This follow-up book to "The Emperor of all Maladies: A Biography of Cancer" chronicles some of the greatest biomedical advancements of the past 50 years: lab-synthetized protein drugs, the Human Genome Project, and gene therapy. A common angle in all these advancements is that while they've been (so far) overwhelmingly good for civilization[5], there were also plenty of scientists and writers who cautioned us to watch out, tread carefully, go more slowly, warning of perils beyond our comprehension. And at the same time the pioneers would grumble and argue that the cautioners are just jealous and driven to mediocrity, and make it anyway. Those pioneers often found themselves outcasts of academia, in successful but ruthless new companies.
But the thing about the people who said to slow down, is they have a way of sounding compelling even in hindsight. And two of them: Paul Berg and James Watson, have found themselves on both sides of these arguments. The author, physician Siddhartha Mukherjee, was a student of Paul Berg–and he tends to side with Berg and the cautionary establishment, but at the same time acknowledge the paradox of how things have shaken out. When it comes to this question of progress, this book doesn't really have a thesis, so this review is an epilogue absent in the main text.
I'm going to focus on the medical progress angle of the book, but medicine is a side-show in Muckerjee's survey of genetics. The overall book has less technology and more philosophy. As Mukherjee writes:
"Three profoundly destabilizing scientific ideas ricochet through the twentieth century: the atom, the byte, the gene...Each represents the irreducible building block of a larger whole: the atom, of matter; the byte, of digitized information; the gene, of heredity and biological information...The atom tantalizes us with the prospect of controlling matter and energy. The gene tantalizes us with the prospect of controlling our bodies and our fates."
The rest of the book, or: What are drugs next to our demons?
What Mukherjee is most interested in is the genetics of mental conditions and eugenics. He dedicates the book
"To Priyabala Mukherjee (1906-1985), who knew the perils;
to Carrie Buck (1906-1983), who experienced them."
Priyabala was his grandmother, who raised a manic-depressive son and fiercely nursed another schizophrenic son for decades of his adulthood. Schizophrenia is often genetic, found in family trees alongside mania. Before the genetic theory gained traction, the Freudian explanation was that paranoia is a psychological response to domineering mothers, which...actually Priyabala sounds domineering in this book. Funny how two theories can explain one datapoint.
Priyabala's fourth son, the author's father, himself has a late-onset genetic illness–see why the book subtitle is "intimate history?" And Mukherjee's mother has an identical twin with a very different life. This family-history organizing arc of the book reads like the highest-quality school admissions essay, except without the "and that's why I want to study medicine" at the end.
Hereditary mental illness–that is the most opinionated heart of this sprawling, intermittently editorializing book. When he describes one scientist's effort to catalogue genetic diseases, Mukherjee writes:
"[McKusick's] fourth insight is so pivotal to this story that I have separated it from the others...He understood that gene mutations are just variations...A mutation doesn't imply disease, nor does it specify a gain or loss of function...The definition of disease rests on the specific disabilities caused by an incongruity between an individual's genetic endowment and his or her current environment–between a mutation, the circumstances of a person's existence, and his or her goals for survival or success. It is not mutation that ultimately causes disease, but mismatch.
The mismatch can be severe and debilitating. A child with the fiercest variation of autism who spends his days rocking monotonously in a corner, or scratching his skin into ulcers, possesses an unfortunate genetic endowment that is mismatched to nearly any environment or any goals. But another child with a different variant of autism may be functional in most situations, and possibly hyperfunctional in some (a chess game, say, or a memory contest). His illness is situational; it lies in the incongruity of his specific genotype and his specific circumstances.”
I'll irresponsibly take this idea further. Perhaps Mukherjee himself is hyperfunctional from mild variants of his uncles' diseases. How does he write 500-page books filled with creative analogies on top of his many job responsibilities: as a professor and physician and founder of several companies? Book-tour interviewers have asked him, and this nephew of a manic man says he writes books in fits, where "once I'm in the zone, I can pull out 40 pages a day[6]."
Mildly-variant or not, Mukherjee is as protective as his grandmother was towards those with mental illness. The peril of thinking otherwise is draconian eugenics, and ending up like Carrie Buck. Carrie Buck didn't have a Priyabala to protect her. She was the diagnosed "Moron, middle-grade" daughter of a "low-grade moron." In 1924, a judge sent her to the Virginia State Colony for Epileptics and Feebleminded. The superintendent of the colony set Carrie up as the poster child for hereditary mental disability, and the Supreme Court agreed he could sterilize her, as "three generations of imbeciles is enough." Wait, three generations? Turns out Carrie had a daughter by rape before the colony sterilized her. The daughter, despite also being classified an imbecile, made honor roll in school. Mukherjee purposefully makes these diagnoses sound like an inexact science at best.
Promotional photo to cast the Bucks as hereditary imbeciles. Do they look to you like they should have the right to children? Eight court justices thought not.
Those are the ethical perils. On to the promised biomedical pioneers.
Science Race 1: recombinant DNA, Genentech, and insulin
Several medical conditions occur because the body doesn't make enough of a particular protein. The cures for these conditions seem to progress through three steps.
1) Identify which protein is needed.
Some of the steps toward these discoveries make me flinch and think "the ends justify the means," as if I were a monk furiously muttering his mantra during a crisis of doubt. Scientists discovered the role of insulin in the 1800s by taking the pancreas out of a living dog, watching with mild curiosity as it peed itself to death, and then days later noticing the flies swarming around puddles of what turned out to be the sugariest pee they'd ever seen. They realized the dog couldn't digest the sugar without the pancreas.
2) Distill the protein from a pile of livestock organs or blood donors.
This is generally an expensive, inefficient process. On top of the cost, it can be unsanitary. The medicine was deadly to many hemophiliacs getting blot-clotting proteins from blood donors during the first years of the AIDS epidemic. But this was the only way anybody knew how to get patients the proteins they needed, until some fundamental science from the 60s and 70s suggested there could be a better way, leading to:
3) Trick a fast-growing bacteria[7] into mass-producing the protein in a lab.
To accomplish this, you have to change the bacteria's DNA. You cut open the bacteria DNA, add in the DNA sequence corresponding to the chemical makeup of the protein, and stitch the bacteria together again. Then when the bacteria divides and multiplies, it will also multiply the DNA code to make the protein with it. My friend says that doing this is easy today–it’s just a matter of mixing some beakers together and leaving the mixture out on a shelf while the bacteria multiply. Paul Berg invented this technique in 1968 and called it "recombinant DNA" because the technique combines the DNA from two species.
But it took ten years after Berg's discovery before anybody used the recombinant DNA to make medicines. A startup company, Genentech, was the first to grow insulin and blot-clotting proteins in a lab, and Genentech's lab-synthesized medicines were cheaper and safer than the older ones. Why did it take so long, and why was a company the first to do it? At least partly it was because Paul Berg decided that recombinant DNA could be dangerous, and got the NIH to restrict any university lab that studied it.
I'm making Paul Berg sound like an anti-libertarian villain here, at worst a hoarder jealously guarding his discoveries lest others overtake him. But he's not a villain in Mukherjee's telling (remember that Mukherjee worked in Berg's lab). He comes across as smart and reasonable.
Berg's be-careful crusade began when his star student announced at a conference her plans to combine a hamster-cancer-causing virus with E-coli DNA. In response, a worried professor wrote to Berg that this sounded dangerous: "You can stop splitting the atom; you can stop visiting the moon; you can stop using aerosol...But you cannot recall a new form of life. [The new genetic hybrids] will survive you and your children and your children's children." Why did Berg's student choose a cancer-causing virus instead of insulin? Had it even occurred to any scientists to use gene-cloning to make medicines, or was the VC founder of Genentech actually the first to propose it? Certainly the hamster virus (a much smaller DNA sequence than insulin) was less technologically challenging to make. Mukherjee is vague on how scientists transitioned from playing with viruses to fathoming the medical applications of recombinant DNA. That vagueness makes Berg's next move more understandable.
Berg deliberated and eventually redirected his student to other projects. Reflecting on his thought process many years later, he said "I knew the risk was little, but I could not convince myself that there would be no risk...I must have realized that I'd been wrong many, many times in predicting the outcomes of an experiment, and if I was wrong about the outcome of the risk, then the consequences were not something that I would want to live with." But Berg didn't just [temporarily] forbid the work in his own lab. He organized a highly-publicized conference where geneticists would decide how to police themselves when using the new technology. The conference, Asilomar II, was as vicious as you'd expect. Out of it came a document ranking the biohazards of different genetically-altered organisms and different containment rules for each ranking. Any lab receiving NIH grants had to abide by these rules.
Compared to the NIH-funded university labs, a privately-funded startup had more leeway around the Asilomar rules. That's how the struggling venture capitalist Swanson was able to recruit two professors into Genentech. Swanson even used the Asilomar handbook to get their phone numbers. The professors he recruited, Boyer and Cohen, were catching up with Berg's lab, and were deeply unhappy with the new rules. They had a higher risk tolerance than Berg, although they weren't infinitely cavalier. For instance, Boyer had a son who could benefit from gene-cloned human growth hormone, but he wasn't audacious enough to make it himself and feed his son the concoction–yet. Through Genentech, he would eventually synthesize the growth hormone, although there were many other hormones to figure out how to synthesize first.
It took Genentech almost three years to clone insulin. Five years later, they were also the first to clone the blot-clotting gene. Both times they were racing against academics to be the first. Genentech won the insulin race because the university group competitors had to follow the Asilomar rules, slowed down in high-security biohazard containment areas.
—————
I like the insulin story, because of its ties to recent politics. Bernie Sanders became an Independent politician the year Genentech first cloned insulin; 40 years later he'd be telling campaign stories about how insulin is so expensive in America that some diabetics are rationing it. Considering Genentech did win a narrow race to clone insulin, they do seem to be replaceable billionaires. In a world without Asilomar, how expensive is 21st century insulin in America?
Drawing upon other recent politics, perhaps the Asilomar rules weren't strict enough. Many scientists have wondered whether the NIH-funded lab in Wuhan leaked the COVID-19 pathogen.
In fairness to Berg, while recombinant DNA has been an incredible gift, its potential remains scary. I want the technology to be simple enough that somebody could make recombinant insulin without sophisticated supply chains, but not so simple that anybody could synthesize a new, deadly lifeform.
Is Berg a hero or villain? Based on the morals of time-travel stories, I should be at least as hesitant to go back in time and prevent Asilomar from happening as Berg was hesitant to allow his uncertain future to arrive unhindered. One of the first glorious steps in Berg's uncertain future was:
Science Race 2: Human Genome Project
The next goal of genetics was to sequence the whole human genome. This would tell us what human DNA sequence is normal and healthy, and from comparison against that baseline, determine who has mutations. Everybody wanted it, even at a prospective price tag of $1 per letter in the sequence. But geneticists disagreed about how to accomplish it.
On the one side was the Human Genome Project, a publicly-funded effort that wanted to sequence DNA the most methodical way, letter-by-letter in order. On the other side was Craig Venter, a disaffected NIH scientist who started a company to apply the so-called "shotgun sequencing" method instead. To fund the company, Venter promised he would patent the most important-looking bits of the genome, the bits that could be evaluated for future blockbuster medicines.
Venter's approach horrified many scientists, including James Watson, the Nobel-winning co-discoverer of the double-helix structure of DNA. Watson hated Paul Berg's cautionary approach to recombinant DNA, but now he hated Venter's cavalier approach to DNA sequencing. Maybe Watson was consistent, in feeling that research should be unrestricted, and that the outputs of research should also be unrestricted (rather than patented). But the other criticism of Venter was that shotgun sequencing would miss much of the genome. Shotgun sequencing meant splitting the DNA into bits, sequencing all the little bits, and trying to arrange them together again. "What if you found only the letters p..un..y out of 'profundity?' You'd get the exact opposite meaning."
In the end, Venter finished and published his shotgun sequence the same time the Human Genome Project published their more-expensive, more-thorough letter-by-letter sequence. The Human Genome Project alleged that Venter relied on their more thorough sequence to sort his shotgun sequence. Although in Venter's defense, shotgun sequencing is more accepted today. So Science Race 2 was a tie between the more cavalier private participant and the more careful, more expensively-funded public participant.
Science Race 3: gene therapy
If recombinant DNA enables us to tweak virus DNA to make human proteins, and the Human Genome Project shows us the normal, healthy human DNA sequence, then gene therapy puts those pieces together. We can diagnose people for gene mutations, and give them a transfusion of a recombinant virus altered with the corrected human gene.
The previous two "science races" were literal races to the same finish line between a private startup and a publicly-funded institution. The private company won the recombinant-insulin race, and the human genome sequence was a tie. The gene therapy example in this book was not a public-private race, but the participants hurried as if it was.
In 1999, a private company won FDA approval to run human trials of its gene therapy medicine. Their medicine treated people who were missing the protein that allowed them to safely eat protein-rich food (burgers, peanuts, milk...). The medicine was a common-cold virus altered with the human DNA sequence that made the protein-digesting protein. The teenager Jesse Gelsinger enthusiastically signed up for the 18th spot on the trial, and died days after the gene transfusion.
Jesse's autopsy concluded that he died from a hyperactive immune response to the gene-carrying virus. Because the virus was a common cold, Jesse was likely already exposed, and had high levels of antibodies to the virus from before his transfusion. But nobody tested Jesse's blood for antibodies before his treatment, and nobody told the Gelsingers that one of the monkeys who received the earlier-phase, higher-dosage trials died. Jesse's father was furious, and the FDA shut down all human gene therapy trials for the next decade.
If we're keeping score, the private company lost this third science race. In all three races, the private company was the more cavalier participant, and the U.S. government the one who wanted to progress more carefully. And the final score: 1 win, 1 tie, 1 loss.
The only subsequent gene therapy example in this book was for hemophiliacs, and they already had a medicine available from Science Race 1. So I should back off my earlier claim that the outcome of these races was overwhelmingly positive. But most gene therapy progress is newer than this book–there are hopes for gene therapy replacing chemotherapy for cancer treatment. Mukherjee himself is involved in one of these gene-therapy companies.
Two kinds of cautiousness
Mukherjee lives for medical progress, yet sympathizes with those who wanted to proceed carefully. He is angrier about the cavalier approach to the gene therapy study that killed Jesse Gelsinger than about the fallout, the next decade of quashed trials. And he notes how Paul Berg's hesitancy about recombinant DNA did prevent some early excited scientists from running dangerous-sounding experiments.
Mukherjee takes this side partly because he doesn't believe that these moratoriums truly paused scientific progress. During both moratoriums, scientists refined their techniques, making discoveries that enabled later projects to proceed more swiftly and reliably. And we still got synthetic insulin, and have since resumed gene therapy trials. In Mukherjee's telling, science under moratorium is a camped army that still sends out scouts, the better to avoid ambush when they resume their march.
But I suspect Mukherjee is more generally disposed to take the cautious side. He writes sentences like “Fortunately, technical barriers intervened before the ethical mayhem had a chance to become unmoored.” Circling back to the theme of mental illness, Mukherjee says he doesn't want to know what genetic diseases are in store for his own daughters:
"In the wake of the monumental studies on the genetics of familial schizophrenia, I have often wondered about sequencing my [and my family's] genome. The technology exists: my own lab is equipped to do it…[but] if the history of the last century taught us the dangers of empowering governments to determine genetic "fitness", then the question that confronts our current era is what happens when this power devolves to the individual."
There is a Great and Secret Knowledge hidden somewhere in the World that will grant us enormous powers once we have discovered it.[8] Do you strain everything to hunt for it, or are you afraid of what you may find? Do you fear only for your own feeble sanity, or for what others may do, what casualties all may suffer from their search? Mukherjee fears for both, though the hunt consumes him.
In closing
Before this reread of The Gene, I felt that Mukherjee was the best writer on the vast world of modern medicine. That in 500 pages he could advance my understanding of the scope of the field from a high-school biology level to a VC or division chief's level, and with a more engaging storyline than a college textbook. That if I reread carefully, I could appreciate the reasoning that led scientists to each discovery. But several of the explanations remain fuzzy no matter how much I stare at Mukherjee's words, and the topics still seem to jump around randomly. At several points Mukherjee offhandedly mentions things as if they are well-worn discussions, but does not elaborate. Maybe my golden memory came from Berg's blurb on the book jacket: "'The Gene' is a magnificent synthesis of the science of life." Once again Berg plays the villain misleading us.
But it's hard to efficiently survey the life's work of so many brilliant people, which is why I neglected to review sections of Mukherjee's book. I skipped the bits about Darwin and Mendel, Watson and Crick and Franklin, and the brief ending about stem cells, CRISPR, and designer babies. Also, knowing how Bayesian priors determine how generously we listen to people, I've omitted his saucier coverage of race, sex, and epigenetics. And Nazis. If you, too, want a heftier feel for the ethics and biotech surrounding genetics than comes from essays, I still recommend this book as a relatively efficient and enjoyable journey into the vast fog.