Dissertation committee:
Μιχαήλ Κουτσιλιέρης, Καθηγητής, Ιατρική, ΕΚΠΑ
Αθανάσιος Τζιούφας, Καθηγητής, Ιατρική, ΕΚΠΑ
Γεώργιος Κόλλιας, Καθηγητής, Ιατρική, ΕΚΠΑ
Μιχαήλ Βουλγαρέλης, Αναπληρωτής Καθηγητής, Ιατρική, ΕΚΠΑ
Ευσταθία Καψογεώργου, Επίκουρη Καθηγήτρια, Ιατρική, ΕΚΠΑ
Κλειώ Μαυραγάνη, Επίκουρη Καθηγήτρια, Ιατρική, ΕΚΠΑ
Αναστάσιος Φιλίππου, Επίκουρος Καθηγητής, Ιατρική, ΕΚΠΑ
Summary:
Mesenchymal and epithelial cells consist two of the main cell types in vertebrates such as humans. Mesenchymal cells constitute a quite heterogeneous population which consists of cells that are more (e.g. osteoblasts, chondroblasts) or less (e.g. mesenchymal stem cells) differentiated, with varying capabilities for differentiation in other cell types. In literature, the term “mesenchymal cells” usually refers to cells like mesenchymal stem cells, fibroblasts and myofibroblasts, which are not terminally differentiated/specialized but exhibit some plasticity both regarding their capabilities for further differentiation as a response to different stimuli and regarding to their functions.
It is now well known that mesenchymal cells play a key role in functions that are essential both for the preservation of homeostasis (e.g. during wound healing) and in the context of a disease (e.g. inflammation or cancer). However, the great heterogeneity of mesenchymal cell types that is observed among different tissues as well as inside the same tissue makes the detection and clarification of the specific roles that each one of these cells plays in health and disease a quite difficult task. In parallel though, there is increased interest in targeting mesenchymal cells as a treatment for different diseases for which the conventional treatments have failed or are ineffective. It is therefore crucial to study and understand deeper the different roles that mesenchymal cells play according to the context and the conditions in which they act.
This was also the aim of this project which consists of two parts. On the one hand, to study the role of mesenchymal cells in somatic growth, and on the other hand to study their role in cancer and specifically in colorectal cancer.
Regarding somatic growth, it is known (and has been described in detail) that it is mainly controlled by the axis growth hormone releasing hormone – growth hormone – insulin-like growth factor 1 (GHRH - GH - Igf1). Additionally, it is known that in the level of a tissue or a cell, mesenchymal cells play an important role in supporting the survival and proliferation of many other cell types (like the epithelial cells) as well as in the generation of the whole tissue since they provide a scaffold on which the tissue will be formed. However their role in somatic growth has not been studied in literature.
The unexpected expression of the human growth hormone gene in the collagen VI expressing mesenchymal cells of our new transgenic mice (called TgC6hp55) which exhibited enhanced somatic growth gave us the opportunity to study the role of tissuespecific expression of the growth hormone from the mesenchyme. More specifically, the human growth hormone was found to be expressed in many different tissues like brain, heart and skeletal muscle, as well as (although in smaller quantities) in lungs, kidneys, spleen, adipose tissue (both white and brown), intestine, and was also detected in the circulation of the transgenic mice inducing a variety of phenotypic changes in them.
Initially, the transgenic mice exhibited enhanced somatic growth (increased body length and body weight) starting already from the 3rd week of age. The increased body weight could be attributed to the proportionally increased weight of their internal organs and lean mass and also to a disproportional increase in white adipose tissue which points towards a mild obesity-like phenotype. In the same time, probably due to the expression of the human growth hormone in their brain (mainly in their pituitary and hypothalamus), the mice showed hyperphagia as well as increased metabolism and enhanced mobilization of carbohydrates as a source of energy during night (when nocturnal animals have increased energy demands), with the daily energy surplus being positive and increased in comparison to their wt controls. This observation could explain their obesity-like phenotype we described before. Additionally, regarding their metabolic profile, the TgC6hp55 mice presented a phenotype resembling the initial stages of diabetes mellitus in humans since they were insulin resistant but remained normoglycemic due to the increased insulin production they had (hyperinsulinemia).
In accordance with the previous findings, histological examination of the transgenic mice revealed the progressive accumulation of lipids in their livers which developed in a moderate macrovesicular hepatic steatosis, as well as decreased storage of hepatic glycogen. These findings were followed by alterations in the expression of genes that regulate lipid, protein and carbohydrate metabolism in the liver. Moreover, the adipocytes of the white adipose tissue were also enlarged due to the increased accumulation and storage of lipids inside them. Finally, we detected alterations in a series of physiological parameters of the transgenic mice like biochemical parameters (serum albumin, total protein, triglycerides, iron etc.), cytological parameters (hematocrit, hemoglobin, white blood cells etc.) as well as in markers of bone metabolism (bone mineral density, alkaline phosphatase etc.).
The inhibition of human growth hormone activity through the pharmacological blockade of its receptor led to an inversion of the main phenotypic characteristics of the TgC6hp55 mice, thus confirming the etiological relation between the mesenchymally produced human growth hormone and the observed phenotype. The fact that the Igf-1 serum levels were unchanged in the TgC6hp55 mice despite the presence of the human growth hormone in their circulation leads to the conclusion that probably the main mechanism of action of this hormone in the TgC6hp55 mice was through its local expression by the mesenchyme of different organs and its subsequent paracrine actions in them, assisted by some additional systemic effects. Simultaneously, since the human growth hormone can activate also the prolactin receptor of the mouse, it is possible that some secondary features of the observed phenotype could be attributed to this mechanism.
As a conclusion, our study showed that the local expression of growth hormone from the collagen VI expressing mesenchymal cells can act both on a local and on a systemic level thus affecting not only somatic growth but also various other physiological parameters of the mouse.
On the other hand, regarding colorectal cancer, it is known that apart from the cancer cells per se, the tumor “microenvironment” plays also a very important role in the initiation and development of the disease. This microenvironment consists of different cell types like mesenchymal, endothelial and immune cells, which through the production of different molecules have the ability to affect the cancer cells and enhance tumorigenesis. For example, it has been shown that cyclooxygenase-2 (Cox-2) that is produced by tumors, through the production of prostaglandin E2 (PGE2), enhances colorectal carcinogenesis via many different mechanisms (enhancement of the proliferation and survival of cancer cells, avoidance of apoptosis, promotion of angiogenesis, avoidance of the immune surveillance etc.). Interestingly, more and more publications suggest as a source of this Cox-2 the mesenchymal cells that are located underneath the intestinal crypts. Therefore, the aim of our study was to examine the role of Cox-2 that is produced by the intestinal myofibroblasts in colorectal carcinogenesis.
For this reason, we used the CollagenVI-Cre mice in order to genetically delete, in a tissue-specific way, the Cox-2 from the intestinal myofibroblasts in a very common and well characterized mouse model of colorectal carcinogenesis, the APCmin/+ mouse. Normally these mice develop spontaneously dozens or hundreds of adenomas/tumors in their intestine and therefore have a life expectancy of up to 6 months. However, after the deletion of Cox-2 from their intestinal myofibroblasts (APCmin/+ColVICreCox2f/f mice), the number of the developed adenomas/tumors was decreased more than 50%. This reduction, despite not being accompanied by a reduction in tumor size, led to a decreased disease severity and an important elongation of the life duration of these mice. The fact that the APCmin/+ColVICreCox2f/f mice showed fewer microadenomas already from the initial stages of carcinogenesis, suggests that the Cox-2 coming from the intestinal myofibroblasts probably affects the initiation of carcinogenesis and not its development.
From a mechanistic point of view, mesenchymally produced Cox-2 seems to affect proliferation of the intestinal epithelial cells even in the seemingly “healthy” intestine without affecting their apoptosis. Additionally, it seems not to affect (neo)angiogenesis or immune infiltration of the tumor.
Finally, on a molecular level, we found that Cox-2 coming from the intestinal myofibroblasts affects the b-catenin pathway as well as a series of stem cell markers, thus enhancing intestinal carcinogenesis.
As a conclusion, our study confirmed in vivo the important role that the Cox-2 signaling in the intestinal myofibroblasts plays in the procedure of intestinal carcinogenesis, as well as the crosstalk that exists between mesenchyme and intestinal epithelium. However, further studies are needed to elucidate the exact molecular pathways that drive this disease. Moreover, we must be aware of one important drawback of our study. Genetic deletion of Cox-2 using the CollagenVI-Cre mice, on the one hand is not 100% efficient among CollagenVI+ intestinal myofibroblasts, and
on the other hand does not affect CollagenVI- intestinal myofibroblasts. Therefore, in the APCmin/+ColVICreCox2f/f mice there are still intestinal myofibroblasts that keep expressing Cox-2. Complete (100%) genetic deletion of Cox-2 form the CollagenVI intestinal myofibroblasts or even deletion from other types of intestinal myofibroblasts, could give us even better results and help enlighten thespecific roles that these cells play in intestinal carcinogenesis.In general, extra experiments that are already running in our lab or in our collaborator’s labs will be very helpful for the better understanding of the pathophysiology of colorectal cancer, so that we can develop more effective and targeted (with less side-effects) treatments as well as markers for better diagnosis and prognosis for patients with this devastating disease.
