Murine Tissue-Engineered Stomach Demonstrates Epithelial Differentiation

Journal of Surgical Research 171, 6–14 (2011) doi:10.1016/j.jss.2011.03.062

ASSOCIATION FOR ACADEMIC SURGERY Murine Tissue-Engineered Stomach Demonstrates Epithelial Differentiation Allison L. Speer, M.D., Frederic G. Sala, Ph.D., Jamil A. Matthews, M.D., and Tracy C. Grikscheit, M.D.1 Developmental Biology and Regenerative Medicine Program, The Saban Research Institute, Children’s Hospital Los Angeles, Los Angeles, California Originally submitted January 8, 2011; accepted for publication March 22, 2011

and cellular mechanisms of this regenerative process. Delineating the mechanism of how tissue-engineered stomach develops in vivo is an important precursor to its use as a human stomach replacement therapy. Ó 2011

Background. Gastric cancer remains the second largest cause of cancer-related mortality worldwide. Postgastrectomy morbidity is considerable and quality of life is poor. Tissue-engineered stomach is a potential replacement solution to restore adequate food reservoir and gastric physiology. In this study, we performed a detailed investigation of the development of tissue-engineered stomach in a mouse model, specifically evaluating epithelial differentiation, proliferation, and the presence of putative stem cell markers. Materials and Methods. Organoid units were isolated from <3 wk-old mouse glandular stomach and seeded onto biodegradable scaffolds. The constructs were implanted into the omentum of adult mice. Implants were harvested at designated time points and analyzed with histology and immunohistochemistry. Results. Tissue-engineered stomach grows as an expanding sphere with a simple columnar epithelium organized into gastric glands and an adjacent muscularis. The regenerated gastric epithelium demonstrates differentiation of all four cell types: mucous, enteroendocrine, chief, and parietal cells. Tissueengineered stomach epithelium proliferates at a rate comparable to native glandular stomach and expresses two putative stem cell markers: DCAMKL-1 and Lgr5. Conclusions. This study demonstrates the successful generation of tissue-engineered stomach in a mouse model for the first time. Regenerated gastric epithelium is able to appropriately proliferate and differentiate. The generation of murine tissue-engineered stomach is a necessary advance as it provides the transgenic tools required to investigate the molecular

Elsevier Inc. All rights reserved.

Key Words: tissue-engineering; regeneration; stomach; gastric; development; differentiation; proliferation; DCAMKL-1; Lgr5; postgastrectomy.

INTRODUCTION

Gastric cancer remains the second largest cause of cancer-related mortality worldwide [1]. The standard operation for early or advanced gastric cancer is a radical gastrectomy including a D2 lymph node dissection. Although this has contributed to an improved survival rate (>90%) for early gastric cancer patients, postgastrectomy morbidity is considerable and quality of life is poor [2]. More than 70 reconstructive procedures have been proposed since Schlatter’s initial gastrectomy in 1897 in an effort to address an inadequate food reservoir and lack of gastric physiology [3–8]. Studies reviewing these surgical strategies, including pouch formation and the jejunal interposition (Longmire’s procedure), have found conflicting evidence regarding quality of life improvement for these patients [7]. Jones and Cohen suggest that a Hunt-Lawrence (HL) jejunal pouch achieves the best long-term results in the rare congenital condition of microgastria; however, only 12 of the 59 reported cases have undergone a HL jejunal pouch [8]. Laparoscopic gastrectomy for gastric cancer has also gained popularity with the goals of minimizing surgical insult and maximizing quality of life, without affecting survival [2]. Tissue-engineered stomach is an attractive solution postgastrectomy because it offers a more exact replacement of the organ with restoration of an

1 To whom correspondence and reprint requests should be addressed at Developmental Biology and Regenerative Medicine Program, The Saban Research Institute, Children’s Hospital Los Angeles, 4650 W. Sunset Blvd. Mailstop #100, Los Angeles, CA 90027. E-mail: [email protected].

0022-4804/$36.00 Ó 2011 Elsevier Inc. All rights reserved.

6

SPEER ET AL.: EPITHELIAL DIFFERENTIATION OF ENGINEERED STOMACH

adequate food reservoir and appropriate gastric physiology. Additionally, the use of autologous cells would avoid the problems associated with transplanted organs: donor supply, risk of other epithelial cancer development, rejection, and the potential complications of lifelong immunosuppressive therapy. Syngeneic and autologous tissue-engineered stomach have been previously generated in Lewis rats [9–12] and Yorkshire swine [13], respectively. Although these studies demonstrated a regenerated epithelium organized into gastric glands, epithelial differentiation and proliferation were not comprehensively analyzed. Furthermore, investigation into the mechanism of formation of tissue-engineered stomach has been limited as transgenic animals are rare or unavailable in those species. In this study, we performed a more detailed investigation of the development of tissue-engineered stomach in a mouse model, specifically evaluating epithelial differentiation, proliferation, and the presence of putative stem cell markers. MATERIALS AND METHODS Animals All animals were housed in the Saban Animal Care Facility of the Saban Research Institute at Children’s Hospital Los Angeles, Los Angeles, California, and all experimental protocols were in accordance with and approved by the Institutional Animal Care and Use Committee. Animals were maintained in a temperature-regulated environment on a 12-h light-dark cycle and given access to chow and water ad libitum. All recipient mice were inspected daily for the duration of the study. Wild-type mice included CD1 mice from Charles Rivers Laboratory and C57Bl/6 mice from The Jackson Laboratory. Lgr5EGFP mice were maintained on a C57Bl/6 background (The Jackson Laboratory). Nonobese diabetic, severe combined immunodeficiency, g chain deficient (NOD/SCID g) mice were obtained from The Jackson Laboratory.

7

isolate the glandular stomach, consisting of simple columnar epithelium. The glandular stomachs were opened and each washed several times with 4
C HBSS, sedimenting between washes, to remove debris and mucous. Each lavaged stomach was minced into <1 mm3 pieces with sterile scissors. Tissue fragments were digested enzymatically with 0.12 mg/mL dispase (Invitrogen) and 800 U/mL collagenase type 1 (Sigma-Aldrich) at 37
C on an orbital shaker for 20 min. Tissue fragments were further digested mechanically using a 10 mL pipette for trituration. Digestion was stopped with 4
C 4% sorbitol (Sigma-Aldrich), 10% Fetal Bovine Serum (FBS) (Invitrogen) in high glucose Dulbecco’s modified Eagle medium (DMEM) (Invitrogen). After centrifugation at 500 rpm for 10 min, the supernatant was poured off, the pellet resuspended in 4
C 10% FBS in DMEM, and then centrifuged at 800 rpm for 5 min. The resulting pellet contained the isolated organoid units, multicellular clusters of epithelium and mesenchyme (Fig. 1A). These organoid units were then seeded onto the scaffolds at a density of 2.34 3 106 6 0.80 3 106 cells per scaffold (Fig. 1B).

Implantation Adult syngeneic (n ¼ 5) or NOD/SCID g (n ¼ 10) mice (both females and males) served as recipients (Table 1). NOD/SCID g mice were exposed to a single dose (350 cGy) of total body radiation immediately prior to implantation. Inhaled 1%–5% isoflurane was used to anesthetize the mice. The recipients underwent an upper midline vertical incision and the omentum was exposed. The seeded scaffold was wrapped completely in the omentum, secured with a 5-0 monocryl suture (Ethicon, New Brunswick, NJ, USA), and placed back into the abdominal cavity (Fig. 1C). The abdominal muscles and skin were closed with 4-0 vicryl suture (Ethicon) in two layers. Postoperative pain was controlled with 2 mg/kg ketofen (Fort Dodge Animal Health, Fort Dodge, IA, USA) administered subcutaneously immediately after implantation. The mice were also given 3 mL of sulfamethoxazole and trimethoprim oral suspension (200 mg/40 mg/5 mL; Qualitest Pharmaceuticals, Inc., Huntsville, AL, USA) per 300 mL water bottle from the day of implantation until postoperative d 7. Animals were euthanized by exposure to CO2 in an inhalation chamber

Scaffold Construction Cylindrical microporous biodegradable polymer scaffolds (5 mm length, 4 mm width, 3 mm height) were made from a 2 mm thick, nonwoven, polyglycolic acid (PGA) felt (bulk density 50 mg/cm3, porosity > 95%, Biofelt; Concordia Medical, Warwick, RI, USA) formed on a glass mandril and coated with 5% poly-L-lactic acid (PLLA) (Durect Corporation, Cupertino, CA, USA) in chloroform (Sigma-Aldrich, St. Louis, MO, USA). We have found that PLLA infiltrates the PGA mesh and provides structural rigidity. After sterilization with 100% ethanol, the polymer tubes were coated with collagen type I solution (0.4 mg/mL, Sigma-Aldrich) for 20 min at 4
C, rinsed with phosphate buffered saline (PBS), and stored in a dessiccator in order to avoid premature hydrolysis and degradation of the polymer.

Organoid Unit Isolation Nonfasted, <3 wk-old, neonatal wild-type or transgenic mice (both females and males) were euthanized by exposure to carbon dioxide (CO2) in an inhalation chamber. The stomachs were harvested, the omentum removed, and placed in Hank’s balanced salt solution (HBSS) (Invitrogen, Carlsbad, CA, USA) on ice. The forestomach, consisting of keratinized stratified squamous epithelium, was removed to

FIG. 1. Tissue-engineered stomach technique. (A) Organoid units, multicellular clusters of epithelium and mesenchyme, are isolated from <3-wk-old mouse glandular stomach per protocol. (B) Organoid units are loaded onto a biodegradable scaffold. (C) The seeded scaffold is implanted into the omentum of an adult mouse. (D) Tissue-engineered stomach grows in vivo and is harvested at designated time points. (Color version of figure is available online.)

8

JOURNAL OF SURGICAL RESEARCH: VOL. 171, NO. 1, NOVEMBER 2011

TABLE 1 Tissue-Engineered Stomach Characteristics Syngeneic

Recipient

Wk

Cells/HPF

M

E

C

P

1 2 3 4 5

C57Bl/6 CD1 NOD/SCID CD1 C57Bl/6

1 4 4 8 8

567 203 55 186 444

þ þ þ þ þ

þ þ – – þ

– – – – –

– – – – –

Allogeneic

Recipient

Wk

Cells/HPF

M

E

C

P

1 2 3 4 5 6 7 8 9 10

NOD/SCID NOD/SCID NOD/SCID NOD/SCID NOD/SCID NOD/SCID NOD/SCID NOD/SCID NOD/SCID NOD/SCID

1 4 4 4 4 4 4 8 12 12

684 271 774 1020 280 523 639 614 293 600

þ þ þ þ þ þ þ þ þ þ

þ þ þ þ – þ þ þ þ þ

þ – þ þ – – þ þ – –

– – – þ – – þ – – –

Recipient ¼ type of mouse into which the seeded scaffold was implanted; wk ¼ number of weeks post-implantation (harvest time); cells/HPF ¼ average number of epithelial cells per high power field; M ¼ mucous cells; E ¼ enteroendocrine cells; C ¼ chief cells; P ¼ parietal cells; þ ¼ cells present; – ¼ cells absent. at designated time points: 1 (n ¼ 4), 4 (n ¼ 20), 6, (n ¼ 1), 8 (n ¼ 3), and 12 (n ¼ 2) wk for tissue harvest (Fig. 1D).

Histology and Immunofluorescence The tissue-engineered stomachs were fixed in 10% formalin and embedded in paraffin. Serial sections were cut at 5 mm of thickness. Histology slides were stained with hematoxylin-eosin per standard protocol. Additional samples were stained with Alcian blue solution per protocol [14]. The remaining slides were prepared for immunofluorescence. These slides were dehydrated to water and then an antigen retrieval step was performed by boiling the slides in a microwave for 12 min in 10 mM sodium-citrate buffer pH ¼ 6.0. The slides were incubated with the primary antibody diluted in universal blocking solution with 2% goat serum overnight at 4
C. The primary antibodies used included: rabbit anti-chromogranin AþB (1:100; ABcam, Cambridge, MA, USA), sheep anti-pepsinogen II (1:100; ABcam), mouse anti-H/K ATPase (1:100; ABcam), mouse Cy3-coupled anti-a-smooth muscle actin (SMA) (1:300; Sigma-Aldrich), mouse anti-desmin (1:50; DAKO, Carpinteria, CA, USA), mouse anti-proliferating cell nuclear antigen (PCNA) (1:100; Vector Laboratories, Burlingame, CA, USA), rabbit doublecortin and calcium/calmodulin-dependent protein kinase-like-1 (DCAMKL-1) (1:50; Abgent, San Diego, CA, USA), and mouse anti-green fluorescent protein (GFP) (1:200; ABcam). Cy3 or FITC conjugated goat anti-rabbit or -mouse, or Cy3 conjugated donkey anti-sheep secondary antibodies were used. Slides were mounted using Vectashield with DAPI (Vector Laboratories) as mounting medium. Adult native glandular stomach served as positive and negative controls.

engineered stomach specimen. If a large amount of epithelium was present, two (n ¼ 4) or three (n ¼ 4) HPFs were evaluated. Two HPFs were counted for each adult native glandular stomach. The average and standard deviation were calculated using Microsoft Excel software. P value was determined by using Student’s t-Test (Microsoft Excel software). P value <0.05 was considered statistically significant.

RESULTS Tissue-Engineered Stomach Grows as an Expanding Sphere

After implantation into the omentum of syngeneic or irradiated NOD/SCID g mice for 1, 4, 8, or 12 wk, the constructs grew to form a sphere of tissue-engineered stomach (Fig. 2) with a lumen filled with thick, white-yellow mucous. On histologic examination, half of all implants harvested (15 out of 30 implants) regenerated an epithelium at 1 (n ¼ 2), 4 (n ¼ 8), 8 (n ¼ 3), or 12 (n ¼ 2) wk. The remaining 15 implants harvested did not demonstrate an epithelium, but rather consisted of fibrous tissue, inflammatory infiltrate, and degrading polymer, and were filled with thick yellow fluid and debris. Tissue-Engineered Stomach Mucosa Consists of a Simple Columnar Epithelium Organized into Gastric Glands

Hematoxylin and eosin staining of paraffin embedded sections of each tissue-engineered stomach demonstrated a simple columnar epithelium similar to that of native stomach (Fig. 3). Portions of this tissueengineered stomach epithelium were simply a flat epithelium that appeared to be migrating to completely cover the inner lumen of the tissue-engineered stomach. Other areas of the tissue-engineered stomach epithelium consisted of rudimentary invaginations resembling early gastric gland formation. Finally, additional regions of the tissue-engineered stomach epithelium demonstrated more complete gastric gland formation with a base, neck, isthmus, and pit resembling the normal architecture of the native glandular stomach. Histologic variability of the tissue-engineered stomach

Cell Counting Immunohistochemistry was performed on paraffin-embedded sections of tissue-engineered stomach (n ¼ 15) and adult native glandular stomach (n ¼ 3) with PCNA primary antibody and Cy3 goat antimouse secondary antibody. The total number of epithelial cells and PCNA positive cells were counted per high power field (HPF ¼ 203 magnification). At least one HPF was counted for each tissue-

FIG. 2. Tissue-engineered stomach grows as an expanding sphere. Gross photograph of a biodegradable scaffold measuring 5 3 4 3 3 mm and 2 mm thick, and a tissue-engineered stomach harvested at 4 wk measuring 9 3 7.5 3 5 mm. (Color version of figure is available online.)

SPEER ET AL.: EPITHELIAL DIFFERENTIATION OF ENGINEERED STOMACH

9

FIG. 3. Tissue-engineered stomach mucosa is a simple columnar epithelium organized into gastric glands. Hematoxylin and eosin staining demonstrates that (A) tissue-engineered stomach histology at 4 wk is similar to that of (B) native glandular stomach (scale bars ¼ 50 mm). (Color version of figure is available online.)

epithelium is further demonstrated by the difference in the total number of epithelial cells present per high power field (HPF) with a range of 55 to 1246 and an average of 540 6 313. Adult native glandular stomach epithelium had less variability with a range of 660 to 858 and an average of 779 6 84. Interestingly, the type of histology present and the total number of epithelial cells present per HPF in tissue-engineered stomach were independent of implantation time. All three types of histology described above were found in tissue-engineered stomach harvested at all four time points: 1, 4, 8, and 12 wk (data not shown). Tissue-Engineered Stomach Demonstrates Epithelial Differentiation

All 15 tissue-engineered stomach specimens were analyzed for the presence of the four terminally differentiated gastric epithelial cell types using immunohistology

techniques (Table 1). Tissue-engineered stomach was compared with native glandular stomach. Mucous cells, which secrete the protective mucous layer coating the gastric mucosa, were identified by Alcian blue staining throughout the epithelium of all 15 tissue-engineered stomach specimens (Fig. 4A). Enteroendocrine cells produce hormones that control gastric secretion and motility and were demonstrated using chromogranin A antibody (Fig. 4B) in the majority of tissue-engineered stomach specimens (12 out of 15). Notably, the three specimens lacking enteroendocrine cells had incompletely developed epithelium present on histology with a small number of epithelial cells ranging from 55 to 280 total epithelial cells per HPF and an average of 174 6 113, well below the amount seen in native glandular stomach. Chief cells are responsible for secreting pepsinogen, the precursor to the digestive enzyme pepsin, and stained positive for pepsinogen antibody in onethird (5 out of 15) of the tissue-engineered stomach

FIG. 4. Tissue-engineered stomach demonstrates epithelial differentiation. (A)–(D) Tissue-engineered stomach harvested at 4 wk. (E)–(H) Native glandular stomach. (A) and (E) Alcian blue staining reveals mucous cells. (B) and (F) Enteroendocrine cells are detected by chromogranin A. (C) and (G) Chief cells are identified with pepsinogen at the base of the gastric glands. (D) and (H) Parietal cells are marked by H/ K ATPase (panels ¼ 203, insets ¼ 403, scale bars ¼ 50 mm). (Color version of figure is available online.)

10

JOURNAL OF SURGICAL RESEARCH: VOL. 171, NO. 1, NOVEMBER 2011

specimens (Fig. 4C). When present, chief cells were located in their normal location at the base of the developing gastric glands. Interestingly, the five tissueengineered stomach specimens with chief cells had a more well-developed epithelium than the other 10 specimens. Their histology recapitulated that of the native glandular stomach. These five specimens with pepsinogen positive chief cells also had a total number of epithelial cells per HPF comparable to native glandular stomach with an average of 749 6 299. Finally, hydrogen potassium adenosine triphosphatase (H/K ATPase) antibody detected acid secreting parietal cells in two out of 15 tissue-engineered stomach specimens (Fig. 4D). Of note, these two specimens had a well-developed epithelium with gastric gland formation and all four differentiated cell types. Tissue-Engineered Stomach Epithelium Proliferates at a Rate Similar to Native Glandular Stomach and Demonstrates the Expression of Two Putative Gastrointestinal Stem Cell Markers: DCAMKL-1 and Lgr5

In order to investigate if the regenerating tissueengineered stomach epithelium had differences in proliferation from native glandular stomach, immunohistochemistry was performed to detect expression of

proliferating cell nuclear antigen (PCNA). PCNA positive cells were identified in their normal location in the proliferative zone of the isthmus [15, 16] in both tissueengineered stomach (Fig. 5A) and native glandular stomach (Fig. 5B). We determined the percentage of PCNA positive cells per total epithelial cells per HPF and found no statistically significant difference (P value ¼ 0.2) between the proliferation rate in tissue-engineered stomach (45.4% 6 18.3%) and native glandular stomach (38.1% 6 4.7%). Additionally, the expression of two putative gastrointestinal stem cell markers, doublecortin and calcium/calmodulin-dependent protein kinase-like1 (DCAMKL-1) and Leucine-rich repeat-containing G-protein coupled receptor 5 (Lgr5), were evaluated in the tissue-engineered stomach epithelium by immunohistochemistry. DCAMKL-1 was expressed in native glandular stomach controls (Fig. 5B) and at all four harvest time points (1, 4, 8, and 12 wk) in 10 of the 15 tissue-engineered stomach specimens (Fig. 5A). Of note, double immunofluorescence staining for both PCNA and DCAMKL-1 demonstrated the quiescence of the DCAMKL-1 positive (PCNA negative) cells. Lgr5 expression was identified using the transgenic reporter line Lgr5EGFP [17] as donor cells for organoid unit isolation and an irradiated NOD/SCID g mouse served as the host for implantation (n ¼ 1). Lgr5 expressing cells

FIG. 5. Tissue-engineered stomach epithelium demonstrates proliferation and the expression of two putative gastrointestinal stem cell markers: DCAMKL-1 and Lgr5. Proliferating cell nuclear antigen (PCNA) detected progenitor cells in their normal location in the isthmus of both (A) tissue-engineered stomach at 4 wk and (B) native glandular stomach. Tissue-engineered stomach and native glandular stomach demonstrated expression of two putative gastrointestinal stem cell markers. (A) and (B) DCAMKL-1 positive cells were identified scattered throughout the epithelium of the gastric glands, whereas (C) and (D) Lgr5 was expressed at the base of the gastric glands in both tissueengineered stomach and native glandular stomach, respectively (scale bars ¼ 50 mm). (Color version of figure is available online.)

SPEER ET AL.: EPITHELIAL DIFFERENTIATION OF ENGINEERED STOMACH

were positive for green fluorescent protein (GFP) on immunofluorescence staining and found at the base of the gastric glands in the epithelium of tissue-engineered stomach (Fig. 5C) at 4 wk post-implantation and in native glandular stomach (Fig. 5D). Tissue-Engineered Stomach Develops a Muscularis

The muscularis was identified on H&E staining in all 15 tissue-engineered stomach specimens and confirmed with double immunofluorescence staining for a smooth muscle actin (SMA) and desmin in two specimens (Fig. 6). The muscularis, although present beneath the submucosa, did not recapitulate the inner circular and outer longitudinal layers normally observed in the native glandular stomach. The muscularis mucosae, a thin layer of smooth muscle just beneath the lamina propria, stained positive for SMA in both tissue-engineered stomach harvested at 4 wk and native glandular stomach. DISCUSSION

The earliest tissue-engineering of the gastrointestinal tract has been accomplished in Lewis rat models [9–12, 18–21] and Yorkshire swine [13] because of their larger size, resilience for successful survival surgeries, and historic precedent. However, investigation into the mechanism of formation of tissue-engineered organs has been limited as transgenic animals are rare or unavailable in those species. This study demonstrates the successful generation of tissue-engineered stomach in a mouse model for the first time. Generation of murine tissue-engineered stomach is a necessary advance for our tissue-engineering technique as it provides an armamentarium of transgenic tools to investigate the molecular and cellular mechanisms of formation of tissue-engineered stomach. Understanding the mecha-

11

nism of how tissue-engineered stomach develops in vivo is an important precursor to its use as a human stomach replacement therapy. Previous studies of syngeneic and autologous tissueengineered stomach in Lewis rats [9–12] and Yorkshire swine [13], respectively, have demonstrated a regenerated gastric epithelium from organoid units using a similar technique. However, epithelial differentiation was not investigated in detail and proliferation was not evaluated. In this study, we conducted a more comprehensive analysis of the development of tissue-engineered stomach in a mouse model, specifically examining epithelial differentiation, proliferation, and the presence of putative stem cell markers during regeneration. In native glandular stomach epithelium, progenitor cells in the upper gland neck and isthmus give rise to three second-order progenitor cell lineages: prepit, preparietal, and preneck cells [22–24]. These then differentiate into the four types of gastric epithelial cells, including two types of mucous cells, surface mucous cells and mucous neck cells, as well as enteroendocrine, chief cells, and parietal cells [25]. Surface mucous cells and mucous neck cells secrete a protective viscous mucous layer that stains positive with Alcian blue [26]. As previously demonstrated in tissue-engineered stomach in Lewis rats [10–12] and Yorkshire swine [13], we also identified mucous cells in all 15 of our tissue-engineered stomach specimens regardless of the type of histology present and independent of implantation time. Surface mucous cells were in their proper location in the superficial epithelium of the gastric pit, and mucous neck cells were appropriately located in the neck and base of the gastric glands. Enteroendocrine cells produce a variety of hormones that control gastric secretion and motility. They are usually found scattered throughout the glandular stomach epithelium in the base, neck, and isthmus of

FIG. 6. Tissue-engineered stomach demonstrates a muscularis. The muscularis mucosae, a thin layer of smooth muscle just beneath the lamina propria, stained positive with a smooth muscle actin (SMA) in (A) tissue-engineered stomach harvested at 4 wk and (B) native glandular stomach. The muscularis, seen below the submucosa, stained positive for SMA and desmin in both as well (scale bars ¼ 50 mm). (Color version of figure is available online.)

12

JOURNAL OF SURGICAL RESEARCH: VOL. 171, NO. 1, NOVEMBER 2011

the gastric gland [16, 27]. Grikscheit et al. found appropriate gastrin staining in their tissue-engineered stomach epithelium generated in Lewis rats [9]. A specific type of enteroendocrine cell, the G cell, normally produces gastrin. We elected to identify multiple types of enteroendocrine cells using chromogranin A, an acidic glycoprotein that is widely expressed in endocrine and neuroendocrine cells [28]. Enteroendocrine cells were present in the majority (12 out of 15) of our tissueengineered stomach specimens in the proper location. The limited amount of underdeveloped epithelium in the three specimens lacking enteroendocrine cells likely explains their absence. Pepsinogen is the zymogen to the enzyme, pepsin, which digests food proteins into peptides. This proenzyme is produced by chief cells that are normally located at the base of the gastric glands in the antrum [27]. We found chief cells in their appropriate location in one-third (5 out of 15) of our tissue-engineered stomach specimens. These five specimens appeared to have well-developed gastric glands resembling those of native glandular stomach. We attribute the dearth of chief cells in the remaining 10 specimens to their underdeveloped histology. Parietal cells are located in the corpus and produce the enzyme H/K ATPase, which is responsible for acid secretion in the stomach [27]. One prior study in Lewis rats demonstrated the presence of the proton pump antibody (H/K ATPase a subunit) in tissue-engineered stomach at 4 wk after implantation, but they noted it was at a lower level than the native stomach [11]. We detected parietal cells in two of our 15 tissue-engineered stomach specimens at 4 wk after implantation using antibody to the b subunit of H/K ATPase. Notably, these two specimens had well-developed epithelium with gastric glands and all four differentiated cell types. Our regenerated gastric epithelium proliferated at a similar rate to native glandular stomach. Notably, this progenitor cell population was located in the proper location, the isthmus of the gastric glands, when the epithelium was well-developed [15, 16, 25]. We had expected the proliferation rate of our tissue-engineered stomach to be higher than that of native glandular stomach since it was undergoing regeneration. While there was a trend towards a higher proliferation rate in the tissue-engineered stomach, it was not statistically significant. Of note, some individual specimens had proliferation rates as high as 83.6%, but these did not correlate with implantation time or type of histology. Perhaps this can be explained by our sectioning technique. Native glandular stomach was sectioned in a controlled manner allowing for reproducible histology with minimal variation. However, tissue-engineered stomach histology is quite variable because sectioning is at random since these specimens grow as expanding spheres.

Two distinct cell populations have recently been proposed as adult intestinal stem cells [29, 30]. DCAMKL-1 has been evaluated in the stomach in two studies and was expressed in immature quiescent cells in the isthmus of normal gastric glands where proliferative cells and putative progenitor cells are thought to reside [31, 32]. Lgr5 expression is identified in two to four cells at the base of the pyloric glands [33]. We found DCAMKL-1 positive, PCNA negative cells expressed in our tissue-engineered stomach, although these were located throughout the gastric gland and not just in the isthmus. Furthermore, we identified Lgr5 positive cells at the base of the gastric glands in a tissue-engineered stomach harvested at 4 wk. Both of these putative gastric stem cells and the progenitor cells of the isthmus may be imperative for our tissueengineered stomach, specifically during processes such as regeneration, homeostasis, and tissue repair after injury. Further investigation to define their specific roles during these processes in tissue-engineered stomach is required before its future use as a replacement graft. There are several explanations for the variability in histology and incomplete differentiation of our tissueengineered stomach. First, the presence of progressive differentiation in our specimens that is associated with the development of the gastric glands suggests that perhaps tissue-engineered stomach requires more time in vivo to fully mature. Gut development in the mouse is not complete at birth and usually continues until the end of the third postnatal week [15]. Moreover, differentiation of our tissue-engineered stomach epithelium is reminiscent of developmental stages in embryonic mouse stomach organogenesis. Nyeng et al. determined the presence of various markers of differentiated stomach cell lineages by RTPCR from E14.5 to E18.5 [27]. This included chromogranin A for maturing endocrine cells, pepsinogen C for zymogenic chief cells, and H/K ATPase for parietal cells. They found that chromogranin A was present at low levels at E15.5, doubled in expression on E16.5, and remained constant for the remainder of embryonic development. Pepsinogen was weakly detectable at E15.5 but increased dramatically on E16.5. H/K ATPase was undetectable until E16.5 and peaked at E17.5. We observed a similar developmental pattern in our tissue-engineered stomach during regeneration. Specimens with underdeveloped histology had only mucous cells and occasionally enteroendocrine cells. Tissue-engineered stomach with well-developed gastric epithelium and more complete gland formation demonstrated the presence of mucous cells, enteroendocrine cells, chief cells, and, occasionally, parietal cells. Furthermore, although we isolated organoid units from the entire glandular stomach including the corpus,

SPEER ET AL.: EPITHELIAL DIFFERENTIATION OF ENGINEERED STOMACH

antrum, and pylorus, perhaps our tissue-engineered stomach is regenerated from only one region, the antrum. This would explain the paucity of parietal cells in some of our well-developed tissue-engineered stomach samples since chief cells are usually found in the antrum and parietal cells in the corpus [27]. Finally, another plausible explanation for the paucity of chief cells in our tissue-engineered stomach specimens is the concomitant lack of parietal cells. Parietal cells secrete numerous growth factors including sonic hedgehog [34], amphiregulin [35], and epidermal growth factor receptor ligands [25], which promote epithelial proliferation and differentiation. The loss of parietal cells leads to inhibition of mucous neck cell to chief cell differentiation. It is a reasonable theory that the absence of parietal cells in the majority of our tissue-engineered specimens has reduced the differentiation of chief cells. Our tissue-engineered stomach mouse model affords several advantages. First, it can serve as a versatile in vivo tool for the study of gastric epithelial regeneration, stem cell behavior, and molecular causes of gastric cancer that may complement existing cell lines and other in vitro long-term gastric epithelium culture systems. Our technique is efficient since epithelium forms in 50% of harvested implants. Furthermore, our model offers the advantage of investigating epithelialmesenchymal interactions that are paramount during gut development [36–39]. Finally, as mentioned previously, our mouse model provides the opportunity to use transgenic mouse lines. For example, fibroblast growth factor 10 (FGF10) produced by gastric mesenchyme and acting on fibroblast growth factor receptor 2b (FGFR2B) in gastric epithelium has been found to regulate appropriate proliferation and differentiation during stomach development, and is essential for stomach morphogenesis [36]. In fact, Barker et al. demonstrated that the addition of FGF10 was necessary for budding and expansion of cultured gastric gland epithelial units [33]. Interestingly, while Spencer-Dene et al. successfully identified mucous producing cells and endocrine cells in the stomachs of Fgfr2b/ and Fgf10/ mice, there was a reduced abundance of chief cells and an absence of parietal cells, suggesting that differentiation of the gastric mucosa in these mutants was severely compromised [36]. We observe a similar phenomenon in our tissue-engineered stomach. We speculate that regenerating tissue-engineered stomach from transgenic mice overexpressing Fgf10 may improve the development and differentiation of our gastric epithelium. The ability to manipulate specific genes to understand their influence on tissue-engineered stomach regeneration and development will be crucial to transitioning this technique to humans and optimizing it for potential replacement therapy.

13

ACKNOWLEDGMENTS This study was supported by grants from the California Institute for Regenerative Medicine (RN2-00946-1, TG2-01168).

REFERENCES 1. Parkin DM, Bray F, Ferlay J, et al. Global Cancer Statistics. CA Cancer J Clin 2005;55:74. 2. Kim HH, Hyung WJ, Cho GS, et al. Morbidity and mortality of laparoscopic gastrectomy versus open gastrectomy for gastric cancer: An interim report–a phase III multicenter, prospective, randomized Trial (KLASS Trial). Ann Surg 2010;251:417. 3. Herfarth C, Schlag P, Buhl K. Surgical procedures for gastric substitution. World J Surg 1987;11:689. 4. Beal M, Briggs JD, Longmire WP Jr. Use of a jejunal segment to replace the stomach following total gastrectomy. Am J Surg 1954;88:194. 5. Scott HW, Law DH, Gobbel WG, et al. Clinical and metabolic studies after total gastrectomy with a Hunt-Lawrence jejunal food pouch. Am J Surg 1968;115:148. 6. Nakane Y, Okumura S, Akehira K, et al. Jejunal pouch reconstruction after total gastrectomy for cancer. Ann Surg 1995; 222:27. 7. Hoksch B, Ablassmaier B, Zieren J, et al. Quality of live after gastrectomy: Longmire’s reconstruction alone compared with additional pouch reconstruction. World J Surg 2002;26:335. 8. Jones VS, Cohen RC. An 18-year follow-up after surgery for congenital microgastria—case report and review of literature. J Pediatr Surg 2007;42:1957. 9. Grikscheit TC, Srinivasan A, Vacanti JP. Tissue-engineered stomach: A preliminary report of a versatile in vivo model with therapeutic potential. J Pediatr Surg 2003;38:1305. 10. Maemura T, Shin M, Sato M, et al. A tissue-engineered stomach as a replacement of the native stomach. Transplantation 2003; 76:61. 11. Maemura T, Shin M, Ishii O, et al. Initial assessment of a tissue engineered stomach derived from syngeneic donors in a rat model. ASAIO J 2004;50:468. 12. Maemura T, Ogawa K, Shin M, et al. Assessment of tissueengineered stomach derived from isolated epithelium organoid units. Transplant Proc 2004;36:1595. 13. Sala FG, Kunisaki SM, Ochoa ER, et al. Tissue-engineered small intestine and stomach form from autologous tissue in a preclinical large animal model. J Surg Res 2009;156:205. 14. Available at: http://www.ihcworld.com/_protocols/special_stains/ alcian_blue.htm. Accessed: January 7, 2011. 15. Karam SM, Li Q, Gordon JI. Gastric epithelial morphogenesis in normal and transgenic mice. Am J Physiol 1997;272:G1209. 16. Barker N, Clevers H. Leucine-rich repeat-containing G-proteincoupled receptors as markers of adult stem cells. Gastroeneterology 2010;139:1691. 17. Barker N, van Es JH, Kuipers J, et al. Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature 2007; 449:1003. 18. Grikscheit TC, Ogilvie JB, Ochoa ER, et al. Tissue-engineered colon exhibits function in vivo. Surgery 2002;132:200. 19. Grikscheit TC, Ochoa ER, Srinivasan A, et al. Tissue-engineered esophagus: Experimental substitution by onlay patch or interposition. J Thorac Cardiovasc Surg 2003;126:537. 20. Grikscheit TC, Ochoa ER, Ramsanahie A, et al. Tissueengineered large intestine resembles native colon with appropriate in vitro physiology and architecture. Ann Surg 2003;238:35. 21. Grikscheit TC, Siddique A, Ochoa ER, et al. Tissue-engineered small intestine improves recovery after massive small bowel resection. Ann Surg 2004;240:748.

14

JOURNAL OF SURGICAL RESEARCH: VOL. 171, NO. 1, NOVEMBER 2011

22. Karam SM, Leblond CP. Dynamics of epithelial cells in the corpus of the mouse stomach. I. Identification of proliferative cell types and pinpointing of the stem cells. Anat Rec 1993;236:259. 23. Karam SM, Leblond CP. Dynamics of epithelial cells in the corpus of the mouse stomach. II. Outward migration of pit cells. Anat Rec 1993;236:280. 24. Karam SM, Leblond CP. Dynamics of epithelial cells in the corpus of the mouse stomach. III. Inward migration of neck cells followed by progressive transformation into zymogenic cells. Anat Rec 1993;236:297. 25. Nozaki K, Weis V, Wang TC, et al. Altered gastric chief cell lineage differentiation in histamine-deficient mice. Am J Physiol Gastrointest Liver Physiol 2009;296:G1211. 26. Lambert R, Martin F, Andre C, et al. Apical localization of glycoproteins in gastric mucous cells. Am J Dig Dis 1968;13:941. 27. Nyeng P, Norgaard GA, Kobberup S, et al. FGF10 signaling controls stomach morphogenesis. Dev Biol 2007;303:295. 28. Mouland AJ, Bevan S, White JH, et al. Human chromogranin A gene. Molecular cloning, structural analysis, and neuroendocrine cell-specific expression. J Biol Chem 1994;269:6918. 29. May R, Riehl TE, Hunt C, et al. Identification of a novel putative gastrointestinal stem cell and adenoma stem cell marker, doublecortin and CaM kinase-like-1, following radiation injury and in adenomatous polyposis coli/multiple intestinal neoplasia mice. Stem Cells 2008;26:630. 30. Sato T, Vries RG, Snippert HJ, et al. Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature 2009;459:262. 31. Kikuchi M, Nagata H, Watanabe N, et al. Altered expression of putative progenitor cell marker DCAMKL1 in rat gastric

32.

33.

34.

35.

36.

37.

38.

39.

mucosa in regeneration, metaplasia, and dysplasia. BMC Gastroenterol 2010;10:65. Zhang Y, Huang X. Investigation of doublecortin and calcium/ calmodulin-dependent protein kinase-like-1 expressing cells in the mouse stomach. J Gastroenterol Hepatol 2010;25:576. Barker N, Huch M, Kujala P, et al. Lgr5þve stem cells drive selfrenewal in the stomach and build long-lived gastric units in vitro. Cell Stem Cell 2010;6:25. Zavros Y, Waghray M, Tessier A, et al. Reduced pepsin A processing of sonic hedgehog in parietal cells precedes gastric atrophy and transformation. J Biol Chem 2007;282:33265. Abe S, Sasano H, Katoh K, et al. Immunohistochemical studies on EGF family growth factors in normal and ulcerated human gastric mucosa. Dig Dis Sci 1997;42:1199. Spencer-Dene B, Sala FG, Bellusci S, et al. Stomach development is dependent on fibroblast growth factor 10/fibroblast growth factor receptor 2b-mediated signaling. Gastroenterology 2006;130:1233. Koike T, Yasugi S. In vitro analysis of mesenchymal influences on the differentiation of stomach epithelial cells of the chicken embryo. Differentiation 1999;65:13. Fukada K, Sakamoto N, Narita T, et al. Application of efficient and specific gene transfer systems and organ culture technique for the elucidation of mechanisms of epithelial-mesenchymal interaction in the development gut. Dev Growth Differ 2000; 42:207. Sakamoto N, Fukuda K, Watanuki K, et al. Role for cGATA-5 in transcriptional regulation of embryo chicken pepsinogen gene by epithelial-mesenchymal interactions in the development chicken stomach. Dev Biol 2000;223:103.