Minocycline

Limits on what this shows

• Null findings: Counter-direction preclinical literature describes minocycline as mitochondria-protective (reduced calcium uptake and ROS in [9]; further neuroprotection papers such as Wang 2004 and Motaghinejad 2023 were seen in a PubMed listing and not opened [16]). Net effect in people is unknown.
• At-risk subgroups: Not defined. Whether people with mitochondrial disease respond differently was not searched in a form that returned data.

Regulatory status

Neither label retrieved mentions mitochondria [1][2].

Population evidence

No human study with a mitochondrial endpoint was located. Xiang 2026 [8] (Discover Oncology; abstract via OpenAlex): a chemical screen identified minocycline as selectively toxic to chemo-resistant nasopharyngeal carcinoma cells; it inhibits mitochondrial translation, reducing activity of respiratory complexes I and IV and impairing oxygen consumption; the effect weakens under anoxia; knockdown of mitochondrial elongation factor Tu mimicked it; in a xenograft model minocycline reduced tumor growth and impaired mitochondrial function in tumor-derived cells. Concentrations were not in the abstract. Garcia-Martinez 2010 [9] (Biochemical Pharmacology; PubMed abstract via a summarizing fetch): rat cerebellar granule neurons and isolated brain mitochondria; minocycline partially depolarized mitochondria by about 5 to 30 mV, markedly inhibited state 3 respiration, raised state 4 oxygen uptake, and inhibited NADH-cytochrome c reductase and cytochrome c oxidase but not succinate-cytochrome c reductase, suggesting selective complex I and IV inhibition; the authors say effective concentrations were higher than those that block inflammation (values not in the abstract).

Class extrapolation

CLASS EXTRAPOLATION (Tier C, capped at yellow): Moullan 2015 [10] reports that tetracyclines such as doxycycline disrupt mitochondrial translation across cells, worms, flies, mice and plants. Cohort test: the abstract I opened (via OpenAlex) names doxycycline as the example and does not name minocycline, so minocycline is not confirmed to be in that cohort; it is not used for any grade here, only as context. The minocycline-specific findings above are independent of it.

Mechanism

Present: tetracycline-class inhibition of the mitochondrial ribosome and translation (named as the target in [8] by EF-Tu knockdown) and, in isolated mitochondria, direct inhibition of complexes I and IV [9].

Counter-evidence held on record

Counter-evidence and confounding Research systems use high concentrations and cell lines or isolated tissue; the xenograft study is an oncology model. The two studies come from unrelated groups and one mechanism, so they count as consistent preclinical support, not as human evidence. Both abstracts were read through summarizing tools. Framing caution Do not state that minocycline damages mitochondria in patients. This is a preclinical signal with a protective counter-signal, and no nutrient was measured.

Limits on what this shows

• Null findings: None for people. Tetracycline itself lacked the antioxidant effect in the same assays [11].

Regulatory status

No oxidative stress statement in the labels read [1][2].

Population evidence

Kraus 2005 [11] (Journal of Neurochemistry; PubMed abstract via a summarizing fetch): minocycline and reference antioxidants protected mixed neuronal cultures in an oxidative stress assay; in rat brain homogenate lipid peroxidation and DPPH radical-scavenging assays minocycline, unlike tetracycline, was an effective antioxidant with potency similar to vitamin E. No concentrations or IC50 values were in the abstract. Garcia-Martinez 2010 [9]: minocycline abrogated NMDA-induced ROS formation in neurons and the ROS rise triggered by 75 micromolar calcium in isolated brain mitochondria. A PubMed query combining minocycline with oxidative stress, malondialdehyde and antioxidant terms (130 results, 10 shown) returned mostly psychiatric and neurology reviews and no study of oxidative stress markers in minocycline-treated people [16].

Mechanism

Present, reverse direction: direct radical scavenging [11] and reduced calcium-driven mitochondrial ROS [9]. A competing, unmeasured possibility is more mitochondrial ROS from respiratory chain inhibition (see Mitochondrial burden); no ROS measurement under that condition was retrieved.

Counter-evidence held on record

Counter-evidence and confounding In vitro scavenging at research concentrations does not predict in vivo antioxidant capacity at clinical plasma levels (peak 2.1 to 5.1 mcg/mL after one 100 mg capsule [1]). The conflicting mitochondrial signals mean the net redox direction in people is unknown. This entry shares one study [9] with the mitochondrial entries. Framing caution Not an antioxidant depletion claim and not a claim that minocycline is a dietary antioxidant. Reverse-direction chemistry only.

Limits on what this shows

• Null findings: None found for systemic iron. A PubMed search for minocycline with hyperpigmentation and iron deposition returned 11 records (10 shown): case reports and reviews of minocycline pigmentation and an ultrastructural study; no iron status study [16].
• At-risk subgroups: Patients who develop pigmentation, typically after prolonged therapy (frequency and exposure window not retrieved).

Regulatory status

Both labels list pigmentation: the capsule label reports skin, mucous membrane, oral cavity and thyroid discoloration; the extended-release label lists tissue hyperpigmentation affecting skin, nails, teeth, sclerae and other organs [1][2]. Neither mentions iron as a component of the pigment. Iron appears in the labels only as an agent lowering absorption [1][2].

Population evidence

Sato 1981 [12] (Journal of Investigative Dermatology; PubMed abstract via a summarizing fetch): lesional skin from minocycline-treated patients with blue-black hyperpigmentation examined by transmission electron microscopy and energy-dispersive x-ray microanalysis. Electron-dense iron-containing particles were found in siderosomes and dermal histiocytes, occasionally among collagen; the authors concluded that iron-laden particles in dermal histiocytes were primarily responsible for the discoloration, though some siderosomes contained melanosome complexes. Patient number was not stated in the text retrieved. No study of minocycline and iron absorption, serum iron, ferritin or hemoglobin in treated people was located.

Mechanism

Present for the pigment: the abstract identifies iron in the deposits. Why iron accumulates (chelate formation, hemosiderin from microhemorrhage, melanin complexes) was not established in the text read. A chelation effect on dietary iron absorption is conceivable from the tetracycline class chemistry but was not tested here.

Counter-evidence held on record

Counter-evidence and confounding The finding is histologic and describes where the pigment contains iron, not whether the body is iron replete or depleted. A skin deposit is a drug adverse reaction (Hard Rule 2). The Okada 1993 pigment granule characterization (listed, not opened) may describe the composition further [16]. Framing caution Not an iron deficiency and not an iron overload claim. It does not support iron supplementation or iron restriction. Pigmentation is a clinical reaction to report to the prescriber.

Limits on what this shows

• Why no grade: Dual condition met as far as this review could test: no opened Tier B or C source supports even a yellow grade for copper, and no circulating consumer claim for copper was found to rebut (consumer audit tooling-limited). This is a disclosed gap, not a finding of safety. Copper is kept separate from Iron (Hard Rule 7). A later revision should open the Magnan 1984 and 1986 papers and re-run a consumer-claim search.

What was searched

Labels (both read): no copper statement [1][2]. First review: PubMed "minocycline copper" (14 results, 10 shown): two French-language minocycline copper(II) complex records (Magnan 1984, 1986; no abstract available, not opened), a 2026 tigecycline photodegradation paper (copper effects described for tigecycline only) [15], a yeast antioxidant study (Angrave 2001; no minocycline-copper experiment stated) [14]; PubMed "minocycline zinc copper serum levels": 0 results. Repair pass 2026-10-09: PubMed "minocycline ceruloplasmin" (3 results, listing summary only): a rat study of minocycline in iron-induced brain injury (PMID 26463975), a 2016 case report of aceruloplasminemia improved by minocycline (PMID 27175663; an inherited ceruloplasmin deficiency treated with the drug, so disease-to-drug, not a drug effect on copper, and no copper values in the abstract) and a 1977 French-language side-effect screening record (PMID 143905) with no copper result in the abstract; none opened in full. PubMed "Metal ion-tetracycline interactions in biological fluids" (4 records, the Brion series): the minocycline-naming parts cover calcium, magnesium and zinc, none copper [16][17]. PubMed "tetracycline copper serum humans" (5 results): a copper nanocluster sensor paper, bacterial genomics, an albumin-binding spectroscopy study, a doxycycline hepatitis case where iron and copper studies were reported normal or negative as part of a work-up (doxycycline, not minocycline, not a drug-effect measurement), and a Wilson disease hemolysis case treated with unnamed tetracyclines; none minocycline copper status. One malformed Boolean query returned the PubMed home page and was discarded.

Channel outcomes

PubMed listing pages worked after spaced waits (3 to 5 minutes) and with simple queries; several requests, including OpenAlex single-record lookups, returned 429 with a request not to refetch and were not repeated. The web search tool returned a message that its budget was used up, so no consumer-claim search was possible a second time. Genuine gap in human data for the sources reached; partly tooling-limited because the Magnan abstracts were unavailable and the consumer claim audit could not be run.

Limits on what this shows

• Null findings: Gallo 2025 [7] found no detectable change at 40 mg daily for 16 weeks. In Zaura 2015 [6] the minocycline fecal effect was short-lived relative to clindamycin (up to 4 months) and ciprofloxacin (up to 12 months) per the summarized text; no minocycline-specific result at 2, 4 or 12 months was reported in the text retrieved.
• At-risk subgroups: Not defined by these trials (healthy volunteers only). Older adults and people with prior antibiotic exposure are general C. difficile risk groups per class labeling, not a minocycline-specific finding.

Regulatory status

Labels carry a Clostridioides difficile associated diarrhea warning that notes it can occur more than two months after the last antibacterial dose, and a superinfection (including fungal) overgrowth precaution [1][2]. No other microbiome statement was retrieved.

Population evidence

Zaura 2015 [6]: two-center randomized placebo-controlled trial in 66 healthy adults (UK and Sweden) given clindamycin, ciprofloxacin, amoxicillin, minocycline or placebo, with saliva and feces at baseline, immediately after exposure and at 1, 2, 4 and 12 months (16S, selected shotgun metagenomes). In the UK cohort 15 participants were randomized to minocycline and 10 completed; 14 were randomized to placebo and 13 completed (figures as given by a summarizing fetch of the PMC text). Minocycline produced the only significant fecal diversity reduction in that cohort, directly after exposure; fecal composition differed from placebo for up to 1 month; predicted metagenome functions showed 102 KOs up and 81 down; tetracycline-class efflux genes (Tet*) and several other resistance genes rose in feces. On re-reading the PMC full text: the 5 UK minocycline participants who did not complete were lost to follow-up right after baseline; the text gives no numeric diversity value, P value or CI for the fecal reduction (the values sit in Figure S1C, not reproduced); of the predicted fecal KOs immediately after exposure, 102 rose and 81 fell significantly versus placebo, and 3 of the rising KOs were antibiotic-resistance associated. The Table S1 caption states that only genera or higher taxa present in at least 20% of samples are shown, with significance judged by Mann-Whitney test at P below 0.005 (red cells decreases, green increases versus placebo); the table body, and the minocycline dose and duration (given in Supplementary Text S1), could not be retrieved. Participants received minocycline or amoxicillin per the Methods summary. Gallo 2025 [7]: multicenter double-blind randomized placebo-controlled trial, 60 healthy adults randomized 2:1 to low-dose minocycline (DFD-29, 40 mg once daily) or placebo for 16 weeks; 38 on drug and 19 on placebo were microbiota-evaluable; stool, skin and vaginal samples at baseline and weeks 4, 8 and 16 (culture and 16S). The abstract reports no significant difference from placebo in microbial abundance at any site, in minocycline resistance (MIC90) or in opportunistic microbiota; no effect sizes in the record read. A second PubMed record read (uid lookup) confirmed the same content: DFD-29 is 40 mg minocycline extended-release capsules (indication context: rosacea), the conclusion states no detectable effects on skin, GI tract or vaginal microflora over 16 weeks, and the record gives no p values and no named primary endpoint; the full text was not opened.

Mechanism

Present as intended pharmacology: minocycline is an antibacterial and the label reports urinary and fecal recovery of one-half to one-third that of other tetracyclines [1], so some drug reaches the gut lumen. The two trials differ in dose, duration, formulation and endpoint, which is the most likely reason they disagree; this is an inference, not a tested explanation.

Other human data

A trial of oral minocycline for rosacea measured skin microbiota, not gut (Zhang 2023, listed in PubMed, not opened) [16]. A rat hypertension study that used minocycline to alter gut microbiota (Yang 2015, listed, not opened) is animal data [16].

Counter-evidence held on record

Counter-evidence and confounding Zaura 2015 is small (10 minocycline completers in the UK cohort), used healthy volunteers and a short exposure, and its exposure details were not retrievable. Gallo 2025 used a low sub-antimicrobial dose in a different formulation and its primary endpoint was not named in the record. The two trials therefore do not test the same exposure, and the conflict is not resolved. Neither measured a nutrient or a functional microbial product such as vitamin K or B vitamin synthesis. Why not red Two independent human cohorts exist and both measured the microbiome directly, but their results conflict, effect sizes were not retrieved, and neither links the change to a nutrient or clinical outcome. Framing caution This is a bounded finding about fecal microbiome composition after a short course in healthy volunteers. It is not a nutrient deficiency, not proof of dysbiosis in people on long-term acne or rosacea therapy, and not a reason to add probiotics or any supplement.

Limits on what this shows

• Null findings: No measurement showing unchanged or reduced ATP in minocycline-exposed human cells or people was found.

Regulatory status

No mention in the labels read [1][2].

Population evidence

None in humans. Surrogates only: Garcia-Martinez 2010 [9] (state 3 respiration markedly inhibited, state 4 oxygen uptake raised, complexes I and IV inhibited, rat brain mitochondria) and Xiang 2026 [8] (impaired oxygen consumption and lower complex I and IV activity in carcinoma cells). A PubMed query combining minocycline, mitochondria, ATP and cytochrome c returned no records; a broader query on minocycline neuroprotection and mitochondria (39 results, 10 shown) listed no human study and no visible ATP measurement [16].

Mechanism

Present as for mitochondrial burden. Cells can compensate with glycolysis, so lower respiration does not establish lower ATP.

Counter-evidence held on record

Counter-evidence and confounding This entry and Mitochondrial burden rest on the same two studies and one mechanism (see the overlap map). It is one line of evidence, not two. Framing caution Not a claim of low energy or low ATP in patients. A "low energy" statement about minocycline users has no support in the sources read.

Limits on what this shows

• Null findings: No oxygen saturation, hematocrit or red-cell mass effect of routine minocycline use was reported in any source read.

Regulatory status

Capsule label adverse reactions: agranulocytosis, hemolytic anemia, thrombocytopenia, leukopenia, neutropenia, pancytopenia and eosinophilia have been reported [1]. Extended-release label postmarketing section: hemolytic anemia, thrombocytopenia and eosinophilia; no hematologic events in its clinical-trial table [2]. No frequency was retrieved.

Population evidence

Kudoh 1994 [13] (case report; abstract via OpenAlex): a 2-year-old boy developed sudden anemia and hemoglobinuria after minocycline; Western blotting showed a specific IgG antibody against minocycline, interpreted as immune complex mediated intravascular hemolysis. Hemoglobin values and outcome were not in the abstract. A PubMed query combining minocycline with hemolytic, hemolysis, methemoglobinemia and hemoglobin terms (108 results, 10 shown) listed this case and an in vitro photohemolysis study of tetracyclines (listing only, not opened) and no hemoglobin outcome study [16].

Mechanism

Absent as a nutrient pathway. Hemolysis here is an immune-mediated adverse reaction. Any effect on hemoglobin through iron would run through the Iron entry, which has no human systemic data.

Counter-evidence held on record

Counter-evidence and confounding Anemia is not iron deficiency and hemolysis is not an oxygen-delivery nutrient finding (Hard Rules 2 and 7). A single case report cannot estimate frequency. Framing caution Adverse-reaction note for clinicians. It does not support any nutrient or iron supplementation claim.