Interception of the development of self tolerance in fetal lambs

Eur. J. Immunol. 1989.19: 1387-1392 Interception of the development of self tolerance in fetal lambs 1387

Peter McCullagh Interception of the development of self tolerance in fetal Department of Immunology, John lambs Curtin School of Medical Research, Australian National University, Canberra Investigation of the nature of immunological self tolerance has usually relied upon

experimental protocols in which the tolerant state is interrupted in mature animals with the production of autoimmune disease. While such research has improved the understanding of those processes operative in overt autoimmunity, it has not been informative in relation to events associated with the establishment of self tolerance. Any description of this state which is to be based on observation will necessitate the use of experimental systems that permit observation of animals during the development of self tolerance. The present experiment entailed intervention approximately one third of the way through the gestation period of fetal lambs. An earlier experiment had established that 54-day fetal lambs would accept allografts of adult skin. This indicated that the capacity to discriminate between self and non-self had not been acquired at that age. Fetuses at this stage of gestation were submitted to either partial or total removal of the thyroid gland. The excised tissue was then implanted in nude mice for periods of 5 to 9 weeks. It was subsequently replaced subcutaneously, either in the original donor or in another fetus at a comparable stage of gestation. At postmortem examination, several weeks later, self implants in lambs from which the thyroid gland had been completely removed displayed autoimmune thyroiditis of varying degrees of severity. However, self implants in partially thyroidectomized animals were uniformly free from autoimmune manifestations. This implied that these reactions had not been directed against contaminating murine tissues in the implants replaced in completely thyroidectomized lambs. All allogeneic implants were subject to vey- heavy lymphocytic infiltration, usually with accompanying necrosis consistent with allograft rejection. This was taken as an indication that hypothyroid fetal lambs had become immunocompetent by the time of thyroid reimplantation. Spontaneous immunological reactivity against reimplanted self thyroid tissue by thyroidectomized lambs was interpreted as a failure to acquire the capacity for self recognition as a result of antigen deprivation.

1 Introduction

Understanding of the mechanism of self tolerance may suggest approaches that will extend the range of situations in which it is possible to induce tolerance of non-self determinants. In particular, the experimental demonstration that full development of tolerance to self determinants requires exposure of the developing immune system to them might shed light on cellular interactions required for tolerance induction. One means of investigating the induction of self tolerance has been to interfere with the access of particular self antigens to the fetal immune system until its maturity precludes de novo induction of tolerance.

The first experimental attempt to interfere with acquisition of self tolerance involved removal of the hypophysis from tree frog larvae and its replacement in the donors after metamor- phosis [l]. Failure of acceptance of the replaced self tissue in this experiment suggested that self tolerance required exposure of the immune system to antigen. However, similar experiments, in which frogs and toads (Xenopus) successfully accepted reimplanted hypophysis and eye anlagen that had been removed at the larval stage, cast doubt on the interpretation of the earlier results [2 ] . Subsequently, rats in which thyroid development had been disrupted by exposure to 13'1 in

[I 73111

Correspondence: Peter McCullagh, Department of Immunology, John Curtin School of Medical Research, Australian National University, GPO Box 334, Canberra, A.C.T. 2601, Australia

fetal life were observed to mount spontaneous responses resembling experimental allergic thyroiditis against syngeneic thyroid grafts [3].

While demonstration of thyroid autoimmunity in rats in which thyroid development had been modified supports the conten- tion that development of self tolerance requires early exposure to self antigens, considerable advantage was envisaged in investigating this phenomenon in a larger species with a much longer gestation period. The feasibility of surgical access to fetal lambs before and during acquisition of self tolerance might permit not only the excision of an organ but also its replacement with antigenically distinguishable tissue.

Before any attempt can be made to interfere with the establishment of self tolerance in the fetal lamb, it is essential to identify the time during development up until which induction of tolerance of foreign determinants remains feasible. In the absence of this information, failure to interrupt development of self tolerance because of late intervention might be misin- terpreted as the absence of any requirement for exposure to self determinants during tolerance induction. In contrast with rodents, fetal lambs at 55 days gestation exhibit extended survival during the first year of postnatal life [ 5 ] .

The present study records the development of autoreactivity directed against thyroid tissue in fetal lambs thyroidectomized at 53-55 days gestation. This autoreactivity was demonstrated by reimplanting the extirpated thyroid gland between 92 and 117 days gestation.

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1989 0014-2980/89/0808-1387$02.50/0

1388 P. McGullagh

2 Materials and methods 2.1 Sheep

Adult Merino ewes were used. Ovulation was induced by the placement of an intravaginal flugestone acetate (40 mg) sponge (Chronogest, Laboratoire Pharmaceutique, Porges, Paris, France) and its removal 10 to 12 days later accompanied by the intramuscular injection of 500 IU of serum gonado- trophin (Folligon, Intervet, Australia). Ewes were immedi- ately placed with harnessed rams and the date of their joining was noted. They were subsequently run with harnessed rams bearing a distinctively colored crayon to detect any returning animals. Pregnant ewes were brought into the laboratory 3 to 5 days before their scheduled operation and were maintained on lucerne hay and water ad. lib.

Eur. J. Immunol. 1989.19: 1387-1392

2.2 Surgery

After starving sheep overnight, surgery was performed. Thiopentone induction was followed by halothane anesthesia. The shaved abdomen was opened via a paramedian incision and the uterus was exteriorized. Uterine muscle and fetal membranes were incised using cautery and the fetal head and neck were delivered. Any overflow of amniotic fluid was reserved. Following a midline cervical incision in the fetus and displacement of the sternohyoid muscles, one or both lobes of the thyroid gland were freed from the surrounding tissues by blunt dissection and removed. The fetal wound was closed with interrupted 5/0 atraumatic silk sutures. The fetus and any amniotic fluid that had been collected (with the addition of penicillin) were replaced in utero and the membranes and uterine muscle were reconstituted with a series of purse string sutures using size 6 silk. The uterine wound was inverted by means of widely spaced sutures through the serosa to reduce the risk of subsequent adhesions.

When thyroid tissue was to be re-implanted in a fetus, a small incision was made in uterine muscle and membranes overlying an ear. After partial removal of the ear from the uterus, the thyroid tissue was implanted S.C. on its dorsal surface. Post- mortem examination was performed from 1 to 4 weeks after thyroid implantation. Both larynx and ear were removed and fixed en bloc in formol saline. Sections were stained with hematoxylin and eosin.

2.3 Excised thyroid glands

These were maintained in nude mice [Swiss nu/N+(Nude)] until reimplantation. Under light ether anesthesia, thyroid tissue was introduced S.C. into the flank of each mouse. Large lobes were bisected before implantation to improve their via- bility. Thyroid tissue was recovered, when required, after mice had been killed with chloroform. The fetal lamb tissue could be readily shelled out from the S.C. tissue of the mouse. Any macroscopically evident fibrous tissue was trimmed from the thyroid before this was replaced in a fetal lamb.

2.4 Experimental plan

Each experiment began with the surgical removal of part or all of the thyroid gland from a fetal lamb of 53 to 55 days gestational age (gestation occupies from 145-150 days in the sheep). This thyroid tissue was maintained in the S.C. tissue of a nude mouse for 6 to 8 weeks and was then replaced in the original donor, or in another fetal lamb, during the fourth month of gestation. Three variations of this basic experiment

were undertaken: (a) bilateral removal of the thyroid gland and the eventual replacement of this tissue in the ear of the donor fetus. The completeness of thyroid removal was ascer- tained at autopsy. Fetuses in which the persistence of cervical thyroid tissue (usually a substantial portion of a lobe) was microscopically confirmed at this time were reclassified in group (b). This experimental protocol was intended to detect whether autoimmunity would be manifest against reimplanted self tissue. (b) Partial removal of the thyroid gland was undertaken, either intentionally, as a hemithyroidectomy, or inad- vertently when a substantial portion of a lobe remained in sku. Replacement of the recipient’s own thyroid tissue was performed in these animals to provide an indication of the extent to which any immune reaction against the implant could be directed against passaged tissue of murine origin. (c) Partial or total thyroidectomy of some fetal lambs was followed by implantation of thyroid allografts, first to ensure that the fetus had attained sufficient immunological maturity to reject allografts by the time of reimplantation. This would exclude de novo induction of tolerance by reimplanted self tissue. The second aim was to detect whether any autoimmune reactions observed against reimplanted self thyroid may have been depressed nonspecifically as a consequence of hypothyroidism.

3 Results 3.1 Survival after surgical procedures

A total of 51 fetal lambs carried by 44 ewes were entered in this experiment. Of these fetuses, 34 were submitted to bilateral and 17 to partial thyroidectomy. Reduction in numbers of fetuses available to receive subsequent implants of thyroid tissue occurred both as a result of fetal loss and of failure to recover from nude mice the thyroid tissue corresponding to particular fetuses. Fetal loss was not unexpected given the fragile nature of the animals (a typical 54-day fetal lamb would have a crown-rump length of 8 cm and a body weight of approximately 30 gm, Fig. 1). However, post-thyroidectomy survival rates improved considerably as the experiment pro- gressed and technique improved. Twenty fetuses that had undergone bilateral and nine that had been subjected to unila- teral thyroidectomy survived to the fourth month of gestation. Surviving fetuses, for which the corresponding self thyroid tissue could not be recovered from nude mice, were used as allograft recipients. Thyroid tissue, the fetuses corresponding to which had died, was used as allografts. After reimplantation of thyroid tissue, 25 out of 29 fetuses survived until the time of examination. Recovery of small thyroid implants from the S.C. tissues several weeks after placement was difficult in some instances. Implanted thyroid tissue was successfully identified at autopsy in 17 of these fetuses and the results in succeeding sections are derived from those animals. Allografts which had undergone rejection were usually easy to identify on account of edema. Bilaterally thyroidectomized fetal lambs were char- acterized macroscopically by reduced growth rate and a complete absence of wool at the time of autopsy in the fourth month of gestation.

3.2 Maintenance of fetal thyroids in nude mice

When removed from fetal lambs, thyroid glands consisted of discrete cords of cells in which follicle formation was about to begin (Fig. 2a and b). Over the course of the 6- to 8-week period during which they remained implanted in nude mice, thyroid maturation occurred. By the time of reimplantation


nude mouse, was observed in some grafts that were not reimplanted. There was no indication of inflammatory cell infiltration in any of the thyroid tissue recovered from nude mice.

3.3 Response of completely thyroidectomized fetal lambs to

Fetal lambs which had been subjected to bilateral thyroidectomy at 54 and 55 days gestational age received their own thyroid glands as S.C. implants. As summarized in Table 1, a consistent response was observed in all 6 fetuses. The most conspicuous feature observed in the implants of self thyroids was the variation in intensity of the inflammatory process throughout the gland. As shown in Fig. 4a and b, areas of retained, comparatively normal thyroid follicles were interspersed with other areas of total lymphocytic replacement. Necrosis was not observed in any of the implants. Lymphatic vessels with numerous lymphocytes were prominent. In areas between residual thyroid follicles and complete lymphoid infiltration, remnants of disintegrating thyroid follicles were visible among the infiltrate (Fig. 4b). As noted in Table 1 and in its footnotes, marked lymphocytic infiltration was observed in autologous implants in two fetuses which were found to retain very small remnants of thyroid tissue in the cervical region, despite morphological evidence of severe hypothyroidism at autopsy. This may indicate that a total absence of a self determinant may not be necessary for autoimmunity against it to develop.

3.4 Response of partially thyroidectornized fetal iambs to

The results of replacing self thyroid tissue in fetal lambs which had previously been submitted to partial thyroidectomy are summarized in Table 2. Lymphocytic infiltration of the thyroid parenchyma was not observed in any of these grafts. Lym- phocytes were present in small numbers in the fibrous capsules of some. Several grafts showed a central fibrous replacement

reimplantation of self thyroid tissue

Figure I. Surgical removal of thyroid gland from fetal lamb at 54 days gestation. Note trachea free from thyroid tissue and differing size of extirpated lobes of thyroid gland (arrowed). Magnification X 4.

the tissue had been completely transformed into follicles containing colloid (Fig. 3). Owing to the small size of the implants it was not feasible to biopsy them in addition to effecting their replacement in the donor. However, central fibrosis, presum- ably as a result of ischemic necrosis after placement in the

reimplantation of self thyroid tissue

Table 1. Response of completely thyroidectomized fetal lambs to surgical reimplantation of autologous thyroid tissue

Sheep Gestational age Age of autologous Interval when bilaterally thyroid before autopsy thyroidectomized reimplantation (days)

(days) (days)

Fcaturcs at autopsy

54

19 152

412R G5"'

G58

GWb'

G6LIb)

54

54 54

55 54

54

54

54

97 92

107 99

1 us 105

I05

28

14 12

20 21

16

16

20

No cervical thyroid. Irrcgular foci of lymphocytic infiltration interspersed with intact thyroid tissue in implant. N o cervical thyroid. Heavy infiltration of implant.

N o cervical thyroid. Patchy lymphocytic infiltration of implanted thyroid.

No cervical thyroid. Heavy lymphocytic infiltration of implant. No cervical thyroid. Implant rcplaccd by hcavs lymphocytic

infiltration containing disintcgrating thyroid follicles. Some intact areas of thyroid tissue.

No ccrvical thyroid. Majority of implant rcplaccd by lymphocytic infiltration.

Small residual cervical thyroid (see footnote). Implant hcavily infiltrated with lymphocytes.

Sniall residual cervical thyroid (see footnote). Implant hcavily infiltrated with lymphocytes.

a) Also received several allogeneic thyroid implants (see Table 3). b) These two fetuses were observed to be extremely hypothyroid at postmortem examination as judged by gross morphology. G6U had very

sparse cover with wool fibers while G68 was completely bald. Examination of the cervical region revealed a small residuum of thyroid tissue, occupying less than one low power microscopic field, in each case.

1390 P. McGullagh Eur. J. Immunol. 1989.19: 1387-1392

Figure 2. (a) Normal thyroid tissue from fetal lamb at 53 days gestation. Note cords of cells with tendency to grouping in a follicular arrangement, but without any collid formation. (b) Normal thyroid gland from fetal lamb at 60 days gestation. Note grouping of many of cells into follicles with the appearance of colloid in some. Magnifica- tion x 330.

which was free from any infiltration. This resembled that noted above in some of the thyroid implants recovered from nude mice and not implanted in fetal lambs. Thyroid parenchyma in this groups of implants, while free from lymphoid infiltration, commonly deviated from normal as regards fol-

Figure 4. (a) Implant of self thyroid tissue into right ear of bilaterally thyroidectomized fetus G5 (see Table 1). Postmortem performed 23 days after implantation. Note retention of some areas of thyroid tissue, albeit with erosion of margins by lymphocytic infiltrate [see (b) for more detail]. In other areas, total replacement of thyroid tissue by lymphoid tissue, with development of follicles, has occurred. Magnifi- cation x 29. (b) Higher power of section photographed in (a). Note isolated thyroid follicles (arrowed) surrounded by lymphocytic infiltrate. Magnification x 130. (c) Allogeneic thyroid implant in left ear of bilaterally thyroidectomized fetus G5 (details in Table 3). Note complete replacement of thyroid tissue by lymphocytic infiltrate with formation of primary lymphoid follicles. Dilated lymphatic vessels in implant are arrowed. Magnification x 54.

Figure 3 . Thyroid implant, originally derived from fetal lamb at 54 days gestation, recovered after 7-week interval in a nude mouse. Note differentiation of thyroid follicles, presence of fibrous capsule and absence of lymphocytic infiltrate. Magnification X 78.


3.5 Response of incompletely and completely thyroidectomized fetal lambs to implantation of allogeneic thyroid tissue

Figure 5. Self thyroid gland replaced in hemithyroidectomized donor 512L (see Table 2 for experimental details) after 9 weeks in nude mouse. Examination undertaken 14 days later. Note retention of thyroid follicular architecture and absence of lymphocytic infiltrate within gland. Thin fibrous capsule and septae present. Note presence of mild cellular infiltration of capsule and also presence of dilated lymphatic vessel (arrowed). Magnification x 75.

licular size and structure. This may have reflected the abnormal environment of the nude mouse in which differentiation had occurred. Abnormally high levels of TSH in the thyroidectomized recipient fetuses may have also been contributory. Fig. 5 shows an example typical of this group of implants.

The results of implanting thyroid allografts in seven fetal lambs that had previously had a portion or all of their thyroid gland removed are summarized in Table 3. All implants showed major changes ranging from confluent lymphoid infiltration, sometimes with formation of giant cells, to central necrosis with peripheral infiltration. Fig. 4c illustrates an example of the former response in a totally thyroidectomized host. No difference in seventy of reaction was noticeable between implants in partially and totally thyroidectomized fetuses examined some 3 weeks after placement. This should not be taken to imply that immunological function had necessarily attained normal levels in severely hypothyroid lambs, only that it remained adequate to sustain an allograft response over this period.

4 Discussion If fetal lambs were subjected t o complete thyroidectomy at 54/55 days gestation they reacted to the reimplantation of self thyroid tissue 6 to 8 weeks later with heavy lymphoid infiltration that resulted in partial replacement of the grafted thyroid parenchyma. In contrast, if only a portion of the thyroid gland had been removed, fetuses did not react in this way against the reimplanted portions. Although the experimental systems dif- fered substantially, these observations are open t o a similar interpretation as that drawn by Eishi and McCullagh [3]. They observed the condition of syngeneic thyroid tissue grafted into

Table 2. Response of partially thyroidectornized fetal lambs to surgical reimplantation of autologous thyroid tissue

Sheep Gestational age at Age at thyroid Interval to autopsy first operation reimplantation (days)

(days) (days)

77 58 1021 15 412L 55 107 20

417L 56 114 17

5 12"' 53 116 14

61" 54 117 26

Features at autopsy

Healthy implant. No infiltration. Some old central fibrosis. Variable thyroid follicle size in implant. Completely free from

inflammation. Implant atropic with old fibrotic areas and few follicles. No

infiltration. Well preserved thyroid with pronounced fibrous capsule. No

infiltration. Thyroid tissue very well preserved and free from infiltration. Some

lymphocytes present in fibrous capsule.

a) Also received an implanted allogeneic thyroid gland (see Table 3).

Table 3. Response of incompletely and completely thyroidectomized fetal lambs to implantation of allogeneic thyroid tissue

61 Partial thyroidectomy 55 117 26 Allograft reaction with necrosis. 376 Partial thyroidectomy 55 96 15 Continuous lymphocyte infiltration and giant cells.

433R Partial thyroidectomy 55 98 21 Florid reaction. Central necrosis with peripheral

433L Partial thyroidectomy 55 98 21 Total replacement of implant by lymphoid tissue. 441L Partial thyroidectomy 55 105 6 Central necrosis surrounded by vigorous

512 Partial thyroidectomy 53 116 14 Violent allograft reaction with necrosis of implant. G5 Complete thyroidectomy 54 99 21 Three different allografts implanted in separate

locations. All replaced by lymphocytic infiltration. including giant cells, but without necrosis.

No necrosis.

lymphocytic infiltration.

lymphocytic infiltration.

1392 P. McGullagh

rats, the thyroid development of which had been disrupted by administration of 13'1 in utero. The syngeneic grafts were subject to lymphocytic infiltration and other changes closely resembling experimental allergic thyroiditis induced by immunization against thyroid tissues.

Interpretation of the present experiment requires assumptions on several points. The first point concerns the nature of the pathological process in the implants. Its classification as immunological is based exclusively on its dominant feature of lymphoid infiltration. This interpretation was supported in experiments with rats by the similarity in appearance between thyroid tissue grafted into rats exposed to I3'I in utero and tissue that was subject to experimental allergic thyroiditis. However, the morphology of autoimmune thyroiditis in the sheep is not available for comparison with that of thyroid implants. Exami- nation of allograft responses in thyroid implants placed in fetal lambs revealed a more severe and rapid reaction than that occurring in autologous implants in completely thyroidectomized hosts. Thus, necrosis was never observed in the latter. A similar difference in severity of reaction was observed between thyroid allografts and autografts implanted in 13'1- treated rats (unpublished observation). The apparent difference in severity of reaction against allogeneic and autologous thyroid implants in the present experiments could be attribut- able to the much greater range of antigenic targets borne by the former. However, in view of the relatively limited number of experimental animals and their examination at single points in time, the significance of the difference is questionable and observations could be due to divergent timing of reaction.

Eur. J . Immunol. 1989.19: 1387-1392

A second assumption is that the reaction against reimplanted autologous thyroid tissue was directed against parenchymal thyroid cells rather than against contaminating blood borne cells derived from the intermediate mouse host. The possibility must be considered that what was apparently an autoimmune response represented a xenograft reaction against passaged cells. This possibility, recognized in the original experiments on self tolerance in amphibia [12], does not appear to explain the widespread lymphocytic infiltration of autologous grafts in the present experiment. Hemithyroidectomized lambs would have been as likely as their completely thyroidectomized equivalents to receive contaminating tissue from the intermediate host and to react against it, yet infiltration was absent from the parenchyma of autologous grafts replaced in partially thyroidectomized fetuses. Assuming that the fibrous tissue encapsulating sheep thyroids recovered from nude mice was of murine origin, the sparse foci of lymphocytes sometimes observed in it after reimplantation may have represented a weak response to antigens on tissue from the intermediate host.

As in the experiment of Eishi and McCullagh [3] the target antigen of the inferred autoimmune response in the present experiment has not been identified. These authors clearly showed that thyroglobulin was not the target determinant of the autoimmune response. Complete removal of the thyroid gland in the present experiment, as distinct from infliction of radition damage on it, may have resulted in autoreactivity in the fetal lamb being directed against a broader range of deter-

minants than occurred in fetal rats. If the range of morphological indications of varying secretory activity observed in thyroid implants in the present experiment was reflected in differing expression of potential self determinants, the antigenicity of different implants may also have vaned substantially.

Under these circumstances, variation between individual fetuses in the nature and extent of the autoimmune process might be anticipated. The occurrence of reactions against autologous thyroid implants in two fetuses (G60 and G68) which, although severely affected by hypothyroidism, retained some detectable thyroid tissue in the cervical region raises the ques- tion of the magnitude of self determinant expression that is required in order to induce self tolerance. Similar residual thyroid tissue, albeit grossly abnormal, remained in rats that manifested thyroid autoimmunity as a result of exposure to I3lI in utero [3]. In neither that report nor in the present experiment was the target determinant of autoimmunity identifiable. Consequently, it is impossible to distinguish between complete failure of damaged or residual thyroid tissue to express the target determinant and the alternative that its expression was not at a level sufficient to induce tolerance.

The findings of Illig [6] suggest that complete absence of a self determinant may not be required for autoimmunity to develop. Illig reported the occurence of autoimmunity against growth hormone in pituitary dwarves who, nevertheless, had detectable levels of this hormone in the blood. The present experiment fails to distinguish whether active induction of autoimmunity occurred solely as a consequence of antigen deprivation or whether deprivation merely conferred a liability to development of autoimmunity following subsequent exposure to that determinant.

The main aim of the present experiment was to establish that autoreactivity in response to interference with organ maturation could be demonstrated in a second mammalian species. A longer term objective was to test the feasibility of interception of acquisition of self tolerance in a larger species as a prelimi- nary to untertaking further experiments not practicable in rodents. One such experiment, in progress, will examine whether tolerance of organ-specific thyroid self determinants can be induced by immediate substitution of identical tissue from an allogeneic donor for the ablated organ and compare the relative requirements for simultaneous induction of tolerance of self and foreign determinants. The possible extent to which mechanisms are shared between these two manifestations of tolerance could thus be determined.

.

Received December 12, 1988; in revised form February 20, 1989.

5 References

1 Tnplett, E. L., J. lmmunol. 1962. 89: 505. 2 Rollins-Smith, L. A. and Cohen, N., Nature 1982. 299: 820. 3 Eishi, Y. and McCullagh, P., lmmunology 1988. 46: 280. 4 Silverstein, A. M., Prendergast, R. A. and Kraner, K. L., J. Exp.

5 McCullagh, P., Transplantation 1988. 46: 280. 6 Illig, R., J . Clin. Endoc. Metab. 1970. 31: 679.

Med. 1964. 119: 955.

Documents

Interception of the development of self tolerance in fetal lambs