Control of Neonatal Tolerance to Tissue Antigens by Peripheral T Cell Trafficking

See allHide authors and affiliations

Science  13 Nov 1998:
Vol. 282, Issue 5392, pp. 1338-1341
DOI: 10.1126/science.282.5392.1338


Self tolerance is acquired by the developing immune system. As reported here, particular properties of the neonatal tissue contribute to this process. Neonatal skin, but not adult skin, was accessible for naı̈ve CD8 T cells. In mouse bone marrow chimeras generated at different ages, recent thymic emigrants were tolerized to a skin-expressed major histocompatibility complex class I antigen only during a neonatal period but not during adulthood. Blockade of T cell migration neonatally prevented tolerance induction. Thus, T cell trafficking through nonlymphoid tissues in the neonate is crucial for the establishment of self tolerance to sessile, skin-expressed antigens.

Differences in tolerance induction during the neonatal and adult periods of life have fascinated immunologists since the pioneering work of Billingham, Brent, and Medawar (1). Neonatal mice, in contrast to adults, develop lifelong tolerance to allogeneic skin grafts when exposed to allogeneic cells of the same donor strain; hence, self tolerance is actively acquired. The newborn immune system can also mount an immune response when challenged (2). Thus, there appear to be quantitative but not qualitative differences among the cells generating an immune response (2).

Although these investigations have focused on systems in which mobile antigen-presenting cells pick up antigen and carry it to lymphoid organs for T cell recognition, the role of differential T cell migration in tolerance induction to sessile self antigens expressed exclusively on extrathymic tissues is undefined. Large-scale trafficking of virgin T cells through extralymphoid tissues has been observed in fetal sheep, in contrast to the restricted circulation in the adult animal (3). To test whether differential T cell migration through neonatal versus adult tissue would influence tolerance induction to tissue-specific self antigens, we used a transgenic mouse model expressing the major histocompatibility complex (MHC) class I antigen Kb under control of the constitutively expressed keratin IV promoter on skin keratinocytes. These 2.4KerIV-Kb mice exhibit peripheral tolerance characterized by the presence of Kb-specific T cells that do not reject Kb-positive grafts and hence are functionally tolerant (4).

We first investigated whether Kb expressed on keratinocytes in the adult mouse could induce T cell tolerance. The following bone marrow (BM) chimeras were established: Adult 2.4KerIV-Kbmice on the Rag-2–/– background (5) were reconstituted with T cell–depleted BM cells from Kb-specific T cell receptor (Des-TCR) transgenic mice (Des-TCR → 2.4KerIV-Kb.Rag-2–/–). Efficient reconstitution visualized with the anticlonotypic antibody Désiré-1 (6) was obtained 2 to 3 months after transfer of Des-TCR BM cells (Fig. 1). Des-TCR → Rag-2–/–chimeras were included to test the Kb responsiveness of Des-TCR T cells in the absence of the Kb transgene. Tolerance was assessed by injecting chimeric mice with Kb-positive P815 tumor cells (4) 2 to 3 months after reconstitution. P815.Kb tumors are rejected by single-transgenic Des-TCR mice and accepted by tolerant double-transgenic Des-TCRx2.4KerIV-Kb mice (7). The chimeric mice that had been reconstituted as adults rejected the P815.Kb tumor (Table 1, group A) and were therefore not tolerant. This was confirmed in another assay system. After grafting Des-TCR → 2.4KerIV-Kb.Rag-2–/– chimeras with skin autografts from their tails onto their lateral thoracic walls, rejection became visible in all seven recipients after 14 days (8). These findings show that tolerance to the Kb self antigen is not induced when the antigen is expressed by mature peripheral tissue in the adult environment.

Figure 1

T cell phenotype of 2.4KerIV-Kb.Rag-2–/– and Rag-2–/– mice 10 to 14 weeks after reconstitution with Des-TCR BM cells. (A to C) Two-color immunofluorescence of B cell–depleted splenocytes (5) from 5-week-old (A), 15-day-old (B), and neonatally (C) reconstituted Des-TCR → 2.4KerIV-Kb.Rag-2–/– BM chimeras stained for CD8 (vertical axis) and the clonotypic Des-TCR (horizontal axis). (D) Another group of 5-week-old thymectomized 2.4KerIV-Kb.Rag-2–/– mice were grafted with thymic lobes from neonatal 2.4KerIV-Kb transgenic mice before BM inoculation. (E) Untreated Des-TCRx2.4KerIV-Kb double-transgenic mice. (Fto I) Corresponding control groups of Rag-2–/–mice were injected at 5 weeks (F), 15 days (G), on the first day after birth (H), or after thymectomy and grafting at the age of 5 weeks with a thymus from neonatal 2.4KerIV-Kb transgenic mice (I). (J) Des-TCR single-transgenic mice.

Table 1

Tumor acceptance in BM chimeric mice. Recipients were reconstituted at various ages, with (group D) or without (groups A, B, and C) thymectomy and engraftment with a neonatal 2.4KerIV-Kb thymus. Mice in groups A, B, and C were injected subcutaneously with 105 P815.Kb tumor cells 10 to 14 weeks after reconstitution with Des-TCR BM cells (5). Mice in group D were thymectomized and grafted with a neonatal thymus from 2.4Ker-Kb mice at the age of 5 weeks, followed by reconstitution with Des-TCR BM cells; P815.Kbcells were injected subcutaneously 6 weeks later (5). Mice in group E were nonchimeric controls injected with P815.Kbcells at the age of 8 to 12 weeks.

View this table:

Is there a critical time window during early development for tolerization to the skin-expressed Kb antigen? We injected 2.4KerIV-Kb.Rag-2–/– mice with Des-TCR BM cells either at birth or at the age of 15 days (Fig. 1, B and C) (5) and assessed tolerance after 2 to 3 months. In the neonatal chimeras, Des-TCR T cells were already found in the peripheral blood on day 19 after reconstitution and rapidly increased afterward. Tolerance was observed in these chimeras (Table 1, group C). This result is consistent with the observation that Kb-specific tolerance is induced in Des-TCRx2.4KerIV-Kb mice expressing Kb on keratinocytes and the Des-TCR constitutively from birth. In contrast, 2.4KerIV-Kb.Rag-2–/– mice reconstituted at 15 days of age rejected the P815.Kb tumor cells (Table 1, group B). Thus, tolerance induction to skin-expressed Kbantigen appears to occur only during an early period after birth.

Support for this view was also derived from the following experimental approach: Elimination of Des-TCR T cells with the anticlonotypic antibody until day 15 after birth abolished tolerance induction in Des-TCRx2.4KerIV-Kb mice. Six of seven mice that had received the antibody treatment rejected P815.Kb tumor cells grafted as adults, compared to 1 of 18 that did not receive antibody (9). The Désiré antibody-treated mice displayed on day 18 less than 1% and on day 24 only 4 to 5% Des-TCR T cells in the peripheral blood. Thus, the reconstitution and the antibody treatment experiment indicate a time window of about 3 to 4 weeks after birth as a rough approximation for the tolerance-sensitive period.

Tolerance induction in the neonate, in contrast to that in the adult, could be attributable either to intrinsic properties of T cells leaving the neonatal versus adult thymus or to characteristics of the target tissue in the newborn mouse. To distinguish between these possibilities, we constructed thymus-BM chimeras by replacing the endogenous thymic rudiment of 5-week-old 2.4KerIV-Kb.Rag-2–/– or Rag-2–/– mice with the thymus of newborn 2.4KerIV-Kb animals, followed by reconstitution with Des-TCR BM cells (5) (Fig. 1, D and I). These chimeras rejected P815.Kb cells and were therefore not tolerant. Two conclusions can be drawn from these results. First, a neonatal 2.4KerIV-Kb thymus does not elicit Kb-specific tolerance in Rag-2–/– mice, thereby excluding the possibility of transient Kbexpression in the neonatal thymus. Second, the age of the thymic tissue in which the T cells develop does not control their capacity to become tolerized by peripheral tissue antigens. Instead, the maturity of the peripheral tissue appears to be critical. Because the neonatal thymus graft was not irradiated, low numbers of recent thymic emigrants could already be detected in the periphery a few days after transplantation. Such T cells represent neonatal T cells emigrating during the tolerance-sensitive time window mentioned above.

Various characteristics of the neonatal tissue may account for tolerance induction in the neonatal Des-TCR → 2.4KerIV-Kb.Rag-2–/– chimeras. Differential Kb expression on keratinocytes of the neonatal and adult skin could be excluded by immunohistological analysis (10) and skin transplantation. The survival time of grafts from neonatal and adult 2.4KerIV-Kb.Rag-2–/–donors transplanted on Des-TCR mice was similar (11). Alternatively, an enhanced trafficking of naive T cells through peripheral tissues in the neonate may facilitate efficient encounter of the skin-expressed Kb antigen. Therefore, we investigated the homing pattern of lymphocytes in newborn and adult 2.4KerIV-Kb mice by injecting radiolabeled lymphocytes from Des-TCR mice (12). Four hours after transfer, 38% of51Cr-labeled T cells could be detected in the adult spleen, versus less than 1% in the neonatal spleen, similar to data in newborn rats (13). Instead, inoculated T cells accumulated in nonlymphoid tissues of neonatal recipients, such as lung and skin (Fig. 2). For example, 14% of injected lymphocytes entered the skin of neonatal 2.4KerIV-Kb mice, compared to only 2.3% in the skin of adult recipients; this ratio was independent of the number of injected cells. Intradermal trafficking of lymphocytes at the single-cell level, observed by means of vital fluorochromes (14), confirmed that neonatal skin is more accessible for naive T cells than adult skin.

Figure 2

Enhanced migration of Des-TCR T cells into the skin of newborn 2.4KerIV-Kb mice (solid bars) versus 5-week-old 2.4KerIV-Kb mice (open bars).51Cr-labeled B cell–depleted spleen and lymph node cells from Des-TCR mice were injected intravenously into six or seven mice per group, and various organs and remaining body were tested 4 hours after transfer (12). Data are shown as percentage of total counts per minute (measured per mouse) and are representative of three experiments.

The intradermal trafficking of naı̈ve T cells in the newborn animal may have important implications for self-tolerance induction. Therefore, we asked whether blockade of T cell interaction with endothelia in the neonatal skin during the tolerance-sensitive phase would prevent tolerance induction. E- and P-selectin have been suggested to function as skin-selective adhesion molecules for T cells at sites of inflammation (15). In an orientation experiment, a single dose of antibodies to E- and P-selectin (20 μg), simultaneously given with 51Cr-labeled T cells to neonatal 2.4KerIV-Kb mice, reduced homing to the skin by 39% after 12 hours but did not influence the migration pattern to other tissues. Therefore, newborn Des-TCRx2.4KerIV-Kb mice were treated for 15 days with antibodies and tested for tolerance at the age of 6 weeks (16). Most of these animals rejected Kb-positive tumor grafts, indicating that tolerance was abrogated, whereas newborn mice that were injected with control antibodies or received no injection accepted the P815.Kb tumor (Table 2) (17, 18). There was no sign of an impaired thymic maturation caused by the selectin monoclonal antibody (mAb) treatment, because the T cell phenotype was unchanged (10) and the respective T cells were functionally active in rejecting the P815.Kb tumor (Table 2). Thus, blocking T cell migration to the neonatal skin prevented tolerance induction to Kbexpressed on keratinocytes.

Table 2

Effect of E- and P-selectin mAbs on Kb-specific tolerance induction (16).

View this table:

Because recognition of cutaneous antigens by naı̈ve CD8 T cells is apparently limited to the perinatal period, maintenance of self-protection in the adult must be controlled by alternative mechanisms, as naı̈ve T cells continue to leave the thymus in adult life. Further studies will be required to determine whether regulation as described in various animal models (19) or other processes might account for the inability of the naı̈ve anti-Kb T cells in the adult Des-TCRx2.4KerIV-Kb mice to reject grafts. Recent reports indicate that self antigens expressed by intact peripheral organs may be presented to CD8 T cells in the draining lymph nodes after uptake by dendritic or similar cells via an exogenous class I–restricted pathway. However, this phenomenon of cross-presentation was only observed when the tissue antigen was expressed in large amounts in the kidney and pancreas (20). In the present mouse model, however, we used a “foreign” MHC class I antigen and a corresponding T cell receptor recognizing only the intact Kb antigen but not a processed form of that antigen. Hence, our results exclude the possibility that antigen-presenting cells of the skin pick up the Kb antigen and present it in the neonate but not in the adult animal. The results indicate that the tolerance induction described here is dependent on migration of T cells into the skin. It is therefore different from cross-presentation and may represent a mechanism for tolerance induction to cutaneous antigens in situations where the expression of tissue antigens is too low for cross-presentation.

  • * Present address: Ontario Cancer Institute/Amgen Institute, Toronto, Ontario M5G 2C1, Canada.


View Abstract

Navigate This Article