Molecular markers can be applied to study the complex embryonic transition of the Müllerian duct epithelium into the squamous epithelium of the vagina and ectocervix on the one hand and into the columnar epithelium of the endocervix on the other [1, 17, 22, 23]. Also, the stage at which an early or primitive squamocolumnar junction is formed can be studied by applying such markers by immunophenotyping. (Cyto)keratins are differentiation markers for the different types of epithelial cells in the cervix and have been applied to epithelial neoplasms to determine cell lineage and differentiation and for differential diagnosis [26–30]. For example, antibodies directed to keratin 5 (k5) and k17 can be used to detect squamous differentiation, while k7 and k8 are characteristic of glandular differentiation. The protein p63, a p53 homologue and a transcription factor, is required for normal development and epithelial cell differentiation in the vagina and endocervix [2, 23]. As a marker, p63 is used for recognition of basal squamous cells and other types of progenitor cells in epithelial tissues. BCL2 is a marker for cell survival and an indicator of protection against apoptosis and can be used as such to recognize stem cells [31].

Two studies complement each other in describing the spatiotemporal distribution of these relevant markers in the developing lower uterine tract, including the developing cervix and vagina [1, 22]. In weeks 14–15, the squamous markers p63 and k17 are found in the developing lower and upper vagina (vaginal anlage), while k8 is negative [22]. BCL2 was shown to be strongly positive in basal cells of this epithelium. The columnar epithelium of the cervical segment showed scattered BCL2-positive cells and was described to be negative for k8 at this stage. The human Müllerian duct epithelium between weeks 16 and 18 of gestation is strongly positive for k7 and k8, typical of primitive columnar epithelium. k17 and p63 cannot be detected immunohistochemically in this period, while k5 and BCL2 are weakly positive (see Fig. 1.2, left panels) [1]. Fritsch et al. [22] have shown that the upper vaginal Müllerian epithelium is negative or only weakly positive for k8 at this early stage, while being positive for k17 and p63, both indicators of squamous differentiation. In weeks 19–25 a primitive squamocolumnar junction can be recognized as a transition between the uterine cavity and the solid cord epithelium. This transition becomes recognizable on the basis of the immunostaining results with the aforementioned antibodies (see Fig. 1.2; middle panels). A solid cord-like structure with a stratified appearance merges with the columnar Müllerian epithelium. In the solid cord the typical markers for programming of squamous differentiation, i.e., k5, k17, and p63, are strongly positive, with additional reactivity for k7 and k8. The basal cell compartment in this solid cord epithelium is weakly positive for k7, k8, and k17. The Müllerian columnar epithelium is strongly positive for k5, k7, k8, and BCL2, but negative for k17. At this stage basal cells can be recognized in the Müllerian duct epithelium as p63-positive cells. These cells were identified as reserve cells (RC) by Martens et al. [1]. It should be noted that Fritsch et al. [22] detected such p63-positive basal cells already in weeks 14–15 of gestation. While these progenitor cells become k17 positive at later stages, they are still negative for this marker at week 22. BCL2 was strongly expressed in the mesenchymal stroma of the Müllerian vagina and Müllerian duct epithelium. A gradient in staining intensity of several markers can be recognized at the site where the solid cord and the Müllerian columnar duct epithelium merge. k5 and p63 expression decreases in the cranial direction, while k7 and k8 expression decreases in the caudal direction (see Fig. 1.2, right panels). With increasing gestational age, the length of the solid cord reduces, the vaginal epithelium moves cranially, and, around week 24–25, a clear-cut squamocolumnar junction is detectable by conventional histology and is located in the endocervical canal.

It has been shown that during the first semester of embryonic development, the differentiation processes described above are initialized by cellular stimuli from the urogenital sinus or surrounding mesenchymal tissue. Strong BCL2 positivity is detected in the mesenchymal tissue surrounding the solid cord, while the surrounding tissue of the Müllerian duct is weakly BCL2 positive or even negative. This supports the view that in the human situation, the Müllerian vagina and Müllerian duct epithelium are under different stromal stimuli [22, 23, 32, 33].

In the period between weeks 29–40 of gestation, the differentiation markers clearly demarcate the squamous and glandular epithelia, and the squamocolumnar junction has been firmly established. Also the immunomarker profile for the reserve cells becomes gradually comparable to the newborn and adult expression pattern. In these basally located cells of the developing glandular endocervical epithelium, the expression of p63 and k5 becomes clearly visible, while the beginning of expression of k17 and BCL2 can be seen as weak staining in the fetal reserve cells. As illustrated in Fig. 1.3, immunostaining results in a 32-week-old embryo show that the squamocolumnar junction is clearly demarcated by k7 and k8 positivity in the columnar epithelium and k5 and p63 positivity in the squamous epithelium.

Close to the squamocolumnar junction (see Fig. 1.3b, box 1), reserve cells beneath k7 and k8 positive cells can be recognized. These are typically k5 and p63 positive and share the k7 and k8 positivity of columnar epithelium, although to a lesser extent (see Fig. 1.3b, box 2). More cranially from the squamocolumnar junction, the Müllerian duct epithelium is seen in which k5- and p63-positive and k5- and p63-negative cells are still intermingled (see Fig. 1.3b, box 4). Here, the reserve cells have not yet reached their basal position in the columnar epithelium. Close to the SCJ of this embryo, metaplasia is present (see Fig. 1.3b, box 3) in which reserve cells beneath columnar epithelium are programmed to differentiate into squamous epithelium. This is a process that is typical of the adult situation in which columnar epithelium in the TZ undergoes metaplasia to form squamous epithelium. During fetal development this metaplasia may be induced by maternal estrogens.

In the adult situation reserve cells beneath normal columnar epithelium typically combine markers for squamous (p63, k5, and k17) and glandular (k5 and k8) differentiation (see Fig. 1.4a–c). Furthermore these cells are protected against apoptosis by a high concentration of BCL2 (see Fig. 1.4d). In nearly all routine biopsies that include the squamocolumnar junction, reserve cells can be recognized by immunostaining either as stretches of cells or as single cells [34].

Based on the immunophenotyping studies described above, we propose a model for the hierarchical order of cell lineages developing during embryological and fetal growth, in which Müllerian cell types differentiate into endocervical columnar cells and reserve cells (see Fig. 1.5) [1, 22]. To date there is a consensus that the reserve cell is the progenitor for the squamous cell formed during metaplasia in the endocervix. However, whether the reserve cell in the adult situation is a remnant of the embryological population or is formed de novo from a columnar cell is still debated (see below).

The two types of epithelia that meet at the squamocolumnar junction each have their own progenitor cells in the adult situation. It can be anticipated that these progenitor cell compartments comprise (1) a real stem cell fraction, from which (2) the basal layer of the squamous epithelium and (3) the reserve cells underlying the glandular epithelium arise. In the ectocervical squamous epithelium, a (4) transient-amplifying (parabasal) cell compartment arises upon proliferation of basal cells.

Extensive and strong Ki67 staining is seen in these parabasal cells, while the basal cells show only very low proliferative activity, as indicated by the fact that only sporadic cells are Ki67 positive in this compartment (see Fig. 1.4l). These basal cells are, however, heavily protected against apoptosis by strong expression of BCL2 (see Fig. 1.4k) [31]. On the contrary, the highly proliferative parabasal cells are negative for BCL2. These mutually exclusive expression patterns of Ki67 and BCL2 indicate that the processes of extensive proliferation on the one hand, and protection from apoptosis on the other, are meticulously separated within the squamous epithelium. These immunophenotyping results are supported by earlier findings that the generation cycle of basal cells is about 30 days, while under physiological conditions the generation time for parabasal cells is about 3 days [36]. In both the endocervical columnar cells, as well as reserve cells beneath the columnar epithelium (see Fig. 1.4a–d), very little proliferative activity is observed under normal physiological conditions. This indicates that the turnover of the endocervical epithelium is apparently low compared to the ectocervical epithelium. The reserve cells, which are recognized throughout the entire endocervical canal up to the endocervix-isthmus junction, show strong positivity for BCL2, while the columnar epithelial cells are BCL2 negative (see Fig. 1.4d) [34, 37, 38]. These observations suggest that the reserve cells are part of a pool of progenitor cells protected against apoptotic cell death and required to ensure survival of the endocervical epithelium. Both columnar and reserve cells, however, can proliferate independently as has been shown in regenerating endocervical epithelium, although the reserve cells are not required a priori for normal regeneration of secretory columnar epithelium [39]. Whether the reserve cells should be regarded as the default progenitor cell for the columnar epithelium, or vice versa, the columnar epithelium as progenitor for the reserve cells can be questioned. In the case of microglandular hyperplasia, it has been proposed that reserve cells are created in adulthood during specialized columnar proliferations [35].

The combined phenotype of the reserve cells expressing both markers of glandular (k7, k8) and squamous differentiation (p63, k5, and k17) is a strong indicator for the consensus that reserve cells are also the progenitors for (mature and immature) squamous metaplasia [14, 28]. As can be seen in Fig. 1.4f–m, remnants of squamous metaplasia express k7, k8, and k17, which are not found in the normal ectocervical squamous epithelium.

Recently, it was reported that an immature cuboidal progenitor endocervical cell type can be recognized by high levels of k7 expression, situated close to the squamocolumnar junction. After infection by an oncogenic HPV type, this cell type may develop into a high-grade squamous intraepithelial lesion [4, 40] (see also section “Squamous Metaplasia”).

Currently no specific markers are available for the real stem cells in the columnar and squamous epithelium. The fact, however, that both the reserve cell and the basal cell of the squamous epithelium express BCL2 and p63 could be an indicator that the real stem cell of the uterine cervix could also express these two markers. The cells originating in the fetus around week 20, which have begun to express p63 and BCL2, could act as the first stem cells of the cervix.

The area between the original and the new squamocolumnar junction is defined as the TZ. This TZ can be visualized by colposcopic inspection and is also the area in which approximately 90% of (pre)neoplastic lesions develop [11–13, 41]. The TZ is a dynamic entity formed during puberty and is the area where the glandular epithelium is replaced by metaplastic squamous epithelium [42]. Furthermore the presence of endocervical glands underlying the squamous epithelium is an indicator of the position of the TZ [43]. The last gland can serve as a landmark for the position of the original squamocolumnar junction [13]. The junction between the metaplastic squamous epithelium and the glandular cells defines the new squamocolumnar junction and the cranial limit of the transformation zone. At birth and during the premenarchal years, the squamocolumnar junction resides at or very close to the external os [5]. During puberty the endocervical mucosa everts onto the ectocervix as a result of hormonal stimulation and swelling of the stroma of the cervix. Reproductive hormones also influence the production of ectopy during late fetal life and pregnancy and as a result of the use of oral contraceptives. Ectopy is modified over time by squamous metaplasia and epithelialization, low pH, trauma, and possibly cervical infection [12, 42, 44]. As a result of this eversion, the squamocolumnar junction becomes located on the ectocervix, and the exposed endocervical mucosa (ectropion) shows the gradual replacement of the columnar epithelium by squamous epithelium. This represents a protective response to the exposure of the glandular epithelium to the vaginal environment [12, 42, 45, 46]. Reserve cells beneath the columnar epithelium are the progenitor cells for the newly formed squamous epithelium. A consequence of this development is the formation of a new squamocolumnar junction (see Fig. 1.1e). During late reproductive life and after the menopause, decreasing hormone levels lead to shrinkage of the cervix, and the new squamocolumnar junction comes to lie in the endocervical canal [12].

Three different types of squamous metaplastic lesions can be recognized in the cervix, i.e., (1) immature squamous metaplasia, (2) mature squamous metaplasia, and (3) atypical immature metaplasia [13, 42, 45, 47–50]. There are two histogenetic mechanisms by which the endocervical mucosa is replaced by squamous epithelium [5]. The first is the direct ingrowth of squamous epithelium in the direction of the endocervix, which is referred to as squamous epithelialization. The other route is through proliferation of subcolumnar reserve cells (reserve cell hyperplasia) and their subsequent maturation into squamous epithelium. Both mechanisms result in a squamous epithelium overlying endocervical mucus-producing glands [44]. In the first phase, reserve cells proliferate and stratify. Subsequently, these cells undergo a squamous differentiation process that is at first incomplete, with persistence of the columnar epithelium, which is often seen as a residual layer on the surface [45]. Later, metaplastic cells can mature to keratinocytes that are indistinguishable from the suprabasal cells of the pluristratified epithelium, resulting in mature squamous metaplasia.

Typical immunohistochemical staining patterns for p63, k5, k7, k17, and BCL2 in immature squamous metaplasia are shown in Fig. 1.6, while Fig. 1.7 shows a schematic overview of the changes that this marker profile undergoes during the formation of the metaplastic lesions. The biomarker expression pattern of immature metaplasia strongly resembles that of the reserve cell, with the exception of BCL2, which is significantly reduced in intensity as compared to that of the reserve cell. Although the precise mechanisms underlying the induction of squamous metaplasia are still obscure, it seems that cytokines and growth factors present in the metaplastic microenvironment might alter the transcription factor profile of reserve cells. It has been proposed that metaplastic transformation results from the release of cytokines and other soluble factors by both epithelial and inflammatory cells [45].

Atypical immature metaplasia is a poorly reproducible diagnosis, which spans a morphological spectrum and has features both of metaplasia and atypia. It is difficult to distinguish from low-grade cervical neoplasia. These lesions may arise as a result of reactive or inflammatory processes or through high-risk HPV infections of true precursors, resulting in cervical carcinoma [46, 49, 51–54].

In general, the intact cervical epithelium is resistant to viral infections. However, chemical or mechanical disruption of the integrity of this epithelium enables HPV entry. As has been analyzed in experimental animal systems, a simple brush with a Cytobrush cell collector or the application of a spermicide (nonoxynol-9) resulted in abrasions that enabled binding of the virus to the basement membrane prior to its transfer to the basal cells [55]. Adsorption to the basal surface of the epithelial cells and reestablishment of contact with the basement membrane during repair of the damaged epithelium might further promote preferential infection of basal cells. It has been found that heparan sulfate in the extracellular matrix and on the surface of these cells acts as an initial attachment receptor for HPV. It was furthermore postulated that the adult reserve cells with CD49f (α 6-integrin) expression could be preferentially targeted by high-risk HPV types during cervical carcinogenesis (see also Chap. 2 for the biology of HPV infection) [9, 55, 56].