Intestinal apoptotic changes linked to metabolic status in fasted and refed rats

Intestinal apoptosis and expression of apoptosis inducers – the cytokines TNFα, TGFβ1 – and the intestinal transcription factor Cdx2, were studied according to two different metabolic and hormonal phases which characterize long-term fasting: the long period of protein sparing during which energy expenditure is derived from lipid oxidation (phase II), and the later phase characterized by a rise in body protein utilization and plasma corticosterone (phase III). Apoptosis was further studied in 2, 6 and 24h refed rats. Morphological apoptotic events were observed by environmental and conventional scanning electron microscopy and TUNEL test was used to characterize the final stages of apoptotic death. The gene and protein expressions of TNFα, TGFβ1, and Cdx2 were measured. Apoptotic events and TNFα, TGFβ1, and Cdx2 gene and protein expressions did not vary significantly during phase II compared to normally-fed animals. However, phase III fasting induced a delay in intestinal epithelial apoptosis, along with a 92%, 58% and 25% decrease in TNFα, TGFβ1 and Cdx2 mRNAs, respectively. The amounts of TNFα, TGFβ1 and Cdx2 proteins decreased by 70%, 36% and 25%, respectively. Apoptosis was restored rapidly after 2h refeeding following phase III, accompanied by a significant increase in TNFα, TGFβ1 and Cdx2 mRNA and protein levels, compared to the phase III fasting values. The concomitant decreases in cytokines and Cdx2 and in apoptotic cells during phase III suggest the preservation of enterocytes during this critical fasting period in order to optimize nutrient absorption as soon as food is available and thus, to rapidly restore body mass.

Author keywords: apoptosis, body reserves depletion, corticosterone, cytokines

Three distinct metabolic phases have been described during a fast (7, 21, 30, 32). During the first phase (phase I), glycogen stores are completely exhausted and fat stores are progressively used. This phase lasts only a few hours in rats. The second phase (phase II) corresponds to a phase of economy with the mobilization of fat stores for energy expenditure whereas body proteins are efficiently spared. The third phase (phase III) is characterized by an increasing protein catabolism caused by a rise in plasma corticosterone levels. During this later phase, there is a clear change in behavior which promotes food foraging, therefore anticipating a lethal depletion of energy stores (30). In rats, fasting during either phase II or phase III induces a significant decrease in the size of intestinal villi and thus, in the small intestinal mass (13). Also, during the phase of protein sparing (phase II), cell proliferation in crypts and cell migration along the crypt-villus axis decrease, whereas apoptosis at the tip of the villi is maintained (22). In phase III however, when the rat has reached a critical threshold in body reserves utilization, intestinal cell proliferation and migration increase significantly compared to phase II fasted and normally-fed rats. In parallel, we observed a delay in apoptotic events at the tip of the villi (22). The increase in cell proliferation and the preservation of absorptive cells are concomitant with a rise in locomotor activity reflecting the search for food (30) and may prepare the mucosa to nutrient absorption as soon as food is available. The aim of the current work is to further evaluate the intestinal apoptotic changes linked to the metabolic state of fasted rats.

During the normal course of the development of the animal, apoptosis is activated as a response to specific endogenous factors such as TNFα and TGFβ1 (15, 36); the caudal-related homeodomain protein Cdx2 has also been shown to be critical in these processes (3). Apoptosis can be induced by exposure to cytotoxic compounds, hypoxia, viral infection, changing levels of specific hormones, and fasting. A short period of starvation (less than 3 days) has been shown to increase apoptosis in the small intestine mucosa (19, 27, 28). However, other authors (34) have related a decrease in apoptosis following a 3-day starvation period. The effects of longer periods of fasting and of refeeding on apoptotic events in the small intestine remain almost unknown and intestinal apoptosis has never been studied by other authors in relation to whole body metabolism.

In this study, we evaluate the effects of phase II and phase III fasting and of refeeding on apoptotic events in the intestinal epithelium. Apoptosis was studied at the level of cell morphology by Scanning Electron Microscopy (SEM), and cell biology using the TUNEL test. The concomitant variations of proapoptotic cytokines (TNFα, TGFβ1) and of the intestine specific transcription factor Cdx2 were analyzed by RT-PCR in order to determine gene expressions, and by western blot in order to quantify TNFα, TGFβ1 and Cdx2 protein levels. Finally, TGFβ1 and Cdx2 protein distribution along the crypt-villus axis was investigated by immunohistochemical detection.

Results

Discussion

References

Male Wistar rats weighing 350g were obtained from Iffa-Credo (Lyon, France). The animals were housed individually in leucite cages with a wire mesh floor to minimize coprophagia, and were maintained at 23 °C with a 12-hour light period. They had free access to the control diet (A03 pellets from UAR, Epinay-sur-Orge, France) consisting of 23% (by mass) protein, 51.1% carbohydrates, 4.3% fat, 4% cellulose, 5.6% minerals, and 12% water. The rats had free access to water throughout the experiments. They were weighed every day between 9.00 and 10.00 a.m. Our experimental protocol followed the Centre National de la Recherche Scientifique (CNRS) guide for care and use of laboratory animals.

Animals

Experimental procedures

After a one week acclimatization, normally-fed rats served as control animals (Ctrl, n=15), whereas the other rats were food-deprived. The fasting phases were determined by calculating the specific daily rate of body mass loss dM/Mdt (g/kg/day) for each animal (dM represents the loss of body mass during dt=t₁−t₀ and M is the rat body mass at t₀). This calculation permitted a daily monitoring of the physiological status for each animal through fasting. With this monitoring, all the animals survived the prolonged starvation procedure and could be successfully refed. The first fasting phase (phase I) lasted only a few hours and was characterized by a rapid decrease in dM/Mdt. The specific daily body mass loss then reached a steady rate (approximately 55g/kg/day) representing phase II and finally, strongly increased which was characteristic of phase III.

The phase II fasting period lasting between one and six days (for 350g rats), a first group of rats was killed in the fourth day in phase II (P2r0, n=15). Three other groups were refed during 2h (P2r2, n=5), 6h (P2r6, n=5), or 24h (P2r24, n=5) following phase II, and then killed. Four additional groups continued fasting until the second day of phase III, reaching on average eight days of fasting. One group was killed in phase III without refeeding (P3r0, n=15), whereas the three others were killed after refeeding for 2h (P3r2, n=5), 6h (P3r6, n=5), or 24h (P3r24, n=5) following phase III.

The animals were killed between 9.00 and 10.00 a.m. The jejunum was removed, weighed, and cut into segments. These segments were then treated separately depending on the analysis considered (CSEM, ESEM, immunohistochemistry, RT-PCR, western blotting).

Plasma parameters

Blood samples were collected immediately after sacrifice to measure plasma concentrations of urea and corticosterone in all experimental groups. To assess the metabolic state for each animal, plasma urea was determined with a Urea Nitrogen Kit (Sigma Diagnostics, St Louis, USA), according to the manufacture’s guidelines. The plasma concentration of corticosterone was determined with an Enzyme Immunoassay Kit (Assay Designs Inc., Ann Arbor, USA).

Environmental Scanning Electron Microscopy

Samples of jejunal mucosa were placed in 1% glutaraldehyde and buffered saline for 30s to block mucus secretion, and then viewed with a Philips XL-30 ESEM. The pressure in the sample chamber was maintained at 5 Torr and the temperature at 4°C, in order to hold a relative humidity rate of 80% at the sample surface.

Conventional Scanning Electron Microscopy

Jejunal mucosal samples were fixed for 2h at 4°C in 5% glutaraldehyde in 0.05M cacodylate buffer (pH=7.4), post-fixed 1h in 1% osmium tetroxide, and dehydrated. Pieces were then dried with hexamethyldisilazane. After being mounted on stubs, samples were coated with gold and examined with a Philips XL-30 ESEM.

TUNEL assay procedure

The TUNEL assay (in situ Cell Death Detection Kit, fluorescein) was performed according to the manufacturer’s specifications (Roche, USA). Each experiment set up by TUNEL reaction mixture without terminal transferase served as negative control. Samples pre-treated with DNAse (2mg/mL) served as positive controls. Sections were examined with a fluorescent microscope (Zeiss Axioplan) equipped with the appropriate filter set (450–590nm bandpass excitation filter). TUNEL-positive cells with fully condensed nuclei were counted in a minimum of 30 villus sections per animal and in 5 animals per group and expressed as a percentage of the total epithelial cells.

RT-PCR analysis

Total RNA from jejunal mucosa was extracted by the method of Chomczynski and Sacchi (8). cDNA was synthesized from 2μg total RNA in 25μL reaction buffer (Finnzymes OY, Finland) containing 1μL of avian myeloblastosis virus reverse transcriptase (Finnzymes OY, Finland), oligo(dT)₁₇ primer (50pM, Eurogentec, Belgium) and 0.2mM of deoxynucleotide triphosphate (Promega, France), (60min at 42°C). cDNA was then amplified by Polymerase Chain Reaction (PCR) using specific primers (Table I). The PCR reactions were carried out in 10μL Red’Y’StarMix (Eurogentec, Belgium) diluted to a final volume of 20μL, 50pM of each primer and 1μL of the cDNA mixture. cDNAs were amplified for a determined number of cycles (Table I) as follows: hot starting for 10min at 95°C, denaturation for 45s at 94°C, annealing at 50°C for 45s, and elongation at 72°C for 45s. PCR products were resolved on a 3% agarose gel and visualized by ethidium bromide staining. The staining intensity was evaluated using the Molecular Analyst software (Bio-Rad Laboratories, USA). Results were expressed as relative densitometric units, normalized to the values of a phosphoribosomal protein mRNA used as internal control.

Western blotting and immunohistochemistry

Western blotting and immunohistochemistry were conducted as described previously (23). The primary antibodies used were polyclonal rabbit anti-rat TGFβ1 [anti-LC-(1-30)] (generously provided by Dr. K. C. Flanders), monoclonal mouse anti-human Cdx2 (BioGenex, San Ramon, Ca, USA) and polyclonal rabbit anti-human TNFα (Genzyme Diagnostics, Cambridge, USA).

Cdx2 antibody could be detected on paraffin sections after incubation with Alexa ₄₈₈-conjugated goat anti-mouse IgG (Molecular Probes), whereas TGFβ1 immunolocalization was performed using the standard extravidin-biotin-peroxidase complex technique and diamino-benzidine coloration.

Statistical analysis

Data are presented as mean values +/− SEM. Statistical comparisons of experimental data were performed by one-way analysis of variance (ANOVA) and Tukey post-hoc test by using the software Sigmastat (Jandel). The level of statistical significance was set at P<0.05.

Evolution of the metabolic and hormonal status

Plasma urea concentration (Table II) did not vary between control, phase II fasted rats and refed rats following phase II. A phase III fast induced a 3.7-fold significant increase in urea concentration. Uremia then decreased in rats after 6h refeeding following phase III. Plasma corticosterone concentration (Table II) showed a 13-fold increase in P2r0 fasted animals compared to controls, but it was not significant. After refeeding following a phase II fasting period, plasma corticosterone concentration was lower compared to control values. In phase III fasted rats, plasma corticosterone concentration exhibited a 370-fold significant increase compared to controls, and a 29-fold increase compared to P2r0 fasted rats. Refeeding following phase III induced a rapid decrease in plasma corticosterone.

Apoptotic events

Environmental Scanning Electron Microscopy makes it possible to examine fresh, lightly fixed biological specimens (neither dehydrated, nor gold coated) without artifacts linked to sample manipulation and treatment. A morphological study of apoptotic events at the tip of the villi was therefore possible for our 9 experimental groups (figure 1, a to i). These observations were made along the jejunum for 5 animals per group. Apoptotic zones were observed at the villi tips in control (figure 1a) and phase II fasted rats (figure 1b), but very few apoptotic events could be seen in phase III fasted rats (figure 1c) and in rats refed 2 and 6h following phase II (figure 1, d and e). In animals refed 24h following phase II (figure 1f), and in those refed following phase III (figure 1, g to i), the intestinal villi tips looked like those in controls. Observations made at higher magnifications with Conventional Scanning Electron Microscopy (figure 1, j to l) showed apoptotic cells being extruded from the epithelium and losing their microvilli in control (figure 1j) and in phase II fasted rats (figure 1k) but not in phase III fasted rats (figure 1l).

There were no differences in the number of TUNEL-positive cells at the tips of the villi between control and phase II fasted rats (figure 2, a, b and j). Only a few apoptotic cells could be seen in animals refed 2h and 6h after a phase II fast (figure 2, d, e and j). Numerous TUNEL-positive cells could be seen again after 24h refeeding following phase II (figure 2f and j). In phase III, almost no apoptotic cells could be detected at the tip of intestinal villi (figure 2c and j). However, after refeeding following phase III, the apoptotic index was the same as in control animals (figure 2, g to j).

Changes in TNFα and TGFβ1 gene and protein expressions

TNFα and TGFβ1 were expressed in the small intestine mucosa (figures 3 and 4). However, TNFα was only faintly expressed, as indicated by the high number of PCR cycles needed to detect its mRNA (Table I).

TNFα transcripts and protein level (figure 3A and B) did not vary during a phase II fast compared to control animals. However, TNFα mRNA levels were very low in animals refed 2h and 6h after phase II compared to controls and phase II fasted rats, but no significant difference in the amounts of TNFα protein could be detected. A phase III fast strongly reduced TNFα gene expression and protein level. After 2h refeeding following phase III, the amount of TNFα mRNA exhibited a 7.5-fold increase compared to the control value, and a 94-fold increase compared to the phase III value. It then decreased during 6h and 24h refeeding following phase III, but was still higher than in control rats. TNFα protein level rose to control value after only 2h refeeding following phase III.

TGFβ1 (figure 4A and B) gene expression and protein level did not vary during a phase II fast compared to controls, but a significant increase was observed between animals refed 2h following phase II and control and phase II fasted animals. TGFβ1 gene expression and protein level were then lowered to control values in 6h and 24h refed rats following phase II. During phase III fasting, TGFβ1 mRNA level and protein amount were decreased by 58% and 36%, respectively. After refeeding following phase III, TGFβ1 gene expression and protein level increased and rapidly reached the control values.

TGFβ1 was localized in the cytoplasm of intestinal epithelial cells, and particularly in the apical part of the enterocytes (figure 4C). It was observed in the upper third of the intestinal villi in control and refed rats, but was weakly expressed in phase III fasted rats.

Modulations of Cdx2 expression and localization

Cdx2 gene expression and protein level (figure 5A and B) did not change during a phase II fast. However, a significant increase was observed after 2h refeeding following phase II. After 6h refeeding, Cdx2 gene expression and protein level were then lowered to control values. A phase III fast induced a 25% decrease in Cdx2 mRNA level and protein amount. After refeeding following phase III, Cdx2 gene expression and protein level increased and rapidly reached the control values.

Cdx2 was localized in the nuclei of intestinal cells and showed a decreasing gradient along the crypt-villus axis (figure 5C). A decrease in the expression of Cdx2 protein was observed in phase III fasted animals compared to controls. The number of strongly labeled nuclei increased in animals refed 2h following either phase II, mainly in the crypt compartment, or phase III, all over the crypt to villus axis.