Small organic probes as amyloid specific ligands – Past and recent molecular scaffolds

1. Introduction 

Protein misfolding diseases, also referred to as amyloidoses or proteinopathies, are diseases of disparate etiologies characterized by extracellular or intracellular proteinaceous deposits in tissues and organs [1]. These deposits, termed “amyloid” or inclusion bodies, result from misfolding and/or partial unfolding of proteins, followed by the formation of protein aggregates. At least 25 different proteins have been reported to form disease-associated amyloid in vivo [1], and there is evidence that any given polypeptide can be induced to form amyloid in vitro under appropriate conditions [2]. From a biophysical perspective, amyloid deposits consist of fibrils with a diameter of 7–10nm. These fibrils consist of a number of protofilaments and structural details of the fibril morphology can be visualized by transmission electron microscopy (TEM) or atomic force microscopy (AFM) [3], [4]. Structural studies of amyloid with high-resolution methods, such as X-ray diffraction, have shown that the protein or peptide molecules are arranged so that the polypeptide chain forms β-strands that run perpendicular to the long axis of the fibril, and novel structural insight and molecular details have been provided by solid-state nuclear magnetic resonance (NMR) spectroscopy and by single crystal X-ray diffraction analysis of small amyloid-like peptide fragments [5], [6], [7], [8]. A particularly significant characteristic of the structures determined with these techniques is that they are remarkably similar even for polypeptides having no sequence homology, suggesting that many amyloid fibrils share similar core structures. Hence, most peptides form comparable aggregated β-sheet rich fibrillar assemblies, although heterogenic structures for specific types of fibril can be observed as the alignment of adjacent strands and the separation of the sheets might be slightly different due to the specific nature of the side-chain packing.

The shape shifting nature of proteins or peptides when being converted to amyloid, i.e. going from a native fold to β-sheet rich fibrillar assemblies, makes it hard to specifically identify amyloid with conventional immunochemical reagents such as antibodies. Firstly, it is hard to separate the aggregated form of the protein from the native protein as most antibodies recognize a specific sequence of a distinct protein independently of the fold of the protein. Secondly, most antibodies poorly enter the compact beta-sheet accumulations of amyloid and many amyloids also incorporate immunoglobulins and complement-derived opsonins. Hence, antibodies can give rise to both false-negative and false-positive stains, leading to misdiagnoses of amyloidoses [9], [10]. However, conformational antibodies specific for a diversity of aggregated states of proteins, including soluble amyloid oligomers, fibrillar oligomers or amyloid fibrils, have recently been reported [11], [12], [13], [14]. Such conformational specific antibodies show great promise for being used in clinical diagnostics of amyloidoses and for studying the underlying molecular and pathological events of these diseases.

A second class of molecules that can be used for visualization and identification of protein deposits is small hydrophobic agents. The most common small amyloid ligands are derivatives of Congo red or Thioflavins and in contrast to antibodies these dyes do not bind to specific proteins but are rather selective for protein aggregates having an extensive cross β-pleated sheet conformation and sufficient structural regularity. Molecular scaffold based on these conventional amyloid ligands will be described in more detail in the following section. In addition, a novel class of amyloid imaging agents, luminescent conjugated polythiophenes (LCPs), will also be discussed in more detail, as these novel dyes seems to identify and differentiate a broader subset of protein aggregates than conventional dyes.

2. Conventional amyloid ligands 

Congo red (Fig. 1), an aromatic sulfonated azo dye, was introduced more then 80 years ago and its gold-green birefringence under polarized light has been the gold standard for amyloid detection ever since [15], [16]. Nowadays, a large number of amyloid ligands have been synthesized and most of these agents are derivatives of Congo red or Thioflavine. As shown by Klunk and co-workers, the key structural feature of Congo red might be the two acidic functional groups and the spacing between them [17]. Chrysamine G (Fig. 1), a lipophilic analogue of Congo red which also has two acidic functional groups with the same amount of spacing between them as seen for Congo red, has also shown high binding affinity to protein aggregates [18]. Chrysamine G has also been evaluated for in vivo imaging of amyloid but brain uptake of the compound in mice was limited [19]. As a consequence, considerable efforts to develop novel derivatives with high blood–brain-barrier permeability based on the Chrysamine G or Congo red scaffolds have been made. In this regard, Methoxy-X04 (Fig. 1), an optimized Chrysamine G derivative which lacks the carboxylic acid groups, has proven to be a good candidate for performing in vivo imaging of protein aggregates in transgenic mouse models by multiphoton microscopy [20]. Removing the carboxylic acid groups had little effect on the binding affinity for the protein aggregates, showing that the planar hydrophobic backbone is crucial for binding to amyloid. Furthermore, the brain uptake of [11C]MeO-X-04 was shown to be sevenfold higher than that of the related carboxylic acid derivative Me-X-34 [20]. However, the level of brain uptake of [11C]Me-X04 was still insufficient for using the ligand in positron emission tomography (PET) studies in humans. Hence, Congo red and Chrysamine G derivatives have shown to be less than ideal from a clinical perspective and further optimization of these derivatives is limited, as the molecular frameworks of these dyes are rather rigid. As a consequence, recent research has shifted towards the utilization of other types of molecular frameworks for the development of amyloid imaging agents.

View Large Image Download to PowerPoint

Fig. 1. Examples of different amyloid imaging agents. Molecular scaffolds generate from Congo red (left panel), Thioflavine T (middle panel) and miscellaneous molecular scaffolds (right panel).

Thioflavine T (ThT, Fig. 1) and Thioflavine S (ThS) are other organic dyes often used for in vitro characterization of amyloid fibril formation or histopathological staining of amyloid plaques [21], [22]. Although ThT contains a positively charged benzothiazolium unit, whose ionic charge is unfavorable for brain uptake, several amyloid imaging agent utilizing ThT as a scaffold have been developed. Klunk and co-workers developed a series of neutral ThT derivatives containing uncharged benzothiazole instead of benzothiazolium for PET imaging [23], [24], [25]. All of these neutral derivatives showed higher binding affinities towards protein aggregates than the charged compound ThT. One of these compounds, Pittsburgh Compound-B (PIB) (Fig. 1), has also been evaluated in human patients with Alzheimer’s disease (AD) and the results suggested that PET imaging with PIB provides quantitative information on amyloid deposits in living subjects [25]. Other neutral ThT derivatives for single photon emission computed tomography (SPECT) imaging have also been developed by Kung and co-workers (TZDM, Fig. 1) [26]. These derivatives showed specific binding to protein aggregates at subnanomolar concentrations and this study also included iodinated analogues based on the Chrysamine G scaffold [26]. The same research group has also continued to develop a variety of ThT derivatives where the benzothiazolium is replaced with benzoxazole [27], benzofuran [28] or imidazopyridine (IMPY, Fig. 1) [29].

In addition to the Congo red and Thioflavine derivatives mentioned above, other chemical scaffolds have been utilized for the development of amyloid specific ligands [30]. Kung and co-workers have reported novel amyloid ligands based on a stilbene scaffold (m-I-stilbene, Fig. 1) that could be used for visualization of protein aggregates with PET or SPECT imaging [31], [32], [33]. Furthermore, vinylbenzoxazole derivatives were recently reported as promising candidates for imaging of amyloid [34], [35]. These derivatives resemble both ThT and stilbene amyloid ligands as they contain not only a benzoazole and an aromatic ring in their structures (ThT derivatives), but also a double bond between them (stilbene derivatives). One of the amyloid ligands reported, [11C]BF-227 (Fig. 1), demonstrated high binding affinities for protein aggregates, excellent brain penetration and rapid clearance from the brain, suggesting that this amyloid ligand may be a promising candidate for use in clinical diagnostics of proteinopathies [35]. A fluorinated derivative of 2-(1-(6-(dimethylamino)naphthalen-2-yl)ethylidene) malononitrile ([18F]FDDNP, Fig. 1) has also been shown to be a promising PET radioligand for imaging AD brain pathology [36]. The binding affinity of this compound to protein aggregates was also excellent and competition assays with other conventional amyloid ligands indicated that the binding site for FDDNP on protein aggregates was different from those of Congo red and ThT. However, this indication of a high affinity binding site separated from ThT has been contradicted by three subsequent studies [37], [38], [39]. Furthermore, a variety of molecular frameworks, including derivatives of fluorene, biphenyl thiophene, and biphenyltriene, have also been utilized for the development of amyloid imaging agents [30]. Interestingly, a novel attractive molecular scaffold including two thiophene moieties were recently reported and these amyloid imaging ligands were designed for use in near infra red fluorescent imaging of protein aggregates [40], [41]. Indeed, molecular scaffolds consisting of repetitive thiophene moieties seem to have an excellent affinity for protein aggregates. Specifically LCPs, have recently proven to be a novel interesting class of fluorescent molecules for investigating protein aggregates and proteinopathies.

3. Luminescent conjugated polythiophenes (LCPs) 

Electro- and optically active conjugated polymers are a class of molecules that has been utilized for a wide range of applications, such as solar cells and light emitting diodes. Additionally, the molecular similarity between conjugated polymers and biological polymers offer a great possibility to create simple versatile biosensors, as these two classes of molecules are able to form strong complexes with each other due to multiple non-covalent interactions [42]. Many of these biosensors exploit the conjugated polymer’s inherent optical properties and to achieve such polymers, e.g. polymers exhibiting fluorescence with high quantum efficiency, a polymer backbone consisting of substituted thiophene rings or fluorene building blocks is preferable.

The optical processes in conjugated polymers are highly influenced by the conformation of the polymer backbone and the separation and aggregation of polymer chains. The optical transitions of conjugated polymers are believed to occur on different parts of the same polymer chain, intrachain events, or between adjacent polymer chains, interchain events. The intrachain events are mainly dependent on the conformation of the polymer backbone and the interchain processes occur as nearby polymer chains come in contact with each other leading to stacking of the aromatic ring systems, such as the thiophene rings. These phenomena were recently shown for a LCP (also reported as conjugated polyelectrolyes, CPEs) with a zwitter-ionic side chain functionality, POWT [43]. Due to a change of the net charge of the side chain, the conformational freedom of the thiophene backbone was restricted, leading to different emission profiles from the LCP. The same principle has also been utilized for assigning distinct conformations of synthetic peptides [44]. Nilsson and co-workers used the optical properties of POWT for spectral discrimination of synthetic peptides with opposite charge having a random coil conformation or a four-helix-bundle motif. Hence, the flexible thiophene backbone was restricted upon binding to a distinct conformational state of a molecule. The ability to afford an optical fingerprint corresponding to a distinct conformational state of a specific molecule sets LCPs apart from conventional fluorescent dyes. Most conventional techniques are limited by their reliance on detecting a certain biomolecule, whereas the LCPs are identifying a specific structural motif or a distinct conformational state of a biomolecule. Hence, the LCPs offer the possibility to monitor the biochemical activity of biological events on the basis of a structure–function relationship rather than on a molecular basis. The unique conformational-sensitive optical properties of LCPs have proven to be a great asset for studying protein misfolding and aggregation, as binding of LCPs to distinct forms of protein aggregates gives rise to specific optical features for the LCPs. This is an improvement compared to the other small amyloid imaging agents discussed earlier, as these probes only change in optical feature whether they are free in solution or upon binding to amyloid fibrils.

4. LCPs for exploring protein aggregation and amyloid fibril formation in vitro 

LCPs with their unique structural related optical properties have proven to be a powerful tool to study amyloid fibrillation events in vitro. Recently, novel methods for the detection of formation of amyloid fibrils in bovine insulin and chicken lysozyme based on conformational changes of the anionic polythiophene derivative, PTAA (polythiophene acetic acid, Fig. 2), were reported [45]. Depending on the conformation of the protein, different emission spectra from PTAA were observed. Upon binding to the native monomeric form of insulin, PTAA emitted light with emission maximum of 550 nm, whereas PTAA bound to amyloid fibrils of insulin emitted light with lower intensity and the emission maximum was red-shifted to 580 nm. Hence, a specific optical fingerprint was achieved for the β-sheet containing amyloid fibrils. Furthermore, when plotting the ratio of the intensity of the emitted light at 550 nm and 580 nm, the formation of insulin fibrils could be monitored. The kinetic plot showed a lag phase, followed by a growth phase and a plateau phase, which are characteristic for the formation of amyloid fibrils when Congo red absorbance or ThT fluorescence are used as the read-out [45]. The detection of insulin fibrils was also observed by absorption and visual inspection, which might be useful for the development of simple screening methods for the detection of amyloid fibrils [45].

View Large Image Download to PowerPoint

Fig. 2. Chemical structures of luminescent conjugated polythiophenes (LCPs) and examples of their use as amyloid imaging agents. (a) Chemical structures of the first generation (left panel) and second generation (right panel) of LCPs. (b) Spectra of polythiophene acetic acid (PTAA) (left) and tPOWT (right) bound to native insulin (blue) or insulin amyloid fibrils (red). (c) Spectrum (left) and fluorescence image (right) of PTAA bound to amyloid deposits in tissue samples. Some representative PTAA stained amyloid deposits are indicated with white arrows. (d) Fluorescence images of prion deposits, CWD (left) and sheep scrapie (right) being stained by PTAA. Some representative PTAA stained prion deposits are indicated with white arrows.

Under the conditions used [45], PTAA was interacting with both the native form of insulin and the fibrillar form of insulin. Although, these forms were easily distinguished by distinct fluorescent signatures from PTAA, LCPs selective for amyloid fibrils would be favorable. A second generation of LCPs was recently synthesized by Konradsson and co-workers [46], [47]. Except for having ionic side chain functionalities on all thiophene moieties, these molecules also include unsubstituted thiophene rings which give rise to a greater conformational freedom of the polythiophene backbone (Fig. 2). This molecular scaffold was achieved through controlled synthesis of a symmetrical trimer building block that was further assembled through randomized polymerization. Interestingly, the chain length distribution of these randomly polymerized materials was shown to be rather narrow and around 80–90% of the material had a well-defined chain length [46], [47]. A zwitter-ionic LCP, PONT, also known as tPOWT (Fig. 2), showed a huge increase in the emission intensity and a blue-shift of the emission maximum to 560 nm upon binding to insulin amyloid fibrils [46]. The spectrum for PONT mixed with native monomeric insulin showed a weaker intensity with a maximum at 600 nm and resembled the spectrum for PONT free in solution, indicating that the interaction between PONT and native insulin was absent. The kinetics of the amyloid fibrillation was also followed by plotting the ratio of the intensity of the emitted light at 560 and 600 nm. Similar to the observations with PTAA, the characteristic three different phases of amyloid fibril formation were seen [46]. The increased amyloid specificity for the second generation of LCPs was also confirmed in a recent study comparing the monomer-based LCPs with their trimer-based counterpart [47]. Hence, by chemical design of a trimeric thiophene scaffold, LCPs with enhanced discrimination between the amyloid-like fibrillar state and the corresponding native protein was achieved. So far, LCPs have been used to distinguish between the native form of proteins and the amyloid fibrillar form of proteins. However, further chemical development of novel LCPs might offer a novel approach to discriminate between different conformational structures observed during the amyloid formation processes.

5. LCPs for histological characterization of protein aggregates including amyloid 

As described above, LCPs can be utilized as molecular tools for studying the amyloid fibril formation in vitro. However, those in vitro systems only contained the desired molecules. Recent studies have also shown that LCPs can be used as amyloid specific dyes for histological staining of tissue sections and that LCPs appear to provide additional information regarding the pathological events of proteinopathies, compared to conventional techniques [48], [49], [50], [51], [52].

Nilsson et al. exhibited the proof of concept by utilizing the LCPs as amyloid specific dyes in tissue samples [48]. Under distinct conditions PTAA, POMT and PONT (Fig. 2) were shown to selectively stain a plethora of amyloid deposits in formalin fixed tissue sections (Fig. 2). The negatively charged PTAA was amyloid specific under alkaline (pH 10) staining conditions, whereas the staining with the cationic and the zwitter-ionic LCPs were performed at acidic pH (pH 2). All the LCPs stained amyloid deposits associated with systemic amyloidoses and local amyloidoses, such as Type 2 diabetes and Alzheimer’s disease (AD). Similar to results obtained on in vitro formed amyloid fibrils [45], PTAA showed a red-shifted spectrum with a maximum around 580 nm upon binding to amyloid deposits, whereas deposits stained by PONT emitted light with a more green-yellowish hue [46], [48]. Hence, upon binding to amyloid deposits in tissue samples, the rotational freedom of the thiophene rings and the geometry of the thiophene backbone were restricted, leading to a specific emission profile from the LCP similar to what was observed on pure amyloid fibrils in solution [45], [46], [47], [48]. Furthermore, some results indicated that PTAA emits light of different colors upon binding to different amyloid subtypes and that smaller protein deposits that were undetected with conventional dyes was visualized by PTAA [48].

As mentioned, above, the conformational induced change of the fluorescence is a distinct intrinsic property seen for LCPs that cannot be achieved by sterically rigid conventional amyloid ligands based on the molecular scaffold of Congo red or Thioflavins. Thus, LCPs could offer the possibility to obtain a specific spectroscopic signature for individual protein aggregates. Instead of looking at the total amount of protein aggregates, heterogenic populations of specific protein aggregates could be observed. These assumptions have also been verified with experimental data using transgenic mouse model having AD pathology and transgenic mice infected with distinct prion strains [49], [50], [51], [52].

Upon application of several LCPs to transgenic mouse models having AD pathology, a striking heterogeneity in the characteristic plaques composed of the A-beta peptide was identified [49]. LCP staining of brain tissue sections revealed different sub-populations of A-beta deposits, observed as protein aggregates with different colors. The spectral features of LCPs enabled an indirect mapping of the architecture of individual amyloid deposits, as the different colors of the LCPs are associated with different conformations of the polymer backbone [43]. This notion allowed the authors to hypothesize how senile plaques are formed. Since the interior of large amyloid deposits seemed disordered compared to the exterior, it appeared plausible that diffuse aggregated A-beta matured into a rigid plaque exterior [49]. It should however be noted that very small amyloid deposits in young mice showed a LCP fluorescence, indicative of a highly compact fibrillar structure. Hence, it is equally plausible that A-beta amyloid deposits assemble from the outside in, i.e. from rigid fibril association forming a disordered plaque center [49]. Furthermore, PTAA was recently used for spectral separation of A-beta deposits associated with the Swedish/Arctic mutation or only Swedish mutation in the amyloid precursor protein (APP), indicating that different point mutations of the amyloidogenic protein can give rise to distinct morphologies of protein deposits [51]. Overall, these findings might lead to novel ways of diagnosing AD and the LCP technique could provide a new method for studying the pathology of the disease in a more refined manner. LCPs might be particularly useful for identifying distinct toxic entities giving rise to cell death and loss of neurons, or for establishing a correlation between the type of protein deposits and the severity of AD. However, further studies of the molecular interaction between well-characterized in vitro produced protein aggregates with distinct structures and LCPs will be necessary in order to correlate the spectroscopic read out from the LCP with a distinct molecular structure of the protein deposit. Although the achievement of obtaining certain spectroscopic LCP signatures from heterogenic populations of protein aggregates is beneficial compared to conventional amyloid imaging probes, correlation of the spectroscopic signature to a specific form of the aggregated protein will be necessary in order to gain novel insight into the pathological process of the disease. However, the LCPs could be useful for comparison of heterogenic protein aggregates in well-defined experimental systems, such as transgenic mouse models.

Heterogenic protein aggregates are also found in other proteinopathies, such as the infectious prion diseases. Prions can occur as different strains and the prion strain phenomenon is believed to be encoded in the tertiary or quaternary structure of the prion aggregates. This belief was also verified when protein aggregates in brain sections from mice infected with distinct prion strains were stained by LCPs [50]. The LCPs bound specifically to the prion deposits and different prion strains could be separated due to alternative staining patterns of LCPs with distinct ionic side chains. Furthermore, the anionic LCP, PTAA, emitted light of different wavelengths when bound to protein deposits associated with a distinct prion strain (Fig. 2). By calculating ratios of the intensity of the emitted light at certain wavelengths, prion aggregates associated with distinct prion strains were easily distinguished from each other, verifying the usefulness of spectral properties of LCPs for classification of protein deposits. The LCP technique was also recently used for specific spectral identification of protein deposits seen for a de novo generated prion strain [52]. These protein deposits were not stained by any of the conventional amyloid ligands, ThT and Congo red, indicating that LCPs could be used to identify a subset of protein deposits normally undetectable by conventional methods [52]. Whether this observation was due to an increased sensitivity of the LCPs compared to other amyloid ligands or if LCPs identify a broader range of structurally diverse protein deposits compared to conventional amyloid dyes has yet to be clarified.

In the first study of prion strains [50], it was also shown that the optical fingerprint observed for PTTA bound to protein deposits related to distinct prion strains most likely occurred due to a structural variance of the protein deposits. By taking recombinant mouse prion protein (mPrP) and converting it into two different types of amyloid fibrils by using varying conditions for fibrillation, Sigurdson et al. were able to show that the emission profile of PTAA could be used to distinguish these two fibril preparations [50]. As these two preparations of fibrils were chemically identical, having the same protein (mPrP) and being dialyzed against the same buffer, the spectral differences seen for PTAA were most likely due to structural differences between the fibrils. The same procedure has also been performed to distinguish A-beta 1–42 amyloid fibrils grown under different conditions in vitro [49] and amyloid fibrils built from reduced or intact bovine insulin [53]. In these studies, the zwitter-ionic LCP, tPTT [49], or the cationic LCP, POMT [53], were used, showing that distinct LCPs can be used to differentiate protein aggregates of diverse origin. Hence, LCPs have so far been shown to provide indirect structural insights regarding the morphology of individual protein deposits. Additionally, LCPs could be used as a complementary amyloid imaging agent for the characterization of protein deposits associated with individual prion strains [50], [52]. These findings may be of great value, as phenomena similar to those occurring in prion strains may be much more frequent than is now appreciated, and may extend to additional proteinopathies.

6. Future prospects of LCPs 

Unexpectedly, conjugated polymer materials that originate from electronics and solar cells have proven useful to gain novel insights into the molecular mechanism of protein aggregation and the pathology of proteinopathies [54]. Apparently, the unique optical properties of LCPs will be useful to gain more information regarding the molecular details of the protein aggregation process and the pathological events underlying proteinopathies. However, there is still a great extent of basic research that needs to be performed to take fully advantage of the technique and to develop LCP based molecular scaffolds that can be used for diagnostic purposes of these diseases.

As the LCPs have proven to stain a greater diversity of protein aggregates than Thioflavine T and Congo red in histological tissue samples [50], [52], the development of LCP based molecular scaffolds that can be used for in vivo imaging of protein aggregates would be of great interest. All the previously reported LCPs have a molecular weight between 1000 and 5000Da and will not likely cross the BBB, so chemical design of smaller well-defined oligomeric thiophene derivatives are necessary. These dyes could be utilized in multiphoton imaging applications for in vivo imaging of protein aggregates, as previously reported LCPs have been shown to work efficiently in this imaging setup [48]. Such molecules will be excellent for monitoring the molecular pathology of protein misfolding diseases and disease progress upon experimental treatment in transgenic mouse models. Nevertheless, multiphoton imaging is not suitable as a clinical diagnostic tool for proteinopathies, so radiolabeled versions of novel thiophene scaffolds useable for PET or SPECT will be advantageous. Hence, there are very clear hurdles, including radiolabeling strategies, basic toxicological and pharmacokinetical studies, that such novel thiophene based molecules will have to pass in order to be useful in the clinical setting.

Future chemical design of novel well-defined oligomeric thiophene scaffolds will certainly include combinatorial approaches for optimizing the thiophene core structure. Such chemical design may also provide more effective binders for distinct classes of protein aggregates and LCPs having different and well-defined chain lengths would also be of interest in order to understand the conformation induced optical properties of the LCPs. A focus must also be turned to the molecular details regarding the selective binding site of LCPs to specific protein aggregates. Therefore, a general understanding of photo-physical processes of LCPs and design rules in the synthesis of LCPs will be important for continued progress in the development of novel amyloid imaging based on a repetitive thiophene scaffold.

7. Conclusions 

As described in this review, a diversity of molecular scaffolds has been utilized for the development of smart amyloid imaging agents. However, there is still a need for novel molecular tools, as many basic scientific questions regarding the mysterious molecular and pathogenic events of protein misfolding diseases remain elusive. Secondly, imaging agents for accurate clinical diagnosis and for following the progression of these diseases are highly relevant, as successful therapeutic solutions are lacking today to a large extent due to difficulties in reliable and relevant quantitative monitoring methods. Hopefully, the present knowledge about amyloid imaging agents will lead to the development of novel smart imaging agents that will solve these issues.