In acute graft-versus-host disease (GVHD) alloreactive donor T cells selectively damage skin, liver, and the gastrointestinal tract while other organs are rarely affected. The mechanism of this selective target tissue infiltration is not well understood. We investigated the importance of alloantigen expression for the selective organ manifestation by examining spatiotemporal changes of cellular and molecular events after allogeneic hematopoietic cell transplantation (allo-HCT). To accomplish this we established a novel multicolor light sheet fluorescence microscopy (LSFM) approach for deciphering immune processes in large tissue specimens on a single-cell level in 3 dimensions. We combined and optimized protocols for antibody penetration, tissue clearing, and triple-color illumination to create a method for analyzing intact mouse and human tissues. This approach allowed us to successfully quantify changes in expression patterns of mucosal vascular addressin cell adhesion molecule–1 (MAdCAM-1) and T cell responses in Peyer’s patches following allo-HCT. In addition, we proofed that LSFM is suitable to map individual T cell subsets after HCT and detected rare cellular events. We employed this versatile technique to study the role of alloantigen expression for the selective organ manifestation after allo-HCT. Therefore, we used a T cell receptor (TCR) transgenic mouse model of GVHD that targets a single peptide antigen and thereby mimics a major histocompatibility complex (MHC)-matched single antigen mismatched (miHAg-mismatched) HCT. We transplanted TCR transgenic (OT-I) T cells into myeloablatively conditioned hosts that either express the peptide antigen ovalbumin ubiquitously (βa-Ova) or selectively in the pancreas (RIP-mOva), an organ that is normally not affected by acute GVHD. Of note, at day+6 after HCT we observed that OT-I T cell infiltration occurred in an alloantigen dependent manner. In βa-Ova recipients, where antigen was ubiquitously expressed, OT-I T cells infiltrated all organs and were not restricted to gastrointestinal tract, liver, and skin. In RIP-mOva recipients, where cognate antigen was only expressed in the pancreas, OT-I T cells selectively infiltrated this organ that is usually spared in acute GVHD. In conditioned RIP-mOva the transfer of 100 OT-I T cells sufficed to effectively infiltrate and destroy pancreatic islets resulting in 100% mortality. By employing intact tissue LSFM in RIP-mOva recipients, we identified very low numbers of initial islet infiltrating T cells on day+4 after HCT followed by a massive T cell migration to the pancreas within the following 24 hours. This suggested an effective mechanism of effector T cell recruitment to the tissue of alloantigen expression after initial antigen specific T cell encounter. In chimeras that either expressed the model antigen ovalbumin selectively in hematopoietic or in parenchymal cells only, transplanted OT-I T cells infiltrated target tissues irrespective of which compartment expressed the alloantigen. As IFN-γ could be detected in the serum of transplanted ovalbumin expressing recipients (βa-Ova, βa-Ova-chimeras and RIP-mOva) at day+6 after HCT, we hypothesized that this cytokine may be functionally involved in antigen specific OT-I T cell mediated pathology. In vitro activated OT-I T cells responded with the production of IFN-γ upon antigen re-encounter suggesting that IFN-γ might be relevant in the alloantigen dependent organ infiltration of antigen specific CD8+ T cell infiltration after HCT. Based on these data we propose that alloantigen expression plays an important role in organ specific T cell infiltration during acute GVHD and that initial alloreactive T cells recognizing the cognate antigen propagate a vicious cycle of enhanced T cell recruitment that subsequently culminates in the exacerbation of tissue restricted GVHD. / In der akuten Graft-Versus-Host Disease (GVHD) infiltrieren allogene Spender T Zellen Haut, Leber und den Magen-Darm-Trakt des Empfängers und attackieren das Gewebe. Andere Organe sind dagegen interessanterweise nur selten betroffen. Die Mechanismen dieser selektiven Organinfliltration sind bisher weitestgehend unbekannt. In meiner Dissertationsarbeit untersuchte ich den Einfluss der Alloantigenexpression auf die selektive Organmanifestation während der GVHD. Um komplexe Immunprozesse die nach allogener Stammzelltransplantation auftreten, besser zu verstehen, entwickelten wir eine Lichtblattmikroskopietechnik (LSFM) die zelluläre und molekulare Veränderungen im intakten Gewebe detektieren kann. Wir etablierten eine neuartigen mehrfarben LSFM-Methodik, die es ermöglicht, Immunprozesse in großen Gewebsstücken von Maus und Mensch in Einzelzellauflösung dreidimensional darzustellen. Dazu kombinierten und optimierten wir Protokolle, um eine Penetration von Antikörpern tief in das Gewebe sowie die Aufklärung und die dreifache Beleuchtung des Gewebes zu ermöglichen. Diese Methode erlaubte uns die erfolgreiche Quantifizierung der Proteinexpression des Adressins mucosal vascular addressin cell adhesion molecule–1 (MAdCAM-1) als auch die Quantifizierung der T Zell Antwort im intakten Peyer’s Plaque nach allogener hämatopoetischer Transplantation (HCT). Weiterhin konnten wir die Methode zur Untersuchung der Migration unterschiedlicher T Zell-Subpopulationen nach HCT erfolgreich einsetzen und konnten einzelne, organinfiltrierende Zellen detektierten und quantifizieren. Wir benutzten die LSFM Methode um den Einfluss der Alloantigenexpression auf die selektive Organmanifestation zu studieren. Dazu verwendeten wir ein Transplantationsmodell, in dem der Haupthistokompatibilitätskomplex übereinstimmt (MHC-matched) und eine Diskrepanz nur in einem einzelnen Peptid Antigen (miHAG-mismatch) zwischen Spender und Empfänger bestand. Wir transplantierten T Zell Rezeptor (TCR) transgene (OT-I) T Zellen in myeloablativ bestrahlte Empfänger, die das Peptidantigen Ovalbumin entweder in allen Geweben (βa-Ova) oder selektiv in der Bauchspeicheldrüse (RIP-mOva) exprimieren. Die Bauchspeicheldrüse ist ein Organ, das normalerweise nicht von der akuten GVHD betroffen ist. An Tag 6 nach allogener HCT waren alle Organe die das Alloantigen exprimieren auch von Spender T Zellen infiltriert. In myeloablativ bestrahlten RIP-mOva Empfängern reichten bereits 100 transferierte OT-I T Zellen aus, um Alloantigen-exprimierende pankreatische Inselzellen zu zerstören. Dies führte zu einer Mortalität von 100% der Empfänger und spricht für eine sehr effiziente Alloantigendetektion und Gewebsinfiltration durch die Spender T Zellen. Um die Kinetik der Organinfiltration der Spender T Zellen detailliert zu untersuchen, verwendeten wir die neue Lichtblattmikroskopietechnik, welche die Analyse intakter Organe ermöglicht. In RIP-mOva Empfängern identifizierten wir erste wenige Spender T Zellen im Pankreas an Tag 4 nach Transplantation, gefolgt von einer massiven Pankreasinfiltration durch Spender T Zellen innerhalb von 24 Stunden. Dies deutet auf eine gezielte Rekrutierung der Spender T Zellen nach erstem Antigenkontakt in das Gewebe mit Alloantigenexpression. Um zu untersuchen, ob die Alloantigenexpression vom parenchymalen Gewebe oder aber durch hämatopoetische Zellen zur spezifischen Organinfiltration führt, transplantierten wir OT-I T Zellen in chimäre Empfänger, in denen das Alloantigen entweder nur im Gewebsparenchym oder ausschließlich von hämatopoetischen Zellen exprimiert wird. An Tag 6 nach der allogenen HCT fanden wir Spender T Zellen in allen Geweben, unabhängig davon welches Empfängerzellkompartment das Alloantigen präsentierte. Wir detektierten hohe IFN-γ-Werte im Serum von Ovalbumin exprimierenden Empfänger (βa-Ova, βa- Ova-Chimären und RIP-mOva). Weiterhin fanden wir, dass nach erneutem Kontakt mit dem spezifischen Alloantigen, OT-I T Zellen die in vitro aktiviert wurden, IFN-γ produzierten. Wir schließen aus diesen Beobachtungen, dass für die antigenabhängige Gewebeinfiltration IFN-γ wichtig ist. Zusammenfassend postulieren wir, dass die Alloantigenexpression im Gewebe eine wichtige Rolle in der organspezifischen Infiltration durch Spender T Zellen spielt, und dass T Zellen die Alloantigen spezifisch erkennen, dafür verantwortlich sind, dass weitere Effektor-T Zellen in das Gewebe rekrutiert werden.

As graft source for lymphoma or leukemia treatment, hematopoietic stem and progenitor cells (HSPCs) have been the focus of translational medicine for decades. HSPCs are defined by their self-renewing capacity and their ability to give rise to all mature blood cells. They are found anchored to a specialized microenvironment in the bone marrow (BM) called the hematopoietic niche. HSPCs can be enriched by sorting them based on the presence of the surface antigen CD34 before clinical or tissue engineering use. As these cells represent a minority in most graft sources and the amount of applicable cells is limited, ex vivo expansion-cultures were established using cytokine cocktails or small molecules. However, in vitro culture of HSPCs as suspension-cultures result in heterogeneous cell populations with undefined cellular identities. In the BM niche, HSPCs are not exclusively maintained by cytokines but also by cell-matrix adhesions mediated by integrins (ITGs). Thus, β1 and β2 ITGs were found to promote initial contact of HSPCs with mesenchymal stromal cells (MSCs) and ITGβ3 expression was shown to be a marker for long-term repopulating HSPCs in vivo. Consequently, ex vivo remodeling of the BM niche using co-cultures of HSPCs and niche cells like MSCs came into spotlight and was proven to be a promising tool for stem cell expansion. However, in clinical and research applications, direct contact of two cell populations necessitates HSPC post-culture purification. To address these problems, we established a novel culture method for remodeling the BM extra cellular stroma in vitro wherein we used decellularized extracellular matrix (ECM) scaffolds derived from immortalized mesenchymal stromal cells (SCP-1). Such scaffolds were found to be highly reproducible and served as in vitro niche for HSPCs by being more effective for the expansion of CD34+ cells, compared to classical suspension cultures. ECMs were shown to consist of multiple proteins including fibronectins, collagens, and a major niche chemokine responsible for BM homing and retention of HSPCs in vivo, namely, stromal derived factor 1 (SDF-1). SDF-1 is known to be secreted by MSCs and is anchored to matrix proteins. This reveals that ECM scaffolds produced by SCP-1 cells are a naïve reconstructed microenvironment. When CD34+ cells were seeded, only around 20% of the cells adhered to the provided ECM scaffold. These cells recognized SDF-1 via C-X-C chemokine receptor type 4 (CXCR-4), as shown by laser scanning confocal microscopy. Thus, adhesive sides as they are present in the BM niche are provided. However, CD34+ cells isolated from G-CSF mobilized peripheral blood of healthy donors were found to be heterogenous with respect to adhesion capacity. Nonetheless, it was similar to HSPC co-cultures with SCP-1 cells as feeder layer. Therefore, we separated and analyzed two cell fractions, the adherent (AT-cells) and the non- adherent supernatant (SN-cells) cells. Other signals provided by the BM extracellular stroma to HSPCs are physical cues that control HSPC fate. HSPCs sense these physical features through focal contacts and accordingly remodel their morphological and biomechanical properties. Using real-time deformability cytometry (RT-DC) to uncover biomechanical phenotypes of freshly isolated HSPCs, SN-cells, AT-cells, and classical suspension cultured HSPCs in plastic culture dishes (PCD) were analyzed. We found freshly isolated cells to be less deformable and small. AT-cells displayed actin polymerization to stress fibers, and exhibited a stiffer mechanical phenotype compared to PCD-cultured or SN-cells. This might constitute the first hint of functional adaptation. Integrins are known to establish mechanosensing focal contacts. Thus, we analyzed ITG surface expression and identified ITGαIIb, ITGαV, and ITGβ3 to be enriched on AT-cells compared to freshly isolated cells or SN-cells. Active integrins need to form heterodimers consisting of one α- and one β subunit. Interestingly, the identified ITGs exclusively interact with each other to form RGD peptide receptors. RGD is a tripeptide consisting of the amino acids arginine, glycine, and aspartic acid and was identified as an adhesion sequence within fibronectin and other extracellular proteins. Consequently, we could confirm an important role for ITGαVβ3 in HSPC- ECM interaction with respect to adhesion and migration. However, we also identified ITGβ3 expression on a subset of CD34+ cells either freshly isolated or ECM cultured cells, as a marker for erythrocyte differentiation. These findings demonstrate that, in vitro, the ECM compartment acts as a regulator of HSPC fate and portray mechanical recognition as a potent driver of differentiation. In this context, targeted modulation of ECM scaffolds could enhance cell-ECM interactions and accelerate stem cell expansion or differentiation. These modulations could also provide further insights into HSPC-niche regulation. We demonstrate that ECMs derived from osteogenic differentiated SCP-1 cells increase HSPC expansion but do not lead to increased cell adhesion. As ECM adhesion preliminary alters HSPC function, we aimed at developing ECM scaffolds with increased adhesion capacity. Using lentiviral transduction, we generated a stable knock down of fibulin-1 in SCP-1 cells. Fibulin-1 is an ECM protein known to form anti-adhesion sites with fibronectin. However, we failed to increase adherent cell numbers or enhance HSPC expansion in the fibulin-1 knock down ECMs. Taken together, SCP-1 cell-derived ECM protein scaffolds provide an in vitro niche for HSPCs capable of stem cell expansion. Integrin mediated signaling altered the biomechanical and functional properties of HSPCs and hints at suspension cultures as being inappropriate to study the physiological aspects of HSPCs. Targeted modulation of ECM scaffolds could theoretically generate suitable ex vivo environments with the capacity to gain detailed insight into HSPC regulation within their niche. This will enhance the functionality of new biomaterials and will lead to improved regenerative therapies like BM transplantation.

As graft source for lymphoma or leukemia treatment, hematopoietic stem and progenitor cells (HSPCs) have been the focus of translational medicine for decades. HSPCs are defined by their self-renewing capacity and their ability to give rise to all mature blood cells. They are found anchored to a specialized microenvironment in the bone marrow (BM) called the hematopoietic niche. HSPCs can be enriched by sorting them based on the presence of the surface antigen CD34 before clinical or tissue engineering use. As these cells represent a minority in most graft sources and the amount of applicable cells is limited, ex vivo expansion-cultures were established using cytokine cocktails or small molecules. However, in vitro culture of HSPCs as suspension-cultures result in heterogeneous cell populations with undefined cellular identities. In the BM niche, HSPCs are not exclusively maintained by cytokines but also by cell-matrix adhesions mediated by integrins (ITGs). Thus, β1 and β2 ITGs were found to promote initial contact of HSPCs with mesenchymal stromal cells (MSCs) and ITGβ3 expression was shown to be a marker for long-term repopulating HSPCs in vivo. Consequently, ex vivo remodeling of the BM niche using co-cultures of HSPCs and niche cells like MSCs came into spotlight and was proven to be a promising tool for stem cell expansion. However, in clinical and research applications, direct contact of two cell populations necessitates HSPC post-culture purification. To address these problems, we established a novel culture method for remodeling the BM extra cellular stroma in vitro wherein we used decellularized extracellular matrix (ECM) scaffolds derived from immortalized mesenchymal stromal cells (SCP-1). Such scaffolds were found to be highly reproducible and served as in vitro niche for HSPCs by being more effective for the expansion of CD34+ cells, compared to classical suspension cultures. ECMs were shown to consist of multiple proteins including fibronectins, collagens, and a major niche chemokine responsible for BM homing and retention of HSPCs in vivo, namely, stromal derived factor 1 (SDF-1). SDF-1 is known to be secreted by MSCs and is anchored to matrix proteins. This reveals that ECM scaffolds produced by SCP-1 cells are a naïve reconstructed microenvironment. When CD34+ cells were seeded, only around 20% of the cells adhered to the provided ECM scaffold. These cells recognized SDF-1 via C-X-C chemokine receptor type 4 (CXCR-4), as shown by laser scanning confocal microscopy. Thus, adhesive sides as they are present in the BM niche are provided. However, CD34+ cells isolated from G-CSF mobilized peripheral blood of healthy donors were found to be heterogenous with respect to adhesion capacity. Nonetheless, it was similar to HSPC co-cultures with SCP-1 cells as feeder layer. Therefore, we separated and analyzed two cell fractions, the adherent (AT-cells) and the non- adherent supernatant (SN-cells) cells. Other signals provided by the BM extracellular stroma to HSPCs are physical cues that control HSPC fate. HSPCs sense these physical features through focal contacts and accordingly remodel their morphological and biomechanical properties. Using real-time deformability cytometry (RT-DC) to uncover biomechanical phenotypes of freshly isolated HSPCs, SN-cells, AT-cells, and classical suspension cultured HSPCs in plastic culture dishes (PCD) were analyzed. We found freshly isolated cells to be less deformable and small. AT-cells displayed actin polymerization to stress fibers, and exhibited a stiffer mechanical phenotype compared to PCD-cultured or SN-cells. This might constitute the first hint of functional adaptation. Integrins are known to establish mechanosensing focal contacts. Thus, we analyzed ITG surface expression and identified ITGαIIb, ITGαV, and ITGβ3 to be enriched on AT-cells compared to freshly isolated cells or SN-cells. Active integrins need to form heterodimers consisting of one α- and one β subunit. Interestingly, the identified ITGs exclusively interact with each other to form RGD peptide receptors. RGD is a tripeptide consisting of the amino acids arginine, glycine, and aspartic acid and was identified as an adhesion sequence within fibronectin and other extracellular proteins. Consequently, we could confirm an important role for ITGαVβ3 in HSPC- ECM interaction with respect to adhesion and migration. However, we also identified ITGβ3 expression on a subset of CD34+ cells either freshly isolated or ECM cultured cells, as a marker for erythrocyte differentiation. These findings demonstrate that, in vitro, the ECM compartment acts as a regulator of HSPC fate and portray mechanical recognition as a potent driver of differentiation. In this context, targeted modulation of ECM scaffolds could enhance cell-ECM interactions and accelerate stem cell expansion or differentiation. These modulations could also provide further insights into HSPC-niche regulation. We demonstrate that ECMs derived from osteogenic differentiated SCP-1 cells increase HSPC expansion but do not lead to increased cell adhesion. As ECM adhesion preliminary alters HSPC function, we aimed at developing ECM scaffolds with increased adhesion capacity. Using lentiviral transduction, we generated a stable knock down of fibulin-1 in SCP-1 cells. Fibulin-1 is an ECM protein known to form anti-adhesion sites with fibronectin. However, we failed to increase adherent cell numbers or enhance HSPC expansion in the fibulin-1 knock down ECMs. Taken together, SCP-1 cell-derived ECM protein scaffolds provide an in vitro niche for HSPCs capable of stem cell expansion. Integrin mediated signaling altered the biomechanical and functional properties of HSPCs and hints at suspension cultures as being inappropriate to study the physiological aspects of HSPCs. Targeted modulation of ECM scaffolds could theoretically generate suitable ex vivo environments with the capacity to gain detailed insight into HSPC regulation within their niche. This will enhance the functionality of new biomaterials and will lead to improved regenerative therapies like BM transplantation.:List of contents I List of figures IV List of tables VI Abbreviations VII 1 Introduction 1 1.1 The stem cell microenvironment 3 1.1.1 The cellular endosteal bone marrow microenvironment 6 1.1.1.1 Mesenchymal stem/stromal cells 7 1.1.1.2 Hematopoietic stem and progenitor cells 8 1.1.2 Extracellular bone marrow microenvironment 10 1.1.2.1 Extracellular matrix 11 Chemokines and Cytokines 12 Cell adhesion to ECM 13 1.2 Native ex vivo ECM scaffolds 16 2 Aim of the study 19 3 Materials and methods 21 3.1 Materials 21 3.1.1 Chemicals and reagents 21 3.1.2 Kits 23 3.1.3 Media 24 3.1.4 Antibodies 24 3.1.5 Primers, sh-RNA sequences, and vectors 25 3.1.6 Equipment 26 3.1.7 Software 27 3.2 Methods 27 3.2.1 Cell preparation and culture 27 3.2.1.1 Mesenchymal stromal cells 27 3.2.1.2 Hematopoietic stem cells 28 3.2.1.3 Single cell picked clone 1 (SCP-1) cells 28 3.2.2 Generation of surface immobilized ECM preparations 29 3.2.2.1 Surface functionalization 29 3.2.2.2 ECM preparation 29 3.2.3 Flow cytometry and fluorescent activated cell sorting 30 3.2.4 Cell cycle analyses 30 3.2.5 Proliferation analyses 31 3.2.6 Colony forming unit cell assay (CFU-GEMM) 31 3.2.7 Migration assays 31 3.2.7.1 Transwell migration 31 3.2.7.2 Live cell migration 32 3.2.8 Confocal laser scanning microscopy 32 3.2.9 Real-time deformability cytometry (RT-DC) 32 3.2.10 Molecular biological methods 33 3.2.10.1 RNA isolation, reverse transcription, and PCR 33 3.2.10.2 Lentiviral shRNA transduction 34 3.2.10.3 Western blot 35 3.2.10.4 ELISA 36 3.2.11 Statistical analysis 37 4 Results 38 4.1 Extracellular matrix scaffolds for HSPCs 38 4.1.1 ECM properties 39 4.1.2 HSPC survival in ECM and PCD cultures 40 4.1.3 HSPC expansion in ECM and PCD cultures 41 4.2 HSPC morphological and mechanical adaptation to ECM 44 4.2.1 Actin polymerization and polarization 45 4.2.2 Biomechanical phenotype 46 4.3 Bioactive SDF-1 is incorporated in ECM scaffolds 49 4.3.1 CXCR4 polarization towards ECM 50 4.4 HSPC integrin expression and migration 52 4.4.1 Integrin surface expression on HSPC subsets 52 4.4.2 Focal contact formation 53 4.4.3 Integrin activation via ECM adhesion 55 4.4.4 Clonogenicity of ECM cultured HSPCs 57 4.4.5 HSPC migration when attached to ECM scaffolds 60 4.4.5.1 Reduced migratory behavior via ITGαVβ3 inhibition 61 4.4.5.2 SDF-1 induces migration but not adhesion 64 4.5 Targeted modulation of ECM scaffolds 65 4.5.1 Fibulin-1 knock down in SCP-1 cells 66 4.5.2 HSPC support of fibulin-1 reduced ECM scaffolds 70 5 Discussion 73 5.1 SCP-1 cells as a source for ECM scaffold production 74 5.2 Cell adhesion and focal contact formation 75 5.3 HSPC multilineage potential 78 5.4 ECM scaffold modulation 79 6 Summary 83 7 Zusammenfassung 86 Bibliography 89 Danksagung 108 Anlagen 110 Erklärung zur Eröffnung des Promotionsverfahrens [Formblatt 1.2.1] 110 Erklärung zur Einhaltung rechtlicher Vorschriften [Formblatt 1.1] 110