hIAPP has a strong tendency to form toxic oligomers and fibrils that lead to pancreatic β-cell apoptosis and eventually the onset of T2DM 
. It is clear from our results that BPA promotes hIAPP aggregation in a dose-dependent manner, which is supported by the significantly accelerated aggregation lag time as well as the enhanced fluorescence intensity that reflects the orderly β-structures formed (). CD data further confirms the accelerated transition of hIAPP from unordered structure to β-structure in the presence of BPA (Fig. S2
). The helical intermediates are thought to play a role in hIAPP aggregation 
, the transition from helix to β-sheet structure may cause the reduction of helical structures as what we observed (Fig. S2
It is well established that the toxic hIAPP oligomers can disrupt the islet β-cell membrane and lead to permeabilization. In MTT study, we observed the cytotoxicity of hIAPP on INS-1 cells rose sharply with the addition of BPA (). The observed strong cytotoxicity by BPA alone also agree with previous studies which suggested BPA itself also disrupt the cell function through stimulating the estrogen-receptor and several other apoptosis-related pathways 
. It is interesting to note that the increased toxic effect of hIAPP in combination with BPA was only observed at high BPA to hIAPP ratios but not at the lower ratios. The CDI was calculated to explore the potential BPA and hIAPP interaction, and it was found that the two compounds showed a synergistic exacerbation of cototoxicity especially at high molar ratios of BPA.
Further analysis was conducted to distinguish the direct molecular toxic effects of BPA on live cells from its interaction with hIAPP. Dye leakage assays were performed to monitor exclusively the membrane disruption property of hIAPP in the presence of BPA. The hIAPP oligomers have been proven to bind and penetrate membranes more efficiently than monomers and are regarded as an important causative factor to β-cell death 
. Serious dye leakage was observed in the hIAPP group and was dose-dependently enhanced in the presence of BPA (). It is interesting to note that no evident hIAPP aggregation was identified within a short incubation time at a low concentration (1 µM) as suggested by the ThT and CD results (data not shown), whereas hIAPP at this concentration immediately caused significant membrane disruption in dye leakage assays. This may be explained by a recent report that hIAPP form oligomers much faster in the presence of membrane structures 
. These data suggested that BPA significantly increases the ability of hIAPP to disrupt membranes.
Oxidative stress induced cytotoxicity was another mechanism underlying amyloid-related β-cell apoptosis besides direct membrane disruption. Amyloid formation has been reported to associate with ROS generation 
. hIAPP oligomers may form pores on membrane and lead to permeabilization of lipid bilayers 
. The generalized increase in membrane permeability results in intracellular calcium elevation which disrupts mitochondrial function and finally increases ROS generation 
. In this study, ROS accumulation is also observed in INS-1 cells treated with hIAPP. The ROS levels rose significantly in the presence of BPA and hIAPP, while BPA by itself has little effect on ROS levels (), suggesting BPA has a synergistic effect on the ROS production related to hIAPP amyloid formation. A summary of possible molecular scheme is provided for the toxic effects of BPA on the formation of hIAPP amyloid which further result in cell damage ().
A schematic representation of hIAPP aggregation pathway.
The influence on human health of BPA exposure is regarded as an accumulative process because of its widespread penetration in daily life 
. The BPA tolerable intake has been set as 50 µg/kg/day 
, but adverse effects at lower BPA concentrations in animal studies have been demonstrated, which may lead to the requirement of a new risk assessment for BPA 
. BPA concentrations in human blood (serum and plasma) are in the range of 0.3–4.4 ng/ml (1.3–19.4 nM) in developed countries 
. In contrast, physiological circulating concentrations of hIAPP are below 10 pM in fasted non-diabetic people and rise up to over 20 pM after a meal 
, suggesting the physiological ratio of BPA to hIAPP may actually be much higher than those used in the present study. Moreover, since BPA exposure is a continuous and accumulative process, it is logical to expect long-term BPA exposure may be accompanied with accelerated hIAPP amyloid formation and β-cell apoptosis, and eventually a higher risk of T2DM. In a recent report, BPA at near physiological concentration also showed direct toxicity 
. Due to the detection sensitivity limitation of existing biophysical technologies, in the present study, we tested the interaction between hIAPP and BPA at much higher concentrations in vitro
. Therefore the system we used may be considered as a model that simulates physiological interactions at accelerating rates. Future biophysical study with novel experimental methods which can tackle hIAPP and BPA interaction at physiological conditions will be important.
In summary, our data provide the evidence that BPA exposure concentration-dependently accelerates the toxic amyloid formation, exacerbates the toxic membrane disruption of hIAPP and promotes the levels of toxic ROS generated by hIAPP in vitro. Our study suggest that in addition to direct biological effect, long-term BPA exposure may also have adverse effects on hIAPP amyloid formation that eventually contribute to the onset of T2DM. The results may provide a new angle on how BPA exposure influences the risk of diabetes from hIAPP aggregation related pathogenesis. Moreover, since BPA also possess other important biological effects including estrogen-like function, it will be interesting to explore the effect of BPA exposure on physiological functions of hIAPP, for example, hIAPP secretion and hIAPP related insulin resistance. In addition to hIAPP, a great variety of amyloidogenic proteins such as amyloid β peptide and α-synuclein, are known to form extracellular amyloid deposits that induce human diseases including Alzheimer's disease and Parkinson's disease 
. It will be of future interest to study how BPA exposure may affect the misfolding of those amyloidogenic proteins.