The distribution of immune cells and macrophages in the endometrium of women with recurrent reproductive failure I: Techniques.

Recurrent miscarriage affects approximately 1% of the population and in half of these cases no cause is found. Abnormally functioning immunocompetent cells, including natural killer (NK) cells, in the endometrium, are thought to be responsible for many such cases and treatment trials including oral prednisolone and intravenous immunoglobulins are now underway. Despite these encouraging developments, there is neither adequate standardization of counting uterine NK cells nor consensus as to what constitutes an abnormal level. To address this issue, immunohistochemistry was used to examine the presence and distribution of selected immune cells and macrophages in the endometrium from 222 women who had a routine endometrial biopsy for investigation of recurrent miscarriage or IVF failure, at various stages of the menstrual cycle, and accessioned prospectively over a 7-month period. Biopsies were examined by H+E and immunostained for CD8(+) T-cells, CD163(+) macrophages, CD56(+) NK cells, and CD57(+) cells. Cell numbers (expressed as immunopositive cells per mm(2)) were determined in the stroma of the functional layer of endometrium and the relative concentrations of some cell types (CD163(+) macrophages, CD56(+) NK cells) were expressed as a percentage of all stromal cells. Routine H+E sections revealed 12 patients with focal "endometritis" without plasma cells. CD8(+) T-cells showed focal perivascular aggregates in most instances, and non-random but scattered cells in all cases, with a twofold increase in the luteal phase. CD163(+) cells were distributed evenly throughout the superficial endometrial stroma and also present as single or clustered macrophages within the lumens of superficial glands, mostly in the luteal phase. CD56(+) NK cells showed "diffuse" but variable distribution throughout the functional layer and perivascular aggregates of various sizes in two thirds of cases. Raw cell counts were low and relatively stable in the proliferative phase, but increased somewhat during the first half of the secretory phase, while in the second half of secretory phase they increased six to tenfold. Percentage counts rose from approximately 5% of stromal cells in the early part of the secretory phase of the cycle to over 35% in premenstrual endometrium. CD57(+) cells were present in very low numbers in most cases. The study illustrates the complexity and variability of immune cell infiltration of endometrium. We stress the need for strict counting protocols and attention to histological criteria if any immunological perturbations potentially responsible for recurrent reproductive failure are to be identified. Reference ranges for individual cell types are only valid for individual "days" of a normalized menstrual cycle.