Is islet amyloid polypeptide indeed expressed in the human brain?

This study aims to study the association between pancreatic islet amyloid polypeptide (IAPP) and Alzheimer's disease neuropathological change (ADNC) in brain biopsies obtained from subjects with idiopathic normal pressure hydrocephalus (iNPH) and in post‐mortem (PM) brain samples obtained from aged individuals.


INTRODUCTION
Alzheimer's disease (AD) is the most common clinical diagnosis of dementia. Age, sex, educational level, apolipoprotein ε4 status, and cardiovascular disease are some risk factors associated with AD [1].
Islet amyloid polypeptide (IAPP), also called amylin, is a hormone synthesised and secreted by pancreatic β-cells [8,9]. IAPP is involved in insulin and glucagon homeostasis, regulation of bone mass, and blood pressure as well as satiety and gastric emptying. In pathological conditions, IAPP is the main component of islet amyloid, contributing to β-cell loss and finally T2DM. This is due to the molecular structure of IAPP and its ability to form β-sheets that aggregate into amyloid fibrils [8][9][10]. These features are also common to other amyloidforming proteins in humans [10].
Amyloid β (Aβ), one of the elements of AD neuropathological change (ADNC), is prone to aggregate within the grey matter of the brain as well as in the meningeal and cortical blood vessels [11]. The level of Aβ pathology is not associated with cognitive impairment, as it can be detected in the brains of young subjects as well as in aged individuals that do not experience cognitive decline and dementia [6,12,13].
During the last decade, several groups have studied the association between Aβ and IAPP, as both proteins show amyloidogenic properties and are primarily detected in the aged [2,14]. Molecular interaction and the cross-seeding mechanism between Aβ and IAPP have been described in in vitro studies and animal models [15][16][17].
Surprisingly, some studies describe associations between IAPP and hyperphosphorylated τ (HPτ), the other hallmark lesion of ADNC [19][20][21]. The presence of Aβ and HPτ deposits in pancreatic β-cells has also been described in a few studies but has not been reproduced by others [15,19,22,23].
Idiopathic normal pressure hydrocephalus (iNPH) is a neurological condition, affecting older individuals, caused by altered cerebrospinal fluid (CSF) circulation, presenting with cognitive impairment, gait disturbances, and urinary frequency and incontinence. The only treatment strategy available is a ventriculoperitoneal shunt (VPS) insertion that normalises the CSF flow and reverses the symptoms [24]. During such curative surgery, a brain biopsy can be obtained from the area of the shunt channel. Numerous studies describe ADNC in biopsies from iNPH patients; moreover, when present, it is associated with a worse shunt response and progression to AD [25][26][27][28][29][30].
The aim of this study was to assess the presence of IAPP in brain tissue obtained during VPS insertion from iNPH subjects with and without ADNC. This allowed us to assess the protein content in the human brain tissue without changes caused by the agonal state, PM delay (PMD), or a long fixation time. These results were compared with those observed while assessing the pathology in PM brain tissue from aged subjects with or without ADNC.

Study subjects
iNPH subjects The brain biopsies were obtained from the right frontal lobe during curative VPS insertion as previously described [24,28,30] [31]. The subjects included those displaying none or various levels of Aβ and/or HPτ in their brains. Two samples, measuring 2 mm in diameter, were obtained from the amygdala of each brain following a previously described method [32]. The TMA blocks were cut into 4 μm thick sections for further analysis.
To validate the protocols for the IHC stains while using IAPP antibodies (Abs), a tissue sample was obtained during the autopsy from the pancreas of an 85-year-old male.

Immunohistochemistry
The IHC stains were performed using the automatic platform, Dako Autostainer Plus (Agilent Technologies, Santa Clara, CA, USA) with the Dako EnVision Flex detection system according to the manufacturer's instructions. The Abs and the pretreatment strategies used are summarised in Table 1.

Key points
• Several studies have suggested an association between type 2 diabetes mellitus (T2DM) and Alzheimer's disease (AD) as the pancreatic islet amyloid polypeptide (IAPP) has been detected in human brains.
• We studied the expression of IAPP and its association with AD markers in curative, surgical brain biopsies from subjects with idiopathic normal pressure hydrocephalus and in post-mortem (PM) brain samples obtained from aged individuals.
• There was no specific expression of IAPP in human brain tissue in either of our cohorts; hence, there was no association between IAPP and Alzheimer's disease neuropathological change.
• Non-specific IAPP reactivity was observed due to a methodological pitfall.
• According to our results, after studying the surgical and PM brain samples, there is no association between AD neuropathological change and IAPP.

Assessment of the samples
All samples were assessed using light microscopy at Â20 to Â400 magnification. The pathology within a sample was assessed and dichotomised as present or not, except for the polyclonal IAPP Ab, where the compartmentalisation of the staining was also noted.

Statistical analysis
The statistical analyses were performed using IBM SPSS statistic software, version 28 (IBM Corp., NY, USA). To describe the cohort, means and standard error of means (m ± SE) were used. Non-parametric test, Mann-Whitney U test (MWU), was applied to study the differences between the groups. The correlations between the different variables were defined using a non-parametric Spearman's rho two-tail test.

RESULTS
The IAPP Abs were tested on PM pancreatic tissue from an 85-year-old male with cardiovascular disease, cerebrovascular disease, renal failure, and a clinical diagnosis of dementia but no known T2DM. The cause of death was cardiac arrest. The PMD was 24 h.
The pancreas displayed extensive autolytic changes, with only a few viable β-cell islets. Immunoreactivity for both monoclonal and polyclonal IAPP Abs was seen in the pancreatic β-cells; however, the polyclonal Ab also displayed background staining, as illustrated in Figure 1.   is, no significant correlation (p = 0.071, Spearman's rho = 0.2). Contrary to the above, no IAPP reactivity was seen in any of the samples when using the monoclonal IAPP Ab, as shown in Figure 2D.

iNPH subjects
In total, 113 samples from the iNPH subjects were included in  Figure 2E.

DISCUSSION
Several studies have suggested an association between T2DM and AD in vitro, in animal models and in PM human tissue [15][16][17][18][19]. Here, for the first time, we assessed the expression of the pancreatic protein IAPP in surgical brain biopsies obtained during a curative shunt insertion procedure from iNPH subjects with or without ADNC.
Additionally, we evaluated the IAPP expression in PM brain samples from the amygdala region, from 77 subjects with or without ADNC.
In our iNPH cohort, including subjects between 54 and 86 years old, Aβ was present in 50% and HPτ in 47% of the cases, and combined Aβ/HPτ pathology was seen in 32% of the cases. The frequency of the pathology is congruent with what has previously been described in the majority of iNPH cohorts [28][29][30]33]. These results, however, are lower compared to our previous studies due to the different assessment strategies and the selection criteria of the cohort [27,34].
In the TMA, including 77 CS from the amygdala, Aβ pathology was detected in 69% of the cores and HPτ in 91%. This is in line with the stepwise progression of the pathologies, as the Aβ primarily affects the cortical regions of the brain; when seen in the amygdala, the pathology has already reached Thal phase 2 [11]. In contrast, the HPτ can be seen in the amygdala early on in the disease process and can be detected in younger individuals [13,35].
We assessed the immunoreactivity of two different IAPP Abs, one monoclonal (clone: R10/99) and one polyclonal (T4149), with the latter being used in previous studies assessing PM brains [18,19]. The performance of these Abs was tested on pancreatic autopsy tissue.
Despite a defined agonal state due to cardiac arrest and a PMD of 24 h, the pancreas displayed extensive autolysis, with only a few remaining β-cell islands. Both Abs labelled pancreatic islands. The polyclonal Ab also displayed background staining. The agonal state is a factor to consider when assessing PM tissue, as prolonged death is associated with a lower pH, hence increased autolysis [36].
Post-mortem, the body's organs undergo autolysis caused by the deposition of proteolytic enzymes. The pancreas is recognised as undergoing autolysis soon after death, as it is rich in digestive enzymes, resulting in morphological changes and altered protein expression as judged by IHC [37,38]. In line with the literature, our case displayed marked autolysis within the pancreatic islet 24 h after death [37]. In Sweden, an autopsy cannot be performed without consent from a close relative; thus, the PMD is often 1 day or more.
The information regarding the agonal state has not been provided in previous publications, and PMD is only given in one study assessing the IAPP pathology in the pancreas and brain [18,19].
No IAPP pathology was seen within the brain tissue in either the iNPH or the PM cohorts when using the monoclonal IAPP Ab. Of note, when applying the polyclonal IAPP Ab, different expression was seen in various cellular compartments and structures within the PM brain tissue. In line with what has been reported previously, the IAPP labelling was seen as granular structures in the neuronal soma, extracellular plaque-like aggregates as well as inclusions in the vessel walls [15,18,19]. Since we could not detect any specific IAPP expression with the monoclonal Ab, the positive structures visualised by the polyclonal Ab have to be interpreted as non-specific.
IHC is used daily in research and clinical diagnostics [39,40].
There are numerous pitfalls to consider when using IHC. Factors such as PMD/cold ischaemia time, type of fixative, fixation time, type of paraffin, storage time, section thickness, antigen retrieval technique, mode of staining (manual or automatic), and detection system should be considered when performing and assessing the IHC staining outcome [41,42]. Additionally, the choice of an antibody is crucial.
Monoclonal Abs are highly specific, as they recognise only one epitope on an antigen, thus reducing the risk of non-specific binding and hence immunoreactivity. In contrast, the polyclonal Abs recognise several epitopes, resulting in a higher affinity to an antigen but also non-specific staining caused by binding to similar epitopes leading to what is in general known as cross-reactivity [41].
In a study from 2021, Rees and colleagues revealed strong crossreactivity of several polyclonal IAPP Abs with a neuropeptide, calcitonin gene-related peptide (CGRP), which is widely expressed in the brain and shares 40%-50% of the amino acid sequence with IAPP [43].
Previous studies assessing the IAPP expression in PM brain have used polyclonal IAPP Abs [15,18,19]. The polyclonal IAPP Ab used by us and in the previous studies on this topic (T4149, Peninsula Laboratories) recognises the full length of the IAPP protein, thus including the sequence recognised in CGRP leading to labelling that is interpreted, by us, as non-specific [43]. Of note, the monoclonal IAPP Ab used by us (clone R10/99) recognises an 11-amino acid sequence (7-18) of the IAPP protein.
We assessed the performance of a monoclonal Ab, which provides specific labelling of IAPP. IAPP monoclonal Ab staining was observed in the pancreas as expected. However, there was no staining at all in the surgical brain biopsies, which lacked the pre-analytical alterations caused by PMD and long fixation time. IAPP labelling with the monoclonal was also absent from the PM brain samples, altered by various pre-analytical factors. The ADNC was detected in both our cohorts and, according to our results, did not influence the IAPP immunoreactivity of either of the Abs used. As polyclonal Abs are more unreliable, the expression seen in our PM cohort was interpreted as non-specific. Thus, based on our results, we cannot confirm that specific IAPP expression is observed in brain tissue or vascular structures, as previously reported.
IAPP, like insulin, can cross the blood/brain barrier, but a seeding mechanism of pathologic IAPP protein from the blood/plasma to the brain was not confirmed in an animal model, further questioning the reported role of IAPP pathology within the brain [44,45].
Several studies performed in vitro or in animal models have suggested a connection between AD and T2DM and have reported an association between IAPP and Aβ or even in a single case with HPτ [2,16,17,21,46]. These studies are interesting on a molecular level, as interactions between the different altered proteins promoting the pathological processes are increasingly recognised in neurodegenerative diseases [47]. IAPP is an amyloidogenic protein, prone to aggregate and with a similar structure to other amyloids. This makes it an interesting protein to study with regard to protein alteration and seeding [10]. The connection between IAPP and ADNC has been seen in in vivo studies as well as in animal models and the PM human brain, albeit using the polyclonal IAPP Abs; thus, the results are not reliable [15,16,18,19,46].
The association of T2DM and a clinical diagnosis of dementia of AD type has also been suggested in several epidemiological studies of aged individuals. However, this association has not been confirmed in large neuropathological cohorts when a definite diagnosis of the pathological type of dementia was available [4-7, 14, 48]. One explanation for the outcome might be the cognitive decline caused by mixed pathologies that increases with age and is not acknowledged until the final PM examination of the brain is carried out [6,47,49]. T2DM is associated with vascular disease and vascular changes within the brain, but it has not been associated with ADNC in numerous PM studies [4][5][6][7]14]. One study revealed that the only brain proteinopathy associated with pancreatic IAPP in T2DM subjects was transactive DNA binding protein 43, the cause of limbic-predominant age-related TDP encephalopathy [23,50]. This is certainly intriguing as it connects T2DM with cognitive impairment and brain pathology.

IS IAPP INDEED EXPRESSED IN THE BRAIN?
In summary, we had the opportunity to assess the expression of IAPP in surgical brain biopsies of 113 iNPH subjects. This is unique, as most studies regarding IAPP in the brain are performed on PM tissue, animal models or in vitro. Here, we could assess protein expression in the human brain without pre-analytical confounding factors affecting the epitopes to be labelled. We also assessed the IAPP in a cohort including 77 PM samples from the amygdala region, as previous studies have been performed on PM brains. Both cohorts included samples with and without ADNC. There was no immunoreactivity for IAPP in either of the cohorts when using the monoclonal IAPP Ab. In contrast, the IAPP protein reactivity was visualised in different structures within the brain samples when using the polyclonal Ab, in line with what has been reported in previous publications [15,18,19]. We interpreted the reactivity of the polyclonal IAPP Ab as non-specific due to the total lack of IAPP reactivity while using the specific monoclonal Ab. Of note, polyclonal IAPP Abs have been reported to cross-react with a neuronal protein, leading to false-positive outcomes [43].
In conclusion, we could not verify any association between ADNC and IAPP in brain tissue in either of our cohorts; hence, there seems to be no support for a causative association between ADNC and T2DM, in line with previous neuropathological studies [4][5][6][7]. An association has been reported based on molecular studies and in animal models, but not when studying human brain tissue. A significant factor leading to a false-positive outcome may be the choice of the Ab. A reliable, specific Ab is crucial when assessing a surgical specimen as the interpretation of the outcome can affect the final diagnosis and thereby the choice of treatment. To produce reliable results, pre-analytical and analytical variables have to be taken into consideration [41].
Both authors read and approved the final manuscript.

CONFLICT OF INTEREST STATEMENT
The authors have no conflict of interest to report.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

ETHICS STATEMENT
The study has been approved by the regional Ethical Committee of