Main

The newborn is limited in its ability to defend against acid-base disturbances(1–4). Limitations in distal, as well as proximal, nephron H+/HCO-3 transport ability have been suggested as underlying causes for this susceptibility(3–8). Previous transport studies from our laboratory have shown that isolated perfused neonatal CCD do not secrete HCO-3 in vitro, in contrast to mature CCD from alkaline ash-fed rabbits(6). On the other hand, the OMCD secretes protons at rates approaching that of the mature segment(6).

The CCD is comprised of two cell types, principal cells and intercalated cells; the latter mediate H+/HCO-3 transport and contribute to the final regulation of urinary acidification(9, 10). Intercalated cells have an alkaline cell pH and are rich in carbonic anhydrase, mitochondria, and H+ pumps; they are capable of vectorial H+ secretion into the luminal fluid(A-intercalated cells) or HCO-3 secretion (B-intercalated cells)(9–13).

The HCO-3-secreting intercalated cells of the cortical collecting duct have an apical Cl/base exchanger and, in the rabbit, bind PNA and MAb B63 to the apical surface(14–17). The presence of these two subtypes of intercalated cells in the fully differentiated CCD accounts in part for the ability of this segment to secrete net HCO-3in vitro while the rabbits ingest an alkaline ash diet and net H+ after in vivo acid treatment(12, 18, 19). In addition, there are intercalated cells that appear to have some characteristics of both A and B types(14, 17, 20, 21) and have been called “hybrid” intercalated cells by some investigators(14, 17) and γ cells by others(20); their contributions to net HCO-3 transport have not yet been established.

OMCD and IMCD consistently secrete H+ (absorb net HCO-3(18, 22). Rarely, if ever, have HCO-3-secreting intercalated cells been found in the OMCD or IMCD of either the adult rat or rabbit(17, 23–26). Moreover, intercalated cells are not commonly found in the rabbit IMCD(13, 26), indicating that H+ secretion is accomplished by other cells therein.

Fluorescent functional studies have shown that neonatal intercalated cells of the rabbit cortical collecting duct do not exhibit as alkaline a cell pH, as many acidic cytoplasmic vesicles, or as many mitochondria as observed in mature intercalated cells(27, 28). Moreover, intercalated cells have not been identified in the neonatal outer cortex(27, 28). These data would indicate that HCO-3-secreting intercalated cells are either not present or not functionally developed in the neonatal CCD(14, 29, 30). On the other hand, the neonatal OMCD more closely resembles the mature segment(27, 28), as would be expected from the transport studies(6). These studies would suggest that early in life H+-secreting intercalated cells are more differentiated than HCO-3-secreting cells. The purpose of the present study is to better characterize the maturation of H+- and HCO-3-secreting intercalated cells in the collecting duct using cell-specific markers and immunohistochemistry and correlate the findings with previous transport and cell physiologic studies.

METHODS

Animals. New Zealand White rabbits were purchased from Hazleton-Dutchland Farms (Denver, PA) or from Charles River (Wilmington, MA). Pregnant dams were allowed to deliver in our animal quarters to provide newborns (1-7 d of age) and 3 wk olds (22-27 d of age). Adult females (2.0-2.5 kg) and pregnant dams were fed standard laboratory rabbit chow (Purina Mills HF 5326, Richmond, IN) and allowed free access to tap water. The pups were fed by and raised with their mothers. At least three different litters were used in these studies.

Kidney sections. Rabbits were anesthetized with pentobarbital(30 mg/kg of body weight, i.v.). In the adult the left renal artery was cannulated with a polyethylene 160-190 (Clay Adams, Parsippany, NJ) catheter, perfused with PBS for 10-15 s to eliminate erythrocytes, and perfusion fixed with periodatelysine-paraformaldehyde for 15 min at a pressure of 150 cm water(31). Similar infusions were administered to newborn and 3-wk-old rabbits via a polyethylene 50 catheter inserted into the aorta after ligating all but the renal arteries. The perfusion pressure was 100 cm water in these younger animals(14). Kidney cortex and medulla were cut into 1-2-mm thick cubes and immersed in the same fixative for 2 h at 4°C. After fixation, the tissue was osmotically protected by immersing in 1 M sucrose in PBS at 4°C until it sank (3-4 h)(14). It was embedded in Tissue-Tek OCT compound (Miles, Elkhart, IN) and frozen at-80°C. Cryostat sections (2 μm) were cut and transferred to poly-L-lysine (500 μg/mL; Sigma Chemical Co., St. Louis, MO) coated glass slides for immunocytochemistry and/or peanut lectin labeling.

We focused particularly on cells in the CCD, OMCD, and IMCD. Kidneys from at least three different animals from three different litters were used for immuno- and peanut lectin labeling. All cells in focus in a given field were analyzed; at least two different fields per zone of kidney were examined(outer cortex, mid-inner cortex, outer medulla, inner medulla).

Immuno- and lectin labeling. Sections were immersed in 1% BSA in PBS for 20 min for blocking. The following MAb were used to identify H+/HCO-3 transport epitopes on collecting duct intercalated cells: E11, an antibody to the 31-kD subunit of the vacuolar H+-ATPase(32), provided by Dr. Stephen Gluck; IVF12, an antibody to the membrane domain (anion transporting region) of human erythrocyte band 3 and a marker of H+-secreting intercalated cells(24), provided by Dr. Victor Schuster; and MAb B63 to PNA-binding B-intercalated cells(14, 16). To intensify the band 3 (IVF12) and H+-ATPase (E11) labeling, slides were usually immersed in 6 N guanidine hydrochloride, 50 mM Tris-HCl (pH 7.4) at room temperature for 15 s before application of these antibodies(33). MAb ST.9(34) to principal cells of the rabbit was used to provide information regarding the development of principal cells in the neonatal kidney.

All MAb were supplied as supernates from hybridoma cultures and were applied undiluted to tissue sections for 1 h at 37°C. After three rinses in PBS for 5 min each, sections were exposed for 1 h at room temperature to FITC conjugated to rabbit anti-mouse Ig (IgG + IgA + IgM) (Zymed; San Francisco, CA) at a dilution of 1:100, followed by 3 more 5-min rinses in PBS. Background staining of proximal tubules (arrows) was consistently found despite using a variety of different secondary antibodies (see Fig. 1a). Glomeruli (G) and collecting ducts(not shown) exhibited much less background fluorescence (see Fig. 1a).

Figure 1
figure 1

MAb E11 (H+-ATPase). (a) Adult: FITC secondary antibody staining without E11 antibody shows significant background in proximal tubules (arrows) which was difficult to eliminate. A glomerulus(G), negatively labeled, is in the center. This background was observed with all antibodies counterstained with FITC secondary. (b) Newborn: superficial cortex is not stained. The arrowhead at the left denotes the boundary of the outer cortex; the arrow shows an S-shaped vesicle.

Full size image

Other sections were exposed to 50-100 μg/mL rhodamine conjugated to PNA(Vector; Burlingame, CA) for 1 h at room temperature, followed by three rinses of 5 min each. The PNA was often used to counterstain sections previously labeled with MAb and fluorescein secondary(14). To characterize the distribution of Ell (H+-ATPase) labeling in PNA-identified cells, we examined approximately 300 cells in 25 cross-sections of cortical collecting ducts from newborn, 3-wk-old, and adult rabbit kidneys.

Because the resolution of the immunofluorescence does not clearly discriminate between cytoplasmic and membrane staining, we identified patterns as being predominately apical, basolateral, or cytoplasmic (diffuse/nonpolar). Note that the adult rabbit CCD rarely shows basolateral H+-ATPase staining(17), in contrast to that of the rat(25, 35, 36).

Imaging of sections. Slides were coverslipped using Cytoseal 60(Stephens Scientific, Riverdale, NJ). They were examined using a Nikon Diaphot inverted epi-fluorescence microscope equipped with a Nikon FG camera body attached to the 2.5 × camera port. Images were generally photographed at a real magnification of × 125, processed, and printed on 4 × 5 inch paper; the real magnification of one photograph (Fig. 4d) was × 265. We used a 50 × water immersion objective (1.0 numerical aperture) and exposed fluorescent images at 1600 ASA and push processed in high contrast developer to provide the best quality photographs. In order to determine the percentage of band 3-positive cells we also obtained some corresponding phase contrast images. T-Max 400 film (Eastman Kodak, Rochester, NY) was used for these photographs.

Figure 4
figure 4

MAb B63. (a) Newborn: superficial cortex stained with MAb B63. Occasional cells in the vicinity of S- and comma-shaped bodies stained intensely and without polarity (arrows). The most superficial boundary of the kidney is seen at the wide arrow.(b) Newborn: mid CCD showing weak apical staining in several cells(arrows). Background staining of proximal tubules was evident in the absence of primary antibody. (c) Three weeks: mid CCD showing more intense apical staining of some cells (arrow). Weak background staining of proximal tubules was evident in the absence of primary antibody.(d) Adult: mid CCD from adult showing apical staining(arrows) of some cells which was more intense than that seen at younger ages. This figure was prepared and printed at higher magnification.

Full size image

RESULTS

The expression of these epitopes as a function of maturation and location within the kidney is summarized in the Table 1.

Table 1 Maturation of collecting duct epitopes
Full size table

Intercalated Cell Phenotype

E11 (H+-ATPase) . Newborn. In the superficial(outer) cortex there was no labeling by MAb E11 within the S- or comma-shaped bodies or developing collecting ducts (Fig. 1b,arrow). In the mid to deep cortex there were two patterns of E11 labeling. Most cells (47/59 or 80%) showed a generally cytoplasmic staining pattern (Fig. 2,a, arrows, and b, wide arrow). However, in 20% of the positive cells an apically accentuated appearance was noted in the CCD (Fig. 2b, thin arrows). For these and subsequent photomicrographs, at this power of magnification, it could not be established whether or not the staining was at or near the apical membrane. The labeling of the neonatal collecting duct cells was usually less intense and the number of labeled cells fewer than that observed in the 3-wk and adult kidney cortex (see later).

Figure 2
figure 2

MAb E11 (H+-ATPase). (a) Newborn: cells in the mid CCD bound MAb E11 (arrows) in a cytoplasmic pattern without preference to apical or basolateral staining. (b) Newborn: some cells in the mid CCD showed a predominately apical H+-ATPase staining pattern (thin arrows). Other cells showed a more cytoplasmic pattern (wide arrow). Proximal tubular staining at the top shows high background, because it was also seen in comparable sections in the absence of primary antibody. (c) Newborn: apical staining in the OMCD (arrows). (d) Newborn: apical staining in the neonatal initial IMCD (arrow). (e) Three weeks: approximately half the E11-identified cells in mid CCD showed apical staining(thin arrow), whereas some cells showed a cytoplasmic pattern(wide arrows). (f) Three weeks: apical staining in the OMCD (arrows). (g) Three weeks: apical staining in the IMCD (arrow). (h) Adult: approximately two-thirds of the E-11 identified cells in the CCD showed apical staining (thin arrow), whereas others showed more cytoplasmic staining (wide arrow). The proximal tubular background on the right was seen in comparable sections in the absence of primary antibody. (i) Adult: several cells in the outer stripe of OMCD showed apical staining (arrows).(j) Adult: a few cells in the initial IMCD showed apical staining(arrow).

Full size image

In the outer medulla, most labeled cells (31/40 or 78%) showed apical staining (Fig. 2c, arrows). The few positive cells of the initial IMCD also showed apical staining (20/21 or 95%;Fig. 2d, arrow). Rarely E11-labeled cells were seen in the terminal inner medullary collecting duct.

Three weeks old. S- and comma-shaped bodies were no longer seen in the superficial cortex, but some cells in the superficial CCD showed H+-ATPase labeling. The patterns of staining were not grossly different between outer and mid-inner cortex and were both apical (27/50 or 54%;Fig. 2e, thin arrow) and cytoplasmic patterns(23/50 or 46%; Fig. 2e, wide arrows).

In the outer medulla more than 90% (39/42) of positive cells showed an apical pattern (Fig. 2f, arrows). Those few cells staining with E11 MAb in the IMCD generally showed apical patterns(11/15 or 73%; Fig. 2g, arrow).

Adult. H+-ATPase staining in the mature CCD was mainly apical (10/15 or 66%; Fig. 2h, thin arrow) or cytoplasmic patterns (4/15 or 27%; wide arrow). Predominately apical H+-ATPase labeling was observed in the outer medulla (11/13 or 85%;Fig. 2i, arrows). Some positive cells were found in the initial IMCD, and they were apically labeled, as well (20/22 or 91%;Fig. 2j, arrow).

IVF12 (Band 3) . Newborn. In the superficial cortex there were no MAb IVF12-positive cells. In the mid to deep CCD there were cells with basolateral labeling (9/44 or 20%; Fig. 3a,wide arrow), as well as others with a cytoplasmic (but not apically accentuated) pattern of labeling (35/44 or 80%; arrow). Similarly, both cytoplasmic (22/32 or 69%) and basolateral patterns (10/32 or 31%) were observed in the OMCD (Fig. 3b, arrow) and IMCD(not shown).

Figure 3
figure 3

MAb IVF12 (band 3-like anion exchanger). (a) Newborn: labeling of mid CCD showed typical basolateral staining (wide arrow) as well as cytoplasmic labeling of other cells (arrow).(b) Newborn: basolateral staining (arrow) was sometimes observed in cells of the outer stripe of OMCD. Background staining of proximal tubules was evident in the absence of primary antibody. (c) Three weeks: basolateral (thin arrow) and nonpolar circumferential(wide arrow) staining in the CCD. Background staining of proximal tubules was evident in the absence of primary antibody. The staining of the structure with the large lumen, probably a vessel, is likely to be due to autofluorescence or nonspecific reactivity of the secondary antibody.(d) Adult: basolateral staining of several cells (arrows) in the outer stripe of the OMCD. (e) Three weeks: basolateral staining of cells (arrows) in the initial IMCD.

Full size image

Three weeks old and adult. In the CCD both basolateral (12/23 or 52%; Fig. 3c, thin arrow) and cytoplasmic (11/23 or 48%; Fig. 3c, wide arrow) patterns of staining were observed to approximately the same extent; however, the intensity was generally greater, as well as the number of identified cells, compared with that seen in the newborn. In the OMCD the basolateral pattern was almost exclusively observed (30/34 or 88%; Fig. 3d,arrows). Occasional cells in the initial IMCD showed basolateral(Fig. 3e, arrows) or cytoplasmic labeling to approximately the same extent. Band 3-positive cells were not seen in the terminal IMCD.

Cell counts. Band 3-positive cells (as a percentage of total cells counted in the collecting duct) in the midcortex was 9/71 or 13% in the newborn, 5/30 or 17% in the 3 wk old, and 12/72 or 17% in the adult. In the mature outer cortex there were 15/121 or 12% positive for band 3. These data for the mature kidney are comparable to the 11% band 3-positive cells previously found by Schuster et al.(24) in unspecified regions of the rabbit kidney cortex.

In the outer medulla there were 19/87 or 22% band 3-positive cells in the newborn, 23/65 or 35% in the 3 wk old, and 51/138 or 37% in the adult. The percentage of positive cells in the mature outer medulla was similar to the 43% previously found by Schuster et al.(24). In both cortex and medulla there was a maturational increase in both number and intensity of staining of A-intercalated cells.

MAb B63 to HCO-3-secreting intercalated cells . Newborn. In the superficial cortex there were occasional cells in or near the S- and comma-shaped bodies which stained diffusely in a nonpolar pattern for MAb B63 (Fig. 4a, arrows). Such labeled cells were not seen along the superficial CCD or in the ampulla. Deeper in the cortex along the CCD there appeared cells that labeled apically(Fig. 4b, arrows), although the intensity of staining was less than that observed in more mature kidneys (see next). Indeed, as shown previously(14), we had used 5-15-μm sections in the newborn to show adequate staining with this antibody. No staining was observed in the OMCD or IMCD (not shown).

Three weeks old and adult. MAb B63-positive cells expressing apical staining were observed in the CCD throughout the entire cortex in both 3-wk-old and adult rabbit kidneys (Fig. 4,c and d, arrows). The number and intensity of labeled cells increased with maturation (Fig. 4, compare b, c, andd). No staining was observed in the OMCD or IMCD (not shown).

PNA . Newborn. PNA staining was not observed in the superficial CCD; however, along and around the ampullary bud and neck structures and S-shaped bodies there was positive staining(Fig. 5a, arrows), as has been previously shown by Minuth et al.(37). Also, near the S- and comma-shaped bodies there were occasional cells showing a cytoplasmic staining pattern (Fig. 5b, arrow). Deeper in the cortex along the CCD there were cells with apical PNA labeling (Fig. 5c, arrows). There was no staining in the OMCD or IMCD (not shown). Three weeks old and adult. There was apical PNA staining of approximately one-third of all the cells along the CCD of the 3 wk old(Fig. 5d, wide arrows). This staining appeared more intense than that seen in the newborn CCD but less than that of the adult(31). There was no staining in the OMCD or IMCD. The number of PNA labeled cells increased with maturation (Fig. 5, compare c and d), as shown previously in isolated CCD(14).

Figure 5
figure 5

PNA. (a) Newborn: staining of the outlines of S-shaped bodies and ampullary structures in the outer (superficial) cortex(arrows). The most superficial boundary of the cortex is seen at the lower left corner (arrowhead). (b) Newborn: occasional cells staining nonpolar diffusely or cytosolically were found in or near the S-shaped bodies (arrow). The most superficial boundary of the cortex is seen at the top of the figure. (c) Newborn: deeper in the cortex the CCD showed apically labeled PNA-positive cells (arrows).(d) Three weeks: the midcortex shows more intense apical staining of some cells along the lumen of the CCD (wide arrows).

Full size image

Principal Cell Phenotype

MAb ST.9 . Newborn. In the superficial cortex there was no labeling with MAb ST.9. In the mid to deep cortex, several cells in the cross sections of CCDs showed faint apical staining (Fig. 6a, arrowheads). Observing down through the kidney there was an axial increase in number and intensity of staining of labeled cells through the outer medulla, but the intensity was not nearly as great as observed in the older animals (see later).

Figure 6
figure 6

MAb ST.9. (a) Newborn: some cells in the mid CCD stained apically with MAb ST.9 (arrowheads). (b) Newborn: intense apical and particulate staining of most cells along the OMCD(arrows). (c) Newborn: some cells were stained in the IMCD(arrow). The faint lateral staining is believed to be nonspecific.(d) Adult: intense apical staining was found along most cells of the CCD (arrowhead); some cells, presumed to be intercalated cells, were negative (arrow). Weak background staining of proximal tubules was evident in the absence of primary antibody. (e) Adult: several cells intensely stained in the IMCD (arrow).

Full size image

In the OMCD many cells labeled positively with MAb ST.9(Fig. 6b, arrows); staining was more intense than that observed in the cortex, and some particulate or vesicular labeling was evident. Some cells in the IMCD expressed apical and faint lateral staining (Fig. 6c, arrow), but there was no staining in the terminal inner medulla.

Three weeks old and adult. In the cortex through to the outer medulla, most collecting duct cells were stained with MAb ST.9 on the apical membrane (Fig. 6d, arrowhead). Some cells, staining negatively (Fig. 6d, arrow), were presumed to be intercalated cells(34). The staining in both 3 wk old and adult was more intense than observed in the newborn and showed somewhat less particulate labeling as well. In the initial IMCD there was staining of a few cells per cross-section (Fig. 6e,arrow), but no positive cells were seen in the terminal IMCD (not shown).

Double Labeling of Kidney Cortex Sections with PNA

E11 (H+-ATPase) . Newborn. In the neonatal CCD 94% (302/320) of cells identified with E11/PNA were positive for both PNA(apical staining, Fig. 7a, arrowhead) and H+-ATPase (apical or cytoplasmic staining, Fig. 7b,arrowhead), possibly resembling the B-intercalated cell phenotype(17). Only 5% of cells showed apical H+-ATPase staining without PNA labeling (presumed A-intercalated cell phenotype,Fig. 7,a and b, arrows), and only 1 of 320 cells (0.3%) stained positively for PNA but negatively for H+-ATPase.

Figure 7
figure 7

Double Labeling with MAb E11 and PNA. (a andb) Newborn: double staining with PNA (a) and E11(H+-ATPase) (b) revealed occasional cells with apical H+-ATPase and negative PNA label (arrows), whereas most identified cells were labeled apically with PNA and apically, or cytoplasmically with H+-ATPase antibody (arrowheads).

Full size image

Three weeks old and adult. In the older animals there were slightly fewer dual-labeled cells (possibly B-intercalated cell phenotype) in the mid CCD (82% in the 3 wk old and 84% in the adult), and there were 2-fold more H+-ATPase positive/PNA negative A-intercalated cell phenotypes(12% in both the 3 wk old and the adult). Approximately 4% of identified CCD cells were PNA positive but H+-ATPase negative and could not be further classified.

MAb B63. Newborn. Occasionally individual cells near the S-shaped bodies were observed with cytoplasmic staining for both MAb B63(Fig. 8,a and c, arrows) and PNA(Fig. 8,b and d, arrows). Cells of the mid to deep CCD labeled apically with both MAb B63 and PNA (Fig. 8,e and f, wide arrows, respectively). There was rarely any deviation from this 1:1 correspondence.

Figure 8
figure 8

Double Labeling with MAb B63 and PNA. (a andb) Newborn: an individual cell in or near the S-shaped vesicle was identified with both MAb B63 (a) and PNA (b)(arrows and arrowheads). Staining was essentially cytosolic. (c and d) Neonate: another kidney section showing an isolated cell stained with both MAb B63 (c) and PNA(d) (arrows). In this and the preceding (a andb) pair, the staining of cells in and around the S- and comma-shaped vesicles by both MAb B63 and PNA is likely to be nonspecific. Staining is probably due to the FITC secondary in the case of MAb B63 because peroxidase secondary failed to show staining there. The PNA staining of the vesicles was quite variable and also not likely to be specific, as previously shown by Minuth et al.(37). (e andf) Neonate: double staining of cells along a mid CCD showing a 1:1 correspondence between MAb B63 (e) and PNA (f)(wide arrows) at the apical membrane. (g and h) Adult: double staining of cells along mid CCD showing a 1:1 correspondence between MAb B63 (g) and PNA (h) (wide arrows) at the apical membrane.

Full size image

Three weeks old and adult. A 1:1 correspondence of apical B63 and PNA staining patterns along the CCD was almost always observed(Fig. 8,g and h, wide arrows, respectively), as shown previously for the mature CCD(16).

DISCUSSION

Two major types of cells have been previously identified in the rabbit collecting duct, intercalated cells and principal cells. The majority, principal cells, are primarily involved in Na+, K+, and water transport. Intercalated cells transport H+/HCO-3 and can be further separated into A- or α-type (which have apical H+-ATPase and basolateral band 3-like anion exchangers and secrete protons) and B- or β-type (which have apical anion exchangers and PNA binding sites and secrete HCO-3)(12, 14, 15, 17, 21, 38, 39). There are also γ-types which have anion exchangers on both apical and basolateral membranes and bind PNA but have not been further characterized immunocytochemically(20). Finally, there are also other “hybrid” types which have features of one or more subtypes of intercalated cells(14, 17, 21). In the rabbit it is not clear whether antibodies to H+-ATPase can clearly distinguish A from B intercalated cells(17, 23). On the other hand, it is likely that HCO-3-secreting cells in the rabbit kidney cortex can be identified by binding to PNA and MAb B63, with negative staining for band 3(14, 16, 17, 24).

Some of the confusion regarding the subtypes of intercalated cells in the rabbit kidney has been recently clarified by Weiner et al.(40). Using fluorescent intracellular pH measurements and Cl- substitutions, they showed that all intercalated cells possessed basolateral Cl-/base exchangers but only that of the A cell was inhibitable by the anion exchange inhibitor 4,4′-diisothiocyanatostilbene-2,2-disulfonic acid. Under experimental conditions which inactivated the apical anion exchanger, these investigators were able to show that all B-intercalated cells had, in addition to apical Cl-/base exchangers, basolateral Cl-/base exchange activity, as well. Thus, they concluded that there were only two major intercalated cell subtypes, A and B; it was not necessary to invoke a γ-type to accommodate their experimental results(40).

To characterize the maturation of the major cell types in the collecting duct, we compared kidney sections from newborn (1 wk old), 3-wk-old, and adult rabbits using a panel of MAb (to H+-ATPase, band 3 anion exchanger, B-intercalated cells, and principal cells) and PNA, a lectin that identifies B- and γ-intercalated cells in the rabbit kidney(12, 15, 17, 20, 21).

In a previous study in developing rat CCD Holthofer(41) showed that carbonic anhydrase immunoreactivity characteristic of intercalated cells appeared soon after birth and the number of positive cells, and their intensity of staining increased during the first postnatal month. Band 3 glycoprotein was not apparent at the basolateral membrane until after 5 d of age and then gradually increased, but by 15 d there were still few band 3-positive cells. Adult numbers of band 3-positive cells were not observed until 30 d postnatally.

In a more recent study Kim et al.(42) observed carbonic anhydrase II and H+-ATPase staining in subpopulations of cells in the cortex as well as in the medulla of 18-d-old fetal rats. Carbonic anhydrase II and H+-ATPase staining was found primarily in connecting tubules, medullary collecting ducts, and papillary surface epithelium, but CCD were not labeled. Staining of inner CCD was weak and outer CCD did not become labeled until 1-2 wk postnatally. Band 3 staining became evident in the 3-d-old rat kidney, primarily in the connecting tubule and OMCD.

In the present study in rabbits we observed classical, as well as nonpolar, staining patterns for both H+-ATPase and band 3 in the mid to deep cortex of the neonatal (1 wk or 1-7 d old) kidney. Both apical and cytoplasmic H+-ATPase patterns and both basolateral and cytoplasmic band 3 patterns were observed in the CCD and OMCD of newborn rabbits. However, the intensity of staining and number of labeled cells in both CCD and OMCD increased with maturation. Regarding band 3, there was no apical accentuation of nonpolar staining. We cannot rule out an artifact of sectioning or subcellular localization of band 3 to account for this appearance.

It is possible that these MAb are labeling primarily intracellular vesicular H+-ATPase and band 3-related proteins(17, 36, 43, 44). It is not clear from these observations as to whether or not a subpopulation of such nonpolar cells ultimately develops a specific polarity or gives rise to daughter cells that are polarized. But the observations in more mature kidneys indicate slightly smaller percentages of cells with nonpolar staining with these antibodies, consistent with the development of increasing polarity in a population of intercalated cells.

In the superficial cortex of the neonate there was no immunocytochemical evidence of intercalated-like cells in the CCD; no staining was observed with antibodies to H+-ATPase, anion exchanger, or to the surface of B-intercalated cells. These findings agree with ultrastructural, immunohistochemical, and fluorescent studies published previously which failed to identify intercalated cells in the developing neonatal CCD(14, 28, 42, 45) and pH-sensitive fluorescent staining techniques which failed to show selective uptake by individual intercalated cells of the immature CCD(27). Furthermore, using PNA in isolated perfused CCD, we previously showed an increase in number of labeled cells per mm of CCD as well as an increase in length of apical profiles with maturation(14). On the other hand, PNA staining of ampullary structures in the nephrogenic zone has been noted previously in the newborn rabbit by Minuth et al.(37). These findings indicate that with maturation of the CCD many (principal and A-intercalated) cells lose galactose residues on their surface glycoproteins, whereas such residues are reacquired and become prominent on B- and γ-intercalated cells.

The finding by Minuth et al. of PNA and wheat germ agglutinin co-labeling in the ampullary neck indicates that principal and intercalated cell markers might be expressed within a single cell type(37, 46). They have suggested that intercalated cells might derive from principal cells or from a common precursor cell. We previously had speculated that intercalated and principal cells in the CCD were derived from a poorly differentiated cell in the outer cortex(28). Kim et al.(42) recently suggested that intercalated cells might be derived from undifferentiated cells both in the nephrogenic blastema as well as in the ureteric bud.

These conclusions differ from those presented by Narbaitz et al.(29), who showed in newborn rats that A-type intercalated cells appear before B cells in the CCD, although this was not confirmed in a subsequent study(30). Also, they differ from thein vitro findings by one of us in cultured rabbit collecting duct cells that suggest that B-intercalated cells can give rise to both A-intercalated cells and principal cells in vitro, whereas principal cells cannot convert to intercalated cells(47).

Kim et al.(42) has shown that B intercalated cells seem to appear at the same time as A cells in the developing rat kidney, suggesting that the two types of intercalated cells might be derived from undifferentiated precursors, rather than from each other. We have shown in the rabbit that the intensity of intercalated cell labeling increased in the OMCD and mid/inner CCD with maturation, but new labeling appeared in the outer CCD by 3 wk postnatally (see Table 1 and below). These findings indicate that the postnatal increase in number and intensity of staining of outer CCD intercalated cells would result primarily in differentiation of B cells with the development of net transepithelial HCO-3 secretion(6).

With a significant number of differentiated A cells in the OMCD, the newborn is capable of transepithelial net H+ secretion(6). During maturation there is only a small increase in this flux(6), suggesting that the A cell is relatively more differentiated than the B cell at birth. We have previously shown that the CCD of the newborn rabbit, unlike that of the adult rabbit, fails to secrete HCO-3 in vitro and, indeed, may absorb HCO-3(6). In support of the observed flux data are the findings above of significant numbers of A-intercalated (band 3-positive or apically stained H+-ATPase positive/PNA negative) cells in neonatal CCD. Moreover, OMCDs from newborn rabbits absorb HCO-3 in vitro at ≈70% of the rate of the mature animal(6), indicating that high rates of HCO-3 absorption in the OMCD plus little or no HCO-3 secretion in the CCD both contribute to an observed hypochloremic metabolic alkalosis of the newborn rabbit(6). These data plus the findings described herein might also suggest that A-intercalated cells are more differentiated than B cells in the newborn rabbit kidney.

One result that was unexpected was our finding of individual cells in or near the S- and comma-shaped vesicles staining diffusely or circumferentially with PNA and MAb B63. How these cells, which were not found in structures with lumens, relate to the development of intercalated cells in the CCD and connecting tubule is not clear at the present time. The finding of these cells with cytoplasmic staining for PNA and MAb B63 was not considered an artifact because of the more polar staining observed deeper in the cortex in CCD in the same sections, and because of the relative frequency of the occurrence (≈1 cell per section). It is possible that the PNA and B63 epitopes appear on the apical membrane only after the cells' attachment to the basement membrane of the CCD and establishment of apical and basolateral poles. Perhaps cellular interconversion(12, 47) or possibly asymmetrical cell division(28, 48) could give rise to the different cell types of the mature mammalian CCD. In keeping with the foregoing models of renal development(29, 47), a common precursor cell could mediate the development of the cellular heterogeneity of the CCD. Further studies are needed to address this important question in developmental biology.

Deeper in the cortex, those cells with the intercalated cell phenotype have been shown previously to be functionally immature as indicated by the lack of acidic vesicles in the cytoplasm, lower mitochondrial potential, a lower cell pH, less apical Cl--HCO-3 exchange, fewer cells concentrating pH sensitive dyes, shorter apical binding profiles for PNA, and failure to secrete net HCO-3 in the newborn period(6, 14, 27, 45). In other studies, thicker sections were needed to show comparable intensities of PNA and B63 staining compared with mature kidney(14). We found deeper in the neonatal cortex that the staining with each of the antibodies and with PNA was much less intense than observed in older animals and generally fewer cells were labeled. Furthermore, there were more nonpolar patterns in the immature kidney CCD and OMCD. These findings would predict less than mature rates of H+/HCO-3 transport in the neonatal segments(6).

We have previously shown that the OMCD of the newborn rabbit is relatively mature, as seen using fluorescent dyes probing acidic cytoplasmic vesicles, mitochondrial potential, and cell pH(14, 27). However, the OMCD of the newborn showed slightly fewer polar patterns of H+-ATPase (78%) and band 3 (31%) and fewer band 3 labeled cells (22%) than did more mature OMCD (≈90, ≈90, and 37%, respectively). These observations would be consistent with our previous ultrastructural findings showing intercalated cells of the outer medulla to have shorter apical perimeters, fewer vesicular profiles, and smaller mitochondrial volume percents than mature outer medullary intercalated cells(28). One could thereby account for a somewhat lower rate of HCO-3 absorption in the neonatal OMCD compared with the mature segment(6).

Our studies go further to indicate that A-intercalated cells appear early in development in the IMCD, as has been demonstrated previously by ultrastructure in the rat(29). We found H+-ATPase and band 3 labeling in the initial inner medulla of the newborn. With maturation, intercalated cells were still occasionally found in the initial inner medulla. Whether these cells might be eliminated by apoptosis or programmed cell death, as suggested by Kim et al.(42) remains to be determined.

The findings of cells staining with apical H+-ATPase antibodies and PNA would confirm the existence of hybrid intercalated cells in both immature and adult rabbit kidneys(17). It is generally believed that A-intercalated cells show apical H+-ATPase, basolateral band 3 and no PNA labeling, whereas B-intercalated cells in the rabbit show diffuse H+-ATPase and apical PNA labeling(17, 24). It is possible that the resolution at the fluorescence level cannot provide the sensitivity to characterize all types of intercalated cells at the membrane level(23, 28, 36) or from the aspect of vectorial H+/HCO-3 transport(14, 20, 40). Recent ultrastructural studies by Verlander et al.(23) identify three patterns of H+-ATPase localization in the rabbit CCD. Rarest were apical patterns such as seen for rabbit OMCD as well as for A intercalated cells of rat. A somewhat larger group of cells stained over the basolateral membrane and cytoplasmic vesicles throughout the cell, as has been described for B intercalated cells of the rat. But the great majority of rabbit CCD intercalated cells showed H+-ATPase staining over cytoplasmic vesicles and not along any membrane, as found in the present study. Although many of these cells would likely be positive for PNA labeling, the empirical marker for HCO-3-secreting intercalated cells of the rabbit CCD(17, 23), it is not clear as to whether they are functionally active or quiescent.

The findings of nonpolarized H+-ATPase staining, as well as“hybrid” intercalated cells (PNA/band 3 positive, band 3/diffuse H+-ATPase, and apical H+-ATPase/PNA positive) in both immature and mature kidneys suggest that they could be transitional or quiescent cells that might be activated in response to acid-base perturbations. Indeed, maternal metabolic alkalosis is associated with an increase in percentage of B-type intercalated cells in the CCD of newborn rats(30), indicating that initial differentiation of intercalated cells is responsive to maternal acid-base disturbances. Moreover, the mature kidney may show substantial intercalated cell remodeling in response to metabolic acidosis(11, 12, 49). Further studies are necessary to address this issue in regard to development.

Data on the maturing principal cells are quite limited. Previous studies have shown that Na,K-ATPase, the abundant enzyme of principal cells, begins to polarize to the basolateral membranes hours after birth(41). Some (presumably intercalated) cells become devoid of Na,K-ATPase reactivity by the first postnatal week. Minuth et al.(50) noted that the collecting duct ampulla had no immunoreactive staining with an antibody to the α-subunit of Na,K-ATPase, whereas there was both apical and basolateral staining of the proximal portion of the developing CCD. Progressing toward the papilla, mature OMCDs and IMCDs were strongly labeled on the basolateral membrane, as expected for a differentiated epithelial cell.

Using the principal cell antibody MAb ST.9, we showed little, but primarily apical, staining of the superficial CCD. Deeper in the cortex, the staining intensity was greater over the apical membrane and apical vesicles. Apical and basolateral labeling such as observed previously for Na,K-ATPase by Minuthet al.(50), was not detectable in these sections. Intense staining was observed in the OMCD, as had been demonstrated for Na,K-ATPase by Minuth et al.(50); however there was no staining in the terminal IMCD at any age. Additional studies with other principal cell antibodies are needed to better characterize this developmental process.

In summary, we have shown in the developing rabbit kidney that the acquisition of differentiated markers takes place during the first 3 wk of postnatal life. With differentiation and maturation we observed more numbers of intercalated cells and cells with stronger labeling to H+-ATPase, band 3, B-intercalated cell and principal cell antibodies and to PNA. The A intercalated cells appeared to display a marker phenotype more closely resembling the adult pattern than B cells early in life. There was little intercalated cell labeling in the maturing terminal medulla. A surprising finding was the detection of cytoplasmic patterns of staining of PNA/B63-positive cells in and near the S- and comma-shaped vesicles of the developing nephron. This result requires confirmation and additional study.