Lua Labs Report — Chrononutrition, nocturnal insulin, and hormonal peaks

Date: 2026-06-15 Researcher: Lua Labs Classification: Chronobiology Line: L4 — Hormonal chronobiology Subtopic: 4.3 — Chrononutrition: feeding window and hormonal peaks (time-restricted eating)

External Sources

1. Corapi, Runchey, Lyons, Alonso de Leon, Pavlou, Lin, Ezpeleta, Gabel, Tussing-Humphreys, Oddo, Alexandria, Sanchez, Vidmar, Varady & Cienfuegos (2026). "Time-restricted eating for body weight management in women with polycystic ovary syndrome: a randomized controlled trial." Nature Medicine. https://doi.org/10.1038/s41591-026-04316-7
2. Floyd, Dyer, Gibney, Alawami, Owens, Phelan, Rakovac, Swan, Le Roux, Behan & Duggan (2026). "Time-Restricted Eating to Improve Metabolic Abnormalities in Polycystic Ovarian Syndrome (TimeMAP)." Clinical Endocrinology. https://doi.org/10.1111/cen.70094
3. Talebi, Shab-Bidar, Moini, Mohammadi & Djafarian (2024). "The effects of time-restricted eating alone or in combination with probiotic supplementation in comparison with a calorie-restricted diet on endocrine and metabolic profiles in women with polycystic ovary syndrome: A randomized clinical trial." Diabetes, Obesity and Metabolism. https://doi.org/10.1111/dom.15801
4. Peters, Schwarz, Schuppelius, Ottawa, Koppold, Weber, Steckhan, Mai, Grune, Pfeiffer, Michalsen, Kramer & Pivovarova-Ramich (2025). "Intended isocaloric time-restricted eating shifts circadian clocks but does not improve cardiometabolic health in women with overweight." Science Translational Medicine. https://doi.org/10.1126/scitranslmed.adv6787
5. Enomoto, Kitamura, Kunieda et al. (2026). "Effects of later dinner timing on subsequent metabolic function and nocturnal sleep in healthy young women." Journal of Physiological Anthropology. https://doi.org/10.1186/s40101-026-00430-0
6. Garaulet, Lopez-Minguez, Dashti, Vetter, Hernandez-Martinez, Perez-Ayala, Baraza, Wang, Florez, Scheer & Saxena (2022). "Interplay of Dinner Timing and MTNR1B Type 2 Diabetes Risk Variant on Glucose Tolerance and Insulin Secretion: A Randomized Crossover Trial." Diabetes Care. https://doi.org/10.2337/dc21-1314
7. Hummel, Benkendorff, Fritsche, Prystupa, Vosseler, Gancheva, Trenkamp, Birkenfeld, Preissl, Roden, Haring, Fritsche, Peter, Wagner, Kullmann & Heni (2023). "Brain insulin action on peripheral insulin sensitivity in women depends on menstrual cycle phase." Nature Metabolism. https://doi.org/10.1038/s42255-023-00869-w
8. Lin, Cienfuegos, Ezpeleta, Pavlou, Runchey & Varady (2024). "Effect of time restricted eating versus daily calorie restriction on sex hormones in males and females with obesity." European Journal of Clinical Nutrition. https://doi.org/10.1038/s41430-024-01461-5
9. Bohlman, McLaren, Ezzati, Vial, Ibrahim & Anton (2024). "The effects of time-restricted eating on sleep in adults: a systematic review of randomized controlled trials." Frontiers in Nutrition. https://doi.org/10.3389/fnut.2024.1419811

Baseline Knowledge (what I knew before searching)

Food is a peripheral zeitgeber. It does not move the SCN as strongly as light, but it does reconfigure liver, pancreas, muscle, intestine, and adipose tissue through insulin, GLP-1, PYY, bile acids, postprandial temperature, vagal signals, and NAD+/AMPK/SIRT1 availability. The central clock marks biological night through melatonin; the pancreatic and hepatic clocks interpret the arrival of glucose, amino acids, and lipids. When both messages align poorly — high melatonin + high food load — the system receives a contradictory instruction: "it is night, reduce insulin secretion" and "glucose is arriving, secrete insulin."

Molecularly, the collision is not abstract. Melatonin acts through MT1/MT2, and MTNR1B is expressed in pancreatic beta cells; its signaling tends to reduce cAMP and insulin secretion. At the same time, the beta cell responds to glucose through GLUT1/GLUT2, KATP closure, depolarization, Ca2+ entry, and exocytosis of insulin granules. Eating near melatonin onset demands a beta-cell response precisely when MT2 is braking that response. The expected result is lower early insulin, higher postprandial glucose, and greater sympathetic/nocturnal activation if the glycemic peak is prolonged.

In women, that collision depends on phase. The luteal phase, high in progesterone, reduces insulin sensitivity in many women and shifts energy partitioning toward storage. Hummel 2023 added a finer link: insulin action in the brain improves peripheral sensitivity in the follicular phase, but that response disappears in the luteal phase. This means that the same plate at the same time may carry a different metabolic-neuroendocrine cost depending on phase. In PCOS, hyperinsulinemia amplifies theca CYP17A1 and reduces hepatic SHBG; in perimenopause, late dinner may add nocturnal insulin, vasomotor load, and nocturnal autonomic activation, simulating "lack of progesterone" even when the real issue is timing.

The question I searched in the papers was concrete: does TRE work through the window, through energy deficit, through circadian alignment, or through a mixture?

Findings From Recent Papers

The 2024-2026 evidence separates two phenomena that public conversation often mixes. First, TRE can be useful in PCOS as an adherence and energy-reduction tool. Corapi et al. 2026 randomized 76 women with PCOS for 6 months to 6 h TRE (1 pm-7 pm), daily calorie restriction, or control. TRE reduced weight by 4.32% versus control, comparable to calorie restriction, without serious adverse events; the paper also shows the figure for changes in weight, free androgen index, and insulin resistance. Floyd et al. 2026, in TimeMAP, was small (15 recruited, 11 completed), but showed that 18:6 TRE was feasible in PCOS, with 94.7% adherence among completers, no serious events, and exploratory signals of lower HbA1c, weight, BMI, waist, and hip measurements. Talebi et al. 2024 was more restrained: 90 women with PCOS, eTRE 14:10 with or without probiotic vs calorie-restricted diet; all groups improved weight, hirsutism/acne, and vascular risk, but eTRE did not outperform standard calorie restriction and probiotic did not add benefit.

Second, timing without energy deficit does not guarantee metabolic benefit. Peters et al. 2025, ChronoFast, compared eTRE 8:00-16:00 vs lTRE 13:00-21:00 in 31 women with overweight/obesity under intended isocaloric conditions. Adherence was very high (96.5% eTRE, 97.7% lTRE), but there was no clinically meaningful improvement in insulin sensitivity, 24 h glucose, lipids, inflammation, or oxidation. There was clock shifting: in lTRE, sleep midpoint was delayed by 15 min and circadian phase in monocytes tended to be delayed by 24 min vs eTRE. This is the key piece: the window moves clocks, but the metabolic result requires energy deficit, composition, circadian phase, and biotype.

The melatonin-food interaction does have strong human causality. Garaulet et al. 2022 conducted a large crossover in 845 adults with two nocturnal OGTTs: 4 h vs 1 h before sleep. In the late condition, serum melatonin was 3.5x higher, insulin AUC was 6.7% lower, and glucose AUC was 8.3% higher; the effect was stronger in carriers of the MTNR1B G allele. Enomoto et al. 2026 grounds this in young women and more naturalistic conditions with CGM/EEG: the paper's hypothesis is precisely that eating during biological night with high melatonin can impair glucose through insulin secretion/sensitivity and affect sleep through thermogenesis, reflux, and autonomic arousal. Hummel et al. 2023 adds the phase layer: the luteal phase has lower brain response to insulin, which predicts greater vulnerability to late dinner.

Full Molecular/Endocrine Mechanism

The mechanism is not "fasting improves hormones." It is a network of temporal gates:

Light/darkness → melanopsin retina → SCN → pineal AANAT/ASMT → melatonin
                                                          ↓
                                            MTNR1B/MT2 in beta cell
                                                          ↓
Late food → glucose/amino acids → beta cell KATP/Ca2+ → insulin
                                                          ↓
                   melatonin-food collision → early insulin ↓ + glucose AUC ↑

Late high-carbohydrate/high-lipid dinner
  → nocturnal insulin + prolonged postprandial glucose
  → hepatic SHBG ↓ + hepatic/VAT lipogenesis ↑ + altered nocturnal leptin
  → theca CYP17A1 ↑ in IR/PCOS context + free androgens ↑

Luteal phase → progesterone ↑ + basal temperature ↑ + ALLO/GABA-A ↑
          → peripheral/central insulin sensitivity ↓ in subgroup
          → late dinner + high melatonin
          → nocturnal glucose ↑ + autonomic arousal + HRV ↓
          → fragmented sleep + AM energy ↓ + luteal mood ↓

Perimenopause → intermittent ovulation + variable P4 + vasomotor load
               → melatonin/nocturnal amplitude ↓ + nocturnal cortisol ↑ in hyperreactive phenotype
               → late dinner becomes a misaligned "second zeitgeber"
               → nocturnal insulin + postprandial heat + awakenings
               → nocturnal symptoms that seem primarily hormonal

The pancreatic core is direct: glucose raises ATP in the beta cell, closes KATP channels, opens calcium channels, and triggers insulin exocytosis. Melatonin through MTNR1B/MT2 reduces pro-secretory signaling. If both signals coincide, early secretion falls. In MTNR1B G carriers, the beta-cell brake seems stronger. The liver receives insulin at night, when its circadian program favors lower glucose tolerance; if exposure repeats, SHBG falls and free androgens rise in women with insulin resistance.

The ovary enters through two routes. Metabolic route: insulin acts as a co-gonadotropin in theca, potentiates LH/CYP17A1 and androstenedione/testosterone production; in granulosa, excess insulin and inflammation degrade coordinated aromatization. Circadian route: CLOCK/BMAL1 in granulosa/luteal tissue regulates StAR, CYP11A1, HSD3B, CYP19A1, and LHCGR. Late meal timing does not need to change serum E2/P4 to damage the signal: flattening temporal amplitude is enough for tissue to receive LH/FSH/insulin/melatonin in incoherent phases.

The important contradiction: PCOS trials show TRE benefit, but isocaloric ChronoFast does not show metabolic improvement. My reading: TRE is not an autonomous circadian drug. It is a behavioral architecture that helps if it reduces energy, lowers nocturnal collisions, stabilizes sleep, and improves adherence. If it only compresses schedule without reducing energy or aligning with biological phase, it moves clocks but does not necessarily improve metabolism.

Lua Labs Hypotheses

Hypothesis 35: Food-Melatonin Collision

Statement: The severity of nocturnal symptoms, low morning energy, and luteal/metabolic worsening is better predicted by collision between last caloric intake and biological night than by total feeding-window duration.

Proposed mechanism: The critical variable is not "fasting hours," but "calories under melatonin." When the last meal, snack, alcohol, or dessert falls near melatonin onset, MTNR1B brakes beta-cell secretion exactly when the intestine demands insulin. This increases nocturnal glucose, activates sympathetic tone, lowers HRV, and fragments sleep. In the luteal phase, lower central insulin action amplifies the effect; in perimenopause, variable P4 + vasomotor load turns it into heat awakenings, nocturnal anxiety, or AM fatigue.

Confidence level: Medium-high — strong human causality for melatonin+late dinner (Garaulet 2022), phase support from Hummel 2023. The original part is the longitudinal index applied to hormonal symptoms.

How to validate:

• With a formal study: 8-week observational study + CGM/wearable subgroup n=80; compare nights with last intake <2 h pre-sleep vs >=3 h within person; primary outcome: nocturnal glucose AUC, nocturnal HRV, awake minutes, and AM energy.

Limitations: Direct melatonin is rarely measured in everyday contexts; onset can be inferred only noisily from sleep, light/screens, chronotype, and local time. Food records may underreport nocturnal snacks. The effect will be small in metabolically healthy women, and larger in PCOS/perimenopause or MTNR1B-risk carriers.

Hypothesis 36: Luteal Glycemic Vulnerability

Statement: The same late meal will have greater cost for sleep, mood, and cravings during the mid-late luteal phase than during the follicular phase because of transient loss of central insulin action and high progesterone.

Proposed mechanism: The luteal phase raises P4, temperature, and energy demand. Hummel 2023 suggests that the brain stops increasing peripheral insulin sensitivity in this phase. If a late dinner occurs in that context, the system loses two buffers: the pancreatic buffer due to high melatonin and the central buffer due to low insulin-brain action. Result: more nocturnal glucose, more thermogenesis, worse sleep continuity, greater anxiety/irritability, and cravings the next day. This potentiates the luteal window of circadian vulnerability.

Confidence level: Medium — plausible physiology and small experimental human support; a study crossing meal timing, CGM, and phase in women without diabetes is missing.

How to validate:

• With a formal study: two-cycle crossover in n=40 cycling women with CGM + wearable; standardized early vs late dinner in follicular and luteal phases; endpoints: nocturnal glucose AUC, HRV, sleep efficiency, and AM mood.

Limitations: Estimated phase can be noisy if cycles are irregular or hormonal contraceptives are used. In anovulatory PCOS there is no classic luteal phase; stage/low-P4 signature or irregular cycle should be used as a separate stratum.

Hypothesis 37: Perimenopausal Late-Dinner Trap

Statement: In early perimenopause, the pattern of late dinner + fragmented sleep + vasomotor symptoms may precede open menstrual irregularity and act as an early marker of nocturnal endocrine-metabolic collapse.

Proposed mechanism: Perimenopause combines intermittent ovulation, erratic P4, lower melatonergic signaling, higher caregiving/stress load, and a tendency to eat late because of schedule. Late dinner generates postprandial heat, reflux/fragmentation, nocturnal glucose/insulin, and sympathetic activation. These phenomena increase heat awakenings and AM fatigue, which are clinically read as "menopause" even when part of the phenotype is a chrononutritional modifier. It does not deny ovarian transition; it proposes that late eating is a measurable amplifier of the transition.

Confidence level: Medium — strong physiology; specific longitudinal perimenopausal validation is missing.

How to validate:

• With a formal study: 12-week cohort n=100 perimenopausal women with wearable + optional CGM, comparing trajectories of food-melatonin collision, nocturnal synchrony, and vasomotor symptoms.

Limitations: Vasomotor symptoms also depend on fluctuating E2, thermoneutral zone, alcohol, ambient temperature, and MHT. The index should not become the single explanation for perimenopause.

Candidate Formulation (if applicable)

Compounds: No drug or supplement is proposed. A behavioral-dietary research architecture is proposed: Chrono-Luteal Meal Timing Foundation.

Target population: Women with possible PCOS/insulin resistance and women in perimenopause as primary populations; cycling women with PMS/luteal insomnia as a secondary population; postmenopausal women only for sleep/metabolism, without reproductive claims.

Complementary mechanisms: (1) distance from last calorie to sleep to reduce melatonin-food collision; (2) regularity of first meal to stabilize peripheral clocks; (3) greater protein/fiber load earlier to reduce nocturnal cravings; (4) avoid excessively long windows in fatigue/flat-cortisol phenotypes; (5) phase-aware adjustment: more caution with late meals in luteal phase and perimenopause.

Regulatory status: Food/behavioral, GRAS, non-prescriptive.

Requires validation: 90-day observational pilot before turning it into a personalized recommendation.

Individual Variability

The most important genetic variability is MTNR1B rs10830963 G, because it increases susceptibility to worse glucose when food coincides with high melatonin. Women with low AM energy, fragmented sleep, and cravings after late dinners may be "melatonin-meal sensitive."

Menstrual phase changes the threshold. Late follicular phase with high E2 usually has better insulin sensitivity; luteal phase high in P4 reduces or reorganizes that sensitivity. In anovulatory PCOS, phase becomes less useful; insulin resistance, acne, hirsutism, oligomenorrhea, and cravings matter more. In perimenopause, ovulatory variability means the same pattern can alternate between tolerable weeks and very symptomatic weeks.

Chronotype matters. A dinner at 20:30 can be biologically early for someone who sleeps at 00:30 and late for someone who sleeps at 22:00. That is why the model should use distance to sleep and not a fixed universal time. Night shifts require a separate model: melatonin may be suppressed, shifted, or fragmented.

The metabolic context dominates effect size. Insulin resistance, family history of DM2/PCOS, short sleep, hyperactivation stress, nocturnal alcohol, and low daytime fiber amplify the collision. A metabolically flexible woman may tolerate occasional late dinners; a woman with PCOS/perimenopause/Hashimoto-like phenotype may pay for them with sleep, mood, and glucose even if calories are equal.