That’s really super, supergirl (part 2)

6 12 2018

I spent a fair amount of time in the last post going over the technical aspects of gene transfer, in large part because so much of the concern about the prematurity of Jiankui He’s work centers on those technical aspects. (Ed Yong at The Atlantic offers a terrific roundup of those and other concerns; he also collected a snapshot of initial responses to He’s announcement.)

I want to focus on two, different, questions about germline gene transfer: the point of such gene editing, and what it means for those who’ve been so edited.

First, the point. If your concern is with preventing single-locus (Mendelian) disorders, there’s a much more straightforward way of dealing with the risk: test the embryo for the presence of the lethal genes discard those which test positive (two lethal recessives or one lethal dominant).

This process, known as preimplantation genetic diagnosis, is available at most, if not all, fertility clinics in the US, and generally costs in the low thousands; in fact, some of those who don’t have fertility problems choose to make use of assisted reproductive techs (ART) precisely so the embryos can be tested prior to transfer. If you know you and your partner are carriers for, say, cystic fibrosis, it’s a lot easier simply to transfer “healthy” embryos than to try to edit affected ones.

The upshot is that germline gene transfer for single-locus disorders is unnecessary. (The exception might be for those who create embryos which are all affected, but even then, it might be easier to create new embryos than to edit out problem genes.)

This brings up the question, then, for what germline gene transfer could be necessary, and it appears, at this point, nothing; it could be used only for improvement or enhancement.

Now, I should point out that “enhancement” is looked upon with some suspicion by many, many bioethicists, so using that term is. . . provocative. Still, it’s not unwarranted: He altered a normal (or non-disease) gene in order to enhance the offspring’s resistance to HIV. While it’s questionable as to whether the twins will actually have that greater resistance, the clear intent was create people with a capability they would not otherwise have had.

Otherwise known as enhancement.

I am a skeptic of genetic enhancement, not least because most of our traits are complex or multifactorial. Do you know how many genes are involved in your height? Over 700. Any guesses as to how many are involved in, say, intelligence? Your guess would, at this point, be as good as anyone else’s.

Furthermore, many of our genes are pleiotropic, which means that a single gene may be associated multiple traits. And let’s not even get into epigenetics, which is the study of the process by which environmental factors affect gene expression.

All of this means that attempts to edit our genomes in order to enhance the traits so many express an interest in enhancing (eg, height, intelligence, athletic ability) will not be straightforward. This doesn’t mean that all such edits will fail, but that success is likely far off.

There are some traits which are less complicated, traceable to one or a few genes, so it may be possible to fiddle with those genes, but even then there’d be concerns, as there is with the CCR5 gene He edited, that boosting one aspect of the gene’s expression (resistance to HIV) can cripple another (resistance to West Nile virus).

That germline gene editing may not, strictly speaking, be necessary, doesn’t necessarily mean there’s no point at all to it. Even an enhancement skeptic like me can recognize that not every use is automatically terrible, or that, in the case of an environmental disaster or pandemic, it could actually become necessary for species survival.

But we’re a long way away from knowing enough that such use can currently be justified.

Which brings me to the second point: what happens to those little girls, Nana and Lulu? Are they to be research subjects for the rest of their lives? Will their parents be required to offer them up for study? Will they ever be able to say no to such study? How much of their lives be known? Will they have any control over information about them?

And what about their offspring? Will their own children have to be studied? If, as seems probable, Nana and Lulu are mosaics, then there would certainly be interest in the inheritance of those mixed genomes.

If He’s work is not to be a complete waste, the girls should be studied. But how to balance the need/desire for knowledge about his experiment with their human rights and dignity? After all, they didn’t sign up for any of this.

I should point out that in some ways their birth parallels that of Louise Brown, the first IVF baby. No one knew if creating a human embryo outside of the body and then transferring it back to a woman would result in a healthy child, or whether IVF-offspring would themselves be fertile. (It wasn’t until Louise’s younger sister Natalie, also IVF-conceived, gave birth did we know that IVF babies could make babies the old-fashioned way.) In vitro fertilization (and a variation, intracytoplasmic sperm injection) was an experiment which could have ended in horror; that it didn’t has had the effect of minimizing just how great a leap it was.

So what will happen with Lulu and Nana? If all seems well with them, does that make it all okay? If not, then not?

And, again, how will we know? One of the criticisms of the fertility industry is just how much isn’t known: there is no database of children conceived via ART, nor of women who’ve taken fertility drugs. Yes, it is possible to do research on the health of these women and kids, some of which indicate increased risks to health. Is that work sufficient? It’s been 40 years, and it seems mostly okay; is that good enough?

We can’t go back and retroactively require that all ART babies be surveilled—I’m certainly not suggesting that—but would it make sense, going forward with gene-edited people, to have some way to keep tabs on their health?

Y’all know I’m a privacy crank, so even suggesting some sort of life-long surveillance makes my teeth itch, but if such research is to continue, then a necessary part of that research is information about the people who participated in it. Given that a central tenet of human subjects research protection is the right to withdraw from any study at any time, there’s no ethical way to require people who’ve been gene-edited to submit to lifelong study; it is not out of the question, however, to ask.

Anyway, back to Nana and Lulu, two new people who were created as a science experiment which many of us decry. Would it have been better had they never been born? Better, certainly, had He not plowed past the many cautions to mess with the embryos, but now that the girls are here, well, best to welcome them to the human race.

Advertisements




That’s really super, supergirl (part 1)

28 11 2018

“What’s so bad about superbabies?”

It’s a question I ask my 300-level bioethics students after we’ve spent 12 or so weeks discussing eugenics, genetics, stem cell technology, somatic cell nuclear transfer, and the varieties of assisted reproductive technologies. They’ve also written their first set of research papers—some of which take up the question of just how humans could be enhanced (“should” is taken up in the second paper).

Most of them tend to recoil from the question, even if they can’t exactly say why, while a few take a Hell, yeah! Why not? approach to enhancement.

And, oh yes, we spend a fair amount of time trying to figure out just what counts as “enhancement.” Therapy or treatment, I note, is generally accepted by bioethicists as Good, but enhancement is looked upon with considerably more suspicion. Normal is good, well is good; better than well? Ehhhh.

Which is a long way of saying that there are a lot of us who are primed to react to the news that germline genetic engineering has not only been attempted, but accomplished, in human beings for the first time.

Germline genetic engineering (or transfer) is distinguished from somatic-cell gene transfer in that any genetic changes will be passed down to offspring—which is to say, the germ cells (sperm & eggs) are themselves changed. The most efficient way to achieve this is to alter the cells of an early embryo (a 2-3 day-old morula or 4-6 day-old blastocyst): at this stage the cells are undifferentiated, so if you are able to insert altered (“recombinant”) genetic material (rDNA or rRNA) into all of these cells, as they divide and specialize they will carry the altered material into every cell—including, of course, the germ cells.

One of the great challenges in gene transfer of whatever sort has been getting rDNA/rRNA into enough cells to affect function; that the morula and blastocyst have so few cells bypasses this problem. Another issue has been inserting rDNA/rRNA into the correct place in the genome—and only that one place in the genome—to affect genetic expression. (There are other issues, especially regarding the vectors , or delivery vehicles for recombined sequences, but I’ll skip over these for now.)

Okay, so you’ve heard of CRISPR, yes? Clustered regularly interspersed short palindromic repeats? Why, of course, Terri; who hasn’t filled their days wondering about clustered regularly interspersed short palindromic repeats? Skipping (almost) alllllll of the technical stuff, CRISPR is a huge advance over other forms of gene transfer and gene editing: by using a guide enzyme, eg, Cas9, researchers are able precisely to target a specific sequence. . . .

Well, shit, I’m losing you, aren’t I? I start with superbabies and now I’m talking about “targeting specific sequences.” So, some basics:

Almost all of our cells (exc: red blood cells) contain a nucleus, which in turn contains chromosomes, which are histones (a protein) wrapped in a very tight spool of DNA.

DNA comprises a sugar-phosphate double-helix connected by the nucleotide base-pairs (bp) adenine (A), which is always paired with thymine (T), and guanine (G), which is always paired with cytosine (C). Determining the exact order of these base pairs on a strand of DNA is “sequencing”; a sequence will look like this: ATTCCAAGGGGTAACAATTCGACCTGAT. . . .

A strand of DNA is about 6 feet long and is mostly noncoding: it has no direct role in protein synthesis. (I should note that the non/functionality of this noncoding or junk DNA is disputed, which is neither here nor there for this discussion but thought I should mention it anyway. If you don’t understand what this means, don’t sweat it.)

A very small portion of the DNA, less than 2%, comprises genes, or coding sections of DNA. Genes are the Action Jacksons of DNA, the functional units of heredity, involved in protein synthesis, and, unlike in the noncoding sections, if something is messed up with your genes it can affect how you function. The average gene size anywhere from 10-27K bp (yes, disputed), with the largest, dystrophin, over 2.3M bp.

So: DNA is the set, genes are the (very small) subset. Most members of our species have about 19,000 genes scattered amongst 3.2 billion base-pairs and across 46 chromosomes. (Some of us have more chromosomes, some have fewer, but most of us have 46.) Some chromosomes have thousands of genes, some (well, the Y chromosome) has fewer than 100.

This sounds like I’m going off track, but, really, there’s a point behind all of this.

I like to compare DNA to a road, genes to cities and towns, and chromosomes to different regions of the US: some chromosomes are like the northeast, packed with many genes, and others (the Y chromosome) are like South Dakota, with those city-genes few and far between.

To change gene NYC, say, you need to be able to get a package of rDNA or rRNA to the right place on the sequence, and then (to continue the analogy), you have to get that package not just to the right city (gene), but to the right borough, to the right street, the right address, the right apartment, and to the right room in that apartment.

This is “targeting.”

The other issue is to get the right package to the right city in the right borough, etc., in a sufficient number of cells, with “sufficient” often meaning “millions and millions.” (Again, with an embryo, this would appear to be less of an issue.) Oh, and once the package is delivered, it needs to be opened and turned on, i.e., integrated into the gene and made functional.

So, targeting, sufficiency, and functionality have all been obstacles to successful gene transfer. In fact, only one gene transfer product, involving immunotherapy, has been approved for use in the US, and that only happened last year.

Now CRISPR, CRISPR is highly efficient at targeting and integration, and is often referred to as a “cut-and-paste method” of gene editing: it uses an enzyme to guide the package to the targeted sequence, snip out the old sequence and insert the new. Another analogy: think of the “find and replace” function in a word processing document. You type the old word (say, “Chris”) into find, the new word (“Kris”) into replace, and hit go.

Much, much MUCH more efficient than manually looking for “Chris” and replacing it with “Kris.”

Great, yeah? Wellllll, not always. One of the problems with CRISPR is off-targeted mutations. Say you only want to replace Chris with Kris on page 1, but it find-and-replaces throughout the entire 20-page document. Or maybe it replaces “chris” in words like “Christmas” or “christen” or “christallmighty!” with “Kris”.

That would be bad for your document; it would be even worse for your genes.

Also—and here’s where the analogy to find-and-replace breaks down—the cut-and-paste doesn’t work every time. CRISPR efficiency is much higher than that for other methods, but it’s not 100 percent. This means that some cells will take up the changes and other cells won’t, resulting in mosaicism, or an organism which contains more than one genome.

Mosaicism is also a natural phenomenon, and not necessarily a dangerous one, but it could at the very least affect the efficacy of the gene transfer and the function (aka health) of the organism.

Okay, that’s enough for tonight. Next: whatsamatta with what He did?