A fish intestinal barrier model to assess the interaction with chemicals in vitro

The intestinal epithelium acts as vital gate keeper between the fish and its surrounding environment. A single layer of connected epithelial cells forms a selective barrier, which, among other essential functions, is important for nutrient uptake and defence against pathogens and putatively toxic chemicals. Despite its recognised significance, large knowledge gaps exist regarding the functionality of the fish intestine in general and the epithelium in specific, in large part due to the unavailability of suitable model systems. This thesis exploits a recently established in vitro fish intestinal cell barrier model from rainbow trout (Oncorhynchus mykiss), which is based on the so far only available fish intestinal epithelial cell line, RTgutGC. Motivated by the fact that the RTgutGC-based barrier model allows unprecedented opportunities for experimentation, this thesis engaged in embedding this cell line model into an in vitro testing strategy for assessing chemical transfer into fish. The focus was set on generally difficult-to-test chemicals, i.e. chemicals that are hydrophobic and/or volatile. For this purpose fragrances were selected, which combine chemicals in regular use with the described properties. Moreover, the response of the model epithelium to pathogenic stimulation was explored, in order to lay the groundwork for extending the in vitro testing strategy to different states of immune challenge in the future.

To initiate this research, fragrance cellular toxicity was assessed to derive non-toxic concentrations as prerequisite for transfer experiments and for in vitro to in vivo toxicity extrapolation of fish acute toxicity, which was both successfully achieved. Chemical transfer experiments were subsequently carried out with a set of fragrance chemicals using a purposefully designed transfer chamber, which reduces losses due to binding and evaporation of chemicals. The obtained transfer data were combined with in silico kinetic models to distinguish mechanisms of transfer and enable reliable result interpretation. Based on the mechanistic model, transfer processes were identified to involve intracellular accumulation and biotransformation, combined with paracellular transport and active transport processes. This outcome is contradictory to the current assumptions used in bioaccumulation models that chemical uptake is logKOW-dependent.

While chemical transfer assessment was conducted under optimal cellular conditions, the in vivo intestine is likely exposed to pathogens activating the immune system and potentially impairing barrier tightness. Indeed, also the RTgutGC cell barrier model was found to respond to bacterial and viral stimulation in a manner previously shown in vivo. Toll-like receptor-mediated regulation of cytokines and interferon were recorded within hours of exposure while long-term exposure resulted in an, albeit small, loss of epithelial tightness.

To conclude, this thesis comprises a detailed exploration of the RTgutGC-based fish intestinal cell model in terms of its physical and immunological barrier properties and demonstrates the value of this in vitro model for fish gut physiology, immunology and toxicology. This unique barrier model has the potential to successfully span mechanistic knowledge on gut functionality to practical application, such as fish feed development and chemical hazard assessment, while contributing to the refinement, reduction and even the potential replacement of animal testing

Tissue engineering complex and mechanically active microenvironments for studies in asthma and cancer

Alice Anna Tomei

The goals of tissue engineering include recapitulating specific tissue functions for regenerative medicine and developing in vitro models of human tissues to study human physiology and pathophysiology and for testing and screening drugs before expensive clinical trials. Success of these goals relies on the capability to recapitulate the complex interactions between cells and their microenvironments in vitro, which requires interdisciplinary approaches. Tissue engineering has evolved greatly in the past two decades, with notable progress in replacing function in tissues such as cartilage, bone, arteries, skin, and pancreatic islets. What is lacking, however, are examples of, and design principles for, engineering complex microenvironments to understand disease progression. The two particular examples addressed in this thesis are (1) airway remodeling in asthma and (2) tumor invasion and metastasis in cancer. These diseases involve complex interactions between different cell types as well as extracellular matrix remodeling, which, together with the mechanical environment that affects cell-cell and cell-matrix interactions and that can change in a diseased state (or in some cases induce the disease state), collectively cause the pathology. Such disease states are difficult to study in vivo because (a) one cannot easily tease out mechanisms or have much control over these complex interactions in an animal model, and (b) there are many differences between animal and human cells. The aim of this thesis was to engineer mechanically dynamic microenvironments for the study of two disease processes: viral infection in asthmatic remodeled airways and cancer metastasis to the lymph node. These involved multidisciplinary approaches and drew tools and methods from mechanobiology, materials science, cell biology, and molecular biology. First, we developed and characterized a relevant 3D in vitro model of the bronchial mucosa with a functional epithelial barrier, including improved epithelial differentiation, ciliogenesis, and mucociliary function and showed for the first time that dynamic compression, such as that seen in asthmatic bronchoconstriction, enhances viral infection of epithelial and subepithelial cells using lentiviral technology and may help explaining why asthmatic patients are more susceptible to viral infection. However, the strain device had deficiencies in reproducibility, ease of use, time to assemble, and most importantly, homogeneous distribution of strain (because collagen and other natural extracellular matrices are mechanically weak since they have >99% water and cannot retain residual strain). To address these problems, we engineered a completely different dynamic strain device, which includes an elastic porous superstructure to contain the collagen and which can modulate stress/strain to the entire culture system. We characterized this device and showed that we could impose dynamic strains to 3D soft, hydrated cultures over relatively long periods of time and with variable amplitude and frequency in a uniform way. The device has the potential to be used in a wide spectrum of applications, including the possibility to study the correlation between mechanostimulation and metastatic potential of tumor cells in the context of lung metastasis of many cancer types. Finally, we used this material as a supporting structure for tissue-generated stress such as that generated by myofibroblasts, and in particular the ones that build up the lymph node stroma in the paracortex. The importance of these cells relies on their ability to form conduits for chemokine transport to high endothelial venules, where lymphocytes enter the node, the ability to form a platform for directing immune cell migration and interactions, and the ability to secrete two chemokines that are important to regulate lymphocyte adhesion and migration: CCL19 and CCL21. Specifically, we found that lymph node stromal cells use this supporting composite to remodel and reorganize their surrounding matrix into physiologically relevant bundles. By giving these cells the relevant stimuli (matrix microenvironment and flow), we could recapitulate certain features of the lymph node microenvironment, including a physiological stromal structure and chemokine profile, that may be critical for understanding immune and cancer cell interactions there. Using this model we demonstrated for the first time that slow flow through the lymph node stroma can itself drive cell and stromal organization and upregulates CCL21 production. Finally, we incorporated this engineered lymph node model into a larger model of the tumor-lymphatic-lymph node pathway to model the relative modulation of tumor growth, migration within a lymphatic "channel" towards the lymph node, and behavior within this lymph node environment. We used human lymphatic endothelial cells and a mass of transduced tumor cells with engineered chemokine profiles. We showed how CCR7 signaling correlates with the ability of tumor cells to metastasize to the lymph node and to interact with lymph node stromal cells once they reach the node. Our findings demonstrate how a multidisciplinary tissue engineering approach allows creating complex microenvironments in vitro that include both physical and chemical cues and allows recapitulating tissue function in vitro and investigating dynamic pathophysiological diseases like asthma and cancer metastasis. It combines both the development of these model systems and some examples of their usefulness to gain insight into asthma pathophysiology and cancer biology.

EPFL2008

Biophysical Regulation of Lymphatic Vessel Function

Dimana Miteva

Lymphatic capillaries collect interstitial fluid and dendritic cells from the periphery and deliver them to the lymph nodes for immune surveillance and tolerance maintenance. Upon injury and inflammation, lymphatic drainage can increase rapidly, and thus the lymphatic capillaries experience a broad range of fluid stresses and need to respond rapidly to such changes. Our understanding of their functional biology, adaptive ability and response to pathological changes in the biophysical environment is still very limited. This thesis focuses on how the lymphatic capillary endothelium senses and responds to changes of its biophysical environment with respect to its function, including fluid and cell transport, in the context of physiological and pathophysiological conditions. Using in vitro and in vivo models, we demonstrate that lymphatic endothelium is exquisitely sensitive to transmural flow, modulating both fluid and cell transport functions even in the absence of inflammatory cytokines. Flow increased lymphatic permeability and dendritic cell (DC) transmigration, which were coincident with changes in gene and protein expression of factor known to be involved in these, including chemoattractant chemokines, adhesion molecules and junctional proteins. These data provide the first evidence that lymphatic endothelium can regulate its transport functions in response to flow cues. They also show for the first time that transmural flow can, in the absence of inflammatory cues like TNF-a, drive expression of DC adhesion molecules like ICAM-1 and chemokines like CCL21 by lymphatic endothelium. Based on these findings we suggest that lymphatic flow is an important mediator of lymphatic function, particularly with respect to DC recruitment and trafficking to the lymph node. Furthermore, the response of the lymphatic endothelium to flow may be specific to the type of low, as lymphatic endothelial cells (LECs) might react differently to luminal shear flow, at physiological and pathophysiological levels, compared to transmural flow. To address this we used laminar shear stress chambers to expose only the apical surface of the lymphatic endothelial cells (LECs) to shear stress and demonstrated that lymphatic endothelium is indeed sensitive to luminal flow, fine-tuning its expression of adhesion molecules to modulate DC adhesion according to the shear stress. Specifically, at extremely low shear, adhesion was enhanced, unlike at higher shear, adhesion was downregulated. This is consistent with the notion that DC adhesion to LECs is only needed for entry but not transport in conducting vessels, since only the absorbing capillaries have low shear stresses. The interactions between DCs and LECs were strongly mediated by CCL21 and its receptor CCR7. These findings suggest that luminal shear stress acts as an active mechanosignal in LECs to potentially facilitate the traffic of immune cells inside of the lymphatic vessel to reach the lymph node and trigger an immune response. Additionally, we demonstrated that lymphatic vessels respond differentially to immunogenic and tolerogenic inflammatory stimulus. We used in vitro models to evaluate lymphatic uptake and lymphatic participation in dendritic cell transmigration in the presence of various inflammatory stimuli. We found that lymphatic endothelium exposed to TNF-a or complement activation were more conductive to dendritic cell transmigration, while LPS and TGF-β had opposite effect, affecting DCs to decrease DC transmigration. We further presented some insights in similarities in tumor and dendritic cells intravasation into inflamed lymphatic capillaries. These findings imply that lymphatic activation and participation in immune cells transport is not a universal phenomenon, but is strongly dependent on the nature of the inflammation. In conclusion, this work highlights the critical role of biophysical environment in lymphatic endothelial function, and how exquisitely sensitive the lymphatic capillaries are to their extracellular environment and how they use multiple cues to sense inflammation and danger, modulating their transport functions accordingly to regulate the delivery of antigens and immune cells to the lymph nodes.

EPFL2010

The contribution of the host stress response to the pathogenesis of Pseudomonas entomophila in the Drosophila intestine

Sveta Chakrabarti

Recently, several studies done in Drosophila have revealed that efficient and rapid recovery from bacterial infection in the gut is possible only when bacterial clearance by the immune system is coordinated with repair through renewal of the epithelium damaged by infection. In this thesis, I have analyzed how the entomopathogenic bacterium Pseudomonas entomophila affects the immune response and epithelium renewal. A microarray analysis first showed that P. entomophila is recognized by the Drosophila immune system since ingestion of this bacterium stimulated the expression of antimicrobial peptide genes by the Imd pathway. In addition, stress and damage related pathways were strongly induced by P. entomophila, which correlate with its capacity to inflict severe damage. While antibacterial genes were induced in the gut following infection with P. entomophila, the immune response was not productive due to a general inhibition of translation that affected all the newly synthesized transcripts. Furthermore, the blockage of translation also inhibited the repair program by blocking the production of JAK-STAT ligands (upd3, upd2) and growth factors, which stimulate the intestinal stem cells proliferation and differentiation. I next analyzed the pathways that link cellular damage to reduction of translation. Using a genetic approach, I showed that inhibition of translation was induced by two signaling pathways: i) the phosphorylation of elongation initiation factor 2α (eIF2α) by the stress kinase GCN2 and ii) the inhibition of the TOR pathway by the AMPK kinase. Both kinases sense metabolic deprivation, suggesting that cellular damages induced by P. entomophila induce a state of “starvation”. Inhibition of translation is usually an adaptive cellular response to adjust the metabolism to the energy status of the cells. The observation that GCN2-deficient flies survived better than wild-type flies to P. entomophila indicates that pathogenesis is linked to an over-activation of stress pathways that usually help to endure the consequence of an infection. In this in vivo model of infection, I also showed that the reduction of translation was a consequence of cellular damage to the intestine caused by host-derived reactive oxygen species (through the activity of Duox) and by the direct action of a pore-forming toxin produced by the pathogen. As a consequence of this translational arrest, flies succumbed P. entomophila infection because they were unable to repair gut damage. Finally, I showed that inhibition of translation also had a strong influence on innate immune responses observed upon P. entomophila infection. The specific activation of a systemic immune response (antimicrobial peptides produced by the fat body) observed upon P. entomophila infection could be recapitulated by feeding flies with a non-lethal pathogen and an inhibitor of translation. Hence, I showed translation inhibition could be an important feature that shapes the immune response. The p38 MAPK family is involved in stress and immunity in both mammals and Drosophila. In Drosophila, three p38-MAPK-encoding genes, p38a, p38b and p38c, have been identified but their function in the gut immune response is poorly characterized. In the second part of my thesis, I have analyzed the role of these three p38 MAPKs in Drosophila intestinal immunity, especially p38c, which is strongly enriched in the gut. I first confirmed that the three p38 are involved kinases are involved in the defense to oral infection with pathogenic (but non-lethal) bacteria such as Erwinia carotovora 15. Surprisingly, p38c mutant flies were more resistant to P. entomophila and did not show the reduction in translation observed in wild-type flies. Interestingly, the transcription of the ROS producing enzyme Duox was reduced in p38c raising the hypothesis that p38 contributes to P. entomophila pathogenicity by activating Duox. Furthermore, I have obtained preliminary data indicating that p38c and the transcription factor ATF-3 act in the same pathway contributing the host immune response and lipid metabolism. Together, my thesis reveals the complex cross-talks between stress and immune pathways during microbial infection. While, it shows that stress pathways usually contribute to endure the damage caused by infection, excessive activation of these pathways can contribute to pathogenesis.

EPFL2013