a well-dispersed population of small tubule fragments. These could then be separated on Percoll gradients, as described by Vinay et al., to yield a highly homogenous population of viable proximal tubule fragments. The resulting confluent monolayers, grown under defined conditions, display many of the characteristic features of vertebrate proximal tubules.

Morphologically, these cells typify a polarized transporting epithelium, with cuboidal cell shape, tight junctions, lateral interspaces (seen at higher magnification), and numerous microvilli. As has been previously described for avian proximal tubules (26, 31), numerous mitochondria can be seen scattered throughout the cytoplasm, contrary to the pattern seen in mammalian cells, where these organelles are localized to deep basal infoldings. Also seen in these cells are apically situated vacuoles, another characteristic of proximal tubule cells (26, 31).

Assays were performed on Percoll-separated fractions for known proximal and distal marker enzymes to monitor the relative enrichment of proximal fragments. The results show that specific activity of glucose-6-phosphatase, a proximal tubule enzyme, was significantly enriched in the PT or proximal fraction and significantly reduced in the F1 or mixed/distal fraction compared with the initial homogenates (Table 1). Furthermore, specific activity of HK, a distal cytoplasmic enzyme, was significantly reduced in the PT fraction but unchanged in the F1 material. Taken together, these results indicate that the starting material for seeding these cultures is highly homogenous in proximal tubule segments, in agreement with the visual observations of PT band fragments (Fig. 1A).

For both of these enzymes, specific activities in the whole kidney homogenate agree well with previously reported values for 0- to 10-day-old chicks (25). Furthermore, the specific activity of HK in the PT fraction of ~2 nmol · min¹ · mg¹ agrees well with rat F4 values reported by Gesek et al. (11). Vinay et al. (28) report somewhat higher values, having assayed this enzyme at 37°C rather than at room temperature. Using a different tubule isolation procedure, Toutain et al. (27) found much lower HK activity in initial proximal isolates (~0.2 nmol · min¹ · mg¹) with, however, a marked increase in specific activity in cultured cells over a 16-day period.

Glucose-6-phosphatase activities reported in the literature seem to show much higher variability. The specific activity of glucose-6-phosphatase in the PT fraction reported in the present study is ~3-fold higher than that reported for rat cortex (7) and 10-fold greater than activities in rat proximal tubule isolates (11, 27). The much higher values reported here and by Shen and Mistry (25) may be related to the important role attributed to the avian kidney in gluconeogenesis from amino acid substrates (29).

In addition to these enzyme data, the cultured monolayers in the present study also expressed high levels of a brush-border membrane marker -GT. A compartmental assay based on in situ generation of a soluble colorimetric product demonstrated a threefold higher activity of this enzyme in apical than in basolateral cell membranes (Fig. 4), thus indicating a high degree of polarized differentiation. Although generally considered a proximal brush-border enzyme, -GT has also been localized, by immunocytochemical methods, to basolateral membranes of rat proximal tubules (20). It was postulated that basolateral localization may be important in the extraction of glutathione by proximal tubule cells. The finding of basal -GT activity in the present study (Fig. 4) may suggest a similar function in avian kidneys.

The electrophysiological characteristics of these monolayers are representative of vertebrate proximal nephron segments. Transepithelial resistances, although somewhat higher than those seen in mammalian proximal tubules or primary cultures (3, 23), are nevertheless in the range of values seen in other vertebrate species studied (10, 12). There have been no prior measurements of avian proximal tubule resistance, although micropuncture studies of proximal tubules in the European starling have shown low, lumen-negative transepithelial potentials, consistent with a low-resistance epithelium (16).

Previous studies with brush-border membrane vesicles have demonstrated the presence of an SGLT in chick proximal tubules (21). In the present study, proximal cultures incubated in the presence of 5 mM glucose (mucosal and serosal) had twofold higher currents than those incubated in zero glucose (Table 2). Mucosal addition of 2 mM -MG, a nonmetabolizable substrate of the SGLT, stimulated current by ~50% (2.9 µA/cm²) in glucose-free monolayers, whereas prior addition of 2 mM glucose to the serosal side had a minimal effect on the I_sc, indicating a specific stimulation of current attributable to apical Na⁺-glucose cotransport. Furthermore, mucosal addition of 1 mM phloridzin, a specific inhibitor of SGLT, reduced the I_sc by 7.5 and 3.5 µA/cm² in the presence of 5 mM glucose or 2 mM -MG, respectively (30-50%). Thus these cultures appear to express a well-defined transport activity characteristic of vertebrate proximal tubules.

A somewhat surprising observation was the partial inhibition of -MG-stimulated current by basolateral phloridzin. Because glucose itself had almost no effect on I_sc when added to the basolateral side, this effect of phloridzin is interpreted as organic anion secretion, i.e., delivery of basolateral phloridzin to the apical side via an organic anion transport system and subsequent apical inhibition of the SGLT. Net renal secretion of phloridzin has been demonstrated in chickens, as well as in dogs and the aglomerular fish Lophius americanus (5). In dogs, phloridzin secretion was inhibited by p-aminohippurate infusions, suggesting a common transport pathway for these organic anions.

In addition to glucose-stimulated current, these monolayers also express an amino acid-stimulated I_sc and a component of amiloride-sensitive current. The nature of this latter component is unknown. Amiloride-sensitive Na⁺ channels, although generally considered characteristic of high-resistance epithelia, have also been demonstrated in late proximal segments of the rabbit (33) and in LLC-PK₁ cells, a cell line derived from porcine late proximal or S3 segments (8). Of interest is the finding that 10⁶ M amiloride also inhibited gradient-driven Na⁺ uptake into chick renal brush-border membrane vesicles (22), although this result cannot distinguish between possible effects on Na⁺ channels and an Na⁺/H⁺ exchanger. Amiloride also affects numerous other transport processes, although generally at much higher concentrations than those used in the present study.

In birds, net tubular secretion of urate has been demonstrated in proximal segments by micropuncture (17) and in vitro perfusion techniques (6). The latter experiments, however, also revealed a high paracellular permeability to urate and, consequently, a marked dependence of urate secretion on luminal flow. In the present study, cultured monolayers also demonstrated an ability to secrete urate, as shown by the steady-state concentration gradient developed over a 2-h incubation period. The small size of the gradient is consistent with a large backleak, as would be expected under steady-state conditions.

The degree of cellular heterogeneity of these cultures remains to be established. Although monolayers display properties typical of vertebrate proximal tubules, these segments are known to be composed of multiple cell types. This heterogeneity can arise from internephron and axial, intranephron variation (i.e., analogous to S1, S2, and S3 segments of mammalian nephrons). Between 10 and 30% of the nephrons in avian kidneys contain simple loops of Henle arranged in bundles called medullary cones (4). The remaining nephrons lack loops of Henle but display considerable heterogeneity in overall length, glomerular size, and folding patterns (26, 31). Within individual nephrons, axial heterogeneity has been noted in overall tubule diameter and in cytoplasmic staining intensity, with two apparent segments recognized (26). Despite this potential cell heterogeneity, the overall characterization of these monolayers is consistent with avian proximal tubule function. Thus this model represents an important new approach for the study of avian proximal tubule transport processes.