Journal of Neurocytology 21, 623-634 (1992)
MHC-positive, ramified macrophages in the normal and injured rat peripheral nervous system S. M O N A C O * ,
J. G E H R M A N N ,
G. R A I V I C H a n d G. W. KREUTZBERG;
Department of Neuromorphotogy, Max-Planck-Institute for Psychiatry, Am Ktopferspitz 18A, D-8033 Martinsried, Germany Received 14 January 1992; revised 2 April 1992; accepted 14 April 1992
Summary Resident endoneurial macrophages form a prominent, but little recognized component of the PNS. We have studied immunocytochemicallythe distribution, morphology and immunophenotype of endoneurial macrophages in several normal peripheral nerves of the rat. In addition, we investigated the macrophage response following crush injury of the sciatic nerve. Resident endoneurial macrophages had a ramified morphology with processes oriented parallel to the long axis of nerve fibres. They were positive for several monocyte/macrophage markers such as ED1, ED2 and the recently-described MUC 101 and MUC 102 antibodies. They furthermore expressed the complement type three receptor, the CD4 antigen and MHC class I and II molecules. These results were consistent in all the peripheral nerves studied. In addition, 1000rad of y-irradia tion led to a strong reduction in the number of MHC class II-positive ramified cells in the peripheral nerves similar to that observed in other peripheral organs such as the heart. A considerable percentage of resident rnacrophages in the PNS and/or their precursor cells are therefore radiosensitive and could be related to the lineage of dendritic cells. Following crush injury, ED1-3-, OX-42-, MUC 101- and MUC 102-positive round macrophages were observed from 24 h postlesion onward at the site of trauma. In the distal part, they were observed to form strings of round, foamy macrophages probably involved in myelin phagocytosis. In contrast, the number of MHC class II-positive resident macrophages was only slightly increased at the site of trauma and in the distal part. These cells transformed from a ramified to a round morphology, but did not appear as typical strings of foamy macrophages. These results demonstrate that the PNS is provided with a resident macrophage population analogous in many respects to microglial cells in the CNS. These constitutively MHC class II-positive PNS microglial-like cells could act as the major antigen-presenting cells in the peripheral nerve. They may thus constitute a local immune defense system of the PNS with a function similar to that of microglial cells in the CNS.
Introduction In the mammalian CNS microglial cells are antigenically and functionally related to cells of the monocyte/ macrophage lineage; they thus represent the resident, intrinsic macrophage of the brain (Streit et al., 1988; Perry & Gordon,. 1988; Graeber & Streit, 1990). They comprise between 5% and 20% of the total glial cell population (Lawson et aI., 1990; Peters et al., 1991). Under normal conditions they have a typical, ramified morphology and are rapidly activated in response to even subtle pathological stimuli (Streit et al., 1988; Graeber & Streit, 1990; Gehrmann et al., 1991a,b). There is evidence that the PNS is also provided with resident macrophages that share some properties with other tissue macrophages (Hughes et aI., 1987; Perry & Gordon, 1987; Stevens et aI., 1989; Vass & Lassmann, 1990) and CNS microglial cells (Gehrmann et al.,
1991a,b). They comprise 1% and 4% respectively of the total endoneurial cell population in the normal peripheral nerve (Oldfors, 1980; Schubert & Friede, 1981). However, the immunophenotype, morphology and functional properties of these cells in the PNS await a more precise characterization. After peripheral nerves are crushed macrophages rapidly invade the site of trauma, but also the distal part of the nerve which undergoes Wallerian degeneration (Olsson & Sj6strand, 1969; Perry & Gordon, 1987; Stoll et al., 1989). The presence of activated macrophages in the injured peripheral nerve is functionally important. They phagocytose the myelin debris (Beuche & Friede, 1984, 1986; Stolt et aL, 1989), participate in the production of mitogenic factors for Schwann cells (Baichwal et al., 1988) and fibroblasts
* On leaveof absence fromthe Instituteof Neurology,Universityof Verona, Italy. To whom correspondenceshouldbe addressed. 0300-4864/92 $03.00 +.12 9 1992 Chapman and Hail Ltd
624 (Leibovich & Ross, 1976; Martinet et al., 1986; Raivich et al., 1990) a n d induce the synthesis of e n d o n e u r i a l n e r v e g r o w t h factor b y secreting interleukin-1 (Lindh o l m et al., 1987, 1988). The origin of m a c r o p h a g e s has b e e n in turn a s s i g n e d to h a e m a t o g e n o u s (Olsson & Sj6strand, 1969; Beuche & Friede, 1984) or e n d o g e n ous cells only (Berner et al., 1973). To characterize r e s i d e n t m a c r o p h a g e s in the rat PNS further as a n a l o g o u s to microglial cells in the CNS, e n d o n e u r i a l , resident m a c r o p h a g e s w e r e studied imm u n o c y t o c h e m i c a l l y u n d e r n o r m a l conditions a n d d u r i n g Wallerian d e g e n e r a t i o n . Several m o n o c l o n a l antibodies w e r e u s e d w h i c h recognize m o n o c y t e / m a c r o p h a g e m a r k e r s s u c h as the c o m p l e m e n t t y p e three r e c e p t o r a n d m a j o r histocompatibility c o m p l e x (MHC) antigens.
Materials and methods
Surgery and tissue collection Adult male Wistar rats weighing 200-250 g were used for all the experiments. For immunocytochemical investigation of resident macrophages in normal animals, three untreated, healthy rats were used. After decapitation under deep aether anaesthesia L5 motor and sensory roots, L5 dorsal root ganglia (DRG) and the sciatic, tibial, facial and vagal nerves were rapidly dissected, embedded in O.C.T. II tissue compound (Miles, Elkhart, Minnesota, USA) and snapfrozen in liquid nitrogen. For studying the macrophage reaction following Wallerian degeneration, the right sciatic nerve was crushed under aether anaesthesia in the gluteal region for 30 s with a jeweller's forceps. Three rats per time point were killed at 0, 1, 2, 4, 7, 14, 21 and 42 days after the operation. The sciatic nerve was removed for a length of about 30 mm, encompassing a 5-10 m m segment proximal to the crush site and the distal part, embedded in O. C.T. Compound (Miles, Elkhart, USA) and snap-frozen in liquid nitrogen. In addition, six 8-week-old male Wistar rats, anaesthetized with chloral hydrate, were irradiated by exposure to 1000 rad (given over i min) from a gamma-irradiation source (Stabilon, Siemens, FRG) according to the protocol given by Hart and Fabre (1981). Six naive littermates were used as controls. Ten days after irradiation the irradiated and control rats were killed; L5 motor and sensory roots, L5 DRG, the sciatic nerves, the tibial nerves and the heart were removed and rapidly frozen. Longitudinal cryostat sections (10 b~m)of the above specimens were serially cut and collected on gelatine-coated glass slides and stored at - 7 0 ~ until immunocytochemical examination.
Immunocytochemistry Cryostat sections were fixed in 3.7% formalin for 5 min followed by acetone (50% 2 min, 100% 2 rain and 50% 2 min) at room temperature. The sections were then rehydrated in 0.01 M phosphate-buffered saline (PBS) (pH 7.4) for 5 rain and 0.01M PBS containing 0.1% bovine serum albumin. After blocking with 2% normal horse serum (DAKO, Hamburg, FRG), adjacent sections were incubated over-
M O N A C O , G E H R M A N N , RAIVICH and KREUTZBERG night at 4 ~C with the following monoclonal antibodies for cell typing (working dilutions in 0.01 M PBS are given in parentheses): (1) MRC OX-42 (Serotec, UK), recognizing the rat complement receptor type three (Robinson et al., 1986) (1:6400); (2) MUC 101, recognizing two proteins of 116 and 95 kDa on rat microglia (Gehrmann & Kreutzberg, 1991) (1 : 1600); (3) MUC 102, recognizing two proteins of 62 and 70 kDa on rat microglia (Gehrmann & Kreutzberg, 1991) (1 : 1000); (4) MRC OX-6 (Serotec, UK) directed against MHC class II antigens (McMaster & Williams, 1979) (1:6400); (5) MRC OX-18 (Serotec, UK) directed against MHC Class I antigens; (6) the W3/25 antibody, recognizing the rat CD4 antigen (Perry & Gordon, 1987) (1 : 6400); (7) ED1, ED2, ED3 (Serotec, UK) recognizing monocytes/macrophages (Dijkstra et al., 1985) (1:6400). Subsequent antibody detection was carried out using a biotinylated horse anti-mouse secondary antibody (Serotec, UK) and the Vectastain ABC-Elite kit (Vector Labs, Burlingame, CA, USA) with 3,3'-diaminobenzidine as peroxidase substrate following a previously described protocol (Gehrmann et al., 1991a). Immunocytochemical controls included omission of the primary antibody or replacement of the primary antibody by an irrelevant antibody of the same IgG subclass. The number of OX-6- and EDl-positive macrophages was evaluated quantitatively in ten randomly chosen areas in the sciatic nerve of irradiated and control rats. Immunostained cells were counted twice in fields of 0.7 m m 2 at x 10 primary magnification. The values obtained were uncorrected cell counts (means + SEM) and were evaluated for statistical significance using Student's t-test.
Results THE NORMAL PERIPHERAL NERVE In all the p e r i p h e r a l n e r v e s investigated, e n d o n e u r i a l cells w e r e strongly positive for M H C class II antigens, r e c o g n i z e d b y t h e OX-6 a n t i b o d y (Fig. 1). T h e y w e r e f u r t h e r m o r e positive for M H C class I antigens, recognized b y the OX-18 antibody, a n d w e r e stained b y several m o n o c l o n a l antibodies such as ED2, M U C 101, M U C 102 a n d OX-42 w h i c h recognize different m o n o c y t e / m a c r o p h a g e antigens (Fig. 1). A few e n d o n e u r i a l cells preferentially a r o u n d b l o o d vessels w e r e labelled b y ED1 (Fig. 1). Endoneurial, ramified cells further e x p r e s s e d the CD4 antigen, r e c o g n i z e d b y the W3/25 a n t i b o d y (Fig. 1). In contrast, ED3 i m m u n o r e a c t i v i t y Table 1. Quantitative evaluation of immunostained macrophages in the sciatic nerve of control and irradiated rats.
Marker
Control
Irradiation
OX-6 ED1
57.07 + 0.74 51.94 + 5.05
7.69 + 0.66* 51.77 _+ 4.04
The mean values are given as the number of immunostained cells per mm2tissue section + SEMfrom 5 animals (n = 5). * irradiation versus control p ~ 0.001
Resident m a c r o p h a g e s in the PNS
625
Fig. 1. Stai!ning of resident, ramified macrophages with several monoclonal antibodies in the normal periphera~ nerve. The respective antibodies are given in the figure. (a,c-f) Sciatic nerve. (b) Vagus nerve, x 760.
626
MONACO, GEHRMANN, RAIVICH and KREUTZBERG
Fig. 2. ~-Irradiation experiment. (a,c,e) Control rats. (b,d,f) Irradiated rats. (a,b) Heart, OX-6 immunoreactivity (indicating the presence of MHC class II-antigens). MHC class II-positive cells are almost completely depleted in the heart following irradiation. (c,d) Sciatic nerve, ED2 immunoreactivity. (e,f) Sciatic nerve, OX-6 immunoreactivity. The number of ED2-positive macrophages remains unchanged following irradiation (d) whereas the number of MHC class II-positive, ramified cells is strongly reduced in the irradiated animals (f) similar to the heart (b). x 200.
was not observed on these endoneurial cells. The i m m u n o p h e n o t y p e of resident macrophages is summarized in Table 2. I m m u n o s t a i n e d cells had an elongate cell b o d y from which crenellated, uni-or bipolar processes arose. These processes were longitudinally oriented along the axis of nerve fibres (Fig. 1). The m o r p h o l o g y as well as the i m m u n o p h e n o t y p e of the labelled cells was consistent in all the peripheral a n d cranial nerves examined. In contrast to resident macrophages, Schwann cells and endothelial cells were not observed to express M H C antigens either
u n d e r normal or pathological conditions (for comparison see below). X-RAY IRRADIATION Irradition (1000 rad) induced a strong reduction in the n u m b e r of M H C class II-positive, dendritic cells from the heart as previously reported (Hart & Fabre, 1981) (Fig. 2a,b) whereas the n u m b e r of ED2-positive cells remained unchanged. The n u m b e r of MHC class II-positive cells was also strongly reduced in the peripheral nerves following irradiation. Quantitative
Resident macrophages in the PNS
627
Fig. 3. MUC 101 (left) and OX-6 (right) immunoreactivities of adjacent longitudinal sections of normal (0 days) and crushed (1--42 days) sciatic nerves. The site of crush is in the middle of each micrograph, x 18. Note the rapid and strong increase in MUC 101 immunoreactivity at the site of crush compared to the relatively low levels of OX-6 immunoreactivity, indicating the presence of MHC class II molecules.
evaluation showed that almost 90% of the MHC class II-positive cells (p < 0.001) were depleted in the sciatic nerve following y-irradiation (Table 1). Similar to the situation in heart, the number of ED2-positive cells in the irradiated sciatic nerve remained unchanged (Table 1) although they appeared to be slightly hypertrophic. The expression of other cellular markers such as the CD4 antigen, the MUC 101, MUC 102 determinants and the complement type three receptor was not affected by the irradiation.
WALLERIAN DEGENERATION
The site of crush Numerous MUC 101-positive cells were observed at the site of the crush, I day postlesion (Fig. 3, left side). They had a round morphology and were mainly observed around endoneurial blood vessels and along the injured perineurium. These cells were also labelled by the MUC 102, ED1 and OX42 antibodies (data not shown). By day 2 through day 4 postlesion the number
628 and staining intensity of MUC 101, MUC 102, OX-42 and EDl-positive cells further increased at the site of crush. They were then also stained by the ED3 antibody which otherwise did not stain macrophages in the peripheral nerve (data not shown). From day 4 onward, immunostained cells were also observed in the adjacent segments proximal and distal to the site of crush and were arranged in strings of rounded cells. Staining of these strings of macrophages reached a maximum around two weeks postlesion. After six weeks the staining intensity was only slightly increased as compared to control nerves. Although some of the labelled cells still displayed a round, stout morphology after six weeks, most cells again displayed a slender, ramified morphology. In contrast, the number of MHC class II-positive resident macrophages was decreased at the site of crush at 1 day postlesion (Fig. 3, right side). From 2 days postlesion onward, the number and staining intensity of MHC class II-positive macrophages was slightly increased at the site of crush including the closely adjacent proximal and distal segments. This reaction reached a maximum around 4 days postlesion. At this time point MHC class II-positive cells further transformed from the ramified to a round morphology. Strings of MHC class II-positive macrophages, however, similar to the distribution of, for example, MUC 101- or OX-42-positive cells, were not observed. From two weeks onward, the number of MHC class II-expressing cells had returned to control levels. These cells again displayed a ramified morphology. The distal part A first noticeable increase in the number of MUC 101-positive macrophages was observed in the distal part at 4 days postlesion (Fig. 4, left side). These cells were also stained by the MUC 102, OX-42, ED1 and ED2 antibodies. They transformed from their slender, ramified morphology to a round morphology resembling phagocytic cells. Seven days postlesion, many transition forms between ramified and phagocytic macrophages were seen. In addition, small, round cells most likely representing invading monocytes were-occasionally intermingled with these large, foamy macrophages. From two weeks onward, the immunostained foamy macrophages were arranged in strings of rounded cells. This string-like arrangement was most evident six weeks after the crush injury, the latest time point studied (Fig. 4, left side).
MONACO, GEHRMANN, RAIVICH and KREUTZBERG In contrast, the number of MHC class II-positive cells was only slightly increased in the distal part at 4 days postlesion (Fig. 4, right side). Some of these MHC class II-positive cells had a round morphology. However, MHC class II immunoreactivity was not observed on the strings of large, foamy macrophages as revealed for example by MUC 101 immunoreactivity (cf Fig. 4, left side). Forty-two days postlesion, MHC class II-positive cells showed different morphologies depending on their localization in the injured sciatic nerve (Fig. 5a). In the proximal part, the number of these cells was increased. They had a ramified morphology, but were hypertrophic (Fig. 5b). In a segment 1.5 cm distal from the site of crush, MHC class II-positive cells appeared as ramified cells with a density comparable to that in control nerves (Fig. 5c). In a segment more distal than 1.5 cm from the site of crush, the number of MHC class II-positive cells was not increased, but they had a round morphology with few, short processes (Fig. 5d). Discussion
This study shows that, in the normal rat PNS, resident macrophages express several monocyte/macrophages antigens similar to other tissue macrophages and CNS microglial cells (Dijkstra et al., 1985; Streit et al., 1988; Perry & Gordon, 1988). This extends previous observations of ED2- or W3/25-positive spindle shaped cells in the rat peripheral nerve (Hughes et aI., 1987; Stevens et al., 1989) and F4/80-positive cells in the mouse PNS (Perry et al., 1990). Their ramified morphology is reminiscent of that of resting microglial cells in the CNS, particularly in the white matter (Streit et al., 1988; Lawson et al., 1990; Graeber & Streit, 1990). Resident macrophages in the normal rat PNS Resident macrophages in the PNS expressed several monocyte/macrophage markers such as the complement type three receptor and the ED1 and ED2 determinants which are characteristic of cells of the monocyte/macrophage lineage (Dijkstra et aI., 1985). ED3, however, was not found on resident macrophages in the PNS. Resident macrophages were furthermore labelled by the monoclonal antibodies MUC 101 and 102 which recognize rat microglial cells in the CNS and other mononuclear cells in the peripheral organs (Gehrmann & Kreutzberg, 1991). Morover, they showed a strong, constitutive MHC class I and II expression and were positive for the CD4
Fig. 4. MUC 101 (left) and OX-6 (right) immunoreactivities in adjacent longitudinal sections of the distal part of normal (0 days) and crushed (4M2 days) sciatic nerves. The micrographs were taken from segments about 1.5 cm distal from the site of crush, x 200. MUC 101-positive cells form strings of round macrophages, most prominently 6 weeks postlesion, whereas the number and morphology of OX-6-positive cells undergoes only relatively mild and transient changes.
631
Resident macrophages in the PNS Table 2. Immunocytochemical staining characteristics of resting microglial cells in the CNS and resident macrophages in the PNS of the rat.
Antibody
Structure/cell recognized
CNS microglia
PNS macrophage
OX-42 ED1 ED2 ED3 MUC 101
CR3 receptor Macrophage antigen Macrophage antigen Macrophage antigen Microglial and perivascular cells, 116 and 95 kDa antigen Microglial cells, 62 and 70 kDa antigen CD4 antigen MHC class II antigen (Ia) MHC class I antigen
++ -
++ + ++ -
_+
++
++ _+ • •
++ ++ ++ §+
MUC102 W3/25 OX-6 OX-18
no staining. • only a small number of cells stained. + welldiscernable. ++ strong staining. -
antigen. In the PNS, their i m m u n o p h e n o t y p e thus resembles closely that of satellite cells in the dorsal root ganglion ( G e h r m a n n et al., 1991a). In the CNS, however, the i m m u n o p h e n o t y p e of resident PNS macrophages differs to some extent from t h a t of resting microgtial cells (for comparison see Table 2). Resting microglial cells are also positive for the complement type three receptor, the MUC 101- and MUC 102-determinants, but express the ED determinants, the CD4 antigen and M H C antigens at m u c h lower levels (Streit et al., 1988; Perry & Gordon, 1987; Graeber et al., 1989; Vass & Lassmann, 1990; G e h r m a n n et aI., 1991a). The constitutive M H C class II antigen expression by resident macrophages in t'he PNS is in contrast to the results of previous studies which described only a few M H C class II-positive cells either in the e p i n e u r i u m (Craggs & Webster, 1985) or preferentially in a perivascular position in the e n d o n e u r i u m of peripheral nerves and spinal nerve roots (Lassmann et al., 1986; Vass & Lassmann, 1990). A recent preliminary report, however, confirms the high constitutive M H C class II expression by resident macrophages t h r o u g h o u t the PNS of different species (Griffin et al., 1991). M H C class II products (Ia) are antigen-binding glycoproteins which are expressed on the surface of accessory cells presenting antigen to T-lymphocytes. Accessory ceils can be divided into two groups: (a)
professional accessory cells such as dendritic cells which constitutively express high levels of Ia-antigen, and (2) facultative accessory cells such as peripheral macrophages and microglial cells which constitutively express only low levels of Ia-antigens (for review see Klinkert, 1990). The constitutive high level of Iaantigen expression on resident macrophages in the PNS as well as their ramified m o r p h o l o g y suggest a possible relationship to dendritic cells. M H C class II-positive, ramified macrophages were almost completely depleted from the PNS following 1000 rad of y-irradiation, which has been s h o w n to deplete effectively other tissues from dendritic cells due to their low residency time (Hart & Fabre, 1981). The results of the present s t u d y indicate that a high percentage of M H C class II-positive macrophages and/or their precursors are radiosensitive. In contrast, the n u m b e r of ED1positive macrophages remained u n c h a n g e d following irradiation, which w o u l d suggest a heterogeneity of PNS macrophages in terms of lineage and functional properties. In spite of their morphological, antigenic and functional similarities, however, it is still debatable w h e t h e r CNS microglial cells, PNS resident macrophages and l y m p h o i d dendritic cells might in fact belong to the same cellular entity or represent distinct cell lineages. Although the irradiation experim e n t offers only a single line of evidence, the results of this experiment could indicate an analogy to the data
OX-6 immunoreactivity indicating the presence of MHC class II antigens in a longitudinal section of a sciatic nerve 42 days after crush injury. (a) Overview. The arrows indicate the site of crush. The distal stump is below the arrows, x 12. (b) Proximal part. The number of MHC class II-positive macrophages is increased as compared to the distal part (cf. (c,d)), but they have a ramified morphology. (c) Distal part, segment 1.5 cm distal from the site of crush. MHC class-lI positive cells have a density and morphology similar to the normal nerves (for comparison see Fig. 1). (d) Distal part, segment more than 1.5 cm distal from the site of crush. The number of stained cells is not increased, but they have a round, stout morphology. (b-d) • 760. F i g . S.
632 obtained by Hart and Fabre (1981) for several peripheral organs suggesting that at least a part of the resident macrophages in the PNS may be related to the lineage of dendritic cells. MHC class II-positive, resident macrophages would seem to represent a local immune defense system in the PNS analogous to microglial cells in the CNS (Streit et at., 1988; Graeber & Streit, 1990). The nature of antigen-presenting cells (APC) in the PNS, however, is still controversial. In addition to macrophages, Schwann cells have been shown in vitro to act as APCs (Wekerle et al., 1986), while their rote in vivo as APCs is debatable. Under pathological conditions, in particular in experimental autoimmune neuritis, macrophages but not Schwann cells express Ia molecules (Schmidt et aI., 1990). Based on these in vivo observations, resident macrophages would seem to qualify as the main intrinsic APCs of the PNS. Macrophages after WalIerian degeneration As demonstrated by Olsson and Sj6strand (1969), haematogenous macrophages rapidly invade the injured sciatic nerve mainly at the site of trauma. Activated macrophages then serve a variety of functions which are related to the phagocytosis of the degenerating myelin, stimulation of Schwann cell proliferation and possibly the ensuing nemite outgrowth (Perry et aI., 1987; Baichiwal et aI., 1988; Raivich et aI., 1990, 1991). Recent studies using a mouse mutant, C57BL/OIa, in which Wallerian degeneration is slow, purport to show a major role of haematogenous macrophages in the phagocytosis of the myelin sheath during rapid Wallerian degeneration in the PNS (Lunn et aI., 1989). In addition, the recruitment of haematogenous macrophages can be prevented by antibodies against the complement receptor type three (Lunn et al., 1989). In contrast, the involvement of resident PNS macrophages in Wallerian degeneration is still debated. In vitro, marked myelin phagocytosis occurs in nerves kept in organ cultures without added monocytes (Hann-Bonnekoh et al., 1989), suggesting that resident macrophages are the main source of mononuclear phagocytes under these conditions. Using a bone marrow chimera approach, Perry and colleagues (1990) have discussed the possibility that these resident macrophages could contribute to some extent to the slow Wallerian degeneration observed in C57BL/ OIa mice. It should be stressed that our immunocytochemistry does not allow the unequivocal distinction between extrinsic haematogenous and endogenous resident macrophages under conditions of blood-nerve barrier damage. The anti-rat monocyte/macrophage antibodies so far described recognize differentiation markers which can be expressed in principle on both types of cells (Dijkstra et al., 1985; Robinson et al., 1986;
MONACO, GEHRMANN, RAIVICH and KREUTZBERG Gehrmann & Kreutzberg, 1991). Several lines of evidence, however, suggest that the macrophage response in the injured sciatic nerve is differentially regulated in the proximal and the distal part as well as at the site of trauma. At the site of trauma MUC 101-, MUC 102-, OX-42-, and ED1-3-positive monocytes/ macrophages were detected as early as 1 day postlesion. Based on their morphology, tl~eir immunophenotype and the documented breakdown of the blood-nerve barrier (Olsson, 1966), they should represent haematogenous, invading myelomonocytic cells. In the distal part, resident, ramified MUC 101-, MUC 102-, OX-42-, and ED1-3-positive macrophages assumed a round morphology" similar to that of foamy macrophages. At the late stages of Wallerian degeneration studied, they formed strings of round to square shaped cells as described recently for transferrin receptor- and EDl-positive macrophages in the injured sciatic nerve (Raivich et al., 1991). The expression of complement type three receptor on these macrophages has been found essential in myelin phagocytosis both in vitro (Brueck & Friede, 1990) and in vivo (Lunn et aI., 1989). In contrast, the number of MHC class II-positive resident macrophages was only slightly increased in the distal part and at the site of trauma. In contrast to the MUC 101-, MUC 102-, OX-42- and ED1-3-positive foamy macrophages, MHC class II-positive macrophages only transiently lost their ramified morphology to become round cells, and were never arranged in strings of foamy macrophages. Thus the contribution of resident, MHC class IIpositive cells to myelin phagocytosis in Wallerian degeneration seems to be limited. In view of the absence of T lymphocyte involvement in Wallerian degeneration, MHC class II expression on resident macrophages appears to be tess important than, for example, in T cell-mediated autoimmune neuritis. However, our results do not exclude the possibility that some MHC class II-positive cells lose this marker as they transform into phagocytic macrophages. Therefore the differential expression of cellular markers on macrophages in the peripheral nerve could be related both to different activation states as well as to cellular heterogeneity of resident PNS macrophages. Heterogeneity of tissue macrophages is in fact welldocumented in other tissues (Springer & Unkeless, 1984): A recent study on the expression of transferrin receptor indicates that different populations of cells of the mononuclear phagocyte system indeed exist in the injured sciatic nerve (Raivich et aI., 1991). In summary, resident, ramified macrophages in the PNS show to some extent morphological and immunophenotypical similarities with microglial cells in the CNS suggesting a role as 'PNS microglial cells'. These resident macrophages are strategically located to respond rapidly to nerve injury as well as to act as the major APCs in the PNS.
R e s i d e n t m a c r o p h a g e s in t h e P N S
Acknowledgements W e a p p r e c i a t e t h e t e c h n i c a l a s s i s t a n c e of D i e t m u t e Bfiringer, Irmtraud Milojevic and Karin Brflckner. We
633 thank Martin Reddington and Wolfgang Klinkert ( D e p a r t m e n t of N e u r o i m m u n o l o g y ) for r e a d i n g t h e m a n u s c r i p t . Ralf G o l d ( D e p a r t m e n t of N e u r o i m m u nology) has helped with the irradiation experiments.
References BAICHIWAL, R. R., BIGBEE, J. w. & DEVRIES, G. H. (1988) M a c r o p h a g e - m e d i a t e d myelin-related mitogenic factor for cultured Schwann cells. Proceedings of the National Academy of Sciences (USA) 85, 1701-5. BERNER, A., TORVIK, A, & STENWIG, A. E. (1973) Origin of macrophages in traumatic lesions and Wallerian degeneration in peripheral nerves. Acta Neuropathologica 25, 228-36. BEUCHE, W. & FRIEDE, R. L. (1984) The role of non-resident cells in Wallerian degeneration. Journal of Neurocytology 13, 767-96. BEUCHE, W. & FRIEDE, R. L. (1986) Myelin phagocytosis in Wallerian degeneration d e p e n d s on silica-sensitive, bg/ bg-negative and Fc-positive monocytes. Brain Research 378, 97-106. BRUECK, W. & FRIEDE, R. L. (1990) Anti-macrophage CR3 antibody blocks myelin phagocytosis by macrophages in vitro. Acta Neuropathologica 80, 415-8. CRAGGS, R. I. & WEBSTER, H. DEF. (1985) Ia antigens in the normal rat nervous system and in lesions of experimental allergic encephalomyelitis. Acta Neuropathologica 68, 263-72. DIJKSTRA, C. D., DOPP, E. A., JOLING, P. & KRAAL, G. (1985) The heterogeneity of mononuclear phagocytes in l y m p h oid organs: distinct macrophage subpopulations in the rat recognized by monoclonal antibodies ED1, ED2 and ED3. Immunology 54, 589-99. GEHRMANN, J. & KREUTZBERG, G. W. (1991) Characterization of two n e w monoclonal antibodies directed against rat microglia. Journal of Comparative Neurology 313, 409-30. GEHRMANN, J., MONACO, S. & KREUTZBERG, G. W. (1991a) Spinal cord microglial cells and DRG satellite cells rapidly r e s p o n d to transection of the rat sciatic nerve. Restorative Neurology and Neuroscience 2, 181-98. GEHRMANN, J., SCHOEN, S. W. & KREUTZBERG, G. W. (1991b) Lesion of the rat entorhinal cortex leads to a rapid microglial reaction in the dentate gyrus. A light and electron microscopical study. Acta Neuropathologica 82, 442-55. GRAEBER, M. B., STREIT, W. J. & KREUTZBERG, G. W. (1989) Identity of ED2-positive perivascular cells in rat brain. Journal of Neuroscience Research 22, 103-6. GRAEBER, M. B. & STREIT, W. J. (1990) Microglia: i m m u n e network in the CNS. Brain Pathology 1, 2-5. GRIFFIN, J. W., LOBATO, C. & GLASS, J. D. (1991) The resident m a c r o p h a g e system of peripheral nerves. Journal of Neuropathology and Experimental Neurology 50, 321. HANN-BONNEKOH, P. G., SCHEIDT, P. & FRIEDE, R. L. (1989) Myelin phagocytosis by peritoneal macrophages in organ cultures of m o u s e peripheral nerve. A n e w model for studying myelin phagocytosis in vitro. Journal of Neuropathology and Experimental Neurology 48, 140-53.
HART, D. N. J. & FABRE, J. W. (1981) Demonstration and characterization of Ia-positive dendritic cells in the interstitial connective tissues of rat heart and other tissues, but not brain. Journal of Experimental Medicine 154, 347-6l. HUGHES, R. A., ATKINSON, P. F., GRAY, I. A. &TAYLOR, W. A. (1987) Major histocompatibility antigens and lymphocytes subsets during experimental allergic neuritis in the Lewis rat. Journal of Neurology 234, 390-5. KLINKERT, W. E. F. (1990) L y m p h o i d dendritic accessory cells of the rat. Immunological Reviews 117, 103-20. LASSMANN, H., VASS, K., BRUNNER, C. & WISNIEWSKI, H. M. (1986) Peripheral nervous system lesions in experimental allergic encephalomyelitis. Acta Neuropathologica 69, 193-204. LAWSON, L. J., PERRY, V. H., DR1, P. & GORDON, S. (1990) Heterogeneity in the distribution and m o r p h o l o g y of microglia in the normal adult mouse brain. Neuroscience 39, 151-70. LEIBOVICH, S. J. & ROSS, R. (1976) A macrophaged e p e n d e n t factor stimulates the proliferation of fibroblasts in vitro. American Journal of Pathology 84, 501-13. LINDHOLM, D., HEUMANN, R., MEYER, M. & THOENEN, H. (1987) Interleukin-1 regulates nerve growth factor (NGF) synthesis in the non-neuronal cells of the rat sciatic nerve. Nature 330, 658-9. L1NDHOLM, D., HEUMANN, R.,HENGERER, B. &THOENEN, H. (1988) Interleukin 1 increases stability and transcription of m R N A encoding nerve growth factor in cultured rat fibroblasts. Journal of Biological Chemistry 263, 1634851. LUNN, E. R., PERRY, V. H., BROWN, M. C., ROSEN, H. & GORDON, S. (1989) Absence of Wallerian degeneration does not hinder regeneration in peripheral nerve. European Journal of Neuroscience 1, 27-33. MARTINET, Y., BITTERMAN, P. B., MORNEX, J- F., GROTENDORST, G. R., MARTIN, G. R. & CRYSTAL, R. (1986) Activated h u m a n monocytes express c-sis protooncogene and release a mediator showing PDGF-like activity. Nature 319, 158-60. MCMASTER, W. R. & WILLIAMS, A. F. (1979) Identification of Ia glycoprotein in rat t h y m u s and purification from rat spleen. European Journal of Immunology 9, 426-33. OLDFORS, A. (1980) Macrophages in peripheral nerves. A n ultrastructural a n d e n z y m e histochemical study on rats. Acta Neuropathologica 49, 43-9. OLSSON, Y. & SJOSTRAND, I. (1969) Origin of macrophages in Wallerian degeneration of peripheral nerves demonstrated autoradiographically. Experimental Neurology 23, 102-12. PERRY, V. H. & GORDON, S. (1987) Modulation of CD4 antigen on macrophages and microglia in rat brain. Journal of Experimental Medicine 166, 1138-43.
634 PERRY, V. H., BROWN, M. C. & GORDON, S. (1987) The macrophage response to central and peripheral nerve injury. Journal of Experimental Medicine 165, 1218-23. PERRY, V. H. & GORDON, S. (1988) Macrophages and microglia in the nervous System. Trends in Neurosciences 11, 273-7. PERRY, V. H., BROWN, M. C., LUNN, E. R., TREE, P. & G O R D O N , S. (1990) Evidence that very slow Wallerian degeneration in C57BL/OIa mice is an intrinsic property of the peripheral nerve. European Journal of Neuroscience 2, 802-8. PETERS, A., PALAY, S. L. & WEBSTER, H. DEF. (1991) Microglia. In The Fine Structure of the Nervous System: The Neurons and Supporting Cells (edited by PETERS, A., PALAY, S. & WEBSTER, DEF, H.) pp. 304-11. New York, Oxford: Oxford University Press. RAIVICH, G. & KREUTZBERG, G. W. (1987) Expression of growth factor receptors in injured nervous tissue. II. Induction of specific platelet-derived growth factor binding in the injured nervous system is associated with a breakdown in the blood-nerve barrier and endoneurial interstitial oedema. Journal of Neurocytology 16, 701-11. RAIV1CH, G., HELLWEG, R., GRAEBER, M. B. & KREUTZBERG, G. W. (1990) The expression of growth factor receptors during nerve regeneration. Restorative Neurology and Neuroscience 1, 217-23. RAIVICH, G., GRAEBER, M. B., GEHRMANN, J. & KREUTZBERG, G. W. (1991) Transferrin receptor expression and iron uptake in the injured and regenerating rat sciatic nerve. European Journal of Neuroscience 3,919-27. ROBINSON, A. P., WHITE, T. M. & MASON, D. W. (1986) Macrophage heterogeneity in the rat as delineated by two monoclonal arttibodies MRC OX-41 and MRC OX-42, the
M O N A C O , G E H R M A N N , R A I V I C H and KREUTZBERG latter recognizing complement receptor type 3. Immunology 57, 527-31. SCHMIDT, B., STOLL, G., HARTUNG, H. P., HEININGER, K.. B., SCHAEFER, B. & TOYKA, K. V. (1990) Macrophages but not Schwann cells express Ia antigen in experimental a u t o i m m u n e neuritis. Annals of Neurology 28, 70-7. SCHUBERT, T. & FRIEDE, R. L. (1981) The role of endoneurial fibroblasts in myelin degradation. Journal of Neuropathology and Experimental Neurology 40, 134--54. SPRINGER, T. A. & UNKELESS, J. C. (1984) Analysis of macrophage differentiation and function with monoclonal antibodies. Contemporary Topics in Immunobiology 13, 1-23, STEVENS, A., SCHABET, M., SCHOTT, K. & WIETHOLTER, H. (1989) Role of endoneural cells in experimental allergic neuritis and characterisation of a resident phagocytic cell. Acta Neuropathologica 77, 412-9. STOLL, G., GRIFFIN, J. W., LI, C. Y. & TRAPP, B. D. (1989) Wallerian degeneration in the peripheral nervous system: participation of both Schwann cells and macrophages in myelin degradation. Journal of Neurocytology 18, 671-83. STREIT, W. J., GRAEBER, M. B. & KREUTZBERG, G. W. (1988) Functional plasticity of microglia: a review. Glia 1, 301-7. VASS, K. & LASSMANN, H. (1990) Intrathecal application of interferon gamma. Progressive appearance of MHC antigens within the rat nervous system. American Journal of Pathology 137, 789-800. WEKERLE, H., SCHWAB, M., LININGTON, C. & MEYERMANN, R. (1986) Antigen presentation in the peripheral nervous system: Schwann cells present endogenous myelin autoantigens to lymphocytes. European Journal of Immunology 16, 1551-7.
MHC-positive, ramified macrophages in the normal and injured rat peripheral nervous system S. M O N A C O * ,
J. G E H R M A N N ,
G. R A I V I C H a n d G. W. KREUTZBERG;
Department of Neuromorphotogy, Max-Planck-Institute for Psychiatry, Am Ktopferspitz 18A, D-8033 Martinsried, Germany Received 14 January 1992; revised 2 April 1992; accepted 14 April 1992
Summary Resident endoneurial macrophages form a prominent, but little recognized component of the PNS. We have studied immunocytochemicallythe distribution, morphology and immunophenotype of endoneurial macrophages in several normal peripheral nerves of the rat. In addition, we investigated the macrophage response following crush injury of the sciatic nerve. Resident endoneurial macrophages had a ramified morphology with processes oriented parallel to the long axis of nerve fibres. They were positive for several monocyte/macrophage markers such as ED1, ED2 and the recently-described MUC 101 and MUC 102 antibodies. They furthermore expressed the complement type three receptor, the CD4 antigen and MHC class I and II molecules. These results were consistent in all the peripheral nerves studied. In addition, 1000rad of y-irradia tion led to a strong reduction in the number of MHC class II-positive ramified cells in the peripheral nerves similar to that observed in other peripheral organs such as the heart. A considerable percentage of resident rnacrophages in the PNS and/or their precursor cells are therefore radiosensitive and could be related to the lineage of dendritic cells. Following crush injury, ED1-3-, OX-42-, MUC 101- and MUC 102-positive round macrophages were observed from 24 h postlesion onward at the site of trauma. In the distal part, they were observed to form strings of round, foamy macrophages probably involved in myelin phagocytosis. In contrast, the number of MHC class II-positive resident macrophages was only slightly increased at the site of trauma and in the distal part. These cells transformed from a ramified to a round morphology, but did not appear as typical strings of foamy macrophages. These results demonstrate that the PNS is provided with a resident macrophage population analogous in many respects to microglial cells in the CNS. These constitutively MHC class II-positive PNS microglial-like cells could act as the major antigen-presenting cells in the peripheral nerve. They may thus constitute a local immune defense system of the PNS with a function similar to that of microglial cells in the CNS.
Introduction In the mammalian CNS microglial cells are antigenically and functionally related to cells of the monocyte/ macrophage lineage; they thus represent the resident, intrinsic macrophage of the brain (Streit et al., 1988; Perry & Gordon,. 1988; Graeber & Streit, 1990). They comprise between 5% and 20% of the total glial cell population (Lawson et aI., 1990; Peters et al., 1991). Under normal conditions they have a typical, ramified morphology and are rapidly activated in response to even subtle pathological stimuli (Streit et al., 1988; Graeber & Streit, 1990; Gehrmann et al., 1991a,b). There is evidence that the PNS is also provided with resident macrophages that share some properties with other tissue macrophages (Hughes et aI., 1987; Perry & Gordon, 1987; Stevens et aI., 1989; Vass & Lassmann, 1990) and CNS microglial cells (Gehrmann et al.,
1991a,b). They comprise 1% and 4% respectively of the total endoneurial cell population in the normal peripheral nerve (Oldfors, 1980; Schubert & Friede, 1981). However, the immunophenotype, morphology and functional properties of these cells in the PNS await a more precise characterization. After peripheral nerves are crushed macrophages rapidly invade the site of trauma, but also the distal part of the nerve which undergoes Wallerian degeneration (Olsson & Sj6strand, 1969; Perry & Gordon, 1987; Stoll et al., 1989). The presence of activated macrophages in the injured peripheral nerve is functionally important. They phagocytose the myelin debris (Beuche & Friede, 1984, 1986; Stolt et aL, 1989), participate in the production of mitogenic factors for Schwann cells (Baichwal et al., 1988) and fibroblasts
* On leaveof absence fromthe Instituteof Neurology,Universityof Verona, Italy. To whom correspondenceshouldbe addressed. 0300-4864/92 $03.00 +.12 9 1992 Chapman and Hail Ltd
624 (Leibovich & Ross, 1976; Martinet et al., 1986; Raivich et al., 1990) a n d induce the synthesis of e n d o n e u r i a l n e r v e g r o w t h factor b y secreting interleukin-1 (Lindh o l m et al., 1987, 1988). The origin of m a c r o p h a g e s has b e e n in turn a s s i g n e d to h a e m a t o g e n o u s (Olsson & Sj6strand, 1969; Beuche & Friede, 1984) or e n d o g e n ous cells only (Berner et al., 1973). To characterize r e s i d e n t m a c r o p h a g e s in the rat PNS further as a n a l o g o u s to microglial cells in the CNS, e n d o n e u r i a l , resident m a c r o p h a g e s w e r e studied imm u n o c y t o c h e m i c a l l y u n d e r n o r m a l conditions a n d d u r i n g Wallerian d e g e n e r a t i o n . Several m o n o c l o n a l antibodies w e r e u s e d w h i c h recognize m o n o c y t e / m a c r o p h a g e m a r k e r s s u c h as the c o m p l e m e n t t y p e three r e c e p t o r a n d m a j o r histocompatibility c o m p l e x (MHC) antigens.
Materials and methods
Surgery and tissue collection Adult male Wistar rats weighing 200-250 g were used for all the experiments. For immunocytochemical investigation of resident macrophages in normal animals, three untreated, healthy rats were used. After decapitation under deep aether anaesthesia L5 motor and sensory roots, L5 dorsal root ganglia (DRG) and the sciatic, tibial, facial and vagal nerves were rapidly dissected, embedded in O.C.T. II tissue compound (Miles, Elkhart, Minnesota, USA) and snapfrozen in liquid nitrogen. For studying the macrophage reaction following Wallerian degeneration, the right sciatic nerve was crushed under aether anaesthesia in the gluteal region for 30 s with a jeweller's forceps. Three rats per time point were killed at 0, 1, 2, 4, 7, 14, 21 and 42 days after the operation. The sciatic nerve was removed for a length of about 30 mm, encompassing a 5-10 m m segment proximal to the crush site and the distal part, embedded in O. C.T. Compound (Miles, Elkhart, USA) and snap-frozen in liquid nitrogen. In addition, six 8-week-old male Wistar rats, anaesthetized with chloral hydrate, were irradiated by exposure to 1000 rad (given over i min) from a gamma-irradiation source (Stabilon, Siemens, FRG) according to the protocol given by Hart and Fabre (1981). Six naive littermates were used as controls. Ten days after irradiation the irradiated and control rats were killed; L5 motor and sensory roots, L5 DRG, the sciatic nerves, the tibial nerves and the heart were removed and rapidly frozen. Longitudinal cryostat sections (10 b~m)of the above specimens were serially cut and collected on gelatine-coated glass slides and stored at - 7 0 ~ until immunocytochemical examination.
Immunocytochemistry Cryostat sections were fixed in 3.7% formalin for 5 min followed by acetone (50% 2 min, 100% 2 rain and 50% 2 min) at room temperature. The sections were then rehydrated in 0.01 M phosphate-buffered saline (PBS) (pH 7.4) for 5 rain and 0.01M PBS containing 0.1% bovine serum albumin. After blocking with 2% normal horse serum (DAKO, Hamburg, FRG), adjacent sections were incubated over-
M O N A C O , G E H R M A N N , RAIVICH and KREUTZBERG night at 4 ~C with the following monoclonal antibodies for cell typing (working dilutions in 0.01 M PBS are given in parentheses): (1) MRC OX-42 (Serotec, UK), recognizing the rat complement receptor type three (Robinson et al., 1986) (1:6400); (2) MUC 101, recognizing two proteins of 116 and 95 kDa on rat microglia (Gehrmann & Kreutzberg, 1991) (1 : 1600); (3) MUC 102, recognizing two proteins of 62 and 70 kDa on rat microglia (Gehrmann & Kreutzberg, 1991) (1 : 1000); (4) MRC OX-6 (Serotec, UK) directed against MHC class II antigens (McMaster & Williams, 1979) (1:6400); (5) MRC OX-18 (Serotec, UK) directed against MHC Class I antigens; (6) the W3/25 antibody, recognizing the rat CD4 antigen (Perry & Gordon, 1987) (1 : 6400); (7) ED1, ED2, ED3 (Serotec, UK) recognizing monocytes/macrophages (Dijkstra et al., 1985) (1:6400). Subsequent antibody detection was carried out using a biotinylated horse anti-mouse secondary antibody (Serotec, UK) and the Vectastain ABC-Elite kit (Vector Labs, Burlingame, CA, USA) with 3,3'-diaminobenzidine as peroxidase substrate following a previously described protocol (Gehrmann et al., 1991a). Immunocytochemical controls included omission of the primary antibody or replacement of the primary antibody by an irrelevant antibody of the same IgG subclass. The number of OX-6- and EDl-positive macrophages was evaluated quantitatively in ten randomly chosen areas in the sciatic nerve of irradiated and control rats. Immunostained cells were counted twice in fields of 0.7 m m 2 at x 10 primary magnification. The values obtained were uncorrected cell counts (means + SEM) and were evaluated for statistical significance using Student's t-test.
Results THE NORMAL PERIPHERAL NERVE In all the p e r i p h e r a l n e r v e s investigated, e n d o n e u r i a l cells w e r e strongly positive for M H C class II antigens, r e c o g n i z e d b y t h e OX-6 a n t i b o d y (Fig. 1). T h e y w e r e f u r t h e r m o r e positive for M H C class I antigens, recognized b y the OX-18 antibody, a n d w e r e stained b y several m o n o c l o n a l antibodies such as ED2, M U C 101, M U C 102 a n d OX-42 w h i c h recognize different m o n o c y t e / m a c r o p h a g e antigens (Fig. 1). A few e n d o n e u r i a l cells preferentially a r o u n d b l o o d vessels w e r e labelled b y ED1 (Fig. 1). Endoneurial, ramified cells further e x p r e s s e d the CD4 antigen, r e c o g n i z e d b y the W3/25 a n t i b o d y (Fig. 1). In contrast, ED3 i m m u n o r e a c t i v i t y Table 1. Quantitative evaluation of immunostained macrophages in the sciatic nerve of control and irradiated rats.
Marker
Control
Irradiation
OX-6 ED1
57.07 + 0.74 51.94 + 5.05
7.69 + 0.66* 51.77 _+ 4.04
The mean values are given as the number of immunostained cells per mm2tissue section + SEMfrom 5 animals (n = 5). * irradiation versus control p ~ 0.001
Resident m a c r o p h a g e s in the PNS
625
Fig. 1. Stai!ning of resident, ramified macrophages with several monoclonal antibodies in the normal periphera~ nerve. The respective antibodies are given in the figure. (a,c-f) Sciatic nerve. (b) Vagus nerve, x 760.
626
MONACO, GEHRMANN, RAIVICH and KREUTZBERG
Fig. 2. ~-Irradiation experiment. (a,c,e) Control rats. (b,d,f) Irradiated rats. (a,b) Heart, OX-6 immunoreactivity (indicating the presence of MHC class II-antigens). MHC class II-positive cells are almost completely depleted in the heart following irradiation. (c,d) Sciatic nerve, ED2 immunoreactivity. (e,f) Sciatic nerve, OX-6 immunoreactivity. The number of ED2-positive macrophages remains unchanged following irradiation (d) whereas the number of MHC class II-positive, ramified cells is strongly reduced in the irradiated animals (f) similar to the heart (b). x 200.
was not observed on these endoneurial cells. The i m m u n o p h e n o t y p e of resident macrophages is summarized in Table 2. I m m u n o s t a i n e d cells had an elongate cell b o d y from which crenellated, uni-or bipolar processes arose. These processes were longitudinally oriented along the axis of nerve fibres (Fig. 1). The m o r p h o l o g y as well as the i m m u n o p h e n o t y p e of the labelled cells was consistent in all the peripheral a n d cranial nerves examined. In contrast to resident macrophages, Schwann cells and endothelial cells were not observed to express M H C antigens either
u n d e r normal or pathological conditions (for comparison see below). X-RAY IRRADIATION Irradition (1000 rad) induced a strong reduction in the n u m b e r of M H C class II-positive, dendritic cells from the heart as previously reported (Hart & Fabre, 1981) (Fig. 2a,b) whereas the n u m b e r of ED2-positive cells remained unchanged. The n u m b e r of MHC class II-positive cells was also strongly reduced in the peripheral nerves following irradiation. Quantitative
Resident macrophages in the PNS
627
Fig. 3. MUC 101 (left) and OX-6 (right) immunoreactivities of adjacent longitudinal sections of normal (0 days) and crushed (1--42 days) sciatic nerves. The site of crush is in the middle of each micrograph, x 18. Note the rapid and strong increase in MUC 101 immunoreactivity at the site of crush compared to the relatively low levels of OX-6 immunoreactivity, indicating the presence of MHC class II molecules.
evaluation showed that almost 90% of the MHC class II-positive cells (p < 0.001) were depleted in the sciatic nerve following y-irradiation (Table 1). Similar to the situation in heart, the number of ED2-positive cells in the irradiated sciatic nerve remained unchanged (Table 1) although they appeared to be slightly hypertrophic. The expression of other cellular markers such as the CD4 antigen, the MUC 101, MUC 102 determinants and the complement type three receptor was not affected by the irradiation.
WALLERIAN DEGENERATION
The site of crush Numerous MUC 101-positive cells were observed at the site of the crush, I day postlesion (Fig. 3, left side). They had a round morphology and were mainly observed around endoneurial blood vessels and along the injured perineurium. These cells were also labelled by the MUC 102, ED1 and OX42 antibodies (data not shown). By day 2 through day 4 postlesion the number
628 and staining intensity of MUC 101, MUC 102, OX-42 and EDl-positive cells further increased at the site of crush. They were then also stained by the ED3 antibody which otherwise did not stain macrophages in the peripheral nerve (data not shown). From day 4 onward, immunostained cells were also observed in the adjacent segments proximal and distal to the site of crush and were arranged in strings of rounded cells. Staining of these strings of macrophages reached a maximum around two weeks postlesion. After six weeks the staining intensity was only slightly increased as compared to control nerves. Although some of the labelled cells still displayed a round, stout morphology after six weeks, most cells again displayed a slender, ramified morphology. In contrast, the number of MHC class II-positive resident macrophages was decreased at the site of crush at 1 day postlesion (Fig. 3, right side). From 2 days postlesion onward, the number and staining intensity of MHC class II-positive macrophages was slightly increased at the site of crush including the closely adjacent proximal and distal segments. This reaction reached a maximum around 4 days postlesion. At this time point MHC class II-positive cells further transformed from the ramified to a round morphology. Strings of MHC class II-positive macrophages, however, similar to the distribution of, for example, MUC 101- or OX-42-positive cells, were not observed. From two weeks onward, the number of MHC class II-expressing cells had returned to control levels. These cells again displayed a ramified morphology. The distal part A first noticeable increase in the number of MUC 101-positive macrophages was observed in the distal part at 4 days postlesion (Fig. 4, left side). These cells were also stained by the MUC 102, OX-42, ED1 and ED2 antibodies. They transformed from their slender, ramified morphology to a round morphology resembling phagocytic cells. Seven days postlesion, many transition forms between ramified and phagocytic macrophages were seen. In addition, small, round cells most likely representing invading monocytes were-occasionally intermingled with these large, foamy macrophages. From two weeks onward, the immunostained foamy macrophages were arranged in strings of rounded cells. This string-like arrangement was most evident six weeks after the crush injury, the latest time point studied (Fig. 4, left side).
MONACO, GEHRMANN, RAIVICH and KREUTZBERG In contrast, the number of MHC class II-positive cells was only slightly increased in the distal part at 4 days postlesion (Fig. 4, right side). Some of these MHC class II-positive cells had a round morphology. However, MHC class II immunoreactivity was not observed on the strings of large, foamy macrophages as revealed for example by MUC 101 immunoreactivity (cf Fig. 4, left side). Forty-two days postlesion, MHC class II-positive cells showed different morphologies depending on their localization in the injured sciatic nerve (Fig. 5a). In the proximal part, the number of these cells was increased. They had a ramified morphology, but were hypertrophic (Fig. 5b). In a segment 1.5 cm distal from the site of crush, MHC class II-positive cells appeared as ramified cells with a density comparable to that in control nerves (Fig. 5c). In a segment more distal than 1.5 cm from the site of crush, the number of MHC class II-positive cells was not increased, but they had a round morphology with few, short processes (Fig. 5d). Discussion
This study shows that, in the normal rat PNS, resident macrophages express several monocyte/macrophages antigens similar to other tissue macrophages and CNS microglial cells (Dijkstra et al., 1985; Streit et al., 1988; Perry & Gordon, 1988). This extends previous observations of ED2- or W3/25-positive spindle shaped cells in the rat peripheral nerve (Hughes et aI., 1987; Stevens et al., 1989) and F4/80-positive cells in the mouse PNS (Perry et al., 1990). Their ramified morphology is reminiscent of that of resting microglial cells in the CNS, particularly in the white matter (Streit et al., 1988; Lawson et al., 1990; Graeber & Streit, 1990). Resident macrophages in the normal rat PNS Resident macrophages in the PNS expressed several monocyte/macrophage markers such as the complement type three receptor and the ED1 and ED2 determinants which are characteristic of cells of the monocyte/macrophage lineage (Dijkstra et aI., 1985). ED3, however, was not found on resident macrophages in the PNS. Resident macrophages were furthermore labelled by the monoclonal antibodies MUC 101 and 102 which recognize rat microglial cells in the CNS and other mononuclear cells in the peripheral organs (Gehrmann & Kreutzberg, 1991). Morover, they showed a strong, constitutive MHC class I and II expression and were positive for the CD4
Fig. 4. MUC 101 (left) and OX-6 (right) immunoreactivities in adjacent longitudinal sections of the distal part of normal (0 days) and crushed (4M2 days) sciatic nerves. The micrographs were taken from segments about 1.5 cm distal from the site of crush, x 200. MUC 101-positive cells form strings of round macrophages, most prominently 6 weeks postlesion, whereas the number and morphology of OX-6-positive cells undergoes only relatively mild and transient changes.
631
Resident macrophages in the PNS Table 2. Immunocytochemical staining characteristics of resting microglial cells in the CNS and resident macrophages in the PNS of the rat.
Antibody
Structure/cell recognized
CNS microglia
PNS macrophage
OX-42 ED1 ED2 ED3 MUC 101
CR3 receptor Macrophage antigen Macrophage antigen Macrophage antigen Microglial and perivascular cells, 116 and 95 kDa antigen Microglial cells, 62 and 70 kDa antigen CD4 antigen MHC class II antigen (Ia) MHC class I antigen
++ -
++ + ++ -
_+
++
++ _+ • •
++ ++ ++ §+
MUC102 W3/25 OX-6 OX-18
no staining. • only a small number of cells stained. + welldiscernable. ++ strong staining. -
antigen. In the PNS, their i m m u n o p h e n o t y p e thus resembles closely that of satellite cells in the dorsal root ganglion ( G e h r m a n n et al., 1991a). In the CNS, however, the i m m u n o p h e n o t y p e of resident PNS macrophages differs to some extent from t h a t of resting microgtial cells (for comparison see Table 2). Resting microglial cells are also positive for the complement type three receptor, the MUC 101- and MUC 102-determinants, but express the ED determinants, the CD4 antigen and M H C antigens at m u c h lower levels (Streit et al., 1988; Perry & Gordon, 1987; Graeber et al., 1989; Vass & Lassmann, 1990; G e h r m a n n et aI., 1991a). The constitutive M H C class II antigen expression by resident macrophages in t'he PNS is in contrast to the results of previous studies which described only a few M H C class II-positive cells either in the e p i n e u r i u m (Craggs & Webster, 1985) or preferentially in a perivascular position in the e n d o n e u r i u m of peripheral nerves and spinal nerve roots (Lassmann et al., 1986; Vass & Lassmann, 1990). A recent preliminary report, however, confirms the high constitutive M H C class II expression by resident macrophages t h r o u g h o u t the PNS of different species (Griffin et al., 1991). M H C class II products (Ia) are antigen-binding glycoproteins which are expressed on the surface of accessory cells presenting antigen to T-lymphocytes. Accessory ceils can be divided into two groups: (a)
professional accessory cells such as dendritic cells which constitutively express high levels of Ia-antigen, and (2) facultative accessory cells such as peripheral macrophages and microglial cells which constitutively express only low levels of Ia-antigens (for review see Klinkert, 1990). The constitutive high level of Iaantigen expression on resident macrophages in the PNS as well as their ramified m o r p h o l o g y suggest a possible relationship to dendritic cells. M H C class II-positive, ramified macrophages were almost completely depleted from the PNS following 1000 rad of y-irradiation, which has been s h o w n to deplete effectively other tissues from dendritic cells due to their low residency time (Hart & Fabre, 1981). The results of the present s t u d y indicate that a high percentage of M H C class II-positive macrophages and/or their precursors are radiosensitive. In contrast, the n u m b e r of ED1positive macrophages remained u n c h a n g e d following irradiation, which w o u l d suggest a heterogeneity of PNS macrophages in terms of lineage and functional properties. In spite of their morphological, antigenic and functional similarities, however, it is still debatable w h e t h e r CNS microglial cells, PNS resident macrophages and l y m p h o i d dendritic cells might in fact belong to the same cellular entity or represent distinct cell lineages. Although the irradiation experim e n t offers only a single line of evidence, the results of this experiment could indicate an analogy to the data
OX-6 immunoreactivity indicating the presence of MHC class II antigens in a longitudinal section of a sciatic nerve 42 days after crush injury. (a) Overview. The arrows indicate the site of crush. The distal stump is below the arrows, x 12. (b) Proximal part. The number of MHC class II-positive macrophages is increased as compared to the distal part (cf. (c,d)), but they have a ramified morphology. (c) Distal part, segment 1.5 cm distal from the site of crush. MHC class-lI positive cells have a density and morphology similar to the normal nerves (for comparison see Fig. 1). (d) Distal part, segment more than 1.5 cm distal from the site of crush. The number of stained cells is not increased, but they have a round, stout morphology. (b-d) • 760. F i g . S.
632 obtained by Hart and Fabre (1981) for several peripheral organs suggesting that at least a part of the resident macrophages in the PNS may be related to the lineage of dendritic cells. MHC class II-positive, resident macrophages would seem to represent a local immune defense system in the PNS analogous to microglial cells in the CNS (Streit et at., 1988; Graeber & Streit, 1990). The nature of antigen-presenting cells (APC) in the PNS, however, is still controversial. In addition to macrophages, Schwann cells have been shown in vitro to act as APCs (Wekerle et al., 1986), while their rote in vivo as APCs is debatable. Under pathological conditions, in particular in experimental autoimmune neuritis, macrophages but not Schwann cells express Ia molecules (Schmidt et aI., 1990). Based on these in vivo observations, resident macrophages would seem to qualify as the main intrinsic APCs of the PNS. Macrophages after WalIerian degeneration As demonstrated by Olsson and Sj6strand (1969), haematogenous macrophages rapidly invade the injured sciatic nerve mainly at the site of trauma. Activated macrophages then serve a variety of functions which are related to the phagocytosis of the degenerating myelin, stimulation of Schwann cell proliferation and possibly the ensuing nemite outgrowth (Perry et aI., 1987; Baichiwal et aI., 1988; Raivich et aI., 1990, 1991). Recent studies using a mouse mutant, C57BL/OIa, in which Wallerian degeneration is slow, purport to show a major role of haematogenous macrophages in the phagocytosis of the myelin sheath during rapid Wallerian degeneration in the PNS (Lunn et aI., 1989). In addition, the recruitment of haematogenous macrophages can be prevented by antibodies against the complement receptor type three (Lunn et al., 1989). In contrast, the involvement of resident PNS macrophages in Wallerian degeneration is still debated. In vitro, marked myelin phagocytosis occurs in nerves kept in organ cultures without added monocytes (Hann-Bonnekoh et al., 1989), suggesting that resident macrophages are the main source of mononuclear phagocytes under these conditions. Using a bone marrow chimera approach, Perry and colleagues (1990) have discussed the possibility that these resident macrophages could contribute to some extent to the slow Wallerian degeneration observed in C57BL/ OIa mice. It should be stressed that our immunocytochemistry does not allow the unequivocal distinction between extrinsic haematogenous and endogenous resident macrophages under conditions of blood-nerve barrier damage. The anti-rat monocyte/macrophage antibodies so far described recognize differentiation markers which can be expressed in principle on both types of cells (Dijkstra et al., 1985; Robinson et al., 1986;
MONACO, GEHRMANN, RAIVICH and KREUTZBERG Gehrmann & Kreutzberg, 1991). Several lines of evidence, however, suggest that the macrophage response in the injured sciatic nerve is differentially regulated in the proximal and the distal part as well as at the site of trauma. At the site of trauma MUC 101-, MUC 102-, OX-42-, and ED1-3-positive monocytes/ macrophages were detected as early as 1 day postlesion. Based on their morphology, tl~eir immunophenotype and the documented breakdown of the blood-nerve barrier (Olsson, 1966), they should represent haematogenous, invading myelomonocytic cells. In the distal part, resident, ramified MUC 101-, MUC 102-, OX-42-, and ED1-3-positive macrophages assumed a round morphology" similar to that of foamy macrophages. At the late stages of Wallerian degeneration studied, they formed strings of round to square shaped cells as described recently for transferrin receptor- and EDl-positive macrophages in the injured sciatic nerve (Raivich et al., 1991). The expression of complement type three receptor on these macrophages has been found essential in myelin phagocytosis both in vitro (Brueck & Friede, 1990) and in vivo (Lunn et aI., 1989). In contrast, the number of MHC class II-positive resident macrophages was only slightly increased in the distal part and at the site of trauma. In contrast to the MUC 101-, MUC 102-, OX-42- and ED1-3-positive foamy macrophages, MHC class II-positive macrophages only transiently lost their ramified morphology to become round cells, and were never arranged in strings of foamy macrophages. Thus the contribution of resident, MHC class IIpositive cells to myelin phagocytosis in Wallerian degeneration seems to be limited. In view of the absence of T lymphocyte involvement in Wallerian degeneration, MHC class II expression on resident macrophages appears to be tess important than, for example, in T cell-mediated autoimmune neuritis. However, our results do not exclude the possibility that some MHC class II-positive cells lose this marker as they transform into phagocytic macrophages. Therefore the differential expression of cellular markers on macrophages in the peripheral nerve could be related both to different activation states as well as to cellular heterogeneity of resident PNS macrophages. Heterogeneity of tissue macrophages is in fact welldocumented in other tissues (Springer & Unkeless, 1984): A recent study on the expression of transferrin receptor indicates that different populations of cells of the mononuclear phagocyte system indeed exist in the injured sciatic nerve (Raivich et aI., 1991). In summary, resident, ramified macrophages in the PNS show to some extent morphological and immunophenotypical similarities with microglial cells in the CNS suggesting a role as 'PNS microglial cells'. These resident macrophages are strategically located to respond rapidly to nerve injury as well as to act as the major APCs in the PNS.
R e s i d e n t m a c r o p h a g e s in t h e P N S
Acknowledgements W e a p p r e c i a t e t h e t e c h n i c a l a s s i s t a n c e of D i e t m u t e Bfiringer, Irmtraud Milojevic and Karin Brflckner. We
633 thank Martin Reddington and Wolfgang Klinkert ( D e p a r t m e n t of N e u r o i m m u n o l o g y ) for r e a d i n g t h e m a n u s c r i p t . Ralf G o l d ( D e p a r t m e n t of N e u r o i m m u nology) has helped with the irradiation experiments.
References BAICHIWAL, R. R., BIGBEE, J. w. & DEVRIES, G. H. (1988) M a c r o p h a g e - m e d i a t e d myelin-related mitogenic factor for cultured Schwann cells. Proceedings of the National Academy of Sciences (USA) 85, 1701-5. BERNER, A., TORVIK, A, & STENWIG, A. E. (1973) Origin of macrophages in traumatic lesions and Wallerian degeneration in peripheral nerves. Acta Neuropathologica 25, 228-36. BEUCHE, W. & FRIEDE, R. L. (1984) The role of non-resident cells in Wallerian degeneration. Journal of Neurocytology 13, 767-96. BEUCHE, W. & FRIEDE, R. L. (1986) Myelin phagocytosis in Wallerian degeneration d e p e n d s on silica-sensitive, bg/ bg-negative and Fc-positive monocytes. Brain Research 378, 97-106. BRUECK, W. & FRIEDE, R. L. (1990) Anti-macrophage CR3 antibody blocks myelin phagocytosis by macrophages in vitro. Acta Neuropathologica 80, 415-8. CRAGGS, R. I. & WEBSTER, H. DEF. (1985) Ia antigens in the normal rat nervous system and in lesions of experimental allergic encephalomyelitis. Acta Neuropathologica 68, 263-72. DIJKSTRA, C. D., DOPP, E. A., JOLING, P. & KRAAL, G. (1985) The heterogeneity of mononuclear phagocytes in l y m p h oid organs: distinct macrophage subpopulations in the rat recognized by monoclonal antibodies ED1, ED2 and ED3. Immunology 54, 589-99. GEHRMANN, J. & KREUTZBERG, G. W. (1991) Characterization of two n e w monoclonal antibodies directed against rat microglia. Journal of Comparative Neurology 313, 409-30. GEHRMANN, J., MONACO, S. & KREUTZBERG, G. W. (1991a) Spinal cord microglial cells and DRG satellite cells rapidly r e s p o n d to transection of the rat sciatic nerve. Restorative Neurology and Neuroscience 2, 181-98. GEHRMANN, J., SCHOEN, S. W. & KREUTZBERG, G. W. (1991b) Lesion of the rat entorhinal cortex leads to a rapid microglial reaction in the dentate gyrus. A light and electron microscopical study. Acta Neuropathologica 82, 442-55. GRAEBER, M. B., STREIT, W. J. & KREUTZBERG, G. W. (1989) Identity of ED2-positive perivascular cells in rat brain. Journal of Neuroscience Research 22, 103-6. GRAEBER, M. B. & STREIT, W. J. (1990) Microglia: i m m u n e network in the CNS. Brain Pathology 1, 2-5. GRIFFIN, J. W., LOBATO, C. & GLASS, J. D. (1991) The resident m a c r o p h a g e system of peripheral nerves. Journal of Neuropathology and Experimental Neurology 50, 321. HANN-BONNEKOH, P. G., SCHEIDT, P. & FRIEDE, R. L. (1989) Myelin phagocytosis by peritoneal macrophages in organ cultures of m o u s e peripheral nerve. A n e w model for studying myelin phagocytosis in vitro. Journal of Neuropathology and Experimental Neurology 48, 140-53.
HART, D. N. J. & FABRE, J. W. (1981) Demonstration and characterization of Ia-positive dendritic cells in the interstitial connective tissues of rat heart and other tissues, but not brain. Journal of Experimental Medicine 154, 347-6l. HUGHES, R. A., ATKINSON, P. F., GRAY, I. A. &TAYLOR, W. A. (1987) Major histocompatibility antigens and lymphocytes subsets during experimental allergic neuritis in the Lewis rat. Journal of Neurology 234, 390-5. KLINKERT, W. E. F. (1990) L y m p h o i d dendritic accessory cells of the rat. Immunological Reviews 117, 103-20. LASSMANN, H., VASS, K., BRUNNER, C. & WISNIEWSKI, H. M. (1986) Peripheral nervous system lesions in experimental allergic encephalomyelitis. Acta Neuropathologica 69, 193-204. LAWSON, L. J., PERRY, V. H., DR1, P. & GORDON, S. (1990) Heterogeneity in the distribution and m o r p h o l o g y of microglia in the normal adult mouse brain. Neuroscience 39, 151-70. LEIBOVICH, S. J. & ROSS, R. (1976) A macrophaged e p e n d e n t factor stimulates the proliferation of fibroblasts in vitro. American Journal of Pathology 84, 501-13. LINDHOLM, D., HEUMANN, R., MEYER, M. & THOENEN, H. (1987) Interleukin-1 regulates nerve growth factor (NGF) synthesis in the non-neuronal cells of the rat sciatic nerve. Nature 330, 658-9. L1NDHOLM, D., HEUMANN, R.,HENGERER, B. &THOENEN, H. (1988) Interleukin 1 increases stability and transcription of m R N A encoding nerve growth factor in cultured rat fibroblasts. Journal of Biological Chemistry 263, 1634851. LUNN, E. R., PERRY, V. H., BROWN, M. C., ROSEN, H. & GORDON, S. (1989) Absence of Wallerian degeneration does not hinder regeneration in peripheral nerve. European Journal of Neuroscience 1, 27-33. MARTINET, Y., BITTERMAN, P. B., MORNEX, J- F., GROTENDORST, G. R., MARTIN, G. R. & CRYSTAL, R. (1986) Activated h u m a n monocytes express c-sis protooncogene and release a mediator showing PDGF-like activity. Nature 319, 158-60. MCMASTER, W. R. & WILLIAMS, A. F. (1979) Identification of Ia glycoprotein in rat t h y m u s and purification from rat spleen. European Journal of Immunology 9, 426-33. OLDFORS, A. (1980) Macrophages in peripheral nerves. A n ultrastructural a n d e n z y m e histochemical study on rats. Acta Neuropathologica 49, 43-9. OLSSON, Y. & SJOSTRAND, I. (1969) Origin of macrophages in Wallerian degeneration of peripheral nerves demonstrated autoradiographically. Experimental Neurology 23, 102-12. PERRY, V. H. & GORDON, S. (1987) Modulation of CD4 antigen on macrophages and microglia in rat brain. Journal of Experimental Medicine 166, 1138-43.
634 PERRY, V. H., BROWN, M. C. & GORDON, S. (1987) The macrophage response to central and peripheral nerve injury. Journal of Experimental Medicine 165, 1218-23. PERRY, V. H. & GORDON, S. (1988) Macrophages and microglia in the nervous System. Trends in Neurosciences 11, 273-7. PERRY, V. H., BROWN, M. C., LUNN, E. R., TREE, P. & G O R D O N , S. (1990) Evidence that very slow Wallerian degeneration in C57BL/OIa mice is an intrinsic property of the peripheral nerve. European Journal of Neuroscience 2, 802-8. PETERS, A., PALAY, S. L. & WEBSTER, H. DEF. (1991) Microglia. In The Fine Structure of the Nervous System: The Neurons and Supporting Cells (edited by PETERS, A., PALAY, S. & WEBSTER, DEF, H.) pp. 304-11. New York, Oxford: Oxford University Press. RAIVICH, G. & KREUTZBERG, G. W. (1987) Expression of growth factor receptors in injured nervous tissue. II. Induction of specific platelet-derived growth factor binding in the injured nervous system is associated with a breakdown in the blood-nerve barrier and endoneurial interstitial oedema. Journal of Neurocytology 16, 701-11. RAIV1CH, G., HELLWEG, R., GRAEBER, M. B. & KREUTZBERG, G. W. (1990) The expression of growth factor receptors during nerve regeneration. Restorative Neurology and Neuroscience 1, 217-23. RAIVICH, G., GRAEBER, M. B., GEHRMANN, J. & KREUTZBERG, G. W. (1991) Transferrin receptor expression and iron uptake in the injured and regenerating rat sciatic nerve. European Journal of Neuroscience 3,919-27. ROBINSON, A. P., WHITE, T. M. & MASON, D. W. (1986) Macrophage heterogeneity in the rat as delineated by two monoclonal arttibodies MRC OX-41 and MRC OX-42, the
M O N A C O , G E H R M A N N , R A I V I C H and KREUTZBERG latter recognizing complement receptor type 3. Immunology 57, 527-31. SCHMIDT, B., STOLL, G., HARTUNG, H. P., HEININGER, K.. B., SCHAEFER, B. & TOYKA, K. V. (1990) Macrophages but not Schwann cells express Ia antigen in experimental a u t o i m m u n e neuritis. Annals of Neurology 28, 70-7. SCHUBERT, T. & FRIEDE, R. L. (1981) The role of endoneurial fibroblasts in myelin degradation. Journal of Neuropathology and Experimental Neurology 40, 134--54. SPRINGER, T. A. & UNKELESS, J. C. (1984) Analysis of macrophage differentiation and function with monoclonal antibodies. Contemporary Topics in Immunobiology 13, 1-23, STEVENS, A., SCHABET, M., SCHOTT, K. & WIETHOLTER, H. (1989) Role of endoneural cells in experimental allergic neuritis and characterisation of a resident phagocytic cell. Acta Neuropathologica 77, 412-9. STOLL, G., GRIFFIN, J. W., LI, C. Y. & TRAPP, B. D. (1989) Wallerian degeneration in the peripheral nervous system: participation of both Schwann cells and macrophages in myelin degradation. Journal of Neurocytology 18, 671-83. STREIT, W. J., GRAEBER, M. B. & KREUTZBERG, G. W. (1988) Functional plasticity of microglia: a review. Glia 1, 301-7. VASS, K. & LASSMANN, H. (1990) Intrathecal application of interferon gamma. Progressive appearance of MHC antigens within the rat nervous system. American Journal of Pathology 137, 789-800. WEKERLE, H., SCHWAB, M., LININGTON, C. & MEYERMANN, R. (1986) Antigen presentation in the peripheral nervous system: Schwann cells present endogenous myelin autoantigens to lymphocytes. European Journal of Immunology 16, 1551-7.
