Dry Fasting — The Record

Pathway atlas · node 38 of 41

Epithelial-mesenchymal transition (EMT)

What fasting does to this node, by modality, by system and by grade. Disconfirming evidence is reported first.

Founder's role line (READ): "Developmental and morphogenetic program (embryogenesis, wound healing, fibrosis) in which epithelial cells acquire migratory mesenchymal characteristics." Key components (READ): SNAI1/2 (Snail/Slug), TWIST1, ZEB1/2, CDH1 (E-cadherin), CDH2 (N-cadherin)

DRY status

DRY: none located. Queries run:

MEASURED: neither opened human dry-fasting study measured SNAI1, SNAI2, TWIST1, ZEB1, ZEB2, CDH1 or CDH2. Stated precisely, because both studies discuss EMT: Mindikoglu 2022 cites outside literature that TUBB4 downregulation with TUBB3 upregulation drives EMT in colon cancer and that FLOT2 overexpression drives EMT in hepatocellular carcinoma, and in that study TUBB4B went up at the end of the fast and FLOT2 went down one week after. Neither EMT nor any EMT marker was measured. Those are discussion citations, not findings, and they do not constitute DRY evidence for this node.

Disconfirming and null first

The only fasting-intervention evidence located for this node runs AGAINST the direction a reader would expect, and it is the first thing stated.

Graded rows

#ClaimModalitySpecies / systemTierGradeSourceTag
38.1A once-weekly 24 h fast for 4 weeks increased Snail and vimentin and decreased E-cadherin in tumor tissue sampled in the post-fast overeating phaseIF-ADF family (deviation stated: once weekly, not alternate-day; the brief has no label for a weekly single-day fast)male BALB/c mouse, subcutaneous CT26 colon tumor, in vivoA1CKusuoka 2017, Int J Oncol 52(2):413-423MEASURED
38.2The same protocol increased lung micrometastasis 1.8-fold and 6-fold versus matched non-intermittent dietsIF-ADF family, same deviationsameA1CKusuoka 2017MEASURED
38.3The achieved calorie reduction was about 1.5 percent, not the intended 14 percent, so the exposure is a fast-overeat cycle and not caloric restrictionas abovesameA1CKusuoka 2017, authors' own characterizationMEASURED and READ
38.4Continuous 30 percent CR did not increase lung micrometastasis relative to the intermittent armCRsameA1CKusuoka 2017MEASURED
38.5Fasting changes TGF-beta, HIF and Wnt signaling, which are canonical EMT inducers, therefore fasting should modulate EMT. Hops: fasting to those upstream nodes (not measured here, those are other slices), those nodes to EMT (measured in non-fasting systems). Two unverified hops. Not a fasting effect.n/ahops shownMCderivation stated, no source assertedinference, shown and labeled

Conflict left unresolved

None located, because there is only one fasting-intervention measurement. The conflict that does exist is internal to that one study: EMT markers rose, yet the intended restriction was not achieved and the continuous-CR comparator showed no metastatic increase. Reported unresolved.

Open questions


← 37. DNA methylation machinery All 41 nodes 39. Integrin / FAK adhesion signaling →


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.