The model uterus on Kristin Myers’s computer screen begins to grow and stretch, reflecting the changes that take place over the course of a pregnancy. In her rendering, the tissue is initially colored blue, indicating low levels of mechanical stress. As the uterus enlarges and the simulation fast-forwards through gestational time, spiky patches of green and yellow appear, followed by streaks of orange pulsating around the cervix, indicating the intense forces that will pull it open and allow the fetus to enter the world.
Myers’s simulation is part of an effort to ask fundamental questions about pregnancy that virtually no scientific papers have addressed. In a typical non-pregnant woman, the uterus has a volume carrying capacity of about 10 milliliters (two teaspoons). By the end of pregnancy, it will have grown and stretched to accommodate five liters (21 cups) of fluid as well as the fetus itself. How does that happen? How much does the uterus stretch, and how much does it grow by adding new material? What is maternal anatomy like at each gestational stage? “No one has this data,” Myers says. “It’s simply not in the literature.”
Myers, who earned her PhD in mechanical engineering in 2008, hopes that by studying the changes in female anatomy during pregnancy and the mechanical forces at play, she can help clinicians assess pregnant women’s risk of preterm labor. She says that in analyzing the structure, strength, and stretch of the uterus and cervix—and building a corresponding computational model—her lab is “doing standard engineering work, just applying it to something radically interesting.”
Filling in the gap
The lack of scientific data on such a fundamental area of human health is especially astounding given the size of the patient population. Approximately 61 million women in the United States are of childbearing age (15 to 44), and by the time they reach their early 40s, around 86% of women have become mothers. Worldwide, the total number of births per woman is 2.4, according to the World Bank’s latest data. But pregnancy has largely been seen as a condition to be endured, not studied. Myers notes that with few women among those deciding how to allocate research dollars, women’s health issues—pregnancy among them—have typically been underfunded. “Historically, women were expected to ‘tough it out’ when it came to pregnancy,” she says. “For the most part, the sentiment was if someone is facing a pregnancy problem, then that is their problem, not a societal problem.”
Myers is one of a small number of MIT alumni and faculty bringing a new perspective to pregnancy and uterine health by approaching it as engineers. Melissa Moore, PhD ’89, is taking on preeclampsia, a form of maternal high blood pressure that can threaten mothers’ lives and necessitate early, emergency C-sections. Linda Griffith, the School of Engineering Teaching Innovation Professor of Biological and Mechanical Engineering, is studying endometriosis and other disorders that can result in significant disability, as well as infertility. Katharina Ribbeck, an associate professor of biological engineering, is exploring whether changes in the composition of cervical mucus can predict whether women are at heightened risk of preterm labor. (See “The science of slime,” MIT News, January/February 2019.) “There aren’t many of us,” Myers says. She adds, “I can’t stress it enough to young trainees: if you want to differentiate yourself and come into a field where there are a large number of unexplored, highly tractable questions, come into pregnancy.”
The mechanics of pregnancy
Myers first got interested in studying preterm birth as an MIT grad student under Simona Socrate, a principal research scientist and senior lecturer in mechanical engineering. As a visiting professor at Tufts in 2000, Socrate had begun thinking about the issue when Michael House, a high-risk obstetrician there, suggested to her that preterm birth could result from a kind of mechanical failure of the cervix. The problem is surprisingly common: roughly 10% of babies in the United States are born preterm, meaning at least three weeks before their due date, which puts them at higher risk for complications such as respiratory distress, vision and hearing loss, and cognitive impairment. So when Myers joined Socrate’s lab, she set out to explore a basic question: How strong is the cervix? She obtained cervical samples from women who had undergone hysterectomies at Tufts, and then she and her advisor ran traditional mechanical tests. “We put the samples in our load frame and pushed and pulled on them and measured how much they pushed and pulled back,” she says. “Is the cervix the strength of a rubber band? Is it the strength of a piece of wood? We were getting a quantitative sense for the first time.” (As it turns out, the pregnant cervix is indeed “like a rubber band when you pull it in tension,” Myers says. The non-pregnant cervix, on the other hand, is “stiff and tough like a piece of meat.”)
After finishing her PhD in 2008 and completing a postdoc, Myers started a lab at Columbia, where she continued to investigate the uterus and cervix. In 2015, when pregnant with her own daughter, she had an “aha moment” during the fetal anatomy scan, realizing that it would be possible to design a study in which obstetricians added brief maternal anatomy scans to the ultrasounds they were already performing. “I was a patient, and that helped me to design a study that was patient friendly,” she says. Teaming up with obstetricians at Columbia, she began to collect ultrasound measurements of the size and shape of the uterus at different points during pregnancy. She also came up with a way for them to document changes in cervical stiffness during regular exams (they use an aspiration probe—a long, thin wand that applies gentle suction—to measure the pressure needed to deform a patient’s cervix by four millimeters). As Myers is finishing a study of high-risk women, tracking these anatomical changes over the course of their pregnancies, she’s also recruiting patients for a parallel study of low-risk women whom she will follow until their babies are delivered.
Myers has already begun to notice structural differences between the two groups that could prove useful in assessing women’s risk of preterm labor. In designing her study, she assigned women to the high-risk cohort partly because they had smaller cervical lengths. But she has observed that these women also seem to have smaller cervical diameters. She hypothesizes that such a cervix may be smaller in volume, and “there is simply less material to hold it together.” In addition, the high-risk women seem to have thinner walls in the lower part of the uterus, suggesting that it’s bearing a larger-than-usual portion of the “load of pregnancy” well before birth. “We believe it is a combination of factors that truly place you at high risk for preterm birth,” Myers says. “Your cervix must be structurally inferior, and the load of pregnancy must have shifted.” But she stresses that these hypotheses must be tested in a clinical study.
Of course, to build a predictive model of pregnancy, she also needs extensive data on the mechanical properties of the uterus and cervix. This is work she began at MIT and has continued to pursue at Columbia. One morning in July, a graduate student in her lab examined a sample of pink uterine tissue, obtained from a woman who had just undergone a C-section followed immediately by a hysterectomy. The sample was one centimeter square and seven millimeters thick; the graduate student moved it delicately from a saline bath to a small platform underlying a large probe. The probe descended, and by measuring how much pressure was required to create a temporary indentation, Myers was able to quantify the tissue’s stiffness. “We specialize in doing this work with fresh tissue,” says Myers, noting that her samples often have blood vessels and irregularities like fibroid tissue that need to be accounted for. In another part of Myers’s spacious new lab, which she moved to in February, is an Instron machine, likely familiar to anyone who has spent time in a mechanical engineering lab. This machine has various appendages that stretch and compress material in order to reveal its mechanical properties. The data from such tests feed into Myers’s computational model as well, improving its predictive potential.
No single factor is responsible for preterm birth, Myers says, and ultimately she hopes to collaborate with obstetricians and other researchers to develop a risk calculator that will allow clinicians to predict, on the basis of ultrasound scans, mechanical tests, and clinical history, whether particular women are more or less likely to deliver before their due date. When a woman is about to go into labor, her cervix dilates and softens, which is why cervical stiffness is one of the variables Myers considers. The size of the cervix is also significant, since longer cervices generally require more force to open (although some women with soft or small cervices deliver babies at full term). Women’s risk also depends on the stiffness and integrity of the fetal membranes that connect to the uterine wall. “That connection actually helps hold up the baby, so you can have a structure underneath it softening and remodeling, but if there’s no mechanical load on the cervix it won’t open,” she says. (Myers is studying the properties of the fetal membranes in nonhuman primates, as well as in samples from women who required hysterectomies immediately after a C-section. She is also gathering data on fetal weight and amniotic fluid volume using ultrasound; and in the future, she hopes to assess pelvic floor strength, which may prove relevant to her model.)
Myers’s goal is to generate hypotheses about which factors are most important in assessing individual patients’ risk and to help build a tool that will let physicians visualize a wide range of mechanical scenarios. Eventually, for high-risk patients, it may be possible to use mechanical modeling to figure out what kind of intervention would be most appropriate as well. This would be valuable since the approaches that physicians have traditionally prescribed—including bed rest or a surgical stitch to hold a woman’s cervix closed—have not been proved effective in recent studies.
Another major reason that babies enter the world before they reach full term is a condition called preeclampsia, which affects approximately 3% to 7% of pregnancies in the United States. These expectant mothers—most of whom have not had a history of hypertension—develop severe high blood pressure and compromised kidney function, making it necessary to deliver the baby early by C-section. Babies born to women with preeclampsia tend to receive less nourishment during gestation. They also incur the risks associated with preterm birth itself, including vulnerability to infection and to neurological conditions like cerebral palsy. Melissa Moore, who did her PhD and postdoctoral work at MIT, has explored how preeclampsia arises—and how it might be treated using experimental strands of RNA.
Moore’s interest in preeclampsia began in 2003, when she was pregnant with her daughter and developed the condition herself. She was hospitalized and placed on bed rest when she was 28 weeks pregnant and then admitted to Beth Israel Hospital in Boston at 30 weeks, hoping to delay delivery for as long as possible. (Even an extra week of gestation can reduce a baby’s risks: babies born at 29 or 30 weeks have a 90% chance of survival, whereas those born at 31 weeks have a 95% chance of survival. Babies who stay in the womb for longer are also at lower risk of needing a ventilator or of having long-term neurological problems.) While Moore was in the hospital, a researcher named Ananth Karumanchi approached her to ask if he could have a sample of her placenta after the baby was delivered. He had just published a paper showing that a particular protein, called soluble Flt-1, is overexpressed in the placentas of women with preeclampsia, perhaps because of an issue with RNA processing. As it turns out, Moore is an expert on RNA processing, and the two discussed a potential collaboration. Nothing came of it in the short run. But a few years later, when Moore joined the faculty at the University of Massachusetts medical school and moved to a different town, Karumanchi’s daughter became her own daughter’s best friend in preschool.
Since then, Moore, Karumanchi, and eventually another collaborator, Anastasia Khvorova, have done extensive work on soluble Flt-1. Normally expressed by the placenta at the end of pregnancy, it is an important part of the process that causes blood vessels to break down and the placenta to separate from the uterine wall, which is necessary for delivery. In preeclampsia, however, the placenta produces large amounts of soluble Flt-1 at an earlier stage in pregnancy. This often means that the fetus is not receiving enough nutrients because the blood vessel connections are inadequate. It also means that excess quantities of soluble Flt-1 enter the mother’s circulation, causing an unwanted breakdown of blood vessels in areas like the kidney. That is why women with preeclampsia have compromised kidney function and protein in their urine as well as high blood pressure.
Moore and her collaborators found a way to lower the concentration of soluble Flt-1 in maternal blood, reversing some of the symptoms of preeclampsia. In animal models of the condition, they injected siRNA, or small interfering ribonucleic acid, which can reduce the expression of specific proteins, and showed that it prevented the placenta from making so much soluble Flt-1. This resulted in lower maternal blood pressure. By knocking down “the protein that was making mom sick,” says Moore, “we thought we could ameliorate her symptoms and keep the baby in a bit longer.” In December 2018, she and her team published a paper in Nature Biotechnology showing that in studies with mice, siRNA treatment could lower the amount of soluble Flt-1 in circulation by as much as 50%. They also showed that in studies with baboons, one dose of the treatment could not only reduce the amount of damaging protein in circulation but normalize the mother’s blood pressure and significantly reduce the concentration of protein in her urine, suggesting an improvement in kidney function.
Moore is now working to secure funding for clinical trials, which has been at least as difficult as the research itself. Although the Gates Foundation and NIH funded the preclinical work, larger sums are required for the next phase of work, and companies and investors tend to shy away from scientific experiments involving pregnant women for fear of incurring financial liability, she says: “It’s very frustrating.” More than 70,000 maternal deaths and 500,000 fetal deaths around the world each year are attributable to preeclampsia, and siRNA therapy, which can be administered easily by injection, holds great promise. “We’ll probably start a not-for-profit entity, a mission-based organization to make this happen,” she says.
The etiology of endometriosis
MIT professor Linda Griffith, also a passionate advocate for women’s health research, serves as the scientific director of MIT’s Center for Gynepathology Research and is a member of the advisory board of the Society for Women’s Health Research. (See “The practical activist,” MIT News, September/October 2014.) Her own lab focuses on endometriosis and adenomyosis, in which the tissue lining the uterus, known as endometrial tissue, proliferates inappropriately. In endometriosis, it spreads outside of the uterus, affecting other structures such as the fallopian tubes or ovaries. In adenomyosis, the endometrial tissue invades the muscle layer of the uterus itself. Griffith herself suffered for years from undiagnosed endometriosis, experiencing pain that was at times debilitating. She estimates that around 15% of women have endometriosis or adenomyosis, which may cause infertility in 30% to 50% of those affected—as well as severe discomfort, disability, and absences from school and work. For those who do become pregnant, endometriosis and adenomyosis increase the chances of miscarriage, preterm birth, and a condition called placenta previa, in which the placenta forms over the cervix, increasing the risk of dangerous bleeding during labor and delivery. Developing better treatments for endometriosis and adenomyosis would improve women’s quality of life and affect their ability to get pregnant and stay pregnant, Griffith says.
In a recent project, Griffith homed in on a group of proteins called proteases. Under normal circumstances, proteases break down the endometrial lining of the uterus once a month, causing women to menstruate. In endometriosis and adenomyosis, however, the proteins become overactive and break down the uterine walls inappropriately, which in turn opens the door to invasive, opportunistic growth. Proteases tend to operate in networks, and it can be challenging to figure out which proteins are actually turned on and which are merely present, Griffith says. Her lab recently developed a new method for analyzing cells and determining which proteases are active at a given moment. “We figured out how to answer this question at a systems level,” she says. “We have a recipe in which we grind up the tissue and measure activity.” Griffith says her technique will benefit not only other researchers interested in uterine health but those seeking to understand protease networks in the context of other diseases, like cancer.
Griffith is also growing endometrial organoids, which will serve as model systems for studying the causes of endometriosis as well as potential treatments. And she is encouraging undergraduates to sharpen their own engineering thinking in the context of uterine health. A group of students is looking at the relationship between the vaginal microbiome and the development of adenomyosis, for instance. Griffith is also working to promote additional formal research at MIT on how sex chromosomes affect cells and tissues, and the resulting implications for health and society. In the meantime, she’s encouraged that the Media Lab is planning “There Will Be Blood,” a hackathon focused on menstruation that will take place in fall 2020.
“If there’s one message,” she says, “it is that studying women’s biology, including factors that make it possible to sustain a pregnancy, is important to members of the MIT community and is a crucial part of our mission.”
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.