Founder's components: PIK3CA, PIK3R1, PTEN, AKT1/2/3, MTOR, TSC1/2, RHEB, RICTOR, STK11. This is the single most heavily claimed node in the fasting literature. It is also where the strongest human nulls sit.
DRY status
DRY: none located, for any component and for IGF-1. No dry-fast study measured PIK3CA, PTEN, AKT, MTOR, TSC1/2, RHEB, RICTOR or STK11, and none measured IGF-1, IGFBP-1, IGFBP-3 or insulin. Papagiannopoulos et al. 2013 (5-day absolute food and water deprivation, n=10) measured anthropometry, hemodynamics, glucose, osmolality, electrolytes and renal function, and the paper states no hormonal profile was run. Papagiannopoulos-Vatopaidinos et al. 2020 (5-day dry fast, n=10) measured cortisol, ACTH, ADH, renin, angiotensin II, aldosterone, erythropoietin and catecholamines, and measured no insulin, no IGF-1 and no mTOR component. Koppold et al. 2021 (19-day daytime dry fast, n=34) measured hydration, glucose metabolism and monocyte circadian transcripts, and no node component. Mähler et al. (daytime dry fast, n=11 men) measured tissue microdialysis and interpreted insulin sensitivity from glucose oxidation without reporting a measured insulin value. Queries run: the four listed under node 1, plus "fasting mimicking diet IGF-1 randomized trial Longo results", "prolonged fasting humans IGF-1 reduction randomized trial IGFBP-1", Arabic "صيام رمضان دراسة الالتهاب السيتوكينات عامل النمو IGF-1 نتائج", Persian "روزه داری رمضان مطالعه التهاب سیتوکین IGF-1 نتایج". Empty for DRY on every one.
Disconfirming, nulls and contradictions first
- Eight weeks of 18 to 20 hour daily time-restricted eating did not move IGF-1, IGFBP-1 or IGFBP-3. Akasheh et al. 2024, an exploratory analysis of an 8-week randomized controlled trial, n=35 TRE (eating window roughly 1 pm to 7 pm) versus n=14 no-advice control. The authors' conclusion is that TRE did not affect IGF1, IGFBP1 or IGFBP3. HOMA-IR and 8-isoprostane did change over time; IGFBP2 increased; leptin fell only in the subgroup losing more than 3.5 percent of body weight. This is the highest-tier human test of the IGF-1 claim under a daily fasting window, and it is null. [TRE, H1, grade B+]
- One and six years of severe calorie restriction did not change IGF-1 in humans, against the rodent expectation. Fontana et al. 2008: severe CR without malnutrition did not change IGF-1 or the IGF-1:IGFBP-3 ratio, explicitly in contrast to rodent studies. What did lower IGF-1 was protein restriction: in moderately protein-restricted individuals total and free IGF-1 were significantly lower, and in six CR volunteers reducing protein from about 1.67 to 0.95 g/kg/day for three weeks lowered serum IGF-1 from 194 to 152 ng/mL. [CR, H3, grade B+]
- A 40 hour fast did not change plasma IGF-1. Larsen et al. 2006, n=6, vastus lateralis biopsies and blood at 3, 15 and 40 h: plasma IGF-1 unchanged, plasma myostatin unchanged, and no significant change in myoD, myogenin, Myf5, myostatin, atrogin-1 or MuRF-1 gene expression. [PF, H3, grade C]
- Calorie restriction reduced mTORC1 in the niche cell and not in the stem cell. Yilmaz et al. 2012: calorie intake regulates mTORC1 in Paneth cells but not in intestinal stem cells. A whole-tissue or whole-body statement that "fasting lowers mTORC1" is not what was measured; the effect was cell-type specific and acted non-cell-autonomously. [CR, A1, grade C]
- UNRESOLVED CONFLICT ON IGF-1, BOTH SIDES NAMED. FMD lowers IGF-1 and TRE does not, and CR alone does not. Side one: Wei et al. 2017, 100 adults randomized to FMD 5 days per month for 3 months versus unrestricted control, with crossover and 71 completing three cycles, reports reduced IGF-1 among the effects; Mohammadzadeh et al. 2025 pooled 11 randomized trials and found IGF-1 reduced by a weighted mean difference of 19.211 ng/mL, 95 percent CI -32.986 to -5.437, with no significant effect on the other cardiovascular risk factors examined; Rahbar et al. 2019 found IGF-1 significantly decreased after Ramadan in 34 men. Side two: Akasheh et al. 2024 (TRE, null), Fontana et al. 2008 (CR, null), Larsen et al. 2006 (40 h, null). The conflict is not adjudicated here. Fontana names protein intake as the determinant in humans; FMD is low in protein as well as calories; stating that protein explains the split would be inference, and it is logged as an open question rather than a finding.
- A second unresolved point inside the supporting side. Mohammadzadeh et al. 2025 pooled FMD trials specifically. Whether the IGF-1 effect belongs to fasting, to the protein restriction inside FMD, or to the weight loss, is not separable from the pooled estimate as retrieved.
Graded rows
| # | Modality | System | Tier | Grade | Finding | Source |
|---|---|---|---|---|---|---|
| 2.1 | PF (72 h) | Human skeletal muscle, n=8 men, crossover | H3 | B | mTOR phosphorylation decreased about 50 percent; downstream 4EBP1, ULK1 and rpS6 phosphorylation reduced; the insulin-stimulated mTOR/rpS6 response was blunted after the fast. LC3B-II rose about 30 percent, p62 degraded about 10 percent, MAFbx and MURF1 unchanged. THE ONLY HUMAN MEASUREMENT OF THIS NODE'S OWN COMPONENTS UNDER FASTING IN THIS SLICE. | Vendelbo et al. 2014, PLoS ONE 9(7):e102031, PMID 25020061, DOI 10.1371/journal.pone.0102031 |
| 2.2 | TRE (18-20 h daily, 8 wk) | Human, n=35 vs n=14 control, RCT secondary analysis | H1 | B+ | IGF1, IGFBP1 and IGFBP3 not affected. IGFBP2 increased. HOMA-IR and 8-isoprostane changed over time. NULL ON THE HEADLINE LIGAND. | Akasheh et al. 2024, Nutrients 16(20):3476, PMID 39458471, DOI 10.3390/nu16203476 |
| 2.3 | FMD (5 d/month x3) | Human, n=100 randomized, crossover | H1 | B+ | IGF-1 reduced, alongside reduced body weight, trunk and total fat and blood pressure. Post hoc, at-risk participants improved more on BMI, blood pressure, fasting glucose, IGF-1, triglycerides, LDL and CRP. Numeric IGF-1 delta not retrieved. | Wei et al. 2017, Sci Transl Med 9(377):eaai8700, PMID 28202779, DOI 10.1126/scitranslmed.aai8700 |
| 2.4 | FMD (pooled) | Human, 11 RCTs, 12 arms | H1 | B+ | IGF-1 WMD -19.211 ng/mL, 95% CI -32.986 to -5.437. Also HbA1c -8.589 mmol/mol (-12.389 to -4.769), SBP -4.148 mmHg (-7.584 to -0.711), DBP -2.263 mmHg (-4.162 to -0.364). No significant effect on other risk factors. PROSPERO CRD42024569426. | Mohammadzadeh et al. 2025, Diabetol Metab Syndr 17(1):137, PMID 40287774, DOI 10.1186/s13098-025-01709-5 |
| 2.5 | CR (1 y and 6 y) | Human | H3 | B+ | Severe CR did not change IGF-1 or IGF-1:IGFBP-3, in contrast to rodents. Protein restriction lowered total and free IGF-1; 194 to 152 ng/mL over three weeks in six CR volunteers cutting protein from 1.67 to 0.95 g/kg/day. CR IS NOT FASTING AND IS LABELED. | Fontana et al. 2008, Aging Cell 7(5):681-687, PMID 18843793, DOI 10.1111/j.1474-9726.2008.00417.x |
| 2.6 | RAMADAN (1 month) | Human, n=34 men, before and after | H3 | C | IGF-1 significantly decreased after Ramadan. IL-2, triglycerides, total cholesterol and LDL-C also decreased; HDL-C unchanged. Weight, waist and energy intake also decreased, so the IGF-1 change is not separated from them. No numbers or p-values retrieved. | Rahbar et al. 2019, Int J Prev Med 10(1):7, PMID 30774841, DOI 10.4103/ijpvm.ijpvm_252_17 |
| 2.7 | PF (40 h) | Human skeletal muscle, n=6 | H3 | C | Plasma IGF-1 unchanged. Plasma myostatin unchanged. No significant change in myogenic or atrogenic gene expression at 15 or 40 h. | Larsen et al. 2006, Ann Nutr Metab 50(5):476-481, PMID 16931880, DOI 10.1159/000095354 |
| 2.8 | CR | Mouse intestine | A1 | C | CR reduced mTORC1 signaling in Paneth cells and not in intestinal stem cells. Rapamycin mimicked the stem-cell benefit; forced Paneth-cell mTORC1 activation during restriction cancelled it; increased Paneth-cell Bst1 and cyclic ADP ribose mediated the effect. CR percent and duration not retrieved. | Yilmaz et al. 2012, Nature 486(7404):490-495, PMID 22722868, DOI 10.1038/nature11163 |
| 2.9 | PF (7 d, water only) | Human plasma, n=12 (5 F, 7 M) | H3 | B | About 3,000 plasma proteins measured before, daily during and after. More than 1,000 responded significantly, consistently across volunteers, in nine distinct response profiles. Systemic change became evident ONLY AFTER 3 DAYS. Mean weight loss 5.7 plus or minus 0.8 kg. No node component measured; listed here because it sets the duration threshold any PF claim must respect. | Pietzner et al. 2024, Nat Metab 6(4):764-777, PMID 38429390, DOI 10.1038/s42255-024-01008-9 |
| 2.10 | PF (26 h post-meal) | Human subcutaneous adipose, n=11; mouse 16 h | H3 | C | INSULIN SIGNALING genes down-regulated in human adipose, along with triglyceride and fatty acid synthesis, glycolysis, glycogen synthesis, TCA cycle, oxidative phosphorylation, mitochondrial translation and proteasomal degradation. Effects less pronounced in humans than mice and many genes responded in opposite directions between species. | Defour et al. 2020, Physiol Genomics 52(10):451-467, PMID 32866087, DOI 10.1152/physiolgenomics.00083.2020 |
Open questions
- Whether the FMD-versus-TRE split on IGF-1 is a protein-intake effect, a calorie-deficit effect, a weight-loss effect or a duration effect. Fontana names protein; no retrieved trial isolates it inside an FMD arm.
- Pietzner et al. 2024 show nothing systemic moves until day 3 of a water-only fast. Every null in this node that ran shorter than 72 h (Larsen 40 h, Elliott 36 h) is therefore consistent with the node being real and the exposure being too short. That is a dose-of-time problem, not a resolution, and it cuts both ways: it equally undermines any claim built on a sub-72-hour protocol.
- The human species contrast is explicit and unflattering in two independent places: Fontana (CR does not lower human IGF-1 as it does rodent) and Defour (adipose fasting response weaker in humans, many genes opposite in direction). Any statement that transports a rodent mTOR result into a human claim has to clear both.
← 1. RTK / RAS / MAP-kinase All 41 nodes 3. JAK / STAT →
dryfasting.com has been published since 2014. This evidence record was compiled on 22 November 2026. Every row names the fasting modality the study actually ran, the system it was measured in and the grade. Where no source carries a DOI or PMID the row says so rather than supplying one. Nothing is sold on this site.