Luteal phase defect: the possibility of an immunohistochemical diagnosis

FERTILITY AND STERILITY

Vol. 51, No.2, February 1989

Copyright " 1989 The American Fertility Society

Printed in U.S.A.

Luteal phase defect: the possibility of an immunohistochemical diagnosis

Mourad W. Seif, M.R.C.O.G.* John D. Aplin, Ph.D.*t C. Hilary Buckley, M.D., F.R.C. Path.:j: University of Manchester at St. Mary's Hospital, Manchester, United Kingdom

Monoclonal antibody D9Bl has been shown to bind to a carbohydrate epitope associated with high molecular weight secretory sialoglycoprotein(s) of human endometrium. The authors demonstrate that, in a group of 28 patients diagnosed on the basis of histopathologic assessment as exhibiting luteal phase defect, 68% reveal significantly diminished expression of the epitope. Furthermore, histologic assessment of the secretory activity in single glands does not correlate simply with the level of expression of the epitope. Of specimens from patients complaining of infertility, 92% showed defective production of the D9Bl epitope. Thus, D9Bl immunohistochemistry provides a new molecular criterion for defective endometrial function. Fertil Steril51:273, 1989

The spectrum of luteal phase defect (LPD) encompasses several abnormalities. The morphology of the endometrium is dependent on, and closely reflects, the abnormal pattern of ovarian hormone secretion. There is, however, a possibility that the structural and functional abnormalities observed in the tissue also are influenced, to some extent, by the functional capacity of the endometrium. The incidence, diagnosis, and management of LPD has met with much controversy. Various reports have suggested the concept of LPD as a major clinical entity and demonstrated its contribution to infertility1 and early pregnancy wastage. 2 The condition has been associated with a variety of clinical disorders, including cases of hyperprolactinemia,3 menstrual disorders, 4 endometriosis, 5 and even in association with ovulation induction therapy. 6 Received June 20, 1988; revised and accepted October 21, 1988. *Department of Obstetrics and Gynaecology. t Reprint requests: John D. Aplin, Ph.D., Department of Obstetrics and Gynaecology, University of Manchester, St. Mary's Hospital, Whiteworth Park, Manchester M13 OJH, United Kingdom. Department of Pathology.

*

Vol. 51, No.2, February 1989

Criteria for diagnosis of LPD vary among investigators. Accordingly, estimates of its contribution to the diagnosis of infertility range from 3.57 to 20%. 1 Despite the alleged rarity of this condition among infertile women in some reports, 7 the possibility remains that LPD is more common where it is largely unsuspected and undetected. Hence, it is crucial to reach an agreement on the diagnosis before management is initiated. Assessment of luteal adequacy by measuring midluteal serum progesterone (P) is less invasive than an endometrial biopsy, and avoids the technical difficulties of inadequate sampling or misinterpretation of endometrial biopsies. However, the correlation between these two diagnostic methods has been questioned and proven unsatisfactory.s.-10 Moreover, definition of the lower limit of the normal range for midluteal plasma P has varied among investigators. 10 However, endometrial biopsy remains the cornerstone of diagnosis that has been used to determine the adequacy of endometrial secretory transformation in response to changes in hormonal milieu. Of greatest importance in the diagnosis are the criteria on which adequacy is based. It should be confirmed by at least two examinations that endome-

Seif et al. D9Bl in luteal phase defect

273

trial development lags 2 days behind the expected date according to ovulation and the onset of menses.U· 12 Alternatively, the two studies should demonstrate morphologic asynchrony between the glands and the stroma. 13 Furthermore, because of the lack of correlation between hormonal and histologic diagnosis, it is suggested10 that both midluteal phase P level and the late secretory phase endometrial biopsy should be assessed in infertile women in order to diagnose LPD. More recently, data14 has demonstrated the value of prolactin (PRL) as a metabolic marker in the assessment of the adequacy of secretory transformation of the endometrium. There is an increasing demand for the contribution of biochemistry in the assessment of the functional adequacy in the endometrium through metabolic markers and particularly, protein synthesis. 15 We have previously demonstrated the presence and cycle dependency of a secretory epitope in the endometrium by monoclonal antibody D9B1. 16 The sialylated oligosaccharide nature of this epitope has been demonstrated using a series of selective enzymatic degradationsP Binding of the antibody has shown to be a novel parameter in characterization and standardization of the normal endometrial response to ovarian hormones. 16 This study represents an immunohistochemical assessment of abnormalities of endometrial secretory transformation using monoclonal antibody D9B1 in comparison with the conventional histologic assessment using standard criteria. 11

MATERIALS AND METHODS

Twenty-eight samples of endometrial curettings, obtained from different patients undergoing investigation for infertility, menorrhagia, or irregular menstruation, and in whom luteal phase insufficiency had been previously diagnosed, were recruited from the tissue bank at St. Mary's Hospital, Department of Pathology. All had been given a diagnosis of LPD based on the histologic examination of the single endometrial specimen included in this study. The diagnosis was based on standard histologic criteria, which included asynchronous development of the endometrial glands and stroma, endometrial development that lagged 2 or more days behind the expected date according to the onset of menstruation, or heterogeneity in the differentiation of glands and their secretory activ-

274

Seif et al.

D9Bl in luteal phase defect

ity through different fields of the endometrial biopsy. Thin sections (5 ~J.m) were obtained from the specimens, which had been placed, immediately on excision, in Bouin's solution and fixed for 2 to 24 hours at room temperature. None of the specimens were held unprocessed; all material was processed within 24 hours and embedded in paraffin wax (56 to 58.C). Paraffin blocks had been stored at room temperature for a period of 2 to 3 years before this study; however, there were no observed changes in antigenic immunoreactivity with time in other studies performed using specimens that had been stored for up to 4 years. Monoclonal antibody D9B 1 was isolated as previously described. 16•18 The antibody was shown by enzyme-linked immunoadsorbent assay (ELISA) to be an immunoglobulin (lg) M. 16 Sections were stained using an indirect immunoperoxidase technique. 16 Briefly, the sections were incubated with hybridoma supernatant containing monoclonal antibody D9B1 as the first antibody, and the binding was localized with peroxidase-conjugated rabbit antimouse lg (Dakot Denmark) and the substrate N,N'-diaminobenzidine (Fluka, Glossop, UK). Each run contained appropriate controls. The positive control was a selected case of normal endometrium in the early secretory phase where the epitope has been shown to be maximally expressed. 16 The negative control involved omission of incubation with D9B1 antibody. The results were consistent in all runs as well as with results obtained from freshly collected and processed specimens. The results also were consistent with staining obtained, in other studies, by immunofluorescence using acetone-fixed cryosec.tions. Results were assessed according to the patient's menstrual history and date oflast menstrual period, and with the guideline of the semiquantitative method of scoring stained glands that has been previously reported. 16

RESULTS

Thirteen of the samples (46%) were from women complaining of infertility, whereas in the remainder (15 cases), the main complaints were of menorrhagia or irregular menstruation. Only 1 of the patients received hormone therapy before or at the time of the biopsy. She was infertile and receiving clomiphene citrate (CC) for induction of ovulation. On this regimen, she menstruated regularly every

Fertility and Sterility

28 days, and had an endometrial biopsy on the 23rd day of the cycle. In only one sample was there evidence of an intrinsic pathologic process, a~d this was endometritis. Histochemical evaluation of each endometrial sample included (1) uniformity of the expression of D9B1 epitope throughout the tissue; (2) the semiquantitative measurements previously described; 16 (3) and the location of cytoplasmic epitope as being suggestive of specific biosynthetic function in the early days of the luteal phase. Relative to the distribution of D9B1 epitope at a given phase of the normal cycle 16 (as detected in specimens exposed to the same conditions of fixation, processing, storage, and staining), cases of LPD showed one or more of the following patterns (Fig. 1A): nonuniform, patchy distribution of D9B1 epitope, significant diminution of epitope expression according to the semiquantitative assay, and a lack of cytoplasmic immunoreactivity in all or the majority of cells in each gland (Fig. 1B). Heterogeneity between glands and between cells within the same gland was characteristic. Limited quantities of immunoperoxidase complex could sometimes be detected either in the lumens of some of the glands in a sample or at the apical surface of occasional cells in some glands. This was associa ted with lack of the epitope in other locations, and probably indicates previous temporary secretory activity. The above observations were typical of 19 (68%) of the histopathologically diagnosed cases of LPD. However, within this subgroup, and especially in those in which histologic diagnosis was based on the presence of heterogeneity of secretory differentiation (Figs. 2A and B), the histologic assessment of secretory activity did not correlate with the level of expression of the D9B1 epitope within the same gland (Fig. 1B). Moreover, the D9B1 epitope could be detected in many small and straight glands in which the histologic evidence of secretory transformation was negligible. Immunohistochemical examination of the 13 specimens from patients complaining of infertility demonstrated defective production of the D9B1 epitope in 12 cases (92%), including the patient who was on CC therapy. In this case, LPD was diagnosed histologically on the basis of delayed stromal maturation and exhaustion of glandular secretory activity, despite the uniformity of glandular transformation in the midsecretory phase. Risto-

Vol. 51, No. 2, February 1989

Figure 1 (A), Endometrium from the 19th day of the cycle stained by the indirect immunoperoxidase technique using McAB D9B1, and counterstained with hematoxylin. Uneven peroxidase product (black) is present in the lumen of some glands. Heterogeneity between glands is evident. The distribution is characteristic of LPD (X43). (B), A gland from a day 21 biopsy (28-day cycle). The presence of early secretory activity contradicts the given day; Despite the uniform subnuclear vacuolation and the presence of luminal secretion, epitope is expressed by very few cells, and little of the abundent luminal secretion is immunopositive (X535). Luteal phase inadequacy is diagnosed in both cases on the basis of marked diminution of D9B1 epitope.

chemical examination revealed clearly that the rnajority of glands expressed no D9B1 epitope, whereas some showed evidence of previous expression limited to parts of the gland lumen but not as-

Seif et al.

D9Bl in luteal phase defect

275

was obtained. This subgroup included 1 case of infertility and 8 cases of menstrual disorder. The correlation between the histologic and immunohistochemical assessment was further investigated in individual glands. Glands in which all of the cells were uniformly transformed and showed subnuclear vacuolation in a uniform pattern could

Figure 2 Endometrium (day 20 of a 28-day cycle). (A), Morphologic (hematoxylin and eosin) evidence of variability in glandular maturation indicates the diagnosis of LPD. (B), Adjacent section from the same specimen stained by indirect immunoperoxidase technique using McAB D9Bl. A minimal amount of product is seen in some gland lumens. The expression of the epitope bears no relation to the glandular maturity (Xl08).

sociated with any cytoplasmic or apical locations of the epitope. In 9 of the 28 cases (32%) in the study, D9Bl immunohistochemistry did not provide any evidence ofbiosynthetic or secretory inadequacy (Fig. 3A). The epitope was detected in a uniform pattern, in sufficient quantity relative to the phase of cycle given, and evidence of cytoplasmic staining

276

Seif et al. D9Bl in luteal phase defect

Figure 3 (A), Endometrium (day 26 of a 28-day cycle). Here, irregular ripening was the basis for diagnosing LPD histologically. D9Bl epitope is uniformly expressed by all glands and abundant in the cytoplasm (see Fig lA). The case represents one of the nine in which immunohistochemistry using McAB D9Bl failed to demonstrate evidence ofLPD. (X43) (B), Endometrium (day 26 of a 28-day cycle). A group of glands is illustrated, all showing adequate secretory activity. Note the variability ofD9Bl epitope expression relative to the total quantity ofluminal secretion (X108).

Fertility and Sterility

Table 1 The Correlation Between Histologic and Immunochemical Diagnosis of LPD Immunochemical diagnosis ofLPD Criteria for histologic diagnosis ofLPD

No. cases

Asynchrony of glandular/stromal development Inadequate development Variability of glandular development

6

4

13

8

Total

+ 2 5

9

7

2

28

19

9

be almost totally (Fig. 1B) or partially negative for D9B1 epitope. Cellular expression ofD9B1 epitope did not correspond to the subnuclear vacuoles seen in the cytoplasm of the glandular epithelium in the early secretory phase, nor did its detection in the glandular lumens necessarily correspond in intensity to the quantity of intraluminal secretions as seen in the corresponding haematoxylin- and eosin-stained sections. Luminal secretions varied in their content of epitope (Fig. 3B). The correlation between histologic and immunohistochemical diagnosis was further examined by reference to the main criterion on which morphologic diagnosis of LPD was based. The results are illustrated in Table 1. Where the histologic diagnosis was mainly based on inadequate development, the correlation with the histochemical demonstration of LPD was lowest. Inadequacy was diagnosed when evidence of poor transformation had been demonstrated and the sample was abnormal for any cycle, or in some cases, when the development of the tissue lagged behind the stated phase of the biopsy cycle. Conversely, the correlation between the two techniques was maximum when the tissue showed heterogeneous glandular development. In fact, one of the two cases in which immunohistochemistry did not correlate with standard histology, was in the late secretory phase when heterogeneity appears in glandular staining in the normal cycle. 16 DISCUSSION

This study represents an attempt to characterize the LPD on the basis of expression of an endometrial secretory epitope detected by monoclonal antibody D9Bl. Previously we have reported the

Vol. 51, No.2, February 1989

presence of this component in endometrium from secretory phase. 16 Its cycle dependency has been demonstrated by a semiquantitative immunohistochemical technique. Also, further investigations 19 have demonstrated the correlation between the luteinizing hormone (LH) peak and its first appearance, cellular locations, and levels of expression. The results presented here suggest a significant diminution in the production of this epitope in the study group. In the cases included, LPD was diagnosed histologically on the basis of a single assessment endometrial biopsy and not by repetitive assessment. In fact, data suggest an increased incidence of LPD when single cycles are evaluated histologically as compared with repeated biopsy. 20 However, the purpose of this study was to compare two evaluation techniques, the conventional single histologic assessment and D9B1 immunohistochemistry, in the same cycle. Of the 28 cases included, 68% (19) showed evidence of defective expression of the D9B1 epitope, as demonstrated by both qualitative and semiquan" titative approaches. The deficiency was more prevalent (92% [12]) when the complaint was infertility. Though the case numbers are rather small, this significant reduction in expression of the D9B1 epitope in infertility cases raises the question ofthe role of this secretory component in the evaluation of endometrial receptivity to implantation. This is particularly pertinent in light of the observation that maximal amounts of secreted epitope are present at LH +6 and LH +7, 19 that is, at about the time of implantation. No evidence of defective D9B1 epitope expression was obtained in nine (32%) of the cases included. The main bulk of this subgroup consists of cases (5) in which the histologic diagnosis of LPD was based mainly on the evidence of inadequate secretory differentiation of the tissue. Various reasons could account for this result. Timing of the biopsy specimen is often difficult; in many cases, it remains unknown how early the biopsy has been taken in relation to the onset of the subsequent menstruation, and investigators have shown that dating is more accurate when biopsies are timed from the LH peak. 21 Moreover, Davidson et al. 22 have remarked on the frequency with which biopsy specimens are dated behind the actual day of the menstrual cycle, consequently questioning the clinical significance ofthe diagnosis. The increased incidence of LPD when diagnosed on the basis of

Seif et al. D9Bl in luteal phase defect

277

single sample examination has been previously mentioned. Furthermore, inadequacy constitutes an abnormality in which the morphologic assessment can be most subjective. Thus, although diagnosis of LPD may in some cases be subject to uncertainty, there remains a significant group (68%) within the cases selected for this study in which a deficiency has been detected in endometrial secretory activity. Results obtained using normal tissue specimens timed from the LH peak showed a controlled temporal progression in which the epitope appears in the basal cytoplasm at day LH +3, is then translocated to the cell apex and subsequently into the gland lumen, where it reaches a maximum at day LH+6. 19 Therefore, although immunolocalization studies cannot provide definitive evidence of biosynthesis, the circumstantial evidence is strong that this epitope originates in the epithelium. It will be interesting to compare results obtained using D9Bl with other molecular indexes of secretory function. 15 Evidence presently available suggests that the menstrual cycle dependency of the D9Bl epitope differs from that of other known stage-specific endometrial antigens (see Aplin 15 for a review). It clearly will be important to investigate further the nature of the hormonal controls. Our current biochemical evidence suggests that the major polypeptide species carrying the D9Bl epitope has an apparent molecular weight of approximately 275,000P This does not correspond to any of the presently known secretory products of endometrium. On the basis of this study, the possibility exists that, in some patients, LPD arises as an endometrial defect rather than inadequacy of folliculogenesis and ovulation. This endometrial functional insufficiency may be related to inappropriate protein synthesis, which may in turn affect the process of implantation. 23 It is interesting in this light to note that normalization of endometrial histology when alternative regimens of therapy were applied was not associated with an improvement in conception rate. 24 These ideas might be relevant to the correlation between habitual abortion and LPD. 22 The endometrial biopsy should be timed to maximize the yield of information about endometrial function and adequacy in relation to fertility. Traditionally, tissue samples have been obtained in the late secretory phase. In view of the expression of the D9Bl epitope in early secretory phase and its peak in midsecretory phase, it is possible that en-

278

Seif et al.

D9Bl in luteal phase defect

dometrial function and abnormalities may be more objectively demonstrated by immunohistochemical assay in the preimplantation period. This is also the phase of the cycle in which the morphologicalterations that occur in preparing the tissue to support an implanting blastocyst are most closely regulated and synchronized. 25 Late secretory phase tissue carries the disadvantages that the uniformity of glandular morphology 25 and biosynthetic activity 16 characteristic of the preimplantation tissue is considerably relaxed, giving a more heterogeneous tissue structure. However, pseudodecidual changes in endometrial stroma in late cycle give rise to a new pattern of protein biosynthesis and associated alterations in morphology; these changes also are potentially useful in diagnosis. More specific information on the function of individual secretory products will, in the foreseeable future, allow the biopsy to be timed in order to answer more specific questions of diagnosis tailored to the needs of the patient.

REFERENCES 1. Sillo-Seidl G, Dallenbach-Hellweg G: Uterusschleim-

haut-Polypen und Sterilitat. Fortschr Med 92:825, 1974 2. Horta JLH, Fernandez JG, De Soto LB: Direct evidence of luteal phase insufficiency in women with habitual abortion. Obstet Gynecol 49:705, 1977 3. Seppala M, Hirvonen E, Ranta T: Hyperprolactinemia and luteal insufficiency. Lancet 1:228, 1976 4. Scomegna A, Dmowski WP: Dysfunctional uterine bleeding. In Clinical Obstetrics and Gynecology, Vol16, Edited by HJ Osofsky, G Schaefer. Hagerstown, Maryland, Harper & Row Publishers, 1973, p 242 5. Ayers JWT, Birenbaum DL, Menon KMJ: Luteal phase dysfunction in endometriosis: elevated progesterone levels in peripheral and ovarian veins during the follicular phase. Fertil Steril4 7:925, 1987 6. Garcia J, Jones GS, Wentz AC: The use of clomiphene citrate. Fertil Steril 28:707, 1977 7. Israel SL: Menstrual Disorders and Sterility. New York, Roeber, 1959 8. Shepard MK, Senturia YD: Comparison of serum progesterone and endometrial biopsy for confirmation of ovulation and evaluation of luteal function. Fertil Steril 28:541, 1977 9. Rosenfeld DL, Chudow S, Bronson RA: Diagnosis of luteal phase inadequacy. Obstet Gynecol56:193, 1980 10. Shangold M, Berkeley A, Gray J: Both midluteal serum progesterone levels and late luteal endometrial histology should be assessed in all infertile women. Fertil Steril 40: 627, 1983 11. Noyes RW, Hertig A, Rock J: Dating the endometrial biopsy. Fertil Steril1:3, 1950 12. Noyes RW: Uniformity of secretory endometrium. Obstet Gynecol 7:221, 1956

Fertility and Sterility

13. Tredway D, Mishell DR, Moyer DL: Correlation of endometrial dating with luteinizing hormone peak. Am J Obstet Gynecol117:1030, 1973 14. Ying YK, Walters CA, Kuslis S, Lin JT, Daly DC, Riddick DH: Prolactin production by explants of normal, luteal phase defective, and corrected luteal phase defective late secretory endometrium. Am J Obstet Gynecol 151:801, 1985 15. Aplin JD: Cellular biochemistry of the endometrium. In Biology of the Uterus, Edited by RM Wynn, WP Jollie. New York, Plenum Press. In press 16. Seif MW, Aplin JD, Foden LJ, Tindall VR: A novel approach for monitoring the endometrial cycle and detecting ovulation. Am J Obstet Gynecol. In press 17. Aplin JD, Hoadley ME, Seif MW: Hormonally regulated secretion of keratan sulphate by human endometrial epithelium. Biochem Soc Trans. In press 18. Aplin JD, Seif MW: Basally located epithelial cell surface component identified by a novel monoclonal technique. Exp Cell Res 160:550, 1985 19. Smith R, Seif MW, Rogers AW, Li TC, Cooke ID, Aplin JD: The endometrial cycle: the expression of a secretory

Vol. 51, No.2, February 1989

20.

21.

22.

23.

24.

25.

component correlated with the luteinising hormone peak. Hum Reprod. In press Haciski R, Maroulis GB, BlakeR, Hatch R, Holt J: Inadequate luteal phase: decreased endometrial response to progesterone? (Abstr.) Fertil Steril39:392, 1983 Li T-C, Rogers AW, Lenton EA, Dockery P, Cooke 1: A comparison between two methods of chronological dating of human endometrial biopsies during the luteal phase and their correlation with histological data. Fertil Steril48:928, 1987 Davidson BJ, Thrasher TV, Seraj IM: An analysis of endometrial biopsies performed for infertility. Fertil Steril 48: 770, 1987 Daly DC, Walter CA, Soto-Albors CE, Riddick DH: Endometrial biopsy during treatment of luteal phase defects is predictive of therapeutic outcome. Fertil Steril40:305, 1983 Witten Bl, Martin SA: The endometrial biopsy as a guide to the management of luteal phase defect. Fertil Steril 44: 460,1985 Dockery P, Li TC, Rogers AW, Cooke ID, Lenton EA: The ultrastructure of glandular epithelium in the timed endometrial biopsy. Hum Reprod. 3:826, 1988

Seif et al. D9Bl in luteal phase defect

279