Peter Attia MD409 ‒ Inside modern drug development: the science, economics, and regulatory hurdles
CHAPTERS
- 0:01 – 3:36
Lloyd Klickstein’s journey: from rheumatology and NIH science to Novartis translational medicine
Peter and Lloyd set the stage with Lloyd’s background as an MD-PhD physician-scientist and his transition from academic immunology to industry drug development. Lloyd describes Novartis Institutes’ early push to formalize “translational medicine” as a bridge between biology and clinical impact.
- •MD-PhD training, rheumatology practice, and basic science on immune adhesion molecules
- •Long tenure at Brigham and Women’s Hospital before moving to industry
- •Novartis Institutes’ founding mission under Mark Fishman and Dan Vasella
- •Translational medicine as an organizing principle for R&D
- 3:36 – 5:42
How new drugs really start: unmet needs, new indications, and building regulatory paths
Lloyd explains his patient-first approach: begin with clinical need rather than molecules. He contrasts incremental drug improvements with “quantum step” innovation that may require creating new disease definitions and regulatory endpoints from scratch.
- •Drug discovery begins with identifying what medicine lacks (unmet indications)
- •Incremental vs step-change innovation and why industry biases toward lower-risk increments
- •Challenges when an indication lacks established diagnostic codes or regulatory precedent
- •Creating awareness among physicians, payers, and regulators is part of “inventing” the indication
- 5:42 – 8:27
Novartis’ New Indication Discovery Unit: systematically mapping 7,000 unmet conditions
Lloyd recounts building a structured process at Novartis to catalog unmet clinical indications and prioritize them into actionable “buckets.” This creates a portfolio approach to finding under-addressed diseases, including healthy aging and renal disease where endpoints are tough.
- •Compiling ~7,000 recognized unmet indications as a starting framework
- •Grouping indications into strategic buckets (e.g., healthy aging, renal, fibrotic diseases)
- •Importance of endpoints and regulatory feasibility when choosing targets
- •Bimagrumab ultimately emerges from this new-indication exploration
- 8:27 – 10:26
Why frailty and sarcopenia matter: mortality risk and the problem of defining the disease
The conversation turns to frailty and sarcopenia as major drivers of late-life morbidity and mortality. Lloyd explains why frailty-related nursing home placement can be worse than many cancers and why the field still struggles to operationalize sarcopenia definitions.
- •High mortality associated with frailty and institutionalization
- •Sarcopenia definitions: low muscle mass plus impaired function (grip, gait speed, stair climb, etc.)
- •Why treating/preventing frailty could create outsized societal value
- •Tradeoff: huge impact potential but harder regulatory path
- 10:26 – 25:25
Modalities and IP: small molecules vs biologics, exclusivity, patents, and trade secrets
Lloyd breaks down drug modalities (chemicals vs biologics vs devices) and how intellectual property protection actually works in pharma. They discuss patents as time-limited monopolies, exclusivity differences, and how companies layer patents—or sometimes use trade secrets.
- •Small molecules (chemicals) vs biologics (antibodies, proteins/peptides, receptors) vs devices
- •Patent basics: 20 years from filing, typically ~10–15 years of post-launch exclusivity
- •Regulatory exclusivity differences for small molecules vs biologics
- •Trade secrets in pharma (e.g., Armour Thyroid, Acthar Gel) and why they can matter
- •“Patent thickets” (Humira example): composition, formulation, device, dosing, route, etc.
- 25:25 – 36:35
Choosing sarcopenia vs orphan disease strategy—and why measuring falls derailed an early program
Peter and Lloyd explore strategic choices: target a massive but ill-defined indication (sarcopenia) or start with an orphan disease for faster approval. Lloyd then explains his initial attempt to prevent falls—and the fundamental failure: falls are hard to measure objectively and reliably.
- •Two strategies: big indication first vs orphan disease first for regulatory speed
- •Falls are multifactorial (weakness, vision, attention, proprioception, reactivity, dizziness)
- •Older adults underreport falls; clinical systems mostly capture injurious falls
- •Accelerometer pendant study in nursing home residents: poor sensitivity/specificity
- •Program halted because the endpoint (falls) couldn’t be measured robustly
- 36:35 – 48:02
Scientific rationale for bimagrumab: myostatin + activin biology and why follistatin wasn’t the answer
Lloyd lays out the muscle-size control system and why myostatin blockade alone isn’t enough in humans. Bimagrumab’s advantage is receptor-level inhibition that covers myostatin and key activins (notably activin A). They also address gray-market follistatin hype and the half-life problem.
- •Myostatin as an inhibitor of muscle growth; knockout animals show dramatic hypertrophy
- •In humans, activins contribute substantially—blocking receptor captures more biology
- •TGF-β superfamily signaling via SMADs; impact on protein synthesis/turnover pathways
- •Follistatin as an endogenous inhibitor but limited by short half-life and dosing practicality
- •Fc-fusion concept: extend half-life and simplify purification (Enbrel as historic example)
- 48:02 – 1:00:38
Engineering bimagrumab: antibody discovery, affinity constraints, screening assays, and developability
This chapter details how bimagrumab was built: why an antibody (not a small molecule) was required due to ultra-high-affinity ligand–receptor interactions. Lloyd walks through phage display, reporter assays, receptor subtype uncertainty (IIA vs IIB), and early developability filters.
- •Need for picomolar-range binding to outcompete myostatin/activin ligand affinity
- •Antibody discovery via phage display rather than mouse immunization
- •Reporter (luciferase) assay used because receptor expression was too low for standard flow cytometry detection
- •Goal: block signaling regardless of whether ligand is myostatin or activin
- •Early-stage “developability”: stability, shelf life, manufacturability, and predictable properties
- 1:00:38 – 1:05:26
From candidates to proof of concept: cross-species issues, mouse hypertrophy, and early safety thinking
Lloyd explains the transition from in vitro success to in vivo demonstration, including species cross-reactivity requirements. They discuss dramatic hypertrophy and performance gains in mice, immunogenicity complications requiring a murinized surrogate, and early considerations about doping misuse and toxicology signals.
- •Need for activity in at least one tox species; otherwise create surrogates
- •Bimagrumab immunogenic in mice but not rats → murinized version used for mouse studies
- •Rodents showed very large hypertrophy (~20–30%) and improved strength/running performance
- •Human responses later proved much smaller (4–8%) highlighting translation limits
- •Early coordination with WADA to enable anti-doping assays once abuse potential is clear
- 1:05:26 – 1:18:04
IND-enabling path: non-human primates, toxicology, DMPK, capital checkpoints, and failure economics
The discussion broadens into what it takes to reach an IND: toxicology design, exposure margins, reversibility/monitorability of risks, and manufacturing scale-up. Lloyd and Peter also explain how big-company governance and small-company fundraising differ while facing similar scientific gating decisions.
- •Non-human primates often required for antibody tox/species relevance
- •Toxicology package: clinical labs, imaging as needed, necropsy and histopathology, recovery periods
- •Key risk criteria: seriousness, reversibility, monitorability; idiosyncratic liver tox as common program-killer
- •DMPK: distribution, metabolism, pharmacokinetics to predict human exposure and duration
- •Big-company stage gates vs startup fundraising cadence; estimated ~$20–30M to IND + Phase 1 value inflection
- 1:18:04 – 1:24:50
IND requirements, GMP manufacturing, and why gray-market peptides are not real drugs
Lloyd defines the IND as permission to dose humans and emphasizes manufacturing as a core regulatory pillar. He explains GMP’s documentation, purity, sterility, and comparability requirements—and uses it to critique the risks of online “research peptides” and the BPC-157 evidence vacuum.
- •IND/CTA: regulatory permission to administer experimental drugs to humans
- •GMP ensures label accuracy, purity, sterility, and traceable documentation across supply chain steps
- •Clinical-to-commercial comparability: what’s tested must match what’s sold
- •Gray-market peptides: unknown identity/contaminants, analogous to street-drug uncertainty
- •BPC-157 critique: not encoded in genome, lacks reproducibility and known receptor/clear mechanism
- 1:24:50 – 1:50:08
First-in-human design: healthy vs patient volunteers, sentinel dosing, and Phase 1 readouts for bimagrumab
They discuss how Phase 1 populations are selected to balance interpretability with ethical risk, including Lloyd’s “lightning strike” risk heuristic. Historic disasters (TGN1412) illustrate why sentinel dosing exists. For bimagrumab, older healthy volunteers were chosen and early side effects characterized.
- •Prefer “clean” populations for signal clarity, but avoid exposing healthy people to undue risk
- •Heuristic: serious adverse risk should be below ~1 in 100,000 per year for healthy volunteers
- •TGN1412 cytokine storm case changed practice: sentinel dosing, not dosing many subjects simultaneously
- •Bimagrumab Phase 1: older healthy volunteers; assessed muscle mass/strength and biomarkers
- •On-target adverse effects: cramps/spasms, diarrhea (often first-dose), acne more in younger cohorts
- 1:50:08 – 1:57:37
Phase 2 reality check at Novartis: muscle mass increases, limited function gains, and the nutrition interaction
Lloyd summarizes Novartis’ extensive Phase 2 exploration across many indications and the core limitation: muscle size reliably rose, but functional endpoints improved only modestly. They discuss parallels with androgens/SARMs and highlight evidence that higher protein intake enhanced hypertrophy and that bimagrumab prevented diet-related lean mass loss.
- •Novartis ran ~16 Phase 2 studies across indications; consistent hypertrophy, modest function change
- •Human muscle mass gain (~4–8%) far smaller than rodent hypertrophy, affecting functional translation
- •Meta-analysis: small ~9-meter improvement in 6-minute walk; other function tests used variably
- •Nutrition study: more protein → more muscle gain; bimagrumab mitigated muscle loss under low protein
- •Missed co-development opportunity: Novartis nutrition unit sale disrupted integrated drug+nutrition strategy
- 1:57:37 – 2:10:27
Metabolic pivot: fat loss and A1C improvements, spin-out to Versanis, and the BELIEVE + semaglutide strategy
The surprising metabolic effects—fat mass reduction and improved glycemic control—reframed bimagrumab’s opportunity. Lloyd describes licensing the Phase 2–ready asset, the initial investor skepticism around obesity, and how semaglutide’s breakthrough shifted strategy toward combination therapy and a large factorial trial design.
- •Type 2 diabetes study: ~0.7–0.8% absolute A1C reduction plus fat mass decrease
- •Novartis out-licensed/spun out despite signal; Adietum/Versanis pursued sarcopenic obesity
- •Fundraising headwinds pre-semaglutide; raised ~$70M (wanted ~$100M)
- •Post-semaglutide “skate to the puck”: reposition to combination with incretin standard of care
- •BELIEVE design: open-label semaglutide logistics, placebo-controlled bimagrumab; full factorial (nine arms)
- 2:10:27 – 2:20:40
BELIEVE results, acquisition by Eli Lilly, unexpected LDL rise, and what comes next
Lloyd reports headline BELIEVE outcomes: large weight loss in the high-dose combination arm and bariatric-surgery-level fat loss, with more modest functional gains. They address an unexpected LDL increase likely due to on-target liver biology and discuss the strategic implications for obesity treatment as induction/maintenance. The chapter closes with Lilly’s acquisition and ongoing trials.
- •BELIEVE expanded to 72 weeks treatment with 48-week primary endpoint; ~507 participants; DEXA in all
- •High-high combo: ~22–23% body weight loss; ~45.7% reduction in body fat mass (surgery-like)
- •Function: grip strength changes small/variable; broader function tests considered
- •Safety signal: ~20% LDL increase—likely on-target hepatic activin receptor biology; monitorable/treatable but clinically important
- •Company acquired by Eli Lilly; ongoing combination studies listed on clinicaltrials.gov; open questions about durability after withdrawal
- 2:20:40 – 2:35:04
Beyond bimagrumab: selective mTORC1 inhibition and a new thesis for pharmacologic cancer prevention
In the closing segment, Peter and Lloyd pivot to two forward-looking areas: geroprotection via mTORC1-selective inhibition and prevention-first medicine. Lloyd explains why TORC1 selectivity is the core technical hurdle and then outlines a novel cancer-prevention approach: activating (gently) endogenous tumor-protective pathways inferred from drugs that inadvertently cause cancer.
- •mTOR: likely conserved geroprotective biology, but human testing is hard; TORC1 selectivity remains central challenge
- •Rapalogs: TORC1-selective but chronic exposure can suppress TORC2; intermittent dosing is uncertain without tissue biomarkers
- •Measurement problem: meaningful mTOR biology may require tissue (e.g., liver) sampling; fasting responses differ with age in rodents
- •Cancer prevention concept: drugs that cause cancer may inhibit protective pathways; carefully turning such pathways on could prevent cancers
- •Proposed initial test case: high-risk skin cancer cohorts; parallel evidence for melanoma sensitivity and prevention feasibility
