Convergent mitochondrial toxicity of fluorinated agents in oncological anaesthesia: sevoflurane, 5-fluorouracil, and the ACOD1/itaconate immune axis

Introduction

The biochemistry of fluoroacetate-mediated aconitase inhibition—first described by Peters in 1952 as the prototype of lethal synthesis (1)—provides the mechanistic foundation for the framework developed in this Perspective. Fluorocitrate, the toxic metabolite generated from fluoroacetate, inhibits aconitase 2, the mitochondrial enzyme that converts citrate to cis-aconitate in the tricarboxylic acid (TCA) cycle (2–4). The direct product of this reaction, cis-aconitate, is the obligate substrate for ACOD1 (aconitate decarboxylase 1, formerly IRG1)—the sole enzyme responsible for itaconate biosynthesis. Any agent capable of inhibiting aconitase will therefore deplete the cis-aconitate substrate for ACOD1 and suppress itaconate synthesis, with downstream immunological consequences.

The ACOD1/itaconate axis has emerged over the past decade as a pivotal immunometabolic regulator in cancer biology, with roles in tumour immune surveillance, NLR family pyrin domain-containing protein 3 (NLRP3) inflammasome control, and nuclear factor erythroid 2-related factor 2 (Nrf2) pathway activation (5–7). We propose that two agents routinely co-administered in oncological surgery create convergent pressure on aconitase 2: 5-fluorouracil (5-FU), whose hepatic metabolism to fluoroacetate and subsequent aconitase inhibition via fluorocitrate is experimentally established (8, 9), and sevoflurane, the predominant volatile anaesthetic worldwide, whose CYP2E1-mediated metabolism generates fluorinated intermediates for which further biotransformation to aconitase-inhibitory species cannot be excluded (10–12). The timing of this convergence—the perioperative window in fluoropyrimidine-treated oncological patients—places it precisely when intact anti-tumour immune surveillance is most critical.

The biochemical basis: lethal synthesis and aconitase inhibition

As described by Peters (1952), lethal synthesis converts a non-toxic precursor into a toxic metabolite using the organism’s own enzymatic machinery (1). Fluoroacetate—whether exogenous (rodenticide ‘1080’) or generated endogenously from fluoropyrimidines—condenses with oxaloacetate in the TCA cycle to yield fluorocitrate, of which only the (2R,3R)-erythro-2-fluorocitrate diastereomer inhibits aconitase (2–4). Because biological fluorocitrate synthesis generates all four diastereomers non-selectively, aconitase inhibition is inherently partial—never complete (13, 14)—a feature that defines the sublethal perioperative scenario and distinguishes it mechanistically from classical acute fluoroacetate poisoning.

Brain and glial tissue exhibit disproportionate sensitivity to fluorocitrate, reflecting preferential astrocytic uptake of tricarboxylic acids and high metabolic demand (15). This selectivity underpins the well-documented syndrome of type B lactic acidosis and encephalopathy associated with 5-FU, attributed definitively to Krebs cycle inhibition by hepatically generated fluoroacetate (9).

5-Fluorouracil: an established source of aconitase-inhibitory fluoroacetate

Arellano et al. (1998) demonstrated by ¹⁹F NMR and enzymatic assay that 5-FU undergoes hepatic β-oxidation to yield free fluoroacetate in quantities sufficient to inhibit aconitase and produce measurable TCA cycle perturbation (8). The clinical correlate—type B lactic acidosis, encephalopathy, and cardiac arrhythmia appearing 4–72 h after infusion—is well characterised (9, 16). Although 5-FU has a plasma half-life of 8–20 min, its metabolic activation continues for days; perioperative scheduling typically places surgery within 4–8 weeks of the last chemotherapy cycle, squarely within the activation window. Capecitabine, FOLFOX (folinic acid, fluorouracil, and oxaliplatin), FOLFIRI (folinic acid, fluorouracil, and irinotecan), and—most critically—hyperthermic intraperitoneal chemotherapy (HIPEC) protocols using intraperitoneal 5-FU share this risk profile.

Sevoflurane: CYP2E1 metabolism and evidence of aconitase perturbation

Sevoflurane is metabolised primarily by cytochrome P450 2E1 (CYP2E1) to hexafluoroisopropanol (HFIP) and formaldehyde (11, 17). Karpiński, Szulc & Szyfter (2006), in a systematic analysis of CYP2E1-mediated metabolism of inhalation anaesthetics, identified CYP2E1 as the dominant isoform responsible for sevoflurane defluorination and noted that the fluorinated metabolites generated by this pathway have the potential for further biotransformation (10). The specific downstream products and their aconitase-inhibitory potential require direct mass-spectrometric verification—an experiment not yet performed in humans. Compound A, the CO₂ absorber degradation product of sevoflurane, undergoes cysteine conjugate β-lyase-catalysed biotransformation to a thionoacyl fluoride capable of generating fluoroacetyl-equivalent adducts with mitochondrial proteins (18, 19), providing an additional mechanistic route to aconitase perturbation.

Wei et al. (2021) conducted a prospective observational study of plasma metabolomic changes in 500 patients undergoing abdominal surgery under sevoflurane anaesthesia, with full profiling in 114 (12). A dual elevation of plasma citrate and cis-aconitate—a metabolomic pattern consistent with aconitase perturbation—was documented in every profiled patient, with dose-dependence on end-tidal sevoflurane concentration and correlation with cumulative minimum alveolar concentration-hours (MAC-hours). It should be noted that simultaneous citrate and cis-aconitate elevation may also reflect mitochondrial stress, altered TCA flux from surgical trauma, or hypoxia-related shifts; fluorocitrate-mediated inhibition represents one mechanistic interpretation of this pattern, and the identity of the responsible inhibitory species has not been directly established. Geisser et al. (2003) provided independent support in the form of insulin-independent perioperative hyperglycaemia (plasma glucose 4.1 ± 0.3 to 7.0 ± 1.0 mmol·litre⁻¹), consistent with citrate-mediated allosteric inhibition of phosphofructokinase-1 downstream of the aconitase step, though metabolic stress-related explanations cannot be excluded (20) (Table 1).

Table 1. Pathways to aconitase inhibition by fluorinated agents. HTn, 4-hydroxy-trans-aconitate; HFIP, hexafluoroisopropanol.

The ACOD1/itaconate immune axis: roles in tumour immunity

ACOD1 is the sole enzyme responsible for itaconate biosynthesis, catalysing the decarboxylation of cis-aconitate to itaconate in the mitochondrial matrix. Under inflammatory conditions, itaconate accumulates to millimolar concentrations in macrophages, where it exerts pleiotropic immunomodulatory effects through at least three distinct molecular mechanisms (5).

First, itaconate alkylates Kelch-like ECH-associated protein 1 (KEAP1), stabilising Nrf2 and activating an antioxidant transcriptional programme that constrains inflammatory gene expression and supports cellular redox homeostasis (5). Second, itaconate inhibits succinate dehydrogenase (SDH), preventing succinate accumulation and the consequent activation of hypoxia-inducible factor 1α (HIF-1α)-driven pro-inflammatory transcription. Third, itaconate directly alkylates NLRP3 at Cys548, inhibiting inflammasome assembly and the downstream release of IL-1β and IL-18 (7).

In the tumour microenvironment, ACOD1 activation in TAMs is driven by NF-κB signalling from tumour-derived stimuli. The resulting itaconate inhibits TET dioxygenase activity in TAMs, dampening inflammatory gene expression and reducing CD8⁺ T-cell recruitment to the tumour bed (6). Simultaneously, ACOD1 in tumour-infiltrating neutrophils (TINs) confers Nrf2-mediated ferroptosis resistance, promoting pro-metastatic neutrophil functions (21). Mak et al. (AACR 2024) demonstrated that the IRG1/ACOD1 axis mediates resistance to cancer immunotherapy, with ACOD1 expression in tumour-infiltrating myeloid cells correlating with anti-PD-1 treatment failure (22). Table 2 summarises the key ACOD1/itaconate functions and their predicted perioperative consequences.

Table 2. ACOD1/itaconate immune functions in the tumour microenvironment and predicted perioperative consequences of itaconate depletion. TAMs, tumour-associated macrophages; TINs, tumour-infiltrating neutrophils; NLRP3, NLR family pyrin domain-containing protein 3; NK, natural killer.

Perioperative ACOD1/itaconate depletion: a paradoxical immunological vulnerability

The perioperative depletion of cis-aconitate through aconitase inhibition creates a paradoxical immunological state that operates at two distinct levels, with opposing implications. At the tumour microenvironment level, reduced itaconate synthesis in TAMs partially de-suppresses TET dioxygenase activity; in principle, this could partially restore local inflammatory gene expression and CD8⁺ T-cell recruitment. However, the net effect on anti-tumour immunity is unlikely to be beneficial: itaconate also sustains ACOD1-dependent pro-tumorigenic neutrophil survival, and the complex remodelling of TAM polarisation resulting from itaconate loss is context-dependent and not predictably anti-tumorigenic. At the systemic level, by contrast, the depletion is unambiguously detrimental: loss of the itaconate-mediated brake on NLRP3 inflammasome activation predisposes to exaggerated systemic inflammatory responses, cytokine release, and sepsis risk—particularly relevant in major oncological surgery where infectious complications are a leading source of perioperative mortality. In summary, the paradox is that itaconate depletion may simultaneously reduce pro-tumorigenic TAM-derived immune suppression while increasing systemic inflammatory susceptibility; both consequences carry clinical risk in the perioperative oncological setting.

These ACOD1-mediated consequences are compounded by direct immunosuppressive effects of volatile anaesthetics independent of the fluoroacetate pathway. Volatile agents suppress natural killer (NK) cell cytotoxicity—documented for halothane and isoflurane but not propofol (23)—and upregulate HIF-1α in cancer cell lines, potentially promoting tumour cell survival (24). Together, volatile anaesthesia and the likely contribution of fluoropyrimidine-mediated aconitase perturbation produce a layered immune vulnerability—ACOD1/itaconate depletion, NK cell suppression, and HIF-1α upregulation—converging at the moment of surgical resection.

Clinical and pharmacogenomic implications

Given the likely contribution of volatile anaesthetics to perioperative immune vulnerability in fluoropyrimidine-treated patients, propofol-based total intravenous anaesthesia (TIVA) remains recommended as the preferred technique in this population, eliminating the volatile anaesthetic contribution to aconitase perturbation and NK cell suppression. This recommendation is supported by meta-analytic evidence consistently favouring TIVA for oncological outcomes (25–27). The ENIGMA-II randomised controlled trial (RCT; Sessler et al., 2019) found no overall difference in breast cancer recurrence (28), but its design did not stratify by fluoropyrimidine exposure or ACOD1 genotype—the subgroup predicted by this framework to be at greatest risk. The foundational retrospective observation by Exadaktylos et al. (2006) linking anaesthetic technique to cancer recurrence remains relevant as the first empirical evidence of this relationship (29).

Pre-operative ACOD1 genotyping—specifically rs2009009 and related variants affecting promoter activity—is proposed alongside dihydropyrimidine dehydrogenase (DPYD) genotyping as a dual genetic risk profile for anaesthetic technique selection. DPYD genotyping is increasingly incorporated into pre-chemotherapy screening to identify patients at risk of severe 5-FU toxicity; ACOD1 genotyping would extend this approach to the perioperative setting.

Discussion

This Perspective integrates four independent lines of evidence: the established biochemistry of fluorocitrate-mediated aconitase inhibition; the experimentally proven fluoroacetate pathway of 5-FU; the emerging ACOD1/itaconate immune axis in oncology; and prospective clinical metabolomic evidence of aconitase perturbation in sevoflurane-anaesthetised patients. The novelty lies in their synthesis, which reveals a mechanistically coherent immunological vulnerability specific to the fluoropyrimidine-treated oncological perioperative patient.

The most important limitation of this framework is the absence of direct chemical identification of a fluoroacetate-equivalent metabolite of sevoflurane in humans. Karpiński et al. (2006) established that CYP2E1-dependent defluorination of sevoflurane generates fluorinated metabolites capable of further biotransformation (10), but the identity of any resulting aconitase-inhibitory species has not been directly characterised. The Wei et al. (2021) metabolomic findings (12)—while consistent with aconitase perturbation—are correlational and cannot exclude alternative explanations such as mitochondrial stress or altered TCA flux secondary to surgical trauma and anaesthesia. A prospective study directly comparing plasma fluoroacetate levels and itaconate concentrations after sevoflurane vs. propofol anaesthesia, stratified by CYP2E1 inducer status, would provide the direct mechanistic evidence currently lacking.

The ACOD1/itaconate mechanism differs fundamentally from previously described immunological effects of volatile anaesthetics. NK cell suppression by volatile agents has been documented for two decades (23) and is non-specific with respect to patient chemotherapy background. The ACOD1/itaconate mechanism, by contrast, is specifically potentiated by fluoropyrimidine exposure and is therefore invisible to studies that do not stratify by prior chemotherapy. This contextual specificity may explain why this mechanism has not previously been identified in the extensive literature on anaesthetic technique and oncological outcomes.

Conclusions

Sevoflurane and 5-FU may converge on aconitase perturbation through complementary mechanisms in fluoropyrimidine-treated oncological patients, depleting the cis-aconitate substrate for ACOD1 and suppressing itaconate synthesis at a critical perioperative window. The functional consequences—altered TAM polarisation, NLRP3 disinhibition, reduced NK cell cytotoxicity, and potential impairment of anti-PD-1 immunotherapy efficacy—constitute a layered immune vulnerability specific to this patient population that current anaesthetic safety frameworks are not designed to detect. Prospective metabolomic evidence supports the occurrence of aconitase perturbation during sevoflurane anaesthesia; direct mechanistic verification and immunological translation require prospective validation with ACOD1/itaconate endpoints. Until such data are available, propofol-based TIVA remains recommended for major oncological surgery in fluoropyrimidine-treated patients, and pre-operative DPYD + ACOD1 dual genotyping is proposed as a pharmacogenomically rational framework for anaesthetic technique selection.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Ansgar Brambrink, Columbia University, United States

Reviewed by: Guilherme Ayres Rossini, Butantan Institute, Brazil