Total parenteral nutrition (TPN) impairs small intestine development and is associated with barrier failure, inflammation, and acidomucin goblet cell expansion in neonatal piglets. We examined the relationship between intestinal goblet cell expansion and molecular and cellular indices of inflammation in neonatal piglets receiving TPN, 80% parenteral + 20% enteral nutrition (PEN), or 100% enteral nutrition (control) for 3 or 7 days. Epithelial permeability, T cell numbers, TNF-α and IFN-γ mRNA expression, and epithelial proliferation and apoptosis were compared with goblet cell numbers over time. Epithelial permeability was similar to control in the TPN and PEN jejunum at day 3 but increased in the TPN jejunum by day 7. By day 3, intestinal T cell numbers were increased in TPN but not in PEN piglets. However, goblet cell expansion was established by day 3 in both the TPN and PEN ileum. Neither TNF-α nor IFN-γ mRNA expression in the TPN and PEN ileum correlated with goblet cell expansion. Thus goblet cell expansion occurred independently of overt inflammation but in association with parenteral feeding. These data support the hypothesis that goblet cell expansion represents an initial defense triggered by reduced epithelial renewal to prevent intestinal barrier failure.
- total parenteral nutrition
- goblet cells
- barrier function
gastrointestinal atrophy is a common outcome in humans and animals receiving total parenteral nutrition (TPN) (4, 14, 21, 22, 25, 39); however, the contributory mechanisms have not been fully established. In TPN-fed piglets, we observed elevated inflammatory indices, including CD8+ and CD4+ T lymphocyte (T cell) expansion in the small intestine (SI), in conjunction with altered epithelial development and goblet cell distribution. Specifically, a significant expansion of the acidomucin (sulfo- and sialomucins) goblet cell lineage was observed only in inflamed intestinal regions (22). The chemotype and spatial distribution of acidomucin goblet cells typically develops in parallel with bacterial colonization in the intestine (10, 35, 42). Relative to neutral mucins, acidic mucins (especially sulfomucins) resist enzymatic degradation by both bacterial glycosidases and host proteases (20, 42). Sulfation of terminal carbohydrate moieties common to intestinal mucins (e.g., N-acetylglucosamine) inhibits bacterial growth in vitro compared with nonsulfated carbohydrates (6). Acidic mucins are also highly expressed in the intestine during fetal development (16, 32), indicating that they may be of particular importance defensively during the neonatal period, when the acquired immune system is immature and not fully functional (5). Collectively, these observations point to both the dynamic and protective nature of intestinal acidomucin goblet cell populations.
The present study was designed to determine the extent to which the coincident increase in the number of T cells and acidomucin goblet cells (goblet cell expansion) is related in the TPN-nourished SI. We hypothesized that goblet cell expansion is either directly induced by T cell signals (i.e., proinflammatory cytokines) or indirectly induced by secondary inflammatory signals from supportive mesenchymal or neighboring epithelial cells responding to inflammation. In this case, inflammation would follow compromised barrier function resulting from gastrointestinal atrophy. The provision of even minimal enteral nutrition improves intestinal mucosal integrity (38). Accordingly, we examined the temporal relationship between T cell and goblet cell expansion in neonatal piglets receiving TPN or partial enteral nutrition (PEN) vs. total enteral nutrition (TEN). To further examine potential mechanisms contributing to goblet cell expansion, local expression of the proinflammatory cytokines TNF-α and IFN-γ and common indices of intestinal epithelial development including proliferation, apoptosis, and enteroendocrine cell differentiation were measured in conjunction with epithelial permeability.
MATERIALS AND METHODS
Animals and study design.
Experimental procedures were approved by the Institutional Animal Care and Use Committee of the University of Illinois (Urbana-Champaign, IL). Piglets (n = 38) <24 h old were obtained from the University of Illinois swine facilities. Initially, six piglets were killed to provide baseline (day 0) data. The remaining piglets were assigned to three dietary treatment groups: TEN (100% enterally fed, n = 12), TPN (100% parenterally fed,n = 11), or PEN (80% parenterally and 20% enterally fed, n = 9). The enteral formula was a nutritionally complete commercial porcine milk replacer formula (Advance Baby Liqui-Wean; Milk Specialties, Dundee, IL). Fresh formula was provided three times daily, and formula not voluntarily consumed by the PEN piglets was delivered by orogastric gavage. The TPN regimen was formulated to provide adequate amounts of all nutrients necessary for normal neonatal pig growth and was identical to that used by Park et al. (40). The solution was prepared daily as a 3:1 admixture in 1-l Viaflex bags (Baxter Healthcare, Clintec Nutrition Division, Deerfield, IL) under sterile conditions using 8.5% amino acid, 50% dextrose, and 20% Intralipid solutions (Baxter Healthcare). The solution provided 3,752.3 J energy and 53.1 g protein/l with 50% nonprotein energy as fat and the remaining 50% as dextrose. Vitamins, minerals, and trace elements were added daily to the infusate before administration. Catheter surgery, nutrient administration, piglet housing, and general maintenance were followed exactly as described previously (40). The general health, weight gain, formula intake, and volume of TPN infused for each piglet were recorded daily.
TPN and PEN piglets were killed by an injection of pentobarbitol sodium (80 mg/kg body wt), and TEN piglets were killed by electrocution and exsanguination on day 3 (TEN, n = 6; TPN,n = 8; PEN, n = 5) and day 7(TEN, n = 6; TPN, n = 3; PEN,n = 4) of the experiment. The SI was dissected as described previously (22). Segments were flushed with ice-cold saline to remove luminal contents and then weighed and measured. Sections of jejunum, ileum, and colon were frozen in liquid nitrogen for RNA analysis and were fixed in Carnoy's or Bouin's solutions for histological analysis. Sample collection was similar for piglets killed at day 0.
Carnoy's-fixed cross-sections of jejunum, ileum, and colon (3) were embedded in paraffin, sectioned onto glass slides (Fisher Scientific, Pittsburgh, PA), and stained with Toluidene blue. Villus height and width and crypt depth were measured in 10 well-defined villi and crypts in the jejunum and ileum. Colon crypt depth was measured in 10 well-defined crypts in the colon. Measurements were performed using Image-1 software (Universal Imaging, Westchester, PA) and a Nikon Diaphot microscope (Fryer, Carpentersville, IL).
Tissue conductance measurements.
Jejunal and colonic samples were opened longitudinally and mounted in modified Ussing chambers (Physiologic Instruments, San Diego, CA). Mucosal and serosal surfaces (0.5 cm2) were exposed to 4 ml of oxygenated (95% O2-5% CO2) Krebs buffer (pH 7.4) maintained at 37°C with a recirculating water reservoir. After equilibration, spontaneous transmural potential difference (mV), short-circuit current (mA), and conductance (Ω/cm2) were measured by using established techniques (29). The modified Ussing chamber apparatus was connected to a dual-channel voltage/current clamp (VCC MC2; Physiologic Instruments) with a computer interface allowing for real-time data acquisition and subsequent analysis (Acquire and Analyze software; Physiologic Instruments).
Lamina propria CD8+ and CD4+ T cells were detected as previously described (22). Proliferating cells and enteroendocrine cells were identified via immunohistochemistry by using antibodies directed against proliferating cell nuclear antigen (PCNA; Refs. 24 and 30) and chromogranin A (54), respectively. All tissue sections were deparaffinized to double-distilled water (ddH2O) and processed for antigen retrieval by incubating the sections in 0.05% saponin (Sigma, St. Louis, MO) for 30 min at room temperature. Sections were then washed with PBS without Ca2+ (GIBCO, Grand Island, NY) and blocked in either 10% sheep serum (for PCNA) or 10% rabbit serum (for chromogranin A) for 30 min at room temperature. Sections were incubated with either an anti-PCNA antibody (PC-10; Santa Cruz Biotechnology, Santa Cruz, CA) for proliferating cells or an anti-chromogranin A antibody (C-20, Santa Cruz Biotechnology) for enteroendocrine cells for 30 min at room temperature. Finally, sections were washed and incubated with either a goat anti-mouse IgG-FITC-conjugated secondary antibody for PCNA or a rabbit anti-goat IgG-FITC-conjugated secondary antibody for chromogranin A for 30 min at room temperature in the dark. A nucleic acid counterstain of 4′,6′-diamidino-2-phenylindole dihydrochloride hydrate (DAPI; Sigma) was included in the secondary antibody solutions for nuclear visualization. All incubations were carried out in a humidified chamber to minimize evaporation of antibody solutions. After being washed, all sections were treated according to manufacturer's instructions with SlowFade antifade (Molecular Probes, Eugene, OR) and coverslipped. Only antibody-positive cells appearing in the epithelial layer of villi or crypts were counted.
RNA extraction from tissues.
Total RNA was extracted from the ileum by using TRIzol reagent (Invitrogen, Carlsbad, CA) following the manufacturer's instructions. RNA quality was determined by loading aliquots onto a 1% agarose gel to check the integrity of 28S and 18S rRNA. The RNA was treated with RNase-free DNase RQ1 (Promega, Madison, WI) according to the manufacturer's instructions and stored at −80°C for subsequent analysis.
Quantitative real-time PCR for detection of proinflammatory cytokines.
Expression of the proinflammatory cytokines TNF-α and IFN-γ was determined by quantitative real-time PCR. Equal quantities of RNA per sample, as determined spectrophotometrically at 260 nm, were reverse transcribed to cDNA by using a GeneAmp PCR System 2400 thermocycler (Applied Biosystems, Foster City, CA) with a final reaction volume of 50 μl containing 1,000 ng RNA, 10 μl 5× PCR buffer, 1 mM MgCl2, 40 μM dNTP, 1.25 μl RNAsin, 4 mM DTT, 1.25 μl random hexamers, and 0.75 μl MultiScribe reverse transcriptase from a GeneAmp Gold RNA PCR Core kit (Applied Biosystems). The reaction cycle consisted of a 10-min incubation at 25°C followed by a 20-min incubation at 42°C, after which the cDNA was stored at 4°C. Subsequent quantitative PCR analysis was performed in a GeneAmp 5700 Sequence Detection system (Applied Biosystems) in a final reaction volume of 25 μl containing SYBR Green PCR Master mix (Applied Biosystems), 0.5 μl (each) primer, and 5 μl cDNA template. The primers used to detect porcine TNF-α mRNA were pTNF-α forward (5′-GCTGTACCTCATCTACTCCC-3′) and pTNF-α reverse (5′-TAGACCTGCCCAGATTCAGC-3′), which generates a 291-bp fragment (53). For detection of porcine IFN-γ mRNA, the primers pIFN-γ forward (5′-ATTTTGAAGAATTGGAAAGAGG-3′) and pIFN-γ reverse (5′-AAATTCAAATATTGCAGGCAGG- 3′) were used, generating a 368-bp fragment (17). Both TNF-α and IFN-γ primers were previously validated by sequencing resultant amplicons for each set both in house and by others (17,53). Standard curves generated for each primer set by using pig spleen RNA as a control were used to calculate mRNA concentrations. TNF-α and IFN-γ mRNA concentrations were normalized by GAPDH mRNA expression levels. Ratios of cytokine mRNA concentrations were generated by comparing TPN and PEN values to TEN values within days. Data are presented as the percentage of change in cytokine expression relative to TEN piglets by day.
TUNEL assay for the identification of apoptotic crypt cells.
Apoptotic crypt cells were identified by terminal transferase deoxyuridine nick-end labeling (TUNEL) by using the In Situ Cell Death Detection Fluorescein kit (Boehringer Mannheim, Mannheim, Germany). The staining procedure was performed with slight modifications to the manufacturer's instructions. Tissue sections were deparaffinized to ddH2O and then processed for antigen retrieval by heating in a microwave at 375 W in 0.1 M sodium citrate (pH 6.0) for 5 min. Sections were washed and then incubated with the TUNEL reaction mixture for 60 min at 37°C in a dark, humidified chamber. A DAPI counterstain was included in the reaction mixture (1:200 of a 50 μg/ml stock solution). The sections were washed, treated with SlowFade antifade solution, and coverslipped. A section was incubated with DNase to serve as the positive control (DNase RQ1). A slide incubated without the terminal transferase served as the negative control. Cells were considered to be apoptotic if they stained positive with fluorescein and exhibited nuclear morphology consistent with that of apoptotic cells (chromatin clumping and condensation) (48).
Goblet cell histochemistry.
Acidic and neutral chemotype goblet cells were identified by staining with Alcian blue-periodic acid-Schiff reagent (3). Frosted microscope slides (Fisher Scientific) supporting piglet jejunum and ileum sections were deparaffinized to ddH2O. The tissues were stained with Alcian blue (Sigma) in 3% acetic acid (pH 2.5) for 5 min, washed in running tap water, and rinsed in ddH2O. The tissues were incubated in 1.0% periodic acid for 10 min, rinsed as before, stained with Schiff's reagent (Sigma) for 10 min, and then incubated in 2.5% sodium metabisulfite for 5 min and rinsed in running tap water for 10 min. The tissues were counterstained for 30 s with Harris's hematoxylin (Sigma) and washed in tap water for 10 min.
Sulfomucin and sialomucin goblet cells in the jejunum, ileum, and colon were distinguished by staining with a high-iron diamine (HID) solution followed by an Alcian blue stain. Mounted tissue sections were deparaffinized and stained for 16 h with a HID solution (3). Following HID staining, tissues were washed in running tap water for 5 min and stained with Alcian blue (pH 2.5) for 5 min followed by another 5-min wash in tap water. After final washes, all tissues were dehydrated to xylene and mounted with Permount (Fisher Scientific) on 1.5-mm-thick coverslips.
Histochemical sections were observed with a Nikon Optiphot-2 microscope (Nikon, Melville, NY) equipped for epifluorescence (except where noted otherwise), and digital images were captured by using Image Pro Plus software, version 3.0 (Media Cybernetics, Silver Spring, MD).
Data were analyzed using SAS statistical software (version 8.01; SAS Institute, Cary, NC). A one-way analysis of variance using the general linear model procedure compared values among day 0;day 3 TEN, TPN, and PEN; and day 7 TEN, TPN, and PEN groups, with the main effect being treatment. When a significant main effect was observed, differences between groups were determined by using the least significant difference test (18). A value of P < 0.05 was considered statistically significant. Instances in which P ≤ 0.1 are discussed as trends.
Body weight and caloric intake.
Body weight gain was similar among TEN, TPN, and PEN piglets throughout the experiment. At day 3, PEN piglet energy intake was significantly lower compared with day 3 TEN piglets, whereasday 7 PEN and TPN piglet energy intake was significantly lower compared with day 7 TEN piglets (Table1). SI weight (g/kg) was greater (P < 0.05) in TEN piglets compared with TPN and PEN piglets on both day 3 and day 7 (Table2). Colon weights were similar among groups throughout the study.
Villus height was similar in the baseline and day 3 andday 7 TEN jejunum (Table 3). At day 3, villus height was reduced (P < 0.05) in the TPN jejunum compared with day 3 TEN values, andday 3 PEN was similar to both. However, by day 7, relative to day 7 TEN, villus height was lower (P < 0.05) in the TPN and PEN jejunum, which were similar to each other. Jejunum crypt depth was similar among baseline and day 3 TEN, TPN, and PEN groups, although by day 7, crypt depth was significantly lower in both parenteral groups compared with day 7 TEN values. A trend of decreased villus height was seen in the ileum of both parenteral groups compared with TEN values across time. Day 3 ileal crypt depth was lower (P < 0.05) in both parenteral groups compared withday 3 TEN values and remained reduced (P < 0.05) at day 7 relative to TEN values. Crypt depth tended to be greater (P < 0.1) in the day 3 andday 7 TEN colon relative to all other groups, which were similar throughout the study.
Relative to day 0 (baseline) values, jejunal epithelial conductance was increased (P < 0.05) at day 3in TEN animals, whereas colonic epithelial conductance increased (P < 0.05) twofold over the 7-day period of growth (Table 4). At day 3, the measures of epithelial conductance in the TPN and PEN jejunum and colon were similar to TEN values, but by day 7, epithelial conductance in the TPN jejunum was fivefold greater (P< 0.05) than day 7 TEN values. This progression was not observed in the PEN jejunum. Colonic epithelial conductance did not vary among treatments within days.
T cell populations.
CD8+ and CD4+ T cell counts in TEN ileal villi and crypts as well as in the colon did not vary significantly fromday 3 to day 7 (Fig.1). However, numbers of each T cell subtype increased significantly in the TPN ileum (except crypt CD4+ T cells) and colon by day 3 (Fig. 1,A and C–F) and remained elevated (P < 0.05) at day 7 relative to TEN values. With the exception of the elevated CD8+ T cell counts in the PEN ileum (villi and crypts) at day 7, T cell counts were similar in the TEN and PEN ileum and colon throughout the study. Crypt CD4+ T cell counts were numerically but not statistically greater in the TPN and PEN ileum relative to the TEN ileum (Fig. 1 E). Crypt CD8+ T cell counts atdays 3 and 7 in the TPN ileum and at day 7in the PEN ileum were significantly greater than TEN values (Fig.1 B).
The numbers of villus CD8+ and CD4+ T cells were significantly greater at both day 3 and 7 in the TPN ileum relative to the PEN ileum (Fig. 1, A andD). Crypt T cell counts were numerically lower in the PEN ileum at both day 3 and 7 relative to the corresponding TPN ileum values; however, these differences were not significant (Fig.1, B and E).
Quantitative RT-PCR of the proinflammatory cytokines TNF-α and IFN-γ.
At day 3, both TNF-α and IFN-γ mRNA concentrations were numerically lower in the TPN and PEN ileum but not statistically different from TEN values (Fig. 2). Ileal TNF-α mRNA concentrations were similar among the three diet groups atday 7. IFN-γ mRNA concentrations were numerically greater (P < 0.1) in the day 7 TPN ileum relative to TEN values, whereas in the day 7 PEN ileum, IFN-γ mRNA expression was similar to TEN values.
PCNA staining for proliferating cells.
PCNA+ epithelial cells were localized in intestinal crypts, and their numbers were not affected by mode of nutrition during the first 3 days of treatment (Fig. 3). Byday 7, the number of PCNA+ crypt cells was significantly elevated in the TEN ileum relative to baseline (day 0). For TPN, a progression toward reduced epithelial proliferation was noted, in that the number of PCNA+ crypt cells was numerically lower at day 3 and significantly reduced atday 7 relative to TEN values (Fig. 3). Day 3 TEN and PEN values were similar, although by day 7, the number of PCNA+ cells in the PEN ileum was also significantly decreased compared with day 7 TEN values. The threefold difference in the number of PCNA+ cells at day 7between the TEN and TPN ileum was associated with both a greater number of PCNA+ cells in the TEN ileum and a lower value in the TPN ileum. An intermediate number of PCNA+ cells was observed in the day 7 PEN ileum relative to day 7TPN and TEN values.
TUNEL assay for apoptotic cells.
Few TUNEL-positive epithelial cells were observed in ileal crypts regardless of treatment. At day 3, apoptotic crypt cell counts were similar in the TEN, TPN, and PEN ileum and not different from baseline. Apoptotic crypt cell counts in the TEN ileum progressively increased (P < 0.1) from baseline today 7, but this trend did not reach significance (Fig.4). The numbers of apoptotic cells in TPN and PEN ileal crypts were similar over time.
Chromogranin A staining for enteroendocrine cells.
Similar numbers of villus enteroendocrine cells were seen in the baseline and day 3 TEN ileum; however, a trend of fewer villus enteroendocrine cells was observed in the day 7 TEN ileum compared with all other groups, which appeared similar over time (Fig. 5). The number of enteroendocrine cells in ileal crypts was similar among treatment groups and did not vary from baseline over the 7-day study period.
Goblet cell populations.
Representative histochemical staining of goblet cells expressing acidic sulfo- and sialomucins and neutral mucins in the SI is shown in Fig.6. Villus sulfomucin goblet cell counts were similar in the baseline and day 3 TEN jejunum but decreased (P < 0.05) in the day 7 TEN jejunum (Table 5). Day 3villus sulfomucin goblet cell counts in the TPN jejunum were significantly greater than day 3 PEN values; however, TEN counts were intermediate to both TPN and PEN values. By day 7, villus sulfomucin goblet cell counts were greater (P < 0.05) in the TPN jejunum relative to day 7TEN and PEN values, which were similar. Similar numbers of crypt sulfomucin goblet cell counts were observed among treatment groups over time in the jejunum. Also, little variation was seen in jejunal villus sialomucin goblet cell counts among treatment groups. Crypt sialomucin goblet cell counts were similar in the baseline, day 3, andday 7 TEN jejunum. By day 3, crypt sialomucin goblet cell counts in the TPN jejunum increased (P < 0.05) twofold compared with TEN values, and PEN values were intermediate to TEN and TPN. At day 7, the number of crypt sialomucin goblet cells in the TPN jejunum remained elevated (P < 0.05) relative to TEN values and was significantly greater than PEN values. Day 7 crypt sialomucin goblet cell counts in the TEN and PEN jejunum were similar.
Villus neutral goblet cell counts were decreased (P < 0.05) in the day 3 and day 7 TEN jejunum compared with baseline values but were similar to one another. At day 3, villus neutral goblet cell numbers were similar among TEN, TPN, and PEN; however, day 7 TPN villus neutral goblet cell counts were greater (P < 0.05) than both day 7TEN and PEN values, which were statistically equivalent. The number of crypt neutral goblet cells in the jejunum was similar among treatment groups.
Villus sulfomucin goblet cell numbers were similar in the TEN ileum atdays 3 and 7 (Table 5). At day 3, villus sulfomucin goblet cell counts were doubled (P < 0.05) in the TPN ileum and had increased by 50% in the PEN ileum compared with TEN values, such that day 3 PEN values were statistically intermediate to TEN and TPN. At day 7, villus sulfomucin goblet cell counts remained elevated (P < 0.05) in the TPN ileum relative to the TEN ileum. Villus sulfomucin goblet cell counts in the PEN ileum at day 7 were numerically intermediate to day 7 TEN and TPN values, although not statistically different from either group. After 1 wk, crypt sulfomucin goblet cell counts were significantly lower in the TPN and PEN ileum compared with day 7 TEN values. Villus sialomucin goblet cell numbers were increased (P < 0.05) in the TPN ileum at days 3 and 7 compared with TEN ileum counts, which were similar to PEN ileum values. Crypt sialomucin goblet cell counts were similar in the baseline, day 3, and day 7 TEN and PEN ileum; however, crypt sialomucin goblet cell counts were significantly increased in theday 3 TPN ileum compared with day 3 TEN values.
Villus neutral goblet cell counts were similar in the baseline andday 3 TEN ileum but were significantly decreased in theday 7 TEN ileum relative to baseline values (Table 5).Day 7 villus neutral goblet cell counts doubled (P < 0.05) from day 3 values in the TPN ileum, although they were not significantly different from day 7TEN ileum values. Day 7 villus neutral goblet cell counts in the PEN ileum were fourfold greater (P < 0.05) than day 3 PEN values and were significantly greater than day 7 TEN values. Crypt neutral goblet cell counts were significantly decreased from baseline values in the TEN ileum atday 7. Lower crypt neutral goblet cell counts were apparent (P < 0.05) in the day 3 TPN ileum compared with day 3 TEN values, although by day 7, crypt neutral goblet cell counts in the TPN ileum had normalized to day 7 TEN levels. At day 3, the PEN ileum had fewer (P < 0.05) crypt neutral goblet cells compared with the day 3 TEN ileum but had similar values to the day 3 TPN ileum. By day 7, differences were not seen in crypt neutral goblet cell counts among groups.
Sulfomucin goblet cell counts were similar in the baseline, day 3, and day 7 TEN colon (Table 5). At day 3, sulfomucin goblet cell levels were similar in the TEN, TPN, and PEN colon as well. By day 7, only the PEN colon demonstrated reduced (P < 0.05) sulfomucin goblet cells compared with day 7 TEN values. Sialomucin goblet cells were not found in the baseline colon but appeared by day 3 in the TEN colon and were similar to day 7 TEN values. Differences in sialomucin goblet cell counts were not seen among the day 3TEN, TPN, and PEN colon. By day 7, sialomucin goblet cell counts were increased significantly in the PEN colon compared with the TEN and TPN colon. After 1 wk, the TPN colon had significantly reduced sialomucin goblet cell counts compared with day 7TEN values.
The present data demonstrate acidomucin goblet cell expansion in the SI of PEN-nourished neonatal piglets concomitant with a reduction in epithelial cell proliferation before significant increases in T cell populations. These observations indicate that acidomucin goblet cell expansion is a distinct, T cell-independent epithelial event associated with parenteral nutrition, because goblet cell expansion was observed in the SI of both parenterally fed (TPN and PEN) groups. Thus acidomucin goblet cell expansion may reflect a compensatory mechanism to bolster barrier function during states of reduced epithelial renewal.
Numerous studies have reported the modulation of goblet cell number and mucin composition and secretion by a diverse range of luminal insults including varied diet, surgery, and altered microbiota (13, 37,46, 47, 51). Previously, we demonstrated increased numbers of acidomucin goblet cells in the ileum of neonatal piglets maintained on TPN for 1 wk (22). Our current observations of fewer villus acidomucin goblet cells in the PEN relative to TPN SI agree with studies in rats demonstrating the attenuating effects of enteral nutrients on the TPN intestine (43). However, providing 20% enteral nutrients to neonatal piglets for 7 days had little effect on SI weight, DNA and protein content, morphology, and S-phase activity, indicating the inadequacy of 20% enteral nutrition to fully prevent the adverse effects of parenteral nutrition on the SI (2). Also, several studies have reported altered intestinal microbial profiles by restriction or variation of enteral nutrients (10, 31, 34, 47). In this regard, we observed selection for the opportunistic pathogen Clostridium perfringens and other mucolytic bacterial species in the ileum of TPN-fed neonatal piglets (11). Enrichment of mucolytic bacteria in an intestine lacking exogenous nutrients may potentiate mucus degradation and pose a greater threat to the underlying epithelium, because mucins likely provide the primary nutrient source for microbial metabolism with parenteral feeding. Goblet cell expansion could, in turn, counter increased mucolytic activity through regeneration of the mucosal barrier. Also, acidomucins (primarily sulfomucins) restrict mucolysis due to their resistance to enzymatic degradation (20, 42). For example, sulfated terminal carbohydrate moieties were shown to the inhibit growth ofPseudomonas aeruginosa in vitro relative to nonsulfated carbohydrates (6). Together, these observations characterize acidomucin goblet cell expansion as a multifaceted defense that is upregulated in the parenterally nourished intestine.
The proinflammatory cytokines TNF-α and IFN-γ are pleiotropic, in that they increase epithelial cell DNA synthesis (57) and paracellular permeability by weakening epithelial tight junctions, separately and synergistically (36, 45, 56), and IFN-γ also promotes intestinal epithelial cell migration (12). Although elevated T cell counts were observed in the parenterally fed ileum, IFN-γ mRNA expression was only numerically greater in the TPN ileum after 1 wk relative to TEN values, and TNF-α mRNA concentrations remained similar to TEN values throughout the study. The T cell and cytokine data together indicate that goblet cell expansion is independent from direct (cytokines) or indirect (surrounding cells responding to cytokines) T cell influence. Nonetheless, epithelial permeability was increased in association with parenteral nutrition in the SI and therefore may reflect barrier compromise.
Indeed, our current and previous (22) observations of CD8+ and CD4+ T cell expansion in the TPN SI predict antigen translocation across the mucosal barrier. Mucosal barrier integrity depends on epithelial renewal (epithelial proliferation and migration; Refs. 8 and 49) and epithelial tight junctions. Parenteral nutrition has been associated previously with reduced intestinal epithelial cell proliferation and migration in rats (26, 41) and piglets (Ref.2 and unpublished results), potentially increasing the risk of barrier compromise. Here, tissue conductance (a common index of epithelial permeability) was increased fivefold in the TPN jejunum compared with TEN values at day 7, indicating a significant compromise in barrier function. These data are consistent with studies on TPN-fed rats (9, 19) and reflect changes in the paracellular route because nearly 85% of tissue conductance in the SI occurs by this pathway (15a). In addition, Deitch et al. (9) demonstrated TPN-induced loss of intestinal barrier function contributing to Escherichia coli translocation in rats. Others have correlated bacterial translocation with TPN-fed animal models as well (15, 33, 38). Conversely, epithelial permeability in the PEN jejunum at day 7 was significantly less than TPN values, emphasizing the moderating effects of enteral nutrients on the parenterally fed intestine. Similarly, Sax et al. (43) showed reduced bacterial translocation and improved intestinal permeability in rats receiving 25% PEN vs. TPN alone. Colonic epithelial permeability did not differ significantly among groups, indicating that colonic barrier function was unaffected by parenteral nutrition. Although T cell counts were elevated in the TPN colon relative to TEN values, they were only one-tenth that of TPN ileal villus values and not suggestive of an overt inflammatory response. The colon's resistance to the adverse effects of TPN may relate to its greater density of acidomucin goblet cells relative to the SI (27), a phenotype associated with greater microbial density (10, 35, 42).
The mechanisms regulating epithelial proliferation and lineage commitment in the intestine are incompletely characterized. Cheng and Leblond (7) proposed a model in which “oligomucous” goblet cell progenitors are derived from undifferentiated columnar stem cells located in the crypt base. More recently, separate columnar and oligomucous progenitor pathways in the SI have been described, of which the columnar lineage is favored, accounting for the predominance of villus columnar epithelial cells (1). In support of distinct secretory and absorptive lineages, expression of the basic helix-loop-helix transcription factor, Math1, was shown to be requisite for secretory cell lineage commitment (55). Acidomucin goblet cell expansion could be explained by increased proliferation of oligomucous progenitor cells, reduced goblet cell loss, or both. However, our observations suggest that goblet cell expansion did not result from increased proliferation of oligomucous progenitors, because epithelial cell proliferation was numerically decreased by day 3 in TPN piglets and crypt goblet cell numbers were similar among groups at day 3, independent of diet. Yet the coincidence of reduced epithelial proliferation and goblet cell expansion highlights the possibility that the goblet cell lineage is favored when epithelial renewal is reduced. Also in support of this possibility, in contrast to goblet cells, enteroendocrine cell counts were similar among treatment groups over time, although a trend of decreasing villus enteroendocrine cell counts in the day 7TEN ileum was observed.
Apoptosis tended to be greater in epithelial crypts in theday 7 TEN ileum relative to baseline; however, the numbers of apoptotic crypt cells in the TPN and PEN ileum were similar to baseline throughout the study. This raises the possibility that “selective sparing” (52) of oligomucous progenitors from apoptosis (over other lineages) might explain goblet cell expansion. Indeed, the goblet cell product intestinal trefoil factor (ITF), which has been shown to inhibit apoptosis of intestinal epithelial cells in vitro (28, 50), may promote selective sparing of goblet cells. Increased ITF mRNA expression has been positively correlated with reduced epithelial proliferation and greater goblet cell retention in the jejunum of methotrexate-treated rats (52). In addition, previous observations of decreased epithelial migration in TPN piglets (Ref. 2 and unpublished results) and similar crypt goblet cell counts among groups at day 3 support the hypothesis that goblet cell expansion results from a longer residence time for villus goblet cells rather than greater influx of newly differentiated goblet cells. Evidence for greater goblet cell residence time has been reported in rabbit distal intestine and was attributed to slower goblet cell migration relative to columnar cells (23, 24). The combination of selective sparing and reduced goblet cell turnover during times of reduced epithelial renewal would clearly enrich the intestinal epithelium with apoptosis-resistant cells that are also crucial for maintenance of barrier function.
In summary, acidomucin goblet cell expansion appears to be a distinct epithelial response that precedes T cell expansion in the SI of parenterally fed neonatal piglets. We propose that acidomucin goblet cell expansion represents a primary defense triggered by compromised epithelial renewal to prevent barrier failure in the intestine (Fig.7). This mechanism helps to maintain mucosal barrier homeostasis through enriching the epithelial monolayer with longer-lived, apoptosis-resistant cells, which contribute sulfomucins that resist bacterial enzymatic degradation.
We thank Dr. Bart Deplancke and Dr. Peter Reeds for their review of the manuscript. We thank Dr. Frederico Zuckermann for kindly providing the anti-pig CD4 and anti-pig CD8 antibodies.
↵* J. E. Conour and D. Ganessunker contributed equally to this work.
This work was supported in part by United States Department of Agriculture (USDA) Agricultural Experiment Station projects ILLU-50–0343 (S. M. Donovan) and ILLU-35–0914 (H. R. Gaskins) and USDA National Research Initiative grant 98–35206–6429 (H. R. Gaskins).
Address for reprint requests and other correspondence: H. R. Gaskins, Univ. of Illinois, 1207 W. Gregory Drive, Urbana, IL 61801 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
July 11, 2002;10.1152/ajpgi.00097.2002
- Copyright © 2002 the American Physiological Society