A single biomarker may not yield a reliable binary clinical answer: B cell lymphoma 6 testing and resulting interventions for endometriosis in infertility care

Abstract

Objective

To study the associations of B cell lymphoma 6 (BCL6) in endometrial endometriosis, and to determine a practice’s experience testing BCL6 in patients with recurrent pregnancy loss/recurrent implantation failure (RPL/RIF).

Design

Retrospective chart review.

Subjects

Endometrial samples were collected at the time of laparoscopy from 19 reproductive-aged individuals with known stage II–IV endometriosis. Endometrium biopsies were also collected from eight healthy controls, at a single academic center.

Exposure

Results using standard immunohistochemical staining for BCL6 on endometrial biopsies from October 2019 to March 2023 were obtained. For the in vitro fertilization patients with a positive BCL6 result (HSCORE≥1.4), postbiopsy interventions on the basis of the positive result were then compared with positive patients who did not receive treatment for their positive result of BCL6 test.

Main Outcome Measures

Outcomes for the embryo transfer cycle immediately after that biopsy/intervention were analyzed. Descriptive statistics with independent t-tests for continuous data, and χ² or Fischer’s exact tests for categorical data were used. A two-sided P value of <.05 was considered statistically significant.

Results

We found no significant upregulation of BCL6 or downstream target gene expression changes in endometriosis relative to the control group, indicating no significant single-cell differences in the BCL6 endometriosis pathway among endometrial samples from these patients with known endometriosis. A total of 33 RPL/RIF patients were treated on the basis of a positive result vs. those who were not treated. These two groups also did not have any statistically significant differences in their age, body mass index, history of live births, previous pregnancy losses, or HSCOREs. Of the patients treated for their positive BCL6 results, 20 (86%) were treated with leuprolide acetate, 9 (39%) with an aromatase inhibitor, and 1 (4%) was surgically managed with a laparoscopy in preparation for their frozen embryo transfer. There were no differences in implantation rate, miscarriage rate, or ongoing pregnancy rate among the patients who were and were not treated for their positive BCL6 result.

Conclusion

The relationship between BCL6 and endometriosis may not be the golden solution to a single marker focus for testing, or for subsequent therapy planning among infertility patients. Clinicians should be cautious in using single marker testing for diagnosing and treating endometriosis in patients without strong additional therapeutic evidence.

Focus on the role and implications of endometrial pathology for early implantation and pregnancy has increased in recent years. Endometriosis is a very common disease associated with infertility at the level of the endometrium (1, 2, 3). The association has been attributed to the disruption of the hormonal milieu, alteration of pelvic anatomy, and modulation of inflammatory and immune responses, all of which affect the tissue architecture of the endometrium and can impair implantation. The exponential promise in uncovering previously inaccessible and clinically applicable answers about the impact of genetic and microenvironmental alterations on infertility and pregnancy outcomes has led to the development of various diagnostic testing, such as endometrial receptivity assays, microbiome, and B cell lymphoma 6 (BCL6) testing. The BCL6 is one such test which is currently in use as a standalone clinical diagnostic endometrial endometriosis marker for subsets of patients with unexplained infertility, specifically with either recurrent pregnancy loss (RPL) or recurrent implantation failure (RIF) [4]. A positive test result has been proposed to corelate with a patient diagnosis of endometriosis, and treatment on the basis of a positive result has previously been shown in limited data to potentially improve pregnancy outcomes for embryo transfer cycles, without large scale validation (5, 6).

The BCL6 gene is a protein-coding transcription factor housed on chromosome three and has numerous tissue specific functions (7). It is implicated in differentiation, inflammation, apoptosis, and cell cycle control pathways. Although BCL6 is primarily known for its responsibility in the development of hematologic malignancies, there is basic scientific evidence that it plays some role in the inflammatory pathogenesis of endometriosis and progesterone resistance via a Kirsten Ras oncogene (KRAS) activation alongside the Sirtuin 1 (SIRT1) gene (4, 8, 9). The BCL6 protein has a molecular mass of 78,846 Daltons with 706 amino acids and functionally forms ternary complexes for various downstream gene expression elements (10). Its homodimer protein attribute is what is tested clinically for endometrial endometriosis. However, BCL6 cell expression in the endometrium remains vastly undercharacterized, with new data showing high BCL6 levels in all human endometrium samples by flow cytometry (11).

Research to study the genetics behind the pathophysiology within the dynamic endometrial tissue in a translational manner has been previously limited in the foundational study tools used for downstream patient testing development. However, modern multiple omics and single-cell technology have allowed for the precise characterization of molecular signatures for endometriosis (12, 13). The aim of this study is to use an established high-resolution single-cell transcriptome data to examine the specific associations of BCL6 in endometrial endometriosis, and to consider those translatable findings in a contextual clinical retrospective review of patients tested for BCL6 in the infertility setting.

Methods and materials

Ovarian lesions, wide local peritoneal excisional lesions, and endometrial samples were collected at the time of laparoscopy from 19 reproductive-aged individuals with known stage II–IV endometriosis. Endometrium biopsies were also collected from eight healthy controls, at a single academic center (13). This primary study’s sample pools include the following groups’ surgically obtained tissue samples: Ctrl (control/no endometriosis endometrium), EuE (endometrium), EcP (peritoneal lesion), EcPA (tissue adjacent to peritoneal lesion), and EcO (ovarian lesion). The specimens were processed and analyzed in an on-campus biomedical research institute, the Jackson Laboratory for Genomic Medicine. This primary study was approved by the Ethics Committee of the Institutional Review Board at the University of Connecticut Health Center, The Jackson Laboratory, and the Human Research Protection Office of the US Department of Defense and conducted according to all relevant ethical regulations regarding human participants. Single-cell ribonucleic acid sequencing (RNA-seq), cumulating to 122,000 cell samples, with hyperplexed antibody imaging mass cytometry was used to spatially localize cell types. Further ethics declarations, detailed demographics, processing, and experimentation for the original study’s patient sample pools can be referenced in the previously described methodology by Tan et al. (13). Exploratory and comparative uniform manifold approximation and projection (UMAP) analyses of this established cellular atlas, dedicated to BCL6 in endometrial samples, were conducted for this secondary analysis.

At the same academic center’s affiliated Reproductive Endocrinology and Infertility practice, a retrospective chart review was simultaneously performed to pool clinical correlative BCL6 testing results with the associated single-cell basic science findings. With institutional review board approval, the medical records for all infertility patients who underwent luteal phase endometrial BCL6 testing during in vitro fertilization (IVF) care were reviewed from October 2019 to March 2023. Presence or absence of endometriosis was determined by reviewing patient intake forms, clinical notes, operative notes, pelvic imaging, and any pathology reports in each chart.

During this period, a single expert gynecology-specific pathologist confronted all staining requests, with two other gynecologic pathologists serving as intermittent internal integrity and quality controls. Endometrial biopsies were fixed in 10% neutral buffered formalin, paraffin-embedded, and sectioned for hematoxylin and eosin-stained sections and immunostaining. Sections cut at 4–5 micrometers were placed on Fisherbrand Superfrost Plus microscope slides and stained with BCL6 antibodies, clone LN22 (Leica Biosystems), utilizing a Bond III (Leica Biosystems) automated strainer. Lymph nodes were used as a positive external control. The BCL6 staining of the endometrial epithelium was scored with histologic scores (HSCOREs), utilizing the following equation for immunohistochemical staining: HSCORE = ∑Pi(i+1)/100, where i = intensity of staining with a value from 1–3 (weak–strong) and Pi is the percentage of epithelial cells staining at each intensity ranging from 0%–100% (14). A normal HSCORE was defined by a receiver operator characteristic analysis of biopsies in patients with and without endometriosis, ≤1.4 (9).

For this study, RPL was defined as two or more prior pregnancy losses and RIF failure was defined as two or more prior failed blastocyst embryo transfers (ET). Only the first result for patients who underwent duplicate or postintervention testing was included. Both RPL and RIF patients who met the inclusion criteria with a positive BCL6 result (HSCORE≥1.4) were compared with RPL and RIF patients with a negative result (HSCORE<1.4). For the IVF patients with a positive BCL6 result, interventions after biopsy on the basis of a positive result were then compared with BCL6 positive patients who did not receive treatment for their positive BCL6 result. Possible treatment options in this setting for endometriosis were provider and patient-individualized, with shared management decisions made in the setting of a positive test result. Interventions included leuprolide acetate, letrozole, or an excisional laparoscopy. Embryo transfer outcomes for the cycle immediately after the biopsy and/or intervention were analyzed. Descriptive statistics with independent t-tests for continuous data, and χ² or Fischer’s exact tests for categorical data were used. A two-sided P value of <.05 was considered statistically significant.

Results

The demographic data and patient characteristics for the laparoscopic cases used for single-cell analysis are summarized in Table 1. Figure 1 and supplemental figures depict single-cell expression analyses across various sample types, utilizing UMAPs, violin plots, and dot plots. Across all cell types, BCL6 expression was consistently low expressed globally (Fig. 1a–c); however, stromal and myeloid cell populations exhibited relatively higher levels (Fig. 1d). Within the myeloid cell lineages, notably, infiltrated macrophages demonstrated the highest level of expression (Fig. 1e and f), although there was only a modest difference between epithelial and myeloid cells. A colorimetric bar in Figure 1e illustrates standardized BCL6 expression for infiltrated macrophages, which is relatively low. Despite these findings, no significant differences in BCL6 expression were observed across the different tissue sample groups.

Table 1.

Demographics/characteristics for the laparoscopic cases used for single-cell analysis.

Patient characteristic at the time of laparoscopy (mean ± SEM or n, %)	Endometriosis n = 19	Controls n = 8	P value
Age (y)	33.95 ± 1.1	34.13 ± 2.05	.94
Previous live births (n)	0.32 ± 0.15	1.13 ± 0.35	.02
rASRM stage 2	1 (5%)	n/a	n/a
rASRM stage 3	2 (11%)	n/a	n/a
rASRM stage 4	16 (84%)	n/a	n/a

Note: n = total sample size, n/a = not applicable, rASRM = revised American Society for Reproductive Medicine, SEM = standard error of measurement.

Figure 1.

Single-cell gene expression analysis was performed on surgically confirmed healthy (Control) and endometrium biospecimen of patients with endometriosis. Plots A and B illustrate major cell populations identified through their transcriptomic signature, and their distribution across control and endometriosis groups. The BCL6 gene expression is shown at the global cell population level (C) and per major cell population for both groups (D). The BCL6 expression in myeloid cell populations (E) and in the “infiltrated macrophage” subpopulation (F).

The violin plots in Figure 1 demonstrate that standardized BCL6 expression is consistently low across all cell populations (epithelial, endothelial, stromal, myeloid and lymphoid; Fig. 1d). Although there are small relative differences in gene expression, the absolute differences between sample population groups are minimal, particularly among stromal, endothelial, and lymphoid cell populations (Fig. 1d). Although there is a mild difference in cell population density for the epithelial and myeloid groups, the absolute difference in expression between endometrial endometriosis samples and controls remains low. Notably, even among infiltrated macrophages, BCL6 expression does not remain specific to endometrial endometriosis samples (Fig. 1f).

Downstream targets for the BCL6 gene were analyzed to further assess this gene’s impact on endometriosis. Supplemental Figure 1 (available online) depicts the mean standardized expression of prominent downstream gene targets for BCL6 in endometrial epithelial and stromal cells, including IRF4, PRDM1, CDKN1A, FAS, STAT1, CDKN1B, ATR, CHEK1, TP53, CADD45A, SUB1, FCER2, EID1, MDB1, and CBX3 (15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31). Notably, only a small fraction of cells (approximately 5%) expressed BCL6 across all major cell populations in patients with known endometriosis (Supplemental Fig. 1a). However, there was no significant gene expression upregulation or downregulation in downstream targets relative to the control group, indicating no significant single-cell differences in the BCL6 endometriosis pathway among endometrial samples from these patients with known endometriosis (Supplemental Figs. 1 and 2).

A total of 102 luteal phase endometrial samples from infertility patients were sent to pathology for analysis during the study period. Of those, 80 patients met the inclusion criteria for RPL or RIF and 33/80 had a positive HSCORE. A summary of the results for RPL/RIF patients with a positive vs. negative BCL6 testing result is outlined in Table 2. There were no differences in baseline characteristics or outcomes for those 80 patients.

Table 2.

Summary of the results for the 80 RPL/RIF patients with a positive vs. negative BCL6 testing result.

RPL/RIF patient characteristic at the time of the biopsy (mean ± SEM or n, %)	BCL6 positive n = 33	BCL6 negative n = 47	P value
Age (y)	35.2 ± 0.76	35.6 ± 0.59	.64
BMI (kg/m2)	29.2 ± 1.19	27.0 ± 0.81	.12
Known history of endometriosis (n, %)	3 (9%)	6 (13%)	.73
Previous live births (n)	0.33 ± 0.11	0.57 ± 0.10	.13
Previous pregnancy losses or failed ETs (n)	3.21 ± 0.17	3.00 ± 0.18	.41
Previously failed >1 euploid ET (n, %)	17 (52%)	32 (68%)	.13
First postbiopsy cycle ET outcomes (if available)	n = 22/33	n = 42/47
Implantation rate (n, %)	14 (64%)	29 (69%)	.66
Miscarriage rate (n, %)	3 (14%)	9 (21%)	.22
Ongoing pregnancy/delivery rate (n, %)	11 (50%)	20 (48%)	.86

Note: BCL6 = B cell lymphoma 6, BMI = body mass index, ET = embryo transfer, n = total sample size, RPL = recurrent pregnancy loss, RIF = recurrent implantation failure, SEM = standard error of measurement.

A summary of the results for the 33 RPL/RIF patients who were treated on the basis of a positive result vs. those who were not treated is outlined in Table 3. These two groups also did not have any statistically significant differences in their age, body mass index (BMI), history of live births, previous pregnancy losses, or HSCOREs. Of the patients treated for their positive BCL6 results, 86% were treated with leuprolide acetate, 39% with an aromatase inhibitor, and 4% were surgically managed with a laparoscopy in preparation for their frozen embryo transfer (FET). There were no differences in implantation rate, miscarriage rate, or ongoing pregnancy rate between patients who were treated for a positive BCL6 result and those who were not.

Table 3.

Summary of the results for the 33 RPL/RIF patients who were treated on the basis of a positive result vs. those who were not treated.

RPL/RIF patient characteristic at the time of the BCL6 positive biopsy (mean ± SEM or n, %)	Not treated n = 10	Treated n = 23	P value
Age (y)	36.93 ± 0.99	34.4 ± 0.70	.06
BMI (kg/m2)	29.69 ± 2.91	28.96 ± 1.19	.78
Known history of endometriosis (n, %)	1 (10%)	2 (9%)	.90
Previous live births (n)	0.89 ± 0.52	0.30 ± 0.15	.24
Previous pregnancy losses or failed ETs (n)	2.90 ± 0.28	2.90 ± 0.15	.96
Previously failed > 1 euploid ET (n, %)	6 (60%)	14 (61%)	.96
HSCORE	2.78 ± 0.24	3.17 ± 0.17	.20
Postbiopsy treatment
Leuprolide acetate for 1–3 months (n, %)		20 (86%)
Letrozole with FET (n, %)		9 (39%)
Laparoscopy (n, %)		1 (4%)
Postintervention ET outcomes (if available)	n = 5/10	n = 17/23
Implantation rate (n, %)	3 (60%)	11 (65%)	.85
Miscarriage rate (n, %)	1 (20%)	2 (12%)	.64
Ongoing pregnancy/delivery rate (n, %)	2 (40%)	9 (53%)	.61

Note: BCL6 = B cell lymphoma 6, BMI = body mass index, ET = embryo transfer, FET = frozen embryo transfer, HSCORE = histologic score, n = total sample size, RPL = recurrent pregnancy loss, RIF = recurrent implantation failure, SEM = standard error of measurement.

The remaining 22/102 patients without a history of RPL or RIF had undergone the endometrial biopsy as a patient request, or as a provider recommendation in cases of poor prognosis or prospective final ET cycles.

Discussion

Our objective was to outline a basic science perspective of limitations to the role of BCL6 alone in the endometrium as a clinical fertility screening test using single-cell analyses of endometriosis tissues, and to comment on applicable correlated clinical outcome data. There were no differences in patients' age, BMI, known history of endometriosis, number of previous live births, number of previous pregnancy losses, number of previous failed ETs, or HSCORE for BCL6 among the positive patients who underwent a postresult intervention vs. the BCL6 positive patients who did not. Although the data are limited by the small sample size, there was also no difference in any of the groups’ patient characteristics or pregnancy outcomes.

In fact, our high-resolution single-cell gene expression data show that BCL6 expression is not specifically enriched in the endometrium for patients with known endometriosis. Although our clinical data are limited by the sample size, the lack of sensitivity is supported by the resulting high false negative rate. Approximately 12.8% (6/47) of patients with a negative BCL6 result had a known history of a laparoscopy with histopathological confirmed endometriosis. Their mean ± standard error of the mean HSCORE was 0.31 ± 0.14. Our data also show a potential high false positive rate for the biopsy result: 24.2% (8/33) of patients with a positive BCL6 result had no evidence of endometriosis on a recent laparoscopy by a highly skilled minimally invasive gynecologic surgeon, with their mean HSCORE being 3.0 ± 0.29. More patients (12.8%) with known endometriosis were BCL6 negative than positive (9.1%).

Interestingly, BCL6 gene expression is detected in stromal and epithelial cells, and is generally expressed at higher levels among myeloid cell populations, particularly monocytes and infiltrated macrophages. The BCL6 is known to be involved in biological inflammatory pathways, and so these cell populations offer confounders when using BCL6 based testing for a clinical endometriosis diagnosis via an endometrial biopsy. These cell lineages can be enhanced in the endometrial tissue of patients with other pathologies for RPL/RIF, such as chronic endometritis, a recent pregnancy loss itself, adenomyosis, uterine fibroids, or a recent surgery (32, 33, 34, 35, 36, 37, 38). In fact, our data show that BCL6 is enhanced even in known normal control patients’ endometrial infiltrated macrophages. We also importantly identified downstream target genes from BCL6 which confirm by their standardized expressions that there is not a difference at the high-resolution single-cell level between endometrial biopsies from patients with endometriosis and those from control patients.

Conclusion

Although some studies have claimed patient benefit to single marker BCL6 testing, more research in matched patients in the same cycle phase is certainly needed to fully understand the complex relationship between all specific genes and proteins, as well as their involved pathways, in endometriosis. The clinical portion of this study used immunohistochemical staining for protein levels, and the single-cell portion examined BCL6’s gene expression. However, it is likely that the answer in eventual nonsurgical diagnostics to identify biomarkers for targeted therapies will necessitate a comprehensive multimodal testing panel. The relationship between BCL6 and endometriosis may not be the golden solution to a single marker focus for testing, or for subsequent therapy planning among infertility patients. Interventions on the basis of this single testing result such as surgery, gonadotropin-releasing hormone agonists or aromatase inhibiting medication additions have not yet been shown to benefit patient outcomes and may delay care. Clinicians should minimally recognize fundamental basic science limitations and biases in single marker testing on whole endometrial tissue samples for endometriosis to be able to further appropriate clinically applicable patient results for infertility treatment planning.

Declaration of Interests

A.T. has nothing to disclose. A.U. has nothing to disclose. Y.T. has nothing to disclose. D.E.L. has nothing to disclose. M.S. has nothing to disclose. E.T.C. has nothing to disclose. C.B. is a member of the Advisory Board in Ferring Pharmaceuticals.