Prof. A. Bukovsky, MD, PhD, DSc, Institute of Biotechnology, Academy of Sciences of the Czech Republic, Prague, Czech Republic. E-mail: firstname.lastname@example.org
Morphostasis (tissue homeostasis) is a complex process consisting of three circumstances: (1) tissue renewal from stem cells, (2) preservation of tissue cells in a proper differentiated state and (3) maintenance of tissue quantity. This can be executed by a tissue control system (TCS) consisting of vascular pericytes, immune system–related components – monocyte-derived cells (MDC), T cells and immunoglobulins and autonomic innervation. Morphostasis is established epigenetically, during the critical developmental period corresponding to the morphogenetic immune adaptation. Subsequently, the tissues are maintained in a state of differentiation reached during the adaptation by a ‘stop effect’ of MDC influencing markers of differentiating tissue cells and presenting self-antigens to T cells. Retardation or acceleration of certain tissue differentiation during adaptation results in its persistent functional immaturity or premature ageing. The tissues being absent during adaptation, like ovarian corpus luteum, are handled as a ‘graft.’ Morphostasis is altered with age advancement, because of the degenerative changes of the immune system. That is why the ageing of individuals and increased incidence of neoplasia and degenerative diseases occur. Hybridization of tumour stem cells with normal tissue cells causes an augmentation of neoplasia by host pericytes and MDC stimulating a ‘regeneration’ of depleted functional cells. Degenerative diseases are associated with apoptosis. If we are able to change morphostasis in particular tissue, we may disrupt apoptotic process of the cell. An ability to manage the ‘stop effect’ of MDC may provide treatment for early post-natal tissue disorders, improve regenerative medicine and delay physical, mental and hormonal ageing.
Morphostasis (tissue homeostasis) refers to the maintenance of a proper differentiated state in distinct tissues after morphogenesis. Many studies indicated an involvement of the immune system–related cells in wound healing and tissue regeneration, and we attempted to include a role of vascular pericytes, autonomic innervation (AI) and immune system memory in a complex Tissue Control System (TCS). The TCS appears to regulate proper differentiation of tissue-specific cells, enable a maintenance of tissue quantity and play a role in distinct longevity and ageing of particular tissues (reviewed in ). The tissues vary in the extent of their differentiation. Normal kidney tubular epithelium consists of a single layer of epithelial cells and contains no intraepithelial monocyte-derived cells (MDC) or T cells . Small intestine epithelium contains T cells but not intraepithelial monocyte-derived dendritic cells (DC) . Stratified squamous epithelium of the uterine ectocervix contains intraepithelial T cells and DC . Corpus luteum (CL) of the ovary differentiates with the assistance of vascular pericytes and exhibits ageing and regression after homing of T cells and MDC among luteal cells .
The immune system plays an important role in morphostasis by its action in the surveillance against foreign (non-self) antigens, but its tolerance to self-antigens within the adaptive immune system is usually viewed as an anergy because of the deletion of autoreactive T and B lymphocytes [6, 7], i.e. no action. New approaches to the fundamental regulations within the immune system, and emerging role of the immune system in tissue physiology, may influence significantly future development in basic and applied immunology.
In 1922, Alexis Carrel has shown that leucocyte extracts, like embryonic tissue extracts, possess the power of increasing rate of multiplication of fibroblasts in vitro . R. Geoffrey Burwell in 1963  suggested that immune system may play an important role in morphostasis and in 1980, Isaiah J. Fidler suggested that participation in host immune responses is one of the many functions of lymphocytes and that lymphocytes may function as trephocytes and regulate the growth of various organs .
Trephocytes were determined as leucocytes-stimulating nutrition and growth of tissue-specific cells, i.e. representing a food for these cells. Mitotic divisions and degeneration of mouse lymphocytes within the cells of intestinal epithelium were described . The engulfment and digestion of living trephocytes were observed, and additional cases of cellular cannibalism among invertebrates and vertebrates were brought forward. The conclusion was reached that cellular cannibalism is a normal and apparently fairly common form of nutrition of the animal cell . Other forms of non-apoptotic cell death exist that occur by cell-in-cell invasion, including a homogeneous entosis in epithelial cells or internalization of NK cells by tumour cells (reviewed in ). In contrast to cellular cannibalism, entosis, or cellular internalization, the naturally occurring active process of cell suicide known as programmed cell death (or apoptosis) plays a crucial role in animal development and homeostasis. We observed apoptosis of intraepithelial T cells and monocyte-derived DC accompanied by p53 and ras expression . This process may be required not to simply feed epithelial cells but rather to stimulate their advanced differentiation by growth factors and cytokines released from T cells and DC undergoing apoptotic suicide.
Investigation of the immune system after partial organ resection has shown that lymphatic system is specifically activated. Lymphocytes ‘primed’ by a partial hepatectomy or unilateral nephrectomy initiated and maintained the growth of liver or kidney until the morphostasis by means of the original size of the organs was restored. Transfer experiments have shown that these lymphocytes stimulate the growth of liver and kidney in otherwise intact recipients as well .
Therefore, there is a mechanism which initiates and maintains a ‘priming’ of lymphocytes after partial hepatectomy resulting in the stimulation of liver regeneration, and this mechanism is silenced when the original size of the liver is reached. It appears that the size of each tissue is regulated by autonomic innervation, because elimination of limited areas of the cephalic neural crest in stage 9 or 10 chick embryos markedly reduced the size of the thymus gland [15, 16].
In a set of articles between 1995 and 2007, Jamie Cunliffe suggested that the core function of the immune (or morphostatic) system is to dispose of tissue debris (mess/non-mess discrimination) and restore order. Because the debris of degenerating cells provides a rich source of nutrients for micro-organisms, it has been argued that the major goal of the immune system could be to rapidly dispose of such debris. This strategy could lead to a ‘denial-of-nutrient-substrate’ that starves invading bacteria out of contention and, in consequence, suppresses infection [17–20].
More recently, Sprent and Cho  indicated that, for T cells, a limited degree of autoreactivity is beneficial for the immune system. In our opinion, such limited degree of interaction of the immune system with self is represented by a support of differentiation of unaltered self-tissue cells . Therefore, it is assumed that the immune system has a dual function: (1) eliminates non-self-cells and substances, including infections, by immune surveillance and (2) actively promotes tissue-specific differentiation of self-cells by an ‘immune maintenance of self,’ i.e. morphostasis.
Monocyte-derived cells are virtually present in all tissues, and T cells are physiologically present in several types of normal epithelial tissues, including the intestinal tract, respiratory tract, genitourinary tract epithelium and the skin . The T cells in epithelial tissues are T cell antigen receptor (TCR) αβ of thymic origin and γδ mostly of extrathymic origin . The question arises, which of these two distinct T cell populations play a particular role in immune surveillance and morphostasis. Microbial deprivation in germ-free mice preferentially influences the intestinal intraepithelial αβ T cell population to decrease and become non-cytolytic but has a little effect on the pool size or characteristics of intestinal intraepithelial γδ T cells .
As indicated later, however, the problems to be solved still are what the differences between self and non-self cells and substances actually are. Morphostatic rules are expected to be determined during the immune adaptation (IA) also known as a ‘critical developmental period’ [25, 26], and their duration for certain tissue during subsequent life depends on a particular tissue appearance in early ontogeny . In other words, the earlier certain tissue evolves, the longer its function is expected to be supported . In addition, self-tissues absent during IA, such as the ovarian CL, are during adulthood handled as a graft .
Available data indicate that the morphostasis is a complex event. It can be executed by a diversified tissue-specific involvement of vascular pericytes, immune system–related components (MDC, T cells and immunoglobulins) and AI and is determined during the early development of each individual.
Morphostasis and the tissue control system theory
Our studies in the late 1970s [29–32] and early 1980s [33, 34] resulted in the concept of a wider role of the immune system–related cells and vascular pericytes, the so called TCS, in the regulation of differentiation of tissue-specific cells . The TCS theory was refined when the role of AI in the regulation of quantitative aspects in tissues was added . The theory was advanced further by studies of TCS in morphostasis of the ovary [5, 37–44], placenta , stratified squamous epithelium [4, 46] and cancer [47–49]. In addition, we investigated a role of the IA in tissue longevity [25, 27, 28, 50] and a role of MDC and T cells in the induction of asymmetric division of somatic stem cells [51–53]. Complex aspects of immune physiology in tissue regeneration and ageing, tumour growth and regenerative medicine were recently analysed [1, 54–57].
Basic ‘tissue control unit’
The TCS consists of vascular pericytes, immune system–related cells (MDC, T and B lymphocytes) and AI. Basic tissue control unit (TCU) is associated with post-capillary venules (marked PCV in Fig. 1A) showing association of pericytes (P), endothelial cells (En) and primitive MDC (pM; CD14+) . In post-capillary venules, the intercellular spaces of pericytes are wide enough for leucocytes to pass through, and the occasional extravasation of leucocytes through venule walls can be seen under electron microscopy . Pericytes provide a cellular basis for the role of neural crest in thymus development and possibly in its maintenance . Distributed by the microvasculature throughout the organism, the pericytes represent pluripotent adult stem cells and provide an ingenious system to assure the maintenance, physiological repair and regeneration of organs . Pericytes show similar findings on gene expression pattern with neuronal cells and pericytes, and some neuronal cells originate from neural crest . In vitro, after a combination of sex steroids, the pericytes differentiate into neural and neuronal cells , which supports their proposed stem cell neuroendocrine properties . They also share Thy-1 differentiation antigen with brain and AI, and AI regulates number of pericytes and their function (black arrowheads, Fig. 1A) [63, 64].
The AI accompanies vessels in normal tissues but is absent in neovascularization of the ovarian CL  and invasive cancer . Lack of vascular innervation in these tissues is accompanied by a high release of Thy-1 differentiation protein, a morphoregulatory molecule associated with cell differentiation [37, 67], from vascular pericytes . On the other hand, in innervated structures, such as ovarian follicles, the Thy-1 release is enabled only in growing structures, but inhibited in resting ones. These observations suggest that Thy-1 release is inversely correlated to AI inhibitory presence (AI−/AI+ in Fig. 1A). By other words, resting ovarian follicles are ready to grow when the AI stops to inhibit an activity of their vascular pericytes .
The activity of TCU is enabled by a lack of AI inhibition and initiated by a request (blue arched arrow, Fig. 1A) of a tissue-specific cells, e.g. epithelial (Ep), towards perivascular CD14+ primitive MDC (pM) to stimulate tissue stem cell division. Prior stimulation of epithelial cells, primitive MDC may interact with pericytes (yellow up-down arrow) to seek whether the activity of the unit is not inhibited by AI, because the activity of MDC is regulated by tissue fibroblasts [68, 69].
During regeneration, the tissue MDC produce cytokines and chemotactic factors attracting fibroblasts and endothelial cells and often activate them to produce additional mediators-stimulating angiogenesis and maturation of tissue-specific cells [70–77]. Tissue MDC secrete numerous growth factors and cytokines-stimulating tissue regeneration, reconstruction and wound healing [71, 75, 78]. These include members of the tumour necrosis factor (TNF) super family such as TNFa, TNF-like weak inducer of apoptosis (TWEAK), lymphotoxin beta, members of the interleukin (IL)-6-like family including IL-6, oncostatin M and interferon gamma (IFNc). Invading MDC play a major role in the liver progenitor cell response to chronic liver injury, and knockout studies suggest that a combination of cytokines is important .
It has also been shown that depletion of MDC in one testis of adult rats selectively abolished differentiation of Leydig cells from mesenchymal precursors , such as vascular pericytes . Recent findings indicate that CD14 MDC in circulation are involved in a variety of physiological functions other than innate and acquired immune responses, such as repair and regeneration of tissues .
Figure 1B shows CD14+ MDC (brown colour) in lamina propria (lp), entering (arrows) through the basement membrane (white dotted line) the basal epithelial layer (b; epithelial stem cells = stage 1 of differentiation) of the stratified squamous epithelium. Association of MDC with stem cells (yellow arrowhead) causes their division (red arrowheads). One of the two post-mitotic cell daughters passes basal/parabasal interface (black dotted line) and enters (upper red arrowhead and white arrowheads) parabasal epithelial layer (pb), representing an early epithelial differentiation = stage 2. Inset shows Ki67+ post-mitotic cells (arrowheads) in the lower parabasal layer; arrows indicate numerous Ki67+ cells in the adjacent lamina propria. Accumulation of CD14+ MDC among basal epithelial cells was also described in the human nasal mucosa .
Thy-1 vascular pericytes
Another mesenchymal cell type involved in tissue regeneration are tissue fibroblasts and derived vascular pericytes. They secrete a hepatocyte growth factor-scatter factor, a plasminogen-like protein and potent mitogen, motogen and morphogen for hepatocytes and other cells, that is also thought to play a role in tissue regeneration [83–88]. Pericytes also secrete TGF-beta1, PDGF and IL-10, and support mitogenic activity of progenitor cells by EGF and TGFα .
Early differentiation of tissue cells is initiated by interaction of MDC with pericytes in TCU (green arched arrow, Fig. 1A). The pericytes release intercellular Thy-1+ vesicles (ICV) migrating by a chemotaxis through the basement membrane and collapsing into intercellular spikes after reaching the target cells [4, 35]. Similar event is executed towards differentiation of endothelial cells .
Recognition at the cell surface
Most of the molecules involved in the TCS pathway belong to the immunoglobulin (Ig) superfamily of molecules. It has been suggested that the involvement of Ig-related molecules in tissue interactions is more primitive than their involvement in the immune system and the immune functions evolved from the sets of molecules mediating tissue interactions . One of them, the Thy-1 differentiation protein, consists of a single Ig domain and represents the most primitive and ancestral member of the Ig-superfamily. The Ig-related molecules have a diversity of functions, but in most cases the common denominator is recognition at the cell surface . Also, the only function of Thy-1 differentiation protein and other Ig-related molecules is to mediate recognition, with the consequences of recognition being because of the differentiated state of the cells. It requires that the correct ligand and receptor are expressed on the appropriate cells at the right time .
Staining for Thy-1 differentiation protein (Fig. 1C) shows pericytes (p) associated with microvasculature adjacent to the basement membrane (white dotted line). The pericytes secrete intercellular vesicles (ICV, white arrowheads), which migrate through the basement membrane among basal epithelial cells (Thy-1 ICV route, yellow arrowheads) above (red arrowheads) the b/pb interface (black dotted lines). After reaching the post-mitotic parabasal cells, the Thy-1+ ICV collapse into empty intercellular spikes (ICS, black arrowhead; see also Fig. 1A). Hence, targets for Thy-1 vesicles are parabasal cells adjacent to the b/pb interface, i.e. epithelial cells expressing Ki67 and entering differentiation (yellow arrow, stage 2 to stage 3 transition).
The intercellular Thy-1 vesicles have been shown by immunoelectron microscopy to exhibit Thy-1 surface expression and to contain a substance lacking Thy-1 staining  – shown in inset, Fig. 1C, where the arrowhead shows Thy-1+ ICV surface and arrow indicates ICV opening and release of the content. The Thy-1− content of ICV may consist of growth factors and cytokines released by pericytes [62, 83–88] and stimulating differentiation of particular epithelial cells.
Targeted delivery of growth factors
The Thy-1 ICV may represent an elaborated paracrine mechanism via Thy-1 recognition at cell surface, the so called ‘targeted delivery’ of differentiation promoting substances , by which certain growth factors (vesicle content) are delivered to certain type-/stage-specific target cells expressing receptor for Thy-1 ligand. However, the receptor for Thy-1 has not yet been identified. One possibility is that the Ki67+ cells entering differentiation are the targets for Thy-1+ ICV. Also, there is a lack of expression of major histocompatibility complex (MHC) class I molecules in epithelial stem cells adjacent to the basement membrane, but strong staining of parabasal cells . Hence, MHC class I molecules could be involved in the recognition of the Thy-1 ligand, because they have the MHC I-associated peptide, which contains a tissue-specific signature . From this point of view, such tissue-specific peptides could be released to attract by chemotaxis the migration of Thy-1+ vesicles to target tissue cells. However, this implies that there should also be tissue-specific moieties in Thy-1 molecules.
In ovarian follicles, the Thy-1 vesicles secreted by pericytes associated with follicular basement membrane specifically migrate and release their content among granulosa cells and not among the adjacent theca cells , while Thy-1 vesicles secreted by pericytes of thecal vessels migrate and release their content among the thecal cells [38, 49]. Accordingly, in stratified squamous epithelium, the Thy-1 vesicles from pericytes in the lamina propria migrate through basal layer of epithelial cells and release their content after reaching relatively distant parabasal layer (Fig. 1C).
Targeted delivery of some growth factors or cytokines stimulating many types of tissues to particular tissue cells by intercellular Thy-1 vesicles could be enabled by the tissue specificity of Thy-1 glycoprotein carbohydrate moieties  possibly recognizing tissue-specific peptides. It has been found that each tissue creates unique sets of glycoforms, i.e. the same peptides but carrying oligosacharides that differ either in structure or in site location or both . The ‘targeted delivery’ of differentiation promoting substances is much more effective in delivering cytokines to particular targets when compared with generally functioning and elementary endocrine or paracrine signalling.
Homing of intraepithelial DC precursors and T cells
Interaction of CD14+ MDC with endothelial cells in TCU (red arched arrow, Fig. 1A) may stimulate endothelial signalling for homing of T lymphocytes (T, Fig. 1A) and monocyte-derived dendritic cell precursors (DCP, also known as veiled cells), which may differentiate into mature dendritic cells (DC, Fig. 1A) in the intermediate layer of stratified squamous epithelium (see complete TCS pathway below). Peripheral T and B lymphocytes, i.e. lymphocytes found outside the ‘central’ lymphoid organs (thymus and bone marrow), are not only in the classical lymphoid tissues (spleen and lymph nodes), but are also associated with epithelia and subepithelial tissues of the skin, intestinal tract, respiratory tract and genitourinary tract. Although interstitium of these tissues exhibits many lymphocytes, few data on such cells are available. In the liver, interstitial T lymphocytes and MDC are activated during liver regeneration, and they may be responsible for regulation of normal cell regeneration [79, 94]. On the other hand, intraepithelial lymphocytes (IEL), which are predominantly T lymphocytes (CD3+) with CD8+ cells predominating most, were much more intensively studied.
Both T cell TCR αβ and TCR γδ T lymphocytes were found within intraepithelial T lymphocytes populations, the former representing mainly T lymphocytes differentiated within the thymus and the latter T lymphocytes mainly of extrathymic origin . Regarding the thymus, recent observations indicate that very immature CD4(−)CD8(−)CD3(−)CD44(+)CD25(int) thymocytes, which have not yet rearranged their TCR, leave the thymus, migrate to the blood, colonize the gut, reconstitute CD8 αα thymic intraepithelial T lymphocytes, and this cell set is fully responsible for the generation of the CD8 αα pool. They complete TCR rearrangements, and the TCR-αβ/γδ lineage commitment must occur in the gut .
Nevertheless, γδ intraepithelial T lymphocytes were found to exist in athymic mice, and in mice with severe combined immunodeficiency. They develop in irradiated thymectomized mice reconstituted with T lymphocyte-depleted bone marrow, and the selection mechanism, such as clonal deletion, operative in thymus-derived T lymphocytes do not apply to the γδ intraepithelial T lymphocytes [22, 23, 96, 97]. Intestinal γδ intraepithelial T cells develop in mice lacking thymus, all lymph nodes, Peyer’s patches and isolated lymphoid follicles but not gut cryptopatches, which are pivotal birthplace of mature T cells such as the thymus-independent intestinal intraepithelial T cells . The cryptopatches play a role in early extrathymic maturation of intestinal intraepithelial T cells, and their absence is accompanied by a lack of γδ intraepithelial T lymphocytes .
During evolution and ontogeny, the gut-associated lymphoid tissue develops before the development of the thymus, and it may be expected that the gut-associated lymphoid tissue represents more ‘primitive’ lymphoid compartment. The finding that γδ intraepithelial T lymphocytes have a restricted repertoire is in line with the idea of a ‘primitive’ function. The tissue-committed intraepithelial T lymphocytes recognize tissue-specific antigens, and they undergo differentiation under the influence of intraepithelial MDC. They are also influenced by epithelial cells, intercellular matrix and soluble mediators – cytokines and immunoglobulins . Many lymphokines produced by intraepithelial T lymphocytes were detected that directly influence epithelial function (IFN-γ, TNF-α and TGF-β) and intraepithelial T lymphocytes exhibit alloreactivity and virus-specific cytotoxicity, provide B cell help and regulate epithelial cell differentiation and function [22, 100]. For instance, TGF-β1 may function in coordination of the rapid cell turnover typical for the intestinal epithelium , and TGF-α and TGF-β may play a significant role in the regulation of the balance between proliferative and differentiated cell compartments in the intestinal epithelium through both autocrine and paracrine mechanisms [102, 103] and intestinal wound healing and epithelium repair .
Altogether, the interstitial and intraepithelial T lymphocytes may play a role in the immune surveillance, and influence renewal, regeneration, and function of parenchymal and epithelial cells. They are represented by two distinct T lymphocyte populations: (1) thymic T cells governing the immunity and (2) more primitive extrathymic T lymphocytes appearing earlier in phylogeny and ontogeny and involved in regulation of renewal and function of epithelial cells.
Intraepithelial MDC and T cells
The HLA-DR+ (DR+) DCP and CD8+ T cells interact themselves and accompany differentiation of epithelial cells. Epithelial lamina propria (lp, Fig. 2A) contains clusters (dotted/dashed lines) of numerous DR+ DCP lacking CD1a of mature DC (see unstained cells in lamina propria, Fig. 3C) and CD8+ T cells interacting themselves (orange and blue arrows). These clusters resemble cryptopatches in the lamina propria of the small intestine, which are small clusters containing two main cell types, dendritic type cells and early extrathymic T cells .
As indicated earlier, cryptopatches are pivotal birthplace of mature T cells such as the thymus-independent intestinal intraepithelial T cells . The cryptopatches play a role in early extrathymic maturation of intestinal intraepithelial T cells and their absence is accompanied by a lack of γδ intraepithelial T lymphocytes . Because selection mechanism operative in thymus-derived T lymphocytes do not apply to the γδ intraepithelial T lymphocytes [22, 23, 96, 97], the immature T cells in lamina propria can be stimulated by DCP to mature into T cells which are able to recognize and home into the adjacent epithelial layer.
Altogether, we hypothesize that epithelial lamina propria contains clusters of DCP and T cells, resembling intestinal cryptopatches. These lamina propria resident clusters may serve for self-renewal of epithelium-committed DCP and γδ T cells. Arrows in inset Fig. 1B show numerous Ki67+ cells in epithelium lamina propria.
Some DR+ DCP enter the epithelium (red arrowhead, Fig. 2A) through the basement membrane (dotted line). CD8+ T cells entering the epithelium (yellow arrowheads) significantly accumulate (white open arrows) in the basal epithelial layer (b). These T cells may contribute to the asymmetric division of epithelial stem cells, i.e. one of the resulting cell daughters remains a stem cell, and the other contributes to the tissue-specific differentiating cells, as described for ovarian stem cells (OSC) .
After interaction with intraepithelial DCP (solid white arrows, Fig. 2A), isolated T cells migrate through the parabasal layer (green arrows) towards the parabasal (pb)/intermediate (im) interface (dashed line). The DCP in the mid-parabasal layer (mpb; stage 3 of epithelial cell differentiation) secrete large amounts of DR+ substances (red arrow; see also white arrows in Fig. 3A) among epithelial cells. At the pb/im interface, the T cells interact with DCP again (solid white arrowheads, Fig. 2A, B) and show an activation characterized by HLA-DR expression (solid yellow arrows, Fig. 2A, B) , which can be mediated by IL-2 T cell and MDC activation factor [107, 108] derived from epithelial cells . The CD8 T cells entering intermediate layer show apoptotic fragmentation (open yellow arrows, Fig. 2A, B; see also arrowheads, Figs. 3B and 4A, B). This accompanies differentiation of DC from DCP (yellow arrowhead, Fig. 2B) and indicates that these intraepithelial T cells are of extrathymic origin, because such T cells show undue susceptibility to apoptosis .
Red arrowheads in Fig. 3A show DR+ DCP entering epithelium from the lamina propria. DCP in the mid-parabasal layer show DR release (white arrows). This site-specific DR secretion may be accompanied by cytokine and growth factor release [71, 75, 78, 79] promoting differentiation of particular epithelial cells. The DR release is particularly evident when the intraepithelial DCP are compared with DCP in the lamina propria (orange arrowheads). This suggests that intraepithelial DCP may contribute specifically to the stage 3 differentiation of epithelial cells (circled 3) and make them ready to proceed to stage 4, which may be IgM dependent. Blue arrowheads show DR+ DCP entering the intermediate layer. Note that some DC show a diminution of DR expression (white arrowheads).
Figure 3B shows clusters of CD8+ T cells in lamina propria (red doted lines), accumulation of T cells in the basal epithelial layer (black arrows) and migration of individual T cells (white arrow) towards the pb/im interface (dashed line). The T cells entering intermediate layer undergo apoptotic fragmentation (arrowheads). This may stimulate differentiation of DC (see Fig. 2B), contribute to stage 5 differentiation of epithelial cells (circled 5) and make the epithelial cells ready to proceed to stage 6.
CD1a marker of mature DC is not detectable in the lamina propria (lp, Fig. 3C), but is expressed by DCP throughout the parabasal layer (arrows), by mature DC in the lower and mid-intermediate layer (black arrowheads) and by fragments of DC in the upper intermediate layer (white arrowheads). The mature DC may contribute to stage 6 differentiation of epithelial cells (circled 6), because they show a release of CD68 (white arrowheads, Fig. 3D) prior fragmentation (yellow arrowheads, Fig. 3D).
The CD68 MDC are present under the epithelium basement membrane in the lamina propria (orange arrowheads, Fig. 3D). Intraepithelial DCP do not express CD68 of differentiating MDC until they reach stage 4 epithelial cells under the pb/im interface (arrows, Fig. 3D), where they interact with and activate the CD8 T cells (Fig. 2A, B).
Altogether, double colour immunohistochemisry indicates that epithelium lamina propria contains resident clusters of DCP and T cells resembling intestinal cryptopatches. They may serve for self-renewal of epithelium-committed DCP and γδ T cells. In the stratified squamous epithelium, the intraepithelial mesenchymal cells show distinct behaviour in distinct layers and sublayers of epithelial cells, possibly because of their mutual interactions and response to distinct substances secreted by particular epithelial cells. Consequently, intraepithelial mesenchymal cells appear to stimulate differentiation of certain epithelial cells as evidenced from secretion of their markers. Finally, the intraepithelial T cells and DC degenerate, while the epithelial cells continue to differentiate.
Autoreactive IgMs and IgG
Autoantibodies, or natural antibodies, are present in the serum of healthy individuals and animals, and it has been proposed that the antigens recognized by natural antibodies and by conserved T cell reactivities are essential for the selection of natural B- and T cell repertoires and for the maintenance of tolerance to self (reviewed in ). The natural antibodies are almost exclusively immunoglobulins (Ig) M, but IgG can also be detected . The IgM antibodies appear earlier in phylogeny and ontogeny than IgG. Shark immunoglobulins resemble mammalian IgM in structure and gene similarity .
IgM may comprise as much as 50% of serum proteins in the shark. By contrast, in humans, the IgM comprises <5%. Human IgM natural antibodies show little dependence on thymic function for expression and tend to increase with age. Sharks and humans possess IgM antibodies that react with thyroglobulin and single-stranded DNA. Affinity-purified natural shark antibodies to thyroglobulin or single-stranded DNA constitute small fractions of total IgM. They illustrate extensive cross-reactivity comparable to that shown by polyspecific IgM natural antibodies produced by human B cells (CD5+) that appear early in ontogeny [112, 113].
In human sera, reactivity of natural IgM and IgG to epidermal keratins varies among normal healthy individuals, with most often binding to suprabasal layers and stratum corneum . Naturally occurring antibodies and autoantibodies mediate tissue injury only after an organ has been subjected to a stressor such as ischaemia . Natural IgM antibodies appear in the absence of apparent antigenic stimulation, are secreted by the long-lived self-renewing B1 subset of B cells and a large proportion of the natural antibodies is polyreactive to phylogenetically conserved structures, such as nucleic acids, heat shock proteins, carbohydrates and phospholipids .
When available, the autoantibodies may contribute to the regulation of epithelial cell differentiation by binding to certain epithelial cells. The IgMs are present throughout the stratified squamous epithelium (Fig. 3E versus 3F) and show increased affinity (binding) to certain epithelial sublayers (arrowheads, Fig. 3E). IgMs may regulate early (IgM1, stage 4, Fig. 3E), mid (IgM2, stage 7) and late differentiation (apoptosis) of epithelial cells (IgM3, stage 9).
Apoptosis in stratified squamous epithelium
Under normal conditions, apoptosis is a programmed cell death, and it occurs as a physiological process that helps to remove excess or unwanted cells or tissues such as ovarian CL of the menstrual cycle, unless pregnancy occurs. An alteration of the ovarian CL apoptosis in some climacteric women results in an alteration of the cyclic ovarian function . Apoptosis is a normal condition occurring at the surface of stratified squamous epithelium and results in the clearance of used cells. Apoptosis is caused by an activation of the caspase family proteases. One of the activation mechanisms are autoantibodies penetrating into the cells. The antibody catalyses hydrolysis of single- and double-stranded DNA, and cell death induced by the antibody is inhibited by the pan-caspase inhibitor [117, 118].
Figure 3F shows that autoreactive IgG binds to the surface of epithelial cells in the lower superficial layer (white arrowhead). This may cause a preapoptotic ageing of surface epithelial cells (stage 8 of differentiation). Apoptosis at the epithelium surface is associated with intracellular penetration of IgM and IgG (black arrowheads, Fig. 3E, F).
Critical role during apoptosis is played by a p53 and ras protein expression, which activates the RAF/MEK/ERK apoptotic pathway . Arrows in Fig. 4A and detail panel B show normal CD3+ T cells under the pb/im interface (dashed line), and arrowheads indicate fragmentation of T cells above the interface. No T cells are present at the epithelium surface (panel C). Arrows in panel D show normal DC, and arrowheads and detail panel E indicate DC fragmentation in the upper intermediate layer. No DR+ cells are present at the epithelium surface (panel F). Note that when compared to Figs. 2A and 3A, the Fig. 4D shows no DR release from DCP in the mid-parabasal layer. This suggests that for a given time, there is no need for a renewal of the mid-parabasal epithelial cells in this case.
Staining for p53 shows immunopositive cell fragments above the pb/im interface (blue arrowheads, Fig. 4G) corresponding to the site-specific apoptosis of T cells (see panels A and B) and in the upper intermediate layer (black arrowheads – see also detail in panel H) corresponding to the site-specific apoptosis of DC (see panels D and E). Surface epithelial cells showed p53 expression in the fragmented nuclei (arrows, panel I). Ras protein was also detected in the cell fragments above the pb/im interface (blue arrowheads, panel J) and in the upper intermediate layer (black arrowheads – see detail in panel K). Surface epithelial cells also showed ras in the fragmented nuclei (arrows, panel L). These observations indicate that the fragmentation and apoptosis of intraepithelial T cells and DC and surface epithelial cells in squamous epithelium are accompanied by p53 and ras expression.
Complete TCS pathway reflects immune system phylogeny and ontogeny
Stratified squamous epithelium of uterine ectocervix exhibits differentiation from basal stem cells (SC, Fig. 5A) into the young (parabasal), mature (intermediate) and aged (superficial) cells, divided by three interfaces (b/pb, pb/im and im/s). The stem cells represent a stage 1 (s1) of epithelial cells. Based on the TCS interactions, the layer of young cells can be divided into the lower (s2), mid (s3) and upper (s4) layers and the layer of mature cells can be divided into the lower (s5), mid (s6) and upper (s7) layers. The aged cells consist of lower (s8) and upper (s9) cell layers (Fig. 5A).
Our observations (see Figs. 1-4) suggest that transitions to the higher differentiated state (red arrows, Fig. 5B) are dependent on distinct TCS elements recognizing particular (stage-specific) markers on epithelial cells (indicated by triangles with distinct patterns). Stem cell division is induced by primitive CD14+ macrophages (PM, Fig. 5B), and asymmetric division (s1–s2) may require an assistance of T cells (T), s2-s3 transition is enabled by vascular pericytes secreting intercellular Thy-1 vesicles, s3–s4 by DCP secreting HLA-DR, s4–s5 by IgM binding (IgM1), s5–s6 by IEL, s6–s7 by DC secreting CD68, s7–s8 by another IgM binding (IgM2) and s8–s9 by IgG. The apoptotic surface cells show an additional binding of IgM (IgM3).
During complete TCS pathway, three cell types are destined to undergo apoptosis (marked >X), intraepithelial T cells, DC and surface epithelial cells. Hence, complete TCS pathway appears to require a suicide of the immune system-derived cells to stimulate advanced differentiation of epithelial cells.
The surface of the squamous epithelium, which could be exposed to the environmental threats, does not show a presence of T cells and MDC (see Fig. 4C, F) but a strong surface and cytoplasmic binding of IgM and IgG (Fig. 3E, F). There is a possibility that these antibodies are naturally occurring monoclonal polyspecific autoantibodies capable of binding multiple and apparently unrelated antigens, including bacterial antigens and toxoids . If so, they may represent a first line of defence against bacterial environmental threats. In addition, normal epidermal stratum corneum exhibits a binding of IgM and IgG autoantibodies isolated from normal human sera to keratins . Moreover, mucosal surfaces are protected by IgG and pentameric IgM [121, 122]. If such autoantibodies are polyspecific, the entire body external surface and mucosal internal surfaces are similarly protected.
A hierarchy of TCS-related elements in the stratified epithelium resembles their appearance during phylogeny and ontogeny. Monocytes/tissue macrophages exist virtually in all animals, including invertebrates . In humans, primitive macrophages first develop in yolk sac haematopoiesis, and monocyte-derived cells are the first differentiated blood cells found in the liver sinusoids by the 7th week of the fertilization age . In the brain, they appear quite early, during the second month of embryogenesis .
Thy-1 differentiation glycoprotein serving for targeted delivery (recognition at cell surface) of growth substances released from pericytes was detected in invertebrates [126, 127], and embryonic blood vessels are composed of endothelial cells and pericytes that organize and expand into highly branched conduits . More differentiated MDC (DCP/veiled cells) presenting antigens to peripheral blood leucocytes and T cells capable of mixed leucocyte reaction are ubiquitous among the vertebrates [129, 130], and the immune response in mammals does not mature until the late prenatal or early post-natal period [131, 132]. When compared to IgG, the IgM molecules appear earlier in phylogeny and ontogeny [113, 133].
Altogether, these observations indicate that a hierarchy of TCS elements reflects their phylogenetic and ontogenetic appearance (see Figs. 1-5), with the MDC being the first, because they accompany proliferation of tissue stem cells, and the IgG being the last involved in the regulation of tissue cell differentiation, because they accompany apoptosis of tissue-specific cells.
Autoimmunity and regulation of tissue differentiation
Monocyte-derived DC are considered to play a role in resistance to infection and cancer and cause autoimmunity, allergy and graft rejection. Maturing DC loaded ex vivo with microbial or tumour antigens may cause disease treatment, and TGF-beta, retinoic acid, rapamycin or steroids may have a tolerogenic effect on DC loaded with antigens .
Nevertheless, it remains unknown why the presence of DC and T cells is beneficial for advanced differentiation of stratified epithelium cells, while their interaction with pancreatic beta cells causes autoimmune type 1 diabetes. The autoimmunity is usually tissue-specific, and its cause could simply be viewed as a stimulation of differentiation of tissue-specific cells into a stage higher than required for normal function of particular tissue.
While the stratified squamous epithelium exhibits all nine stages of epithelial cell development, many other normal tissues show an interruption of tissue cell differentiation at certain stage, i.e. they are stimulated to differentiate to the stage optimal for their normal function and prevented to differentiate above it. An interaction of TCS towards differentiation of particular tissue is supposed to depend on two events:
First, not all tissue-committed immune system–related cells i.e. DCP, T cells and B cells producing autoreactive IgM and IgG may be available for particular tissue. For instance, growing ovarian follicles show an association of activated vascular pericytes and HLA-DR+ MDC with the follicular basement membrane. During follicular regression, the MDC invade granulosa cell layer, but no T cells are present. On the other hand, ovarian CL regression is initiated by invasion of T cells and terminated by binding of IgG to luteal cells [5, 49].
Secondly, the interaction of immune system–related and other TCS components, including autoantibodies, depends on the expression of appropriate ligands on tissue-specific cells (see distinct triangles in Fig. 5B). For instance, Fig. 1C shows that the Thy-1+ vesicles release their contents after reaching parabasal cells, but not during their migration through the basal layer. This indicates that basal cells do not express a ligand for Thy-1+ vesicles. In addition, Fig. 3F shows that autoreacive IgG binds to epithelial cells in the superficial layer, but not to the less differentiated cells. This indicates that less differentiated epithelial cells do not express a ligand for autoreactive IgG.
Stop effect of MDC
The MDC were postulated to play a dominant role in the regulation of differentiation of tissue-specific cells . Intraepithelial MDC may influence expression of ligands on epithelial cells. For instance, DCP secrete HLA-DR among epithelial cells in the mid-parabasal cells (stage 3; arrows, Fig. 3A), and resulting upper parabasal layer shows a binding of IgM1 (stage 4, Fig. 3E). Also, DC secrete CD68 in the mid-intermediate layer (stage 6; white arrowheads, Fig. 3D), and resulting upper intermediate layer shows a binding of IgM2 (stage 7, Fig. 3E). Similarly, IgM binds to luteal cells in the young CL (YCL) but not to the mature luteal cells and then binds again to luteal cells in the regressing CL . In Alzheimer disease, IgG type brain reactive antibodies were found in 57% of patients, and such antibodies were also found in 30% of normal controls. The IgG deposits were detected in brains of patients with Alzheimer disease but not in brains of normal ageing controls , with particularly strong binding to amyloid perivascular plaques . This suggests that patients with Alzheimer disease exhibit IgG-binding ligands in the brain which are absent in normal ageing brains. As a consequence of a lost stop effect of the brain MDC (microglia), the neuronal cells may undergo apoptosis and degenerate into the stage 9 after IgG binding, like the superficial layer in the stratified epithelium (see Figs. 3F and 4I, L).
In addition, the DCP (veiled accessory cells) or DC can leave the periphery and transport via afferent lymph the information to the draining lymph nodes, where they modulate reactivity of T cells and production of immunoglobulins , and DC apoptosis regulates through multiple pathways the balance between tolerance and immunity .
Therefore, it depends on the properties of veiled cells whether autoreactive T cells and antibodies towards certain tissue are produced. From this point of view, the T cells and B cells are effectors of veiled accessory cells. Intraepithelial DCP interact with homing T cells and may cause their activation (DR expression) and apoptosis at certain layer of the epithelium (pb/im interface; Fig. 2A, B). This is accompanied by a differentiation of epithelial cells in the lower intermediate layer (stage 5, Fig. 3B) and differentiation of DCP into DC accompanying stage 6 differentiation (Fig. 3C, D; see also Fig. 5).
Development, differentiation and proliferation of resident MDC are regulated by the tissue microenvironment, including the in situ production of macrophage growth factors in both foetal and adult life, and resident MDC are a long-lived and self-proliferating and renewing population [139, 140]. In particular tissue environments or under particular stimuli MDC can downregulate immune response development. This is not only important for preventing overactivation of the immune system but also for ensuring tolerance against self [139, 141].
In stratified epithelium, we did not find any Ki67+ cells, except for post-mitotic epithelial cells in the parabasal layer (arrowheads, inset in Fig. 1B). However, adjacent lamina propria contained numerous Ki67+ cells (arrows). The lamina propria contains clusters of many DR+ DCP, some of which enter epithelium. It is possible that a long-lived and self-proliferating population of resident MDC [139, 140] is constituted by the DR+ DCP in the epithelium lamina propria or in the stroma adjacent to tissue-specific cells in other tissues.
Resident self-renewing MDC may exhibit a ‘stop effect’ interrupting the complete TCS pathway towards differentiation of tissue cells at a particular stage . This ‘stop effect’ can be encoded in self-renewing resident MDC by the end of the IA. The stop effect determines the highest state of epithelial cell differentiation allowed for a particular tissue. Figure 6A indicates that skeletal muscle, brain and pancreatic beta cells should be stimulated to differentiate into stage 3, but a ‘stop effect’ towards moderately differentiated tissue cells (‘moderate’ SE) exerted by DCP will prevent tissue cells to enter stage 4 of differentiation (for cell differentiation stages see Fig. 5). As discussed earlier, the stop effect at DCP can act here in two ways, e.g. (1) prevent production of autoreactive IgM1, and/or (2) prevent expression of IgM1-binding ligands on tissue cells. Therefore, a lack of IgM1 binding will prevent the cells to differentiate from stage 3 to stage 4.
Figure 6B shows ‘high’ stop effect exerted by intestinal DC, which will allow stage 4 differentiation of intestinal epithelium with a presence of intraepithelial T cells and epithelial IgM binding in the gut [122, 142]. Intestinal DC extend their dendrities between the tight junctions of intestinal epithelial cells and are very important for maintaining intestinal homeostasis . Stop effect of intestinal DC may prevent DC activation, which can cause inflammatory epithelial cell damage and bowel or coeliac disease [144, 145].
Figure 6C shows ‘low’ stop effect exerted by primitive CD14 MDC, which reflects vaginal epithelium atrophy appearing in post-menopausal women, where epithelial cells lacking oestrogen stimulation differentiate into the (para)basal cells only . This may be caused by an inability of primitive CD14 MDC in the vaginal lamina propria to activate Thy-1+ vesicle release from vascular pericytes (see Fig. 1A, C). In women with menstrual cycles, the oestrogens appear to by-pass the ‘low’ stop effect, and stratified squamous vaginal epithelium shows a complete differentiation with apoptosis of the upper superficial cells under high oestrogen levels . However, the ‘low’ stop effect of adult oestrogen-dependent tissues, including vagina, is lost after the oestrogen exposure during the critical developmental period in mice and rats (post-natal days 0–7). Persistent vaginal cornification is induced in mice by a short post-natal treatment with oestrogen, and this is not prevented by later ovariectomy, antioestrogenic steroids, or transplantation into ovariectomized normal mice .
Mouse vaginal stromal cells showed marked proliferative response to oestrogen treatment during the critical developmental period (post-natal days 0 and 5) but not at post-natal days 20 and 70. However, vaginal epithelial cells showed marked proliferative response to oestrogen treatment at post-natal days 20 and 70 but not during the critical developmental period. Microarray analysis revealed that the oestrogen induced gene expression pattern in vaginae is relatively low on post-natal day 0, but when adjacent to the end of the critical developmental period (post-natal day 5), the gene expression is closer to the post-natal days 20 and 70 . These observations suggest that oestrogen-independent permanent cornification of the vaginal epithelium is not necessarily caused by a permanent alteration of the vaginal epithelial cells but can be caused by oestrogen effect on vaginal stromal cells, including MDC, during the critical developmental period.
Setting the stop effect during IA for particular tissue ‘lower than required’ can cause a persisting tissue immaturity, like in a permanent muscular dystrophy (Fig. 6D versus 6A). The Duchenne muscular dystrophy exhibits a deficiency of a sarcolemmal-associated proteins  and appear to be caused by an inability of vascular pericytes to stimulate proper differentiation of muscle cells . On the other hand, if the stop effect is set higher than required (Fig. 6E versus 6A), the homing of T cells to pancreatic beta islets will occur resulting in type 1 diabetes mellitus because of the differentiation of pancreatic islet beta cells being higher than required (stage 4 instead of stage 3).
Stop effect of MDC and immune adaptation during the critical developmental period
Tissue-specific stop effects of MDC are expected to be established epigenetically during IA (critical developmental period) ending at about 6 months of intrauterine life in humans and 7 days of post-natal life in mice and rats .
Epigenetic events of MDC involved in the process may include alteration in DNA transcription pattern by histone acetylation and methylation and correlations with mRNA. Epigenetic mechanisms result in an alteration of cellular phenotype (gene transcription) without altering the genotype. Therefore, epigenetic alterations during critical period of development are not genetically transmitted but are involved in the programming of the health or disease in the given individual. Gene transcription is influenced by modification of surrounding histone proteins and methylation pattern of DNA. Specific modulators influence histone-modifying effector enzymes .
Overfeeding during IA resulting in a metabolic syndrome phenotype (obesity, hyperleptinaemia, hyperglycaemia, hyperinsulinaemia and an increased insulin/glucose ratio) is caused by an alteration of the DNA methylation pattern . Aberrant histone lysine methylation patterns that change chromatin structure can promote dysregulated gene transcription and disease progression.
As indicated earlier oestrogen injection during IA permanently alters vaginal stromal cell rather than epithelial cell properties. Diabetic conditions mimicked by culturing THP-1 MDC in high glucose relative to normal glucose lead to dynamic changes in histone H3 lysine acetylation  and methylation . Relevance to human diabetes was demonstrated by noting that histone H3 lysine 4 dimethylation at the coding and promoter regions of two candidate genes was significantly greater in blood MDC of patients with diabetes relative to normal controls similar to the THP-1 MDC data. Moreover, regular mRNA profiling with cDNA arrays revealed correlations between mRNA and and lysine 9 dimethylation levels . In addition, lymphocytes of patients with type 1 diabetes displayed a distinct profile of chromatin histone H3 lysine 9 dimethylation when compared to lymphocytes from control heathy subjects .
These observations indicate that gene transcription of MDC is altered in vitro by high glucose conditions. In women exhibiting high glucose levels during pregnancy, such alteration of MDC gene transcription can be encoded in the foetus during IA and persist thereafter, as evidenced from the altered circulating MDC in patients with diabetes versus normal controls. Because the MDC regulate properties of T cells, the gene transcription is similarly affected in T cells of patients with diabetes. This results in the inflammatory gene transcription with alteration of the pancreatic beta cells and persistent diabetic condition.
In animal models, inhibition of ovarian differentiation by oestrogens during IA in post-natal rat females causes persisting ovarian immaturity in adult rats . This suggests that the muscular dystrophy (Fig. 6D) could be caused by a retardation of muscle development during adaptation.
On the other hand, stimulation of advanced ovarian differentiation by androgens during IA in post-natal rats results in a persistence of prematurely aged ovaries in adult rats . This suggests that the type 1 diabetes (Fig. 6E) could be caused by a stimulation of advanced pancreatic beta cell differentiation during adaptation.
Both types of ovarian dysfunctions cannot be induced after termination of the IA, e.g. in 10-day old post-natal rat females [25, 50]. However, there are differences whether the ovarian development in mice is inhibited by oestrogens earlier (post-natal days 0–3) or later (days 3–6) before the end of the IA. In earlier injected females, the persisting ovarian failure is permanent but the later injected animals begin to ovulate at 8 months of age .
This suggests that persisting ovarian immaturity can result in a delay of normal ovarian function when the ovarian development is inhibited shortly before the end of the IA. Therefore, one may expect that the ‘stop effect’ has a tendency to ‘shift up’ with age advancement. This shift up of the stop effect with age can cause, however, an increased incidence of autoimmune/degenerative diseases, where the stop effect is higher than required (autoimmunity) or lost (degenerative diseases) – see Fig. 6F.
Consequences of cell differentiation during IA for the tissue function outcome are summarized in Fig. 7 showing four possible variants (A–D). Retardation during immune adaptation (IA, Fig. 7A) will result in persisting immaturity, where the ‘stop effect’ is set lower than required for functional cells in particular tissue. Interestingly, muscular dystrophy does not affect all muscles of the body, but different types of muscular dystrophy show different symmetric areas of muscles affected by a muscle weakness . This may be because of the selective retardation of certain muscles during particular period of their development and subsequent self-renewal of corresponding resident MDC carrying the ‘stop effect’ causing persistent muscle immaturity.
Differentiation of functional cells (Fig. 7B) – tissues vary in an optimal stage of tissue-specific cell differentiation required for proper tissue function – see Fig. 6A–C, will enable the tissue cells to be preserved in functional stage and prevented from ageing. Acceleration during IA (Fig. 7C) will cause persisting premature ageing because of the autoimmunity. The higher then required ‘stop effect’ still prevents a tissue from complete regression, which occurs in tissues like ovarian CL lacking ‘stop effect,’ because such tissue was absent during IA (Fig. 7D). Dashed arrow in Fig. 7 indicates a tendency of the ‘stop effect’ to ‘shift up’ with the age advancement.
Determination of tissue functional longevity
The heart differentiates into functional tissue from early stages of ontogeny and IA (‘Long’ IA, Fig. 8) and can function for 100 years in humans. The ovary, which differentiates later, can function only half of that period. The ovarian foetal primordial follicles (fpf) begin to differentiate during 4th month of intrauterine life (‘Moderate’ IA). After termination of IA (dashed line), the ageing primordial follicles (apf) persist until menarche, and normal ovarian function is limited by follicular renewal (fr) during the prime reproductive period, until 35–40 years of age ). The immune system shows a beginning of significant functional decrease between 35 and 40 years of age in humans  and concomitantly ovarian follicular renewal wanes . Thereafter, ageing primordial follicles are utilized until exhausted, and physiological menopause occurs at about 50 years.
‘Shorter’ IA towards the ovary may occur when ovarian differentiation is delayed. Consequently, the follicular renewal is shorter and results in ovarian dysfunction known as a Premature Ovarian Failure (POF). Patients with POF have been found to have abnormalities in the function of MDC and DC , suggesting a relationship between immune system and POF.
More delayed ovarian differentiation may cause ‘Short’ IA resulting in the lack of the follicular renewal and primary amenorrhoea. Absence of ovarian CL during IA causes their cyclic degeneration, except during pregnancy, which is accompanied by immune suppression .
Lessons from immune maintenance of ovarian function
A number of studies in distinct animal species indicated that immune cells play an important role in cyclic ovarian activity and suggested a possibility that the immune system participates in the regulation of ovarian function [32, 160]. The ovary exhibits cyclic renewal, differentiation and regression of its functional structures and is therefore a unique model for better understanding of the immune system role in these processes. The ovarian function is also affected relatively early during the female life, along with the early stages of immune senescence. Altogether, studies of the role of the immune system in ovarian homeostasis could contribute to a better understanding of the immune system role in morphostasis in general.
Differentiation of human ovarian stem cells into granulosa cells
Ovarian stem cells in human ovaries (osc, Fig. 9A) are capable to differentiate into two distinct cell types required for follicular renewal: (1) precursors of follicular granulosa cells and (2) ovarian germ cells [38, 51]. In adult humans, the OSC differentiate from cytokeratin (CK)-positive fibroblast cells in tunica albuginea (+fb, panel B) transformed into a transitory type (fb/osc) with the assistance (arrowhead, panel C) of strongly DR+ DCP (DRhigh). The fb/osc cells form the OSC (arched arrow, panels B and C) in the presence of less DR+ DCP (DRlow, panel C).
To form granulosa cell precursors, a bilaminar layer of OSC (arrow, panel D) is overgrown by ovarian tunica albuginea (ta); note a lack of OSC at the ovarian surface (arrowhead). The OSC layer descends into the upper ovarian cortex (uc, panel E), where it is fragmented into small nests of primitive granulosa cells (arrowheads, panel E) [161, 162]. The nests descend into the lower cortex (lc and yellow arrowheads, panel F) and cover the interior (n, panel G) of the vascular lumen (vl) to catch intravascular oocyte (o) by an extending arm (a). The oocytes captured by epithelial nests form new primordial follicles showing CK+ Balbiani body within the oocyte (white arrowheads, panel F) .
These observations indicate that the development of granulosa cell precursors from OSC requires an involvement of DR+ DCP. The number of granulosa cell nests appears to determine the number of newly formed primordial follicles during adulthood, because superfluous germ cells degenerate in medullary vessels . A lack of primitive granulosa cell nest formation will cause a lack of follicular renewal in POF, regardless of eventual formation of new ovarian germ cells.
Involvement of MDC and T cells in development of germ cells from ovarian stem cells by asymmetric division
The asymmetric division of OSC in adult human ovaries (white arrowhead, Fig. 10A–D) is initiated by interaction of CD14 MDC with OSC and release of CD14 (arrows, panel B). Resulting cell daughters are still accompanied by CD14 MDC (yellow arrowheads, panels A and B), but one of them develops into secondary germ cell (sgc and red asterisks, panels A–D) in the presence of CD8 T cell (red arrowheads, panels A and C). The other daughter remains in the pool of OSC (osc and yellow asterisks, panels A–D). In panel D, double colour immunohistochemistry shows that resulting OSC daughter keeps the CK staining (CK, blue colour) and developing germ cell loose CK and expresses PS1 meiotically associated antigen (PS1, brown colour). Similar asymmetric division was detected in adult rat ovaries – panel E shows emerging rat germ cell (red asterisk) expressing zona pellucida protein (ZP, magenta colour) in the presence of lymphocyte type cell (black asterisk) expressing leucocyte common antigen (LCA, brown colour) [38, 42, 51]. These data indicate that formation of new germ cells in adult mammalian females requires an asymmetric division of OSC with an involvement of CD14 MDC and CD8 T cells.
Asymmetric division of OSC is also present in mid-pregnancy human foetal ovaries (white arrowhead, panel F). The OSC daughter keeps the MHC class I expression (yellow asterisk), while the developing germ cell is MHC class I depleted (red asterisk). Symmetric division of MHC class I–negative germ cells follows (yellow arrowhead, panel F) and is accompanied by CD14 MDC (yellow arrowhead panel G). Asymmetric division is accompanied by activated CD8 T cells expressing HLA-DR (panels H and I) .
Taken together, the CD14 MDC appear to promote a symmetric division of OSC and secondary germ cells but an involvement of CD8 T is also required for asymmetric division of OSC.
Developing secondary germ cell attains amoeboid shape (dashed line, panel F) and enters the ovarian cortex to form primordial follicles (pf, panel J). In foetal and adult human ovaries, growth of resting follicles (rf, inset, panel J) is initiated by activation of vascular pericytes releasing Thy-1 differentiation protein among granulosa cells (arrowhead, inset). Growing follicles (gf, panel J) are accompanied by DR+ MDC (asterisk), penetrating the follicular basement membrane (dashed line). The activated MDC exhibit extensions (white arrowheads) among granulosa cells and towards (yellow arrowhead) the oocyte nucleus (on) [49, 52].
Therefore, once primordial follicles are developed, some of them are selectively activated towards further growth, possibly by disinhibition of the Thy-1+ pericyte activity by autonomic innervation regulating function of pericytes . Continuation of follicular development also requires activation of perivascular MDC showing direct contacts with granulosa cells and growing oocyte.
The MDC accompany migration of secondary germ cells in adult human ovaries
In adult human ovaries, CD14 MDC also accompany (white asterisk, Fig. 11A) symmetric division of germ cells (gc and arrows) in tunica albuginea (ta) under the OSC. The germ cells leaving the tunica albuginea and entering the upper ovarian cortex (uoc) are guided by CD14 MDC (arrowhead). The migrating amoeboid germ cells are accompanied by DR+ MDC (arrowhead, panel B) and show perinuclear DR expression (arrow). They associate with the cortical vasculature (panel C) visualized by a strong MHC class I expression on vascular endothelium (ve) and enter vascular lumen (vl, panel D) lined by Thy-1+ pericytes (arrows) . The vascular transport in adult human ovaries enables germ cells to reach granulosa cell nests lining vascular lumen in distant lower ovarian cortex (about 1000 μm apart) and form new primordial follicles (see Fig. 9F, G) .
These observations indicate that MDC play a pivotal role during migration of germ cells towards microvasculature in the upper ovarian cortex, and ovarian venules serve for a transport of germ cells to granulosa cell nests in a distant lower ovarian cortex.
Totipotency of ovarian stem cells
In vitro, the OSC are capable to differentiate into distinct cell types, including epithelial cells, fibroblasts, neural cells and germ cells . With the assistance of satellite cell (black arrowhead, Fig. 11E) providing additional organelles (black arrow) and neuronal type cell (white arrow) with dendritic extension (white arrowhead), the germ cell differentiated into large (150 μm) oocyte with germinal vesicle breakdown (gvbd) and thick ZP membrane lacking expression of ZP proteins. Such oocytes eventually differentiated into four-cell parthenogenetic embryos (panel F) expressing deleted azoospermia–like (DAZL) protein – inset shows a DAPI staining with two isolated nuclei (arrowheads) and a dividing cell (arrow). Also observed were DAZL+ parthenogenetic blastocysts (panel G) consisting of blastocoele (bc), trophectoderm (te) and inner cell mass (icm), the latter producing large quantities of DAZL+ embryonic stem cells (esc). Ovarian stem cells were capable to differentiate into oocytes regardless of the patient’s age and including ovaries of patients with POF [55, 164–166].
These observations indicate that totipotent OSC are capable to produce normal oocytes in vitro, because spontaneous parthenotes can also develop during culture of normal non-inseminated oocytes .
Involvement of TCS in ovarian CL development and regression
During menstrual cycle, unless pregnancy occurs, the ovarian CL life is limited to the following phases: young CL developing after ovulation (YCL in Fig. 12), mature CL (MCL) about 1 week later and CL regression (RCL) till the end of the menstrual cycle. During the next menstrual cycle, the terminally regressing structure shows an amorphous character called the corpus albicans (CAlb). In the case of pregnancy, the CL persists as a CL of pregnancy (CLP).
Figure 12A shows that CL development is associated with a release of large quantities of Thy-1+ ICV (arrow) from vascular pericytes (p). After reaching luteal cells, the ICV collapse into Thy-1+ spikes (arrowhead). In MCL, the activity of pericytes towards luteal cells diminishes but is still apparent (arrow and arrowhead, panel B) within the microvasculature (mv). In CLP, weakly active pericytes (arrow, panel C) accompany microvasculature. During CL regression, an alteration of pericyte morphology is apparent (arrow, panel D), and this is accompanied by a homing of CD8 T cells (inset) causing regression of luteal cells (arrowhead). The CD3 and CD4 T cells were also involved . Amorphous CAlb contains remnants of degenerated vascular pericytes (panel E).
Therefore, Thy-1 pericytes appear to be required for the CL development and maintenance in the functional state, and their regression causes homing of T cells and regression of luteal cells.
Autoreactive IgM shows week reactivity with microvasculature and strong binding to luteal cells in YCL (panel F). In contrast, no IgM binding is apparent in the MCL (panel G). In CLP, the IgM binding to microvasculature but not to luteal cells is apparent (panel H). During CL regression, IgM binding to both the regressing vessels and luteal cells is present (panel I) and similar character persists in the CAlb (panel J).
Collectively, the early binding of autoreactive IgM may stimulate differentiation of luteal cells and represent a parallel to IgM1 binding to young differentiating epithelial cells in the squamous epithelium (stage 4, Fig. 3E). Binding to regressing luteal cells and amorphous CAlb may contribute to the ageing and apoptosis of luteal cells (see stages 7 and 9, Fig. 3E). Accordingly, like in the ageing stratified epithelium, the CL regression was also accompanied by IgG binding to luteal cells .
CD14 MDC associated with microvasculature showed CD14 release in YCL (arrow, Fig. 12K), stimulating proliferation of endothelial cells and pericytes during CL neovascularization, and MDC in regressing CL showed DR expression and dendritic morphology .
Altogether, the involvement of TCS in ovarian CL differentiation resembles TCS role in the differentiation of squamous epithelial cells. The CL regression is, however, accompanied by a detrimental effect of T cells on luteal cells. This contrasts the T cell apoptosis in the stratified epithelium, which suggests an extrathymic origin of these intraepithelial T cells. It is more likely that T cells infiltrating CL during regression are of thymic origin. This is supported by an observation that CL regression is affected in climacteric women , possibly because of the age-associated involution of the thymus. Advanced thymic involution has a considerable impact by inhibiting T cell immune response because of the disintegration of thymic epithelial network evident in mice as early as at 1 year of age . In humans, linear patter of thymic involution begins at 35–40 years of age , and the climacteric symptoms in premenopausal women begin to occur at ≥46 years of age . From the beginning of advanced thymic involution, both the thymic cortex and the medulla begin to show the same progressive loss of tissue that, unlike in mice, is replaced by fat tissue . The climacteric symptoms are more severe in women with higher progesterone (PG) levels  caused by hyperplasia of luteal cells in persisting CL lacking invasion of T cells .
Taken together, the physiological luteal regression, which is essential for regular cyclic ovarian function, may be caused by thymic T cells directed against luteal cells. They are released from the thymus to destroy MCL ‘graft’ because such T cells escape negative selection (clonal deletion) because of the lack of luteal cells during IA towards self-antigens. With advancement of thymic involution, the release of these T cells autoreactive against the luteal cells diminishes, and persisting luteal cells are stimulated towards hyperplasia by vascular pericytes because of the lack of AI in the CL.
Augmentation of cancer
An evidence for the role of disruption of morphostasis in malignancy has been discussed by the means that cancer occurs much more frequently when the morphostatic influences fail . Once emerged, however, the neoplasia is augmented by tissue fibroblast–related mesenchymal cells , such as vascular Thy-1+ pericytes, and also by the immune system–related elements (MDC and T cells) stimulating proliferation and differentiation of normal tissues [1, 49], possibly because of the hybridization of malignant with normal tissue cells .
Hybridization of tumour cells with normal host cells
Malignant cells have an enhanced ability to hybridize with various types of normal cells . It is commonly believed that such fusions generally lead to the ‘normalcy’ of resulting malignant/normal cell hybrids . In reality, the behaviour of malignant/normal cell hybrids shows a marked diversity, depending on the cell type involved. For instance, fusion of myeloma cells with B lymphocytes, widely used for production of monoclonal antibodies, results in a hybridoma still capable of producing ascites and killing the host.
Possible pathways of hybridization of malignant and normal cells [175, 176] is shown in Fig. 13. The malignant stem cell may fuse with normal stem cell or other fusogenic cells within the tumour. Tumour tissues contain fusogenic cells, which include tissue-specific stem cells, tumour cells and cancer stem cells, as well as recruited monocytes/macrophages and bone marrow-derived stem cells [176, 177]. Resulting heterokaryon shows the presence of the malignant and normal cell nuclei. The cell identity can be modified by transdifferentiation, fusion or nuclear transfer . After deregulated cell division, resulting cell daughters are malignant/normal cell hybrids (Fig. 13), because they contain a set of malignant (black) and a set of normal cell chromosomes (blue). Fusion of malignant cells with distinctly different type of normal cells results in less stable hybrids than fusion with relevant type of normal cells. Continued division of malignant/normal cell hybrids may gradually lose normal chromosomes and renew their hybridization potential (dashed arrows, Fig. 13). Because dividing cells lie in a proximity of tumour microvasculature, the malignant cells renewing the hybridization potential may enter the vessels and cause distant metastasis in tissues suitable for hybridization of normal stem cells with particular malignant stem cells.
In the carcinoma/normal epithelium hybrids (relevant cell types), the malignant potential is greatly reduced, whereas in the carcinoma/normal mesenchyme hybrids (distinctly different cell types) the malignant phenotype is expressed dominantly. In both cases, the dominant in vitro and in vivo phenotype, but not behaviour, is that of the normal parental cells. In the carcinoma/epithelium hybrids, the more differentiated epithelial phenotype is dominant, and in the carcinoma/fibroblast hybrids, the mesenchymal phenotype is dominant .
The cell identity can also be modified by epithelial to mesenchymal transition. The loss of epithelial characteristics and the acquisition of a mesenchymal-like migratory phenotype are essential to the development of invasive carcinoma and metastasis . An alternative for epithelial to mesenchymal transition is a fusion of tumour cells with bone marrow-derived cells, and MDC in particular, which could lead to the metastatic phenotype .
Cancer cell fusion with normal tissue stem cell can produce a malignant phenotype exhibiting an oncogenic resistance, such as increased drug resistance and malignancy. It has been demonstrated that breast stem cells can fuse spontaneously with breast cancer cells, and this results in stable hybrid cells, which possess an increased proliferatory activity, increased expression of multidrug resistance transporters and antiapoptotic proteins, as well as BRAC1 and BRAC2 molecules that have been associated with a poor prognosis in breast cancer . However, beside those effects promoting malignancy of cancer and normal cell hybrids, such hybrids may express cell surface markers of normal cell, resulting in tumour augmentation by TCS [47–49]. Indeed, hybrids of human monocyte and mouse melanoma show increased motility, enhanced metastatic potential, tend to be super melanotic (malignant cell parent phenotype) but express some human genes (normal cell parent phenotype) .
Augmentation of cancer by the TCS
In experimental oncology, the suspension of the spleen cells injected in immunologically incompetent animals has been shown to stimulate tumour growth . Hence, immune/mesenchymal cells, which are required for regeneration and function of normal tissues, may facilitate malignant disease.
We studied an involvement of the TCS in ovarian cancer when compared to the CL management. Figure 12L shows extensive secretion from vascular pericytes (p) of Thy-1+ ICV (arrow) collapsing into spikes when reaching adjacent malignant cells (black arrowhead) and endothelial cell sprouts (e, and white arrowhead). This resembles behaviour of Thy-1 pericytes in the YCL (panel A). The IgM binding to microvasculature in ovarian cancer and lack of binding to malignant cells (panel M) were similar to that of the CLP (panel H). High activity of CD14 MDC in ovarian cancer (arrows, panel N) in release of CD14 associated with multiplication of cells in normal tissues was similar to that in YCL (panel K). However, the CD8+ T cells (black arrow, panel O) showed release of CD8 after entering the ovarian cancer (white arrowhead) and underwent regression among cancer cells (arrowheads, inset panel O). This contrasted with regression of luteal cells mediated by CD8+ T cells in the regressing CL (inset, panel D). Therefore, TCS conditions that maintain CL and prevent its regression during pregnancy show striking similarity in cancer augmentation, with one exception – there are no stem and dividing luteal cells within the CL, but cancer cells associated with tumour microvasculature permanently proliferate. The apoptosis of homing T cells in the cancer indicates that they are rather extrathymic T cells susceptible to apoptosis .
Ovarian cancer cells may hybridize with CD68+ MDC
Fusion of tumour cells with MDC was detected in numerous animal and human cancers . A question arises which of the MDC types is likely to hybridize with cancer stem cells in vivo and provide normal/cancer cell hybrids. Primitive CD14 MDC (arrowhead, Fig. 14A) associate with microvasculature (mv) and secrete CD14 among cancer cells (arrow) but no CD14 expression is apparent on malignant cells. The interaction of MDC with stem cells is accompanied by a proliferation of malignant cells evidenced by Ki67 expression (brown colour, panel B), like in the stratified epithelium (see Fig. 1B and arrowheads in its inset). DR+ DCP secreted large amounts of DR among malignant cells (arrow, Fig. 14C), suggesting an attempt to initiate early differentiation of malignant cells – compare with distinct DR release among mid-parabasal cells in Figs. 2A and 3A. However, no DR expression is apparent on malignant cells. These observations suggest that neither CD14+ MDC nor DR+ DCP hybridize with ovarian malignant cells in vivo.
However, staining for CD68 (MDC maturation marker) of more differentiated MDC showed moderate or strong expression on virtually all ovarian malignant cells (Fig. 14D) and strong staining of perivascular MDC (m). Detail panel E shows strongly CD68+ perivascular MDC (m) showing a tropism with membrane apposition (white arrowheads) to unstained perivascular cell (dotted line), possibly representing malignant stem cell (s). Resulting hybrids divide (dashed lines) and produce moderately CD68+ cancer cells (dashed/doted line). Black arrowheads indicate strongly counterstained nucleoli in a dividing CD68-negative cancer stem cell.
These observations suggest that CD68-negative ovarian cancer stem cells divide under the influence of the host CD14+ MDC, post-mitotic malignant cells fuse with CD68+ MDC, and resulting hybrids utilize differentiation promoting substances released by DR+ MDC to make conditions for continuing proliferation. However, because of the inability of malignant cells to differentiate into the functional stage, the tissue-committed TCS continues to stimulate cancer stem cell proliferation. Moreover, because of the CD68 expression of normal MDC on malignant/normal cell hybrids, the immune system is unable to recognize them as a target for the immune surveillance. CD68 expression is significantly increased in the invasive margin of colorectal cancer . Of particular importance for the cancer augmentation might be an observation that the tissue-homing CD68+ MDC express a tissue-specific microRNA .
Most problematic aspect in the treatment of malignant disease is a lack of specific targeting of growing tumour cells or malignant cells remaining after the surgery. Attempts made in this direction, including local treatments, appear to fail. Current therapies of advanced cancers are usually based on cytostatic drugs and approaches such as surgery and irradiation, which greatly affect normal host tissues. Thirty to fifty per cent of patients with advanced cancer die, in spite of the enormous and continuously expanding research devoted to the cancer. Because of the lack of tumour-specific antigens in most human cancers, the immunotherapy of malignant disease has not been yet successful.
A better understanding of the events occurring at the host-tumour interface and the role of the host organism in promoting cancer development and growth may revolutionize cancer therapy and prevention. This may require a revolutionary switch in the perception of the essence of malignant disease. One should realize that the survival, proliferation and metastasis of malignant cells is completely dependent on the behaviour of the host mesenchymal cells and MDC in particular. In contrast to the traditional strategy and current developments, which are directly oriented on proliferating malignant and endothelial cells and reduction in tumour cell mass (surgery, elimination of proliferating cells, induction of apoptosis, reduction in vascular supply), we may do better by targeting the host mechanisms enabling cancer development and promotion. Alloimmunization has been proposed to prevent cancer development . For an established cancer, a development of drugs capable of preventing cell fusion [177, 186] should also be considered.
There, however, might be novel strategies for the targeting the activity of mesenchymal cells in ‘tissue control units’ associated with cancer. An essential issue in cancer therapy is to prevent proliferation of malignant cells by CD14+ primitive MDC. CD14 is a lipopolysaccharide receptor , and association of CD14 MDC with proliferating normal and cancer cells is apparent. Hence, there might be some role of CD14 in promotion of tumour cell multiplication. In addition, CD14 MDC have been shown to inhibit activation of NK cells, which recognize and kill tumour cells . It is possible that temporary targeting of CD14 MDC, e.g. by CD14 antibody or other approaches affecting primitive MDC, may result in the regression of unstable TCU in malignant tissues (lack of innervation), activation of NK cells and subsequent regression of malignant stroma and elimination of tumour cells .
Recent observations indicate that REIC/Dkk-3 protein, which is significantly downregulated in cancer cell lines and clinical cancers, caused advanced differentiation of CD14 MDC into a dendritic phenotype and suppressed cancer growth with dendritic cell and CD8 T cell accumulation . These data indicate that a stimulation of CD14+ primitive MDC into more differentiated phenotype indeed causes a suppression of cancer growth.
Regenerative medicine and morphostasis
Within the field of regenerative medicine, there is a considerable interest in cellular therapy, such as grafting of stem cells into the affected tissue to induce renewal and repair of tissue defects with new functional tissue-specific cells. The stem cells can be isolated from the embryonic or adult tissues and propagated in vitro in the presence of mitogenic growth factors prior to use. Additional sources are umbilical cord blood, amniotic epithelial cells, bone marrow stem cells and mobilized peripheral blood CD133+ cells. The pluripotent adult stem cells could be stimulated to differentiate into particular phenotype or direct reprogramming of a fully committed differentiated cell from one lineage into another could be utilized thereby bypassing steps of either dedifferentiation or reversion to a pluripotent state. Once the proper stem or reprogrammed cells are obtained they are introduced into the body either systemically or locally to stimulate repair of particular tissue (reviewed in ).
An approach alternative to the organ-/tissue-specific stem cells utilized for functional engraftment to the particular sites by the regenerative medicine (topic therapy) could be a ‘systemic regenerative treatment’ with utilization of common drugs having a low molecular weight with the ability to cross a blood-tissue barrier, such as the blood-brain barrier. Sex steroids may have the potential to stimulate the proliferation and differentiation of existing neural stem cells (NSC). They easily pass the blood-brain barrier and can bind to abundant sex steroid receptors in the brain areas important for the regulation of emotions, cognition and behaviour . However, it still remains to be determined whether utilization of individual sex steroids alone might be efficient in prevention or treatment of neurodegenerative diseases and traumatic neurological injuries. To address this question, we studied the effect of sex steroids and their combinations on the transdifferentiation of pluripotent human ovarian epithelial stem cells and porcine granulosa cells into NSC and neuronal cells.
Age-related loss of androgens in man is associated with symptoms that include depression and impaired cognitive function. Lower free testosterone (TS) levels can be detected 5–10 years prior to the diagnosis of Alzheimer’s disease. Animal experiments suggest that both oestrogens and androgens can play a protective role in preventing neurodegeneration. In clinical trials to date, the most studied sex steroids have been androgens. Androgen therapy, however, has no apparent effect on cognitive performance, including verbal and non-verbal memory in elderly patients. Besides in the gonads, some steroids are also synthesized in adrenals and in the brain; the latter are sometimes termed neurosteroids (reviewed in ).
Ovarian epithelial cells transdifferentiate in vitro into neural and neuronal cells
Our initial observations of cultured OSC indicated that they are capable to differentiate spontaneously into distinct cell types in a single well of tissue culture, including neuronal cells (inset, Fig. 15A ). Additional experiments utilized sex steroid combinations in ovarian cell cultures.
Ovarian epithelial stem cells differentiate into large epithelial cells lacking expression of SSEA-1 and neural cell adhesion molecule (NCAM). A few epithelial cells in untreated cultures showed moderate staining for Thy-1 differentiation protein (Fig. 15A) and similar expression of SSEA-4. Addition of individual gonadotropins (FSH and/or hCG), EGF, Estradiol (E2), PG and TS alone, or E2 + PG and E2 + TS combinations showed no change in either cell morphology or immunohistochemical staining. No changes were observed in control cultures, including those with the sex steroid vehicle .
On the other hand, utilization of TS mixed with PG 1 day after E2 pretreatment produced a marked effect 1 h after the treatment. There was a transdifferentiation of epithelial cells into small cells, a portion of which strongly expressed SSEA-1 (arrowhead, panel B), a glycoconjugate of NSC and precursor cells . An asymmetric division resulting in SSEA-1+ and SSEA-1− daughter cells is shown in panel C; note stained early extensions (arrow) associated with the SSEA-1+ cell. Three hours after the treatment, neuronal type cells developed and strongly expressed Thy-1 antigen (white arrowhead, panel D), a GPI-anchored protein abundantly expressed by neurons . Many cells exhibited development of Thy-1+ extending processes (black arrowhead), characteristic for neuronal differentiation.
Stage-specific embryonic antigen-4 (SSEA-4) is commonly used as a cell surface marker to identify the pluripotent human embryonic stem cells and SSEA-4 expressing cells enriched in the neural stem/progenitor cell fraction . It was strongly expressed in neuronal cell bodies (black arrowhead, panel E) but not extensions (white arrowhead). Control immunohistochemistry produced no staining of neuronal or other cell types. In phase contrast observations, large number of putative NSC was found to be detached from the chamber bottoms and were found floating in the centre of the wells (panel F). These cells showed bubble-like anchoring extensions (arrowhead, panel G), which apparently serve for the attachment of seeded non-neuronal cells. Because such putative NSC did not attach again, they may be ready for homing where needed.
Similar transdifferentiation was also observed in porcine ovarian granulosa cell cultures. In vivo, the pluripotent ovarian epithelial stem cells have been found to give rise to two distinct cell types in the ovaries, the ovarian germ and granulosa cells [38, 51]. Porcine ovarian granulosa cells characteristically differentiated in vitro into cells of fibroepithelial phenotype.
The treatment of long-lasting porcine granulosa cell cultures on Day 0 with EGF, hCG and E2 alone, or with the mixture of PG + TS had no apparent effect on the occurrence of neural/neuronal cells, but a combination of E2 + PG + TS produced a small increase on the subsequent day. No changes were observed in the vehicle treated cells. Additional steroid treatment on Day 1 consisted of PG + TS. There was a marked conversion of fibroepithelial into the neural/neuronal cell types in cultures pretreated with E2 + PG + TS or PG + TS a day before. The neural/neuronal cells showed SSEA-4 expression in the porcine neuronal cell bodies, and strong NCAM expression was detected in small stem-like cells. Stage-specific embryonic antigen-1 staining was apparent in the neuronal cell types but not in fibroepithelial cells .
The presence of NSC with anchors after combined steroid treatment in human ovarian stem cell cultures suggests that such adult stem cells could eventually be utilized for the systemic or topic autologous regenerative treatment of neuronal disorders in women. Our observations indicate that once developed, many NSC are released into culture medium and are easy to collect. Our observations also indicate that in human females the ovarian stem cultures could be readily obtained from ovaries regardless of the patient’s age, frozen, recultivated, transdifferentiated into neural/neuronal cells by sex steroid combination and utilized for autologous stem cell therapy.
Finally, in addition to uses in regenerative medicine of in vitro-derived NSC, sex steroid mixtures can stimulate proliferation and differentiation of NSC derived from the CNS and cause transdifferentiation of potential sources of NSC. The adult bone marrow and peripheral blood are of particular interest. They could be a natural source of NSC precursors, as they are a source of haematological and non-hematological stem cell lineages capable of migrating to particular tissues [195, 196]. New NSC can differentiate from peripheral blood precursors in vitro  and possibly under the influence of sex steroids in vivo. The circulating NSC can home within the CNS  and can be stimulated to differentiate further by sequential steroid treatment, i.e. E2 followed by PG + TS .
Vascular smooth muscle cells transdifferentiate in vitro into neural and neuronal cells and vascular smooth muscle stem cells
In a subsequent study , we tested the effect of Neurobasal/B27 neuron culture medium with 2 mm Glutamax (B27 medium), which is utilized for the preservation of hippocampal neuronal cells in vitro because it substitutes the role of astrocytes , on cultured OSC. Within 3 h, there was a direct transdifferentiation of epithelial into interconnected neuronal cells (Fig. 15H, phase contrast of living culture). Such cells strongly expressed Thy-1 neuronal glycoconjugate (panel I), similarly to cultures exposed to sex steroid combinations. These observations indicate that sex steroid combinations are similarly effective as B27 medium.
Subsequently, we have chosen primary culture of human vascular smooth muscle cells (SMC), the non-epithelial mesenchymal cells known to express sex steroid receptors [199–201], to test them as a control cell type regarding their morphology and expression of NSC and neuronal markers. Unexpectedly, however, the vascular SMC subjected to sex steroid combinations also transdifferentiated into NSC and differentiating neuronal cells .
Cultured human vascular SMC exhibited moderate expression of Thy-1 differentiation protein (Fig. 15J), which is characteristic for human vascular pericytes . The cultured cells did not express cytokeratin 5 (CK-5, upper inset), but strongly expressed alpha smooth muscle actin (αSMA, lower inset). Utilization of single steroid treatment alone did not change the SMC morphology and Thy-1 expression. However, one day after the treatment with PG + TS (d1 PT, panel K), a direct transdifferentiation of SMC into neuronal type cells emerged with strong Thy-1 expression (white arrowhead). Note also enhanced Thy-1 expression in remaining SMC (arrow, compare with panel J). Open arrowhead indicates the SMC which is already in the process of transdifferentiation into neuronal cell.
The situation which we called a ‘brain in vitro’ feature  is shown in panel L, where all SMC were transdifferentiated into connected neuronal type cells (vimentin stain). The emerging neural/neuronal cells showed strong SSEA-1 expression. Panel M shows an asymmetric division resulting in SSEA-1+ (black arrowhead) and SSEA-1− (white arrowhead) daughters like in transdifferentiated OSE cultures – see panel C.
Neural nuclei protein expression in cultures 1 day after E2 + PG + TS (EPT, panel N) treatment was associated with neuronal cell bodies (black arrowhead) but not extensions (white arrowhead). Control immunohistochemistry produced no staining of neuronal (arrowhead, panel O) and other cell types. Neuronal cells showed stage-specific embryonic antigen 4 (SSEA-4, panel P) in their bodies (black arrowhead) but not extensions (white arrowhead), like in ovarian stem cell cultures – see panel E. Neuronal cell bodies and extensions expressed neural cell adhesion molecule (NCAM; arrowheads, panel Q) characteristic of later stages of neuronal differentiation . Neural cell adhesion molecule is used for the isolation of human embryonic stem cell-derived neurons .
Utilization of B27 medium without steroids caused a fast (within 3 h) transdifferentiation of SMC into neuronal cells (panel R, phase contrast of living culture) like in ovarian epithelial cell cultures (see panel H). These data clearly demonstrate that both the SMC and ovarian epithelial cells have a potential to transdifferentiate into neuronal cells in the medium stimulating preservation of neurons and neuronal differentiation of embryonic stem cells [198, 202].
Treatment with E2 + PG + TS caused fast (within 3 h) conversion of SMC into stem type cells (black arrowheads, panel S) exhibiting bubble-like anchors (white arrowheads) signalling a readiness for settlement. These cultures within several days showed a re-establishment of the vascular smooth muscle cell type (panel T). This indicates that sex steroid combinations could cause transdifferentiation of SMC into neuronal and vascular smooth muscle stem cells.
As a positive control for CK-5 staining, we used a culture of amniotic epithelial cells (panel U) which did not express αSMA (negative control, panel V) . Amniotic culture responded well by transdifferentiation to neuronal cells after B27 treatment (panel X) but poorly after sex steroid treatment (panel Y) – compare with panel K (unpublished data).
17β estradiol in vascular SMC cultures after TS, PG and TS + PG treatment
Our earlier study  indicated that pretreatment with E2 is essential for the transdifferentiation of OSC and granulosa cells into NSC and neuronal cells after PG + PT treatment, but some cultures were capable to transdifferentiate after PG + PT alone, possibly because of the conversion of TS (and PG) by aromatase. We investigated a possibility that the PG, TS or both could be converted into E2 in vascular SMC cultures. The medium with 10% FBS and spent control culture medium showed low E2 (E2 pg/ml 1050 ± 150 and 1477 ± 742 SEM, respectively). Cultures after TS treatment showed significant production of E2 (22086 ± 4650), and PG pretreatment was even more efficient (71920 ± 9090). Most efficient was the PG + TS pretreatment (198202 ± 40088) .
These data confirm that the E2 is produced in cultures pretreated with PG and TS. This explains that although E2 is essential for the effect of PG + TS combination, it is present because of the PG and TS conversion by vascular SMC. In vivo, the age advancement is associated with a diminution of circulating oestrogens. However, in post-menopausal women and in ageing men, the E2 is produced in a number of extragonadal sites, which include vascular SMC and numerous sites in the brain . Vascular SMC express aromatase and 17beta-hydroxysteroid dehydrogenase type I, which converts estrone, the major product of aromatase, to a potent oestrogen E2 . Therefore, like in vitro, the PG and TS treatment in vivo may also be sufficient to induce transdifferentiation of vascular SMC into vascular stem cells and neural/neuronal cells, possibly just in the sites, where this is needed.
Interestingly, the high circulating levels of PG + TS combination are unlikely to occur either in females or in males and E2 + PG or E2 + TS combinations were not capable to transdifferentiate ovarian epithelial cells or vascular SMC into stem and neuronal cells. Our observations [54, 56] suggest that either E2 + PG + TS treatment or PG + TS treatment and E2 produced by targeted cells expressing aromatase+17beta-hydroxysteroid dehydrogenase type I could be effective.
Experimental studies indicate that large doses of PG for up to several days after injury limited CNS damage, reduced loss of neural tissue and improved functional recovery (reviewed in ). This is probably because of the combined effect of PG and locally produced E2 from PG. However, treatment with PG + TS with even more enhanced E2 production and transdifferentiation of vascular SMC into neural/neuronal cells (Fig. 15K–M, Q) could be substantially more efficient by causing regeneration of injured areas of CNS.
Vascular SMC stem cells
An ability of vascular SMC to transdifferentiate into neural and neuronal cells is of particular importance because, unlike ovarian cells, the vascular SMC accompany as pericytes all vessels, including CNS microvasculature in both females and males. In addition, a combination of all three sex steroids caused fast conversion of differentiated vascular SMC into stem type cells, and these cells differentiated back into mature vascular SMC. Vascular SMC, also known as pericytes, regulate endothelial cell properties and contribute to the stability and maintenance of blood vessels, and they have also been suggested to be pluripotential and serve as precursors for a variety of other cell types . They accompany capillaries and post-capillary venules  and function as mesenchymal stem cells in various tissues, including bone marrow stromal stem cells [208–211].
Pericytes represent perivascular niche of mesenchymal stem cells and reside in virtually all post-natal organs and tissues. They represent multipotent progenitor cells which can give rise to mesenchymal and epithelial cells in vitro and in vivo, and their differentiation potential is related to the tissue of origin. All mesenchymal stem cell populations derived from distinct tissues, including bone marrow, spleen, muscle, aorta, vena cava, kidney, lung, liver, brain and thymus, showed expression of alpha SMA suggesting their relationship to pericytes .
Altogether, systemic or local (spinal cord injury) treatment with sex steroid combinations could stimulate transdifferentiation of resident vascular SMC to improve neurodegenerative, traumatic and ischaemic neurological disorders and vascular diseases without utilization of in vitro developed stem cells.
However, the morphostatic maintenance of tissues declines with age, particularly because of the degenerative changes in the immune system. This is accompanied by an increased incidence of tissue dysfunctions (type 2 diabetes mellitus, neurodegenerative diseases). From this point of view, one may expect that the regenerative medicine is more likely to be successful in acute/traumatic rather than in chronic tissue disorders , unless we will be able to manage the attenuated or lost stop effect of MDC in certain tissues of ageing individuals (see Fig. 6F and orange arrowheads, Fig. 7).
In conclusion, available data indicate that morphostasis is a complex event requiring renewal from stem cells, preservation of tissue-specific cells in a proper differentiated state and regulation of tissue quantity. This can be executed by TCS consisting of immune system–related components, vascular pericytes and autonomic innervation. Morphostasis is established epigenetically, during morphogenetic IA, i.e. during the critical developmental period. Subsequently, the tissues are maintained in a state of differentiation reached during the adaptation by a ‘stop effect’ of MDC. Alteration of tissue differentiation during the critical developmental period causes persistent alteration of tissue function. Morphostasis is altered with age advancement, because of the degenerative changes of the immune system. That is why the ageing of individuals and increased incidence of neoplasia and degenerative diseases occur. The TCS causes augmentation of cancer growth by a stimulation of ‘regeneration’ of depleted functional cells because of the hybridization of tumour stem cells with normal tissue cells. Chronic degenerative diseases may be less responsive to regenerative medicine because of the defects in the TCS-mediated morphostasis, such as an alteration of the ‘stop effect’ of MDC. The ability to manage stop effect of MDC may, in reality, provide treatment for early post-natal tissue disorders, such as muscular dystrophy and type 1 diabetes, and delay physical, hormonal and mental ageing of an individual, including delay of menopause in women, sexual dysfunction in ageing men, and age-associated high blood pressure, type 2 diabetes and autoimmune diseases. Degenerative diseases may be associated with tissue cell apoptosis. If we are able to change morphostasis, we may disrupt apoptotic process of the cell. It may also improve chances of regenerative medicine to treat chronic degenerative diseases.