Download PDF Engineering for Women's Health April 25, 2022 Volume 52 Issue 1 The articles in this issue describe the latest technologies for detection of breast and other cancers, approaches to reduce the incidence of premature births, and remote monitoring for pregnancy, a development of particular interest as the pandemic discouraged many people from going to a doctor’s office or hospital. The Menstrualome: Bringing Precision Medicine to Reproductive Care Tuesday, March 29, 2022 Author: Ridhi Tariyal and Stephen K. Gire The availability of accessible, objective, and precise diagnostics is critical for delivery of adequate female reproductive care. A lack of accessible, precise diagnostic tools in women’s health significantly impacts the lives of female-born individuals and hampers the delivery of care by practitioners. Available ob-gyn tools are anchored in a clinical setting, costly, and not always clearly understood by the patient. We describe an effective, simple, inexpensive, in-home tool for individual monitoring and diagnosis of reproductive health. Challenges in Healthcare Access for Women Diagnosis of uterine diseases such as cancer, fibroids, and endometriosis requires tools available only in clinical settings, such as magnetic resonance imaging (MRI), ultrasound, and endometrial tissue biopsy. For example, the gold standard diagnosis of endometriosis depends on laparoscopic surgery and pathological assessment of excised tissue by a skilled technician. The debilitating condition, in which uterine-like tissue attaches to other organs, affects 10–15 percent of all women, causing pain and sometimes infertility (Parasar et al. 2017). The invasiveness and cost of a confirmatory laparoscopy lead to delays in diagnosis for individuals who are hesitant to undergo surgery or who are un- or underinsured. Even the mundane bedrock of gynecological preventive care, the pap smear, involves an office visit with stirrups, a speculum, and a cervical scraping. The compounding effect of in-office procedures, high claims burden, and required specialist skills is that access to reproductive care is cumbersome for many patients. Diagnostic Challenges Non-Disease-Specific Symptoms Similarities in presentation of reproductive disorders make differential diagnosis challenging. Pelvic pain, heavy menstrual bleeding, and reduced fecundity are ubiquitous features of most uterine disorders, such as endometriosis, adenomyosis, fibroids, and endometrial cancer. In-office procedures, high claims burden, and specialist skills needed mean that access to reproductive care is cumbersome for many patients. Abnormal uterine bleeding (AUB), a condition where cyclic bleeding is longer than average or unpredictable, is a hallmark of endometrial hyperplasia and raises concerns of endometrial cancer. However, polycystic ovarian syndrome, fibroids, benign polyps, and even early pregnancy can present with AUB, making a patient’s decision to seek care confusing and a doctor’s approach to diagnosis challenging. With so many possible alternatives to a shared set of symptoms, the availability of accessible, objective, and precise diagnostics becomes critical for delivery of adequate female reproductive care. Variability in Interpretation of Test Results Myriad testing modalities in gynecological practice are operator dependent and subject to variability by skillset and interpretation by a specialist (Buchweitz et al. 2005). Tests that require interpretation of imaging, such as confirmation of adenomyosis by MRI or ultrasound, are inherently subjective. As an example, a high-risk strain of HPV (human papillomavirus) infection or an abnormal pap smear requires a colposcopy to assess cervical epithelium for precancerous or cancerous cells. But the effectiveness of the procedure is highly dependent on the experience of the operator examining the patient (Xue et al. 2020). Endometrial biopsies to assess hyperplastic lesions (rapid growth of organ tissues) for endometrial cancer often provide inadequate specimens that lead to decreased confidence in negative results (Clark et al. 2002). Moreover, the classification of female reproductive disorders is still based on interpretation of gross anatomical features, rather than more objective, molecular-based diagnostics. While the morphology of lesions, plaques, scars, endometriomas, and deep infiltrating nodules in patients with endometriosis are used to classify different types and stages of the disease, the accuracy of these findings depends on a skilled pathologist. Similarly, an imaging-based assessment of the uterus requires interpretation to determine the presence of pathology in the muscle wall of the uterus (myometrium) and a diagnosis of adenomyosis (Chapron et al. 2020). Substantial reliance on the identification of anatomical features of a disease leads to subjectivity and associated variability in diagnoses. Lack of Effective Biomarkers A biomarker’s sensitivity and specificity are crucial standards in assessing the effectiveness of a test.[1] But circulating biomarkers (those in blood or saliva) have shown inadequate sensitivity and specificity for endometriosis to warrant clinical adoption (Nisenblat et al. 2016). Furthermore, attempts at developing global molecular biomarkers that are not dependent on identifying known features of a disease have exhibited low specificity. CA125, a protein biomarker initially developed for the detection of ovarian cancer, shows high variability for other cancers as well (Zhang et al. 2019) and also correlates with chronic conditions such as osteoporosis (Akinwunmi et al. 2018). Thus, although elevated levels of CA125 in serum are a useful factor to augment clinical decision making, they are not a standalone solution for accurate diagnosis of cancer (Muyldermans et al. 1995). A New Noninvasive, Self-Administered Diagnostic Tool NextGen Jane seeks to address many of these challenges in female reproductive health by developing tools that improve access to care, minimize operator subjectivity, and provide precise, molecular profiling of disease states. We do this by creating a unique catalogue of data that we call the “menstrualome,” a discrete dataset of expansive molecular profiles of cells from the reproductive tract acquired via a tampon and informed by deep phenotypic annotation. Noninvasive Biopsy A critical feature of the menstrualome, and a necessary component for many precision diagnoses, is access to tissue. A tampon-based sampling system provides access to cells from the reproductive tract that would otherwise be accessed only through an invasive acquisition method. Menstrual effluence is analogous to a rich natural biopsy of diverse cell types that can be interrogated for markers of disease. This natural biopsy is not simply a redundant source of whole blood but a unique milieu of blood, cells from the uterine lining, and local microbial organisms of the endometrium. Menstrual effluence provides a novel alternative to assay the health of the endometrium and other reproductive organs and holds promise over traditional, circulating systemic biomarkers in peripheral blood. Database for Molecular Analysis In our lab at NextGen Jane, we have undertaken the task of charting the menstrualome and cataloguing the rich matrix of cells that can be leveraged for molecular diagnosis. By applying next-generation sequencing technologies to better understand the composition of menstrual effluence, we can begin to understand the potential such a biospecimen offers. With longitudinal samples collected across multiple menstrual cycles (cervicovaginal and menstrual samples collected via a tampon, and peripheral blood specimens through traditional blood draw) from the same patients, a pattern of cell composition begins to emerge. The abundance of gene expression from each sample type enables identification of cell-specific expression profiles using vast repositories of cell-specific gene signatures to identify prominent cell types in the data. Analysis of Differences in Cell Composition By looking at each individual cell type in the endometrium, researchers can begin to understand the unique molecular signature of both abundant and rare cells that make up this highly regenerative tissue. Study of these tissues at the single-cell level throughout the menstrual cycle can yield an accurate picture of how tissue remodels in the uterus during the four phases of the cycle (menstrual, follicular, ovulatory, and luteal), influencing reproductive outcomes from fertility to longevity. One of the most important contributions to understanding of the uterine lining across the cyclic transformation of the endometrium is a recent single cell analysis (Wang et al. 2020). Tellingly, this mapping exercise did not generate transcriptomic data to characterize the endometrium during menstruation as it has historically been difficult to collect samples during menstruation. Cell-specific expression profiles can be identified using vast repositories of cell-specific gene signatures to identify prominent cell types in the data. The changing composition of cells from three sample types drives most of the differences between the samples collected during a menstrual cycle (figure 1). Cervicovaginal and menstrual samples were collected in NextGen Jane’s tampon system, and venous (peripheral) blood was collected during menstruation by traditional phlebotomy. RNA was extracted from each sample and sequenced on an Illumina shotgun sequencer. After sequence read filtering and alignment to the human genome, expression of each gene was counted based on the number of reads aligned to each transcript. Read counts were then parsed through a database of cell-specific gene signatures using the University of California, San Francisco, webtool Xcell,[2] which has 64 unique cell types from 1822 human cell type transcriptomes. FIGURE 1 Heatmap of cellular composition of cervicovaginal, menstrual, and peripheral (venous) blood, based on 277 samples collected from 80 individuals at various times of their menstrual cycle. Rows denote the presence in each tissue of specific cell types (examples are listed for cervicovaginal and menstrual cells) over a menstrual cycle: “row max” (rose) denotes overabundance, “row min” (aqua) underabundance. Large seas of aqua represent complete absence of those cell types in the sample. For example, venous blood is completely devoid of epithelial and most stromal cell types (bottom half of the chart), since the only nucleate cells in venous blood are immune cells. Cervicovaginal samples contain mostly cells from the vaginal wall and specialized immune cells of the vaginal cavity. Enriched cells listed for each sample type represent abundant cells. Each cell type was given an abundance rank based on its unique signature in the heterogeneous tissue collected, and al, and venous blood each have a unique milieu of cell types. Both peripheral blood and cervicovaginal samples include cells integral to specific immune responses (macrophages, plasma cells, and neutrophils), suggesting a highly attuned local immune environment in the vaginal cavity. While menstrual fluid and peripheral blood share some common immune cell types (e.g., T cells, macrophages, and dendritic cells), they differ dramatically in cell composition, from uterine-specific immune cells to cells lining blood vessels and specific reproductive organs. The rich composition of cells in menstrual and cervicovaginal samples represents tissues from various organs in the reproductive tract. During ovulation we see tissue-specific transcripts for ovarian and fallopian tube cells, while menstrual samples are enriched for tissue-specific transcripts of the endometrium. The unique composition of cells in menstrual effluence explains the potential utility of the menstrualome for diagnostic development. By targeting specific cell types in our samples, and by timing collections to coincide with a natural enrichment of cells of interest, we can improve signal-to-noise and begin to identify signatures of disease. Demonstrated Diagnostic Uses of Tampons The utility of this sample collection method motivates the question of why a tampon-based test was not previously developed for clinical use, although both standard clinical assays and research-driven inquiry have used tampons to aid in diagnosis of specific conditions. They have served as a binary indicator of rupture of placental membranes during pregnancy (amnio-dye test), and an early study demonstrated the superiority of tampons as a self-administered sample collection method for testing of sexually transmitted infections in remote settings (figure 2). FIGURE 2 Visual summary of two applications of tampon use at the point of care. Left: In the clinic, to detect premature rupture of the placental membrane indigo carmine dye is injected into the amniotic sac. A tampon is placed in the vaginal cavity and checked every 30 minutes for the presence of blue dye, which would indicate a placental rupture and the leak of amniotic fluid into the vaginal cavity. Reported in Medina and Hill (2006). Right: In research, tampons were used as a sample collection method for the detection of gonorrhea and chlamydia in remote regions of Australia. They showed greater sensitivity (97.2%) in detecting the presence of these diseases compared to other methods. Reported in Knox et al. (2002). In the clinic, the indigo carmine test is used to detect premature rupture of the placental membrane (Medina and Hill 2006). A blue dye is injected into the amniotic sac, and then a tampon placed in the vaginal cavity is checked every 30 minutes for the presence of the dye, which would indicate a placental rupture that is causing amniotic fluid to leak into the vaginal cavity. In research, tampons were used as a novel sample collection method for the detection of gonorrhea and chlamydia in remote regions of Australia (Knox et al. 2002). Through both polymerase chain reaction analysis and culture analysis, tampons showed greater sensitivity in picking up the presence of gonorrhea and chlamydia compared to in-clinic cervical swabs, at-home swabs, and urine tests. It was hypothesized that the tampon’s greater surface area and absorbent fibers aided in the collection and recovery of pathogens. More recently, two pivotal studies in noninvasive cancer diagnostic development showed tampons as effective methods for the collection of cervicovaginal cells (exfoliated from the cervix and uterine cavity) for the detection of ovarian and endometrial cancers (figure 3). FIGURE 3 Tampon use for detection of cancer. Left: A tampon is used to collect samples from women preparing for surgical removal of a pelvic mass. Cells from the tampon are assayed for the presence of P53 gene mutations indicative of cancer. Mutations are detected at 0.01%–0.07%. Reported in Erickson et al. (2014). Right: Cells collected from a tampon are used to assess genomic changes associated with endometrial cancer. Epigenetic analysis of methylation sites—HOXA9, RASSF1, and HTR1B—reveals the presence of cancerous cells in the tampon, with 92% sensitivity and 86% specificity. Reported in Sangtani et al. (2020). One study used tampons to collect samples from women preparing for surgical removal of a pelvic mass (Erickson et al. 2014). Cells from these tampons were assayed for the presence of P53 gene mutations indicative of cancer. These mutations were detected at 0.01–0.07 percent, meaning that ovarian/fallopian tube cells were present in the collection in about 1 in every 1500–10,000 cells collected. In contrast, cancer cells circulating in blood are much rarer, counting for 1–10 cells for every 10 million cells collected (Müller Bark et al. 2021). Another study used cells collected from tampons to assess genomic changes of endometrial cancer (Sangtani et al. 2020). Epigenetic analysis of methylation sites identified the presence of cancerous cells in the tampon. A panel of three genes with differential methylation in patients with endometrial cancer showed a 92 percent sensitivity and 86 percent specificity to identify endometrial cancer. Advantages for Reproductive Health Care NextGen Jane’s tampon-based system aims to address the gap in sample collection and motivate a deeper interrogation of a medically relevant but underexplored biological substrate. The system has been designed to overcome some of the natural impediments to menstrual sample collection: The self-administered device can be used in the privacy of an individual’s home, obviating the need for an in-clinic visit for sample collection. The preservation medium eliminates the cold chain, allows for long-term storage and transport of the sample in ambient temperatures, and facilitates the experience of users trying to get samples to a lab for analysis. In the lab, NextGen Jane has optimized processing of each unique, heterogeneous sample for various molecular pipelines, yielding terabytes of data per sample for multiple downstream analyses. NextGen Jane also catalogues deep phenotypic data (about how DNA, environment, and behavior affect an individual’s life) to contextualize the molecular data. The power of any big data analysis in medicine is only as good as the self-reported and clinical characterization of the patient’s health state. In addition to cataloguing clinical parameters through engagement with physicians, NextGen Jane collects in-depth self-reported health context to remedy the persistent problem of missing data in women’s health. The problem is exaggerated in scientific inquiry about female reproductive health as important questions often go unasked because of stigma, shame, or embarrassment surrounding topics such as menstruation, fertility, and pregnancy loss. Understanding and treatment of female reproductive health are thus hampered by missing data due to both a lack of answers and a lack of critical questions. NextGen Jane attempts to improve phenotypic capture by asking these critical questions. We do this by gathering information about a patient’s experience, their observations of their menstrual cycle, their history of disease and health care, their environment, and assessment of mental health, pain, abuse, and other factors on standardized, clinically validated metrics. A reference range for useful parameters—length of cycle, perceived pain, amount of blood loss—will enable a standard by which anomalies of menstruation can be identified early. And patients can be empowered by understanding how their menstrual cycles compare with the typical features of the cycles of their peers matched by age, ethnicity, and birth control method. Conclusion An inability to effectively capture and collect menstrual samples has created a critical lacuna in the reproductive lexicon. A tampon-based test, paired with next-generation sequencing, can add resolution to the diagnosis of disease categories currently defined by similar anatomical features. With the menstrualome, Women can engage more accessibly and less expensively with their reproductive health. Practitioners will have access to both the patient context and the technology to make sophisticated diagnoses without sophisticated, invasive, or subjective tools. Physicians will be better able to disambiguate the symptoms of reproductive disorders without specialized procedures, leading to earlier diagnoses in more generalist settings. Reproductive disorders can be identified more easily and with a more useful molecular classification that may enable better, more precise therapeutic options. The scope of impact of the menstrualome will be as expansive and deep as the scale and diversity of patients recruited and samples collected. The utility of the device is not limited to one disease or a single chapter in an individual’s life. From the onset of menstruation to postmenopause, many diseases and conditions can diminish quality of life, obstruct reproductive goals, and worsen mortality rates. For effective adoption of molecular technologies in female reproductive health, the genomic revolution that transformed oncological care over the last decade needs to be applied to precision reproductive medicine. Access to tissue local to the reproductive tract, ambitious multibiomarker class sequencing (transcriptomics, small RNA, methylation, microbiome), and integration of self-reported and clinical health data are essential to this endeavor. The ability to collect and collate this information at scale will usher in the future for women’s health. References Akinwunmi BO, Babic A, Vitonis AF, Cramer DW, Titus L, Tworoger SS, Terry KL. 2018. Chronic medical conditions and CA125 levels among women without ovarian cancer. Cancer Epidemiology, Biomarkers & Prevention 27(12):1483–90. Buchweitz O, Wülfing P, Malik E. 2005. Interobserver variability in the diagnosis of minimal and mild endometriosis. European Journal of Obstetrics, Gynecology, and Reproductive Biology 122(2):213–17. Chapron C, Vannuccini S, Santulli P, Abrão MS, Carmona F, Fraser IS, Gordts S, Guo S-W, Just P-A, Noël J-C, and 4 others. 2020. Diagnosing adenomyosis: An integrated clinical and imaging approach. Human Reproduction Update 26(3):392–411. Clark TJ, Mann CH, Shah N, Khan KS, Song F, Gupta JK. 2002. Accuracy of outpatient endometrial biopsy in the diagnosis of endometrial cancer: A systematic quantitative review. British Journal of Obstetrics and Gynecology 109(3):313–21. Erickson BK, Kinde I, Dobbin ZC, Wang Y, Martin JY, Alvarez RD, Conner MG, Huh WK, Roden RBS, Kinzler KW, and 4 others. 2014. Detection of somatic TP53 mutations in tampons of patients with high-grade serous ovarian cancer. Obstetrics and Gynecology 124(5):881–85. Knox J, Tabrizi SN, Miller P, Petoumenos K, Law M, Chen S, Garland SM. 2002. Evaluation of self-collected samples in contrast to practitioner-collected samples for detection of Chlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis by polymerase chain reaction among women living in remote areas. Sexually Transmitted Diseases 29(11):647–54. Medina TM, Hill DA. 2006. Preterm premature rupture of membranes: Diagnosis and management. American Family Physician 73(4):659–64. Müller Bark J, Kulasinghe A, Hartel G, Leo P, Warkiani ME, Jeffree RL, Chua B, Day BW, Punyadeera C. 2021. Isolation of circulating tumour cells in patients with glioblastoma using spiral microfluidic technology: A pilot study. Frontiers in Oncology 11:681130. Muyldermans M, Cornillie FJ, Koninckx PR. 1995. CA125 and endometriosis. Human Reproduction Update 1(2):173–87. Nisenblat V, Bossuyt PMM, Shaikh R, Farquhar C, Jordan V, Scheffers CS, Mol BWJ, Johnson N, Hull ML. 2016. Blood biomarkers for the non-invasive diagnosis of endometriosis. Cochrane Database of Systematic Reviews CD012179. Parasar P, Ozcan P, Terry KL. 2017. Endometriosis: Epidemiology, diagnosis, and clinical management. Current Obstetrics and Gynecology Reports 6(1):34–41. Sangtani A, Wang C, Weaver A, Hoppman NL, Kerr SE, Abyzov A, Shridhar V, Staub J, Kocher J-PA, Voss JS, and 5 others. 2020. Combining copy number, methylation markers, and mutations as a panel for endometrial cancer detection via intravaginal tampon collection. Gynecologic Oncology 156(2):387–92. Wang W, Vilella F, Alama P, Moreno I, Mignardi M, Isakova A, Pan W, Simon C, Quake SR. 2020. Single-cell transcriptomic atlas of the human endometrium during the menstrual cycle. Nature Medicine 26:1644–53. Xue P, Ng MTA, Qiao Y. 2020. The challenges of colposcopy for cervical cancer screening in LMICs and solutions by artificial intelligence. BMC Medicine 18:169. Zhang M, Zhang Y, Fu J, Zhang L. 2019. Serum CA125 levels are decreased in rectal cancer but increased in fibrosis-associated diseases and in most types of cancers. Progress in Molecular Biology and Translational Science 162:241–52. [1] Sensitivity and specificity refer mathematically to the accuracy of a test to reveal the presence or absence of a disease. [2] https://xcell.ucsf.edu/ About the Author:Ridhi Tariyal is cofounder and CEO of NextGen Jane; Stephen Gire is cofounder and chief scientific officer.