Mitochondrial pearling: membrane biophysics, Ca²⁺-dependent regulation, mtDNA nucleoid dynamics, and emerging links to mitochondrial quality control and neurological disorders

Abstract

Mitochondrial dynamics has long been interpreted primarily through fission and fusion, yet tubular mitochondria can also undergo rapid pearling, a phenomenon in which elongated mitochondria reorganize into a beads-on-a-string morphology while retaining a continuous imaged contour. The occurrence and biological relevance of mitochondrial pearling require careful study. The dimensionless tension–bending ratio used to organize these observations is a heuristic analogy to single-membrane tubes, not a validated quantitative model of the mitochondrial double membrane. Evidence does not yet establish a continuous sequence from pearling through coordinated outer- and inner-membrane scission to mitophagy or intercellular mitochondrial transfer; those links are therefore presented as hypotheses and testable predictions. We use the provisional term “candidate disease-associated sustained pearling phenotype” only for within-study events that meet dynamic pearling criteria and show longer duration or delayed/failed reversal relative to appropriately matched controls. No universal duration threshold or validated pearling-defined disease entity currently exists. Event duration, wavelength, un-pearling kinetics, separate outer- and inner-membrane continuity, and the fate of individual pearls should be measured together to determine whether sustained events are incidental, adaptive, or causally involved in disease.

1. Introduction: from beads-on-a-string mitochondria to an expanded framework of mitochondrial dynamics

Mitochondria support cellular metabolism and continually remodel their membranes. Most work over the past 3 decades has interpreted those changes through fission and fusion. Drp1 and its receptors drive fission, whereas Mfn1/2 and OPA1 mediate fusion of the outer and inner membranes (Giacomello et al., 2020; Tábara et al., 2025). That model accounts well for network fragmentation, complementation between organelles, exchange of genetic material, and cell-fate control. It is less suited to rapid, periodic shape changes that leave the membrane topologically intact.

Beaded mitochondria were already being described in the early twentieth century. Lewis and Lewis reported a beads-on-a-string appearance in 1915 (Lewis and Lewis, 1915). Decades later, Webster observed focal mitochondrial accumulation and beaded remodeling in axons during Wallerian degeneration (Webster, 1962). Similar appearances have since been noted in Alzheimer’s disease, ischemia–reperfusion, aging neurons, calcium overload, and experiments involving osmotic perturbation (Zhang et al., 2016; D’Alessandro et al., 2025). Without a common physical account, however, authors generally interpreted them as consequences of injury, intermediates before fission, or descriptive features of axonal pathology. Pearling was not yet recognized as a mode of mitochondrial dynamics in its own right.

Recent work has recast these observations. Sturm et al. described mitochondrial pearling as a Rayleigh–Plateau-type instability set by membrane tension and bending elasticity (Sturm et al., 2025). Live-cell experiments by Landoni et al. then showed frequent, reversible pearling under physiological conditions and linked it to regular mtDNA-nucleoid spacing (Landoni et al., 2026). A contemporaneous perspective placed pearling alongside fission and fusion as another dimension of mitochondrial dynamics (van den Ameele and Prudent, 2026). Related instabilities in the axonal plasma membrane, platelet-producing extensions, and other membrane tubes broaden the physical context (Griswold et al., 2025; Léon et al., 2025). Pearling has already entered the vocabulary of mitochondrial dynamics. In this Review, we separate direct observations from neighboring mechanisms and more speculative links, then ask which disease models can discriminate among them.

Across the evidence reviewed here, morphology, membrane topology, and pathway attribution must be evaluated separately. Periodic beading is not synonymous with fragmentation or swelling; continuity of an imaged matrix or outer-boundary contour does not by itself prove that both mitochondrial membranes remain continuous; and a pathway that shares Ca2+ regulation with pearling is not necessarily mediated by pearling. Box 1 defines the required dynamic observations and quantitative readouts.

Box 1. Criteria for distinguishing mitochondrial pearling from similar morphologies.

A beaded, granular, or focally enlarged mitochondrion can arise through several processes that look alike in a still image. Reliable classification requires membrane-continuity data, scission status, the order of events, reversibility, and quantitative kinetics. Morphology alone is insufficient.

• Pearling/mitochondrial pearling: A tubular mitochondrion develops periodic or semiperiodic swellings along its long axis while the imaged contour remains continuous. Reversal to a tubule is common and strengthens identification, but reversibility is not part of the definition because an event may undergo downstream scission or clearance before reversal can be observed. Conventional matrix or outer-boundary imaging does not establish simultaneous continuity of the outer mitochondrial membrane (OMM) and inner mitochondrial membrane (IMM); these topological features should be reported separately.

• λ_p, τ_p, and un-pearling rate: λ_p denotes the characteristic pearling wavelength or interpearl spacing; τ_p denotes the duration of an individual pearling event; and the un-pearling rate denotes the rate at which a beaded structure returns to a tubular morphology after Ca²⁺ withdrawal or relief of stress.

• Fission: Fission ends with topological scission of the mitochondrial membranes, producing physically separate fragments. Drp1, endoplasmic reticulum (ER) contact sites, actin, and associated receptors may participate. A short tubule or fragment records only the endpoint. It does not reveal how it formed and, on its own, is not evidence of pearling.

• Fragmentation: This term describes a network outcome in which long mitochondria are replaced by shorter fragments. Increased fission, impaired fusion, or postinjury breakage can all produce it. A fragmented network, by itself, does not demonstrate pearling.

• Swelling: Expansion of matrix volume enlarges a mitochondrion and is often accompanied by osmotic imbalance, mitochondrial permeability transition pore (mPTP) opening, cristae disruption, or loss of membrane potential. A focal swelling may resemble one pearl. Without a periodic spatial pattern, continuity/topology information, and temporal evolution, it should be classified as swelling or indeterminate morphology; failure to reverse alone does not distinguish swelling from pearling that has progressed to another fate.

• Mitochondria-on-a-string (MOAS): A morphological description of beads-on-a-string mitochondria in disease or stress settings. MOAS can motivate further analysis of pearling dynamics, but static morphology alone does not establish that pearling occurred, that reversal failed, or that individual pearls followed an abnormal fate.

• Candidate disease-associated sustained pearling phenotype (provisional): a pearling event dynamically confirmed in a disease or stress model that, within the same study, retains the imaged contour and shows longer duration (τ_p) or delayed/failed reversal than controls matched for cell type, stimulus, and observation window. No universal τ_p threshold is established. OMM and IMM continuity, swelling, rupture, fragmentation, and terminal depolarization must be assessed separately.

1.1. Literature search and evidence-assignment framework

Search strategy. This narrative Review was updated by searching PubMed from database inception to 12 August 2026 in Title/Abstract fields using three concept blocks: (1) mitochondria* AND (pearling OR beading OR “beads-on-a-string” OR “beads on a string” OR “mitochondria-on-a-string”); (2) mitochondria* AND (pearling OR constriction) AND (calcium OR cristae OR nucleoid* OR mitophagy OR transfer); and (3) “pearling instability” AND (membrane OR vesicle OR tubule). Crossref was used to verify DOI and bibliographic metadata, and reference lists of key direct studies were searched manually.

Eligibility and evidence assignment. English-language, peer-reviewed primary studies were prioritized. We included time-resolved mitochondrial imaging studies that could resolve event onset, periodicity, continuity, duration, or reversal; static pearling-like observations when explicitly labeled as morphological evidence; mechanistic studies of Ca²⁺, cristae, nucleoids, fission, mitophagy, or transfer when they informed a testable link; and non-mitochondrial membrane-tube studies only for the underlying physical analogy. We excluded generic fragmentation or swelling without periodic beading or topological information, non-peer-reviewed abstracts, and non-mitochondrial pearl-like structures from biological claims. Reviews and perspectives were used for context rather than as substitutes for primary evidence.

Evidence level was assigned to each claim, not globally to each paper, according to directness of observation and temporal/causal proximity to pearling. Level A required dynamic capture of pearling itself; Level B comprised static or temporally undersampled pearling-like morphology; Level C comprised adjacent mechanisms without direct capture of a pearling event; and Level D comprised extrapolation or prediction. Because the literature is heterogeneous and the purpose is conceptual synthesis rather than effect-size estimation, no meta-analysis or formal risk-of-bias score was performed.

These definitions distinguish states by dynamics and topology rather than single-frame appearance. The table below lists the minimum observation required for each state, features that are insufficient on their own, and recommended measurements. Reversibility is a measurable outcome of pearling, not an absolute definitional requirement.

State	Minimum evidence	Insufficient on its own	Key measurements
Physiological pearling	Time-resolved periodic or semiperiodic beading with continuity of the imaged contour during the event; reversal supports identification but is not mandatory if downstream scission or clearance intervenes	A single frame; or a matrix-only/outer-boundary contour used as proof that both OMM and IMM remain continuous	λp; τp; un-pearling rate; membrane-potential recovery; OMM and IMM continuity reported separately
Fission	Direct observation of topological membrane scission or complete separation into two fragments	Short tubular or punctate mitochondria alone	Scission event; Drp1/Mff/Fis1; ER contact site
Fragmentation	Increased fragmentation at the network scale	No demonstration of periodic interpearl spacing or reversal	Network length; branching; fragment count
Swelling	Matrix expansion; loss of membrane potential; evidence of mPTP opening or osmotic change	Rounded mitochondria or focal enlargement alone	Matrix volume; Δψm; mPTP; cristae disruption
Candidate disease-associated sustained pearling phenotype (provisional)	Within-study confirmation of pearling and continuity plus longer τp or delayed/failed reversal relative to matched controls; no universal duration threshold is established	Static MOAS alone; an abnormal adjacent pathway; a single endpoint; or comparison with unmatched controls	τp; recovery half-time; λp; individual-pearl fate; separate OMM/IMM continuity; exclusion of swelling, rupture, and fragmentation

The A–D hierarchy is applied at the claim level. Level A evidence directly records pearling onset and at least the imaged contour continuity, with wavelength (λ_p), duration (τ_p), or reversal when available. Level B evidence is static or temporally undersampled MOAS/beads-on-a-string morphology. Level C evidence concerns an adjacent mechanism—such as Ca²⁺ handling, cristae remodeling, mitophagy, Miro1-dependent transport, or lipid homeostasis—without direct capture of pearling. Level D is an extrapolation that requires both continuous imaging and causal perturbation. Only Level A can support claims about pearling kinetics; Levels B–D are described as morphological clues, adjacent mechanisms, or hypotheses, respectively.

2. Biophysical basis of mitochondrial pearling

Rayleigh–Plateau instability occurs when surface tension drives a sufficiently long-wavelength disturbance in a liquid column. Classically, perturbations longer than the cylinder circumference grow and eventually divide the column into regularly spaced droplets (Plateau, 1873; Rayleigh, 1878; Bar-Ziv and Moses, 1994; Nelson et al., 1995). A membrane tube is governed by a related but richer energy balance that includes membrane tension, bending rigidity, and tube radius. Work by Bar-Ziv, Moses, Nelson, Seifert, and colleagues showed experimentally and theoretically that lipid-bilayer tubes can settle into a stable pearled configuration (Bar-Ziv and Moses, 1994; Nelson et al., 1995; Goldstein et al., 1996; Bar-Ziv et al., 1998).

A mitochondrion adds several layers of mechanics to this membrane-tube problem. Its outer and inner membranes differ in composition and organization. Cardiolipin, respiratory-chain complexes, adenosine triphosphate (ATP) synthase, and genome-anchoring sites make the effective bending rigidity of the inner membrane unlike that of the outer membrane. Cristae also bear mechanical load; they are internal elastic supports rather than passive folds. The matrix contributes a further variable because it contains Ca²⁺, Mg²⁺, H⁺, phosphate, proteins, and mtDNA nucleoids. Ion movement can therefore translate an osmotic change into a rapid change in membrane tension.

The dimensionless ratio Σ = σR²/κ_eff is used here as a heuristic parameter adapted from single-membrane tube models, not as a quantitatively validated constitutive model of mitochondria. In this notation, σ is membrane tension, R is tube radius, and κ_eff is a coarse-grained parameter that summarizes bending resistance, cristae support, and membrane coupling. A critical value has not been measured in living mitochondria, and no unique κ_eff can yet be assigned to the coupled OMM–IMM system. Double-membrane coupling, area–volume constraints, transmembrane pressure differences, spatially varying cristae, and external forces can all shift the instability threshold. The ratio therefore organizes qualitative predictions—higher tensile load or radius favors instability, whereas stronger elastic support opposes it—but should not be used to infer a threshold, bead size, or wavelength quantitatively. Ion flux, lipid remodeling, and cytoskeletal force provide possible routes into this regime (Griswold et al., 2025; Sturm et al., 2025), and their contributions may overlap.

Topological scission separates fission from pearling (Figure 1). Fission ends in physical separation, whereas pearling first produces multiple swellings and necks along the mitochondrial contour. Electron-microscopy studies of canonical fission provide useful structural context: platinum-replica EM shows actin arrays at mitochondrial constrictions, and cryo-electron tomography can resolve mitochondrial membrane morphology (Friedman et al., 2011; Korobova et al., 2013; Cho et al., 2017; Chakrabarti et al., 2018; Yang and Svitkina, 2019; Mageswaran et al., 2023; Sturm et al., 2025; Landoni et al., 2026). Although EM provides high structural resolution, the cited studies did not capture time-resolved pearling, plausibly because such events are dynamic and may be sparse. Cryo-ET nevertheless remains valuable for resolving membrane morphology during captured events, including OMM and IMM continuity, cristae organization, and the distribution of membrane-associated macromolecules. Complementary live-cell imaging is required to distinguish an elongated continuous neck from completed scission and to establish event timing; a matrix or OMM contour alone should not be treated as proof of IMM continuity. Pearling is not simply incomplete fission; the relationship between pearling, neck maturation, and any later scission is an experimental question.

FIGURE 1.

Mitochondrial pearling: biophysics and distinction from fission. (A) Rayleigh–Plateau instability provides a qualitative physical analogy for pearling of membrane tubes. Tension (σ), bending resistance (κ), tube radius (R), and wavelength (λ) interact to permit periodic constrictions; the mitochondrial double membrane and cristae add unmodeled complexity. (B) Fission ends in topological scission, whereas pearling initially preserves continuity of the imaged contour and is often reversible if no downstream conversion intervenes. The schematic does not establish simultaneous OMM and IMM continuity and is not to scale.

Cristae architecture is expected to influence the pearling threshold, but its effect should be framed as a mechanical hypothesis rather than a measured stiffness law. OPA1 and the mitochondrial contact site and cristae organizing system (MICOS) maintain inner-membrane organization; their disruption is associated with altered cristae, nucleoid organization, and increased susceptibility to beaded morphology (Frezza et al., 2006; Cogliati et al., 2013; Anand et al., 2014; Fry et al., 2024; Sturm et al., 2025). These observations are consistent with reduced or spatially heterogeneous effective support, but they do not yield a calibrated κ_eff or prove that cristae disruption alone causes pearling.

Cristae are distributed nonuniformly along the mitochondrial long axis, so κ_eff is unlikely to be spatially uniform. Local differences in crista density, orientation, and membrane contacts could shift the preferred neck positions and produce deviations in pearl size or spacing. Coupling between the OMM and IMM could also increase the effective resistance to bending relative to either membrane alone, but neither the magnitude of this increase nor a universal coupling factor is known. Measurements of λ_p and τ_p should therefore be analyzed together with local crista architecture rather than interpreted through a uniform-cylinder model.

3. Ca²⁺ as an important trigger of mitochondrial pearling

Ca²⁺ connects mitochondrial morphology, metabolism, and cell fate. Mitochondrial Ca²⁺ handling combines ER–mitochondrial contacts with transport through voltage-dependent anion channels (VDACs), the mitochondrial calcium uniporter (MCU) complex, and the mitochondrial Na⁺/Ca²⁺/Li⁺ exchanger (NCLX). Bulk cytosolic Ca²⁺ is about 100 nM at rest, but local concentrations can rise sharply during receptor activation, synaptic firing, muscle contraction, or stress. At mitochondria-associated membranes (MAMs), the inositol 1,4,5-trisphosphate receptor (IP3R)–glucose-regulated protein 75 (GRP75)–voltage-dependent anion channel 1 (VDAC1)–MCU axis exposes mitochondria to high-Ca²⁺ signals released from the ER across nanometer-scale gaps (Csordás et al., 2010; Rizzuto et al., 2012; Giorgi et al., 2015). Once Ca²⁺ reaches the intermembrane space, MCU/mitochondrial calcium uniporter b (MCUb) conducts it into the matrix, with mitochondrial calcium uptake proteins 1/2/3 (MICU1/2/3) and essential MCU regulator (EMRE) regulating the complex (Baughman et al., 2011; De Stefani et al., 2011; Patron et al., 2014) (Figure 2).

FIGURE 2.

Evidence-stratified framework connecting mitochondrial pearling with Ca²⁺ handling, adjacent pathways, and hypothetical downstream fates. (A) Direct evidence. ER-to-mitochondria Ca²⁺ transfer at mitochondria–ER contact sites (MERCs) is illustrated through the IP3R–GRP75–VDAC1 complex, followed by matrix uptake through MCU. Direct time-resolved studies support rapid, reversible pearling after mitochondrial Ca²⁺ loading. Ca²⁺-independent mechanical forces could induce pearling but their nature seems unknown as of now. (B) Adjacent mechanisms. PINK1/Parkin–LC3 quality control, Miro1-dependent mitochondrial transport, and excessive-Ca²⁺ injury involving mPTP opening, swelling, and membrane rupture are shown as established or mechanistically adjacent processes. Their shared regulation by Ca²⁺ or mitochondrial stress does not establish mediation by pearling. (C) Hypotheses. A continuous pearled tubule could become a candidate mitophagy substrate or a unit for TNT or large-carrier transfer only after neck scission. Such conversion would require preserved double-membrane integrity; OMM and IMM continuity, neck ultrastructure, and post-scission integrity remain to be validated separately. No complete sequence from pearling through scission to mitophagy or carrier-mediated transfer has been recorded. Solid green denotes direct evidence, solid gray denotes adjacent mechanisms, and dashed red denotes hypothetical relationships. The schematic is conceptual and not quantitative. Abbreviations: ER, endoplasmic reticulum; GRP75, glucose-regulated protein 75; IMM, inner mitochondrial membrane; IP3R, inositol 1,4,5-trisphosphate receptor; LC3, microtubule-associated protein 1 light chain 3; MCU, mitochondrial calcium uniporter; MERCs, mitochondria–ER contact sites; Miro1, mitochondrial Rho GTPase 1; mPTP, mitochondrial permeability transition pore; OMM, outer mitochondrial membrane; PINK1, PTEN-induced kinase 1; TNT, tunneling nanotube; VDAC1, voltage-dependent anion channel 1.

A Ca²⁺ signal reaches the pearling machinery through a sequence of physical changes. After cytosolic Ca²⁺ rises, MAM microdomains and the MCU complex promote Ca²⁺ entry into the matrix. The accompanying influx of counterions and water increases matrix osmotic pressure and membrane tension, eventually driving Σ above the pearling threshold. Rapid light-sheet imaging by Landoni et al. showed that Ca²⁺ influx can induce pearling within seconds and that the process can occur independently of the canonical fission machinery (Landoni et al., 2026). This observation provides a more direct physical explanation for many previous reports of mitochondrial constriction during calcium overload.

Ca²⁺-associated pearling should be separated from Ca²⁺-associated injury. Matrix Ca²⁺ entry, counterion movement, and water influx provide a plausible route to a rapid increase in osmotic pressure and membrane tension, and direct imaging supports a class of reversible events after Ca²⁺ influx. By contrast, sustained high-conductance mPTP opening can produce swelling, cristae disorganization, IMM fragmentation, or OMM rupture (Wolf et al., 2017; Bernardi et al., 2022). Such injury is not evidence of sustained pearling unless time-resolved imaging first verifies the pearling pattern and continuity and then shows prolonged τ_p or delayed reversal relative to matched controls. Mechanically induced events without a detectable Ca²⁺ increase further show that Ca²⁺ is an important trigger, not a universal requirement.

Recovery from pearling begins with removal of matrix Ca²⁺ and restoration of membrane homeostasis. NCLX extrudes matrix Ca²⁺, although direct evidence that it relaxes membrane tension or speeds un-pearling is still lacking (Palty et al., 2010; Kostic et al., 2015). Transient, low-conductance openings of mPTP could relieve osmotic pressure, whereas persistent high-conductance opening marks irreversible injury (Basso et al., 2005; Bernardi et al., 2022). Return to a tubular network will also depend on cardiolipin maintenance, clearance of lipid peroxides, the balance of long and short OPA1 isoforms, and repair of cristae.

Ca²⁺ also regulates pathways adjacent to pearling. Ca²⁺ oscillations can initiate PINK1–Parkin-associated quality control (Yu et al., 2021), and the EF-hand domains of Miro1 contribute to Ca²⁺-sensitive mitochondrial transport (Ahmad et al., 2014). These findings demonstrate shared upstream Ca²⁺ regulation; they do not show that pearling mediates mitophagy or mitochondrial transfer. Figure 2 therefore depicts these pathways as adjacent evidence rather than as branches of a pearling-controlled fate cascade.

The take-home message is therefore threefold: Ca²⁺ influx can trigger a class of rapid pearling events, and its subsequent withdrawal can accompany their reversal; Ca²⁺ independently regulates downstream quality-control and transport pathways in which a role for pearling remains untested; and Ca²⁺-independent mechanical force can also induce pearling. A single Ca²⁺ dose-fate hierarchy is not established.

4. Candidate sustained pearling as a provisional operational framework

Candidate disease-associated sustained pearling is a provisional operational phenotype, not a distinct disease entity or a synonym for MOAS, fragmentation, or swelling. No validated absolute τ_p cutoff exists. Classification therefore requires a within-study comparison showing that a dynamically confirmed pearling event lasts longer or reverses more slowly than in controls matched for cell type, stimulus, and imaging window, while alternative outcomes are excluded. To our knowledge, no neurological disease model currently satisfies this full set of criteria; current disease associations should be treated as candidates for validation.

5. Direct functional evidence: pearling and mtDNA nucleoid redistribution

Mammalian cells contain numerous copies of mtDNA. These genomes are not freely dispersed but are packaged into nucleoid structures approximately 100 nm in size. Super-resolution imaging shows that most nucleoids contain one or a few mtDNA copies and are closely associated with the inner membrane and cristae (Garrido et al., 2003; Brown et al., 2011; Kukat et al., 2011; 2015; Bogenhagen, 2012). In healthy cells, nucleoids often exhibit a semiregular distribution, with neighboring nucleoids commonly separated by 1–2 μm. Random diffusion alone cannot readily explain this spatial precision.

Before the work of Landoni et al., three main models were used to explain nucleoid distribution: coupling of mtDNA synthesis and fission at ER–mitochondrial contact sites (Lewis et al., 2016), transport of nucleoids along the inner membrane regulated by ATPase family AAA domain-containing protein 3A (ATAD3A) (Zhao et al., 2019; Ishihara et al., 2022), and ER-mediated dynamic stretching of mitochondria that promotes long-range nucleoid movement (Qin et al., 2020). These models account for transport, replication, and partitioning but do not fully explain why steady-state spacing has a relatively stable characteristic length. Pearling offers a geometric explanation for this feature.

Landoni et al. showed that the characteristic wavelength of physiological pearling is approximately 1 μm, closely matching the spacing between nucleoids (Landoni et al., 2026). A single pearling event can displace nucleoids within a tubule toward the centers of adjacent pearls. After un-pearling restores a tubular morphology, the nucleoids can remain at their new positions. Repeated brief pearling events could therefore maintain nucleoid distribution in a dynamic steady state. In other words, nucleoid spacing may not be a purely static structural property but a pattern continuously reset by Ca²⁺–pearling cycles.

Cristae and OPA1 status connect pearling mechanics to mtDNA organization, but the relationship is not monotonic. Reduced OPA1 or MICOS function can weaken or spatially disorder inner-membrane support and is associated with altered pearling and nucleoid spacing (Fry et al., 2024; Sturm et al., 2025; Landoni et al., 2026). The same crista disruption that lowers a local mechanical threshold may also remove the relatively uniform support and membrane landmarks needed for regular bead spacing and reproducible nucleoid repositioning. More frequent or easier beading therefore does not predict more organized nucleoids. Patient-derived OPA1-associated autosomal dominant optic atrophy cells show disrupted cristae and nucleoid organization (Macuada et al., 2024), but dynamic τ_p, local crista architecture, and un-pearling still need to be measured together.

Direct evidence currently ends with pearling-associated nucleoid redistribution. Proposals that a pearl subsequently isolates a damaged nucleoid for focal fission, nucleoid-phagy, or mitophagy remain hypotheses (Liu et al., 2024b; Newman et al., 2024). Existing nucleoid-clearance studies concern cytosol-exposed mtDNA–TFAM complexes or other damage contexts, not recognition of a nucleoid enclosed within an intact pearl. Likewise, cGAS–STING activation requires mtDNA access to the cytosol; no continuous sequence from pearling to membrane breach, mtDNA release, and cGAS–STING activation has been shown. These downstream links should be tested rather than inferred from spatial compatibility.

6. Pearling before scission-dependent downstream processes: evidence and predictions

Pearling generates a patterned but still connected tubule. Mitophagic engulfment or intercellular transfer of an individual pearl would require at least one additional event: coordinated topological scission that preserves the relevant membrane architecture. No study has followed the same mitochondrion through a complete pearling→neck scission→mitophagy or pearling→neck scission→carrier loading→recipient uptake sequence. The following subsections therefore separate established downstream biology from pearling-specific predictions.

Established mitophagy studies show that depolarized mitochondrial material can recruit PINK1/Parkin or receptor-mediated machinery and be engulfed after network remodeling (Narendra et al., 2008; 2010; Twig et al., 2008; Geisler et al., 2010; Lazarou et al., 2015; Pickrell and Youle, 2015; McWilliams et al., 2016). Reports that mitochondrial division during mitophagy can occur without Drp1 (Yamashita et al., 2016; Burman et al., 2017) broaden the possible scission mechanisms but do not identify pearling as an upstream step. Size compatibility between pearls and autophagic substrates is therefore insufficient evidence of a causal interface.

Structural studies provide contextual, not pearling-specific, evidence. In situ cryo-electron tomography has visualized spherical mitochondrial fragments and phagophore engulfment after depolarization (Rose et al., 2025), but those snapshots did not establish a preceding pearling event or continuity at pearl necks. The decisive experiment remains continuous live imaging followed by correlative ultrastructure on the same mitochondrial segment.

Hypothesis and prediction: if pearling contributes to mitophagy, it should reproducibly precede local loss of membrane potential, coordinated OMM/IMM scission, recruitment of PINK1/Parkin or receptor-mediated machinery, LC3-positive engulfment, and lysosomal delivery on the same mitochondrial segment. Selective suppression or enhancement of pearling should then change this sequence without simply altering Ca²⁺ load, global fission, or mitochondrial injury. Until these criteria are met, pearling and mitophagy should be treated as potentially parallel responses.

Mitochondria-derived vesicles (MDVs) should not be placed on a simple size continuum with pearls. Canonical MDVs are generally substantially smaller than a ∼1-μm pearl and can contain one or two mitochondrial membranes, depending on cargo and route (Soubannier et al., 2012). A pearl remains part of a connected double-membrane organelle until proven otherwise. A pearling-to-MDV model would therefore require direct demonstration of neck remodeling, membrane provenance, coordinated or selective OMM/IMM scission, and vesicle trafficking; none is currently available.

6.1. Intercellular mitochondrial transfer as a lower-evidence hypothesis

Intercellular mitochondrial transfer can occur through tunneling nanotubes (TNTs), extracellular vesicles, migrasomes, or localized cell fusion (Rustom et al., 2004; Spees et al., 2006; Islam et al., 2012; Pasquier et al., 2013; Phinney et al., 2015; Hayakawa et al., 2016; Torralba et al., 2016; Dong et al., 2017; Jiao et al., 2021; Saha et al., 2022; Borcherding and Brestoff, 2023; Guan et al., 2024; Liu et al., 2024a; Palese et al., 2025). Transfer has been observed from mesenchymal stem cells to injured epithelial cells, from astrocytes to neurons, and in tumor–immune systems (Islam et al., 2012; Phinney et al., 2015; Hayakawa et al., 2016; Saha et al., 2022; Baldwin et al., 2024). These routes differ in cargo capacity, membrane architecture, and biogenesis. Their existence does not demonstrate a shared loading process or a requirement for pearling.

Pearling alone does not create an independently transportable unit because adjacent pearls remain connected. Transfer of a pearl-sized unit would require coordinated neck scission that preserves OMM and IMM integrity, followed by loading into a carrier capable of accommodating an intact mitochondrion. We do not assign ESCRT proteins as candidate executors of this step because current evidence does not place them at the IMM or demonstrate coordinated OMM/IMM scission during pearling. The mechanism and membrane topology are open questions.

No study has continuously recorded pearling, neck scission, carrier loading, and recipient uptake. A discriminating experiment would perturb pearling while preserving global mitochondrial transport and network integrity, then use continuous donor–recipient imaging and correlative ultrastructure to determine whether intact double-membrane units enter TNTs or large mitochondria-containing carriers. A temporal association alone would be insufficient; selective inhibition and rescue are required.

Miro1 is an adjacent transport regulator, not evidence of a pearling-dependent loading step. Manipulating Miro1 alters mitochondrial transfer from mesenchymal stem cells and recipient-cell rescue (Wang et al., 2011; Ahmad et al., 2014), but these studies did not record pearling. Simultaneous imaging of Miro1, donor mitochondrial motion, pearling, neck topology, and recipient uptake is needed before Miro1 can be placed in a pearling-to-transfer sequence.

7. Candidate disease-associated sustained pearling: priority models and evidentiary boundaries

Beads-on-a-string mitochondria and abnormalities in pearling-related pathways occur in neurological disease, but similar morphology does not imply a shared mechanism. To our knowledge, no neurological disease model has yet demonstrated the full within-study sequence required for a candidate sustained phenotype: dynamic pearling, continuity of the imaged contour, longer τ_p or delayed/failed reversal relative to matched controls, and exclusion of swelling, rupture, and fragmentation. Disease associations are therefore presented as priorities for testing rather than as established examples. Table 1 summarizes candidate disease contexts for mitochondrial pearling and details their mechanistic relevance, current supporting evidence, and key validation needs. Instead, it reports only the current evidence type, why each context is informative, and the minimum event-level validation. Unlisted criteria should not be interpreted as satisfied.

TABLE 1.

Candidate neurological contexts and minimum validation for sustained pearling.

Candidate context	Current evidence (not pearling confirmation)	Why the context is informative	Minimum event-level validation and boundary
Alzheimer’s disease (AD)	MOAS or beads-on-a-string morphology and Aβ/tau, MAM, and mitochondrial Ca2+ abnormalities provide adjacent morphological and mechanistic evidence	Tests whether a static disease morphology maps to a reversible or sustained dynamic event	Acquire live-cell λp, τp, un-pearling, and separate OMM/IMM continuity in matched neurons; exclude swelling, rupture, and fragmentation. MOAS alone is not confirmation (Du et al., 2008; Wang et al., 2009; Calì et al., 2012; Zhang et al., 2016; Fang et al., 2019; Calvo-Rodriguez et al., 2020)
Parkinson’s disease (PD)	PINK1/Parkin quality-control defects and α-synuclein-related MAM and Ca2+ abnormalities provide adjacent mechanistic evidence	Tests whether quality-control recruitment follows a dynamically verified pearling event	Track the same mitochondrion from pearling through scission and PINK1/Parkin or LC3 recruitment, with perturbation and rescue. Adjacent pathway abnormalities are not confirmation (Calì et al., 2012; Lazarou et al., 2015; Pickrell and Youle, 2015; Palese et al., 2025)
Huntington’s disease (HD)	Altered mPTP regulation and mitochondrial Ca2+ sensitivity provide adjacent mechanistic evidence	Tests whether an altered injury threshold delays reversal or instead causes swelling	Measure τp and un-pearling together with Δψm, matrix volume, membrane continuity, and mPTP state; distinguish pearling from osmotic expansion (Quintanilla et al., 2013; Gu et al., 2024)
Amyotrophic lateral sclerosis (ALS)	Mitochondrial dysfunction, oxidative stress, and impaired axonal transport define a vulnerable disease context but do not demonstrate pearling	Tests whether long-axon vulnerability selectively changes recovery kinetics	Use continuous axonal imaging with matched stress and prespecified event criteria. Mitochondrial vulnerability alone does not establish sustained pearling (Smith et al., 2019; Rizea et al., 2024)
CMT2A/MFN2-associated axonopathy	MFN2 mutations alter morphology, bioenergetics, and axonal transport; mitofusin activation modifies some phenotypes	Provides a genetic entry point to test transport and fusion effects on recovery without equating fusion with un-pearling	Measure λp, τp, un-pearling, and topology in patient-derived neurons, with MFN2 perturbation and rescue. This is a testing context, not a confirmed example (Franco et al., 2022; Zanfardino et al., 2023)
OPA1/ADOA and cristae pathology	OPA1-dependent crista remodeling and abnormal nucleoid distribution may alter effective mechanical support	Provides a genetic entry point for testing crista-dependent pearling thresholds	Establish temporality between crista change and event kinetics in patient cells; measure OMM/IMM continuity and use OPA1 or OMA1 perturbation and rescue (Frezza et al., 2006; Cogliati et al., 2013; Anand et al., 2014; Fry et al., 2024; Macuada et al., 2024)
Wallerian-like axonal degeneration	Early axonal mitochondrial beading and SARM1-linked NAD+/NMN dysregulation provide adjacent morphological and mechanistic evidence	Acute onset permits separation of early beading from downstream axonal disintegration	Continuously image onset, reversal, membrane topology, and axonal integrity during SARM1 perturbation. Historical beading alone is not dynamic pearling evidence (Webster, 1962 ; Osterloh et al., 2012 ; Salvadores et al., 2017 ; Figley et al., 2021 ; Ulshöfer et al., 2022)
Ischemia–reperfusion/stroke	Acute Ca2+ and reactive-oxygen-species loads, OPA1 cleavage, mPTP opening, and poststroke transfer provide adjacent mechanisms.	Tests the transition from a reversible event to swelling, rupture, or fragmentation	Image before and after stress withdrawal with λp, τp, Δψm, matrix volume, and OMM/IMM continuity; exclude late injury morphology (Hayakawa et al., 2016; Bernardi et al., 2022; Fry et al., 2024)
Aging	Age-related drift in Ca2+ homeostasis, OPA1, mitophagy, Miro1/TNT function, and lipid homeostasis defines a broad context only	Tests whether cumulative homeostatic decline selectively delays reversal.	Compare age-stratified matched cohorts using the same stimulus, imaging window, and event annotation; do not treat aging as a single trigger (Sun et al., 2015 ; Sun et al., 2016 ; López-Otín et al., 2023)

The A–D evidence hierarchy delimits the claims that can be made. Level A can support event-level statements about pearling; Levels B and C identify morphological or mechanistic entry points; Level D propositions offer testable predictions. The central disease-focused question is not whether multiple disorders share one pearling mechanism, but whether any selected model first satisfies the operational criteria and then yields a causal chain to dysfunction.

The highest-priority models should offer a clear genetic or acute-injury entry point and permit single-mitochondrion time-series imaging. First, OPA1/ADOA provides a direct mechanistic entry point linking cristae support, κ_eff, nucleoid distribution, and the pearling threshold (Frezza et al., 2006; Cogliati et al., 2013; Anand et al., 2014; Fry et al., 2024; Macuada et al., 2024). If patient-derived cells or induced pluripotent stem cell (iPSC)-derived neurons show longer τ_p, altered λ_p, and failed reversal, and these phenotypes can be corrected by manipulating the OPA1–OMA1 zinc metallopeptidase (OMA1) axis or stabilizing cristae, the model would gain strong support. Second, Charcot–Marie–Tooth disease type 2A (CMT2A)/mitofusin 2 (MFN2) is suitable for testing whether network morphology and axonal transport are coupled to recovery by un-pearling. Fusion and un-pearling are topologically distinct processes, so MFN2 deficiency cannot be assumed to cause failed reversal. MFN2 mutations affect mitochondrial morphology, bioenergetics, and axonal transport, and small-molecule mitofusin activators provide tools for causal intervention (Franco et al., 2022; Zanfardino et al., 2023). Third, Wallerian-like axonal degeneration and ischemia–reperfusion models provide an acute time axis that can distinguish brief pearling triggered by Ca²⁺/reactive oxygen species (ROS) from failed reversal, swelling, and downstream cell death (Webster, 1962; Osterloh et al., 2012; Salvadores et al., 2017; Figley et al., 2021; Bernardi et al., 2022; Ulshöfer et al., 2022; Fry et al., 2024).

Alzheimer’s disease (AD), Parkinson’s disease (PD), Huntington’s disease (HD), amyotrophic lateral sclerosis (ALS), and aging remain important, but current evidence makes them more appropriate as adjacent mechanistic contexts than as established examples of disease-associated sustained pearling. In AD, amyloid-β, tau, altered MAM signaling, and mitochondrial Ca²⁺ overload all disrupt mitochondrial homeostasis. MOAS is a useful morphological clue, but classification as sustained pearling still requires dynamic evidence of membrane continuity, prolonged τ_p, and failed reversal (Du et al., 2008; Wang et al., 2009; Calì et al., 2012; Zhang et al., 2016; Fang et al., 2019; Calvo-Rodriguez et al., 2020). PD offers a different mechanistic context, centered on PINK1/Parkin dysfunction, α-synuclein, and intercellular spread. Per-pearl imaging is needed to determine whether PINK1/Parkin, LC3, or Miro1 influences the fate of an individual pearl (Calì et al., 2012; Lazarou et al., 2015; Pickrell and Youle, 2015; Palese et al., 2025). In HD, increased mPTP susceptibility could prolong osmotic and membrane stress after a pearling trigger; in ALS, disrupted axonal transport and organelle quality control could impede restoration or clearance of affected mitochondria; and during aging, cumulative redox, proteostatic, and bioenergetic decline could weaken the homeostatic recovery required for un-pearling. These are distinct upstream routes by which un-pearling capacity might be reduced, but each remains a testable hypothesis until event-resolved pearling is measured in the relevant model (Quintanilla et al., 2013; Sun et al., 2015; Sun et al., 2016; Smith et al., 2019; López-Otín et al., 2023; Gu et al., 2024; Rizea et al., 2024).

From a therapeutic perspective, the objective should not be to abolish pearling itself. Physiological pearling may support mtDNA distribution and focal quality control; abnormal duration, frequency, and fate divergence are the more relevant targets. At present, mPTP modulators, mitophagy enhancers, regulators of the ubiquitin-specific peptidase 30 (USP30)/PINK1 axis, allosteric MFN2 activators, and OPA1/OMA1 modulators are therefore better regarded as tools for testing disease-associated sustained pearling readouts or as candidate interventions, not as established treatments that specifically target sustained pearling (Andreux et al., 2019; Fang et al., 2019; Franco et al., 2022; Antico et al., 2025; Kazi et al., 2025). The working model can advance from a correlated phenotype to a classifiable and actionable pathological mechanism only if selective alteration of pearling—while controlling for concurrent changes in Ca²⁺, membrane potential, swelling, and fission—demonstrates causal relationships of τ_p, λ_p, and un-pearling rate with membrane-potential stability, respiration, the precision of mtDNA distribution, neurite integrity, and cell survival, and if rescuing pearling dynamics restores functional outcomes.

8. Measurement challenges, validation strategies, and scope of mitochondrial pearling

The next step in pearling research is not to accumulate further descriptions of static morphology but to establish reproducible dynamic readouts and use them to test links between pearling and nucleoid distribution, quality control, intercellular transfer, and disease phenotypes. Three connected questions must be addressed: How can individual pearling events be captured reliably? Which mechanistic links have direct support? Under what conditions should current interpretations be restricted or rejected?

8.1. Methodological bottlenecks and measurement framework

The first major limitation is the mismatch between the spatial scale of pearl necks and nucleoids and the speed of onset and reversal. A tiered workflow is appropriate: rapid multi-channel light-sheet imaging for morphology, matrix Ca²⁺, and membrane potential; targeted super-resolution imaging for necks and nucleoids; and event-triggered correlative cryo-CLEM, cryo-FIB-SEM, or cryo-ET for ultrastructure. Matrix, OMM, and IMM reporters should be interpreted separately because a continuous matrix or OMM contour does not prove that both membranes are continuous. Event-triggered cryo-ET can complement and validate these fluorescence-based continuity assignments by directly resolving the OMM, IMM, and cristae at the captured event site. Correlative localization should link the tiers, and phototoxicity, photobleaching, probe overexpression, and missed-event rates should be prespecified quality-control measures (Chen et al., 2014; Mahamid et al., 2016; Schermelleh et al., 2019; Gwosch et al., 2020).

The second bottleneck is the lack of direct measurements of physical parameters. In addition to interpearl spacing, duration, frequency, and un-pearling rate, studies should separately record the osmotic pressure difference ΔΠ and hydrostatic pressure difference ΔP and estimate the effective mechanical properties of the mitochondrial inner membrane. The fluorescence lifetime of Flipper-type probes can provide an indirect readout of membrane-lipid packing and changes in tension, but their use in the inner mitochondrial membrane requires calibration for subcompartment targeting, lipid composition, and curvature dependence. Combining these measurements with matrix Ca²⁺, membrane potential, and artificial-intelligence-assisted morphological tracking could produce a state map for individual mitochondria. Open standards for event annotation should also record onset, peak, end of reversal, number of pearls, neck radius, membrane continuity, positions of neighboring nucleoids, and the fate of each pearl (Colom et al., 2018; Carpenter et al., 2006; Caicedo et al., 2017; Stringer et al., 2021).

A third question is whether un-pearling is driven solely by restoration of ionic and mechanical parameters or also requires active network regulation. OPA1/OMA1, MICOS, ATAD3A, cardiolipin metabolism, MCU/MICU, NCLX, and mPTP are candidates, but each requires independent mechanistic rationale and direct validation. CRISPR screens could use post-withdrawal τ_p, the decay rate of a normalized pearling index, the proportion of failed reversals, and membrane continuity as joint primary readouts. Perturbations that produce a reproducible pearling phenotype should then undergo secondary measurements of membrane potential, nucleoid distribution, and single-pearl fate. USP30 is a tractable candidate because its deubiquitinase activity opposes PINK1/Parkin-dependent mitophagy, allowing tests of whether quality-control signaling alters pearling reversal independently of downstream clearance. PHB1/2 should be tested separately as inner-membrane scaffolds that may influence cristae organization and effective membrane support; OPA1 processing and cristae architecture should be measured alongside pearling dynamics. By contrast, the proposed SLP2-cardiolipin microdomain remains a hypothesis because direct localization and perturbation data linking it to pearl-neck formation or reversal are currently lacking. The proposed SLP2–cardiolipin microdomain, in contrast, still lacks direct supporting evidence (Kazi et al., 2025; Rose et al., 2025).

8.2. Evidence base and validation sequence for core propositions on mitochondrial pearling mechanisms

Table 2 places core propositions regarding mitochondrial pearling mechanisms alongside their current claim status, unresolved gaps, and the experiments most likely to distinguish among competing explanations. Candidate sustained pearling should be tested first in models with a clear causal entry point, compatibility with continuous imaging, and opportunities for mechanistic rescue.

TABLE 2.

Evidence base, claim status, and validation sequence for the core propositions.

Core proposition	Current evidence level	Principal gap	Next experiment
Ca2+ can trigger rapid, reversible mitochondrial pearling	A (direct dynamic evidence in selected systems)	Generality across cell types, Ca2+ loads, and cristae states remains unresolved	Simultaneously measure matrix Ca2+, membrane potential, λp, τp, and un-pearling rate while manipulating MCU, NCLX, and mPTP
Rapid pearling can redistribute mtDNA nucleoids and influence spacing in directly imaged models	A (model-specific direct evidence)	Existing evidence is concentrated in specific models; validation is needed in neurons and patient-derived cells	Track retention of nucleoid positions after individual pearling events during OPA1, ATAD3A, or MICOS manipulation and in patient-derived cells
Pearling may create geometry that precedes scission-dependent mitophagy	C for adjacent mitophagy mechanisms; D for the pearling link	No same-mitochondrion sequence from pearling through scission, PINK1/Parkin or receptor recruitment, engulfment, and lysosomal delivery has been demonstrated	Continuously image morphology, OMM/IMM continuity, membrane potential, scission, PINK1/Parkin or receptor recruitment, LC3, and lysosomal delivery on the same mitochondrial segment
After coordinated OMM/IMM scission, a pearl could hypothetically become a unit for TNT or large-carrier transfer	C for adjacent transfer mechanisms; D for the pearling link	No continuous pearling→neck scission→carrier loading→recipient uptake sequence has been recorded; scale matching is insufficient	In MSC–epithelial, astrocyte–neuron, and tumor–immune cocultures, continuously track pearl-neck scission, double-membrane integrity, entry of units into TNTs or large mitochondria-containing vesicles, and arrival in recipient cells
A candidate disease-associated sustained pearling phenotype may exist in selected neurological models	B/C; no disease model currently satisfies the full dynamic criteria	No validated absolute τp threshold exists, and most disease evidence is static or mechanistically adjacent	Use within-study matched controls and prespecified event-level criteria before testing causality or generalizing across diseases

OPA1/ADOA, CMT2A, and acute axonal injury or ischemia can serve as priority models for determining whether longer pearling duration (τ_p) or failed reversal is reproducible and whether these changes can be corrected by targeted interventions. Only after reproducible event-level evidence is obtained should the scope expand to complex neurodegenerative diseases and population genetics. Population-scale associations can nominate disease settings worth studying, but they cannot verify a pearling event or establish its causal role.

Pearling is also found outside mitochondria. Related transitions occur in the axonal plasma membrane, platelet-producing extensions, endosomal tubules, and reconstituted membrane tubes (Bar-Ziv and Moses, 1994; Nelson et al., 1995; Gao et al., 2018; Griswold et al., 2025; Léon et al., 2025). Comparisons across these systems can test whether λ_p scales with tube radius, how internal supports shift the instability threshold, and whether Ca²⁺, tension, and lipid composition act as a recurring control set. Pearling across these systems, including mitochondrial pearling, therefore provides a clear interface between cell biology and soft-matter physics.

8.3. Boundaries of applicability for disease-associated sustained pearling

Whether disease-associated sustained pearling becomes an explanatory disease mechanism depends on the reproducibility of its dynamics, spatial scale, and functional consequences across independent models. The following results would directly limit the scope of the interpretation.

First, disease-associated differences are not reproducible. If preregistered, adequately powered, and appropriately matched studies fail to identify reproducible and practically meaningful between-group differences in pearling frequency, τ_p, or un-pearling, sustained pearling should not be generalized as a shared mechanism across diseases.

Second, proposed pearl-level fates are not spatially resolved. If PINK1/Parkin, LC3, Miro1, scission, or carrier loading does not reproducibly localize to specified pearls or necks before the outcome occurs, fate should not be assigned to an individual pearl. The more conservative interpretation would be that pearling and the molecular pathway are parallel, tubule-wide responses.

Third, pearling scale is not stably related to nucleoid spacing. If the relationship between characteristic wavelength (λ_p) and nucleoid spacing cannot be reproduced across cell types, cristae states, or patient-derived models, the interpretation of pearling as a geometric ruler for nucleoids should be restricted to experimental systems with direct supporting evidence.

Fourth, pearling follows irreversible injury. If imaging with sufficient temporal resolution shows that mPTP opening, mitochondrial swelling, or membrane rupture usually precedes pearling, then beads-on-a-string morphologies in acute injury are more likely manifestations of terminal damage than consequences of sustained reversible pearling or failed un-pearling.

Fifth, altering pearling does not affect downstream function. If selective manipulation of pearling, after controlling for Ca²⁺ load, membrane potential, swelling, and overall fission and fusion, does not affect nucleoid distribution, mitophagy, intercellular mitochondrial transfer, or functional outcomes, the proposed causal relationship lacks support. Its scope should then be narrowed or, if necessary, the interpretation abandoned.

9. Discussion

Current evidence identifies mitochondrial pearling as a rapid membrane transition influenced by tension, bending resistance, matrix osmotic pressure, Ca²⁺ signaling, and external force, with direct evidence for mtDNA nucleoid redistribution in selected systems. The single-membrane instability ratio is a heuristic, and simultaneous OMM/IMM topology at pearl necks remains unresolved. Continuous imaging and causal perturbation are still lacking for pearling-dependent mitophagy or intercellular transfer; these links require an intervening scission step and are presented as hypotheses. Disease-associated sustained pearling is likewise a provisional within-study phenotype with no validated absolute τ_p threshold and no currently established neurological example. Future studies should first capture pearling, membrane topology, and reversal in the same time series, then test whether selective manipulation of pearling changes nucleoid organization or a downstream scission-dependent sequence without confounding Ca²⁺ load, swelling, or global fission. This evidence-stratified framework preserves pearling as a distinct dynamic process while preventing adjacent pathways and static disease morphologies from being overinterpreted as direct consequences.

Funding Statement

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the Natural Science Foundation of Sichuan Province (Grant No. 2025ZNSFSC0674).

Author contributions

YC: Conceptualization, Writing – original draft. FT: Data curation, Writing – original draft. XL: Conceptualization, Supervision, Writing – review and editing.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. Generative artificial intelligence (DeepSeek) was used solely to translate portions of the manuscript from Chinese into English and to assist with language polishing and document formatting. The authors subsequently reviewed, edited, and verified all AI-assisted content and take full responsibility for the accuracy, integrity, and final content of the 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.