Frontline Science: Myeloid-derived suppressor cells (MDSCs) facilitate maternal-fetal tolerance in mice

Abstract

During successful pregnancy, a woman is immunologically tolerant of her genetically and antigenically disparate fetus, a state known as maternal-fetal tolerance. How this state is maintained has puzzled investigators for more than half a century. Diverse, immune and nonimmune mechanisms have been proposed; however, these mechanisms appear to be unrelated and to act independently. A population of immune suppressive cells called myeloid-derived suppressor cells (MDSCs) accumulates in pregnant mice and women. Given the profound immune suppressive function of MDSCs, it has been suggested that this cell population may facilitate successful pregnancy by contributing to maternal-fetal tolerance. We now report that myeloid cells with the characteristics of MDSCs not only accumulate in the circulation and uterus of female mice following mating but also suppress T cell activation and function in pregnant mice. Depletion of cells with the phenotype and function of MDSCs from gestation d 0.5 through d 7.5 resulted in implantation failure, increased T cell activation, and increased T cell infiltration into the uterus, whereas induction of MDSCs restored successful pregnancy and reduced T cell activation. MDSC-mediated suppression during pregnancy was accompanied by the down-regulation of L-selectin on naïve T cells and a reduced ability of naïve T cells to enter lymph nodes and become activated. Because MDSCs regulate many of the immune and nonimmune mechanisms previously attributed to maternal-fetal tolerance, MDSCs may be a unifying mechanism promoting maternal-fetal tolerance, and their induction may facilitate successful pregnancy in women who spontaneously abort or miscarry because of dysfunctional maternal-fetal tolerance.

Keywords: G-CSF; T cell activation; immune suppression; implantation; myeloid cells; pregnancy.

Figure 1.. Circulating immune-suppressive granulocytic MDSCs (PMN-MDSCs) increase in pregnancy.

Female BALB/c mice were caged overnight or in the morning with male C57BL/6 mice. The following morning or afternoon, respectively, females were monitored for vaginal plugs (gestation d E0.5). (A) Circulating blood leukocytes were stained for MDSC markers CD11b, Gr1, Ly6C, and Ly6G. MDSCs are CD11b⁺Gr1⁺; PMN-MDSCs are CD11b⁺Ly6G^hiLy6C^low/−; monocytic MDSCs are CD11b⁺Ly6G^low/−Ly6C^hi. n = 4 mice for each time point. (B) Pregnancy-induced MDSCs (P-MDSC; >82.5% CD11b⁺Gr1⁺) and non-MDSCs (CD11b⁻Gr1⁻ cells) were purified from d E15.5 pregnant mice (n = 2) or from tumor-bearing mice (T-MDSCs, >90% CD11b⁺Gr1⁺ cells) and were cocultured with peptide-activated transgenic T cells at a ratio of 1:1 MDSC:T cells. Data are the means of 4 replicates from 1 of 2 independent experiments. Data were assessed for statistical significance with the Student’s t test. **P < 0.01. Error bars represent sd.

Figure 2.. MDSCs are essential for implantation and subsequent pregnancy.

Female BALB/c mice were caged overnight with male C57BL/6 mice. Plugged females were either control or Ab-depleted for MDSCs at the indicated time points. Mice were followed for delivery of live offspring. Data are pooled from 3 independent experiments in which females were mated and plugged over the course of several days. Numbers of mice per group are indicated in the figure. Statistical significance was determined with the Fisher’s exact test. 1, According to the Jackson Laboratory, the normal pregnancy rate for BALB/c mice is 31–44% of plugged females (http://jaxmice.jax.org/jaxnotes/archive/501d.html; §, Groups that are not significantly different from each other and are significantly different from groups without § (P < 0.0001). Groups without § are not significantly different from each other.

Figure 3.. Depletion of MDSC causes resorption of concepti and recruitment of CD3⁺ cells in the uteri of d E7.5 pregnant female mice.

BALB/c females were mated with C57BL/6 males, and plugged mice were depleted for MDSC or treated with an irrelevant Ab on E0.5 and E4.5. Mice were sacrificed on E7.5 and their uteri removed. (A) Uteri from plugged MDSC-depleted, control-depleted, and nonpregnant mice. Black arrows indicate viable concepti; red arrows indicate resorbing concepti. Not all concepti are indicated. (B and C) Uteri of (A) were sectioned and stained by immunohistochemistry for MDSCs (anti-Gr1 mAb plus hematoxylin) (B) or CD3⁺ cells (anti-CD3 mAb plus hematoxylin) (C). Boxed regions were further examined under ×10, ×25, and/or ×40 objectives. (B) Arrows indicate MDSCs. (C) Arrows indicate CD3⁺ cells. Only a few CD3⁺ cells are indicated. (C) Graph indicates the means ± sd of CD3⁺ cells per ×40 field. Boxed areas are the magnified regions. Images are from representative mice from a cohort of 5 mice per treatment.

Figure 4.. G-CSF induces immune-suppressive MDSCs in vivo.

(A) G-CSF increases MDSCs in mice. Mice were given PBS or 250–1000 μg/kg G-CSF for 4 d, bled on d 5, and the circulating leukocytes were stained for the MDSC markers Gr1 and CD11b. (B) G-CSF–induced Gr1⁺CD11b⁺ cells are immune suppressive. TCR transgenic, Ag-specific DO11.10 or CLN4 splenocytes plus cognate peptide were cocultured with blood or splenic MDSCs from G-CSF–treated or 4T1–tumor-bearing BALB/c mice and were analyzed for activation by incorporation of [³H]thymidine. Data are means ± sd of quadruplicates from 1 of 2 independent experiments. (C) G-CSF–induced MDSCs persist in vivo. BALB/c females were inoculated i.p. with 500 µg/kg G-CSF on 3 successive d, and MDSC levels in the blood were measured 3 and 6 d after the last dose of G-CSF. Data are means ± sd of 3 mice/group and are representative of 3 independent experiments. Statistical analysis is by ANOVA. *P < 0.05; **P < 0.01. Error bars represent sd.

Figure 5.. Induction of MDSCs in mice impaired for maternal–fetal tolerance restores successful pregnancy by inhibiting T cell activation.

Female BALB/c mice were caged overnight or in the morning with male C57BL/6 mice. The following morning or afternoon, respectively (d E0.5), plugged females were placed on water ± 2 mg/ml 1-MT ± G-CSF (1000 µg/kg/d) ± MDSCs or control Ab depletion ± CD4 plus CD8 T cell depletion, or untreated, and followed for delivery of live offspring. Mice depleted for CD4⁺ or CD4⁺ and CD8⁺ T cells were started on T cell–depleting Abs before mating. Data are pooled from 2–5 independent experiments in which multiple females were mated and plugged over the course of several days. Numbers of mice/group are indicated in the figure. Statistical analysis was with Fisher’s exact test. 1, According to the Jackson Laboratory, the normal pregnancy rate for BALB/c mice is 31–44% of plugged females (http://jaxmice.jax.org/jaxnotes/archive/501d.html; §, Groups that are not significantly different from each other and are significantly different from groups without § (P < 0.004). Groups without § are not significantly different from each other.

Figure 6.. Pregnancy-induced MDSCs down-regulate L-selectin on circulating, naïve T cells and prevent T cell activation in the periphery.

(A) Pregnancy-induced MDSCs down-regulate L-selectin on naïve, circulating T cells. Pregnant, not-pregnant, and plugged–Gr1-depleted BALB/c female mice were bled, and their leukocytes analyzed for the percentage of MDSCs, and their CD3⁺ cells were analyzed for L-selectin. (Left top panels) L-selectin staining of CD3⁺ cells from a representative mouse from each treatment group. (Right top panel) Average L-selectin levels for 3 not-pregnant, not-plugged; 3 pregnant; and 5 plugged, MDSC-depleted mice. (Bottom panel) T cell level of L-selectin vs. MDSC level. Each circle represents an individual mouse. (B) C57BL/6 mice were administered G-CSF on d 1–3. One group was depleted with Gr-1 Ab on d 1 and 3. Mice were sacrificed on d 4, and blood, spleen, and superficial inguinal, brachial, and axillary lymph nodes were analyzed for the percentage of MDSCs (Gr1⁺CD11b⁺ cells). CD3⁺ T cells were gated and analyzed for L-selectin. Data are means ± sd of 5 mice/group. *P < 0.05; **P < 0.01. Error bars represent sd. FS, forward scatter.

Figure 7.. G-CSF–induced MDSCs down-regulate L-selectin on peripheral T cells and reduce T cell activation.

(A) MDSCs in pregnant mice reduce T cell activation in vivo. PBMCs were harvested from not-plugged, not-pregnant; pregnant; and plugged, MDSC-depleted mice. CD3⁺ T cells were gated and analyzed for T cell activation markers CXCR3, CD44, and CCR7. Data are means ± sd pooled from 3 not-pregnant, not-plugged; 3 pregnant; and 5 plugged, MDSC-depleted mice. Statistical analysis was with Student’s t test. (B) C57BL/6 (CD45.1) mice were administered G-CSF on d 1–4, adoptively transferred with transgenic OT-1 T cells (CD45.2; SIINFEKL specific) on d 2, and given cognate peptide on d 3. One group was also depleted with Gr1 Abs. Mice were sacrificed on d 5, and blood, spleen, and lymph nodes were analyzed for MDSCs and for activated (CXCR3⁺) CD3⁺ OT1 T cells. Upper panels are representative of individual mice; lower panels are the means ± sd of 3 independent experiments using 5 mice/group. OT1 cells were either transferred into CD45.1 C57BL/6 mice or OT1 cells were violet or CFSE-labeled and transferred into C57BL/6 CD45.2 mice. Statistical analysis was with Student’s t test. *P < 0.05. Error bars represent sd.