Founder's key components: BECN1, ATG family, MAP1LC3, SQSTM1/p62, ULK1, mTOR.
DRY status: none located. This is the one node for which a dedicated DRY query was executed: Russian, сухое голодание аутофагия Филонов исследование человек. Channel productive, literature empty. Returns were books and popular pages (Filonov, Сухое лечебное голодание: мифы и реальность; gastronom.ru; habr.com). No peer-reviewed human or animal study measuring any autophagy marker under absolute dry fasting was located in any channel. The English DRY query was not executable (search budget exhausted) and is a retrieval debt, not a finding.
Species precision on the 2016 Nobel Prize, settled from the primary source
[ROW A2 / not a fasting study / not graded] The 2016 Nobel Prize in Physiology or Medicine was awarded to Yoshinori Ohsumi, official motivation "for his discoveries of mechanisms for autophagy" (nobelprize.org, 2016 summary and press release). The press release states that Ohsumi "used baker's yeast to identify genes essential for autophagy". Species: Saccharomyces cerevisiae, which the press release's own reference list misspells as "cervisiae". The release adds that "Soon it became clear that virtually identical mechanisms operate in our own cells", and it makes no claim of any human intervention result. It links disrupted autophagy to Parkinson's disease, type 2 diabetes, cancer and genetic disease as associations, and notes drug development is ongoing. Therefore: the prize was for yeast genetics, tier A2, and it was not awarded for fasting. Any copy that implies a Nobel Prize validates fasting-induced autophagy in humans is false on the primary record.
DISCONFIRMING FIRST. This is the core of the node.
1. The one human study that measured autophagic FLUX found a post hoc p = 0.04 in one of three arms. [ROW H1 / TRE + CR / C] Bensalem J, Teong XT, Hattersley KJ, Hein LK, Fourrier C, Dang LVP, Singh S, Liu K, Wittert GA, Hutchison AT, with Heilbronn LK and Sargeant TJ also named in the statement of responsibility. "Intermittent time-restricted eating may increase autophagic flux in humans: an exploratory analysis." Journal of Physiology 2025;603(10):3019-3032. Retrieved from the University of Adelaide institutional repository; DOI and PMID were not in the retrieved record and are a retrieval debt.
- Design: 121 humans with obesity, randomized to standard care, calorie restriction, or intermittent fasting plus time-restricted eating (iTRE), 6 months, measured at 2 and 6 months.
- Method: autophagic flux, as the flux of MAP1LC3B isoform II in peripheral blood mononuclear cells in whole blood. This is the only human flux measurement located in this slice.
- At 2 months: between-group differences in change from baseline were not significant.
- At 6 months: iTRE versus standard care P = 0.04, described by the authors as a post hoc analysis.
- The authors note the result may be partly driven by a tendency for autophagy to DECREASE in the standard-care group. The effect is therefore partly a control-group drift, not purely an intervention rise.
- Calorie restriction versus standard care: not significant. The CR arm, which lost weight on the same trial, did not show the effect.
- A correlation between change in autophagy and change in blood triglycerides was uncorrected for multiple comparisons.
- The paper's own framing: autophagy is stimulated by nutrient restriction in animal studies but this had not been shown in humans, and the title says "may increase".
- Grade C is assigned despite the H1 tier, because his scheme grades a single weak study as C and this is one exploratory post hoc comparison inside one trial, with the two other comparisons null.
2. The strongest human tissue study at 72 hours contradicts itself, and the authors say so. [ROW H3 / PF / B] Vendelbo et al. 2014, PLoS ONE 9(7):e102031 (full citation under node 14). 72 h fast with tap water permitted, randomized crossover, n = 8 healthy men, vastus lateralis.
- LC3B-II increased with fasting, approximately 30 percent, main effect P < 0.05.
- p62/SQSTM1 ALSO increased with fasting, approximately 10 percent, main effect P < 0.05, and was unchanged by insulin.
- ULK1 Ser757 phosphorylation decreased with fasting (P < 0.05) and increased with insulin, which is consistent with mTOR releasing ULK1.
- BECN1, other ATG proteins, BNIP3 and cathepsins were not measured.
- A reporting inconsistency inside the paper: the text says LC3B-II was reduced by insulin while the Figure 4 legend says increased.
- The authors state that LC3B-II alone does not measure flux, that the p62 data conflict with the LC3B-II data, and that they could not conclude whether autophagy was induced or inhibited.
- Why this is the single most important row in the slice: p62/SQSTM1 is an autophagy substrate. If flux rose, p62 should fall. Here both LC3B-II and p62 rose. The two markers cannot both be explained by one increase in flux. The best human skeletal-muscle data at 72 hours of fasting are internally contradictory, and the investigators who produced them declined to call it induction.
3. Two assays in the same human samples disagree. [ROW H3 / TRE / B] Kröpfl et al. 2022, Apoptosis 27(9):730-739 (full citation under node 15). 14 to 15 h overnight fast, n = 8.
- By flow cytometry: LC3B positive monocytes per microliter rose with fasting, 121 +/- 69 versus fed 35 +/- 29, overall p = 0.026, fasting versus fed p = 0.029.
- By western blot on the same samples: LC3B-II/I ratio 1.14 +/- 0.55 with fasting versus fed set to 1, NOT significant. p62 0.62 +/- 0.06 versus fed, NOT significant.
- LC3B-II/I and p62 were not associated with each other (r = 0.43, p = 0.15).
- LC3B positive percentage of mononuclear cells: no difference between conditions.
- Lymphocytes showed a lower autophagy signal than monocytes in every condition (p = 0.016 fasting, 0.038 fed, 0.020 exercise), which is the paper's actual conclusion and is a cell-type finding, not a fasting-induction finding.
- The abstract says p62 was measured by flow cytometry while the Methods describe p62 only by western blot.
- Reading: the direction of the human autophagy signal under a short fast depends on which assay is used, in the same blood, in the same subjects.
4. The field's own researchers published a protocol in December 2024 to measure, for the first time, what popular writing already asserts. [ROW H1 / PF / no result] Masedunskas A, de Ciutiis I, Hein LK, Ge A, Kong YX, Qi M, Mainali D, Rogerson-Wood L, Kroeger CM, and others including Fontana L. Nutrients 2024;16(24):4297, DOI 10.3390/nu16244297, published 12 December 2024. Article type: Study Protocol. No results. The author list is truncated on the page and is a retrieval debt.
- Trial: PROFASTA, randomized two-period two-sequence crossover, Charles Perkins Centre and RPA clinic, University of Sydney.
- Intervention: three-day water-only fast at home, approximately 84 h including a 12 h overnight fast before the clinic visit, with up to two 250 mL servings per day of vegetable broth (110 kJ each) permitted if a participant struggles to continue. One arm adds a single glycogen-depleting cycling session on day 1.
- Primary outcome: autophagic flux in PBMCs, measured in whole blood by adding a lysosomal inhibitor, following the method of Bensalem et al.
- Exploratory outcomes include p62/SQSTM1, NBR1 and OPTN. LC3B-II/beta-actin was used only for the sample size estimate.
- Enrollment is stated inconsistently in the protocol itself: 24 in the abstract, 25 in Section 3.1, 20 in the Figure 1 legend, with 25 targeted to allow 20 percent dropout.
- The authors state that the correlation between tissue-specific autophagy and whole-blood measurements still needs to be established, and that future work will need to examine tissue-specific autophagy activation.
- Reading: as of this publication, autophagic flux under a multi-day water-only fast had not been measured in humans. The trial exists to do it. This is the cleanest available answer to "what has actually been measured in humans".
5. The most-cited human "fasting increases autophagy" datapoint is blood mRNA, in 11 people, for 4 days. [ROW H1 / TRE / C] Jamshed H, Beyl RA, Della Manna DL, Yang ES, Ravussin E, Peterson CM. "Early Time-Restricted Feeding Improves 24-Hour Glucose Levels and Affects Markers of the Circadian Clock, Aging, and Autophagy in Humans." Nutrients 2019;11(6):1234, DOI 10.3390/nu11061234, published 30 May 2019.
- Randomized controlled crossover, 4 days per arm, 3.5 to 5 week washout. 18 enrolled, 11 completed both arms (7 men, 4 women).
- eTRF eating window 08:00 to 14:00 (6 h) versus control 08:00 to 20:00 (12 h). The morning sample follows an 18 h fast.
- Tissue: whole blood in Tempus tubes. mRNA only. No protein. No flux.
- LC3A increased 22 +/- 5 percent in the morning, p = 0.001.
- ATG12 increased 5 +/- 2 percent in the evening, p = 0.04, and the authors state this was no longer significant after multiple-comparison adjustment.
- All other genes unchanged, p >= 0.13. SQSTM1/p62 is not mentioned in the retrieved text; the full gene list is in Table S1, which was not retrieved and is a retrieval debt.
- Authors' own limitation, quoted: "gene expression data are known to be limited in nature and not necessarily reflective of changes in protein levels", and their claim is hedged as eTRF "may also increase autophagy".
- Reading: a 22 percent rise in one transcript in whole blood, in 11 people, after 4 days, with its companion gene losing significance to multiplicity correction. This is the headline behind a very large amount of popular writing.
Supporting human evidence, such as it is
[ROW H3 / PF / C] Nunes et al. 2024, Nutrients 16(1):128. Human duodenal biopsies examined by transmission electron microscopy showed autophagic vacuoles and dense bodies in the apical region, interpreted as probable primary lysosomes, with microvillus shortening, bending and disruption, dilated and vesiculated smooth endoplasmic reticulum, heterogeneous mitochondria, dilated intercellular spaces and basement-membrane detachment; junctional complexes unchanged. After refeeding, "residual bodies probably resulting from the autophagic activity". This is morphological only, with no molecular marker, and the population is PEG-referred patients with intake below 50 percent of needs for at least a month, not a defined fast. Reference check performed: no study cited by this paper addresses autophagy in human intestinal mucosa under fasting. The electron microscopy is the authors' own.
[ROW H3 / starvation, not a fasting protocol / C] Human hepatocyte autophagy in severe starvation, three records located and not opened: Rautou et al, "Acute liver cell damage in patients with anorexia nervosa: a possible role of starvation-induced hepatocyte autophagy", Gastroenterology 2008, article S0016-5085(08)00903-7; Kheloufi et al, "Liver Autophagy in Anorexia Nervosa and Acute Liver Injury", BioMed Research International 2014, DOI 10.1155/2014/701064; Restellini et al, "Severe Starvation-Induced Hepatocyte Autophagy as a Cause of Acute Liver Injury in Anorexia Nervosa: A Case Report", Case Reports in Hepatology 2013, DOI 10.1155/2013/749169. These are human electron-microscopy observations in pathological starvation with liver injury, not in any fasting protocol, and they are retrieval debts. They are listed because they are the strongest human morphological autophagy evidence in existence and because the context in which it appears is organ injury.
What the marker proves versus what the marker is claimed to prove
Stated from opened sources only.
- LC3-II (MAP1LC3 on the founder's component list). Its steady-state amount is a balance between autophagosome formation and autophagosome clearance. Vendelbo et al. state in their own paper that LC3B-II alone does not measure flux. The PROFASTA protocol's method confirms the operational consequence: to measure flux you must compare paired samples with and without a lysosomal inhibitor, which is why the assay is run ex vivo on whole blood rather than in the living subject. A rise in LC3-II is consistent with more autophagosomes being made OR with autophagosomes failing to be cleared. It is claimed to prove the first. It proves neither on its own.
- p62/SQSTM1 (on the founder's component list). It is an autophagy substrate, so increased flux predicts a fall. In the 72 h human muscle study it ROSE alongside LC3B-II. In the 14 to 15 h human blood study it did not move and did not correlate with LC3B-II/I (r = 0.43, p = 0.15). In the two human datasets where both markers were measured, p62 and LC3 never told a single consistent story.
- ATG and LC3A transcripts. Jamshed et al. measured mRNA in whole blood and stated in print that gene expression is not necessarily reflective of protein. Transcript change is two steps away from a degradation event: transcript, then protein, then flux.
- Electron microscopy. The only human morphological evidence located comes from a malnourished clinical population (Nunes 2024) and from anorexia nervosa with liver injury (retrieval debts). Autophagic vacuole counts are a snapshot, not a rate.
- Flux. Measured once, in blood, in 121 people, over 6 months of iTRE, with a post hoc p = 0.04 in one arm, a null CR arm, a null 2-month timepoint, and the authors noting the control group drifted down.
Tissue and duration coverage, with the denominator stated
Human tissues in which any autophagy marker has been measured under a fasting modality, from this sweep: skeletal muscle (1 study, n = 8, 72 h), whole blood transcripts (1 study, n = 11, 4 days of a 6 h window), PBMC and monocytes by flow and western (1 study, n = 8, 14 to 15 h), PBMC flux (1 study, n = 121, 6 months iTRE), duodenal morphology (1 study, 30 baseline and 14 paired, chronic low intake). Total: 5 human datasets. Combined n across them: 8 + 11 + 8 + 121 + 30 = 178. Human tissues with zero located fasting autophagy measurement: liver, brain, heart, kidney, adipose. Durations measured in humans: 14 to 15 h, 18 h, 4 days of a 6 h window, 72 h, 6 months of iTRE. 84 h is registered (PROFASTA) and unreported.
Node 16 verdict, two grades because there are two different claims
- "Fasting-type nutrient restriction changes autophagy-associated markers in humans": grade B, best tier H1. This is his B definition exactly: human mechanistic and biomarker evidence. It is supportable, and it is not the same sentence as the next one.
- "Fasting increases autophagic flux in humans": grade C, best tier H1 (exploratory, post hoc). One weak study. The brief's rule applies: this is a single weak result and must not read as settled.
- "Fasting induces autophagy in humans" as written in popular sources: not supportable at any grade on the evidence located. The mouse and yeast literature is not what the sentence claims, the human static markers are mutually contradictory, and the one human flux measurement is exploratory.
Open questions
1. PROFASTA results. Until they publish, no human study has reported autophagic flux under a multi-day water-only fast. 2. Open Bensalem et al. 2021, "Measurement of autophagic flux in humans: an optimized method for blood samples", to record the lysosomal inhibitor used and the method's stated limits. Retrieval debt; the host (aminer.org) returned only site metadata. 3. Open "Training state and skeletal muscle autophagy in response to 36 h of fasting", J Appl Physiol 2018, and "Muscle fiber type-specific autophagy responses following an overnight fast and mixed meal ingestion in human skeletal muscle", Am J Physiol Endocrinol Metab 2022. Both are human, both are directly on this node, and journals.physiology.org refused this session. Retrieval debts. 4. Open "Increasing autophagy with fasting and diet: separating fact from fiction", locator PMC13673093. By its title it is the closest existing review to this node's question. Retrieval debt. 5. Does whole-blood PBMC flux track tissue autophagy in humans? The PROFASTA authors state it is not yet established.
← 15. Apoptosis (intrinsic and extrinsic) All 41 nodes 17. NRF2 / KEAP1 (oxidative-stress response) →
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.