Tolerance or Rejection: A Delicate Balance as Judged by Exposure

of Heart-Transplanted Rats to the Immunomodulator LinomideÒ

Z. QI,* G. TUFVESON² & H. EKBERG³

Departments of *Experimental Research and ³Vascular and Renal Diseases, Lund University, University Hospital, MalmoÈ, Sweden; ²Department

of Transplantation Surgery, Uppsala University, Academic Hospital, Uppsala, Sweden

(Received 4 March 1999; Accepted in revised form 22 June 1999)

Qi Z, Tufveson G, Ekberg H. Tolerance or Rejection: A Delicate Balance as Judged by Exposure of Heart-

Transplanted Rats to the Immunomodulator LinomideÒ. Scand J Immunol 1999;50:394±398

When applied in rodent transplant models most immunosuppressive drugs yield adequate graft protection for

as long as the drug is given, and permanent graft survival is often induced. The immunomodulator,

LinomideÒ, previously shown to stimulate T cells and prevent apoptosis, usually reduces or abolishes both

tolerance induction and the graft-protective effect of the immunosuppressive drug. By chance, we observed

that LinomideÒ alone exerted a modest but unequivocal graft-protective effect in the BN to WF strain

combination. This ®nding was analysed by simple genetic mapping of rat strains. Untreated WF recipients

kept BN grafts for a median of 8 days, whereas LinomideÒ treatment prolonged graft survival to 12.5 days

(P� 0.0001). In control groups (DA to LEW, BN to LEW, DA to WF and WF to BN), median graft survival

was 5.5±7 days irrespective of whether LinomideÒ was given. However, the BN to F1 (LEW ´ WF)

combination also manifested slightly longer graft survival in the presence of LinomideÒ. F1 (BN ´ WF) to

WF grafts survived a median of 15 days without LinomideÒ and 46 days with LinomideÒ treatment. Both in

the presence and absence of LinomideÒ, two of the control graft combinations [F1 (BN ´ DA) to WF and F1

(BN ´ WF) to BN] manifested 6±7-day graft survival. Taken together, our results suggest a delicate balance

between unresponsiveness and rejection, while a single agent (LinomideÒ) may either cause on its own long-

term survival of allografts in one setting or rejection despite optimal immunosuppression in another setting.

Dr H. Ekberg, Department of Vascular and Renal Diseases, University Hospital, S-205 02 MalmoÈ, Sweden

INTRODUCTION

Graft tolerance and rejection have traditionally been regarded as

two diametrically opposed end points of alternative immune

responses differing from each other in most of the events and

cytokine patterns involved. In contradiction, Shoskes and Wood

[1] wrote in an overview on indirect presentation of MHC

antigens, commenting on a former report by Butcher and

Howard [2], that `there may be a delicate balance between

unresponsiveness and activation that is regulated by T cells

recognizing alloantigen by the indirect pathway'. Here, we report

a unique drug that promotes such balance towards unresponsive-

ness in one rat strain combination and towards rejection in all

other contexts despite optimal immunosuppression.

For many years we have used the immunomodulator, roquini-

mex (LinomideÒ), in rodent transplant models to prevent the

induction of tolerance, which is often accomplished following

immunosuppressive treatment of only short or a medium duration,

for a review, see Tufveson et al. [3]. In the presence of LinomideÒ,

the graft-protective effects of most immunosuppressants are

either lost completely or manifestly reduced. In cardiac transplant

models, for instance, exposure to LinomideÒ has been reported to

abolish all of the immunosuppressive effect of cyclosporine A

(CsA) [4] or of prednisone [5], to reduce the effect of deox-

yspergualine in a dose-dependent manner [6] and to eliminate

almost all of the effect of tacrolimus [7], mycophenolate mofetil

[8] or le¯unomide [9] and its analogues [10]. Polyclonal anti-

body-mediated immunosuppression with antithymocyte globu-

line (ATG) is also impaired by LinomideÒ [11], however, the

effect of monoclonal anti-CD4 antibody is not [12].

Although the mode of action of LinomideÒ is incompletely

understood, our in vivo ®ndings strongly suggest its main target

to be effector cells ± probably effector T cells, as suggested by

Wanders et al. [13]. LinomideÒ itself does not accelerate cardiac

graft rejection (compared with no treatment), presumably

because graft rejection is an immunological process already

Scand. J. Immunol. 50, 394±398, 1999

q 1999 Blackwell Science Ltd

occurring at the maximum rate. In other experimental settings,

LinomideÒ has been shown to enhance natural killer (NK) cell

activity [14], delayed-type hypersensitivity and antibody produc-

tion [15], and to stimulate polyclonal T-cell activation [16].

LinomideÒ may also aggravate experimental autoimmune dis-

ease [17, 18] as well as graft-vs.-host disease after semisyngeneic

small bowel transplantation [19, 20]. In apparent contrast, it may

prevent the development of autoimmunity if given prophylacti-

cally [17, 21±24]. Recently, interleukin (IL)-2-induced T-cell

proliferation has been shown to be enhanced by LinomideÒ [25].

Taken together, most available data have suggested LinomideÒ

to be at least a powerful T-cell stimulator. In all likelihood, not

only effector T cells but also T cells with a suppressive effect are

stimulated by the drug.

The core ®ndings of LinomideÒ elimination of the effects of

various immunosuppressive drugs have been repeated at several

laboratories and in several different strain combinations with

consistent results. It was therefore highly unexpected when we

found that one group of heart-transplanted rats (BN to WF)

treated with LinomideÒ alone manifested a 75% improvement in

median graft survival, compared with the no treatment group (14

vs. 8 days; P� 0.0002) [26].

This hitherto unique ®nding prompted us to undertake further

analysis of this issue by means of simple genetic mapping, to

consider the implications of these ®ndings vis-aÁ-vis current con-

cepts of rejection, immunosuppression and tolerance induction.

MATERIALS AND METHODS

Isogenic Brown Norway (BN), Wistar Furth (WF), Dark Agouti (DA)

and Lewis (LEW) rats were obtained from Mùllegaard Breeding &

Research Centre A/S (Denmark), and breeding was set up at our

laboratory, including that of F1 hybrids. Those used in the present

study were as follows: male rats of the BN (RT1n), WF (RT1u), DA

(RT1avl), and LEW (RT1 l) strains, and the F1 hybrids BN ´ DA,

BN ´ WF and LEW ´ WF. The animals were preconditioned for at

least 1 week prior to transplantation, and were housed in standard

cages under controlled light/dark cycles, fed standard laboratory diet

and given free access to water. Transplant recipients weighed 180±240 g

and donors 100±160 g. The study design was approved by the Research

Ethics Committee of Lund University, and all procedures were per-

formed in accordance with the Good Laboratory Practice code published

by the National Board of Health and Welfare in Sweden (SoSR-GLP,

RoA no. 31/1974).

The transplants were scheduled as shown in Table 1, 6±12 transplants

being performed in LinomideÒ and untreated subgroups in each series of

Tolerance or Rejection ± A Delicate Balance 395

q 1999 Blackwell Science Ltd, Scandinavian Journal of Immunology, 50, 394±398

Table 1. Cardiac allograft survival in different rat strain combinations in LinomideÒ-treated and untreated

subgroups

Linomide Graft survival (days) Median P-value

1. BN to WF

ÿ 6 7 7.5 8 9 9.5 8.0

� 10 10 11 11.5 11.5 13.5 14 14 18 19 12.5 0.0001

2. BN to LEW

ÿ 6 6 6 6 6 6.5 7 7 8 6.0

� 7 7 7 7 7 8 7.0 NS

3. DA to WF

ÿ 6 6 6 6 6 6 6.0

� 6 6 6 6 6 7 6.0 NS

4. DA to LEW

ÿ 5 6 6 6 6.5 6.5 6.5 6.0

� 4 6 6 6 7 7 6.0 NS

5. BN to (LEW ´ WF) F1

ÿ 6 7 7 8 9 9 7.5

� 8 8 10.5 11 11.5 12.5 11.0 0.03

6. (BN ´ WF) F1 to WF

ÿ 9 11 11 12 12 15 15 16 16 17> 100> 100 15.0

� 13 13 14 15 20 23 69 93 > 100> 100> 100> 100 46.0 0.05

7. (BN ´ WF) F1 to BN

ÿ 6 7 7 7 8.5 9 7.0

� 4.5 5 6 6 6.5 7 6.0 NS

8. (BN ´ DA) F1 to WF

ÿ 6 6 6 6 6 6 6.0

� 6 6 6 6 7.5 9 6.0 NS

9. WF to BN

ÿ 6 6 7 7 7 7 7.0

� 5 5.5 5.5 5.5 6.5 7 5.5 NS

experiments. PerfadexÒ (Pharmacia & Upjohn, Lund, Sweden) was used

for cold perfusion. Hearts were transplanted with a nonsuture cuff

technique, anastomosing the graft aorta to the recipient common carotid

artery and the graft pulmonary artery to the recipient jugular vein at the

right side of the neck. The graft veins were all ligated. Total ischaemia

time never exceeded 15 min. Graft survival of cardiac allografts was

evaluated by palpation twice daily, and recorded as present until

cessation of a palpable heart beat.

LinomideÒ (Pharmacia & Upjohn, Lund, Sweden) was dissolved in

water to a concentration of 64 mg/ml and administered orally with a

feeding catheter once daily at a dose of 160 mg/kg from day 0 until the

day of rejection.

Graft survival in various groups was compared using Kaplan±Meier

survival plots and Breslow±Gehan Wilcoxon test.

RESULTS

In the BN to WF transplantation, median allograft survival was

12.5 days in the LinomideÒ-treated subgroup but only 8 days in

the untreated subgroup (P� 0.0001) (Table 1, 1). However,

LinomideÒ treatment did not yield signi®cant prolongation of

graft survival when the same donor strain was used with another

recipient strain (i.e. BN to LEW), or when the same recipient

strain was used with another donor strain (i.e. DA to WF).

LinomideÒ treatment also failed to yield any prolongation of

median graft survival when DA rats were again used as the donor

strain but with LEW rats as the recipients (Table 1, 4).

To determine whether the graft survival prolongation effect

of LinomideÒ in BN to WF transplantation might have been

due to recipient (i.e. WF) properties, BN allografts were trans-

planted to F1 (LEW ´ WF) hybrids (Table 1, 5); median graft

survival was 11 days in the LinomideÒ-treated subgroup but only

7.5 days in the untreated subgroup (P� 0.03). Thus, the graft

survival prolongation effect of LinomideÒ, seen in BN to WF

transplants, was reduced slightly (median 12.5 vs. 11, range 10±

19 vs. 8±12.5 days) by the addition of a LEW haplotype in the

recipients.

In the F1 hybrid (BN ´ WF) to WF transplant combination

(Table 1, 6; i.e. where the presence of foreign antigen in the

donor organ was reduced by 50%, compared with the BN to WF

combination), median graft survival was further increased to

46 days in the LinomideÒ-treated subgroup, one-third (4/12) of

the animals becoming tolerant. Although animals in the untreated

WF subgroup likewise were slow to reject their F1 (BN ´ WF)

grafts (compared with the rejection of BN grafts by the untreated

subgroup in the BN to WF combination), median graft survival

was only 15 days, and only two of the 12 rats became tolerant.

However, when F1 hybrid (BN ´ WF) donor hearts were trans-

planted to BN recipients instead, LinomideÒ treatment yielded

no prolongation of graft survival (Table 1, 7).

Finally, the graft survival prolongation effect of LinomideÒ

was also absent in the F1 hybrid (BN ´ DA) to WF combination

(i.e. where the foreign antigen challenge of the BN to WF

combination was altered by the addition of 50% DA), as it was

when the original combination was reversed and WF hearts were

transplanted into BN hosts (Table 1, 8 and 9).

DISCUSSION

Previous ®ndings by our group and others have shown LinomideÒ

to stimulate T cells [25] and prevent apoptosis of committed

T cells under certain conditions [27, 28]. At present, it is not

known whether these two effects are separate or interrelated.

Nevertheless, it seems attractive to explain the counteraction of

most immunosuppressive drugs by LinomideÒ as a consequence

of its ability to stimulate IL-2-induced proliferation of T cells and

its ability to block tolerance induction, by preventing apoptosis.

T-cell apoptosis may represent one of the steps outlined by the

Oxford group [29±31] resulting in tolerance.

To sum up, we have found LinomideÒ itself to prolong graft

survival in the BN to WF rat strain combination. Technically this

can be described as a de®ciency in the BN graft because

(BN ´ DA) F1-hybrid hearts are rejected at a normal rate by

WF rats. Molecules expressed in the BN graft may be very

weakly immunogenic or they might induce concomitant suppres-

sion in the speci®c recipient. The BN de®ciency can only be seen

in combination with WF rats as recipients, because LEW rats

reject BN hearts within a normal time span. To some extent these

results therefore support the notion that BN rats produce widely

accepted grafts [32, 33], however, both the choice of donor strain

and also that of the recipient strain was shown to be a determi-

nant of graft survival. As the graft survival prolongation effect of

LinomideÒ was manifest in the BN to WF combination, negli-

gible in the BN to LEW combination, and intermediate in the BN

to F1 (LEW ´ WF) combination, the WF strain would appear to

possess some property or genetic advantage, lacking in the LEW

strain and predisposing it to tolerance, the effect of which can be

enhanced by an immunostimulator (in this case LinomideÒ).

This was particularly evident when the challenge of foreign

antigens was weakened by transplanting hybrid donor organs to

rats of one of the hybrid parent strains, as in the F1 (BN ´ WF) to

WF combination, which was characterized by a median graft

survival of 46 days in the present study.

The original concept of high- and low-responder rat strains

does not explain our ®ndings. Initially, Howard and Butcher [2]

described the inability of PVG RT1.c to reject the major

histocompatibility complex (MHC) class I incompatible RT1.Aa

graft. Others later de®ned the low-responder recipient as one

having CD8� T cells unable to proliferate independently of CD4�

T-cell `help' [34, 35]. In this respect, WF and LEW are high

responders and PVG RT1.c and DA are low responders.

Recently, we have shown that PVG to DA, but not DA to PVG,

is a low-responder rat strain combination with tolerance induced

following a brief course of either OX-38 [12] or Le¯unomide

[36]. However, LinomideÒ on its own did not affect graft survival

in either of these two strain combinations. Induction of speci®c

unresponsiveness by CsA in PVG to DA allografts was demon-

strated previously by Hall et al. [37] and shown to depend on

CD4� suppressor cells with a fragile balance between suppression

and rejection depending on the absence or presence of cytokines

(i.e. IL-2) and alloantigens.

Direct inoculation of donor antigen into the thymus and a brief

396 Z. Qi et al.

q 1999 Blackwell Science Ltd, Scandinavian Journal of Immunology, 50, 394±398

course of immunosuppression may induce tolerance of cardiac

transplants in rat strain combinations such as BN to WF [38], but

has been unsuccessful in DA to WF and DA to LEW [39].

Several other strain combinations have been investigated and the

induction or absence of tolerance, unrelated to the CD8� T-cell

subset ability of independent proliferation, constitute yet another

de®nition of high and low responders, which is in parallel to data

from the present study.

Perhaps the strangest ®nding was that some animals develop

permanent survival with a drug previously shown to block the

development of tolerance (induced by cyclosporin) [13]. Thus

one drug (LinomideÒ), shown to be a powerful T-cell stimulator,

can promote tolerance induction on its own in one setting, F1

(BN ´ WF) to WF, but inhibit this event despite optimal immuno-

suppression in all other systems tested. This may well be due

to the anti-apoptotic effect described recently. If the effect of

LinomideÒ is assumed to be dual (i.e. T-cell stimulation and

prevention of apoptosis of active clones), it is reasonable to

suppose that the quality of tolerance and the tolerance induction

mechanism will differ from one situation to another, in variation

of rat strain combinations and induction modalities. In one system,

namely that of ATG-induced tolerance, the cause of tolerance

may well be suppressor cells [40]. However, this tolerance induc-

tion is severely hampered by LinomideÒ [11]. It is tempting to

interpret our present ®ndings as suggesting that suppression can

be stimulated with the immunostimulator LinomideÒ. This sug-

gests the `decision' by nature to reject or accept a graft to be a very

delicate one. Indeed, something along these lines has recently

been suggested in a quite different context by Calne [41, who

used the term `window of opportunity' to describe a phenomenon

based on immunosuppression given intermittently to allow for a

potential `suppressor' cell expansion. Naturally, this theory lacks

full experimental backing, and our observation does not consti-

tute unequivocal evidence in its support, but both our results and

the `window of opportunity' theory are consistent with the

existence of a very delicate natural balance between unrespon-

siveness and rejection.

REFERENCES

1 Shoskes DA, Wood KJ. Indirect presentation of MHC antigens in

transplantation. Immunol Today 1994;15:32±8.

2 Butcher GW, Howard JC. Genetic control of transplant rejection.

Transplantation 1982;34:161±6.

3 Tufveson G, Gannedahl G, Johnsson C, et al. New immunosuppres-

sants: testing and development in animal models and the clinic: with

special reference to DSG. Immunol Rev 1993;136:99±109.

4 Wanders A, Larsson E, Gerdin B, et al. Abolition of the effect of

cyclosporine on rat cardiac allograft rejection by the new immuno-

modulator LS-2616 (Linomide). Transplantation 1989;47:216±17.

5 Gerdin B, Wanders A, Tufveson G. Rat cardiac allografts protected

with cyclosporin A are rejected in the presence of LS-2616 (Lino-

mide). Transplant Proc 1989;21:853±5.

6 Gannedahl G, Wanders A, Carlsson M, et al. 15-Deoxyspergualin

potentiates the immunosuppressive effect of cyclosporine A in a rat

heart allograft model. Transplant Proc 1993;25:768±9.

7 Bersztel A, Wanders A, Wadstrom J, et al. Deoxyspergualin is

synergistic with cyclosporin A, but not with FK506 in a rat heart

allograft transplantation model. Transplant Proc 1995;27:3547.

8 OÈ straat Qi Z, Olausson M, et al. Mycophenolate mofetil, azathio-

prine and cyclophosphamide enhanced ef®cacy combined with

cyclosporine in rat cardiac transplantation. Scand J Immunol 1997;

45:343±8.

9 OÈ straat Qi Z, Tufveson G, et al. The effects of le¯unomide and

cyclosporin A on rejection of cardiac allografts in the rat. Scand J

Immunol 1998;47:303±9.

10 Qi Z, Ekberg H. Malononitrilamides 715 and 279 prolong rat cardiac

allograft survival, reverse ongoing acute rejection, inhibit allo-

speci®c antibody synthesis and interact positively with cyclosporin.

Scand J Immunol 1998;48:379±88.

11 Olausson M, Lukes D, Norrby J, et al. Anti-thymocyte globulin and

cyclosporin A are synergistic in an experimental transplantation

tolerance model in the rat. Scand J Immunol 1998;49:38±44.

12 Qi Z, Riesbeck K, OÈ straat OÈ , et al. Single dose anti-CD4 monoclonal

antibody for induction of tolerance to cardiac allograft in high- and

low-responder rat strain combinations. Transplant Immunol 1997;

5:204±11.

13 Wanders A, Vogt P, Karlsson Parra A, et al. Evidence that LS-2616

(linomide) causes acute rejection of rat allografts protected by

cyclosporine but not of long-term surviving allografts. Transplanta-

tion 1991;52:234±8.

14 Kalland T, Alm G, Stalhandshe T. Augmentation of mouse natural

killer cell activity by LS 2616, a new immunomodulator. J Immunol

1985;134:3956±61.

15 Spetz Hagberg AL, Larsson Sciard EL. Decreased levels of patho-

genic IgG anti-DNA antibodies in autoimmune mice after linomide

treatment. Res Immunol 1989;140:517±25.

16 Larsson EL, Joki A, Stalhandske T. Mechanism of action of the new

immunomodulator LS-2616 on T cell responses. Int J Immunophar-

macol 1987;9:425±31.

17 Kleinau S, Larsson P, BjoÈrk J, et al. Linomide, a new immuno-

modulatory drug, shows different effects on homologous versus

heterologous collagen-induced arthritis in rats. Clin Exp Immunol

1989;78:138±42.

18 BjoÈrk J, Kleinau S. Paradoxical effects of LS-2616 (Linomide)

treatment in the type II collagen arthritis model in mice. Agents

Actions 1989;27:319±21.

19 Johnsson C, Tufveson G. MC 1288 ± a vitamin D analogue with

immunosuppressive effects on heart and small bowel grafts. Trans-

plant Int 1994;7:392±7.

20 Johnsson C, Tufveson G. The immunomodulator LS-2616 (lino-

mide) more effectively than sensitization of the donor enhances

graft-versus-host reaction after small bowel transplantation. Trans-

plant Immunol 1994;2:56±60.

21 Tarkowski A, Gunnarsson K, Stalhandske T. Effects of LS-2616

administration upon the autoimmune disease of (NZB ´ NZW) F1

hybrid mice. Immunology 1986;59:589±94.

22 Spetz Hagberg AL, Goldschmidt TJ, Stalhandske T, et al. Ameliora-

tion of intrathymic T cell development and peripheral T cell reac-

tivities in autoimmune mice undergoing therapy with a novel

immunomodulator. Int Immunol 1990;2:645±50.

23 Karussis DM, Lehmann D, Slavin S, et al. Treatment of chronic-

relapsing experimental autoimmune encephalomyelitis with the

synthetic immunomodulator linomide (quinoline-3-carboxamide).

Proc Natl Acad Sci USA 1993;90:6400±4.

24 Gross DJ, Sidi H, Weiss L, et al. Prevention of diabetes mellitus in

Tolerance or Rejection ± A Delicate Balance 397

q 1999 Blackwell Science Ltd, Scandinavian Journal of Immunology, 50, 394±398

non-obese diabetic mice by Linomide, a novel immunomodulating

drug. Diabetologia 1994;37:1195±201.

25 Condiotti R, Slavin S, Barak V, et al. The novel immunomodulator,

Linomide, stimulates interleukin-2-induced human natural killer

(NK) cell and PHA-stimulated T cell proliferation from normal

donors. Leuk Res 1996;20:57±63.

26 OÈ straat OÈ , Qi Z, Gannedahl G, et al. Moderate additive immuno-

suppressive effect of thalidomide combined with cyclosporin A in rat

cardiac transplantation. Transplant Immunol 1996;4:241±6.

27 Gonzalo JA, Gonzalez Garcia A, Kalland T, et al. Linomide inhibits

programmed cell death of peripheral T cells in vivo. Eur J Immunol

1994;24:48±52.

28 Zamzami N, Marchetti P, Castedo M, et al. Reduction in mitochon-

drial potential constitutes an early irreversible step of programmed

lymphocyte death in vivo. J Exp Med 1995;181:1661±72.

29 Dallman MJ, Wood KJ, Hamano K, et al. Cytokines and peripheral

tolerance to alloantigen. Immunol Rev 1993;133:5±18.

30 Darby CR, Bushell A, Morris PJ, et al. Nondepleting anti-CD4

antibodies in transplantation. Evidence that modulation is far less

effective than prolonged CD4 blockade. Transplantation 1994;57:

1419±26.

31 Bushell A, Morris PJ, Wood KJ. Induction of operational tolerance

by random blood transfusion combined with anti-CD4 antibody

therapy. A protocol with signi®cant clinical potential. Transplanta-

tion 1994;58:133±9.

32 Kamada N, Davies HS, Wight D, et al. Liver transplantation in the

rat. Biochemical and histological evidence of complete tolerance

induction in non-rejector strains. Transplantation 1983;35:304±11.

33 Murase N, Starzl TE, Tanabe M, et al. Variable chimerism, graft-

versus-host disease, and tolerance after different kinds of cell and

whole organ transplantation from Lewis to brown Norway rats.

Transplantation 1995;60:158±71.

34 Hall BM. Cells mediating allograft rejection. Transplantation

1991;51:1141±51.

35 Ilano AL, McConnell MV, Gurley KE, et al. Cellular basis of

allograft rejection in vivo. V. Examination of the mechanisms

responsible for the differing ef®cacy of monoclonal antibody to

CD4� T cell subsets in low- and high-responder rat strains.

J Immunol 1989;143:2828±36.

36 OÈ straat OÈ , Qi Z, Tufveson G, et al. The effects of le¯unomide and

cyclosporin A on rejection of cardiac allografts in the rat. Scand J

Immunol 1998;47:236±42.

37 Hall BM, Gurley KE, Pearce NW, et al. Speci®c unresponsiveness in

rats with prolonged cardiac allograft survival after treatment with

cyclosporine. II. Sequential changes in alloreactivity of T cell

subsets. Transplantation 1989;47:1030±3.

38 Hamashima T, Stepkowski SM, Smith S, et al. Induction of trans-

plantation tolerance by a single intrathymic injection of 3M KCl-

extracted donor histocompatibility antigens combined with two

doses of anti-rat alpha/beta-T cell receptor monoclonal antibodies.

Transplantation 1994;58:105±7.

39 Alfrey EJ, Wang X, Lee L, et al. Tolerance induced by direct inocu-

lation of donor antigen into the thymus in low and high responder

rodents. Transplantation 1995;59:1171±6.

40 Olausson M, MjoÈrnstedt L, Wramner L, et al. Characteristics of the

induction phase of antithymocyte-globulin-induced heart allograft

tolerance in the rat. Int Arch Allergy Appl Immunol 1988;86:131±

8.

41 Calne RY. An opportunity in organ transplantation. Nature Med

1995;1:20±2.

398 Z. Qi et al.

q 1999 Blackwell Science Ltd, Scandinavian Journal of Immunology, 50, 394±398