#148 - Richard Miller, M.D., Ph.D.: The gold standard for testing longevity drugs: the ITP

KEYPOINTS

• 01

  The Hayflick limit sent aging research down a side road

  Normal human cells in culture divide about 50 times and then stop. Miller says many excellent cell biologists spent their careers on this 'Hayflick system' believing they were studying aging. He thinks it has little to do with why organisms age.

• 02

  The key question is whether aging can be slowed, and it can

  Before the 1990s many believed aging 'is what it is.' Cynthia Kenyon's 1993 daf-2 work in C. elegans showed that single genes (daf-2 resembles the IGF receptor, daf-16 resembles FOXO) can greatly extend lifespan. That came the same year a Nature paper argued no single gene mutation could do so. Miller argues that aging is at least partly a single process a drug might slow, not a 'spaghetti bowl' of 30–50 separate ones. He once had a friendly fight about this with George Martin.

• 03

  The ITP is built for statistical power and generality

  The Interventions Testing Program runs each drug at three sites: Jackson Labs, Miller's lab in Michigan and Randy Strong's lab at UT San Antonio. Each year each site has 50 males and 50 females per drug and 100 controls of each sex. That's enough to detect a roughly 8–10% lifespan change with 80% power. Anyone can suggest a compound, and a steering committee, advised by five outside scientists, chooses.

• 04

  Genetically heterogeneous mice are the point

  About 90% of mouse research uses a single inbred genotype, usually 'Black 6', where every mouse is identical. In the ITP's four-way cross, no two of the more than 20,000 mice are the same, though any two share half their genes. Results are more likely to generalize. The NIA stopped supplying such mice for lack of demand.

• 05

  Rapamycin works even when started late, in both sexes

  In the second cohort (started 2005), 90–95% of oral rapamycin was destroyed in the stomach. Reformulating it with a coating took a year, so treatment began at 19–20 months, roughly a 60-year-old human. By then 15–20% of males had already died. Rapamycin still extended lifespan strongly in both sexes and remains the only drug with a very strong effect in females. Peter cites 90th-percentile lifespan gains of about 9% in males and 14% in females.

• 06

  Acarbose and canagliflozin probably work by blunting glucose spikes

  Acarbose, suggested by David Allison and Daniel Smith as a calorie-restriction mimic, blocks starch digestion and works mainly in males. Canagliflozin, an SGLT2 inhibitor, also blocks peak glucose and also extends life mostly in males. That supports the view that cutting post-meal glucose peaks, not eating less, is the mechanism.

• 07

  17-alpha-estradiol helps only males

  This weak, non-feminizing form of estrogen extended male lifespan, and Peter and Miller compare a 14.4 ppm dose with a lower one. In another lab, males started at 16 or 20 months got stronger muscles and better glucose tolerance; females didn't. Female mice already outlive males, and their estrous cycles stop around 8–11 months.

• 08

  Negative results are valuable

  Resveratrol failed at doses about 3–10 times higher than in David Sinclair's original paper (Miller wasn't sure of the exact multiple). Green tea extract, curcumin and methylene blue also failed. Miller notes that drinking enough wine to match the mouse dose would take about 600 bottles a day. No ITP drug has significantly shortened lifespan.

• 09

  Starting late and measuring the tail are tougher, fairer tests

  Starting a drug when 35–40% of animals have already died asks a lot of it. Miller considers age at the 90th percentile of survival more informative than percentage change in median lifespan, because it shows whether a drug delays death in already-old animals. Rapamycin, 17-alpha-estradiol and acarbose all still work when started in middle age. In rapamycin mice, tissue changes in organs such as the adrenal gland resembled those of younger controls.

• 10

  Metformin's failure in mice isn't the final word

  Metformin alone didn't extend lifespan in the ITP; it worked only when paired with rapamycin, which may just be the rapamycin. Miller lists reasons: mice and people are organized differently, it was tested at one dose with continuous dosing, and a different dose or schedule might work. Observational data showing diabetics on metformin outliving non-diabetics justify the human TAME trial, he says.