409 ‒ Inside modern drug development: the science, economics, and regulatory hurdles

Episode Overview

• Physician-scientist Lloyd Klickstein, Ph.D., M.D., joins host Peter Attia, M.D., to trace how a novel therapeutic drug moves from an unmet clinical need and basic pathway biology through discovery, preclinical testing, regulatory hurdles, and human clinical trials.
• The narrative centers on the 20-year development of bimagrumab (BYM338), an activin receptor antagonist originally developed at Novartis to treat age-related frailty and muscle wasting, which later pivoted into a metabolic and obesity therapy at Versanis Bio before being acquired by Eli Lilly.
• Along the way, the discussion demystifies pharmaceutical economics, patent law versus trade secrets, the compounding math of clinical trial risk, and why objective clinical endpoints (such as measuring falls in the elderly) can make or break an entire drug program.
• The episode concludes with critical evaluations of Good Manufacturing Process (GMP) standards versus grey-market peptides like BPC-157, age-dependent changes in mTOR regulation during fasting, and a novel cancer-prevention strategy utilizing the ribotoxic stress pathway.

Key Concepts

• Speaker Backgrounds and Disclosed Business Ties: Peter Attia, M.D., is the host of The Drive podcast and a physician with no specific sponsors or business ties disclosed. Lloyd Klickstein, Ph.D., M.D., practiced as a rheumatologist for ~10 years and directed an NIH-funded basic science lab at Brigham and Women’s Hospital before joining the Novartis Institutes >20 years ago, where he co-founded the New Indication Discovery Unit and worked on the clinical development team for bimagrumab (BYM338). Klickstein later served as an advisor to Aditum Bio; was the founding CEO, CMO, President, CSO, and board member of Versanis Bio (which licensed bimagrumab from Novartis in Feb 2021, raised $70M—though they wanted $100M—from Atlas Venture and Medicxi plus a Series B after pitching 53 investors, and was acquired by Eli Lilly); previously worked at resTORbio on mTORC1 inhibitors; currently works on Khosla Therapeutics (collaborating with the Broad Institute on a cancer prevention/melanoma compound); and advises another unnamed company.
• Unmet Clinical Needs and the Mortality of Frailty: At Novartis under Mark Fishman, Klickstein's New Indication Discovery Unit compiled a list of ~7,000 unmet clinical indications (spanning 10 to 20 pages of paper) and organized actionable ones into 7 or 8 buckets, including healthy aging, ENT, renal diseases, and fibrotic/liver diseases. Addressing age-related frailty was a high priority because, at the time Klickstein was practicing and researching frailty, if a frail elderly patient had to go to a nursing home, the 3-year mortality rate approached 90%—worse than most cancers (though Klickstein hedges that it might be a little better now, it is still not good).
• The Scale, Economics, and Compounding Risk of Drug Development: Developing a widely used new drug requires thousands of people, takes many years, and costs >$1 billion—comparable in time, cost, and labor to building the world's largest skyscraper, the Burj Khalifa. According to the Tufts Center for the Study of Drug Development, the total cost per approved drug today is $2, $3, or $4 billion (up from ~$1 billion and 10 years two decades ago). Furthermore, risk compounds across development: if a program involves 20 separate risks and each carries a 90% probability of success, multiplying them together yields a 0% chance of overall success. While failing in Phase 3 is the second-worst outcome in drug development, the absolute worst outcome is succeeding in Phase 3 and failing commercially.
• Patents vs. Trade Secrets and Evergreening: A patent grants a 20-year monopoly from the filing date in exchange for full public disclosure of how to make the drug, which typically yields 10 to 15 years of practical market exclusivity post-launch before the drug becomes free for anyone to manufacture forever. By contrast, older biologics like Armour Thyroid (desiccated thyroid hormone) and Acthar Gel (ACTH purified from pig pituitaries) were protected by trade secrets. When Acthar Gel was pulled during the bovine spongiform encephalopathy (BSE) scare, a small company bought it, re-commercialized it for infantile seizures, and jacked up the price 100- or 1,000-fold, pricing rheumatologists out of using it for acute gout. In the modern patent system, companies sometimes abuse "evergreening"—for example, AbbVie patented every detail around Humira (drug, formulation, salts, auto-injector, indication, dose, route of administration, and manufacturing) staggered over 5- to 10-year intervals to extend patent life.
• The Endpoint Problem in Measuring Elderly Falls: Even with a clear clinical need, a drug program cannot proceed without an objective, reliable clinical endpoint. While Attia cites a previous podcast guest's research showing great toe force generation relative to body weight is a major predictor of falls, Klickstein hedges that "I would argue that they don't really know that," pointing out ~11 different causes of falls, including weakness, dizziness, vision, attention, and proprioception. To measure falls objectively, Klickstein, geriatrician Lou Lipsitz, and Kiran Dole conducted a 6-month published study using a pendant triaxial accelerometer on 60 elderly nursing home residents (ages 80s to >100) who had fallen at least once in the prior 6 months. Out of 117 actual fall events (defined as staff finding a resident on the floor), the device detected only 17% of real falls, and only 17% of device-reported falls were actual falls, forcing Novartis to shut down its fall-prevention drug program.
• Myostatin, Activin, Follistatin, and Fc Fusion Biology: Discovered by Se-Jin Lee, myostatin inhibits muscle growth; blocking it in mice, dogs, and Belgian Blue cattle produces massive "Arnold Schwarzenegger" muscle hypertrophy. In humans, muscle size is regulated by both myostatin and activins (mostly activin A) signaling through TGF-beta superfamily receptors (Activin Type IIA and IIB) and SMAD transcription factors ("similar to mothers against decapentaplegic"), which regulate muscle protein turnover proteins MuRF1 and MAFbx (atrogin), discovered by David Glass and colleagues. Although follistatin is an endogenous inhibitor of this pathway used in grey-market gene therapy and bodybuilding, it has a very short half-life, and Attia notes that systemic dosing would require several doses a day costing >$1 million/month. Another biologic strategy, the Fc fusion protein (developed by Brian Seed at Mass General and licensed by Immunex to create etanercept/Enbrel), attaches a soluble receptor to an antibody's Fc region to extend blood half-life and allow easy purification on Protein G/A columns; notably, Enbrel was first tested in sepsis, where it made patients worse, before succeeding in rheumatoid arthritis.
• Discovery and Preclinical Development of Bimagrumab (BYM338): Led by Chris Lu and David Glass at Novartis in collaboration with MorphoSys (using phage display to screen thousands of antibodies), the team chose to target the activin receptor rather than circulating ligands. Because myostatin and activins bind the receptor at low nanomolar / high picomolar concentrations, the therapeutic needed low picomolar affinity—achievable with a biologic antibody rather than a small molecule. Because Activin Type II receptors are expressed at levels too low for flow cytometry, the team engineered cells with a firefly luciferase reporter gene that glowed when myostatin/activin bound, screening for antibodies that stopped the glowing. This yielded bimagrumab, which is Type IIB preferential but also hits Type IIA, working across all muscle fiber types. Identifying 2 to 10 candidate antibodies took years. Because bimagrumab was immunogenic in mice (though not rats), mouse studies required a "murinized" construct (CDD866), which increased rodent muscle mass by 20% to 30% (more than myostatin knockouts) alongside increases in strength and running speed. In weekly rat and non-human primate toxicology studies, rats exhibited cardiac hypertrophy, but this normalized when adjusted for their massive body and muscle size increase and reversed upon drug withdrawal.
• Preclinical Attrition, Costs, and Phase 1 Safety Standards: About 30% of candidate therapeutic antibodies advance from discovery into human testing because antibodies rarely have off-target toxicity, and funding IND-enabling studies plus Phase 1 for a biologic like bimagrumab costs ~$20 to $30 million in today's dollars. When transitioning to Phase 1 human trials, serious adverse events in healthy volunteers occur about once every 10 years across the industry—such as the TeGenero incident >20 years ago (where an agonistic anti-CD28 antibody dosed in 6 people at once caused acute cytokine release syndrome, leading to sentinel dosing rules) and the Bial small-molecule disaster. Phase 1 studies of bimagrumab tested single and ~3 doses up to 50–100 mg/kg in older healthy volunteers, revealing monitorable, reversible on-target adverse effects: muscle spasms/cramps, acne (more common in younger people), and GI symptoms/diarrhea (mostly first-dose related).
• Human Muscle Mass vs. Function and Metabolic Pivot: Across ~16 Phase 1 and Phase 2 studies at Novartis, bimagrumab reliably increased human muscle mass by 4% to 8% (8% max), but unlike in rodents, this hypertrophy did not automatically translate into major strength or performance gains without resistance training; a meta-analysis of sarcopenia studies showed only a 9-meter increase in 6-minute walk distance. However, in a nutrition study across 3 levels of protein-calorie intake (0.5x RDA, 1x RDA, and 1.5x or 2x RDA / up to ~1.2 g/kg), eating more protein resulted in more muscle built, and bimagrumab completely prevented muscle loss in the 0.5x RDA group. Crucially, Novartis's final study—dosing 10 mg/kg monthly for 48 weeks (12 doses) in Type 2 diabetics—produced expected muscle mass gain alongside substantial fat mass loss and a 0.7% to 0.8% absolute decrease in HbA1c.
• Versanis Bio and the Phase 2b BELIEVE Trial: Following the metabolic signal in diabetics, Versanis Bio licensed bimagrumab and ran mouse studies showing additive weight/fat loss and lean mass preservation when bimagrumab was combined with semaglutide, tirzepatide, or liraglutide. They then launched the BELIEVE trial (published in Nature Medicine), a 9-arm full factorial design (low/high dose open-label semaglutide, low/high dose IV placebo-controlled bimagrumab, 4 combination arms, and placebo) in 507 overweight/obese patients (including metabolic syndrome and pre-diabetics) treated for 72 weeks (with a primary endpoint at 48 weeks and a 6-month off-drug follow-up for 104 weeks total, with follow-up results scheduled for presentation at EASD in September). At 72 weeks, the high-dose combination group lost 22% to 23% of starting body weight and 45.7% of starting body fat—equivalent or superior to bariatric surgery—while preserving/building lean mass, though grip strength showed only a small, variable improvement and patients experienced an unexpected on-target ~20% increase in LDL cholesterol due to activin receptors in the liver. Eli Lilly subsequently acquired Versanis Bio and is running a combination study with tirzepatide.
• Selective mTOR Inhibition and Cancer Prevention via Ribotoxic Stress: Regarding longevity pathways, Klickstein believes mTORC1 inhibition will probably be geroprotective in humans with a modest effect size, though rapalogs downregulate mTORC2 with sustained exposure (which resTORbio managed by combining a catalytic and allosteric inhibitor). Counterintuitively, while fasting downregulates liver mTOR in young rodents as expected, fasting does not downregulate mTOR in old rodents. In his current work at Khosla Therapeutics, Klickstein investigated why multi-kinase inhibitors like sorafenib (used for renal cell and hepatocellular carcinoma) cause basal and squamous cell skin cancers in ~10% of older patients: they inhibit a sensing kinase that triggers ribotoxic stress (a cell-death pathway also targeted irreversibly by diphtheria toxin, sarcin, and ricin). By developing a tool compound to gently and controllably turn on this ribotoxic stress pathway, Klickstein aims to prevent ~50% of cancers, and screening the compound against 1,000 cancer cell lines at the Broad Institute revealed that melanomas are extraordinarily sensitive to this mechanism.

Quotes

• At 0:08:45 - "If a patient had to go to a nursing home—not wanted to go or whatever, but had to go—the three-year mortality rate approached 90%. So being a frail elderly person who has to go to a nursing home was worse than most cancers." - Klickstein explaining why age-related frailty and sarcopenia were prioritized as major unmet clinical needs at Novartis.
• At 0:13:51 - "Making a new drug that's used by many people takes longer, is more expensive, and takes as many people as building the biggest skyscraper in the world. Think, you know, the Burj Khalifa." - Klickstein illustrating the massive organizational, financial, and human scale required to bring a therapeutic to market.
• At 0:14:46 - "At the end of the drug's patent life, it is freely available for anybody to make for the rest of eternity. That's the deal you make with getting a patent. A patent is essentially a monopoly on being able to make, use, and sell the drug in exchange for telling everybody how to do it." - Klickstein detailing the fundamental social contract of pharmaceutical patent law.
• At 0:17:27 - "It was a rheumatology smart missile. A patient could come in with acute gout and be miserable; you give them one shot of that and it gives them endogenous steroid taper over a period of several days. It was great." - Klickstein describing the clinical utility of Acthar Gel before its trade-secret status enabled a 100- to 1,000-fold price hike.
• At 0:24:31 - "If you have, say, 20 risks you're taking in a drug development program, whether it's target and format and bioavailability, everything, and all of them have maybe a 90% chance of success, you multiply all that out 10 or 20 times, it's zero." - Klickstein demonstrating how compounding probabilities across multiple steps destroy most drug discovery programs.
• At 0:25:10 - "The worst outcome in drug development is failing in Phase 3. Actually, that's probably not true. The worst outcome is succeeding in Phase 3 and failing commercially." - Klickstein highlighting the financial and strategic realities of late-stage clinical development.
• At 0:33:46 - "However, the device was awful, in that I think it had about—it detected 17% of the real falls, and only 17% of the device actuations where it said somebody fell were falls. So it was just not useful." - Klickstein recounting the results of his 6-month triaxial accelerometer study in 60 high-risk nursing home residents.
• At 0:35:06 - "So we had to stop our whole program for making drugs to prevent falls because we couldn't measure them." - Klickstein explaining how the lack of an objective clinical endpoint halted Novartis's fall-prevention program.
• At 0:39:31 - "In humans, it turns out that it's more complex. It's myostatin plus activins, mostly activin A." - Klickstein clarifying why blocking myostatin alone is insufficient for maximum muscle hypertrophy in humans compared to rodents.
• At 1:03:22 - "It doesn't work as well in humans. Because just skipping all the way ahead and we'll come back, humans get about 4 to 8% increase in muscle mass." - Klickstein contrasting the 20% to 30% muscle mass gains seen in bimagrumab-treated mice with the 4% to 8% gains seen in human trials.
• At 1:04:35 - "The first thing that any company does when they're working on a drug that has the potential for abuse is we work with the—with WADA, the World Anti-Doping Agency, to make sure that they can screen for these things." - Klickstein explaining how pharma companies proactively prevent performance-enhancing drug abuse in athletics as soon as a molecule works in mice.
• At 1:07:41 - "The two big things we look for are whether it's monitorable, whether it's reversible. So if we have irreversible cardiac toxicity with a therapeutic, that's usually the end, for example, or neurologic." - Klickstein outlining the core toxicology criteria required for a drug to advance into human testing.
• At 1:23:02 - "I think it's a big mistake to buy these peptides from, you know, from fly-by-night manufacturers. Conceptually, it's no different than going and—than a drug user going and buying some opioid from a street-corner drug dealer. You have no idea what's in that." - Klickstein warning against purchasing non-GMP "research-grade" peptides online.
• At 1:23:56 - "All of the data for that peptide come from one investigator who's the only person published on it. And remember, the fundamental tenet of science is if it's real, it's reproducible. This has not been reproduced. What the hell is it? It's not encoded in the human genome, so it's not a human peptide." - Klickstein dismantling the scientific validity of the popular grey-market peptide BPC-157.
• At 1:24:45 - "Despite how much grift there is in the health and wellness industry, I'm putting BPC-157 on the podium, at least." - Attia concurring with Klickstein's assessment of BPC-157's lack of reproducible human evidence.
• At 1:30:03 - "I don't want to expose people to a risk greater than that of a lightning strike in a year... So the risk of being struck by lightning in the US in a year is about 1 in 100,000." - Klickstein sharing Bob Schmouder's ethical rule of thumb for acceptable serious adverse event risk in Phase 1 healthy volunteer trials.
• At 1:53:31 - "The other very cool thing we found is that, as you would expect, if you have half the recommended daily amount of protein-calorie nutrients, you lost muscle mass. But bimagrumab prevented that." - Klickstein discussing how activin receptor blockade protects lean mass even during severe protein-calorie deficit.
• At 1:55:59 - "The SARMs were extensively studied, is that you can make muscles larger; they don't get stronger in the absence of resistance training. So it's not unique to the activin receptor antagonist pathway." - Klickstein explaining that pharmacological muscle hypertrophy across multiple drug classes requires mechanical training to yield functional strength.
• At 2:13:28 - "In the high-dose combination group, the body weight loss at 72 weeks was 22, 23% of starting body weight... But the fat loss was 45.7% of starting body fat. Now, that's what you get with bariatric surgery." - Klickstein sharing the 72-week results of combining bimagrumab and semaglutide in the BELIEVE study.
• At 2:24:36 - "In young rodents, with fasting, they downregulate mTOR, as you would expect. In old rodents, they didn't. So it 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." - Klickstein revealing a counterintuitive preclinical finding that challenges assumptions about fasting and mTOR in older populations.

Takeaways

• Avoid purchasing "research-grade" peptides (such as retatrutide or BPC-157) from online grey-market vendors; operating outside Good Manufacturing Process (GMP) regulations—which ensure purity, activity, sterility, and documentation across chemical synthesis, formulation, packaging, and distribution—carries risks comparable to buying street opioids.
• Apply rigorous scientific skepticism to wellness trends like BPC-157 by checking whether a compound is encoded in the human genome, has a known receptor, and possesses findings reproduced beyond a single investigator.
• Do not rely on pharmacological muscle-builders alone for functional performance; whether using activin receptor antagonists, SARMs, or IGF-1 agonists, pair interventions with resistance training to translate muscle hypertrophy into actual strength.
• Optimize dietary protein intake (up to 1.5x or 2x the RDA, or ~1.2 g/kg as tested in bimagrumab trials) when trying to build muscle, as higher protein-calorie nutrition directly amplifies hypertrophy even when using potent anabolic biologics.
• When evaluating or designing clinical trials, validate that your primary endpoint can be measured objectively and accurately before investing in drug discovery, avoiding pitfalls like the pendant accelerometer that missed 83% of real falls.
• For biopharma leaders and clinical trialists, enforce the "lightning strike standard" (a <1 in 100,000 risk of a serious adverse event) and utilize sentinel dosing when administering novel mechanisms to healthy Phase 1 volunteers.
• In preclinical toxicology screening, prioritize identifying whether adverse effects are both monitorable and reversible (such as physiological cardiac hypertrophy in rodents that normalizes to body weight and reverses off-drug) rather than terminating promising compounds prematurely.
• Proactively engage with regulatory bodies like the World Anti-Doping Agency (WADA) during the preclinical phase if developing therapeutics with high potential for athletic abuse, ensuring screening assays exist years before commercialization.
• When utilizing or developing incretin mimetics (like semaglutide or tirzepatide) for obesity, look to combination therapies that block the activin/myostatin pathway to preserve lean muscle mass and deepen fat loss toward bariatric-surgery levels (45.7% fat loss).
• Monitor lipid panels—specifically LDL cholesterol—when inhibiting activin receptors, as blocking these receptors in the liver can trigger an on-target ~20% increase in LDL cholesterol.
• Reconsider the assumption that intermittent fasting downregulates mTOR in older adults the same way it does in younger individuals, given preclinical evidence that old rodents fail to downregulate liver mTOR during fasting.
• In oncology trial design, leverage high-risk patient populations—such as adults aged 50+ with ≥5 prior skin cancers who carry a 50% annual recurrence risk—to test preventative therapeutics quickly and cost-effectively in small cohorts of 100 to 120 patients over 1 year.