Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by extracellular amyloid plaques, intracellular tau tangles, and neuronal loss. Developing imaging tools for these hallmarks is important for the diagnosis and treatment of AD. Some imaging markers are currently used in clinics, such as amyloid plaque positron emission tomography (PET) radiotracers, tau protein PET radiotracers, and magnetic resonance imaging (MRI) of hippocampal atrophy, especially in the late stage of the disease. However, amyloid-beta (Aβ) has several forms, such as non-toxic monomeric Aβ (mAβ) and toxic oligomeric Aβ (oAβ), and the misfolding of mAβ precedes the formation of oligomers (Huang and Liu, 2020).

Oligomeric forms of amyloid precursor protein (APP) cleavage products play a pivotal role in the early stages of AD (Thomas et al., 2016; Chae et al., 2020). To facilitate the early diagnosis of AD patients, it is essential to detect oAβ rather than amyloid plaques. oAβ imaging tool because oAβ is considered to be the most toxic and pathogenic form of Aβ in AD, as it can impair synaptic function, induce neuroinflammation, and trigger neuronal death, leading to cognitive decline and dementia (Viola et al., 2022; Pyun et al., 2023). However, current imaging tools for AD diagnosis mainly target amyloid plaques, which are aggregates of fibrillar Aβ that accumulate in the brain. Amyloid plaques do not correlate well with disease progression and are not present at the earliest stages of the disease. Therefore, imaging tools that can detect oAβ in vivo would be more useful for early diagnosis, monitoring, and treatment of AD (Viola et al., 2022; Pyun et al., 2023). Recently, we developed an MRI contrast agent to detect oAβ based on the conjugation of gadolinium (Gd)-dodecane tetraacetic acid (DOTA) to a DNA aptamer called ob5 which can bind to a given target with high affinities (Kim et al., 2023). We demonstrated the ability of a Gd-DOTA-ob5- cyanine5.5 (cy5.5) contrast agent to detect oAβ using fluorescence imaging in cell experiments, and ex-vivo and in-vivo experiments using APP/PS/ApoE knockdown (KD) AD mouse model, also known as 3xTg AD mice (Kim et al., 2023).

To study the effects of oAβ on the brain, some researchers have developed mouse models that represent oAβ deposits and toxicity. Different types of mouse models can be used to mimic oAβ pathology, depending on the methods and targets of genetic engineering or treatment. The APP/PS AD mouse is a well-known model that has oAβ from more than 10 weeks of age, as well as amyloid plaques in the brain at later stages. The 3xTg AD mouse, which is APP/PS/ApoE knockout (KO), is the model that is often used. Another mouse model is APP/PS/Tau AD. Both these mouse models have been used to evaluate oAβ deposits in the brain.

Whenever a novel MRI contrast agent is developed, it is important to evaluate the MRI signal changes in different mouse models and the time-dependent signal changes after injection of the contrast agent. Therefore, the objectives of this study were 1) to evaluate the signal enhancement using our developed Gd-DOTA-ob5 aptamer MRI contrast agent in 3xTg, APP/PS/Tau, and control C57BL/6 mouse models for early diagnosis of AD using MRI and 2) to evaluate the time-dependent signal changes after injection of the contrast agent.

We developed a novel MRI contrast agent by conjugating a Gd-DOTA-DNA aptamer, named ob5, which was constructed by amide bond formation, for the detection of oAβ deposits in the brain (Jahng and Kim, 2017; Kim et al., 2023). We used the Selective Evolution of Ligands by Exponential enrichment (SELEX) method (Burke and Gold, 1997; Horii et al., 2010) to prepare selective targeting nucleotides that showed high binding affinity to the oAβ conformation. After preparing Gd-DOTA as previously described (Kielar et al., 2010; Huynh et al., 2017), we synthesized Gd-DOTA-ob5 as described previously (So et al., 2006; Dudeffant et al., 2017; Aminololama-Shakeri et al., 2019; Li and Meade, 2019; Kim et al., 2023) and evaluated its ability to detect oAβ deposits in the brain using MRI.

All MRI experiments were performed on a 9.4 T animal MRI scanner (94/20 USR, Bruker, Ettlingen, Germany). A quadrature birdcage RF resonator with an inner diameter of 40 mm was used for signal transmission and reception. To acquire in vivo MR images using our Gd-DOTA-ob5 contrast agent, animals were anesthetized with 1.5%–2.0% inhalational isoflurane in oxygen. Animals were subjected to respiratory gating at a rate of 30–50 breaths/min (SA Instruments, Stony Brook, NY, United States) without monitoring cardiovascular gating.

To evaluate MRI signal changes in the whole brain and specific brain areas before and after intravenous injection in the tail vein of the Gd-DOTA-ob5 contrast agent, T1-weighted images were acquired before and after injecting our contrast agent. After injection of the contrast agent, T1-weighted images were repeatedly scanned at approximately 5-minute intervals until 25 min after injection. T1-weighted images were obtained by a rapid acquisition with relaxation enhancement (RARE) spin-echo pulse sequence with the following parameters: TR = 467.7 ms, TE = 6.7 ms, FOV = 20 × 20 mm², NSA = 3, RARE factor = 1, frequency direction = left-right, pixel BW = 0.4807 kHz, matrix size = 200 × 200, number of slices = 30 with slice thickness = 0.7 mm, and voxel size = 0.10 × 0.10 × 0.7 mm³. The scan time for each measurement was 4 min 40 s.

Furthermore, T2-weighted images were also acquired with a TurboRARE pulse sequence before and after injecting the Gd-DOTA-ob5 contrast agent with the same imaging parameters as T1-weighted images except the following parameters: TR = 2,500 ms, TE = 25 ms, NSA = 4, RARE factor = 8, pixel BW = 0.3367 kHz, matrix size = 128 × 128, and voxel size = 0.156 × 0.156 × 0.7 mm³. The scan time for each measurement was 2 min 40 s. The T2-weighted images after the contrast injection were acquired at the end of each experiment for each animal.

DICOM images were converted into the NIFTI image format using PMOD software (version 3.8, PMOD Group, Graubünden, Switzerland). We preprocessed NIFITI images using the Statistical Parametric Mapping (SPM12) software (Wellcome Trust Centre for Neuroimaging, UCL Institute of Neurology, London, United Kingdom; www.fil.ion.ucl.ac.uk/spm) according to the following steps. First, all the images were magnified by 10 for the SPM Display function. Second, T1-weighted images after contrast injection were co-registered with those before the injection of the contrast agent (CA). Third, the T2-weighted image after contrast injection was co-registered with those before the injection of the CA. Fourth, T2-weighted images were co-registered with T1-weighted images before the injection of the CA. Fifth, the T2-weighted image and other images were spatially normalized into the C57BL/6 mouse brain template (Hikishima et al., 2017) Finally, the signal intensity of the post-contrast T1-weighted images was adjusted in each voxel by pre-contrast enhanced images T1-weighted image the following equation as S(t)=(Spost-Spre)*100/Spre, where Spost is the signal intensity at post = 5, 10, 15, 20, and 25 min after injection of contrast agent and Spre is the signal intensity at before contrast injection. We drew noise areas in the image before the injection of the CA for each mouse by MRIcro software. Then, the signal-noise ratio (SNR) of all T1-weighted images was calculated by the following equation: SNR = SI*100/Smean noise, where SI is the signal intensity at each scan-time point and Smean noise is the mean value of its noise.

The details of the immunofluorescence imaging to image oAβ deposits were described in our previous paper (Kim et al., 2023). Briefly, mice were sacrificed and perfused with 0.9% saline. The brains were removed and post-fixed with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer saline (PBS: pH 7.4) at 4°C overnight. The brains were cryo-sectioned into 25 µm-thick sections with a cryostat microtome (Leica CM 1950; Leica Instruments, Nussloch, Germany). Cryosections (25 µm) were mounted on Superfrost Plus slides (Thermo Scientific), fixed for 10 min with 4% PFA in 30% sucrose in PBS, washed, and blocked with 10% BSA (bovine serum albumin) + 0.04% Triton X-100 in 0.05 M PBS for 1 h. Free-floating sections were incubated with 4% bovine serum albumin in PBS for 1 h, then reacted with monoclonal anti-β-amyloid antibody (Bam10)-QD525 and ob5 aptamer-conjugated QD565 at 4°C overnight. The monoclonal antibody BAM-10 (MA1-91209) (ThermoFisher, Waltham, United States) recognizes the epitope for the N-terminus (1–12 amino acid residues) of Aβ1-42 for amyloid plaques in the brains of AD mouse models (Lee et al., 2009). Confocal images were acquired on a spinning disk confocal imaging system (LSM710; Carl Zeiss AG, Oberkochen, Germany).

The IF analysis revealed the deposition pattern of mAβ and oAβ in the brain tissue of AD mouse models (Figure 7). The IF images of 16-week-old and 36-week-old 3xTg mice differed remarkably from those of non-Tg mice in each age group. Moreover, the amyloid deposition patterns in the brain of 16-week-old 3xTg AD or APP/PS/Tau AD mice were considerably different from those of 36-week-old 3xTg AD or APP/PS/Tau AD mice. In the non-Tg control group, there was almost no expression of mAβ and oAβ. However, in the 3xTg mice, red fluorescence due to oAβ was very strong, and green fluorescence due to mAβ was slightly strong, in the 16-week-old and 36-week-old mice, (Figure 7A).

On the other hand, to determine whether the signals were due to mAβ and oAβ and whether they were colocalized in the 36-week-old 3xTg mouse, we detected fluorescent anti-mAβAb-QD525 that showed strongly blue or green fluorescence under a confocal microscope and the fluorescent QD565-conjugated ob5 aptamer that showed slightly red fluorescence (Figure 7B). The results confirmed that mAβ and oAβ, which are a paralytic protein, were deposited at the same location in the 36-week-old 3xTg mouse, but oAβ deposits were more prominent using QD565-conjugated ob5 aptamer in the 16-week-old 3xTg mice. This finding is consistent with our previous observation that, in 3xTg mice, ApoE KO mutant accumulated more abundantly within brain tissues than control mice. These features indicate that the 3xTg AD mice are a suitable model for investigating oAβ deposits even before the accumulation of amyloid plaques at young ages.

We performed in vitro neuropathological fluorescent staining of mAβ and oAβ deposits in slices of brain tissue from 3xTg and APP/PS/Tau AD mice and control mice to evaluate the affinity of these monoclonal Aβ antibody (green) and ob5-cy5.5 (red) probes. As shown in Figure 7, the presence and distribution of mAβ were consistent with the results of staining adjacent slices using anti-mAβ (6E10) (Figures 7A,B), while no oAβ staining was found when using ob5 (Figure 7A) in C57B6BL mice, which further confirmed that ob5 had no affinity to mAβ. Furthermore, intense labeling of oAβ by ob5 was strongly observed in the brain slices of both AD mice (Figure 7A). More interestingly, Figure 7B showed slight staining of mAβ and oAβ in the cortex region of the normal and AD Tg mice (Figure 7B). Ex vivo intensity studies indicated that Figure 7A has high ex vivo oAβ deposits in mouse brain tissues of 16-week-old mice, and the fluorescence intensity decreased by an average of 55% in 36-week-old mice with two AD mice. There was a significant difference in the clearance profile after the intensity of 16-week-old mice between AD Tg and wild-type control mice, and the brain of AD mice exhibited synaptic alteration in parallel with intraneuronal deposits of Aβ oligomers without formation of amyloid plaques during early aged stage, which might be strongly attributed to the high affinity of ob5 to oAβ deposits in both AD mice.

We demonstrated that the T1-enhanced signal by the oligomeric amyloid-β targeted contrast agent was higher in 16-week-old AD models compared to the 36-week-old models. Our previous research proved that the oligomeric amyloid-β targeted contrast agent could visualize oligomeric amyloid-β in the 3xTG AD model (Kim et al., 2023). This suggests that the higher T1-enhanced signal in 16-week-old AD models indicates a greater presence of oligomeric amyloid-β in brain regions compared to the 36-week-old AD models. Recent studies have reported that the formation of oligomeric amyloid-β and Aβ plaques are pathologically interconnected in AD (Huang and Liu, 2020). Oligomeric amyloid-β is formed by the binding of a few amyloid-β proteins, which can later develop into amyloid plaques commonly found in the brains of AD patients (Lesne and Kotilinek, 2005; Sun et al., 2015; Mroczko et al., 2018). Numerous studies have developed MRI contrast agents and diagnostic radiopharmaceuticals for early AD diagnosis in animal brains (Bort et al., 2014; Patil et al., 2015; Zhang et al., 2015; Dudeffant et al., 2017; Dong et al., 2020; Luo et al., 2020), but to our knowledge, there were no MRI contrast agents or diagnostic radiopharmaceuticals available for detecting early oligomeric amyloid-β. Therefore, based on our findings, we suggest that the oligomeric amyloid-β targeted contrast agent can provide important information for the development of early diagnosis and treatment strategies for AD.

We confirmed a higher contrast effect in the hippocampus of the AD model compared to the control group through ROI-based analysis. Also, after the administration of the contrast agent, a pronounced statistical difference was observed only in the hippocampus of the AD model compared to the normal group, with no contrast effect differences in the cerebral cortex. This result differs from the typical amyloidosis process. Generally, in the early stages of AD, the accumulation of amyloid oligomers is observed mainly in specific areas of the cerebral cortex, such as the frontal and temporal lobes, and then extends to the limbic system like the hippocampus (Reilly et al., 2003; McGowan et al., 2006; Kimura and Ohno, 2009; Willuweit et al., 2009). Our findings, showing differences only in the hippocampus and not in typical amyloidosis, suggest a correlation with previous research indicating that cognitive impairment in long-term memory retention, which occurs in 16-week-old 3xTG AD, is related to amyloid-β accumulation, even though amyloid-β plaques are not present in the hippocampus and amygdala of 16-week-old 3xTG AD models (Billings et al., 2005). Furthermore, in 6-month-old 3xTG AD models, Aβ plaques first form in the cerebral cortex (Oddo et al., 2003). Therefore, in 36-week-old 3xTG AD models, a reduced amount of oligomeric amyloid-β may than amyloid plaques result in a lower contrast effect.

Although a high contrast effect was not observed in the cerebral cortex, the pronounced difference shown by our contrast agent in the hippocampus is important for evaluating AD models. The hippocampus plays a crucial role in memory formation and storage and shows high neural and metabolic activity, making it more vulnerable to oxidative stress and microenvironment changes (Venkateshappa et al., 2012; Watts et al., 2018; Rao et al., 2022). Therefore, the generation and accumulation of Aβ, especially in its oligomeric form, which is highly neurotoxic and can induce neuronal cell death and inflammatory responses, may be accelerated with the progression of AD (Sun et al., 2015). Consequently, observing AD pathology in the hippocampus at an early stage is extremely important, and among these, the evaluation of oligomers using contrast agents can be a useful indicator.

After the administration of the contrast agent, time-dependent variations in the contrast effect in the hippocampus differed between the 3xTG AD and APP/PS/Tau AD animal models. In the 3xTG AD model, the signal increase after contrast agent administration was sustained up to 25 min. However, in the APP/PS/Tau AD model, the peak contrast effect was observed at around 10 min, followed by a continuous decrease. Although no statistical difference was found in ROI-based analysis between 3xTG AD and APP/PS/Tau AD, the variation in the contrast effect suggests it may reflect the characteristics of the animal models. In the early pathological stages of AD, oligomeric amyloid-β tends to accumulate more around cells rather than inside them, showing the formation of aggregates in the extracellular space (Mroczko et al., 2018; Huang and Liu, 2020). Our previous research indicated that the late enhancement of the contrast effect was due to oligomeric amyloid-β circulating around amyloid plaques, and the rapid increase and decrease in the contrast effect were attributed to oligomeric amyloid-β in the brain’s extracellular space (Kim et al., 2023). Therefore, the initial rapid increase in contrast effect in both 3xTG AD and APP/PS/Tau AD could be due to an increase in oligomeric amyloid-β in the extracellular space in the early stages of AD. Additionally, the ongoing increase in the contrast effect in 3xTG AD over time and the decrease in APP/PS/Tau AD suggest a difference in the amount of oligomeric amyloid-β circulating around the amyloid plaques, indicating a more dominant formation of amyloid plaques in 3xTG AD compared to APP/PS/Tau AD Hence, our contrast agent proposes the flexibility in evaluating differences due to amyloid oligomers, suggesting its translational potential across model species.

The interaction between oligomeric amyloid-beta (Aβ) and a DNA aptamer involves a specific binding mechanism where the aptamer selectively recognizes and binds to the Aβ oligomer. The interaction of the process between them may be explained as the following: First, DNA aptamers are single-stranded DNA molecules that can fold into unique three-dimensional structures (Jeddi and Saiz, 2017). These structures are designed to have high affinity and specificity for their target molecules, in this case, the Aβ oligomer (Zhou and Rossi, 2017; Sharma et al., 2022). Second, when the DNA aptamer encounters an Aβ oligomer, it undergoes a conformational change that allows it to fit snugly around the oligomer. This is often due to the formation of complementary shapes and charge interactions between the aptamer and the specific epitopes on the Aβ oligomer (Guo et al., 2023). Upon binding, the aptamer may wrap around the Aβ oligomer or form a pocket that encloses it (Deng et al., 2020). This conformational change is critical as it brings the aptamer’s nucleotides into proximity with the Aβ oligomer, allowing for hydrogen bonding and electrostatic interactions. Third, the binding is stabilized by various non-covalent interactions, such as hydrogen bonds, van der Waals forces, and electrostatic attractions (Yu et al., 2023). These interactions ensure that the aptamer remains tightly bound to the Aβ oligomer. The specificity of the aptamer for the Aβ oligomer is due to the precise arrangement of its nucleotides, which are selected through an iterative process known as SELEX (Systematic Evolution of Ligands by EXponential enrichment) (Li and Lee, 2019). This process yields aptamers that are highly specific to their target molecules. Finally, the binding of the aptamer to the Aβ oligomer can inhibit the oligomer’s pathological interactions with cellular components, potentially preventing the toxic effects associated with Alzheimer’s disease (Zheng et al., 2022). For instance, an electrochemical aptasensor developed for Aβ oligomers utilizes a double-stranded DNA as a “conductive spring” (Deng et al., 2020; Ren et al., 2020). Upon the binding of the aptamer to the Aβ oligomer, the conformation of the aptamer changes, leading to a measurable change in the electrical signal, which can be correlated to the concentration of Aβ oligomers. The specificity and sensitivity of DNA aptamers make them suitable for detecting Aβ oligomers, which are considered key biomarkers and therapeutic targets in the pathology of Alzheimer’s disease (Ren et al., 2020; Zheng et al., 2022).

This study had several limitations. First, we did not compare the developed contrast agent with a commercially available MRI contrast agent, which is currently not used for AD diagnosis. This limits the generalizability of our findings and the potential clinical applications of our contrast agent. Second, we only used two types of AD mouse models, 3xTG and APP/PS/Tau, while there are several other types of animal models for AD such as 5xFAD. Therefore, the new developed contrast agent should be evaluated with more AD model mice to assess its validity and reliability across different models. Third, we did not include amyloid-beta PET, which is usually targeted to amyloid plaques rather than oligomeric amyloid-beta. Therefore, a direct comparison between the two imaging modalities might not be adequate to show the distribution of oligomeric amyloid-beta in the AD model mice brain. A multimodal imaging approach might be more informative and comprehensive for detecting AD pathology. Finally, the number of mice in each model was relatively small, which might reduce the statistical power and the sensitivity of our analyses. Therefore, we recommend a more extensive study with more mice in each model and more AD models to confirm and extend our results.