Routine Syndecan-1 Immunohistochemistry Aids in the Diagnosis of Chronic Endometritis Open Access

The classic morphologic features of chronic endometritis include superficial stromal edema; a stromal inflammatory infiltrate including lymphocytes, leukocytes, and plasma cells with a glandular leukocytic infiltrate; and increased stromal density.1 However, not all cases of chronic endometritis present these classic morphologic findings. Specimens that show late menstrual or early proliferative endometrium, a mononuclear cell infiltrate, abundant stromal mitoses, stromal cell proliferation or plasmacytoid stromal cells, or a pronounced predecidual reaction in late secretory endometrium may mimic the histologic findings of chronic endometritis or even interfere with the search for plasma cells.2–4 Additional factors including inexperience, fatigue, and time constraints may contribute to a missed diagnosis. This necessitates a more accurate assessment of these samples in order to avoid unnecessary surgeries in patients with abnormal uterine bleeding4,5 or repeated biopsies in patients with infertility,5 and in human immunodeficiency virus (HIV)–infected patients in whom morbidity from a hysterectomy may be avoided.6 The diagnosis of chronic endometritis may at times be academic particularly with the majority of endometrial samplings performed in clinics.

Syndecan-1 is a cell surface proteoglycan that is expressed on plasma cells7 and on keratinocytes,8 but it is not expressed by mononuclear cells, lymphocytes, or endometrial stromal cells. It is used extensively in flow cytometry to identify plasma cells7 and is also used as a marker to identify both malignant and benign plasma cells in paraffin-embedded tissue.3,5 Its function is to mediate cell migration and proliferation as well as cell-cell adhesion and cell–extracellular matrix adhesion.9–11 Syndecan-1 also participates in growth factor activities in that it acts as a receptor for heparin-binding growth factors.12 

Previously, we reported that the antibody to syndecan-1 was expressed on plasma cells and was a useful adjunct in the diagnosis of chronic endometritis.3 In order to test the feasibility of this conclusion, we prospectively performed immunohistochemistry for syndecan-1 on all endometrial curettage specimens with a non-neoplastic diagnosis and also attempted to obtain clinical follow-up. Our hypothesis was that syndecan-1 immunohistochemistry would increase the sensitivity of plasma cell detection, thus aiding in the diagnosis of chronic endometritis.

Forty-seven endometrial curettage specimens with a nonmalignant diagnosis were prospectively selected from the patient population at the University of Arkansas for Medical Sciences (UAMS). Three levels were cut and mounted on 3-aminopropyltrethoxy-silane (APS)–coated slides the same day as the original hematoxylin-eosin (H&E) sections were cut. These were archived and then batched. The slides were deparaffinized prior to undergoing antigen retrieval by heat treatment in DAKO Target Retrieval solution (DAKO, Carpinteria, Calif). The endogenous peroxidase activity was quenched with 0.3% peroxidase, and nonspecific binding was quenched with horse serum block (Oncogene Research Products, Cambridge, Mass). The DAKO Large Volume LSAB2 Alkaline Phosphatase Kit (DAKO) was used with 1:100 dilution of B-B4 (Serotec, Raleigh, NC), a mouse anti-human antibody that recognizes an epitope of human syndecan-1 (CD138).12 Sections were incubated with biotinylated secondary antibody for 30 minutes and then with streptavidin alkaline phosphatase for 30 minutes. The signal was visualized with DAKO Fast Red Substrate System (DAKO) yielding a red end-product at the site of the target antigen. Glandular structures were present in all samples and were used as the internal control. The negative control consisted of a 1:100 dilution of a nonspecific isotype-matched immunoglobulin G₁ mouse antibody, MCA928 (Serotec), which was substituted for B-B4.

Three levels of each biopsy were evaluated for the presence of plasma cells stained for immunoreactivity with syndecan-1. Plasma cells were identified by the characteristic clockface chromatin within an eccentrically placed nucleus and presence of a perinuclear halo. These cells showed strong syndecan-1 immunoreactivity on the cytoplasmic cell membrane. Also included were plasma cells that were immunoreactive with syndecan-1, which on H&E showed only the classic clockface chromatin nucleus without an eccentrically placed nucleus or perinuclear halo. Plasma cells in which only the membrane and cytoplasm were visible with H&E sections but which showed membranous syndecan-1 immunoreactivity were also included.3 

The biopsy levels were graded as “negative” when no plasma cells stained with syndecan-1, “few plasma cells” when 5 or fewer plasma cells were present, and “positive” when more than 5 plasma cells were present. When plasma cells were identified by syndecan-1 immunohistochemistry, the original H&E slides were re-examined. In all cases, plasma cells were identified on the original H&E sections. Follow-up for all patients who had plasma cells present in the endometrial curettage specimen, as evidenced by staining with syndecan-1, was obtained when possible. This consisted of a chart review to obtain the history, treatment, and outcome.

The cases of chronic endometritis showed superficial mucosal stromal edema, increased stromal density with a leukocytic infiltrate within glands and stroma, and plasma cells within the stroma on the H&E stain. Syndecan-1 immunohistochemistry showed staining of the cell surface in the glandular epithelium and cell surface immunoreactivity of the plasma cells (Figures 1 and 2).

Syndecan-1 immunohistochemistry of 47 endometrial curettage specimens (age range, 28–77 years; average age, 43 years) yielded 20 patients for which at least 1 biopsy level contained more than 5 plasma cells. In 7 of these patients, the diagnosis of chronic endometritis had been rendered on the original H&E biopsy, whereas the remaining patients were given the diagnoses of tubal metaplasia (1), secretory endometrium (4), proliferative endometrium (4), menstrual endometrium (1), endometrial polyp (1), secretory endometrium with endometrial polyp (1), and endometrial polyp with exogenous hormone effect (1). In the single case diagnosed as tubal metaplasia, only 1 of the 3 levels showed more than 5 plasma cells present. In the 4 cases diagnosed as secretory endometrium, 2 cases showed more than 5 plasma cells present on all 3 levels, and 1 case showed more than 5 plasma cells present on 2 levels with 1 level graded as few plasma cells. The remaining case revealed more than 5 plasma cells present on only 1 level; however, 1 of the levels was discarded due to poor technique. In the 4 cases of proliferative endometrium, 2 showed more than 5 plasma cells present on all 3 levels, 1 showed more than 5 plasma cells present on 2 levels with 1 of the levels being discarded due to poor technique, and a single case showed more than 5 plasma cells present on 1 level. The single case of menstrual endometrium showed more than 5 plasma cells present on all 3 levels as did the endometrial polyp with exogenous hormone effect (1). The 2 remaining cases of endometrial polyps showed more than 5 plasma cells present on 2 levels with 1 of the levels being discarded due to poor technique (Table 1).

The 7 cases of chronic endometritis diagnosed on the H&E sections showed a variety of findings with syndecan-1 immunohistochemistry. Two cases showed more than 5 plasma cells present on all 3 levels, whereas a single case showed more than 5 plasma cells present on 2 levels with 1 of the levels being discarded due to poor technique. One case showed more than 5 plasma cells present on 2 levels, 1 showed more than 5 plasma cells present on a single level, 1 showed more than 5 plasma cells present on a single level with 1 of the levels being discarded due to poor technique, and 1 showed fewer than 5 plasma cells present on a single level (Table 2).

The 13 cases that were not diagnosed as chronic endometritis on the original H&E slides yet showed presence of plasma cells by syndecan-1 immunohistochemistry were treated in various ways. The single case of tubal metaplasia showed marked improvement when treated with Megace. Three of the 4 cases of proliferative endometrium were treated by total abdominal hysterectomy with bilateral salpingo-ophorectomy (TAH/BSO) (2) and Megace (1); no clinical information was available for the final case. Two of the 4 cases of secretory endometrium were treated by TAH/BSO (1) and cyclic Povera (1); no clinical information was available for the other 2 cases. The single case of secretory endometrium with an endometrial polyp was treated by TAH/BSO, but had previously been treated with Flagyl, Megace, and Provera. Failing medical management, she went on to have a TAH/BSO; however, her pelvic pain continues postoperatively. The 2 additional cases with endometrial polyps on biopsy did not have any clinical follow-up. The single case of menstrual endometrium failed hormonal management, was treated with TAH/BSO, and was doing well on her postoperative visit (Table 1).

The 7 cases of chronic endometritis were treated primarily with antibiotics. Clinical information is available for only 4 cases; 3 showed marked improvement and 1 was eventually treated with TAH/BSO.

Syndecan-1 immunohistochemistry increased the sensitivity for detection of the plasma cells in endometrial curettage specimens. In cases where the clear-cut features of plasma cells were not evident on the H&E slides, syndecan-1 immunohistochemistry enabled the identification of plasma cells and differentiated them from plasmacytoid stromal cells.

Immunoreactivity for syndecan-1 was present in the endometrial glands, particularly at the basal aspect. Syndecan-1 is expressed in epithelial tissue,12 being found in squamous epithelium12 as well as gastric epithelium (personal observation). Normal endocervical columnar cells stain with syndecan-1, whereas subcolumnar reserve cells show variable weak staining with syndecan-1.12 In the current study, the syndecan-1 expression evident in the basal aspect of the endometrial glands may be due to the modulating effect of syndecan-1 on growth factors in that cell surface proteoglycans interact with the extracellular matrix.12 

Syndecan-1 immunoreactivity was present in all cases of chronic endometritis diagnosed on the H&E sections. Staining for syndecan-1 was evident on at least 1 level in all 7 cases; the number of plasma cells present was fewer than 5 in 2 of the 7 cases. Thus, we recommend that multiple levels on a single slide be stained for syndecan-1. Presence of any plasma cells in the characteristic milieu is then diagnostic of chronic endometritis. Presence of more than 5 plasma cells on 3 levels, in the absence of a known cause such as an endometrial polyp, would suggest the diagnosis of chronic endometritis. The cases in which the characteristic milieu is not present and in which fewer than 3 levels showed presence of more than 5 plasma cells or 1 or more levels showed fewer than 5 plasma cells are problematic. Communication with the clinician may reveal a clinical suspicion of chronic endometritis. In such instances, the diagnosis of chronic endometritis could be rendered. Otherwise, a comment in the diagnostic report may be added to indicate that chronic endometritis could not be ruled out. This is important as chronic endometritis may be masked by features of late menstrual or early proliferative endometrium, a mononuclear cell infiltrate, abundant stromal mitoses, stromal cell proliferation or plasmacytoid stromal cells, or a pronounced predecidual reaction in late secretory endometrium. In the current study, if the presence of plasma cells had been detected in at least 2 of the 4 cases given the diagnosis of proliferative endometrium, a trial of antibiotics may have prevented the morbidity of an operative procedure. In the case diagnosed as secretory endometrium who underwent a TAH/BSO, a trial of antibiotics may have resulted in improvement of her symptoms and obviated the need for an operative procedure. In a similar way, the single case diagnosed as menstrual endometrium who failed hormonal management and underwent a TAH/BSO may have had resolution of her symptoms with antibiotic therapy. The 2 cases with endometrial polyps may have experienced definitive cure by the curettage itself, thus obviating the need for further therapy.

Immunohistochemistry with syndecan-1 is recommended in all cases of infertility, in cases of dysfunctional uterine bleeding with no other cause for abnormal uterine bleeding, in HIV-positive patients with abnormal uterine bleeding and no other underlying cause, and in cases that have the characteristic milieu of chronic endometritis in which plasma cells cannot easily be found on H&E sections. The diagnosis of chronic endometritis in infertile patients may result in a successful pregnancy outcome,4 and treatment in HIV-infected patients may prevent morbidity from a hysterectomy.6 

In conclusion, the recognition of chronic endometritis is important as treatment with antibiotics may lead to resolution of the symptomatology as well as prevent morbidity associated with a hysterectomy. Syndecan-1 may aid the recognition of plasma cells in the milieu of chronic endometritis when characteristic features of plasma cells are not easily found. It may also aid to the recognition of plasma cells in cases that are masked by a mononuclear cell infiltrate, features of late menstrual or early proliferative endometrium, stromal cell proliferation with plasmacytoid stromal cells, abundant stromal mitoses, or a pronounced predecidual reaction in late secretory endometrium.