Meet the fetal surgeon forging CRISPR’s next frontier: curing diseases in the womb

SAN FRANCISCO — Outside, the August sun wasn’t yet visible through the thick folds of fog blanketing the San Francisco skyline. Its warmth did not reach the operating room tucked into the sprawling Parnassus Heights hospital complex. In there, the light was all cold and blue fluorescence washing over the sea of scrub caps huddled around an anesthetized young woman on a gurney. From one corner of the crowded room, a medical student named Tippi MacKenzie watched, eyes widening, as the woman’s uterus was gently lifted out of her open abdomen and an incision was made to expose the legs and backside of the fetus inside.

At 23 weeks, it was barely the size of a mango, made even smaller by the massive, glistening purple tumor protruding from its tailbone. As the doctors began the painstaking process of removing the mass — an orchestrated flurry of fingers tying knots, scalpels slicing flesh, cautery pens searing blood vessels shut — the room grew thick with anticipation. Sacrococcygeal teratomas weren’t usually cancerous, but these kinds of tumors steal the fetus’ blood supply and strain its heart, in most cases, fatally.

Here at the University of California, San Francisco, a team of surgeons led by a doctor named Michael Harrison, had spent the past few years developing techniques to try to fix those dismal outcomes, by operating on the tumors prior to birth. His team had tried the procedure three times before. None of the fetuses had survived. This one seemed to be going better, or so MacKenzie thought. The tumor, all 16 ounces of it, had been cut free. Then she noticed Harrison had the fetus cupped in his gloved hand, squeezing it at regular intervals. It dawned on her suddenly; he was performing CPR.

The team around him injected the fetus with epinephrine, then atropine. They placed warming pads on its exposed limbs. Five minutes after it had stopped, a heartbeat returned. As the room collectively exhaled, a nurse reached back through the crowd with a vial of blood drawn from the fetus and handed it to MacKenzie. She sprinted to the elevator, rode it to the lab on the 16th floor, hurried back down, and read out the numbers on the slip of paper in her hand. Still panting as she watched Harrison seal the fetus back inside the uterus, MacKenzie knew this is what she wanted to do with her life. “Here you have some horrible deadly malformation and with surgery you go in and you just fix it,” she said, snapping her fingers. “It’s just so definitive and so vital.”

Five weeks later, the woman delivered a baby girl who would go on to become a healthy child with a normally functioning heart.

That August day in 1996 was just one of many milestones for Harrison, who retired 10 years later after having earned global recognition as “the father of fetal surgery.” For MacKenzie, it was a singular experience that set her on a path that would propel her, nearly three decades on, to the forefront of the next revolution in medicine.

If UCSF is known for birthing the field of fetal surgery, UC Berkeley, located a short drive across the Bay Bridge, is famous in biomedical circles for pioneering CRISPR gene editing, the most powerful DNA-manipulating tool ever invented. What MacKenzie envisions — the future she is now preparing for — is the convergence of these technologies into a whole new field of medicine dedicated to curing diseases before birth: fetal genome surgery. This is surgery without scalpels or sutures, just a syringe pushing particles containing CRISPR into the vein that connects a pregnant person to the fetus. Once inside the fetus’ cells, CRISPR’s molecular scissors snip away the string of problematic DNA, stopping a catastrophic genetic disease before it really starts. If successful, it would fundamentally and forever change the practice of fetal and maternal care.

That’s the idea, anyway. There are still big questions that have to be answered — about what risks fetal genome surgery runs for the mother’s health and the likelihood of unintended mutations in the fetus, including those that could be passed down. Mackenzie, 53, is now building a research team focused on solving these issues in the lab. Over the next decade, she hopes to lay out a road map for how in utero applications of CRISPR— as well as other bespoke molecular medicines — might be safely and ethically tried for the first time in humans.

“She’s basically making the field,” Beltran Borges, a postdoc in her lab, told STAT. “There are not many people right now doing what she does.”

While it may sound like science fiction, many researchers believe that with the success of potentially curative CRISPR therapies like Casgevy, which was approved last year for sickle cell disease, the next logical step is to move genome editing earlier in the human life span — into the womb, where its curative potential is the greatest. This tantalizing prospect recently led the National Institutes of Health to begin funding projects that are carving a path to clinical trials. Getting permission from the Food and Drug Administration to test a fetal therapy is an enormous challenge, even more so when it involves an emergent technology like CRISPR. Yet, if anyone can do it, scientists told STAT, it’s MacKenzie.

“Tippi MacKenzie’s work is not just a sketch on a napkin, it’s a beacon,” said Fyodor Urnov, a genome editing expert and the scientific director of the Innovative Genomics Institute at UC Berkeley. The idea of therapeutically altering the DNA of a fetus in utero had been floating, opaquely, in the ether since the 1990s, shrouded by issues both technical and ethical in nature. It’s like the fog that rolls through San Francisco; no one in the gene-editing community had been able to see a way through, he said. “Literally, there’s now a little light across the bay at UCSF saying, ‘come this way.’”

Long before she could imagine editing the billions of letters of human DNA, MacKenzie set out to master the 88 keys of the piano. Born Tippi Ciçek, she moved with her family to the United States from Turkey in 1981, when she was 11 years old, after her father was hired as a professor of computer science at the University of Delaware. She spoke the language some, thanks to her mother, who was an English teacher, but the transition wasn’t easy.

A shy and serious child, she spent her hours outside of school coaxing Bach fugues from the Steinway grand that her parents took out a 15-year payment plan to purchase for her. She missed the chaotic, cramped commute by bus and dolmuş to the Istanbul Conservatory, the smell of her teacher’s perfume as they worked the same two bars over and over for hours. “Spending intense time working on the craft of perfecting a page of music, most kids hate that,” MacKenzie said. “I was riveted by it.”

Although she also excelled in math and science, everyone in her family assumed she was bound for the life of a professional pianist. By the time she was in high school, her father was driving her two hours to New York City every Saturday to take lessons at Juilliard. But she learned there that her love was with practice, not performance. And when she crunched the numbers on how many music students actually make it as musicians, she decided on something more practical: a degree in biochemistry (from Harvard), then medical school at Stanford.

During a surgical rotation, her ears perked up when she heard another medical student talk about what Harrison was up to at UCSF and arranged to shadow him for a month. After his car broke down, MacKenzie drove him to hospitals all across the Bay Area, scrubbing into cases and handing him instruments. Studying his movements day after day as he repaired a hole in a developing diaphragm or rerouted a faulty fetal urinary tract, she felt that familiar thrill of refinement through repetition.

In fetal surgery, the slightest squeeze or tilt of a scalpel can be the difference between life and death. It wasn’t a piano concerto, but learning to apply just the right amount of pressure with this finger or rotate one’s wrist an imperceptible couple of degrees — that was a type of craft she knew how to perfect.

Her talent, even then, was obvious to Harrison. “I thought she would be a star,” he told STAT. “But no one else did.”

Back then, he explained, surgeons were all cut from a particular cloth. Decisive. Athletic. Macho. All ego and sinew, and nearly all of them, men. MacKenzie, in contrast, came across as eager but awkward, which Harrison took as evidence of her keen intelligence. He thought she would make a brilliant researcher. So he was surprised when she moved to Brigham and Women’s Hospital in Boston for a surgery residency. “In those days, it was not common for even a bright young woman to go into surgery,” he said. “That was a big deal.”

In Boston, she continued to surprise the surgeons around her. Like when she nabbed a fellowship spot in the lab of Alan Flake at Children’s Hospital of Philadelphia. He was doing headline-grabbing stuff. A few years before she arrived, his team had published a breakthrough report; they had cured a child with X-linked severe combined immunodeficiency, or “bubble boy” disease, before birth by injecting the mother with infusions of blood stem cells from the father. The cells settled into the developing fetus’ bone marrow, eventually giving rise to a normal immune system.

After this early success, though, Flake and his colleagues struggled to use stem cell transplants to treat other types of diseases. They turned to the newly invented tools of gene therapy — engineered viruses that could slip a stretch of DNA into human cells. Kathy High, a pioneering hematologist whose efforts would lead to the first approved gene therapy, worked out of a lab just a few floors below Flake’s. MacKenzie began research injecting fetal mice with copies of the healthy hemophilia gene, packaged inside viruses produced by High and another collaborator, Jim Wilson at the University of Pennsylvania.

But during her first year there, tragedy struck. A young man named Jesse Gelsinger was injected with a large dose of some of those gene-shuttling viruses designed in Wilson’s lab in an attempt to treat his rare metabolic disease. His body’s immune system responded to the viral invasion with lethal force. Four days after being treated, Gelsinger died. Lawsuits and investigations followed. The gene therapy center Wilson ran at Penn was disbanded, and he received a five-year ban from conducting clinical trials. High and Flake and other researchers carried on, but money and industry interest dried up.

While all that was going on, MacKenzie was busy starting a family in addition to her career as a fetal surgeon. She’d set a goal of using her time in Philadelphia to get married, have a child, and finally fix her teeth, which had been crooked since childhood. “I didn’t care if it happened in any conventional order,” she said. “I just needed to get it done.” In the spring of 2000, her braces came off, the same week she and John MacKenzie, whom she’d met in medical school, were wed. Two months later, she found herself pregnant with her oldest daughter, Emma.

They spent the next few years juggling parenthood with more training — fetal surgery for her, pediatric radiology for him. In 2007, when they all moved to the Bay Area (Steinway included) for MacKenzie to start a lab and perform surgeries in the hospital’s fetal treatment center, gene therapy wasn’t on her research agenda. Instead, she returned to the mystery of the disappointing in utero stem cell transplants. Like when a pesky passage of notes still don’t sound quite right, she felt the itch to figure out why the transplants didn’t work more broadly.

In collaboration with Flake’s lab, she hypothesized that the problem was that the stem cells always came from the male parent, because nobody had wanted to harvest bone marrow from a pregnant woman. Studies they did in mice showed that cells from the mother trafficked into the fetus and rejected the transplants that came from fathers. That suggested an obvious solution: use maternal cells instead.

In the fall of 2017, after a decade of research to tease out those details, they launched the world’s first clinical trial of a fetal stem cell therapy.

They set out to see if infusing a large dose of a mother’s stem cells, along with several blood transfusions, could rescue fetuses diagnosed with a fatal form of alpha thalassemia. The inherited blood disorder is caused by a mutation in one of the hemoglobin genes, which saps developing tissues of oxygen and leads to severe swelling in the heart, a condition known as hydrops that medical textbooks describe as a “harbinger of fetal demise.”

Families carrying fetuses with this form of alpha thalassemia are usually told they have two agonizing options: end the pregnancy, or continue, at risk of severe complications for the mother, despite nearly assured fetal death.

That’s what Nichelle Obar, 46, and her husband Chris Constantino, 43, heard from a genetic counselor in their home state of Hawaii, when an ultrasound during the second trimester of their pregnancy revealed hydrops and a genetic test confirmed it was caused by alpha thalassemia. The counselor also told them about the UCSF study, which she’d learned about in a webinar only the day before. It was around Obar’s birthday, and the couple had planned to take a quick vacation to Maui. Instead they flew to San Francisco to meet with MacKenzie’s team.

“I remember looking at all the information they gave us and reading a lot about large animals,” recalled Obar. “It was like, ‘OK, but are there human participants?’” MacKenzie told her no. They’d be the first. That raised the stakes for Obar, a county administrator, and Constantino, a grocery clerk, who’d lived their whole lives in a small town on the north shore of Kauai. While they were deliberating, an ultrasound showed that fluid had begun to move from the fetus’ heart to its belly, a bad sign.

One of MacKenzie’s research partners, Elliott Vichinsky, laid out the options and the risks of each. An expert in alpha thalassemia who has treated hundreds of patients with the disease, he explained that the fetus’ heart had formed normally. It was just working overtime. They decided to go through with the experimental procedure, despite the risks and knowing that nothing was guaranteed. “We did it because there was that hope that our baby could survive,” Obar said.

As the pregnancy progressed, Vichinsky tried to prepare them for the birth — alpha thal babies often come out silently, tinged purple or blue because of a lack of oxygen. But in the early hours of Feb. 1, 2018, their daughter entered the world with a powerful cry and a full head of jet-black hair. They named her Elianna, after the UCSF nurse who had administered Obar’s first in utero transfusion.

MacKenzie and her team were thrilled. In the neonatal ICU, Elianna was recovering well from a case of jaundice and her heart looked strong and healthy. But when the researchers analyzed her cells and DNA, they realized that the stem cell treatment hadn’t worked as hoped. Lacking oxygen, Elianna’s own bone marrow had gone into overdrive, expanding to try to rev up production of red blood cells. There wasn’t any room for the maternal stem cells to go. As they enrolled other patients and had more successful births, they continued to see the same, disappointing trend. After six patients, MacKenzie made the call to stop the trial early.

“It was the worst kind of host-microenvironment to be transplanting the cells in,” she said. What had worked — at least to buy some time — were the blood infusions they’d done in the lead-up to the procedure. Those red blood cells, while not a forever solution, given they only live a few months, had provided the fetuses with enough oxygen to survive to birth. Elianna and the other children now had a chance, but they would need regular transfusions for the rest of their lives, or until they could get a bone marrow transplant.

As MacKenzie was coming to grips with the fact the treatments had failed, she was approached about taking a different kind of therapy into the womb for the first time. Ultragenyx, a Bay Area biotech, had created a synthetic enzyme that the FDA had just approved to treat a genetic lysosomal storage disorder called MPS7. Patients with MPS7 carry a mutation that prevents them from making this enzyme, without which toxic sugars accumulate and damage the brain, heart, and other tissues.

With only 150 or so patients worldwide, the disease is considered ultra-rare, but MacKenzie suspected its true incidence was higher. At UCSF, she and her colleagues had begun sequencing the DNA of fetuses who developed hydrops and discovered that the most common genetic defect driving their fatal heart swelling was the MPS7 mutation.

Some patients that survived to birth and beyond saw improvements to their heart and lung function after they received Ultragenyx’s enzyme. But for others, by the time they got the treatment, too much damage had already been done. Ultragenyx wondered if treating earlier, during fetal development — when neurons had not yet been sealed off behind the blood brain barrier — might change that. And MacKenzie’s lab already had a mouse model they could use to test the hypothesis.

It worked, even better than they had expected.

MacKenzie’s team showed that enzymes fed through a mother mouse’s uterine vein could migrate to the brain cells of her pups, clearing out the toxic sugars and preventing dangerous inflammation. The results convinced the FDA to let her move forward with an international clinical trial, and in an unusual vote of confidence for the approach, the agency allowed her to try in utero enzyme replacement not just for MPS7, but for the handful of lysosomal storage diseases for which approved enzyme therapies existed.

That trial is still ongoing, slowed down in its early stages by the Covid-19 pandemic. But in 2021, Canadian doctors at a hospital in Ottawa tried using MacKenzie’s protocol on a 37-year-old woman pregnant with her third child. The first two had died of Pompe disease before their third birthdays. After six prenatal infusions of an enzyme treatment, Ayla was born with no signs that the mutation she’d inherited had started to weaken her heart or any other muscles. As in the alpha thalassemia trial, however, Ayla’s health was still fragile — without a dose of the enzyme every week, her disease would progress.

“This was meaningful, but it wasn’t a cure,” MacKenzie said. Enzymes and blood infusions were just Band-Aids, temporary patches that required regular updates. A cure would require debugging the genetic source code spreading through a fetus’ rapidly dividing cells. She now believed such a thing was possible, but making it a reality for her patients and all the hopeful parents who regularly sent her desperate emails would mean giving something up.

By that time, she had spent 25 years, nearly half her life, gaining membership to one of medicine’s most elite clubs. There are approximately 300,000 physicians living and working in the U.S. Fewer than 100 of them are pediatric surgeons who operate on fetal patients. The niche subspecialty demands delicate hands, unflappable focus, and a nuanced understanding of developmental biology. More than anything, though, it requires guts. Across hundreds of surgeries and two clinical trials, MacKenzie had shown she had all of those things in spades. Now she showed she also had the courage to walk away.

A month after Ayla was born, MacKenzie hung up her scrub cap. Her name came down from the operating room on call schedule. In their place, she stepped into a new role as the director of the Broad Center of Regeneration Medicine and Stem Cell Research, a position that would enable her to put the science of developing curative fetal therapies front and center.

“This field has so much potential to change medicine,” MacKenzie said. “Curing patients through individual surgeries is an incredible privilege. But I realized that if I spent my time doing that, I wouldn’t be able to focus on something that has the potential to have a much bigger impact.”

A year before MacKenzie joined the UCSF surgery staff, the phone rang in Jennifer Doudna’s seventh-floor office at UC Berkeley. On the line was Jillian Banfield, a microbiologist on campus whose lab had developed a method to sequence the DNA of all the microorganisms living in a pinch of soil or an acidic drop of mine drainage. Using that method, her team kept finding the same repeating genetic sequences called CRISPR crop up across disparate branches of the bacterial family tree. Banfield wanted to know what they did. Doudna agreed to use her skills in genetics and biochemistry to find out.

In 2012, Doudna and her collaborator on the project, Emmanuelle Charpentier, then at Umeå University in Sweden, revealed that CRISPR was a two-part system that bacteria used to recognize and cut DNA. And through a bit of engineering to one of those parts, they could direct it to cut any piece of DNA they wanted. What bacteria had evolved as an ancient defense mechanism against viruses, Doudna and Charpentier had repurposed into a tool for operating on life’s most minute anatomy: DNA.

While not the first of its kind, CRISPR was by far the most exacting and easiest genome-editing tool to wield inside the cells of any living organism. Unlike gene therapy vectors, which insert their therapeutic stretch of DNA at random, CRISPR systems can be designed to make their changes predictably, in only the gene of interest. That had the potential to alleviate one of the longest-standing concerns about in utero gene therapy: the possibility of cancer.

Back in January 1999, in response to the very first requests to begin clinical trials of fetal gene therapy, the National Institutes of Health Recombinant DNA Advisory Committee held a meeting to discuss the scientific, legal, ethical, and societal implications of the procedure. Cancer was raised as a potential risk because a fetus’ rapidly dividing tissues would be extra susceptible to instances where insertion of a piece of DNA disrupts a part of the genome that regulates cell growth.

The committee’s members were also concerned about viruses used in gene therapy embedding themselves into the stem cells that give rise to sperm and eggs. Any changes to the DNA of these so-called germline cells could be passed down to future generations — and because of the social and ethical implications of making inheritable genetic modifications in humans, they are prohibited in many countries, including the U.S. Then there was the omnipresent danger when you’re doing anything with viruses; that even in their weakened, hollowed-out state, something on their surface would trip the fetal or the maternal immune system, triggering a dangerous inflammatory response.

Until these risks could be minimized, the committee concluded, it wasn’t safe to pursue fetal gene therapy trials. That decision looked prescient nine months later, when despite testing in mice, monkeys, baboons, and one human patient, gene therapy claimed its first victim in Gelsinger, when the virus triggered a massively fatal immune reaction. As the field receded into a decade-long dark age, the physician-scientist who’d first proposed fetal gene therapy, W. French Anderson, was convicted of sexually abusing a minor and sentenced to 14 years in prison.

Amid the failures and scandals, the few U.S. research labs still working on understanding how to make fetal therapies safe kept their heads down and did what they could with what little money they could scrape together.

Around the time that Doudna’s team was unraveling the machinery of CRISPR, a bioengineer at Wake Forest University named Graça Almeida-Porada came to MacKenzie to vent. Her proposal to study fetal cell and gene therapies for hemophilia in sheep — a step toward clinical trials — had been rejected by the NIH. The reason, she was told, was one grant reviewer’s comment that “in utero therapy will never be a reality.” The perception was that it would never work. In order to convince funders otherwise, the two women decided, their community had to get organized.

In 2014, they hosted the first-ever meeting of iFetis, the International Fetal Transplant and Immunology Society, in San Francisco. It was a shoestring affair, with only enough funding to pay for hotels and coffee and maybe a dinner, MacKenzie said. “People had to fly on their own dime.” But fly they did; about 30 people that first year, and more than 50 the next. “There was clearly both a need and an interest,” she said.

At the group’s annual meeting in 2018, MacKenzie assembled a panel of international experts to revisit the 1999 NIH committee’s recommendations, to assess if enough technological progress had been made to start to think about fetal gene therapy clinical trials again. In the intervening 20 years, the field of gene therapy had begun to make a comeback. The FDA had just approved the first three gene therapy products for use in the U.S.— Kymriah, Yescarta, and Luxturna — and the agency was considering more than 700 requests for new gene therapy trials.

After spirited discussion, the panel reached a verdict. The future looked bright indeed. There were still hurdles — still questions about how to keep DNA changes out of the germline or not ring any maternal immune alarm bells  — but now there were new tools to solve those, tools like CRISPR and mass-producible mRNA and lipid nanoparticles. Combined, they allowed scientists to slip invisible instructions through the blood, past immune cells, and into tissues where they would become precise, DNA-slicing micromachines.

A few months later, one of the panelists, Bill Peranteau, a fetal surgeon at the Children’s Hospital of Philadelphia, published an electrifying report with Kiran Musunuru, a gene editor at the University of Pennsylvania. Together, they had used a form of CRISPR known as base editing to alter the DNA of laboratory mice in the womb, and eliminated an often-fatal liver disease before the animals had even been born — with no ill effects for the mouse mothers. Strikingly, these mice did much better than pups who were treated shortly after birth.

“We were hoping to just get enough editing for them not to die,” Musunuru told STAT. “To see the genome-edited mice not just do better if treated from the get-go but grow and thrive and look like normal mice was pretty illuminating. It was like, ‘Wow, hey, we could really do this.’”

That year, the NIH launched a $190 million push to improve the efficacy of genome editing and enable delivery of genome editors into a wide range of tissues and organs. Five years later, when the NIH added a phase two of the initiative, focused on moving these technologies into the clinic, they selected a proposal from Peranteau and Musunuru that included in utero gene editing. The $26.5 million project involves developing CRISPR systems aimed at treating three diseases, all inborn errors of metabolism that are currently picked up in newborn screening: PKU, tyrosinemia, and Hurler syndrome. In parallel experiments in mice and monkeys, the researchers plan to test versions of therapy for both postnatal and in utero treatment.

It’s not that they couldn’t have gotten funding just for the in utero work. Chris Boshoff, the program manager for the NIH’s Somatic Cell Genome Editing Consortium told STAT that such projects were explicitly welcomed. It’s just more practical.

“We think there could be a substantial benefit for treating tyrosinemia and Hurler syndrome prior to birth, but there’s no way any regulatory agency is going to allow something prenatally that has not already been validated postnatally,” Musunuru said. He suspects the FDA will want to follow patients for at least a few years to make sure nothing unexpected emerges before greenlighting the first trial of CRISPR in the womb.

That’s how the NIH envisions it would work too. “We believe that if we support the IND-enabling study for a particular disease, the next step would be to go in utero,” Boshoff said. “We hope to see it go in that direction.” He stressed that intervention in utero isn’t appropriate for every disorder, but the NIH is particularly invested in fetal treatments that can meaningfully alter the trajectory of disease, where, as in neurological conditions, problems start during the early phases of brain development.

Take, for example, spinal muscular atrophy, an inherited neuromuscular disease caused by a defect in a gene that makes SMN, a protein that motor neurons need to survive. In 2019, the FDA approved a gene therapy for the disease, Zolgensma, that has been hailed as a miracle drug. In clinical trials, children who received a single injection of Zolgensma before the age of 2 survived, breathed on their own, and hit major motor milestones. Those diagnosed prenatally or at birth, who received the treatment in their first month of life, showed the biggest improvements.

For many kids, gene therapy hasn’t been a cure, and in some cases, it hasn’t even been enough to live a significantly more normal life. Part of the reason, neurologists told Reuters, is because children are getting Zolgensma too late. Their neurons had already begun to die off before they were born.

It’s these kinds of stories that have swayed people like Matthew Porteus, who until recently considered himself a skeptic of fetal genome editing, that there’s a real need for it. “What convinced me is the fact that there are so many diseases where already at birth irreversible organ damage has occurred, so the only opportunity is to treat in utero,” he said.

For a long time, Porteus, a pediatric hematologist at Stanford School of Medicine who has worked in the field of gene editing since the ’90s and is a scientific founder of CRISPR Therapeutics, couldn’t get past the idea of exposing healthy people to potentially toxic CRISPR components to treat the fetus developing inside them. “Moms will rightly and wrongly jump into flaming cars to save their children,” he said. “How do you protect the mom from taking undue risks on behalf of her fetus?”

One part of that is building gene editors that are safe to infuse into people directly. Encouraging data released over the last few years from CRISPR company Intellia have shown that’s possible to do. But the other is having thoughtful, experienced people constructing clinical trials that are transparent about all the risks, both known and unknown. And for that, Porteus couldn’t think of a better person than MacKenzie.

“She has the knowledge and the technical skills, but more than that, she’s open and listening and integrating feedback from the community in a great way,” he said. She also has just the right amount of impatience. “Part of her wants this all to happen tomorrow and part of her knows this is a years-, if not decades-long process.”

On a gloomy, rain-soaked day in London last March, MacKenzie took the stage at the Third International Summit on Human Genome Editing to tell an auditorium filled with the world’s leading CRISPR luminaries about her vision for taking that technology into the womb. Speaking softly, with a voice strained from an abating cold, she explained her intent was to alter the DNA inside the somatic cells of a fetus in the second or third trimester of pregnancy.

This was not, she emphasized, to be conflated with embryo editing. “This is my most important slide,” she said, showing an image of a floating sphere of cells with a bright red X through it.

A collective nod of acknowledgment rippled through the room. At the previous summit, in 2018, a little-known Chinese scientist named He Jiankui stunned the world by revealing that he had created the world’s first “CRISPR babies,” twin girls whose genomes had been changed while they were IVF embryos in a dish. It was an experiment that crossed ethical boundaries and would have been illegal in more than 70 countries, including the U.S. The ensuing scandal and the intense ethical debate it ignited still clung like a stain to the emerging field.

To be successful, MacKenzie knew that she would have to show it was possible to edit a fetus without accidentally altering its germline. What she didn’t say was that her team had just discovered that completely avoiding germline edits might be more difficult than she had expected.

Galvanized by the stories of spinal muscular atrophy patients who hadn’t received Zolgensma early enough, and inspired by researchers in Turkey who had successfully used fetal gene therapy to treat the disease in mice, MacKenzie decided to see if the findings could be extended to sheep. In one experiment, her team used a reporter gene — coding for proteins that would glow under a microscope if successfully expressed. What MacKenzie wanted to know was, if they were to deliver a CRISPR gene editor using a viral vector like the one in Zolgensma, in utero, where would it go?

Almost everywhere, it turned out. The genetic instructions went where the researchers had hoped — the brain and spinal cord and muscle cells all glowed. But also, in lesser amounts, they went to locations where they shouldn’t, including the gonads, MacKenzie told STAT. These findings have been presented at conferences but not yet published in a peer-reviewed journal.

Borges, the postdoc who conducted the analyses of all these tissues, remembered the relief he felt when additional testing showed that in male lambs, it was the support cells around the sperm-producing cells that glowed green. That meant their germlines were still untouched. But the same was not true when he looked in the ovaries of the female lambs. “There was clear overlap,” he said. It wasn’t much, but it was enough to be worried.

When he took the results to MacKenzie, she was floored. “It’s completely surprising,” she said. As part of the approval process, all gene therapies and gene-editing-based medicines have to be rigorously tested for the potential to contaminate the germline and be passed down to future generations of offspring. Study after study of these drugs in lab animals have failed to turn up evidence that AAV viruses used in these therapies deliver their genetic payloads into the DNA of sperm or eggs. But few studies have rigorously asked that question for in utero application of such therapies.

“We were the only ones not afraid to look,” MacKenzie said. The results didn’t raise as big of a concern for gene therapy because the AAV vector showed a very low probability of integrating in germ cells. But if you use AAV to deliver a gene editor, “regardless of integration, you’re going to have a heritable event,” MacKenzie said. “So it was really using AAV for gene editing that worried me so much.”

She was so concerned that a month after the London summit, she gathered together a group of gene editing experts, lawyers, and bioethicists at UCSF, and for two days they discussed the data and what it meant for moving forward with fetal genome surgery.

The data showed that the aspects of the uterine environment that are appealing for performing fetal genome surgery — stem cells that are more readily accessible to viral vectors, a tolerant immune system that won’t react to them — also increase the odds that the germline, in particular the fetal germline, could be inadvertently altered. What MacKenzie wanted to know was, if you know there’s a risk, even a teeny-tiny risk, can you really say it’s unintentional?

“Germline genome editing is no-go territory, and the worry here is that you would stray into that territory and unintentionally engage in something that’s taboo,” said Benjamin Hurlbut, a bioethicist at Arizona State University who attended the meeting. “That’s a legitimate and important worry.”

But he sees MacKenzie’s project, despite its futuristic sheen, as essentially a conventional practice of medicine. “There’s a patient that’s going to be severely sick and suffer profoundly but for an intervention,” he said. Embryo editing, in contrast, brings a patient into being for the sake of making the intervention. “Conceptually it’s a very important distinction,” he said.

Alta Charo, a bioethicist and legal scholar who was also in attendance, noted that although the risks are higher in a fetus, heritable germline modification is still probably a remote possibility, because of all the things that have to happen for those changes to be passed down. Women produce about a million eggs over their lifetimes, and if some of those get edited, what are the odds that those eggs get fertilized and implant and become a baby that becomes an adult that goes on to have their own children? “It’s a lot of probabilities multiplied against one another,” she said.

Others, like Porteus, argued that it’s not enough to assume the possibility is remote. Doctors and patients and ethicists need to be able to calculate the risk-benefit profile of editing the fetal genome, he said. “If it’s a one in a million chance but 100% of babies will die without it, maybe that’s a calculus that’s worth it. If it happens every time, maybe not. There is very little in the way of facts about any of this right now.”

MacKenzie came away from the meeting assured, and determined to start filling in those gaps.

Since then, her lab has launched a number of new projects — including a deeper investigation of why the female sheep cells were impacted but not the male ones, and a collaboration with a gene editing company to assess, in mice, whether lipid nanoparticles, a nonviral method for delivering gene therapies into cells, are any better at staying out of the cells that will become sperm and eggs.

Working with Mary Norton, her co-director at the UCSF Center for Maternal-Fetal Precision Medicine, MacKenzie is also pushing the boundaries of what’s possible to learn about a developing fetus, to diagnose genetic conditions earlier in a pregnancy. For the last nine months, anytime ultrasound flagged a potential problem with a pregnancy at UCSF, cells collected from the amniotic fluid have gone to the hospital’s in-house sequencing lab to be cracked open, the DNA extracted, and the 3 billion letters of the genome read out. So far, in about 20% of cases, they are finding a genetic abnormality that allows them to make a diagnosis.

The goal of all of this work is to determine whether there’s a window during development when, with the right tools, it’s possible to find something wrong with a fetus and safely treat it with an injection of DNA-altering molecules that can get into the cells that matter without hitting ones that could carry those changes into future generations. MacKenzie isn’t shy about the fact that it could take many years to figure that out, and that stumbling blocks could still pop up out of nowhere.

It’s not lost on her that it took two decades of experiments in the lab for Harrison to get the science to a place where he felt ready to try fetal surgery in the first human. Two decades of people telling him it would never work and the risks were too high and that he might as well give up.

“When you’re doing these whacko things for the first time, you have to expect to fail, and you have to learn as much as you can from the failures,” Harrison said. “That’s just part of the game. It’s not a game that we’re used to as surgeons or doctors. We like to win. We don’t do things that aren’t going to be successful. Tippi gets that. And her genius is seeing what’s coming next. She really is the future.”