|Home | About | Journals | Submit | Contact Us | Français|
A major challenge for diagnosing and monitoring the progression of amyloid-based diseases is the capability to distinguish between amyloid deposits that are associated with related, but distinctly different, diseases. Here, we demonstrate that Amino Naphthalenyl-2-Cyano-Acrylate (ANCA)-based probes can fluorescently discriminate between different types of amyloid deposits in brain. This discriminating behavior is due to the stabilization of the ground versus excited states of these probes as a function of the polarity of their microenvironment (i.e. within the binding pocket on the amyloid). This property makes it possible for the first time to estimate the inherent static relative permittivity of the binding pocket of each amyloid within tissue. The capability to selectively follow the deposition of specific amyloids in tissue may provide important information for therapeutic development that is not readily accessible from currently available technology.
Amyloid plaque accumulation in the brain is the hallmark of many neurodegenerative disorders, including Alzheimer’s (AD) and Creutzfeldt-Jakob (CJD) disease.1,2 Approaches to clinically diagnose and monitor the progression of these diseases include targeting of amyloid deposits with small molecule imaging agents.3-5 Along these lines, an amyloid-labeling probe for use in positron emission tomography (PET) has recently been shown to be a clinically useful diagnostic agent for AD.6,7 While PET imaging represents a great first step toward the diagnosis of neurodegenerative diseases, this technique is limited by the binary nature of the radioactive signal that does not allow for discrimination between amyloids associated with different, but closely related, diseases. Alternative techniques that can discern between different types of amyloids may offer important information necessary to develop effective treatment strategies that are tailored to specific diseases.
Fluorescence-based imaging of amyloids has emerged as a potentially lower cost, more accessible, and nonradioactive alternative method to PET.8-16 In principle, fluorescence-based-probes offer an advantage over their radio-active counterparts since their fluorescence profile could be used to distinguish different amyloids in tissue. Here, we report the ability of Amino Naphthalenyl-2-Cyano-Acrylate (ANCA)-based probes (for a generic structure see Table 1) to fluorescently discriminate between deposits in brain derived from amyloid-β (Aβ) peptides associated with AD or from prion (PrPSc) proteins associated with prion disease (Figure 1).
We previously reported that ANCA probes (1-3, Table 1) could label Aβ plaques in human brain tissue sections from AD patients.15 Brain sections stained with these probes showed good overlap of fluorescence with a mouse monoclonal anti-human Aβ IgG, demonstrating that these probes could mark the location of Aβ deposits with good specificity compared to surrounding background tissue. The ANCA probes fluoresce by a photon-induced dipole mechanism,14,17 where the electron rich amine functionality, attached to the naphthalene moiety, can donate electron density to the electron deficient cyano ester through the π-system of the ANCA scaffold (Table 1)-15,18 The emission of these compounds is enhanced upon restriction of the rotation around the single bonds between the donor (amine) and acceptor (cyano ester) groups. The ANCA probes, therefore, exhibit a large enhancement in fluorescence properties upon binding to amyloid substrates compared to the weaker fluorescence of the free compounds when in solution (see Figure S1 in the SI).
To explore the ability of the ANCA probes to fluorescently label amyloid deposits from proteins other than Aβ in tissue, we compared fluorescence micrograph images of plaques derived from Aβ (Figure 1A) and PrPSc (Figure 1B) stained with 1 in frozen brain sections (i.e., unfixed) from mouse models for AD and prion disease.19-21 Importantly, 1 exhibited excellent capability to label amyloid plaques in both tissues and displayed a visually observable difference in the color emitted upon staining of the plaques. Probes 2 and 3 exhibit similar properties (see Figure S3 in SI).
Table 1 displays that the ANCA probes bound to amyloid deposits exhibit a difference in maximal emission wavelength (λmax) of ~20 nm, depending on the type of amyloid protein present in the plaque. Remarkably, the variability the observed λmax between plaques was quite narrow (± 2-3 nm from inspection of ≥ 25 plaques in each sample). Moreover, Aβ plaques stained with 1 from different parts of the brain (e.g., hippocampus, hypothalamus, cortex) from the same mouse all exhibited essentially the same emission λmax (± 3 nm). Additionally, prion plaques stained with 1 from the corpus callosum of different mice also exhibited essentially the same emission λmax (± 2 nm). In contrast, we did not observe such fluorescence discrimination upon staining of the plaques with Congo red (see Figure S4 in the SI). These surprising results prompted us to investigate the origin of fluorescence discrimination by 1-3.
Previous studies in solution have shown that the small differences in the environment of proteins22-26 and lipids27 can affect the Stokes shift of bound fluorophores.28 If the time dependence of fluorescence polarization of probes 1-3 can be neglected—that is, if we assume that any reorientation molecules within the binding pocket of the amyloid proteins is fast compared to the lifetime of the excited state of the probes—the dependence of the polar environment on the frequency of the absorbance and fluorescence emission spectra of 1-3 can be approximated by the Ooshika-Lippert. Mataga (OLM) equation:27,29,30
where Δv is the average frequency of the Stokes shift,31 μe and μg are the dipole moments of the molecules in the excited state and ground state respectively, h is Plank’s constant, a is the approximate radius of the molecules assuming a spherical cavity with a rigid dipole at the center, ε0 is the static relative permittivity (i.e., dielectric constant) of the dielectric continuum surrounding the molecules, and n is the refractive index of the medium.
Since the frequency of absorption for molecules 1-3 remains relatively constant when measured in different solvents of various polarities, we found an approximately linear relationship between the Stokes shifts of 1-3 and their observed λmax for fluorescence emission (see Figure S5 in the SI). This linear correlation allows us to use only the λmax of emission to analyze the effect of the polar environment within the binding pockets of the amyloids on the fluorescence properties of 1-3.32 Equation 1 could, therefore, be simplified to:27
where C1 and C2 are constants that reflects several inherent properties of the ANCA probes.
Figure 2A-C (open circles, ○) plots the dependence of 1/λmax for fluorescence emission of 1-3 in solvents of various polarities (see Figure S6 in the SI) as a function of the static relative permittivity (ε0) and refractive index (n) of each sol-vent according to eq. 2. The observed linear behavior shown in Figures 2A-C suggests that the OLM equation (eq. 1) describes fairly well the dependence of the observed fluorescence of compounds 1-3 on the polarity of their surrounding environment. These results suggest that the origin of the observed fluorescence discrimination between the two different types of amyloid deposits using probes 1-3 could indeed be simply explained by small differences in the polar environment within the binding pockets of the amyloid proteins.
In order to gain some additional insight into the nature of the polar environment within the binding pockets of the amyloid tissues, we replotted the dependence of 1/λmax for fluorescence emission of 1-3 in different solvents as a function of ε0 only, while ignoring the refractive index of the solvent (Figure 2A-C, open squares, ).27 We found that the linear fit of 1/λmax versus (ε0-1)/(2ε0+1) was consistently better compared to when the refractive index was included as in eq. 2 (see Figure S7 in the SI). This suggests that the static relative permittivity (ε0) dominated the dependence of the polar environment on the λmax for fluorescence emission of 1-3.33 This simple relationship between λmax and ε0, hence, made it possible to estimate the static relative permittivity of the binding pockets of the two different amyloids in tissue by the observed emission maxima of plaques stained with 1-3 (Figure 2D). Satisfyingly, the estimated ε0 values for the two amyloids were in fairly good agreement (given the assumptions and approximations made in equations 1 and 2), regardless of which ANCA probe was used for measurement. The data reveals that both amyloid binding pockets are relatively hydrophobic as expected, but that the permittivity within the binding pocket of prion plaques is roughly two-fold greater than the permittivity within the binding pocket of Aβ plaques. To provide a reference point for calibration, we estimate the dielectric constants of the binding pocket of Aβ and prion plaques to be roughly similar to diethyl ether (ε0 =4.27) and tetrahydrofuran (ε0 = 7.52), respectively.34
Although small differences in the polar environment within the binding pockets of the amyloids can reasonably account for the observed fluorescence discrimination of the amyloids with probes 1-3, we tested another plausible explanation for this observation. We considered that acid-base interactions between the amyloid and the probes could affect the electron donor capability of the amine group of these probes, thus affecting their observed emission profiles. To investigate this possibility, we examined the excitation and emission profiles of 1-3 as a function of pH in free aqueous solution (Figure 3).
Figure 3A,C shows that probes 1 and 2 exhibit a sharp 100-125 nm change in excitation maxima at a pH of ~4.2 and ~1.7, respectively. These changes presumably reflect the protonation of the nitrogen donor of the piperidine in 1 or morpholine in 2, and provide an estimate of the pKa’s of the protonated forms of these groups. A much smaller change in excitation λmax (~ 10 nm) was observed for 3 at a pH of ~6.9, which presumably reflects the pKa of the protonated N-methyl tertiary amine. More importantly, the emission profiles of 1-3 (Figures 3B, D, and F) in acidic (pH 3.8) versus neutral (pH 7.8) solutions indicate that only probe 1 should exhibit a strong (~125 nm) difference in emission λmax when bound to amyloid plaques, if the origin of fluorescence discrimination was due to an acid-base interaction.35 Since we observe a difference (~20 nm) in emission λmax between Aβ versus prion plaques upon staining with all three probes 1-3, it is unlikely that the origin of fluorescence discrimination is due to acid-base interactions between the probes and the proteins within the binding pocket of the amyloid deposits.
In conclusion, we present a set of fluorescent amyloid-binding probes that can report a different color of fluorescence emission when bound to different types of amyloid deposits in tissue samples. The origin of this fluorescence discrimination most likely arises from the sensitivity of these probes to the polar environment within the binding pocket of an amyloid plaque. Subtle changes to the structure (i.e., electron donor moiety) in these ANCA probes make it possible to tune the spectral window of fluorescence discrimination. The strong correlation between fluorescence emission of the probes and polarity of the environment afforded estimates of the static relative permittivity of the binding pocket in the amyloid plaques. Analysis of amyloid deposits in tissue revealed that the prion deposits exposed a significantly more polar environment to the probes than the Aβ deposits. This previously unreported, fundamental difference between these two types of amyloids make it possible to distinguish them in tissue by simple inspection of the fluorescence emission of molecules that target them. Given the recent findings that amyloiddeposition in the brain of AD patients may be paralleled by deposition of amyloids in the retinal tissue36-41 or lens42,43 of the eye, optical inspection of amyloid deposits in the eye may represent an exciting opportunity for diagnosing and monitoring the progression of amyloid-associated diseases. Fluorescent amyloid imaging agents such as the ANCA compounds may, therefore, have practical ophthalmic applications in living patients to help distinguish between closely-related diseases where the symptoms and pathological characteristics show many similarities.44,45 Efforts to examine the scope of these probes for discriminating amyloid deposits associated with diseases other than Alzheimer’s and prion diseases are currently underway.
Financial support from the NIH [CA 133002 (EAT) and R01NS069566 (CS)] is gratefully acknowledged. We also thank the NSF for financial support [CHE-9709183, CHE-0741968, and CHE-0847530 (JY)]. We thank Dr. Christina C. Capule for her help with the binding studies. We also thank Dr. Edward Koo for the brain samples from AD transgenic mice.
ASSOCIATED CONTENT Details for the synthesis of compounds 1-3, experimental protocols for the staining of plaques in neuronal tissue, and for binding and fluorescence studies. This information is available free of charge via the Internet at http://pubs.acs.org.