|Home | About | Journals | Submit | Contact Us | Français|
Alzheimer's disease is a devastating disease, not only for the patient but particularly for their families and their loved ones. What typically begins innocently enough as fairly subtle memory loss, progresses relentlessly over a period of approximately 7−10 years, until it ultimately erodes all higher cognitive functions and robs us of what many of us hold most dear, our identity, our ability to interact with the outside world. Currently, estimates indicate that over 27 million individuals are affected by the disease worldwide (1). In the United States alone, over 4 million individuals have the disease (2). Unfortunately, unless a cure is found, or a means to otherwise prevent this disease or significantly slow its progression, the number of affected individuals in the United States is expected to triple by 2050 due to the aging of the baby boomer generation(2). This enormous increase in the number of affected individuals is likely to have dire consequences on the already overburdened health care system in this country.
Based on the numbers alone, the identification of novel therapeutic and/or preventative agents is clearly of considerable importance. To rationally develop these agents, an understanding of the etiology and pathogenesis of this complex disease is necessary. Towards this end, there have been many hypotheses advanced over the last century ranging from abnormal phosphorylation of tau, unconventional infectious agents, trace element neurotoxicity, growth factor deficiency, excitatory amino acid insult, altered calcium homeostasis, free radical toxicity, and deficits in energy metabolism, to altered protein processing resulting in abnormal Aβ accumulation (reviewed in part by Markesberry, in (3). Ultimately, any plausible hypothesis concerning the etiology and pathogenesis of AD must take into account not only the neurological and neuropathological features of AD but also the known genetic risk factors and causative mutations, and the heightened risk associated with advanced age.
One hypothesis that attempts this, and that has received significant attention particularly over the last two decades (for better or for worse), has been the amyloid hypothesis (4-6). While many iterations of this hypothesis currently exist, simply put the initial hypothesis stated that the “deposition of amyloid β protein (AβP), the main component of the plaques, is the causative agent of Alzheimer's pathology...”(5). In this chapter we will examine the evidence for this hypothesis, as well as its potential limitations particularly for the development of novel therapeutic and preventative agents.
The first evidence for the amyloid hypothesis came from neuropathological assessments of brains isolated from patients with Alzheimer's disease. In the earliest neuropathological examinations published by Alois Alzheimer himself, he described the neuropathology in two patients(7-9). The examinations revealed a diffuse atrophy primarily of the cerebral cortex. Staining of the brains isolated from these two patients demonstrated the presence of two types of lesions. The first type, now known as neurofibrillary tangles (NFTs), was only observed in the initial patient examined. This lesion was described to be a twisted coil of fibrils derived from degenerating cerebral cortical cells. The second type of lesion, now known as senile plaques, was present in both of the cases examined, although to differing degrees. These plaques were found throughout the cerebral cortex and were characterized by a central core surrounded by a more diffuse halo.
We now know that these classical senile plaques are complex, extracellular lesions. They are generally associated with degenerating neuronal processes, have activated microglia intimately associated with the central deposit, and are surrounded by reactive astrocytes. These deposits are found throughout the neocortex and hippocampus in patients with Alzheimer's disease(10). The central deposit in classical senile plaques is structurally similar to the deposits seen in a group of diseases referred to as amyloidoses wherein there is extracellular deposition of proteins with a beta-pleated sheet conformation (reviewed by Sipe, 1992(11), 1994(12)). Over 15 different polypeptides have been identified as the primary proteinaceous components of the amyloids that are deposited in various tissues in the clinically diverse amyloidoses.
The central location of the plaque core within this pathology led to the speculation that whatever comprises the core may play a pivotal role in the disease process itself. In a landmark finding in 1984, Glenner and Wong published the purification and sequence of the primary proteinaceous component of amyloid isolated from meningeal vessels obtained from AD brains(13). By comparing samples from 6 AD cases and 3 controls, they identified a unique protein present in the AD cases. Fractionation of this protein on a sizing column gave a peak corresponding to an approximate molecular weight of 4200 daltons and amino acid analysis and sequencing revealed a novel amino acid sequence that we now refer to as the β-amyloid peptide or Aβ.
It had been previously established that individuals with Down syndrome who live past the age of 50 have neuropathological changes like those in Alzheimer's disease (reviewed by Mann, 1988(14)). In a follow up to their original finding, Glenner and Wong isolated and analyzed the cerebrovascular amyloid from patients with Down syndrome(15). In this report they established that the amino acid sequence of cerebrovascular amyloid in Down syndrome is essentially identical to that observed in patients with AD. Given the similarity of the amyloid deposited in AD and Down syndrome, Glenner and Wong proposed that there was a common pathogenic process involved. Since Down syndrome results from trisomy of the 21st chromosome, this meant that AD pathology could be produced by increased expression of a gene or genes on chromosome 21 (15).
Following these initial reports, amino acid sequencing of the amyloid isolated from senile plaques themselves, obtained from AD and Down syndrome brains, was reported by other groups(16,17). These reports established that the amino-terminal sequence and amino acid composition of plaque core amyloid is essentially identical to that of cerebrovascular amyloid isolated from AD or Down syndrome brains, except for the presence of ragged NH2 termini(17).
To isolate the gene encoding Aβ, Kang and colleagues utilized degenerate primers targeted against amino acids 10−16 of the peptide to screen a cDNA library constructed from the brain of a five month old fetus(18). In these experiments they isolated a clone encoding a protein 695 amino acids long that contained the Aβ sequence beginning 99 amino acids from the carboxyl end of the protein. This protein is now known as β-amyloid protein precursor (βAPP). Essentially simultaneous reports from other groups also identified βAPP as the precursor to Aβ(19-21). Southern blot analysis of mouse/human cell hybrids revealed that the gene encoding βAPP is located on the 21st chromosome(18). As discussed above, Glenner and Wong had suggested that overexpression of a gene or genes on chromosome 21 should be sufficient to cause AD pathology, based on the similarities of neuropathology between individuals with Down syndrome and typical late onset Alzheimer's disease. Now it has been determined that the gene encoding the Aβ peptide, that is deposited as an invariant feature in all forms of Alzheimer's disease, is localized to chromosome 21 where increased expression could presumably increase the amount of Aβ present. Subsequent studies by Tamaoka's group showed that Aβ levels were significantly increased in plasma isolated from patients with Down syndrome when compared to control individuals, indicating that an increased copy number of βAPP does result in increased levels of Aβ in humans(22).
A basic description of the metabolism of βAPP is necessary to understand how the familial Alzheimer's disease (FAD) linked mutations, discussed in the following section, can influence the accumulation of Aβ. The Aβ peptide sequence is embedded in the βAPP protein, indicating that 2 separate proteolytic cleavages are required to generate Aβ from its precursor. The N-terminus of Aβ is generated by cleavage of βAPP by β-secretase, producing a 99 amino acid C-terminal fragment (CTF) of βAPP that can be further cleaved by γ-secretase to release Aβ. γ-Secretase generates 2 major Aβ species, 40 and 42 amino acids in length, termed Aβ40 and Aβ42. βAPP can also be cleaved within the Aβ domain by α-secretase, an action which precludes Aβ generation. The β- and γ-secretase cleavages will be discussed further as we describe the discovery of the 3 genes linked to familial early-onset AD and the mechanisms by which they elevate Aβ levels.
Perhaps some of the strongest evidence for a critical role for Aβ in Alzheimer's disease has come from an analysis of the genetic mutations that cause Alzheimer's disease. In addition to trisomy 21 causing neuropathology that is essentially identical to that seen in typical late onset Alzheimer's disease, Alzheimer's disease can be inherited as a fully penetrant, autosomal dominant trait in certain families (reviewed by (23), (24), (25), and (26). In these families, the clinical and neuropathological presentation of the disease is essentially identical to typical late onset Alzheimer's disease, but the age of onset is earlier, typically in the 50s. Mutations in three distinct genes, on three separate chromosomes have been identified that cause Alzheimer's disease in these families. These are the βAPP gene on chromosome 21(27-31), the presenilin 1 gene on chromosome 14(32), and the presenilin 2 gene on chromosome 1(33). Each of these is reviewed in greater detail in the chapter by Taner. For the sake of the discussion of the amyloid hypothesis, however, it is worthwhile reviewing some of these mutations here.
The first mutation shown to cause Alzheimer's disease was a point mutation in the βAPP gene itself that was found in a single family. This mutation results in a substitution of the more hydrophobic amino acid isoleucine for valine at position 717 (V717I), which is immediately carboxyl to the Aβ sequence(28). In other families, additional mutations at this position were subsequently identified which result in the substitution of phenylalanine (V717F)(27) or glycine (V717G)(29) for valine. Following the identification of mutations at position 717 in the βAPP gene, a double mutation at position 670/671 was identified in a large Swedish family with a mean age of onset of 55 years(30). The 670/671 double mutation results in a substitution of asparagine and leucine for the lysine and methionine immediately preceding the N-terminus of Aβ (K670N/M671L). In context, the identification of causative mutations for Alzheimer's disease, not only within the βAPP protein itself but immediately adjacent to the cleavage sites needed to liberate the Aβ peptide from its precursor protein provided additional, immediate support for the amyloid hypothesis.
To investigate the hypothesis that the mutations identified in the βAPP gene would alter the amount of Aβ peptide being produced, several groups turned their attention to the analysis of βAPP metabolism and extracellular Aβ accumulation in model systems (34-37). Analysis of the concentration of total Aβ in the conditioned medium of transfected cells expressing these FAD-linked mutations indicated that the Swedish mutation caused a several-fold increase in the amount of Aβ accumulated extracellularly(34,35). Analysis of the C-terminal fragments and secreted forms of βAPP (sAPP) in cells transfected with the Swedish mutation showed elevations in CTFβ and sAPPβ indicating that the increased Aβ concentration observed with this mutation is likely due to enhanced β secretase cleavage(38). Using the same experimental paradigm, no significant differences in total Aβ, CTFs, or sAPPs were observed, however, in cells transfected with the 717 mutations(35).
Pioneering work by Peter Lansbury and his colleagues being conducted at that time had shown that the carboxy-terminal length of the Aβ molecule was critically important in determining the rate at which Aβ fibrils form(39-41). Using synthetic peptides, they showed that Aβ ending at position 42 formed fibrils far more rapidly and at lower concentrations than Aβ ending at position 40. As the deposition of Aβ in the form of amyloid fibrils represents an invariant feature of AD, Younkin and colleagues proposed that the 717 mutations might be acting to selectively increase secretion of Aß42 (35). In a landmark finding Younkin's group showed that secreted Aβ42, which normally constitutes only a fraction of total secreted Aβ, is significantly increased in the medium of cells expressing the 717 mutations(37). Thus, both the Swedish mutation and the 717 mutations increase the concentration of Aβ, in particular Aβ42.
Perhaps one of the greatest tests to date of the amyloid hypothesis came with the analysis of mutations that also cause early onset Alzheimer's disease, but that do not reside in the βAPP gene, or even on chromosome 21 for that matter. These were the presenilin mutations. Again, these are covered in detail in the proceeding chapter by Taner. In context, the reason why these mutations were so critical of a test for the amyloid hypothesis was that there was no a priori reason to think that they had anything to do with βAPP processing. In fact, they were just as likely to directly influence tau, synapse loss, energy metabolism, or a host of other possibilities that would be consistent with alternate theories regarding the etiology and pathogenesis of Alzheimer's disease.
That is not, however, how they would turn out. In fact, studies performed by Younkin and colleagues showed that in both cultured medium from primary fibroblasts and plasma isolated from patients with either presenilin 1 or presenilin 2 mutations, Aβ levels were elevated, particularly Aβ42 levels, similar to the 717 mutations in βAPP. Follow-up studies by a number of groups examining the influence of these mutations on Aβ levels in either transfected cells or in the brains of animals transgenic for these mutations confirmed these changes.
When the presenilins were discovered as FAD-linked genes in 1995, their functions were unknown and most AD researchers could not have fathomed that they would soon be found to be intimately associated with APP metabolism. In 1997, Selkoe and colleagues showed that APP and presenilin interact in mammalian cells, as evidenced by their co-immunoprecipitation(42). In fact, over the next 5 years the work of a number of labs demonstrated conclusively that the presenilins are the catalytic component of the multi-protein complex that is γ-secretase(43-45). The relationship between βAPP, the presenilins and AD is now clear: Aβ is generated by β-secretase and γ-secretase (presenilin complex) cleavage of the βAPP protein. All of the mutations identified in βAPP, presenilin 1, and presenilin 2 that cause early-onset familial Alzheimer's disease give rise to increases in Aβ levels, particularly Aβ42 levels, or otherwise perturb the ratio of Aβ42 to Aβ40 levels (neuron paper/Golde paper) in ways likely to foster Aβ aggregation and deposition.
Aging is clearly the most significant risk factor for Alzheimer's disease, and Aβ levels begin to increase in the brains of many cognitively normal people between the age of 40 and 80 (46,47). According to the study of consecutive autopsy cases by Funato et al., insoluble Aβ42 in particular accumulates with age in the cortex and precedes senile plaque formation (46). Compared to normal aged brain, AD brain had higher levels of soluble and insoluble Aβ42, Aβ40, and a higher degree of N-terminally truncated or modified Aβ. Similar correlations between Aβ levels and age in cognitively normal individuals were reported by Morishima-Kawashima and colleagues, with significant increases in Aβ accumulation beginning after the age of 40 (47). In both studies, insoluble Aβ concentration was logarithmically related to plaque density, and a critical threshold of about 100 pmol/g insoluble Aβ42 was required for immunocytochemical detection of senile plaques. In the latter study, carriers of the apolipoprotein E ε4 allele, a strong risk factor for AD, were found to accumulate Aβ at an earlier age than non-carriers (47).
As already discussed, elevations in Aβ concentration that are likely to enhance aggregation and deposition have been detected as a consequence of expression of all of the FAD-linked mutations analyzed to date, and in Down syndrome. Importantly, these elevations can been detected in plasma and in fibroblast conditioned medium isolated from presymptomatic individuals (36,48) and in transgenic animals prior to deposition (49,50) suggesting that these changes are early and are not simply an epiphenomenon associated with end-stage Alzheimer's disease. In addition it appears that Aβ levels increase during aging both in humans and in animal models, with age being the largest risk factor for the development of the disease.
The question then becomes whether or not these elevations in Aβ play a central, causal role in the pathogenesis of the disease or whether these changes merely represent a very good marker of the underlying disease process, that in and of itself is relatively benign. Any well intentioned observer will recognize that both of these possibilities remain open today, and probably will until multiple approaches to lower Aβ levels either fail or show significant improvement in the clinics. Having said that, there have been a great number of studies attempting to show that these elevations in Aβ levels are not likely without consequence.
The extent of correlation between the neuropathological lesions in Alzheimer's disease patients and the severity of dementia has been an area of considerable debate and continues to consistently be used as an argument against the amyloid hypothesis. As is true with any correlative function, a correlation can be a good indicator of a causal relationship, but close correlation is certainly not definitive proof of causality. For example, a well correlated change can simply be an inconsequential, reliable biomarker of another process that is causative. With that in mind, some of the earliest correlative studies showed significant correlations between plaque numbers and the extent of dementia (51). Several other studies, however, have reported that the number of neurofibrillary tangles and neuropil threads are a far better indicator of the degree of dementia in at least cases of “pure AD”(52,53). One of the more comprehensive recent analyses undertaken, at least with respect to the extent of variables examined, was published by Cummings and Cotman(54). In this study they found that the number of plaques, neurofibrillary tangles, and dystrophic neuritis all significantly correlated with dementia severity, as well as the area occupied by Aβ and PHFs. Having said that, it is clear that there are at least some individuals that have extensive amyloid deposition that are cognitively normal. For example, in a study by Markesbery's group, significant AD-like pathology (both plaques and tangles) were found in the brains of a substantial number of elderly, well educated, cognitively normal individuals(55). Studies such as these have led some to argue that the amyloid hypothesis must therefore be wrong. In response, some amyloid theory proponents have adjusted the hypothesis accordingly to accommodate and now argue for pre-amyloid like aggregates of Aβ, such as Aβ oligomers, as the causative agent in the disease process. Regardless of which turns out ultimately to be correct, it is important to point out that the development of Alzheimer's disease is clearly a reasonably long process and with nearly one half of the population susceptible to the disease if they live long enough, it is perhaps not surprising that individuals can be found with significant neuropathology that are cognitively normal. This is certainly the case in other neurodegenerative diseases as well, such as Parkinson's disease where approximately 70% of the dopaminergic neurons in the subtantia nigra are lost prior to the development of clinical symptoms.
If alterations in Aβ are necessary and sufficient to play a causal role in AD pathogenesis, then Aβ should be able to elicit, either directly or indirectly, the neuropathological and cognitive changes observed in patients with Alzheimer's disease. Furthermore, mechanisms must exist that can explain the prevalence of the disease in the aging population and in people carrying causative mutations and known genetic risk factors. Evidence gathered over the last several years has continued to build an increasingly stronger case that the alterations in Aβ observed in the genetic forms of AD are not likely without consequence and can potentially account for many of these observations. In this section we will review evidence for the neurotoxicity of abnormal Aβ species. While neurotoxicity was initially attributed to the fibrillar species of Aβ deposited in plaques, recent data implicates soluble Aβ oligomers as well, which may form prior to plaque deposition and cause neuronal dysfunction that may facilitate many of the downstream pathological events in AD. Since these soluble oligomers exist in equilibrium with fibrillar Aβ as deposition progresses, the neuronal loss, inflammation, and other pathology seen in the vicinity of plaques may be due to the oligomers, the plaques, or a combination of the two.
Soluble, synthetic Aβ peptides were shown by Yanker et al. to be neurotrophic at low concentration to undifferentiated hippocampal neurons in culture, and toxic at higher concentrations to mature neurons (56). Subsequently, the neurotoxicity of Aβ was shown to be dependent on its aggregation state (57,58). Stable Aβ aggregates were highly toxic to primary neurons, and partial reversal of aggregation resulted in a loss of toxicity. Similar results were found in in vivo studies, with microinjection of fibrillar, but not soluble, Aβ causing neurotoxicity in the cerebral cortex of aged rhesus monkeys (59). Interestingly, neurotoxicity was dependent not only on the aggregation state of Aβ, but also on both the age and species of the animal model used. Plaque-equivalent concentrations of fibrillar Aβ resulted in extensive neuronal loss, as well as tau phosphorylation and microglial activation in the brains of aged monkeys, but were not toxic to young adult monkeys or aged rodents. Much higher concentrations of Aβ were required to elicit neurotoxicity in young adult monkeys and in rodents (59-61). These results may help to explain the vulnerability of the elderly to AD, as well as the difficulty of generating a rodent model that faithfully reproduces all of the neuropathological features of the disease.
In vitro, fibrils are believed to form via the progression from Aβ monomers to low molecular weight oligomers, to intermediate species called protofibrils that assemble into mature fibrils (62). The data indicating that Aβ fibrils are neurotoxic and can elicit other AD characteristics including tau phosphorylation led to the hypothesis that disrupting fibrils might be therapeutically beneficial. But because the disruption of insoluble Aβ fibrils could result in an accumulation of protofibrils and other soluble oligomers, experiments were carried out to investigate whether these lower-level aggregates were apparently non-toxic, like Aβ monomers, or whether they might elicit neurotoxic effects, like fibrils. Data generated over the past several years has convincingly demonstrated that not only are Aβ oligomers neurotoxic, in many assays they are even more toxic than fibrils (63).
Soluble oligomers range from dimers and trimers to dodecamers, also called Aβ-derived diffusible ligands (ADDLs) (64,65). The smaller SDS-stable oligomers are produced by a number of cell lines and have been detected in human brain and cerebrospinal fluid. Similarly, the larger ADDLs are not merely an artifact of the in vitro assembly of Aβ, as structurally indistinguishable Aβ oligomers are present in soluble extracts of AD brain at average levels 12-fold higher than in control brains (66). ADDLs, whether formed in vitro or purified from AD brain, bind specifically to synapses in differentiated hippocampal neuronal cultures (63). This evidence for specific neuronal attachment, coupled with the fact that both ADDLs and lower molecular weight Aβ oligomers have been shown to be potent inhibitors of long-term potentiation (LTP), a model of synaptic plasticity and memory, provides a rational explanation for early memory loss in AD, and in animal models of AD as well (64,67,68) .
As previously discussed, a common criticism of the amyloid hypothesis was that in some studies, plaque burden correlated poorly with severity of dementia in AD. The discovery of soluble oligomers as neurotoxic Aβ species led to an examination of the relationship between soluble Aβ concentration and clinical and pathological severity. The result was a strong correlation between soluble Aβ and markers of disease severity including synaptic loss (69,70). Two additional lines of evidence support the hypothesis that soluble Aβ oligomers are the primary toxic entity in the brain, at least in animal models.
First, impaired synaptic transmission and cognitive function are seen prior to overt amyloid deposition in mouse models of AD (71-73). In the widely used APP transgenic mouse model Tg2576, a partial decline in memory occurs around 6 months, prior to amyloid deposition. Cognitive function then remains stable over the next 7−8 months, even though plaque deposition progresses and becomes significant over this time period. Finally, a further decline in cognitive function is detected at ages greater than 15 months. The initial memory decline followed by the period of stability was perplexing in terms of the lack of correlation with the course of amyloid plaque deposition in this model. This led Lesné, Ashe, and colleagues to conduct a detailed biochemical analysis of Aβ complexes in the brains of these mice during the time period when the first behavioral deficits are detected (72). Soluble, extracellular-enriched extracts from the forebrain of 6-month-old Tg2576 contained SDS and urea stable Aβ complexes with molecular weights theoretically corresponding to trimers and multiples thereof, up to a molecular weight of 56 kDa. Only the 56 kDa (theoretical dodecamer) and 40 kDa (theoretical nonamer) species appeared for the first time at 6 months. Both correlated inversely with memory performance, with the 56 kDa form (termed Aβ*56) showing the strongest correlation. The levels of both the 40 and 56 kDa Aβ complexes remained stable on average during the subsequent period of cognitive stability in the mice. To more directly test whether Aβ*56 causes cognitive impairment, the complexes were purified from Tg2576 brain extracts and were then injected into the lateral ventricle of rats. Aβ*56 caused a transient decrease in spatial memory in rats, supporting the hypothesis that this complex could be responsible for the onset of memory deficits in the Tg2576 mouse model (72). Whether Aβ*56 is structurally identical to the 56 kDa ADDLs derived from AD brain (66) is an intriguing question that remains to be determined.
Second, therapeutic interventions which lower the level of soluble Aβ or disrupt Aβ assembly in animal models, often in the absence of detectable changes in plaque load, ameliorate cognitive deficits (74-79). This effect is not unique to a single therapeutic approach, and has been observed with such divergent strategies as Aβ immunization, acute γ-secretase inhibition, and oligomer neutralization. Recently, Lee and colleagues showed that short-term passive immunization of aged Tg2576 APP transgenic mice with a conformation specific Aβ antibody that preferentially recognizes dimers, soluble oligomers, and certain amyloid deposits, resulted in significant improvements in spatial learning and memory without affecting amyloid burden (77). These results are similar to those obtained by independent groups using different Aβ antibodies and different transgenic lines (75,76), and support the hypothesis that the neutralization of toxic Aβ species can reverse cognitive deficits in mice. This hypothesis has also been tested by McLaurin, St. George-Hyslop and colleagues using a completely different experimental paradigm, but with similar results. Cyclohexanehexol stereoisomers, which inhibit Aβ aggregation and favor the disassembly of fibrils, can prevent Aβ-oligomer induced toxicity in cultured primary neurons and hippocampal slices, and oligomer-induced memory deficits in rats (79). When administered orally to TgCRND8 APP transgenic mice from 6 weeks of age (pre-deposition) to 4−6 months (significant amyloid deposition), scyllo-cyclohanehexol showed a dose-dependent improvement in spatial learning accompanied by decreases in both amyloid burden and Aβ oligomers (78). Synaptic loss was ameliorated at 6 months, as was accelerated mortality in the treated mice.
Perhaps the most important implication of these studies is that the cognitive impairment in these models is not permanent. To what degree this applies to the human condition is unknown, since the profound neuronal loss in AD is absent in AD mouse models. Nonetheless, reducing soluble Aβ levels or altering a toxic conformation may be a less ambitious goal than clearing plaques. The true test for the amyloid hypothesis of AD, as well as the specific notion that soluble oligomers mediate Aβ toxicity, awaits the further development of Aβ-targeted therapies and their progression to clinical trial.
This work was supported by Grants from the National Institutes of Neurological Diseases and Stroke (NIH 5R01NS042192 and NIH 5R01NS048554) and by the Mayo Foundation for Medical Education and Research.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.