Abstract
Background Mixed endocrine-exocrine tumors display histologic features of endocrine and glandular differentiation. Unlike in collision tumors, the two components are thought to arise from a monoclonal precursor. Evidence from molecular testing supports the monoclonal theory and suggests that the exocrine component may give rise to the endocrine component but not vice versa.
Case Report We report a case of an adenomatous polyp in the large intestine that had groups of endocrine cells arising from the crypt bases of the adenomatous (exocrine) epithelium. To our knowledge, ours is only the second report of an adenomatous polyp in which groups of microcarcinoid endocrine cells were recognized. The histologic findings in our case correlate with the molecular findings described in mixed endocrine-exocrine tumors.
Conclusion Our description may represent the primordial stage of a mixed endocrine-exocrine neoplasm.
INTRODUCTION
Mixed endocrine-exocrine neoplasms of the gastrointestinal (GI) tract are rare tumors that demonstrate two heterogeneous histologic components, namely endocrine and glandular (exocrine) components. The endocrine component can be poorly differentiated or well differentiated (ie, carcinoid tumor). Similarly, the exocrine component can be adenomatous or carcinomatous. The endocrine component comprises one-third to one-half of the tumor bulk and can either admix intimately and diffusely with the exocrine component (combined tumor) or exist in a separate area with a transition zone between the two components (composite tumor).1 Alternatively, endocrine-exocrine collision tumors are composed of two topographically distinct components and as such are not considered mixed.2 Mixed tumors may demonstrate malignancy that usually involves the exocrine component, whereas simultaneous growth of an adenoma and a carcinoid is rare.3 Considering the degree of heterogeneity in grade and differentiation, different combinations have been reported (Table). Therefore, it is not surprising that some of these tumors will exhibit a benign behavior, while others may be aggressively invasive and metastatic (with distant metastases being endocrine or exocrine).4,5-8 One report described distant lymph node metastases containing both components with a transition zone, identical to the original tumor.9 These tumors can emerge anywhere in the GI tract, including the esophagus (usually in the setting of Barrett esophagus),10 stomach,9 small intestine,8 cecum,4 colon,11,12 rectum,13,14 and anal canal.15 They can occur with other GI conditions, such as atrophic gastritis16 or familial adenomatous polyposis syndrome.17 The origin of these tumors and whether both components share a mutual precursor have been subjects of debate.
Examples of Different Combinations of Exocrine and Endocrine Components in Terms of Differentiation and Grade
In 2006, Pulitzer et al described four cases of adenomatous polyps in the large intestine with concomitant microcarcinoid, an entity that had not been described previously, and discussed its histologic and immunohistochemical attributes.18 In this article, we report a case that further characterizes this entity and its histogenesis.
CASE REPORT
A 72-year-old man underwent colonoscopy screening. Two polyps, a 0.3 cm sessile polyp in the ascending colon and a 2.0 cm pedunculated polyp in the descending colon, were discovered.
Histology
Evaluation of the hematoxylin and eosin stained slides of the descending colon polyp revealed tubulovillous adenoma with no high-grade dysplasia (Figure). Careful examination of the base of the crypts to rule out microscopic invasion revealed multiple minute foci of morphologically distinct groups of cells that had many features of well-differentiated endocrine (carcinoid) tumors. These cells were arranged in nests, cords, irregular clusters, and single cells and appeared to superficially invade the muscularis mucosa. They had relatively abundant eosinophilic, granular cytoplasm and bland, round, central nuclei with finely stippled chromatin. Most of the nuclei had a small central nucleolus. No mitotic figures were identified within these groups of cells. These foci were present in multiple areas throughout the polyp and did not appear to alter or disrupt the overall configuration of its architecture, an observation that was also made by Pulitzer et al.18 The largest focus was approximately 3 mm in diameter. The most interesting phenomenon was that these nests appeared to be budding directly from the base of the crypts where the adjacent adenomatous epithelium appeared to have undertaken similar histologic differentiation, as it had abundant eosinophilic cytoplasm, very similar to the cells described above. In the Figure (photomicrograph C), we capture this smooth transition from the usual adenomatous epithelium to adenomatous epithelium with more abundant eosinophilic cytoplasm and finally to the nests of endocrine cells that superficially infiltrate the muscularis propria. The ascending colon polyp was a regular tubular adenoma with no unusual features.
A: On low magnification, a small endocrine component might be overlooked (hematoxylin and eosin, 4×). B: On medium magnification, groups of cells with endocrine features can be seen (hematoxylin and eosin, 20×). C: An area of transition between usual adenomatous epithelium (arrow) to adenomatous epithelium with more abundant eosinophilic cytoplasm (arrowhead) to nests of endocrine cells (star) (hematoxylin and eosin, 40×). D: Endocrine cells and the adjacent adenomatous epithelium stain positive with synaptophysin, while the usual adenomatous epithelium is negative (hematoxylin and eosin, 40×).
Immunohistochemistry
The microcarcinoid component stained positively with synaptophysin and neuron-specific enolase and negatively with chromogranin. Interestingly, an identical pattern of staining was evident in the adjacent adenomatous epithelium with abundant eosinophilic cytoplasm.
Follow-up
The patient underwent a follow-up colonoscopy a year later that revealed two more polyps in the large intestine, both of which were tubular adenomas with no unusual features.
DISCUSSION
Two hypotheses discuss the origin of mixed endocrine-exocrine tumors. In the first hypothesis, these lesions could be understood as two distinct tumors derived from two different precursors that coincidentally arise in the same location. This hypothesis could account for the histogenesis of collision-type tumors that are by definition excluded from the mixed tumor category. The second hypothesis postulates a monoclonal origin for the tumors: the two components may arise from a pluripotent epithelial stem cell capable of bidirectional differentiation.2 In this section, we discuss several origin-related issues, including the embryologic monoderivation of epithelial and endocrine cells in the GI tract, the close relationship between glandular tumors and endocrine cells, and the genetic/molecular studies of these tumors. Last, we attempt to correlate these aspects with the findings of our case.
Since the beginning of the 20th century, the origin of gut endocrine cells has been controversial. There were two competing theories. The first postulated that gut endocrine cells are endodermally derived like the rest of the GI epithelial cells.5 The second theory accounted for the cytochemical and ultrastructural similarities between these cells and other extra-GI endocrine cells (eg, parafollicular thyroid cells). All endocrine cells that display these features within and outside the GI tract were designated collectively as amine precursor uptake and decarboxylation cells, and the neural crest was thought to be a possible origin.19
In 1990, Thompson et al found that endocrine cells have the same clonal origin as other gut cell lineages, using an in situ DNA hybridization technique in chimeric mice.6 Today, it is widely accepted that GI endocrine cells share the endodermal stem cell pool with other epithelial cells of this tract.20,21 This finding is relevant because it provides strong evidence that the stem cells that give rise to the GI tract epithelial cells also give rise to the endocrine cells within it. Therefore, these cells are capable of dual differentiation toward endocrine and exocrine phenotypes.
The association between endocrine and exocrine cells goes beyond sharing a common progenitor. Several studies have shown the frequent presence of endocrine cells in colorectal adenomas, ranging from 59%-76%.22,23 These cells were so numerous in some adenomas that they were regarded as an intrinsic part of the tumor.23 Adenomas are monoclonal proliferations,24 and the presence of endocrine cells within them is a clue that their precursor clone is capable of differentiation toward the endocrine phenotype while giving rise to adenomatous exocrine cells.
The ability of a clone to differentiate into both endocrine and exocrine phenotypes is suggestive of a monoclonal origin for mixed neoplasms; however, this differentiation ability is not considered sufficient evidence of a monoclonal origin. Several studies have investigated this theory at the molecular and genetic level. Furlan et al studied five mixed endocrine-exocrine neoplasms and one collision tumor using allelotyping analysis, a technique that offers valuable information when studying a lesion that has components with different histologic phenotypes or grades.2 Allelotyping analysis detects consistently prevalent genetic alterations that are considered an early event in tumorigenesis, regardless of the histologic phenotype. This high prevalence implies that a precursor cell underwent these changes and that every component that displays them must have risen from this precursor (monoclonal origin). According to this model, additional genetic alterations restricted to specific histologic phenotypes are interpreted as events occurring later in the progression of the tumor. The same principle can be applied to tumors with one histologic phenotype but different grades of differentiation.
Furlan et al found that the endocrine and exocrine components in all five mixed neoplasms exhibited common genetic alterations, supporting the monoclonal theory. In addition, the endocrine component exhibited secondary alterations that were not present in the exocrine component and thus represented a later event. Therefore, the authors concluded that differentiation may occur from an exocrine to endocrine direction but not vice versa. Furlan et al also noted that in some cases the endocrine component was present in association with a minimal component of adenoma, suggesting that differentiation into an endocrine phenotype may have occurred at an early stage. The collision tumor in their study had different endocrine and exocrine genetics, suggesting a biclonal precursor.2 Similar findings were demonstrated by other investigators using different methods.25,26
In our case, small groups of endocrine cells were seen budding from the crypts of the larger adenomatous (exocrine) component, distinguishing the tumor from collision tumors and also suggesting that endocrine cells were differentiating from exocrine cells and not vice versa. Furthermore, this lesion was still in its early stage; it mainly consisted of a small adenoma without high-grade dysplasia and had a microcarcinoid component. Although our report is limited by its single case size, as well as the lack of molecular confirmation to detect consistent, prevalent genetic alterations, we believe that this case may represent a very early precursor of a mixed endocrine-exocrine tumor that has not been described before. In addition, our case correlates with previous findings about this group of neoplasms. If such tumors remain untreated, the adenoma component may evolve into an invasive adenocarcinoma, and the microcarcinoid component may turn into a well-differentiated endocrine neoplasm (carcinoid vs malignant carcinoid). The lesion would meet the criteria for a mixed endocrine-exocrine tumor once the microcarcinoid component comprises one-third of the tumor bulk. If adenocarcinoma arises, it may completely destroy the adjacent endocrine component, especially if this component is in its early stages. Interestingly, grade progression from well-differentiated to poorly differentiated is very rare in endocrine tumors.27
In our opinion, the incidence of mixed endocrine-exocrine neoplasms could be more frequent than reported. Underreporting may be attributed to several reasons. For example, an early small endocrine component may be overlooked (Figure, photomicrograph A) or misinterpreted as squamous metaplasia or microinvasion.18 Moreover, such a component may be destroyed by locally growing adenocarcinoma as hypothesized above, or an adenomatous component may be destroyed by locally growing poorly differentiated endocrine carcinoma. The end result in both scenarios would be a tumor with one histologic type, leaving no evidence that it arose as a mixed neoplasm.
CONCLUSION
In this case report, we provide histologic and immunohistochemical evidence to describe a mixed endocrine and exocrine tumor of the colon with an early precursor form.
This article meets the Accreditation Council for Graduate Medical Education and the American Board of Medical Specialties Maintenance of Certification competencies for Patient Care and Medical Knowledge.
Footnotes
The authors have no financial or proprietary interest in the subject matter of this article.
- © Academic Division of Ochsner Clinic Foundation