In 1999, a 36-year-old woman in Australia decided to try in vitro fertilization after seven years of attempting to become pregnant. Doctors stimulated her ovaries with hormone shots, performed minor surgery to harvest eggs, which were fertilized in a petri dish and then transferred into her uterus. In four separate attempts, a total of eight embryos made from this same batch of eggs were transferred into her body. All failed to implant.

The woman opted to try again. Researchers at her fertility clinic, Melbourne IVF, made nine new embryos but this time decided to check the chromosomes in each one. Nearly all embryos with too few or too many copies of some chromosome-rather than the normal complement of 23 pairs-will fail to implant or will miscarry shortly after implantation. (Down’s syndrome, by far the most common such abnormality among surviving infants, is caused by three copies of chromosome 21.) To assess the chromosomes, a team of scientists led by geneticist Leeanda Wilton removed one cell from each three-day-old embryo. To each cell they added fragments of DNA-labeled with fluorescent tags-that bound to complementary sequences on the five chromosomes most susceptible to abnormalities. The tags showed that all but two of the removed cells had either an extra or a missing copy of a chromosome. The researchers then transferred the two seemingly normal embryos. Neither took.

The woman went through yet a third cycle. This time, Wilton and her coworkers, taking no chances, turned to a new technique called comparative genomic hybridization to analyze each embryo’s entire set of chromosomes. Only one of the five embryos appeared normal-chromosomally speaking. “It wasn’t a particularly great-looking embryo,” Wilton says. Still, the lone normal embryo was transferred. Nine months later, the woman gave birth to a healthy daughter.

So-called preimplantation genetic diagnosis is now common practice at dozens of in vitro fertilization clinics around the world. The procedure, which was introduced in the early 1990s as a way to determine whether an embryo had inherited the genes that cause fatal diseases such as cystic fibrosis and Huntington’s, can now identify more than 50 different genetic diseases. And as the Australian case demonstrates, the technology is playing an ever expanding role in the success of fertility clinics, allowing physicians to detect severe chromosomal abnormalities and carefully choose embryos for implantation.

But this flood of new genetic information on the unborn is reaching far beyond the world of in vitro fertilization clinics. Amniocentesis and chorionic villus sampling, the two techniques that since the 1970s have been mainstays in the diagnosis of chromosomal problems in fetuses, both now incorporate a panoply of tests for inherited genetic diseases. At the same time, researchers are developing safer, noninvasive means of assessing fetal health, two of which have made inroads in clinics around the world: ultrasound screening of body features associated with genetic abnormalities, and blood tests for proteins and hormones that are markers for Down’s syndrome and other diseases.

All of these new tests vary in risk or usefulness, and some remain highly controversial. Yet they all serve the same purpose: to provide parents with the information needed to select an embryo or to determine the fate of a fetus on the basis of its genetic makeup. “The changes to come are going to be even more profound as we get into more sophisticated genetic testing,” says Rebecca Smith-Bindman, a radiologist at the University of California, San Francisco.

Ultimate Preview

Although studies have proven that chromosomal abnormalities in offspring become more prevalent as a woman ages, researchers have widely varying estimates of how frequently errors occur. Santiago Munn and colleagues at the Institute for Reproductive Medicine and Science of Saint Barnabas in Livingston, NJ, have shown that up to 70 percent of embryos produced in their clinic for in vitro fertilization-typically for older women with fertility problems-have chromosomal abnormalities that would likely prevent them from properly implanting. But traditional preimplantation genetic diagnosis techniques have failed to have a dramatic impact on success rates.

Comparative genomic hybridization-the technique that worked for Leeanda Wilton in Melbourne-could change that. A shortcoming of standard preimplantation testing is that it can analyze only a handful of chromosomes from each embryo. Wilton and coworkers have shown that analyzing all of an embryo’s chromosomes with comparative genomic hybridization detects 75 percent more errors.

Originally developed to detect abnormal amounts of DNA in cancer cells, comparative genomic hybridization compares the number of chromosomes in an embryonic and a normal cell. Researchers remove a single cell from an embryo, labeling its chromosomes with a green fluorescent tag and attach a red fluorescent tag to the chromosomes taken from the normal cell. Next, researchers add both the green and red chromosomes to a set of chromosomes on a glass slide, where they bind to complementary sequences. Software analyzes the fluorescent signals (see “Comparative Genomic Hybridization,” below). If the embryo is healthy, each set of reference chromosomes should have an equal amount of red and green. If there’s more red than green in, say, chromosome 14, it indicates that the embryo is missing a copy of that chromosome and likely will fail to implant.

Comparative Genomic Hybridization

Detecting for chromosomal defects, researchers label DNA from an embryo with a green tag, and DNA from a normal cell with a red tag. Both are added to a slide, where they compete in binding to template DNA. The high ratio of red to green indicates a missing copy of chromosome 14.

But in addition to simply looking at numbers of chromosomes, preimplantation genetic diagnosis techniques can look for specific genes. This involves using DNA probes designed to marry themselves to parts of a given gene. In 1999 Yury Verlinsky, who directs the Reproductive Genetics Institute in Chicago, used this technique to help a woman undergoing in vitro fertilization select an embryo that did not have the gene for a heritable life-threatening blood disease. An international uproar ensued because the parents also picked an embryo that had the same immune system genes as their daughter, who was five years old at the time and had the disease. The parents hoped that doctors could safely transfer stem cells-which have the ability to differentiate into any type of cell in the body-harvested from the umbilical cord of their newborn to help their ailing daughter. A baby was born in 2000, stem cells were transfused and the older sibling is doing well.

Verlinsky, who has strongly defended the ethics of this intervention, argues that qualms about such genetic diagnoses will subside as the technique continues its inexorable progress from cases of concerned parents who carry known disease genes to those of parents with no known genetic-risk factors but who have difficulty conceiving a baby. “It will probably become routine in all in vitro fertilization cycles,” he predicts. It would offer women the option of discarding embryos that have genetic or chromosomal defects. “It’s the only scientific proven way to select an embryo. Everything else is absolutely irrelevant.”

Playing the Odds

Detailed genetic diagnostic tests may soon become routine for women undergoing in vitro fertilization. But for most women who become pregnant, the tools for detecting genetic defects still offer risks and uncertainties.

It is a warm London morning in June, and Kypros Nicolaides is bouncing from room to room in the Fetal Medicine Centre, a clinic just down the block from the city’s majestic Royal Academy of Music. An obstetrician and gynecologist who runs the Harris Birthright Centre for Fetal Medicine at King’s College Hospital and, one day a week, this private clinic, Nicolaides has his own intimate relationship with sound: he has developed an international reputation for pushing the bounds of prenatal ultrasound.

In an examination room, Nicolaides greets a 31-year-old woman, 13 weeks pregnant, who has been anxiously awaiting him. As a technician moves an ultrasound transducer over the woman’s belly, Nicolaides points to a monitor that shows the grainy image of a fetus. He then has the technician focus in on the folds of skin behind the fetus’s neck.

Over the past 10 years, Nicolaides has convinced many of his colleagues that looking for fluid in this region, in a procedure called a nuchal translucency screen, can in the first trimester reveal cases of Down’s syndrome as well as other even more severe syndromes that result from extra copies of chromosomes 13 and 18. The ultrasound shows that this woman’s fetus has a slightly larger than normal nuchal translucency measurement, but Nicolaides assures her that it’s nothing to worry about.

On a separate computer screen, software developed by Nicolaides’ foundation graphically shows the risk this fetus has of a chromosomal abnormality. By the woman’s age alone, the risk factor is one in 650. The nuchal translucency slightly increased the risk to one in 600. But earlier that morning, Nicolaides’ team drew blood from the woman and analyzed it for a protein associated with these chromosome problems and for levels of a pregnancy hormone that serves as a marker for abnormalities. With the blood tests factored in, her risk plummets to 1 in 1,000, making her look like a woman in her early 20s. But then the woman tells Nicolaides that she had previously conceived a baby with Down’s syndrome, a pregnancy that she had terminated. Plugging that into the software, Nicolaides calculates her risk to be one in 180.

The woman and her husband decide then and there that they want to have a chorionic villus sampling performed. Nicolaides explains that this procedure, in his hands, carries a risk of one in 200 or so, maybe higher, of causing a miscarriage. They weigh the options but are resolute. Nicolaides, under guidance of the ultrasound, inserts a catheter into her abdomen and removes fetal cells from the villus-the area that links the part of the placenta known as the chorion to the mother’s uterus. The cells are then sent to the lab for chromosome analysis.

Both chorionic villus sampling and its diagnostic cousin amniocentesis can detect Down’s syndrome, spina bifida and many other disorders with 99 percent accuracy. Adding to their value, they have a low “false positive” rate: when they report that a problem exists, they are hardly ever mistaken. But because, on average, both of these invasive tests cause miscarriages up to 1.5 percent of the time, they are also extremely dangerous.

Indeed, in the opinion of Nicolaides and others, many more women than ought to undergo amniocentesis and chorionic villus sampling, given the risks of the procedures and the odds of detecting a fetal abnormality. This concern has driven the development of Nicolaides nuchal translucency test and several noninvasive blood screens. Vivienne Souter, an obstetrician and gynecologist at the University of Washington, sees great value in these new technologies. “If we can get better screening with higher sensitivity and lower false positivity,” says Souter, “we can potentially save babies’ lives.”

In the United States, a multicenter trial is now recruiting 40,000 pregnant women as subjects and will perform the most rigorous analysis yet of nuchal screens and various blood markers used in both the first and second trimester. Fergal Malone, a perinatologist at Columbia University and the coprincipal investigator of the study, says he expects to have results by the winter of 2003. “If we choose the right combination of screening tests, we could have detection rates as high as 90 to 95 percent with a one percent false-positive rate,” says Malone. “That would be a huge advantage.”

Many clinicians contend that, as ever increasing computing power improves the resolution of ultrasound, screening of the nuchal region and other anatomical features will have an increasing ability to reveal problems. “Fifteen years ago, it was a very blurry image,” says Beryl Benacerraf, a radiologist at Diagnostic Ultrasound Associates in Boston. She published a report in 1985 about the promise of nuchal screening. “As it’s gotten better and better, the images look more like Kodak pictures.”

Fishing for DNA

One of the most promising prenatal diagnostic tests in the works might one day eliminate altogether the need for amniocentesis or chorionic villus sampling. Its strategy: testing the mother’s blood for fetal DNA. For years, Diana Bianchi, a neonatologist and medical geneticist at the Tufts-New England Medical Center in Boston, and others had hoped that fetal cells found in maternal blood might provide a noninvasive means for assessing a baby’s genetic makeup. But, Bianchi explains, the cells proved too few in number and thus too difficult to find. In 1997 a team led by Dennis Lo of the Chinese University of Hong Kong first reported the discovery of fetal DNA floating freely in maternal serum and plasma. Since then, preliminary evidence from Bianchi’s group and other researchers has shown that this fetal DNA can be assessed for chromosomal or gene defects.

Bianchi cautions that the approach has several critical limitations right now-not the least of which is that researchers have determined how to fish out only male fetal DNA, which is easier to see because of its distinctive Y chromosome. But she predicts that given the speed with which this technology has moved forward, in five years it may have developed far enough to actually turn up in the clinic. “Fetal DNA opens a whole area of new research that is going to be huge,” predicts Bianchi.

As ever more information pours out about the relationship between genes and diseases, and as screening tools become more sophisticated, prospective parents will inevitably face harder decisions about the types of babies they want to bring into the world. New tools such as DNA microarrays could soon enter the arena of preimplantation and prenatal genetic diagnosis, allowing physicians to screen DNA for a host of genetic diseases simultaneously.

Prenatal screens and diagnostics cannot eliminate all of the uncertainties about the health of a newborn. They can dramatically reduce them, though, and that is the aim of the technologies. But the same diagnostics also raise vexing ethical questions. As the technology continues to advance, parents will gain the ability to determine whether their offspring have genes that “predispose” them to diseases such as Alzheimer’s or breast cancer. In the end, these future tests could themselves introduce ethical ambiguities that no amount of science and technology can overcome.

Genetic Diagnostic Tools