Pane et al. 2026 — Probiotic detoxification of Cd, Cr, Hg, and Pb in SHIME and gut ex vivo models

Pane and colleagues assessed three gut-derived probiotic lactobacilli strains — Lactiplantibacillus plantarum LP14, Lactobacillus crispatus LCR04, and Lactobacillus acidophilus LA12 — for their capacity to detoxify a mixture of inorganic cadmium, chromium(III), mercury(II), and lead(II) salts across three escalating model systems: a static in vitro co-culture / post-growth assay, the SHIME® dynamic gastrointestinal simulator (gastric, small-intestinal, and short-term proximal colonic phases), and a gut ex vivo system (GEVS) using murine small intestine. Strain- and metal-specific sequestration emerged: LP14 and LCR04 reduced heavy-metal bioavailability in the colonic SHIME compartment whereas LA12 showed minimal SHIME detoxification despite broad in vitro activity, and all three strains attenuated heavy-metal-induced epithelial barrier disruption and pro-inflammatory cytokine induction in the GEVS. The study contributes mechanism-of-action data on probiotic-mediated gut bioremediation of inorganic Cd, Cr(III), inorganic Hg(II), and Pb(II); it does not measure contamination of any consumer product.

Key numbers

Heavy-metal stock concentrations used in the in vitro co-culture and post-growth tests (paper p. 5, §3.1 Results; salts: CdSO4, CrCl3, Hg(NO3)2, PbCl2):

• Cd(II): 1.1 mg/L final
• Cr(III): 17.6 mg/L final
• Hg(II): 8.8 mg/L final
• Pb(II): 14.9 mg/L final

Heavy-metal reduction (% w/w, mean ± SD, n = 3 biological replicates) by lactobacilli strains in the in vitro assay (Figure 1, paper p. 6):

Authors’ narrative (paper p. 6): Cd was reduced by all three strains in both co-culturing and post-growth conditions. Cr reduction was highly strain-specific in co-culturing — substantial by LA12, virtually no reduction by LP14 or LCR04 — while in the post-growth test Cr reduction was comparable across all three strains. Hg was not appreciably removed by LP14 in either test, whereas LA12 and LCR04 displayed high Hg removal in co-culture (less pronounced post-growth). Pb was “essentially not reduced by any of the strains in either test” (p. 6).

SHIME® bacterial growth across compartments (Figure 2, paper p. 7; ¹⁰log AFU/reactor and CFU/reactor, mean ± SD, n = 3 SHIME experiments):

• Inoculum bacterial counts in SHIME at start: 10.14 (LP14, LA12) or 10.15 (LCR04) ¹⁰log/reactor (SD < 0.1).
• Cell concentration remained stable (~10.1–10.3 ¹⁰log AFU/reactor) for all three strains through SI start, DUO end, JEJ end, and ILE end.
• After 24 h colonic incubation, LP14 and LCR04 displayed up to almost 2 ¹⁰log increase in AFU/reactor (LP14 reaching ~12.0, LCR04 reaching ~11.7); LA12 showed no significant growth in the colonic phase (~10.4).

SHIME® detoxification in the colonic compartment (Figure 3, paper p. 7; differential recovery of HMs by ICP-MS in cell pellets vs. cell-free supernatants after 24 h colonic incubation, n = 3 SHIME experiments):

• LA12 did not promote a significant variation in HM content in the supernatant or pellet relative to the SHIME blank, indicating no detectable colonic detoxification by this strain (text p. 6).
• Pb: predominantly accumulated by LCR04, with roughly 45 percentage-points higher recovery from the cell pellet relative to the control experiment (text p. 7).
• Cd and Cr: markedly accumulated by both LP14 and LCR04 (~20 to 40 percentage-points higher recovery from cell pellets vs. control) (text p. 7).
• Hg: the most recalcitrant metal — only a modest increase (~10 percentage-points) in metal associated with cell pellets was observed for LP14 and LCR04 relative to controls (text p. 7).
• Statistical significance levels for cell-pellet / supernatant differences vs. blank (per Figure 3 caption, reproducing the paper’s wording verbatim from p. 4 §2.4.5): * 0.01 < p ≤ 0.05; ** 0.01 < p ≤ 0.001; *** 0.001 < p ≤ 0.0001. The paper’s printed ** bound is non-monotonic with the * bound; reproduced as printed rather than corrected.
• End-of-ileum samples for all strains showed HM levels in the supernatant similar to the SHIME blank, indicating no HM detoxification under small-intestinal simulated conditions (text p. 6, “data not shown”).

Heavy-metal concentrations used in the GEVS ex vivo experiments (paper p. 8, §3.3): Cd 0.69 mg/L, Cr 31.2 mg/L, Hg 5.2 mg/L, Pb 6.9 mg/L (selected after preliminary range-finding indicated the in vitro / SHIME concentrations were too toxic for GEVS tissue viability).

GEVS endpoints (Figure 4, paper p. 9; mean ± SD, two independent biological experiments, two biological samples per condition each, technical duplicates; mRNA fold-change relative to control):

• FITC-Dextran (FD4) permeability: in the HM-mix (MIX) condition, all three lactobacilli panels showed a significant increase in permeability versus control (CTRL); pre-incubation with each strain’s cell-free supernatant (CFS) restored physiological permeability (LP14 panel: ** p ≤ 0.01 CTRL vs MIX, ** p ≤ 0.01 MIX vs MIX+LP14; LCR04 panel: *** p ≤ 0.001 CTRL vs MIX, *** p ≤ 0.001 MIX vs MIX+LCR04; LA12 panel: *** p ≤ 0.001 CTRL vs MIX, *** p ≤ 0.001 MIX vs MIX+LA12).
• Tight-junction gene expression in the MIX condition: claudin-2 (CLD-2), claudin-15 (CLD-15), and occludin (OCL) mRNA were dysregulated; pre-incubation with each strain’s CFS restored CLD-2 and CLD-15 levels (statistically significant across all three strains); OCL restoration was statistically significant only with LA12 (paper p. 8, §3.3).
• Pro-inflammatory cytokine mRNA (IFNγ, TNFα): MIX strongly elevated both (**** p ≤ 0.0001 vs CTRL across all three panels); pre-incubation with each strain’s CFS reduced both back toward control (significance levels: LP14 panel **** for IFNγ and ** for TNFα MIX vs MIX+LP14; LCR04 panel **** for IFNγ and * for TNFα; LA12 panel **** for IFNγ and **** for TNFα).
• Anti-inflammatory cytokine IL-10 mRNA: MIX down-regulated IL-10; pre-incubation with each strain restored or partially restored IL-10 (LP14 panel: **** p ≤ 0.0001 MIX vs MIX+LP14; LCR04 panel: ** p ≤ 0.01; LA12 panel: * p ≤ 0.1, per the Figure 4 caption’s atypical convention — see Verification notes).

Methods (brief)

Heavy metals were dosed as inorganic salts (CdSO4 Sigma 481,882; CrCl3 Sigma 230,723; Hg(NO3)2 Sigma 104,439; PbCl2 Sigma 268,690) dissolved in deionized water and sterilized by 0.22 µm filtration; stock solutions were matrix-matched to the calibration standards. Heavy-metal content was quantified by Inductively Coupled Plasma Mass Spectrometry on a Thermo Scientific iCAP RQ ICP-MS (RF power 1,550 W; collision gas flow 4–4.6 L/min) after closed-vessel digestion with concentrated nitric acid and H2O2 under reflux on a hot plate (20 min, to clear/transparent endpoint). The instrument measures total elemental concentration; the paper does not perform Cr(III)/Cr(VI) speciation or methylmercury/inorganic-Hg speciation. All experimental Cr was introduced as Cr(III) (from CrCl3); all experimental Hg was introduced as inorganic Hg(II) (from Hg(NO3)2). All HM experiments and quantifications were run in triplicate and results are expressed as reduction (%) by comparing levels before and after HM contact with bacterial cells, with initial concentrations in the stock solution as positive control. Bacterial viability was monitored by SYTO 24 / propidium iodide flow cytometry on a BD Accuri C6 Plus with FlowJo analysis (results in active fluorescent units, AFU). The SHIME® system used a dynamic upper-GI simulation (gastric pH 4.6 → 3.0 over 120 min via 0.5 M HCl dosing; small intestinal phase pH 5.5 → 6.5 → 7.5 over duodenal/jejunal/ileal sub-phases via 4.8 g/L NaHCO3 dosing; pancreatic juice with 15.9 g/L oxgall, 10.6 g/L pancreatin, 6.45 g/L trypsin, 8.1 mL chymotrypsin solution) followed by a short-term proximal-colon simulation (24 h, pH 6.5–5.8 maintained, anaerobic, 37 °C, 90 rpm) with sterile-filtered human fecal inoculum. The gut ex vivo system used resected small intestine from 13-day-old C57BL/6J mice cultivated under luminal flow (99 µL/h Iscove’s Modified Dulbecco’s Medium with 20% KnockOut serum replacement) in a silicone-based six-chamber rig under humidified O2 (95%) and CO2 (5%); tissues were stimulated for 5 h with Iscove’s medium alone (CTRL), the HM mix (MIX), or MIX with cell-free supernatants of each strain after 16 h MRS co-incubation with the HM mix at 1:10 (MIX + CFS LP14/LCR04/LA12). Endpoints: FITC-Dextran (FD4) permeability across the tissue, qPCR of CLD-2, CLD-15, OCL, IFNγ, TNFα, and IL-10 normalized to L34, and AlamarBlue tissue viability. Statistical analysis used Student’s t-test (in vitro and SHIME) or ANOVA (GEVS) via GraphPad Prism 7.

Limitations explicitly named by the authors (paper p. 11, §5): the SHIME® model lacks a complex competitive endogenous microbiota; the HM mixtures use defined inorganic-salt concentrations rather than dietary-bound or organic/methylated forms; the GEVS experiments do not dissect per-metal contributions or distinguish barrier reinforcement from sequestration; no in vivo data; absence of colonic ex vivo tissue limits compartment-level alignment with the SHIME finding that colonic conditions are where detoxification is most evident.

Implications

• Certification (HMTc): Contributes mechanism-of-action data for probiotic-mediated reduction of luminal Cd, Cr(III), inorganic Hg(II), and Pb(II) bioavailability under simulated GI conditions; does not generate occurrence data for any food matrix or consumer product, so it does not feed any product-category percentile pool. The Pb result is the most distinctive single finding (no in vitro reduction yet a ~45 percentage-points colonic-pellet accumulation by LCR04 under SHIME conditions), supporting the authors’ contention that live colonic metabolism — not cell-surface biosorption alone — drives Pb sequestration by L. crispatus LCR04 in this system.
• Microbiome: Direct metal–microbiome interaction paper. Adds to the body of evidence that specific lactobacilli species can immobilize ingested heavy metals in the colonic compartment via cell-pellet association and that pre-incubation with strain-derived metabolites (cell-free supernatants) protects intestinal epithelium from HM-induced tight-junction dysregulation and pro-inflammatory cytokine induction. Strain- and metal-specific behavior — LA12 active in simple in vitro assays but inactive in SHIME, LP14 and LCR04 reciprocally active in SHIME but not on Pb in vitro — is the load-bearing nuance for any downstream WikiBiome cross-link.
• App / Courses: Not directly applicable; this paper does not characterize a food, a supplement, or an exposure route. May surface in advanced course material on gut-microbiome-mediated detoxification mechanisms.

Wiki pages this source may touch

• cadmium
• chromium
• mercury
• lead

Verification notes

• Speciation flags (locked at ingest): the paper performs total-element quantification by ICP-MS and does not speciate. The experimental Cr is exclusively Cr(III) from CrCl3 (not Cr-VI), and the experimental Hg is exclusively inorganic Hg(II) from Hg(NO3)2 (not MeHg). Frontmatter uses Cr and tHg per the system-prompt rule that unspeciated quantifications default to the total form. The Methods section above states the actual species used in the experiment to preserve the chemistry context for downstream readers.
• No brand-firewall implications: the only corporate names in the paper are reagent / instrument / reference-material vendors (Sigma Aldrich, Thermo Scientific iCAP RQ, BD Accuri C6 Plus, Carl Roth, Merck, Chem-lab, ProDigest, GraphPad Prism 7) which are scientific reproducibility content and are retained per the Part 12 Exception 2 (locked 2026-05-17). No food / supplement / consumer-product brands appear in the source.
• No HMTc threshold framing: the paper proposes no certification thresholds and the Implications section does not introduce any.
• No consumer-product or food-ingredient match in the current taxonomy: this is a bioremediation mechanism paper using inorganic salt mixtures in gut simulators and a murine intestinal explant; ingredients: [] and products: [] are correct, not a missing-slug case.
• matrices: [simulated-gastrointestinal, in-vitro-digestion, intestinal-epithelium, gut-microbiome] reuses bare-string matrix tokens that already appear in the corpus for SHIME / GEVS / bioaccessibility studies; no new matrix terms invented.
• Jurisdictions are author-affiliation codes (Italy: Probiotical / University of Piemonte Orientale; Belgium: Ghent University CMET; Netherlands: Bron Consultancy) — the experiments are bench / ex vivo, not site-specific population sampling, so country codes describe where the science was performed rather than a sampling jurisdiction.
• Audit subagent (2026-06-03) flagged four findings against the source: (1) ❌ asterisk-inequality bounds in the SHIME significance bullet (“0.05 < p ≤ 0.01” is mathematically impossible) — verified against paper p. 4 §2.4.5; corrected to the paper’s verbatim “0.01 < p ≤ 0.05 (*)” and explicitly noted that the paper’s printed ** bound is non-monotonic; (2) ⚠️ FITC-Dextran significance level on the LCR04 GEVS panel — verified against Figure 4A; corrected from ** to *** for the MIX vs MIX+LCR04 comparison; (3) ⚠️ Figure 4’s atypical * = p < 0.1 convention (the paper’s caption diverges from the standard p < 0.05 cutoff) — applied to the IL-10 LA12 line and flagged here so downstream readers do not silently inherit the standard threshold; (4) ⚠️ minor rounding of SHIME pancreatic juice oxgall from the paper’s 15.9 g/L to 16 g/L in Methods — verified against paper p. 3 §2.4.1; corrected to 15.9 g/L. Audit verdict pre-revision: REVISE; post-revision: PROMOTE.

Page history

The five most recent substantive edits to this page. The full version history lives in git; when DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.