‒ Inside modern drug development: the science, economics, and regulatory hurdles behind bringing new medicines to patients | Lloyd Klickstein, M.D., Ph.D.

Lloyd Klickstein spent two hours with Peter Attia walking through how a drug actually gets built, tested, and killed or commercialized, using bimagrumab as the running case. Klickstein is a physician-scientist who ran Novartis's New Indication Discovery Unit and now leads Costlab Therapeutics. The conversation is an insider's education, not a consumer guide to any medication.

How targets get chosen

Klickstein's team catalogued roughly 7,000 unmet clinical indications and sorted them into buckets: healthy aging, fibrotic diseases, renal diseases, and others. His personal rule was to pursue large indications even when the regulatory path was harder, because more people are helped. That calculus led him to muscle loss and obesity.

Antibodies like bimagrumab get developed when a small molecule can't do the job. Myostatin and activin bind their receptor at extremely tight (low nanomolar) affinity, and only a biologic can out-compete them. Small molecules also bind off-target in ways that produce unpredictable toxicity; antibodies are specific enough that they rarely do.

The bimagrumab story

Bimagrumab blocks the receptors that myostatin and activin use to suppress muscle growth. In older humans with muscle loss, it produced a reliable 4 to 8 percent increase in muscle mass. That sounds meaningful until you see that the same compound produced 20 to 30 percent gains in mice, and the human gains did not translate into reliable improvements on functional strength tests. Without resistance training layered on top, more muscle on a scan did not mean more capable in the body. Novartis out-licensed the drug in 2021.

What changed the story was testing it in type 2 diabetics, where it produced about a 0.7 to 0.8 percentage point absolute reduction in HbA1c (a standard measure of blood sugar control over roughly three months) alongside significant fat loss. That led to the BELIEVE trial: 507 patients, 48-week primary endpoint, 72 weeks of treatment, nine arms combining bimagrumab with semaglutide (Ozempic/Wegovy). The highest-dose combination arm produced roughly 22 to 23 percent total body weight loss and 45.7 percent fat mass loss from baseline. Klickstein calls it the first medical therapy producing fat loss comparable to bariatric surgery. One caveat in the data: LDL cholesterol rose about 20 percent. Eli Lilly acquired Versanis Bio, the spinout that ran the trial, and is now pairing bimagrumab with tirzepatide (Mounjaro/Zepbound) in follow-on studies.

Patents, exclusivity, and what drugs cost

A drug patent runs 20 years from filing, but by the time a drug reaches market, practical exclusivity is typically 10 to 15 years. AbbVie extended Humira's life by stacking dozens of patents on formulation, auto-injector, dosing, and manufacturing process on top of the composition patent. The Tufts Center for the Study of Drug Development puts total cost per approved drug at $2 to $4 billion today, a figure that includes the cost of failed programs and imputed capital costs; it's contested, but the order of magnitude is real.

Gray-market peptides

Klickstein is direct about compounds sold online with "research use only" labels. An Investigational New Drug application (IND) is FDA permission to give a compound to humans; it requires toxicology data, pharmacokinetics, and proof of manufacturing quality (GMP, Good Manufacturing Practice). None of that applies to gray-market peptides. On BPC-157 specifically: every published study comes from a single investigator, the peptide is not encoded in the human genome, no specific receptor has been identified in peer-reviewed literature, and the findings have not been independently replicated. Klickstein calls it "the poster child for the greatest grift of the entire health and wellness industry" and compares buying unregulated peptides to buying an opioid off a street corner.

mTOR and intermittent fasting in older people

Klickstein believes selective mTORC1 inhibition probably extends healthy lifespan in humans, given that the effect is conserved from yeast to dogs. He's honest that the human effect size will be modest and may never be directly proven in a trial. The more practically interesting point: in old rodents, fasting does not downregulate mTOR the way it does in young rodents. Klickstein flags this as a reason to question whether intermittent fasting produces the same metabolic benefit in older people. That is not a settled finding, but it is a legitimate scientific concern and worth watching.

One figure to discount

Klickstein cites a three-year nursing home mortality rate "approaching 90 percent" to argue that frailty deserves aggressive treatment. He immediately hedges it himself, noting he hasn't looked at the data recently. Peer-reviewed estimates for newly admitted nursing home residents generally land in the 50 to 70 percent range depending on age, diagnosis, and era. His directional point about frailty being life-threatening is sound; that specific number is likely overstated.

Peter Attia and physician-scientist Lloyd Klickstein do a deep, two-hour tour of how a drug goes from a medical need on a whiteboard to a patient's bloodstream, using bimagrumab — an antibody that blocks myostatin and activin receptors to build muscle and, unexpectedly, strip fat — as the running case study. The episode is an insider's education on drug development economics, clinical trial design, and the biology of muscle loss and obesity; it is not a consumer guide to taking any medication, but it gives readers a sophisticated framework for evaluating what drug news actually means.

What was covered

• How drug targets are chosen. Lloyd Klickstein describes building Novartis's New Indication Discovery Unit by cataloguing roughly 7,000 unmet clinical indications and grouping them into buckets — healthy aging, fibrotic diseases, renal diseases, and others — then funding projects in under-served areas. His personal rule: pursue large indications to help as many people as possible, even if the regulatory path is harder.

• Small molecules vs. biologics. Klickstein explains that small molecules (chemicals) can bind off-target, producing unpredictable toxicity; antibodies are so specific they rarely do. That specificity is also why bimagrumab required an antibody: myostatin and activin bind their receptor at low nanomolar (extremely tight) affinity, and only a biologic can out-compete them.

• Patent economics and exclusivity. A drug patent runs 20 years from filing, but by the time a drug reaches market, practical exclusivity is 10–15 years. Companies like AbbVie layered dozens of patents on Humira — composition, formulation, auto-injector, dosing, manufacturing process — to extend protection. Trade secrets (Armour Thyroid, Acthar gel) are a legal alternative but rare in modern pharma.

• The IND process, GMP manufacturing, and gray-market peptides. An IND (Investigational New Drug application) is FDA permission to give a compound to humans; it requires toxicology data, pharmacokinetics, and proof of GMP (Good Manufacturing Practice) manufacturing. Klickstein is blunt that peptides sold online as "research use only" — he singles out BPC-157 by name — lack reproducible science, known receptors, or quality controls, comparing buying them to buying an opioid on a street corner.

• Bimagrumab's clinical arc. Phase 2 in sarcopenic patients produced a reliable 4–8% increase in muscle mass in older adults (versus 20–30% in mice) but no robust improvement in functional strength tests, leading Novartis to out-license the drug in 2021. A later study in type 2 diabetics showed a ~0.7–0.8% absolute reduction in HbA1c (a standard measure of blood sugar control) alongside fat loss. The BELIEVE trial — 507 patients, 48-week primary endpoint, 72-week treatment, nine arms combining bimagrumab and semaglutide (Ozempic/Wegovy) — found up to 22–23% total body weight loss and 45.7% fat mass loss in the highest-dose combination group, comparable to bariatric surgery. An unexpected ~20% rise in LDL cholesterol was observed. Eli Lilly acquired Versanis Bio (the spinout) and is now running combination studies with tirzepatide (Mounjaro/Zepbound).

• mTOR inhibition and geroprotection. Klickstein says he believes selective mTORC1 inhibition probably extends healthy lifespan in humans given its conservation across species (yeast to dogs), but the effect size will be modest and may never be directly proven. He raises a provocative point: in old rodents, fasting does not downregulate mTOR the way it does in young rodents, which calls into question whether intermittent fasting produces the same benefit in older people.

• Cancer prevention as the next frontier. Klickstein's current company, Costlab Therapeutics, is pursuing a novel approach: activating a stress-signaling pathway (ribotoxic stress) that drugs like sorafenib accidentally suppress, causing skin cancers in ~10% of patients. His hypothesis is that gently turning this pathway on could prevent cancers. A collaboration with the Broad Institute testing the compound on 1,000 cancer cell lines found melanoma is the most sensitive tumor type.

Notable claims & predictions

• Lloyd Klickstein: "If a patient had to go to a nursing home… the three-year mortality rate approached 90%." He uses this to justify why treating frailty aggressively — like cancer — is warranted.

• Lloyd Klickstein: In the BELIEVE trial's highest-dose combination arm (bimagrumab + semaglutide), fat mass fell 45.7% from baseline. "This is the first medical therapy that gives fat loss equivalent to or superior than bariatric surgery."

• Lloyd Klickstein: Bimagrumab produces a reliable 4–8% increase in muscle mass in older humans but does not reliably improve strength or functional performance tests without resistance training — the same limitation seen with IGF-1 agonists and SARMs.

• Lloyd Klickstein on gray-market peptides: "It's no different than going and buying some opioid from a street corner drug dealer. You have no idea what's in there." On BPC-157 specifically: all published data come from a single investigator, the compound is not encoded in the human genome, has no known receptor, and has not been independently replicated — "the poster child for… the absolute greatest grift of the entire health and wellness industry."

• Lloyd Klickstein: Old rodents do not downregulate mTOR during fasting the way young rodents do. "That makes me call into question the whole concept of intermittent fasting in older people because I don't know if it'll do the same thing."

• Lloyd Klickstein on drug development cost: The Tufts Center for the Study of Drug Development puts total cost per approved drug at $2–4 billion today, up from the older "$1 billion" figure.

Fact check

Claim: Three-year nursing home mortality "approaches 90%." This figure is higher than most peer-reviewed estimates, which place 3-year mortality for newly admitted nursing home residents in the 50–70% range depending on age, diagnosis, and study era. Klickstein himself qualifies it as something he "hasn't looked at recently" and concedes "it might be a little better now." Listeners should treat this as a directionally correct but likely overstated figure.

Claim: 45.7% fat mass loss with high-dose bimagrumab plus semaglutide is "equivalent to or superior than bariatric surgery." Klickstein says this finding was published in Nature Medicine. Bariatric surgery fat-loss figures vary widely by procedure and follow-up duration, so the comparison is plausible but context-dependent. The claim is not verifiably false, but readers should note that bariatric surgery outcomes are measured over years with long-term follow-up data that a 72-week trial cannot match.

Claim: BPC-157 has no known receptor and is not encoded in the human genome. This is accurate as of current scientific consensus: BPC-157 is a synthetic peptide not found in the human proteome, and no specific receptor has been identified in peer-reviewed literature. The single-investigator publication record is also well documented.

Claim: The risk of being struck by lightning in the U.S. in a year is about 1 in 100,000. National Weather Service data put annual lightning strike risk closer to 1 in 500,000 to 1 in 1,000,000. Klickstein's figure of 1 in 100,000 overstates the risk by a factor of 5–10. This matters because he uses it as his personal threshold for exposing healthy volunteers to experimental drugs — a higher true risk would make his standard more permissive than he implies.

Claim: The Tufts figure for cost per approved drug is "$2–4 billion." Klickstein cites the Tufts Center for the Study of Drug Development accurately in name; the organization has published estimates in that range, though those figures are contested (they include imputed cost of capital and failed programs, and the Tufts center has historically received pharmaceutical industry funding). The range is real but not universally accepted as the definitive number.

No other factual claims clear the bar for a strong false/misleading call.