Centrosome amplification in type 2 diabetes mellitus: Putative mechanisms, clinical research implications, and investigational therapeutic perspectives

Abstract

Type 2 diabetes mellitus (T2DM) is a chronic metabolic disorder characterized by progressive pancreatic β-cell failure, systemic insulin resistance, and multiorgan complications driven by sustained hyperglycemia, lipotoxicity, oxidative stress, and chronic low-grade inflammation. Centrosome amplification (CA), defined as a numerical increase of centrosomes (typically ≥ 3 mature centrosomes or ≥ 4 centrioles per cell), is distinct from centrosome structural defects, fragmentation, or primary ciliary dysfunction. Accumulating preclinical and clinical evidence suggests a bidirectional association between T2DM and CA: The pathological microenvironment of T2DM induces CA in multiple cell types via several reported signalling cascades, including the protein kinase B-reactive oxygen species-Rho-associated coiled-coil-containing protein kinase 1/14-3-3 protein sigma axis, Wnt6-frizzled 4-β-catenin pathway, and peroxisome proliferator-activated receptor γ-spindle and kinetochore-associated protein 1 transcriptional regulatory axis; in turn, CA may contribute to β-cell dysfunction, systemic insulin resistance, and the progression of diabetic macrovascular and microvascular complications. This narrative review aims to summarize the molecular regulation of centrosome homeostasis, the inducing factors and signaling networks underlying T2DM-associated CA, and the cellular mechanisms whereby CA may participate in T2DM pathogenesis and target-organ damage. Notably, much of the mechanistic data are derived from tumour-derived cell lines, and direct in vivo evidence from key metabolic tissues remains scarce, while clinical evidence supporting CA as a biomarker is still limited. We further synthesize available evidence from in vitro studies, animal models, and clinical cohorts, discuss investigational therapeutic strategies targeting CA for T2DM and its complications, and highlight current limitations and future research directions. This work provides a conceptual framework for exploring novel cellular mechanisms of T2DM and for developing centrosome-targeted interventions that remain to be validated for disease-modifying effects.

Key Words: Centrosome; Centrosome amplification; Type 2 diabetes mellitus; Pancreatic β-cell; Insulin resistance; Diabetic complications; Biomarkers; Oxidative stress; Rho-associated protein kinase 1

Core Tip: In vitro studies indicate that the diabetic microenvironment can induce Centrosome amplification (CA) via multiple interconnected signalling cascades, including protein kinase B-reactive oxygen species(Rho-associated coiled-coil-containing protein kinase 1/14-3-3 protein sigma, Wnt6-frizzled 4, protein kinase C-p38 and peroxisome proliferator-activated receptor γ-spindle and kinetochore-associated protein 1 axes. Putative links between CA and pancreatic β-cell failure, systemic insulin resistance, as well as diabetic macro-and microvascular complications remain largely hypothetical, owing to limited direct in vivo evidence from human metabolic tissues. Peripheral-blood-mononuclear-cell CA correlates with glycated haemoglobin and complication severity within a single observational clinical cohort; these associations cannot support prognostic inference, and it remains purely an exploratory research biomarker and is not ready for clinical application. Modulation of centrosome homeostasis represents a pre-clinical investigational direction; therapeutic benefits of polo-like kinase 4 inhibitors, Rho-associated coiled-coil-containing protein kinase blockers and antidiabetic agents related to CA suppression still await formal experimental verification.

INTRODUCTION

Type 2 diabetes mellitus (T2DM) is the most prevalent chronic metabolic disorder worldwide. According to epidemiological estimates from the 10^th edition of the International Diabetes Federation Diabetes Atlas (2021)[1], 537 million adults aged 20-79 lived with diabetes globally in 2021, with prevalence projected to reach 783 million by 2045; over 90% of all diabetic cases correspond to T2DM. Persistent hyperglycaemia drives a spectrum of multisystem complications affecting the cardiovascular and cerebrovascular systems, kidneys, retina and nervous system, thereby markedly increasing disability and all-cause mortality. T2DM has therefore emerged as a major public health challenge and a substantial threat to human health[2]. The pathophysiology of T2DM is multifaceted, with insulin resistance in peripheral tissues and progressive pancreatic β-cell dysfunction representing its central pathogenic features. Glucolipotoxicity, chronic low-grade inflammation, oxidative stress and other interrelated processes collectively contribute to disease onset and progression. Nevertheless, many aspects of its underlying mechanisms remain incompletely understood, limiting the development of precise prevention strategies and targeted therapeutic interventions[3].

The centrosome is the principal microtubule-organizing centre in animal cells, consisting of a pair of orthogonally arranged centrioles embedded within a dense pericentriolar matrix. Traditionally, its canonical function has been regarded as the organization of the mitotic spindle, the coordination of faithful chromosome segregation and the preservation of genomic stability. However, recent studies have shown that the centrosome also fulfils critical roles in non-dividing cells, participating in the establishment of cell polarity, intracellular trafficking, the anchoring of signalling molecules and primary ciliogenesis. It is therefore increasingly recognized as an essential subcellular structure for maintaining cellular architecture and functional homeostasis[4]. Under physiological conditions, centrosome duplication is tightly coupled to cell-cycle progression, such that each interphase cell contains only one centrosome. When this regulatory programme is disrupted, cells may acquire two or more centrosomes, a phenomenon termed centrosome amplification (CA). Numerical CA is distinguished from centrosome structural defects, centriole fragmentation, and primary ciliary dysfunction, which represent separate classes of centrosomal abnormalities with distinct pathogenic mechanisms. Initially regarded as a defining feature of cancer cells, CA has become a major focus in oncology research, as it can drive tumorigenesis and disease progression by inducing chromosomal instability, disrupting epithelial polarity and promoting invasive cellular phenotypes.

As research into the subcellular mechanisms of metabolic disease has advanced, the link between centrosomal dysfunction and disturbances in glucose and lipid metabolism has become increasingly apparent. Several monogenic disorders involving centrosomal or basal-body structural proteins provide direct genetic evidence for a causal relationship between these processes. Patients with Alström syndrome harbour mutations in Alström syndrome protein 1, centrosome-and basal-body-associated protein (ALMS1), which encodes a protein localized to the centrosome and ciliary basal body; clinically, the disorder is characterized by obesity, severe insulin resistance, early-onset T2DM and dyslipidaemia[5,6]. Pathogenic variants in PCNT, encoding the centrosomal scaffold protein pericentrin, cause microcephalic osteodysplastic primordial dwarfism and are associated with severe insulin resistance and diabetes[7,8]. In addition, proteome of centriole 1 homologue A (POC1A) deficiency causes short stature, onychodysplasia, facial dysmorphism, and hypotrichosis (SOFT) syndrome and may be accompanied by hyperinsulinaemia, impaired glucose tolerance, insulin resistance, and hepatic steatosis[9-12]. Notably, these monogenic disorders involve congenital structural or functional defects of the centrosome-cilium apparatus, which differ from the acquired numerical CA observed in common sporadic T2DM. Evidence from these genetic conditions supports the biological plausibility of CA involvement in metabolic dysregulation, but does not directly prove that acquired CA drives T2DM pathogenesis.

In pancreatic β-cells, the centrosome and its nucleated microtubule network constitute a critical structural foundation for the regulation of insulin secretion. Glucose-stimulated insulin secretion hinges on the directed trafficking, docking and exocytic release of secretory granules from intracellular storage compartments to the plasma membrane; as the core hub for microtubule nucleation, the centrosome directly mediates the spatial orchestration of granule transport. Accumulating evidence demonstrates that pericentrin, a core pericentriolar protein, is abundantly expressed in quiescent β-cells. Beyond its canonical localization at the centrosome, pericentrin distributes along the surface of cytoplasmic insulin granules, where it modulates intracellular clustering and membrane docking of secretory granules to shape the efficacy of glucose-stimulated insulin secretion. Downregulation of pericentrin disrupts the rhythmic pattern of insulin secretion, thereby precipitating impaired glucose tolerance in mice[13]. Dynamic remodeling of the microtubule network represents a pivotal determinant of insulin secretion control. Elevated glucose enhances the accessibility of secretory granules to the plasma membrane by triggering microtubule depolymerization, whereas aberrantly increased microtubule density under diabetic conditions constrains granule trafficking and exacerbates β-cell secretory dysfunction[14]. Collectively, these findings indicate that structural and functional aberrations of the centrosome may impair the secretory competence of pancreatic β-cells, representing a putative, understudied mechanism underlying the progression of T2DM.

Emerging clinical and preclinical investigations have reported altered centrosome homeostasis under T2DM-related metabolic stress. Wang et al[15] provided the first clinical evidence that the frequency of CA in peripheral blood mononuclear cells (PBMCs) is markedly higher in patients with T2DM than in age-matched healthy controls. Complementary in vitro studies further corroborated that hallmark metabolic insults of T2DM-including elevated glucose, hyperinsulinaemia and free fatty acids-act synergistically to trigger CA, a phenotype mechanistically coupled to activation of the protein kinase B (AKT)-reactive oxygen species (ROS) signalling axis, which drives upregulated expression and centrosomal translocation of both Rho-associated protein kinase 1 (ROCK1) and 14-3-3 protein sigma (14-3-3σ)[15]. Subsequent bodies of work have progressively delineated the multilayered regulatory network underlying diabetes-associated CA. Secreted Wnt6 engages the plasma membrane receptor frizzled 4 (FZD4) to activate canonical Wnt signalling, which functions upstream of the ROCK1/14-3-3σ complex to govern CA[16]. Proliferating cell nuclear antigen (PCNA) promotes CA through ROCK1 activation, whereas the c-Jun N-terminal kinase 1 (JNK1)-signal transducer and activator of transcription 3 (STAT3) pathway imposes negative feedback suppression by targeting 14-3-3σ; these two counterbalancing axes collectively preserve centrosome numerical homeostasis. The physical interaction between nucleophosmin (NPM) and histone H2B (H2B), along with the cellular myelocytomatosis oncogene (c-Myc)/cellular Jun proto-oncogene (c-Jun) transcriptional axis downstream of protein kinase C alpha (PKCα)-p38 signalling, have also been implicated as contributory mediators of diabetes-induced CA[17,18]. In addition, the peroxisome proliferator-activated receptor γ (PPARγ)-spindle and kinetochore-associated protein 1 (SKA1) axis has been identified as a novel transcriptional regulatory cascade that mediates diabetes-induced CA, extending the mechanistic link between metabolic nuclear receptors and centrosome dysregulation[19]. Furthermore, bioactive agents exemplified by resveratrol have been shown to attenuate diabetes-related CA by blunting the abovementioned signalling cascades, furnishing preliminary experimental rationale for centrosome-targeted therapeutic interventions in T2DM[20].

Despite robust evidence linking CA to T2DM, the current body of literature is constrained by several fundamental gaps and limitations. Most empirical data originate from studies of peripheral blood cells and tumour cell lines, whereas the prevalence, spatiotemporal dynamics and functional consequences of CA in metabolically critical cell types-including pancreatic β-cells, hepatocytes and adipocytes-remain largely uncharacterized. Furthermore, direct causal evidence remains elusive as to whether CA is merely an epiphenomenon secondary to the diabetic pathological milieu, or conversely acts as a pathogenic driver that reciprocally accelerates β-cell exhaustion, insulin resistance and the progression of diabetic complications. Finally, the downstream molecular cascades through which CA modulates metabolic homeostasis have yet to be systematically delineated.

This narrative review aims to synthesise current evidence on the biological hallmarks of CA, its upstream regulatory networks in the diabetic microenvironment, and its putative roles in T2DM-associated islet dysfunction, insulin resistance, and target-organ damage. It summarises current research advances, highlights core unresolved questions, and proposes prospective research directions for investigating centrosomal mechanisms in T2DM. Key clinical and preclinical studies are summarized in Table 1.

Table 1 Summary of key studies investigating centrosome amplification in diabetes and metabolic disorders.

Ref.	Experimental model	Sample size/cell line	CA definition	Diabetic experimental condition	CA-related quantitative outcome	Main findings	Evidence type
He et al[16], 2020	In vitro cell model	HCT116 human colon carcinoma cells	≥ 3 centrosomes per cell (γ-tubulin immunofluorescence)	High glucose + high insulin + FFA metabolic stress	Increased percentage of cells harbouring supernumerary centrosomes under diabetic-like stimulation	Wnt6-FZD4-β-catenin signalling acts upstream of ROCK1/14-3-3σ to drive diabetes-associated CA	Tumour-derived cell-line direct evidence
Liu et al[19], 2019	In vitro cell model	HCT116 human colon carcinoma cells	≥ 3 centrosomes per cell (γ-tubulin immunofluorescence)	High glucose + high insulin + FFA metabolic stress	Elevated proportion of CA-positive cells under diabetic-like conditions upon PPARγ overexpression	Diabetic-like stimuli up-regulate PPARγ; PPARγ transcriptionally activates SKA1 to promote diabetes-related CA	Tumour-derived cell-line direct evidence
Lu et al[17], 2019	In vitro cell model	HCT116 cells	≥ 3 centrosomes per cell	High glucose + high insulin + FFA metabolic stress	PCNA overexpression further elevates CA frequency under diabetic-like stress; JNK1-STAT3 activation reduces the percentage of CA-positive cells	PCNA promotes whereas JNK1-STAT3 suppresses diabetes-associated CA via targeting the ROCK1/14-3-3σ complex	Tumour-derived cell-line direct evidence
Xu et al[18], 2022	In vitro cell model	HCT116 human colon carcinoma cells	≥ 3 centrosomes per cell (γ-tubulin immunofluorescence)	High glucose + high insulin + FFA metabolic stress	Enhanced CA frequency upon diabetic-like stress, diminished after disruption of NPM-H2B interaction	NPM-H2B complex translocates from nucleus to centrosome under metabolic stress and mediates CA induction	Tumour-derived cell-line direct evidence
Wu et al[20], 2020	In vitro cell pharmacological intervention model	HCT116 human colon carcinoma cells	≥ 3 centrosomes per cell (γ-tubulin immunofluorescence)	High glucose + high insulin + FFA ± resveratrol treatment	Resveratrol dose-dependently reduced percentage of CA-positive cells under diabetic-like stress	Resveratrol attenuates diabetes-triggered CA by inhibiting PKCα-p38-c-MYC/c-JUN signalling cascade	Tumour-derived cell-line direct evidence (pharmacological assay)
Kushner et al[51], 2014	In vitro primary cell functional assay	Primary HUVECs	Supernumerary centrosomes (immunofluorescence)	No diabetic stimulus; experimental artificial centrosome over-abundance	Supernumerary centrosome-bearing cells display defective migration phenotype	Excess centrosomes disrupt endothelial cell polarity and migratory capacity	Primary-cell in vitro direct evidence
Huang-Doran et al[7], 2011	Human monogenic genetic study	Patients carrying biallelic PCNT loss-of-function mutations	Centrosomal structuralfunctional defect (not numerical acquired CA)	No experimental metabolic intervention; natural human germline mutation	Not applicable (structural defect, no quantitative CA count)	Pericentrin congenital deficiency causes severe human insulin resistance independent of adiposity	Human genetic indirect mechanistic evidence
Wang et al[15], 2018	Clinical cross-sectional cohort + in vitro cell model	30 T2DM patients, 30 age-matched healthy controls; HUVECs	≥ 3 centrosomes per cell (γ-tubulin immunofluorescence)	Clinical: Patients with established T2DM	PBMC CA frequency: 2-3-fold higher in T2DM patients vs controls; elevated CA rate observed in metabolically-stressed cultured endothelial cells	CA frequency in PBMCs is markedly increased in T2DM; glucolipotoxic stress induces CA via AKT-ROS-ROCK1/14-3-3σ signalling axis	Human observational-cohort direct evidence + primary-cell in vitro direct evidence
				In vitro: High glucose + high insulin + FFA stimulation

Studies are grouped into two broad categories: (1) Upper panel: Experiments investigating acquired numerical centrosome amplification under diabetic-like conditions; and (2) Lower-panel entry: Human monogenic disease reflecting congenital centrosome structural-functional defects, representing indirect mechanistic background evidence and not evidence for acquired numerical centrosome amplification in sporadic type 2 diabetes mellitus. CA: Centrosome amplification; FFA: Free fatty acids; FZD4: Frizzled-4; ROCK1: Rho-associated coiled-coil-containing protein kinase 1; PPARγ: Peroxisome proliferator-activated receptor gamma; SKA1: Spindle-and kinetochore-associated complex subunit 1; PCNA: Proliferating cell nuclear antigen; JNK1: C-Jun N-terminal kinase 1; STAT3: Signal transducer and activator of transcription 3; NPM: Nucleophosmin; H2B: Histone H2B; PKCα: Protein kinase C alpha; c-MYC: Cellular myelocytomatosis oncogene; c-JUN: Cellular Jun proto-oncogene; HUVECs: Human umbilical vein endothelial cells; T2DM: Type 2 diabetes mellitus; PBMCs: Peripheral blood mononuclear cells; AKT: Protein kinase B; ROS: Reactive oxygen species.

METHODS

This is a narrative minireview, and formal systematic-review-level quantitative meta-analysis was not performed.

Literature search

Electronic literature searches were performed in PubMed and Web of Science Core Collection. The final search execution date was June 30, 2026 (to be replaced with the actual date before formal submission), with no strict restriction on the earliest publication year. Representative search keywords and keyword combinations included: CA, centrosome, T2DM, insulin resistance, diabetic complications, oxidative stress, β-cell. Manual snowball reference screening was conducted by checking the reference lists of retrieved full-text articles to capture additional relevant publications. Only peer-reviewed full-text articles published in English were included for synthesis. Preprints, conference abstracts, editorials, and unpublished raw data were excluded. Non-English-language publications were not retrieved for this article.

Study selection and evidence appraisal

Literature screening followed a two-step process: Initial title-abstract screening, followed by full-text evaluation for potentially eligible records. Studies were prioritised according to the following evidence-level hierarchy: (1) Human clinical observational and cohort studies; (2) In vivo animal experimental studies; (3) Experiments using primary human or rodent metabolic cells; (4) Experimental data obtained from transformed tumour-derived cell lines; and (5) Human monogenic disease genetic evidence related to centrosome-cilium dysfunction. Formal risk-of-bias assessment tools were not applied, consistent with standard practice for narrative mini-reviews. Results originating from tumour-derived cell lines are explicitly commented on within the main manuscript text to alert readers to model-specific limitations. Representative key clinical and pre-clinical studies are summarised in Table 1.

Important limitations of this narrative synthesis should be acknowledged. Despite multi-database retrieval and snowball reference checking, complete capture of all relevant publications cannot be guaranteed, and inherent risks of selective citation exist for non-systematic narrative reviews. Readers should interpret synthesised mechanistic and translational viewpoints with these constraints in mind.

CENTROSOME BIOLOGY AND CA

Structure and duplication cycle of the centrosome

In mammalian cells, the centrosome is a membrane-less organelle approximately 1 μm in diameter, composed of a pair of orthogonally arranged centrioles embedded within a dense proteinaceous matrix known as the pericentriolar material (PCM)[21]. Each centriole is a cylindrical structure, approximately 500 nm in length and 250 nm in diameter, built from nine triplet microtubule blades. The two centrioles are functionally asymmetric. The more mature mother centriole bears distal and subdistal appendages, which mediate microtubule anchoring and enable the centriole to serve as a basal body for primary cilium assembly. By contrast, the daughter centriole remains structurally immature and acquires appendages only during the subsequent cell cycle[22]. The PCM is a stratified molecular scaffold composed of proteins including the γ-tubulin ring complex, pericentrin and CDK5RAP2. It is responsible for microtubule nucleation and anchoring, and its composition is dynamically remodelled throughout the cell cycle[23]. Proteomic studies have identified more than 400 proteins that localize to the centrosome, forming an intricate interaction network that coordinates cell-cycle progression, cell division and signal transduction[24].

Centrosome duplication is tightly coupled to the cell cycle and occurs only once per cycle, thereby ensuring the faithful inheritance of a single centrosome by each daughter cell. This duplication programme proceeds through four sequential stages: Centriole disengagement, procentriole assembly, centriole elongation and centrosome maturation[25]. At the G1/S transition, the paired centrioles lose their orthogonal configuration and disengage, a licensing step that permits centriole duplication. Subsequently, a single procentriole grows perpendicularly from the lateral wall of each mother centriole, a process initiated and governed by a core regulatory module comprising PLK4, STIL and SAS-6[26]. During S and G2 phases, procentrioles progressively elongate and recruit additional PCM components. Upon entry into mitosis, centrosomes undergo maturation: The PCM expands markedly, microtubule-nucleating capacity increases, and the two centrosomes separate and migrate to opposite poles of the cell to assemble a bipolar mitotic spindle[27]. This highly ordered process is orchestrated by an elaborate regulatory network involving kinases such as CDK2, PLK4 and Aurora A, together with the ubiquitin-proteasome degradation pathway, thereby preserving stringent control over centrosome number[28].

Definition and causes of CA

Numerical CA refers to the presence of centrosomes or centrioles exceeding the number expected for the corresponding cell-cycle stage. It is commonly operationalized as more than two centrosomes in a mitotic cell or more than four centrioles after completion of normal duplication, although the exact criterion depends on the marker and cell-cycle context[25,29,30]. CA arises principally through three mechanisms. First, excessive centriole biogenesis can occur when multiple rounds of centriole duplication take place within a single cell cycle, often as a consequence of dysregulated PLK4 activity or perturbation of other duplication regulatory factors. Second, cytokinesis failure generates tetraploid cells, accompanied by a concomitant doubling of centrosome number. Third, centriole fragmentation, caused by structural damage, can produce incomplete centriole-derived fragments that ectopically nucleate microtubules and thereby form additional microtubule-organizing centres[29]. It is important to distinguish numerical CA from centrosome structural defects, primary ciliary dyskinesia, and other centrosome-related abnormalities, as these entities have distinct molecular aetiologies and functional consequences. The functional impact of CA also depends on cell cycle status: In proliferating cells, CA primarily disrupts mitotic spindle formation and chromosomal instability, whereas in post-mitotic terminally differentiated cells, CA mainly alters interphase microtubule organization, intracellular trafficking, and ciliary function.

A broad range of cellular stresses can induce CA, including DNA damage, viral infection, oncogene activation and oxidative stress[31]. In the context of metabolic disease, nutrient excess, glucolipotoxicity and inflammatory stress have emerged as previously underappreciated drivers of CA, directly linking metabolic dysfunction to the regulation of centrosome homeostasis[15]. Disruption of any component of the centrosome-duplication machinery may compromise this homeostatic system and lead to aberrant centrosome accumulation.

Functional consequences of CA beyond mitosis

The best-characterized consequence of CA is its contribution to tumorigenesis. Supernumerary centrosomes promote the formation of multipolar mitotic spindles, leading to chromosome mis-segregation, aneuploidy and chromosomal instability, thereby driving malignant transformation and tumour progression[32]. Even when extra centrosomes are clustered to form a bipolar spindle, the clustering process itself increases the frequency of chromosome-segregation errors, a phenomenon termed centrosome clustering-mediated chromosomal instability[33].

Importantly, CA also exerts profound effects in interphase cells, independent of overt mitotic defects. Excess centrosomes enhance global microtubule-nucleating activity, altering the dynamics and spatial organization of the microtubule network and disrupting intracellular vesicle trafficking and cell polarity[34]. Studies in epithelial cells have shown that CA compromises tissue architecture, promotes invasive behaviour and rewires signalling outputs by redistributing signalling molecules along an expanded microtubule network[35]. In terminally differentiated, non-dividing cells-including neurons, endocrine cells and podocytes-centrosome abnormalities primarily impair cellular function by perturbing intracellular transport, cytoskeletal organization and primary cilium formation, rather than by inducing mitotic errors[30]. These interphase-specific effects are particularly relevant to metabolic disease, as many key metabolic cell types are terminally differentiated or exhibit only limited proliferative capacity.

GENETIC EVIDENCE LINKING CENTROSOME DYSFUNCTION TO GLUCOSE METABOLIC HOMEOSTASIS

The most compelling evidence that centrosome function participates in the regulation of glucose metabolism derives from human genetics. A series of monogenic disorders caused by mutations in genes encoding centrosomal or basal-body proteins converge on a shared phenotype of profound metabolic derangement, most notably severe insulin resistance and early-onset diabetes. These naturally occurring “experiments of nature” demonstrate that the structural and functional integrity of the centrosome is indispensable for maintaining systemic glucose homeostasis. Importantly, this body of evidence reflects human genetic data for congenital centrosome structural-functional defects and does not constitute direct evidence for acquired numerical CA in sporadic T2DM.

Alström syndrome

Alström syndrome is a rare autosomal-recessive ciliopathy caused by biallelic loss-of-function mutations in ALMS1, which encodes a large protein localized to the centrosome and the basal body of the primary cilium[5]. ALMS1 has critical roles in centrosome maintenance, intracellular protein trafficking and the regulation of ciliary function. Alström syndrome is a multisystem disorder, in which metabolic dysfunction represents one of the most prominent and life-threatening manifestations. Nearly all affected individuals develop early-onset obesity, severe insulin resistance and hyperinsulinaemia during childhood, and more than 70% progress to overt type 2 diabetes before the age of 30 years[6]. Additional metabolic features include dyslipidaemia and hepatic steatosis, yielding an overall phenotype that closely resembles metabolic syndrome. Hearn and colleagues demonstrated that ALMS1 is expressed in key metabolic tissues, including pancreatic islets, adipose tissue and liver, and exhibits a subcellular distribution centred on the centrosome. These observations provide direct support for centrosomal and basal-body dysfunction as an important driver of insulin resistance and diabetes pathogenesis[5].

Pericentrin deficiency

Pericentrin, encoded by PCNT, is a core scaffold protein of the PCM and is essential for centrosome maturation, microtubule nucleation and mitotic spindle assembly. Homozygous mutations in PCNT cause microcephalic osteodysplastic primordial dwarfism type II, a disorder characterized by severe intrauterine and postnatal growth restriction, microcephaly and skeletal dysplasia[8]. Notably, metabolic abnormalities are almost universal among individuals with microcephalic osteodysplastic primordial dwarfism type II: More than 80% develop severe insulin resistance, approximately 50% are diagnosed with diabetes during adolescence or early adulthood, and many also exhibit dyslipidaemia and non-alcoholic fatty liver disease. Huang-Doran et al[7] performed detailed metabolic phenotyping of patients with PCNT mutations and showed that pericentrin deficiency can cause severe human insulin resistance independently of body weight and fat distribution. This finding establishes that congenital centrosome structural defects can drive systemic insulin resistance and diabetes in humans, but does not represent direct evidence for acquired numerical CA in sporadic T2DM.

POC1A deficiency and SOFT syndrome

POC1A is a centriolar protein involved in centriole biogenesis, length control and basal-body function. Homozygous mutations in POC1A cause SOFT syndrome, a rare form of primordial dwarfism named for its cardinal features: Short stature, onychodysplasia, facial dysmorphism and hypotrichosis[9,36]. Beyond abnormalities in growth and development, affected individuals commonly develop metabolic complications, including hyperinsulinaemia, impaired glucose tolerance, systemic insulin resistance and hepatic steatosis[10,11]. In 2024, Perge et al[12] reported longitudinal metabolic data from patients with POC1A deficiency, showing that insulin resistance emerges early in life and progresses with age, with a substantial proportion of adults ultimately developing clinically overt type 2 diabetes.

Other centrosomal and ciliary disorders with metabolic phenotypes

Bardet-Biedl syndrome is a ciliopathy frequently associated with obesity and insulin resistance. However, Bardet-Biedl syndrome dysfunction primarily affects ciliary trafficking and should not be equated with acquired numerical CA. These disorders therefore provide indirect biological support for a link between the centrosome-cilium apparatus and metabolic homeostasis, rather than direct evidence that CA causes sporadic T2DM[37]. In addition, human genetic studies have linked loss-of-function variants in centrosomal scaffold genes CEP19 and CEP250 to severe obesity and impaired glucose homeostasis. Notably, these variants drive congenital centrosome structural defects rather than acquired numerical CA observed in sporadic T2DM.

Collectively, these human genetic data indicate that congenital defects in centrosome structure or function can precipitate persistent and severe metabolic disease, thereby establishing a causal relationship between centrosome integrity and glucose metabolic homeostasis. However, these monogenic conditions involve developmental and structural abnormalities of the centrosome-cilium apparatus, which are distinct from the acquired numerical CA observed in sporadic T2DM. This genetic foundation supports the biological plausibility that acquired CA may contribute to metabolic dysfunction, but direct human evidence for pathogenic acquired numerical CA in sporadic T2DM remains absent.

MOLECULAR MECHANISMS OF CA IN THE DIABETIC MICROENVIRONMENT

Building on genetic evidence, both clinical observations and in vitro studies have shown that the diabetic microenvironment can directly induce CA. The T2DM pathological microenvironment is defined by a constellation of metabolic insults: Hyperglycaemia, hyperinsulinaemia, excess free fatty acids, advanced glycation end products (AGEs), and pro-inflammatory cytokines. These stimuli converge on multiple interconnected signalling pathways to drive CA, forming an integrated network of core pro-amplification axes, endogenous negative feedback loops, and downstream functional effector complexes. The core signalling networks mediating diabetes-associated CA are schematically illustrated in Figure 1. This section summarizes the molecular cascades that mediate diabetes-associated CA.

Figure 1 Integrated signaling networks mediating centrosome amplification under diabetic conditions. Under the pathological microenvironment of type 2 diabetes mellitus, multiple metabolic stressors - including high glucose, insulin, palmitic acid/free fatty acids, and advanced glycation end products - activate converging signaling cascades that collectively drive centrosome amplification (CA). The core pro-CA axes include: (1) The proliferating cell nuclear antigen-Rho-associated protein kinase 1 (ROCK1) axis, which promotes assembly of the centrosomal ROCK1/14-3-3σ complex via upstream reactive oxygen species/protein kinase B signaling; (2) The Wnt6/plasma membrane receptor frizzled 4/β-catenin pathway, in which secreted Wnt6 binds to the membrane receptor plasma membrane receptor frizzled 4 and activates nuclear β-catenin to transcriptionally upregulate ROCK1 and 14-3-3σ; (3) Diabetic stimuli promote centrosome amplification by enhancing ROCK1 binding to dynactin subunit 2 and kinesin family member 2A; (4) The protein kinase C alpha-p38 mitogen-activated protein kinase-cellular myelocytomatosis oncogene/cellular Jun proto-oncogene transcriptional cascade, which activates the transcription of centrosome duplication-related genes; (5) The nucleophosmin-histone H2B complex pathway, where the nuclear nucleophosmin-histone H2B complex translocates to the centrosome and disrupts centrosome duplication control; (6) The peroxisome proliferator-activated receptor gamma-spindle-and kinetochore-associated complex subunit 1 regulatory axis, which modulates centrosome homeostasis via transcriptional upregulation of spindle-and kinetochore-associated complex subunit 1; (7) the advanced glycation end products-receptor for advanced glycation end-products-Krüppel-like factor 5-mouse double minute 2 signaling axis, which facilitates CA through mouse double minute 2-mediated degradation of p53 and Rb; and (8) The c-Jun N-terminal kinase 1-signal transducer and activator of transcription 3-14-3-3σ pathway serves as an endogenous negative feedback loop to restrain excessive CA. Symbols: Solid arrow (→) = promotion, activation, or transcriptional upregulation; blunt-ended line (⊣) = inhibition. T2DM: Type 2 diabetes mellitus; FFA: Free fatty acids; PCNA: Proliferating cell nuclear antigen; ROCK1: Rho-associated coiled-coil-containing protein kinase 1; DCTN2: Dynactin subunit 2; KIF2A: Kinesin family member 2A; JNK1: C-Jun N-terminal kinase 1; STAT3: Signal transducer and activator of transcription 3; PPARγ: Peroxisome proliferator-activated receptor gamma; SKA1: Spindle-and kinetochore-associated complex subunit 1; PKCα: Protein kinase C alpha; c-Myc: Cellular myelocytomatosis oncogene; c-Jun: Cellular Jun proto-oncogene; AGEs: Advanced glycation end-products; RAGE: Receptor for advanced glycation end-products; KLF5: Krüppel-like factor 5; MDM2: Mouse double minute 2; NPM-H2B: Nucleophosmin-histone H2B complex; p-JNK1: Phosphorylated c-Jun N-terminal kinase 1; p-STAT3: Phosphorylated signal transducer and activator of transcription 3; p-PKCα: Phosphorylated protein kinase C alpha; Rb: Retinoblastoma protein; NPM: Nucleophosmin; AGEs: Advanced glycation end-products.

Core pathway: The AKT-ROS-ROCK1/14-3-3σ signalling axis

Wang et al[15] first established an in vitro model of diabetes-induced CA using combined treatment with high glucose, high insulin and palmitic acid, and identified the AKT-ROS-ROCK1/14-3-3σ signalling axis as a central mediator of this process. Mechanistically, diabetic metabolic stress activates the serine/threonine kinase AKT, which in turn perturbs mitochondrial function and increases intracellular ROS production. Accumulation of ROS upregulates the expression of ROCK1 and the scaffold protein 14-3-3σ, and promotes their translocation from the cytoplasm to the centrosome. The centrosomal accumulation of the ROCK1/14-3-3σ complex disrupts the normal licensing and control of centrosome duplication, ultimately resulting in CA. Pharmacological inhibition of AKT, scavenging of ROS, or knockdown of ROCK1 or 14-3-3σ effectively suppresses diabetes-induced CA, confirming the central role of this pathway[15].

In HCT116 colon cancer-cell models, diabetic pathological stimuli further promote CA by facilitating the interaction between centrosomal ROCK1 and its binding partners dynactin subunit 2 as well as kinesin family member 2A[38]. Disruption of either ROCK1-dynactin subunit 2 or ROCK1-kinesin family member 2A protein-protein interaction can attenuate CA induced by glucolipotoxic stress. Of note, these mechanistic observations are exclusively derived from transformed tumour-cell-line experiments and remain to be validated in non-transformed primary cells.

Upstream regulation by Wnt6/FZD4

Subsequent work by He et al[16] identified the secreted ligand Wnt6 as a key upstream trigger of diabetes-associated CA. They showed that diabetic stimuli induce cellular secretion of Wnt6, which binds to the membrane receptor FZD4 and activates canonical Wnt/β-catenin signalling. Activated Wnt signalling transcriptionally upregulates ROCK1 and 14-3-3σ and facilitates their recruitment to the centrosome, thereby driving CA. Neutralization of extracellular Wnt6 with a specific antibody, or genetic ablation of FZD4, markedly attenuates diabetes-induced CA, indicating that Wnt6/FZD4 signalling acts upstream of the ROCK1/14-3-3σ complex[16]. This finding further suggests that diabetes-associated CA may involve a paracrine mechanism, whereby metabolically stressed cells secrete Wnt ligands that induce centrosomal abnormalities in neighbouring cells.

PCNA as a positive regulator and JNK1-STAT3 as a negative-feedback pathway

The regulation of diabetes-associated CA reflects a finely tuned balance between pro-amplification and inhibitory signals. Wang et al[15] demonstrated that PCNA, classically known for its roles in DNA replication and repair, can positively regulate CA by directly binding to and activating ROCK1. This interaction enhances formation of the ROCK1/14-3-3σ complex and promotes its recruitment to the centrosome[17].

In contrast, the JNK1-STAT3 pathway was reported to restrain diabetes-associated CA by modulating 14-3-3σ and the centrosomal ROCK1/14-3-3σ complex[17]. The net effect is therefore inhibition of CA. Under physiological conditions, the pro-amplification activity of PCNA is counterbalanced by the inhibitory action of JNK1-STAT3, thereby preserving centrosome-number homeostasis. In the diabetic milieu, however, chronic metabolic stress may overwhelm this negative-feedback circuit and shift the equilibrium towards sustained CA[17].

Notably, the abovementioned pathway analyses were conducted in the HCT116 human colon cancer cell line[17]. Transformed cancer cell lines inherently harbour genomic instability and dysregulated centrosome dynamics, which may not fully recapitulate the centrosomal regulatory landscape in primary metabolic cells such as pancreatic β-cells, hepatocytes, or podocytes. Therefore, these findings should be interpreted with caution, and validation in non-neoplastic primary cells or in vivo models is required to confirm their relevance to T2DM pathogenesis.

Nuclear-to-centrosomal translocation of the NPM-H2B complex

Dynamic redistribution of nuclear proteins represents an additional layer of centrosome regulation. Xu et al[18] found that NPM, a multifunctional protein primarily localized to the nucleolus but also present at the centrosome, has a critical role in diabetes-associated CA. Under diabetic conditions, the interaction between nuclear NPM and histone H2B is strengthened, leading to translocation of the NPM-H2B complex from the nucleus to the centrosome. This redistribution disrupts the normal control of centrosome duplication, probably by altering the stability or activity of centrosome regulatory proteins. Disruption of the NPM-H2B interaction markedly reduces diabetes-induced CA, revealing a previously unappreciated mechanism by which metabolic stress couples nuclear protein dynamics to centrosome homeostasis[18].

The PKCα-p38 mitogen-activated protein kinase-c-MYC/c-JUN transcriptional cascade

The PKCα-p38 mitogen-activated protein kinase (MAPK) pathway is a canonical stress-responsive signalling cascade that also mediates diabetes-induced CA. Wang and colleagues showed that diabetic metabolic stress activates PKCα, which subsequently phosphorylates and activates p38 MAPK. Activated p38 MAPK increases the expression of the oncogenic transcription factors c-MYC and c-JUN, which in turn transcriptionally activate downstream genes involved in centrosome duplication, culminating in CA[20]. This pathway links metabolic and inflammatory stress signals to transcriptional control of the centrosome biogenesis machinery.

PPARγ-SKA1 regulatory axis

The PPARγ-SKA1 axis represents an additional transcriptional regulatory module involved in diabetes-associated CA[19]. Under diabetic metabolic stimuli, upregulated nuclear PPARγ acts as a transcription factor to bind the promoter region of the SKA1 gene and upregulate its expression. Elevated SKA1 protein participates in the regulation of centrosome homeostasis and contributes to CA occurrence by modulating microtubule-kinetochore attachment and centrosome dynamics. This pathway extends the mechanistic link between metabolic nuclear receptors and centrosome dysregulation in T2DM.

Additional mediators: AGEs and inflammatory cytokines

Beyond glucolipotoxicity, AGEs and pro-inflammatory cytokines are well-established drivers of diabetic tissue injury[39-44]. However, there are no published experimental studies that directly measure CA as an experimental endpoint to verify whether AGE-receptor for advanced glycation end-products (RAGE)-Krüppel-like factor 5 (KLF5)-mouse double minute 2 (MDM2) or tumor necrosis factor-alpha (TNF-α)-nuclear factor-kappa-B (NF-κB) signalling can drive acquired numerical CA. Any proposed functional connection between these inflammatory cascades and CA remains an unvalidated scientific hypothesis. AGEs accumulate in diabetic tissues and promote diabetic complications[40]. It is well-established that AGE-RAGE signalling triggers intracellular oxidative-stress responses, including nicotinamide adenine dinucleotide phosphate, reduced form-oxidase-dependent ROS production[41]. In unrelated disease contexts, AGE-RAGE signalling can induce transcriptional up-regulation of KLF5, which further modulates MDM2 expression[42]. MDM2 mediates ubiquitination-dependent degradation of p53 and Rb, two tumour suppressors that constrain centriole over-duplication. While these individual molecular events have been separately reported in existing literature, the complete hypothetical signalling cascade from AGE-RAGE-KLF5-MDM2 towards centrosome over-duplication has not been functionally interrogated in experiments measuring CA phenotypes. Similarly, chronic low-grade inflammation constitutes a hallmark of T2DM[43]. The pro-inflammatory cytokine TNF-α triggers downstream NF-κB-dependent inflammatory signalling[44]. Even though ROS-ROCK1/14-3-3σ mediates CA in tumour-cell-line assays[15], whether TNF-α-NF-κB signalling functionally perturbs centrosome number homeostasis has not been directly tested. Further dedicated in vitro experiments quantifying CA endpoints are required to test this proposed mechanistic link.

PATHOGENIC ROLES OF CA IN T2DM AND ITS COMPLICATIONS

Having outlined how the diabetic milieu can induce CA, we next consider how this aberration may, in turn, contribute to the onset and progression of T2DM and its complications. It is critical to draw a conceptual distinction between published observations of physiological-pathological behaviour of the centrosome-microtubule machinery and pathogenic phenotypes specifically stemming from acquired numerical CA in T2DM; data from the former cannot be uncritically applied to the latter. Functional consequences of CA depend heavily on cell proliferative status. Accordingly, the pathogenic roles of CA are discussed separately for proliferating vascular/interstitial cells and terminally differentiated metabolic parenchymal cells, as schematically summarized in Figure 2. Most mechanistic reasoning within this section represents scientific hypotheses, given the scarcity of direct human or primary-cell evidence documenting CA itself in key metabolic organs.

Figure 2 Distinct pathogenic effects of centrosome amplification in proliferating vs terminally differentiated cells. The functional consequences of centrosome amplification (CA) are highly dependent on cell cycle status. In proliferating vascular cells (endothelial cells and vascular smooth muscle cells), CA primarily induces mitotic defects including multipolar spindle formation, chromosomal instability and aneuploidy, as well as cellular senescence and dysregulated proliferation/migration, which collectively drive diabetic macrovascular complications and microvascular endothelial injury. In terminally differentiated metabolic parenchymal cells (pancreatic β-cells, hepatocytes, adipocytes, and podocytes), CA exerts pathogenic effects mainly through disrupting interphase microtubule organization, intracellular vesicle trafficking, primary ciliogenesis, and Golgi positioning, contributing to β-cell dysfunction, systemic insulin resistance, and diabetic kidney disease. Solid black arrow: Supported by in vitro primary-cell experimental evidence; Grey dashed arrow: Hypothesized pathogenic cascade without direct evidence from metabolic tissues. VSMC: Vascular smooth muscle cell; CA: Centrosome amplification; CIN: Chromosomal instability.

CA impairs pancreatic β-cell function

Pancreatic β-cell failure is a central determinant of T2DM progression. As the principal microtubule-organizing centre of β-cells, the centrosome has an essential role in insulin biosynthesis, vesicular trafficking and glucose-stimulated insulin secretion[13,14]. Pericentrin and centrosome-associated microtubule networks govern insulin-granule organization, trafficking, and docking required for glucose-stimulated insulin secretion[13,14]. These results describe centrosome-related structural-functional biology, rather than phenotypes stemming from acquired numerical CA.

Direct experimental evidence documenting the existence or functional impact of acquired numerical CA in human or primary diabetic β-cells remains absent. Drawing on known interphase phenotypes of supernumerary centrosomes from other cellular models, it is biologically plausible that excess centrosomes could disrupt microtubule network architecture and thereby disturb insulin-granule trafficking and secretory kinetics in β-cells. In experimental cellular systems, supernumerary centrosomes can nucleate extra cilia and disturb primary-cilium-dependent signal transduction[45]. Whether comparable events take place in diabetic β-cells remains unknown. Primary-cilium defects impair β-cell stimulus-secretion coupling in rodent experimental models[46]; nevertheless, primary-ciliary dysfunction must not be conflated with numerical CA. Whether acquired CA alters β-cell performance via ciliary-related perturbations is still untested.

Centrosomal dysfunction promotes peripheral insulin resistance

Insulin-regulated glucose handling in adipose tissue and skeletal muscle relies on properly coordinated cytoskeleton-dependent vesicle trafficking[47-49]. Microtubules participate in insulin-receptor trafficking, intracellular signal relay and glucose transporter type 4 (GLUT4) vesicle transport; microtubule-dependent GLUT4 trafficking is disrupted in insulin-resistant skeletal muscle[49]. These findings underline the biological importance of cytoskeletal integrity for peripheral insulin action, yet they reflect general microtubule biology and do not serve as evidence for acquired numerical CA.

Hypothetically, CA-driven remodelling of interphase microtubule networks could interfere with GLUT4 vesicle trafficking in adipocytes and skeletal muscle and lower insulin-stimulated glucose uptake. This conjecture has not been directly interrogated in primary diabetic metabolic tissues. For hepatocytes, the potential contribution of numerical CA to hepatic insulin resistance is even less well characterised, because hepatic insulin resistance originates predominantly from impaired insulin signalling and enhanced gluconeogenesis rather than GLUT4-mediated glucose uptake. At present, CA should only be treated as a candidate mechanistic hypothesis linking cytoskeletal disturbance to peripheral insulin resistance and cannot be regarded as an established causal factor.

CA accelerates diabetic macrovascular complications

Cardiovascular disease represents the leading cause of mortality among patients with T2DM, and endothelial dysfunction constitutes an early initiating event of diabetic vascular injury. Hyperglycaemia can induce CA in cultured human vascular endothelial cells[50]. Experimental generation of supernumerary centrosomes in vitro disrupts endothelial-cell migration via centrosome scattering and polarity defects[51]. In cultured human endothelial cells, AGEs trigger CA alongside elevated pro-inflammatory cytokine release, pointing to a mutually reinforcing feedback loop within this in vitro experimental setting[50].

Pericentrin deficiency in smooth muscle cells has been shown to augment atherosclerosis through altered cholesterol biosynthesis and endoplasmic-reticulum stress[52]. This phenotype arises from congenital centrosome structural deficiency, not acquired numerical CA in T2DM; it therefore delivers only indirect background information for vascular pathology.

Centrosomal abnormalities in diabetic microvascular complications

Diabetic nephropathy is a major microvascular complication of T2DM and a leading cause of end-stage kidney disease. Podocyte injury is central to nephropathy progression[53]. Polarized microtubule cytoskeletons are required for podocyte process formation and maintenance[53]. This observation highlights the biological significance of cytoskeletal organization in podocytes but does not establish the presence or pathogenic role of acquired numerical CA in diabetic kidney disease. Direct evidence demonstrating that high glucose induces numerical CA in mature podocytes is currently lacking. Accordingly, a pathogenic role for CA in diabetic nephropathy remains hypothetical and demands validation in primary podocytes, human diabetic kidney tissue, and tissue-specific in vivo models.

For diabetic retinopathy, which is characterised by endothelial dysfunction and pericyte loss, CA-associated endothelial injury is observable in cell-culture experiments. Nevertheless, no direct clinical or in vivo animal evidence links CA to retinal microvascular damage in diabetic retinopathy. This mechanistic connection constitutes a scientific hypothesis.

CLINICAL IMPLICATIONS: CA AS A BIOMARKER

Persistent elevations in CA across the course of T2DM, coupled with its close association with disease severity and complication burden, position CA as a potential exploratory candidate biomarker. Existing human direct evidence for CA is restricted to observational studies of PBMCs; evidence originating from target metabolic organs is absent.

CA in peripheral blood cells as a surrogate biomarker

PBMCs, readily accessible via routine venipuncture, represent an attractive surrogate tissue for biomarker exploration. Wang et al[15] reported higher CA frequency in PBMCs from patients with T2DM compared with age-matched non-diabetic controls. In view of the limited sample size and absence of independent large-cohort replication, PBMC CA should presently be treated as an exploratory research measure rather than a clinically validated biomarker.

Associations with disease severity and micro-/macrovascular complications

The magnitude of CA in PBMCs correlates with well-established clinical markers of T2DM severity, as reported in a clinical observational cohort[15]. Positive correlations exist between CA frequency and fasting plasma glucose, glycated haemoglobin, as well as disease duration; higher CA levels are observed in patients with suboptimal glycaemic control[15]. Furthermore, this crosssectional cohort also observed that patients presenting with end-organ complications (coronary artery disease, diabetic nephropathy, diabetic retinopathy) exhibited significantly higher CA than age-matched patients without such complications[15]. Notably, these observational findings are exclusively derived from a singlecohort crosssectional analysis[15]. These associations cannot establish causality, nor confirm whether CA precedes complication onset, and do not support prognostic inference for clinical complications. Collectively, these data indicate that PBMC CA levels reflect the overall burden of metabolic stress and end-organ injury in T2DM.

Potential utility in risk stratification and prognostic evaluation

Given mechanistic insights from in vitro experiments and observational correlative findings from a single human cohort[15], CA holds only hypothetical translational potential for clinical risk stratification. Quantification of PBMC CA might theoretically enable earlier identification of patients at elevated risk of end-organ injury and facilitate timely intensified interventions; however, such prognostic utility remains completely unvalidated and cannot be inferred from existing crosssectional data[15]. Nevertheless, longitudinal clinical studies are still missing to validate whether baseline CA levels prospectively predict major adverse cardiovascular events, progressive renal function decline or other adverse clinical endpoints. CA could theoretically serve as a pharmacodynamic biomarker to monitor responses to new glucose-lowering or disease-modifying therapies, yet this idea remains untested.

Multiple practical barriers stand in the way of clinical translation: Standardized detection workflows, validation in large prospective cohorts, and population-specific reference ranges must all be established. At present, CA is only an exploratory biomarker candidate without proven clinical utility. Even with these caveats, CA remains a biologically plausible, mechanistically grounded candidate biomarker for risk stratification and prognostication in T2DM.

Several analytical and biological confounders complicate PBMC-CA measurement, including variable ex vivo blood-sample handling conditions, heterogeneity across immune-cell subpopulations, and the absence of standardized immunofluorescence scoring criteria for centrosome counting. These sources of variation constitute substantial practical obstacles before clinical translation can be contemplated.

At present, the hypothetical clinical-research use-case for PBMC-CA would be risk-stratification for diabetic end-organ complications. Critical prerequisites still need to be fulfilled: Standardized high-throughput detection workflows, population-specific reference ranges, head-to-head comparison against established biomarkers such as glycated haemoglobin and estimated glomerular filtration rate, and large prospective-cohort data demonstrating incremental predictive value over conventional clinical variables. Immunofluorescence-based CA quantification is technically demanding and not readily implementable within low-resource laboratory settings, and relevant population-stratified data across diverse ethnic and comorbidity subgroups remain absent.

THERAPEUTIC OPPORTUNITIES TARGETING CA

Targeting CA represents an investigational research direction for T2DM and its complications. All supporting evidence is pre-clinical. No animal-in vivo or human intervention trials have measured CA as a primary experimental endpoint. Therefore, observed in vivo therapeutic benefits of agents described below cannot be attributed to CA suppression at this stage; CA-targeted therapeutic potential remains entirely hypothetical. Attenuation of diabetes-related CA has only been demonstrated in cellular in vitro models.

NATURAL BIOACTIVE COMPOUNDS

Naturally-occurring polyphenols and anti-inflammatory phytochemicals have attracted interest as potential CA modulators. Resveratrol can suppress diabetes-triggered CA in cellular models in a dose-dependent fashion[20]. Mechanistic in vitro work shows resveratrol inhibits the PKCα-p38 MAPK cascade and reduces downstream c-Myc/c-Jun expression[20]. Its ROS-scavenging activity may further contribute to CA suppression within cell-culture systems[20]. These observations are confined to cellular models; there is no in vivo or clinical evidence proving resveratrol improves diabetic outcomes by modulating CA.

Small-molecule inhibitors targeting CA pathways

Small-molecule inhibitors against key CA-regulating mediators represent a potential pre-clinical strategy. ROCK inhibitors such as fasudil produce beneficial metabolic and renal outcomes in diabetic animal models[54,55]. Importantly, those animal studies did not quantify CA as an experimental readout. Consequently, it remains speculative whether correction of CA contributes to the in vivo protective effects of ROCK inhibition. Although tumour-cell-line data support ROCK-1 suppression can limit CA[15], phenotypic improvement observed in animal experiments cannot be causally linked to CA-modulation without dedicated CA measurements in vivo.

Direct pharmacological inhibition of centriole duplication remains an oncology-derived experimental concept rather than an established strategy for metabolic disease. However, because these agents interfere more directly with cell division, they are generally associated with greater toxicity and a broader side-effect profile. For chronic metabolic diseases such as T2DM, where long-term safety is a paramount concern, such inhibitors may therefore be unsuitable for routine long-term use, and may only be considered for short-term intervention or for the treatment of specific severe complications. Overall, the therapeutic potential of CA-targeted small molecules in T2DM remains entirely hypothetical, with substantial safety and translational hurdles to be addressed.

Oxidative stress is a major component of T2DM and diabetic complications[56], and ROS participates in diabetes-associated CA in cellular models[15]. Nevertheless, there is no clinical evidence that nonspecific antioxidant supplementation improves diabetic outcomes by correcting CA.

Lifestyle and metabolic interventions

Lifestyle modification including regular exercise and weight management is a cornerstone T2DM clinical care[57]. Exercise improves glycaemic control and lowers oxidative stress and chronic low-grade inflammation[57]. However, no human or animal studies have evaluated whether physical activity alters CA abundance in peripheral blood or metabolic tissues.

Similarly, caloric restriction and weight loss improve insulin sensitivity and reduce systemic metabolic stress[58], and could theoretically ameliorate cellular conditions permissive for CA. Effective glycaemic control with standard glucose-lowering therapies may also reduce centrosome abnormalities indirectly by alleviating glucotoxicity, lipotoxicity and oxidative stress[59,60]. Taken together, these findings suggest that attenuation of CA may represent one of the downstream cellular mechanisms through which lifestyle and metabolic interventions exert their protective effects in diabetes, although direct confirmation is still needed.

CONCLUSION

Over the past decade, accumulating data have revealed a correlative association between CA and T2DM. Genetic studies of rare monogenic centrosomal disorders confirm that intact centrosome function is essential for maintaining glucose metabolic homeostasis; nevertheless, these congenital structural defects cannot be extrapolated to acquired numerical CA observed in sporadic common T2DM. Although the diabetic microenvironment can induce CA in multiple in vitro experimental systems, putative pathogenic roles whereby CA may contribute to β-cell dysfunction, insulin resistance, and the progression of diabetic complications remain largely hypothetical, as direct in vivo evidence from human key metabolic tissues is still scarce. Importantly, for sporadic common T2DM, CA presently constitutes only a correlative phenomenon and mechanistic hypothesis rather than an established pathogenic driver. Despite promising preliminary observational data from peripheral-blood mononuclear-cell cohorts, CA cannot yet be regarded as a practically deployable clinical biomarker. Similarly, therapeutic strategies targeting CA are purely pre-clinical investigational concepts and are not ready for real-world clinical use.

Despite these advances, several critical knowledge gaps remain. First, most existing evidence derives from in vitro studies in tumour cell lines and peripheral blood cells; direct in vivo evidence of CA in key metabolic cell types-including pancreatic β-cells, hepatocytes, adipocytes and podocytes-from patients with diabetes, and its dynamic changes over the course of disease, remains scarce. Second, the causal relationship remains incompletely defined. While genetic studies have firmly established that congenital centrosome dysfunction is sufficient to drive metabolic disorders, it remains unresolved whether acquired CA acts as an active driver of T2DM progression or merely represents a secondary byproduct of chronic metabolic stress. Clarifying this causal relationship requires in vivo validation using animal models with tissue-specific, inducible CA. Third, the downstream molecular mechanisms by which CA impairs metabolic function remain incompletely elucidated. Beyond microtubule and trafficking defects, whether CA alters cellular metabolism, gene expression and organellar function through alternative pathways warrants further in-depth investigation. Fourth, clinical translation of CA as a biomarker and therapeutic target remains in its infancy. Before any prospects for precision-oriented diagnosis or therapy can be considered, rigorous validation of analytical performance, clinical utility, and intervention safety must be completed. Standardised, high-throughput assays for CA detection need to be developed, and their prognostic value and therapeutic utility must be validated in large, prospective clinical cohorts.

Looking ahead, advances in super-resolution microscopy, single-cell analysis, spatial proteomics and CRISPR-based genome editing will enable interrogation of the role of CA in T2DM at unprecedented resolution. Future research should prioritise the following directions: (1) Characterising the landscape of CA in metabolic tissues from patients across different stages of T2DM; (2) Establishing causality using mouse models with tissue-specific CA; (3) Delineating the full spectrum of downstream effects through which CA modulates cellular metabolism; (4) Developing safe, selective CA-targeted therapies suitable for long-term use; and (5) Validating the clinical biomarker utility of CA in large prospective cohorts. Addressing these questions will deepen our fundamental understanding of T2DM pathogenesis and lay the pre-clinical groundwork for potential future translational applications, without implying immediate readiness for clinical precision-oriented management.