Both the prevalence and incidence of T2D have increased greatly over the last few decades worldwide (Guariguata et al., 2014[132]; Ozougwu et al., 2013[234]; Reed et al., 2021[262]). T2D used to be a rare disease, with < 1 % of the population in the USA and China being affected in 1958 and 1980, respectively (Ginter and Simko, 2012[129]; Reed et al., 2021[262]; Weng et al., 2016[327]). Additionally, in some nations, it has become increasingly common for children and adolescents to develop T2D (Onge et al., 2015[231]; Reinehr, 2013[266]), with an unprecedented ~80 % of new cases in Japan not being adults (D'Adamo and Caprio, 2011[84]). Hence, T2D has shifted from being a historically rare adult disease to becoming a major global medical bane that can develop before adulthood (Reed et al., 2021[262]). The major increase in T2D prevalence and incidence was first observed in developed countries like the USA, but this increase has also been mirrored in developing countries (Ginter and Simko, 2012[129]; Guariguata et al., 2014[132]). In the future, the increased global rise in T2D cases is estimated to be largely due to a rise in prevalence, with its incidence largely stabilising in some nations (Reed et al., 2021[262]). Figure 2(Fig. 2) (References in Figure 2: Ong et al., 2023[230]) shows how the age-standardised prevalence rates of T2D varied between countries in 2021 and projected data for 2050.

Given that an estimated 90 % of people with diabetes are overweight, T2D pathogenesis is induced by excessive caloric diets with sedentary lifestyles, although it is still not clear how mechanistically (Reed et al., 2021[262], 2024[263]). The strong positive correlation between T2D risk and increasing BMI has led to the widely accepted notion that T2D is induced by poor diet and lifestyle (Reed et al., 2021[262]). However, it remains unclear why individuals of normal weight or even underweight subjects develop this condition (Mathews and Mathews, 2012[202]; Reed et al., 2021[262]; Zhou et al., 2016[351]). Studies have found that in Asian populations, many subjects are not overweight, with one study finding that > 50 % of the individuals were not overweight (BMI under 25 Kg/m²) in Japan (George et al., 2015[126]; Ma and Chan, 2013[199]; Mathews and Mathews, 2012[202]). Additionally, obese and overweight subjects have been reported to have had a lower mortality rate than normal-weight subjects due to T2D-associated cardiovascular pathology, termed “the obesity paradox” (Costanzo et al., 2015[77]; Hainer and Aldhoon-Hainerová, 2013[135]). Interestingly, studies have also shown that glycaemic control is worse in non-overweight T2D people (Coleman et al., 2014[73]; George et al., 2015[126]). Therefore, the existence of non-overweight T2D individuals with worse prognosis challenges the notion that T2D pathology is solely induced by lifestyles/diets that promote weight gain/obesity. The current consensus is that the islet beta-cell dysfunction is greater in non-overweight subjects and systemic insulin resistance is less pronounced (Coleman et al., 2014[73]; George et al., 2015[126]; Reed et al., 2021[262]), and as insulin resistance is induced by excessive nutrient storage systemically due to excessive diets/sedentary lifestyles, this seems logical. However, it is unclear/paradoxical why islet beta-cell dysfunction is present to a greater extent in non-overweight people, as islet beta-cell pathology in T2D is accepted to be promoted by excessive ingestion of nutrients resulting in weight gain (Reed et al., 2021[262]). Therefore, assuming this notion applies to all T2D individuals, increased islet beta-cell dysfunction should be observed in overweight subjects in comparison, which has not been reported by several studies (Coleman et al., 2014[73]; George et al., 2015[126]; Reed et al., 2021[262]).

Given that > 72 % of the global overweight population, estimated to have exceeded 2.8 billion in 2022 (WHO, 2025[331] ), have not been diagnosed with T2D (WHO, 2024[330]) and many overweight individuals do not even develop insulin resistance (Reed et al., 2021[262]), with one study finding that 66 % of individuals with a BMI ≥ 30-35 did not exhibit insulin resistance (Ferrannini et al., 1997[113]), demonstrates that diabetic pathology is not exclusively induced by excessive diets and poor lifestyles. Genetics has been identified as increasing susceptibility to T2D. The lifetime risk of developing T2D with one affected parent is ~40 % and ~70 % if both parents are affected (Ahlqvist et al., 2011[5]; Chen et al., 2011[64]; Prasad and Groop, 2015[252]). It is currently unclear if the increased susceptibility of relatives of T2D to develop this disease could be induced by them having similar diets and lifestyles, as well as genetics (Chen et al., 2011[64]; Poulsen et al., 2009[251]). One study demonstrated the association of genetics with T2D seemingly independently of diet/lifestyle, as disease prevalence and incidence were higher in the twin population compared to the singleton population; both of these populations had similar BMI scores (Poulsen et al., 2009[251]). Genome-wide association studies (GWAS) have found > 700 variants that confer small-to-modest susceptibility to T2D, as odds ratios reported for any individual loci do not exceed 1.6 to date, which implies there are currently untested variants that confer increased susceptibility 24256 (Imamura and Maeda, 2024[154]; Prasad and Groop, 2015[252]). In addition to unhealthy lifestyles and genetics, other risk factors have also been identified for T2D, such as alcohol consumption, smoking and psychological stress (Merabet et al., 2022[209]; Mozaffarian et al., 2009[216]; Reed et al., 2021[262]).

During weight gain, individuals develop systemic insulin resistance, which then initially induces islet beta-cells to increase their insulin content and secretion to compensate and maintain euglycemia (Fonseca, 2009[116]; Linnemann et al., 2014[187]; Mehran et al., 2012[208]). Over time, usually over a period of several years or more, the islet-beta cells are seemingly unable to cope with the increased burden and individuals insidiously progress to prediabetes, where blood glucose levels rise due to declining islet beta-cell function, but not to such an extent that this warrants T2D diagnosis (Bansal, 2015[27]; Tabák et al., 2012[300]). Although lifestyle interventions, such as weight loss and exercise, can delay or prevent progression to T2D or even reverse prediabetes, an estimated 70 % of individuals with persistent prediabetes eventually progress to T2D (Bansal, 2015[27]; Tabák et al., 2012[300]; Tuso, 2014[314]). T2D manifests clinically when symptoms appear due to hyperglycaemia (Olokoba et al., 2012[228]; Reed et al., 2021[262]). Individual's blood glucose levels rising to > 7 mmol during fasting and > 11 mmol two hours after 75g glucose ingestion warrant T2D diagnosis; subjects typically then require lifelong treatments to try to achieve euglycemia (Bansal, 2015[27]; Davidson et al., 2000[86]). In T2D, people's fasting insulin levels in the circulation are estimated to decrease by ~50 % and postprandial levels in subjects are estimated to be between 25-40 % of what is observed in diabetes-free weight-matched controls (Woerle et al., 2012[334]). Typically, T2D is a progressive disease, and ~75 % of subjects progress to requiring more than one pharmacological agent to alleviate hyperglycaemia within ten years of diagnosis (Reed et al., 2021[262]; Taylor, 2013[303]).

Although the widely accepted notion is that the T2D disease phenotype is induced by a heterogeneous combination of systemic insulin resistance and impaired insulin secretion, emerging evidence suggests that multiple other pathological mechanisms contribute to the disease pathogenesis/phenotype (Reed et al., 2021[262]; Sanches et al., 2023[274]). Islet beta-cell apoptosis is known to be increased in T2D individuals (Butler et al., 2003[54]; Khin et al., 2023[174]; Reed et al., 2021[262]), with greater levels of apoptosis observed in those with a BMI < 25 Kg/m² (Butler et al., 2003[54]), which likely additionally contributes to the insulin deficit in T2D. Islet amyloid polypeptide (IAPP) plaques in pancreatic islets, formed by misfolding of IAPP, have been found in the majority, > 90 %, of T2D individuals (Gupta and Leahy, 2014[134]; Smith et al., 2022[289]). Currently, it is thought that extracellular plaques of IAPP do not induce islet beta-cell pathology, but rather intracellular oligomers of IAPP produce cytotoxic effects, which contribute to the insulin deficit in T2D (Alrouji et al., 2023[10]; Gupta and Leahy, 2014[134]; Matveyenko and Butler, 2006[203]). Cytotoxic effects of intracellular IAPP reported by in vitro studies include increased generation of reactive oxygen species (ROS) and induction of endoplasmic reticulum (ER) stress, which likely contribute to both impaired insulin secretion and increased apoptosis in people with T2D, assuming the situation is similar in in vivo settings (Alrouji et al., 2023[10]; Casas et al., 2007[60]; Pilkington et al., 2016[247]). The precise mechanisms by which these plaques promote T2D in individuals are currently unclear (Smith et al., 2022[289]). Autophagic processes usually protect islet beta-cells from intracellular accumulation of IAPP oligomers, so it seems that these processes are impaired in T2D (Gupta and Leahy, 2014[134]; Shigihara et al., 2014[286]). T2D-free individuals also have IAPP deposits detected in their islets, but with a greatly reduced incidence compared to T2D subjects (Butler et al., 2003[54]; Reed et al., 2021[262]), which implies that the presence of these plaques alone does not induce disease.

Islet alpha-cell dysfunction, where glucagon is undesirably released into the circulation in T2D during the postprandial period, further worsens the hyperglycaemia (Moon and Won, 2015[214]; Sherwin et al., 1976[284]). It has been shown that the postprandial hyperglucagonemia can contribute to up to 50 % of the hyperglycaemia in individuals with T2D after glucose ingestion. T2D is, therefore, now considered to be a bihormonal disorder (Dinneen et al., 1995[97]; Moon and Won, 2015[214]; Reaven et al., 1987[260]; Shah et al., 2000[281]; Sherwin et al., 1976[284]; Shigeno et al., 2021[285]). Islet beta-cells are unresponsive to the incretin action of gastric inhibitory peptide (GIP), even at supraphysiological concentrations in T2D (Højberg et al., 2008[149]; Reed et al., 2021[262]). Hence, the impaired incretin effect also contributes to insufficient postprandial insulin secretion in T2D. Chronic islet inflammation is also associated with T2D, which is thought to contribute to islet beta-cell dysfunction, evidenced by alleviation of the disease phenotype by anti-inflammatory therapies in several studies (Cuenco and Dalmas, 2022[79]; Donath et al., 2009[98]; Eguchi and Nagai, 2017[108]). It has been demonstrated in both rodent (Ehses et al., 2009[109]; Sauter et al., 2008[277]) and human (Larsen Claus et al., 2007[180]; Rissanen et al., 2012[268]) studies that interfering with the IL-1 pathway reduced hyperglycaemia and improved islet beta-cell function. Additionally, extrapancreatic pathology, besides systemic insulin resistance, contributes to the disease phenotype; research suggests that some of these pathologies manifest during prediabetes, implying they are involved in T2D aetiology and pathogenesis (Basu et al., 2013[32]; Phielix and Mensink, 2008[244]; Reed et al., 2021[262]). Impaired fat catabolism and metabolic inflexibility in the skeletal muscle of T2D subjects have been observed, which contribute to the disease phenotype, as well as reduced insulin suppression of hepatic gluconeogenesis and glycogenolysis during prediabetes (Basu et al., 2013[32]; Phielix and Mensink, 2008[244]). The extrapancreatic pathologies, besides systemic insulin resistance, involved in exacerbating the T2D phenotype are reviewed in more detail elsewhere (Reed et al., 2021[262]).

Overall, the incidence and prevalence of T1D increase each year globally, and there were an estimated 9.2 million patients worldwide in 2024 (Bell and Lain, 2025[34]). Historically, the majority of T1D subjects were diagnosed during childhood and adolescence (Haller et al., 2005[136]; Karvonen et al., 2000[172]; Maahs et al., 2010[200]), but recent studies have found that adult-onset T1D has become increasingly common, even more common than childhood-onset (Leslie et al., 2021[183]). Interestingly, the incidence of T1D has increased greatly in the modern world globally, and in Finland, the incidence has increased > 4-fold since the early 1950s (Virtanen and Knip, 2003[323]). Between 1990 and 1994, the incidence of T1D was determined in children ≤ 14 years in 50 countries (Karvonen et al., 2000[172]; Maahs et al., 2010[200]); T1D incidence was much higher in Caucasian populations, with incidence rates of > 36/100,000 per year in Finland and Sardinia. In contrast, in China and Venezuela, T1D incidence was drastically lower, an estimated 0.1/100,000 per year. Therefore, although T1D has a global clinical presence, the associated healthcare burden of this disease varies greatly between nations and ethnicities (Ogle et al., 2025[226]). Surprisingly, subsequent studies have reported enormous increases in T1D incidence in low-risk countries such as China (Liu et al., 2021[188] ; Ogle et al., 2025[226]; Weng et al., 2018[328]). One study reported T1D incidence between 2010-2013 in China was 1.93/100,000 for 0-14 years, which is an approximately 19-fold increase from what was previously found (Weng et al., 2018[328]). Another study found that T1D incidence in China was 4.21/100,000 for the same age group in 2017 (Liu et al., 2021[188]), and overall T1D incidence in China went from 2.72 in 2007 to 3.6/100,000 in 2017. A recent study estimated that T1D incidence for subjects under 15 years in 2025 was 4.8 and 4.9/100,000 for China and Venezuela, respectively (Ogle et al., 2025[226]). These incidence rates are nearly 50-fold higher than those reported for these countries during 1990-1994. The drastic increase in T1D incidence in these low-risk countries over the last few decades warrants thorough investigation. It has been suggested that the earlier studies may have underestimated T1D incidence in China (Liu et al., 2021[188]), which could account for at least part of the reported increase. Figure 3(Fig. 3) (References in Figure 3: Ong et al., 2023[230]) shows how the age-standardised prevalence rates of T1D varied between countries in 2021 and projected data for 2050 are also shown.

Autoantibodies against insulin (IAA), glutamic acid decarboxylase (GADA), islet antigen 2 (IA-2A) and zinc transporter 8 autoantibodies (ZnT8A) are biomarkers of islet beta-cell autoimmunity (Hazime et al., 2025[142]; Jia and Yu, 2024[168]). Islet beta-cell autoimmunity is detected clinically by the presence of one or more of these four autoantibodies, and individuals can be consistently positive for these antibodies for months to years before stage 3 T1D manifests (Addissouky et al., 2024[4]; Bosi et al., 2017[44]; Jacobsen et al., 2020[159]). Individuals positive for multiple autoantibodies usually progress to the clinical phase of T1D more rapidly (Bonifacio and Ziegler, 2025[42]; Jamjoom et al., 2024[161]; Ziegler et al., 2013[353]). Screening for all four of these antibodies allows for a > 94 % detection rate of autoimmunity in recent-onset diabetes cases (Lampasona et al., 2010[179]; Long et al., 2012[195]). However, a subset of T1D subjects do not display seropositivity for any of these autoantibodies (Patel et al., 2021[238]). Interestingly, some individuals, known as slow progressors, remain T1D-free after > 10 years of being consistently positive for 2 or more autoantibodies (Achenbach et al., 2013[3]; Gillespie and Long, 2019[128]; Long et al., 2018[196]). Therefore, whilst detection of these autoantibodies demonstrates the presence of type 1 diabetic autoimmunity in subjects, they are not always reliable indicators of how long the preclinical asymptomatic phase lasts before onset of the clinical phase. Additionally, the presence of these antibodies can be transient, and titres can fluctuate in subjects for currently unclear reasons (Achenbach et al., 2005[2]; Phillip et al., 2024[245]; Sørgjerd, 2019[290]). Not all individuals who develop single autoantibodies test positive for them on follow-up screening or progress to develop multiple antibodies or T1D (Vehik et al., 2016[321]). Positivity for the autoantibodies post-diagnosis has been shown to decline with time (Long et al., 2021[193] ; Williams et al., 2022[333]), with one study reporting that < 30 % of subjects had 2 or more autoantibodies > 15 years after diagnosis (Williams et al., 2022[333]). In contrast, individuals can also test positive for one or more of the autoantibodies post-diagnosis (Decochez et al., 2000[91]; Sørgjerd, 2019[290]). Given these observations, it is clear that the humoral autoimmune response has a heterogeneous nature during the preclinical and clinical disease phases for currently unknown reasons. Whilst slow progressors are established in the literature 1 (Achenbach et al., 2013[3]; Gillespie and Long, 2019[128]), it is suggested here that individuals, after testing positive consistently for one or more of the autoantibodies, could be classified as rapid or intermediate progressors if they develop T1D within 5 years or between 5 and 10 years after initial seroconversion, respectively.

Genetic susceptibility plays a major but not exclusive role in T1D development, with multiple susceptibility genes being identified (Stefan et al., 2014[296]). Concordance rates between identical twins are estimated to be between 30-50 % by several studies (Hyttinen et al., 2003[153]; Stefan et al., 2014[296]; Verge et al., 1995[322]), which indicates the importance of both genetics and environmental factors in conferring T1D susceptibility. The DR3-DQ2 and DR4-DQ8 haplotypes of the human leukocyte antigen (HLA) class II genes are widely accepted as the major inherited risk factors for developing T1D (Nygård et al., 2021[223]; Sticht et al., 2021[298]; Zajec and Trebušak Podkrajšek, 2022[341]; Zhao et al., 2021[348]). An estimated 90 % of T1D individuals have one of these high-risk haploty[348]pes and approximately 40 % of subjects have both: in the healthy population, 30 % have one haplotype and 3 % have both (Zajec and Trebušak Podkrajšek, 2022[341]). An estimated > 95 % of individuals who have one of these high-risk haplotypes do not develop T1D (Giwa et al., 2020[130]). Interestingly, these haplotypes predict seroconversion and not age of onset, clinical progression, or C-peptide loss post-diagnosis (Carré et al., 2021[59]; Zajec and Trebušak Podkrajšek, 2022[341]). The Australian Melbourne prediabetes study reported that these high-risk haplotypes were more prevalent in individuals who developed multiple persistent autoantibodies than in individuals who were positive for one or transiently positive for one or more (Colman et al., 2000[74]). To date, more than 40 single-nucleotide polymorphisms (SNPs) outside the HLA region have been associated with T1D susceptibility, although these SNPs only have a small effect on disease risk in comparison to the strong influence of the HLA class II haplotypes (Barrett et al., 2009[30]; Nygård et al., 2021[223]; Törn et al., 2014[312]). Certain HLA class I alleles have also been shown to influence T1D risk (Zajec and Trebušak Podkrajšek, 2022[341]). HLA-A*24 predicts rapid progression to T1D in the presence of islet beta-cell autoimmunity in individuals with the DR3-DQ2 and/or DR4-DQ8 haplotypes (Jiang et al., 2021[170]; Long et al., 2013[194]; Mbunwe et al., 2013[204]; Noble and Erlich, 2012[221]; Zajec and Trebušak Podkrajšek, 2022[341]). Interestingly, one study also found that HLA-A*24 confers T1D risk in individuals who have neither high-risk haplotype (Honeyman et al., 1995[151]). HLA-B*39 also confers additional risk in individuals with high- and low-risk haplotypes and is also associated with early age at onset (Mikk et al., 2014[212], 2017[211]; Reijonen et al., 1997[265]). HLA-B*18 has also been shown to contribute to T1D progression in islet autoantibody-positive individuals with the DR3-DQ2 haplotypes (Mbunwe et al., 2013[204]). Interestingly, the HLA class II haplotype DRB1*15-DQB1*0602 is present in an estimated 20 % of the population but only occurs in around 1 % of people with T1D, and is therefore considered a protective haplotype for T1D (Lambert et al., 2004[178]; Pugliese et al., 1995[253]).

Given the difference in T1D incidence between ethnic groups, it is reasonable to assume that genetic ethnic differences may account for these observations. However, migration studies have demonstrated the importance of geographical/environmental factors, as when populations from low-risk ethnic groups immigrate to Western nations, the risk of their children developing T1D increases greatly (Harron et al., 2011[139]; Hjern et al., 2012[148]; Hussen et al., 2013[152]; Umpierrez and Ali, 2025[315]). Additionally, the strong increase in incidence globally over the last few decades cannot be exclusively attributed to changes in genetic susceptibility, further providing evidence that environmental factors must be playing an increased role in promoting T1D pathogenesis. For unknown reasons, T1D onset is more common during late autumn, winter and early spring (Maahs et al., 2010[200]; Turtinen et al., 2022[313]). One recent study found that the seasonality of T1D manifestation differed between younger and older children, indicating that age and seasonality interact synergistically to promote disease risk (Turtinen et al., 2022[313]). It has also been shown that infants born between November and February are less likely to develop T1D than infants born in April-July (Maahs et al., 2010[200]; Vaiserman et al., 2007[318]). Interestingly, improved hygiene conditions correlate with greater susceptibility to T1D; there is an inversely proportional relationship between T1D incidence and both infectious disease incidence and poor sanitation (Bach, 2021[24]; Bach and Chatenoud, 2012[25]). A rodent study supported this observation, as it was found that non-obese diabetic mice reared in poor sanitary conditions had decreased risk of T1D, and that deliberately infecting them with various pathogens prevented T1D manifestation (Bach and Chatenoud, 2012[25]). However, some studies have found evidence suggesting that viral infection by enteroviruses may promote T1D pathogenesis (Giwa et al., 2020[130]), with studies finding that among children diagnosed with T1D, > 60 % of them had positive coxsackievirus serology (Clements et al., 1995[70]; Friman et al., 1985[120]). Additionally, one study found that coxsackievirus B1 increased the risk of detection of two or more islet autoantibodies in T1D-free children (Laitinen et al., 2014[177]). In contrast, other studies have failed to show any link between enteroviral infections and T1D (Knip and Simell, 2012[176]).

Pancreatic inflammation (insulitis) was initially observed in pancreas samples from patients with recent-onset T1D nearly a century ago (Stansfield and Warren, 1928[294]). In the later part of the 20^th century, the cellular composition was elucidated: the vast majority of cells present in the insulitis were lymphocytes, which outnumbered the macrophages by a ratio of 10:1 (Foulis et al., 1991[117]; Gepts, 1965[127]). In recent decades, histological evidence for cytotoxic T cells killing islet beta-cells has emerged and the importance of antigen-specific T cells has been described (Coppieters and von Herrath, 2009[76]; Itoh et al., 1993[158]; Willcox et al., 2009[332]). The current notion is that cytotoxic T cell killing of islet beta-cells in a cell-mediated adaptive immune response, via induction of apoptosis, accounts for the majority of T1D pathogenesis, although the innate immune system is thought to also contribute by encouraging inflammation and T cell migration into the islets (Addissouky et al., 2024[4]; Pirot et al., 2008[248]; Richardson et al., 2011[267]; Yoon and Jun, 2005[339]; Zajec and Trebušak Podkrajšek, 2022[341]). The exact mechanisms of islet beta-cell death in T1D are still debated, but it is thought that following binding of a CD8+ T cell receptor to an autoantigen/HLA complex that it recognises, apoptosis is induced in islet beta-cells by perforin and granzyme activity, as well as Fas binding to Fas ligand, which both result in activation of various caspases and release of cytochrome c from the mitochondria (Long et al., 2012[195]; Richardson et al., 2011[267]; Taylor, 2024[304]; Tomita, 2017[311]; Uno et al., 2007[316]; van Belle et al., 2011[319]). Interestingly, the cellular composition of immune cells associated with the autoimmunity present in the islets can vary between subjects, with some cases having regulatory T cells, B lymphocytes and natural killer cells present (Dotta et al., 2007[100]; Richardson et al., 2011[267]; Willcox et al., 2009[332]). In healthy subjects, HLA class II is not expressed in pancreatic endocrine cells, but low levels of class I are (Richardson et al., 2011[267]). Preceding any inflammation or immune cell influx into the pancreatic islets, HLA class I is upregulated in the majority of islet beta-cells, and subsequently, hyperexpression of HLA class II then follows in a minority of these cells (Bottazzo et al., 1985[45]; Foulis et al., 1991[117]; Richardson et al., 2011[267]; Thomas et al., 1998[307]). It is thought that the increased antigen presentation to T cells then encourages autoimmunity, but it is currently unclear why the increased expression of HLA class I and II occurs (Akhbari et al., 2025[7]; Benkahla et al., 2021[37]; Richardson et al., 2011[267]).

The pathogenesis of T1D then progresses insidiously, as cytotoxic T cells move into the islets and begin attacking islet beta-cells presenting autoantigens, and macrophages are present in small numbers. However, the number of islet beta-cells declines slowly and individuals remain asymptomatic (Addissouky et al., 2024[4]; Richardson et al., 2011[267]). This ongoing autoimmunity can be detected by screening for the aforementioned autoantibodies (Hazime et al., 2025[142]; Jia and Yu, 2024[168]). Over a period of time, which can vary greatly between individuals, progression to late-stage islet beta-cell autoimmunity occurs, which is characterised by a more diffuse inflammatory infiltrate in and around the islets with both increased numbers of cytotoxic T cells and B lymphocytes, and significantly decreased numbers of islet beta-cells (Richardson et al., 2011[267]; Uno et al., 2007[316]). The number of macrophages remains at a similar level in late-stage islet beta-cell autoimmunity and it is thought they contribute to beta-cell demise by releasing pro-inflammatory cytokines such as TNFα and IL-1β (Richardson et al., 2011[267]; Uno et al., 2007[316]; Willcox et al., 2009[332]). It remains unclear whether the T1D autoantibodies mediate any clinically relevant islet beta-cell damage/death during the pre-clinical phase (Addissouky et al., 2024[4]; Richardson et al., 2011[267]; Tomita, 2017[311]; Yoon and Jun, 2005[339]). When islet beta-cell numbers become low during late-stage islet beta-cell autoimmunity (usually 70-80 % destroyed), they can no longer prevent hyperglycaemia, and T1D manifests clinically (stage 3 T1D) (Cnop et al., 2005[72]; Richardson et al., 2011[267]). Interestingly, emerging evidence suggests that whilst the vast majority of islet beta-cells are destroyed in subjects, some persistent low-level beta-cell function is usually present (Odyjewska et al., 2025[225]; Oram et al., 2019[232]). Studies have demonstrated that greater residual islet beta-cell function is associated with lower rates of post-diagnosis complications (Harsunen et al., 2023[140]; Steffes et al., 2003[297]). Currently, it is mechanistically unclear how these cells evade autoimmune destruction or if they formed pre- or post-diagnosis (Oram et al., 2019[232]). Understanding how individuals can maintain a small population of islet beta-cells and their function could yield new therapeutic strategies for T1D.

LADA is a form of diabetes that shares characteristics of both T1D and T2D (Ravikumar et al., 2023[259]). LADA diagnosis requires individuals to be over 30 years, exhibit milder T1D symptoms at diagnosis, not need insulin treatment for > 6 months after diagnosis and be positive for at least one of the T1D-associated autoantibodies (Buzzetti et al., 2020[55]; Ravikumar et al., 2023[259]). Given that these individuals typically present with symptoms more akin to T2D, they are often misdiagnosed as T2D (O'Neal et al., 2016[229]; Ravikumar et al., 2023[259]). A recent study estimated that 8.9 % of the global diabetic population had LADA (Ramu et al., 2023[258]). A previous study found a very similar observation, with 9.2 % of the T2D population and 0.6 % of the total population having LADA (Qi et al., 2011[254]). It could be argued that LADA is a milder form of T1D, given that autoantibodies are present and subjects typically progress sooner to requiring insulin treatment than T2D individuals (Sørgjerd et al., 2021[291]; Zhou et al., 2024[352] ). Although the progression to insulin dependence can vary greatly between LADA subjects, from months to many years (Sørgjerd et al., 2021[291]). However, an important distinguishing feature of LADA is that individuals exhibit insulin resistance (Chiu et al., 2007[67]; Ravikumar et al., 2023[259]), which implies similar pathology to T2D, despite the reduced frequency of obesity and metabolic syndrome in the LADA population (Buzzetti et al., 2020[55]; Ravikumar et al., 2023[259]; Zhou et al., 2024[352]). The risk of developing LADA has been associated with being overweight and adopting sedentary lifestyles, similar to T2D, but genetic predisposition similar to T1D has also been documented (Sørgjerd et al., 2021[291]). Hence, LADA has been referred to as type 1.5 diabetes, given the shared pathological features and risk factors between both of the major forms of diabetes (Ravikumar et al., 2023[259]). Interestingly, a similar situation to LADA occurs with individuals aged between 15 and 30, known as LADY; these subjects also do not require insulin treatment for > 6 months after diagnosis (Pampapathi et al., 2024[235]). Studies have found that 7.9-9 % of phenotypic T2D subjects in this age range test positive for islet autoantibodies associated with T1D, and this is associated with reduced islet beta-cell function and lower insulin resistance (Cheng et al., 2021[66], 2022[65]).

Over recent years, it has become increasingly evident that people with T2D exhibit islet beta-cell autoimmunity usually associated with T1D and LADA/LADY (Ravikumar et al., 2023[259]; Reed et al., 2021[262]). Studies have demonstrated that islet-reactive cytotoxic T cells and islet beta-cell autoantibodies are present in a subset of T2D subjects (Itariu and Stulnig, 2014[157]). The prevalence of islet autoantibodies has been estimated to be between 5-30 % in the T2D population by studies (Brooks-Worrell et al., 2013[49]). A study conducted in 2000 found that 12 % of elderly T2D elderly individuals were positive for either GADA or IA-2A, and the presence of these autoantibodies was associated with impairment of acute-phase insulin secretion, revealed by an oral glucose tolerance test (Pietropaolo et al., 2000[246]). More recently, a study found that 11 % of the study population diagnosed with T2D had one or more of GADA, IAA, or ZnT8A, and these individuals presented with worse glycaemic control in comparison to the T2D autoantibody-negative population (Moosaie et al., 2021[215]). To further support the notion that the presence of islet autoantibodies exacerbates disease phenotype in T2D, another study found that a much higher percentage of autoantibody-positive participants were using insulin in comparison to autoantibody-negative individuals (Lomeli et al., 2025[192]). Although it is unknown if the participants in these studies were positive for any islet autoantibodies when they were diagnosed (Brooks-Worrell et al., 2013[49]; Lomeli et al., 2025[192]; Moosaie et al., 2021[215]; Pietropaolo et al., 2000[246]); if they were, then they could have met the criteria for a diagnosis of T1D, LADA or LADY. Any subsequent development of islet autoantibodies in established T2D subjects warrants consideration if they should then be additionally diagnosed with T1D, LADA or LADY, or be classified as having 'double diabetes', especially if their diabetic phenotype has worsened in tandem; this is a plausible scenario, as studies have demonstrated that detection of humoral islet autoimmunity is associated with worse glycaemic control and increased likelihood of dependency on insulin treatment (Brooks-Worrell et al., 2013[49]; Lomeli et al., 2025[192]; Reed et al., 2021[262]). Surprisingly, one study found that islet-reactive T cells were present in > 60 % of 36 phenotypic T2D subjects, who were all obese and had no history of ketoacidosis or insulin treatment, and > 30 % were also positive for at least one islet autoantibody (Brooks-Worrell et al., 2011[51]). The presence of islet-reactive cytotoxic T cells correlated with diminished islet beta-cell function in this study, but autoantibody presence alone did not. The cytotoxic T cell reactivity to ≥ 4 islet autoantigens was considered positive for T1D-associated cytotoxic T cell autoimmunity here. Interestingly, only 7/36 subjects in this study tested negative for any T cell reactivity to islet antigens, with 7/36 testing positive for 2 or 3, but not considered positive for T cell T1D-associated autoimmunity.

A subsequent study found that phenotypic T2D subjects who tested positive for islet-reactive T cells had improved islet beta-cell function with rosiglitazone treatment, which coincided with down-regulation of islet-specific T cell responses (Brooks-Worrell and Palmer, 2013[50]). This further demonstrated that islet-reactive T cells promote islet beta-cell pathology in T2D. Seemingly healthy controls can test positive for T cell reactivity to islet antigens, even multiple, and more than T1D subjects, but it is currently unclear why/how they do not have the T1D phenotype (Herold et al., 2009[147]). Confusingly, a subset of T1D subjects here did not test positive for any T cell reactivity to islet antigens, but it is possible they could be positive for antigens not tested in that study. A more recent, larger-scale study found that cellular islet autoimmunity was present, with reactivity to ≥ 4 islet autoantigens, in 41.3 % of participants with T2D currently on metformin treatment, and humoral autoimmunity, positivity for GADA, IA-2A and ZnT8A, was present in 13.5 % (Brooks-Worrell et al., 2022[48]). Both humoral and cell-mediated autoimmunity were present in 5.3 % of participants in that study. In the same study, many participants had reactivity to 1-3 autoantigens but did not meet ≥ 4 to be determined as T cell positive. Interestingly, T cell-mediated autoimmunity, but not humoral autoimmunity, was also associated with worse islet beta-cell function and glycaemic control. The observation here that humoral autoimmunity did not appear to worsen disease phenotype contradicts findings from other aforementioned studies (Lomeli et al., 2025[192]; Moosaie et al., 2021[215]; Pietropaolo et al., 2000[246]). Rodent studies have also generated evidence for autoimmunity being involved in T2D pathogenesis, as B lymphocyte-deficient New Zealand obese mice, an obese polygenic mouse model of obesity and T2D, did not develop T2D (Haskell et al., 2002[141]), which implies that B lymphocytes are essential for T2D manifestation in these mice. Another study found that B lymphocytes from T2D subjects were unable to secrete the anti-inflammatory cytokine IL-10 and hypersecreted IL-8, suggesting that these cells encourage inflammation and worsen the disease phenotype (Jagannathan et al., 2010[160]).

These observations clearly demonstrate that T1D-associated autoimmunity is present in T2D and has some impact on disease severity. Although they raise some important questions, such as why T2D individuals with T1D-associated cytotoxic T cells do not have the more severe disease phenotype of T1D and whether LADA/LADY diagnosis also needs to take into account if individuals are positive for these T cells as well as autoantibodies. By currently unclear mechanisms, T2D individuals' islet beta-cells are seemingly able to avoid the severe autoimmune destruction usually brought on by T1D-associated cytotoxic T cells. Perhaps the increased expression of HLA seen in T1D does not occur in this scenario (Richardson et al., 2011[267]). Discovery of new autoantigens associated with diabetic autoimmunity could raise the detection rates of islet beta-cell autoimmunity in all forms of diabetes, in both the pre-clinical and clinical phases. The transient nature of islet autoantibodies (Achenbach et al., 2005[2]; Phillip et al., 2024[245]; Sørgjerd, 2019[290]; Vehik et al., 2016[321]; Williams et al., 2022[333]) needs to be considered when screening T2D subjects, since if an individual tests negative, this does not mean they have not been positive for one of these autoantibodies before and/or they may develop positivity in the future. Detection of islet autoreactive T cells can also be difficult to monitor in individuals, as 3/36 subjects who initially tested negative were found to be positive in subsequent checks over six months (Brooks-Worrell et al., 2011[51]). The current notion that subjects need T cell reactivity to ≥ 4 islet autoantigens in order to be determined as having T cell T1D autoimmunity needs revision, as this may underestimate the presence of islet autoimmunity, as ideally, no positivity should be present in individuals with no T1D-associated pathology. The observations that seemingly healthy controls can have positivity for islet antigens, even multiple, > 4, does not mean that pathology is not present, as these individuals might be in the preclinical diabetic phase, and may not ever progress to clinical disease for currently unclear reasons. Further understanding the regulation and contribution of islet beta-cell autoimmunity to disease pathogenesis and severity, as well as improving its detection, will likely enable the development of more effective therapeutic strategies for T1D, T2D and LADA/LADY.

Inappropriate glucagon secretion is well known to contribute to the disease phenotype in T2D (Reed et al., 2021[262] ). However, in recent years, inappropriate glucagon activity has also been found in T1D and has attracted scientific attention (Bisgaard Bengtsen and Møller, 2021[39]; Guo et al., 2021[133]). Similar to what is observed in T2D, T1D subjects exhibit inappropriate/excessive postprandial hyperglucagonemia, which further exacerbates hyperglycaemia (Guo et al., 2021[133]; Zhang et al., 2024[345]). One study found that postprandial plasma glucagon levels increased by 160 % five years after T1D diagnosis (Fredheim et al., 2015[119]). Another study found that T1D individuals with high postprandial glucose tended to have high postprandial glucagon levels (Urakami et al., 2011[317]). In contrast, during hypoglycaemia in some, typically long-duration T1D subjects, glucagon activity is absent and unable to raise blood glucose (Bisgaard Bengtsen and Møller, 2021[39]). Therefore, in T1D, excessive glucagon activity worsens the disease phenotype during the postprandial period and diminished glucagon activity is unable to raise blood glucose during hypoglycaemia (Bisgaard Bengtsen and Møller, 2021[39]; Guo et al., 2021[133]). Given the findings from various studies (Bisgaard Bengtsen and Møller, 2021[39]; Guo et al., 2021[133]; Reed et al., 2021[262]), T1D and T2D should be viewed as bihormonal disorders; further understanding how altered islet alpha-cell activity and glucagon secretion manifests mechanistically in these disorders will likely enable new therapeutic options.

Systemic insulin resistance is a well-established pathological feature of T2D (Reed et al., 2021[262]). However, insulin resistance has also been observed in people with T1D, and these individuals are classified as having 'double diabetes' (Bielka et al., 2024[38]). Insulin resistance manifests in an estimated 25-30 % of T1D individuals (Chaudhary et al., 2025[63]). It is not surprising that T1D subjects can develop systemic insulin resistance post-diagnosis due to complications associated with the intense insulin treatment and associated weight gain and/or if they adopt excessive diets/unhealthy lifestyles (Apostolopoulou et al., 2025[16]; Bielka et al., 2024[38]). The pathology that promotes insulin resistance in overweight T1D and T2D subjects is likely similar. Curiously, though, insulin resistance has also been observed in lean T1D subjects (Donga et al., 2013[99]), which suggests that the pathogenesis of insulin resistance cannot be attributed only to excessive weight gain. Understanding mechanistically how insulin resistance manifests in non-overweight T1D and T2D subjects may lead to the development of new therapeutic strategies. Insulin resistance manifestation during the pre-clinical phase of T1D has also been identified to encourage progression to the clinical phase, independently of other risk markers such as weight and age (Apostolopoulou et al., 2025[16]; Fourlanos et al., 2004[118]). These observations suggest that insulin resistance is also important in T1D pathogenesis as well as T2D. If induction of insulin resistance occurs in overweight and non-overweight individuals via similar mechanisms and has similar clinical implications in both major forms of diabetes before and after diagnosis, then this would further challenge the current distinction between them.

IAPP plaques are present in the islets of the majority of individuals with T2D and intracellular oligomers of IAPP are thought to promote islet beta-cell dysfunction in this disease (Reed et al., 2021[262]). Interestingly, one study found that 11 % of recently diagnosed T1D subjects had abnormally high plasma concentrations of IAPP (Paulsson et al., 2014[239]). A subsequent study then found that 2 out of 6 recent-onset T1D pancreas biopsies had both intra- and extracellular amyloid present, associated with dying islet beta-cells (Westermark et al., 2017[329]). Another more recent study identified that islet amyloidosis was present in all three donors with T1D (Beery et al., 2019[33]). Therefore, given these observations, it seems that islet amyloid is also present in at least a subset of T1D subjects and could contribute to disease pathology in a similar manner to T2D. Evidence has emerged that IAPP and its precursor forms contain several epitopes that act as autoantigens and can activate T cells in NOD mice and human T1D (Denroche and Verchere, 2018[93]), implying that islet IAPP encourages autoimmunity in the islets of T1D subjects. Genetic deletion of IAPP did not alter diabetes progression in NOD mice, indicating that this is not an essential autoantigen in T1D (Baker et al., 2013[26]). Developing treatments that prevent/degrade islet amyloid accumulation may assist in alleviating the T1- and T2D phenotypes by improving islet beta-cell function.

Overall, the global incidence and prevalence of AD have increased over recent decades (Li et al., 2022[186]; Safiri et al., 2024[271]). Globally, the number of dementia patients has increased by an estimated 117 %; in 2016, 43.8 million versus 20.3 million in 1990 (GBD 2016 Dementia Collaborators, 2019[123]). Incidence and prevalence rates increase with increasing age in dementia, and therefore, given the global rise in the ageing population and increased life expectancy, the disease burden is expected to massively increase in the future (Li et al., 2022[186]). By 2050, it is estimated that there will be 152 million people living with dementia globally (GBD 2019 Dementia Forecasting Collaborators, 2022[124]). Given that AD is the most common form of dementia and accounts for an estimated 60-80 % of all dementia cases (Kamatham et al., 2024[171]), it is a major global healthcare burden and a concern for the future. The presence of dementia in 2019 varied between regions/nations, with East Asia, North Africa and the Middle East all reporting an estimated > 700 cases and South Asia and Western Sub-Saharan Africa reporting an estimated < 500 cases per 100,000 (Safiri et al., 2024[271]). Interestingly, marked increases in dementia prevalence have been found in 2021 compared to 1990 in some countries (Safiri et al., 2024[271]; Zhi et al., 2025[349]), with the total prevalence rate increasing in China from 342.1/100,000 in 1990 to 1194.2/100,000 in 2021 (Zhi et al., 2025[349]). After age standardisation, the dementia prevalence rate in China increased by 30.6 % for females and 27.3 % for males in 2021. Figure 4(Fig. 4) (References in Figure 4: Safiri et al., 2024[271]) shows how the global prevalence of AD varied in 2019.

AD has a multifactorial and incompletely understood aetiology, involving genetics, ageing, the presence of other diseases and environmental factors (Breijyeh and Karaman, 2020[47]; Armstrong, 2019[18]; Safiri et al., 2024[271]). Ageing is considered the primary risk factor for AD, with an estimated 1 in 10 individuals aged ≥ 65 having the disease (Liu et al., 2024[190]). The incidence of AD doubles every 5 years after age 65, with an estimated one-in-three individuals aged over 85 developing the disease (Isik, 2010[156]). Approximately 95 % of AD patients are diagnosed after age 65 (Loi et al., 2023[191]), known as late-onset AD (LOAD), which clearly demonstrates an old age association with AD risk. However, AD can manifest in individuals as young as 40 and the prevalence and incidence of dementia manifesting in individuals under 65, known as early-onset AD (EOAD), has increased in recent decades (He et al., 2025[144]). Family history is the second greatest risk factor for AD, with an estimated 35-60 % of EOAD individuals having a first-degree relative with dementia (Safiri et al., 2024[271]). Twin and family studies indicate that genetic factors play a role in an estimated 80 % AD cases (Gatz et al., 2006[122]). EOAD is associated with variants of the amyloid precursor protein (APP), presenilin-1 (PSEN1) and presenilin-2 (PSEN2) genes, which account for 18-50 % of familial EOAD cases (Cai et al., 2015[56]; Tanzi, 2012[302]). LOAD risk is associated with the e4 allele of the apolipoprotein E APOE gene (Tanzi, 2012[302]), and homozygous subjects for e4 have an 8-12 fold increased risk of developing AD (Altmann et al., 2024[11]). These four genes account for 30-50 % of AD inheritability (Tanzi, 2012[302]). Although GWAS have identified more than 80 susceptibility loci (Andrews et al., 2023[15]), reported odds ratios for variants are modest, typically ranging from 1.1 to 2.4 (Andrews et al., 2023[15]; Shen and Jia, 2016[283]). Environmental and behavioural factors have also been identified to increase AD risk, including smoking, excessive alcohol consumption, sedentary lifestyle, and excessive diet/obesity (Safiri et al., 2024[271]). Lifetime smoking was associated with a 70 % greater risk of AD development (Barnes and Yaffe, 2011[29]; Durazzo et al., 2014[104]) and more physically active subjects had a 45 % reduced risk (Hamer and Chida, 2009[137]). Several studies have found that midlife obesity increases the risk of dementia and AD by ≥ 60 % (Albanese et al., 2017[8]; Kivipelto et al., 2005[175]). Health conditions such as cardiovascular disease (CVD) and diabetes also increase the risk of developing AD. Both T1D and T2D raise the risk of developing AD by around 50 % (Cholerton et al., 2016[68]; Li et al., 2025[184]). Hypertension and associated risk factors account for 2-9.6 % of dementia cases globally (Brain et al., 2023[46]). One study found that after 5.4 years, participants with CVD had a higher risk of dementia and AD than those without CVD (Newman et al., 2005[217]).

AD pathogenesis involves multiple intertwined mechanisms, including accumulation of amyloid-beta (Aß) plaques, neurofibrillary tangles (NFTs), neuroinflammation, synaptic dysfunction, and neuronal loss/atrophy (Kamatham et al., 2024[171]). AD is a prominent protein-conformational disease, largely induced by incorrect processing, altered conformations and polymerisation of normally physiologically beneficial proteins (Kamatham et al., 2024[171]; Tiwari et al., 2019[308]). AD is characterised histopathologically by extracellular aggregates of Aβ plaques and intracellular aggregations of NFTs, composed of hyperphosphorylated microtubule-associated tau (DeTure and Dickson, 2019[94]; Tiwari et al., 2019[308]).

Amyloid formation is thought to initiate AD pathology (Fan et al., 2020[112]; Kamatham et al., 2024[171]; Tiwari et al., 2019[308]). Aß plaques are thought to first manifest in the basal, temporal, and orbitofrontal neocortex regions of the brain, and as the disease progresses, these plaques also form throughout the neocortex, hippocampus, amygdala, diencephalon and basal ganglia, cerebellar cortex, striatum and thalamus (Planche et al., 2022[249]; Tiwari et al., 2019[308]). Aß aggregation is thought to cause neuronal damage/death by several mechanisms, such as excitotoxicity, synaptic disruption, oxidative stress, impaired intracellular transport and mitochondrial dysfunction (Kamatham et al., 2024[171]; Tiwari et al., 2019[308]; Zhang et al., 2023[347]). Interestingly, several studies have found that cognitively unimpaired elderly subjects can have significant Aß deposition, while others have found that a positive correlation exists between deposition and cognitive decline and dementia severity (Aizenstein et al., 2008[6]; Cummings and Cotman, 1995[81]; Dickson et al., 1992[95]; Nielson et al., 1996[220]; Wang et al., 2024[325]). It is currently unclear why Aß deposition promotes pathology in some but not others, and how the aberrant processing of amyloid precursor protein by Β-site APP cleavage enzyme 1 (BACE1) and γ-secretases to form Aß is initiated in the disease state remains mechanistically unclear (Aizenstein et al., 2008[6]; Jansen et al., 2015[162]; Kamatham et al., 2024[171]; Tiwari et al., 2019[308]). NFTs containing hyperphosphorylated tau are thought to form after Aß deposition, as the pathological Aß deposits cause kinases to be released, which then hyperphosphorylate tau and thereby cause it to be oligomerised (Kamatham et al., 2024[171]; Tiwari et al., 2019[308]). This then leads to the aggregation of insoluble tau filaments forming NFTs in neurons, leading to loss of signalling between neurons and impaired cytoskeletal function (Jiang et al., 2025[169]; Tiwari et al., 2019[308]). The tau aggregations are abundant components of NFTs found in the brains of people with AD (Kamatham et al., 2024[171]). Additionally, AD-associated tau has been suggested to act as a prion, whereby it induces misfolding of normal tau in neighbouring neurons, leading to progressive tau pathology across the brain (Clavaguera et al., 2015[69]; Kamatham et al., 2024[171]).

The Aß plaques and NFTs are also speculated to drive chronic neuroinflammation, which further induces neuronal damage and in turn, encourages further Aß deposition (Kamatham et al., 2024[171]; Tiwari et al., 2019[308]; Wong-Guerra et al., 2023[335]). The ongoing neuronal toxicity due to these plaques and NFTs induces migration of microglia and astrocytes into the affected areas of the brain, with the intention of these cells to remove the plaques (Deng et al., 2024[92]; Kamatham et al., 2024[171]). Aß plaques are known to activate toll-like receptors directly on microglia, causing the release of proinflammatory cytokines, and so these plaques directly promote neuroinflammation (Heneka et al., 2015[145]; Tiwari et al., 2019[308]). In AD, the immune response is thought to be pathological overall and fails to adequately clear the environment of plaques and NFTs, resulting in a chronic inflammatory state, which further promotes neurotoxicity (Kamatham et al., 2024[171]; Tiwari et al., 2019[308]; Wong-Guerra et al., 2023[335]). Mitochondrial dysfunction is also induced directly by the accumulation of Aß peptides (Kamatham et al., 2024[171]). Aß intracellular oligomers can accumulate within the inter-mitochondrial membrane, leading to reduced ATP synthesis and increased ROS production (Anandatheerthavarada et al., 2003[14]; Averchuk et al., 2023[23]; Kamatham et al., 2024[171]). Reduced ATP metabolically strains the neurons, and increased ROS levels promote cytotoxicity via oxidative stress (Kamatham et al., 2024[171]; Tiwari et al., 2019[308]). Aß-induced mitochondrial damage initiates a vicious cycle, as impaired mitochondrial function further promotes the Aß pathology (Anandatheerthavarada et al., 2003[14]). AD-associated tau impairs mitochondrial transport along neuronal axons, resulting in further oxidative stress and reduced energy production throughout the cell (Shahpasand et al., 2012[282]). In turn, mitochondrial dysfunction and oxidative stress then further encourage tau pathology (Du et al., 2022[102]). Additionally, oxidative stress also contributes to the formation of Aß plaques and NFTs, further exacerbating the neurodegeneration (Roy et al., 2023[269]).

Thus, the formation of Aß plaques initiates a cascade of pathological events that gradually impair neuronal functioning and insidiously promote neurotoxicity (Kamatham et al., 2024[171]; Tiwari et al., 2019[308]). The Aß plaques, NFTs, neuroinflammation and mitochondrial dysfunction interact synergistically to induce progressive AD pathology during the pre-clinical and clinical phases (Kamatham et al., 2024[171]; Parnetti et al., 2019[237]; Tiwari et al., 2019[308]; Zhang et al., 2024[344]). After initiation of AD pathology, the rate of progression can vary greatly between subjects, with some individuals remaining in the pre-clinical asymptomatic phase for > 20 years (Caselli et al., 2020[61]; Jia et al., 2024[167]; Zhang et al., 2024[344]). Additionally, the duration of mild cognitive impairment, which manifests prior to AD diagnosis, and clinical AD also varies substantially between subjects (Zhang et al., 2024[344]). Not all individuals with mild cognitive impairment progress to AD, which is not always progressive and is sometimes reversible, and can be caused by a variety of factors, such as diabetes and depression (Dunne et al., 2020[103]; He et al., 2024[143]; Salemme et al., 2025[273]). Diagnosis of AD occurs through a combination of cognitive tests, blood tests, brain scans and physical examination, reviewed elsewhere (Bomasang-Layno and Bronsther, 2021[41]; Safiri et al., 2024[271]). AD progresses from the mild stage associated with memory loss, speech impairment and mood changes to the severe stage where patients' motor functions and day-to-day activities are impaired, and eventually patients become bedridden and need assistance with basic tasks such as eating (Clifford et al., 2024[71]; Kamatham et al., 2024[171]). Typically, patients survive 3-10 years after diagnosis, and life expectancy is influenced by age at onset, in addition to other factors such as existing comorbidities (Brück et al., 2025[53]; Zanetti et al., 2009[342]). Death in AD subjects usually arises from the manifestation of other age-related diseases and/or from post-diagnosis complications such as pneumonia (Manabe et al., 2019[201]; Todd et al., 2013[309]; Tolppanen et al., 2020[310]). Figure 5(Fig. 5) (References in Figure 5: Averchuk et al., 2023[23]; Deng et al., 2024[92]; DeTure and Dickson, 2019[94]; Du et al., 2022[102]; Jansen et al., 2015[162]; Jiang et al., 2025[169]; Kamatham et al., 2024[171]; Planche et al., 2022[249]; Roy et al., 2023[269]; Shahpasand et al., 2012[282]; Tiwari et al., 2019[308]; Wong-Guerra et al., 2023[335]; Zhang et al., 2023[347]) summarises AD pathogenesis.

JR was a recipient of the Knowledge Economy Skills Scholarship (KESS) II UK PhD studentship. We thank the members of the VK laboratory.

The authors declare that they have no conflict of interest.

Artificial intelligence was not used in the preparation of this manuscript.

JR conceptualised and produced the draft manuscript and figures. SCB and VK contributed equally to reviewing and editing the draft manuscript. All authors have reviewed and approved the final version of the manuscript.