Innate Response to Focal Necrotic Injury Inside the Blood ...

Innate Response to Focal Necrotic Injury Inside theBlood-Brain Barrier1

Jiyun V. Kim and Michael L. Dustin2

We have studied the initial innate immune response to focal necrotic injury on different sides of the mouse blood-brain barrierby two-photon intravital microscopy. Transgenic mice in which the promoter of the myeloid isoform of lysozyme drives GFP wereused to track granulocytes and monocytes. Necrotic injury in the meninges, but not the brain parenchyma, recruited GFP! cellswithin minutes that fully surrounded the necrotic site within a day. Recently, it has been suggested that microglial cells andastrocytes cooperate to mount a distinct response to laser injury behind the blood-brain barrier. We followed the microglialresponse in heterozygous knockin mice in which GFP replaces CX3CR1 coding sequence. Prior to injury, microglial cell bodieswere immobile over days, but moved to the laser injury site within 1 day. We followed astrocytes, which have been proposed tocooperate with microglial cells in response to focal injury, using transgenic mice in which glial fibrillary acidic protein promoterdrives GFP expression. Before injury fine astrocyte processes permeate the parenchyma. Astrocytes polarized toward the injuryin an ATP, connexin hemichannels, and intracellular Ca2!-dependent process. The astrocytes network established a cytoplasmicCa2! gradient that preceded the microglial response. This is consistent with astrocyte-microglial collaboration to mount this innateresponse that excludes blood leukocytes. The Journal of Immunology, 2006, 177: 5269–5277.

I n mammals, the innate immune system responds rapidly toapoptotic and necrotic cell death to limit damage to tissuesand escape of causal pathogens (1, 2). Apoptosis due to pro-

grammed cell death is generally noninflammatory whereas factorsreleased from necrotic cells can be proinflammatory in the absenceof pathogen-derived innate signals (2–4). The response to necroticinjury involves neutrophil and monocyte infiltration to suppressinfection and initiate repair (5, 6). The role of neutrophils andmonocytes in necrotic injury in the cerebral cortex is less welldefined (6, 7).

The blood-brain barrier (BBB)3 separates blood leukocytes thatwould normally be a primary cell type responding to necrotic in-jury from the brain parenchyma where necrotic cell death mighttake place in response to infection, toxin action, excitotoxicity, ortrauma (8). The BBB is composed of two layers. The first layer iscomposed of microvascular endothelial cells, which have abundanttight junctions. The second layer is the glia limitans that is formedby glial foot processes (8). The perivascular space between theendothelial cell and astrocyte-derived basement membranes is con-tinuous with the subarachnoid space of the meninges and is pop-ulated by perivascular macrophages, some of which expressCD11c and behave like immature dendritic cells (9–11). Largenecrotic injuries destroy both neural tissue and the BBB and hence

blood leukocytes play a major role in surveillance of these sites forinfection and tissue repair (12, 13). Experimental necrotic injuriescan be induced behind the intact BBB using a focused laser beam(14, 15).

Microglial cells are long-lived phagocytic cells of the myeloidlineage that populate the parenchyma of the CNS (11, 16, 17).Recent two photon laser scanning microcopy studies have dem-onstrated dramatic responses of microglial cells to laser-inducedfocal necrotic injury (14, 15). This process requires endogenousATP acting through metabotropic P2YRs (14). Astrocyte networkswere implicated in this response through experiments with inhib-itors of connexin hemichannels, which are known to be involved inATP-induced ATP release to relay signals (14). However, astro-cytes labeled with sulforhodamine 101 showed minimal morpho-logical changes in response to laser injury (15). The involvementof conventional leukocytes in this process has not beeninvestigated.

In this study, we examine the cellular response to necrotic injuryoutside and inside the BBB using two-photon laser scanning in-travital microscopy in the same thinned skull surgical preparation.We visualize blood leukocytes, microglial cells, and astrocytes us-ing different transgenic or knockin mouse strains in which the re-spective endogenous cell types express GFP. Necrotic injury out-side the BBB induces rapid leukocytic infiltration that walls off thenecrotic area, essentially forming a granuloma. In contrast, therewas no leukocyte infiltration behind the BBB unless direct vascu-lar injury was induced along with the necrotic tissue injury. By24 h after parenchymal injury behind the BBB, the microglial re-sponse also results in cell body translocation and formation of agranuloma-like lesion in which cell bodies surround the injury, butlack granulocytes and monocytes. We also observed a morpholog-ical polarization of the astrocyte cytoplasm toward the injury andthe formation of a cytoplasmic Ca2! gradient in the astrocyte pro-cesses. The astrocyte polarization and Ca2! response is blocked bythe same agents, which include ATP and gap junction blockers,that inhibit the microglial motility response. These observationssupport the model that astrocytes and microglial cells cooperate to

Molecular Pathogenesis Program, Skirball Institute of Biomolecular Medicine, De-partment of Pathology, New York University School of Medicine, New York, NY10016

Received for publication December 29, 2005. Accepted for publication July 28, 2006.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by National Institutes of Health and Dana Foundationgrants (to M.L.D.), Irene Diamond Foundation (to M.L.D.), and National CancerInstitute Training Fellowship Grant CA009161-31 (to J.V.K.).2 Address correspondence and reprint requests to Dr. Michael L. Dustin, MolecularPathogenesis Program, Skirball Institute of Biomolecular Medicine, Department ofPathology, New York University School of Medicine, 540 First Ave, New York, NY10016. E-mail address: [email protected] Abbreviations used in this paper: BBB, blood-brain barrier; YFP, yellow fluorescentprotein; FFA, flufenamic acid.

The Journal of Immunology

Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$02.00

generate an effective innate response to necrotic injury that ex-cludes blood leukocytes.

Materials and MethodsTransgenic mice

CX3CR1!/gfp mice used to visualize ramified microglial cells and perivas-cular cells (18) were a gift from D. Littman (New York University Schoolof Medicine, Howard Hughes Medical Institute, New York, NY); FVB/N-Tg(glial fibrillary acidic protein (GFAP)-enhanced GFP (eGFP))14Mes/Jused to image astrocytes (19) and Tg(Thy-1H-eYFP) mice used to imageneurons (20) were obtained from The Jackson Laboratory; Tg(LysM-eGFP) used to image granulocytes, monocytes, and macrophages (21) werea gift from T. Graf (Albert Einstein College of Medicine, Bronx, NY). Allmice were housed in specific pathogen-free conditions and treated in ac-cordance with Institutional Animal Care and Use Committee protocols ofNew York University School of Medicine.

Surgery and experimental procedures

Thinned and open skull intravital window surgeries were performed aspreviously described (14, 15). Anesthetized mice were maintained on 37°Cwarming plates and the intravital window was warmed through an objec-tive heater to maintain 37°C. Laser ablation and mechanical injury wereperformed as previously described (14). The size of injury induced is typ-ically "15 !m in diameter. The yellow fluorescent protein (YFP) signal inTg(Thy1H-eYFP) mice was used to define the depth at which brain pa-renchyma starts in the barrel cortex. Vascular or perivascular spaces weredefined by injection of 655 nm emitting quantum dots (R&D Systems) intothe blood or subarachnoid space, respectively. Quantum dots excite at thesame Ti-Sapphire wavelength as GFP (22).

Drugs were applied through open skull as previously described (14).Cell-permeant Ca2! chelator BAPTA-AM (23) and cell permeant Ca2!

indicator Fluo-4-AM (Molecular Probes) were applied and washed as pre-viously described (24, 25).

Two-photon microscopy

GFP, YFP, and quantum dots were excited with a mode locked TsunamiTi-Sapphire laser tuned to 920 nm (Spectraphysics) connected to a Bio-RadRadiance multiphoton microscope. Fluo4-AM was excited at 840 nm.Stacks of image planes were acquired using step sizes 1–3 !m to a depthof 100–200 !m using #40 or #60 water dipping objectives. The GFP andYFP signals were separated using a cyan fluorescent protein/YFP filtercube (22, 26). Time-lapse images were acquired at 1- to 1.5-min intervals.

Data analysis

Microglial and astrocyte process movements were followed using Volocitysoftware (Improvision) and intensities were quantified using ImageJ($http://rsb.info.nih.gov/ij/%) (14, 22, 26). Single-plane pseudocolor imageswere rendered with MetaMorph software (Molecular Devices). Cell counts

were performed within three-dimensional volumes to calculate cells/mm3.Values of p were determined using two-tailed t test.

ResultsLeukocytes response to focal injury

To determine the effect of tissue injury outside and inside the BBB,we used an intravital microscopy preparation in which the parietalbone of the skull is thinned to "30 !m (27). Two-photon laserscanning microscopy can then be used to image through thethinned bone (Fig. 1A, blue) into the meninges, which includes thedura and the subarachnoid space (Fig. 1B), and the cerebral cortexparenchyma (Fig. 1, C and D). When polyethylene glycol-coatedquantum dots were injected into the subarachnoid space, they dif-fused into the perivascular spaces surrounding large vessels in thecerebral cortex (Fig. 1, B–D). This perivascular space is outside theglia limitans.

We imaged blood leukocytes using transgenic mice in which themyeloid-specific lysozyme promoter drives expression of eGFP ingranulocytes, monocytes, and macrophages, but not parenchymalmicroglial cells (14). Initially, we focused on the effects of injuryoutside the BBB by focusing just beyond the thinned skull into themeninges. Before injury, there were scattered eGFP! cells presentin the meninges and lined meningeal blood vessels (Fig. 1E). It ispossible that the marginated eGFP! cells were a consequence oftrauma from the bone thinning process. To induce injury, the laserwas tuned to maximum power at 800 nm and held in place for 1 s.This is a nonlinear process rather than simple heating so the vol-ume in which cells are killed is tightly restricted to "15-!m di-ameter circular volume at the laser focus. Focal injury induced inthe meninges caused a rapid recruitment of eGFP! cells to theinjury site within 30 min (Fig. 1, F and G, and supplemental movie1).4 Many of these cells were drawn from the intravascular poolthat was commented on previously. The median speed of eGFP!

cell movement during this process was 5.23 & 1.78 !m/min (range2.11 to 9.03, n ' 30 cells). The speeds were bimodal with popu-lations centered at 7 and 3 !m/min. After a day, the injury wasentirely walled off by the eGFP! cells in a granuloma-like struc-ture (Fig. 1H). Therefore, necrotic injury outside the BBB inducesrapid blood leukocyte extravasation and convergence on theinjury site.

4 The online version of this article contains supplemental material.

FIGURE 1. Blood leukocyte response to focal injuryoutside and inside the BBB. A–D, Perivascular and sub-arachnoid spaces are continuous. Quantum dots (white)injected into the subarachnoid spaces (SAS) outline thecontinuity of SAS with the perivascular space (PVS)under the thinned skull (blue) and are excluded from theparenchymal spaces where the neurons reside (red).Scale bar, 60 !m. E–H, LysM-eGFP-transgenic mouse.Leukocyte response to meningeal injury. Laser ablationat the surface of the brain causes an immediate recruit-ment of eGFP! cells from postcapillary venules. Timepoints are before injury (BL), 3, 30 min, and day 1.Scale bar, 50 !m (also see supplemental movie 1). I–L,Leukocyte response to brain parenchyma injury. Injuryin the brain parenchyma causes no recruitment to theinjury site (BL, 3, 30 min), but some perivascular ac-cumulation of eGFP! cells is induced (day 1) (see TableI and supplemental movie 1).

5270 COORDINATED RESPONSE OF MICROGLIA AND ASTROCYTES

We next asked whether laser injury in the parenchyma recruitedblood or perivascular leukocytes through the BBB. After the skullthinning, there were a few perivascular eGFP! cells (6 per 0.02mm3, range 2–11, n ' 4). These cells were extensively spread,sessile, and displayed probing movements consistent with theiridentification as macrophages (Fig. 1I, supplemental movie 1).Elongated eGFP! blood leukocytes were often observed flowingrapidly through small capillaries that were detectable with intra-vascular quantum dots (supplemental movie 1). The cerebral cor-tex parenchyma did not contain eGFP! cells (Fig. 1I). As a con-trol, we also imaged mice subjected to the thinned skull procedure,but without laser injury and the basal pattern did not change overdays (Table I). Parenchymal laser injury did not lead to recruitmentof eGFP! cells to the injury focus (Fig. 1, J and K, and supple-mental movie 1) or at 1 day (Fig. 1L and supplemental movie 1).Parenchymal laser injury recruited eGFP! cells to the perivascularspaces in the vicinity of the injury between days 1 and 7 andsubsided to control levels by day 10 (Table I (n ' 4 animals)).Thus, laser injury behind the BBB was weakly detected through anincrease in perivascular leukocytes over a period of days, but thesecells did not enter the parenchyma.

Microglia response to focal injury

Do microglial cells in the parenchyma induce a granuloma-likeencapsulation of the injury focus in the parenchyma? This is animportant question because organotypic culture of brain slices re-vealed that the soma of microglial cells can move within the tissue(28), while the initial response to focal injury in the intact brain isrestricted to the processes of the ramified microglial cells (14). Tovisualize microglial cells, we imaged cytoplasmic eGFP inCX3CR1gfp/! knockin mice (18). We determined the position ofmicroglial cell bodies over a period of days in the intact CNS usingadult CX3CR1gfp/! # Tg (Thy1H-YFP) mice. The Tg(Thy1H-YFP) mice were used to visualize the neural scaffold (27). Weimaged CX3CR1! ramified cells at day 0 and day 3 in the samevolume as defined by the highly stable pattern of neural processesand dendritic spines. Among 62 ramified microglial cells examinedin three separate fields, none of these cells displayed a positionalchange of the soma over 3 days (Fig. 2, A and B). The processtermini of the ramified microglial cells oscillated with a mean

speed of 1.73 !m/min (range 0.22–6.49; n ' 8 animals; p ( 0.01).Thus, ramified microglial cell bodies were sessile in vivo.

We next examined the impact of focal parenchymal injury onthe movement of microglial processes and soma in the first hourand then over longer periods. Sterile focal injuries were created inthe parenchyma of CX3CR1gfp/! # Tg(Thy-1H-eGFP) mice withthe laser. Red quantum dots were used to mark the vasculature toavoid induction of perivascular injury. As previously reported, theresponse at 30 min after injury was based entirely on movement ofprocesses, not soma (Fig. 2, C and D, and supplemental movie 2).We quantified the movement of individual microglial processesearly in the response and found that they moved with a median rateof 0.82 !m/min (range 0.57–1.97; n ' 7 animals). Thus, thesetethered processes move "6-fold slower than leukocytes converg-ing on an injury site in the meninges.

We had previously reported no microglial soma movement to-ward the injury site over "9 h of imaging (14). However, when weexamined microglial cells 24 or 72 h after injury, we found thatmicroglial cells with amoeboid and short ramified processes con-verged on the injury sites (n ' 5 animals) (Fig. 2, E and F, andsupplemental movie 2). In contrast, the neural processes remainedin place (Fig. 2F). The bodies of the microglial cells reached theinjury site by 24 h and conversion to amoeboid shape peaked at 3days, when the injury site had lost most of its autofluorescence andwas instead brightly marked by GFP! microglial cell bodies,which could be distinguished by movement and fluorescence spec-trum from the injury itself (Fig. 2F). The injury focus was resolvedby day 10 (data not shown). This observation suggests that theinjured tissue may have undergone phagocytosis by microglialcells. In addition, the activated microglial cells in the region sur-rounding the injury also relocated. There was a 19% increase in thenumber of microglial cell bodies in 20–40 !m radius from theinjury center (range !14 to !25, n ' 4 animals) and a 20% de-crease in 40–60 !m radial space (range )15 to )30%; n ' 4animals), strongly supporting the soma convergence toward theinjury. There was no evidence that CX3CR1 cells were recruitedfrom *150 !m from the injury site because there was no differ-ence in the cell number within the 150 !m radius between the timeimmediately after injury and 3 days later (60 & 3 cells/0.01 mm3,58 & 6 cells/0.01 mm3, n ' 4 animals). Thus, while perivascular

Table I. Perivascular recruitment of eGFP! cells to parenchymal laser injury in LysM-eGFP-transgenicmicea

Day

0 1 4 7 10 14

Animal 1Injury side 4 6 10 16 0 0Control side 4 5 11 0 0 0




a LysM! cells were counted in a volume of 0.02 mm3 at 60 !m below the bony surface where the brain parenchyma starts.Two thinned skull windows were generated, one for each hemisphere. Laser injury was induced on one side in a parenchymalzone and no injury on the opposite side as a control. The cells were counted on day 0 and several days following injury. Althoughthe number of eGFP! cells involved was small compared to the meningeal injury and the peak days ranged from days 1 to 7between the animals, there were on average 5.3 more cells/0.02 mm3 recruited in the perivascular areas to injury side than thecontrols during days 1 and 7, with an overall statistical significance of p ( 0.05 (two-tailed t test). No statistical significance wasreached between the injured and controls side on days 0 and 14.

5271The Journal of Immunology

macrophages and monocytes also express CX3CR1 there were nochanges in cell number in the vicinity of the injury that wouldsuggest recruitment of cells that were not present in the paren-chyma at day 0.

We conclude from these studies that microglial cells respond toinjury in two phases: an acute phase of process movement (0–30min) and a slower phase of soma movement (1–3 days).

Astrocyte response to focal injury

Astrocytes were examined to determine whether they undergomorphological or physiological responses to injury that are con-sistent with the proposed role in ATP-dependent amplification ofsignaling required for the rapid microglial response (14). We wereparticularly interested in this as a paradigm for tissue cell partic-ipation in guiding innate responses of myeloid cells. We usedTg(GFAP-eGFP), whose eGFP signal was detectable up to 100!m from the surface of parenchyma (19, 28, 29). Using brain slicepreparations of this mouse, Kirchhoff and colleagues (29) demon-strated fine filopodia-like branches from the apical astrocyte pro-cess that undergo extension/retraction cycles. In our in vivo im-aging, these fine processes filled the space in a dense mannergenerating a “mossy” effect where it was not possible to identifyindividual processes (146 & 11 cells/0.01 mm3; n ' 7 animals,Fig. 3A). The most striking aspect of the astrocyte processes is thatunlike the microglial processes that surveyed a large volumethrough active probing of the parenchyma, the astrocyte processesfilled the parenchyma to such as extent that only small movementsnot readily detectable with two photon microscopy through thethinned skull would be necessary to survey all extracellular spacein the parenchyma (Fig. 3A, supplemental movie 3).

The ability to visualize astrocyte processes using a bright cyto-plasmic marker led us to reinvestigate the morphological responseof astrocyte processes to focal injury in vivo. We determined theastrocyte response to laser damage in Tg(GFAP-eGFP) mice fromimmediately post injury to 48 h. After laser injury, we observed a46% (n ' 6) decrease in GFP fluorescence in a halo-like zonearound the bright, autofluorescent injury site within the first fewseconds to minutes following the injury (Fig. 3B, supplementalmovie 3). This reduction in fluorescence was not due to photo-bleaching because experiments in both CXCR3gfp/! and Thy-1-eYFP Tg mice did not display photobleaching in the same region(data not shown). Because it was hard to resolve individual pro-cesses, the reduced fluorescence could be due either to a retractionof some processes or a reduction in cytoplasmic volume within thesame number of processes. It was clear that some processes remainin contact with the injury site. This result is consistent with theearlier reported decrease in sulforhodamine 101 fluorescence nearthe laser injury site (15). The same study noted no recovery ofsulforhodamine fluorescence. In contrast, we observed postinjurychanges in GFP fluorescence associated with thickening of a fewprocesses in halo zone within 5–10 min (Fig. 3, C and D). Theformation of these thick processes was maximal by 30 min, whichcorresponded to the time frame for the microglial processes toreach the injury focus (Fig. 3, E–G). Astrocytes are polarized cellswith one major projection corresponding to the apical pole (29).Thus, the process thickening toward the laser injury site is mostlikely a reorientation of neighboring astrocytes toward the injury.This was followed by a slower phase of recovery between 30 minto 48 h, by which time the GFP signal in the fine processes wasfully recovered except where excluded from a tight cuff around the

FIGURE 2. Microglial cell body movement in re-sponse to injury. A and B, Microglial cell bodies aresessile. CX3CR1gfp/! # Tg(Thy-1H-YFP) mice. Num-bered microglial cell bodies (green) positions on day 0(A) and day 3 (B) were imaged relative to the stableneural scaffold (red) through the thinned skulled in theabsence of injury. Scale bar, 50 !m. C and D, Acutemicroglial reaction to parenchymal injury. CX3CR1gfp/!

mouse. Microglial cells (green) 1 min (C) and 30 min(D) after focal laser injury (yellow, arrowhead). Vesselsare highlighted by quantum dots (red). Scale bar, 50 !m.E and F, Late microglial cell response to injury. After 1day (E), the microglial cell bodies have reached the laserinjury site. Microglial cell bodies are still present aroundthe reduced injury site (F). The neural scaffold inCX3CR1gfp/! # Tg(Thy-1H-YFP) (red) does notchange over the same period (F). Scale bar, 50 !m (seesupplemental movie 2).


injury site, most likely corresponding to the microglial cell bodies(Fig. 3G, supplemental movies 3). Using Tg(GFAP-eGFP) to markastrocyte cytoplasm, we observed a distinct cytoplasmic polariza-tion of astrocytes toward the injury.

A more detailed analysis of this polarization process in responseto injury revealed that the large astrocyte processes remain in con-tact with the injury site based on visible eGFP! strands (Fig. 4,A–F, supplemental movie 3, and data not shown). The enlargedprocesses displayed irregular, saccadic motions consistent withtheir viability (n ' 3 animals). Because there appeared to be manysmall viable processes that continued to permeate the halo aroundthe injury it is not clear how a few processes are selected to en-large. This might be explained by a cell polarity mechanism thatcan reinforce initial selection of a dominant process.

Signaling mechanisms involved in the astrocyte response

We next asked whether astrocyte polarization requires a gradientof extracellular ATP as reported for microglial cells (14). Wetested pharmacological interventions that had previously beenshown to block the microglial response (30–33). ATP is a putativechemoattractant that requires formation of a gradient. Applicationof large amounts of ATP through an opening in the skull obscuresthis gradient (14). The ATP gradient can also be destroyed bydegrading endogenous ATP with apyrase (14). Applications of 5mM ATP, UTP, or CTP to the cortex in vivo before laser injury didnot inhibit the initial reduction of eGFP fluorescence around theinjury site (n ' 3 animals) (Fig. 4, G–J, K–N, and O–R, supple-mental movie 3). However, ATP and UTP, but not CTP, abolishedthe polarization response (Fig. 4, G-J, K–N, and O–R). Therefore,ATP signaling, possibly through P2Y2 (34), was required for thepolarization of cytoplasm to astrocyte processes connected to theinjury site. We further found that the ATP-degrading enzymeapyrase also blocked polarization (Fig. 4, S–V). Unlike the micro-glial reaction, which was restored soon after washout of apyrase

(14), astrocyte polarization was not restored after washout ofapyrase (n ' 3, each). A second injury in an adjacent location afterapyrase washout progressed normally with intact astrocyte polar-ization, demonstrating that apyrase was effectively removed. Thus,astrocyte polarization to the injury was dependent upon earlyevents after injury and did not take place when the early responseswere blocked. Furthermore, astrocyte polarization as definedabove was not necessary for the microglial movement to the injury,which does recover fully after apyrase washout.

We next asked whether astrocyte polarization was dependentupon connexin hemichannels (33). Connexin hemichannels ex-pressed by astrocytes are implicated in release of ATP to relay thesignal to adjacent cells such as microglial cells and in the entry ofCa2! into the cytoplasm of the astrocytes to mediate activation-related signals in the astrocyte network (23, 32). The connexinhemichannel inhibitor flufenamic acid (FFA) reversibly blocks mi-croglial process convergence to laser injury (14). We found thatFFA abolished astrocyte polarization after laser injury (Fig. 4,W–Z, n ' 3). Although the removal of FFA restores microglialprocess convergence (14), the astrocyte polarization was not re-stored after FFA washout. However, astrocyte polarization wasseen in new injuries after FFA washout (n ' 3, each, data notshown). We also tested the effect of BAPTA-AM, an intracellularCa2! chelator that would block the Ca2! signal, on astrocyte polar-ization after laser injury. BAPTA-AM blocked the astrocyte polariza-tion response (Fig. 4, AA–DD) and also blocked the microglial re-sponse (data not shown). Rapid reversal of BAPTA-AM was notpossible because the active form is a highly charged molecule thatbecomes trapped in the cytoplasm of cells after de-esterification andis expelled slowly following washout. The accumulated evidence in-dicates that the astrocyte cytoplasmic polarization is a component ofan astrocyte response. This response takes place in parallel to themicroglial response, but astrocyte polarization of eGFP to the injurywas not necessary for microglial movement to this site.

FIGURE 3. Astrocyte polarization to injury. All im-ages in Tg(GFAP-eGFP) mice with eGFP signal onpseudocolor scale as indicated. A, Before injury, astro-cyte processes fill the parenchymal volume except forblood vessels- baseline (BL). Scale bar, 25 !m. B–G,Series of time points from a 1-h time lapse and on day2 after laser injury. Upon a focal laser injury, a halo oflow GFP signal forms around the autofluorescent injurysite (B, 1 min), followed by polarization of GFP fluo-rescence in the form of enlarged processes (C–F, 10, 20,30 min, and 1 h). By day 2, GFP signal is further re-covered in the halo area with thickened processes com-pared with the baseline with a residual area with no GFPsignal around the laser injury corresponding to the lo-cation of microglial call bodies. Scale bar, 25 !m (seesupplemental movie 3).


Cytoplasmic calcium gradient forms surrounding the injury site

We next asked whether we could detect function of the many as-trocytic processes that remained in the halo region. ExtracellularATP triggers astrocytic network signaling through calcium in vitro(35) and ATP injection has been shown to elicit calcium waves inbrain slice studies (32). Relative Ca2! concentration in the astro-cyte cytoplasm can be detected with Fluo-4 AM applied to thecerebral cortex through an open skull preparation (24). AlthoughFluo-4AM is taken up by all cells in the cortex the majority of thesignal comes from astrocytes, which as described above, fill mostof the volume of the brain and are more active in taking up andde-esterifying the dye (24). Fluo-4 AM signal has been shown tobe in the astrocytes cytoplasm selectively by sulforhodamine 101double labeling (36). The experiments were performed inCX3CR1gfp/! mice because this allowed monitoring of the micro-glial response directly and the GFP! microglial processes occupyso little of the volume of the cortex that they did not interfere withthe measurement of the Fluo-4 signal (Fig. 5A). Immediately afterthe injury the Fluo-4 fluorescence in the injury site had a similarprofile to the GFP fluorescence in the Tg(GFAP-eGFP) mice withstrands of signal in the halo (20 !m from the bright autofluores-cence at the injury (n ' 3 animals) (data not shown). Fluo-4 flu-orescence increases specifically near the injury as early as 1 min

following injury, preceding the microglial reaction (Fig. 5, B–F,supplemental movie 4) (n ' 3 each). The Fluo-4 fluorescenceformed a gradient that can be observed to decline *5-fold over adistance of 100 !m from the injury site (Fig. 5, B–F). At theresolution of our imaging acquisition rate and within the field size(300 !m # 300 !m), the Ca2! signaling was not in the form oftemporal waves as previously reported in tissue culture (23) orbrain slice (32), but in vivo we observed a standing gradient ex-tending over the astrocyte network that remained in place over theduration of microglial movement with maximal increase during thefirst 60 min (Fig. 5, B–F). Bathing the open cortex with 5 mM ATP(Fig. 5, G–I, supplemental movie 4) or with 300 U/ml apyrasebefore injury abolished the Ca2! gradient, but the Ca2! gradientwas restored with apyrase washout (Fig. 5, K–O) (n ' 3, each) (1).The recovery of the Ca2! signal is readily detectable, but lower inmagnitude than the signal resulting from a fresh injury. The re-covery of microglial reaction is also slower after washout ofapyrase than in a fresh injury (data not shown). CTP preincubation,in contrast had no effect on the injury induced Ca2! gradient (datanot shown). Thus, ATP is required for Ca2! gradient formation.Fluo-4 AM itself slowed the microglial response by 2-fold, con-sistent with a mild Ca2! buffering effect of Fluo-4AM. FFA andFluo-4 AM together appeared to cause toxicity to microglial cells

FIGURE 4. Astrocytic polarization to laser injury re-quires an ATP gradient, gap junction hemichannels, andcytoplasmic Ca2!. All open skull images of Tg(GFAP-eGFP) mice. A–F, Astrocyte cytoplasm polarizes towardlaser injury (!). The astrocyte processes (gray scale) donot lose contact with the injured site (A and B). A fewprocesses polarize through enlargement with saccadicmotions (white arrowheads) over a period of 30 min.Scale bar, 10 !m. G–R, Preincubation with 5 mM ATP(G–J) and UTP (K–N), but not CTP (O–R) abolishedastrocyte polarization. S–V, Preincubation with 300-U/ml apyrase also abolished astrocyte polarization.Scale bar, 10 !m. W–Z, Effect of FFA on astrocyte po-larization. Preincubation of 1 mM FFA (reversiblehemichannel inhibitor) abolished polarization. AA–DD,Effect of BAPTA-AM on astrocyte polarization. Prein-cubation with 100 !M BAPTA-AM (intracellular cal-cium chelator) abolished the astrocyte polarization.Scale bar, 10 !m (see supplemental movie 3).


based on loss of GFP fluorescence in CX3CR1gfp/! mice so thiscombination could not be used to study the role of connexinhemichannels in the generation of the Ca2! gradient (data not shown).

We conclude that focal injury generates a standing Ca2! gradi-ent in the astrocyte network that depends upon extracellular ATPand gap junction hemichannels.

DiscussionIn this study, we have compared the innate cellular responses tofocal necrotic injury induced outside or behind the BBB. Outsidethe BBB LysM-eGFP! cells (mostly neutrophils) rapidly leave thevascular space and surround the injury focus. Behind the intactBBB there was no invasion of the parenchyma by LysM-eGFP!

cells, but a robust albeit slower response by microglial cells andastrocytes. Our results support a model first proposed by Davalos

et al. (14) in which astrocyte networks guide microglial cells to theinjury (Fig. 6, A and B). Although these responses are very distinctin terms of the responding cell populations, the end result of anecrotic lesion fully surrounded by cell bodies of phagocytic cellsis similar. Although myeloid cells did not invade the CNS paren-chyma, we show that they were recruited in the nearby perivascu-lar areas over days (Fig. 6C).

There are a number of mechanisms that could account for theexclusion of LysM-eGFP! cells from the brain parenchyma duringresponses to necrotic injury. One is that the hydrodynamics ofblood flow in the brain may be consistent with neutrophil margin-ation in the meninges, but not parenchymal microcirculation. It hasbeen noted that parenchymal microcirculation is very high velocitycompared with meningeal microcirculation (37). A second possi-bility is that the intact glia limitans may not allow blood leukocytes

FIGURE 5. Cytoplasmic Ca2! gradient radiatesfrom injury site. A–F, The Ca2! signal (Fluo4-AM flu-orescence) and microglial cells (CX3CR1-eGFP fluo-rescence) were followed using same fluorescence chan-nel. Because microglial cells are discrete the Fluo-4fluorescence could be monitored between the microglialcell processes. The Fluo-4 signal rises around the laserinjury site continuously for 30 min (pseudocolor scale).Ca2! gradient formation precedes the movement of themicroglial process (red arrowheads) to the injury focus(red asterisk) (see supplemental movie 4). Scale bar, 50!m. G–I, Effect of ATP on calcium gradient formation;5 mM ATP preincubation inhibits formation of calciumgradient around the injury focus (red asterisk) and themicroglial reaction (supplemental movie 4). Scale bar,50 !m. J, Ca2! gradient formation. Fluo-4 fluorescenceprofile from linear segment indicated by the yellow linein B and H. F/F0 (where F0 is the fluorescence at time 0min after injury) shows that there is a rapid formation ofa Ca2! gradient with a maximum "5-fold decrease overa distance of 100 !m from the injury site (high) to theedge of the field (low). Pretreatment with 5 mM ATPblocks Ca2! gradient formation (black, 1 min, lightblue, 15 min, gray, 30 min). K–O, Effect of apyrase andwashout of apyrase on Ca2! gradient formation. Prein-cubation with 300 U/ml apyrase inhibits Ca2! gradientformation. Upon washout, the Ca2! gradient formsaround the injury site (red asterisk). Scale bar, 50 !m(see supplemental movie 4).


to enter the parenchyma. This is consistent with the observationthat more than 1 week after laser injury many LysM-eGFP! cellsaccumulate only in the perivascular spaces (Fig. 6C). These cellsmay have migrated in from the subarachnoid space in response tosignal from the resolving necrotic injury. The only barrier thatwould prevent these cells from entering the parenchyma from theperivascular space is the glia limitans. Finally, the coordinatedmovement of the microglia and astrocytes may act functionally asa second line BBB in intercepting signals to blood leukocytes thatkeep these chemotactic signals at a low level in the vasculature andwork with the BBB to reduce neutrophil and monocyte entry.

The astrocyte/microglial axis forms a first-line innate immunemechanism in the parenchyma to detect mediators released fromdead or injured cells using the ATP signaling system (4, 38). Al-though leukocytes can also use ATP as a signaling molecule (33,39, 40), it is not clear whether ATP is used in the leukocyte re-sponse in the meninges. The meninges are an important site forinfections by viruses and bacteria so understanding this responsewill be important. We limit our detailed focus on further under-standing the collaboration between microglial cells and astrocytebehind the BBB, which serves as a model for tight collaboration oftissue cells with myeloid cells to guide cell migration.

We confirm the rapid movement of microglial projections tofocal injury as previously published (14, 15). We extended ourintravital observations to days and found the microglial cell bodiesdirectly surround the injury focus as early as 24 h (Fig. 6A). Thismore extensive response may facilitate phagocytosis of necroticmaterial from the injury site and containment of pathogens thatmight be a cause of cell necrosis. We found that the astrocytesundergo changes in morphology and Ca2! signaling that are fullyconsistent with their playing a key role in this process, althoughnone of the experiments we have performed directly demonstratea causal link between astrocyte signaling and microglial responses.Astrocytes polarize their cytoplasm toward the injury site in par-

allel to the convergence of the microglia, but astrocyte polarizationwas not needed for microglial process movement toward the injurysite (Fig. 6A). The astrocyte cytoplasm generates a Ca2! gradienttoward the injury focus that is well-correlated with the microglialresponse (Fig. 6B). Our results establish a strong link in that onlyastrocytes are known to have gap junction hemichannels thatwould be blocked by FFA, but it is possible that FFA could havesome direct effect on microglial cells. Similarly, ATP, apyrase, andBAPTA-AM may all have effects directly on microglial cells.More specific genetic tools are needed to forge a causal link betweenthe astrocyte Ca2! gradient and microglial responses, but our resultshere provide motivation for future effort in this direction.

Previous organ and cell culture studies demonstrated a Ca2!

wave in the astrocytic network in response to ATP injection in ahemichannel but not gap junction-dependent manner (32, 33). Incontrast to these brain slice studies, we report for the first timeformation of a standing gradient of cytoplasmic Ca2! in the as-trocyte network around the injury site. It is possible that the levelof damage in our model is not sufficient to induce Ca2! waves asin an epilepsy model (25). The Ca2! gradient formed around focalinjury sites was well-correlated with the rapid response of micro-glial processes, which is only observed in the in vivo setting. It isnot clear how this Ca2!gradient is maintained in the extensivelyinterconnected astrocyte network and whether the triggering of thegradient requires a direct astrocyte injury besides the ATP releasefrom the injured cell(s).

The movement of the astrocyte cytoplasm toward the injury sitemay represent a classical polarization response. Astrocytes havebeen studied as a model for cell polarization (41). Neural cells arederived from neuroectoderm and thus are epithelial in origin withdistinct apical and basolateral surfaces. At baseline, astrocyte cal-cium activity is lower and the cell bodies and their major processextend in random orientations in the brain parenchyma. Astrocytepolarization toward sites of injury may allow them to contribute to

FIGURE 6. Working hypothesis and model. A, Morphological response of glia. Four time points are illustrated in a clockwise fashion starting from thetop left panel: before, 1 min, 30 min, and 1–3 days after a focal laser injury (yellow). Before: At baseline, microglia (green) and astrocytes (orange) activelysurvey their environment with their processes without soma displacement. 1 min: Upon injury ATP is released (gray clouds) and a P2YR-dependent ATPrelay signal is activated. Some astrocyte processes remain attached to the injury focus at 30 min: There is a polarization of the astrocyte cytoplasm andmicroglial processes reach the laser injury site; 1–3 days: The injury focus is covered with microglial cell bodies. B, Ca2! gradient formation after injury.The same four time points as in A are illustrated with the Ca2! signals in pseudocolor scale within the astrocytes; 1 min: a Ca2! gradient begins to formaround the injury site preceding the microglial response in the strands of astrocyte processes that remain attached to the injury focus; 30 min: the Ca2!

gradient continues to increase rapidly up to 30 min and subside over several days (1–3- days). C, Blood leukocyte cell response to parenchymal injury.Blood leukocytes remain in the perivascular spaces and recruited into the perivascular space after parenchymal injury.


the process of containing the injury, although this polarization was notrequired for the microglial convergence that leads to full containmentof the lesion by microglial cell bodies. Alternatively, the enlargementof these processes toward the injury site may be a response to injurythat does not involved polarization machinery in the cell. Other mark-ers of astrocyte polarity could be investigated to resolve this.

Focal laser injury is different from other models like photoco-agulation to model stroke or freezing to model larger necrotic in-juries in terms of the small and controlled size of injury ("volumeof a single cell) as well as minimal damage to the BBB near theinjury (14, 15, 42–44). Therefore, this study highlights the criticalrole played by the microglia and astrocytes in the innate responseto tissue injury that may include active maintenance of the BBB inthe day-to-day trauma-ridden environment faced by mammals. Wedid not examine larger injuries that span the vasculature and resultin loss of the BBB. In larger injury models such as in stroke andtrauma, the mechanical breakdown of the BBB will enable initialinfiltration of leukocytes with inflammatory tissue damage (6).Even in large necrotic injuries with reperfusion the role of bloodleukocytes may be limited (6, 7). Although neutrophils may play arole in reperfusion injury in stroke by damaging the vasculature,our results suggest that they may not enter live brain parenchymawhere the astrocyte-microglial reaction is capable of walling offinjuries. Indeed, in brain slice studies, activated microglial cells arefound all around the edge of the tissue (45).

AcknowledgmentsWe thank Guy Shakhar, Dimitrios Davalos, Jason Lieberthal, WenbiaoGan, Dan Littman, Juan Lafaille, and the members of the Dustin laboratoryfor helpful discussions. We thank Jianzhong Li for help with Fig. 2E. Wethank Dan Littman and Thomas Graf for their gift of mutant and transgenicmice used in our study.

DisclosuresThe authors have no financial conflict of interest.

References1. Cohen, J. J., R. C. Duke, V. A. Fadok, and K. S. Sellins. 1992. Apoptosis and

programmed cell death in immunity. Annu. Rev. Immunol. 10: 267–293.2. Aderem, A., and D. M. Underhill. 1999. Mechanisms of phagocytosis in mac-

rophages. Annu. Rev. Immunol. 17: 593–623.3. Edinger, A. L., and C. B. Thompson. 2004. Death by design: apoptosis, necrosis

and autophagy. Curr. Opin. Cell Biol. 16: 663–669.4. Shi, Y., J. E. Evans, and K. L. Rock. 2003. Molecular identification of a danger

signal that alerts the immune system to dying cells. Nature 425: 516–521.5. Tyler, M. P., A. M. Watts, M. E. Perry, A. H. Roberts, and D. A. McGrouther.

2001. Dermal cellular inflammation in burns: an insight into the function ofdermal microvascular anatomy. Burns 27: 433–438.

6. Clark, R. K., E. V. Lee, R. F. White, Z. L. Jonak, G. Z. Feuerstein, andF. C. Barone. 1994. Reperfusion following focal stroke hastens inflammation andresolution of ischemic injured tissue. Brain Res Bull. 35: 387–392.

7. Hayward, N. J., P. J. Elliott, S. D. Sawyer, R. T. Bronson, and R. T. Bartus. 1996.Lack of evidence for neutrophil participation during infarct formation followingfocal cerebral ischemia in the rat. Exp. Neurol. 139: 188–202.

8. Lassmann, H., F. Zimprich, K. Vass, and W. F. Hickey. 1991. Microglial cells area component of the perivascular glia limitans. J. Neurosci. Res. 28: 236–243.

9. Peters, P., S. L. Palay, and H. D. Webster. 1991. The Fine Structure of theNervous System. Oxford University Press, New York.

10. Engelhardt, B., and R. M. Ransohoff. 2005. The ins and outs of T-lymphocytetrafficking to the CNS: anatomical sites and molecular mechanisms. Trends Im-munol. 26: 485–495.

11. Aloisi, F., F. Ria, and L. Adorini. 2000. Regulation of T-cell responses by CNSantigen-presenting cells: different roles for microglia and astrocytes. Immunol.Today 21: 141–147.

12. Ghirnikar, R. S., Y. L. Lee, and L. F. Eng. 1998. Inflammation in traumatic braininjury: role of cytokines and chemokines. Neurochem. Res. 23: 329–340.

13. Fitch, M. T., C. Doller, C. K. Combs, G. E. Landreth, and J. Silver. 1999. Cellularand molecular mechanisms of glial scarring and progressive cavitation: in vivoand in vitro analysis of inflammation-induced secondary injury after CNS trauma.J. Neurosci. 19: 8182–8198.

14. Davalos, D., J. Grutzendler, G. Yang, J. V. Kim, Y. Zuo, S. Jung, D. R. Littman,M. L. Dustin, and W. B. Gan. 2005. ATP mediates rapid microglial response tolocal brain injury in vivo. Nat. Neurosci. 8: 752–758.

15. Nimmerjahn, A., F. Kirchhoff, and F. Helmchen. 2005. Resting microglial cellsare highly dynamic surveillants of brain parenchyma in vivo. Science 308:1314–1318.

16. Hickey, W. F. 2001. Basic principles of immunological surveillance of the nor-mal central nervous system. Glia 36: 118–124.

17. Ransohoff, R. M., P. Kivisakk, and G. Kidd. 2003. Three or more routes forleukocyte migration into the central nervous system. Nat. Rev. Immunol. 3:569–581.

18. Jung, S., J. Aliberti, P. Graemmel, M. J. Sunshine, G. W. Kreutzberg, A. Sher,and D. R. Littman. 2000. Analysis of fractalkine receptor CX3CR1 function bytargeted deletion and green fluorescent protein reporter gene insertion. Mol. CellBiol. 20: 4106–4114.

19. Nolte, C., M. Matyash, T. Pivneva, C. G. Schipke, C. Ohlemeyer, U. K. Hanisch,F. Kirchhoff, and H. Kettenmann. 2001. GFAP promoter-controlled EGFP-ex-pressing transgenic mice: a tool to visualize astrocytes and astrogliosis in livingbrain tissue. Glia 33: 72–86.

20. Feng, G., R. H. Mellor, M. Bernstein, C. Keller-Peck, Q. T. Nguyen, M. Wallace,J. M. Nerbonne, J. W. Lichtman, and J. R. Sanes. 2000. Imaging neuronal subsets intransgenic mice expressing multiple spectral variants of GFP. Neuron 28: 41–51.

21. Faust, N., F. Varas, L. M. Kelly, S. Heck, and T. Graf. 2000. Insertion of en-hanced green fluorescent protein into the lysozyme gene creates mice with greenfluorescent granulocytes and macrophages. Blood 96: 719–726.

22. Shakhar, G., R. L. Lindquist, D. Skokos, D. Dudziak, J. H. Huang,M. C. Nussenzweig, and M. L. Dustin. 2005. Stable T cell-dendritic cell inter-actions precede the development of both tolerance and immunity in vivo. Nat.Immunol. 6: 707–714.

23. Mulligan, S. J., and B. A. MacVicar. 2004. Calcium transients in astrocyte end-feet cause cerebrovascular constrictions. Nature 431: 195–199.

24. Hirase, H., L. Qian, P. Bartho, and G. Buzsaki. 2004. Calcium dynamics ofcortical astrocytic networks in vivo. PLoS Biol. 2: 494–499.

25. Tian, G. F., H. Azmi, T. Takano, Q. Xu, W. Peng, J. Lin, N. Oberheim, N. Lou,X. Wang, H. R. Zielke, J. Kang, and M. Nedergaard. 2005. An astrocytic basisof epilepsy. Nat. Med. 11: 973–981.

26. Lindquist, R. L., G. Shakhar, D. Dudziak, H. Wardemann, T. Eisenreich,M. L. Dustin, and M. C. Nussenzweig. 2004. Visualizing dendritic cell networksin vivo. Nat. Immunol. 5: 1243–1250.

27. Grutzendler, J., N. Kasthuri, and W. B. Gan. 2002. Long-term dendritic spinestability in the adult cortex. Nature 420: 812–816.

28. Petersen, M. A., and M. E. Dailey. 2004. Diverse microglial motility behaviorsduring clearance of dead cells in hippocampal slices. Glia 46: 195–206.

29. Hirrlinger, J., S. Hulsmann, and F. Kirchhoff. 2004. Astroglial processes showspontaneous motility at active synaptic terminals in situ. Eur. J. Neurosci. 20:2235–2239.

30. Verderio, C., and M. Matteoli. 2001. ATP mediates calcium signaling betweenastrocytes and microglial cells: modulation by IFN-". J. Immunol. 166:6383–6391.

31. Honda, S., Y. Sasaki, K. Ohsawa, Y. Imai, Y. Nakamura, K. Inoue, andS. Kohsaka. 2001. Extracellular ATP or ADP induce chemotaxis of culturedmicroglia through Gi/o-coupled P2Y receptors. J. Neurosci. 21: 1975–1982.

32. Schipke, C. G., C. Boucsein, C. Ohlemeyer, F. Kirchhoff, and H. Kettenmann.2002. Astrocyte Ca2! waves trigger responses in microglial cells in brain slices.FASEB J. 16: 255–257.

33. Beyer, E. C., and T. H. Steinberg. 1991. Evidence that the gap junction proteinconnexin-43 is the ATP-induced pore of mouse macrophages. J. Biol. Chem. 266:7971–7974.

34. Gallagher, C. J., and M. W. Salter. 2003. Differential properties of astrocytecalcium waves mediated by P2Y1 and P2Y2 receptors. J. Neurosci. 23:6728–6739.

35. Wang, Z., P. G. Haydon, and E. S. Yeung. 2000. Direct observation of calcium-independent intercellular ATP signaling in astrocytes. Anal. Chem. 72:2001–2007.

36. Wang, X., N. Lou, Q. Xu, G. F. Tian, W. G. Peng, X. Han, J. Kang, T. Takano,and M. Nedergaard. 2006. Astrocytic Ca2! signaling evoked by sensory stimu-lation in vivo. Nat. Neurosci. 9: 816–823.

37. Engelhardt, B., P. Vajkoczy, and M. Laschinger. 2003. Detection of endothelial/lymphocyte interaction in spinal cord microvasculature by intravital videomi-croscopy. Methods Mol. Med. 89: 83–93.

38. Matzinger, P. 2002. The danger model: a renewed sense of self. Science 296:301–305.

39. Wilkin, F., X. Duhant, C. Bruyns, N. Suarez-Huerta, J. M. Boeynaems, andB. Robaye. 2001. The P2Y11 receptor mediates the ATP-induced maturation ofhuman monocyte-derived dendritic cells. J. Immunol. 166: 7172–7177.

40. Clifford, E. E., K. A. Martin, P. Dalal, R. Thomas, and G. R. Dubyak. 1997.Stage-specific expression of P2Y receptors, ecto-apyrase, and ecto-5+-nucleoti-dase in myeloid leukocytes. Am. J. Physiol. 273: C973–C987.

41. Hall, A. 2005. Rho GTPases and the control of cell behaviour. Biochem. Soc.Trans. 33: 891–895.

42. Ling, C., M. Sandor, M. Suresh, and Z. Fabry. 2006. Traumatic injury and thepresence of antigen differentially contribute to T-cell recruitment in the CNS.J. Neurosci. 26: 731–741.

43. Kuroda, S., E. Yamada, and K. Kani. 1996. A new method to selectively injurethe optic nerve using argon-laser photocoagulation. Jpn. J. Ophthalmol. 40:344–355.

44. Carmichael, S. T. 2005. Rodent models of focal stroke: size, mechanism, andpurpose. NeuroRx. 2: 396–409.

45. Dailey, M. E., and M. Waite. 1999. Confocal imaging of microglial cell dynamicsin hippocampal slice cultures. Methods 18: 222–230.


Innate Response to Focal Necrotic Injury Inside the Blood ...

Documents

Transcript of Innate Response to Focal Necrotic Injury Inside the Blood ...