Characterization of Tolerance Induction through Prenatal Marrow Transplantation: The Requirement for a Threshold Level of Chimerism to Establish Rather than Maintain Postnatal Skin Tolerance

Abstract

Hematopoietic chimerism resulting from prenatal marrow transplantation does not consistently result in allotolerance for unidentified causes. In a C57BL/6-into-FVB/N murine model, we transplanted T-cell-depleted adult marrow on gestational day 14 to elucidate the immunological significance of chimerism towards postnatal tolerance. Postnatally, chimerism was examined by flow cytometry, and tolerance by skin transplantation and mixed lymphocyte reaction. Regulatory T cells were quantified by FoxP3 expression. Peripheral chimerism linearly related to thymic chimerism, and predicted the degree of graft acceptance with levels >3% at skin placement, yielding consistent skin tolerance. Low- and high-level chimeras had lower intrathymic CD3^high expression than microchimeras or untransplanted mice. Regardless of the skin tolerance status in mixed chimeras, donor-specific alloreactivity by lymphocytes was suppressed but could be partially restored by exogenous interleukin-2. Recipients that lost peripheral chimerism did not accept donor skin unless prior donor skin had engrafted at sufficient chimerism levels, suggesting that complete tolerance can develop as a consequence of chimerism-related immunosuppression of host lymphocytes and the tolerogenic effects of donor skin. Thus, hematopoietic chimerism exerted immunomodulatory effects on the induction phase of allograft tolerance. Once established, skin tolerance did not fade away along with spontaneous regression of peripheral and tissue chimerism, as well as removal of engrafted donor skin. Neither did it break following in vivo depletion of increased regulatory T cells, and subcutaneous interleukin-2 injection beneath the engrafted donor skin. Those observations indicate that the maintenance of skin tolerance is multifaceted, neither solely dependent upon hematopoietic chimerism and engrafted donor skin nor on the effects of regulatory T cells or clonal anergy. We conclude that hematopoietic chimerism generated by in utero hematopoietic stem cell transplantation is critical to establish rather than maintain postnatal skin tolerance. Therefore, the diminution of hematopoietic chimerism below a threshold level does not nullify an existing tolerance state, but lessens the chance of enabling complete tolerance.

Introduction

Based upon the observation that sharing of fetal blood is sufficient to enable skin graft acceptance (4,32), acquired immune tolerance to allografts was first demonstrated in Medawar and colleagues' study using preemptive exposure of donor cells (3). Subsequently, tolerance was produced by marrow cell infusion in irradiated adult recipients prior to or at the time of organ transplantation (18,29,31). With the discovery of trace donor leukocytes in the tissues or blood of long-surviving organ recipients in 1992, Starzl et al. (38,39) and Wood and Sachs (46) proposed that immunological tolerance achieved either by bone marrow or solid organ transplantation is linked by a common dependence on the presence of hematopoietic chimerism. Thereafter, this concept has influenced our thinking about tolerance and rejection of organs in modern transplantation immunology, fueling high expectations for experimental manipulations of donor cell chimerism to improve tolerance induction.

There is no shortage of experimental strategies for tolerance induction through the creation of mixed chimerism in postnatal life (45). Despite years of effort with animal studies, questions remain as to the relationship of hematopoietic chimerism to tolerance. In rat chimeras with skin tolerance, the abolition of chimerism by antidonor antisera caused the rejection of engrafted donor skin within 12 h after serum injection (28). This hyperacute graft rejection is most likely procured by direct inimical effects of antibodies to the alloantigens expressed on vascular endothelia of the graft (27) rather than by the abolition of chimerism. Further studies conducted in mice (22,35) showed that the elimination of donor cells from established mixed chimeras with antidonor major histocompatibility complex (MHC) antibody caused the failure of skin acceptance rather than the rejection of engrafted skin. This observation was interpreted as the recipients' inability to maintain tolerance due to the loss of chimerism, based on the assumption that there had been existing tolerance resulting from hematopoietic chimerism. However, it is not always the case that hematopoietic chimerism yields complete tolerance and the acceptance of tissue grafts (25,30), especially with microchimerism (11,17,33). Moreover, chimerism and skin tolerance may be maintained by different mechanisms (48). Evidence also shows that graft tolerance can persist despite the loss of peripheral chimerism (20,21,23,37). It is worth mentioning that the preconditioning or immunosuppressive programs required for postnatal bone marrow transplantation potentially interfere with the assessment of hematopoietic chimerism and graft tolerance because they can delay the rejection of skin grafts (14) or suppress the elimination of donor leukocytes (46).

In utero transplantation (IUT) provides a unique approach to generating allogeneic hematopoietic chimerism as well as inducing postnatal tolerance to donor-specific grafts without artificially suppressing the host's immune system. However, conflicting experimental observations pertaining to tolerance and chimerism remain in the setting of IUT (5,12). This prompted us to study the levels of peripheral chimerism required for postnatal skin tolerance, clarify the causal relationship between chimerism and donor-specific tolerance, and investigate possible peripheral mechanisms involved in immune tolerance after IUT of bone marrow.

Materials and Methods

Mouse Husbandry

Inbred FVB/N (H-2^q, Thy1.1), C57BL/6 (H-2^b, Thy1.2), and C3H (H-2^k) mice were purchased from National Laboratory Animal Center (Taipei, Taiwan) at the age of 6–8 weeks, and housed in the Animal Care Facility at Chang Gung Memorial Hospital under the standard guidelines from Guide for the Care and Use of Laboratory Animals and with the approval of the Committee on Animal Research. Recipient FVB/N females were caged with FVB/N males in the afternoon and checked for vaginal plugs the following morning. The day when the plug was observed was designated as day 0 of the pregnancy.

Bone Marrow Cell (BMC) Preparation and IUT

Adult BMCs from C57BL/6 mice (8–24 weeks of age) were harvested by flushing the tibias and femurs with phosphate-buffered saline (PBS) using a 26-gauge needle. Light-density BMCs were obtained by layering BMCs over NycoPrep 1.077A (Nycomed, Pharma AS, Oslo, Norway) and centrifuging at 600 x g for 25 min. Light-density BMCs were then depleted of T cells by anti-CD3∊ FITC (BioLegend, San Diego, CA) and anti-FITC microbeads (Miltenyi Biotec, Auburn, CA) to obtain T-cell-depleted BMCs that contained less than 0.5% of CD3⁺ cells and 2.22–4.74% of CD11c⁺ cells. Within 3 h after preparation, IUT of 5–10 × 10⁶ T-cell-depleted BMCs in 5–10 μl PBS was performed as previously described (7,8). Recipients with any clinical sign of graft-versus-host disease were disqualified for enrollment.

Harvests of Peripheral Blood and Tissues

Regular blood sampling was performed via the tail tip after restraint and thymic biopsy through a suprasternal incision under anesthesia. Harvests of spleen, bone marrow, lymph node, thymus, and peritoneal cells demanded a sacrifice procedure. To avoid contamination by cells from different tissues, peripheral blood, if necessary, was first sampled from axillary or neck vessels. Then, peritoneal cells were obtained by flushing the peritoneal cavity with 10 ml PBS. Subsequently, BMCs were harvested by flushing the tibias and femurs with PBS. At last, the spleen, lymph node, and thymus (or its remnant) were removed, washed with PBS, and dissociated by passage through 70-μm cell strainers (BD Biosciences, Bedford, MA) to obtain cells for analyses.

Analyses of Chimerism

Samples were depleted of red cells using ACK buffer, pH 7.2–7.4, consisting of 0.15 M NH₄Cl, 1.0 mM KHCO₃, and 0.1 mM Na₂EDTA (Sigma Chemical Co., St. Louis, MO). Cells were first incubated with anti-mouse CD16/32 antibody (Clone 93, BioLegend) to lessen nonspecific Fc-mediated binding of monoclonal antibodies, and then stained with anti-H-2K^q FITC (BioLegend) and anti-H-2K^b PE (BioLegend). A negative control consisted of anti-H-2K^q FITC and mouse IgG2a PE (BioLegend) to define background staining. Thymic chimerism was further examined by staining the thymocytes with anti-Thy1.1 FITC (Clone OX-7, BioLegend) and anti-Thy1.2 PE (Clone 30-H12, BioLegend). Lineage analyses for thymic chimerism were carried out using anti-H-2K^b PE and either FITC-labeled anti-CD3∊, -CD4, -CD8, -CD45R, -CD11b, -CD11c, or-Gr1 (BioLegend). After gating out dead cells using propidium iodide, 100,000 events were acquired for analyses of chimerism levels and 50,000 events for lineage analyses. The sensitivity of chimerism analyses by flow cytometry is 0.01%.

Skin Transplantation

Skin transplantation was performed within 3 days after the examination of chimerism, using tail skins from FVB/N (syngeneic), C57BL/6 (donor-specific), and C3H mice (third-party). After the removal of dressings on day 7, grafts were monitored daily for the first month and then at least twice a week thereafter until both C57BL/6 and C3H skin grafts were rejected or C57BL/6 skin engrafted for at least 4 months after skin transplantation. Engrafted skin was defined by good hair growth and examined by hematoxylin-eosin staining after its removal. Rejection was defined as when ≤20% of the original graft remained. A tolerant state was defined by skin engraftment for at least 120 days.

Mixed Lymphocyte Reaction (MLR) and Anergy Evaluation

Splenic lymphocytes were enriched with NycoPrep 1.077A density gradient. Responders were lymphocytes from FVB/N mice and stimulators were 3,000 cGy irradiated lymphocytes from FVB/N, C57BL/6, and C3H mice. Briefly, 2 × 10⁵ responder cells were cultured in 200 μl RPMI-1640 medium (Life Technologies, Grand Island, NY) for 4 days with 6 × 10⁵ stimulator cells at 37°C in a humidified, 5% CO₂-containing incubator. Then, [³H]thymidine (ICN Biomedicals, Costa Mesa, CA) was added at a final concentration of 1 μCi/well. The next day, cells were harvested for counting incorporated ³H in a liquid scintillation counter (1450 Microbeta Plus counter, Wallac, OH). Stimulation indices were calculated using mean cpm divided by mean background cpm (responder cells only).

Clonal anergy was evaluated by supplementing mixed lymphocyte cultures with interleukin-2 (IL-2, PeproTech, Rocky Hill, NJ) at 2 and 10 ng/ml. Additionally, some mice with long-term skin acceptance received subcutaneous injection of 4 μg IL-2 beneath the engrafted donor skin, followed by 2 μg IL-2 daily for 5 consecutive days. Engrafted skin was then monitored for 30 days.

Quantification and Functional Test of Regulatory T Cells (Tregs)

Peripheral blood was sampled and depleted of red cells using ACK lysing buffer. Flow cytometric analyses were performed after intracellular staining of FoxP3 using Mouse Treg Flow^™ Kit (BioLegend). Tregs were quantified by the expression of FoxP3 and CD25 in CD4⁺ cells. Real-time PCR was also used to evaluate FoxP3 expression in splenic CD4⁺ cells enriched by positive selection using Mouse CD4 MicroBeads (Miltenyi Biotec). Total mRNA were extracted from enriched CD4⁺ cells (>95%) with RNeasy Mini Kit (Qiagen K.K., Tokyo, Japan). First-strand cDNA were reversely transcribed from mRNA templates of each sample in equal amount, using MMLV High Performance Reverse Transcriptase Kit with an oligo(dT) primer (EPICENTRE Biotechnologies, Madison, WI). Then, target gene expression was quantified by TaqMan® Gene Expression Assays (Applied Biosystems, Foster City, CA), consisting of unlabeled PCR primers for FoxP3 (Assay ID: Mm00475165_m1) or GAPDH (Assay ID: Mm999 99915_g1) and TaqMan® MGB probe (FAM^™ dye labeled). Normalized value for FoxP3 mRNA expression in each sample was calculated as the relative quantity of FoxP3 divided by the relative quantity of GAPDH.

The inhibitory function of Tregs on alloreactivity was evaluated by MLR. Tregs were enriched from splenocytes of tolerant animals using Mouse CD4⁺CD25⁺ Regulatory T-cell Isolation Kit (Miltenyi Biotec). MLR was conducted in 200 μl RPMI-1640 medium that contained 5 × 10⁴ responder cells from untransplanted FVN/B mice and 2 × 10⁵ irradiated stimulator cells from FVB/N, C57BL/6, or C3H mice with addition of 5 × 10⁴ Tregs from tolerant mice. Proliferation was measured on day 4 as described for MLR.

In Vivo Depletion of Tregs

For in vivo Treg depletion, mice with long-term skin tolerance and loss of peripheral chimerism were intraperitoneally injected with 0.25 mg anti-CD25 (Clone PC61, BioLegend). Tregs were quantified by intracellular FoxP3 staining 1 day before and 1 week after PC61 injection.

Intravenous Injection of Naive Host Splenic Lymphocytes in Tolerant Mice

Naive lymphocytes were enriched from untransplanted FVB/N splenocytes, using NycoPrep 1.077A density gradient. Cells were injected via tail vein in mice with skin tolerance. Engrafted syngeneic and donor skins were then monitored daily.

Statistical Analyses

Because the distribution of donor cell levels was obviously skewed to the right (positive skew), nonparametric tests were used to measure the significance of differences in chimerism levels between two (Mann-Whitney U-test) or more (Kruskal-Wallis H-test) independent groups. Wilcoxon signed-ranks test was used to compare paired medians of chimerism levels at two different time points from recipients of test groups. Spearman's rank correlation coefficient (rho) was computed to evaluate the bivariate correlation of thymic and peripheral chimerism. The equality of means for MLR data, thymic lineage expressions, and Treg proportions was examined by Student's t-test between two independent groups, or one-way ANOVA among three or more groups with post hoc multiple comparison test by least significant difference. As for survival analysis of skin transplants, the survival time was defined by estimating the length of time from the date of skin transplantation to the date of graft rejection or the date of last follow-up. Otherwise, the graft was considered censored at last follow-up. When a mouse died before skin rejection was observed, it was counted as a withdrawal and treated the same way as cases lost to follow-up. Plots of survival time were constructed by Kaplan-Meier method. The log rank test was employed to compare survival curves. Differences were regarded as significant in all tests when p < 0.05.

Results

Analyses of Thymic Chimerism

Thymic biopsy was performed to examine thymic chimerism using donor H-2K^b antigen within 1 day after peripheral chimerism was measured at the age of 1 month. A linear relationship was found between peripheral and thymic H-2K^b+ chimerism (Fig. 1A). Thymic chimerism is estimated at about one third of peripheral chimerism in terms of H-2K^b+ donor cells. T cells (CD3^high, 4⁺, or 8⁺) were the major donor cell type engrafted in the thymus (Fig. 1B), accounting for 50–70% of donor cells in low- or high-level chimeras (Fig. 1C). Regardless of donor or host origin, CD45R, CD11b, CD11c, and Gr1 were usually expressed in <0.5% of thymic cells (Fig. 1B, D). Interestingly, higher levels of peripheral chimerism correlated with lower percentages of thymic H-2K^b-CD3^high cells (Fig. 1D). Overall thymic CD3^high expression was significantly lower in low- and high-level chimeras than in microchimeras and untransplanted controls (Fig. 1E).

Figure 1.

Relationship of thymic chimerism to peripheral chimerism and lineage analyses of thymic donor cells. (A) The scatterplot of peripheral versus thymic H-2K^b+ chimerism is constructed (n = 128) with Spearman's rho of 0.724 (p < 0.001). A fit line (solid) applies to the scatterplot with confidence intervals of 95% (dotted lines). (B) Percentages of H-2K^b+Lin⁺ cells in the thymuses of low- and high-level chimeras are shown as numbers of H-2K^b+Lin⁺ cells divided by total thymic cells acquired. (C) In seven mixed chimeras with thymic chimerism of 2.90–13.01%, the phenotypic distribution among H-2K^b+ donor cells is depicted as numbers of H-2K^b+Lin⁺ cells divided by total H-2K^b+ donor cells. (D) Percentages of H-2K^b-Lin⁺ cells in the thymuses of micro, low- and high-level chimeras are shown as numbers of H-2K^b-Lin⁺ cells divided by total thymic cells acquired. Significant differences in the percentage of H-2K^b-CD3^high cells were observed among them (one-way ANOVA with multiple comparison). (E) Total thymic CD3^high expression is displayed with significant difference among micro-, low-level, high-level chimeras, and untransplanted controls (one-way ANOVA with multiple comparison). All bar data are shown as 95% confidence interval for mean.

Durability of Chimerism, and Association Between Skin Transplantation Timing and Skin Graft Survival

We first collected 48 recipients with peripheral chimerism of 0.01–8.75% at the age of 1 month to determine the durability of their chimerism by monthly flow cytometric examination (Fig. 2A). These data were then used to categorize 147 recipients with first month peripheral chimerism of <10% into three groups of <0.5%, 0.5–1.99%, and ≥ 2% as M1(<0.5), M1(0.5–1.99), and M1(≥2), respectively. To evaluate whether the timing of skin transplantation could influence skin graft survivals, we further subgrouped the recipients in each group arbitrarily by the timing of skin transplantation at the ages of 1 (ST1), 2 (ST2), or 6 months (ST6).

Figure 2.

Durability of peripheral chimerism, and the influence of skin transplantation timing on graft survival. (A) Monthly analyses revealed that peripheral chimerism of <0.5% at 1 month old (M1) faded away by M2 [M2(-), n = 18] and chimerism of 0.5–1.99% by M4–6 [M4–6(-), n = 12 including one with a level of 2.32%]. Chimerism levels of ≥ 2% at M1 sustained circulating donor cells over M6 [M6(+), n = 18]. (B) Peripheral chimerism levels are shown in boxplots. Each chimerism group had similar first month chimerism levels for its subgroups [p = 0.377 for M1(<0.5) in blue, p = 0.927 for M1(0.5–1.99) in red, and p = 0.724 for M1(≥ 2) in green], (C) but significantly lower chimerism levels for ST2 or ST6, compared with ST1 (p < 0.001 for all groups). Circles are outlier values. (D) In Kaplan-Meier survival analyses of donor skin grafts, M1(≥ 2)ST1 had better graft survivals than M1(≥ 2)ST2 (p = 0.009) and M1(≥ 2)ST6 (p = 0.001). M1(0.5–1.99)ST1 and M1(0.5–1.99)ST2 also compared favorably in graft survivals with M1(0.5–1.99)ST6 (p = 0.002 and 0.033, respectively). However, there was no difference in graft survivals among M1(≥ 2)ST2, M1(≥ 2)ST6, M1(0.5–1.99)ST1, and M1(0.5–1.99)ST2. Neither was there significant difference among M1(0.5–1.99)ST6, M1(<0.5)ST1, and M1(<0.5)ST2 despite that they all had better graft survivals compared with untransplanted controls (p < 0.001).

Chimerism levels at 1 month of age were similar among skin transplant subgroups (Fig. 2B), but significantly decreased at the second and sixth month time points of skin transplantation (Fig. 2C). Paired comparison of donor cell levels for each recipient between the first and second or sixth month further indicated that the decline in chimerism was significant (p = 0.001–0.006). Earlier skin placement benefited the survivals of donor skin grafts in groups M1(0.5–1.99) and M1(≥2) (Fig. 2D), but not in group M1(<0.5) despite that two recipients in subgroup M1(<0.5)ST1 had long-term skin tolerance (Table 1). All syngeneic skin grafts survived well during the period of graft monitoring, but third-party skin grafts were rejected within 2–3 weeks without significant difference among any of the subgroups. These findings indicate that skin graft survivals in recipients with transient chimerism are affected by the timing of skin transplantation, but it is limited by chimerism levels at 1 month of age.

Table 1.

Summary of Recipients With Skin Tolerance

	M1(<0.5)		M1(0.5–1.99)			M1(≥2)
	ST1	ST2	ST1	ST2	ST6	ST1	ST2	ST6
Number of tolerant mice with skin survivals >120 days	2	0	4	5	0	16 *	8	3
Number of tolerant mice that lost chimerism by 12 months of age	2	—	4	5	—	8	5	0
Age at the loss of chimerism (months)	2,3	—	3–7	3–6	—	4–8	5–10	—
Number of cases with secondary skin transplantation	—	—	—	5 †	—	—	2	1 ‡

*

Seven with detectable chimerism remained below 12 months of age at the end of this study.

†

Two died within 7 days after secondary skin transplantation due to anesthesia complications and a tight dressing.

‡

Loss of peripheral chimerism at the age of 13 months.

Relationship of Peripheral Chimerism Levels at the Time of Skin Transplantation to Skin Survival

A scatterplot illustrates the relationship between chimerism levels at the time of skin transplantation and graft survival times (Fig. 3A). Survival analyses (Fig. 3B) showed that chimerism levels of >3% consistently correlated with long-term graft acceptance of >120 days, but a level of 0.2–3% only stood a 34% chance of long-term skin tolerance. Recipients with a chimerism level of <0.2% never accepted grafts for >60 days, but their transient graft survivals were still superior to untransplanted controls. This indicates that recipients with peripheral chimerism dropping below 0.2% or even becoming undetectable remained in a hyporesponsive state to the donor. Therefore, the immune response of mixed chimeras to the donor could be presaged as “reliable tolerance,” “possible tolerance,” and “hyporesponsiveness” based upon peripheral chimerism of >3%, 0.2–3%, and <0.2% at the time of graft placement, respectively.

Figure 3.

Relationship of peripheral chimerism at the time of skin transplantation to graft survival times. (A) The scatterplot of donor cell levels at skin placement versus graft survival times shows that skin tolerance required minimal peripheral chimerism of 0.2% at skin transplantation, and was reliably present with peripheral chimerism of >3%. Recipients that died unexpectedly before donor skin rejection was observed are marked as triangles. (B) The probability of skin tolerance (>120 days) in relation to circulating donor cell levels at skin placement is further depicted in Kaplan-Meier survival curves (p < 0.001 for all pairwise comparisons except groups 0.01–0.19 vs. undetectable with p = 0.457).

Follow-up of Peripheral Chimerism and Reappraisal of the Tolerant State by Secondary Skin Transplantation

In total, 38 mixed chimeras tolerated donor skin for more than 120 days (Table 1). By the age of 12 months, 24 tolerant mice lost circulating donor cells, but kept syngeneic and donor-specific skins in good shape (Fig. 4A). Eight of them (Table 1) had their engrafted skins removed (Fig. 4A), and received secondary skin transplants 1 month later, when circulating donor cells remained undetectable, to assess the role of engrafted donor skin in tolerance maintenance. Two mice died within 7 days after skin retransplantation (Table 1). The remaining six had long-term acceptance (>120 days) of secondary syngeneic and donor-specific skin grafts (Fig. 4A), but showed accelerated rejection of secondary third-party C3H skin grafts compared with primary third-party grafts (p = 0.021, data not shown). Engrafted syngeneic, primary, and secondary donor skins had no histological evidence of rejection (Fig. 4B). Engraftment of secondary donor skins did not restore blood cell chimerism. In contrast to these results, 12 mixed chimeras (0.05–2.48% donor cells) with prior rejection of donor skin also had their tolerance reevaluated by secondary skin transplantation. They all rejected secondary donor skins with similar kinetics to primary donor grafts (p = 0.334, data not shown).

Figure 4.

Immunological memory of skin tolerance and alloreactivity of host lymphocytes in MLR. Primary skin transplantation was performed in a representative mixed chimera. (A) Engrafted syngeneic and donor-specific (C57BL/6, black) skin grafts (1 × 1 cm) had good hair growth (on posttransplant day 141, left panel). Neovascularization was present in the engrafted donor skin at its removal (arrow, middle panel). The recipient exhibited long-term acceptance of secondary syngeneic and donor-specific skin grafts (on posttransplant day 150, right panel). (B) Histological examinations show no evidence of mononuclear cell infiltration in engrafted syngeneic (200x, left panel), primary (200x, middle panel), and secondary donor skins (200x, right panel). Engrafted donor skin displays marked melanin deposition in the basal layer (arrows) of epidermis. (C) Stimulation indices (SI) to C57BL/6 stimulators significantly differed among all the groups (p < 0.001, one-way ANOVA). Multiple comparisons reveal that chimeras without skin grafting, chimeras with graft rejection, and peritoneal chimerism(+) had significantly lower SI than untransplanted controls (*p < 0.001), but comparable SI to recipients with skin tolerance (p = 0.371–0.890). (D) In four mixed chimeras (graft survivals of 32, 54, >120, and >120 days) with perceived unresponsiveness to C57BL/6 stimulators as assessed by MLR (p = 0.016, compared with four untransplanted mice by t-test), the supplement of IL-2 in MLR culture caused significant amplifications of SI to FVB/N (p < 0.001, one-way ANOVA), C57BL/6 (p = 0.005), and C3H (p < 0.001) stimulators, as observed in their untransplanted counterparts (p = 0.006, p = 0.001, and p < 0.001, respectively). The SI amplification by IL-2 was does responsive to C3H stimulators, but not to FVB/N and C57BL/6 stimulators. In mixed chimeras, the amplified levels of donor-specific SI by IL-2 was unable to catch up with those in control animals (p = 0.004 and 0.007, t-test), but at a level comparable to control SI in the absence of IL-2 (p = 0.432 and 0.188). All bar data are shown as 95% confidence interval for mean.

Alloreactivity of Host Lymphocytes

Proliferative responses of recipients' lymphocytes in four experimental groups were evaluated by MLR with untransplanted mice as positive controls and recipients with skin tolerance as negative controls (Fig. 4C). Chimeras without skin grafting, which had 0.10–0.46% chimerism at 1 month of age, were sacrificed for MLR with chimerism levels of 0.01–0.18% by 2 month of age. Chimeras with graft rejection (on day 21–30), which received skin transplantation with first month chimerism levels of 0.13–0.66%, were sacrificed by 2 months of age. Recipients without skin transplantation were grouped as peritoneal chimerism(-) when there was no evidence of chimerism in any tissue examined, but grouped as peritoneal chimerism(+) when only peritoneal donor cells were detected (0.43–2.25%) at sacrifice by 2 month of age. All groups exhibited syngeneic nonreactivity but third-party alloreactivity. Low-level chimerism rendered host lymphocytes unresponsive to donor alloantigens even in the cases of prior donor skin rejection, or specific peritoneal lodgment of donor leukocytes. These findings indicate that the lack of response of host lymphocytes in MLR does not fully predict the status of tolerance for allogeneic skin grafting.

Exogenous IL-2 only partially restored the alloreactivity of mixed chimeras in MLR (Fig. 4D). However, subcutaneous injection of IL-2 beneath engrafted donor skins up to an accumulated dose of 14 μg in two tolerant mice failed to cause graft rejection within an observation period of 30 days.

Analyses of Tissue Chimerism

Twelve skin-tolerant mice (Table 2, group A) that lost peripheral chimerism were sacrificed to examine donor cells in other lymphoid or hematopoietic tissues. Among them, three recipients, sacrificed at the ages of 7, 20, and 22 months, showed no detectable donor cells in any tissue examined. These findings argue against the need for persistent tissue chimerism to maintain a tolerant state once skin grafts have taken. Trace donor cells could sometimes be detected in the thymus, lymph node, and peritoneum but were entirely undetectable in the bone marrow and spleen. In comparison with recipients that had rejected donor skins in the categories of “possible tolerance” (Table 2, group B) and “hyporesponsiveness” (Table 2, group C), tolerant mice (group A) exhibited similar or even lower donor cell levels in all tissues examined. Therefore, chimerism in any lymphohematopoietic tissue alone was not sufficient to ensure a durable state of skin tolerance.

Table 2.

Comparison of Tissue Chimerism Between Tolerant and Hyporesponsive Mice

	Group A (n = 12) *	Group B (n = 13) †	Group C (n = 21) ‡
Peritoneum
% Donor cells	0–7.87	0.03–17.47	0–2.25
Mean/median	1.34/0.24	3.46/0.69	0.31/0.11
Undetectable (No.)	5	0	8
Thymus
% Donor cells	0–0.82	0–1.75	0–0.84
Mean/median	0.13/0.03	0.42/0.22	0.10/0.01
Undetectable (No.)	3	3	4
Spleen
% Donor cells	0	0–0.67 ¶	0–0.07
Mean/median	0/0	0.10/0.01	0/0
Undetectable (No.)	12	6	19
Lymph node
% Donor cells	0–0.26	0–1.86	0–0.42
Mean/median	0.04/0	0.27/0.06	0.02/0
Undetectable (No.)	9	5	15
Bone marrow
% Donor cells	0	0–0.50	0–0.29
Mean/median	0/0	0.08/0	0.02/0
Undetectable (No.)	12	8	15
Thymocytes (Thy1) §
% Donor cells	0–0.38	0–5.67	0–0.75
Mean/median	0.06/0.01	1.39/0.24	0.09/0
Undetectable (No.)	3/9	4/11	11/19

*

Mice tolerated donor skin for >120 days, but eventually lost peripheral chimerism. Nine were sacrificed at the age of 7–11 months and three at the age of 19–22 months.

†

Mice rejected donor skin after skin transplantation at peripheral chimerism of 0.23–2.09% and were sacrificed at the age of 1.5–3.5 months after donor skin rejection.

‡

Mice rejected donor skin after skin transplantation at peripheral chimerism of 0.01–0.15% and were sacrificed at the age of 1.5–2.5 months after donor skin rejection.

§

Three in group A and two each in groups B and C were not examined for Thy1 thymocytes.

¶

p = 0.022, Mann-Whitney U-test, compared with group A.

Tregs in Mice with Skin Tolerance

We collected mice with skin tolerance but loss of peripheral chimerism, and mice with donor skin rejection to analyze Treg numbers among their CD4⁺ T cells. Tolerant mice had significantly higher FoxP3 expression among peripheral and splenic CD4⁺ T-cell populations than nontolerant and untransplanted mice (Fig. 5A–C), suggesting increased Tregs in tolerant mice. Functional assays showed that Tregs from tolerant mice inhibited proliferative alloresponses in a nonspecific manner (Fig. 5D). However, in vivo depletion of Tregs (Fig. 5E) in three tolerant mice failed to cause rejection of engrafted donor skins within an observation period of 30 days. In contrast, intravenous infusion of 5–7.5 × 10⁷ naive host lymphocytes (23.96% T cells) abolished the tolerant state as evidenced by rapid rejection of engrafted donor skins within 11–19 days (Fig. 5F) as well as failure to accept secondary donor skins transplanted at least four months after lymphocyte infusion.

Figure 5.

FoxP3 expression in tolerant mice. (A) FoxP3 expression is concurrent with expression of CD25 on peripheral CD4⁺ T cells. (B) Tregs (CD25⁺FoxP3⁺CD4⁺) were quantified in untransplanted mice (control), recipients with skin rejection (rejection), and recipients with skin tolerance (tolerance). The expression of CD25⁺FoxP3⁺ in circulating CD4⁺ cells significantly differed among the three groups (p < 0.001, one-way ANOVA) with a significantly higher ratio in tolerant mice than in the other two groups. (C) FoxP3 mRNA expression in splenic CD4⁺ T cells also significantly differed among the three groups (p < 0.001, one-way ANOVA), wherein tolerant mice showed higher FoxP3 expression than control and rejection groups. (D) Tregs from tolerance mice exhibited inhibitory effects on lymphocyte proliferation of untransplanted FVB/N mice to both C57BL/6 and C3H stimulators in MLR. All bar data are shown as 95% confidence interval for mean. (E) In vivo Treg depletion was achieved by intraperitoneal injection of PC61, but failed to cause rejection of engrafted donor skin in three mice with long-term skin tolerance. (F) Intravenous infusion of naive host lymphocytes led to rapid rejection of engrafted donor skin on day 19.

Discussion

As the strategy of hematopoietic chimerism is contemplated for therapeutic induction of donor-specific tolerance, a critical question important to design of such a therapy is what level of donor cell expression is needed to achieve tolerance. Our data show that peripheral chimerism of >3% consistently conferred lasting skin tolerance. Thus, postnatal therapy based upon prenatal marrow transplantation appears feasible within a window of opportunity afforded by the presence of sufficient circulating donor cells. The gray zone of success in tolerance induction within 0.2–3% peripheral chimerism may explain the conflicting association between tolerance induction and low-level chimerism observed in previous studies (5,12,34), and also fits with the earlier observation that chimeric mice could tolerate donor skin with the presence of residual donor cells following incomplete abolition of chimerism by antidonor MHC antibodies (22). Taniguchi et al. (41) reported that peripheral chimerism of >10% is required for graft tolerance. This threshold level was likely to be an overestimate because there was a significant delay in skin grafting of 3–4 months after marrow transplantation. Robust skin tolerance also relates to splenic chimerism of 10% in mice, but the chance for graft acceptance drops to half with approximately 5% chimerism (14). Measuring splenic chimerism is less clinically applicable because it necessitates an invasive procedure. Moreover, a major concern in these studies is the difficulty in evaluating the confounding effects of preconditioning or immunosuppressive regimens on chimerism levels and test graft survivals.

Previously we have demonstrated that only durable peripheral chimerism lasting at least 6 months stood a chance of inducing tolerance to skin grafts placed at 6 months of age (6). Herein, we extend this observation and show that the timing of skin placement influenced graft acceptance with the critical parameter being the chimerism level at skin placement as opposed to a higher chimerism level at an earlier point in life prior to skin grafting. As a result, high-level chimeras still rejected donor skins once they had lost chimerism by the time skin grafting was performed (42). Two possible mechanisms can be envisioned for this observation: tolerance fades along with a decline or loss of blood chimerism or complete tolerance fails to develop at lower or undetectable levels of chimerism. Although the belief has long been that peripheral chimerism is required for tolerance maintenance (22,35), engrafted donor skins in our study persisted despite spontaneous regression of peripheral chimerism. This argues against the necessity of circulating donor cells for enduring tolerance. Kanamoto et al. (19) reported that engrafted H-2^d skin was rejected after depletion of H-2^b/d donor cells with H-2^a host lymphocytes specifically sensitized to H-2^b haploidentical alloantigens but naive to H-2^d. Although skin rejection was ascribed to depletion of chimeric cells, our study shows that naive host lymphocytes could still cause skin rejection. Thus, the observed break of tolerance results from MHC incompatibility between the injected lymphocytes and engrafted donor skin rather than depletion of chimeric cells. In fact, enduring graft tolerance following the loss of peripheral chimerism was observed in nonhuman primates and in humans treated with simultaneous marrow and renal transplants (20,21,37), and ascribed to alloantigens present in the form of the surviving organ grafts that helped to maintain tolerance (20,37,45). In this study, the removal of engrafted primary donor skin in tolerant mice that had lost peripheral chimerism failed to break the state of tolerance. The accelerated rejection of secondary third-party C3H skin suggested sensitization of recipients' immunity to H-2^k alloantigens, and proved the recipient's immunocompetence after skin retransplantation. Our results support that the failure of primary skin acceptance after spontaneous or experimental depletion of peripheral chimerism was not attributable to the loss of tolerance but rather to the incapability of establishing sufficient tolerance at lower or undetectable chimerism levels. Therefore, complete tolerance does not develop until successful skin engraftment has taken place at a sufficient level of hematopoietic chimerism.

In this regard, long-term complete tolerance resulted from the immunological consequences of hematopoietic chimerism and skin grafting. This argument is strengthened by the fact that recipients losing peripheral chimerism did not accept donor skin unless there was prior donor skin engraftment. Primary skin grafting made recipients capable of accepting secondary skin transplants even in the absence of detectable chimerism. Hematopoietic chimerism not only rendered host lymphocytes specifically unresponsive to donor alloantigens in MLR but also prevented hosts from rejecting secondary donor skin at an accelerated speed. These results support the concept that donor cell chimerism attenuates or abolishes donor-specific T-cell alloreactivity so as to facilitate the initial induction of tolerance through the tolerogenic effects of donor skin. It essentially agrees with the proposition that inhibition of initial graft rejection by sufficient immunosuppression may allow the tolerogenic properties of organ allografts to eventually prevail (2,15,40). It appears that blood chimerism is not solely responsible for tolerance induction, but mechanistically has important immunomodulatory effects on tolerance induction in immunocompetent mice, as opposed to an immunosuppressed rat model of cyclosporine-induced tolerance (24). These observations mirror the phenomenon that transfusion or adoptive transfer of donor leukocytes with solid organ transplants induces prolonged allograft survivals instead of long-term graft tolerance (10,36,43,44,49).

Clonal deletion is well documented following tolerance induction through IUT (16), and occurs effectively at a level as low as 1% intrathymic donor cells (42), which correlates well with a threshold of 3% peripheral chimerism in this study. The reduced CD3^high expressions in the thymus of tolerant mice may represent another aspect of intrathymic T-cell deletion. These results, together with a linear correlation between peripheral and thymic chimerism, strongly support the rationale for peripheral chimerism as a surrogate marker of immune tolerance. Thus, a relevant effect of blood chimerism on tolerance induction appears to be its role in central tolerance induction, but persistent peripheral chimerism is not relevant to the maintenance of skin tolerance. An important question raised is whether enduring skin tolerance relies upon chimerism outside the peripheral circulation. Specific peritoneal lodgment of donor cells is commonplace following IUT (7–9). It reduces donor-specific alloreactivity of host lymphocytes in MLR, but only minimally prolongs survivals of donor skin grafts (6). Although tolerant mice could have a trace of donor cells in the thymus at 7–8 months old (6), longer term follow-up in this study did not reveal detectable donor cells in any tissue examined. Hyporesponsive mice with delayed donor skin rejection did not always lose tissue chimerism. These animals could keep a vestige of donor cells, comparable to or even at higher levels than tolerant animals. These observations indicate a dissociation between tissue chimerism and enduring skin tolerance, arguing against the idea that low-level tissue chimerism is solely responsible for the maintenance of complete tolerance.

As a vital mediator of peripheral tolerance in many systems, Tregs are regarded as essential for tolerance maintenance in situations where high-level chimerism is not established or sustained in rodent models of postnatal bone marrow transplantation (26). However, reports pertaining to the role of Tregs in the maintenance of hematopoietic chimerism and tolerance have varied greatly, depending upon the protocols or strategies used in their studies (47,48). T-cell clonal anergy may represent another mechanism of peripheral tolerance (1,13,50,51), wherein anergic T cells show profound defects in their immune responses with inability to secret IL-2 and proliferate upon antigen recognition (13). Although tolerant mice in this study exhibited increased expression of Tregs with alloreactivity suppression and partial reversal of alloreactivity by exogenous IL-2 in MLR, neither in vivo Treg depletion nor IL-2 injection beneath the engrafted donor skin caused graft rejection. The dissociation between in vitro and in vivo systems failed to support Tregs or clonal anergy as the sole or major role in the maintenance of skin tolerance. The partial reversal of alloreactivity by IL-2 may have an important implication for dampening of IL-2 signaling, which opens a new direction for further experimental elucidation of tolerance maintenance.

In conclusion, hematopoietic chimerism generated by prenatal marrow transplantation is immunologically relevant to induction of graft tolerance, but not to tolerance maintenance. We propose that hematopoietic chimerism causes a state of donor-specific immunosuppression to facilitate the establishment of complete skin tolerance. These findings encourage further development of fetal hematopoietic transplantation as a preconditioning regime for postnatal cell or organ transplantation to treat birth defects.