Constraints in antigen presentation severely restrict T cell recognition of the allogeneic fetus

Abstract

How the fetus escapes rejection by the maternal immune system remains one of the major unsolved questions in transplantation immunology. Using a system to visualize both CD4+ and CD8+ T cell responses during pregnancy in mice, we find that maternal T cells become aware of the fetal allograft exclusively through "indirect" antigen presentation, meaning that T cell engagement requires the uptake and processing of fetal/placental antigen by maternal APCs. This reliance on a relatively minor allorecognition pathway removes a major threat to fetal survival, since it avoids engaging the large number of T cells that typically drive acute transplant rejection through their ability to directly interact with foreign MHC molecules. Furthermore, CD8+ T cells that indirectly recognize fetal/placental antigen undergo clonal deletion without priming for cytotoxic effector function and cannot induce antigen-specific fetal demise even when artificially activated. Antigen presentation commenced only at mid-gestation, in association with the endovascular invasion of placental trophoblasts and the hematogenous release of placental debris. Our results suggest that limited pathways of antigen presentation, in conjunction with tandem mechanisms of immune evasion, contribute to the unique immunological status of the fetus. The pronounced degree of T cell ignorance of the fetus also has implications for the pathophysiology of immune-mediated early pregnancy loss.

Figure 1. Spatiotemporal patterns of OT-I T cell recognition of fetal/placental OVA.

B6CBAF1 females mated to Act-mOVA or B6 control males were adoptively transferred with CFSE-labeled B6CBAF1 OT-I T cells. (A) OT-I proliferation in secondary lymphoid organs. OT-I cells were visualized by flow cytometry at the end of their indicated 2-day window of exposure. The percentage of cells undergoing more than 1 cell division is noted; such cells were only rarely observed in B6-mated females at any stage of gestation. (B) Midgestational onset of OT-I proliferation in secondary lymphoid organs. Act-mOVA–mated female mice at different stages of gestation were transferred with OT-I cells and then sacrificed 2 days later. Using raw values for proliferation determined as in A, the percentage of OT-I cells in antigen-driven proliferation was calculated by subtracting mean B6-mated control values for each of the 3 secondary lymphoid organ sets in a given experiment from their respective values in individual Act-mOVA–mated mice. The graph shows all data for n = 21 Act-mOVA–mated females over 5 independent experiments. (C) Absence of OT-I reactivity at the maternal/fetal interface in early gestation. Transferred OT-I cells were visualized in the uterine LNs and in dissected decidua/placentas pooled from several implantation sites. Many of these cells lie along the x axis of the plot and so are difficult to see. We therefore note the total number of OT-I cells recovered from these sites. Percentages indicate the fraction of OT-I cells with a CFSE^undilutedCD69^lo naive phenotype (cells in the lower-right quadrant).

Figure 2. OVA shed from the conceptus accumulates on FDCs.

(A–G) OVA accumulation on splenic FDCs. (A–D) E17.5 spleens from Act-mOVA–mated (A and B) or B6-mated (C and D) mice double stained for OVA (red; A and C) and CD35 (green; B and D). Red pulp is apparent as areas of diffuse green staining resulting from rbc rendered artifactually fluorescent by the immunohistochemistry protocol. Data are representative of n = 8–9 mice per group on E15.5–E17.5. Scale bar: 0.5 mm. (E–G) High-power confocal photomicrographs of a double-stained OVA⁺ focus, showing OVA staining (E), CD35 staining (F), and the merged image (G). (H–J) OVA accumulation on LN FDCs. (H) Two subcutaneous LNs from an Act-mOVA–mated mouse on E17.5; a higher magnification of the OVA⁺ focus indicated with an arrow is shown in I. Data are representative of all n = 8 mice analyzed at E15.5–E17.5. (J) An OVA⁺ focus in a uterine LN on E11.5, as such foci begin to appear. Similar foci were seen in n = 3 of 4 mice analyzed. Scale bar: 1 mm (H) and 0.2 mm in (I and J). As in C, LNs from B6-mated mice showed no OVA⁺ foci (data not shown).

Figure 3. Act-mOVA expression at the maternal/fetal interface.

Sections of implantation sites of Act-mOVA (A, C, E, and F) or nontransgenic concepti (B and D) stained with anti-OVA antibodies (red). C–F also show artifactually fluorescent maternal red blood cells (green) to demarcate maternal blood vessels, and DAPI-stained nuclei (blue). (A and B) E9.5 implantation sites show high OVA expression in transgenic concepti (n = 9) confined to trophoblast giant cells (arrowheads) and trophoblasts at the invasive front of the ectoplacental cone (epc; arrow) abutting the maternal decidua (dec). Scale bar: 1 mm. (C–E) Higher-power magnifications show that transgenic trophoblasts expressing high levels of OVA had not yet invaded maternal arteries by E9.5 (C and D) but were lining maternal vessels (arrows) by E10.5 (E; n = 5). Transgenic trophoblasts expressed low levels of OVA in the body of the ectoplacental cone and developing labyrinth (lab), where they also contact maternal blood (compare the purple-tinged color of these structures in C and E, reflecting a combination of OVA staining and the DAPI counterstain, with their blue color in D). Fetal rbc appear blue-tinged due to their nucleation (arrow in D); the asterisk in C marks the vestigial uterine lumen. Scale bar: 0.4 mm. (F) At E13.5, transgenic trophoblasts expressing high levels of OVA were closely associated with or lining (arrowheads) maternal arteries deeper in the decidua (n = 3). By late gestation, high OVA expression was also apparent in the spongiotrophoblast layer (data not shown). Scale bar: 0.2 mm.

Figure 4. Indirect presentation of fetal/placental OVA to maternal OT-I T cells.

(A) OT-I cells on a B6CBAF1 background are able to detect K^b/SIINFEKL complexes in (bm1 × CBA)F₁ mice. Splenocytes from (bm1 × CBA)F₁ or B6CBAF1 female virgins were analyzed 2 days after the injection of CFSE-labeled B6CBAF1 OT-I cells, with or without a concurrent injection of SIINFEKL-loaded B6CBAF1 splenocytes. Histograms were normalized to account for the loss of OT-I cells in (bm1 × CBA)F₁ mice. (B) Lack of OT-I proliferation in (bm1 × CBA)F₁ females mated to Act-mOVA males. Splenocytes from B6- or Act-mOVA–mated (bm1 × CBA)F₁ females were analyzed 2 days after the injection of CFSE-labeled B6CBAF1 OT-I cells on or after E12.5. Data are representative of n = 10 Act-mOVA–mated (bm1 × CBA)F₁ and n = 3 B6-mated mice in 3 independent experiments. Identical results were obtained with LN cells.

Figure 5. Maternal BM3 cells directly alloreactive with paternal H-2K^b show immune ignorance during gestation.

CBA females mated to either CBA males or B6 males were adoptively transferred with CFSE-labeled BM3 T cells in during the first half of gestation (E5.5–E10.5). Splenocytes and implantation sites (pooled decidua/placentas) were analyzed by flow cytometry either 2 days later (early gestation, E7.5–E9.5) or 6–7 days later (late gestation, E13.5–E17.5). Percentages indicate the fraction of BM3 cells with CFSE^undilutedCD69^lo or CFSE^undilutedCD44^lo naive phenotypes (cells in the lower-right quadrant). The total number of BM3 cells recovered from implantation sites is noted. Some mice were also stimulated by the intravenous injection of 5 × 10⁶ B6CBAF1 splenocytes at the times indicated before sacrifice. Implantation sites are not shown for these mice since they showed a dramatic loss of BM3 cells from blood and implantation sites, likely due to their retention within secondary lymphoid organs. Data are representative of n = 4–7 mice per group in 2 independent experiments. In additional experiments, n = 6 CBA females mated to B6 males also failed to show evidence of TCR engagement in transferred BM3 cells when assayed for proliferation alone.

Figure 6. Maternal OT-II CD4⁺ T cells respond to fetal/placental OVA via indirect presentation.

(A) OT-II cell proliferation in secondary lymphoid organs. B6CBAF1 females mated to Act-mOVA or B6 control males were adoptively transferred with CFSE-labeled B6CBAF1 OT-II T cells. OT-I cells were visualized by flow cytometry at the end of their indicated 4-day window of exposure. The percentage of cells undergoing more that 1 cell division is noted; such cells were only rarely observed in B6-mated females at any stage of gestation. Data are representative of n = 6 Act-mOVA–mated mice and n = 6 B6-mated mice over 3 independent experiments. (B) Indirect presentation of fetal/placental OVA. Splenocytes from B6- or Act-mOVA–mated B6 or B6.C-H2^bm12 females were analyzed 4 days after the injection of CFSE-labeled OT-II cells late in gestation (on or after E13.5). Data are representative of 2 independent experiments, encompassing n = 5 Act-mOVA–mated B6.C-H2^bm12 mice, n = 5 Act-mOVA–mated B6 mice, and n = 2–4 mice each in the 2 other mating groups. Identical results were obtained with LN cells.

Figure 7. Defective OT-I cell priming in pregnant mice bearing Act-mOVA concepti.

(A) Impaired cell accumulation. The number of OT-I cells remaining after transfer into Act-mOVA–mated females was calculated as a percentage of total CD8⁺ T cells, then normalized to the number of cells remaining after equivalent time periods in B6-mated mice injected in parallel. Similarly, the number of OT-I cells remaining in virgin females treated with OVA or OVA plus anti-CD40 antibodies and poly(I:C) was normalized to the number remaining in untreated virgins. Each point represents mean ± SEM of n = 3–10 Act-mOVA–mated or n = 2–3 virgin mice. (B) Impaired modulation of CD62L and CD25 expression. OT-I cells were analyzed either 7 days after transfer into E10.5 Act-mOVA–mated or in virgin mice 3 days after treatment with OVA plus anti-CD40 antibodies and poly(I:C). Plots are representative of 3 independent experiments, each with individual mice or pools of n = 2–5 Act-mOVA–mated and n = 1–3 virgin mice per experiment. Plots display similar cell numbers in the 2 groups of mice, even though this does not represent their relative prevalence in vivo. The percentage of OT-I cells with altered marker expression is shown. (C and D) Impaired cytokine expression. Seven days after transfer into E10.5 pregnant mice, OT-I cells in the spleen (C) or uterine LNs (D) were stained for intracellular cytokines. Virgin females were similarly analyzed 7 days after indicated treatments given at the time of OT-I transfer. The percentage of OT-I cells with a CFSE^locytokine⁺ phenotype is shown. Data are representative of 2 independent experiments, each with a pool of n = 4 Act-mOVA–mated mice. The other treatment groups show data representative of individual mice or pools encompassing a total of n = 2–6 mice per group.

Figure 8. Defective OVA-specific cytotoxic responses in Act-mOVA–mated mice.

In vivo cytotoxicity assays were performed by injecting 50:50 mixtures of CFSE^hi SIINFEKL-pulsed B6CBAF1 splenocytes (target cells) and CFSE^lo unpulsed splenocytes (control cells). The ratio (r) was calculated as the number of CFSE^hi to CFSE^lo cells remaining 16 hours later; low ratios imply high levels of SIINFEKL-specific killing. (A) Low levels of SIINFEKL-specific cytotoxicity in Act-mOVA–mated mice receiving OT-I cells. E10.5–E11.5 pregnant or virgin females injected with OT-I cells were treated as indicated, then assayed 6 days later. Data are representative of 2 independent experiments, encompassing n = 4 Act-mOVA–mated mice. (B) Absent SIINFEKL-specific cytotoxicity in Act-mOVA–mated mice with unmanipulated T cell repertoires, unless treated with anti-CD40 antibodies and poly(I:C). E10.5–E13.5 pregnant or virgin females were treated as indicated, then assayed 6 days later. Data are from of 1 of 6 independent experiments. Each histogram shows representative data for a treatment group, except in the case of Act-mOVA–mated females treated with anti-CD40 antibodies and poly(I:C), for which results for the mouse with the highest cytotoxicity are shown. The total number of mice in each group over all experiments is indicated. (C) CTL responses in Act-mOVA–mated females treated with CD40 antibodies and poly(I:C). The percentage of induced SIINFEKL-specific killing was calculated by normalizing the CFSE^hi to CFSE^lo ratio for individual anti-CD40/poly(I:C)-treated mice to the mean value of this ratio for their respective nonadjuvant-treated group. Act-mOVA–mated mice were assayed in 3 independent experiments; the number of Act-mOVA concepti/total concepti for each of these mice is shown.