The problem of replacement and differentiation of the intestinal epithelium: its relation to squamous metaplasia in the uterine cervix.

Abstract

In 1971 and 1972 I received training in cytopathology in the laboratory of the late Dr. James Reagan in Cleveland, Ohio. A few years before my arrival, Dr. Stanley Patten had worked in this laboratory before moving to the Department of Pathology of the University of Rochester in Rochester, New York. As part of my training in cytopathology, I have worked in a number of cytopathology laboratories in the U.S. I visited Dr. Patten's laboratory on several occasions, initially for a full month. At that time we discussed cervical cytopathology, in particular the etiology and morphogenesis of cervical carcinoma. I vividly remember discussions with Dr. Patten on the origin and nature of squamous metaplasia, immature squamous metaplasia, and reserve cell hyperplasia. At that time, the theory of the origin of reserve cells from cervical stromal cells still was the most popular theory. Dr. Patten and I discussed at length what the scientific basis for such an origin could be. An origin of the reserve cells in the coelomic epithelium appeared a logical hypothesis. We did not yet have the immunocytochemical markers that later proved this theory to be correct. Over the years this scientific question has continued to intrigue me. We have performed extensive research to clarify the origin of these cells. The following excerpts from three Ph.D. studies performed in my own laboratory on these subjects are a tribute to Dr. Patten's scientific and personal support to these studies. To illustrate the relationship with this original Ph.D. study, I have listed a few excerpts of Dr. Patten's Ph.D. thesis as well as excerpts from the publications of my own Ph.D. students, which can be considered as a follow-up of his initial line of thinking. By Stanley Fletcher Patten, Jr. B.S., 1949, The George Washington University M.S., 1950, The George Washington University A dissertation submitted to the Faculty of the Graduate Council of The George Washington University in partial satisfaction of the requirements for the degree of Doctor of Philosophy May 27, 1953 "A critical study of the cells composing the intestinal epithelium of the Wistar rat and Urodele was made with respect to their general morphological characteristics as well as their method of replacement in the adult animal. Seven species of North American salamanders, one European species, and one Japanese species were utilized. In addition to the study of replacement of the intestinal epithelium, a survey of the so-called intestinal glands of the Urodele was made with particular emphasis on their function and possible relationship to the crypts of Lieberkühn of higher vertebrates." "Submucosal aggregations of cells or "cell-nests" were observed in the 382 salamanders examined. These aggregations were found to be transient structures having no specific site within the intestinal submucosa, being either attached too or entirely separate from the mucosa and entirely dependent upon the functional state of the intestinal mucosa. Their function appears to be that of epithelial replacement alone, although early differentiation's of the nodal cells may result in the formation of mucous elements." "The apparent absence of endodermal evagination, the relative lack of mitotic activity, and the presence of mesodermal-like cells in various stages of nidal cell formation lend support to the concept that the intestinal epithelium is replaced by cells of mesodermal origin, or "lymphoid-hemoblastic" elements in the case of the Urodele." "Because the cell-nest is a constantly occurring entity within the intestine, it must serve a useful function. As previously stated, these structures have in the past been credited with the function of either epithelial cell replacement or glandular secretion, or a combination of the two. The evidence in the present study leads the writer to the conclusion that the sole function of the cell-nest is that of providing replacement elements for the intestinal epithelium." "The origin of the cell-nest is an interesting problem, particularly when viewed from the phylogenetic aspect. As pointed out by reference to the work of previous investigators, as well as by several observations made in the present study, the Urodele holds a unique position among the vertebrates by the possession of these aggregations of cells." "These structures have not been observed in the fish, reptile, or even in other amphibians of a higher order." "There are two possible explanations which might be considered. The first, which perhaps would be readily acceptable to the embryologist, is the possible withholding of certain primitive endodermal cells during the development of the intestinal epithelium. These primitive cells, by repeated division, could form the definitive nest, and upon exhaustion of the nest, further divisions of a remaining cell could reestablish the cell-nest. This would fit in with the concept of the cell-nest acting as a transient structure as well as into the picture of a fold depleted of nests, since individual epithelial-like nuclei are observed within the stroma of the fold. Without the process of endodermal evagination or the mitotic division of basally located primitive endodermal elements, it is natural to look toward cells arising from other primitive germ layers which might possibly provide replacement cells for the intestinal epithelium, after once forming a cell-nest. The concept of differentiating mitoses and the lack of absolute fixation of cells derived from the primitive germ layers seems to be gaining some prominence today. Cytochemical techniques and tissue culture have added greatly to our knowledge of these processes "there seems to be no fundamental difference between epithelium and connective tissue." "Gruenwald (1942) demonstrated that tissues originating from the mesothelium in the embryo retain, permanently or for a limited period of time, one important characteristic of their parent tissue: a combination of epithelial and mesenchymal properties and potencies." "Thus, these investigators have attempted to break down the classical conception that a definite separation of epithelial and mesenchymal tissues is established soon after the earliest periods of embryonic development. Processes such as the differentiation of the kidney epithelia from a mesenchymal blastema should not be considered merely an exception to the general rule. The complexities of abnormal growth in the gonads or tumors related to the adrenal cortex might be partially eliminated if the clear-cut separation of epithelial and mesenchymal elements was not maintained. From the work done by Gruenwald, it would appear that tissues arising from the coelomic wall can produce an unusually wide variety of pathologic growth due to the wide range of differentiation of these cells." "An undifferentiated cell of mesodermal or mesenchymal origin would be a likely replacement cell for the following reasons: (1) mesodermal cells represent a readily available source of replacement cells because of their frequency within the intestinal stroma, which would be an important consideration due to the scarcity of mitotic divisions; (2) cells of mesodermal origin are relatively undifferentiated and possess the ability to travel varied differentiation pathways; (3) these cells generally possess polymorphic nuclei that are quite basophilic; (4) these cells have been termed "wandering cells"; and (5) mesodermal cells can produce a mucous secretion, although there is some question as the whether it is the same as that produced by the goblet cell. These properties of the mesodermal cell, if their differentiation into cell-nest elements could be accepted, would lend themselves ideally to their role in epithelial replacement." "The evidence accumulated in this study would seem to lend support to the conception of epithelial replacement by cells of mesodermal or mesenchymal origin. The question naturally arises, however, as to the exact type of mesodermal or mesenchymal cell present in the intestine that would be available. Therefore, the presence of certain cells of mesodermal or mesenchymal origin within the submucosa under normal conditions, an increase in these cells in situations calling for a mobilization of the cell-nests, the appearance of transitional forms within the cell-nest, and the virtual absence of mitotic activity all lend support to the supposition that the intestinal epithelium of the Urodele is replaced by differentiated mesenchymal cells." "It is therefore suggested that within the intestine of the Urodele, aggregations of cells formed by the differentiation of cells having a mesenchymal origin constitute the primary means of replacement of the intestinal epithelium." Dr. Patten referred to the origin of the reserve cells, and their role in the renewal and metaplastic change of the cells of the endometrial canal in his monograph, Diagnostic Cytology and Its Histopathologic Basis. The relationship with the theories developed during his thesis work is evident. "The actual mechanism of squamous metaplasia has been the subject of numerous reports since the terms "epidermidization" and "epidermoidization" were applied to this process in the latter part of the nineteenth century. As the earliest stage of squamous metaplasia in the human uterine cervix is characterized by the appearance of a layer of primitive cells beneath the endocervical columnar cells, the area of greatest interest and controversy resides in the origin of these cells. Briefly, the hypotheses concerning their genesis include: 1. an ingrowth of basal cells from the stratum germinativum of adjacent normal stratified squamous epithelium, 2. origin from fetal squamous basal cells in pre-existing stratified squamous lining of the urogenital sinus, 3. origin from undifferentiated fetal rests, 4. origin from endocervical columnar cells and 5. origin from cervical stromal cells. Although Fluhmann rather arbitrarily states that the primitive subcolumnar cells are of epithelial origin and arise above the basement membrane directly from the columnar cells by a process termed "squamous prosoplasia," the evidence is lacking to definitely exclude their origin from stromal cells. Since the lining epithelium in this site is derived embryologically from the mesodermal urogenital ridge and subsequent Müllerian system, it is not embryologic heresay to suggest that the stroma of the uterine cervix may retain certain of its embryonic multipotentialities to supply replacement cells through a poorly defined "basement membrane." Certain cytologic features of the subcolumnar cell and its probable abnormal counterpart in carcinoma in situ would lend support to Song's hypothesis of a stromal origin. However, the argument might persist as to whether migration from the stroma occurs during fetal development, as in the fetal rest hypothesis, or during adult life following proper stimulation. Again, definitive studies on this question are lacking." "The concept of reserve cell hyperplasia as a stage in squamous metaplasia, based on the embryonic rest hypothesis, was initially proposed by Eichholz. Reserve cell hyperplasia as a pathologic entity was reviewed as part of a symposium in Acta Cytologica in 1961, by Von Haam and Old in 1964 and by Song in 1964. "Reserve cell hyperplasia may be defined as the presence of one or more layers of primitive cells, in a subcolumnar position between an overlying endocervical lining epithelium and underlying "basement membrane." In the author's opinion, the earliest form of reserve cell hyperplasia is a single layer of subcolumnar cells. Some authors state this does not represent hyperplasia and require a minimum of two or three layers of primitive subcolumnar reserve cells. However, the proponents of the embryonic rest hypothesis, whether the rest are of epithelial or stromal origin, would have to agree that a significant linebr distribution of a single layer of subcolumnar reserve cells in the endocervical canal, would most likely arise as a result of reserve cell proliferation, this increase in number of cells fitting the definition of hyperplasia." "Returning to the discussion of a possible stromal origin for the reserve cell, at the ultrastructural level, Song has reported the presence of lamellated bodies in the cytoplasm of fetal reserve cells, presumably representing collagen material. He described similar structures in superficial stromal fibroblasts. At the light microscope level, the nuclear configurations of the reserve cell more closely resemble those of the superficial stromal cells as well as small histiocytes in the uterine cervix, suggesting a possible relationship between those cell types." "Reserve cell hyperplasia has been described in the uterine cervix of the developing fetus, in the newborn, less often in children and young adults, and more commonly during pregnancy and later adult life." "In all instances reserve cell hyperplasia was related to the endocervical canal with the majority of the reactions falling within the distribution of the transitional zone. The site of maximum involvement was somewhat proximal or higher in the canal than the site of maximal involvement of immature squamous metaplasia to be described. In this respect, approximately 60% of the involvement in reserve cell hyperplasia was within the distribution of immature squamous metaplasia whereas about 40% was more proximal in the cervical canal. With coexistence of immature squamous metaplasia and reserve cell hyperplasia in the same cervix, a more or less constant anatomical relationship existed with reserve cell hyperplasia lying proximal in all instances. The progressive reduction in the linebr extent of reserve cell hyperplasia in relation to age together with its anatomical distribution, would lend support to the concept that this reaction represents the earliest stage in squamous metaplasia in the uterine cervix." With the increasing availability of immunocytochemical techniques, we have been able to contribute a few stones to the building of the morphogenisis of cervical carcinoma, in particular the function of the reserve cell as a multipotential stem cell originating in the coelomic cavity, and the common pathway of squamous as well as glandular lesions of the uterine cervix have been studied. From the studies of Frank Smedts it became evident that the reserve cells, when appearing as a single layer of subcolumnar cells, show that the multipotential character based on the environment of the uterine cervix may develop into squamous metaplastic cells or glandular cells. It is shown that the development of epithelial abnormalities from these metaplastic cells is paralleled by the loss of specific cellular fractions such as specific subkeratin proteins (From the thesis entitled "Keratin phenotyping of the normal, premalignant and malignant cervix." Nijmegen 1993). "It is believed that reserve cells play a central role in the pathogenesis of cervical intraepithelial neoplasia. Considerable efforts have been undertaken in the study of the exact nature of these cells and the precise role they play in neoplastic transformation. One of the recent approaches that has been chosen is the characterization of cytoskeletal components, in particular keratins, in reserve cell and cervical intraepithelial neoplastic (CIN) lesions. With chain specific monoclonal antibodies against various individual keratins, information concerning the keratin content of reserve cells and other cervical epithelial cells can be easily obtained with immunocytochemical techniques. By comparing the keratin distribution patterns in CIN with that in reserve cells, a possible link between the various stages of cervical carcinogenesis can be studied." Reserve cells are reported to contain keratins 5, 8, 17, 18 and 19, with the presence of keratin 13 in a subpopulation of these cells. Reserve cell proliferations can lead to immature squamous metaplastic epithelium, which can transform into mature squamous metaplastic epithelium. Conversely, the reserve cell proliferation may be atypical and progress into intraepithelial neoplasia. Immature squamous metaplastic epithelium has been found to express keratins characteristic of simple epithelia, such as keratins 8, 18, and 19, as well as keratins found in ectocervical and mature squamous epithelium (i.e., keratins 4, 13, and 14). Expression of keratins 5, 14, and 17 in endocervical subcolumnar reserve cells was detected by means of immunohistochemical methods using polypeptide specific monoclonal antibodies. These particular keratins that were found among others in basal cells also could be detected to a variable extend in metaplastic and dysplastic cervical lesions. In some cases of immature squamous metaplasia all the keratin subtypes were expressed throughout the full thickness of the epithelium. In contrast, in mature squamous metaplasia a compartmentalization of these keratins was observed. Mature squamous metaplastic epithelium showed a keratin distribution pattern comparable to ectocervical squamous epithelium, with the exception of keratin 17, which was only found sporadically in the basal layer of ectocervical epithelium and always was present in the basal cells of mature squamous metaplastic epithelium. During progression of cervical intraepithelial neoplasia a clear increase in the expression of keratin 17 was observed. However keratins 5 and 14 also were expressed. Our results demonstrate that a considerable number of premalignant lesions of the uterine cervix express the same keratins as found in the progenitor reserve cells. Lesions that show no expression of keratin 17 are suggested to form a distinct group that is regressive in nature and does not progress into cervical carcinoma. The keratin content of reserve cells has been shown to be highly complex. Stromal cells as yet have not been shown to contain keratins. Vimentin, which has been found in all stromal cells, has not been observed in reserve cells. In general, transitions from one type of epithelium to another have been shown to be correlated with gradual keratin changes. Therefore it appears highly unlikely that stromal cells migrating through a basal membrane will initiate the expression of up to 10 different keratins. In the transitional phase, when they are in the endocervical stroma, stromal cells contain no cytoskeletal proteins of the keratin type. In general when nonepithelial cells obtain keratins, they normally comprise only a limited subset of keratin 8 and 18. It has been shown that the keratin phenotype of reserve cells is more complex than that of endocervical columnar cells. Thus columnar cells would have to initiate synthesis of a considerable number of differentiation related keratins when transforming into reserve cells. This is an extremely unlikely pathway for the genesis of reserve cell. In contrast the argument that multipotent stemcells (reserve cells) are present in the endocervical canal having originated there during embryonal development, is still valid. The studies by Frank Smedts were initiated in the thesis by Jo Puts entitled "Intermediate filament proteins as markers in histo- and cytopathology, with emphasis on gynecological neoplasia" (Nijmegen, 1986) In the thesis work by Anniek Van Aspert van Erp we have investigated comparable processes in the development of glandular abnormalities of the uterine cervix. We have illustrated the combined occurrence of squamous and glandular lesions, which are much more frequently found than originally thought. This greatly supports the role of the reserve cell as the multipotential stem cell of cervical epithelium and of the abnormalities originating in it. "ENDOCERVICAL COLUMNAR CELL INTRAEPITHELIAL NEOPLASIA ECCIN): Cytomorphologic characteristics and accuracy of diagnosis" (thesis Nijmegen, 1995) "It is a widely accepted hypothesis that the majority of squamous (metaplastic) abnormalities of the cervix develop from "reserve cells." "We support the concept of a common origin for the majority of squamous metaplastic and columnar cell abnormalities because of the frequently combined occurrence of squamous (metaplastic) abnormalities and columnar cell lesions." "We consider the reserve cells to be multipotential Müllerian precursor cells that, depending on the nature of the stimulus and in part on their proximal or distal localization in the endocervical canal, can mature into normal mucus producing columnar cells or through metaplasia and to a variable degree disturbed maturation, into "normal" or more or less abnormal squamous metaplastic epithelium. We support the concept of a common origin for the majority of squamous metaplastic and columnar cell abnormalities because of the frequent combined occurrence of squamous (metaplastic) abnormalities and columnar cell lesions. Also the cooccurrence of squamous and columnar elements in mixed adenosquamous adenocarcinomas could be explained on this basis. The diagnosis of these combined lesions depends upon the awareness of a possibility that both entities may occur in the same cervix. Since the morphologic characteristics of columnar cell lesions are relatively unfamiliar to many pathologists, columnar cell lesions cooccurring with squamous lesions frequently will not be recognized." "In endocervical columnar cell intraepithelial neoplasia (ECCIN) the progression to an invasive adenocarcinoma should be considered as a continuum of changes that gradually increase in severity, finally leading to an uncoordinated and uncontrolled proliferation of the columnar epithelium. Cytologic prediction of a columnar cell atypia is important because of the subsequent surgical therapy. When a conventional biopsy is taken there is a risk that the glandular lesion, which generally is located more proximally in the endocervical canal, is missed and is left in situ." "The frequent combined occurrence of squamous (metaplastic) abnormalities and columnar cell lesions, the finding of columnar cells in squamous (metaplastic) abnormalities and the presence of squamous metaplastic cells in columnar cell lesions as well as the co-occurrence of squamous and columnar elements in mixed adenosquamous carcinomas, supports the hypothesis of a common origin for the majority of squamous metaplastic and columnar cell abnormalities. The diagnosis of these "double" tumors depends upon the awareness of the possibility that both entities may coexist in the same cervix. Since the morphologic characteristics of columnar cell lesions are relatively unfamiliar to many pathologists, columnar cell lesions co-occurring with squamous lesions will frequently not be recognized. In analogy to progressively abnormal squamous (metaplastic) intraepithelial lesions as precursors of squamous cell carcinoma, also in endocervical columnar cell intraepithelial neoplasia the progression to an invasive adenocarcinoma should considered as a continuum of changes that gradually increase in severity, finally leading to an uncoordinated and uncontrolled proliferation of the columnar epithelium." With this work, initiated by Dr. Patten, I hope that some of the puzzles in the morphogenesis of cervical carcinoma have been solved. Apart from our very frequent personal communications, these subjects very often were the basis of scientific discussions with Dr. Patten that were always inspiring, and very frequently gave me clues as to further pathways to be explored.