Expansion of intestinal stem cells associated with long-term adaptation following ileocecal resection in mice

Christopher M. Dekaney, Jerry J. Fong, Rachael J. Rigby, P. Kay Lund, Susan J. Henning, Michael A. Helmrath

Abstract

Sustained increases in mucosal surface area occur in remaining bowel following massive intestinal loss. The mechanisms responsible for expanding and perpetuating this response are not presently understood. We hypothesized that an increase in the number of intestinal stem cells (ISC) occurs following intestinal resection and is an important component of the adaptive response in mice. This was assessed in the jejunum of mice 2–3 days, 4–5 days, 6–7 days, 2 wk, 6 wk, and 16 wk following ileocecal resection (ICR) or sham operation. Changes in ISC following ICR compared with sham resulted in increased crypt fission and were assayed by 1) putative ISC population (SP) by flow cytometry, 2) Musashi-1 immunohistochemistry, and 3) bromodeoxyuridine (BrdU) label retention. Observed early increases in crypt depth and villus height were not sustained 16 wk following operation. In contrast, long-term increases in intestinal caliber and overall number of crypts per circumference appear to account for the enhanced mucosal surface area following ICR. Flow cytometry demonstrated that significant increases in SP cells occur within 2–3 days following resection. By 7 days, ICR resulted in marked increases in crypt fission and Musashi-1 immunohistochemistry staining. Separate label-retention studies confirmed a 20-fold increase in BrdU incorporation 6 wk following ICR, confirming an overall increase in the number of ISC. These studies support that expansion of ISC occurs following ICR, leading to an overall increase number of crypts through a process of fission and intestinal dilation. Understanding the mechanism expanding ISCs may provide important insight into management of intestinal failure.

  • intestinal stem cells
  • intestinal resection
  • crypt fission
  • label retention

intestinal adaptation is a compensatory physiological response that occurs following the loss of mucosal surface area, and is important in restoring the absorptive and digestive capacity. Animal models used to study adaptation demonstrate early increases in proliferation of crypt epithelium leading to enhanced crypt depth, villus height, microvillus surface area, and functional absorptive capacity per unit length of intestine to compensate for the loss of mucosal surface area. Strong evidence supports a role of luminal nutrients (9, 12, 26), gastrointestinal secretions (1, 45), mesenchymal (38), neuronal (39, 40), and humoral factors (41, 45, 46) in regulating the adaptive response (27). A majority of studies in animal models have focused on early adaptive growth responses after bowel resection. Given that sustained adaptive increases in mucosal mass are essential to normal function, a better understanding of the time course and long-term outcome of intestinal adaptation is desirable.

Clinical experience suggests that intestinal adaptation occurs in humans. Many patients with intestinal failure experience decreases in diarrhea and dependence on parenteral nutrition with an apparent increase in nutrient absorption over the first 1–2 years following intestinal loss. These changes are often accompanied by a marked increase in the caliber of the intestine (26). Despite these observations, adaptive changes in intestinal morphology and epithelial homeostasis leading to sustained increases in mucosal surface area are not well characterized in humans. Only a few small studies exist in which morphometric data from humans with intestinal failure have been evaluated. None of these studies have demonstrated changes in proliferation, apoptosis, or crypt depth in either the small bowel or colon compared with normal controls (11, 15, 28, 31, 43). Adaptive growth in human intestine after resection may be underestimated due to the length of time (usually greater than 1 year) after surgery that patients are studied. Also there may be a selection bias in clinical studies if growth is evaluated more frequently in patients with clinical signs of malabsorption who may be unable to adapt.

Two potential mechanisms could explain long-term increases in mucosal surface area seen clinically following intestinal resection. The first is that sustained increases in proliferation occur, resulting in continuous increased epithelial production per individual crypt. This potential mechanism would result in sustained augmentation of surface area by increasing the size, not the overall number of crypts and villi. Alternatively, limited increases following intestinal loss in the numbers of proliferating stem cells per crypt may occur, resulting in crypt fission and the expansion of the overall number of crypts and villi without long-term or sustained increases in proliferation. One goal of the present study was to investigate long-term as well as short-term adaptive responses to bowel resection to distinguish between these mechanisms. We also developed and characterized adaptive growth in our new model of ileocecal resection (ICR) in mice. Our rationale in choosing this model is that the loss of the ileum and proximal colon is more commonly associated with human intestinal failure than is the loss of the proximal small bowel.

In small intestine, the epithelial lining is rapidly regenerated every 4–5 days by continuous proliferation of intestinal stem cells (ISC) and transit amplifying cells residing near the base of crypts. Pioneering studies performed in mice by Cheng and Leblond in the early 1970s (7) suggested that ISC are responsible for the generation of all four major lineages of the small intestinal epithelium. The precise number of ISC within each crypt remains a matter of debate. Most studies suggest the number to be four to six (35), although some studies point to the number being as low as one ISC per crypt (14). Kinetic data derived from [3H]thymidine labeling predict that ISC are slowly cycling (24–30 h) cells that give rise to more rapidly cycling progeny by asymmetric division. These progeny are termed transit amplifying cells and typically have a cycle time of 12 h (33). Transit amplifying cells undergo 4–5 rounds of cell division while migrating up the crypt, prior to differentiating into either absorptive (enterocytes) or secretory lineages (mucous, enteroendocrine, and Paneth cells) toward the top of the crypt (47). Paneth cells are the only lineage that migrate down to the base of crypts (4) rather than onto the villi.

This study was designed to test whether ICR affects the stem cell population in the small intestine. We hypothesized that part of the adaptive response following resection involves the expansion of ISC number. By studying the time course of ISC expansion after ICR, we aimed to establish whether ISC expansion is transient or sustained and define the optimal times when clinical interventions may enhance ISC expansion or, in the future, promote survival of transplanted stem cells.

METHODS

Animals and Experimental Design

The Baylor College of Medicine Institutional Animal Care and Use Committee approved the protocol for this study. Male C57BL/6J mice, 8–12 wk old (weight range 25–31 g) were obtained from the Jackson Laboratories (Bar Harbor, ME). The mice were housed in groups of five and randomly assigned to undergo either sham operation (intestinal transection only, n = 4–8/group) or ICR (n = 5–8/group). All mice were switched to a liquid diet (Micro-Stabilized Rodent Liquid Diet LAD 101/101A, Purina Mills, St. Louis, MO) 2 days before the operation and maintained on this diet over the first 7 days of the experiment to avoid intestinal obstruction, as previously described (23). Recent experience with our ICR model has shown that animals can safely be switched back to regular chow on postoperative day 7 without any risk of intestinal obstruction.

Operative Procedure

Mice were weighed, ear tagged, and clipped the day of the procedure. All operations were performed under sterile conditions and with the aid of an operating microscope (×7 magnification) using inhaled 2% isoflurane and oxygen for anesthesia. Through a midline incision the intestines were eviscerated and the ileocecal junction was identified. All mice received a single dose of piperacillin and tazobactam (100 mg/kg ip) at the start of the procedure. In mice undergoing ICR, the small bowel was divided 12 cm proximal to the ileocecal junction and 1 cm distal to the cecum in the ascending colon. The mesentery of the resected intestine was ligated, resulting in the removal of 12 cm of the intervening ileum, cecum, and proximal right colon. Intestinal continuity was restored by using an end-to-end, single layered anastomosis with interrupted 9-0 monofilament sutures. Sham operations consisted of transection and reanastomosis of the bowel 12 cm proximal to the ileocecal junction. All mice were hydrated with warm intraperitoneal saline and the abdomen was closed. Total operative time for sham operations was ∼20 min and ∼45 min for ICR. Liquid diet and water was provided immediately postoperatively.

Tissue Harvest

At 2–3 days, 4–5 days, 6–7 days, 2 wk, 6 wk, and 16 wk following either sham operation or ICR, mice were weighed and then killed between 6 am and 9 am by cervical dislocation while under isoflurane anesthesia. The entire bowel was removed, and the intestinal contents were gently expressed by cotton swabs and saline flush. The distal 12 cm of intestine in sham-operated animals and 1 cm proximal and distal to the anastomosis in all mice were removed (Fig. 1). The next 1 cm of proximal intestine (jejunum) from all mice was immediately fixed with 10% neutral buffered formalin for perpendicular sectioning. In some mice, the preceding 5 cm of intestine was harvested for flow cytometry, whereas in others that 5 cm of intestine was opened along the antimesenteric border, pinned open on paraffin, and fixed in 10% neutral buffered formalin for longitudinal sectioning. Intestinal length was determined at 16 wk following surgery by suspending the entire small bowel, for 1 min with a constant weight and measuring the length.

Fig. 1.

Cartoon depicting the surgical procedures and the collection of intestinal tissue from sham-operated and ileocecal resection (ICR) animals. In animals that underwent resection, 12 cm of the terminal small bowel as well as the cecum were removed and the remaining small bowel and ascending colon were anastomosed. At the time of tissue collection, the distal 12 cm of intestine in sham-operated animals and 1 cm proximal to the anastomosis in all mice were removed and discarded. The next proximal 1 cm and 5 cm were harvested for histological analysis and flow cytometry, respectively.

Histology and Analysis of Intestinal Morphometrics

Sections were stained with hematoxylin and eosin or subjected to immunohistochemistry (as described below). Histological samples from all time points were analyzed in a blinded manner. To determine crypt depth, crypt width, villus height, and villus width at least 10 well-oriented, full-length crypt-villus units were counted and averaged for each sample. Entire circumference of the intestinal submucosa was used as a measure of intestinal caliber. Circumference of submucosa was measured rather than the entire circumference of the outer serosal surface of the bowel because we reasoned that this was a closer measure of mucosal caliber that would not be influenced as much by any adaptive thickening of the enteric smooth muscle layers. Measurements were performed by using AxioVision 4.4 on digital images acquired with an AxioImager upright microscope coupled to an AxioCam MRc5 digital camera. The density of crypts in both perpendicular and longitudinal sections was determined by counting the number of crypts per 500 μm in five well-oriented sections per animal. The number of crypts per circumference was calculated based on the submucosal circumference and perpendicular crypt density. The relative mucosal surface per unit area (M) was calculated from the formula described by Kisielinski et al. (25): Math where VW is villus width, VH is villus height, and CW is crypt width. To determine proliferative index, mice received an injection of bromodeoxyuridine (BrdU; 120 mg/kg ip) 2 h prior to being euthanized. Following routine BrdU immunohistochemistry, proliferative index was determined by calculating the ratio of BrdU-positive cells to total cells within 10 intact crypts.

Crypt fission was also measured as a potentially important mechanism to expand the functional mucosal mass after resection that has not been extensively studied (3). Hematoxylin and eosin stained longitudinal tissue sections were utilized to determine the percentage of crypt fission from at least 100 intact crypts per animal 2–3 days, 4–5 days, 6–7 days, 2 wk, 6 wk, and 16 wk following surgery. A crypt undergoing fission was defined as a bifurcating crypt with a bisecting fissure creating two (or sometimes more) flask-shaped bases with a shared single crypt-villus junction, as illustrated in Fig. 5 (13). Apoptotic index was determined by quantitating the number of apoptotic bodies, defined as cells demonstrating pyknotic nuclei, condensed chromatin, and nuclear fragmentation, within 25 consecutive well-oriented crypts and expressed as a percentage of the total number of cells in the 25 crypts (32).

Quantification of Intestinal Stem Cells

Because there is currently no universally accepted method for measuring ISC, we performed three independent assays to evaluate the hypothesized expansion of ISCs following ICR.

Label retention of BrdU.

Label retention studies were carried out in a separate group of animals (n = 5–8/group) randomly assigned to either sham operation or ICR. Each mouse received injections of BrdU (60 mg/kg ip; ∼1.5 mg/mouse) twice daily (9 AM and 9 PM) for 7 consecutive days immediately postoperatively. Animals were killed at 6 wk after surgery (5 wk “washout” of BrdU label) and routine BrdU immunohistochemistry was performed on longitudinal sections of jejunum. The percentage of BrdU-positive label-retaining stem cells per crypt were determined by quantifying the number of labeled epithelial cells present within the base of 250 well-oriented crypts per animal. All labeled epithelial cells on high power that contained secretory vesicles were considered Paneth cells and were not included in our quantification.

Isolation of SP cells, a putative stem cell population.

Single mucosal cell suspensions were prepared from 5 cm of jejunum harvested at 2–3 days, 4–5 days, 6–7 days, 2 wk, and 6 wk from sham-operated and ICR mice as previously described with minor modifications (10). For these experiments enzymatic digestion of tissue was performed for 10 min instead of 20 min and mechanical disruption was carried out for 5 min rather than 15 min. Time required to isolate viable single cells from the intestine and perform flow cytometry limited the number of animals that could be studied to five per day; therefore groups of mice were harvested 24 h apart to maintain a minimum of n = 5 per group studied. FACS analysis of CD45-negative side population (SP cells) was performed using the Hoechst 33342 staining method reported by Dekaney et al. (10). To determine the number of CD45-negative SP cells per crypt, we first calculated the total number of CD45-negative SP cells by multiplying the total number of cells acquired after enzymatic and mechanical disruption by the percentage of CD45-negative SP cells obtained from flow cytometry. Next, we calculated total crypt number by multiplying crypt density by the length of tissue collected for isolation of CD45-negative SP cells (5–7 cm). Finally, we divided the total number of CD45-negative SP cells by the total number of crypts to obtain total number of CD45-negative SP cells per crypt.

Msi-1 immunostaining.

Tissue sections were subjected to 30-min antigen retrieval in 0.01 M citric acid buffer pH 6.0 at 100°C. Endogenous peroxidase activity was quenched with 3% H2O2 for 8 min at room temperature. Sections were incubated overnight at 4°C with antibody against Musashi-1 (Msi-1) (kindly provided by Dr. Hideyuki Okano, Department of Physiology, Keio University School of Medicine, Tokyo, Japan) diluted 1:500 in PBS. Sections were then incubated with biotinylated-secondary antibody diluted 1:200 in PBS at RT for 30 min. After being rinsed, sections were incubated in avidin-biotin complex (Vector Laboratories, Burlingame, CA) for 45 min at RT. Sections were developed with Nova red as the chromogen and were counterstained with hematoxylin.

Statistical Analyses

All quantitative results are presented as mean values ± SE. For the morphometry data, the analyses were based on the assumption that the 10 crypts and villi used for counting were selected at random. For each animal the 10 observed crypt, villus, and circumference counts were averaged. The average values were used to generate means ± SE for particular treatment group or time points. Means were compared at each time point using ANOVA with correction for multiple comparisons using the Fisher's procedure. For all parameters a P value of P < 0.05 was considered the level of significance.

RESULTS

Ileocecal Resection Model

All mice were healthy and vigorous throughout the study. Survival following sham operation was 95% (78 survived/82 operated) and 91% following ICR (82 survived/90 operated). No animals died of intestinal obstruction as a result of switching from liquid diet to solid chow on postoperative day 7. For the first week following surgery, loose stools were observed in animals undergoing ICR; stool consistency returned to normal by the second week postsurgery. Sham-operated mice initially lost weight up to the 4–5 days time point and then demonstrated linear weight gain up to 6 wk, at which point their weight remained stable. Mice undergoing ICR lost a greater amount of weight following surgery than observed in sham-operated mice. Weight gain was noted in all ICR mice beyond 4–5 days, yet weight gain equivalent to sham-operated mice was not achieved until 16 wk following surgery, thereby defining our long-term time point in this study. Although specific micronutrient measurements were not obtained in either sham-operated or ICR groups, continued weight gain without any mortality seen beyond 2 wk postoperatively in all mice throughout the remainder of the study suggest overall good health.

Figure 2A shows an image of the dissected intestine in animals after ICR or sham surgery to demonstrate the remarkable increase in intestinal diameter at 16 wk after ICR. Figure 2B shows examples of perpendicular sections of the entire bowel wall to illustrate the progressive increase in circumference of the intestine beginning as early as 7 days after ICR.

Fig. 2.

Gross morphology of sham-operated and ICR intestines. A: photograph of 16-wk intestine demonstrating grossly the increase in intestinal diameter in ICR compared with sham-operated mice. Note that the increase in diameter can be seen throughout the length of adapted bowel. B: low-power (×12.5) cross-sectional view of small intestine from sham and ICR mice at various times after surgery.

The length of small bowel was measured in sham-operated and ICR animals at 16 wk after surgery to assess whether lengthening of the small bowel was a component of the adaptive response. Intestinal length from the unoperated group was 30 ± 0.9 cm; therefore linear growth was defined as increases above 30 cm in sham operated. Since we resected 12 cm of ileum in the ICR group, increases above 18 cm in ICR mice were considered evidence of linear growth. By this method, no significant difference in small intestinal linear growth was observed in sham-operated (8.7 ± 0.4 cm) compared with ICR mice (10.5 ± 1.2 cm) by 16 wk following surgery.

Morphometric Measure of Intestinal Adaptation

Measurements of crypt depth (Fig. 3A) demonstrated significant increases within 2–3 days following ICR compared with sham-operated animals, with peak increases in crypt depth occurring in the first week after ICR. Crypt depth remained higher up to 6 wk following ICR compared with sham-operated animals. By 16 wk after surgery crypt depth did not differ significantly between sham-operated and ICR surgery. Significant increases in villus height occurred within 4–5 days following ICR compared with sham-operated mice (Fig. 3B). The overall adaptive trend differed from crypt depth, in that villus height continued to increase until it peaked at 6 wk, but, like crypt depth, villus height returned to baseline by 16 wk following ICR when there was no difference in villus height between ICR and sham-operated controls.

Fig. 3.

Morphometric comparison of intestinal histology from sham-operated and ICR mice. Comparison of small intestinal crypt depth (A), villus height (B), and total submucosal intestinal circumference (C), from sham-operated and ICR mice 2–3 days, 4–5 days, 6–7 days, 2 wk, 6 wk, and 16 wk following surgery (n = 4–8 animals/group). Based on univariate analysis, significant differences (*P < 0.05) between sham operation and ICR are shown.

Measurement of jejunal circumference from perpendicular sections taken from sham-operated and ICR mice at 2–3 days, 4–5 days, 6–7 days, 14 days, 6 wk, and 16 wk following surgery revealed significant increases in ICR animals as early as 4–5 days following surgery (Fig. 3C). Differences between sham-operated and ICR animals became more profound at 6 wk and exceeded twofold by 16 wk. Two-way ANOVA revealed significant influences of time (P < 0.001) and treatment (P < 0.001) and interaction between time and treatment (P < 0.001), highlighting the continued increase in bowel caliber occurring during these later time points.

To evaluate changes in mucosal surface over a unit length of jejunum, we calculated the mucosal surface area using the formula described by Kisielinski et al. (25) (see methods) and found no significant difference in mucosal surface per unit area between sham-operated and ICR animals by 16 wk postsurgery (7.0 ± 0.2 vs. 8.2 ± 0.4). In addition, no differences in crypt depth, villus height, or density of crypts per unit length (41.6 ± 0.7 vs. 40.1 ± 2.1 crypts/mm) were observed. Therefore, the significant (2.5-fold) increase in the total number of crypts per circumference of jejunum accounted for the overall increase in mucosal surface area following ICR (Fig. 4, A and B).

Fig. 4.

Adaptive responses of remnant small bowel to resection. Comparison of the surface area per unit length (A) and number of crypts per circumference (B) in sham and ICR mice 16 wk following surgery. Significant differences (**P < 0.01) are indicated.

Observation of Crypt Fission

Evaluation of histological sections of jejunum from mice 6–7 days following ICR demonstrated numerous crypts undergoing fission (Fig. 5A). Typically, we observed clusters of bifurcated crypts, with often four or five crypts bifurcating within a short distance of each other. Most crypts undergoing fission appeared to be dividing in a symmetrical pattern, from the base up. Figure 5B further demonstrates distinguishing features of crypt fission at higher magnification: a bisecting fissure creating two flask-shaped bases with a single shared opening at the crypt-villus junction. In addition, incorporation of BrdU at the tip of the common fissure demonstrates increased proliferation occurring in the region. Quantitative analysis of the overall percent of crypts undergoing fission at various times following surgery (Fig. 6) showed significant increases in crypt fission at 6–7 days and 14 days following ICR, compared with sham-operated animals. The increase in crypt fission peaks at 6–7 days and was back to baseline levels by 6 wk.

Fig. 5.

Intestinal histology of representative sections of jejunum 7 days following surgery. A: broad region of an hematoxylin and eosin stained section from ICR animal (×200) demonstrating marked crypt fission (arrows). B: comparison of bromodeoxyuridine (BrdU) labeling in crypts in sham vs. ICR (outlined by bracket). Note that cells at the apex of the common wall of fissioning crypts are positive for BrdU label (arrows).

Fig. 6.

Comparison of crypt fission. Percent of crypts undergoing fission within the jejunum from sham-operated and ICR mice 2–3 days, 4–5 days, 6–7 days, 2 wk, 6 wk, and 16 wk following surgery (n = 4–8 animals/group). Based on univariate analysis, significant differences (*P < 0.05) between sham operation and ICR are shown. Two-way ANOVA indicated significant influence of time and treatment (P < 0.001) on crypt fission and a significant interaction of time and treatment (P < 0.001).

Intestinal Stem Cells

Classic experiments demonstrating long-term retention of labeled thymidine analogs ([3H]thymidine or BrdU) within symmetrically dividing stem cells have been used to demonstrate both location and expansion of ISCs during development and following irradiation (7, 36). After incorporation of the analog, the rapid migration of transit amplifying cells from crypts to villi and turnover of differentiated epithelial cells results in the dilution of BrdU label from the majority of epithelial cells by 5 wk after BrdU administration. We observed marked increases in BrdU labeling of epithelial cells located in the base of crypts (stem cell zone) in mice following ICR. This pattern of staining reflects enhanced acquisition of label during symmetric division associated with expansion of new ISCs that retain the BrdU-labeled strand (36). For this study, we defined a positively labeled ISC as an epithelial cell located within the stem cell region at the crypt base without any evidence of secretory vesicles under high-power magnification. Representative photomicrograph of BrdU immunohistochemistry following ICR (Fig. 7A) demonstrates positive BrdU labeling in both a crypt epithelial cell, considered to be an ISC, and a positively labeled Paneth cell. We acknowledge that newly formed Paneth cells in this region may not have secretory vesicles, but given the fact that labeling occurred 6 wk prior to tissue harvest all labeled Paneth cells should be fully differentiated at this time. Using this method, we observed a 20-fold increase in the number of label-retaining crypt stem cells within the jejunum of mice following ICR compared with sham-operated animals (Fig. 7B). Our evaluation of histological sections 6 wk following 1 wk of BrdU administration also demonstrated occasional positive BrdU labeling in submucosal myofibroblast, few hematopoietic cells within villus mesenchyme, and Paneth cells from both sham-operated and ICR mice. This positive labeling in nonepithelial cells is likely due to slow proliferation or incorporation within hematopoietic stem cells and in Paneth cells due to their prolonged estimated half life of ∼21 days.

Fig. 7.

Intestinal stem cell label retention. A: photomicrograph of BrdU immunohistochemistry taken 6 wk following ICR from a mouse that received 7 days of BrdU injections immediately postoperatively, demonstrating the characteristic ISC label retention (black arrow, ×400). A labeled Paneth cell is indicated by the white arrow. B: comparison of the number of non-Paneth cell label-retaining cells (LRC) per 100 crypts from sham-operated (n = 5) and ICR (n = 4) mice 6 wk following surgery. Based on univariate analysis, significant differences (*P < 0.05) between sham operation and ICR are shown.

We employed an independent method to quantify putative ISCs by flow cytometry recently described by our laboratory (10). This method depends on the isolation of a side population (SP) following staining with the Hoechst 33342 dye and exclusion of hematopoietic cells from this population by sorting with the pan-leukocyte marker CD45. Utilizing this technique, we determined the percent of CD45 negative SP cells within the jejunum from sham-operated and ICR mice 2–3 days, 4–5 days, 6–7 days, 14 days, and 6 wk following operation. This study demonstrated a fivefold increase in the percentage of SP cells in mice at 2–3 days following ICR, with significant increases sustained at 4–5 days compared with sham-operated animals (Fig. 8A). Ultimately, as the adaptive response increased the total number of intestinal epithelial cells, the overall percentage of SP cells returned to baseline. To aid in comparison of this data to our finding following label-retention studies, we expressed the number of SP cells on a per crypt basis. In doing this we found that the number of SP cells per crypt was significantly increased in animals undergoing ICR at 2–3, 4–5, and 6–7 days following surgery compared with respective sham-operated animals (Fig. 8B). By 14 days postsurgery the number of SP cells per crypt in ICR animals was not significantly different compared with sham-operated animals.

Fig. 8.

Evaluation of CD45-negative side population (SP cells) following sham or ICR. A: comparison of the percentage of CD45-negative SP cells in jejunum from sham-operated and ICR mice 2–3 days, 4–5 days, 7 days, 14 days, and 6 wk following surgery. Based on univariate analysis, significant differences (*P < 0.05) between sham operation and ICR are shown. Two-way ANOVA indicated significant influence of treatment (P < 0.001) and time (P < 0.004) and a significant interaction between treatment and time (P < 0.003). B: comparison of the number of CD45-negative SP cells per crypt at the same time points as in A. Based on univariate analysis, significant differences (*P < 0.05) between sham operation and ICR are shown. Two-way ANOVA indicated significant influence of treatment (P < 0.001).

As another measure of stem cells, Msi-1 immunohistochemistry (Fig. 9) was performed on histological sections from sham-operated and ICR mice 7 days following surgery. The location of Msi-1 staining was similar in the base of crypts from both sham-operated and ICR mice, with overall stronger staining observed within bifurcating crypts following ICR. In addition, we commonly observed positive Msi-1 staining at the tip of the common wall in bifurcating crypts (Fig. 9), as has been reported by He et al. (20).

Fig. 9.

Representative photomicrographs of Musashi-1 (Msi-1) immunohistochemistry of jejunum from sham-operated and ICR mice at 7 days postsurgery. Note extensive crypt fission in the ICR tissue including a trifurcated crypt (far right). Note Msi-1 staining in cells at leading edge of the common wall (white arrows). Magnification is ×200. Scale bar equals 50 μm.

DISCUSSION

Adaptation is a compensatory response of the small intestine that occurs following massive loss of mucosal surface area to increase the absorptive and digestive capacity of the remaining bowel. The mechanisms that drive and maintain this response have not been clearly delineated. Our studies define a time line of distinct cellular mechanisms and components of intestinal adaptation in a new mouse model of ICR summarized in Fig. 10. We hypothesized that the adaptive process after ICR would include expansion of intestinal epithelial stem cells. Our data suggest that the effective increase in total crypt number occurs because of significant increases in the rate of crypt fission following ICR. Consistent with our hypothesis, we observed a significant increase in ISC numbers within the first week after ICR as assessed by increases in crypt fission, Msi-1 immunohistochemistry, and long-term label retention of BrdU administered in the first week after surgery. We also observed an immediate (as early as 2–3 days postresection) increase in CD45-negative SP cells, previously shown by our laboratory to be enriched for ISCs (10). To our knowledge these findings are the first direct evidence for early and rapid stem cell expansion after intestinal resection. The observations support that expansion of ISCs occurs in the immediate postsurgical period and has implications for optimal timing of interventions with trophic factors aimed at enhancing the adaptive response. Furthermore, the data suggest that the microenvironment may be most favorable to stem cell expansion in the immediate postresection period and this is relevant to the increasing interest in transplantation of ISCs as a potential mechanism to improve functional mass of intestine after major surgery.

Fig. 10.

Composite diagram showing the sources of intestinal adaptation following ICR. Increases were seen in mucosal circumference (circ.), number of crypts per circumference, and mucosal surface area per unit of circumference. NS, not significant.

Our data indicate that sustained adaptive increases in mucosal surface area in the jejunum are characterized by an increase in the caliber of the remnant small bowel following resection. This is associated with immediate but transient increases in villus height and crypt depth and sustained increases in the number of crypts per circumference in the remaining dilated bowel. In rodent models, intestinal adaptation following massive small bowel resection is characterized by increases in several parameters including villus height and crypt depth (19, 23). Several investigators have demonstrated significant regional differences between the jejunum and ileum occur following resection. The adapting ileum demonstrates greater morphometric changes in crypt depth and villus height (2, 16, 44), allocation of secretory lineages (22), and increases in apoptosis (18) than the jejunum, whereas the jejunum appears to augment the absorptive capacity to a greater extent than the ileum (42). Given these facts, animal models that utilize changes in morphometry as a measure of adaptation typically resect the jejunum and evaluate the adaptive response in the ileum. Furthermore, most studies have reported the adaptive morphometric changes within 7 days following resection, as the few studies that evaluated adaptation at time points out to 28 days have not seen significant differences between these two time points (22). In our model of ICR we evaluated the jejunum and found that both crypt depth and villus height were increased out to 6 wk following resection. By 16 wk postresection, both of these parameters had returned to sham levels. These data suggest that in this model of resection sustained adaptation is not mediated through changes in crypt depth and villus height as is seen at earlier time points.

Interestingly, we found that intestinal caliber, which was significantly greater in ICR animals compared with sham-operated animals as early as 4–5 days postresection, continued to increase in ICR animals and was over two times larger than sham animals by 16 wk postresection. These data coupled with the fact that we observed minimal changes in intestinal length following resection suggest that long-term recovery of lost mucosal surface area, in this model, is mediated by increasing intestinal caliber. Of note, clinically many patients with intestinal failure experience a decrease in diarrhea and an apparent increase in nutrient absorption over the first 1–2 years following intestinal loss. These changes are often accompanied by a marked increase in the caliber of the intestine, as determined by contrast imaging (24). Prior to the present work, the mechanisms responsible for increasing intestinal caliber have not been explored.

One potential mechanism of increasing intestinal caliber is by increasing the number of intestinal crypts. Our data show that significant increases in crypt fission (i.e., synthesis of new crypts) occur at 1 and 2 wk postresection, several weeks before the peak in mucosal circumference. We also observed that as intestinal caliber increased in ICR animals over time there was no change in crypt density (data not shown). On the basis of these data, we hypothesize that the increase in crypt fission leads to the increase in mucosal circumference. Prior reports of crypt fission indicate that this phenomenon is directly associated with neonatal intestinal development and is increased in preneoplastic states such as familial adenomatous polyposis. In the rat neonate, rapid increases in the length and caliber of the intestine occur as crypts expand through fission (8). In mice, low levels (∼5% of crypts) of fission occur at 1 wk, rapidly increasing between 2–3 wk (∼35%) until weaning occurs, reaching adult levels around 6 wk (∼2–5%) that persist throughout the life of the animal (6). Interestingly, regional differences in crypt fission have been described, with much higher levels observed in the jejunum than ileum following irradiation (34). Crypt fission typically begins as an indentation at the base of the crypt, and advances via a vertical split until two new crypts are produced as was observed in our study (8). In some instances, asymmetrical “budding” occurs more commonly in the colon and in regenerative situations, such as following irradiation, where multiple buds can be seen coming off the same crypt. Budding crypts were rarely observed following ICR.

It is generally accepted that crypt fission occurs due to duplication of ISC within crypts (34). Prior to the increase in crypt fission in ICR animals, we observed a rapid, but transient, increase in the percentage of CD45-negative SP cells, a population of cells previously shown in our laboratory to be enriched for ISCs (10). Furthermore, when expressed on a per-crypt basis, we observe that the number of CD45-negative SP cells per crypt is increased immediately following ICR. In fact, we observed a similar ∼20-fold increase in label-retaining cells and the number of CD45-negative SP cells per crypt following ICR. These data suggest that ISC numbers are rapidly increasing in response to ICR. Other studies utilizing label retention following radiation have confirmed increased ISC incorporation during periods of crypt fission (5). We surmise that the rapid expansion of ISCs drives the increase in crypt fission; however, the precise mechanisms responsible for this process are unclear.

Two signaling pathways have been shown to play key roles in regulating normal renewal of the intestinal epithelium. Perhaps most notably, the Wnt/β-catenin pathway has received attention for its role in driving proliferation in intestinal crypts, including the ISCs (29). Numerous studies have demonstrated the importance of the Wnt pathway in maintaining the ISC compartment (29, 30, 37). In contrast, the bone morphogenic protein (BMP) pathway acts as a brake on proliferation within the intestinal crypt. BMP receptor signaling prevents the inactivation of PTEN, a factor that, when active, inhibits PI3K/Akt-dependent activation and stabilization of β-catenin (21). Targeted disruption of BMP signaling by inactivating its receptor, Bmpr1a, within the intestinal epithelium results in aberrant crypt formation which is indicative of increased ISC numbers and crypt fission (21). Similarly, overexpression of Noggin, a BMP inhibitor, also induces aberrant crypt formation (17). Taken together, these data demonstrate that Wnts act as the accelerator and Bmp works as the brake on proliferation of ISCs. Clearly, further studies are necessary to delineate the roles of these pathways in the adaptive response seen in the mouse jejunum following ICR. This model of ICR will be valuable to study the role of different signaling pathways in the multiple cellular mechanisms of adaptation defined in this new model. Furthermore, by delineating the sequential processes of stem cell expansion and increased crypt fission leading to increases in intestinal caliber, we define a new model system to test the effects of different interventions or possible therapies for intestinal failure on different components of the adaptive response.

In summary, increases in the caliber of the remaining small bowel occur quickly following ICR concomitant with significant but transient increases in villus height and crypt depth. The increase in small intestinal caliber is due in part to an increase in the total number of crypts as a result of an increased rate of crypt fission, which is preceded by an increase in the number of ISCs. Only the increase in intestinal caliber is maintained long term, suggesting that in this model of intestinal resection augmenting caliber compensates for lost mucosal mass. These findings have implications for optimal methods to evaluate intestinal function in humans, where analyses of adaptive growth necessarily occur at relatively extended times after resection. Furthermore, understanding the mechanisms that drive ISC expansion will provide a better understanding of ISC biology, and our mouse ICR model provides a particularly useful new system to define these mechanisms in the future.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants P30 DK56338, K08 DK067395, and R01 DK069585.

Acknowledgments

The authors wish to thank Dr. Milton Finegold and the Texas Medical Center Digestive Disease Center Cellular and Molecular Morphology core for histological assistance; Dr. Kristin Kaiser for administrative assistance; Bobbie Antalffy for immunohistochemistry; Jennifer Norris for technical assistance; and Mike Cubbage, Chris Threeton, and the Texas Children's Cancer Center Flow Cytometry Laboratory for assistance with flow cytometry.

Footnotes

  • 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.

REFERENCES

View Abstract