Intravenous or luminal amino acids are insufficient to maintain pancreatic growth and digestive enzyme expression in the absence of intact dietary protein

Abstract

We previously reported that rats receiving total parenteral nutrition (TPN) undergo significant pancreatic atrophy characterized by reduced total protein and digestive enzyme expression due to a lack of intestinal stimulation by nutrients (Baumler MD, Nelson DW, Ney DM, Groblewski GE. Am J Physiol Gastrointest Liver Physiol 292: G857–G866, 2007). Essentially identical results were recently reported in mice fed protein-free diets (Crozier SJ, D'Alecy LG, Ernst SA, Ginsburg LE, Williams JA. Gastroenterology 137: 1093–1101, 2009), provoking the question of whether reductions in pancreatic protein and digestive enzyme expression could be prevented by providing amino acids orally or by intravenous (IV) infusion while maintaining intestinal stimulation with fat and carbohydrate. Controlled studies were conducted in rats with IV catheters including orally fed/saline infusion or TPN-fed control rats compared with rats fed a protein-free diet, oral amino acid, or IV amino acid feeding, all with oral carbohydrate and fat. Interestingly, neither oral nor IV amino acids were sufficient to prevent the pancreatic atrophy seen for TPN controls or protein-free diets. Oral and IV amino acids partially attenuated the 75–90% reductions in pancreatic amylase and trypsinogen expression; however, values remained 50% lower than orally fed control rats. Lipase expression was more modestly reduced by a lack of dietary protein but did respond to IV amino acids. In comparison, chymotrypsinogen expression was induced nearly twofold in TPN animals but was not altered in other experimental groups compared with oral control animals. In contrast to pancreas, protein-free diets had no detectable effects on jejunal mucosal villus height, total mass, protein, DNA, or sucrase activity. These data underscore that, in the rat, intact dietary protein is essential in maintaining pancreatic growth and digestive enzyme adaptation but has surprisingly little effect on small intestinal mucosa.

the exocrine pancreas synthesizes digestive enzymes that are essential for digestion and absorption of dietary carbohydrate, lipid, and protein. Despite the low rate of cell turnover in adult pancreas, the organ is remarkably adaptable to changes in macronutrient intake and has a notably high level of plasticity with regard to digestive enzyme expression (4). Maintenance of pancreatic mass and digestive enzyme content is dependent on the presence of nutrients in the digestive tract, indicating that diet plays a central role in the regulation of these processes (1, 9, 21, 22).

It was long held that pancreatic growth, enzyme synthesis, and zymogen granule exocytosis are primarily stimulated during the intestinal phase of digestion by duodenal fat and peptides inducing cholecystokinin (CCK) release (36). In vivo administration of the CCK receptor antagonist L364,718 was found to block the trophic effects of exogenous injections of CCK in rats, guinea pigs, and hamsters but had no effect on normal pancreatic growth and maintenance in the absence of exogenous CCK (40). A number of studies utilizing the CCK null mouse model have since shown that the pancreatic growth response to dietary protein and amino acids does not require CCK (5, 18, 25), indicating that additional nutrient-sensing pathways likely exist.

With respect to intracellular signaling pathways that modulate acinar protein synthesis, CCK, acetylcholine, and bombesin are known to activate the nutrient-sensing kinase mammalian target of rapamycin (mTOR), via phosphoinositide 3-kinase (PI3K) (2, 38), and, moreover, mTOR activation is required for pancreatic growth and cell division in mice (6). Two of the downstream effectors of mTOR, eukaryotic initiation factor 4E binding protein (eIF4E-BP) and the ribosomal protein S6, are phosphorylated in response to CCK and modulate cap-dependent protein translation in isolated acinar cells. Inhibition of mTOR with rapamycin prevented eIF4E-BP and S6K phosphorylation and protein synthesis (3).

In HeLa cells, amino acids directly modulate the calcium-sensing receptor on the plasma membrane to cause a rise in intracellular calcium and thereby activate mTOR complex 1 via calmodulin and the class III PI3K human vacuolar protein sorting 34 (hVps34) (17), opening the possibility that amino acids in the blood may directly modulate the pancreatic mTOR pathway. Indeed, Sans et al. (35) found that oral gavage of leucine led to an increase in phosphorylation of the translation initiation factors eIF4E-BP1 and S6K in both wild-type and CCK-deficient mice, and this response was inhibited by rapamycin. This same effect was also seen in isolated acinar cells, indicating that branched-chain amino acids can directly activate the mTOR pathway independent of hormone and neurotransmitter synthesis. More recently, Crozier et al. (5) demonstrated that placing mice on a protein-free diet induced significant pancreatic atrophy that was reversed by protein refeeding. Surprisingly, mTOR inhibition by rapamycin attenuated pancreatic regrowth in response to dietary protein refeeding but had no effect on pancreatic digestive enzyme synthesis. These results indicate that the dietary regulation of exocrine pancreatic digestive enzyme synthesis and tissue growth may occur through distinct signaling cascades.

Early evidence that pancreatic digestive enzyme synthesis undergoes differential regulation according to specific changes in macronutrient intake includes studies demonstrating that ingestion of diets that are high in protein, fat, or starch increases protease, lipase, and amylase expression, respectively (11, 16, 31–33). Thus the exocrine pancreas is able to recognize specific macronutrients and regulate digestive enzyme synthesis accordingly. Nutrient sensing is mediated at least in part through signals that originate in the intestine; however, whether there is also direct action of nutrients on acinar cells via the bloodstream is less clear and may be species dependent. The direct action of glucose on acinar cells was demonstrated by a significant increase in pancreatic amylase following intraperitoneal injections or continuous parenteral infusion of glucose in rats (16, 26, 28). Research is less clear on the mechanism by which dietary fat stimulates pancreatic lipase synthesis. Studies have shown that dietary fat results in the secretion of secretin (13), which has been shown to increase pancreatic lipase content (30, 36). Interestingly, however, continuous infusion of intravenous (IV) lipid in rats also resulted in an increase in lipase content in pancreatic tissue (26), suggesting that lipid stimulation of the pancreas is not entirely dependent on an intestinal factor such as a gastrointestinal (GI) hormone. In contrast to carbohydrate and lipid, the ability of protein to stimulate pancreatic protease synthesis may be dependent on the presence of intact protein in the digestive lumen. Intraperitoneal administration or continuous IV infusion of amino acids in rats had no effect on pancreatic chymotrypsinogen or trypsinogen levels (26, 28). However, these studies are ambiguous because in each study rats were maintained with a 15% casein diet, making it unclear whether amino acids in the bloodstream may alter pancreatic protease levels in the complete absence of intact dietary protein.

A recent study by Crozier et al. (5) demonstrating that a protein-free diet in mice results in significant pancreatic atrophy was nearly identical to our previous study (1) showing that rats maintained with total parenteral nutrition (TPN) had significant pancreatic atrophy marked by a loss of zymogen granules and digestive enzymes. This led us to question whether substituting luminal or IV amino acids in place of intact dietary protein would also result in pancreatic atrophy and a reduction in digestive enzymes. In the present study we systematically evaluated the role of intact dietary protein in contrast with oral or IV amino acids in the regulation of pancreatic growth, digestive enzyme expression, and small intestinal growth.

METHODS

Materials.

Isoflurane was purchased from Abbott Laboratories (North Chicago, IL). Enterokinase was purchased from Sigma-Aldrich (St. Louis, MO). Trypsin and chymotrypsin substrates were purchased from Peptides International (Louisville, KY). Anti-lipase antibody was a generous gift from Mark Lowe, Washington University, St. Louis, MO. The Phadebas Amylase Test was purchased from Magle (Lund, Sweden). GelCode Blue Stain Reagent was purchased from Thermo-Fischer Scientific (Rockford, IL). Bio-Rad protein assay reagent was purchased from Bio-Rad Laboratories (Hercules, CA). Hoechst fluorescent DNA stain was purchased from Calbiochem (La Jolla, CA). SYBR Green PCR Master Mix, TURBO DNA-free kit, and RETROscript kit were from Ambion (Foster City, CA), TRIzol from Invitrogen (Carlsbad, CA), and the RNAeasy minikit from Qiagen (Valencia, CA). Sprague-Dawley rats were purchased from Harlan (Madison, WI). Diets were purchased from Harlan Teklad (Madison, WI). Amino acid and electrolyte mix, 8.5% Travarsol, and dextrose were from Baxter (Deerfield, IL), and triglyceride lipid emulsion (Intralipid) was purchased from Pharmacia (Clayton, NC). Vitamins were purchased from Astra USA (Westborough, MA), and trace elements (Multitrace-4) were from American Regent Laboratories (Shirley, NY).

Animals, experimental design, and diet composition.

The University of Wisconsin-Madison Institutional Animal Care and Use Committee approved the animal facilities and protocols. Male Sprague-Dawley rats initially weighing 150–175 g were housed individually in stainless steel cages and acclimated to the facility at 22°C and a 12:12-h light-dark cycle for 6 days before the start of the experiment. Animals had ad libitum access to a pelleted purified diet similar to AIN-93G (8) during this period and unlimited access to water for the duration of the experiment.

Rats were randomized to five treatment groups and fed the indicated diet for 7 days: a complete oral diet as a positive control (n = 7), TPN negative control (n = 7), protein-free diet (n = 7), amino acid-defined diet (n = 7), and IV amino acids with orally consumed protein-free diet (n = 7) (Table 1). All groups underwent surgery for jugular catheter placement after an overnight fast. Those groups not receiving TPN or IV amino acids received saline infusions. Two sets of experiments were initiated on separate dates and utilized animals within a similar weight range. The first set of experiments included the positive oral and TPN negative control animals, a protein-free group, and an IV amino acid group. The second set of experiments was conducted at a later date and contained a TPN control group and a group receiving oral amino acids along with fat and carbohydrate.

Table 1.

Feeding paradigms

Group	Protein	Amino Acids	Fat	CHO
Oral control	Oral		Oral	Oral
TPN control		IV	IV	IV
Protein free			Oral	Oral
Oral amino acids		Oral	Oral	Oral
IV amino acids		IV	Oral	Oral

TPN, total parenteral nutrition; IV intravenous; CHO, carbohydrate.

Three different oral diets were utilized in this experimental design: a complete oral control diet, a protein-free diet, and a defined-amino acid diet. The control diet was designed based on AIN-93G with casein for the protein component, cornstarch, maltodextrin, and sucrose for the carbohydrate component, and soybean oil for the lipid component. The protein-free diet was matched to the control diet on all parameters including energy density, with, however, the protein calories substituted exclusively with cornstarch. The defined-amino acid diet was also matched to the control diet on all parameters except that the protein source was a mixture of amino acids that exactly matched the profile of those provided in the daily TPN provisions. The control diet had 176.1 g protein/kg, while the amino acid diet had 181.0 g of equivalent protein/kg. All rats that received an oral diet were given ∼12 g of food per day to match the previously determined reduction in the ad libitum intake of the protein-free diet; 12 g of the defined-amino acid diet contained the same amount of amino acids present in the daily infusion of TPN and amino acids. Food intake was measured daily. All diets were prepared in pellet form by Harlan Teklad.

The TPN, amino acid solution, and saline were prepared for infusion aseptically. The TPN consisted of a commercial preparation of amino acids and electrolytes, 60% dextrose, 20% long-chain triacylglycerol lipid emulsion, and all essential vitamins and trace elements as detailed previously (8). The TPN solution consisted of (in g/l) 44 amino acids, 180 dextrose, and 28 lipid. For the amino acid infusion group, the same commercial amino acid solution used in the TPN was diluted with saline to provide an identical quantity of amino acids and fluid volume per day as in the TPN group. These rats, however, ate the protein-free diet, and thus had fat and carbohydrate presented to the body via the digestive tract. The infusion of TPN, amino acids, or saline began on day 0 at 1 ml/h and was increased to 1.6 ml/h after 12 h. The final goal rate of 2.2 ml/h was achieved 24 h after surgery to provide full nutritional needs. Daily infusion volumes were recorded for calculation of nitrogen provisions. Animals were euthanized and tissues harvested beginning at 9 AM on day 7.

Pancreatic and jejunal mucosal mass and cellularity.

The pancreas and jejunum (defined as from the ligament of Treitz to 25 cm proximal to the cecum) were removed, flushed with saline at 4°C, blotted dry, and weighed intact. A 5-cm section of the proximal jejunum was cut longitudinally and scraped to obtain the mucosa for determination of wet and dry mass. The pancreas and a second 5-cm (in succession) jejunal mucosal scraping were immediately frozen in liquid nitrogen and stored at −80°C for later analysis. Approximately 300 mg of each frozen pancreas as well as the second 5-cm jejunal mucosal scraping were homogenized in ice-cold buffer containing (in mM) 12 Tris (pH 7.1), 300 mannitol, and 5 EGTA. Protein and DNA concentrations of pancreas and mucosal homogenates were determined by Bio-Rad assay and a fluorometric method (24), respectively.

Lipase expression.

Pancreatic proteins (200 μg) were separated by SDS-PAGE, immobilized to nitrocellulose membranes, blocked in Tris-buffered saline (pH 7.4) containing 5% nonfat dry milk and 3% Tween 20, and incubated with the indicated antibodies for 90 min at room temperature. Immunoreactivity of anti-lipase (1:1,000) or heat shock cognate 70 (HSC70, 1:500) was detected with horseradish peroxidase-conjugated donkey anti-rabbit (1:2,500) and enhanced chemiluminescence. Protein expression was quantified by scanning densitometry using a PDI model DNA 35 scanner interfaced with the protein and DNA image system.

Enzyme assays.

Pancreatic amylase activity was determined by the Phadebas blue starch test. To generate active, measurable trypsin and chymotrypsin, diluted pancreas homogenates (1:25) were incubated with enterokinase (1 ng/μl) for 1 h at 37°C. Samples were loaded onto a 96-well plate, along with trypsin assay buffer containing (in mM) 50 Tris (pH 8.1), 150 NaCl, and 1 CaCl₂ and a fluorescent trypsin substrate (t-butyloxycarbonyl-l-glutaminyl-l-alanyl-l-arginine-4-methyl-coumaryl-7-amide) or a fluorescent chymotrypsin substrate (succinyl-l-alanyl-l-alanyl-l-prolyl-l-phenylalanine 4-methyl-coumaryl-7-amide) at a final concentration of 150 μM. Fluorescence was measured at 355 nm (excitation) and 460 nm (emission) for 10 readings over 10 min, and activity/time was corrected for the amount of total protein in the sample.

Sucrase activity was measured in jejunal homogenates by an established method (7). Sucrase specific activity was determined by calculating the molar amount of glucose produced per minute divided by the protein concentration of the sample.

Pancreas and jejunal mucosal histology.

A portion of each pancreas and 1 cm of jejunum was fixed in 10% buffered formalin, paraffin embedded, and sectioned (5 μm) and stained with hematoxylin and eosin (H & E) for histological and trypsin autofluorescence analysis. Zymogen granules in pancreas sections were evaluated by eosin autofluorescence (37, 39) on a Nikon TE 2000 brightfield microscope using a Hamamatsu Orca Camera, a Plan APO 100X/1.4 oil objective, and Volocity software with excitation and emission at 565 and 617 nm, respectively. Digital images of jejunal mucosal sections were captured with a ×10 objective. Jejunal villus height and crypt depth were measured on at least 10 crypt-villus axes per animal with SigmaScan software (Jandel Scientific, San Rafael, CA).

mRNA expression.

Total RNA was prepared from freshly harvested pancreas with TRIzol reagent and chloroform extraction. RNA was further purified with the RNAeasy mini kit. RNA integrity was confirmed by gel electrophoresis. Trypsin mRNA expression was measured with a two-step reverse transcriptase quantitative polymerase chain reaction (RT-qPCR) with the SYBR Green detection method. RNA was treated with DNase to eliminate genomic DNA and reverse transcribed with random hexomers and the RETROscript kit. cDNA was diluted to 1 ng/μl, and real-time qPCR reactions were made with the SYBR Green PCR Master Mix. Sequences for the forward and reverse trypsinogen primers, respectively, were as follows: 5′-TCT GAT CCT AGC CCT TGT-3′ and 5′-GAT GCG GGA TTT GTA GCA G-3′. Reactions were cycled with an Applied Biosystems 7000 Real-time PCR instrument with the following parameters: step 1, 50°C for 2 min; step 2, 95°C for 10 min; step 3, 40 cycles of 95°C for 15 s, followed by 60°C for 1 min; and step 4, 95°C for 15 s, followed by 60°C for 20 s and 95°C for 15 s (dissociation). GAPDH qPCR was performed under similar conditions and did not change based on the treatment group. Equal efficiency of amplification was verified by a serial dilution technique. Data were analyzed with the 7000 system software, and relative quantification was done with the ΔΔC_t method (where C_t is threshold cycle) with GAPDH as the internal control and the oral control group as the reference control.

Statistical analysis.

SAS (version 9.2, SAS Institute, Cary, NC) and R (version 2.9, Vienna, Austria) were used for the statistical analysis. Differences among treatment groups were evaluated by using multiple comparisons based on the general linear model procedure in SAS. All data are presented as group means ± SE; P ≤ 0.05 was considered statistically significant. Significant differences among treatment groups are denoted by different lower case letters in results and Figs. 1, 2, 4, 6, and 7.

Fig. 1.

Daily weight change, energy intake, and nitrogen provision. A: daily body weight changes calculated from days 2–5. Note that body weight changes were consistent across all groups except those receiving protein-free diets. B: energy intake and nitrogen provision. Note that energy intake in the total parenteral nutrition (TPN) and intravenous (IV) amino acid (AA) groups are elevated because of the constant infusion of nutrients. Values are means ± SE; n = 7 rats/group. Means with different superscripts are significantly different based on a P value ≤ 0.05. CHO, carbohydrate.

Fig. 2.

Intact dietary protein is essential for normal pancreatic growth. Pancreatic wet mass (A), protein (B), and DNA (C) were analyzed on day 7 of each feeding paradigm. Note that the protein-free diet results in significant pancreatic atrophy that is not fully compensated for by oral or IV amino acids. Values are means ± SE; n = 7 rats/group. Means with different superscripts are significantly different based on a P value ≤ 0.05. TBW, total body weight.

Fig. 4.

Oral or luminal amino acids only partially prevent TPN and protein-free diet-induced reductions in digestive enzyme expression. A: amylase activity was measured in pancreatic homogenates and expressed relative to total protein content. B and C: trypsin (B) and chymotrypsin (C) enzyme activity were measured after enterokinase treatment of pancreatic homogenates and expressed relative to total protein content. D: lipase expression was measured by immunoblotting and normalized to heat shock cognate 70 (HSC70) expression, which was unchanged in all treatment groups (not shown). Values are means ± SE; n = 7 rats/group. Means with different superscripts are significantly different based on a P value ≤ 0.05.

Fig. 6.

A 7-day protein-free diet does not cause mucosal atrophy. A: dry weight of a 5-cm jejunal mucosal scraping from each group. B and C: total jejunal mucosal protein (B) and DNA (C) content. Note that the protein-free diet, with or without luminal or IV amino acids, fully prevented the TPN-induced jejunal atrophy. Values are means ± SE; n = 7 rats/group. Means with different superscripts are significantly different based on a P value ≤ 0.05.

Fig. 7.

Jejunal mucosal histology and sucrase activity. A: representative pictures from each treatment group. B and C: cross sections of jejunum were evaluated for villus height (B) and crypt depth (C). Note the reduction in villus height in the TPN group compared with all other groups. D: jejunal mucosal scrapings were also analyzed for sucrase activity. A protein-free diet, with or without luminal or IV amino acids, prevented the TPN-induced reduction in jejunal mucosal sucrase activity. Values are means ± SE; n = 7 rats/group. Means with different superscripts are significantly different based on a P value ≤ 0.05.

RESULTS

Body weights and energy intake.

Studies were conducted with five treatment groups of rats (Table 1), all of which were subjected to surgical cannulation. Two sets of experiments were carried out starting on different days (see methods). Within each set of experiments, there were no significant differences in mean body weight among the treatment groups before surgery. However, mean body weight of rats in the first set of experiments was 10 g greater compared with the second set, and this led to a difference in final body weights when the two sets of experiments were compared. Final body weights in the first set of experiments were as follows (means ± SE): oral control, 189 ± 4^a; TPN control, 183 ± 4^a; protein free, 156 ± 3^b; and IV amino acids, 181 ± 2^a (means with different superscripts are significantly different based on a P value ≤ 0.05). Final body weights in the second set of experiments were as follows (means ± SE): TPN, 170 ± 4^a; oral amino acids, 169 ± 1^a. Aside from the rats in the protein-free group, which were in negative nitrogen balance, there were no significant differences in final body weights among groups in experimental block 1 or groups in experimental block 2. Likewise, daily weight gain among all groups, except the protein-free group, was the same (Fig. 1A). Mean energy intake was significantly greater in the TPN control and IV amino acid groups compared with the oral control, protein-free control, and oral amino acid groups because of the constant infusion of nutrients (Fig. 1B, gray bars). Calculation of nitrogen provision per day among groups indicated a slight but significant reduction in the IV amino acid group versus oral control animals and no nitrogen intake for the protein-free group (Fig. 1B, black bars).

Oral or IV amino acid feeding causes pancreatic atrophy.

In agreement with previous reports (1, 9, 19, 23) pancreatic wet mass expressed as a percentage of total body weight was significantly reduced by 25% in the TPN control group compared with oral control animals (Fig. 2A). Pancreatic mass was also significantly reduced by ∼15% in the protein-free, oral amino acid, and IV amino acid groups compared with oral control animals. Conversely, pancreatic mass of protein-free, oral amino acid, and IV amino acid groups was significantly greater than the TPN control animals. For comparison, no differences in heart mass were determined among all treatment groups (data not shown).

As previously demonstrated (1), pancreatic protein concentration was significantly reduced by 18% in the TPN control group compared with oral control animals. This reduction was also seen in the protein-free and oral amino acid groups; however, the IV amino acid group was unchanged from either oral or TPN control animals (Fig. 2B). Also consistent with previous findings (1, 19, 20), pancreatic DNA levels, when normalized to pancreatic weight, were significantly enhanced in the TPN control group compared with all other groups. However, this result was not obtained when pancreatic DNA levels were expressed as total DNA per pancreas, likely reflecting that the reduction of pancreatic weight elevated the DNA concentration (Fig. 2C).

Oral and IV amino acids are insufficient for optimal digestive enzyme expression.

It was previously shown that autofluorescence of eosin Y in H & E-stained pancreatic sections provides an excellent method for selectively identifying acinar cell zymogen granules (37, 39). Fluorescence analysis of pancreatic sections confirmed our previous report (1) that TPN induces a striking loss of granular staining in the apical cytoplasm of acini indicating a loss of zymogen granules (Fig. 3 and Supplemental Fig. S1).¹ As recently reported in mice (5), feeding rats a protein-free diet also induces a marked reduction in acinar granule staining. Surprisingly, this loss of granular staining was only partially prevented by providing oral or IV amino acids, suggesting that intact dietary protein regulates zymogen granule biogenesis more than dietary amino acids, fat, and carbohydrate.

Fig. 3.

Oral or IV amino acids partially prevent TPN- and protein-free diet-induced reduction in acinar acidic granule staining. Hematoxylin and eosin-stained pancreatic sections were evaluated by eosin autofluorescence. Note the marked reduction in acidic granular staining in sections from TPN and protein-free diet-fed rats compared with orally fed rats and the more moderate staining in sections from the oral and IV amino acid groups. Representative images from each treatment group are shown. Reference bar, 7 μm. Also see Supplemental Fig. S1 for a wide view of the tissue. DIC, differential interference contrast.

Pancreatic digestive enzyme expression was evaluated by measuring enzyme activity of amylase, trypsin, and chymotrypsin in pancreatic homogenates, whereas lipase levels were determined by immunoblotting (Fig. 4). Compared with oral control animals, TPN animals showed a marked 75% reduction in amylase activity (Fig. 4A). Furthermore, rats fed a protein-free diet also had a large 90% reduction in amylase. Interestingly, providing oral or IV amino acids attenuated the loss of amylase induced by the TPN or protein-free diets; however, activity was still 50% less than oral control animals. Similar to amylase, trypsin activity was significantly attenuated in all groups compared with oral control animals, with the TPN and protein-free groups showing 80% and 67% reductions, respectively (Fig. 4B). Animals receiving oral or IV amino acids had greater trypsin activity than the TPN control animals; however, both groups remained significantly lower than the oral control animals by 50% and 40%, respectively.

In contrast with amylase and trypsin, comparatively modest reductions in chymotrypsin (Fig. 4C) and lipase (Fig. 4D) levels were detected among treatment groups. Indeed, there were no changes in chymotrypsin activity among treatment groups, with the exception of a greater than twofold increase in the TPN group compared with all other groups, the explanation for which is unclear (see discussion). Pancreatic lipase levels were significantly reduced by 15% in the TPN control animals compared with oral control animals (Fig. 4D). Similarly, rats fed a protein-free diet or receiving oral amino acids each had 20% reductions in lipase expression compared with oral control animals; however, lipase expression in the IV amino acid group was unchanged from oral control animals and significantly elevated compared with TPN control animals.

To determine whether the reduction in digestive enzyme expression was a result of transcriptional changes, qPCR was conducted to measure trypsinogen mRNA levels (Fig. 5). In contrast to the pronounced reduction in trypsin activity seen in all groups compared with oral control animals, little or no change in mRNA was detected across treatment groups. Indeed, the only statistically significant difference was seen between the oral amino acid and IV amino acid groups, in which trypsinogen mRNA levels increased by ∼175%. However, neither the oral amino acid nor the IV amino acid groups were significantly different from the oral or TPN control groups. Consistent with a recent study in mice maintained on protein-free diets (5), these data suggest that the reduction in trypsin levels caused by a lack of dietary protein occurs mainly at the translational level.

Fig. 5.

Pancreatic trypsinogen mRNA expression. Trypsinogen mRNA expression was measured by real-time quantitative RT-PCR. Data were analyzed with the ΔΔC_t method (where C_t is threshold cycle) with GAPDH RNA as the internal control and expressed relative to the oral control group. Values are means ± SE; n = 3–7.

Dietary fat and carbohydrate prevent mucosal atrophy seen during TPN.

It is well recognized that resting the GI tract during TPN or fasting results in pronounced atrophy of the intestinal mucosa; thus the effects of these dietary regimens on the cellularity and morphology of the small intestinal mucosa were examined. Jejunal mucosa was utilized to represent the small intestinal response because the jejunum is the most active region of dietary absorption. As expected, jejunal mucosal dry mass was significantly decreased by 25% in TPN control rats compared with oral control animals (Fig. 6A). Surprisingly, jejunal dry mass in rats receiving protein-free or oral or IV amino acid diets was unchanged from oral control animals, indicating that luminal fat and carbohydrate are sufficient to support normal mucosal growth for 7 days even with a complete lack of nitrogen intake.

Jejunal protein content was also significantly reduced in the TPN control group compared with all other treatment groups (Fig. 6B). Moreover, no changes in protein content were noted in the protein-free diet or oral or IV amino acid groups compared with oral control animals. Similarly, jejunal mucosal DNA content was significantly reduced in the TPN control group compared with all other treatment groups (Fig. 6C). Although the mucosal DNA content in the protein-free and IV amino acid groups was unchanged from oral control animals, it was significantly increased in the oral amino acid group by 125% of the oral control animals.

Morphological analysis of jejunal mucosa confirmed the TPN-induced atrophy as evidenced by reduced villus height compared with all other groups (Fig. 7, A and B). Rats maintained with the protein-free, oral amino acid, or IV amino acid regimens had normal measures of villus height. Determination of crypt depth revealed that the protein-free group showed a modest but significant reduction compared with all groups except for the IV amino acid group, which was unchanged from all groups (Fig. 7C). Finally, mucosal sucrase activity was also measured as an indicator of mature enterocyte function, indicating that TPN induced a significant decrease in enzyme activity (Fig. 7D). Conversely, sucrase activity for all other treatment groups was significantly higher than the TPN control group but unchanged from oral control animals. Collectively, these data indicate that in the absence of intact dietary protein, luminal carbohydrate and fats are sufficient to sustain jejunal mucosal mass, protein, DNA, villus height, and sucrase activity for up to 7 days.

DISCUSSION

The exocrine pancreas is known to adapt morphologically and biochemically to both the composition and route of macronutrient intake (4). Findings that a nutritionally complete diet administered by TPN induces a reduction in pancreatic mass, total protein, and digestive enzyme expression demonstrate that the presence of nutrients in the intestinal lumen is essential to maintain normal pancreatic growth and digestive enzyme expression. The present results (see Table 2 for a summary of results) showing that luminal stimulation with fat and carbohydrate either alone or in combination with a complete source of amino acids was insufficient to sustain normal levels of pancreatic mass, protein and zymogen granule expression emphasize the critical importance of intact dietary protein for optimal pancreatic growth and maintenance. Moreover, the differential effects on digestive enzyme expression observed when providing luminal fat and carbohydrate in the absence or presence of intact protein, oral amino acids, or IV amino acids clearly solidify the notion that macronutrients generate specific neural and/or hormonal signals within the intestine that regulate pancreatic adaptation to the diet.

Table 2.

Summary of results

	TPN	Protein Free	Oral AA	IV AA
Pancreatic mass, mg/g body wt	↓↓	↓	↓	↓
Pancreatic protein, mg/g pancreas	↓	↓	↓	—
Pancreatic DNA, mg/g pancreas	↑↑	—	—	—
Pancreatic DNA, mg/pancreas	—	↓	↓↓	—
Pancreatic amylase, U/μg protein	↓↓↓	↓↓↓	↓↓	↓↓↓
Pancreatic lipase (lipase/HSC70)	↓	↓	↓	—
Pancreatic trypsin, U/μg protein	↓↓↓	↓↓↓	↓↓↓	↓↓
Pancreatic chymotrypsin, U/μg protein	↑	—	—	—

Difference from oral control: single arrows, <25%; double arrows, ≥25%; triple arrows, ≥50%; dashes, no change. AA, amino acids; HSC70, heat shock cognate 70.

Consistent with the present results, previous studies with rats maintained with TPN reported a 19–25% reduction in pancreatic mass relative to body weight, with 40–50% reductions in total pancreatic protein (9, 19). With respect to digestive enzyme expression, Fan et al. (9) found a 50% reduction in amylase, a 20% decrease in lipase, and, surprisingly, a >50% increase in both trypsinogen and chymotrypsinogen following 7 days of TPN. In contrast, Helton et al. (19) found no change in amylase, a 50% reduction in lipase, and a 40% reduction in trypsinogen after 7 days of TPN. The present data show more dramatic reductions in both amylase and trypsinogen expression with TPN but a more modest reduction in lipase. This discrepancy may be attributed to experimental differences including initial size of the rats, assay methods, and TPN components. Indeed, Helton et al. (19) provided no lipid in the TPN solution, possibly explaining the greater reduction in pancreatic lipase content.

The present data indicate that chymotrypsinogen expression was twofold greater in TPN rats compared with oral control rats. These results are consistent with Fan et al. (9); however, these authors also reported increases in trypsinogen as a result of TPN. Because TPN rats had no luminal stimulation by fat or carbohydrate it is conceivable that there was less ongoing pancreatic secretion that could potentially elevate chymotrypsinogen storage. However, amylase, lipase, and trypsinogen content were all reduced in rats receiving TPN, suggesting rather that rates of synthesis and/or the half-life of these enzymes are different.

Hara et al. (18) reported that rats fed amino acids derived from casein hydrolysates showed less induction of pancreatic protease expression than those fed a high-casein diet. These findings, together with recent evidence that a protein-free diet in mice results in pancreatic atrophy (5), led us to question whether intact dietary protein is the primary signal responsible for normal maintenance of pancreatic growth and digestive enzyme expression. Providing an isocaloric protein-free diet had effects similar to TPN, decreasing mass, protein, and digestive enzyme content. Most interesting, providing a nutritionally complete source of amino acids to keep the animals in positive nitrogen balance allowed us to demonstrate that luminal stimulation with fat and carbohydrate significantly increased both amylase and trypsinogen expression; however, this luminal stimulation did not fully compensate for the lack of intact dietary protein in maintaining pancreatic mass or trypsinogen or amylase expression. In contrast, two previous studies investigating the influence of oral or IV amino acids on pancreas found no effect on trypsinogen or amylase expression (26, 28). Our experimental design is distinct from these earlier studies because our rats were maintained on diets completely devoid of intact protein, whereas their rats were provided a 15% casein diet when IV or intraperitoneal amino acids were administered. Thus it is unlikely that amino acids would stimulate pancreatic growth in the presence of intact dietary protein.

Compared with the dramatic reductions in amylase and trypsinogen, relatively modest changes were detected in chymotrypsinogen and lipase expression with TPN and protein-free diets compared with oral control animals. With the exception of the large TPN-induced increase in chymotrypsinogen (see above), no changes in this protease were observed when providing luminal fat and carbohydrate regardless of nitrogen status. Lipase expression was significantly reduced from oral control animals in animals given TPN or fed protein-free or oral amino acid diets, suggesting that luminal fat was not sufficient to maintain lipase expression. However, lipase expression was unchanged from oral control animals when providing IV amino acids. This may reflect that concentrations of various amino acids in the total nitrogen pool differed between the oral and IV amino acid groups, or that lipase expression was enhanced by the significantly greater energy intake in the IV amino acid group.

The signal by which dietary protein stimulates pancreatic growth and digestive enzyme synthesis in rats was long accepted to be CCK. This was largely due to studies demonstrating that diets high in intact protein or containing soybean trypsin inhibitor resulted in a large increase in plasma CCK and stimulated hyperplastic pancreatic growth (15, 27). It is hypothesized that high levels of intact protein or trypsin inhibitors saturate or inhibit protease activity, respectively, and thereby stimulate CCK release by blocking a negative feedback response from proteases in the intestine (15). The development of CCK knockout mice has since shown that CCK is not essential for either normal pancreatic growth and development or adaptation to dietary protein (25). Thus it appears that intact protein does not exclusively signal through CCK, suggesting that additional neural and hormonal signals participate in orchestrating dietary adaptation of the pancreas.

In addition to the central role of dietary protein, insulin together with elevated blood glucose is also known to stimulate amylase expression (4), and insulin has been shown to strongly stimulate protein synthesis in rat acini (3). Interestingly, insulin levels in this TPN model were previously shown to be elevated compared with postprandial levels obtained from animals maintained on complete oral diets or intragastric infusion of TPN solution (29). Additionally, dietary fat is also known to cause a major increase in plasma secretin and a modest elevation of plasma CCK (13), both of which stimulate pancreatic growth. Thus the normal digestion seen in animals provided with oral or IV amino acids together with dietary carbohydrate and fat suggests some likelihood that signaling via secretin and autonomic nerves was intact in these studies. Because in the rat dietary protein is a major stimulus for CCK secretion, it is unclear whether a lack of CCK release prevented full adaptation when nitrogen was administered in the form of amino acids. Protein refeeding studies in CCK knockout mice (see above) suggest this is not the case (25). Clearly, additional studies using these feeding paradigms in CCK knockout mice should be helpful in approaching this question.

As with the neural and hormonal signals mediating diet-induced pancreatic adaptation, the molecular signaling events that regulate acinar cell digestive enzyme expression are also uncertain. Crozier et al. (5) examined pancreatic growth in mice fed protein-replete chow after a 4-day protein-deficient diet and showed that dietary protein rapidly induced regrowth and digestive enzyme expression. Digestive enzyme adaptation to protein occurred at the translational level, as mRNA levels were unchanged (5). In accordance with these results, we found no significant reduction in trypsinogen mRNA levels among treatment groups by qPCR. Strikingly, Crozier et al. (5) demonstrated that inhibition of mTOR-regulated protein translation with rapamycin attenuated regrowth of the pancreas but did not inhibit the increases in digestive enzyme expression, suggesting that digestive enzyme expression and acinar growth occur through independent mechanisms. Recent studies indicate that cap-dependent protein translation occurs through two distinct mTOR protein complexes, mTORC1 and mTORC2, the latter of which is resistant to rapamycin inhibition (10). Additional studies utilizing more potent active-site inhibitors of mTORC2 should be helpful in identifying potential roles for these complexes in acinar cell adaptation and growth.

Our data demonstrating reduced dry mass, protein, DNA, and sucrase activity in the jejunum as a result of TPN is in agreement with previous studies (8, 20). Interestingly, no difference was found in mucosal dry mass, protein, DNA, or sucrase activity in rats fed a protein-free diet or when amino acids were provided, indicating that luminal carbohydrate and fat can maintain normal mucosal growth for up to 7 days. These data have potential clinical implications when avoiding pancreatic stimulation but maintaining optimal mucosal function is desirable. Furthermore, evidence that trypsinogen expression is greatly diminished when nitrogen is provided in the form of amino acids may be particularly relevant in the treatment of pancreatitis, where trypsin activation may play a pivotal role in the onset of the disease (34). In rodents, the pancreas is less responsive to cerulein in stimulating growth and digestive enzyme synthesis on a low-protein diet compared with a normal or high-protein diet (14), and protein-free diets are protective against experimental pancreatitis (12). Furthermore, we recently demonstrated (23) that rats given TPN were protected against secretagogue-induced pancreatitis. Clearly, further studies identifying both the luminal signals originating in the intestine as well as molecular mechanisms that mediate pancreatic growth and adaptation should provide valuable insight into potential therapeutic strategies aimed at resting the pancreas while maintaining optimal nutritional status.

GRANTS

This work was supported by National Institutes of Health (NIH) Grant DK-07088 and U.S. Department of Agriculture HATCH Grant WISO4958 to G. E. Groblewski and NIH Grant R01-DK-07665 to D. M. Ney. M. D. Baumler and M. C. Koopmann were supported by NIH Training Grant T32 DK-007665.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).