African lungfish (Dipnoi) have the ability to aestivate through the dry seasons that constitute part of tropical life (Fishman et al., 1986; Graham, 1997). Aestivation is a state of dormancy that presents, among other characteristics, exclusive air-breathing, slowdown of the general metabolism, and suppression of kidney and digestive functions (Burggren and Johansen, 1986; Fishman et al., 1986). The adaptive changes in function that occur throughout the aestivation cycles have been mainly studied from the physiological and biochemical points of view (Chew et al., 2003, 2004; Ip et al., 2005; Wood et al., 2005; Amelio et al., 2008; Perry et al., 2008). However, the function of any organ relies on its structure. Surprisingly, the study of structural modifications that accompany functional adaptation to aestivation has received little attention. These modifications do occur. They range from discrete cellular changes to gross histological modifications. Although the decrease in heart function occurs without any significant modification of the heart structure (Icardo et al., 2008), the aestivating kidney shows glomerular collapse and considerable thickening of the filtration barrier (Ojeda et al., 2008). These changes are accompanied by modifications in the activity of nitric oxide (Amelio et al., 2008). Structural changes in the lungs and gills have also been reported during aestivation (Sturla et al., 2002). However, to our knowledge, the study of structural modifications that may occur in the gut of the lungfish during aestivation has not received any attention. On the other hand, several species of the Callichthydae and Cobitididae fish families use a section of the intestine as an air-breathing organ (McMahon and Burggren, 1987), and they are known to be able to survive in mud for hours (see, Graham, 1997). However, this is not equivalent to aestivation and, to our knowledge, modifications in gut structure have not been reported.
The alimentary canal of the African lungfish Protopterus annectens is comprised of a short oesophagus, an intestinal vestibule (a true stomach is not present in lungfish; see Kardong, 2006), a spiral intestine, and a cloaca (Parker, 1892; Purkerson et al., 1975; Rafn and Wingstrand, 1981; Icardo et al., 2010, 2011). The intestinal vestibule, interposed between the oesophagus and the spiral intestine, is a thin-walled sac that appears to serve as a simple passage for food and/or as a food reservoir (Icardo et al., 2010, 2011). The spiral intestine appears to be the single part of the entire gut involved in food processing. It consists of a first chamber with mucosal ridges followed by a longer portion that shows a smooth surface. The two portions are lined with a columnar, pseudostratified epithelium that contains up to six cell types: enterocytes, goblet cells, ciliated cells, dark pigment cells, wandering leukocytes, and endothelial cells pertaining to intraepithelial vessels (Icardo et al., 2011). The functional meaning of some of these cells is unclear as yet. Other cell types could not be identified. The correct structural organization of this epithelium is required for the functional processes associated with digestion. Its structure would likely be modified in the absence of food intake.
In fact, the intestine is a very flexible organ capable to regulate the digestive function to adjust to diet changes and to dietary restrictions (Starck and Rahmaan, 2003; Secor, 2005; Andriamihaja et al., 2010). This is accomplished by structural modifications that range from an increase/decrease of the intestinal mass to more specific cellular and subcellular changes (Waheed and Gupta, 1997; Dunel-Erb et al., 2001; Starck and Beese, 2001, 2002; Hume et al., 2002; Secor, 2008; German et al., 2010). Loss of intestinal mass, downregulation of intestinal performance, and several structural modifications, such as the reduction in microvillous height, are common during fasting episodes. Modifications are more severe in species that, like snakes and aestivating anurans, undergo long periods of fasting in a repetitive manner (see Cramp et al., 2005; Secor, 2005). The gut of the African lungfish P. annectens does not show gross anatomy modifications during aestivation (Icardo et al., 2010). However, we hypothesized that the structure of the epithelium would have to be modified to adjust to the lack of food and water ingest. The study of these modifications is reported here.
MATERIALS AND METHODS
Maintenance of Specimens
The study was performed on a total of 20 juvenile specimens of the lungfish P. annectens (Poll, 1961) with a body weight of 100–150 g. Sex identification within this weight range is not possible due to the lack of distinct anatomical features; however, females represent the main component (up to 98%) of any sampled population of Protopterus (Oliva et al., 1988). The animals were acclimated to laboratory conditions for at least 1 month. They were maintained in plastic aquaria filled with dechlorinated tap water (pH 7.1–7.2), 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 25°C (see Ojeda et al., 2008). The water also contained low concentrations of phosphates (0.10 mM) and sulphates (0.04 mM). During the acclimation period, the fish were fed midge larvae (Bio-Pure Blood Worm, Hikari Sales, CA). Food was withdrawn 96 h before collection of the guts. At this time the guts are empty, indicating a postabsorptive status.
The fish were induced to aestivate at 25−30°C individually in plastic tanks containing a small volume (15 mL) of water, as previously described (Chew et al., 2004; Ip et al., 2005). The water dried up in approximately 6 days. During this time the animals formed a mucus cocoon which enveloped the entire body. The time required for cocoon formation was counted as part of the aestivation process. Initially, seven animals were sacrificed after 6 months of aestivation. Later, four additional animals were sacrificed at 4 months of aestivation to evaluate whether the response to aestivation would vary with time.
Return to Aquatic Conditions After Aestivation
An additional group of fish was reimmersed in water 6 months after the beginning of aestivation. They were not fed during the experimental period. These fish were sacrificed at 6, 10, and 15 days after being returned to the water. Three fish were used for each period.
All fish were killed by a blow to the head and the ventral body wall was opened. The entire gastrointestinal tract was dissected free from the pharynx to the cloaca. Fragments from the intestinal vestibule, the spiral intestine, and the cloaca were sampled and processed for light microscopy or for scanning electron microscopy. A dissection of the entire alimentary canal indicating the sites of sampling has been provided before (Icardo et al., 2011).
For conventional light microscopy, selected gut fragments were fixed in MAW (methanol:acetone:water, 40:40:20), dehydrated in graded ethanol, embedded in Paraplast (Sherwood, St. Louis, USA), and serially sectioned at 8 μm with a Leica RM 2165 microtome. The sections were stained with hematoxylin and eosin for general observations, orcein for the detection of elastic fibers, Sirius red for the detection of collagen, or the periodic acid-Schiff (PAS) reaction for the detection of mucosubstances (see: Sheehan and Hrapchak, 1980). Observations were made with a Zeiss Axioskop 2 plus microscope equipped with an AxioCam HRc digital camera.
Selected gut fragments were fixed in 3% glutaraldehyde in phosphate buffered saline (PBS), pH 7.3, postfixed in 1% osmium tetroxide, dehydrated in graded acetone and propylene oxide, and embedded in Araldite (Fluka, Buchs, Switzerland). Semi-thin sections (1 μm thick) were cut with an LKB III ultratome, stained with 1% toluidine blue, and observed with the Zeiss microscope.
Scanning Electron Microscopy
Selected gut fragments fixed in 3% glutaraldehyde were dehydrated in graded acetone, dried by the critical point method, coated with gold, and observed with an Inspect S scanning microscope (FEI Company).
The intestinal vestibule and the cloaca.
The two ends of the gut are lined with stratified epithelia and appear to lack any digestive function (Icardo et al., 2011). In the intestinal vestibule, the lumen was collapsed and folded throughout the aestivation period (Fig. 1a, inset). Curiously, the modifications of the epithelium were very irregular. At 4 months of aestivation, large areas of the epithelium maintained the stratified organization (Fig. 1a). Many cells occupying the middle region of the epithelium showed dilated cytoplasms. These cells reached the epithelial surface and appeared to release their content into the lumen. Except for an apparent increase in the number of mucous cells, which would indicate an increase in secretory activity, the structure of these areas was similar to that found in freshwater conditions. Structural changes were more apparent in other areas where the entire apical side of the epithelium was occupied by globular cells that showed clear, dilated cytoplasms (Fig. 1b). The cell nuclei were located close to the epithelial surface and the cytoplasms bulged into the lumen. Many of these cells appeared to be desquamating. The two types of areas were irregularly distributed without any apparent preferential location. The basal cells with a dark nucleus, present in freshwater conditions, were not apparent. In addition, all epithelial cells maintained the PAS-positive staining (Fig. 1c) observed in freshwater conditions. The lumen contained large amounts of mucus, cell debris, and isolated cell nuclei. Dead cells with condensed chromatin were observed both in the epithelium and in the thick lamina propria (Fig. 1a). Under the scanning microscope, the epithelial cell surface was either smooth, with attenuation of the microridges and an increase in the number of surface pits (Fig. 1d), or it was dominated by the presence of bulging cytoplasms (Fig. 1e). In the latter areas, the cells appeared to be shedding large portions of the cytoplasm leaving small holes. The apical cell complexes appeared conserved in all cases, and the general structure of the wall was not modified. Strikingly, the epithelium showed a uniform, stratified organization after 6 months of aestivation (Fig. 1f,g). Thus, the areas characterized by the presence of bulging cytoplasms were not observed. The epithelium had lost the typical cobblestone appearance, the cell surface showed a variable number of microridges, and the cells appeared flattened or showed areas of cytoplasm overlapping (Fig. 1h).
The overall structure of the cloaca was maintained throughout aestivation. The wall was folded (Fig. 2a) and the lumen appeared filled with mucus and cell debris. The inner side of the cloaca was lined by a stratified epithelium that showed many cells with dilated cytoplasms (Fig. 2a). Images of cytoplasmic discharge into the lumen were frequent (Fig. 2b). Dead cells were often observed both in the epithelium and in the lamina propria (Fig. 2b). In addition, cells had lost the surface microridges and presented a smooth surface and areas of discrete overlapping (Fig. 2c).
The spiral intestine.
The general organization of the spiral intestine was preserved during aestivation. In the first large chamber, the ridges appeared irregularly deformed (Fig. 3a), but they maintained the connective and vascular core, and the lamina propria appeared intact. We could not detect structural modifications in the muscular component of the wall. In addition, thickness of the external muscle layer remained unchanged (0.74 μm of average thickness in both freshwater and aestivation conditions). The most important changes occurred at the epithelial level. At 4 months of aestivation, the epithelium was almost disintegrated. Epithelial cells appeared to be shedding into the lumen together with a high number of dark pigment cells (Fig. 3b,c). The apical surface of the epithelium was very irregular, the basal boundary was unrecognizable most of the times, and the height of the epithelium was difficult to determine. The epithelial lining could even be lacking (Fig. 3b,c). These modifications affected the apical and basal portions of the ridges (Fig. 3b,c) more than the middle areas (Fig. 3d). None of the normal constitutive cells (enterocytes, goblet cells, and ciliated cells) could be recognized in semi-thin sections. Instead, the epithelium contained a uniform cell population characterized by the presence of a rounded or oval nucleus with heterochromatin clumps and pale cytoplasm (Fig. 3d). Numerous dead cells with a clear nucleus, condensed chromatin, and peripheral chromatin rims were interspersed with the epithelial cells. Curiously, the epithelium and the ridge core showed similar cell phenotypes (Fig. 3b–d). In addition, the number of dark pigment cells had increased. Blood vessels in the ridge core appeared dilated (Fig. 3b). Lymphatic micropumps, characterized as spherical structures consisting of a central vessel surrounded by a clear matrix, were also very apparent (Fig. 3d). The gut lumen was filled with large amounts of mucus, dark pigment cells, cell debris and dead cells with a pycnotic nucleus (Fig. 3b–d). Strikingly, the epithelium appeared stratified at 6 months of aestivation (Fig. 3e–g), showing a homogeneous height, a clear basal boundary, and a more uniform apical surface (Fig. 3e–f). No differences between the apical (Fig. 3e), the basal (Fig. 3f), and the lateral (Fig. 3g) ridge surface could be observed. The epithelium was mostly formed by cells that showed rounded or oval nuclei with heterochromatin clumps. A few epithelial cells showed bulging, dilated cytoplasm filled with secretory material. However, these cells were restricted to the shallow areas between ridges (Fig. 3f). Dead cells were frequently associated with the ridge surface, but were less frequent within the epithelium itself. Overall, the number of dead cells was reduced both in the epithelium and in the ridge core when compared to 4 months of aestivation. Many cells appeared to be desquamating. In addition to white blood cells, mast cells with different granule densities appeared in the epithelium (Fig. 3g). As occurred at 4 months, dark pigment cells were very numerous and the lumen was filled with large amounts of mucus, cell debris, and dead cells. Despite the stratification, the thickness of the epithelium was reduced by 25% as measured at the ridge lateral surface (mean values shifted from 91.84 to 68.5 μm).
The scanning microscope showed dramatic modifications of the ridge structure. At 4 months of aestivation, the epithelium contained round or ovoid cells that were separated by large intercellular spaces (Fig. 4a). Many cells appeared halfway between the epithelium and the lamina propria. Columns of cells contacting each other through slender cell processes were seen extended across the epithelium, from the basal to the apical surface. At 6 months of aestivation, the epithelium appeared stratified and more compacted, and they showed many round cells and dilated vessels (Fig. 4b). At the ridge surface, brush border cells and cells with irregular microvilli had disappeared, and the number of ciliated cells had been reduced (Fig. 4c). Instead, cells at the ridge crest showed protruding cytoplasms and smooth surfaces with only a small number of discrete microvilli (Fig. 4c). Surface holes, desquamating cells, and areas of cell overlapping were common. Often, dead cells were apposed to the ridge surface (Fig. 4d). Along the lateral surfaces of the ridge, most cells appeared flat, the cell surfaces were smooth, and overlapping cells were common (Fig. 4e). As an exception, the shallow areas between ridges showed protruding cytoplasms and areas of cell extrusion (Fig. 4f).
The smooth portion of the spiral intestine followed a slightly different pattern of modifications than the ridge area. At 4 months of aestivation, none of the normal constitutive cells could be identified (Fig. 5a). However, epithelial disintegration was not observed. The epithelium contained cells that showed a dark nucleus with heterochromatin, and dead cells with condensed chromatin (Fig 5a). The height of the epithelium was reduced. In addition, large granulated cells were observed in the vascular plexus underlining the epithelium (Fig. 5a). At 6 months of aestivation, the epithelium appeared stratified, the number of dead cells had been reduced, vessels appeared dilated, and the number of white blood cells and dark pigment cells had increased (Fig. 5b). Numerous mast cells appeared traversing the epithelium. The thickness of the epithelium undergo a small reduction of about 10% (mean values shifted from 108.33 to 97.55 μm, as measured in the external side of the intestine). Under SEM, the epithelium showed rounded or oval cells, large intercellular spaces, and the presence of columns of aligned cells (Fig. 5c; see also Fig. 5d, inset). The inner side of the spiral intestine also showed a stratified epithelium, round or oval cells, and numerous dark pigment cells (Fig. 5d). However, the modifications were less apparent than in the outer side. The smooth portion of the spiral intestine was covered by flat cells with smooth surfaces that overlapped, adopting a roof tile configuration (Fig. 5e). Overlapping was less marked at the inner side of the spiral intestine. Many ciliated cells had disappeared and the remaining ones showed loss of cilia and were covered by the rest of the cells (Fig. 5e, inset). Finally, the entire epithelium of the spiral intestine had lost the PAS-positive staining observed in freshwater conditions (Fig. 5f). PAS-positive staining remained associated with the mucus present at the epithelial surface and in the gut lumen.
The lungfish broke the cocoon when returned to water and started to surface to gulp air. Movements were at first sluggish, but body activity was fully recovered by the end of the experimental period. Modifications were observed in all segments of the alimentary canal. In the intestinal vestibule, at 6, 10, and 15 days after arousal, the epithelium showed areas of activation similar to those observed at 4 months of aestivation (Fig. 6a). Large cells showing a corrugated body were often observed (Fig. 6a, inset). Both in the intestinal vestibule and in the cloaca, cells appeared to be releasing part of the cytoplasmic content, and surface pits and areas of cytoplasmic exposure were common (Fig. 6b). Also, microridges reappeared at the cell surfaces (Fig. 6b).
In the spiral intestine, the ridges showed a regular appearance 6 days after arousal (Fig. 6c). At this time, the epithelium was formed by columnar cells and showed a pseudostratified appearance (Fig. 6d,e). Many cells bulged into the lumen (Fig. 6d,f) showing a supranuclear cytoplasm filled with secretory material (Fig. 6d). Cells also displayed microvilli. However, the microvilli were often very short and did not cover the entire apical surface (Fig. 6f, inset). Ciliated cells were apparent. Epithelial vessels appeared dilated, dark pigment cells were numerous, and mast cells had disappeared. At 10 and 15 days after arousal, the luminal surface showed a mosaic structure (Fig. 6g). However, the cell surface did not fully recover the structural appearance observed in freshwater conditions (see inset of Fig. 4c for comparison). Brush border cells presented small apical surface areas and goblet cells showed a polygonal, smooth surface with a few scattered microvilli (Fig. 6g).
In the smooth portion of the spiral intestine the epithelium also adopted a pseudostratified, columnar appearance 6 days after arousal (Fig. 7a), progressively adopting structural features typical of freshwater conditions. As in the ridge area, many cells showed supranuclear cytoplasms filled with secretory material (Fig. 7b) that was PAS-positive (inset, Fig. 7b). Mast cells had disappeared from the epithelium. However, intraepithelial vessels were dilated and filled with blood cells. The number of dark pigment cells remained elevated (Fig. 7a,b). The epithelial surface recovered a mosaic appearance (Fig. 7c), but goblet cells showed a smooth surface with a discrete number of microvilli. Many ciliated cells displayed a regular tuft, but others showed a reduced number of cilia. In addition, cilia were of different heights (inset, Fig. 7c). A small number of mitoses were observed throughout the spiral intestine (Fig. 7b). It must be underscored that the epithelial surface did not fully recover the freshwater structure during the arousal period studied.
The epithelium of the lungfish gut undergoes profound structural modifications in response to fasting. Several aspects of this response are shared by a wide range of animal groups, from insects to mammals. This commonality agrees with the presence of a character that has been acquired early in evolution. It allows to adjust the structural and functional properties of the gut to food availability and food quality and to the characteristics of the fasting episodes (see Secor, 2005). In birds and mammals, fasting is accompanied by a decrease in intestinal mass and in the surface area of villi (Dunel-Erb et al., 2001; Hume et al., 2002; Starck and Rahmaan, 2003; Karasov et al., 2004). Fasted catfish also undergo significant reduction in intestine mass (German et al., 2010). In species that undergo long episodes of fasting in a repetitive manner, such as snakes and aestivating anurans, these modifications are more severe, reaching up to a loss of 60% of the wet intestine weight (Starck and Beese, 2001, 2002; Secor, 2005; Cramp et al., 2005; Secor, 2008). In P. annectens, there is no apparent specific mass reduction. Furthermore, shortening is not possible since the alimentary canal is a straight organ attached at both ends. Except for some irregularities in the ridges, architectural changes in the gut were not apparent (also, see Icardo et al., 2010). In contrast, most fasting species undergo a reduction in villous surface, number of villi, or changes in villi length (Waheed and Gupta, 1997; Dunel-Erb et al., 2001; Starck and Beese, 2001, 2002; Cramp and Franklin, 2005; Secor, 2008).
In the lungfish, the different segments of the alimentary tract become affected differently. In addition, temporal differences were detected. The transitional epithelium of the intestinal vestibule shows, at 4 months of aestivation, bursts of secretory activity in discrete areas of the epithelium. We could not distinguish any particular localization, nor could we find any indication of structural differences in freshwater conditions (Icardo et al., 2011) that may be responsible for this unequal response. More curiously, the epithelium showed a uniform, close-to-freshwater structure after 6 months of aestivation. A plausible explanation is that this segment needs some time to adjust to aestivation, responding with bursts of activity until reaching a more quiescent state. This is not unique. Lungfish need a period of about two months to adjust respiratory frequency and heart rate to aestivation conditions (Delaney et al., 1974). Furthermore, green-striped burrowing frogs (Cyclorana alboguttata) undergo significant reductions in the mass of the small intestine, together with discrete cellular changes, between 3 and 9 months of aestivation (Cramp et al., 2005). In aestivating lungfish, the final structure of the intestinal vestibule is close to that observed in freshwater conditions reinforcing the suggestion (Icardo et al., 2010, 2011) that this segment is not involved in food processing. The same could be applied to the cloaca even when this segment does not show bursts of activity. Further studies are needed to confirm differences in function along the lungfish gut. However, the structural data indicate that only the segments that appear to be specifically involved in digestion undergo drastic structural modifications. This is restricted to the spiral intestine.
In the spiral intestine, structural differences are observed between 4 and 6 months of aestivation. As occurs with the intestinal vestibule, the period between 4 and 6 months appears to be necessary to reach some kind of structural equilibrium. Furthermore, differences between the ridge area and the smooth portion of the spiral intestine were detected. At 4 months, the pseudostratified columnar epithelium that covers the ridges is almost disintegrated, cell desquamation appears to be massive, and the epithelium and the lamina propria show numerous dead cells. Disintegration of the epithelium and cell death occurs in fasted and food-restricted birds (Karasov et al., 2004) and in fasting rats (Kakimoto et al., 2008). Cell death also occurs in more distant animals like the cockroach Periplaneta Americana (Park et al., 2009). Although cellular changes appear to be less intense in food-restricted than in fasted mammals, lack or decrease of supranuclear vesicles in enterocytes and intense cell desquamation appear to be a common feature in most animal groups subjected to fasting (Dunel-Erb et al., 2001; Starck and Beese, 2001; Karasov et al., 2004). This is accompanied by a decrease in cytoplasmic staining (German et al., 2010). It should be mentioned that the epithelium of the smooth portion of the spiral intestine does not appear disintegrated. Small differences in response could be related to functional differences. Segments with more important digestive functions (see Icardo et al., 2011) may respond more intensely to aestivation. Of note, the inner portion of the spiral intestine, which appears to be less important in digestion (Icardo et al., 2011), undergoes less intense modifications.
Surprisingly, the epithelium of the entire spiral intestine thickens and adopts a stratified appearance at 6 months of aestivation. In addition, many cells are organized into columns and appear to be migrating across the epithelium. In fact, the entire epithelium appears to be renewed by cells that lack secretory vacuoles, are PAS-negative, and they do not show any of the phenotypic characteristics found in freshwater conditions. The origin of these cells is uncertain. In birds and mammals, continuous cell proliferation allows for fast and reversible changes of the epithelium (Starck and Rahmaan, 2003). In the lungfish, the lack of a distinct germinal zone in freshwater conditions (Icardo et al., 2011) and the absence of mitoses during aestivation indicate an extraepithelial origin. The fact that many cells appear halfway between the lamina propria (or the vascular plexus) and the epithelium, and the similarity of the nuclear phenotypes between cells in the epithelium and in the lamina propria, suggest a vascular origin. Although we know nothing about the presence of stem cells in the lungfish, they are not a late acquisition in evolution. For instance, stem cells give rise to various adult cell types in flatworms, cnidarians, and sponges (Mochizuki et al., 2000; Saló and Baguñá, 2002; Juliano and Wessel, 2010). In mammals, haematopoietic stem cells are able to migrate to sites of injury to regenerate damaged tissues (Kucia et al., 2004; Stroo et al., 2009), and epithelial stem cells appear to be responsible for tooth replacement in zebrafish (Huysseune and Thesleff, 2004).
It is worth to mention that cell death, cell desquamation, and transient stratification of the columnar epithelium, also occur in the gut epithelium of tadpoles at climax of metamorphosis (Bonneville, 1963; Schreiber et al., 2005). However, metamorphosis is driven by high levels of thyroid hormone. We do not know how hormone secretion is modified during aestivation. In the lungfish, the presence of intraepithelial mast cells, many of them with empty granules, indicates a possible role of inflammation as part of the cellular response to aestivation. Mast cells are not detected in freshwater conditions (see Icardo et al., 2011).
It should be underscored that aestivation cannot be looked upon as a simple downregulation of metabolic functions. Many cellular activities are upregulated (Icardo et al., 2008; Ojeda et al., 2008), and energy consumption is required to maintain the aestivation state. During fasting, animals rely upon stored energy to meet metabolic demands (Secor, 2008). A fundamental question is how this energy is generated. In the heart, myocardial cells appear to rely upon stored glycogen (Icardo et al., 2008). In the gut, in addition to preventing desiccation, the enormous amount of mucus and cell debris shed into the lumen may be retrieved for energy consumption. Cell debris can be utilized as a source of energy (Park et al., 2009). The liberation of enzymes that accompanies plasma membrane rupture degrades the mucous material and many components such as amino acids may be freed into the lumen (Hume et al., 2002) and may traverse the epithelium by nonspecific ways (Secor, 2008). Hibernating squirrels (Carey, 1990; Carey and Sills, 1992) and aestivating anurans (Secor, 2005) increase glucose uptake between 50% and 90% despite a reduction in the surface of the villi and a substantial decrease in intestinal mass. Similarly, uptake of several amino acids increases in fasting rats (Waheed and Gupta, 1997). In lungfish, the epithelium during aestivation may simply be more permeable, allowing nutrient passage by a mere gradient concentration. The presence of large intracellular spaces would facilitate absorption. Thus, we hypothesize that the material shed into the lumen is reutilized as a source of energy for general body maintenance. Instead of reabsorbing the intestinal tissue, as premigratory birds (Piersma and Gill, 1998) and aestivating frogs (Cramp et al., 2005) appear to do, continuous cell transit and desquamation would add new cellular material that could serve as fuel for consumption. Identification of brush border membrane enzyme activities, and study of the expression of key transporters, are needed to confirm the present hypothesis. However, the facts that epithelial vessels are dilated, and that the lymphatic micropumps are very apparent, suggest that some kind of transport activity is present throughout the epithelium. In the absence of transport activity, vessels and lacteals appear collapsed (Dunel-Erb et al., 2001; Karasov et al., 2004).
Dipnoi constitute the single fish group where the presence of lymphatic vessels and lymphatic micropumps has been documented (Vogel and Mattheus, 1998). An extensive lymphatic system also occurs in the wall of the alimentary canal (Icardo et al., 2011). It was previously suggested that the intraepithelial vessels could be part of the lymphatic system of the gut (Icardo et al., 2011). The fact that they contain red blood cells during aestivation and after arousal casts doubts on that assertion since erythrocytes are not normally found within lymphatics. Further studies are needed to elucidate this matter.
After arousal, the epithelium of the lungfish intestine recovers many of its features when the lungfish is in water. This occurs in about 6 days. The recovery time is longer than that observed in birds and mammals after experimental fasting (Dunel-Erb et al., 2001; Karasov et al., 2004), but similar to that observed after mammalian hibernation (Carey, 1990; Hume et al., 2002). In fact, recovery appears to be slower after long fasting periods (Hume et al., 2002). Of note, the epithelium of the spiral intestine appears stratified at 6 months of aestivation. Epithelia of this kind can easily be transformed into columnar epithelia (Starck and Beese, 2001) without recurring to high mitotic activity that requires elevated energy costs. This is a strategy that appeared unique to the Burmese python (Python molurus bivittatus) (Starck and Beese, 2001), but seems to be shared by lungfish. The difference is that, in the lungfish, epithelial cells have to develop full phenotype characteristics. This makes the structural recovery more remarkable and explains the growing microvilli and cilia, and the progressive acquisition of the PAS-positive staining. Nonetheless, the lungfish gut does not fully attain freshwater features during the arousal period studied here. It should be mentioned that, although the animals may swallow water, they were not supplied with food. In fact, lungfish refuse to eat until at least the beginning of the second week after arousal (unpublished observations). This may indicate that the animals will only feed when some restructuring of the epithelium has taken place, that initial recovery is independent of food intake, and that it relies on internal mechanisms rather than on the availability of nutrients. However, feeding may be necessary for full structural recovery (Cramp and Franklin, 2003). Complete functional recovery may take up to three months (Cramp and Franklin, 2003).
We thank B. Gallardo and R. García-Ceballos for technical assistance.