Morphological Analyses of Senescent Microglia in Human and Rodent Brains under Conditions of Disease and Injury

Permanent Link: http://ufdc.ufl.edu/UFE0022562/00001

Material Information

Title: Morphological Analyses of Senescent Microglia in Human and Rodent Brains under Conditions of Disease and Injury
Physical Description: 1 online resource (104 p.)
Language: english
Creator: Lopes, Kryslaine
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008


Subjects / Keywords: aging, alzheimer, axotomy, dystrophy, ferritin, glia, immunohistochemistry, iron
Neuroscience (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Little is currently known about microglial cells in the aging brain. Recently, microglia have been shown to succumb to senescence-related changes with advancing age, including dystrophy (i.e., degeneration) of their typically branched morphology and alterations in metabolic functions. The cause(s) and consequences of such alterations to microglial cells themselves as well as to the central nervous system as a whole are not known. In this study we characterized the defining morphological features of dystrophic microglia in both human and rodent brains. The central focus of this application was in the assessment of whether aging-related degeneration of microglial cytoplasmic structure is secondary to prolonged iron storage via expression of the iron storage protein ferritin. Since aging in the brain is accompanied by an accumulation of iron in ferritin proteins, particularly L-rich ferritins expressed by microglia, it is hypothesized that aged microglia are susceptible to iron-induced oxidative damage. The findings presented herein shed light not only on the effects of the aging process on microglial structural and functional integrity, but they also introduce ferritin immunohistochemistry as a useful method for detecting microglial cells that are in danger of being lost. Furthermore, our findings support a possible connection between microglial cytoplasmic deterioration and impaired glial neuroprotection. Understanding the underlying mechanisms responsible for the degeneration of microglial cells in aged brains may be important for elucidating the pathogenesis of aging-related neurodegeneration and neurodegenerative diseases.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kryslaine Lopes.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Streit, Wolfgang J.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022562:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022562/00001

Material Information

Title: Morphological Analyses of Senescent Microglia in Human and Rodent Brains under Conditions of Disease and Injury
Physical Description: 1 online resource (104 p.)
Language: english
Creator: Lopes, Kryslaine
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008


Subjects / Keywords: aging, alzheimer, axotomy, dystrophy, ferritin, glia, immunohistochemistry, iron
Neuroscience (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Little is currently known about microglial cells in the aging brain. Recently, microglia have been shown to succumb to senescence-related changes with advancing age, including dystrophy (i.e., degeneration) of their typically branched morphology and alterations in metabolic functions. The cause(s) and consequences of such alterations to microglial cells themselves as well as to the central nervous system as a whole are not known. In this study we characterized the defining morphological features of dystrophic microglia in both human and rodent brains. The central focus of this application was in the assessment of whether aging-related degeneration of microglial cytoplasmic structure is secondary to prolonged iron storage via expression of the iron storage protein ferritin. Since aging in the brain is accompanied by an accumulation of iron in ferritin proteins, particularly L-rich ferritins expressed by microglia, it is hypothesized that aged microglia are susceptible to iron-induced oxidative damage. The findings presented herein shed light not only on the effects of the aging process on microglial structural and functional integrity, but they also introduce ferritin immunohistochemistry as a useful method for detecting microglial cells that are in danger of being lost. Furthermore, our findings support a possible connection between microglial cytoplasmic deterioration and impaired glial neuroprotection. Understanding the underlying mechanisms responsible for the degeneration of microglial cells in aged brains may be important for elucidating the pathogenesis of aging-related neurodegeneration and neurodegenerative diseases.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kryslaine Lopes.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Streit, Wolfgang J.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022562:00001

This item has the following downloads:

Full Text




2008 Kryslaine Oliveira Lopes 2


To my family. 3


ACKNOWLEDGMENTS Completion of this dissertati on was made possible by the contribution of a number of people. I thank my mentor Dr. Wolfgang J. Stre it for nurturing my inte llectual curiosity about glial cells and for encouraging me not to give up after disappointing e xperiments. I thank the other members of my committee, Dr. David Borchelt, Dr. Christiaan Leeuwenburgh, and Dr. John Petitto, for their assistance and suggestions. I am grateful to the Neuroscience office staff members, John Neeley and B.J. Streetman, for ma king sure I had all the necessary paperwork to complete my coursework and to present my research at conferences. I would also like to thank the pa st and present members of the Streit lab for their research assistance. In particular, I w ould like to thank Dr. Chris Ma riani, Dr. Sarah Fendrick, Robby Sweeney and Celeste Karch for their technical he lp and friendship. I am indebted to Dr. QingShan Xue and Nichole Fife for their multitude of assistance and suggestions. I appreciate the help eMalick Njie gave me in various aspect s of my research, from providing antibodies and helpful comments to creating a fun workplace in th e late night/weekend shifts. I am very grateful to my colleague Kelly Miller for her conti nuous encouragement, friendship and for making graduate school an enjoyable experience. Special thanks go to my parents, Roberto and Elizabeth; and my siblings, Rodrigo and Vanessa, for always believing in me. Despite all th e emotional and financial struggles, my family never stopped supporting my choice to stay mile s away from home while I pursued my career interests. I am particularly th ankful to my dad for his long -time commitment to my academic success. Finally, I am principally indebted to Dr. Baler Bilgin for his unwavering support throughout my academic career. This milestone w ould not have been possible without him. 4


TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................7 LIST OF FIGURES .........................................................................................................................8 ABSTRACT ...................................................................................................................................10 CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW............................................................. 12 Microglia: An Overview .........................................................................................................12 Microglial Senescence ............................................................................................................14 Structural Deterioration ...................................................................................................15 Self-Renewable Capacity of Microglia in the Aging Brain ............................................16 Brain Iron Metabolism ............................................................................................................18 Iron-Dependent and -Independent Regulation of Ferritin mRNAs .................................19 Structure and Function of Ferritin ...................................................................................20 Microglia and Iron Dyshomeostasis................................................................................21 Alzheimers Disease............................................................................................................ ...22 Pathology.........................................................................................................................23 Classification of Senile Plaques......................................................................................25 Dystrophic Microglia and Alzhei mer's Disease Pathogenesis........................................26 Facial Nerve Axotomy Model................................................................................................26 Microglial Response to Facial Nerve Axotomy......................................................................27 Project.....................................................................................................................................28 2 MATERIALS AND METHODS...........................................................................................33 Supplier Information........................................................................................................... ....33 Tissue Specimens....................................................................................................................33 Antibody List..........................................................................................................................34 Methods..................................................................................................................................34 Facial Nerve Axotomy....................................................................................................34 Tissue Sectioning and Storage.........................................................................................35 Single Label Immunohistochemistry...............................................................................35 Double Label Immunohistochemistry.............................................................................36 Immunofluorescence.......................................................................................................37 Human samples........................................................................................................37 Rat samples..............................................................................................................37 Morphometric Analyses and Cell Quantification...................................................................38 Human Study...................................................................................................................3 8 5


Rat Axotomy Study.........................................................................................................39 Statistical Analyses........................................................................................................... ......39 3 RESULTS...................................................................................................................... .........43 Morphological Analyses of Senescent Microgl ia in the Human Brain under Conditions of Normal Aging and Neurodegenerative Disease..............................................................43 Ferritin-Positive Microglia Exhibi t Abnormal Morphological Features.........................43 Morphological Characteristic s of Dystrophic Microglia................................................44 Ferritin-Positive Microglia Are Less Abundant than HL A-DR-positive Microglia and Appear Mostly Dystrophic....................................................................................45 Ferritin-Positive Microglia C onstitute a Subset of th e Larger Microglial Pool..............46 Microglial Dystrophic Changes Are Not Du e to Postmortem Tissue Autolysis.............47 Ferritin-Positive Dystrophic Microglia Are Prominent in the Brains of AD patients.....48 Morphological analyses of microglia in the proximity of senile plaques.......................49 Brain specimens used in this study represent early AD..................................................50 Analyses of Microglial Ferritin Immunoreactivity in Young and Aged Rats........................51 Acute Activation of Microglial Cells Does Not Induce Ferritin Expression..................51 Ferritin Immunohistochemistry Labels Pr edominantly Oligodendrocytes in Rat Brains...........................................................................................................................53 Morphological Analyses of Senescent Microgl ia in the Rat Brain under Conditions of Normal Aging and Acute Injury.........................................................................................54 Accumulation of Lipofuscin Granules Are Prevalent in Dystrophic Microglia of the Aging Rat Brain...........................................................................................................54 4 DISCUSSION AND CONCLUSIONS..................................................................................76 Overview of Findings.............................................................................................................76 Only a Subset of Microglial Cells Express L-rich Ferritin Proteins.......................................77 Potential Link Between Iron Storage, Senescence, and Microglial Dystrophy......................78 Microglial Cells Are Vulnerab le to Oxidative Stress Reactions in Aged Brains...................79 Hallmarks of Microglial Degenerati on in Human and Rodent Brains...................................80 Microglial Dystrophy Is Not Due to Postmortem Tissue Autolysis.......................................83 Dystrophic Microglia Are Most Prevalent in the Alzheimer's Disease Brain Than in Age-Matched Non-Demented Control Individuals.............................................................84 Dystrophic Microglia Are Associated with Aging and Not Injury Conditions in Aged Rats......................................................................................................................................86 Ferritin Immunoreactivity Is Not Upregul ated in Activated Microglial Cells.......................89 Concluding Remarks............................................................................................................. .90 LIST OF REFERENCES...............................................................................................................94 BIOGRAPHICAL SKETCH.......................................................................................................104 6


LIST OF TABLES Table page 1-1 Age-related changes in microglial metabolic activity.......................................................29 1-2 Microglial degenerati on in the literature............................................................................30 1-3 Alzheimers disease-related pathological hallmarks develop in a hierarchical manner....31 2-1 Clinical and pathological features of human cases............................................................41 2-2 Profile of postmortem interval (PMI) cases.......................................................................42 2-3 Source, species reactivit y, cell specificity, and dilutions of primary antibodies...............43 3-1 Classification of micr oglial dystrophic characteristics based on ferritin immunohistochemisty........................................................................................................58 3-2 Proportion of HLA-DRand ferritinimmunoreactive microglia that exhibit dystrophic characteristics in human brain specimens........................................................60 3-3 Morphological characteristics of activated and dystrophic microglia...............................67 3-4 Cell counts of immunoreactive microglia in the unoperated and axotomized facial nucleus...............................................................................................................................75 4-1 Comparison of dystrophic morphological characteristics between humans and rodents................................................................................................................................93 4-2 Differential features between dystrophy, apoptosis, and necrosis.....................................94 7


LIST OF FIGURES Figure page 1-1 Transformation of ramified microglia into hypertrophic and dystrophic forms................30 1-2 Classification of senile plaques..........................................................................................3 1 1-3 Facial nerve axotomy model..............................................................................................32 3-1 Morphological characteristics differentiate ferritin-positive microglia from ferritinpositive oligdendrocytes....................................................................................................56 3-2 Representative photomicrographs of the number of microglia immunoreactive for either HLA-DR antigens or ferritin proteins......................................................................57 3-3 Comparison of the average number of immunoreactive and dystrophic microglia stained for either HLA-DR antigens or ferritin proteins....................................................59 3-4 Ferritin-positive microglia constitute a subpopulation of the larger HLA-DR microglial pool................................................................................................................ ...60 3-5 Postmortem interval study.................................................................................................6 1 3-6 Cytorrhectic microglia ar e present in aged but not in young human brain tissues............62 3-7 Microglial cells in Alzheimers diseas e tissues exhibit a bnormal morphological features...............................................................................................................................63 3-8 HLA-DR-positive microglia in the vicinity of senile plaques exhibit mostly a ramified morphology in both HPC and AD tissues...........................................................64 3-9 Ferritin-positive microglia present mostly a deramified profile in both HPC and AD tissues irrespective of the vicinity of senile plaques..........................................................65 3-10 Morphological characteristics of senile plaques in the AD brain specimens under study...................................................................................................................................66 3-11 Ferritin protein levels in crease with age in the rat.............................................................67 3-12 L-ferritin expression is induced by the ag ing process in microg lial cells and not by acute activation conditions.................................................................................................68 3-13 Cellular characterization of ferritin-pos itive and OX-42-positive cells within the rat facial nucleus.....................................................................................................................69 3-14 Ferritin-positive cells are prevalent in the rat hippocampus..............................................70 8


3-15 Double immunofluorescence staining for ferritin and CR3 receptors in the hippocampal, cortical, and cerebellar regions of a 30-month-old rat................................71 3-16 Immunofluorescence staining using Iba1 and ED1 markers in the control and axotomized facial nuclei of both a 3-month-old and a 30-month-old rat ten days postaxotomy.................................................................................................................... ...72 3-17 ED1-immunoreactivity is limited to the perinuclear region of Iba-1-positive microglia...................................................................................................................... ......73 3-18 Autofluorescence of lipofuscin particles us ing ultraviolet, green, and blue excitation lights......................................................................................................................... ..........73 3-19 Lipofuscin granules accumulate in sene scent dystrophic microglia of aged rats..............74 9


Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MORPHOLOGICAL ANALYSES OF SENE SCENT MICROGLIA IN HUMAN AND RODENT BRAINS UNDER CONDITI ONS OF DISEASE AND INJURY By Kryslaine Oliveira Lopes August 2008 Chair: Wolfgang J. Streit Major: Medical Sciences--Neuroscience Little is currently known about microglial cel ls in the aging brain. Recently, microglia have been shown to undergo senescence-rela ted changes with advancing age, including dystrophy (i.e., degeneration) of their typically branched morphology and alterations in metabolic functions. The cause(s) and consequences of such alterations to microglial cells themselves as well as to the central nervous sy stem as a whole are not known. In this study we characterized the defining mor phological features of dystrophi c microglia in both human and rodent brains. The central focus of this appl ication was in the assessment of whether agingrelated degeneration of microglial cytoplasmic st ructure is secondary to prolonged iron storage via expression of the iron storage protein ferritin Since aging in the brain is accompanied by an accumulation of iron in ferritin proteins, particular ly L-rich ferritins expr essed by microglia, it is hypothesized that aged microglia are highly susceptible to ironinduced oxidative damage. The findings presented herein shed light not only on the effects of the aging process on microglial structural and functional integrit y, but they also introduce ferritin immunohistochemistry as a useful method for detecting microgl ial cells that are in danger of being lost. Furthermore, our findings support a possible connection between microglial cytoplasmi c deterioration and impaired glial neuroprotection. Understanding the unde rlying mechanisms responsible for the 10


11 degeneration of microglial cells in aged brains may be import ant for elucidating the pathogenesis of aging-related neurodegenerati on and neurodegenerative diseases.


CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW Microglia: An Overview Microglial cells constitute a functionally dynamic and morphologically heterogeneous cell population within the centr al nervous system (CNS; Hanisch and Kettenmann 2007). Ramified microglia, which are th e most prevalent in the healthy CNS, are characterized by highly branched processes arising from the cell body in every direction. These processes engage in pinocytosis and are highly motile, continually going through renewal cycles as they inspect the CNS homeostatic status (Booth and Thomas 1991; Nimmerjahn et al., 2005; Davalos et al., 2005). Typically, ramified microglia are uniform ly spaced in the hea lthy brain parenchyma (Streit, 2005) although on occasion they have also been shown to contact neuronal cell bodies, astrocytes, and blood vessels (Nimmerjahn et al ., 2005). The implication is that in the normal, healthy brain, ramified microglia are constantly monitoring the well-bein g of neighboring cells as well as the status of their extracellular milie u. In response to a disturbing stimulus, the first immediate reaction of microg lial cells is to project all of their ramified processes in the direction of the stimulus epicenter, forming a shield as a first line of defense (Davalos et al., 2005). Taken together, these findings strongly suggest that the highly branched morphology of microglia serves an essential purpose in CN S protection and normal functioning. Microglia can react to a number of endogenous (e.g., abnormal protein aggregation, neuronal dysfunction/death) and exogenous (e.g., blood brain barri er [BBB] disruption by injury/ infection) signals. When the disturbing stimulus persists, local ramified microglia become activated. That is, they lose their fine branches, enlarge their processes (hypertrophy), increase expression of immunophenotypic markers, prolifer ate, and increase secretion of cytokines and other inflammatory mediators, including hydrogen peroxide (H2O2) and superoxide (Colton and 12


Gilbert, 1987; Perry et al., 1995). Activation of microglial cells is an elaborate, graded process, whereby precise immune effector functions are matched to the speci fic stimulus at hand (Streit, 2002; Schwartz et al., 2006). More over, activation is under strict control and occurs only in response to strong stimuli (Kreutzberg 1996). In extreme levels of activation, when cellular debris must be cleared (i.e., phagocytosed and de graded), activated microglia metamorphose into rounded brain macrophages. The acquired functi ons of activated microglia are matched by changes in morphological c onfiguration, going from ramifi ed over rodlike to rounded morphology depending on the nature of the activ ating agent and accompanied by changes in gene expression profiles. Cumu latively, these observations emphasize the significance of microglial cytoplasmic structure for the prope r execution of their cellular functions. Recently, several histological studies in hum an brain have identified microglial cells exhibiting an aberrant cytoplasmic structure, termed dystrophic microgl ia (Conde and Streit 2006; Ferraro 1931; Lopes et al. 200 8; Simmons et al. 2007; Streit 2002; Streit 2004; Streit et al. 2004; von Eitzen et al. 1998; Wierzba-Bobrowicz T et al. 2004). Their morphology is marked by degenerative changes, such as thinning and loss of distal branches, fragmentation of cytoplasm (cytorrhexis), and the formati on of rounded swellings (spheroids) along major processes. This type of microglia is more prevalent in the ag ed brain, and thus microglial dystrophy is thought to be a reflection of cellular sene scence and ongoing degeneration (F lanary et al. 2007; Streit 2006; Streit et al. 2004). It is not yet known what functional cha nges result from the degeneration of microglial cytoplasmic structure. It is possibl e that microglial dystrophy leads to an impairment of their normal cellular functioning, possibly resulting in a si gnificant reduction of their normal neurotropic support functions. 13


Microglial Senescence Neurons, oligodendrocytes and astrocytes orig inate from the neuroectoderm. In contrast, microglia arise from myeloid progenitors deri ved from the hemangioblastic mesoderm (Streit 2001). Reminiscent of their monocytic lineage is the ability of microglia to proliferate and undergo self-renewal. The prolifera tive capacity of microg lia is relevant to the study of microglia in the aging brain because it underl ies their senescence potential: micr oglia, as mitotic cells, have a finite replicative lifespan. Upon reaching a criti cal shortening of telomeres, the tandem-repeat sequences at the end of chromosomes, the cell ente rs the senescent state. Analyses of telomere length revealed that microglia beco me senescent with mitotic stimuli in vitro (Flanary and Streit 2004) and with normal aging in vivo (Flanary and Streit 2003). The primary consequence of this replicative senescence is the cessation of cell division through the cell s arrest in the G1-S boundary of the cell cycle (Yanishevsky et al. 1974). Cells that have entered the senescent state remain metabolic active, although there are gradual changes in cellular stru cture and function. For instan ce, senescent cells accumulate lipofuscin granules in their cytosol (Brunk and Te rman 2002; Sitte et al. 2001; Vogt et al. 1998). Lipofuscin is a mixture of autofluorescent lysoso mal lipo-pigments and proteins that cannot be excreted or degraded (Gary and Woulfe 2005; Seehafer and Pearce 2006). The amount of lipofuscin in mitotic cells is depe ndent on both its rate of formation and its rate of dilution by cell division (Sitte et al. 2001). Recen tly, microglia have been show n to progressively accumulate lipofuscin-containing dense bodies in their cyto plasm (Sierra et al. 2007; Xu et al. 2008; Yamasaki et al. 2007), thus providi ng further evidence for their re plicative senescence potential with advancing age. Other age-related change s in microglial metabolic functioning include a 14


decrease in proteolytic activity, an increased pr oduction of pro-inflammatory mediators, and an apparent primed state of activation. Th ese results are summarized in Table 1-1. Structural Deterioration The morphology of senescent microglia is quite different from the ramified appearance of resting microglia or the bushy morphology of hypertrophic ( activated) microglia (Figure 1-1) (Streit 2006; Streit et al. 2004) As a result, to emphasize th eir degenerative cytoplasmic structure, these cells were named dystrophic mi croglia (Streit et al. 2004 ). The most pervasive feature of microglial cytoplasmic degeneration is the presence of atypical, tortuous processes, often found in deramified micr oglia (v Eitzen et al., 1998; Streit et al., 2004a; WierzbaBobrowicz et al., 2004). Microglia l dystrophy is also manifested by the formation of spheroids (various-sized bulbous formations) on branches, either singly or in succession (beading); by fusion with other microglial cells (clusters), indi cative of loss of normal contact inhibition; by fragmentation (cytorrhexis) of processes (i.e., severing of main branches); and by atrophy (i.e., nearly complete loss of cytoplasmic structure) The observation that dystrophic microglia are found scattered randomly in the brain parenchy ma, often alongside ramified microglial cells, suggests that they may form a subset of olde r (senescent) microglia (Streit et al. 2004). In regards to their incidence, microglial ce lls exhibiting dystrophi c characteristics have been identified in the normal aging brain (Conde a nd Streit 2006; Flanary et al. 2007; Streit et al. 2004), as well as in several neurol ogical disorders, including Alzh eimers disease (AD) (Ferraro 1931; Streit 2002; Streit et al. 2004), Creutzfeldt-Jakob diseas e (von Eitzen et al. 1998), Huntingtons disease (HD) (Simmons et al. 2007 ), and schizophrenia (W ierzba-Bobrowicz T et al. 2004). Degenerative changes have also been reported in rat primar y microglial cultures exposed to AD-associated amyloiddeposits (Korotzer et al. 1993), and in murine disease 15


models of HD (Ma et al. 2003; Simmons et al. 2007) and amyotrophic late ral sclerosis (Fendrick et al. 2007). Recent findings on the characteriza tion of microglial dystrophy in various animal species are summarized in Tabl e 1-2. Given that microglial f unctioning is highly dependent on their morphological state (Kreut zberg 1996), it is likely that th ese morphological alterations serve as an impediment to the normal surveillanc e function of these cell s (Davalos et al. 2005; Nimmerjahn et al. 2005), which could lead to dele terious consequences for the CNS as a whole. Self-Renewable Capacity of Microglia in the Aging Brain Iron is indispensable for DNA synthesis and therefore it pl ays a crucial role in cell proliferation rate. As a cofact or of the enzyme ribonucleotid e reductase, iron mediates the production of deoxy-ribonucleotides from the corresponding ribonucleotides (the rate-limiting step in the synthesis of DNA pr ecursors and, therefore, a key control point in DNA synthesis) (Hoffbrand et al. 1976). Numer ous studies have shown that iron limitation arrests cell proliferation (Le and Richardson 2002). In the ag ed brain, microglia proliferate more vigorously than in younger brain tissues (Conde and Streit 2 005). A plausible explana tion for this increased proliferation rate is the age-related increase in microglial intracellular ir on levels. The observed increase in microglial proliferation rate likely reflects an imbalance in microglial self-regulatory mechanisms. Inevitably, higher levels of microglia l proliferation in the aged brain would result in further propagation of microglial se nescence in the affected tissue. An interesting question is how microglia can reach replicative senesc ence in the normal aging brain that is devoid of obvious inflammatory/proliferative stimuli such as infection, neurological trauma or extracellular pr oteinaceous aggregates. Although the underlying mechanisms of microglial prolifer ation are well-understood in specific injury models (e.g., facial nerve axotomy) and in cell culture systems, litt le is known about how normal aging may promote 16


microglial replicative senescence. It is possible, th at in addition to stimuli inherent to the CNS, alterations in systemic functioning also play a role in microglial senescence. One such mechanism may be linked to the bodys response to stressful events. Microglia express receptors for stress-related hormones (Wang et al. 2002) and respond to physical/emotional stress and related hormones with proliferation (Nair and Bonneau 2006; Wang et al. 2003) and morphological and functional activation (Sugama et al. 2007). Similarl y, microglial activation also occurs in response to closed head trauma (Schmidt et al. 2005), syst emic infection (Lemstra et al. 2007; Semmler et al. 2005), and even diet (e.g., cholesterolrich diet) (Crisby et al. 2004; Streit and Sparks 1997). The accumulation of thes e processes during a lifetime, in addition to age-related changes in neural homeostasis, is likely to lead to senescence of at least a subpopulation of microglial cells. Microglial senescence was once considered an immaterial process because microglia are self-renewable cells (Lawson et al. 1992). Besides mitosis, micr oglial replenishment can occur via recruitment of a sp ecific subpopulation of bone marrow-deri ved cells, which migrate into the neural parenchyma and differentiate into micr oglia (Mildner et al. 2007). However, the recruitment of microglial myeloi d progenitors across the BBB has been shown to decline with age (Simard et al. 2006), albeit th e change is not highly substantial. Nonetheless, the mechanisms of microglia replenishm ent described herein might put a heavy load on a system already compromised by old age. As a result, senescent microglia accumulate in the aged CNS. Higher incidences of senescent microglia may contribute to aging-related neurodegeneration and neurodegenerative diseases by at least two mechanisms: (i) Hype ractivation or hypersensitivity to stimuli and/or misregulation of the activation pro cess, leading to an increased production of toxins, such as reactive oxygen species (ROS); and (ii) Hypoac tivation and undernourishment of 17


neighboring neurons in both the quiescent and disturbed CNS. Eith er one of these two possibilities would produce a devastating effect on the welfare of neur al cells, potentially exacerbating cytotoxic signals already present in the tissue or directly ensuing cellular death. Brain Iron Metabolism Iron is an essential nutrient requ ired for the proper functioning of virtually all cells. It is crucial for oxygen transport (function and biosyn thesis of hemoglobin), energy production as a cofactor of several enzymes in the mitochondr ial electron transport chain, immune function, cellular growth (i.e., RNA and protein synthesis), and proliferation (i.e., DNA synthesis) (Lieu et al. 2001). In the CNS, iron participates in add itional roles, including: 1) The synthesis of monoamine (Ramsey et al., 1996; Beard et al., 2003) and GABA (Taneja et al. 1986) neurotransmitters; 2) Normal brain development, specifically in myelin synthesis (Connor et al., 1995), 3) Cognition (Black, 2003); 4) Memory fo rmation, through the development of dendritic spines in the hippocampus (Jorgenson et al., 2003) and many others. These examples highlight the importance of the availability of iron for brain cell viability. However, high levels of free iron in and around cells is very toxic due to its pr opensity to catalyze the generation of the highly reactive hydroxyl radical (OH), which can lead to lethal c onsequences by permanently altering the molecular structure of lipids, DNA and proteins (Crichton et al., 2002). Physiologically, iron exists in one of two oxi dation states: ferrous (Fe2+) iron or ferric (Fe3+) iron. The generation of OH occurs via the oxid ative conversion of Fe2+ to Fe3+ in the presence of H2O2 by means of the Fenton reaction: Fe2+ + H2O2 Fe3+ + OH + OH(Crichton et al., 2002). The coordinated expression of two types of pr otein with high affinity (transferrin) and high capacity (ferritin) for iron is largely responsible for the ma intenance of adequate levels of intracellular iron (Levenson and Ta ssabehji, 2004). Apo (iron free) -transferrin is an iron carrier 18


glycoprotein primarily involved in iron uptake. Each molecule of transferrin (Tf) binds two ferric iron molecules. Iron is taken up by cells primarily after Tf, the principal iron-carrying protein of the plasma, binds to a specific cell surface recept or and the transferrin-receptor complex is internalized (Dautry-Versat et al., 1983). The acidic environment of early endosomes induces a conformational change in Tf resu lting in the release of Tf-bound Fe3+ (Bali et al. 1991). Subsequently, Fe3+ is reduced to Fe2+ by a yet unidentified reducta se, allowing divalent metal transporter 1 (DMT-1; also known as DCT1 or Nramp2) to transfer Fe2+ to the cytosol (Fleming et al. 1998). Apo-Tf and TfR are then recy cled to the plasma membrane, where each can undergo further cycles of iron binding and uptak e. Once iron enters the cell, any portion in excess (i.e., not needed for immediate use in en zymatic reactions) is st ored by ferritin (Ft), a ubiquitous and highly conserved hollow protein shell that constitutes the main intracellular iron storage protein (Torti and Torti 2002). As a cons equence of their inherent functional dichotomy, Ft and TfR expression levels are re ciprocally related in their responses to changes in iron levels. Iron-Dependent and Independent Regulation of Ferritin mRNAs Ferritin and transferrin r eceptor levels are primarily regulated by intracellular iron concentrations at the level of translation (Ponka and Lok 1999; Torti and Torti 2002). The nucleotide sequence of both Ft and TfR mRNAs c ontain a stem-loop struct ure, termed an ironresponse element (IRE), which in the case of Ft mRNA is located in its 5 untranslated region (UTR), while in the TfR mRNA it is located in the 3 UTR. When iron is scarce, the IRE binds a 90-kDa IRE-binding protein (IRE-BP) which has an approximate 30% identical amino acid sequence with the citric acid cycle enzyme aconita se from mitochondria (Kim et al. 1996). In the case of Ft, IRE-BP bound to IRE blocks the in itiation of translation of the Ft mRNA, while inducing TfR mRNA stability and translation (P onka and Lok 1999). When iron levels increase, 19


the IRE-BP binds iron as a 4Fe-4S cluster. Beca use the binding sites of iron and RNA overlap considerably, the result ing IRE-BP-iron complex cannot bind RNA. Thus, in condition of iron surplus, ferritin mRNA is released from the IRE-BP cap and translated to produce ferritin, which sequesters the excess iron. On the other hand, TfR mRNA freed from IRE-BP is rapidly degraded (Ponka and Lok 1999). In addition to iron levels, other factors have be en demonstrated to affect cellular ferritin. Of interest to the present inve stigation are factors produced by microglial cells under activation conditions, including cytokines (Torti et al., 1988 ; Rogers, 1996) and oxidants (Cairo et al., 1995; Mehlhase et al., 2006). Supe roxide in particular has been shown not only to modulate Ft expression but also iron leakage from ferritin (Bolann and Ul vik, 1990; Yoshida et al., 1995). The propensity of labile iron to catalyze the generation of oxygen radicals means that the intracellular concentration and chemical form of the element must be kept under tight control. Structure and Function of Ferritin Apo-ferritin is a large (450 kDa) protein located in the cytoso l of virtually every cell. It consists of a hollow protein shell composed of 24 functionally distinct heavy (H)-chain and light (L)-chain subunits, which store up to 4500 atoms of iron per molecu le of protein (reviewed in Harrison and Arosio, 1996). The H-ch ain has a metal-binding site a ssociated with its ability to catalyse the oxidation of the highly toxic Fe2+ to the less reactive iron form Fe3+ (Bakker and Boyer, 1986, (Lawson et al. 1989). This ferroxida se activity occurs only at the H-chain subunit and it is of great importance because although ferritin can only take up iron in the Fe2+ form, this is a highly reactive cation and mu st be quickly converted to Fe3+ to prevent any iron-induced toxicity. H-rich ferritins take up ir on faster than L-rich in vitro (Lawson et al. 1989; Levi et al., 1988), and are found predominantly in such tissues as heart, kidney and brain (Harrison and 20


Arosio 1996). The L subunit lacks the ferroxidase site but contains addi tional glutamate residues on the interior surface of the protein shell whic h produce a microenvironment that facilitates mineralization of Fe3+ for long-term storage (Santambrogio et al., 1996). L-ferritin subunits are mostly found in tissues principally involved in iron storage, such as the liver and spleen (Harrison and Arosio 1996). Wh ile ferritin is mainly a cytosolic protein, small amounts are present in the nucleus (Surguladze et al., 2004) and in mitochondria (Corsi et al., 2002) as well as in lysosomes where it is degraded (Sibille et al., 1989). Since all cells contain ferritin s composed of both subunits (even though the H:L ratio is celland tissue-specific), ir on enters the protein as Fe2+ by the action of the H-subunit, but is always stored as Fe3+ in the central core via the L-subunit. In order for iron release to occur a reductive conversion back to Fe2+ is required (Jones et al., 1978). Iron mobilization happens more quickly in H-rich ferritins, since most Fe3+ cations are not mineralized and can be easily recruited and released. This is a defining feature behind the cellular composition of ferritin proteins. In the CNS, H-rich ferritins are mostly present in neurons and oligodendrocytes, which are two cell types with high demand for iron participation in a number of enzymatic reactions. On the other hand, L-rich ferritins are primar ily expressed by microglia and oligodendrocytes (Cheepsunthorn et al. 1998; Connor et al. 1994). Because iron is constantly needed for normal myelin production, oligodendrocytes retain a large concentration of iron atoms in both L-rich and H-rich ferritin for their own use. As a conseque nce, microglia are the pr incipal cells responsible for maintaining adequate iron levels w ithin the CNS (Connor and Menzies 1995). Microglia and Iron Dyshomeostasis A common feature of both aging and several of the neurological diseases in which dystrophic microglia are present is iron dyshomeostasis. Iron levels must be tightly regulated to 21


prevent iron-mediated toxicity. Too little iron impedes the normal functioning of cells, whereas too much iron promotes iron-i nduced oxidative stress. Iron metabolism is thought to be disregulated with advancing age because iron progressively accumulates in the aged brain (Bartzokis et al. 1997; Benkovic and Connor 1993; Connor et al. 1990; Roskams and Connor 1994), especially in brain regions affected by age-dependent neurodegenerative diseases such as AD, Parkinsons disease (PD), and HD (Bartzokis et al. 1999; Connor et al 1995; Dexter et al. 1991). This increase in brain iron levels places a large toll on microglia, since prolonged storage of excess iron is performed primarily by micr oglial cells (Connor et al 1994; Connor et al. 1990). The combined effects of microglial sene scence, in which cellu lar functions are at suboptimal conditions, and increased intracellular iron levels may predispose microglial cells to degeneration in the aging brain. Alzheimers Disease First described by Alois Alzheimer in 1907, AD is an incurable, late-onset neurodegenerative disorder that primarily affects areas of th e brain that are essential for cognitive function. As the disease progresses, other brain regions are affected as well. The loss of memory, judgment and emotional stability that AD inflicts on its victim s occurs gradually and inevitably, usually leading to d eath in a severely debilitated, immobile state between four and twelve years after onset. No treatment that retards the progression of the disease is known. AD affects approximately 20 million people worldwide, making it the leading cause of dementia in the elderly (Cummings and Cole 2002). As more pe ople survive to the seventh, eighth, and ninth decades, AD is rapidly becoming an urgent public health problem. The clinical diagnosis of probable AD is based on family history, phys ical examination, neuropsychological testing, laboratory studies and neuroimaging techniques. However, there is no specific laboratory marker 22


to support the definite diagnosis of AD or fo r monitoring the progression of the disease. A postmortem histological examination is often required for confirma tion of its clinical diagnosis. This is because AD is defined pathologically by specific abnormally folded proteins that aggregate extracellularly in senile plaques and as neurofibrillary tangles within affected neurons. In AD, acetylcholine-releasing neurons, w hose cell bodies lie in the basal forebrain, primarily (and selectively) degenerate. Thes e cholinergic neurons provide widespread innervation of the cerebral cortex and related structures and play an important role in cognitive functions, especially memory. Stereological an d biochemical analyses have recently shown, however, that the loss of synaptic density correlates be tter with cognitive decline than either the decrease in neuronal cell numbers or the accumu lation of plaques and tangles (DeKosky and Scheff 1990; Masliah et al. 1989). De spite rigorous efforts to unde rstand the pathogenesis of AD, its etiology remains unknown. Most cases of AD o ccur sporadically, that is, without any known familial predisposition. Pathology Histopathologically, AD is characterized by a generalized atrophy of the cerebral cortex and the widespread appearance of senile pla ques (SP) and neurofibri llary tangles (NFT). Extracellular SPs contain a core of beta-amyloid peptide (A ), which is derived from the proteolytic processing of the amyloidprecursor protein (APP). When this proteolytic cleavage occurs by the actions of both and -secretases, a soluble A peptide with a length between 3942 amino acids is generated, with A 40 being the most common isof orm. The newly generated A peptide forms an -helix structure and remains mostly soluble. For yet poorly understood reasons, in AD the multimeric -secretase complex cleaves APP preferentially at residue 42 instead of 40. The resulting A 42 is highly susceptible to conf ormational changes leading to 23


aggregation into fibrils with an insoluble -pleated sheet. Once A forms a -pleated structure it becomes resistant to degradation resulting in permanent extracellular deposition in the brain parenchyma as the main constituent of senile plaques. Similar to A fibrils, NFTs are composed of misfolded proteins. NFts, howev er, aggregrate within the cell bodies of selectively vulnerable neurons as insoluble paired helical filaments containing hyperphosphorylated microtubuleassociated tau protein. Although these pathological lesi ons constitute the histological hallmarks of the AD brain they can also be found in nondemented elderly people (Arriagada et al. 1992). Thus, for the most part, the distinction between norma l brain aging and AD is quantitat ive rather than qualitative. Usually patients with progressive dementia of th e Alzheimer type have moderately or markedly more mature SPs and NFTs than age-matched nondemented people do. However, the principal difference between demented AD patients and nonde mented elderly people lies in the prevalence and distribution pattern of NFTs and SPs (Thal et al. 2006). In cognitively normal individuals, SPs are usually restricted to the cerebral cortex, the basal ganglia, the thalamus and hypothalamus, while in AD patients SPs are found not only in those areas but also in the mi dbrain, brain stem and cerebellum (Thal et al. 2006). Likewise, NFT pathology is only seen in the pr imary and secondary neocortical areas in AD patients (Braak and Braak 1991). The progressive incursion of anatomical regi ons by AD-related pathologies facilitated the establishment of histological stag ing systems for the NFT (Braak Stages I-VI) and SP (Stages 1-5) pathological e xpansion throughout the neuropil (T able 1-3) (Braak and Braak 1991; Thal et al. 2006). Progressi on through these stages also symbolizes the increase in the severity of cortical destruction. Of note, at early (clinically silent ) stages, plaques first accumulate in the neocortex, while tangles concentrate in the hippocampus. Only at late stages of 24


the disease the two pathologies become interactiv e. Taken together, these findings indicate that although these plaques and tangles are also present in normal aging brain, they are far less severe and less prevalent than in AD. Classification of Senile Plaques The senile plaque is a complex, slowly e volving structure and th e time required to generate fully formed, mature plaques may be years or even decades. For unknown reasons, this maturation seems to occur much more comm only in the symptom-produc ing cerebral cortex than in, for example, the symptom-free cerebellum. SPs can be classified into two representative subtypes (diffuse and neuritic) based primarily on morphological and hist ological characteristics as illustrated in Figure 1-5 (Oide et al. 2006). Amorphous, nonfilamentous deposits of A protein that are sharpl y delineated are called diffuse pla ques. Significantly, most diffuse A protein deposits contain few or no degenerating neurites or reac tive glial cells. The presence of degenerating neuronal processes and reactive astr ocytes and microglia within and around the plaque is often associated with the presence of a homogeneous appearing, darkly stained central core of A protein and plaques containing such are called dense core plaques. Dense core plaques are also called neuritic plaques. In the late 1990s several investigators noticed that diffuse plaques are actually much more abundant th an the classic neuritic plaques (Yamaguchi et al. 1988), occurring in a variety of degenerative and non-degenerative conditions, while neuritic plaques are particularly likely to exist in persons with AD (T hal et al. 2006). Electron microscopic examinations have revealed that much of the tissue in the vicinity of the diffuse plaque is indistinguishable from surrounding normal brain tissue, suggesting a potential innocuous effect at this stage of th e disease (Yamaguchi et al. 1988). 25


Dystrophic Microglia and AD Pathogenesis Although the pathogenesis of AD is not yet unde rstood, several risk factors have been associated with the development of the disease. Age is the strongest known risk factor. Most AD patients are 65 or older and according to the Na tional Institute on Aging the number of patients approximately doubles every 5 years after age 65. Despite the appare nt contribution to AD development, the exact mechanisms by which agi ng contributes to neuronal degeneration remain unresolved. To date several changes in brain homeostasis have been linked to the aging process, a large number of which is linked to microg lial cell functioning, incl uding an increase in oxidative stress markers, the accumulation of iron deposits within the brain parenchyma, and an apparent age-related increas e in microglial activation. It is possible that among the various factors that are involved in AD pathogenesis, a loss of normal function by senescent dystrophic microglia leading for example to an iron metabolic malfunction and/or a loss of ne uronal trophic support, over an ex tended period of time may play a contributory or acceleratory role. This notion, which is based in the microglial dysfunction hypothesis originally proposed by Streit (2002) highlights the signifi cance of microglial functioning for the maintenance of neuronal cell viability. Since in creasing evidence implicates a role of iron in many neurodegenerative disorder s (Dexter et al. 1991; Hallgren and Sourander 1958; Qian and Wang 1998), the study of ferritin-c ontaining microglia in the aged and AD brain can provide important clues about the changes that may lead to iron imbalance in the brain and possibly a model for therapeutic interven tion in case of iron storage dysfunction. Facial Nerve Axotomy Paradigm The cell bodies of facial motoneurons (FMN) are located in the brainstem but axons are projected out of the CNS through th e ipsilateral facial nerve (FN) to innervate the musculature of 26


the face (Fig. 1-2). Corollary to this anatomical structure, the axotomy of the rat FN offers a reliable and useful model in which to study th e underlying principles of microglial activation in vivo as well as microglial responses to neurodegeneration and regeneration. In this model, the FN can be transected, resected or crushed near its exit from the skull at the stylomastoid foramen resulting in retrograde, n on-lethal injury of cell bodies in th e facial motor nucleus (FNu), which can then be microdissected and processed for hi stological examination. The advantages of the FN axotomy paradigm over other in vivo and in vitro models is sevenfold: 1) Remote lesion (responses not due to surgery it self); 2) BBB remains intact ( no influx of peripheral macrophages that could confound the results); 3) High reproduc ibility; 4) Endogenous transient glial responses (as opposed to persistent glial activation in in vitro models); 5) Neuronal regeneration (mild severity); 6) Neuron-glia inter actions (intrinsic microenvironmen t of the brain is maintained); and 7) Analytical strength (contral ateral, unlesioned side serves as an internal control) (reviewed in Moran and Graeber, 2004). This type of le sion is reversible and does not lead to neurodegeneration, resulting primarily in a transien t and graded activation of microglial cells in the vicinity of injured cell bodies (M oran and Graeber 2004; Streit 2000). Microglial Response to Facial Nerve Axotomy The microglial response to FN axotomy pro ceeds though a series of graduated steps, commencing within 24 hours post-injury with hypertrophy of processes re sulting in a bushy appearance and an increase in the expression of th e immune markers, such as type 3 complement receptors (CR3) and major histocompatibility complex (MHC) antigens (Graeber et al., 1988: Streit et al., 1989; Raivic h et al. 1999). At this stage microgl ia are activated and approximately 23 days post-axotomy they start to proliferate forming a sheath around the injured neuronal cell bodies (Fig. 1-3), and their numbers reach maxima l levels after 4-7 days (Graeber et al., 1988b). 27


Activated microglia also begin st ripping the injured neuron of s ynaptic terminals (Blinzinger and Kreutzberg 1968). Once the stimulus has dissipated, microglial cell numbers return to pre-injury levels by programmed cell death mechanisms (G ehrmann and Banati, 1995; Jones et al., 1997). Complete regeneration of motoneurons is depend ent on the severity of the lesion, requiring for instance approximately 2 weeks in the mildest (crush) lesions (Moran and Graeber, 2004). The purpose of using the axotomy model in this inves tigation was two-fold: 1) To determine whether microglial activation by an acute injury (ie, FN axotomy) induces microglial ferritin expression; and 2) To determine whether the appearance of dystrophic microglia is increased after acute microglial activation in aged animals. Project This thesis had two main objectives: 1) To elucidate the morphological similarities and differences between senescent microglia in hum an and rodent brains; and 2) To determine whether long-term iron storage function via expression of L-rich ferritin proteins can be associated with microglial dystrophy, possibly by in creasing their vulnerabili ty to iron-catalyzed oxidative stress reactions. In addi tion, we also wanted to assess the contribution of disease and injury conditions to indices of microglial cytoplasmic degeneration in aged brains. Our findings support a possible connection betw een microglial cytoplasmic dete rioration and impaired glial neuroprotection. We propose that a loss of mi croglial neuroprotectiv e function by dystrophic microglia may contribute to the neuronal dys function and/or death that characterize neurodegenerative diseases, su ch as Alzheimers disease. 28


Table 1-1. Age-related changes in microglial metabolic activity CHANGES IN PROTEOLYTIC ACTIVITY REFERENCE Impaired protein turnover (Stolzing and Grune 2003) Decline in proteasomal function (Stolzing and Grune 2003; Stolzing et al. 2002) Decreased protein synthesis (Stolzing and Grune 2003) Decreased degradation of proteins from apoptotic vesicles (Stolzing et al. 2006) Accumulation of lipofuscin granules (Sierra et al. 2007; Xu et al. 2008) Increase of inclusions, vacuoles, and granularity (Peinado et al. 1998; Sierra et al. 2007) CHANGES IN CELL SURFACE MOLECULE EXPRESSION Increase in MHC class II antigen expression (McGeer et al. 1987; Sheffield and Berman 1998) Increase in CD40, CD45, and CD86 expression (Stolzing and Grune 2003) CHANGES IN CYTOKINE PRODUCTION Increased production of IL-6, TNF, and IFN(Stolzing et al. 2006; Ye and Johnson 1999) Increased production of TNF, IL-1 IL-6, IL-10, and TGF1 mRNA expression (Buckwalter and Wyss-Coray 2004; Sierra et al. 2007) Decreased production of IL-12 (Stolzing et al. 2006) CHANGES IN ROS PRODUCTION Higher basal nitric oxide release (Stolzing and Grune 2003) Higher production of ROS (Heppner et al. 1998; Stolzing et al. 2006) Diminished capacity to evoke oxidative burst (Stolzing and Grune 2003; Stolzing et al. 2006) CHANGES IN ACTIVATION DYNAMICS Ostensibly continuous activation (Dipatre and Gelmann 1997; Overmyer et al. 1999; Rozovsky et al. 1998; Sloane et al. 1999) Increased release of inflammatory mediators (Bodles and Barger 2004; Godbout et al. 2005; Sierra et al. 2007) Insufficient de-activation by astrocytes (Stolzing et al. 2005) 29


Figure 1-1. Schematic representation of transforma tion of ramified microg lia into hypertrophic and dystrophic forms. Microglia b ecome hypertrophic when activated by CNS injury, resulting in the formation of greatly enlarged cytoplasmi c processes. With aging, microglia develop dystrophic cytoplasmic processes ch aracterized by slight enlargemen t, distinct loss of fine branches (deramification), formation of cy toplasmic spheroids, gnarling, beading, and fragmentation. It is rare to see all of these dystrophic changes occurring in a single cell; instead, most dystrophic microglia display just one or two of these charac teristics. (Streit et al., 2004) Table 1-2. Microglial degene ration in the literature Species Method Classification Characteristics of degenerating cells Ref. Human IHC Aging Deramification, cytorrhexis (Streit et al. 2004) AD Deramification, bulbous swellings, cell clusters, cytorrhexis (Ferraro 1931; Streit 2002; Streit 2004) AD, HPC Atrophy, bulbous swellings, cytorrhexis (Flanary et al. 2007) Schizophrenia Deramification, atrophy, cytorrhexis (Wierzba-Bobrowicz T et al. 2004; WierzbaBobrowicz et al. 2005) HD Tortuous processes, cytorrhexis (Simmons et al. 2007) Rat IHC SOD1G93A Deramification, cytorrhexis (Fendrick et al. 2007) Cell Culture A exposed Atrophy, beaded processes, cytorrhexis (Korotzer et al. 1993) Mouse IHC R6/2 model of HD Deramification, bulbous swellings, cytorrhexis (Ma et al. 2003; Simmons et al. 2007) IHC, immunohistochemistry; AD, Alzheimers disease; HPC, High amyloid plaque pathology control; HD, Huntingtons disease 30


Table 1-3. Alzheimers disease-related pathologic al hallmarks develop in a hierarchical manner Stage Senile Plaque Pathology Neurofibrillary Tangle Pathology I Neocortex Transentorrhinal region II Allocortical Areas entorrhinal cortex subiculum/CA1 region Entorrhinal Cortex Ammons Horn III Basal Ganglia Thalamus Hypothalamus Hippocampal Formation IV Midbrain Medulla Oblongata Inferior Temporal Cortex V Pons Cerebellum Neocortex Association Areas Frontal Parietal Occipital VI Occipital Lobe Senile Plaque Pathology: stages 1-3= preclinical st ages; 4-5= clinical stages. Neurofibrillary Tangle Pathology: Braak stages I-II (clinically silent); III-IV= (incipient AD); V-VI= (fully developed AD). Braak Stages V-VI are conventionally used as criteri a for neuropathologic confirmation of the clinical diagnosis of AD. (Based on Thal et al., 2006; Braak and Braak, 1991). Figure 1-2. Classification of sen ile plaques based on Bodian, A and GFAP triple staining. (Oide et al., 2006) 31


Figure 1-3. Schematic diagram of the facial nerve axotomy model. Th e facial nerve nuclei containing motoneuron cell bodies ar e located on either side of the ventral region of the brain stem (dotted-line area). The motoneuron axons loop around the genu of the abducens nucleus in the dorsal region of the brainstem and exit the skull at the stylomastoid foramen to innervate the musculature of the face. In the axotomy procedure, the facial nerve is either crushed or transected after its exit from the skull, inducing a reaction that is propagaded retrogradely to the motoneuron cell bodies. Signals released from the injured neurons activate microglial cells locally, resulting in prolifera tion as well as morphological (hypertrophy) and immunophenotypic alterations in parenchymal microglia within the ipsilateral facial nucleus. 32


CHAPTER 2 MATERIALS AND METHODS Supplier Information Adobe Photoshop (San Jose, CA), Assay Design s (Ann Arbor, MI), Calbiochem (Gibbstown, NJ), DAKO (Carpinteria, CA), Chemicon (Temecula CA), Fisher Scientific (Pittsburg, PA), Media Cybernatics (Silver Spring, MD), Mol ecular Probes (Eugene, OR), MP Biomedicals (Santa Ana, CA), Promega (Madison, WI), SAS In stitute Inc. (Cary, NC), Serotec (Raleigh, NC), Sigma Aldrich (St. Louis, MO), Spot Diagnostic In struments (Sterling Height s, MI), Vector Labs (Burlingame, CA), Wako Chemicals (Richmond, VA). Tissue Specimens Human brain specimens were obtained from the Sun Health Research Center (SHRC) Brain Bank, Sun City, Arizona. Brai n tissues were obtained at au topsy from 24 subjects aged 3497 years. Subject groups were classified based on clinical history and ne uropathological findings as follows: 1Younger (mean age 36.66 2.08 y ears), nondemented individuals (Y) free of amyloidimmunoreactivity; 2Aged (mean age 79.86 8.05 years), nondemented and amyloid-free individuals (ND); 3Aged (mean age 83.43 5.19 years), non-demented individuals with high amyloidburden, designated as high pa thology controls (HPC); and 4Demented individuals (mean age 80.29 11.64 years) that met the clinicopathological criteria of AD, including high amyloid load. There was no significant difference in age between the AD and the elderly nondemented control groups (P > 0.5). The postmortem intervals (PMI) between death and tissue retrieval ranged from 1.5 to 4.83 hours. Age at death, gender, and PMI are shown in Table 2-1. 33


In order to study the poten tial relevance of postmorte m autolysis to microglial degeneration, nine additional brain tissue samp les were obtained from the Kentucky Medical Examiners Office. These specimens ranged in PM I from 3 to 20 hrs as summarized in Table 22. For the FN axotomy study, young adult (3 mont hs; n= 15) and old (30 months; n=15) male Fisher 344-Brown Norway F1 hybrid rats obt ained from the National Institute on Aging were used. Antibody List The list of primary antibodies used in th is study can be found on Table 2-3, along with specificity of the antibody, dilution used, a pplication and source. The following secondary antibodies were used: Biotinylated anti-rabbit and anti-mouse rais ed in goat (1:500; Vector Labs); Alexa Fluor-488 and -568 goat anti-mouse (1:500; Molecular Probes); and Alexa Fluor 488 and -568 goat anti-rabbit (1:500; Molecular Probes). Methods Facial Nerve Axotomy Experiments were done in accordance with the regulations of the University of Florida Institutional Animal Care & Use Committee (IACU C). Under isofluorane anesthesia, the right facial nerve of both groups of ra ts was exposed near its exit from the stylomastoid foramen and crushed with a fine pair of he mostats for 10s. Animals were euthanized at 1, 3, 7, 10, and 14 days postaxotomy (n=3 for each time point) by an over dose of pentobarbital a nd perfused with 4% paraformaldehyde (PFA). The brains were removed, postfixed for 24hr in 4% PFA, and cryoprotected by infiltration with 30% sucrose overnight. 34


Tissue Sectioning and Storage After fixation in 4% PFA for 48 hours at 4C, human brain samples were sectioned at the SHRC in the coronal plane in a vibratome at 50 m prior to shipment to the University of Florida and stored in 30% glycol solution until use. Th e brain areas represented in this study are the superior frontal gyrus, the s uperior, middle and inferior temp oral gyri, the hippocampus and amygdala. From the 50 m free-floating tissue samples obtaine d, equivalent regions of gyri, hippocampi or amygdala were cut from adjacent sections using a razor blade and rinsed with PBS. For the PMI study, serial 50 m-thick sections were cut from formalin (10%)-fixed, tissue blocks of temporal and frontal cortices in a vibratome and rinsed with PBS. For the rat axotomy study, brainstem sections containing both right (axotomized) and left (control) sides of the rat FNu were se rially cut either on a vibratome at 50 m for free floating sections or on a cryostat into 20 m coronal sections., in which case the sections were immediately mounted onto SuperFrost Pl us slides (Fisher Scientific). Single Label Immunohistochemistry After inhibition of endogenous peroxidase with 3% H2O2 for 5 min, rat and human freefloating sections were incubated for 2 hr in blocking buffer composed of 10% normal goat serum (NGS) diluted in PBS, pH 7.4. The sections were then incubated overnight at 4C with an antiserum to rabbit anti-horse spleen (L) ferr itin (Sigma) diluted 1:500 in PBS containing 5% NGS. In the case of human samples, adjacent tiss ue sections were also stained with a mouse monoclonal antibody (mAb) against human HLA-DR (LN-3; MP Biomedicals; 1:500). On the following day, the sections were incubated for 1 hr with biotinylated secondary antibodies raised in goat (anti-mouse or anti-rabbit) diluted 1:500 in 5% NGS/PBS solution. Next, the sections were incubated for 45 min with horseradish pero xidase (HRP)-avidin D (Vector Labs) diluted at 35


1:500 in PBS. Between each step, the sections were rinsed three times in PBS for 5 min. Following peroxidase development with 0.05% 3,3'-diaminobenzidine tetrahydrochloride (DAB; Sigma) and 0.01% H2O2 for 1-15 min, the sections were rinsed in PBS, mounted on subbed slides and air dried. The sections were then de hydrated through ascending series of ethanol and coverslipped out of xylene with Permount (Fisher Scientific). In the human study, a few sections were also processed for double staining immunohi stochemistry. As a negative control, omission of either the primary or sec ondary antibodies yielded no imm unoreactivity. Selected sections were counterstained with eith er hematoxylin or cresyl vi olet, dehydrated through ascending alcohols, cleared in xylenes, and coverslipped with Permount. Double Label Immunohistochemistry In double staining experiments on human br ain specimens, after development with DAB/H2O2 stained sections were incubated overnight at 4C with mouse anti-human -amyloid clone 6F3D (6F3D; Dako; 1:100) in 5% NGS in PBS. 6F3D is a monoclonal antibody reactive to amino acid residue 8-17 of the human -amyloid peptide. Some sections labeled first with LN-3 were also processed for double immunohistochemist ry using Ferritin as a primary antibody in the second reaction. After the same sequences of wash es as for the first label, the antigen was visualized by incubating biotinylated anti-mous e or anti-rabbit seconda ry antibodies (1:500; 1h at RT) followed by a 45 min incubation with HR P-avidin D diluted 1:500 in PBS. The color reaction product for A was developed using a PBS solution containing 0.05% DAB, 0.01% H2O2 and 1% cobalt chloride as the chromogeni c substrate for horseradish peroxidase. This substrate produced a dark black color which c ould be differentiated from the brown color obtained after staining with the first primary antibody. These double labeled sections were then 36


mounted, dehydrated and coverslipped as described for the single la bel procedure. In all cases, elimination of either primary or secondary antibodies yielded no immunoreactivity. Immunofluorescence Human samples Free-floating, adjacent tissue sections from the superior frontal gyri of all groups examined were pre-treated with 0.5% Triton-X in PBS for 15 min followed by 30 min incubation with 0.3% Triton-X/10% NGS blocking solution. The sections were then in cubated overnight at 4C with both LN-3 and Ferritin antibodies at the same concentrations listed above in 0.3% Triton-X/3% NGS/PBS. Next, the sections were ri nsed in PBS and incubated for 1 hr at 37C with goat anti-rabbit IgG conjugate d to Alexa Fluor 568 and goat anti-mouse IgG conjugated to Alexa Fluor 488 diluted 1:500 in 0.3% Triton-X/3% NGS/PBS. Af ter another series of PBS rinses, the sections were immersed in 70% etha nol for 5 min, followed by a 3 min treatment with Autofluorescence Eliminator Reagent (Chemicon). This fluorescence immunohistochemistry counterstain reduces the intensity of the auto fluorescent pigment lipofuscin which accumulates in the cytoplasm of post-mitotic cells with increas ing age. Next, the sections were rinsed with 70% ethanol, mounted onto slid es and coverslipped with V ectashield anti-fading mounting media (Vector Labs). Fluorescent images of tissue sections were obtained using a Zeiss Axioskop-2 Plus fluorescent microscope and photom icrographs were adjusted for contrast and brightness using Adobe Photoshop C2S. Rat samples Frozen sections at the level of the brains tem containing both axot omized and uninjured FNu of young and old rats were processed for double-label immunofluoresce nce. These sections 37


were pretreated with 10% NGS/0.3% Triton-X-100 in PBS for 1 hr at 37C, followed by overnight incubation at 4C in a primary antibody cocktail dilu ted in 5% NGS/0.1% Triton-X100 in PBS. The following primary antibody combinat ions were used: 1) anti-ferritin (1:1000) plus the microglial marker OX-42 (Serotec; 1:50 0); and 2) the macrophage marker ED1 (Dako; 1:300), plus a microglial marker against ionized calcium binding adaptor molecule 1 (Iba-1; Wako Chemicals; 1:500). Following several PBS wa shes, the sections were incubated at room temperature for 2 hr with corre sponding Alexa-Fluor 568 and Alex a-Fluor 488 IgGs each diluted 1:500 in the same buffer solution as for the prim ary antibodies. Next, the sections were washed again in PBS, coverslipped using anti-fading Vectashield mounting media with DAPI (Vector Labs) and photographed with a Zeiss Axioskop 2 Pl us microscope equipped with an RT Color Spot camera model 2.2.1 (Spot Diagnostic Instrume nts). For all staining procedures, omission of primary antibodies resulted in the absence of imm unoreactivity and served as negative controls. Morphometric Analyses and Cell Quantification Human Study The quantitative evaluation of microglial cells immunolabelle d with either anti-HLA-DR or anti-ferritin was performed in all 24 cases on adjacent sections by randomly sampling 10 microscopic fields of gray matter, each measuring 0.3 0.225 mm2. Digital images of fields were acquired through a 20x objective lens with a Zeiss Axioskop 2 Plus microscope using a Spot RT Color CCD camera and Spot RT software (Spot Diagnostic Instruments). Image Pro Plus v.4.6 image analysis software (Media Cybe rnetics) was used to manually calculate the instances (i.e., number of cells per microscopic fi eld) of microglial cells stained with either marker as well as the instances of immunoreactiv e dystrophic microglia. Only microglia with a clearly visible cell body were counted. 38


Incidences of microglial structural abnormalities were evaluated based of a set of predefined criteria (Streit 2006; Streit et al. 2004), including deramification and tortuosity of processes, cluster formation, processes cont aining small but conspicuous spheroid-like structures, and cytorrhexis. Fi gure 3-2 shows representative se ctions used for quantitative analysis of each of the four cohorts in this study. The photomicrographs presented were taken using a higher magnification lens (40x) for bett er visualization of the microglial cytoplasmic structure. All cell counts and morphometric anal yses were performed without knowledge of the case number by coding the specimens. Rat Axotomy Study The quantitative evaluation of lipofuscin (LF)-containing mi croglial cells was performed blind on coded slides using a 40x objective lens. Six adjacent FN u sections per animal at each time point postaxotomy were selected and ten mi croscopic fields per section, each measuring 200 x 400 m2, were randomly sampled within both c ontrol and axotomized sides of the FNu. Because LF residues also accumulate in other neural cells, ImagePro Plus v.4.6 analysis software (Media Cybernetics) was used to manually calc ulate the instances of LF-positive microglia in Iba-1/ED1 double labeled sections by merging di gital photographs of fluorescent material using ultraviolet, blue, and green excitation li ghts (360-460, 440-490, and 490-570 nm, respectively). Only immunoreactive microglial cells with a clearly visible cell body and a DAPI-positive nucleus were counted in order to carry out a consistent semi-quantitative evaluation of the number of cells stained be tween and within groups. Statistical Analyses Data were expressed as the mean standard error of the mean. In the human study, differences among groups were determined by one-way ANOVA followed by Bonferronis t test 39


for multiple comparisons. All data were analy zed using the software program SAS 9.1 (SAS Institute Inc.). For the rat axotomy study, statistical significance among the means was determined using the paired Students t test also using the SAS 9.1 program. In all instances, P values less than 0.05 were consid ered statistically significant. 40


TABLE 2-1 Clinical and pathological features of human cases Case ID Age Gender PMI (h) Group Younger controls (n=3) 1 34 F 3 Y 2 37 M 3.25 Y 3 38 M 2.25 Y Average 36.33 2.83 S.D. 2.08 0.52 Non-demented aged controls (n=7) 4 69 M 2.16 ND 5 73 M 2.25 ND 6 74 M 2 ND 7 81 M 2.75 ND 8 85 M 3.16 ND 9 86 F 2.5 ND 10 91 M 2.5 ND Average 79.86 2.53 S.D. 8.05 0.41 Non-demented high plaque pathology controls (n=7) 11 77 F 3.25 HPC 12 78 M 2.25 HPC 13 82 M 3 HPC 14 83 F 4.83 HPC 15 84 M 3 HPC 16 89 M 1.5 HPC 17 91 M 2.5 HPC Average 83.43 2.90 S.D. 5.19 0.98 Demented high plaque pathology cases (n=7) 18 60 M 3.33 AD 19 73 F 2 AD 20 79 M 2 AD 21 81 M 3 AD 22 85 F 1.66 AD 23 87 M 2.41 AD 24 97 F 1.5 AD Average 80.29 2.27 S.D. 11.64 0.82 Y, younger controls; ND, non-demented aged cont rols; HPC, aged, non-demented high plaque pathology controls; AD, sporadic, Alzheimers disease cases 41


TABLE 2-2. Profile of postmortem interval (PMI) cases Case ID PMI (h) Expired Age (y) Gender 1 3 40 M 2 4 57 M 3 8 60 M 4 9 47 M 5 11 57 M 6 12 15 F 7 14 54 M 8 17.5 19 M 9 20 43 M Table 2-3 Source, species reactivity, cell specific ity and dilution of primary antibodies Antigen (Ab name) Reactivity Specificity Antibody/ dilution Source L-Ferritin Human/rodent microglia/oligodendrocytes Rabbit pAb/ 1:1000 Sigma HLA-DR (LN-3) Human microglia Mouse mAb/ 1:500 MP Biomedicals Iba-1 Rodent microglia Rabbit pAb/ 1:500 Wako Chemicals CR3 (OX-42) Rodent microglia Mouse mAb/ 1:500 Serotec CD-68 (ED-1) Human/rodent macrophages Mouse pAb/ 1:300 Dako Myelin CNPase Human/rodent oligodendrocytes Mouse mAbl/ 1:500 Calbiochem 6F3D Human/rodent -amyloid peptide Mouse mAb/ 1:100 Dako Ab, Antibody; CR3, complement re ceptor type 3; HLA-DR, human leukocyte antigen D-related; mAb, mouse monoclonal antibody ; pAb, polyclonal antibody; 42


CHAPTER 3 RESULTS Morphological Analyses of Senescent Microgl ia in the Human Brain under Conditions of Normal Aging and Neurodegenerative Disease Ferritin-Positive Microglia Exhibit Abnormal Morphological Features The ferritin antibody used in this thesis spec ifically labels the light (L)-chain subunit of ferritin proteins (Kaneko et al. 1989), therefore both microglia and oligodendrocytes could be resolved by light microscopy. Both cell types we re promptly distinguished from one another based on morphological characteristics. Identi fication and distributi on of microglia was confirmed by using LN-3, an antibody speci fic for human leukocyte antigen (HLA-DR) expressed by microglia in both normal and path ological human brain (Miles and Chou 1988). Ftpositive cells exhibiting a dense perinuclear staining, a large r ounded nucleus and a lack of processes (Fig. 3-1A) fitted the morphological crit eria of oligodendrocytes, which was verified by immunostaining for CNPase, a protein expressed exclusively by oligodendrocytes in the CNS (Fig. 3-1D). Other Ft-positive cells, however, presented a highly branched morphology with small, rounded and/or irregularly-s haped nuclei (Fig. 3-1C), all of which are hallmark features of microglial cells as evidenced by immunoreactivity for HLA-DR antigens (Fig. 3-1F). Based on these morphological characteristics Ft-positive o ligodendrocytes could be readily distinguished from Ft-positive microglia (Fig. 1B, 1E). For the purposes of the present study, analysis of ferritin expression was limited to microglial cells. Figure 3-2 shows typical staini ng patterns of microglial cells labeled with either LN-3 or Ferritin antibodies in adjacent ti ssue sections. In all groups ex amined, microglia immunoreactive for HLA-DR exhibited the typical branched mo rphology of ramified microglia with small cell bodies, multiple long proximal branches extending in all directions, and extensive ramification of 43


the distal processes (Fig. 3-2A, -2C, -2E, -2G). Ferritin-positive microglia in the younger control group appeared less ramified, but still maintained mostly a branched morphology (Fig. 3-2B). In contrast, the majority of ferriti n-positive microglia in aged groups (ND, HPC, and AD) exhibited aberrant cytoplasmic structures indicative of dystrophy (Fig. 3-2D, -2F, -2H), which suggests an age-related component. Morphological Characteristics of Dystrophic Microglia In all groups examined, the majority of HLA-DR immunoreactive (IR) microglia exhibited a ramified morphology (Table 3-1). In sharp contrast, most Ft-positive microglial cells exhibited dystrophic changes, including deramification of processes, which produced unilateral ramification patterns and long, single processes de void of distal branches. Any remaining fine processes on dystrophic microglia were often unusually tortu ous and coiled (Table 3-1). In many instances, Ft-positive microglial cells appeared in clusters of two or more cells, suggesting that dystrophic cells ha d lost contact inhibition that normally keeps microglia apart. Another defining characteristic of Ft-positive, dystrophic micr oglia was the formation of spheroid-like structures occurring either sing ly or in succession to produce beading. These changes often coincided with at rophy of processes. A cardinal feature of microglial dystrophy was cytorrhexis, in which cytoplasmic processes were broken up into two or more parts. These defining dystrophic morphological ch aracteristics are depicted in Table 3-1, along with their frequency in the various cohorts examined in th is study. What appears to be the most severe form of cytoplasmic deteriorati on, cytorrhexis, was most widespread in AD tissues than in agedmatched controls, while in younger brains, no examples of microglia l cytorrhexis were encountered. 44


Ferritin-Positive Microglia Are Less Abundant than HLADR-Positive Microglia and Appear Mostly Dystrophic Comparing the total number of stained cells with either marker showed that only a fraction of microglial cells was immunoreactive for L-ferritin (Fig. 3-3A). The difference in the number of ferritin-positive versus HLA-DR-positive microglial cells was highly significant for all the groups examined (P < 0.01). The number of HLA-DR-positive microglial cells increased significantly with age (P < 0.05) (Fig. 3-3A), particularly with AD pathology, in accordance with previous reports (Carpenter et al. 1993; Luber-Narod and Rogers 1988). Although comparison of the number of HLA-DR-positive microglia in the ND and HPC groups showed no significant difference (P > 0.5), the number of HLA-DR-p ositive microglia was significantly different between all other groups (i.e., Y and ND, Y and HPC, Y and AD, HPC and AD groups) (P < 0.01). In contrast to HLA-DR expression, microglial L-chain Ft expression appeared to decrease with advancing age (P < 0.05), although a smalle r subject pool in the young control group (n=3 versus n=7 in aged groups) may preclude a conc lusive comparison of th e number of stained microglial cells among aged groups and young contro ls. In HPC and AD tissues, the number of Ft-IR microglia was slightly higher than in ND ti ssues but this difference did not reach statistical significance. Figure 3-3B shows the number of stained cells quantified in Fig. 3-3A that presented dystrophic changes. Microglial dystrophy was more pr evalent in Ft-positive microglia than in HLA-DR-positive microglia for all groups (P < 0.01) except for the Y subjects (P > 0.5). Table 3-2 shows the percentage of microglial cells stained with eith er LN3 or Ferritin (Fig. 3-3A) that shows dystrophic changes (Fig. 3-3B). In younger brains, about 8% of Ft-IR microglial cells and 4% of HLA-DR positive microglia exhibited dys trophic changes. In ND tissues, approximately 5% of the HLA-DR-IR microglia were found to exhibit dystrophic changes whereas 83% of Ft45


positive microglia appeared dystrophic. In the HPC group, HLA-DR-positive, dystrophic microglia amounted to 11% of the total number of LN-3 stained microglial cells. For this group, approximately 57% of the Ft-IR microglial cells exhibited signs of dystrophy. Mi croglial cells immunoreactive for HLA-DR antigens in the AD tissu es were found to exhibit mostly ramified and hypertrophic profiles, with only about 9% of these cells showing signs of dystrophy. In contrast, 83% of Ft-IR microglial cells in AD tissues possessed a dystrophic morphology. In younger brains, Ft-positive dystrophic microglia were less conspicuous in comparison with Ft-IR dystrophic microglia in aged tissues. Moreover, in regards to the instances of dystrophic microglia immunoreactive for HLA-DR antigens, independent t-tests showed statistically significant differences (P < 0.01) in multiple comparisons between all groups, except for HPC versus AD (P > 0.5) (Fig. 3-3B). Similar comparisons were made for the instances of Ft-positive dystrophic microglia. The only non-si gnificant comparison found was between the ND and HPC groups (P > 0.5) (Fig. 3-3B). Ferritin-Positive Microglia Constitute a Subset of the Larger Microglial Pool Direct evidence for the finding that Ft-pos itive microglia constitute a subpopulation of the overall microglial pool was obtained by fluore scently double labeling microglial cells with both HLA-DR and Ft markers (Fig. 3-4). In immunofluorescence staini ng preparations, HLADR-positive microglial cells outnumbered Ft-positive microglia, as evidenced by comparing single labeling for each marker. By merging color channels, it was confirmed that only a subset of microglial cells express L-Ft by co-localizatio n with HLA-DR antigens. Moreover, in a given microscopic focal plane, Ft immunoreactivity was found to occur throughout the extent of microglial cell processes in a similar manner as for HLA-DR immunoreactivity, suggesting a similar cellular localizat ion for the antigens anal yzed in this study. 46


Microglial Dystrophic Changes are Not Du e to Postmortem Tissue Autolysis Postmortem delay defines the time interval in hours from death to tissue preservation. In order to determine whether or not microglia l dystrophic changes could be caused by tissue autolysis during prolonged post-mortem intervals (PMIs), ferritin immunohistochemistry was performed on the frontal and temporal cortices of 9 forensic cases with postmortem delays ranging from 3 to 20 hours (Table 2-2). The su bjects studied included 8 men and one woman; their mean age at death was 43.56 years with a st andard deviation of 16.52 years. Review of clinical records and general au topsy reports revealed no eviden ce of neurologic disease, drug intake, or metabolic disease. Moreover, temperat ure, mode of fixation an d storage, and staining technique were thor oughly controlled. At a PMI of 3 hrs, Ft-IR microglial cells exhibited a mostly branched morphology (Fig. 3-5A), although distal processes were largely devoid of fine ramifications. Similar morphological characteristics were observed at lo nger PMIs (Fig. 3-5B-D), and most Ft-positive microglia in these tissues displayed deramified and beaded processes as described before (Table 3-1). Instances of deramified microglial cells rema ined constant from shortest to longest PMI. However, non-specific background st aining was found to increase after a PMI of 8 hrs. These results suggest that Ft antigenicity is resistant to postmortem delay autolysis. In contrast, HLADR immunoreactivity, as demonstrated by LN-3 staining, proved very sensitive to prolonged postmortem delay and was of poor quality follo wing immunohistochemist ry (data not shown). Taken together, our results sugge st that the structural change s associated with microglial dystrophy are not associated with postmortem interv al tissue autolysis, bu t instead to inherent metabolic changes. 47


Ferritin-Positive Dystrophic Microglia are Prom inent in the Brains of Alzheimers Disease Patients In our evaluation of microglial dystrophy, it beca me apparent that dystrophic microglia in aged specimens differed in the extent of cytoplasmic degenera tion from those in younger brain tissue samples. Consequently, we set out to analy ze in detail the morphologi cal characteristics of both ramified and overtly dystrophic (i.e., cyto rrhectic) microglia immunoreactive for both HLADR and Ferritin markers in all groups examin ed. In both control and AD tissues, HLA-DRpositive microglia predominantly exhibited a hi ghly branched morphology including distal ramifications (Fig. 3-6A-D). However, a fe w HLA-DR immunoreactive microglia presented signs of dystrophy (Fig. 3-6B). Cytorrhexis, whic h appears to be the most severe form of microglial degeneration, could be observed in HLA-DR-positive microglia but only in aged groups (Fig. 3-6F-H) and not in young controls (F ig. 3-6E). In most cases, only a small number of HLA-DR-positive microglia in younger control subjects presen ted deramification of processes (Fig. 3-6E), which often did not progress into further degenera tion of cytoplasmic structure. Cytorrhexis is thought to progress from deramifi ed, beaded processes (as shown in Fig. 3-6F) into a complete fragmentation of cytoplasmic processes (Fig. 3-6G-H). Ferritin-positive microglial cells ramified to the same extent as HLA-DR-IR microglia were rarely observed. For the most part, Ft-posi tive microglia never fully exhibited the branched morphology of HLA-DR positive microglia. Instead, they appeared in a deramified, atrophied morphology regardless of the age group in question (Fig. 3-6I-L). The majority of L-Ft-positive microglia in aged groups displayed all of the previously established characteristics of dystrophy (Table 3-1), while in young control subjects most Ft-positiv e microglia presented only a deramified morphology (Fig. 3-6M). Cytorrhectic microglia in ferritin-stained sections were 48


more readily identified as well as more pronounced than in LN-3 stained sections and they were largely limited to aged groups (Fig. 3-6N-P and Table 3-1), particularly in AD brain specimens. In contrast to the other two aged groups examined, only 57% of Ft-positive microglia in HPC tissues appeared dystrophic (Fig. 3-3B; Tabl e 3-2). The remainder of Ft-IR microglial cells did not fit the criteria for microglial dystrophy as outlined in Table 3-1, but their processes were often stripped of fine ramifications similar to Ft-positive microglial cells in younger tissues (Figs. 3-6I, 3-6M). AD tissue specimens exhib ited the most severe dystrophic changes compared to the other groups, both for HLA-DR-positive and Ft-positive microglia. Morphological analyses of microglia in the proximity of senile plaques Because microglial cells exhibiting characte ristics of cytoplasmic deterioration were readily noticeable in the AD brai n (Figure 3-7), we subsequently wanted to know whether there was any relationship between AD-re lated histological pathologies (i.e., SP and NFT) and the occurrence of degenerative featur es in microglia. Therefore, to determine whether the incidence of dystrophic microglia was higher in the pr oximity of extracellu lar deposits of A proteins (senile plaques), double immunohistoche mistry was performed for both A (6F3D) and microglia (Ferritin or LN3). Although amyloid de position is considered a prerequisite for the unequivocal diagnosis of AD, they also occur in the brains of non-demented individuals, albeit to a lesser degree and in a more restricted dist ribution (Arriagada et al. 1992). Consequently we chose to analyze and compare the inst ances of dystrophic microglia near A deposits in both cognitively normal (HPC) and de mented (AD) individuals. In both HPC and AD tissues, HLA-DR-positive microglia showed mostly a ramified morphology (Fig. 3-8A-B), even in close proxim ity to SPs (Fig. 3-8C-D). These double labeling experiments corroborated the quantitative analysis of the proportion of dystrophic microglia in 49


HLA-DR immunoreactive microglia (Fig. 3-3B), which indicated that only approximately 10% of HLA-DR-positive microglia in HPC and AD tissues presented signs of degeneration. The dystrophic microglia in these ti ssues showed predominantly mo rphological features of atrophy and spheroid formation. Moreover, the few dystrophi c microglia that could be identified in these tissues appeared randomly distributed throughou t the brain parenchyma. Thus, there was no obvious spatial correlation between HLA-DR-pos itive, dystrophic microglial cells and SPs. Similarly, Ft-positive dystrophic microglia were not concentrated around SPs, but scattered throughout the brain pa renchyma of both HPC and AD ti ssues (Fig. 3-9). Taking into consideration that the in cidence of dystrophic changes is high er in Ft-IR microglia (Fig. 3-3B), most Ft-positive microglia in both AD and HP C tissues appeared morphologically aberrant irrespective of proximity to A deposits. Interestingl y, this result was already apparent in our quantification of dystrophic microglia in Fig. 3-3B; A deposition as a common denominator between HPC and AD tissues appeared to be insu fficient in promoting dystrophic changes, since only roughly half of all Ft-posi tive microglia in HPC tissue samples appeared dystrophic while the large majority (about 90%) of Ft-positive microglia in AD subjects showed signs of dystrophy (Fig. 3-3B). Perhaps because of the hi gher incidence of dystrophic changes in Ft-IR microglia compared to HLA-DR-positive microglia nearly all Ft-positive microglia appeared dystrophic at or near SPs (Fig. 3-3C-D). Brain Specimens Used in This Study Represent Early AD Because we did not observe any correlation in the proximity of microglial cells exhibiting dystrophic features to A deposits in both HPC and AD brain sp ecimens, next we were interested in determining the identity of the SPs present in our brain specimens. We noticed that the 6F3Dimmunoreactive SPs were prevalent mostly in th e neocortical brain regions (SFG, temporal 50


gyrus), particularly in corti cal layers III-V (Fig. 3-10A). Morphologically, the stained SPs appeared amorphous and surrounded by a well-define d outline contour, all of which are defining characteristics of diffuse SPs (Fig. 3-10B). Th is morphological character ization was corroborated by the finding that most microglial cells surr ounding the plaques appeared ramified and not hypertrophic (Fig. 3-10C). We also attempted to stain NFTs in adjacent tissue sections using various established markers of NFTs (CNPase [Calbiochem]; PHF-Tau clone AT8 [Thermo Scientific]; RIP [BD Transducti on labs]) and immunohistochemical procedures with no success, however (data not shown). The predominance of di ffuse SPs in the neocortex and the absence of NFT pathology and neuritic plaques suggest that our brain specimens are staged in the early phase of the disease. Analyses of Microglial Ferritin Immuno reactivity in Young and Aged Rats Acute Activation of Microglial Cells Does Not Induce Ferritin Expression Because dystrophic microglia exhibit some morphological similarities with activated microglial cells (Lopes et al. 2008; Streit 2006), namely deramification and spheroid formation (Table 3-3), we subsequently wanted to dete rmine whether microglial activation induces Ft expression and if perhaps the appearance of dystrophic microglia is increased under injuryinduced activation conditions in aged animals. To this end, we analyzed Ft immunoreactivity in microglial cells activated in vivo by the well-characterized FN axotomy paradigm in both young and aged rats. To examine whether Ft immunoreactivity is upre gulated in activated microglial cells, the right FN of rats was acutely injured while leaving the left FN intact to serve as an internal control. Brain sections from the axotomized rats were then immunostained for L-ferritin. No difference in staining patterns was found between the axotomized and control (contralateral) 51


sides of the rat FNu, regardless of the time interv al following the nerve inju ry or the age of the animal (Figure 3-11). However, Ft immunoreactivity within the FNu increased with age,. At 3 months of age, Ft immunohistoche mistry showed a light staining of both axotomized and control FNu where only a few intensely stained cells could be identified (Figure 311, top panel). In rats aged 30 months, there was a visible increase in Ft background staining in both sides of the FNu, which appeared to spread to adjacent areas accompanied by an increase in the number and distribution of immunoreactive cells (Figure 3-11, bottom panel). In order to find out whether the few Ft-IR cells observed in the FNu were microglial cells, double immunofluorescence staining was performed using both Ft and microglial markers. Microglial cells were visualized by staining with OX-42, a mouse monoclonal antibody (mAb) that specifically recognizes type 3 complement receptors (CR3) expressed by both ramified and activated microglia (Ling et al 1992; Milligan et al. 1991). I nduction of microglial activation was confirmed by changes in cell morphology (hype rtrophy), cell number (proliferation), and immunophenotype (upregulation of CR3 antigens) within the axotomized FNu of both young (Figure 3-12B) and aged (Figure 3-12H) rats compared to the c ontrol side (Figures 3-12A, 312G). This microglial activation response did not induce an upregul ation of Ft, as evidenced by an equivalent number of Ft-posi tive cells in both sides of the FNu (Figures 3-12C-D, 3-12I-J). However, a higher number of Ft-IR cells was observe d in aged rats (Figure 3-12I-J) compared to the younger group (Figure 3-12C-D) in accordance w ith the increased staining shown in Figure 1. In both axotomized and control FNu these Ft -IR cells did not colocalize with OX-42-IR microglia (Figures 3-12E-F, 3-12K-L), although a few did in aged animals (Figure 3-12K-L). Although the majority of Ft-positive cells within the rat FN u did not colocalize with CR3 antigens, their identity could be resolved based on defining mo rphological characteristics. Ft52


positive cells had a larger and rounder nucleus than OX-42-positive microglia as well as fewer (if any) processes stained (Fi gure 3-13), and these morphological features identified them as oligodendrocytes by comparison with previously published oligodendroglial immunohistochemical data using L-ferritin (Che epsunthorn et al. 1998; Cheepsunthorn et al. 2001; Koeppen and Dickson 2001; Ogawa et al. 1994). Ferritin Immunohistochemistry Labels Predom inantly Oligodendrocytes in Rat Brains Our Ft histological studies on human brain specimens indicated that only a subset of microglial cells express L-rich ferritins (see Figs. 3-3 and 3-4). We noticed similar results in FNu sections in rats (Fig. 3-12). In order to determine the validity of our result s in rats, we decided to also stain rat coronal sections that contained the hippocampus, an area previously established to contain a high number of L-Ftpositive cells (Huang and Ong 2005). Our histological analysis corroborated earlier studies in that Ft-pos itive cells were found predominantly in the hippocampal region of rats, irrespective of the age of the animal (Fig. 3-14). Morphological evaluation of these Ft-positive cells under high power magnification revealed cells with a darker perinuclear staining and few stained processes, all of whic h are suggestive of oligodendroglial phenotype (Fig. 3-14). Our histol ogical characterization of LFt-positive cells in the rat hippocampus was supported by our double immunohistrochemisry findings that OX-42-positive microglia do not colocalize with L-Ft markers (Fig. 3-15). In addi tion, Ft-IR cells did not colabel with OX-42-positive microglia in either cortical or cerebellar regions (Fig. 3-15). Moreover, these Ft-IR cells presented a morphology different from OX42-IR microglia, including a rounder and slightly larger perinuc lear regions and fewer, punctate (thin) branches, all of which are hallmark characteristics of oligodendroglial cells. 53


Morphological Analyses of Microglial Sene scence in the Rat Brain under Conditions of Normal Aging and Acute Injury Accumulation of Lipofuscin Granules are Pr evalent in Dystrophic Microglia of the Aging Rat Brain Since the FN axotomy procedure performed represents a mild, reversible injury that does not lead to debris accumulati on, no phagocytic activity by activa ted microglial cells should take place. As stated previously, microglial cells ha ve been shown to senesce with advancing age (Flanary and Streit 2003; Streit 2006), the out come of which remains unknown. Therefore, our next aim was to investigate whether there were any microglial cells in th e aged rat brain that would exhibit an alte red activation response. Since pr ior work had reported microglial hypertrophy in aged rats (Conde and Streit 2006) we wanted to investigate whether these enlarged microglia were expressing lysosomal proteins normally found in phagocytic cells. ED1 is a mouse mAb that recognizes a single chain glycoprotein of 110kDa expressed predominantly on the lysosomal membrane of phagocytic cells (i.e., microglia/macrophages) (Dijkstra et al. 1985). Double labeling immunofluorescence was, theref ore, performed in axotomized rat brain tissues using both ED1 and ionized calcium bindi ng adaptor molecule 1 (Iba-1) markers, the latter of which is a rabbit polyclonal antibody whose antigen is specifically expressed in microglial cells (Ito et al. 1998). As expected an d confirming prior by Graeber et al. (1998), our immunofluorescence analyses revealed that ED1 immunoreactivity is absent in both the unoperated and axotomized FNu of young rats (Fig. 3-16A-B). However, in aged brain, a large fraction of Iba-1-positive microg lia co-localized with ED1 in aged tissues, however (Figure 316C-D). Surprisingly, these ED1-IR microglia were found predominantly and more conspicuously in the control side of the FNu (Fi g. 3-16C) compared to the axotomized side (Fig. 3-16D), indicating that expression of ED1 antigen was unrelated to microglial activation. 54


Upon closer examination of the ED1-positiv e microglia, we noticed that in the nonaxotomized FNu these cells had shorter, less co mplex branches and a stout perinuclear region than ED-1-negative microglia (Fig. 3-16E). Although less rami fied, there was no hypertrophy of major processes as is typical of activated microglial cells In the axotomized FNu, the morphology of ED1-positive microglia was indisti nguishable from ED1-negative microglia (Fig. 3-16F). Lastly, we noticed that ED1 antigens accumulated preferentially in the perinuclear area of the cytoplasm (Fig. 3-17). Interestingly, we also noticed that the ED1-positive signal in microglia was detectable by fluorescence microsc opy across a wide spectr al range (excitation: 350-580 nm; emission: 400-603 nm) (Fig. 3-18), s uggesting that the ED1 positive signal was congruent with autofluorescence. Autofluorescence refers to the intrinsic emission of light from endogenous compounds other than the fluorophore of intere st. It is a common occurrence in aged tissues, due primarily to the accumulation of autofluorescent lipofuscin (LF) granules within long-lived cells (Brunk and Terman 2002; Terman and Brunck 2004). In our an alyses, all ED1-positive signal in microglia coincided with autofluorescent LF granules by exc itation with ultraviolet, blue, and green light as illustrated in Figure 3-18A-C (also see the merged image in Fig. 3-19A). Although LF appeared in all morphological subt ypes of microglia, their intensity and size was more noticeably visible in microglial cells exhibiting signs of dystrophy (Figure 3-19A-C), namely loss of distal, and sometimes proximal, branches compared to LF-negative ramified microglia (Figure 3-19D). Most LF granules in Iba-1-IR microglia accumulated in the nuclear periphery (Figures 3-19A-B). Occasionally, we also found LF pa rticles localized within cytopl asmic spheroids away from the perinuclear cytoplasm (Figure 319C). Spheroid formation in hu man microglia is considered a characteristic trait of dystr ophy. In regards to the anatomi cal localization, LF-containing 55


microglia were found predominantly in the non-a xotomized FNu of aged rats (Table 3-4), reinforcing the idea that LF accumulation occu rs independent of microglial activation. Figure 3-1. Morphological characteri stics differentiate ferritin-positive microglia from ferritinpositive oligodendrocytes. Photomicrographs taken from serial SFG sections from an 85-yearold ND individual. (A) L-ferritin-positive oligode ndrocytes exhibit strong perinuclear staining, few processes and a large nucleus, characteristics that facilitate the differentiation of ferritinpositive oligodendrocytes (B, arrowhead) from ferri tin-positive microglia (B, arrow). In turn, ferritin-immunoreactive microglia pr esent fine cytoplasmic ramifica tions and a small, rounded or irregularly shaped nucleus (C). Oligodendrocyte-like morphology in (A-B) was verified by comparison to cells immunoreactiv e to the oligodendrocyte-specific marker CNPase (D). Likewise, microglial morphology in (B-C) was similar to HLA-DR -positive microglia (F). In gray matter, most ferritin-positive cells present a branched morphology similar to microglia (E). Scale bar = (A-C) 10 m, (D-F) 3 m. 56


Figure 3-2. Representative photomicrographs of the number of microglia immunoreactive for either HLA-DR antigens or ferritin proteins. Microglia immunoreactive for either HLA-DR antigens (A, C, E, G) or L-ferritin proteins (B, D, F, H) in adjacent SFG sections of a 34-year-old control individual (A-B), an 81-ye ar-old control individual (C-D ), an 83-year-old HPC subject (E-F), and an 81-year-old AD subject (G-H). The photomicrographs are representative of the number of microglial cells stained with each ma rker. Note the extensive branched morphology of ramified microglia immunoreactive for HLA-DR. Lferritin-positive microglia exhibit mostly a dystrophic morphology. In all groups, microglia l labeling was less prominent for ferritin expression (B, D, F, H) than for HLA-DR antigen expression (A, C, E, G). Scale bar = (A-H) 20 m. 57


Table 3-1. Classification of mi croglial dystrophic characteris tics based on ferritin immunohistochemistry. Microglial morphology in the temporal cortex of an 81-year-old ND individual (HLA-DR stained) and in an 81-year-old AD patient (Ferritin stained). Frequency refers to the predominance of a given morphology in each cohort under study. Symbols: -, absent; +, rare; ++, common; +++, prevalent. 58


Figure 3-3. Comparison of the average numbe r of immunoreactive (A) and dystrophic (B) microglia stained for either HLA-DR antigens ( open bars) or ferritin pr oteins (closed bars). After scoring the number of immunoreactive microglia for each marker in (A), each immunoreactive microglial cell was evaluated qu alitatively based on a predefined set of morphological criteria for dys trophic changes (B). Valu es are means SEM. p < 0.05 versus HLA-DR for each group. 59


Table 3-2. Proportion of HLA-DRand ferriti n-immunoreactive microglia in human brain specimens that exhibit dy strophic characteristics. Y ND HPC AD HLA-DR 4% 5% 11% 9% Ferritin 8% 83% 57% 89% Figure 3-4. Ferritin immunoreactive microglia cons titute a subpopulation of the larger HLA-DRpositive microglial pool. Immunofluorescence micr oscopy of microglial cells for HLA-DR (green) and ferritin (red) antigens in temporal lobe cortex of a 34-year old Y individual (a-c) and in a 77-year-old HPC subject (d-f). Merged im ages of both channels are included, with yellow representing overlapping signal. No te that not all microglia express ferritin proteins (b, e). Scale bar = (a-f) 20 m. 60


Figure 3-5. Postmortem interval study. Representative photomicrographs of L-ferritin immunoreactive microglia in cortical specimens with ascending postmortem intervals (PMIs): (A) 3 hrs, (B) 8 hrs, (C) 14 hrs, and (D) 20 hrs. No increases in the instances of microglial dystrophy were observed with increasing PMI. Scale bar = (A-D) 10 m. 61


Figure 3-6. Cytorrhectic microglia are present in aged but not in young human brain tissues. Adjacent serial sections of superi or frontal cortex of a 37-year ol d Y individual (A, E, I, M), an 85-year old ND individual (B, F, J, N), an 84-year old HPC subject (C, G, K, O), and an 85-year old AD patient (D, H, L, P) were immunohistoche mically labeled for either HLA-DR (A-H) or L-ferritin antigens (I-P). Instances of ramified microglial cells we re readily visible in HLA-DRstained sections for all groups examined (A-D). A small number of HLA-DR-positive microglia presented a cytorrhectic morphologi cal profile in all ag ed groups (F-H) excep t in young controls (E). Cytorrhexis appears to pr ogress from thin, beaded proce sses (F) to overtly cytoplasmic degeneration (G-H). These HLA-DR-positive dy strophic microglia were observed scattered randomly in the brain parenchyma among ramified microglia and were rela tively few in number. In contrast, ferritin-positive microglial cells never fully exhibited the highly branched morphology of ramified microglia (I-L), includin g in young controls (I). Cytorrhectic microglia were more numerous in ferritin-stained secti ons for all aged groups (N-P) except for in young controls (M), in which ferritin-positive microglia only exhibited a deramified profile. Scale bar = (A-P) 20 m. 62


Figure 3-7. Microglial cells in AD specimens exhibit abnormal morphological features. Photomicrographs of microglial cells immunor eactive for HLA-DR antigens in the frontal cortical gray matter of a 38-y ear-old individual (A) and an 87-year-old AD subject (B). (A) Ramified microglial cells predominate in th e tissue from younger individuals. (B) In the ADtissue, the incidence of dystrophi c microglia increases, although ra mified microglial cells are still visible. Here, signs of dystrophy included deramifi cation, formation of multi cellular clusters, and spheroid formation. The higher incidence of dystr ophic microglia in AD brai n tissues implicates a role for microglial senescence in AD pathogenesis. Scale bar: (A-B) 100 m. 63


Figure 3-8. HLA-DR-positive microglia (brown) in the vicinity of senile plaques (black) present mostly a ramified morphology in bot h HPC and AD tissues. Nearly all HLA-DR immunoreactive microglia around senile plaques exhibited a ramified morphology in the temporal cortex of an 89-year old HPC indi vidual (A,C) and in an 87-year old AD subject (B,D). A few dystrophic microglial cells could be identified in these tissues and they were for the most part scattered randomly in the brain parenchyma irrespective of the location of amyloid deposition. Scale bar = (A,B) 10 m; (C,D) 20 m. 64


Figure 3-9. Ferritin-positive micr oglia (brown) present mostly a deramified profile in both HPC and AD tissues irrespective of th e vicinity of senile plaques (black). Adjacent serial sections labeled for HLA-DR antigens in Figure 7 were labeled for ferritin proteins in the 89-year old HPC (A,C) and 87-year old AD (B,D) subjects. Dystrophic microglial cells could be readil y identified in these tissues. They were distributed randomly in the brain parenchyma, occurring both near and away of amyloid deposits. Ferritin-positive dystrophic microglia within senile plaques often exhibited gnarling and deramification of processes (C) as well as thin, tortuous, beaded processes (D). Scale bar = (A,B) 10 m; (C,D) 20 m. 65


Figure 3-10. Morphological characteristics of senile plaques in the AD brain specimens under study. These photomicrographs were taken from the SFG of a 85 year old AD patient and are representative of other specimens within the AD cohort. (A) Senile plaques immunoreactive for 6F3D (black) concentrate in co rtical gray matter. (B) These plaques appear homogeneous and with a sharply defined rim, all of which are ch aracteristics of diffuse plaques. (C) HLA-DRpositive microglia (brown) maintain a ramified morphology in the proximity (and even within) the plaques. All of these charac teristics pinpoint an early stag e for the AD brain specimens under study. 66


Table 3-3. Morphological Characteristics of Activated and Dystrophic Microglia Dystrophic Microglia Activated Microglia Hypertrophy + + + Atrophy Deramification + + + Beading + + Spheroids + + Fragmentation + Symbols: absent; less pronounc ed; + present; + + often present; + + +, always present. Figure 3-11. Ferritin protei n levels increase with ag e in the rat. Ferritin immunohistochemistry through the brains tem containing the facial nucleus (outlined area) of a 3-mont h-old (top panel) and a 30month-old (bottom panel) rat at 3 days postaxotomy. Light ferritin staining in younger tissues increased in intensity and size with age. No diffe rence in ferritin immunoreactivity was observed between control (ast erisk) and axotomized side s of rat facial nucleus. Scale bar = 500 m applies to both panels. 67


Figure 3-12. L-ferritin expression is induced by the aging process in microglial cells and not by acute activation conditions. Doubl e-immunofluorescence staining of th e facial nucleus region of both young (A-F) and aged (G-K) rats using th e microglial marker OX-42 (red) and ferritin (green). OX-42-immunoreactive micr oglia in the unoperated side of the facial nucleus appear ramified and evenly distributed both in young (A ) and aged (G) rats. Within the axotomized facial nucleus, OX-42-positive microglia are more numerous and hypertrophic (B, H). Few ferritin-positive cells were identified in the rat facial nucleus, irrespecti ve of whether or not motoneurons were injured (arrowheads in C-D, I-J). Most of these ferritin-positive cells did not co-localize with the microglial marker (E-F, K-L) although some colocali zation was apparent but in aged tissues only (K-L, arrows). Yellow colo r in merged images (K-L) defines co-expression. Insets show a higher magnification view of the double labeled cells. Scale bar= 10 m (A-L). 68


Figure 3-13. Cellular characterization of fe rritin-positive (red) and OX-42-positive (green) cells within the rat facial nucleus. The majo rity of ferritin-immunoreactive cells did not colocalize with the microglial marker OX-42. More over, these cells had a larger cell nucleus and fewer processes compared to OX-42-pos itive microglia, fitting the morphological appearance of oligodendro cytes. Scale bar= 10 m 69


Figure 3-14. Ferritin positive cells are prev alent in the rat hippocampus. Ferritin immunohistochemistry on a coronal section of a 3month-old rat at the level of the hippocampal formation. Composite of low ma gnification images (2.5x objective lens) show that the highest density of ferritin-posi tive cells are concentrat ed in the CA3 region of the hippocampus. At higher magnification, the morphology of these cells be come apparent as that of microglia, with small cell bodies and numerous processes. 70


Figure 3-15. Double immunofluorescence staini ng for ferritin and CR3 receptors in the hippocampal, cortical, and cerebe llar regions of a 30-month-old rat. L-ferritin-imunoreactive cells (green) do not colocalize with CR3-positive mi croglia (red) as shown in merged images. FtIR cells exhibit a larger and more densely stai ned perinuclear regions and fewer IR processes compared to OX-42-labeled microglia. 71


Figure 3-16. Immunofluorescence st aining using Iba-1 (green) and ED1 (red) in th e control (A, C) and axotomized (B, D) facial nuclei of a 3-month-old rat (A-B) and a 30-mo old rat (C-D) 10 days postaxotomy. There is no co-labeling betw een Iba-1 and ED1 immunoreactive microglia in younger brains (A-B). However, colo calization of the markers is read ily visible (yellow) in aged tissues (C-D). (E-F), higher magnification of box es in C-D. Note that ED1 immunoreactivity (red) is confined to the perinuclear region of OX-42 positive microglia (green). (E) ED1-positive microglia exhibit a dystrophic morphology (arrowhead) in the control facial nucleus compared to ED1-negative microglia (arrow). (F) The mor phology of ED1-positiv e microglia in the axotomized FN (arrowhead) is indistinguishable from that of ED1-nega tive microglia. Scale bar = 100 m (A-D), 20 m (E-F). 72


Figure 3-17. ED-1 immunoreactivity (red) is lim ited to the perinuclear region (B) of Iba-1positive microglia (green; A). Yellow color in (C) indicates co-loc alization. DAPI-positive nuclear DNA is depicted in (A-C). Scale bar = 10 m (A-C). Figure 3-18. Autofluorescence of lip ofuscin particles using ultravio let (A), green (B), and blue (C) excitation lights (360-460, 490-570, and 440-490 nm, respectively). Scale bar= 20 m (A-C). Lipofuscin granules (arrowhead) can be seen with various wavelengths of light in the same cell. Scale bar= 20 m (A-C). 73


Figure 3-19. Lipofuscin granules accumulate in senescent dystr ophic microglia of aged rats. Fluorescence staining using Iba-1 (green) and ED1 (red) immunohistochemical markers in the non-axotomized facial nucleus of a 30-month-ol d rat. Photomicrographs represent merged images taken with ultraviolet, green, and blue ex citation lights. Cell in (A) is the merged image depicted in Fig. 3-18A-C. The majority of Iba1 immunoreactive microglia (green) that exhibit signs of dystrophy are double labeled for lipofus cin (red), which appears yellow in merged images (A-C). Dystrophic morphologies in thes e cells include an abnormal enlargement of the perinuclear region (arrowhead in A-B), which often occurs in severely deramified microglia, and spheroid formation (C) at distal branches. (D) Representative image of the ramified morphology of lipofuscin-negative, OX-42-positive (gre en) microglia. DAPI-positive nuclear DNA is depicted in (A-D). Scale bar= 10 m (A-D). 74


Table 3-4. Cell counts of immunoreactive microglia in the unoperated and axotomized facial nucleus Young Group Aged Group Control F.N. Axotomized F.N. Control F.N. Axotomized F.N. Iba-1 10.05 0.7 30.83 10.6* 10.55 0.13 30.70 7.73* Lipofuscin 0.3 0.2 0.20 0.11 2.28 0.49** 1.68 0.13** Values are expressed as the mean S.E.M. Symbols: F.N., facial nucleus; p <0.05 compared with the corresponding numbers in the unoperated facial nucleus; ** p <0.001 compared with the corresponding numbers in younger animals. 75


CHAPTER 4 DISCUSSION AND CONCLUSION Overview of Findings The primary aim of this investigation was to compare and contrast degenerative morphological traits in microglia l cells of both human and rode nt brains. Our histological analyses provide evidence that although microgl ial degeneration is not a unique phenomenon to humans, the degree of structural deterioration that microglia unde rgo as a function of senescence in aging brains differs between humans and rats (see Table 4-1). In humans, microglial dystrophy appears to be an elaborated process commenci ng with thinning and deramification of distal processes followed by the formation of bulbous swellings along remaining branches, atrophy, and finally complete degeneration of the microgl ial cytoplasmic structure. We found that the end-stage of degeneration (cytor rhexis) is particularly widesp read in cortical gray matter microglia of the early-stage Alzh eimers disease brain. In turn, dystrophic microglia of aged rats do not develop as many stages and/or severity of degeneration. Furt hermore, ostensibly degenerative microglia in rodents exhibit a consid erable enlargement of the cytoplasmic regions immediately adjacent to the cells nucleus, the principal location for many organelles, including lysosomes. These so-called pot bellies are abse nt in dystrophic microglia of human brains. In regards to similarities, we noticed that dystroph ic microglia in both human and rodent brains present primarily a deramified morphology. In some instances, s pheroid-like structures along the remaining stripped-down processes can be found in both species. In this chapter, I will address and interpret ou r findings taking into account the possibility that overtly dystrophic microglia of aged human brains may facilitate the pathogenicity of agedependent neurodegenerative disorders, such as AD. In light of the differences we found between dystrophic microglia in human versus rat brains (Table 4-1), it seems worthy of note 76


that although rodents are also aff ected by deterioration of cellular and systemic functions that are associated with the normal aging process, no neur odegenerative disorder has ever been shown to occur spontaneously in rodents. Subsequently, I w ill consider here the possibility that one of the reasons why rodents do not develop neurodegenerative diseases is re lated to their lack of overtly dystrophic microglia. Moreover, the findings repor ted herein underscore the importance of a thorough evaluation of the microglia l morphological status in order to appropriately assess their contribution to age-related neurodegeneration and neurodegenerative diseases. As a final point, our immunohistochemical analyses of senescent microglia also reduced the listing of potential causes of cytoplasmic degeneration by providing evidence that microglia immunoreactive for the iron storage protein ferriti n are significantly more likely to ex hibit dystrophic traits with aging than ferritin-negative microglia. The implication here is that the degeneration of microglial cytoplasmic structure may be the outcome of iron-mediated free-radical reactions. Only a Subset of Microglial Cells Express L-rich Ferritin Proteins It has long been known that microglial cells form a heterogeneous cell population within the CNS, both in terms of morphological features and present functional status. More recently, subsets of microglia that sele ctively express certain antigen s have further increased our knowledge of the complexity of the parenchymal microglial pool. For instance, expression of such epitopes as the 5D4 keratan sulfate prote oglycan (Bertolotto et al. 1998), the membranebound scavenger receptor CD163 (Robe rts et al. 2004), the leukocyte chemotactic factor and the leukocyte common antigen (Mittelb ronn et al. 2001) as well as the hematopoietic stem cell marker CD34 (Wirenfeldt et al. 2005) have all been shown to identify a subpopulation of microglial cells in the normal, diseased, and in jured CNS. The immunohistochemical results we report herein support the concept of microglial ce lls as a specialized ce ll population of the CNS 77


by showing that only a subset of microglia participate in long-term iron storage via the expression of L-rich ferritin pr oteins. We hypothesize that the r eason only a subset of microglial cells participate in prolonged iron storage is to minimize the potential for iron-dependent oxidative damage in ferritin-positive microglia. Potential Link between Iron Storage, Senescence and Microglial Dystrophy In our immunohistochemical analyses of hu man brain specimens from elderly and AD subjects, we found that microglial degeneration a ppears to be associated with iron-mediated oxidative stress by providing evidence that L-ferrit in immunoreactive microglia are particularly susceptible to dystrophic changes, especially in the aged brain where more than 60% of L-Ftpositive microglia exhibited signs of cytoplasmic de terioration. Similar findings were obtained in tissue specimens from HD patient s and the r6/2 mouse model of HD (Simmons et al. 2007). From these results we concluded that both longterm iron storage and ce ll senescence contribute to microglial dystrophy. One potential mechanism by which iron storag e and senescence may interact relates to the age-related increase in ferriti n-to-iron ratio. In young animals, when iron levels are relatively low, ferritin expression efficiently maintains iron in a non-reactive form due to the high ferritinto-iron ratio, but as the brain ages and accumula tes iron, ferritin becomes overloaded with iron and increased ferritin levels may not be sufficien t to contain the elevated iron concentrations (Kaur et al., 2007). The iron saturation of ferritin proteins in older animals would likely enhance the risk of iron release during ferritin turnover (Kaur et al., 2007), thus in creasing the probability of iron-induced oxidative damage. On the other hand, microglial dystrophy may be the consequence of improper cellular functioning due to intrinsic senescence mechanisms in ferritinpositive microglia. As reported earlier, microg lia undergo age-related changes in metabolic 78


activity, including a decrease in pr oteolytic activity and a higher production of ROS (Streit et al. 2008). An extension of this premise in the contex t of the current study is the possibility that senescent microglia gradually lose their ability to maintain viable ferritin-bound iron stores and to fight extracellular and intracellu lar stressors, all of which could result in greater vulnerability to degeneration and/or death. The microtubule network is probably the first to be affected by th e unwanted release of iron from ferritin proteins due primarily to two factors: 1) Ferritin molecules bound to microtubules contain twice as many iron atom s compared to unbound ferri tin (Hasan et al., 2006); and 2) OH generated by the reaction of labile Fe2+ with H2O2, which is controlled in large part by microglia, is th e most reactive and short-lived ROS, capable of near instant oxidation of nearby molecules (Kruszewski 2003). Thus, the high iron concentration of microtubule-bound ferritin proteins and their clos e proximity to microtubules increase the susceptibility of the microtubule network to iron-catalyzed oxidative stress reactions This premise may account, at least in part, for the degenerative ch anges observed in microglial cytoplasmic structure, of which their microtubule-rich processes seem to be the most vulnerable. Microglial Cells are Vulnerable to Oxidative Stress Reactions in Aged Brains Because oxidative stress is the outcome of an inadequate balance between production and elimination of ROS, a diminished microglial antioxidant capacity may be necessary in order for microglial dystrophy to occur. Immunocytochemical studies of the antioxidative capacity of microglial cells have revealed that among other neural cells microglia are strongly immunoreactive for glutathione peroxidase (GPx), an enzyme that uses glutathione to reduce H2O2 to water (Hirrlinger et al. 2000; Lindena u et al. 1998). The predominance of this antioxidant system in microglia pinpoints the potentia lly deleterious effects of iron-mediated 79


oxidative stress and the need for endogenous prot ective mechanisms. It is not yet known whether microglial GPx potential is maintained at suffici ent levels in the aged brain to account for the age-related increase in brain iron levels. Despite the lack of evidence for an altered microglial antioxidative potential with aging, oxidative stress markers have been identified in microglia. Specificall y, the reactive carbonyl crotonaldehyde, which is generated during lipid peroxidation, locali zes in microglial cells in the AD brain (Kawaguchi-Niida et al. 2006). Moreover, striatum microglia of aged (> 11 weeks), but not young, R6/2 mice (a mouse model of Huntingt ons disease) was found to be immunoreactive for 8-OHdG, a DNA base oxidized in free radical reactions by OH (Simmons et al. 2007). Taken together, these findings suggest that the microglia l antioxidative capacity may be insufficient in neutralizing ROS within the aging brain, where senescence microglia prevail. In general, every antioxidative protective biom olecule (e.g., GPx) has at least one target oxidant molecule (e.g., H2O2), thus providing an efficient means for ROS detoxification. The exception to this is OH, which is produced fr om the interaction of Fe2+ with H2O2 and is extremely reactive. Mitoc hondrial reactions are the ma in cellular sites of H2O2 production and dystrophic microglia have been shown to c ontain degenerated m itochondria (WierzbaBobrowicz T et al. 2004). Activated microg lial cells are also capable of producing H2O2 by the activity of the enzyme NADPH oxidase at the cell membrane (Twig et al. 2001). As a nonpolar molecule, H2O2 is able to diffuse across membranes. Hence, chronic micr oglial activation may lead to increased H2O2 exposure intracellularly. Taken together these observations suggest that L-ferritin-positive microglia are highly dependent on iron and H2O2 homeostasis to prevent ironinduced oxidative damage. Hallmarks of Microglial Degenerati on in Human and Rodent Brains 80


Our meticulous morphological analyses of immunoreactive microglia under the light microscope revealed important characteristics be tween dystrophic microglia in human and rodent brains. These findings, which ha ve been previously describe d, are summarized in Table 4-1. Here, I will place emphasis on the defining charac teristics that distinguish dystrophic microglia in human brains from dystrophic microglia in rodents. In both species, cytorrhexis, or the fragmentation of the cells cytoplas m into two or more parts, appear s to be the most severe form of cytoplasmic degeneration. Whereas cytorrhectic microglia are rare in r odents of all ages and in tissue specimens from young human adults, th ey are prevalent in aged human brains, especially in AD patients. Cytorrhexis in microglia range in severity from a relatively small nick on a distal process to nearly complete obliteratio n of the cells cytoplasmic structure as shown in Fig. 3-6. The blatantly abnormal morphology of these cells suggests that they are in the process of degeneration, though the underlying mechanis ms by which cytorrhectic microglia execute cellular demise are not presently known. Typically, pathways leading to cell death ar e discussed as discrete, independent entities that are separated dichotomously into either apoptosis or necrosis. Just like the initial morphological characterization of cytorrhexis reported herein, the detection of apoptotic and necrotic cell death mechanisms were initia lly based on cell morphology by using light and electron microscopy (Kerr et al 1972). Nowadays, morphological features that distinguish apoptosis from necrosis as well as their biochemical and physiol ogical characteristics are wellcharacterized. Apoptosis is a form of caspase-mediated cell de ath that exhibit, among other features, chromatin condensation and cytoplasmic fragmentation into apoptotic vesicles (see Table 4-2). Apoptotic cells ar e readily recognized and pha gocytosed by tissue macrophages resulting in an anti-inflammatory outco me. Necrosis, on the other hand, describes the 81


postmortem observation of dead cells that have lost membrane integrity from cell swelling and subsequent lysis (Table 4-2). In vivo necrotic cell death is often associated with extensive tissue damage resulting in intense inflammatory reaction. Although dystrophic microglia share some features with apoptoti c cells and necrotic cells (as summarized in Table 4-2), they do not w holly conform to either cell death pathway. In particular, cytorrhexis (i.e., end-stage dystrophy) affect individual microglial cells (akin to apoptosis) but do not form apoptotic cell bodie s (unlike apoptosis). Moreover, preliminary immunohistochemical studies using the term inal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) method to visualize DNA fragmentation in individual nuclei, which is often used as a defining characteristic of apoptotic cells, yielded only negative results in previous histological preparations from our laboratory (data no t shown). Notably, although tissue debris resulting from either apopt osis or necrosis is phagocytosed by macrophages, no evidence of activated microglia and/or macrophages are observed in the vicinity of dystrophic microglia. In all, because so little is currently known about the biochemical pathways leading to microglial cytorrhexis, it seems unreasonable to classify cyto rrhectic microglia into a specific mode of cell death just yet. Despite the widespread use of a dichotomous classification of cell death into either apoptosis or necrosis, recent obs ervations suggest that the true biological spectrum of cell death is much more diverse than init ially appreciated. Other pathways of cell death iden tified thus far include autophagy, oncosis, and pyroptosis, among others. Today it is al so evident that the molecular processes that mediate cell death ma y actually overlap considerably in dying cells. Futher investigations into mol ecular and biochemical alterations in dystrophic microglia is likely 82


to aid in the elucidation of whether dystrophic microglia are degenerating based on yet unidentified pathways or on a combina tion of known cell death programs. In the aged rodent brain, cy torrhectic microglia are infreque nt. Instead, overtly dystrophic microglia in aged rats exhibit a gross enla rgement of their perinuclear region, which under fluorescence microscopy has been determined to c ontain autofluorescent LF granules. In turn, in our immunohistochemical analyses no LF-positive microglia were de tected in human brains. It is not yet known what the cellular and physiological relevance of this disc repancy is. Labile iron acts as a catalyst in LF formation and accumu lation within the lysosomal compartment of senescence cells (Jolly et al 1995; Terman and Brunck 2004), the same site of normal Ft turnover. It is possible, theref ore, that senescent microglia be come inadequate in maintaining their intracellular ferritin-toiron homeostasis, probably due to an impairment of their degradative pathways (Stolzing a nd Grune 2003) that terminates in labile iron in cidence within lysosomes. For yet unknown reasons, it is possible that ra ts have a propensity to accrue lipid-laden LF particles, and the resulting labile iron from lysosomes may catalyze LF-formation instead of Fenton-type reactions, which is thought to occur in humans. This hypothesis would provide a basis as to why cytorrhectic (i .e., degenerating) microglia are found primarily in humans while LF-positive microglia are predominant in rodent s. Altogether, it seems that even though the morphological features of dystrophic microglia di ffer somewhat between humans and rats, they are still interlinked by their asso ciation with iron chemistry. Microglial Dystrophy Is Not Due to Postmortem Tissue Autolysis Our immunohistochemical analys es revealed that the mor phological signs of microglial degeneration in L-ferritin positive microglia were manifested by shortening, thinning, twisting, 83


and fragmentation of processes. These changes were in accordance with previous reports of microglial degeneration cited in the literature (s ee Table 1-2). In order to determine whether these aberrant morphological char acteristics could actually be an artifact of postmortem tissue autolysis, we compared the morphology of ferritin-positive microglia in tissues with PMI ranging from 3 to 20 hours. Our investigati on provided no correlation between microglial dystrophy and increased postmortem delay, supporti ng the idea that microglial degeneration is a true pathophysiological event. Additional suppor t for this notion is derived from our recent observations in an animal model of amyotrophic late ral sclerosis (where auto lysis is not an issue) demonstrating widespread microglial dystrophy in areas undergoing motoneuron degeneration (Fendrick et al. 2007). Dystrophic Microglia Are Most Prevalen t in the Alzheimers Disease Brain Than in Age-Matched Non-Demented Control Tissues In our immunohistological characterization of dystrophic microglia, it became readily apparent that the incidence and severity of microglial dystrophy was considerably greater in AD tissues compared to young and even age-ma tched controls. Why dystrophic microglia are preferentially localized in the AD brain is presently unknown, however. The predominance of specific dystrophic morphologies (i .e., formation of spheroidlike protuberances [bulbous swellings] along major processes and multicellular clusters) in brain tissues that have extracellular A aggregates in common (HPC and AD tissu es), raised the possibility that the proximity to histopathological hallmarks of AD played a contributory role in the demise of microglial cells. To address this question, we firs t investigated whether these dystrophic changes could be linked to the amyloiddeposition that is present in both of these tissues. 84


Although both HPC and AD tissues contained A deposits, the incidence of L-Ft-positive dystrophic microglia was significantly higher in AD tissu es, which suggests that A deposition alone does not seem to predispose microglial ce lls to degeneration. Furthermore, we found no correlation between dystrophic changes and proxi mity to SPs for both HLA-DRand ferritinpositive microglia. Taken together, it seems likely th at an altered homeostatic milieu, instead of SP accumulation per se, may play a more significant role in the occurrence of dystrophic microglia. Of interest, changes in calcium and ir on homeostasis, as well as higher indices of oxidative stress markers have been reported in AD brains (Selkoe 2004) the significance of which to microglial dystrophy remains unknown. On one side, it is possi ble that senescent microglia contribute to the altered brain homeost asis that occurs in AD tissues due to altered microglial functioning. Alternativel y, senescent microglia may be pa rticularly vulne rable to such changes, which may then predispose this subset of older microglia to degeneration secondarily. In regards to the relation ship between NFTs and dystrophi c microglia, our analysis was hindered by the fact that no pos itive immunohistochemical signal was obtained for NFTs in any of the tissues used in this study. These nega tive data was corroborat ed by evaluating the morphological characteristics of the 6F3D-positive se nile plaques. In all tissues examined, both HPC and AD, we found that the SPs had a diffuse morphology and were distributed primarily in the gray matter region of the SFG. Additionally, we were unable to stain these SPs with Congo Red. In all, these findings sugge st that the AD tissues under study can be situated in the early stages of the disease (s ee Tables 1-3 and 1-4). Since dystrophic microglia are prevalent in early AD, it is possible that microglial degeneration actually cont ributes to AD pathogenesi s and/or progression. Further studies are necessary to delineate the precise effect of degenerating microglia in AD brains. Since AD 85


dementia is best correlated with loss of synapses rather than the accumulation of protein aggregates (DeKosky and Scheff 199 0), it is possible that with ag ing and the ensuing progressive accumulation of dystrophic microglia in the brain parenchyma, results in more and more widespread malnourishment of neurons, which coul d first have an effect on the maintenance of synapses and subsequently on the we ll-being of neurons themselves. The current findings offer additional evid ence for the microglia l dysfunction hypothesis of AD (Streit 2004) by demonstrating a direct relationship between microglial dystrophy and the expression of L-rich ferritin pr oteins. The dysfunction hypothesis stat es that a principal cause for development for aging-associated neurodegenera tion is found in an age-related decline in microglial neuroprotection which occurs because microglial cells themselves are subject to cellular senescence, as evidenced primarily by th e increased incidence of dystrophic microglia in aged brains. The current findings offer a possible mechanism as to why dystrophic changes occur in microglia, namely, that dystrophy is a re sult of iron-mediated oxidative damage. Dystrophic Microglia are Associated with Agin g and Not Injury Conditions in Aged Rats Axotomy by brief constriction of fibre tracts of FMNs produces a mild neuronal injury that induces a transient microglial activation response within the FNu (Graeber et al. 1988a; Graeber et al. 1988b; Moran and Graeber 2004). B ecause there is no cell death (i.e., injured motoneurons regenerate their pe ripheral axons), activated microgl ial cells normally do not reach the advanced phagocytic stage of activation (Streit et al. 2000). In order to determine whether the degree of microglial activation to such reversible neuronal injury is disregulated in dystrophic microglia of aged rats, we analyzed micr oglial immunoreactivity to the macrophage-specific marker ED1 in FNu sections from axotomized young and aged rats. The antigen recognized by ED1, which is the rat homologue of human CD6 8, is expressed on the lysosomal membranes of 86


phagocytic (i.e. myeloid) cells and is often used as a marker of pha gocytic activity (Dijkstra et al. 1985; Flaris and Hickey 1992). Macrophages (con stitutively ED-1-positive) and microglia (facultatively ED1-positive) are f unctionally and developmentally related, expressing a similar repertoire of genes (Sedgwick and al. 1991). First, we noticed occasional ED1-immunoreactiv e cells but mostly in the control FNu of aged rats (not the axotomized side as expected). Aging appears to promote ED1 immunoreactivity in rats after cortical stab injury (Kyrkanides et al. 2001) and stroke (Badan et al. 2003a; Badan et al. 2003b). However, since in the rodent FN axotomy model there is no disruption of the BBB (Moran and Graeber 2004) it seemed unlikely that ED1-positive cells would be present in the axotomized FNu, especially in the uninjure d side. The possibility that the rapid microgliosis observed in the lesioned FNu is dependent on the recruitment of blood-borne progenitors (i.e., ED1-positive macrophages) ha s been recently dismissed in a study using chimeric animals obtained by parabiosis (Ajami et al. 2007). More spec ifically, the authors reported that although rare ED-1-positive cells we re present in the axotomized FNu area, they were located inside blood vessels and not in the brain parenchyma (Ajami et al. 2007). Upon closer examination of the ostensib ly ED1 immunoreactive microglia present in aged rat tissue sections, it wa s noticeable that immunofluorescen ce was achieved irrespective of the fluorophore used, raising the possibility that the positive signal was in fact derived from autofluorescence. Instead of disregarding these data as artifact, we analyzed the characteristics of these seemingly autofluorescent microglia. Base d on spectral properties a nd cellular localization, it was determined that the autofluorescent materi al was derived from intracellular LF granules. Lipofuscin, or age pigment, is a heterogene ous nondegradable lipoph ilic material that accumulate within lysosomal compartments as residual bodies (Brunk and Terman 2002; Terman 87


and Brunk 2004). We concluded, therefore, that the combined effect of staining lysosomal membrane proteins with the autofluorescent signal from lysosomal LF particles led to our conspicuous detection of LF material in microglia l cells of aged rats. Our in-depth analysis of these autofluoresce nt microglia also led to the discovery that lipofuscin-positive microglia, observed primarily in the non-axotomized FNu of aged rats, exhibit dystrophic morphological char acteristics. In particular, LF -positive microglia presented deramification of distal processes, an abnormal swelling of the perinuclear region and spheroidformation, all of which were readily noticeable in the control FNu. Because the axotomized FNu is populated by hypertrophic microglia, which al so exhibit deramification patterns and an enlargement of the cell body, the differentiation between dystrophic and activated microglia was significantly more difficult to achieve. Nonetheless, the finding of LF particles in microglial cells seems to be in itself a strong enough indicator of microglial dystrophy in aged rats as highlighted below. Lipofuscin accumulates progressively with advancing age, which explains why autofluorescent LF signals were absent in young rat brain tissues. Interestingly, we observed only a neglibible number of LF-positive microglia in the axotomized FNu of aged rats. Because LF accumulation is usually seen in cells that have reached a postmitotic phase (Brunk and Terman 2002), our results can be interpreted as a sign that LF-positive microglial cells are, in fact, senescent cells. That is, the low numbers of LF-positive microglia in the axotomized FNu may stem from the fact that upon activation co nditions that demand cell proliferation (e.g., FN axotomy), LF-positive microglia, which are in capable of undergoing ce ll proliferation, are outnumbered by mitotic LF-negative microglia. Alt ogether, our findings sugge st that senescent, dystrophic microglia in aged rats are manifested predominantly as LF-positive microglia. LF 88


granules has been previously shown to accumulate in aging microglia from rodent brains (Sierra et al. 2007; Xu et al. 2008), although this st udy is the first known account of the probable association between LF accumulation w ith microglial senescence and dystrophy. Ferritin Immunoreactivity Is Not Upregul ated in Activated Microglial Cells Because Ft immunohistochemistry has been used in the past to identify activated microglial cells (Grundke-Iqbal et al. 1990; Kaneko et al. 1989) and activated microglia share some morphological similarities with dystrophic microglia, namely atrophy of distal branches (Table 4-1;(Lopes et al. 2008; Streit 2006), a secondary goa l of this thesis was to investigate whether microglial activation induces Ft expression. To answer this question we utilized the FN axotomy paradigm, since the microglial activation response in this model is well-characterized (Graeber et al. 1988a; Graeber et al. 1988b; Moran and Graebe r 2004; Streit et al. 2000). Previously, the rat FNu has been shown to e xpress very low levels of L-Ft mRNA (Han et al. 2002), while to our knowledge the present study is the first to report on L-Ft protein levels in the rat FNu. We hypothesized that if Ft immunoreactivity is ind eed supposed to specifically label activated microglial cells, Ft expression should be upreg ulated during activation conditions. Our histological analysis revealed that microglial Ft expression is not affected by injury-induced activation conditions by comparis on of Ft-staining patterns betw een axotomized and uninjured sides of the FNu from both young and aged rats. The only difference found was a higher level of Ft immunoreactivity in aged animal s. We also noticed that the ma jority of Ft-IR cells found in either side of the rat FNu exhibited morphol ogical characteristics of oligodendrocytes, while only occasionally Ft-positive microglial cell s were identified in the aged rat FNu. Several observations can account for such fi ndings. First, cellula r Ft expression is regulated primarily by iron availability at the posttranslational level, with high iron levels 89


inducing Ft upregulation (Hentze et al. 2004). Since brain iron le vels increase steadily with advancing age (Benkovic and Connor 1993; Focht et al. 1997), the increase in Ft expression with aging is attributed to ferritins role as an in tracellular scavenger of excess iron (Focht et al. 1997). Second, it is possible that in the rat FNu most L-Ft prot eins are expressed primarily by oligodendrocytes instead of microg lial cells due to an intrinsic regi onand species-specific Ft subunit distribution (Benkovic and Connor 1993; Connor et al. 1994). Third, axotomy of the FN has been shown to upregulate the expression of TfR, which are responsible for an increased cellular iron uptake in regenerati ng FMNs (Graeber et al. 1989) TfR expression is inversely related to Ft expression, that is TfR levels in crease in conditions of low iron availability (Harrison and Arosio 1996), thus providing another explanation for the low levels of Ft immunoreactivity in the rat FNu observed in the present study. Overall, we observed that activation conditions were insufficient to upregula te Ft expression levels in microglia. Aging and the resulting increase in brain iron levels, on the other hand, app eared to promote microglial Ft expression within the rat FNu. In all, our data suggests that ferritin immunoreactivity is a poor indicator of activated microglial cells. Concluding Remarks In summary, although ferritin immunohistoc hemistry has been known for its use as a generic microglial marker (Grundke-Iqbal et al. 1990; Kaneko et al. 1989), our current results suggest that ferritin imm unoreactivity could be used more speci fically as a sele ctive marker of those microglial cells that are in danger of being lost. An ex tension of the current findings suggest a possible mechanism as to why dystrop hic changes occur in microglia, namely, that dystrophy is a result of iron-mediated oxidative damage. 90


In our morphological analyses of microglial cells in the normal and diseased human brain, we found that ferritin-positive dystrophic microglia predominated in the aged brain, particularly in postmortem tissue from Alzh eimers disease brains. Because microglial degeneration is widespread in tissues undergoing age-related neurodegeneration associated with disease conditions (i.e., AD), it is speculated that dystrophic micr oglia may participate in the disease process. This hypothesis is based on the probability that the morphological abnormalities that characterize dystrophic microglia likely reflect aberrant functional stat es. More specifically, degeneration of the microglial cytoplasmic structure would clearly be an impediment to the normal surveillance functions of these cells, partic ularly by affecting the st ructural identity of ramified microglia and consequently impairing constitutive neuroprotective functions in such dystrophic microglia. A secondary aim of this thesis was to de termine whether the graded repertoire of microglial activation is misregulated in aged an imals to permit activated microglia to reach the phagocytic stage, which should not occur in response to su ch a reversible neur onal injury. In this study, we also provided evidence that neither Ft nor ED1 immunoreactivity is affected by mildly acute activation conditions in micr oglial cells of either young or aged rats. On the other hand, we report on the discovery of lipid-laden autofluorescen t granules (i.e., lipofus cin) in predominantly dystrophic microglia of aged rats. Collectively, our findings provi de a theoretical basis on wh ich to probe the specific molecular mechanisms that contribute to mi croglial senescence. More specifically, iron dyshomeostasis emerged as a potential culprit in the degenerative potential of microglial cells in both aged human and rodent brains. Further studies are required to investigate the extent to which alterations in ferritin and iron homeostasis may correla te with microglial dystrophy. 91


Table 4.1. Comparison of dystrophic morphological characteristics between humans and rodents DYSTROPHIC MORPHOLOGIES HUMAN RODENT Deramification + + Thinning/Tortuous Processes + Spheroid Formation + Cell Clusters + Cytorrhexis + Lipofuscin Accumulation + Enlargement of perinuclear cytoplasm + Symbols: absent; sometimes present; +, common. 92


Table 4.2. Differential features betwee n dystrophy, apoptosis, and necrosis M M i i c c r r o o g g l l i i a a l l D D y y s s t t r r o o p p h h y y A A p p o o p p t t o o s s i i s s N N e e c c r r o o s s i i s s No loss of membrane integrity Membrane blebbing, but no loss of membrane integrity Loss of membrane integrity Begins with decrease in arborization and volume of cell processes Begins with shrinking of cytoplasm Begins with swelling of cytoplasm and mitochondria Ends with random fragmentation of the cytoplasm (cytorrhexis) Ends with fragmentation of cells into smaller cell bodies Ends with total cell lysis No vesicle formation Formation of membranebound vesicles (apoptotic bodies) No vesicle formation, complete lysis Chromatin condensation: Unknown Chromatin condensation No chromatin condensation Intact Nucleus Non-random DNA fragmentation Random DNA fragmentation Degenerating mitochondria reported Mitochondria becomes leaky due to pore formation involving proteins of the bcl-2 family Disintegration (swelling) of organelles, including mitochondria Caspase-independent Caspase-dependent Caspase-independent Affects individual cells Affects individual cells Affects groups of contiguous cells Unknown if dying cells are phagocytosed Phagocytosis by macrophages Phagocytosis by macrophages Unknown if energy is required Energy (ATP)-dependent Energy (ATP)independent Anti-inflammatory Anti-inflammatory Pro-inflammatory 93


LIST OF REFERENCES Ajami B, Bennett JL, Krieger C, Tezlaff W, Rossi FMV. 2007. Local self-renewal can sustain CNS microglia maintenance and function throughout adult life. Nat Neurosci 10(12):1538-1543. Arriagada PV, Marzloff K, Hyman BT. 1992. Distri bution of Alzheimer-type pathologic changes in nondemented elderly individuals matche s the pattern in Alzheimer's disease. Neurology 42:1681-1688. Badan I, Buchhold B, Hamm A. 2003a. Accelerated glial reactivity to st roke in aged rats correlates with reduced functional rec overy. J Cereb Blood Flow Metab 23:845-854. Badan I, Platt D, Kessler C. 2003b. Temporal dyna mics of degenerative and regenerative events associated with cerebral ischemia in aged rats. Gerontology 49:356-365. Bali P, Zak O, Aisen P. 1991. A new role for the tr ansferrin receptor in the release of iron from transferrin. Biochemistry 30:324-328. Bartzokis G, Beckson M, Hance D, Marx P, Fo ster J, Marder S. 1997. MR evaluation of agerelated increase of brain iron in young adult and older normal males. Magn Reson Imaging 15(29):35. Bartzokis G, Cummings J, Perlman S, Hance DB, Mintz J. 1999. Increa sed basal ganglia iron levels in Huntington's diseas e. Arch Neurol 56(5):569-74. Ben-Porath I, Weinberg RA. 2005. The signals and pathways activating cellu lar senescence. Int J Biochem Cell Biol 37:961-76. Benkovic SA, Connor JR. 1993. Ferritin, transferrin, and iron in selected regions of the adult and aged rat brain. J Comp Neurol 338:97-113. Bertolotto A, Agresti C, Castello A, Manzar do E, Riccio A. 1998. 5D4 keratan sulfate epitope identifies a subset of ramified microglia in normal central nervous system parenchyma. J Neuroimmunol 85:69-77. Blinzinger K, Kreutzberg G. 1968. Displacement of synaptic terminals from regenerating motoneurons by microglial cells. Z Zellforsch 85:145-157. Bodles AM, Barger SW. 2004. Cytokines and the aging brain what we don't know might help us. Trends Neurosci 27:621-626. Braak H, Braak E. 1991. Neuropathological stag ing of Alzheimer-related changes. Acta Neuropathol 82:239-259. 94


Brunk UT, Terman A. 2002. Lipofuscin: mechanisms of age-related accumulation and influence on cell function. Free Rad Biol Med 33(5):611-619. Buckwalter MS, Wyss-Coray T. 2004. Modelli ng neuroinflammatory phenotypes in vivo. J Neuroinflamm 1:10. Carpenter AF, Carpenter PW, Markesbery WR. 1 993. Morphometric analysis of microglia in Alzheimer's disease. J Neuropa th Exp Neurol 52(6):601-608. Cheepsunthorn P, Palmer C, Connor JR. 1998. Ce llular distribution of ferritin subunits in postnatal rat brain. J Comp Neurol 400:73-86. Cheepsunthorn P, Palmer C, Menzies S, Robe rts RL, Connor JR. 2001. Hypoxic/ischemic insult alters ferritin expression and myelination in neonatal rat brains J Comp Neurol 431:382396. Conde JR, Streit WJ. 2005. Effect of aging on the microglial response to peripheral nerve injury. Neurobiol Aging. 27( 10):1451-61. Conde JR, Streit WJ. 2006. Microg lia in the aging brain. J Neur opathol Exp Neurol 65(3):199203. Connor JR, Boeshore KL, Benkovic SA, Menzies SL. 1994. Isoforms of ferritin have a specific cellular distribution in the br ain. J Neurosci Res 37:461-465. Connor JR, Menzies SL. 1995. Cellular management of iron in the brain. J Neurol Sci 134 (Suppl):33-44. Connor JR, Menzies SL, St Martin SM, Mufson EJ. 1990. Cellular distribution of transferrin, ferritin and iron in normal and ag ed brain. J Neurosci 27:595-611. Connor JR, Snyder BS, Arosio P, Loeffler DA, LeWitt P. 1995. A quantitative analysis of isoferritins in select regions of aged, park insonian, and Alzheimer's disease brains. J Neurochem 65:717-724. Crisby M, Ralman SM, Sylven C, Winblad B, Schultzberg M. 2004. Effect s of high cholesterol diet on gliosis in apolipoprotein E knockout mice. Implications for Alzheimer's disease and stroke. Neurosci Lett 369(2):87-92. Cummings JL, Cole G. 2002. Alzheimer disease. J Am Med Assoc 287:2335-2338. Davalos D, Grutzendler J, Yang G, Kim JV, Zu o Y, Jung S, Littman DR, Dustin ML, Gan WB. 2005. ATP mediates rapid microglial response to local brain injury in vivo. Nature Neuroscience 8(6):752-758. DeKosky ST, Scheff SW. 1990. Synapse loss in frontal cortex biopsies in Alzheimer's disease: correlation with cognitive severity. Ann Neurol 27:457-464. 95


Dexter DT, Carayon A, Javoy-Agid F, Wells FR, Daniel SE, Lees A, Jenner P, Marsden CD. 1991. Alterations in levels of iron, ferritin and other trace me tals in Parkinson's disease and other neurodegenerative diseases aff ecting the basal gang lia. Brain 114:1953-1975. Dijkstra CD, Dopp EA, Joling O, Kraal G. 1985. The heterogeneity of mononuclear phagocytes in lymphoid organs: distinct macrophage subpopulations in the rat recognized by monoclonal antibodies ED1, ED2 and ED3. Immunology 54(3):589-599. Dipatre PL, Gelmann BB. 1997. Microglial cell ac tivation in aging and Alzheimers disease: partial linkage with neurofibrillary tangle burden in the hippocampus. J Neuropathol Exp Neurol 56:143-149. Fendrick SE, Xue Q-S, Streit WJ. 2007. Formation of multinucleated giant cells and microglial degeneration in rats expressing a mutant Cu/Zn superoxide dismutase gene. J Neuroinflamm 4:9. Ferraro A. 1931. The origin and formation of se nile plaques. Arch Ne urol Psychiatry 25:10421062. Flanary BE, Sammons NW, Nguyen C, Walker D, Streit WJ. 2007. Evidence that aging and amyloid promote microglial cell se nescence. Rejuv Res 10(1):61-74. Flanary BE, Streit WJ. 2003. Telome res shorten with age in rat ce rebellum and cortex in vivo. J Anti Aging Med 6:299-308. Flanary BE, Streit WJ. 2004. Progr essive telomere shortening o ccurs in cultured rat microglia, but not astrocytes. Glia 45:75-88. Flaris NA, Hickey WF. 1992. Development and char acterization of an ex perimental model of brain abscess in the rat. Am J Pathol 141:1299-1307. Fleming MD, Romano MA, Su MA, Garrick LM Garrick MD, Andrews NC. 1998. Nramp2 is mutated in the anemic Belgrade (b) rat: Evid ence of a role for Nramp2 in endosomal iron transport. Proc Natl Acad Sci USA 95:1148-1153. Focht SJ, Snyder BS, Beard JL, van Gelder W, Williams LR, Connor JR. 1997. Regional distribution of iron, transferrin, ferritin, and oxidatively-m odified proteins in young and aged Fisher 344 rat brains. Neuroscience 79:255-261. Gary DA, Woulfe J. 2005. Lipofus cin and aging: a matter of toxic waste. Sci Aging Knowledge Environ:re1. Godbout JP, Chen J, Abraham J, Richwine AF, Berg BM, Kelley KW, Johnson RW. 2005. Exaggerated neuroinflammation and sickness behavior in aged mice following activation of the peripheral inate imm une system. FASEB J 19:1329-1331. 96


Graeber MB, Raivich G, Kreutzberg GW. 1989. Incr ease of transferrin re ceptors and iron uptake in regenerating motor neurons. J Neurosci Res 23(3):342-345. Graeber MB, Streit WJ, Kreutzberg GW. 1988a. Axotomy of the rat f acial nerve leads to increased CR3 complement receptor expression by activated microglial cells. J Neurosci Res 21(1):18-24. Graeber MB, Tetzlaff W, Streit WJ, Kreutzberg GW. 1988b. Microg lial cells but not astrocytes undergo mitosis following facial nerve axotomy. Neurosci Lett 85:317-321. Graeber MB, Lopez-Redondo F, Ikoma W, Ishikawa M, Imai Y, Nagajima K, Kreutzberg GW, Kohsaka S. 1998. The microglia/macrophage res ponse in the neonatal rat facial nucleus following axotomy. Brain Res 813:241-253. Grundke-Iqbal I, Fleming J, Tung YC, Lassmann H, Iqbal K, Joshi JG. 1990. Ferritin is a component of the neuritic (senile) plaque in Alzheimer dementia. Acta Neuropathol 81:105-110. Hallgren B, Sourander P. 1958. The effect of age on the non-heme iron in the human brain. J Neurochem 3:41-51. Han J, Day JR, Connor JR, Beard JL. 2002. H and L ferritin subunit mRNA expression differs in brains of control and iron-def icient rats. J Nutr 132:2769-2774. Hanisch UK, Kettenmann H. 2007. Microglia: active se nsor and versatile e ffector cells in the normal and pathologic brain. Nat Neurosci 10(11):1387-94. Harrison PM, Arosio P. 1996. The ferritins: molecular properties, iron storage function and cellular regulation. Bioc him Biophys Acta 1275:161-203. Hentze MW, Muckenthaler MU, Andrews NC. 2004. Balancing acts: molecular control of mammalian iron metabolism. Cell 117:285-297. Heppner FL, Roth K, Nitsch R, Hailer NP. 1998. Vitamin E induces ramification and downregulation of adhesion molecules in cultured microglial cells. Glia 22:180-188. Hirrlinger J, Gutterer JM, Kussmaul L, Hamp recht B, Dringen R. 2000. Microglial cells in culture express a prominent glutathione system for the defense against reactive oxygen species. Dev Neurosci 22:384-92. Hoffbrand AV, Ganeshaguru K, Hooton JW, Tatte rsall MH. 1976. Effect of iron deficiency and desferrioxamine on DNA synthesis in human cells. Brit J Haematol 33:517-526. Huang E, Ong W-Y. 2005. Distribu tion of ferritin in the rat hi ppocampus after kainate-induced neuronal injury. Exp Brain Res 161:502-511. 97


Ito D, Imai Y, Ohsawa K, Nakajima K, F ukuuchi Y, Kohsaka S. 1998. Microglia-specific localization of a novel calcium binding protein, Iba1. Brain Res Mol Brain Res 57(1):1-9. Jolly RD, Douglas BV, Davey PM, Roiri JE. 19 95. Lipofuscin in bovine muscle and brain: A model for studying age pigment. Gerontology 41(Suppl 2):283-295. Kaneko Y, Kitamoto T, Tateishi J, Yamaguchi K. 1989. Ferritin immunohistochemistry as a marker for microglia. Acta Neuropathol 79:129-136. Kaur D, Rajagopalan S, Chinta S, Kumar J, Di Monte D, Cherny RA, Andersen JK. 2007. Chronic ferritin expression within murine dopaminergic midbrain neurons results in a progressive age-related neur odegeneration. Brain Res 1140:188-194. Kawaguchi-Niida M, Shibata N, Morikawa S, Uchida K, Yamamoto T, Sawada T, Kobayashi M. 2006. Crotonaldehyde accumulates in glial cel ls of Alzheimer's disease brain. Acta Neuropathol 111:422-9. Kerr JF, Wyllie AH, Currie HR. 1972. Apoptosis: a basic biological phenomenon with wide ranging implications in tissue ki netics. Br J Cancer 26:239-257. Kim HY, LaVaute T, Iwai K, Klausner RD, R ouault TA. 1996. Identification of a conserved and functional iron-responsive element in th e 5-untranslated region of mammalian mitochondrial aconitase. J Biol Chem 271:24226-24230. Koeppen AH, Dickson AC. 2001. Iron in the Halle rvorden-Spatz syndrom e. Pediatr Neurol 25(2):148-155. Korotzer AR, Pike CJ, Cotman CW. 1993. beta -Amyloid peptides induce degeneration of cultured rat microglia. Brain Res 624(1-2):121-125. Kreutzberg GW. 1996. Microglia: a sensor for pathol ogical events in the CNS. Trends Neurosci 19(8):312-318. Kruszewski M. 2003. Labile iron pool: the main determinant of cellular response to oxidative stress. Mutat Res 531:81-92. Kyrkanides S, O'Banion MK, Whiteley PE. 2001. E nhanced glial activatio n and expression of specific CNS inflammation-related molecu les in aged versus young rats following cortical stab injury. J Neuroimmunol 119:269-275. Lawson DM, Treffry A, Artymiuk PJ, Harrison PM Yewdall SJ, Luzzago A, Cesareni G, Levi S, Arosio P. 1989. Identification of the ferroxidase center in ferritin. FEBS Letters 254:207-210. Lawson LJ, Perry VH, Gordon S. 1992. Turnover of resident microglia in the normal adultmouse brain. Neuroscience 48:405-15. 98


Le NT, Richardson DR. 2002. The role of iron in cell cycle progression and the proliferation of neoplastic cells. Biochim Biophys Acta 1603:31-46. Lemstra AW, Groen in't Woud JCM, Hoozeman s JJM, van Haastert ES, Rozemuller AJM, Eikelenboon P, van Gool WA. 2007. Microglia activation in se psis: a case-control study. J Neuroinflamm 4:4. Lieu PT, Heiskala M, Peterson PA, Yang Y. 2001. The roles of iron in h ealth and disease. Mol Aspects Med 22(1-2):1-87. Lindenau J, Noack H, Asayama K, Wolf G. 199 8. Enhanced cellular glutathione peroxidase immunoreactivity in activated astrocytes a nd in microglia during excitocin induced neurodegeneration. Glia 24:252-6. Ling E, Kaur C, Wong W. 1992. Ex pression of major histocompatibility complex antigens and CR3 complement receptors in activated microgl ia following an injection of ricin into the sciatic nerve in rats. Histo Histopathol 7(1):93-100. Lopes KO, Sparks DL, Streit WJ. 2008. Degene ration of microglial cells in the aged and Alzheimer's disease brains is associated with ferritin immunoreactivity. Glia 56( 10): 1048-60. Luber-Narod J, Rogers J. 1988. Immune system associated antigens expressed by cells of the human central nervous system. Neurosci Lett 94(1-2):17-22. Ma L, Morton AJ, Nicholson LFB. 2003. Microglia density decreases with age in a mouse model of Huntington's disease. Glia 43:274-280. Masliah E, Terry RD, DeTeresa RM, Hansen LA. 1989. Immunohistochemical quantification of the synapse-related protein synaptophysin in Alzheimer disease. Neurosci Lett 103:234239. Mazur A, Green S, Saha A, Carleton A. 1958. Mech anism of release of fe rritin iron by xanthine oxidase. J Clin Invest 37:1809-1817. McGeer PL, Itagaki S, Tago H, McGeer EG. 1987. Reactive microglia in patients with senile dementia of the Alzheimer type are positive for the histocompatibility glycoprotein HLADR. Neurosci Lett 79:195-200. Mildner A, Schmitdt H, Nitsche M, Merkler D, Hanisch UK, Mack M, Heikenwalder M, Bruck W, Priller J, Prinz M. 2007. Microglia in the adult brain arise from Ly-6ChiCCR2+ monocytes only under defined host c onditions. Nat Neurosci 10(12):1544-53. Milligan C, Cunningham T, Levitt P. 1991. Diff erential immunochemical markers reveal the normal distribution of brain macrophages and microglia in the developing rat brain. J Comp Neurol 314(1):125-35. 99

PAGE 100

Mittelbronn M, Dietz K, Schluese ner HJ, Meyermann R. 2001. Local distribution of microglia in the normal adult human central nervous system differs by up to one order of magnitude. Acta Neuropathol 101:249-255. Moran LB, Graeber MB. 2004. The facial nerve a xotomy model. Brain Res Brain Res Rev 44(23):154-178. Nair A, Bonneau RH. 2006. Stress-induced elevat ion of glucocorticoids increases microglia proliferation through NMDA receptor activation. J Neuroimmunol 171(1-2):72-85. Nimmerjahn A, Kirchhoff F, Helmchen F. 2005. Re sting microglial cells are highly dynamic surveillants of brain parenchym a in vivo. Science 308(5726):1314-1318. Ogawa H, Nishida A, Mito T, Takashima S. 1994. Development of ferritin-positive cells in cerebrum of human brain. Pediatr Neurol 10:44-48. Oide T, Kinoshita T, Arima K. 2006. Regression stage senile plaques in the natural course of Alzheimer's disease. Neuropath ol Appl Neurobiol 32:539-556. Overmyer M, Helisalmi S, Soininen S, Soininen H, Laakso M, Riekkinen P, Alafuzoff I. 1999. Reactive microglia in aging and dementia: an immunohistochemical study of postmortem human brain tissue. Acta Neuropathol 97:383-392. Peinado MA, Quesada A, Pedrosa JA, Torres MI, Martinez M, Esteban FJ, Del Moral ML, Hernandez R, Rodrigo J, Peinado JM. 1998. Quan titative and ultrastructural changes in glia and pericytes in the parietal cortex of the aging rat. Microsc Res Tech 43:34-42. Ponka P, Lok CN. 1999. The transfer rin receptor: role in health and disease. Int J Biochem Cell Biol 31:1111-1137. Qian ZM, Wang Q. 1998. Expression of iron transport proteins and excessive iron accumulation in the brain in neurodegenerative disord ers. Brain Res Brain Res Rev 27(3):257-267. Roberts ES, Masliah E, Fox HS. 2004. CD163 id entifies a unique population of ramified microglia in HIV encephalitis (HIVE). J Neuropath Exp Neurol 63(12):1255-1264. Roskams AJ, Connor JR. 1994. Iron, transferrin and ferritin in the rat brain during development and aging. J Neurochem 63:709-716. Rozovsky I, Finch C, Morgan T. 1998. Age-relate d activation of microglia and astrocytes: in vitro studies show persistent phenotypes of aging, increase d proliferation, and resistance to down-regulation. Neurobiol Aging 19:97-103. Schmidt OI, Heyde CE, Ertel W, Stahel PF. 2005. Closed head injury -an inflammatory disease? Brain Res Rev 48(2):388-399. 100

PAGE 101

Sedgwick JD, al. e. 1991. Isolation and direct char acterization of resident microglial cells from the normal and inflamed central nervous system. PNAS 88:7438-7442. Seehafer SS, Pearce DA. 2006. You say lipofusci n, we say ceroid: Defining autofluorescent storage material. Neurobiol Aging 27:576-588. Selkoe DJ. 2004. Alzheimer's disease: mechanisti c understanding predicts novel therapies. Ann Intern Med 140:627-638. Semmler A, Okulla T, Sastre M, Dumitr escu-Ozimek L, Heneka MT. 2005. Systemic inflammation induces apoptosis with variable vulnerability of different brain regions. J Chem Neuroanat 30:144-157. Sheffield LG, Berman NEJ. 1998. Microglial expr ession of MHC class II increases in normal aging of nonhuman primates. Ne urobiol Aging 19(1):47-55. Sierra A, Gottfried-Blackmore AC, McEwen BS, Bulloch K. 2007. Microglia derived from aging mice exhibit an altered inflammatory profile. Glia 55:412-424. Simard AR, Soulet D, Gowing G, Julien J-P, Rivest S. 2006. Bone marrow-derived microglia play a critical role in restri cting senile plaque formation in Alzheimer's disease. Neuron 49:489-502. Simmons DA, Casale M, Alcon B, Pham N, Na rayan N, Lynch G. 2007. Ferritin accumulation in dystrophic microglia is an early event in the development of Huntington's disease. Glia 55:1074-1084. Sitte N, Merker K, Grune T, von Zglinicki T. 2001. Lipofuscin accumulation in proliferating fibroblasts in vitro: an indicator of oxidative stress. Exp Neurol 36(3):475-86. Sloane JA, Hollander W, Moss MB, Rosene DL, Abraham CR. 1999. Increased microglial activation and nitration in white matter of the aging m onkey. Neurobiol Aging 20:395405. Stolzing A, Grune T. 2003. Impairment of prot ein homeostasis and de cline of proteasome activity in microglia l cells from adult Wistar rats. J Neurosci Res 71:264-271. Stolzing A, Sethe S, Grune T. 2005. Chronically active: activation of microglial proteolysis in aging and neurodegenera tion. Redox Rep 10(4):207-213. Stolzing A, Wengner A, Grune T. 2002. Degradation of oxidized extracellular proteins by microglia. Arch Biochem Biophys 400:171-179. Stolzing A, Widmer R, Tobias J, Voss P, Grune T. 2006. Tocopherol-mediated modulation of age-related changes in microglial cells: Tu rnover of extracellular oxidized protein material. Free Rad Biol Med 40:2126-35. 101

PAGE 102

Streit WJ. 2000. Microglial response to brain in jury: A brief synopsis. To xicol Pathol 28(1):2830. Streit WJ. 2001. Microglia and macrophages in the developing CNS. Neuro Toxicol 22:619-624. Streit WJ. 2002. Microglia as neuroprotective, immunocompetent cells of the CNS. Glia 40(2):133-139. Streit WJ. 2004. Microglia and Alzheimer's dis ease pathogenesis. J Neurosci Res 77(1):1-8. Streit WJ. 2006. Microglial senes cence: does the brain's immune sy stem have an expiration date? Trends Neurosci 29(9):506-510. Streit WJ, Miller KR, Lopes KO, Njie E. 2008. Microg lial degeneration in th e aging brain bad news for neurons? Front Biosci 13: 3423-38. Streit WJ, Sammons NW, Kuhns AJ, Sparks DL 2004. Dystrophic microglia in the aging human brain. Glia 45(2):208-212. Streit WJ, Sparks DL. 1997. Activation of microglia in the brains of hum ans with heart disease and hypercholesterolemic rabbi ts. J Mol Med 75(2):130-138. Streit WJ, Walter SA, Pennell NA. 2000. Reactive microgliosis. Prog Neurobiol 57(6):563-581. Sugama S, Fujita M, Hashimoto M, Conti B. 2007. Stress induced morphological microglial activation in the rodent brai n: involvement of interleuki n-18. Neuroscience 146(3):13881399. Taneja V, Mishra K, Agarwal KN. 1986. Effect of early iron deficiency in rat on the gammaaminobutyric acid shunt in br ain. J Neurochem 46(6):1670-1674. Terman A, Brunck UT. 2004. Lipofuscin. Int J Biochem Cell Biol 36:1400-1404. Thal DR, Capetillo-Zarate E, Del Tredici K, Br aak H. 2006. The development of amyloid-[beta] protein deposits in the aged brai n. Sci Aging Knowledge Environ 6:1-9. Torti FM, Torti SV. 2002. Regulation of fe rritin genes and protein. Blood 99:3505-3516. Twig G, Jung SK, Messerli MA, Smith PJ, Shirihai OS. 2001. Real-time detection of reactive oxygen intermediates from single mi croglial cells. Biol Bull 201:261-2. Vogt M, Haggblom C, Yeargin J, Christiansen -Weber T, Haas M. 1998. Independent induction of senescence by p16INK4a and p21CPI1 in spontaneously immortalized human fibroblasts. Cell Grow th Diff 9(2):139-46. 102

PAGE 103

von Eitzen U, Egensperger R, Kosel S, Grasbon-Frodl EM, Imai Y, Bise K, Mehraein P, Graeber MB. 1998. Microglia and the deve lopment of spongiform change in Creutzfeldt-Jakob disease. J Neuropath Exp Neurol 57:246-256. Wang W, Ji P, Dow KE. 2003. Cor ticotropin-releasing hormone i nduces proliferation and TNFrelease in cultured rat microglia via MAP kinase signalling pathways. J Neurochem 84(1):189-195. Wang W, Ji P, Riopelle RJ, Do w KE. 2002. Functional expression of corticotropin-releasing hormone (CRH) receptor 1 in cultured rat microglia. J Neurochem 80(2):287-294. Wierzba-Bobrowicz T, Lewandowska E, KosnoKruszewska E, Lechowicz W, Pasennik E, Schmidt-Sidor B. 2004. Degeneration of microglia l cells in frontal a nd temporal lobes of chronic schizophrenics. Fo lia Neuropathol 42(3):157-165. Wierzba-Bobrowicz T, Lewandowska E, L echowicz W, Stepien T, Pasennik E. 2005. Quantitative analysis of activated microglia, ramified and damage of processes in the frontal and temporal lobes of chronic schizophrenics. Folia Neuropathol 43(2):81-89. Wirenfeldt M, Babcock AA, Ladeby R, Lamb ertsen KL, Dagnaes-Hansen F, Leslie RGQ, Owens T, Finsen B. 2005. Reactive microgliosi s engages distinct responses by microglial subpopulations after minor central nervous system injury. J Neurosci Res 82:507-514. Xu H, Chen M, Manivannan A, Lois N, Forrester JV. 2008. Age-dependent accumulation of lipofuscin in perivascular a nd subretinal microglia in experimental mice. Aging Cell 7(1):58-68. Yamaguchi H, Hirai S, Morimatsu M, Shoji M, Ihara Y. 1988. A variety of cerebral amyloid deposits in the brains of the Alzheimer-typ e dementia demonstrated by beta protein immunostaining. Acta Ne uropathol 76(6):541-549. Yamasaki R, Zhang J, Koshiishi I, Sastradipura Suniarti DF, Wu Z, Peters C, Schwake M, Uchiyama Y, Kira J, Saftig P and othe rs. 2007. Involvement of lysosomal storageinduced p38 MAP kinase activation in the ove rproduction of nitric oxide by microglia in cathepsin D-deficient mice. Mol Cell Neurosci 35(3):573-84. Yanishevsky R, Mendelsohn ML, Mayall BH, Cris tofalo VJ. 1974. Proliferative capacity and DNA content of aging human diploid cells in culture: A cy tophotometric and autoradiographic analysis. J Cell Physiol 84:165-170. Ye SM, Johnson RW. 1999. Increased interleukin-6 expression by microglia from brain of aged mice. J Neuroimmunol 93:139-148. 103

PAGE 104

104 BIOGRAPHICAL SKETCH Kryslaine Oliveira Lopes was born in Manaus, Amazonas, Brazil. In her senior year in high school in Manaus, Kryslaine participated in an exchange student program that brought her to the United States. Kryslaine ended up fi nishing both her secondary and undergraduate education in the U.S.A. earning a bachelors degree with honors in biological sciences from Illinois State University in 2001. Kryslaine then joined the neuroscience department at the University of Chicago as a full-time research technician. In 2004, Krysla ine began her doctorate studies in neuroscience at the University of Flor ida College of Medicine under the mentorship of Dr. Wolfgang (Jake) Streit. Upon completion of her Ph.D. program, Kryslaine will join the department of neurosciences at th e University of California-San Di ego as a postdoctoral scholar.