The young couple looked at me expectantly as I re-read the amnio report and tried to decide what to tell them.
“The ultrasound from 15 weeks looks fine,” I stalled, trying to present the good news first.
“What about the amnio?”
“Well, there is something unusual. It’s the Y chromosome. Part of it appears to be flipped. What we call an inversion. It’s a little like if the exits on a highway turned around,” I explained.
The couple looked at each other, puzzled, then back at me, and he reached for her hand.
“What does that mean?” they both asked.
I couldn’t really say. At that time, the mid 1990s, about all we knew about the Y was its diminutive size, its role in determining sex, and that hunks of it were missing in some men who couldn’t father children. We didn’t yet know that Y chromosome DNA is organized like a hall of mirrors, riddled with inverted repeats. (For a brief history of the stunted male chromosome, see “Y Envy”.
“It probably doesn’t mean anything. The first step is to see if you, Dave, also have it. If you do, we’ll know it isn’t dangerous because presumably you’re normal.” Dave smiled.
I doubted he had an abnormal Y, because he was obviously okay in the fertility department.
But Laurie was ahead of me. “And if Dave’s Y is normal? Then what?”
Indeed. I hesitated. “I can search the literature to see if anyone’s reported disease genes near the places on the chromosome that flipped. But our visualization techniques aren’t really good enough to be sure which genes might be involved.” That was an understatement.
The couple left, upset. Dave had his blood drawn, a karyotype (chromosome chart) constructed, and sure enough, his Y was inverted in the same way as his son-to-be’s. The boy is fine.
A few months later, I counseled another couple whose amnio showed an inversion. The woman, like Laurie, had had amniocentesis because she was of “advanced maternal age” – 35 – and therefore statistically more likely to have a fetus with a chromosome abnormality.
But unlike the other case, this couple’s fetus had an inversion of an autosome, a large, gene-dense, non-sex chromosome. That meant either parent could have passed it on – or the chromosome could have flipped anew in the egg or sperm.
The lab had done its best to localize the bands on the chromosome where the DNA molecule had somersaulted, but I knew a band could harbor dozens of genes – these were the days before the genome had been sequenced. So I collected reports of mutations in these regions, knowing the information was too crude to do much more than raise anxiety levels.
“And so what happens if neither of us has the inversion?” asked my patient.
“Let’s hope one of you does have it. But if not, the risk that there’s a related health problem, based on other cases, is about 11%.”
“What? We thought prenatal diagnosis gives an answer! Isn’t that what ‘diagnosis’ means?”
Like Dave, the couple had their chromosomes checked. But unlike his and Laurie’s situation, their inversion was “de novo.” Neither parent had it. We’d watch the pregnancy with ultrasound, but we just didn’t know what might happen. The remainder of the pregnancy was stressful, but that baby turned out okay too.
These two real cases were prophetic, for genetic uncertainty today, even with all the new tools of the post-genome era, remains a problem.
“Genetic counselors face uncertainty on a day-to-day basis. We see mosaics, balanced de novo translocations, inversions, marker chromosomes … the frequency of all of those is about 1 percent,” Ronald Wapner, MD, told me. He’s the Director of the Division of Maternal Fetal Medicine at Columbia University and co-author of 2 of 3 papers published in this week’s New England Journal of Medicine.
Microarrays and Karyotypes - we need them both, for now.
I’m betting that Dr. Wapner’s two papers - first and second - are the ones that make the news. They have the largest numbers and deal with common conditions – autism, developmental delay, intellectual disability, and stillbirth. The two papers use chromosomal microarray (CMA) testing to find copy number variants (CNVs) – which are either missing DNA sequences or extra copies of them, too tiny to show up on traditional chromosome charts.
“A routine karyotype provides little information about autism, and normal results provide little reassurance for concerned families. CMA can’t rule out autism, but it can detect several known causes,” explains Andy Faucett, MS, CGC, Director of Policy & Education at the Genomic Medicine Institute at Geisinger Health System. He’s working with Dr. Wapner in data analysis.
But it’s the third report in the NEJM, with one key case, which struck me, because it reminds me of my long-ago inverted patients.
”Clinical Diagnosis by Whole-Genome Sequencing of a Prenatal Sample” uses a sped-up method of whole genome sequencing to pick up what microarrays won’t, because “rearrangements” – inversions and the chromosome-swaps called translocations – are balanced. They don’t alter the numbers of short DNA sequences, which is the target of CMA.
The research, from the lab of Cynthia Morton, PhD, at Brigham and Women’s Hospital, where she is Director of Cytogenetics, uses genome sequencing to precisely define what goes wrong at the breakpoints of a translocation.
If an inversion is like flipping a series of highway exits, then a translocation is like swapping half of New Jersey with half of Montana. And like a map with new borders, moving parts of chromosomes generates DNA sequences that are not usually next to each other – which is what genome sequencing picks up that microarrays cannot. Sometimes the new borders won’t hurt anyone, but sometimes they’ll disrupt something vital – both in the geographical analogy and in the genome.
The major case in the third report in the NEJM concerns CHARGE syndrome. That’s an acronym of the first letters of a list of serious problems. The first to show up on ultrasound during a pregnancy is the “H,” for heart. And that’s what alerted Zehra Ordulu, M.D, a fellow in Dr. Morton’s lab. (“I’m just a post-doc” is how she introduced herself to me at the American Society of Human Genetics meetings a few weeks ago).
“At 19 weeks, fetal imaging revealed a severe heart defect,” Dr. Ordulu told me. By 30 weeks the upper lip stuck out and by 33 weeks the fetus was floating in an awful lot of amniotic fluid. Plus, the head and stomach were too small and the extremities held in an odd position, something seen in certain syndromes.
The mother-to-be continued to balloon, and when a doctor removed two quarts of the fluid, fetal cells were sent off for karyotyping. And that revealed that one copy of chromosome 6 had swapped parts with chromosome 8. New Jersey and Montana. The microarray test was normal. The researchers then sequenced the fetal genome in other cells saved from the amniotic fluid.
It took another few weeks to check the parents’ chromosomes, and by then, the baby had been delivered and was desperately ill with full-blown CHARGE syndrome. He was taken off life support on the tenth day after birth. (See Medscape Medical News for details).
Should the parents have ended that pregnancy, had they known the full spectrum of problems? That’s a question only for them. But had the balanced translocation been detected earlier with chorionic villus sampling, coupled with the heart defect on ultrasound, led to the genome sequencing that identified the precise mutation (in a gene called CHD7), they would have had a choice. They could have ended the pregnancy or the medical team could have prepared to handle a more serious situation, Dr. Ordulu said.
Genome sequencing will prevail
A turf war seems to be going on in genetic diagnosis.
The news release announcing the microarray papers quotes one of the researchers: “Based on our findings, we believe that microarray will and should replace karyotyping as the standard for evaluating chromosomal abnormalities in fetuses.” But microarrays would have missed the CHARGE baby.
Perhaps we shouldn’t be too hasty in declaring a diagnostic technique obsolete.
I think one of the oldest technologies is still the most telling: ultrasound. Figure 1 in the third NEJM paper shows the fetus’s face and the shrunken right heart ventricle. In my two inversion families, the ultrasounds had been normal.
The articles and editorial in this week’s NEJM show how technologies that hover around birth – prenatal, stillbirth, and pediatrics – complement to get at the same information. And they dance around the elephant in the room: what to do with the information, for that is a personal choice and beyond the scope of a medical journal.
Dr. Morton is brave enough to outline aloud the slippery slope. “If you’re going to do genome sequencing on a newborn, what is the difference doing it 20 weeks earlier? The difference is that you could decide you didn’t want to continue the pregnancy. But the analysis would be essentially the same.”
Yes, context is important. Learning a certain DNA sequence is present during pregnancy can lead to a choice to abort, near the time of birth can guide treatment, and during childhood can refine a diagnosis.
Genetic testing can help patients make decisions if the information means something specific, but often, it doesn’t. In fact, these journeys often bring intense anxiety due to the uncertainty – what one researcher calls “toxic knowledge”. For just as I couldn’t tell my inversion patients whether their fetus’s flipped chromosomes would do harm, so today are parents-to-be today left high and dry with a lab report of finding a “variant of uncertain clinical significance.”
But genetic certainty is impossible, cautions Dr. Wapner, and I agree. Even when we know everyone’s genome sequence, it’ll be a long time, if ever, when we know how all of our genes interact.
For now, while we’re learning about our genetic selves at an ever-increasing pace, but at the same time uncovering even more uncertainty, the scientists and physicians at the frontlines can survive by not looking beyond their roles in getting the information.
Says Dr. Wapner, “There are good reasons to have microarray information as early as possible. We’re not even suggesting that a lot of this information should be used for termination. Some conditions will be treatable, maybe even in utero. We will find different genes and CNVs that contribute to stillbirth, congenital problems, and miscarriages. There won’t be only variants of uncertain clinical significance, but also lots of good information that will help parents, and can help us understand which genes are involved in development.”
Dr. Morton’s group will be sequencing newborn genomes to address two actionable problems: congenital heart disease (like the CHARGE baby) and deafness.
Nearly 80% of newborns who fail hearing exams have blocked ear canals -- vernix (the cheesy crap that’s never on babies delivered in films or on TV) or amniotic fluid, like a kid who’s stayed in the pool too long. Finding, or not finding, a genetic cause – such as mutations in the connexin genes – can bring relief to many families.
But why wait until birth? “You can imagine in the future we won’t be waiting until that baby is delivered to do the sequencing -- it will become part of prenatal evaluation,” says Dr. Morton, returning to the elephant in the room.
The challenge of whether and when to deploy these genetic diagnostic tools of pregnancy and early childhood is going to vanish. Soon. Because we can already sequence a fetal genome (see my blog post on this).
In fact, possibly within 10 years, many a diagnostic work-up will begin with a quick look at the genome. Will we be ready?
The issues raised in this week’s New England Journal of Medicine suggest that before that happens, we have a great deal to think about.