Abstract
The cynomolgus macaque is the most commonly used nonhuman primate in nonclinical toxicity testing, but the macroscopic and microscopic anatomy of the stomach in the cynomolgus macaque is poorly described. To develop a reliable sampling method for histologic evaluation of the cynomolgus macaque stomach in regulatory toxicity studies, the stomachs of control animals were prospectively evaluated using an extensive sectioning pattern. The stomach of the cynomolgus macaque differs from that described for the human stomach and has a prominent fundus that lacks parietal cells. A description of the macroscopic and microscopic anatomy is presented along with a recommended sectioning pattern for nonclinical toxicity studies and discussion of species differences. A thorough understanding of normal anatomy and species comparisons are critical to interpretation of potential toxicity findings and assessment of risk in humans.
Introduction
Gastric toxicity is commonly observed in humans, as oral administration of drugs allows for both local and systemic effects within the stomach (Lewis, 1986). A thorough and consistent evaluation of the stomach in nonclinical species is necessary to demonstrate safety and understand the potential risk of orally administered compounds in humans. The nonhuman primate is often used in nonclinical toxicity testing for its phylogenetic similarity to humans; however, differences in the microscopic anatomy of the stomach of various Old World primates and humans have been reported despite a similar gross appearance (Berger, 1934; Tokii and Tsukamoto, 1953; Kent, 1966; Sheahan and Jervis, 1976; Owen, 1997; Siwek, 1979). In addition, significant variability in the distribution of parietal cells has been observed within the stomach of cynomolgus macaques (
Materials and Methods
Animals
All procedures and care of animals were in accordance with the principles for humane care outlined by the Institute of Laboratory Animal Resources
Cynomolgus monkeys (30 males and 30 females) aged 2.5 to 6 years, from studies with durations of 4 days to 1 month, were used for this study. These monkeys received vehicle only by either the oral, intravenous, or subcutaneous routes of administration. The monkeys were obtained from Covance Research Products, Inc. (Denver, Pennsylvania), Primate Products (Miami, Florida), and Charles River Laboratories (Houston, Texas). Monkeys originated from either Indonesia or Mauritius, and all were captive bred and raised in groups outdoors in their country of origin before shipment to the United States. Monkeys were singly housed in climate-controlled rooms at 64°F to 84° F, with a relative humidity of 30% to 70%. The rooms had approximately 10 to 15 air changes per hour and a photoperiod cycle of 12 hours light/12 hours dark. Monkeys were fed a commercially available complete diet and generally received 8 (male) or 6 (female) biscuits/day; a daily allotment of fresh fruit was also provided. Water was provided ad libitum.
Before necropsy, each monkey was pretreated with an intramuscular or subcutaneous injection of ketamine HCl (approximately 10 mg/kg), anesthetized with an intravenous injection of sodium pentobarbital (starting at approximately 30 mg/kg), and then exsanguinated. The esophagus and duodenum were transected approximately 1 to 2 cm from cardia and pylorus, respectively. The stomach was removed and opened along the greater curvature. The stomach was pinned flat to a corkboard, immersed in 10% neutral buffered formalin, and fixed for 24 to 48 hours at room temperature. The time from exsanguination to immersion fixation was generally less than 45 minutes. Following formalin fixation, photographs of selected stomachs were taken.
Seven longitudinal sections of each stomach representing the greater curvature, lesser curvature, body, fundus, and antrum were trimmed (Figure 1), and a tissue-marking dye (Cancer Diagnostics, Inc., Birmingham, Michigan) was used to identify the proximal end. Tissues were processed to paraffin and 5-micron sections were cut, placed on Superfrost Plus slides (VWR Scientific, West Chester, Pennsylvania), and stained with hematoxylin and eosin (H&E). All slides were evaluated by a single veterinary pathologist (JDV) for the presence and distribution of parietal, chief, and mucous cells.
Histochemistry
Tissues from a subset of the animals (6 males and 6 females) were stained using alcian blue-periodic acid Schiff (AB-PAS) to evaluate mucin expression. Serial sections were deparaffinized and hydrated to distilled water. Sections were stained with AB solution, pH 2.5, for 30 minutes. Sections were rinsed, and a standard PAS reaction was performed. Sections were counterstained with hematoxylin, dehydrated, cleared, and coverslipped.
Immunohistochemistry
Tissues from a subset of the animals (6 males and 6 females) were immunolabeled to accurately identify parietal and chief cells. Slides were placed on the Ventana Discovery System, and all reagents, with the exceptions noted below, were obtained from Ventana (Ventana Medical Systems, Inc., Tucson, Arizona). A serial immunohistochemistry (IHC) procedure was used to multilabel tissue sections. Sections were deparaffinized and rehydrated. Nonspecific staining was blocked with 3% hydrogen peroxide and protein block (DAKO, Carpinteria, California). Parietal cells were immunolabeled with a mono-clonal mouse anti-H+/K+ ATPase (Abcam Ltd., Cambridge, Massachusetts) at a concentration of 10 μg/ml for 1 hour. The primary antibody was labeled with a mouse Ultra MAP-HRP and visualized with 3’3-diaminobenzidine (DAB). Chief cells were immunolabeled with a monoclonal mouse antipepsinogen I (US Biologicals, Swampscott, Massachusetts) at a concentration of 5 μg/ml for 1 hour. Primary antibody was labeled with an antimouse IgG biotinylated secondary (Vector Lab, Burlingame, California), streptavidin-AP, and Fast Red chromagen. Sections were counterstained with hematoxylin, dehydrated, cleared, and coverslipped.
Following evaluation of the dual-labeled IHC described above, several sections of representative stomachs were immunolabeled for parietal cells only to allow for adequate contrast in low-magnification photomicrographs. Slides were processed as described above, and parietal cells were immunolabeled with a monoclonal mouse anti-H+/K+ ATPase (Abcam Ltd., Cambridge, Massachusetts) at a concentration of 10 μg/ml for 1 hour. Antibody was labeled with mouse Ultra MAP-AP and visualized with 5-Bromo-4-chloro-3-indolyl phosphate–nitroblue tetrazolium (BCIP-NBT). Sections were counterstained with nuclear fast red, dehydrated, cleared, and coverslipped.
Results
Macroscopic Evaluation
Within the gastric mucosa, 3 distinct regions could be identified macroscopically both before and after formalin fixation in all animals examined (Figure 2). The fundus was prominent, tan, and extended to approximately the level of the junction of the esophagus and stomach. The body was dark tan, and there was a visible demarcation between the fundus and body (Figure 2). Rugae were present in both the fundus and body, and the lesser curvature appeared thinner than the greater curvature. The antrum was well demarcated and pale tan (Figure 2). The cardia was ill-defined and could not be easily identified macroscopically.
Microscopic Evaluation
The cardia was ill-defined and displayed considerable variability across animals. There were numerous gastric pits (Figure 3, 1A), which were lined by tall columnar epithelial cells that displayed a mixture of AB- and PAS-positive cytoplasmic material (Figure 3, 1C). There was a basal layer of loosely arranged glands that contained PAS-positive cytoplasm (Figure 3, 1C). Within approximately 0.5 cm of the gastroesophageal junction, the glands became more tightly arranged, had a less prominent lumen, and contained variable numbers of parietal and chief cells. There were often lymphofollicular foci present within the lamina propria of the cardia (data not shown).
The fundic mucosa contained numerous deep gastric pits (foveolae), which extended from the surface to the midzonal region (Figure 3, 2A). The surface and pits were lined by tall, columnar epithelial cells that contained a distinct cytoplasmic vacuole that displayed a mixture of AB- and PAS-positive material (Figure 3, 2C). Basally, there was layer of loosely arranged glands with a distinct lumen that occasionally connected with the overlying pits (Figure 3, 2A). The superficial portion of the glandular layer contained a band of PAS-positive cells (Figure 3, 2C), while the basal portion contained numerous chief cells (Figure 3, 2B). Parietal cells were not present within the fundus of most animals (Figure 3, 2B), although individual and small clusters of parietal cells were infrequently identified (data not shown). Occasionally, the gastric pits contained material that immunolabeled for pepsinogen I (Figure 3, 2B). In addition, there were scattered, variably sized lymphoid aggregates and lymphofollicular foci present within the lamina propria. These foci occasionally extended through the muscularis mucosae and into the underlying submucosa (data not shown).
In longitudinal sections that span the fundus and body, a distinct and consistent transition between the fundus and body was observed microscopically in all animals examined (Figure 3, 3A–C, and Figure 4). The gastric pits shorten and the basal glandular layer lengthens and becomes tightly packed (Figure 3, 3A). There was an abrupt appearance of parietal cells that corresponded with the gross demarcation between fundus and body (Figure 3, 3B, and Figure 4). The mucosa within the body contained short gastric pits (Figure 3, 4A) lined by tall, columnar epithelial cells that contained a distinct cytoplasmic vacuole that displayed a mixture of AB- and PAS-positive material (Figure 3, 4C). The glands occupied the majority of the mucosa, were tightly packed, and lacked a distinct lumen (Figure 3, 3A). There was a large zone of parietal cells that extended from beneath the surface toward the basal aspect of the mucosa (Figure 3, 4B). Parietal cells often contained distinct, cytoplasmic, PAS-positive granules, which were more prominent superficially (data not shown). There was a midzonal band of PAS-positive mucous neck cells (Figure 3, 4C). The chief cells were most prominent basally and formed a small band at the junction with the muscularis (Figure 3, 4B), with lesser numbers of chief cells that extended superficially as individual cells or small clusters of cells. The proximal portion of the body contained more chief cells than the distal portions of body, and in several instances, portions of distal body appeared to lack a chief cell layer (data not shown). The greater curvature and lesser curvature were similar in histologic appearance, although the lesser curvature was noticeably thinner than the greater curvature (Figure 3, 4A–C and 5A–C). There were scattered, variably sized lymphoid aggregates and lymphofollicular foci present within the lamina propria, which occasionally extended through the muscularis mucosae and into the underlying submucosa (data not shown).
The antral mucosa contained numerous deep gastric pits, which extended from the surface to the midzonal region (Figure 3, 6A). The surface and pits were lined by tall, columnar epithelial cells that contained a distinct cytoplasmic vacuole that displayed a mixture of AB- and PAS-positive material (Figure 3, 6C). Basally, there was a layer of loosely arranged glands with a distinct lumen that occasionally connected with the overlying pits (Figure 3, 6A). The cytoplasm of the cells within the glandular layer was strongly PAS positive (Figure 3, 6C) and closely resembled the Brunner’s glands within the duodenum. There were rare chief cells present basally, but no parietal cells were identified (Figure 3, 6B). In addition, there were scattered, variably sized lymphoid aggregates and lymphofollicular foci present within the lamina propria. These foci occasionally extended through the muscularis mucosae and into the underlying sub-mucosa (data not shown).
Discussion
An understanding of the normal anatomy of a given species is critical to provide a basis for repeatable sampling and distinguishing normal anatomic variability from a drug-induced finding. Interpreting differences between species is complicated by inconsistent anatomic terminology when referring to the stomach. Typically, the stomach is divided into several macroscopically identifiable regions including the fundus, cardia, body, and antrum (Soybel, 2005). The fundus is the dome-shaped portion that extends above the gastroesophageal junction, while the cardia is a small, ill-defined portion of the gastric mucosa that is present immediately distal to the esophagus (Owen, 1997). The antrum is the distal third of the stomach, with the remaining and largest portion of the stomach representing the body (Soybel, 2005). In humans, the fundus and body are similar histologically with abundant parietal and chief cells, and as a result, this classic gastric phenotype has been termed the fundic mucosa (Berger, 1934). This can present a problem when evaluating a given sampling procedure, as a sample of fundic mucosa may not arise from the fundus. Similar confusion can arise when describing the cardia. The cardia is a small, ill-defined region in humans, characterized by the presence of mucus-secreting cells and a lack of parietal and chief cells (Owen, 1997). This mucosa is often termed the cardiac mucosa and is used in pigs to define a large area of the stomach that occupies not only the classic location of cardia but also the fundus and portions of the body (Kararli, 1995). This presents little problem in humans and in dogs with a similar-appearing fundus and body; however, nonhuman primates and pigs display significant differences in histologic appearance of the fundus and body, and without a clear understanding of the location and histology, interpretation of toxicity data is complicated or impossible.
Despite phylogenetic similarities, Old World primates and humans have differing gastrointestinal anatomy, and differences in the histology of the fundus and body of the stomach have been reported in several nonhuman primate species (Berger, 1934; Tokii and Tsukamoto, 1953; Kent, 1966; Sheahan and Jervis, 1976; Owen, 1997). To date, there has not been a systematic evaluation of the microscopic anatomy of the cynomolgus monkey, which is the most commonly used nonhuman primate in drug-safety evaluation. In this study, an extensive sectioning method allowed for correlation of the gross subdivisions (fundus, cardia, body, and antrum) with the histologic distribution of parietal, chief, and mucus cells. The cynomolgus monkey has a prominent fundus that extends to the level of the gastroesophageal junction. Macroscopically, the fundus can be distinguished from the body, as the fundus is a lighter shade of tan. The fundus of the cynomolgus monkey has prominent gastric pits, lacks parietal cells, has extensive mucous-secreting cells, and is similar in size and appearance to Rhesus, Formosan, and Japanese macaques (Tokii and Tsukamoto, 1953). The body of the cynomolgus monkey is histologically similar to that reported in the human, and although similar to the described “fundic mucosa,” care must be taken in using this term, as the mucosa of the true fundus in cynomolgus monkeys is distinct. The proximal portions of body contained more chief cells than the distal portions, as has been reported in the human, dog, and cat (Hogben et al., 1974). The cardia was similar to that of humans in size and histology but smaller than that reported in the baboon (Sheahan and Jervis, 1976). The antrum was similar in size and consistent in histologic appearance with that reported in other species (Sheahan and Jervis, 1976).
To provide a consistent and thorough evaluation of the stomach in preclinical toxicology studies, a standardized approach to processing and sampling is necessary to minimize variability within and across studies. Typical sections of the stomach evaluated in a nonrodent toxicology study include the gastroesophageal junction with cardia, gastroduodenal junction with pylorus and antrum, and the body. Sections of cardia and antrum pose little difficulty, but care must be taken when sampling the body in the monkey. Often, samples of the body are taken perpendicular to the long axis of the stomach and may be inadvertently taken from the fundus or transition between the fundus and body. Without an understanding of the normal gastric anatomy of the cynomolgus monkey, differences in parietal, chief, and mucus-cell distribution may be incorrectly interpreted as a drug-induced finding. To standardize the sampling methods for the cynomolgus monkey stomach, our lab has adopted the sectioning pattern outlined in Figure 5. This method includes 3 sections taken along the long axis of the stomach. The first section includes the gastroesophageal junction and cardia, but the section is extended distally to include a portion of lesser curvature. The second section includes the gastroduodenal junction and antrum. These 2 sections (approximately 3.0 cm each) can easily be embedded into a single paraffin block. The third sample includes a longitudinal section spanning the fundus and body (approximately 4 to 5 cm) and is embedded in a second block. Samples from smaller monkeys can be embedded easily, but sections from larger monkeys may require the sample to be curved or bent into a “U” shape with the mucosa oriented along the outer surface of the curve. This sectioning pattern is based on the extensive sectioning pattern used in this study. This method allows for rapid and efficient processing in nonclinical toxicology studies, while offering a complete evaluation of all relevant histologic phenotypes in the cynomolgus monkey stomach.
Footnotes
Acknowledgments
The authors thank the UM SA Histology Laboratory for kindly supporting the additional trimming and sectioning of the many stomachs used in this study and Tom Covatta and Beverly Maleeff for outstanding photographic support.