Compensatory changes in villus morphology of lactating Mus musculus in response to insufficient dietary protein Free

Flexibility of the gastrointestinal tract in response to diet has been demonstrated in a variety of vertebrates including snakes, birds and mammals (Secor et al., 1994; Piersma and Lindström, 1997; Starck, 1999; Naya et al., 2007). Flexibility in intestinal morphology may result in changes in nutritional physiology, which in turn can impact reproduction (Raubenheimer et al., 2009). Changes in the width and length of the small intestine in mammals have generally been linked to changes in the amount of ingested food, such as occurs in response to the energetic demands of cold temperatures or lactation, with small intestine mass increasing as the amount of food consumed increases (Boyne et al., 1966; Cripps and Williams, 1975; Derting and Noakes, 1995; Hammond and Kristan, 2000; del Valle et al., 2004; Jaroszewska and Wilczyńska, 2006). Hammond (1997) concluded that lactating female mice increase the size of the small intestine to match the increase in energy demands. A meta-analysis by Naya et al. (2007) demonstrated that in laboratory mice and rats, flexibility of the small intestine was common and was significantly associated with lactation and diet quality. However, Naya's analysis of diet quality was limited by available research in that the only measure of diet quality was variation in fiber content.

Dietary protein is a primary component of diet quality, which is critical to successful growth and reproduction (Bomford and Redhead, 1987; Cameron and Eshelman, 1996; Goettsch, 1960; McAdam and Millar, 1999; Veloso and Bozinovic, 2000; Nel et al., 2015; Bauer et al., 2009). Dietary protein deficiency reduces milk protein levels in mice; however, the reduction in milk protein does not match the reduction in dietary protein consumption, suggesting that mice may have compensatory mechanisms to support milk protein levels (Derrickson and Lowas, 2007). Derrickson (2013) found that when fed isocaloric diets differing in protein, mice consume equal amounts of calories, yet the gastrointestinal tract morphology changes. Lactating mice on a low protein isocaloric diet exhibit gross morphological changes including increased stomach mass, increased duodenal diameter and increased intestinal mass; these changes are similar to the lactation-induced changes in the intestine of mice (Elias and Dowling, 1976; Hammond and Diamond, 1994; Hammond, 1997), but appear to be exhibited to a greater degree on the low protein diet.

In this study, we hypothesized that at the tissue level, the villus exhibits morphological changes in response to low dietary protein similar to the morphological changes induced by lactation. The increase in villus size under conditions of low protein would increase functional capacity, potentially providing greater availability of a nutrient required for successful growth and reproduction. We determined the effect of changing dietary protein levels on villus structure along the length of the small intestine and compared those changes in lactating and non-reproductive mice. Because of the increased protein demands of reproduction and the associated potential negative fitness costs of protein deficiency on pup growth (Derrickson and Lowas, 2007; Chen et al., 2009), we predicted that the effects of dietary protein on intestinal flexibility would be greater in lactating compared with non-reproductive females, with lactating females on the low protein diet exhibiting the greatest increase in villus size. We also predicted that changes in villus structure would vary along the length of the small intestine in a manner similar to that shown during lactation. Specifically, we predicted that any increases in villus height in the duodenum also would be due to changes in enterocyte size (hypertrophy), as suggested by Prieto et al. (1994) for lactating rats. Finally, we compared surface area of lactating and non-reproductive mice on isocaloric low and high protein diets to determine the degree to which dietary protein insufficiency could lead to compensatory changes in surface area.

Twenty primaparous female mice, 6 months of age, ICR strain of Mus musculus, were randomly placed on either the 11.5% protein (low protein, LP) or 23% protein (control protein, CP) isocaloric complete diets containing approximately 15.7 kJ g⁻¹ produced by Harlan Teklad (Madison, WI, USA). Decreases in protein levels were balanced calorically by modifying sucrose levels while keeping fat levels relatively constant (see Derrickson, 2013, for diet composition). The CP diet contained the protein level found to be optimal for growth and reproduction in laboratory mice (Sørenson et al., 2010; Hoevenaars et al., 2012) and was designated as the control diet. The LP diet has previously been shown to result in differing milk protein levels and pup growth rates (Derrickson and Lowas, 2007), and differences in gut morphology (Derrickson, 2013). Ten mice were randomly chosen and paired with males 1 week after placement on the diet while the remaining mice made up the group of non-reproductive females; mice in the four groups were not significantly different in mass at the start of the experiment. Males and females were paired for 4–6 days, the length of one estrus cycle. Female mass, pup mass and food consumption were measured twice weekly throughout the experiment. Lactating females were killed at day 16 of lactation when pups were beginning to ingest solid food; non-reproductive females were killed after an equivalent time period of 6 weeks on the diet. Mice previously were shown to exhibit no difference in food consumption on the two diets (Derrickson, 2013).

Mice were dissected directly after euthanasia. Mass measurements were recorded for the whole animal and for the small intestine after contents were removed by flushing with ice-cold physiological saline. The length of the small intestine was measured by laying the intestine flat with minimal stretching. Sections of 1 cm length were taken from the proximal, mid and distal sections of the intestine for fixation in buffered 10% formalin. For non-reproductive mice, sections were taken from 0–1, 15–16 and 29–30 cm along the small intestine, which averaged 44 cm in length. The small intestines of reproductive mice were longer in length (54 cm); as a result, sections were taken at comparable regions of 0–1, 18–19 and 36–37 cm. Thus, the proximal section was within the duodenum (region proximal to the entrance of the pancreatic and hepatic ducts), whereas the other two sections were taken at approximately one-third and two-thirds of the length of the jejenum/ileum and will be referred to as mid and distal small intestine. The treatment of mice and collection of samples were approved by the Institutional Animal Care and Use Committee of Loyola University Maryland.

Samples were embedded, sectioned at 5 μm thickness, and stained with Hematoxylin and Eosin. Villus height and width, crypt depth and width, intestinal perimeter, and the height and width of enterocytes were measured using ImageJ (Fig. S1). Ten measures of villus height, crypt depth, villus apex and base width were obtained from each individual for each of the three regions; we were not able to obtain 10 measures for all location/individual combinations because of incomplete villi and torn intestines on some samples, but a minimum of 8 measurements was achieved for each villus character except at the distal location, leading to a range of 6–10 measurements for these villus characters in the distal region. Four of the 20 individuals had 0–6 measures for the distal region, one individual had no measures for the mid region and one individual had no measures for the proximal region. In total we used 516 measures each for villus height, crypt depth, villus apex width and villus base width. Enterocyte width was estimated by measuring the total width of a row of 10 enterocytes in the top one-third of the villus at each of the three locations along the intestine. Enterocyte length was measured in each of three cells in the top one-third of the villus at each of the three locations along the intestine.

Surface area of the 10 lactating females was compared across diet and location using the method of Kisielinski et al. (2002), which was modified in two ways. The model of Kisielinski calculates the surface magnification ratio of the unit, M, where a unit consists of the villus surface area and adjacent crypt region. M thus compares the surface area of a unit with the villus present with the surface area if the unit did not contain a villus. The calculation of M assumes a cylindrical villus evenly spaced across an area. In our calculation, we compared the magnification achieved around a unit-wide circumference of the small intestine. We thus modified the method of Kisielinski et al. (2002) by adjusting M by a factor of the perimeter divided by the diameter of the unit bottom to account for changes in the number of villi as the perimeter of the small intestine varies across regions. In addition, because villus shape varies across regions, the average values for apex width and base width were used to calculate an overall average villus width per sample location.

Statistical analyses were performed using commercially available software (JMP Pro version 13.0; SAS Institute Inc., Cary, NC, USA). Repeated measures analyses were performed for villus height with multiple measures on the same individual. Because each of the 20 individuals provided multiple measures for villus height, the repeated measures within each mouse can be used to estimate the random effect within mice, which in turn can be used to partition the variance component into a within and between mouse component. Diet, reproductive status (lactating, L; or non-reproductive, NR) and location (proximal, mid and distal intestine) were included as fixed effects and all interaction terms were entered into the model. F values are reported as F_{fixed effect d.f.,d.f.Den}, where d.f.Den is the estimated degrees of freedom after the random effect variance component is accounted for using the restricted maximum likelihood (REML) method. Fewer replicates were measured for enterocyte width and length; these variables were analyzed by intestinal location with a full model containing diet and reproductive status as main effects.

Surface area effects were analyzed only in lactating females; females on the LP diet had significantly greater intestinal surface area (27%; F_1,23=7.03, P=0.014) than lactating females on the CP diet. The interaction between diet and location was also significant (F_2,23=3.89, P=0.035), with peak surface area magnification exhibited in the proximal small intestine in mice in the LP group, but in the mid intestine in mice in the CP group. Lactating females on the LP diet tended to exhibit intestinal perimeters that were 5–10% greater, depending on the region, than those on the CP diet, but these differences did not rise to significance in any region of the small intestine. This suggests that villus height rather than the number of villus units is primarily responsible for increased surface area.

Villus characteristics varied significantly across the length of the intestine, with reproductive status and with diet. Location strongly affected villus height, with villus height decreasing by nearly 60% from the proximal region to the distal region (F_2,333.6=330, P<0.0001; least-squares means, lsmeans: proximal 386 µm, mid 346 µm and distal 156 µm). Lactating mice overall had villi that were 50% taller than those of non-reproductive mice (F_1,189.5=175, P<0.0001). Diet impacted villus height, with mice on the LP diet exhibiting an 8% increase in villus height across status and location (F_1,189.5=6.93, P<0.01).

Lactation interacted with intestinal location in that villus height increased most in the proximal and mid regions of the small intestine (F_2,333.6=36.3, P<0.0001; Fig. 1). Diet interacted with intestinal location in that the decline in villus height from proximal to distal region was greater in mice on the LP diet than in those on the CP diet (F_2,333.6=14.6, P<0.0001; Fig. 1). Diet, reproductive status and location interacted with villus height (F_2,333.6=12.6, P≤0.0001) with villus height greatest in the proximal region of lactating mice on the LP diet (Fig. 2). Lactating mice on the LP diet exhibited a 24% increase in villus height in this region compared with that in lactating mice on the CP diet. Overall, these results suggest that the impact of diet on villus height differs from the impact of lactation.

Enterocyte width was affected by diet and reproductive status, but not consistently across regions of the intestine. In the proximal region, the most significant effect on enterocyte width was the effect of diet (F_1,19=13.62, P=0.002). Mice on the LP diet exhibited an 18% increase in enterocyte width compared with those on the CP diet. The effect on enterocyte width was even greater in lactating females: females on the LP diet had cells that were 30% wider than those of females on the CP diet (F_1,19=4.64, P=0.04; Fig. 3). Enterocyte width in the mid region of the small intestine was not influenced by diet but was influenced by reproductive status, with enterocyte width increasing in the mid region of the intestine by 24% in lactating mice compared with non-reproductive mice (F_1,17=8.72, P=0.01). The increase in width of the enterocytes translates directly into increased height of the villus as enterocyte height is oriented perpendicular to villus height. In the proximal villus, we saw a 24% increase in villus height, and in that same intestinal region, we saw a similar 29.7% increase in enterocyte width as measured in the top one-third of the villus.

Enterocyte height was affected by both reproductive status and diet, but the pattern of change appeared to differ across the length of the intestine. In the proximal region, enterocyte height was affected by reproductive status, with lactating females presenting enterocytes that were 34% taller than those of non-reproductive females (2.15 µm versus 1.61 µm; F_1,16=5.16, P=0.04). In the distal region, diet influenced enterocyte height (F_1,13=5.24, P=0.04) with mice on the LP diet possessing cells that were 16% taller. This effect was more pronounced in lactating females on the LP diet, which exhibited a 30% increase in enterocyte height in the distal region compared with both non-reproducing females and lactating females on the CP diet (interaction of status and diet, F_1,51=4.88, P=0.05).

The results presented here are consistent with previous findings in our lab at the whole-organ level, which demonstrated that lactating mice on a low protein isocaloric diet exhibit increased tissue production in the region of the proximal small intestine, despite consuming less food (Derrickson, 2013). We could find only one previous study (Campbell and Fell, 1964) that indicated that villus height might increase in response to nutrient deficiency; in that study, villi in the duodenal region of lactating mice qualitatively appeared to increase under food restriction. The increase in intestinal surface area shown in this study provides further support that increased investment in intestinal tissue may occur to compensate for dietary protein deficiency by potentially increasing assimilation efficiency. This result provides further support that the impact of limited dietary protein on milk protein may be ameliorated by compensatory intestinal responses (Derrickson and Lowas, 2007).

Animals have particular intake targets for macronutrients (Raubenheimer and Simpson, 1997), and for mice, that target appears to be about 23% crude protein (Sørenson et al., 2008), which is the level experienced by mice fed our CP diet. When carbohydrate and protein are available in unbalanced ratios, evidence suggests that protein intake regulation takes precedence (Sørenson et al., 2008; Huang et al., 2013). In this study, mice were not able to construct their diet from different food options, but they may have been able to rebalance their intake through increased assimilation of proteins. The changes in morphology were greatest in the proximal region of the small intestine, suggesting that this location may be a target for a compensatory response to decreased protein. However, this study did not address whether the increase in villus size and surface area results in increased uptake of protein or whether the uptake of all nutrients is enhanced (Clissold et al., 2013).

Much research on dietary plasticity has focused on aminopeptidase (e.g. Sabat et al., 1999; Naya et al., 2008a; Zhang et al., 2016), a protein-digesting enzyme found throughout the small intestine (Miura et al., 1983; Naya et al., 2008a). Our finding that protein deficiency localizes morphological flexibility to the proximal small intestine suggests that enterokinase, an enzyme expressed at highest levels in the most proximal part of the duodenum (Nordström and Dahlqvist, 1971; Yuan et al., 1998), may be an important actor in dietary flexibility. Enterokinase is a protease that activates trypsinogen, which subsequently activates other pancreatic enzymes. Enterokinase is specifically localized in the enterocytes (Yuan et al., 1998). Upregulation of enterokinase potentially could improve the efficiency of protein digestion through its direct and indirect activation of other proteases. The increased enterocyte size may reflect the more active production of this (or other) enzymes, thus providing an avenue for lactating female mice to compensate for ingesting a low protein diet.

Our isocaloric protein-restricted diet had higher carbohydrate levels than the control diet in order to maintain energy content per gram of food. The increases that we observed in villus height and enterocyte width in the small intestine may have been a matching response to increased carbohydrate presence rather than a compensatory response to low protein levels. However, Sørenson et al. (2010) (in growing male mice) and Derrickson (2013) (in lactating mice) both found compensatory intestinal responses in situations in which the study animals had elevated protein demands. Similarly, at the cellular level, Diamond and Karasov (1987) found that amino acid transport was upregulated both at very low protein levels and at high protein levels, suggesting that mechanisms exist to match increasing intake and to compensate for deficient intake. Together, these results provide support for the conclusion that deficient protein levels, rather than excess carbohydrate levels, are responsible for the increases that we observed in villus height and enterocyte width in response to the LP diet.

Understanding how dietary composition impacts nutritional physiology has broad implications as human impacts on the environment affect both the quantity and quality of food resources (Birnie-Gauvin et al., 2017). Widely distributed mammals are suggested to exhibit greater flexibility in physiological functions (Bozinovic et al., 2011; Naya et al., 2008b) and some, like M. musculus, successfully reproduce under many environmental conditions (Perrigo, 1987). This study demonstrates that M. musculus, which in captivity has undergone selection for increased reproductive effort, exhibits growth in villus height and intestinal surface area in response to dietary deficiency in protein during lactation. This specific flexibility in the proximal small intestine may provide a mechanism that could allow species to persist in the face of changing dietary profiles, both naturally occurring and anthropogenically induced.

We would like to thank Dr Chris Thompson who was invaluable in helping us to acquire the images and Dr Dave Rivers for use of his scope for the ImageJ analyses. Scott Deamond was instrumental in helping with animal husbandry. Finally, we greatly thank Dr Chris Morrell for his statistical expertise with the repeated measures analysis.