Sage Journals: Discover world-class research

Abstract

[¹⁸F]FDG–PET and β-amyloid-PET are established tools for assessing biomarker status in Alzheimer’s disease. In this study, we evaluated the potential of early-phase [¹⁸F]florbetaben (FBB) PET as a functional proxy for [¹⁸F]FDG–PET in preclinical Alzheimer’s disease models by examining regional perfusion and glucose metabolism in two transgenic mouse lines. Ninety-two APPPS1 (n = 17), APP^SAA (n = 56), and age- and sex-matched wild-type mice (n = 19; 3–12 months, 40% female) underwent static [¹⁸F]FDG–PET (30–60 min p.i.) and dynamic [¹⁸F]FBB–PET (0–60 min p.i.). Standardized uptake values were derived for both [¹⁸F]FDG–PET and [¹⁸F]FBB–PET for the whole brain and 14 Ma–Benveniste–Mirrione atlas regions. We identified the 1–3 min p.i. time window as optimal, yielding the highest concordance with [¹⁸F]FDG (R = 0.53, p < 0.0001) across all regions. Both APPPS1 and APP^SAA mice exhibited significant increases in perfusion (both p < 0.0001) and glucose metabolism (APPPS1: p = 0.0028; APP^SAA: p < 0.0001) compared to wild-type controls. These findings demonstrate that early-phase [¹⁸F]FBB–PET not only mirrors [¹⁸F]FDG–PET-derived metabolic changes but also enables a single-scan assessment of β-amyloid pathology and brain function, thereby reducing the number of required scans and potentially the number of animals per study, and strengthening the translational value of preclinical PET research.

Graphical Abstract

This is a visual representation of the abstract.

Keywords

Alzheimer’s disease early-phase β-amyloid PET FDG PET mouse perfusion

Introduction

Alzheimer’s disease (AD) remains a major medical and societal burden in the 21st century. Recent advances in biomarker research have transformed the diagnosis and monitoring of AD by introducing biologically grounded classification systems. The A/T/N framework categorizes AD pathology based on β-amyloid accumulation (A), tau pathology (T), and neurodegeneration (N).¹ Molecular imaging, particularly PET, plays a central role in detecting these biomarkers in vivo, even before clinical symptoms emerge. β-amyloid-PET and tau-PET identify pathological protein aggregates, while [¹⁸F]fluordeoxyglucose (FDG) PET provides a functional measure of glucose metabolism and neuronal integrity.^2,3 Although these modalities offer complementary insights, they typically require separate imaging sessions, increasing patient burden and resource use. Recent human studies have demonstrated that early dynamic frames of β-amyloid-PET contain perfusion-related information that correlates well with metabolic patterns derived from [¹⁸F]FDG–PET. These early-phase images, acquired within minutes post-injection (p.i.), may thus serve as a proxy for cerebral metabolism and reduce the need for additional scans.^4–6

These previous findings from human studies are promising, but they highlight a key back-translational challenge: determining whether early-phase β-amyloid-PET can reliably reflect cerebral metabolism in preclinical models of AD. Given the widespread use and established role of [¹⁸F]FDG and [¹⁸F]florbetaben (FBB) in preclinical PET imaging, assessing the potential of early-phase [¹⁸F]FBB to serve as a substitute for [¹⁸F]FDG is becoming increasingly relevant for diagnostic and therapeutic research. Moreover, there is an unmet need to enhance the efficiency of preclinical imaging protocols by reducing scan duration, minimizing animal stress, and adhering to the ethical principles of the 3Rs, replacement, reduction, and refinement.⁷ Furthermore, while hypometabolism is a hallmark of neurodegeneration in patients, several AD mouse models exhibit cerebral hypermetabolism, likely driven by plaque-associated and activated microglia rather than neurons.⁸

While early-phase β-amyloid-PET has shown promise as a proxy for glucose metabolism in humans,^6,9,10 its validity in preclinical models remains uncertain, particularly in light of the divergent metabolic signatures observed in mice. Clarifying whether early-phase β-amyloid-PET signals can serve as a proxy for glucose metabolism is therefore essential for interpreting readouts in preclinical imaging of transgenic models of AD. Importantly, while perfusion and glucose metabolism represent distinct physiological processes, blood flow primarily reflecting substrate delivery and metabolism reflecting intracellular energy turnover, they are closely coupled through neurovascular and neurometabolic mechanisms.^11–13 Accordingly, early-phase β-amyloid-PET does not measure metabolism per se but rather provides a perfusion-weighted signal that spatially and functionally correlates with metabolic activity. This study therefore investigates whether such coupling can be leveraged in preclinical models to approximate FDG-derived patterns.

To address this, we conducted a systematic comparison of early-phase [¹⁸F]FBB–PET and [¹⁸F]FDG–PET in two widely used AD mouse models. We aimed to evaluate the extent to which early-phase β-amyloid-PET captures functional brain changes associated with glucose metabolism. By doing so, we sought to assess its potential as a proxy imaging marker in preclinical research, with implications for improving efficiency, adhering to 3R ethical standards, and narrowing the translational gap between small-animal and clinical neuroimaging.

Materials and methods

Animals and study design

All animal experiments were approved by the local animal care committee of the Government of Upper Bavaria (Regierung Oberbayern; ROB-55.2-2532.Vet_02-22-125). They were overseen by a veterinarian and in compliance with the ARRIVE guidelines 2.0 and were carried out in accordance with the U.K. Animals (Scientific Procedures) Act, 1986 and associated guidelines, EU Directive 2010/63/EU for animal experiments. Animals were housed in a temperature- and humidity-controlled environment with a 12 h light-dark cycle, with free access to food (Sniff Spezialdiäten GmbH) and water. Anesthesia prior to tracer application was induced with 3.0% isoflurane and maintained during PET scanning with isoflurane 1.5% delivered via a mask at 3.5 L/min. During all PET acquisitions, animals were positioned on a heated imaging bed to maintain body temperature, and individual physiological parameters (body temperature, heart rate, and respiration rate) were continuously monitored to ensure physiological comparability between groups. Following the experiment, the mice were transferred to fresh temporary cages equipped with food, water, and heating mats for warming. The animals were returned to their home cages only after they had fully regained consciousness. [¹⁸F]FBB (Neuraceq) precursor was provided by LIFE Molecular Imaging GmbH and the synthesis was performed as reported previously.¹⁴ [¹⁸F]FDG was purchased commercially from Advanced Accelerator Applications (AAA).

PET/MRI imaging

All rodent PET procedures followed an established standardized protocol for radiochemistry, acquisition times, and post-processing which was transferred to a novel PET/MRI system.^15,16 All mice were imaged with a 3 T Mediso nanoScan PET/MR scanner (Mediso Ltd., Hungary) using a triple-mouse imaging chamber.¹⁷ Prior to [¹⁸F]FDG imaging, animals were fasted for at least 4 h to minimize baseline glucose levels. Two 2-min anatomical T1 MR scans (sagittal and axial) were performed before tracer injection and PET scan (head receive coil, matrix size 96 × 96 × 22 mm³, voxel size 0.24 × 0.24 × 0.80 mm³, repetition time 677 ms, echo time 28.56 ms, flip angle 90°).

Injected dose was 12.9 ± 2.2 MBq for [¹⁸F]FBB and 16.8 ± 3.1 MBq for [¹⁸F]FDG delivered in 200 μL saline via intravenous injection. [¹⁸F]FDG–PET data were acquired in a static 30–60 min p.i. window, whereas [¹⁸F]FBB–PET was recorded dynamically over a 0–60 min period. List-mode data within a 400–600 keV energy window were reconstructed using a 3D iterative algorithm (Tera-Tomo 3D, Mediso Ltd., Hungary) with the following parameters: matrix size 55 × 62 × 187 mm³, voxel size 0.3 × 0.3 × 0.3 mm³, eight iterations, six subsets. Decay, random, and attenuation corrections were applied. The T1 image was used to create a body-air material map for the attenuation correction.

Image preprocessing and analysis

All images were performed using PMOD (v3.5; PMOD Technologies, Basel, Switzerland). Unified spatial normalization (nonlinear warping, transient 0.6 mm³ Gaussian smoothing of the input image, 16 iterations, frequency cutoff 3, no thresholding) of all original Aβ- and [¹⁸F]FDG–PET images to a previously established template was performed.^17,18 Reconstructed PET data were aligned to individual MR images and analyzed using the whole-brain and predefined regions of interest based on the Mouse Brain VOI Atlas (Ma–Benveniste–Mirrione).¹⁹ For preliminary analysis of [¹⁸F]FBB images, data from the first 1–3 min p.i. were aggregated to generate early-phase images. These were then used for correlation with [¹⁸F]FDG-derived metabolic maps obtained from the static acquisitions. Tracer uptake was quantified using standardized uptake values (SUV) for both [¹⁸F]FDG and [¹⁸F]FBB for each region.

Statistical procedures

In order to evaluate the relationship between early phase [¹⁸F]FBB and [¹⁸F]FDG readouts, Pearson correlation coefficients were computed for regional SUV across 14 brain regions. To assess whether the relationship between early-phase [¹⁸F]FBB and [¹⁸F]FDG uptake followed a linear or non-linear trend, data were fitted using both linear and exponential models. Model performance was compared using the corrected Akaike Information Criterion (AICc), which accounts for model complexity and sample size to identify the best-fitting curve. The model with the lowest AICc value was considered to provide the best fit. Comparisons between the experimental groups APPPS1, APP^SAA, and wild-type (WT) were conducted using one-way analysis of variance (ANOVA). Post-hoc tests were applied where appropriate to identify significant group differences. All analyses were performed in GraphPad Prism 10.1.2 and IBM SPSS Statistics 29, with significance defined as p < 0.05. Data are reported as means ± standard deviation, unless indicated otherwise.

Results

Demographics

Dynamic [¹⁸F]FBB–PET scans (0–60 min p.i.) and static [¹⁸F]FDG–PET scans (30–60 min p.i.) were successfully acquired in all animals (n = 92), including 17 APPPS1, 56 APP^SAA, and 19 age- and sex-matched WT mice.^20,21 A small subset of [¹⁸F]FBB–PET and [¹⁸F]FDG–PET scans acquired in APP^SAA mice (n = 8) originated from a previous investigation conducted by our research group, while all remaining data were newly acquired and have not been published to date.²² Differences in group sizes reflect colony availability. Animals ranged in age from 3 to 12 months, with 40.2% being female. Each mouse underwent both PET/MRI scans on separate days, with an interval of 9 ± 12 days between sessions. Detailed demographic data are displayed in Table 1.

Table 1.

Detailed demographic data.

Demographics	Total	Wild-type	APP^SAA	APPPS1
Animal (n)	92	19	56	17
Sex (M/F)	55/37	12/7	31/25	12/5
Age [¹⁸F]FBB (months, mean ± SD)	5.7 ± 2.0	6.0 ± 1.5	6.0 ± 2.1	4.5 ± 1.7
Age [¹⁸F]FDG (months, mean ± SD)	5.5 ± 2.0	5.8 ± 1.4	5.7 ± 2.1	4.5 ± 1.8
Δ Interval [¹⁸F]FBB– and [¹⁸F]FDG–PET (days, mean ± SD)	9 ± 12	9 ± 6	11 ± 14	7 ± 6
Injected dose of [¹⁸F]FBB (MBq, mean ± SD)	12.9 ± 2.2	13.1 ± 2.0	13.2 ± 2.1	12.5 ± 2.3
Injected dose of [¹⁸F]FDG (MBq, mean ± SD)	16.8 ± 3.1	17.4 ± 2.6	16.5 ± 3.7	16.8 ± 2.8

Optimal time window for early-phase [¹⁸F]FBB imaging

Time-activity curves derived from dynamic [¹⁸F]FBB–PET data showed a rapid increase in signal within the first minutes p.i., followed by a distinct uptake peak within the first 5 min (Figure 1). WT animals reached their peak uptake at 2.1 ± 0.9 min p.i. with a SUV of 1.25 ± 0.28. APP^SAA mice exhibited a peak at 2.2 ± 0.7 min p.i. with a SUV of 1.50 ± 0.31. APPPS1 mice showed a later peak at 3.6 ± 1.3 min p.i. with a SUV of 1.45 ± 0.16. After reaching peak values, the signal declined steadily over the remaining 60-min scan.

Figure 1.

Dynamic β-amyloid-PET imaging: (a) representative averaged axial and coronal [¹⁸F]FBB–PET SUV maps of WTs, APP^SAA, and APPPS1 during dynamic acquisition (0–60 min) projected upon a T1 MRI template and (b) averaged time-activity curves of whole-brain [¹⁸F]FBB–PET across the three groups during the dynamic acquisition.

Comparison of [¹⁸F]FDG–PET (30–60 min p.i.) with individual frames from dynamic [¹⁸F]FBB–PET (p.i.: 6 × 10, 6 × 30, 6 × 60, 10 × 300 s) demonstrated the highest agreement across the whole-brain for frames 7–10 corresponding to 1–3 min p.i., considering all three mouse models (Figure 2(a)). Similarly, the highest agreement in the cortex (frames 7–10), hippocampus (frames 7–10), and subcortical regions including the thalamus (frames 7–10) was observed consistently within the first 5 min p.i., with markedly weaker correlations at later time points. Together with the observed averaged perfusion peak at (t = 2.4 ± 1.0 min p.i.), these findings defined the optimal trade-off between correlation and count statistics within the 1–3 min p.i. window. Detailed statistics, including single frame correlation coefficients for all regions, are provided in Supplementary Table 1.

Figure 2.

Agreement between early-phase β-amyloid–PET and FDG–PET: (a) Pearson correlation coefficients (R) between dynamic [¹⁸F]FBB–PET and static [¹⁸F]FDG–PET were calculated for all mice and plotted over time in four representative brain regions and (b) scatter plot depicts association between early-phase [¹⁸F]FBB SUV (1–3 min) and [¹⁸F]FDG SUV (30–60 min) in the 14 predefined brain regions. Correlation coefficients were computed using partial correlation, accounting for genotype as a cofactor to control for potential group-specific effects.

Relationship between early-phase [¹⁸F]FBB and [¹⁸F]FDG uptake

After identifying the optimal time window, correlations between early-phase [¹⁸F]FBB (1–3 min p.i.) and [¹⁸F]FDG scans were performed. Significant positive associations were observed across target brain regions, including the whole-brain (R = 0.53, p < 0.0001), cortex (R = 0.50, p < 0.0001), hippocampus (R = 0.49, p < 0.0001), and thalamus (R = 0.49, p < 0.0001; Supplementary Figure 1). Furthermore, the combined assessment of the predefined 14 target regions exhibited a strong correlation between early-phase [¹⁸F]FBB and [¹⁸F]FDG uptake (R = 0.50, p < 0.0001), consistently accounting for genotype as a covariate (Figure 2(b)). To evaluate whether this relationship deviated from linearity, we compared linear and non-linear (exponential) fits, which confirmed the linear model as the best fit (ΔAICc = 53.5, R² = 0.25 vs 0.22, F(1,1286) = 425.4, probability >99.9%).

Concordant perfusion and glucose metabolism in AD mouse models

Group comparisons revealed clear genotype-dependent alterations in both perfusion and metabolism. APPPS1 and APP^SAA mice exhibited both visually and quantitatively significantly increased early-phase [¹⁸F]FBB uptake compared to WT controls across all regions examined, including the whole-brain (APPPS1: p < 0.0001; APP^SAA: p < 0.0001), cortex (APPPS1: p = 0.0003; APP^SAA: p < 0.0001), hippocampus (APPPS1: p < 0.0001; APP^SAA: p < 0.0001), and thalamus (APPPS1: p = 0.0004; APP^SAA: p < 0.0001; Figure 3(a) and (b) and Supplementary Figure 2). A similar pattern was observed in the [¹⁸F]FDG–PET-derived metabolism, with consistently elevated uptake in the whole-brain (APPPS1: p = 0.0028; APP^SAA: p < 0.0001), cortex (APPPS1: p = 0.0027; APP^SAA: p < 0.0001), hippocampus (APPPS1: p = 0.0049; APP^SAA: p < 0.0001), and thalamus (APPPS1: p = 0.0012; APP^SAA: p < 0.0001) in both transgenic mouse models (Figure 3(a) and (c) and Supplementary Figure 3).

Figure 3.

Comparison of cerebral perfusion and metabolism across AD mouse models: (a) averaged SUV-scaled axial and coronal maps show early-phase [¹⁸F]FBB and [¹⁸F]FDG uptake in both transgenic AD-models (APP^SAA and APPPS1) and WT mice and (b, c) dot plots represent group-wise comparisons of early-phase [¹⁸F]FBB and [¹⁸F]FDG SUV in four representative target regions using repeated measures ANOVA, followed by post hoc pairwise comparisons with Tukey’s test for multiple comparisons.

Discussion

This study demonstrates that early, perfusion-weighted [¹⁸F]FBB–PET phases can serve as a proxy measure of cerebral metabolism in preclinical models of AD. By directly comparing dynamic early-phase [¹⁸F]FBB–PET and conventional [¹⁸F]FDG–PET in two widely used transgenic mouse models of amyloidosis (APPPS1 and APP^SAA),²⁰ we successfully demonstrated a strong and regionally consistent correlation between early β-amyloid-PET tracer uptake and glucose metabolism. Notably, we identified the highest agreement between both modalities within the 1–3 min p.i. window, coinciding with peak perfusion-weighted values, and indicating that this time interval optimally captures perfusion-related signals that mirror metabolic brain changes. Finally, both transgenic models exhibited significantly increased early-phase [¹⁸F]FBB and [¹⁸F]FDG uptake relative to WT controls, reflecting a concordant pattern of elevated perfusion-weighted and metabolic signals.

The overarching aim of this preclinical study was to determine whether dynamic β-amyloid PET provides insights beyond plaque deposition and can serve as a proxy marker of brain metabolism through early-phase imaging. In human studies, the close relationship between metabolism and cerebral perfusion has been well established in regards to AD.^23,24 Building on this, early-phase acquisitions of β-amyloid-PET have been proposed as functional biomarkers that reflect cerebral perfusion-weighted patterns.^6,23 Consistent with these findings in humans, our results in transgenic AD mouse models demonstrate that early-phase β-amyloid-PET can similarly serve as a proxy marker for glucose metabolism, as measured by in vivo [¹⁸F]FDG–PET.

In particular, our findings demonstrate that both [¹⁸F]FDG–PET and early-phase [¹⁸F]FBB–PET revealed consistently elevated uptake across all examined brain regions in both APPPS1 and APP^SAA mouse models compared to WT controls, which aligns with recent literature.^20,25 Of note, among the two AD mouse models, APP^SAA mice exhibited greater variability and higher peak perfusion-weighted values, reflecting the broader age range and higher average age in this group. In addition, the slightly longer and more variable time intervals between both scan timepoints within this cohort may have contributed to increased measurement variability.

Biologically, the Austrian (T714I) and Arctic (E693G) mutations in APP^SAA mice promote aggregation-prone and proteolysis-resistant Aβ species, leading to earlier and more extensive β-amyloid deposition compared to the Swedish mutation (KM670/671NL) alone. APP^SAA mice develop complex β-amyloid plaque morphologies with both diffuse and cored plaques, accompanied by dystrophic neurites, tau phosphorylation, lysosomal alterations, and robust neuroinflammatory activation, including increased microglial density and elevated TREM2 expression.²⁰ Moreover, the combination of Austrian and Arctic mutations alters the temporal trajectory of disease progression and the age of cognitive decline relative to other APP knock-in models, potentially contributing to broader phenotypic heterogeneity.²⁶

Similarly, a recent study using arterial spin labeling MRI reported increased cerebral blood flow in young and middle-aged APPPS1 mice across both cortical and subcortical brain regions.²⁷ This concordant pattern of increased perfusion and glucose metabolism reflects the well-established coupling between cerebral blood flow and metabolic demand, and suggests that early-phase β-amyloid-PET can reliably capture functional alterations associated with β-amyloid pathology in these models.

Currently, [¹⁸F]FDG–PET represents the gold standard in metabolic imaging, providing direct insight into glucose consumption and neuronal or astroglial activity.^8,15,28,29 However, interpreting these results in preclinical models requires careful consideration of the underlying cellular mechanisms contributing to PET signals. Recent studies have demonstrated that [¹⁸F]FDG–PET signal is strongly influenced by microglial glucose consumption, with Trem2-deficient mice showing reduced [¹⁸F]FDG signal due to attenuated microglial activity.^8,16 Conversely, aging WT mice and β-amyloid models with activated microglia exhibit hypermetabolism, indicating that [¹⁸F]FDG uptake in these animals may reflect neuroinflammation more than neuronal function.^8,18,30 These findings are bolstered by data showing that microglial activation state influences network-level metabolic connectivity in mouse brains.¹⁵

In addition to microglia, astrocytes substantially contribute to FDG uptake. Activation of astroglial glutamate transport has been shown to substantially affect [¹⁸F]FDG signal intensity, indicating that astrocytic glucose metabolism contributes to the coupling between synaptic transmission and cerebral glucose utilization.²⁸

Moreover, human studies have reported reductions in circulating lactate and pyruvate levels with preserved ketone-body concentrations in AD, consistent with impaired glycolytic and oxidative metabolism rather than adaptive substrate switching.³¹ In this context, the elevated [¹⁸F]FDG uptake observed in AD mouse models may also indicate altered glycolytic activity associated with glial activation. These observations highlight that alterations in both energy substrate utilization and cellular metabolism should be considered when interpreting [¹⁸F]FDG–PET findings in preclinical models and patients.

However, not only astroglial activation but also neuronal hyperactivity during early stages of AD can contribute to [¹⁸F]FDG hypermetabolism, as demonstrated in human studies.^32,33 Moreover, Aβ-dependent increases in functional connectivity have been shown to accelerate the spread of tau pathology from epicenters to connected brain regions via neuronal hyperactivity, ultimately resulting in transient increases of glucose metabolism during early AD.³⁴ This concept is further supported by our recent preclinical data of combined β-amyloid-PET and SV2A-PET, revealing localized increases in synaptic density surrounding Aβ plaques.³⁵

As dual-purpose β-amyloid-PET is increasingly leveraged for both pathological and functional imaging, defining an optimal early-phase acquisition window was a key objective of this study to ensure the translational value of early-phase PET metrics. In clinical settings, optimal early-phase windows for β-amyloid-PET tracers align with perfusion peaks: for [¹⁸F]FBB and [¹¹C]PiB, this typically occurs within the first 6–10 min.^10,36–39 The substantially faster bolus transit and mean transit times observed in murine cerebral vasculature, relative to humans, mandate an earlier window to accurately capture peak tracer delivery in preclinical studies and account for species-specific hemodynamic differences.^40–43 Given rodents’ significantly smaller brain volumes and faster circulatory dynamics, our 1–3-min window effectively mirrors the perfusion peak in mice, while maintaining translational consistency.

In addition to selecting an appropriate time window, the choice of quantification method is critical to ensure an accurate interpretation of PET signals in transgenic models. Our decision to use SUV instead of SUVR or kinetic modeling was guided by key practical and physiological considerations specific to small-animal PET. Arterial input function sampling is technically challenging in mice and ethically constrained.⁴⁴ Moreover, [¹⁸F]FDG–PET quantification via SUV has been shown to yield stable and physiologically meaningful results in both anesthetized and awake mice.⁴⁵ In contrast, SUVR requires the presence of a stable reference region, an assumption, that is, difficult to meet in transgenic AD models due to global metabolic alterations. Our approach is consistent with previous [¹⁸F]FDG–PET studies in APPPS1 and APP^SAA mice, which also employed SUV-based quantification.^20,46 In our cohort, we observed pronounced genotype-dependent increases in both perfusion and metabolism across all brain regions examined, reflecting a global effect. These alterations align with the underlying β-amyloid pathology characteristic of the two models used, APPPS1 and APP^SAA. Both, while differing substantially in plaque kinetics and associated physiological responses, are well-established models of progressive amyloidosis, enhancing the generalizability of our findings.^20,47 The widespread nature of these changes further challenges the validity of reference-based normalization approaches and reinforces the suitability of SUV as a robust and unbiased quantification method in this preclinical setting.

Nevertheless, some limitations of this study should be acknowledged. First, the cross-sectional study design may not fully capture the dynamic temporal changes in perfusion and metabolism that occur throughout disease progression. Longitudinal studies will be essential to validate the stability and predictive value of early-phase β-amyloid-PET as a metabolic proxy across different disease stages. Of note, our cohort did not include very old mice or models with advanced neurodegeneration. Therefore, according to the A/T/N framework, we cannot definitively classify the observed metabolic and perfusion-weighted changes as markers of neuronal injury (N). While we demonstrated a strong correlation between early-phase perfusion and glucose metabolism, both parameters were elevated in transgenic mice, likely reflecting a combination of inflammation and compensatory responses rather than neurodegeneration. Future studies should include mouse models exhibiting pronounced neuronal loss and synaptic degeneration, such as aged 5XFAD⁴⁸ or Niemann–Pick type C (NPC) knockout mice,⁴⁹ to further clarify the validity of early-phase amyloid-PET as a true marker of neurodegenerative processes. While our work focused on imaging-based biomarkers aligned with the A/T/N framework,¹ it is important to note that this classification does not include cognitive or behavioral dimensions. Cognitive decline and behavioral alterations are key clinical manifestations of AD, yet they are seldom systematically assessed in preclinical imaging studies. Indeed, behavioral phenotypes such as anxiety-like behavior and memory deficits can emerge early in certain AD models, as reported in 3xTg-AD mice.⁵⁰ The potential mismatch between biomarker-defined stages and cognitive outcomes highlights an important translational challenge. Future studies integrating longitudinal PET imaging with behavioral testing could therefore provide a more comprehensive view of disease progression and therapeutic efficacy in preclinical models.

Conclusion

Early-phase [¹⁸F]FBB β-amyloid-PET captures functional brain alterations with strong regional concordance to [¹⁸F]FDG–PET. Combining pathological and functional information in a single scan enhances the efficiency of preclinical imaging while reducing costs, scan time, and animal burden in line with 3R principles. This integrated approach is particularly valuable for longitudinal preclinical studies and enables a more precise phenotyping of disease models with fewer resources.

Supplemental Material

sj-docx-1-jcb-10.1177_0271678X261420082 – Supplemental material for Assessment of early-phase [18F]florbetaben images as a proxy for brain metabolism in mouse models of Alzheimer’s disease

Supplemental material, sj-docx-1-jcb-10.1177_0271678X261420082 for Assessment of early-phase [18F]florbetaben images as a proxy for brain metabolism in mouse models of Alzheimer’s disease by Robin Gröger, Amelie L Englert, Manvir Lalia, Lis de Weerd, Matteo Rovere, Lea Kunze, Giovanna Palumbo, Karin Wind-Mark, Sonja Fixemer, Sebastian N Roemer-Cassiano, Nicolai Franzmeier, Günter Höglinger, Andreas Zwergal, Sabina Tahirovic, Rudolf A Werner, Matthias Brendel and Johannes Gnörich in Journal of Cerebral Blood Flow & Metabolism

Footnotes

Acknowledgements

We thank Rosel Oos and Karin Bormann-Giglmaier for excellent technical support during small animal PET imaging. We thank Mathias Jucker for providing the APPPS1 mice.

Author contributions

Conceptualization: ST, MB, JG. Methodology: KW-M, RAW, MB, JG. Validation: NF, GH, AZ, ST, RAW, MB, JG. Formal analysis: RG, ALE, ML, LK. Investigation: RG, ALE, ML, LdW, MR, LK, GP, KW-M, SF. Resources: LdW, MR, SF, ST, RAW, MB, JG. Data curation: SNR-C, NF, MB. Writing—original draft: RG, ALE. Writing—review and editing: NF, AZ, ST, MB, JG. Visualization: RG, ALE. Supervision: ST, RAW, MB, JG. Project administration: JG.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: JG is funded by the Munich Clinician Scientist Program (MCSP). NF and MB were funded by the Deutsche Forschungsgemeinschaft (DFG) under Germany’s Excellence Strategy within the framework of the Munich Cluster for Systems Neurology (EXC 2145 SyNergy, ID 390857198). MR thanks the Hector Fellow Academy for support.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: MB is a member of the Neuroimaging Committee of the EANM. MB has received speaker honoraria from Roche, GE Healthcare, Iba, and Life Molecular Imaging; has advised Life Molecular Imaging and GE healthcare; and is currently on the advisory board of MIAC. AZ has received speaker honoraria from Dr. Willmar Schwabe GmbH, AstraZeneca and Pfizer. He got research support from Dr. Willmar Schwabe GmbH; and is currently on the advisory board of Vertigenius.

Ethical considerations

All animal experiments were approved by the local animal care committee of the Government of Upper Bavaria (Regierung Oberbayern; ROB-55.2-2532.Vet_02-22-125).

Consent to participate

Not applicable.

Consent for publication

Not applicable.

ORCID iDs

Lea Kunze

Sebastian N Roemer-Cassiano

Matthias Brendel

Johannes Gnörich

Data availability statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Supplemental material

Supplemental material for this article is available online.

References

Jack

Jr Bennett

Blennow

, et al. A/T/N: an unbiased descriptive classification scheme for Alzheimer disease biomarkers. Neurology 2016; 87: 539–547.

Jack

Jr Knopman

Jagust

, et al. Tracking pathophysiological processes in Alzheimer’s disease: an updated hypothetical model of dynamic biomarkers. Lancet Neurol 2013; 12: 207–216.

Sabri

Sabbagh

Seibyl

, et al. Florbetaben PET imaging to detect amyloid beta plaques in Alzheimer’s disease: phase 3 study. Alzheimers Dement 2015; 11: 964–974.

Aye

WWT

Stark

Horne

, et al. Early-phase amyloid PET reproduces metabolic signatures of cognitive decline in Parkinson’s disease. Alzheimers Dement 2024; 16: e12601.

Völter

Beyer

Eckenweber

, et al. Assessment of perfusion deficit with early phases of [¹⁸F]PI-2620 tau-PET versus [¹⁸F]flutemetamol–amyloid–PET recordings. Eur J Nucl Med Mol Imaging 2023; 50: 1384–1394.

Boccalini

Peretti

Ribaldi

, et al. Early-phase ¹⁸F-florbetapir and ¹⁸F-flutemetamol images as proxies of brain metabolism in a memory clinic setting. J Nucl Med 2022; 64: 266–273.

Russell

Burch

The principles of humane experimental technique. London: Methuen & Co. Ltd., 1959, p.xiv + 238.

Xiang

Wind

Wiedemann

, et al. Microglial activation states drive glucose uptake and FDG–PET alterations in neurodegenerative diseases. Sci Transl Med 2021; 13: eabe5640.

Vanhoutte

Landeau

Sherif

, et al. Evaluation of the early-phase [¹⁸F]AV45 PET as an optimal surrogate of [¹⁸F]FDG PET in ageing and Alzheimer’s clinical syndrome. Neuroimage Clin 2021; 31: 102750.

10.

Myoraku

Klein

Landau

, et al. Regional uptakes from early-frame amyloid PET and ¹⁸F-FDG PET scans are comparable independent of disease state. Eur J Hybrid Imaging 2022; 6: 2.

11.

Phillips

Chan

Zheng

, et al. Neurovascular coupling in humans: physiology, methodological advances and clinical implications. J Cereb Blood Flow Metab 2016; 36: 647–664.

12.

Stackhouse

Mishra

Neurovascular coupling in development and disease: focus on astrocytes. Front Cell Dev Biol 2021; 9: 702832.

13.

Raichle

Mintun

MA.

Brain work and brain imaging. Annu Rev Neurosci 2006; 29: 449–476.

14.

Rominger

Brendel

Burgold

, et al. Longitudinal assessment of cerebral β-amyloid deposition in mice overexpressing Swedish mutant β-amyloid precursor protein using ¹⁸F-florbetaben PET. J Nucl Med 2013; 54: 1127–1134.

15.

Gnörich

Reifschneider

Wind

, et al. Depletion and activation of microglia impact metabolic connectivity of the mouse brain. J Neuroinflammation 2023; 20: 47.

16.

Reifschneider

Robinson

van Lengerich

, et al. Loss of TREM2 rescues hyperactivation of microglia, but not lysosomal deficits and neurotoxicity in models of progranulin deficiency. EMBO J 2022; 41: e109108.

17.

Gnörich

Koehler

Wind-Mark

, et al. Towards multicenter β-amyloid PET imaging in mouse models: a triple scanner head-to-head comparison. Neuroimage 2024; 297: 120748.

18.

Bartos

Kunte

Wagner

, et al. Astroglial glucose uptake determines brain FDG–PET alterations and metabolic connectivity during healthy aging in mice. Neuroimage 2024; 300: 120860.

19.

Hof

Grant

, et al. A three-dimensional digital atlas database of the adult C57BL/6J mouse brain by magnetic resonance microscopy. Neuroscience 2005; 135: 1203–1215.

20.

Xia

Lianoglou

Sandmann

, et al. Novel App knock-in mouse model shows key features of amyloid pathology and reveals profound metabolic dysregulation of microglia. Mol Neurodegener 2022; 17: 41.

21.

Biechele

Monasor

Wind

, et al. Glitter in the darkness? Nonfibrillar β-amyloid plaque components significantly impact the β-amyloid PET signal in mouse models of Alzheimer disease. J Nucl Med 2022; 63: 117–124.

22.

de Weerd

Hummel

Müller

, et al. Early intervention anti-Aβ immunotherapy attenuates microglial activation without inducing exhaustion at residual plaques. Mol Neurodegener 2025; 20: 92.

23.

Tiepolt

Hesse

Patt

, et al. Early [¹⁸F]florbetaben and [¹¹C]PiB PET images are a surrogate biomarker of neuronal injury in Alzheimer’s disease. Eur J Nucl Med Mol Imaging 2016; 43: 1700–1709.

24.

Daerr

Brendel

Zach

, et al. Evaluation of early-phase [¹⁸F]-florbetaben PET acquisition in clinical routine cases. Neuroimage Clin 2017; 14: 77–86.

25.

Poisnel

Hérard

El Tannir

Tayara

, et al. Increased regional cerebral glucose uptake in an APP/PS1 model of Alzheimer’s disease. Neurobiol Aging 2012; 33: 1995–2005.

26.

Blackmer-Raynolds

Lipson

Fraccaroli

, et al. Longitudinal characterization reveals behavioral impairments in aged APP knock in mouse models. Sci Rep 2025; 15: 4631.

27.

Guo

Zhang

, et al. Age‑ and brain region‑associated alterations of cerebral blood flow in early Alzheimer’s disease assessed in AβPPSWE/PS1ΔE9 transgenic mice using arterial spin labeling. Mol Med Rep 2019; 19: 3045–3052.

28.

Zimmer

Parent

Souza

, et al. [¹⁸F]FDG PET signal is driven by astroglial glutamate transport. Nat Neurosci 2017; 20: 393–395.

29.

Bouter

Henniges

Franke

, et al. ¹⁸F-FDG–PET detects drastic changes in brain metabolism in the Tg4–42 model of Alzheimer’s disease. Front Aging Neurosci 2018; 10: 425.

30.

Bouter

¹⁸F-FDG–PET in mouse models of Alzheimer’s disease. Front Med 2019; 6: 71.

31.

Verri

Aquilani

Ricevuti

, et al. Plasma energy substrates at two stages of Alzheimer’s disease in humans. Int J Immunopathol Pharmacol 2018; 32: 2058738418817707.

32.

Habeck

Madison

, et al. Covarying alterations in Aβ deposition, glucose metabolism, and gray matter volume in cognitively normal elderly. Hum Brain Mapp 2014; 35: 297–308.

33.

Rubinski

Franzmeier

Neitzel

, et al. FDG–PET hypermetabolism is associated with higher tau-PET in mild cognitive impairment at low amyloid-PET levels. Alzheimers Res Ther 2020; 12: 133.

34.

Roemer-Cassiano

Wagner

Evangelista

, et al. Amyloid-associated hyperconnectivity drives tau spread across connected brain regions in Alzheimer’s disease. Sci Transl Med 2025; 17: eadp2564.

35.

Kunze

Palumbo

Gnörich

, et al. Fibrillar amyloidosis and synaptic vesicle protein expression progress jointly in the cortex of a mouse model with β-amyloid pathology. Neuroimage 2025; 310: 121165.

36.

Oliveira

FPM

Moreira

de Mendonça

, et al. Can ¹¹C-PiB–PET relative delivery R1 or ¹¹C-PiB–PET perfusion replace ¹⁸F-FDG–PET in the assessment of brain neurodegeneration? J Alzheimers Dis 2018; 65: 89–97.

37.

Rostomian

Madison

Rabinovici

, et al. Early ¹¹C-PIB frames and ¹⁸F-FDG PET measures are comparable: a study validated in a cohort of AD and FTLD patients. J Nucl Med 2011; 52: 173–179.

38.

Rodriguez-Vieitez

Leuzy

Chiotis

, et al. Comparability of [¹⁸F]THK5317 and [¹¹C]PIB blood flow proxy images with [¹⁸F]FDG positron emission tomography in Alzheimer’s disease. J Cereb Blood Flow Metab 2017; 37: 740–749.

39.

Gnörich

Kusche-Palenga

Kling

, et al. Assessment and staging of A/T/N with a single dynamic [¹⁸F]PI-2620 recording. medRxiv 2025; 01.14.25320240.

40.

Wei

Bibic

, et al. Toward accurate cerebral blood flow estimation in mice after accounting for anesthesia. Front Physiol 2023; 14: 1169622.

41.

Hirschler

Munting

Khmelinskii

, et al. Transit time mapping in the mouse brain using time-encoded pCASL. NMR Biomed 2018; 31(2): 3855.

42.

Karbowski

Scaling of brain metabolism and blood flow in relation to capillary and neural scaling. PLoS One 2011; 6: e26709.

43.

Lee

, et al. Whole-brain perfusion mapping in mice by dynamic BOLD MRI with transient hypoxia. J Cereb Blood Flow Metab 2022; 42: 2270–2286.

44.

Kuttner

Luppino

Convert

, et al. Deep-learning-derived input function in dynamic [¹⁸F]FDG PET imaging of mice. Front Nucl Med 2024; 4: 1372379.

45.

Ruch

Gnörich

Wind

, et al. Validity and value of metabolic connectivity in mouse models of β-amyloid and tauopathy. Neuroimage 2024; 286: 120513.

46.

Takkinen

López-Picón

Al Majidi

, et al. Brain energy metabolism and neuroinflammation in ageing APP/PS1–21 mice using longitudinal ¹⁸F-FDG and ¹⁸F-DPA-714 PET imaging. J Cereb Blood Flow Metab 2017; 37: 2870–2882.

47.

Yan

Bero

Cirrito

, et al. Characterizing the appearance and growth of amyloid plaques in APP/PS1 mice. J Neurosci 2009; 29: 10706–10714.

48.

Macdonald

DeBay

Reid

, et al. Early detection of cerebral glucose uptake changes in the 5XFAD mouse. Curr Alzheimer Res 2014; 11: 450–460.

49.

Dinkel

Hummel

Zenatti

, et al. Myeloid cell-specific loss of NPC1 in mice recapitulates microgliosis and neurodegeneration in patients with Niemann–Pick type C disease. Sci Transl Med 2024; 16: eadl4616.

50.

Roda

Esquerda-Canals

Martí-Clúa

, et al. Cognitive impairment in the 3xTg-AD mouse model of Alzheimer’s disease is affected by Aβ-immunotherapy and cognitive stimulation. Pharmaceutics 2020; 12(10): 944.

Assessment of early-phase [ 18 F]florbetaben images as a proxy for brain metabolism in mouse models of Alzheimer’s disease

Abstract

Keywords

Introduction

Materials and methods

Animals and study design

PET/MRI imaging

Image preprocessing and analysis

Statistical procedures

Results

Demographics

Optimal time window for early-phase [18F]FBB imaging

Relationship between early-phase [18F]FBB and [18F]FDG uptake

Concordant perfusion and glucose metabolism in AD mouse models

Discussion

Conclusion

Supplemental Material

sj-docx-1-jcb-10.1177_0271678X261420082 – Supplemental material for Assessment of early-phase [18F]florbetaben images as a proxy for brain metabolism in mouse models of Alzheimer’s disease

Footnotes

Acknowledgements

Author contributions

Funding

Declaration of conflicting interests

Ethical considerations

Consent to participate

Consent for publication

ORCID iDs

Data availability statement

Supplemental material

References

Optimal time window for early-phase [¹⁸F]FBB imaging

Relationship between early-phase [¹⁸F]FBB and [¹⁸F]FDG uptake