Uterine NK cells regulate endometrial bleeding in women and are suppressed by the progesterone receptor modulator asoprisnil

Abstract

Uterine NK cells (uNK) play a role in the regulation of placentation, but their functions in nonpregnant endometrium are not understood. We have previously reported suppression of endometrial bleeding and alteration of spiral artery morphology in women exposed to asoprisnil, a progesterone receptor modulator. We now compare global endometrial gene expression in asoprisnil-treated versus control women, and we demonstrate a statistically significant reduction of genes in the IL-15 pathway, known to play a key role in uNK development and function. Suppression of IL-15 by asoprisnil was also observed at mRNA level (p < 0.05), and immunostaining for NK cell marker CD56 revealed a striking reduction of uNK in asoprisnil-treated endometrium (p < 0.001). IL-15 levels in normal endometrium are progesterone-responsive. Progesterone receptor (PR) positive stromal cells transcribe both IL-15 and IL-15RA. Thus, the response of stromal cells to progesterone will be to increase IL-15 trans-presentation to uNK, supporting their expansion and differentiation. In asoprisnil-treated endometrium, there is a marked downregulation of stromal PR expression and virtual absence of uNK. These novel findings indicate that the IL-15 pathway provides a missing link in the complex interplay among endometrial stromal cells, uNK, and spiral arteries affecting physiologic and pathologic endometrial bleeding.

Figure 1. Expression of PR is altered following treatment with asoprisnil

A: Control secretory endometrium with strong expression of progesterone receptor (PR) in the endometrial stromal cells (S) and negligible expression in glandular epithelium (G). B: PR expression is lost in endometrial stroma (S) and up-regulated in both surface and glandular epithelium (G) in asoprisnil-treated endometrium. Arrows point to spiral arteries. scale bar = 50 microns. The effect of asoprisnil on CD56⁺ uNK cells and spiral arteries C and D: The effect of asoprisnil on CD56⁺ uNK and spiral arteries. Endometrial samples immunostained for CD56⁺ uterine NK (uNK). C: CD56⁺ uNK cells are scattered throughout stroma in placebo-treated control secretory phase endometrium; scale bar = 50 microns D: Virtually no CD56⁺ cells are observed following treatment with asoprisnil; scale bar = 50 microns. E and F: Endometrium samples stained for αSMA expression and collagen. E: αSMA expression in spiral artery after treatment with 25 mg asoprisnil shows thickening of the media; scale bar = 50 microns. F: Collagen in spiral arteries after treatment with 10 mg asoprisnil shows thickening of the arterial wall; scale bar = 50 microns. G: αSMA expression in spiral artery in normal secretory phase endometrium; scale bar = 50 microns H: Collagen in spiral arteries in normal secretory phase endometrium; scale bar = 50 microns

Figure 2. Gene expression following administration of asoprisnil centres on IL-15

Heat map for hierarchical clustering of 245 statistically significant (adjusted p<=0.05) and highly up-regulated or down-regulated (>= 5 fold) genes in comparisons of placebo with 10 mg asoprisnil or placebo with 25 mg asoprisnil. Data were derived from endometrial samples. This heat map represents an overview of gene expression patterns contained amongst the set of significant genes; it is not an independent machine learning analysis of the complete data set. Blue represents low expression, and red represents high expression. Treatment group is shown in the top grey bar: light grey=placebo, dark grey=10 mg asoprsinil, black=25 mg asoprisnil. Hierarchical clustering was performed on genes, not on samples. Dendrogram branches partition genes into sets with similar gene expression across all samples in the study. Row labels are Affymetrix probe IDs followed by official gene symbol.

Figure 3. Biological networks associated with immune cell function and apoptosis predominantly affected by asoprisnil

Most genes within networks are significantly down-regulated (green), only few are up-regulated (red). Central to the network and notably down-regulated is IL-15.

Figure 4

A: IL-15 mRNA is down-regulated after treatment with asoprisnil. Endometrial IL-15 mRNA expression was measured by QRT-PCR in placebo-treated secretory phase endometrial samples and samples from women treated with 10 or 25 mg asoprisnil. IL-15 mRNA expression was significantly down-regulated (p<0.05) by 25 mg asoprisnil compared to placebo/secretory phase samples B: IL-15 expression throughout the menstrual cycle and early pregnancy Total RNA was extracted from endometrial or decidual samples. IL-15 expression levels were measured with QRT-PCR, relative to those of 18S RNA. Samples are arranged by menstrual cycle phase, or first trimester decidua. Menstrual, n=7; proliferative, n=11; secretory, n=16; decidua, n=6. AU = arbitary units. C and D: APCs and stromal cells are the main IL-15 and IL-15RA-expressing cells in first trimester decidua Total RNA was extracted from highly purified decidual NK cells, T cells, APCs, stromal cells and HLA-G⁺ EVT isolated by flow sorting. IL-15 (C) and IL-15RA (D) expression levels were measured by QRT-PCR, relative to those of 18S RNA. Medians and interquartile ranges are shown. Groups that are not significantly different from each other (p>0.05) are indicated with the same lower case letter. Trophoblast is not marked with a letter because the number of samples was insufficient for statistical testing. E: In vitro modelling of secretory endometrium reveals IL-15 concentrations are significantly increased in response to decidualisation. Human endometrial stromal cells were decidualised in vitro and expression levels of IL-15 were measured with QRT-PCR, relative to those of 18S RNA. In decidualised ESCs (DEC), IL-15 is significantly increased compared to vehicle control (Ctrl) after 2 days (n=4, p<0.001), with 8-fold increase in expression. IL-15 is also increased after 1, 4 and 8 days decidualisation (n=4, p<0.01, p<0.001 and p<0.01 respectively).

Figure 5. Asoprisnil treated endometrium displays a striking reduction of uterine NK cells

CD56⁺ cells were counted under blinded conditions in 10 randomly selected 40x objective lens fields of view of endometrial biopsy specimens from women treated with 10 mg (n=11) or 25 mg (n=11) asoprisnil or placebo (n=10) . One-way ANOVA analysis indicates that observed differences in CD56⁺ cell numbers between patient groups was significant (P<0.0001). Asoprisnil significantly reduced CD56⁺ cells relative to the placebo group (*** P<0.001: Tukey’s Multiple Comparison Test).

Figure 6. A-F: Decidual NK cells respond to IL-15 trans-presentation

A and B: Dot plots showing CD56 and CD3 staining before and after magnetic enrichment of CD56⁺ pNK cells (A) and CD56⁺ decidual leukocytes (dNK; B). Percentage of cells in each quadrant is shown. C and D: Magnetically enriched pNK cells (n=6) were cultured for five days either alone or in the presence of 0.5ng/ml recombinant IL-15, 10 ng/ml (excess) IL-15Rα-Fc, or both. After five days, the cells were counted (C), replated at 100,000 cells/well and pulsed with tritiated thymidine (D). E and F: The same procedure was carried out on magnetically enriched dNK cells from unmatched donors (n=7). Cell counts and thymidine incorporation in each set of experiments is presented relative to that when no additions were present. pNK = peripheral blood NK cells; dNK = decidual NK cells

Figure 7. Possible mechanisms of action of asoprisnil on progesterone-mediated effects on the endometrium

PR in endometrial stromal cells senses progesterone (1) and responds by increasing expression of IL-15 and IL-15Rα (2). Trans-presented IL-15 facilitates the development of mature uNK from immature NK (iNK) cells (3), and promotes their expansion. Uterine NK cells then mediate vascular remodelling (4). Asoprisnil may affect arterial remodelling indirectly by blocking PR expression in endometrial stromal cells, resulting in loss of IL-15 trans-presentation, leading to no development of uNK or arterial remodelling (5). An additional direct mechanism of action of asoprisnil is to block PR expression in perivascular smooth muscle cells (6). VSM = vascular smooth muscle cell