Keratan sulphate is a marker of differentiation of ramified microglia

ELSEMER

Developmental Bram Research 86 (1995) 233-241

Research report

Keratan sulphate is a marker of differentiation

of ramified microglia

Antonio Bertolotto ‘-*, Eisa Manzardo a, Marco Iudicello a, Renzo Guglielmone b, Alessandro Riccio ’

Accepted 31 January 1’795

Recently we reported that the keratan sulphate epitope recognised by the monoclonal antibody SD4 is expressed by a population of ramified microglia in adult rats. As ramified microglia is believed to differentiate from ameboii micr@a during postnatal development, we studied the rat blain from birth to Xl postnatal days of lie with the monoclonal antii 5D4. Contrary to all the other microglia markers until now described, keratan sulphate is not expressed by ameboii microglia and by macrophages but appears on the surface of microglia only when the cells are differentiated a%! show ramified pmcesses Tbe keratan sulphate positive cells become evident at different times in different central nervous system areas; the first were locali& in the pyriform cortex and brainstem horn the end of the second postnatal week. These observations suggest that keratan suiphate expression on microglia cells is induced by differentiation and by a restmg fimctional state. Moreover the 5D4 monoclonal antibody showed a strong diffuse posItwe staining of some cortical, thalamic and white matter areas during the first two postnatal weeks. This staining was transient and ii does not seem biologically correlated with the expression of the keratan sulphate on differentiated msroglia. Ideywords: Micro&a;

Macropbage;

Keratao sulphate; Multiple sclerosis; Proteoglycan; Development

1. IntmduetIoE Ramified, or resting, rnicroglia is a Central Nervous &stem (CNS) parenchyma cell population morphologically characterized by many sharply ramZied cell processes [13]. The origin, the antigen pattern and the functions of ramified microglia are controversial and not yet completely defined, but several data indicate that ramified microglia is a CNS resident component of the monocyte-macrophage system [14,23,27,29]. According to thii hypothesis, monocytes enter the CNS antenatally and become amcboid microglia, which differentiates in ramified microglia during the first 2 postnatal weeks [12,27,29]. This differentiation has been suggested, in the rat CNS, bv several immunohistochemical studies that detected the same antigens on the blood monocytes, ameboid microgiia, macrophages and ramified microglia: Common Leukocyte Antigen,

’ Cormsponding a.lhor. c/o Diisione Neu~ologica Univenita’ di Torino, Ospedale S. Laigi Gonaaga, Rrgioae Gonzole 10. I-10043 O~~IISSUIO, Italy. Fax (39,)+.!i) yO26.569 01653806/95/80950 (D 1995 Elsevier Sftence B.V. Ail nahts resewed SSDI 01653806(95)00030-5

,?Y -*.

i &

p2gj,

:$.+A&

Uirtwnmnat~%~ity complexI, _-____-r____

OX-18 (221, Major Histocompatibiity Complex II. OX-6 [9,22], type three complement receptor (CR3), OX-42 [7,21], CD4 [28], vault [71, Gtiffonia simplicifdia B4 isolectin [7,17], ED1 [7]. The differentiation in ramified microglia of the round ameboid microglia is characterised by a down-regulation of the monocyte-macrophage antigens above listed. On the contrary an up regulation or de l~vo expression of cell antigens character& the reaction of ramified microglia to diierent CNS lesion [I&34]. Besides ramified microglia, which is located in the CNS parenchyma, resident CNS macrophages have also been found in the vessel walls [U], in the meninges and in choroid plexuses [28]; moreover a peculiar parenchyma roicroglia population has been lc&iid around small CNS vessels [ill. Recently six monoclonal antibodies (MAbs) against different epitopes of the glycidic chain of keratan sttlphate proteoglycans (KS-PGs) were &ted by immunehistochemistry in the mature rat CNS and one of the cpitopcs, rccognised by the SD4 MAb, WISfound to be

234

A. Eermbtto et a.! /Dewlopmental Brain Ruean% 84 U995) 233-241

selectively expressed by a subset of ramified microgiia [S].ConsequentIy the anti-ES 5D4 MAb was proposed as a new immunohistochemicai microgiia marker KS-PCs seem to be also involved in inhibition of axonai migration; in fact they inhibit neurite growth in vitro [81,they have been immunolocalizedin CNS areas devoid of migrating axons during the embryonic iife [6,33] and between diierent functional units of the neonatal brain 1121. The aims of the present immunohistochemical study on postnatal rat CNS are to investigate the time of SD4 KS epitope appearance at the ceii surface of ramified micmgiia, its expression by macrophages and amehoid microgiia, and its iocahsation in specific CNS areas. 2. Materialsand methods

21.Rats Fisher rats aged 1 (Pl), 4 (P:), 8 @8], 11 (Pll), 15 (PlS], 21 (P211, 26 @26), 30 @30) and 90 @JO] days were anaesthetised with ether and the brain was immediately removed, fixed in Camoy’s fluid (60% ethanol, 30% chioroform, 10% acetic acid] at 4°C overnight, embedded in paraffin and cut 5 pm in the coronal plane. %o animais were utiiised at each time point.

2.2. Antibo&s ES was immunolccaiiied with the SD4 MAb that specificaiiy r ecognhes the sulphatp N-acetyllactosamine units present in both type I (CorneaI]and type II &Wiiage] KS chain [24].5D4 (kind gift of B. Caterson, University of North Caroiina at Chapel Hi& USA) was chosen among a panel of 6 anti-KS MAbs as it selectively stains ramified microgiia in adult rat brain [4]. ED2 MAb, anti rat macrophages, was purchased from Serotech (Oxford, England]. 2.3. Immunoeytochemicalprocedures The Avidin-Biotin Complex (ABC] (Dako, Denmark] was applied on deparaftlniid tissue sections. After inhibition of endogenous peroxidase with 3% hydrogen peroxide for 10 min, the sections were incubated with: (a) primary antibody, overnight at 4°C SD4 anti-KS MAb at 1:lOOOdilution; ED2 1:300, (b) biotin-labelled rabbit anti-mouse Ig (1:200) for 60 min at room temperature; (c) Avidm-Biotin complex for 45 mitt at room temperature. Between each step the sections were rinsed in PBS 0.01 M, pH 7.4 for 10 min. Foiiowhrg peroxidase development with 0.05% diaminobenxicime tetrahydrochIoride (DAB) and 0.01% hydrogen peroxide in 0.5 M Tris buffer pH 7.6 for 5-15

235

min. the sections were rinsed in distilled counterstained with hematoxylin.

water

and

2.4. Control sections and enzymadc digestion with keratanase Control sections were incubated either without primary antibodies or with SD4 dilute< I:750 in PBS containing 10,100 or 1000 pg/ml of type I and type II KS. For pre-digestion, sections were incubated for 2 h at 31°C with keratanase from Pseudomonas sp. (Seikagaku) at 0.05 W./ml in 100 mM Tris-HCI buffer, pH 8.00, containing 10 mM EDTA, 10 mM N-ethylmaleimide, 5 mM PMSF and 0.36 PM pepstatin A.

3. Results Control sections incubated without SD4 MAb or with SD4 MAb diluted 1:750 in PBS containing at least 100 ag/mi of KS type I and type II were negative. Keratanase digestion of tissue sections did not remove any kind of SD4 reactivity. This is not surprising es keratanase does not digest heavily sulphated KS oligosaccharide such as the SD4 epitope [24,26]. TIvo kinds of positive staining were observed on coronal sectkms of postnatal rat CNS Incubated with anti-KS SD4 MAb. The first, found in the cingulate and retrosplenial cortex and in the thalamus, was a transient staining of the white and grey matter that was already present on Pl and disappea:ed during the third postnatal week (Fig. 1, Table 1). The second pattern of staining was permanent, appeared from the end of the second postnatal week and demrated the surface of ramified microglial ceil, from now on called KS-positive (KS + 1 cells (Figs. 2 and 3, Table 1). These cells were identical to those found in mature CNS 141.

3.1. Neocorter On Pl the cingulate cortex and retrosplenial cortex showed diffuse positive staining of the neuropil throughout the cortical mantle; the external cortical layer showed the strongest staining that sharply contrasted with the negative meninge. The staining formed an intermingling network all around the cytoplasm of all the cortical cells. It deiimitated the cingulate and retrosplenial cortex as it decreased at the boundary with fronto-parietal and occipital cortex (Fig. i) where a faint and transitory staining was localiid in the external layer only. The white matter under the cingulate cortex was crossed by positive filaments joining the cd&al radiations to the cingulate cortex. The co&a! and white matter staining was transient, beginning to decrease in intensity during the second postnatal week and disappearing entirely during the fourth week. The first KS + cells were detected on P15 in the fronto-parietal and occipiiai cortex (Fig. ZC,D) and on P21 in the cingulate and retrosplenial cortex. At tbis age their number was exceptionally low, as only a few cells were scattered in every section; on P30 their number was already relevant, although not as high as in the adult animal. Even when the first KS+ cells were detectable they had a morphology identical to those found in mature rats showing long and branching cell processes. No stiting was observed in the cytoplasm of large cells resembling ameboid microglia or macrophages. Any cell with short and thick cell processes was observed, suggesting that the SD4 KS epitope was not detectable on the cell differentiating from ameboid mkroglia into ramified microglia. 3.2. &ifonn

cortex

No staining was detectable until Pll, when two or three positive cells were found in each coronal section.

Table 1 SD4 MAb immuaostainiagin posmatal rat CNS

Brainstem Basal nuclei

Pll PlS

Front+parietal-accipital Cinplate, r.%osplenial

Pll P15 P21

Extrmat corticallayer, Pl5 In all the cortical layen. F21

White maacr

cerebral ccrcbellar fiPpoeamPUS Thahmus Medii eminence

P15 PZl IV1 P21

In sane bundles only. P15 Alveus, PlS In soate nuclei, FZl Persistentin adult animal

236

Also in developingpyriformcortexthe KS + cells show ramitiedmorpbAogy from when they were detectable. On P15 the numberof KS + cells increased(Fig. 2A)

and they were located in all the corticallayers.On P30 their number was almost similar tc that of the adult animal(Fig. 2B).

A. lkrtoloIto cl al. /Lkuelwneiztal BrainResearch84 (1999 233-241

237

Fig. 3. Appearanceof KS + cells in the hippocampus.At Pll (A) no KS + cells arc evident;the section was counterstainedwith hematoaylinto show the hippocampalarchitecture.100x .At P21 few cells an found; the KS + cell shown in Fis B displays its raa~C;~d~~~rpholew.240X . At P90 Q the number of KS + cell is very high. 100Y

.

3.3. Tilalimlrls

lateral ventricles during the embryonic life and the first postnatal weeks. No kind of staining, cell associated or

Yn the thalamus strong staining was evident since Pl. Some thaknic nuclei showed a positive staining that was similar to that found in the cingulate cortex, furmIng a meshwork in the cellular space among neu-

ronal cytoplasms (Fig. 1). Tbe staining was strictly confined to some tbalamic nuclei, whereas other nearby nuclei were completely negative (Fig. 1). Moreover the boundary among some nuclei was c!car~/ outlined by au&KS immunopositivity (Fig. 1). Staining was strong until P15, then it decreased: disappearing in the fourth postnatal week ‘Fire iirsi KS f &is were fogund on P21. 3.4. M&n

eminence

observed in median emineuce was very peculiar because it was piSZ3lt both in tbe postnatal and in the adult brain and because it formed several long bundles that crossed all the median eminencein tbe sagittal plane from the III ventricle ependyma to the external surface.of the brain. T!te staining

3.5. white matter During postnatal development tbe white matter, in particular the corpus calloaum and the callosal radiations, contains numerous amebnid microglia cells that have been shown to differentiate into ramified mieroglia when anti-maerophage MAb or lectins were utilised !22,353_5!. On t&e ,-on*sw -_---, anti-KS 5D4 MAb revealed a different staining distribution. In fact the 5D4 MAb stained, during the first 2 pnatnatal weeks, long and thin filaments running parallei to the axons; the staining was clearly evident aa only same positive bundles were present in the corpus calloaum (Fig. 1); Cmbria and &ears. This faseicutar staining was transient, disappearing at the end of the seoond postnatal week In the corpus calloaum tbe first KS + cella were found on B15; in the cerebellar wbite matter they appeared on P22. They were identical to the KS + celIs present in adult rata, snowing thin and delicate pmcesses both parallel and perpendicular to tbe axonal bundles. During postnatal development anti-KS MAb SD4 never immunolocalixed macrophages or ameboid microglia in white matter or in any other CNS ~giOflS.

3.6. Nwoepitklial

hzye~.pia mater and

CNS verse,%

The neuroepitheiial layer, the CNS vessels and tbe pia mater have been considered areas containing microghal precursors [23]. Neurons and glial cells originate from the neuroepitbe!&! !ayer- a transient CNS structure bordering the

diffuse, was detectable with SD4 MAb in this CNS area. During the first postnatai month no KS + cells were

found in the pia mater, around small intracerebral vessels or in the wall of larger vessels. On the contrary, round or slightly elongated cells, without cell processes elongating in the CNS parenchyma, were localized in the pia mater and in the wali of small intracerebral vessels on serial sections stained with the anti-macrophage ED2 MAb (not shcwn).

On P!Xlvery few KS f celis were immunolocalized in the pia mater and choroid plexuses; on the contrary several macrophages were detected in serial sections incubated with ED2 MM. The exceptional pre ;ence of K8 + cells in the meuinges and choroid plexuses was also observed in other adult rats utilised in previous studies in our laboratory. 3.7 Hippocampus During the first two postnatal weeks diffuse fascicular staining was present in the alveus. This positive staining, evident until Pll, disappeared on P15. With the exception of the transient stainiig of the alveus, the hippocampus did not show any positive staining ~?l?ti!P27-1when few XS + ce!!a were detected (Fig. 3A,B). At the end of the first month their number had greatly increased, almost reaching the level found in adulthood (Fig. 30. 3.8. l3nzin§tem During the first postnatal week the positive staining was restricted to some white matter bundles; no KS + elements were evident. They appeared on Pll, their number increased slowly and reached the number found in the adult on P30.

4. Discussion The KS epitope recognised by the 5D4 hb was found speeiScaUy associated with the cell surface of ramified microglia in the normal CNS of the adult rat and it was proposed as a new marker for ramified microdia [41. The present study was focused on the time of appearance of the i S apitope in ramified microglia, a cell population that differentiates during postnatal rat CNS development [23,27].The progenitor of ramified microglia is not definitively established, A neuroectoderm-,d origin has been mtulated [18,32], although a relevant body of evidence supports the bone marrow origin 1231.According to this latter lqpt%hesis

enter ihe CNS during development and have a different fate: some of them become

mOttocytt%

mucrophages, some others ameboid microglia. This round cell differentiates during the first posmatal weeks into ramified (resting microgha, extending long and branching cell processes. ‘f’hii differentiation ‘uas fol-

lowed, step by step, using carbon labelliig methods [19], histochemical detection of enzymes [20] and, bttmunohistochemically, with lectins and MAbs detecting antigens expressed by the monocyte-macrophage imeage: Common Leukocyte Antigen, 8X-l [22J8], Major Histocompatibility Complex I, OX-18 1221, Major Histocompatibjlity Complex II, OX-6 [22], type three

complement receptor (CR3! OX-42 [7,21], CD4 [28], vault [fl, Gtifinio simpkcifdiaB4 isolectin[7,17],ED1

ance of the KS epitope that ramified microgUa resting functional state l-z.0 ,.i I.‘_ “*%“-“,.~ ““0WII.,W “1 and

mscrophages

peculiar aspect of the hypothesis th croglial differentiation In fact both ameboid active cells that can phagocyte cell debris with au activated state [;Llzl It lished if KS epitope is lost when of the adult anhnal is activated

[7] and unidentified antigens [16]. The trarrformation of ameboid microglia into ramified microglia is characterised by a’ down-regulation of all the antigens and enzymes above listed [U].

The KS epitope recognised by 5D4 MAb showed several aspects completely different from Ae other anti-microglia MAbs until now described: (i) in postnatal development it appears on the cell surface when the microglia is ramified; (ii) its expression is preserved in the adult animal and it is not down-regulated, (iii) the KS epitope is not expressed by brain macrophages or ameboid microglia during the postnatal development.

The KS + ceUs appeared at different times in different CNS a-?&s,according to the degree of mrturation of the nervous structures. The appearance of KS + cells was identical in all the CNS regions: every time Kg + cells were detected they showed long and branching ceil processes, identical to the KS positive cells found in the adult animal. In every CNS area very

few cells were at first found, then, in the following 10 days, their number quickly increased. In adulthood KS + are present in every CNS area, although some regions contain a high number of positive cells and others very few cells [4]. A comparison of the number of cells detected with the microglia marker Grifonia simp!icijMa B4 isolectin indicates that 5D4 MAb im-

munoIocaIiied fewer cells than the lectin. suggesting that KS + cells are a subset, albeit relevant, of microglia [4]. Our dti.a i.ldicate that the KS epitope is expressed by ramified i.e. it has surroundmg

microglia when the cell is differentiated, branching cell processes and when the nervous system has reached a considerable

degree of maturation. The expression of KS seems to be triggered by the resting sta:e of microgUa, a cell population that participates actively in the CNS development [23,27]. The transformation of amebcld microglia into ramified microglia has been considered a regressive phenomenon, as the activity of some enzymes decreases and the expression of several antigens is down-regulated [18,23]. On the contrary the appear-

antigen -epertoire and, likely, functhm. The function and the detailed biochemical stmctwe

Ioca’icn.

of the mieroglial KSPG beariug unknown. Ilo-wever, considering

the 5W KS epitope is that it is located out-

side the cell membrane [41, it is a member of the proteoglycan family and it appears during the first postnatal month, the microgUa1I&PG could interact with the other extracellular matrix components recently found in the CNS that also are pmduccd by CNS cells duriug the flit postnatal month D&5,2525]. The transient staining of some white and grey matter areas seems to be a nettrobiological phenomenon not correlated with the differentiation of ramified mi-

croglia, as KS + cells appear throughout the CNS, including areas where the transient stahring was not found. 5D4 KS epitope could be present in the glyeidie chain of two or more different brain KS-PGsz the first

involved ic microglia differentiation and the second in a diierent neurobiological event happening during CNS development. This hypothesis is strongIy supported by the detection of 5D4 KS epitope in two different KS-PGs extracted from the brain [8,30] Moreover, the presence of a large family of K!l-PCs in the CNS is also indicated by the isolation of a third KS-l% I121 and by the imm~olocaliaation of five dierent KS epitope

in the braa

gional distribution [4] It has KS-PGs are implicated in axonal guidance outgrowth inhibition [8,12,26,31,33].This su been inferred by their localiition in the middle line areas or in the boundaries of embryonic and neonatal UV.3 and by their ability to inhibit nemite outgrowth in culiure [6,8,12,26,31,33]. the transient staining observed

with 5D4 MAb could be in lie with this hypothesis as

240

A. lkr?dot:o et al /Lkt&jmentalBrobi

it is strictly localised to specific developmentally regulated.

and is

We would like to thank Bruce Caterson for the generous gift of anti-KS MAbs. This work was supported by the Italian Association for Multiple Sclerosis OL1.S.M.).

R.B., Gebmumn, J., Scbubcrt,P. and Krcutrber~ G.W., Qtotoaicity of micrqlia, Oh, 7 (1993) 111-118. [Zl Benototto,A, Rocu, G. nod Scbfffer,D., Cboodmitin 4-sulfate pmtm&ao forms an extraccllolar network in human and rat ceotml ?crvxms ? system, J. Naud. SC&100 (1990) 113-123. 131knototto, A., Rocca, G., Caomc, G., Mi&li, A. sod Scbiffer, D., Cboodmitin wlfatt protco&cao swrooods a neomoal subset to bumao and rat CNS, I. N-i. Rec., 29 (1991) ??5-234. 141Beriofotto, A., Caterson, B., Canavese, G., Mi&cli, A. and Scbfffer, D., Moooclooal aotii to keratan solfate immmtobca& ramifkd micro& to mn and cryostat see tions of rat bmfo, I. Hhwhenx Cyto&m, 41(1993~481-4S7. [Sl Bhmi, A., Asber, R sod Peridea, G., Co-localiitioo of lvaluronic acid and cbondmitin sulfate pmtwgiycan in rat cerebral cortex, Brain R&q 579 (1992) 173-177. I61 Burl, MA. awl C&, GJ., Chatrio, ao aotiadbcsive o+xral keratao ndfate pm&o&m, is strocturaUyrelated to MAPIB, J. NeumfM., 2.50994) l-22. M Cbu~aoi,CD., hksra, N.L, lad Rome, XL, Vault im.ooocv fbmmsceoca in the braioz ocw i&h regarding tbe o&h of I&O&, 3. Ncrtmcf.,ll(l990 256-268. ISI Cok, GJ. axl McCabe, C.F., Identification of a dwetopment&y regulated kemtan mdfate pmteo&ao tbat inhibits celt adbesii sod oeorite oottuowtb, Ncrvon, 7 (1991) 1007-1018. 191Cnso, R.I. sod Webster, H. de F., la mtigcos in tbc oomrat rat oervoos system and in leaioo of expcrimcntat alter&c encepbalomyelitis,Acta Nawpathd., 60 (1985)X3-272. 1101Flub ?%A., Dcosmore, T.L. h+olkston, MC. sod Hi&y, W.F., Cbaracterizath of anic@ia sod macmpba;ed to tbe Central Nervoos System of rats: definition of the differential ewesion of molec~ks usins staodard sod xxwel mooockmal a&odii in oonttaf CNS and in four makh of parmtcbymal rwtioo, Glia, 7 (1993)34-U). [ill Gebrmano, J. sod Rreutzbw, G.W., Cbaracterisatiooof two acw monocboal antibodies directed sgsiast rst microglia, J. Canp. Neutul., 313 (1991)409-430. I121Geisert, E.E. sod Bidansct, DJ., A central neww system keratan sulfate pmtco&xn - locahtion to bouodaries in the neonatal rat braio, Dcu. Bmh Rrs., 75 (1993) 163-173. [131Graebcr, MB. and St&, W.G., Mia: immune nehworkin CNS, Emh P&of, 1 WLNJ)2-s. (141Hicke~, W.F. and Kimma, H.. Perivascularmicro&al cells of the CXS arc bone marrow-derivedand present antipen io viva, scimw, 239 wS8) 290-292. [I] Bard,

Pbawch 86 (1995) 233-241

tron microscopic autoradiogaphic study usmg 125~lododeoxywtdine,1. Netmpah/. Exp. Neud.. 45 (1986) 1-19. 1331 Snow, D.M., Steindler, D.A. and Silver, 3.. Molecular and cA!iilar characterization of the glia roof plate of rbe spinal cc:2 and optic tectum: a poaible role for a proteogiycan m the d+vetopmcntof an axon barrier, L&w.Bid. I38 (1990) 359-376. [34] Streit, W.J. and Greaber. M B.. Heterogenify of micro&d and

penvascular cell populatiow insights gained fmm the facial nucleusparzdlgm, G&l, 7 (1993~6%74. :351 Wu. C H . Wen. C Y., Stdeb. J.Y. and L.& Eu.4. A quantitative and morp5ometr.c shtdv of the uansformatioo of amaboid mnoglia Into ramified micro& in the developingcorpuscatlo. sum m rats, J Amt., 181 (1992) 423-430.