Indefinite survival of neural xenografts induced with anti-CD4 monoclonal antibodies

Nnuoscience

Vol 70, No. 3, pp. 775 -7X9, I996 Elsevier Science Ltd Copyright 8 1995 IBRO Prmted in Great Britain. All rights reserved 0306-4522/96 $15.00 + 0.00

0306-4522(95)00400-9

INDEFINITE SURVIVAL OF NEURAL XENOGRAFTS INDUCED WITH ANTI-CD4 MONOCLONAL ANTIBODIES M. J. A. WOOD,*+

D. J. SLOAN,?

K. J. WOODS

and

H. M. CHARLTONI

tDepartment of Human Anatomy. South Parks Road, Oxford OXI 3QX. U.K. $Nuffield Department of Surgery. John Radcliffe Hospital, Oxford OX3 9DU, U.K. Abstract-Xenografts of neural tissue are usually rapidly rejected when transplanted into the central nervous system of adult recipient animals. This study has examined the cell mediated immune response to both concordant (between closely related species) and discordant (between distantly related species) neural xenografts in the mouse, and has investigated the role of the CD4 + and CD8+ T lymphocyte subsets in this process using monoclonal antibodies specific for the CD4 and CD8 cell surface glycoproteins. We have established that: (1) in this model system concordant neural xenograft rejection occurs within 15~311 days; however, xenograft survival can be dramatically prolonged with CD4+, but not CD8+. T lymphocyte depletion; (2) the administration of two successive courses of a high dose of anti-CD4 monoclonal antibody treatment results in indefinite concordant neural xenograft survival: (3) the mechanism by which the high dose anti-CD4 monoclonal antibody therapy appears to function involves the depletion of intrathymic CD4+ cells; (4) anti-CD4 monoclonal antibody treatment enhances discordant neural xenograft survival, to beyond 60 days m many cases. These results demonstrate that CD4+ T lymphocytes are of central importance in the immune response to both concordant and discordant neural xenoantigens. Thus the use of anti-CD4 monoclonal antibody therapy is an effective strategy to prolong significantly the survival of xenogeneic neural transplants. Furthermore this treatment caused no obvious deleterious side-effects. These findings have implications for future cross-species studies in experimental neurobiology and, possibly, in clinical neural transplantation. i(e?, M,ords: neural

transplantation,

xenograft.

CD4, CD8

The central nervous system (CNS) has been traditionally thought of as an immunologically privileged sitc.4.3’ This privilege, however, is not absolute and the rejection of both neural allografts (exchanged within species) and xenografts (exchanged across species) does occur.‘6.‘x 47Early reports suggested that the survival of a proportion of cross-species grafts of mesencephalic dopaminergic neurons could be prolonged without the need for immunosuppression.h but more recent work has demonstrated conclusively that neural xenografts elicit vigorous cellular immune responses in the CNS.‘h.‘X,‘XPresent evidence suggests that the immune response to neural xcnoantigens is more potent than that to alloantigens expressed by neural grafts. The rejection of major histocompatibility complex (MHC) mismatched lateral ventricular neural allografts can take between 120 and 150 days in certain mouse and rat strain combinations.‘“.” whereas neural xenograft rejection has commonly been found to be complete within 20 days.‘h.‘X,‘XThe importance of obtaining prolonged neural xenograft survival arises from the potential of cross-species *To whom correspondence

transplantation as a strategy for the amelioration of neurodegenerative deficits, for example those found in Parkinson’s disease. The immunosuppressive drug cyclosporin A has been shown to prolong the survival of cross-species neural grafts;’ however, long-term use of this drug appears to be required to achieve maximum benefit. The mechanisms responsible for the rejection of neural xenografts are under investigation. Neural xenografts do not appear to undergo hyperacute antibody-mediated rejection, I’.” as defined in the case of vascularized organ xenografts3’ In the latter situation prc-existing or natural antibody is often present. the targets for which appear to reside on donor vascular endothelium,” and consequently xenograft rejection is effected rapidly within 2448 h. The presence or absence of preformed antibody, usually of the IgM subclass, forms the basis for the classification of species combinations as discordant (distantly rclated-preformed antibody present) or concordant (closely related-preformed antibody absent).‘,” However, neural grafts are not vascularized ah inifio and therefore a role, if any, for preformed antibody in the immune response to discordant neural xenografts is unclear. Pre-existing or acquired humoral (antibody-mediated) mechanisms have been found to play only a minor role in the rejection of sk.in’.” and cultured pancreatic islet xenografts.44 Whether such

should be addressed.

Ahhreaicrtims: EAE,

experimental allergic encephalomyelitis; FACS. fluorescent activated cell sorter; IL-2R. interleukin-2 receptor; LCA. leukocyte common antigen; mAB, monoclonal antibody; MHC, major histocompatibility complex; PVC, piebald viral glaxo strain. 115

776

M. J. A. Wood

humoral mechanisms might be involved in the later stages of neural xenograft rejection, once these grafts are revascularized, is not known. Conversely, it is clear that an intact blood-brain barrier can confer some degree of protection from immune attack to xenografts placed in the brain.” In the mouse the rejection of skin xenografts has been shown to be absolutely dependent on the presence of CD4+ T lymphocytes,‘6 as has T lymphocyte proliferation and cytokine production in response to mouse xenoantigens in vitro.” These reports support the hypothesis that mouse T lymphocytes recognize xenoantigens indirectly as processed peptides, similar to the recognition of nominal antigens, and that the CD4+ T cell subset is crucial to this process. However, the cellular requirements for a functional immune response to neural xenoantigens in the brain are not yet clearly understood. Both macrophages, which express CD4 at low levels in rat,” and CD8+ T cells have been typically seen at sites of neural xenograft rejection in the CNS.‘6.‘8 However, the precise cellular mechanism by which neural xenoantigens are presented to the host immune system and a xenoreactive immune response generated is unknown. Several factors may contribute to defective direct immunorecognition of xenoantigens, leading to a requirement for indirect recognition of processed xeno-peptides.” It has been suggested that the accessory molecule interactions or specific cytokine production which are required for T lymphocyte activation might be defective across species. Whether such constraints apply in the context of neural xenoantigens is currently unclear. In this report in ciao studies of the cellular requirements for concordant and discordant neural xenograft rejection are described. We have used monoclonal antibodies (mAbs) against the murine CD4 and CD8 cell surface glycoproteins (expressed principally by helper and cytotoxic T lymphocyte populations, respectively, in this species), selectively to deplete these T cell populations. In this way the specific roles of these T cell subsets have been studied, and the immunosuppressive potential of these mAbs has been investigated. EXPERIMENTAL

Animals

PROCEDURES

and donor tissue

Eight- to 16-week-old C3H/He (C3H strain MHC haplotype H-2k) strain male mice were obtained from the Human Anatomy Department animal research facility and used as recipients in all neural xenograft experiments Donor neuronal tissue consisted of either embryonic day 19 PVC (MHC haplotype RTI’) rat neocortex (for concordant neural xenografts), obtained from the Human Anatomy animal unit, or human fetal forebrain of seven to nine week gestational age (for discordant neural xenografts). The latter tissue was obtained by Dr D. J. Clarke from standard abortions, and in accordance with the ethical guidelines of the Oxford Area Health Authority, U.K. PVG strain adult rats (MHC haplotype RTI’), between eight and 12 weeks of age, were used as recipients for syngeneic control grafts in the concordant neural xenograft experiments.

et crl.

T I.vmphocyle subset depletions mAbs used in the T lymphocyte subset depletions included anti-CD4 (YTA 3.1 and -YTS 191.1, both isotype IeG2b) and anti-CD8 (YTS 169.4. isotvoe IeG2b).‘5.40 These hybridomas were the kind gift 'Professor Herman Waldmann (Department of Pathology, Cambridge). mAb activity and specificity were tested in immunofluorescence assays. The hybridoma cell lines were grown as ascites in three-month-old DA x LOU F, hybrid rats. The IgG fractions were obtained by diethylaminoethyl ion exchange chromatography and purity was determined by sodium dodecyl sulphate polyacrylamide gel electrophoresis analysis. mAbs were dialysed against phosphate-buffered saline and administered at an intraperitoneal dose of 4 pg/g/day from day I (the day before transplantation) to day 9 after transplantation, In later experiments five-fold higher doses (20 pg/g/day) of YTA 3.1 or YTS 169.4 mAbs were administered over the same time course. In some later experiments a second course of mAb therapy was included in the protocol. In these cases 20pg/g/day was given by intraperitoneal injection For a further five days from day 30 to day 34 after transplantation.

of

Transpianrat ion surgery Donor tissue consisted of I-2 mm cubes of neocortical rat tissue (for concordant neural xenografts) or of seven- to nine-week-old human fetal forebrain (for discordant neural xenografts). The tissue was placed into sterile tissue culture medium cooled to 04°C on ice, and then taken up in a stainless steel cannula approximately 3 ~1 in volume. Recipient mice were anaesthetized with thiazine hydrochloride (Rompun, Bayer Ltd, Suffolk, U.K.) and ketamine hydrochloride (Vetalar, Parke Davis Co.. Gwent, U.K.), placed in a stereotaxic frame and the transplant tissue placed in the neostriatum (an intraparenchymal brain site) at the following coordinates; A = 1 mm; L = 2.5 mm; V = 4 mm. (A, anterior from bregma; L, lateral from bregma; V. ventral from dura). Monoclonal

antibody

hiotinylalion

Hybridoma cell lines MRC OX-27 and MRC OX-7 were grown as ascites in three-month-old Balb/c x DBA/Z F, hybrid mice and purified as detailed above. For biotinylation, mAbs were dialysed against 0.1 M sodium hydrogen carbonate buffer (pH 8.6) and then mixed with Biotin succinamide ester (Vector Laboratories, Peterborough, U.K.) dissolved in dimethylsulphoxide to a concentration of I .O mg/ml. The approximate molar ratio of proteinibiotin was 1: 10 (12Opg biotin/l mg protein). After 4 h incubation antibodies were dialysed against phosphate-buffered saline and stored at -20°C. (The MRC OX cell lines were kind gifts of Professor A. F. Williams and Dr D. W. Mason, MRC Cellular Immunology Unit. Oxford.) Immunocytochemistry Groups of transplant recipient animals were sacrificed at IO, 15, 30, 60, 90 or 120 days following surgery. Animals were flushed with intracardiac heparinized saline, brains were removed and quick frozen in OCT embedding compound (Miles, IN, U.S.A.), and 1Opm cryostat sections were prepared for immunostaining by the immunoperoxidase technique of Barclay.> Serial sections were sampled (usually six groups of nine sections each) from throughout the transplant site (anterior, central and posterior neostriatum). Secondary antibody used was a horseradish peroxidase conjugated rabbit anti-rat polyclonal immunoglobulin (Dako, High Wycombe, U.K.), except in the case of the two biotinylated primary monoclonal antibodies where an avidinhorseradish peroxidase conjugate (Vector Laboratories, U.K.) was applied at the second stage. Chromogen development was with diaminobenzidine. Counterstaining was performed with 0.5% Toluidine Blue (BDH, U.K.).

Immune Monoclonul

antibodies

response

The mechanism of action of the high dose YTA 3. I mAb regimen was investigated by fluorescent activated cell sorter (FACS) analysis of various lymphoid compartments. C3H/He mice were either untreated (naive), or treated with the intraperitoneal administration of low (4 wg/g,‘day) or high dose (20 pgg/g,/day) YTA 3.1 mAb therapy for four or IO davs. Mice were killed (on day 5 or day 1I in each of the fair and 10 day treatment pr&ocols, respectively), and splenic. mesenteric lymph node and thymic leukocyte cell suspensions isolated on a histopaque 1083 (Sigma, Dorset. U.K.) gradient. Primary monoclonal antibodies used were: unlabelled (negative control); TIB 122 (LCA, CD45): YTS 169.4 (CD8) and YTS 19 I. 1 (CD4); and these were followed by fluoroscein isothiocyanatc-conjugated goat anti-rat IgG (Dako. U.K.), before analysis on a Cytofluorograph 50L (Ortho Diagnostics Systems. Westwood, MA. U.S.A.).

analysis

717

xenografts RESULTS

,for immunocytochemistr>

MRC OX-27 recognizes a polymorphic determinant of the rat MHC class I antigen found in the RTI’ haplotype;z’ MRC OX-7 is a mAb against the Thy-l. 1 antigen which is found on mature rat CNS neurons, and on thymocytes in the rat and in the mouse AKR strain;29 30Hl2 is a mAb against the Thy-l.2 antigen which is expressed on mature mouse CNS neurons and murine T lymphocytes. except those of the mouse AKR strain;” B7 anti-human Thy-l antigen. exnressed on human CNS neurons45 (kind gift of Dr Rogir hoiris, National Institute of Medical ‘Research, Mill Hill, London); Ml.42 anti-mouse MHC class I$ TIB 122. against mouse leukocyte common antigen (LCA), CD45? KT3 anti-mouse T cell receptor/CD3 complex? YTS 191.1 anti-CD4 and YTS 169.4 anti-CD8,” and TIB 222 an anti-mouse IL-2R mAb.56 Non-specific primary monoclonal antibody (MRC 0X-21. anti-human C3b inactivator). and secondary antibody alone were used as negative controls. (The MRC OX mAbs were kind gifts of Professor A. F. %‘illiams and Dr D. W. Mason, M
Histological

to neural

of‘ neural

xenogrqjkr

Histological analysis was based on sampling at least six series of nine serial sections from throughout the anteroposterior extent of each neural xenograft, as detailed above. Sections from each xenograft were subsequently analysed, and the immune response in each case classified as follows. (1) Xenograft survival; donor Thy-l. 1 expression. Absent donor MHC class I expression and T cell/leukocyte infiltration. (2) Xenograft survival: donor Thy-l.1 expression, donor MHC class I expression. Absent T cell/leukocyte infiltration. (3) Xenograft survival with histological evidence of rejection underway; donor Thy-I.1 expression and T lymphocyte/leukocyte infiltration. with or without donor MHC class I expression. (4) Xenograft rejection; absent donor Thy-l. 1 expression. T lymphocyte/leukocyte infiltration with donor MHC class I expression. (5) Complete xenograft rejection; absent donor Thy-l .I and MHC class I expression. T lymphocyte/leukocyte infiltration. Thus, xenograft survival or rejection was determined in each case by immunostaining with MRC OX-7 mAb against Thy- I, I, a donor-specific rat neuronal marker in the case of concordant neural xenografts, or B7 a mAb against the product of a human donor neuron-specific Thy-l allele, in the case of discordant neural xenografts. In cases where any donor neuronal Thy- I. 1 expression was detected (histological grades l-3), the graft status was classified as xenograft survival. Conversely, where such Thy-l. I immunostaining was absent (histological grades 4 and 5). these xenografts were classified as rejected.

Methodological

The

method

considerations

used

for

the

histological

analysis

of

neural xenografts is described above. The histological classification of xenografts into those surviving (histological grades 1-3) and those rejecting (histological grades 4 and 5) is based upon the presence or absence of immunocytochemically detectable neuronal ThyI 1 expression. This has proven a useful index of graft status in other studies.‘2.54.55 In the present study the use, in particular, of anti-CD4 mAb induced immunosuppression was found to enhance dramatically neural xenograft survival, and therefore the results are more usefully presented in terms of xenograft survival and rejection (based on the graded histological classification), rather than as detailed histological grades (detailed data not shown). It should be noted that this method of classification produces a conservative estimate of xenograft rejection, since a number of xenografts classified as histological grade 3. containing only minimally detectable Thy- I. I expression, were recorded as surviving. It should also bc noted that since the classification of xenograft status is based upon detectable Thy-l .l expression, the designation of xenografts as surviving or rejected refers only to the neuronal (i.e. Thy-l.1 expressing) component of the xcnograft. This approach is of value since it represents a measure of the functional (in cases where graft neuronal function is required) or neuronal longevity of a given graft. Concordant

neural

xenogrqfts

The rat mAbs used for murine T lymphocyte subset depletion in this study have been characterized previously, and > 85% specific T cell subset depletion is achievable when the mAbs are administered m riro at low dose (e.g. 4pg/g/day).” These results were confirmed in the present experiments (Table 3). Greater than 90% CD4 and CD8 T cell depletion was routinely achieved in the lymph node and splenic compartments (data for CD8 not shown). The results shown in Table I, and graphically represented in Fig. I, are from initial transplantation experiments in which low doses of mAbs were used for T lymphocyte subset depletion. Syngeneic control grafts survived indefinitely in all cases. Concordant rat neural xenografts in untreated control recipient mice were rapidly rejected, in most cases (I3 out of 16) within 15 days. This result was also found in anti-CD8 mAb (YTS 169.4) treated animals. Importantly, this anti-CD8 mAb served as an isotype-matched control for the anti-CD4 antibodies (and vice versa) in all experiments. In contrast, treatment with either anti-CD4 mAb (YTA 3.1 or YTS 191.1) at the low dose of 4 pg/g/day for nine days after transplantation produced prolonged neural xenograft survival. At 30 days after transplantation 100% graft survival was found (i.e. all 23 xenografts had immunocytochemitally detectable Thy- 1. I expression). However, xeno-

778 Table

M. J. A. Wood

1.The survival of concordant

rat neural xenografts

in

the mouse Experiment

Day

PVG rat to C3H mouse No treatment

10 1s 30 10 15 30 10 30 60 10 30 60 30 60

YTS 169.4 (anti-CDS) 4ng/g/day Low dose YTS 191.1 (anti-CD4) Low dose YTA 3.1 (anti-CD4) Low dose YTS 191.1 + YTA 3.1 Low dose

Number

% Survival

5 11 11 6 12 10 6 13 9 5 10 14 5 6

20 18 0 33 33 0 100 100 33 100 100 57 100 50

C3H/He mice received PVG rat neocortical neural xenografts into the neostriatum on day 0. Mice were either untreated or received various forms of low dose (4 pggig/ day) anti-CD4 or anti-CD8 mAb therapy for 10 days beginning on day - 1 (the day before transplantation). In animals receiving a combination of anti-CD4 mAbs (YTA 3.1 + YTS 191.1) the total dose was 4 pg/g/day. Grafts were examined on days 15,30 or 60 (column two) and the number of grafts examined is shown in column 3. Xenograft status, i.e. graft survival or rejection, was determined by the presence or absence of immunostaining for donor neuron Thy-l.1 antigen expression with MRC OX-7 mAb (here xenograft survival is expressed as a percentage).

survival declined to 57 and 33% in the YTA 3.1 and YTS 191 .l groups, respectively, at 60 days. Treatment with both anti-CD4 mAbs together was tested (at the same total dose of 4pgg/g/day), since they were raised against non-overlapping determinants on the CD4 molecule and theoretically might have additive effects.40 However, the result in this case was no better than treatment with the YTA 3.1 mAb graft

100

-

80

-

60

-

et al.

Typical immunohistological findings for control concordant neural xenografts are shown in Fig. 2. Intense cellular infiltrates were seen throughout these untreated xenografts by day 10 in all cases. The infiltrate at day 15 comprised numerous LCA+ (shown in Fig. 2D) and T cell receptor positive cells. Smaller numbers of CD8+, CD4+ and IL-2R+ cells were seen. Donor rat neuron-specific Thy-l. 1 expression (MRC 0X-7) was undetectable in nine of 11 cases at day 15 (an example of which is shown in Fig. 2B), and minimally detectable in the remaining two cases (which were histologically classified as grade 3, and therefore surviving). The strongly induced expression of donor-specific MHC class I antigen (MRC 0X-27) was a consistent finding at 10 and 15 days in all control xenografts, i.e. these were all therefore classified as histological grades 24 (shown in Fig. 2A). Expression of this antigen was typically confined to discrete patches within heavily infiltrated xenografts. In contrast, the expression of host MHC class I antigens was heavy and widespread at all timepoints examined (shown in Fig. 2E). No donor Thy-l .l or donor MHC class I expression was detected at day 30 in any case, i.e. all 1 I grafts examined at this timepoint were classified as histological grade 5. In contrast, the histological findings in anti-CD4 mAb (YTS 191.1 or YTA 3.1) treated recipients at 10 and 30 days demonstrated evidence of good xenograft survival (histological grade 1 or 2) in all cases. These grafts were characterized by high levels of uniform donor Thy- 1.1 antigen expression throughout the xenograft tissue, in the absence of donorspecific MHC class I antigen expression and T lymphocyte

infiltration

in most

cases.

However,

;

-

Untreated Anti-CD0

Anti-CD4 YTS

191.1

Anti-CD4 YTA

15

30 Time

Fig. 1. Representation

of data

in Table

at

60 the histological evidence ranged from good graft survival and early indications of rejection (histoday

3.1

60

(days)

1 for concordant

neural

xenografts

in the neostriatum.

immune

response

to neural

xenografts

Fig. 2. Untreated concordant PVG rat neural xenograft in the mouse. Coronal serial sections through the striatum 15 days after xenotransplantation. (A) MRC 0X-27. Discrete patches of donor-specific MHC class I expression are detectable (arrows), confined to the inferior region of the graft. (B) MRC 0X-7. The graft site (g) in the striatum is visible, but donor-specific neuronal Thy-l 1 staining is absent, indicative of graft rejection. (C) 30H 12. Host-specific Thy- 1.2 expression (weak expression in this case) is noted in the host brain (h), and on T lymphocytes within the graft (which express the antigen more strongly. (D) TIB 122. Dense infiltration of LCA+ cells throughout the graft and expression upon cells in the adjacent host brain. (E) Ml .42. Host-specific MHC class I expression within the graft and at perivascular sites in the surrounding host brain (arrow). Scale bar = 250 pm.

780

M. J. A. Wood

logical grade 3) in the minority of instances (14 out of 29 cases), to late (grades 4 and 5) phases of xenograft rejection in most cases (15 out of 29). In the latter cases leukocyte infiltration was found to be heavy, and the presence of CD4+, CDS+ and IL-2R+ cells was detected. Anti-CD4 rnonoclonal antibody treatment results in indefinite concordant neural xenograft survival Further experiments, in which mAb treatment protocols were modified, attempted to prolong concordant neural xenograft survival reliably beyond 60 days. The results of these are shown in Table 2, and are graphically represented in Fig. 3. Importantly, treatment with a single course of the high dose YTS 169.4 anti-CD8 mAb alone, as an isotype matched control mAb, had no beneficial effects on graft survival compared with untreated animals. A single course of a five-fold increased dose (20 pgg/g/day) of YTA 3.1, two courses of low dose (4 pg/g/day) YTA 3.1 mAb or the original singie low dose course of YTA3.1 in combination with anti-CD8 mAb (4pgg/g/day), yielded 83, 100 and 100% concordant xenograft survival, respectively, at 60 days after transplantation. However, in these three experimental groups the proportion of xenografts surviving at 90 days declined to 50, 57 and 56%, respectively. The typical histological features of these surviving xenografts (most of which were classified histological grade 2 or 3) at 90 days are shown in Fig. 4, from an animal receiving a combination of anti-CD4 and anti-CD8 therapy. Most commonly donor MHC class I antigen expression was detectable at low levels; however, expression was entirely absent in a few cases (an example of which is shown in Fig. 4A). This contrasted sharply with the distribution of host MHC class I expression, which was found at moderate

100

et al

Table

2. Effect of modifications antibody treatment upon

Experiment YTS High YTA Low YTA High YTA Low YTA High

to anti-CD4 monoclonal xenograft survival

Day

169.4 (antiCD8) dose, one course 3.1 (antiCD4) dose, two courses 3.1 (anti-CD4) dose, one course 3.1 + YTS 169.4 dose, one course 3.1 (anti-CD4) dose, two courses

Number

15 30 60 90 60 90 60 90 60 90 120

% Survival

5 5 12 I 12 6 10 9 I 8 5

20 0 100 57 83 50 100 56 100 100 100

In these experiments modifications to the mAb treatment protocols were undertaken in attempts to prolong concordant neural xenograft survival. C3H/He mice received PVG rat neocortical grafts into the striatum on day 0. Experimental groups of animals received either one (day - 1 to day 9) or two (day - 1 to day 9 and day 30 to day 34) courses of low (4pgg/g/day) or high (20pg/g/day) dose anti-CD4 (YTA 3.1) or anti-CD8 (YTS 169.4) mAb therapy, as detailed above. Recipient animals were killed on either days 15, 30, 60, 90 or 120 and xenograft survival (expressed as a percentage) was determined by positive immunostaining for donor neuron-specific Thy-l. 1 expression.

levels

in perivascular

locations,

and

on

infiltrating

cells predominantly located at the graft-host border (shown in Fig. 4E). Despite a significant leukocytic infiltrate (shown in Fig. 4D), which comprised many T cells (shown in Fig. 4C), high levels of donor Thy- 1.l expression were commonly seen (shown in Fig. 4B), i.e. most of these grafts (10 out of 12) were classified as histological grade 3. Infiltrating leukocytes and T cells were commonly located in prominent perivascular cuffs and at the graft-host interface, whilst fewer of these cells were detected within the xenograft tissue (shown in Fig. 4C,D). Small numbers

-

80 -

60

-

m

Anti-CD8

m

Anti-CD4 LOW

B

I 40

HIGH

m

m

2

X

1

Anti-CD4 Anti-CD8

+

Anti-CD4 HIGH

20

X

Anti-CD4

X

2

-

O--! 30

60 Time

Fig. 3. Representation

(days)

of data in Table 2, in which modifications protocols were tested.

to the monoclonal

antibody

treatment

Immune

response

to neural

xenografts

Fig. 4. Concordant PVC rat neural xenograft in the mouse striatum. Serial coronal sections 90 days following a single course of combination anti-CD4 and anti-CD8 monoclonal antibody therapy. (A) MRC 0X-27. The xenograft (g) is shown, in this case with minimal donor-specific MHC class I antigen expression. (B) MRC 0X-7. High levels of donor Thy 1.1 expression throughout much of the xenograft, indicative of neuronal survival. (C) 30Hl2. Uniform Thy 1.2 expression is seen in the host striatum (h), and upon scattered T cells in the xenograft, most prominent at the graft-host borders (arrow). (D) TIB 122. LCA+ cells are detectable within the graft but are found predominantly at the graft-host interface, and at perivascular sites (arrow). (E) MI .42. Host-specific MHC class I expression is detectable on vascular structures at the graft border (arrow). with weaker expression noted on scattered cells within the graft or adjacent host brain. Scale bar = 250 lm.

782

M. J. A. Wood et al

.

.

Immune Table

3. Depletion

response

to neural

783

xenografts

of T cell subsets following anti-CD4 treatment in uiuo

monoclonal

antibody

CD4+

% Positive cells CD8+

LCA+

Spleen Naive YTA 3.1 (IOOpg x 4) YTA 3.1 (5OOpg x 4)

19.1 + 2.3 2.4 + 0.4 1.9io.4

5.9 If: 0.7 5.8 k 0.6 5.6 + 0.8

69.9 + 3.9 55.6 + 1.8 59.5 * 2.8

Lymph node Naive YTA 3.1 (IOOpg x 4) YTA 3.1 (5OOpg x 4)

51.1 + 5.8 3.8 f 0.9 2.3 + 0.7

18.4 k 2.9 33.9 * 4.1 34.9 * 5.3

86.2 I 8. I 60.8 i 7.7 60.2 I 5.9

Thymus Naive YTA 3.1 (IOOpg x 4) YTA 3.1 (5OOpg x 4)

43.4 k 5.2 15.1 +2.6 2.1 +0.5

17.1 & 2.9 33.1 * 3.3 18.2 * 3.0

65.2 i 6. I 59.5 & 4. I 54.5 i 6.0

Experiment Lymphoid tissue

C3H/He male mice (four to eight per group) were either untreated or treated for four or IO days with low (4 pg/g/day) or high (20 pg/g/day) dose anti-CD4 (YTA 3. I) mAb therapy beginning on day 0. (The average mouse weight was 25 g.) On day 5 or I I (in the case of treatment for four and IO days, respectively) all mice were killed and leukocytic cell suspensions of spleen, mesenteric lymph nodes and thymus were prepared for FACS analysis. Data for the four-day experiment are shown above. The percentage of cells of CD4+, CD8+ and LCA+ phenotypes is expressed as the mean + S.E.M. The degree of CD4+ cell depletion was determined by comparison with the naive animals in each group. Greater than 90% CD4+ cell depletion was noted in the thymic compartment after the high dose YTA 3.1 treatment. sienificantlv , ., greater (P < 0.001) than the low dose group (unpaired Student’s r-test).

of CD4+ and CD8+ cells were detected within the infiltrate. These observations are suggestive of a chronic immune process. In contrast with these results, two courses of the higher dose (20pg/g/day) YTA 3.1 anti-CD4 mAb treatment resulted in 100% xenograft survival at 90 and 120 days (typical histological findings shown in Fig. 5). In the 120 day group, all (five out of five) neural xenografts demonstrated high levels of donor Thy- 1. I antigen expression-indicative of excellent neuronal survival (shown in Fig. SA). These xenografts typically showed minimal evidence of MHC class I up-regulation of either donor or host origin (shown in Fig. SB,D). Furthermore, the infiltration of leukocytes or T cells (shown in Fig. 5C) into perivascular locations or to the graft-host interface was a very uncommon feature of these xenografts, i.e. most (four out of five) grafts were classified as histological grade 1 or occasionally grade 2. Mechanism monoclonal

of action qf antibody therapy

high

The question of the mechanism high dose (20 @g/g/day) anti-CD4

dose

anti-CD4

body therapy was addressed by examining the degree of CD4+ T cell depletion in various lymphoid compartments after high and low dose anti-CD4 mAb treatment, respectively. Two series of experiments were undertaken, in which mAbs were administered for four and IO days, respectively. The results of this FACS analysis were similar at both four and IO days, and therefore the day 4 results alone are given in Table 3. No differences between high and low dose anti-CD4 therapy were noted in the magnitude of CD4+ T lymphocyte depletion in the mesenteric lymph node and splenic compartments. Greater than 90% CD4+ depletion was found in all cases, as has been previously reported.15 However, the striking finding in these experiments was the profound and significantly greater (P < 0.001) CD4+ T cell depletion in the intrathymic lymphoid compartment which was observed with the five-fold higher dose of YTA 3.1 mAb treatment compared with the lower dose. Discordant

of action of the monoclonal anti-

neural xenogrqfts

To examine whether the findings obtained above in a concordant neural xenograft situation were also

Fig. 5. Concordant PVC rat xenograft in the mouse striatum. Serial coronal sections 120 days after transplantation. following two courses of high dose anti-CD4 (YTA 3.1) mAb treatment. (A) MRC 0X-7. High levels of uniform donor neuron-specific Thy-l .I expression are detectable throughout the graft, indicative of excellent xenograft survival. (B) MRC 0X-27. The xenograft (g) is shown with no detectable evidence of donor-specific MHC class I induction. The anterior commisure (ac) is indicated as a landmark. (C) 30HI2. The graft (g) is noted within the surrounding Thy-l.2 positive host striatum. No evidence of T lymphocyte infiltration is detectable. (D) M1.42. No evidence of elevated host-specific MHC class I expression is detectable within or surrounding the xenograft (g). Scale bar = 250 pm.

M. J. A. Wood

784 Table 4. Discordant

human

neural

Experiment Human to C3H mouse Untreated

et al.

xenograft

survival

Day

Number

15 30 60 I5 30 60 15 30 60 90 60 90

Human to C3H mouse YTS 169.4 (anti-CDS) high dose Human to C3H mouse YTA 3.1 (anti-CD4) high dose

PVC rat to C3H mouse YTA 3.1 high dose

in the mouse

5 6 s 4 4 5 4 5 8 7 5 8

% Survival 20 0 0 25 0 0 100 100 63 17 80 50

The survival of discordant human into C3H/He mouse neural xenografts was compared directly with concordant PVG rat neural xenografts. C3H/He mice received either human fetal forebrain or PVG rat neocortex neural xenografts on day 0. The former groups were untreated or treated with a single course (day - 1 to day 9) of high dose (20 pg/g/day) anti-CD4 (YTA 3.1) or anti-CD8 (YTS 169.4) mAb therapy. Recipients of rat xenografts were treated with a single course of high dose YTA 3.1 therapy. Animals were killed on either days 15,30, 60 or 90, and xenograft survival (expressed as a percentage) was determined by positive immunostaining for donor human or rat neuronspecific Thy-l antigen expression, using the B7 and MRC OX-7 mAbs, respectively.

true of discordant neural xenograft species combinations, experiments were undertaken to compare directly the fate of discordant human neural xenografts with concordant rat neural xenografts in the mouse. Since the supply of human donor tissue for these experiments was limited not all mAb treatment combinations were examined. The data from these experiments are shown in Table 4, and are graphically represented in Fig. 6. Untreated or anti-CD8 mAb treated human neural xenografts were rapidly rejected by days 15-30 in all cases (i.e. 17 out of 19 xenografts were classified as histological grade 4 or 5). This is similar to findings already presented for concordant neural xenografts. In contrast, treatment with a single course of the high dose (20 pg/g/day)

YTA 3.1 mAb prolonged human neural xenograft survival to 60 days in 63% of recipient animals (i.e. five out of eight animals had surviving xenografts). The typical immunohistological features of these surviving xenografts at 60 days are shown in Fig. 7. Vigorous leukocytic infiltration (comprising T cell receptor positive, CD4+, CD8+ and IL-2R+ cells) was detected not only in perivascular locations and at the graft-host border, but also within the xenograft tissue (Fig. 7B,C). These zones of leukocyte infiltration were found to correlate closely with regions of strongest host MHC class I up-regulation (Fig. 7D). Low to moderate levels of donor neuronal Thy-l expression were detected, and this was accompanied by evidence of areas of graft destruction

/

Untreated Discordant

@%%

Anti-CD4 Discordant

m

Anti-CD4 Conoordant

30

60 Time

Fig. 6. Representation

(days)

of data in Table 4 for discordant

xenografts

in the neostriatum

Immune

response

to neural

xenografts

’ .

.

Fig. 7. Discordant human neural xenograft in the mouse striatum. Serial coronal sections 60 days atter xenotransplantation, following a single 10 day course of high dose (20 pg/g/day) anti-CD4 (YTA 3.1) mAb therapy. (A) B7. Uniform human donor-specific Thy-l expression is evident at the graft apex (arrow), indicating some surviving human neural tissue. At the graft base a prominent area of graft destruction and cell infiltration is noted. (B) 30H12. The graft site (g) within the host striatum is noted. Host-specific Thy- 1.2 expression is detected on the surrounding host brain (h) and on T lymphocytes, scattered at the apex but more numerous at the graft base. (C) TIB 122. Numerous LCA+ cells are seen within the xenograft, most prominently at the graft base (arrow) where rejection is underway. (D) M1.42. Strong host-specific MHC class I expression is noted at the graft-host interface, on many cells at the graft base (arrow) and within the adjacent host brain. Scale bar = 250 pm.

786

M. J. A. Wood et al.

in many cases but also other areas of apparently healthy xenograft tissue (Fig. 7A). Thus all five surviving xenografts were classified as histological grade 3. Only 17% of discordant neural xenografts showed detectable levels of donor neuron-specific Thy-1 expression at day 90 (i.e. only one xenograft out of seven was classified as histological grade 3). In contrast, in the same series of experiments 50% of concordant rat neural xenografts (i.e. four out of eight) showed detectable Thy- 1.1 expression, indicating neuronal survival, at this time. DISCUSSION

The essential observation in these experiments is that CD8+ T lymphocytes alone appear unable to mediate the acute rejection of concordant or discordant neural xenografts following CD4+ T lymphocyte depletion. Hence the conclusion that the CD4+ T cell subset plays a crucial, dominant role in the process of neural xenograft rejection. Neural xenograft survival was determined by immunoreactivity for the Thy-l antigen, a developmentally regulated neuronal cell surface glycoprotein of the immunoglobuhn superfamily,32 and a reliable index of graft survival in many previous studies.‘2.43.54,55CD4+ T cell depletion prolonged both concordant and discordant neural xenograft survival dramatically, and in the most successful treatment protocol two courses of high dose anti-CD4 mAb therapy resulted in indefinite concordant neural xenograft survival (i.e. greater than 100 days). These results are consistent with the hypothesis that neural xenoantigens require reprocessing prior to their indirect recognition as processed peptides, for which the CD4+ T cell subset would appear to be essential. The mechanism for the immunorecognition of neural xenoantigens would thus seem to be similar to that described for skin3’ and pancreatic islet44 xenoantigens in the absolute requirement for CD4+ T cells. The precise defect preventing the direct recognition of neural xenoantigens is unclear. It is possible that critical accessory molecule interactions are dysfunctional across species, as has been suggested for the CD4/MHC class II interaction.” Alternatively, the evidence provided by Alter and Bach that primary (i.e. direct) responses to xenoantigens can be reconstituted in vitro by the addition of appropriate responder species cytokines,’ suggests that the absence of cytokines that can function across a species barrier might be an important factor preventing the direct recognition of xenoantigens. Therefore the precise nature of the cytokines (i.e. cytokine type and whether of host or donor origin) produced at a neural xenograft site could be an important determinant of graft outcome, a question which is under investigation in several laboratories at present. There would appear to be several possibilities to account for the beneficial effects upon neural xenograft survival of the higher dose compared with the

lower dose anti-CD4 mAb treatment. Importantly, neither treatment allowed total CD4+ T cell depletion. However, it is now clear that anti-CD4 mAbs may have immunosuppressive effects without the depletion of CD4+ target ceIIs,‘0.41and thus mechanisms for the functional inactivation of CD4+ T cells, as a consequence of anti-CD4 mAb therapy, must exist. Evidence suggests that modulation of the CD4 antigen occurs after anti-CD4 mAb treatment amongst those CD4+ T cells that escape depletion, a consequence of which may be T cell inactivati0n.l” Alternatively, the findings of intrathymic CD4+ T cell depletion in this report suggest an important mechanism of anti-CD4 mAb function at high doses. This observation is in disagreement with the accepted idea of a blood-thymus barrier, thought to produce an antigen free thymic environment.42 However, there is now evidence to suggest that mAbs may have access to the thymic cortex, perhaps via a transcapsular route.35 The means by which cell elimination might occur in the thymus are unclear, but effector mechanisms for mAb induced cell death in the thymus are thought to be absent. ‘I A number of reports have indicated that the CD4 molecule delivers critical signals during thymocyte development,‘s,46 and thus the presence of anti-CD4 mAbs may alter the thymic T lymphocyte selection process. With the observed effects upon thymocyte populations and the indefinite neural xenograft survival produced after treatment with high doses of anti-CD4 mAbs, the generation of specific tolerance (i.e. the specific acceptance of neural xenoantigens by the host immune system) to the neural xenoantigens is an important possible explanation for our findings. Such tolerance has been induced for neural alloantigens with the use of antiIL-2R mAbss5 and anti-CD4 mAbs (M. J. A. Wood et al., unpublished observations). Specific tolerance induction to concordant xenoantigens has recently been reported in an organ xenograft system with the use of combinations of anti-CD4 and anti-CD8 mAbs.” It will now be important to test directly for the presence of tolerance in the present concordant neural xenograft model, by the administration of second xenotransplants and the analysis of graft infiltrating T cells. Why were CD8+ T cells incapable of causing neural xenograft rejection alone? It has been suggested that the rejection of skin xenografts may be independent of CD8+ T cells, relying instead entirely upon CD4+ T cells or other mechanisms.2.‘h However, the observation made in the present and previous studies, ” that CD8+ cells are present at sites of neural xenograft rejection, suggests that they play an important role in the immune response to both concordant and discordant neural xenografts. Since the generation of CD8+ T cell responses is often dependent on CD4+ cells, it is possible that the absence of xenograft rejection in the CDCdepleted animals was because a CD8+ cytotoxic response failed to develop. An interesting observation in the

787

Immune response to neural xenografts present

study

graft survival

was the prolonged following

concordant

both CD4+ and CD8

xeno-

+ T cell

depletion compared with CD4’ cell depletion alone (100 and 57% graft survival at 60 days, shown in Tables 2 and 1, respectively). This suggests that interaction between the two T cell populations does occur in the generation of a functional immune response to neural xenoantigens, and that CD8 + cytotoxic T cell generation and/or perhaps CD8+ T cell entry into brain are CD4-dependent events. Further experiments in which cytotoxic T cell activity is measured directly, and studies for example in MHC class I and II knockout mice,“,‘” will be needed to elucidate more clearly the role played by CD8+ cells in the immune response to neural xenoantigens. A further factor to have bearing on the duration of concordant neural xenograft survival in the present experiments may be the relationship between the rate of graft vascularization and rate of reconstitution of the T lymphocyte population. Earlier studies have shown that reconstitution of the T lymphocyte population reaches approximately 30% by day 30.“” and the present study has indicated that the rate of concordant neural xenograft rejection increases significantly after day 30. This formed the basis of the rationale, in the second phase experiments, to begin the second anti-CD4 mAb course at day 30. On the other hand, revascularization and restoration of the blooddbrain barrier within neural grafts have been found to be complete within 20-30 days after transplantation in most instances.’ Thus, in individual cases where the process of T cell reconstitution was more rapid than that of xenograft revascularization the outcome at 60 days may have been graft rejection. Conversely, despite the presence of donor endothelial cells in neural xenografts which are revascularized by day 30 (M. J. A. Wood and H. M. Charlton, unpublished observations), the reformation of the blood-brain barrier may confer added protection from immune attack thereafter. This suggestion is supported by evidence from experiments in which hyperosmotic disruption of the blood-brain barrier can precipitate the rejection of stable neural xenografts.” Furthermore, this relationship between reconstitution of the T cell population and reformation of the blood-brain barrier may be an important factor to account for the prolonged survival of neural xenografts as compared with skin or pancreatic islet xenografts, which are found to undergo rejection within 60 days after anti-CD4 mAb treatment.3”,‘J In the experiments where a second wave of CD4+ T cell depletion was initiated on day 30, this would appear to ensure that the integrity of the blooddbrain barrier is restored in all cases before normal levels of (potentially xenoreactive) CD4+ T lymphocytes are reconstituted. The observation that discordant human into mouse neural xenografts survived less well than rat into mouse concordant neural xenografts is consistent

with an extensive literature documenting the poor survival of discordant compared with concordant organ and cellular xenotransplants.’ Nevertheless, the survival of human neural xenografts was prolonged following a single course of high dose anti-CD4 mAb therapy, with survival being similar to that noted for discordant skin xenografts following similar treatment.” This suggests that discordant neural xenografts are resistant to antibody-mediated rejection in the early stages. since a proportion of untreated control xenografts showed some evidence of nleuronal survival at 15 days, The most likely reason for this resistance to humoral rejection is the absence of targets for antibody in neural xenografts. since these grafts are not primarily vascularized. Such targets have been shown to exist on the vascular endothelium.77~‘x However. the disparity in survival between concordant and discordant neural xenografts which was noted may result from a susceptibility of the latter group to humoral (antibody-mediated) rejection in the late stages following reconstitution of the vasculature. It is likely that this reconstituted vasculature contains xenogeneic components, and therefore endothelial targets for humoral attack may be present. Despite this suggestion, earlier work has demonstrated prolonged discordant neural xenograft survival and function in a rat model of Park.inson’s disease following anti-IL-2R mAb treatment” and cyclosporin A treatment.‘,14 It therefore seems reasonable to expect that refinements to the anti-CD4 mAb treatment protocol (such as administration of a second mAb course) may have further benefits for graft survival in the present discordant xenograft model. The data presented in this report would support the hypothesis that both concordant and discordant neural xenoantigens are reprocessed and recognized indirectly, and consequently CD4’ T cells are therefore of critical importance. In this respect neural xenoantigens appear little different to the immune system from those expressed on other tissues. There are other known cases where a CD4+ T cell-mediated immune response predominates in brain, the clearest example of which is in the disease experimental allergic encephalomyelitis (EAE). Prevention of this demyelinating disease can be achieved by depleting CD4’ T cells.” and transfer of disease to naive animals can be achieved by passive transfer of activated CD4 + cells.’ Furthermore, in EAE CD8 + cells are thought to play a regulatory role, since in their absence more fi’equent and more severe relapses occur.‘3.26 CONCLUSION

Monoclonal antibodies directed against the: CD4+ T cell subset may result in effective immunosuppression, with significant prolongation of both concordant and discordant neural xenograft survival, and the possible induction of neural xenograft tolerance. This indicates that xenogeneic immune responses in brain are predominantly of the CD4+ T cell-mediated

788 type.

M. J. A. Wood Furthermore,

no evidence

this mAb

treatment

infection

or tumour

course important species future

of

this

study.

implications neuroscience clinical

neural

of any

(for example, formation) These

side-effects

of

a predisposition emerged

results

for

future

research

and,

during

therefore work possibly

to

in

the have cross-

also,

in

et al.

Acknowledgements-We thank Cohn Beesley and Brian Archer for photographic assistance, Elaina Gadsden for excellent technical assistance and Dr D. J. Clarke for obtaining the human fetal tissue. M. J. A. Wood is an E.P.A. Ceuhalosoorin Research Fellow. This work was supported by grants from the Medical Research Council (U.K.) and the Wellcome Trust.

transplantation.

REFERENCES

1. Alter B. J. and Bach F. H. (1990) Cellular basis of the proliferative response of human T cells to mouse xenoantigens. J. exp. Med. 171, 333-338. 2. Auchincloss H. (1988) Xenogeneic transplantation. Transplantation 46, I-20. 3. Barclay A. N. (1981) The localisation of populations of lymphocytes defined by monoclonal antibodies in rat lymphoid tissue. Immunology 42, 593-600. 4. Barker C. F. and Billingham R. E. (1977) Immunologically privileged sites. Adu. Immun. 23, l-54. 5. Baron J. L., Madri J. A., Ruddle N. H., Hashim G. and Janeway C. A. (1993) Surface expression of a4 integrin by CD4 T cells is required for their entry into brain parenchyma. J. exp. Med. 177, 57768. 6. Bjorklund A., Stenevi U., Dunnett S. B. and Gage F. H. (1982) Cross-species neural grafting in a rat model of Parkinson’s disease. Nature 298, 652-654. 7. Broadwell R. D., Charlton H. M., Balin B. J. and Salcman M. (1987) Angioarchitecture of the CNS, pituitary gland, and intracerebral grafts revealed with peroxidase cytochemistry. J. camp. Neural. 260, 47-62. 8. Brundin P., Nilsson 0. G., Gage F. H. and Biorklund A. (1985) Cyclosporin A increases survival of cross-species intrastriatal grafts of embryonic dopamine-containing neurons. _!?xp.-Erai. Res. 60, 204-208. 9. Brundin P.. Widner H.. Nilsson 0. G.. Strecker R. G. and Biorklund A. (1989) Intracerebral xenoerafts of dooamine neurons: the role of immunosuppression and the blood brain barrier. Eip. Biain. Res. 75, 195-267. L 10. Burkhardt K., Charlton B. and Mandel T. E. (1989) The increase in the survival of murine H-2 mismatched cultured fetal pancreas allografts using depleting and non-depleting anti-CD4 monoclonal antibodies, and the further increase with the addition of cyclosporine. Transplantation 47, 77 l-775. 11. Calne R. Y. (1970) Organ transplantation between widely disparate species. Transplant. Proc. 2, 550-553. 12. Charlton H. M., Barclay A. N. and Williams A. F. (1983) Detection of neuronal tissue from brain grafts with anti-Thy- 1.1 antibody. Nature 305, 825-827. 13. Chen Z., Cobbold S., Metcalfe S. and Waldmann H. (1992) Tolerance in the mouse to major histocompatibility complex mis-matched heart allografts, and to rat heart xenografts, using monoclonal antibodies to CD4 and CD8. Eur. J. Immun. 22, 805-8 10. 14. Clarke D. J., Brundin P., Strecker R. E., Nilsson 0. G., Bjorklund A. and Lindvall 0. (1988) Human fetal dopamine neurons grafted in a rat model of Parkinson’s disease: ultrastructural evidence for synapse formation using tyrosine hydroxylase immunocytochemistry. Exp. Brain. Res. 73, 115-126. 15. Cobbold S. P., Jayasuriya A., Nash A., Prosper0 T. D. and Waldmann H. (1984) Therapy with monoclonal antibodies by elimination of T-cell subsets in viva. Nature 312, 5488551. 16. Finsen B., Oteruleo F. and Zimmer J. (1988) Immunocytochemical characterisation of the cellular immune response to intracerebral xenografts of brain tissue. Prog. Brain Res. 78, 261-270. 17. Finsen B. R., Pedersen E. B., Sorensen T., Hokland M. and Zimmer J. (1990) Immune reactions against intracerebral murine xenografts of fetal hippocampal tissue and cultured cortical astrocytes in the adult rat. Prog. Brain Res. 82, 111-128. 18. Freed W. J., Dymecki J., Poltorak M. and Rogers C. (1988) Intracerebral brain allografts and xenografts: studies of survival and rejection with and without systemic sensitization. Prog. Brain Res. 78, 2333241. 19. Grusby M. J., Auchincloss H., Lee R., Johnson R. S., Spencer J. P., Zijlstra M., Jaenisch R., Papaioannou V. E. and Glimcher L. H. (1993) Mice lacking major histocompatibility complex class I and class II molecules. Proc. natn. Acad. Sci. U.S.A. 90, 3913-3917. 20. Grusby M. J., Johnson R. S., Papaioannou V. E. and Glimcher L. H. (1991) Depletion of CD4+ T cells in major histocompatibihty class II-deficent mice. Science 253, 1417-1420. 21. Honey C. R., Clarke D. J., Dallman M. J. and Charlton H. M. (1990) Human neural graft function in rats treated with an anti-interleukin 2 receptor antibody. NeuroReporf 1, 2477249. 22. Jefferies W. A., Green J. R. and Williams A. F. (1985) Authentic T helper CD4 (W3/25) antigen on rat peritoneal macrophages. J. exp. Med. 162, 117-127. 23. Jiang H., Zhang S.-L. and Pernis B. (1992) Role of CD8+ T cells in murine experimental allergic encephalomyelitis. Science 256, 121331215. 24. Kennett M. (1980) Rat IgG2a and anti H-2 haplotypes. In Monoclonal Anribodies (ed. Kennett M.), pp. 1855217. Plenum Press, New York. 25. Kisielow P., Bluthmann H., Staerz U. D., Steinmetz M. and von Boehmer H. (1988) Tolerance in T cell receptor transgenic mice involves deletion of non-mature CD4+8+ thymocytes. Nature 333, 742-746. 26. Koh D.-R., Fung-Leung W.-P., Ho A., Gray D., Acha-Orbea H. and Mak T.-W. (1992) Less mortality but more relapses in experimental allergic encephalomyelitis in CD8-‘mice. Science 256, 1210-1213. 27. Ledbetter J. A. and Herzenberg L. A. (1979) Xenogeneic monoclonal antibodies to mouse lymphoid differentiation antigens. Immun. Rev. 47, 63-90. 28. Mason D. W., Charlton H. M., Jones A. J., Lavy C. B. D., Puklavec M. and Simmonds S. J. (1986) The fate of allogeneic and xenogeneic tissue transplanted to the third ventricle of rodents. Neuroscience 19, 685-694. 29. Mason D. W. and Williams A. F. (1980) The kinetics of antibody binding to membrane antigens in solution and at the cell surface. Biochem. J. 187, l-20.

Immune

response

to neural

xenografts

789

30. Mazerolles F., Auffray C. and Fischer A. (1991) Down-regulation of T cell adhesion by CD4. Human Immun. 31,40&46. 31. Medawar P. (1948) Immunity to homologous grafted skin III. The fate of skin homografts transplanted to the brain, subcutaneous tissue and to the anterior chamber of the eye. Br. J. e.wp. Parh. 29, 58-69. 32. Morris R. J., Beech J. N., Barber P. C. and Raisman G. (1985) Early stages of Purkinje cell maturation demo,nstrated by Thy-l immunohistochemistry on postnatal rat cerebellum. J. Neuroc~tol. 14, 427-452. 33. Moses R. D., Pierson R. N., Winn H. J. and Auchincloss H. (1990) Xenogeneic proliferation and lymphokine production are dependent on CD4 helper T cells and self antigen-presenting cells in the mouse. J. e-yp. Med. 172, 567-575. 34. Nicholas M. K., Ante1 J. P., Stefansson K. and Arnason B. G. W. (1987) Rejection of fetal neocortical neural transplants by H-2 incompatible mice. J. Immun. 139, 227552283. 35. Nieuwenhuis P., Stet R. J. M., Wagenaar J. P. A., Wubbena A. S., Kampinga J. and Karrenbeld A. (1988) The transcapsular route: a new way for (self-) antigens to bypass the blood-thymus barrier. Immun. Today 9, 372-375. 36. Pierson R. N., Winn H. J., Russel P. S. and Auchincloss H. (1989) Xenogeneic skin graft rejection is especially dependent on CD4 T lymphocytes. J. e-up. Med. 170, 991-996. 37. Platt J. L. and Bach F. H. (1991) Discordant xenografting: challenges and controversies. Curr. Opin. Immun. 3,735-739. 38. Platt J. L., Lindman B. J., Chen H., Spitalnik S. L. and Bach F. H. (1990) Endothelial cell antigens recognised by xenoreactive human natural antibodies. Transplanfation 50, 8 17-822. 39. Pollack 1. F., Lund R. D. and Rao K. (1990) MHC antigen expression in spontaneous and induced rejection of neural xenografts. Prog. Bruin Res. 82, 129.-140. 40. Qin S., Cobbold S., Tighe H., Benjamin R. and Waldmann H. (1987) CD4 monoclonal antibody pairs for immunosuppression and tolerance induction. Eur. J. Immun. 17, 1159-I 165. 41. Qin S., Wise M., Cobbold S. P., Leong I., Kong Y. M., Parnes J. R. and Waldmann H. (1990) Induction of tolerance in peripheral T cells with monoclonal antibodies. Eur. J. Immun. 20, 2737-2745. 42. Raviola E. and Karnovsky M. J. (1972) Evidence for a blood-thymus barrier using electron-opaque tracers J. UP. Med. 136, 466-497. 43. Reif A. E. and Allen J. M. (1966) Mouse nervous tissue iso-antigens. Nature 209, 523-529. 44. Ricordi C., Lacy P. E. and Sterbenz K. (1987) Low temperature culture of human islets or in civo treatment with anti L3T4 antibody produces a marked prolongation of islet human-to-mouse xenograft survival. Proc. natn. Acad. Sri. U.S.A. 84, 8080&8084. 45. Ritter M. A., Sauvage C. A. and Delia D. (1983) Human Thyl antigen: cell surface expression on early T and B lymphocytes. Immunology 49, 555-564. 46. Sha W. C., Nelson C. A., Newberry R. D., Kranz D. M., Russell J. H. and Loh D. Y. (1988) Positive and negative selection for antigen receptor on T cells in transgenic mice. Nafure 336, 73376. 47. Sloan D. J., Wood M. J. and Charlton H. M. (1991) The immune response to intracerebral neural grafts. Trends Neurosci. 14, 341-346. 48. Springer T., Galfre G.. Secher D. S. and Milstein C. (1978) Monoclonal xenogeneic antibodies to murine cell surface antigens: identification of novel leukocyte differentiation antigens. Eur. J. Immun. 8, 539-551. 49. Tomonari K. (1988) A rat antibody against a structure functionally related to the mouse T cell receptor/CD3 complex. Immunogenetics 28, 4555458. 50. Vignali D. A. A., Moreno J., Schiller D. and Hammerling G. J. (1992) Species-specific binding of CD4 to the 82 domain of major histocompatibility class II molecules. J. c.yp. Med. 175, 9255932. 51. Waldmann H. (1989) Manipulation of T cell responses with monoclonal antibodies. A. Rev. Immun. 7, 407-444. 52. Waldor M. K., Sriram S., Hardy R., Herzenberg L. A., Lanier L.. Lim M. and Steinman L. (1985) Reversal of experimental allergic encephalomyelitis with monoclonal antibody to a T-cell subset marker. Science 227, 4155417. 53. Winn H. J., Baldamus C. A. and Joooste S. V. (1973) Acute destruction by humoral antibody of rat skin grafted to mice. The role of complement and polymorphonuclear leukocytes. J. exp. Med. 137, 8933910. 54. Wood M. J. A., Sloan D. J.. Dallman M. J. and Charlton H. M. (1992) A monoclonal antibody to the interleukin-2 receptor enhances the survival of neural allografts: a time-course study. Neuroscience 49, 4099418. 55. Wood M. J. A., Sloan D. J., Dallman M. J. and Charlton H. M. (1993) Specific tolerance to neural allografts induced with an antibody to the interleukin 2 receptor. J. esp. Med. 177, 597-603. 56. Zubler R. H., Lowenthal J. W., Erard F., Hashimoto N.. Devos R. and MacDonald H. R. (1984) Activated B cells express receptors for. and proliferate in response to, pure interleukin-2. J. esp. Med. 160, I 170&l 183. (Accepted 4 September

1995)