Aestivation is a state of dormancy or torpor that has developed as a highly specialized adaptation for survival in areas subjected to extreme climatic conditions. The Dipnoi (lungfish) experience the process of aestivation through the dry seasons that characterize tropical life. Lungfish aestivation constitutes a unique process because it represents a drastic change from the aquatic (gill- and air-breathing) to the terrestrial (only air-breathing) habitat (Fishman et al.,1986). Thisfunctional adaptation involves dramatic modifications in hydro-saline control and acid–base balance (DeLaney et al.,1977). During aestivation, the lungfish need to adapt from a situation in which water intake and excretion are the dominant processes, to another state in which water intake is suppressed, and the production of urine decreases dramatically or ceases (DeLaney et al.,1977). Furthermore, the exclusive air-breathing implies a loss of water during respiration. In addition, the cessation of the gill function represents the loss of an important organ in osmoregulation (Greenwell et al.,2003).
The study of the adaptative modifications occurring during lungfish aestivation has mostly focused on changes in general metabolism, and on several aspects of the cardiac and osmoregulatory functions (Janssens,1964; DeLaney et al.,1977; Mesquita-Saad et al.,2002). However, very few studies have focused on the structural modifications that should accompany those functional changes (Sturla et al.,2002). To our knowledge, none has been devoted to the aestivating kidney.
The study of the structural changes that should occur in the lungfish kidney during aestivation is very important. The suppression of urine production must involve profound structural modifications throughout the entire nephron. The idea that these structural changes may be comparable to those which may occur in the vertebrate kidney during the lack of water intake, dehydration, or even renal failure, is appealing. In addition, the lungfish may face cyclic environmental changes that may be similar to those faced by vertebrates in the course of evolution (Vize,2004). Indeed, the evolution of the kidney is essentially the story of the evolution of regulation of the body water content (Smith,1943).
Within the kidney tissue, the renal corpuscle (RC) is the site of primary urine formation. The RC filtration barrier consists of the capillary wall, the glomerular basement membrane (GBM), mesangial cells, and podocytes. The GBM is a charged sieve composed by collagen and noncollagenous glycoconjugates whose permeability depends largely on two factors: the pore size, determined by the nature and organization of its macromolecular components, and the fixed electrostatic charge (Tryggvason and Wartiovaara,2001; Levidiotis and Power,2005). Negatively charged molecules are found in all the layers of the filtration barrier (Latta,1980), being mostly due to the carbohydrate moities of the glycoconjugates. These moities can easily be detected in tissues by means of lectins (Goldstein and Poretz,1986), which serve as markers of several physiological and pathological states (Rademacher et al.,1988). Furthermore, the filtration barrier exhibits lectin-binding patterns that are species-specific (Holthöfer,1983; Laitinen et al.,1989; Ojeda et al.,1993,2003; Ramos-Vara et al.,2004). In addition, it has been suggested that podocytes form an integrated system able to change the properties of the filtration barrier by modifying the spatial arrangement of the foot processes and the synthesis of the BM components (Ojeda et al.,2006). The mesangial cells have also been implicated in the regulation of glomerular filtration (Michael et al.,1970) and in the adaptation to changes in the hydrosaline environment (Hickman,1968).
The aim of the present study is to examine the structural and lectin-binding modifications that occur in the RC of the African lungfish during the process of aestivation. In addition, we have studied the reversal of these modifications during the adaptation to the aquatic environment and the recovery of the body activity.
MATERIALS AND METHODS
The study was performed on 42 specimens of the African lungfish Protopterus dolloi of both sexes, weighing 100–150 g of body mass. The specimens were collected from Central Africa and imported through a local fish farm in Singapore. Identification of the specimens was performed according to Poll (1961). The fish were maintained in plastic aquaria filled with dechlorinated water, containing 0.71 mM Na+, 0.32 mM K+, 0.72 mM Ca++, 0.06 mM Mg++, 2.2 mM Cl−, and 0.2 mM HCO3, at pH 7.0, and at 25°C in the laboratory. Water was changed daily. The specimens were acclimatized to laboratory conditions for at least 1 month. During the adaptation period, the fish were fed midge larvae (Bio-Pure Blood Worm, Hikari Sales USA, CA). A total of eight animals, maintained as indicated above, were used as controls.
Forty-eight hours before the beginning of the experiments, food was withdrawn, giving sufficient time for the gut to be emptied of all food and waste. A first group of 20 animals was induced to aestivate in their aquaria as described by Chew et al. (2004). Specimens were allowed to enter into a state of aestivation individually, in plastic tanks containing a thin film of 10 mL of dechlorinated tap water. The water would dry up in approximately 3–4 days, and the specimens would enter a state of torpor in a layer of dried mucus at day 4–5. For those that were allowed to aestivate for long periods, 1–2 mL of water was sprayed on the surface of the cocoon every 6 days. The fish were killed after 4, 6, 12, 60, and 180 days of aestivation. Four fish were used for each period.
Return to Aquatic Conditions After Aestivation
An additional group of 14 fish were re-immersed in water 180 days after the beginning of aestivation. These fish were killed at 1, 3, and 6 days after being returned to the water. At least four fish were used for each period.
In all cases, the fish were killed by a blow to the head and the ventral body wall was opened. The long columnar kidneys were exposed and the central portion was divided into four parts of approximately 0.5 cm in length (Ojeda et al.,2006).
For conventional light microscopy, small kidney fragments were fixed in 3% glutaraldehyde, post-fixed in 1% osmium tetroxide for 1 hr, dehydrated in graded acetone and propylene oxide, and embedded in Araldite (Fluka, Buchs, Switzerland). Semithin (1 μm) sections were cut with a Leica ultracut UCT, stained with 1% toluidine blue, and observed with a Zeiss III photomicroscope.
Transmission Electron Microscopy (TEM)
For TEM, kidney fragments were fixed in 3% glutaraldehyde in phosphate-buffered saline (PBS), post-fixed in 1% osmium tetroxide for 1 hr, and embedded in Araldite as above. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with a Philips EM 208.
Scanning Electron Microscopy (SEM)
For SEM, selected kidney fragments were fixed in 3% glutaraldehyde. After fixation, a first group of samples was dehydrated in graded acetone, dried by the critical point method, coated with gold, and observed with a Philips SEM 501. A second group of samples was subjected to chemical and mechanical microdissection before desiccation. For chemical microdissection, the specimens were digested either with NaOH (Takahashi-Iwanaga,1992) or KOH (Ushiki and Murakumo,1991). After chemical digestion, the specimens were carefully microdissected under a stereomicroscope with sharpened tungsten needles. The specimens were then processed for SEM as described above.
Kidney slices were immersed in cold ethanol-glacial acetic acid (99:1) (Sainte-Marie,1962) for 1 hr, dehydrated in cold ethanol, transferred to cold xylene, and embedded in diethylene glycol distearate (DGD; Polysciences, Warrington, PA), as previously described (Ojeda et al.,1989). Briefly, the kidney fragments were transferred from the xylene to a 2:1 mixture of xylene/DGD, and then to a 1:2 mixture (10 min each), followed by two changes of 100% DGD (15 min each). During the embedding procedure, the DGD was maintained at 60°C to keep it molten. After infiltration, the kidney fragments were transferred to a flat silicone embedding mould filled with freshly melted DGD. After solidification at room temperature, the DGD blocks were stored at 4°C until sectioning.
Sections were cut at 2 μm by means of a Jeol Jum-7 ultramicrotome with glass knives and a water bath. The sections were removed from the water with a glass rod and transferred to a water drop on a clean slide. The slides were placed on a warm plate (40°C) until dry, and stored at 4°C overnight.
Dewaxed sections were treated with fluorescein isothiocyanate (FITC) -conjugated lectins, 50 μgr/mL in PBS, of different nominal sugar specificities (Table 1), purchased from Vector (Vector, Burlingame, CA) and Sigma (Sigma, St. Louis, MO). All the staining procedures were carried out in the dark at room temperature for 30 min, followed by three washes of 10 min in PBS. The slides were mounted with the antifading medium Vectashield (Vector).
Table 1. Nominal saccharide specificities of thelectins used in this studya
Nominal saccharide specificity
See (Damjanov,1987; Goldstein and Poretz,1986; Sarkar et al.,1981).
The specificity of the lectin binding was tested by preincubation of the lectin conjugate with 0.2 M solutions of the nominal specific sugar (Sigma; Sarkar et al.,1981; Goldstein and Poretz,1986; Damjanov,1987). All the controls were routinely negative.
Selected tissue sections were treated with neuraminidase before lectin staining. Neuraminidase from Clostridium perfringens (Type V, Sigma) was used at concentration of 1U/mL, diluted in acetate buffer for 4 hr at 37°C. This procedure exposes the penultime carbohydrate residues blocked by sialic acid (Uehara et al.1985).
Selected sections were counterstained for 10 min with propidium iodide (1 μg/mL) at room temperature. This fluorochrome selectively marks the nucleic acids (Ockleford et al.,1981), allows to evaluate nuclear changes, and improves the histologic information. The fluorescence studies were performed using a confocal laser microscope (MRC 1024; Bio-Rad) using argon (488nm) and HeNe (543 nm) lasers.
Periodic Acid-Schiff (PAS) Staining
Kidney sections of recovering fishes were PAS stained using a commercial kit (Sigma). The sections were processed according to the staining standard kit protocol and were lightly counterstained with Gill N° 3 hematoxylin (Sigma).
Structural Modifications During Aestivation
The kidneys of the aestivating fish showed several gross morphological changes. A characteristic feature was the dilatation of all the tubular segments (Fig. 1). This was remarkable at the level of the proximal tubule, being evident even to the naked eye. The tubular dilatation obscured the presence of the renal corpuscles, which were difficult to discern. The peritubular capillaries and the vessels of the venous portal system were also affected, showing a narrowed lumen and a small number of blood cells. Thus, aestivation had affected all the structural components of the kidney. We report here the changes that took place in the RC.
A characteristic feature of the P. dolloi kidney is the grouping of the RCs into small clusters (Ojeda et al.,2006). All the RC clusters were affected. Most notably, the RCs showed a marked reduction in size (Fig. 2). The RC collapse was very pronounced in the glomerular tuft, especially at the level of the capillary lumen. Another characteristic feature was the dilatation of the urinary space (Fig. 2). The neck segment was also dilated (Fig. 3), exhibiting a tortuous course that contrasted with the straight course shown by this segment in freshwater fish (compare Figures 3a and 3b). All the RC changes could be observed after the 4th day of aestivation.
Bowman's capsule showed numerous structural modifications (Figs. 4, 5). In freshwater fish, this layer was formed by squamous epithelial cells that exhibited a centrally located nucleus and a solitary cilium (Figs. 4a, 5a). In the aestivating animals, the parietal cells lost their flattened appearance, exhibiting a cuboidal or cylindrical shape and bulging into the urinary space (Fig. 4b). In addition, Bowman's capsule contained a low number of multiciliated cells, similar to those of the neck segment (Fig. 4b). However, the ciliary tuft was less compact and the cilia appeared more irregular (Fig. 4b,c) than those of the normal ciliated cells of the neck segment (see, Ojeda et al.,2006). At 180 days of aestivation, numerous collagen fibrils appeared in the lamina reticularis of the parietal basement membrane (Fig. 5b). Inside the same RC, the extent to which Bowman's capsule was affected varied, and the above-mentioned changes did not occur with the same degree of intensity throughout the capsule. Occasionally, dead cells could be observed in Bowman's capsule, or free in the urinary space (Fig. 4c).
The collapse of the RC was accompanied by the approximation of the podocyte bodies (Fig. 6). In freshwater fish, the podocytes were separated from each other by the long major processes (Fig. 6a). During aestivation, the podocyte bodies approached each other, appearing densely packed (compare Figures 6a and 6b), and establishing numerous focal contacts. In the area of contact, the membrane boundaries became difficult to discern (Fig. 6c) or, even, disappeared at focal points (Fig. 6d,e). Furthermore, after the 12th day of aestivation, approximately 4% of the podocytes were binucleated (Fig. 6f). Occasionally, the podocytes exhibited a nucleus that was up to 2–3 times as large as the nucleus of other podocytes (Fig. 6g). Evidence of nuclear fusion was never observed. In addition, the nuclei of the aestivating podocytes showed an increase in the amount of heterochromatin, and a predominance of the pale fibrillar component in the nucleoli.
The approximation of the podocyte bodies occurred with the progressive loss of the major processes interposed between contiguous cells (Fig. 7a). The remaining major processes broadened and flattened (Fig. 7b), the surface membrane exhibiting numerous blebs. These changes were accompanied by the appearance of abundant lysosomes throughout the podocyte cytoplasm, and by modifications in the distribution of the intermediate filaments (IF; Fig. 8). In freshwater fish, the major processes showed abundant IF bundles (Fig. 8a). However, at the beginning of aestivation, the IF bundles appeared disorganized or were lost in the major processes, appearing well developed in the podocyte body (Fig. 8b). It should be underscored that the main cytoplasm of the podocytes never showed IF bundles in freshwater fish (see Ojeda et al.,2006).
The structural changes experienced by the podocytes also affected the foot processes. Under SEM, the interdigitating foot processes showed a highly regular pattern in freshwater fish (Fig. 9a). During aestivation, the regular pattern was lost, the foot processes varied in size, and the number of filtration slits decreased (Fig. 9b). TEM illustrated additional changes. Contrary to the normal appearance of the filtration barrier (Fig. 9c), the beginning of aestivation was characterized by the presence of foot processes that exhibited finger-like prolongations directed toward the GBM. This gave the GBM a festooned appearance (Fig. 9d). Many foot processes were enlarged (Fig. 10) and established membrane junctions. The filtration slits frequently lacked the slit diaphragms. In addition, the subpodocyte space was lost after the sixth day of aestivation. Consequently, the podocyte bodies and the major processes were directly attached to the GBM (Fig. 6f).
As previously reported (Ojeda et al.,2006), the GBM of P. dolloi is formed by three layers: a lamina rara, a lamina densa, and a subendothelial lamina (Fig. 9c). During aestivation, the GBM increased greatly in thickness (Fig. 9d) and exhibited severe structural modifications. At the beginning of aestivation (first 6 days), the changes mainly affected the subendothelial lamina (Fig. 9d). This lamina showed a large increase in the amount of microfibrils and amorphous material. This occurred concomitantly with the reduction in size or the subendothelial space loss (Figs. 9d, 10a). At 60 days of aestivation the thickness of the GBM had increased up to four times. The most notable feature was the presence of numerous round inclusions (Fig. 10b) consisting of moderately electron-dense granular material (Fig. 10b,c). Numerous coiled, nonstriated fibrils 18–23 nm long and 10–14 nm thick were embedded in the granular component (Fig. 10c). These inclusions were observed on the 6th day of aestivation (Fig. 10a), but were much more numerous at 60 days (Fig. 10b). The inclusions underlay the lamina densa of the GBM, reaching a diameter of up to 800–1,300 nm. Collagen fibers were also present in the GBM, many of them in association with the mesangial cells. Curiously, the round extracellular inclusions almost disappeared at 180 days of aestivation (Fig. 10d). In this period, the GBM was characterized by the presence of large deposits of filamentous material and of a large number of collagen fibers. Often, the collagen fibers formed bundles located in extracellular compartments defined by endothelial or mesangial cell processes (Fig. 10d).
The two types of mesangial cells (see Ojeda et al.,2006) also undergo severe modifications during aestivation. The large mesangial cells showed an increase in both the number and thickness of their cytoplasmic processes. The arrangement of these processes was irregular, frequently crossing each other and forming a dense network embedded in the subendothelial lamina of the GBM (Fig. 11a). This contrasted with the regular palisade-like arrangement displayed by this type of mesangial cells (Fig. 11a) in freshwater animals (also, see Ojeda et al.,2006). The second type of mesangial cells, the small stellate cells, showed more numerous, thicker and shorter cell processes during aestivation than in freshwater conditions (Fig. 11b). All these modifications in the morphology of the mesangial cells could be observed after the fourth day of aestivation.
The glomerular capillaries also underwent modifications during aestivation. At the beginning of aestivation, the luminal face of the capillaries lost it normal regular morphology (see Ojeda et al.,2006), the endothelium exhibiting numerous irregularities, elevations, and depressions (Fig. 12a). Of note, mesangial cell prolongations pushed through the endothelium, protruding into the capillary lumen (Fig. 12b). The mesangial protrusions were mostly covered by the endothelial cells (Fig. 12b). These interactions disappeared in the later stages of aestivation. In freshwater fish, the endothelial surface showed numerous membrane fenestrations (see Ojeda et al.,2006). In aestivation, the distribution of these fenestrations could appear fairly normal (Fig. 12a), be very reduced or, even, have disappeared in large areas. In all cases, the endothelial nuclei bulged into the narrow capillary lumen (Fig. 2b).
Within several hours of being returned to water, the lungfish broke the cocoon and surfaced to gulp air, albeit sluggishly. The body activity fully recovered within several days, but the fish would not feed until 10–14 days after arousal. Throughout the recovery period studied, the kidney tissue was infiltrated by mast cells containing the typical electron-dense membrane-bound granules (Fig. 13a). These granules reacted strongly to PAS (not shown). Under SEM, these cells displayed numerous surface blebs and many slender cell processes (Fig. 13b). The mast cells could be observed 24 hours after reimmersion in water, being initially more abundant around the tubular segments (Fig. 13a). In the RC, the mast cells were located first (1 day) within the glomerular capillaries. Later (3–6 days), most of the mast cells appeared within the urinary space (Fig. 13b). Occasionally, these cells exhibited cell processes that penetrated the podocyte barrier and reached the GBM (Fig. 13b).
On the first day of recovery, the lumen of both the peritubular capillaries and the vessels of the venous portal system attained a diameter similar to that of the freshwater animals. However, the lumen of the glomerular capillaries was still very narrow at 6 days (Fig. 14). The dilatation of the urinary space disappeared on the third day. Throughout the recovery period studied, Bowman's capsule showed a structural appearance (Fig. 14) similar to that of the aestivating animals. The most significant change was the presence of desmosome-like junctions between contiguous parietal cells (Fig. 14).
In this period, the podocyte bodies were less prominent and rounded, and appeared less densely packed (Fig. 15). The separation between contiguous cell bodies allowed the appearance, first (3 days), of wide cell prolongations with numerous microfilaments and, later (6days), of short major processes (Fig. 15). Frequently, the podocytes extended long and slender cell processes over the surface of neighboring podocytes. The ends of these processes showed a morphology similar to that of the “growth cone” of the developing axons (Fig. 16). In addition, most of the podocytes exhibited a short, solitary cilium (Fig. 16), and developed desmosomes between major processes. These features were never observed in freshwater animals (see Ojeda et al.,2006). At 6 days, most of the podocytes recovered the subpodocyte space. The foot processes and the filtration slits progressively become more regular, and slit diaphragms could be clearly observed. However, the subendothelial space remained collapsed (Fig. 17).
Contrary to what occurred with the podocytes, the morphological modifications in the GBM occurred rapidly during the recovery period. The remaining extracellular inclusions disappeared the first day, and the number of collagen fibers was clearly reduced at day 3. Furthermore, the collagen fibers were shorter, did not aggregate into bundles, and showed increased electron density (Fig. 17). At 6 days, the thickness of the GBM was considerably reduced, remaining only twice as thick as in freshwater animals. Throughout the recovery period studied the GBM contained abundant filamentous material, and the subendothelial zone remained collapsed (Fig. 17). The mesangial cells exhibited a morphology similar to that observed in the aestivating fish, and the endothelium of the glomerular capillaries showed numerous surface irregularities similar to those observed at the beginning of aestivation.
Figure 18 shows the distribution of the lectin-binding sites in the different components of the renal corpuscle of P. dolloi in freshwater (Fig. 18a–h) and during aestivation (Fig. 18i–p). In freshwater fish, the podocyte coat reacted faintly to ConA, LEA, WGA and WGAs. The GBM was moderately positive for ConA, LEA, WGA, WGAs, and RCA-I, and faintly positive for MPA and UEA-I. In both cases, the intensity of WGA and WGAs staining was similar, indicating that reactivity to WGA was exclusively due to N-acetyl-D-glucosamine. The endothelial cell coat was faintly positive for ConA and LEA. The parietal basement membrane was moderately positive for ConA and LEA, while the parietal cell coat was only faintly positive for ConA. Curiously, the nucleus of all RC cellular components was faintly positive for MPA and UEA-I. The nuclear fluorescence was located around the heterochromatin and in the interchromatin zones (the heterochromatin domain), and in close relationship with the nuclear membrane. Mesangial cells were only positive for RCA-I. All RC components were negative for SBA and PNA. The use of neuraminidase indicated the absence of cryptic PNA-binding sites. These results are summarized in Table 2.
In aestivating fish, all RC components showed quantitative and/or qualitative modifications in the lectin-binding pattern (Fig. 18i–p). The main modifications affected the podocytes and the GBM. Podocytes lost their positivity for Con A, WGA and WGAs, and displayed increased LEA-positivity. The GBM exhibited increased binding to ConA, LEA, WGA, and MPA, decreased binding to WGAs, and lost its affinity for UEA-I and RCA-I. The differences in staining intensity between WGA and WGAs revealed the appearance of N-acetyl neuraminic acid (sialic acid) in the GBM during aestivation.
The endothelial cell coat lost its LEA reactivity. The parietal cells showed increased ConA positivity both in the cell coat and in the basement membrane. Also, the parietal cell coat became MPA- and LEA-positive, and the parietal basement membrane lost its LEA affinity. The mesangial cells lost their positivity for RCA-I. The nucleus of all the RC cells lost the MPA positivity, but maintained the positivity for UEA-I (see Table 2). All the modifications in the lectin-binding pattern occurred at day 6 of aestivation and were maintained thereafter. The only exception was the loss of affinity for RCA-I in mesangial cells, which was observed at 60 days of aestivation.
The lectin-binding pattern observed during aestivation was not modified through the recovery period studied. As an exception, the reactivity of the GBM to MPA decreased in this period to the level observed in freshwater fish.
In the past 50 years, the emphasis on aestivating African lungfish (for review, see Fishman et al.,1986) was on respiration (DeLaney et al.,1974; Perry et al.,2007) and on nitrogen metabolism (Janssens and Cohen,1968a,b; Chew et al.,2003,2004; Wood et al.,2005; Ip et al.,2005a,b; Loong et al.,2005,2008), and, marginally, on metabolic rate reduction (Smith,1930,1935; Janssens,1964; Perry et al.,2007; Loong et al.,2008). However, the possible occurrence of organic structural modifications during the aestivation period has largely been neglected. Our results demonstrate that all the components of the renal corpuscle of P. dolloi show structural modifications during aestivation. Many of these modifications are reversed during the recovery period studied.
The first noticeable feature of the aestivating kidney is the collapse of the glomerular tuft. Reduction in the glomerular tuft size has been described in mammals subjected to dehydration (Racusen et al.,1984), after angiotensin II or vasopressin administration (Hornych et al.,1972; Racusen et al.,1984), and in freshwater fish adapted to the saline environment (Wendelaar Bonga,1973; Elger and Hentschel,1981; Brown et al.,1983; Gray and Brown,1987). The link common to all these situations is the need to retain water to support the hydro-saline balance. Several mechanisms, acting singly or in combination, can be invoked to explain the decrease in capillary diameter which occurs during lungfish aestivation. One possibility is that it takes place as a consequence of decreased arterial pressure (see Kriz et al.,1994; Ohno et al.,1996). In fact, the blood pressure of P. aethiopicus drops to two thirds of the control values during the first thirty days of aestivation (Delaney et al.,1974). Another possibility is that it is an active mechanism due to arteriolar vasoconstriction. However, renin secretion is unlikely. Most fish classes, including the lungfish (Ojeda et al.,2006), lack a demonstrable juxtaglomerular apparatus. Active contraction of the endothelium and/or of the mesangial cells may be another mechanism. For instance, rat glomerular cells can contract in vitro in response to arginine vasopressin (Ausiello et al.,1980). Although the P. dolloi kidney is responsible for vasopressin (Sawyer et al.,1982), the structural basis of this response are unknown. The down-regulation of nitric oxide (NO) activity may also be implicated. A glomerular decrease in the activity of the endothelial form of nitric oxide synthase during P. dolloi aestivation has recently been reported (Amelio et al.,2008). This situation is similar to that observed during squirrel hibernation (Sandovici et al.,2004).
Another characteristic feature of the renal corpuscle during aestivation is the loss of the flattened appearance of the parietal cells forming Bowman's capsule. The modified lectin-binding pattern in both the parietal cells and the parietal basement membrane suggests that the morphological changes were accompanied by physiological and/or differentiation (phenotype) changes (see: Lis and Sharon,1986; Damjanov,1987). However, the nature of these modifications remains obscure.
The modifications experienced by the podocytes during aestivation constitute a subject of maximum interest. The approximation of the podocyte bodies, the broadening and flattening of the major processes, and the loss of podocyte processes have also been reported in freshwater fish adapted to seawater (Elger and Henstchel,1981; Brown et al.,1983; Gray and Brown,1987), in hibernating animals (Zuasti et al.,1986; Anderson et al.,1990), in dehydrated rabbits (Racusen et al.,1984), and in early postischemic acute failure in rabbits and man (Solez et al.,1981). The loss of podocyte processes appears to be one of main podocyte responses to different physiological and pathological situations (Kerjaschki,1994). The net result is that the podocytes become cuboidal and come to rest directly on the GBM, thus reducing the area of filtration (Salmon et al.,2007). The structural modifications of the podocytes observed here could be attributed to a combination of several factors. The first is the reduction in capillary diameter, which subsequently reduces the mechanical force on the podocytes, inducing reorganization of the cytoskeleton (Coers et al.,1996; Endlich et al.,2001; Petermann et al.,2005). The modifications in GBM composition and structure could also explain the podocyte changes. Podocyte morphology is highly controlled through close interactions between the podocyte surface and the matrix molecules of the GBM (Pavenstädt et al.,2003). These interactions are mediated by transmembrane matrix receptors, which trigger several intracellular signals responsible, among other things, for the organization of the cytoskeleton (Miner and Li,2000; Kobayashi et al.,2001; Pavenstädt et al.,2003). Cytoskeletal organization appears to be one of the main forces necessary to maintain the structure of the podocyte processes (Pavenstädt et al.,2003). As occurs in other vertebrates (see Pavenstädt et al.,2003), the modifications to intermediate filament organization observed in P. dolloi during aestivation could be responsible for the loss of podocyte processes.
There also appears to be a close relationship between podocyte shape and their negative surface charge. Podocytes are endowed with a cell coat that contains glycoproteins with highly anionically charged sugar side chains (Michael et al.,1970; Kerjaschki,1994). In freshwater P. dolloi, the podocyte coat only shows binding sites for lectins of the N-acetylglucosamine group (LEA, WGA, and WGAs), and for ConA. This indicates that, as in the sturgeon RC (Ojeda et al.,2003), the podocyte coat contains N-acetylglucosamine and lacks sialic (N-acetyl neuraminic) acid. During aestivation there is a profound change in the podocyte coat because positivity for Con A, WGA, and WGAs is lost and LEA-affinity is increased. The net loss of negative surface charges correlates with the cell shape changes. The podocyte body becomes cuboidal, and the cell processes retract, when the sialic acid residues and the sulfate groups are reduced (Kerjaschki et al.,1985), or neutralized (Seiler et al.,1977; Andrews,1981).
The presence of nuclear glycoproteins positive for MPA and UEA-I is another interesting feature. This, to our knowledge, has not previously been reported in fish. There is some experimental evidence indicating that lectin-positive glycoproteins intervene in the processes of transcription (Snow and Hart,1998; Hubert et al.,1989). However, the precise role of these proteins in the podocytes of P. dolloi awaits further investigation.
Another interesting observation is the focal loss of the cell membrane between contiguous podocytes. Our TEM and SEM findings are similar to those observed during the fusion of mammalian oocytes (Fulka et al.,1989). The presence of binucleated and of giant podocytes adds further support to the notion that during the aestivation period, spontaneous fusion of some podocytes takes place. Spontaneous cell fusion occurs in vivo as a normal process during development, and also in adult tissues (for review, see: Ogle et al.,2005; Alvarez-Dolado,2007). To our knowledge, podocyte fusion has not previously been reported. The biological implications of this process are unknown. However, cell fusion has been implicated in the repair of damaged tissues (Gussoni et al.,2002), and in reprogramming cells into a more pluripotent status (transdetermination; Terada at al.,2002; Ying et al.,2002).
During aestivation, the thickness of the GBM increases greatly. Thickening of the GBM has also been described in hibernating amphibians (Fenoglio et al.,1996) and in freshwater fish adapted to seawater (Elger and Hentschel,1981; Gray and Brown,1987). Thus, it may be a general response aimed at reducing glomerular filtration rate and coping with dehydration. The thickening of the GBM is accompanied by modifications in the lectin-binding pattern. Thus, there is increased affinity for ConA, MPA, and LEA, indicating increased deposition of mucopolysaccharides and/or glycoproteins. This is in agreement with the increase in the mucopolysaccharide content of the GBM observed in hibernating squirrels (Zimny,1973) and in aestivating frogs (Bayomy et al., 2002). Our results also indicate that the GBM becomes positive for sialic acid during aestivation. Although sialic acid appears to be a universal marker for the GBM (Holthöfer,1983), the GBMs of freshwater fish lack this component. The appearance of sialic acid in the GBM of the aestivating P. dolloi may be related to the terrestrial habitat. This change may represent a general feature of the evolutionary adaptation to land.
The presence within the thickened GBM of rounded inclusions containing coiled fibrils is an interesting feature. Their proximity to the foot processes suggests that they are produced by the podocytes. These cells synthesize GBM components (Pavenstädt et al.,2003) and, in the lungfish, show a prominent Golgi complex in the days before the inclusions appear. The presence of collagen fibers in extracellular compartments formed by endothelial or mesangial cell processes suggests that these cells are implicated in collagen synthesis. The formation of these compartments is typical of cells involved in extracellular matrix synthesis (Trelstad and Hayashi,1979).
The mesangial cells also undergo structural modifications during aestivation. Cell compaction and shortening of the cytoplasmic processes have been also observed in hibernating squirrels (Zimny and Levy,1971). These changes should represent an obstacle to the free passage of molecules through the filtration barrier. Indeed, it has previously been suggested (Hickman,1968) that mesangial cells are responsible for the rapid adaptation from a freshwater to a saline environment. Our results support this hypothesis.
During the recovery period studied, all the RC components tend to return to the control, freshwater situation. However, the degree of recovery depends on the component studied. For instance, parietal cells do not recover their flattened shape, and their lectin-binding patterns are similar to those observed during aestivation. In the recovery period, podocytes show several structural signs of undifferentiated cells, such as the presence of a cilium and desmosomes. These features are similar to those found in immature renal corpuscles (Garrod and Fleming,1990; Reeves et al.,1978). In addition, podocytes exhibit cell processes that resemble developing axons. This observation supports the idea that there are biological features common to the two cell types (Kobayashi,2002). Furthermore, the recovering fish may be a useful model to study the development of the major podocyte processes.
During the recovery period, the GBM shows great reductions in thickness and in collagen content, but the lectin-binding patterns (except reactivity to MPA) and the amount of collagen remain unchanged. This finding could indicate the existence of a delay between the activation of the exoglycosidases that degrade carbohydrates (Grigorova-Borsos et al.,1987) and the metalloproteinases that remodel the extracellular matrix (Catania et al.,2007). The infiltration of the kidney tissue by mast cells is another interesting feature. These cells are involved in extracellular matrix remodelling (for review, see: Trabucchi et al.,1988; Noli and Miolo,2001), and may facilitate the structural normalization of the GBM.
We can assume that, at recovery times longer than those studied here, all the RC components will attain the structure observed in freshwater fish. In fact, the lungfish start urine production shortly after being returned to water (Wood et al.,2005), recover body activity almost immediately, and resume food consumption in approximately 2 weeks. However, it should be stressed that we have used young animals which have most probably experienced the process of aestivation for the first time. We do not know whether the structure is fully recovered, how this may affect the kidney function in the long term, how many aestivations the fish can withstand, or how repeated aestivations may affect fish fitness and survival.
Finally, three main conclusions can be derived from this study. First, all the RC components of the P. dolloi kidney undergo structural modifications during aestivation. Second, these modifications appear to be aimed at achieving a considerable reduction in the filtration coefficient. The suppression of urine excretion appears fundamental to the ability to cope with dehydration. Third, the renal corpuscle is not a fixed composite. Rather, it appears to be a highly dynamic structure capable of modifying its architecture in response to environmental changes.
The authors thank to R. Garcia-Ceballos, M. Mier and B. Gallardo for technical assistance. This study is dedicated to Prof. J.L. Ojeda on the occasion of his retirement.