M ICROGLIA AND MACROPHAGES - Investigation of murine and canine glia cell differentiation in vi

Microglia and macrophages are an essential part of the innate immune system, which also influence the adaptive immune system by e.g. lymphocyte activation or recruitment (SICA et al. 2008). Under physiological conditions microglia and macrophages are responsible for the detection of pathogens, phagocytosis of dying or dead cells as well as for the maintenance of tissue homeostasis (DAVID et al.

2011). Microglia represent the resident phagocytic cells of the CNS. They can be distinguished from activated microglia and monocytes by flow cytometric analysis showing low CD45 expression (DAVID et al. 2011). During inflammation, blood-born macrophages (monocytes) are able to leave the circulation and migrate into the CNS.

In mice inflammatory monocytes are characterized by high expression of chemokine receptor 2 (CCR2), Ly6 (GR1) and a low expression of chemokine receptor 1 (CX3CR1) while resting monocytes showed a contrary expression of these markers (GEISSMANN et al. 2003, MOSSER et al. 2008). Under inflammatory conditions blood-born macrophages and microglia showed a high CD45 expression and cannot be distinguished, neither by morphology nor by antigenic markers (DAVID et al.

2011). Therefore, under inflammatory conditions microglia and blood-born macrophages should be called microglia/macrophages or phagocytes (SHECHTER et al. 2013). Microglia/macrophages have been related to secondary tissue damage in CNS diseases, but they are also necessary for removing cellular debris and therefore exhibit protective/regenerative effects (GIULIAN et al. 1990, POPOVICH et al. 1999, YONG et al. 2009). The beneficial or detrimental behavior of these cells may be caused by different subsets of microglia/macrophages. Recently macrophages were classified into proinflammatory, classically activated cells called M1 and anti-inflammatory, alternatively activated cells called M2 (Table 3; GORDON 2003, KIGERL et al. 2009). M2 macrophages can further be subdivided into M2a, M2b and M2c. Similarly, a polarization of macrophages has been shown in non-CNS diseases like atherosclerosis or bacterial and parasitic infections (DAVID et al. 2011).

General introduction Chapter 2

Table 3: Classification, marker expression and function of M1/M2 cells (modified according to KIGERL et al. 2009, DAVID et al. 2011).

Phenotype Marker Function

CD16

- Killing of intracellular pathogens

- Acidification of phagosome and release of reactive oxygen intermediates and NO

M2a

- Immunity against parasites - Growth stimulation - Tissue repair - Collagen formation

- Recruitment of Th2 cells, basophils and eosinophils

- Pro- and anti-inflammatory function - B cell class switch and antibody production - Recruitment of regulatory T cell

M2c Arginase I

Fizz 1=resistin-like alpha; IL10^high=high expression of interleukin 10; IL12^low=low expression of interleukin 12; iNOS=inducible nitric oxide synthase; MHCII=major histocompatibility complex class II; NO=nitric oxide; SPHK1=sphingosine kinase 1; YM1=chitinase 3-like protein 3.

M2 microglia/macrophagesM1 microglia/macrophages

Differentiation of M1 macrophages is promoted by the influence of lipopolysaccharide (LPS) and/or interferon-γ (IFN-γ; KIGERL et al. 2009). M1 cells produce proinflammatory cytokines such as tumor necrosis factor (TNF), IL-1, IL-6, IL-12 (DAVID et al. 2011) which are necessary for host defense and tumor cell destruction.

However, an over-expression of these cytokines may also induce a severe collateral damage to unaffected, normal tissue (DING et al. 1988). On the other hand, an activation of macrophages under the influence of cytokines such as IL-4 or IL-13 induces the M2 phenotype (DAVID et al. 2011). This group is composed of a heterogeneous subset of macrophages which are described as anti-inflammatory cells expressing IL-10, TGF-β, CD206, arginase I and they down-regulate proinflammatory cytokines (SICA et al. 2006 and 2008, DAVID et al. 2011). It was shown, that M2 macrophages isolated from uninjured spinal cords maintain their phenotype when transplanted into intact spinal cords while transplantation into lacerated spinal cords induced a 20-40% reduction of the M2 phenotype (KIGERL et al. 2009). This observation strongly indicates, that the local environment (transplantation site) determines the phenotype/differentiation of (transplanted) microglia/macrophages. Therefore, a predominance of M1 cells at sites of CNS injury may be expected (KIGERL et al. 2009), whereas the main phenotype of microglia/macrophages in the normal CNS displays M2 polarity (PONOMAREV et al.

2007). Interestingly, a fully differentiated microglia/macrophage population may reversibly change its phenotype and function in response to the local microenvironment (STOUT et al. 2004, MOSSER et al. 2008) or may be influenced by the exogenous application of microglia/macrophage differentiation-inducing molecules like LPS, IFN-γ, IL-4, IL-13 (KIGERL et al. 2009, DAVID et al. 2011).

Chapter 3 Defining the morphological phenotype: 2',3'-cyclic nucleotide 3'-phosphodiesterase (CNPase) is a novel marker for in situ detection of canine but not rat olfactory ensheathing cells

M. Omar, P. Bock, R. Kreutzer, S. Ziege, I. Imbschweiler, F. Hansmann, C.T. Peck, W. Baumgärtner, K. Wewetzer

Abstract

Olfactory ensheathing cells (OECs) are the non-myelinating glial cells of the olfactory nerves and bulb. The fragmentary characterization of OECs in situ during normal development may be due to their small size requiring intricate ultrastructural analysis and to the fact that available markers for in situ detection are either expressed only by OEC subpopulations or lost during development. In the present study, we searched for markers with stable expression in OECs and investigated the spatiotemporal distribution of CNPase, an early oligodendrocyte/Schwann cell marker, in comparison with the prototype marker p75(NTR). Anti-CNPase antibodies labeled canine but not rat OECs in situ, while SCs and oligodendrocytes were positive in both species. CNPase immunoreactivity in the dog was confined to all OECs throughout the postnatal development and associated with the entire cell body, including its finest processes, while p75(NTR) was mainly detected in perineural cells and only in some neonatal OECs. Adult olfactory bulb slices displayed CNPase expression after 4 and 10 days, while p75(NTR) was detectable only after 10 days in vitro. Finally, treatment of purified adult canine OECs with fibroblast growth factor-2 significantly reduced CNPase expression at the protein and mRNA level. Taken together, we conclude that CNPase but not p75(NTR) is a stable marker suitable for in situ visualization of OECs that will facilitate their light-microscopic characterization and challenge our general view of OEC marker expression in situ. The fact that canine but not rat OECs expressed CNPase supports the idea that glia from large animals differs substantially from rodents.

Cell Tissue Res. 2011: 344(3):391-495

Chapter 4 Highly malignant behavior of a murine oligodendrocyte precursor cell line following transplantation into the demyelinated and non-demyelinated central nervous system

F. Hansmann, K. Pringproa, R. Ulrich, Y. Sun, V. Herder, M. Kreutzer, W. Baumgärtner, K. Wewetzer Abstract

Understanding the basic mechanisms that control CNS remyelination is of direct clinical relevance. Suitable model systems include the analysis of naturally occurring and genetically-generated mouse mutants and the transplantation of oligodendrocyte precursor cells (OPCs) following experimental demyelination. However, aforementioned studies were exclusively carried out in rats and little is known about the in vivo behavior of transplanted murine OPCs. Therefore in the present study, we (i) established a model of ethidium bromide-induced demyelination of the caudal cerebellar peduncle (CCP) in the adult mouse and (ii) studied the distribution and marker expression of the murine OPC line BO-1 expressing the enhanced green fluorescent protein (eGFP) 10 and 17 days after stereotaxic implantation. Injection of ethidium bromide (0.025%) in the CCP resulted in a severe loss of myelin, marked astrogliosis and mild to moderate axonal alterations. Transplanted cells formed an invasive and liquorogenic metastasizing tumor, classified as murine giant cell glioblastoma. Transplanted BO-1 cells displayed substantially reduced CNPase expression as compared to their in vitro phenotype, low levels of MBP and GFAP, prominent upregulation of NG2, PDGFRα, nuclear p53, and an unaltered expression of signal transducer and activator of transcription (STAT)-3. Summarized environmental signaling in the brain stem was not sufficient to trigger oligodendrocytic differentiation of BO-1 cells and seemed to block CNPase expression. Moreover, the lack of the remyelinating capacity was associated with tumor formation indicating that BO-1 cells may serve as a versatile experimental model to study tumorigenesis of glial tumors.

Cell Transplant. 2012: 21(6):1161-1175

Chapter 5 Theiler’s murine encephalomyelitis virus induced phenotype switch of microglia in vitro

I. Gerhauser, F. Hansmann, C. Puff, J. Kumnok, D. Schaudien, K. Wewetzer, W. Baumgärtner Abstract

The present in vitro study aimed to define the involvement of astrocytes and microglia in the initial inflammatory response of Theiler's murine encephalomyelitis (TME), a virus-induced mouse model of multiple sclerosis, and whether intralesional microglia exert pro- (M1) or anti-inflammatory (M2) effects following TME virus (TMEV) infection. Therefore astrocytes and microglia were purified from neonatal murine brains and inoculated either with TMEV or mock-solution. Gene expression of IL-1, IL-2, IL-10, IL-12, TNF, TNF receptors (TNFR1, TNFR2), TGFβ1, IFNγ and transcription factors NF-κB (p50, p65) and AP-1 (c-jun, c-fos) were quantified using RT-qPCR at 6, 48, and 240 h post infection (hpi). In addition, IL-1, IL-10, IL-12, TNF and TGFβ1 mRNA transcripts were investigated at 168 hpi in TMEV- and mock-infected SJL/J mice. Overall in vitro astrocytes showed a significant higher amount of viral RNA compared to microglia. In addition, TMEV-infected astrocytes showed higher numbers of IL-1, IL-12 and TNF transcripts at 48 hpi. In microglia high IL-10 and low IL-12 mRNA levels were detected at 48 hpi, while the opposite was the case at 240 hpi. In addition, TNF mRNA was increased in microglia at 240 hpi. In addition, the observed up-regulation of IL-1, IL-12 and IL-10 in the early phase of TME in vivo substantiates the relevance of these cytokines during the disease induction.

Summarized data indicate that TMEV infection of microglia induces a switch from the anti-inflammatory (M2) during the early phase to the pro-inflammatory (M1) phenotype in the later phase of the infection. The simultaneous expression of TNF and its receptors by both cell types might generate autocrine feedback loops possibly associated with pro-inflammatory actions of astrocytes via TNFR1.

J. Neuroimmunol. 2012: 252(1-2):49-55 www.elsevier.com/locate/jneuroim DOI: 10.1016/j.jneuroim.2012.07.018

Chapter 6 Discussion and Conclusions

Inflammation, demyelination, neuronal degeneration, axonal loss and an insufficient or exhausted regenerative capacity represent the hallmarks of demyelinating diseases like MS, canine distemper encephalitis, Theiler’s murine encephalomyelitis or spinal cord injury (COMPSTON et al. 2008, BEINEKE et al. 2009, HANSMANN et al. 2012a, BOCK et al. 2013, ZHANG et al. 2013). For none of these diseases a sufficient therapy leading to a “restitutio ad integrum” is available. Possible therapeutic options are exogenous replacement of damaged cells/tissue or the creation of a beneficial environment at the lesion site favoring endogenous regeneration (KEIRSTEAD et al. 1999, JUNG et al. 2009, LU et al. 2010, DARLINGTON et al. 2011, GRANGER et al. 2012). In this context, the transplantation of non-myelinating cells in order to stimulate remyelination by either modulation of the immune system and/or enhancement of endogenous repair processes represents a new, very promising therapeutic approach (KANG et al.

2012b, WANG et al. 2012). Recently, several studies investigated the therapeutic efficiency of mesenchymal and neural stem cells (UCCELLI et al. 2011, AL JUMAH et al. 2012, KANG et al. 2012a, RYU et al. 2012, SHER et al. 2012, UCHIDA et al.

2012). MSCs have been shown to exert neuroprotective effects by induction of neurotrophic factors and modulation of the immune responses (KANG et al. 2012b, WANG et al. 2012). In addition, MSCs are suggested to be able to replace damaged cells by a process called trans-differentiation (EDAMURA et al. 2012).

If the goal is to repair myelin by transplantation of exogenous cells the following issues should be addressed: Firstly, used cell type and their number and secondly route and time point of application. The selection of appropriate criteria largely depends on the pathogenesis of the respective disease and on the planned time point for the intervention. Therefore, a promising opportunity is the selection of cells that have been shown to produce myelin in vitro and/or in vivo. Most attention has been given to OPCs, oligodendrocytes, olfactory ensheathing cells as well as SCPs and SCs (FRANKLIN et al. 1996b, FRANKLIN 2002a, HALFPENNY et al. 2002,

Discussion and Conclusions Chapter 6

2005, RAISMAN et al. 2007, GRANGER et al. 2012). Each cell type has specific advantages and disadvantages. However, currently no cell type has evolved as universal method of choice and further studies are needed.

Characterization of canine olfactory ensheathing cells as possible candidates for transplantation

Olfactory ensheathing cells, which are non-myelinating glial cells of the olfactory nerve and bulb, represent a suitable cell type for a regenerative approach (FRANKLIN et al. 1996b, SU et al. 2010). A major obstacle for the identification and purification of OECs is their different antigen expression depending on the developmental stage, localization (in vivo, in culture) and culture conditions (WEWETZER et al. 2011). In the present study canine glial cells have been investigated, because the dog represents a suitable translational animal model and canine OECs share many similarities with human OECs (JEFFERY et al. 2005, JEFFERY et al. 2006, WEWETZER et al. 2011, GRANGER et al. 2012). At present, no stable marker for the in vivo identification of OECs is known. Rat OECs were detected using anti-p75^NTR antibodies, however a major disadvantage of this marker is its down-regulation during post-natal development (GONG et al. 1994, FRANCESCHINI et al. 1996). The present study showed that p75^NTR is mainly expressed by perineural cells and only few neonatal OECs in the dog. This leads to the conclusion that p75^NTR is not suitable for the in vivo identification of canine OECs.

CNPase represents a well-established marker for the detection of oligodendrocytes and SCs of many species including rats and dogs (MIRON et al. 2011). In the present study, anti-CNPase antibodies labeled the entire cell body of all canine OECs including their processes during the postnatal development. It seems unexpectedly, that non-myelinating glial cells such as OECs or non-myelinating SCs should express CNPase. However, this enzyme has been shown to exhibit a variety of functions in oligodendrocytes and SCs. An explanation for this observation could be that these cells are differentiating and currently show a premyelinating status. In this context it can be speculated that OECs are closely related to SCs and that both cell types may participate in myelin production in demyelinating diseases.

Furthermore, in this study purified adult canine OECs were treated with FGF-2, which has been previously shown to act as potent OEC and SC mitogen in vitro (TECHANGAMSUWAN et al. 2008, 2009). Under this condition CNPase expression was significantly reduced at the protein and mRNA level. Since CNPase expression is known to be upregulated by OPCs during differentiation and not during proliferation it is presumed that CNPase expression in OECs and SCs is related to cellular differentiation. In addition, dissociation and organotypic slice cultures of canine olfactory bulb were investigated for CNPase and p75^NTR expression. CNPase antigen was detected at 4 and 10 days while p75^NTR was only detectable after 10 days in vitro cultivation. All canine OEC populations lacked or lost their p75^NTR expression during development in vivo. In contrast, long term cultivation of canine OECs leads to an upregulation of this marker in vitro. Therefore, p75^NTR may serve as a useful in vitro marker for the detection of OECs.

Conclusively, CNPase but not p75^NTR can be used as suitable marker for in vivo visualization of canine OECs. It is important to point out that species specific differences regarding the marker expression of cells exist and have to be taken into consideration. The present study demonstrated that canine but not rat OECs express CNPase in vivo. The differences between canine and rodent glia further favors the use of dogs as translational large animal model. Beside this, further studies have to determine whether CNPase can be used as marker for the discrimination between rat OECs and SCs in vivo.

Establishment of an ethidium bromide-induced demyelination model in the murine caudal cerebellar peduncle and in vivo characterization of BO-1 cells The analysis of mechanisms involved in CNS demyelination and remyelination are essential for the establishment of regenerative therapeutic approaches. In this study the species mouse was selected for the following reasons: several natural and genetically induced mouse mutants exist (GRIFFITHS et al. 1990, GRIFFITHS 1996, KLUGMANN et al. 1997, DIMOU et al. 1999, EDGAR et al. 2010) which can be evaluated with this model and secondly the previously established murine

Discussion and Conclusions Chapter 6

investigated in vivo. The caudal cerebellar peduncle (CCP) was chosen as transplantation site because this localization has several advantages: i) it displays a well circumscribed area, ii) most of the axons show a similar diameter and orientation and iii) myelin sheaths have a similar thickness (WOODRUFF et al. 1999). In the present study ethidium bromide concentrations between 0.1% and 0.01% have been applied. It is important to select a toxicant concentration which causes a marked demyelination with limited axonal damage. Ethidium bromide-induced lesions showed a concentration-dependent increase in size. Finally, a concentration of 0.025% ethidium bromide in a volume of 2µl was selected because at this concentration a marked and significant demyelination was shown by myelin basic protein immunohistochemistry and a mild to moderate axonal damage as detected by phosphorylated- and non-phosphorylated neurofilament immunohistochemistry was observed (HANSMANN et al. 2012b). The occurrence of axonal alterations following ethidium bromide application was interpreted as a mouse specific feature because similar studies investigating the caudal cerebellar peduncle of rats lacked significant axonal alterations (WOODRUFF et al. 1999, HANSMANN et al. 2012b).

The next step was the in vivo investigation of BO-1 cells. To ensure the detection of the transplanted BO-1 cells, they were transfected with enhanced green fluorescent protein (eGFP). It was hypothesized that the local environment of the demyelinated CCP would stimulate differentiation of BO-1 cells to remyelinating oligodendrocytes.

Therefore BO-1 cells were stereotactically transplanted into the demyelinated CCP 4 days following ethidium bromide-induced demyelination. Transplanted animals showed a reduced general behavior, weight loss and were euthanized for ethical reasons 17 days post stereotaxic injection of BO-1 cells. Macroscopically and histologically an invasive and liquorogenic metastasizing, murine giant cell glioblastoma was detected in all cell transplanted animals. Furthermore, BO-1 cells were investigated for their marker expression in vivo. BO-1 cells showed a strong expression of eGFP, NG-2, PDFGRα and p53 while only few cells expressed GFAP and MBP. Moreover, the expression of CNPase was markedly reduced compared to the in vitro situation. The reduced expression of CNPase can be interpreted as a reduced degree of differentiation of BO-1 cells in vivo. It can be speculated that the

reason for this may be an insufficient environmental signaling and/or a BO-1 cell intrinsic factor. To confirm that the tumor formation was not the consequence of the ethidium bromide application, BO-1 cells were also transplanted into the normal, non-demyelinated CCP. This revealed that all BO-1 transplanted animals showed tumor formation at the transplantation site.

Conclusively, instead of an expected remyelination by BO-1 cells a murine giant cell glioblastoma occurred in the transplanted animals. Thus BO-1 cells are not suitable for a regenerative therapy. However, they may serve as a versatile experimental model to study tumorigenesis of glial tumors.

Determination/modulation of the phenotype of microglia in TME

The local environment plays an essential role for the blockage of remyelination in demyelinating CNS diseases. In this context the immune system and maybe other glial cells like astrocytes have a major impact on the behavior of microglia and macrophages. Therefore, the present in vitro study aimed to identify the role of astrocytes and microglia in the early inflammatory response to TME. TME represents a well-established, virus-induced animal model of MS. The major goal was to illuminate whether microglia exert pro- (M1) or anti-inflammatory (M2) properties following TMEV infection. In addition, the cytokine profile of astrocytes was measured for the detection of a possible determining role of astrocytes upon the phenotype of microglia. Therefore, astrocytes and microglia were isolated from neonatal SJL/J mice, inoculated with TMEV or mock-solution and gene expression including pro- and anti-inflammatory cytokines, TNF receptors and transcription factors NF-κB (p50, p65) and AP-1 (c-jun, c-fos) was quantified at 6, 48 and 240 hours post inoculation.

Glia cells were isolated from SJL/J mice because this inbred mouse strain is highly susceptible to TMEV infection (DAL CANTO et al. 1996, MECHA et al. 2013). Firstly, microglia and astrocytes were both infected by TMEV with the latter showing a general higher amount of viral RNA. In the early phase post infection (48hpi)

Im Dokument Investigation of murine and canine glia cell differentiation in vitro and in vivo (Seite 31-75)