Respiratory infections and type 1 diabetes: Potential roles in pathogenesis

Abstract Among the environmental factors associated with type 1 diabetes (T1D), viral infections of the gut and pancreas has been investigated most intensely, identifying enterovirus infections as the prime candidate trigger of islet autoimmunity (IA) and T1D development. However, the association between respiratory tract infections (RTI) and IA/T1D is comparatively less known. While there are significant amounts of epidemiological evidence supporting the role of respiratory infections in T1D, there remains a paucity of data characterising infectious agents at the molecular level. This gap in the literature precludes the identification of the specific infectious agents driving the association between RTI and T1D. Furthermore, the effect of severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infections on the development of IA/T1D remains undeciphered. Here, we provide a comprehensive overview of the evidence to date, implicating RTIs (viral and non‐viral) as potential risk factors for IA/T1D.


| INTRODUCTION
Type 1 diabetes (T1D) is a chronic autoimmune condition affecting over nine million worldwide, 1 characterised by the loss of functional pancreatic islet β-cells. This ultimately results in the lifelong dependency on exogenous insulin. [2][3][4] Although the pathophysiology of T1D is well characterised and understood, its aetiology remains unclear. However, it is well established that the mechanisms underlying the development of T1D is multifaceted and likely involves the complex interplay between genetic and environmental factors. 2,5,6 Among the environmental factors associated with T1D, infections with viruses are identified as prime candidate triggers of islet autoimmunity (IA) which precedes most clinical onset of T1D.

| VIRAL AETIOLOGY OF TYPE 1 DIABETES
The reduced prevalence of T1D-associated high-risk human leucocyte antigen (HLA) genotypes among newly diagnosed individuals, increasing global incidence of T1D, 6-9 seasonal variations 6,10 and geographical differences 6,11 in genetically similar individuals as well as the convergence of IA/T1D incidence of migrants to their new country of residence 12,13 all strongly support the growing contribution of environmental factors in the pathogenesis of T1D.
Several hypotheses have been proposed on how environmental factors may influence the progression of T1D. The 'β-cell overload' hypothesis postulates that factors increasing insulin demand such as infection, growth, trauma and other physiological stresses may result in β-cell dysfunction and insulin resistance, instigating and accelerating the development of IA/T1D. 6,[14][15][16] The 'hygiene hypothesis' conversely states that a decrease in childhood infections due to improved hygiene may increase the incidence of autoimmune diseases like T1D. 6,9 The hygiene hypothesis proposes that a lack of childhood infections can limit immune system's exposure to various microorganisms and stunt its development, leading to an inappropriate response to future infections that may cause T1D. 9 Another hypothesis, the 'polio hypothesis', suggests that the decreasing incidence of certain virus infections over time (such as enterovirus or poliovirus infections) has increased the proportion of infants who become infected in the absence of maternal antibodies that could protect against that virus, increasing the risk of complications such as β-cell damage and T1D. 17,18 Among the environmental factors associated with T1D to date, viral infection has been investigated most thoroughly and hypothesised as the prime trigger of IA and progression to T1D, especially in utero and during childhood. 2,13 This is supported by a large body of molecular 6,19,20 and epidemiological [21][22][23][24][25] evidence, and multiple nonmutually exclusive mechanisms have been proposed to explain how viral infections can induce and/or accelerate the development of IA/ T1D. [26][27][28][29] To date, multiple viruses have been associated with T1D. Of the viruses investigated, enteroviruses (EV) have been the most deeply studied and now widely accepted as the prime candidate trigger of IA/T1D. 2,[29][30][31] In total, over 26 different EV types have been associated with IA/T1D, mostly comprised of Enterovirus B (EV-B) species members within the coxsackievirus B and enteric cytopathic human orphan virus (ECHO virus) groups. 2,32 EVs have been detected more frequently in the blood, 22 gut 33,34 and pancreas 24,35,36 of individuals with T1D compared to without, and are associated with an increased risk of T1D in prospective studies. [37][38][39]

| RESPIRATORY TRACT INFECTIONS AND ISLET AUTOIMMUNITY/TYPE 1 DIABETES
Although most research to date on the infectious aetiology of IA/T1D have focussed heavily on viral infections in the gut and pancreas, 2,5,30 respiratory tract infections (RTI), particularly within the first 12 months after birth, 26,[40][41][42] have also been investigated as a potential risk factor for childhood T1D. Both lower RTIs (including pneumonia, bronchitis and bronchiolitis) and upper RTIs (including rhinitis, pharyngitis and laryngitis) have been examined by at least 19 observational studies as potential triggers for IA/T1D development ( Figure 1, Table 1).
Three retrospective case-control and cohort studies reported a significant association between RTIs and T1D, 26,43,44 while two reported no association. 44,45 Limited sampling methods and heterogeneity in study design between studies may have contributed to inconsistent results. These studies relied on insurance claims or medical consultation data to ascertain RTI exposure, which only capture clinically overt symptomatic infections. Hence, such studies are likely to have underestimated the cumulative exposure to RTIs.
Only one retrospective study included molecular testing to confirm the infectious agent, reporting a significant association between laboratory confirmed pandemic influenza A (H1N1) and T1D, but not between clinically diagnosed H1N1 and T1D. 44 All these studies lacked IA testing, precluding the examination of IA as an outcome associated with RTIs.
Prospective birth cohort studies investigating IA as an outcome have reported that early-life RTIs increased the risk of IA. 40,41,[46][47][48] These studies followed genetically at-risk children from birth (as whereas a more developed pancreas would be less susceptible. 53 However, as no other studies have replicated these results, external validation in other prospective cohorts with maternal data and respiratory samples collected longitudinally during pregnancy is needed. Specific respiratory viruses including parechoviruses and influenza virus have been associated with T1D in retrospective studies and animal studies. One mouse study found an association between a strain of parechovirus (Ljungan virus) and T1D. 54 While one Japanese retrospective cohort study reported an increased risk of T1D after the diagnosis of influenza, 55 and an Italian study found increased incidence of T1D diagnoses during the 2009 H1N1 pandemic, 56 most observational studies did not find an association between influenza 57-60 or parechoviruses 61 and T1D in humans. In addition, many EV species replicate in the respiratory tract, and the most common manifestation of EV infection is a common cold-type disease. These EV species include rhinoviruses which are responsible for over 50% of all RTIs, 2 EV-B, 62 and members belonging to Enterovirus C 63 and D, 64    Recently, a machine learning approach was used to rank tissuespecific transcription regulatory effects for single-nucleotide polymorphisms in T1D associated genes, estimating their relative contributions to the development of T1D by integrating T1D case and autoantibody-negative control genotypes with tissue-specific quantitative trait loci (eQTL) data. 70 The investigators found that the largest gene regulatory contribution to the risk of T1D development was made by the rs6679677 eQTL, which is associated with changes to AP4B1-AS1 transcript levels in lung tissues. Therefore, the strongest tissue-specific eQTL effects associated with T1D risk occurred in the lung, supporting the potential contribution of respiratory infections on the development of IA/T1D.

| CORONAVIRUS INFECTION AND TYPE 1 DIABETES
Severe acute respiratory syndrome related coronavirus 2 (SARS-  The relationship between COVID-19 and T1D remains a poorly understood and rapidly evolving area of research, with its long-term diabetogenic effects likely to be unknown until after many years of extensive research. To this end, a global registry (CoviDiab) was established to investigate their interaction. 103 Long-term prospective analysis is needed to decipher any relationship between COVID-19 and T1D.

| CONCLUSION
There is an enormous body of accumulated evidence, both molecular and epidemiological, that support the hypothesised role of viral infections in the development of IA and T1D. By comparison, there remains a major gap in understanding and paucity of data, especially molecular data where infectious agents are characterised at the nucleic acid or protein level, that elucidates the relationship between RTI and IA/T1D. To address this gap, the use of comprehensive metage detection methods, and the prospective collection of respiratory samples and IA testing during pregnancy and early life in large prospective cohorts such as the ENDIA, 68  conceptualisation; preparation; writing; editing.