10

A spoonful of mannose-6-phosphate

Physicians of the Utmost Fame
Were called at once; but when they came
They answered, as they took their Fees,
‘There is no Cure for this Disease.’

HILAIRE BELLOC

Jesse Gelsinger died on 17 September 1999, just a few months after his 18th birthday. Gelsinger had a rare genetic condition that affected his body’s ability to safely dispose of excess nitrogen. He had been a volunteer in a clinical trial that aimed to see whether a virus, which had been modified to carry a normal version of a gene called OTC into the cells of the liver, could safely be given to humans. Sadly, the answer was that it could not — Gelsinger’s body reacted in an unexpected and extreme way. Over the course of four days, his organs progressively failed, and, despite all efforts, he could not be saved.

The clinical trial in question was one of the early attempts to perform gene therapy — a treatment designed to directly replace a gene that is faulty or missing, by inserting a working copy into the DNA of the person affected by the condition. There are formidable technical challenges that stand in the way of gene therapy, and there were some known risks from trying it. But nobody had really thought that a research subject might die in this way. Jesse Gelsinger’s death was a shock to the whole field, and it seemed to many of us in genetics that the very idea of gene therapy might have died with him.

*

At the public announcement of the completion of the Human Genome Project in 2000, both US president Bill Clinton and British prime minister Tony Blair spoke of the ideal that our new understanding of the genome would lead directly to treatments for medical conditions. Both mentioned cancer; Blair also referred to the treatment of hereditary diseases. Over the years that have followed, there has been quite a bit of criticism of the field’s failure to produce cures — a recent newspaper article asked, ‘Was the Human Genome Project a Dud?’, and concluded that yes, it was. Fortunately, this is far from the truth — as we’ve seen, the HGP has unquestionably delivered for people affected by genetic conditions and their families, in all sorts of ways. Treatments for cancers have, in fact, been developed based on knowledge gained from the HGP. By sequencing the genome of a cancer and comparing it with the genome of the same person’s healthy tissues, it is increasingly possible to recognise the specific genetic damage that has driven the development of the cancer. There are many treatments — existing and in development — that work directly to counteract the effects of those changes.

But as for cures for genetic conditions … well, that was always asking for a lot.

The reason it’s hard to cure a genetic condition is that the problem lies so deep within. The cell nucleus is a walled fortress, protected against changes for the very reason that change is dangerous; damage to DNA can kill cells, and those that survive risk becoming cancerous. The idea of gene therapy is generally to replace something that’s missing, but, for many genetic conditions, ‘something missing’ isn’t the problem. Sometimes, the faulty gene is overactive, so the problem is too much rather than too little. Sometimes, the faulty gene causes the cell to make a toxic substance, such as an abnormal protein that builds up inside the cell and poisons it. And sometimes, even if the problem is ‘something missing’, it’s in the form of ‘something missing at a particular time’, such as in the early weeks after conception. If that’s the case, replacing the gene in an adult, or even in a newborn baby, may mean that you are acting far too late to make a difference.

Even if you can get a working copy of a faulty gene inside the nucleus, and then get it inserted into the cell’s genome, and then get it to turn on and function normally — there are risks. It’s hard to control where the new DNA goes; if you’re unlucky, it might wind up somewhere you don’t want it, and cause other problems. Most gene therapy relies on using a modified virus, since there are some types of virus that have already solved the problem of getting their DNA into the cell’s nucleus — that’s part of how they hijack the cell’s machinery to make new viruses. But that means you have to engineer the virus in a way that stops it doing harm, without stopping it from doing the job you need it to do.

So, long before there were successful gene therapies — and as we shall see, there are finally some successful gene therapies — people turned to other approaches to try to treat genetic disease. For instance: if it’s hard to change the DNA inside a cell, why not swap out the cell itself, replacing it with a healthy cell?

*

I know it’s just coincidence. Still, sometimes I see several people with the same very rare condition in a short period of time, and it feels like there must be something in the water — or like the world is out to get me. In the space of a six-week period, I once saw two children whose families were presented with a deadly dilemma.

Both boys were toddlers, and their stories were strikingly similar. A series of seemingly minor problems had led them to see a range of health professionals. They had both needed hernia repairs. They had frequent colds, chest infections, and ear infections. After they turned one, they were slow to learn to walk, which eventually led to them seeing paediatricians. Each of those paediatricians ordered a screening test on urine, and that led to a devastating diagnosis.

Ethan and Angelo both had Hurler syndrome, a condition that affects the lysosome. If mitochondria are the cell’s powerplants, the lysosomes are its recycling centres — structures that contain a set of enzymes that break down materials in the cell that have passed their use-by date. There are 40-odd different enzymes inside the lysosome, each with a different recycling task. One might recycle the aluminium cans of the cell, another its scrap paper, and so on. If you’re born with a deficiency of the paper recycler, loads of used paper are still being delivered to the lysosome by the cell, but nothing happens to it; there’s nowhere for it to go, and it can’t be processed. If you put a tissue sample from someone in this situation under the electron microscope, you can see the lysosomes transformed from small balls, scattered through the cell, into large blobs that grow over time until they can fill the cell completely.

The lysosomal storage disorders manifest in different ways, depending on which enzyme is deficient. Some mainly affect the brain, others the liver and spleen; some affect nerves, or the heart. Hurler syndrome affects most parts of the body. The affected child’s liver and spleen enlarge progressively. The heart valves stiffen and the heart muscle can fail. Cartilage and bone grow abnormally, so that children are very short and can have serious spine and joint problems. The tongue grows larger over time, and facial features become heavier, an appearance that is often (rather unkindly) described as ‘coarse’. Joints stiffen, there can be problems breathing, especially during sleep … and, worst of all, there are progressive effects on the brain. At first, an affected child’s development is normal, but it slows and then stagnates. Finally, as the neurological damage progresses, skills are lost, and, by the time of death, often before the age of ten years, there is profound disability.

This was the grim news I had to give to each child’s parents about their young son.

I also gave them hope, of a sort, and a choice.

Like many genetic conditions, Hurler syndrome is part of a spectrum. With just a little bit of enzyme that works, there can be a less severe picture: the brain might be spared altogether, the first symptoms might start later, and the progression of disease might happen more slowly than in a child with the typical severe form. With just a bit more enzyme again, the picture can be very different indeed. I once met a woman in her 30s who was a little shorter than average, and who told me that she had had severe joint problems — but looking at her, I would never have guessed that she had a variant of Hurler syndrome.105 What’s more, the differences in enzyme function between those who are most severely affected and those with less-severe problems are tiny — one study found that children with the severe form of Hurler syndrome have about a fifth of 1 per cent of normal enzyme activity; those with the in-between form have about a third of 1 per cent; and those with the least severe form, like the woman I met, about 1–2 per cent of normal enzyme activity. It’s very likely that, if you had 3 per cent of normal activity, you would be completely healthy, and virtually certain that you would with 5 per cent.

[1 She had Scheie syndrome, which was once thought to be a separate condition but was later discovered to be effectively a less severe, or attenuated, form of Hurler syndrome. This group of conditions, the mucopolysaccharidoses (try saying that five times quickly) are thus numbered I, II, III, IV, VI, and VII. There’s no V any more.]

This pattern is true for many different conditions that are caused by enzyme deficiencies. The body makes a lot more of most enzymes than it needs. When researchers found this out, they drew the obvious conclusion: we ought to be able to treat people with these conditions, if we could just get a little enzyme back into their bodies.

One of the ways this has been done, and done very successfully, is to manufacture enzyme outside the body and give it to people in the form of regular infusions into a vein. Getting this to work was no simple task. The first condition for which enzyme replacement therapy was developed is called Gaucher syndrome. The idea of treating Gaucher syndrome this way was suggested by Dr Roscoe Brady in the 1960s, and it took until the 1990s before there was a fully-fledged enzyme treatment on the market. Brady and his team spent much of the 1970s extracting tiny amounts of enzyme from placentas, but eventually it became possible to manufacture the enzyme using Chinese hamster ovary cells106 that have been modified to make the human enzyme in large quantities. One of the discoveries that was made along the way was that, to get the enzyme to where it needed to be, a string of sugars attached to the surface of the protein had to finish with one particular sugar, called mannose-6-phosphate — hence the title of this chapter.

[2 I am not making this up.]

Enzyme replacement therapy works well for many storage disorders, up to a point. It’s great for the soft, squishy parts of the body with a good blood supply to deliver the enzyme where it’s needed — the innards, if you will. People with Gaucher disease can have hugely enlarged livers and spleens; the treatment melts those away like ice in the sun. However, for conditions where bones and joints are severely affected, the results tend to be marginal, at best. Worst of all, enzyme can’t get into the brain. For someone with severe Hurler syndrome, giving them enzyme might help with some of the effects of the condition, improving quality of life for a time, but it would not change the long-term outcome.107

[3 For people with the less severe forms of Hurler syndrome, whose brains are not affected, the benefits are clearer.]

There was another option for Ethan and Angelo, a medical roll of the dice. There is a way to get healthy cells into a person’s brain, cells that are capable of producing enzyme that can then be absorbed by the brain cells. The way to do this is far from obvious: you start by poisoning the patient’s bone marrow. Then you replace it, using stem cells from someone else’s marrow. This is bone marrow transplantation,108 and it was developed for treating leukaemia and lymphoma, and, less commonly, some other types of cancer. The concept is simple enough: you have a cancer of the immune system, so you get rid of all of the cells of the immune system completely, and replace them with healthy ones.109 Another application, not surprisingly, is treating disorders of the immune system and other blood conditions. An immune system that doesn’t work properly is replaced with one that will.

[4 More accurately: haematopoietic stem cell transplantation, or HSCT. Haematopoietic stem cells are cells that are capable of making blood cells of all types, and they include cells taken from marrow, but also can be cells taken from the placenta via the umbilical cord after a baby is born (cord blood), or even extracted directly from the blood itself.]

[5 There are other ways this can work. Sometimes, for cancers that don’t involve the bone marrow, the patient’s own marrow is extracted and stored. This allows the use of powerful treatments, such chemotherapy, that would otherwise be fatal because it destroys the bone marrow as a side effect. Once the treatment is over, the patient’s marrow is restored from the backup that has been saved.]

The reason this might work for children with conditions like Hurler syndrome is simple enough: because infection can develop anywhere in the body, white blood cells are needed everywhere. Helpfully, they produce extra enzyme, which other cells can absorb and use. Your brain contains a lot of white blood cells — something like a tenth of all the cells in your brain are a special type of white cell, called microglia. Replace those, and you have a rich supply of enzyme, right where you need it.

There’s a catch. There are several catches. The first is that the treatment is a very big deal. The drugs that are used to kill off the child’s own bone marrow are — not surprisingly — toxic. In part because of this, and in part because of the risks of overwhelming infection while the new immune system establishes itself, there is a substantial risk that the child might die from complications of the bone marrow transplant itself. Those who survive may have lasting side effects from the drugs, or the new immune system may attack their bodies (‘graft versus host disease’). Then, even if the transplant goes perfectly, it takes at least six months for the new white cells to reach high enough levels in the brain to do any good, time in which damage is steadily getting worse. Both boys had early signs that their brains were affected already. The outcomes from a transplant for their brains were uncertain, although it was very likely they would have at least some long-term impairment. And, lastly, this would not be a cure — it would at best change a fatal condition into a chronic one, with gradual worsening of the bone and joint problems in particular.

Faced with a choice like this, what would you do?

In genetics, we make a virtue of non-directive counselling. The idea is that we give people information that empowers them to make their own choices, rather than telling them what to do. Not everyone I see is comfortable with this model — we are used to getting advice from our doctors, after all — and it’s fairly common, in a difficult situation, for people to ask me, ‘Yes, but what would you do?’ There are good reasons why this information may not be helpful. The choice that is best for a married medical professional in his 50s, who already has adult children, may not be best for a 19-year-old single mother, or a 44-year-old couple who are facing a possible problem in what may be their only pregnancy ever. So I don’t often answer this question directly; but, of course, I always have an opinion.

Not this time. I have absolutely no idea what I would do in this situation; the choice seems impossible. Accept that my child has a progressive condition that will lead to his death within the next seven or eight years, and focus on giving him as much happiness as possible in his short life? Or take a chance on a treatment that will certainly cause short-term suffering, might kill him outright, might cause serious, lasting side effects, is unlikely to completely save his brain from the effects of the disease, and will leave him with a lifetime of serious, painful bone and joint problems?

Ethan’s parents made one choice, Angelo’s made the other.

Whose parents made the right choice? Both. Neither. I don’t think there is an objective right or wrong here.

Fortunately, dilemmas like this are rare, but it’s not unusual that a treatment for a genetic condition is only partially effective, changing one condition into another, and, especially, changing a life-shortening condition into a long-term disability. We are getting better at it, and there are already many conditions for which there are specific, targeted treatments. Most of these don’t work directly at the level of the gene, but tackle some aspect of the condition’s biology — aiming to restore balance to some system within the cell, or in the body as a whole. Your body can’t process X and it builds up, causing you harm. So we give you a medication that stops your cells from making any X, and perhaps put you on a low-X diet as well. Your body can’t make Y, so we give you supplements of Y. That sort of thing.

There’s a specific group of conditions, called inborn errors of metabolism — mistakes in the way the body’s chemistry works — that particularly benefit from this kind of approach. For some people,110 it can be as simple as taking a regular large dose of a specific vitamin, to kick a sluggish protein into action or prop up a chemical reaction that isn’t quite working properly. By a strange twist, our patients in this group benefit from the existence of the vitamin and supplement industry. Huge numbers of people take large doses of vitamins that they do not need.111 This creates a market (and competition) that results in those vitamins becoming available to the very few people who actually do need to take them, at much lower prices than would otherwise be the case. If you’re someone who has been conned by the industry into taking unnecessary vitamin supplements, you can take some comfort from the knowledge that you aren’t just creating expensive urine. You’re also helping a group of patients with rare conditions. Thank you!

[6 I don’t want to give the impression that all metabolic conditions can be treated this way. Only a very small subset of people with inborn errors of metabolism have conditions that can be effectively reversed by treatment with vitamins, although there are more who benefit to a smaller degree.]

[7 Of course, there are people who do need them. If you have a vitamin deficiency that was diagnosed by an actual doctor, or have some other medical condition that requires vitamin supplements, for goodness sake keep taking them! And if you are a woman who is planning a pregnancy, it’s really important that you take folic acid, because it will reduce the risk of some serious health problems in the baby, and there are some other vitamins that are worth considering. But if you started taking vitamins or other supplements without medical advice, because an advertisement told you it would ‘boost vitality’ or will ‘make you feel better’ or ‘support your immune system’ or something similar … it’s very likely indeed that you would be better off just eating food. Even a moderately well-balanced diet is likely to contain all the vitamins you need. Taking extra won’t help you.]

Cantú syndrome is a condition for which this type of medical balancing act really ought to work, and recently we gave it a go.

Soon after the Dutch-led discovery of the main cause of Cantú syndrome, Kathy Grange made a different type of discovery. Remarkably, on the same campus as the hospital where she worked, there was a scientist, Colin Nichols, who is internationally renowned for his work on potassium channels — the type of protein that is overactive in people with Cantú syndrome. Neither knew of the other’s existence, but, the same day he read the Dutch papers and saw Kathy’s name as a co-author, Colin was knocking on Kathy’s office door. The pair wrote their first paper together the next year: ‘KATP Channels and Cardiovascular Disease: suddenly a syndrome’. Never mind the first five words: the last three tell the story. Colin had been working in pure science, and, out of the blue, he had a brand-new human condition to sink his teeth into. He got to work immediately.

Colin is tall, lean, blue-eyed, enthusiastic. His Northern English accent seems a little out of place in St Louis, Missouri, where he and Kathy work, but he is very much at home there, having worked at the Washington University School of Medicine in St Louis since 1991. The first time I met Colin was in Utrecht, in the Netherlands, at the first meeting of the Cantú Syndrome Interest Group, an international group of scientists and doctors who share the goal of better understanding the condition and learning how to best to treat it. We were there for a symposium and a research clinic.

At a subsequent meeting of the group, in St Louis, Colin and his trainees presented work on mice that had been genetically engineered to have a version of Cantú syndrome. The condition isn’t exactly the same in mice and humans — it’s hard to spot when a mouse has an excess of fur, for instance — but mostly it’s a pretty good match. This means you can do things such as testing possible treatments on the mice, that you might not be quite ready to try on humans. I was particularly impressed by some work on lymph vessels from the mice. People with Cantú syndrome often have a build-up of fluid in their tissues, called lymphoedema. Colin showed us pictures and videos of lymph vessels from Cantú mice, comparing them with ordinary mice. Normally, these tubes collect fluid that has leaked out from the blood into the body’s tissues, and carry it back to the bloodstream. In healthy mice, the muscle in the wall of the tubes pulsed constantly, squeezing rhythmically to force the fluid towards the heart. The Cantú mice were a different story … their lymphatic tubes sat limply, hardly moving at all. It seemed there was an obvious link with the problem of lymphoedema in humans with Cantú syndrome.

That was interesting — but what made us really sit up and pay attention was what happened when Colin’s student treated the mouse tissues with medicine. There’s a group of drugs called sulphonylureas, which work on a slightly different version of the same channel that’s affected by Cantú syndrome. Instead of being found in blood vessels and lymph vessels, this channel is important in the pancreas. Yes, we’re back to the pancreas, and diabetes. The reason that insulin isn’t the only treatment for diabetes is that there are different reasons why people can develop diabetes. If your pancreas completely fails to make insulin, for any of a number of different reasons, then it’s definitely insulin you need. When Banting and his colleagues worked out how to make a safe, reliable supply of insulin, it was this type of diabetes — called insulin-dependent or type 1 diabetes — that they were treating. In its most common form, this is an immune disease. The body’s immune system mistakes the islet cells of the pancreas for a threat and destroys them.

Non-insulin-dependent or type 2 diabetes is a different kettle of fish. It’s an insidious condition in which the body’s cells become resistant to the action of insulin, and the pancreas gradually loses the capacity to make enough insulin to compensate. Because the islet cells keep the ability to release the insulin they do make, treatments that stimulate this release can be effective. It turns out that, if you block the pancreas’s version of the channel that’s important in Cantú syndrome, extra insulin is released, lowering blood sugar. Sulphonylureas work on this principle: block the channel, allowing the pancreas to release more insulin, and that insulin lowers the patient’s blood sugar. Colin’s student used one of these drugs, called glibenclamide, to inhibit the overactive channels in Cantú mouse lymph vessels. They sprang into life, contracting vigorously as if there were no problem.

If we have a drug that we know works to block some of the effects of Cantú syndrome in mice, and that drug is already registered for use in humans, why not just go ahead and treat all of our Cantú syndrome patients? Well, for a start, we’re worried that it might be dangerous to them. Giving someone who doesn’t have diabetes a blood-sugar-lowering drug might cause them to have low blood sugar, which at worst could be a life-threatening side effect. Also, there is a long history in medicine of people trying treatments that on paper ought to have worked, but which turned out in the real world to be useless, or even dangerous, for unforeseen reasons. Fools rush in where angels fear to treat. If you’re going to try something like this, it has to be done in a careful, controlled way. And if you’re doing that, it would be far better to find a medication that only works on the Cantú channel and leaves the pancreas version alone. Colin is embarking on a search for just such a drug, but such searches can take years and have no guarantee of success.

The one situation in which you might experiment with an untried treatment like this is one in which it seems like you have no other options.

One day in late 2017, I was called by a geneticist called Alan Ma. Alan had been asked to see a baby who had been in intensive care since his birth, several weeks before. Born prematurely, Harry weighed more than expected for a baby who had only made it to 32 weeks of pregnancy. He had a patent ductus arteriosus that had needed surgery. He had a lot of hair, including hair on his forehead that merged with his eyebrows. Alan sent me some photographs, saying, ‘It looks like this boy has Cantú syndrome — what do you think?’

Alan also said that Harry had severe lung disease, and, despite all the usual treatments, he was not getting better. Worryingly, there was a report in the medical literature of an infant with Cantú syndrome who had died from similar lung disease.

It seemed that now might be the time to try glibenclamide in a human with Cantú syndrome.

Harry was in an intensive care unit, so that changes in his blood sugar, or any other — unexpected — responses to treatment could be closely monitored. The balance between risks and possible benefits seemed to favour treating. We discussed the situation with the international group, who agreed, and Alan arranged for me to meet Harry and his parents, so that I could speak with them about the idea.

Before we could try an experimental treatment on a sick baby, we had to be absolutely certain of the diagnosis. The quickest way to do this, it turned out, was to do exome sequencing. We would read the sequence of all 23,000 of Harry’s genes, even though we were interested in just two: ABCC9 and its partner KCNJ8. This was one of the first times an Australian lab had done this as an urgent test, and it was a complete success — in just four days, we had our answer. Harry had a change in his ABCC9 gene that had been seen before in other Cantú children. We had our confirmation. Even better, Colin had already studied this particular change and showed that it could respond to glibenclamide in the laboratory. Alan applied to the hospital for permission to use the drug, and, cautiously, starting with a tiny dose, began the treatment.

Around the world, the Cantú community — who had agreed this was the right thing to do — held their collective breaths.

I wish I could tell you that the treatment was a miracle cure. The truth is more complex than that. Slowly, Alan increased the dose of the drug. Slowly, Harry’s condition improved. His face and limbs had been swollen with fluid; the fluid cleared. His lungs gradually got better, too — not all at once, and not without setbacks, but eventually he was able to leave the intensive care unit, and finally he was discharged from hospital altogether. His blood sugar did drop a couple of times, but only mildly. We saw no evidence that the drug was harming him otherwise. Two years down the track, Harry’s lungs were still working fine, and he was still taking the glibenclamide.

Did the treatment save Harry? In the paper we wrote describing all of this, we said, ‘it is tempting to conclude that glibenclamide was of benefit to our patient’, which I think strikes the right note of caution. I would love to think that the treatment helped with his recovery — and that’s a problem. If you want a particular outcome, it is all too easy to fool yourself about what happened. Perhaps Harry was going to get better anyway. Maybe the drug did nothing, or actually made things slightly worse. Study a single child, and it’s almost impossible to know for certain whether you have made a difference.

You might argue that it hardly matters. We treated a sick baby, his condition improved … who cares whether the treatment made him better or he recovered on his own? The problem is that there is a particular trap that lies in wait for those who treat rare diseases. Once you use a treatment in someone who has a very rare condition, that treatment might become standard practice, even if it doesn’t actually help. It then becomes next to impossible to do the studies that might show whether it really does work or not, because nobody would be willing to have their child (or themselves) potentially given a placebo instead of the treatment that ‘everybody knows you use for (condition X)’.

That’s not to say we should never try drugs on a one-off basis, as we did here. There are things we can learn from an n=1 study (a study of a single patient). We learned that the drug was not obviously harmful to Harry, within a particular range of doses. We gained some initial observations of the problems that might be helped by the treatment (the excess fluid, the problem with his lungs) and also of others that seemed not to be changed at all (the excess hair, for instance). In the ideal world, we would do a controlled study, in which patients with Cantú syndrome were treated, randomly, with either drug or a placebo, with both the doctors and the patients (and their families) unaware of which the patients were receiving. This is the randomised controlled trial, which is the gold standard when you’re trying to find out if a new treatment is effective. But when the numbers are tiny and the need is great, sometimes we have to do the best we can for the child in our care; study the outcome; and report it in the medical literature so that others can learn from our experience, and add to it with theirs.

*

The past few years have seen an explosion of new treatments for genetic conditions, many of which are already at a very advanced stage of readiness, or have even reached the market as drugs that can be prescribed. I recently saw a French doctor, Guillaume Canaud, receive a standing ovation at a European conference,112 after giving a talk about treatment for a complex condition in which there is massive overgrowth of limbs and tissues. Canaud had found out that the drug was being developed as a treatment for cancer, because, in some cancers, the gene involved in the syndrome (PIK3CA) is overactive and contributes to the growth of the cancer cells. He arranged to obtain some from the company that was working on it, and tried it on his patients. He saw dramatic improvement in some of their symptoms. In the very same week as Canaud’s talk, a group led by an Australian, Ravi Savarirayan, published work showing that a new drug could lead to improved growth and other benefits in children with achondroplasia, the most common type of dwarfism. This kind of conjunction is almost the norm these days: practically every conference we go to includes news of new, targeted treatments for genetic conditions.

[8 This is not normal! Polite applause is usually the best a speaker can hope for. A really good talk might be received with polite applause that goes a bit longer than usual.]

Perhaps most exciting of all, gene therapy is back. In one sense, it never went away. Dedicated researchers around the world have never given up on the idea of an actual cure for many different conditions. For those of us outside the field, gene therapy has had something of the feeling of a perpetually moving goalpost, always ten years away from being reality — until suddenly, in the last couple of years, gene therapies have started to hit prime time. It’s a bit like the musician who is an overnight success — overnight, if you ignore the decades of practice and hard work that preceded their breakthrough album.

The basic principles of gene therapy haven’t changed since the 1990s, when Jesse Gelsinger died. The idea is still to get a working copy of a gene into cells that lack that gene; viruses are still used to transport the new gene to where it needs to go. That means inside the cell nucleus, although not necessarily into the DNA of the cell — in some types of gene therapy, the working copy of the gene sits inside the nucleus and is able to function without having to change the cell’s existing DNA. The decades of work it has taken to make this safe and effective have been about finding the right viruses for the job and modifying them to make them safe but still effective. The modifications stop the virus from reproducing itself, either by removing some important genes or by removing the virus’s own genome altogether, so that it’s only the shell of the virus that is used to deliver the replacement gene to the patient’s cells.

In 2017, the Food and Drug Administration in the United States approved the first ever gene therapy for a genetic condition:113 Luxturna treats a severe eye disease, caused by variants in a gene called RPE65. In 2019, the FDA approved a second: Zolgensma is used to treat a progressive neurological condition, spinal muscular atrophy (SMA). There are numerous others on the way — treatments for haemophilia, and for some immune deficiencies. It seems likely that a golden age of gene therapy is just around the corner.

[9 Gene therapy can also be used against cancer — either by directly changing the genes of the cancer cells or by modifying the patient’s immune system so that it can better fight the cancer.]

But … there’s always a but. Luxturna improves vision, but it’s not a cure. Zolgensma has had some dramatic successes, but it can’t reverse damage that has already been done. Michelle Farrar, a paediatric neurologist, and Veronica Wiley, head of the New South Wales newborn screening lab, are leading a study in which newborn babies are screened for SMA. The idea of the study is to try to make the diagnosis before any symptoms develop, and then give gene therapy before damage to the nerve cells has occurred. The hope is that very early treatment may be able to completely cure the condition. We don’t know if this will work, but the idea that it might become possible to take a baby who otherwise would have died in the first year or so of life, give them a single treatment, and cure them, so that they can live a completely normal life, is astonishing; it truly is the stuff of science fiction.

A note of caution is needed. Some of the babies who are identified in the first weeks of life already have symptoms. We don’t know if the effects of the treatment will be permanent, or if treated children might still develop symptoms later on. And even if everything works perfectly, there is bound to be a considerable delay before screening and treatment can become universally available, even in rich countries. We can imagine it might be like the early days of insulin treatment must have been, with a race to get enough of the treatment made and to get it to babies early enough that it will do the most good.

And then there’s the cost.

The first time a patient of mine received enzyme replacement therapy, for a lysosomal storage disorder called Fabry syndrome, I held the bag of fluid containing the drug in my hand while a nurse hooked up the tubing through which we were going to administer it. My patient’s mother pointed out that the contents of that small bag were worth more than her car. This was the first of what will probably be a lifetime of treatments for that boy — now a young man — treatments that are given once a fortnight. The price of a car, every other week. In the case of Zolgensma, you could substitute ‘house’ for ‘car’: the drug costs US$2.1 million for a single dose, the most expensive drug treatment ever made. At least that’s a one-off — over the course of a patient’s lifetime, enzyme replacement therapies work out to be much costlier. There are several of them, including the treatment for Hurler syndrome, that cost hundreds of thousands of dollars per year; over a lifetime, a one-off shot, even at Zolgensma prices, starts to look like a bargain. These high costs are partly because some of these drugs are just very difficult and expensive to make, and partly because the conditions they treat are so rare. If you make a successful asthma drug, it may cost a fortune to develop, but you have good hopes of selling many millions of doses. Treat a condition that affects one person in 100,000 and there’s no way to spread those costs.

All of this means that society may have to face up to some difficult choices. If you have a limited health budget, how best should you spend it? Should you give enzyme replacement therapy to a child with severe Hurler syndrome, knowing that it may improve her quality of life but will not treat her brain disease, and may not greatly lengthen her life? What happens when 20 more such treatments come onto the market?

While some of the new treatments, particularly some types of gene therapy, have the potential to be a cure, it’s likely that most will not. Many of the new treatments may simply turn out to change one problem into another, as with bone marrow transplants for Hurler syndrome. And there are many, many genetic conditions for which there is no realistic prospect of an even partially effective treatment.

It makes you wonder — is there another way?