To our knowledge this is the first study investigating a specific PET radiotracer of 5-HT6 receptors in a population of AD patients with different stages of disease. The aim was to investigate [18F]2FNQ1P binding and quantify 5-HT6 receptor density in AD patients.
Binding assays showed reliable affinity of [18F]2FNQ1P for 5-HT6 receptors. Radioligand saturation binding assays found no significant difference between the “Not” and “High” groups on the KD and the Bmax parameters. The high affinity of [18F]2FNQ1P for 5-HT6 was confirmed by the mean KD for each group (around 2.5 nM). This result is consistent with a previous study which found a KD of approximatively 1 nM [26]. Although Bmax was lower in the “High” than the “Not” group, the difference was non-significant. These results could be explained by the small size of the study population (n = 8). This explanation is frequently put forward in postmortem studies, as access to brains of neuropathologically documented AD subjects is limited. However, further investigations with a larger sample might demonstrate a significant difference, such as was found in the autoradiography assay.
Quantitative autoradiography experiments showed that [18F]2FNQ1P was suitable for visualizing and quantifying striatal 5-HT6 receptor density. In-vitro competition assays with the 5-HT6 antagonist SB–258585 showed displacement of the radiotracer, confirming binding reversibility. Quantification of 5-HT6 receptors revealed significant differences between the control group (i.e., patients without AD neuropathological modifications: “Not”) and patients with a high level of neuropathological modification (“High”). The results also distinguished the control group (“Not”) from all AD stages. On the other hand, no significant differences were observed between the “Not” or ‘ “High” groups and other stages of AD (“Low” and “Int”). Despite these non-significant differences, the results showed a progressive decrease in the 5-HT6 receptor expression according to AD stage, with a strong negative correlation between specific binding of [18F]2FNQ1P and disease stage. Here again, the absence of significant difference between the “High” group and earlier stages of AD can be initially explained by the relatively low number of patients in each group. Larger groups would reinforce the power of the study and could demonstrate significant differences. Another explanation concerns the complexity of assessing Alzheimer’s disease. Because of the large inter-individual variability found in AD, classification can be difficult. This is why many classifications exist, drawn up by different working groups such as the IWG–2 classification of the International Working Group for New Research Criteria for the Diagnosis of Alzheimer’s Disease [32] and, more recently, the “ABC” classification of the National Institute of Aging-Alzheimer’s Association (NIA-AA) [29], used in the present study. This classification considers AD as a continuum, needing to be diagnosed in its early stages. However, these criteria remain to be validated. Lowe and al. applied this classification to the Alzheimer Disease Neuroimaging Initiative (ADNI) cohort [33]. The study showed the weakness of the NIA-AA criteria because of the complexity of interpreting biomarkers. Other studies assessing the applicability of this classification in AD patients with mild cognitive impairment showed the same weakness: variability in biomarker interpretation, lack of measurement standardization, and varying results [34,35]. Furthermore, the “ABC” staging is strongly nested in other diseases contributing to cognitive impairment, such as Lewy body disease, vascular brain injury or hippocampal sclerosis [36]. Thus, patient groups are difficult to distinguish, making it difficult to show significant differences between neighboring groups.
In addition, “ABC” staging is mostly based on anatomopathological criteria. The classification takes account of onset of Aβ or amyloid plaques, neurofibrillary tangles and neuritic plaques, but not of clinical status. However, it is well known that anatomopathological modifications in AD correlate with disease symptomatology. A Symptom onset correlates with brain atrophy [37], hypometabolism [38] and neurofibrillary changes [39]. Thus, the correlation between decreased 5-HT6 receptor density and decrease in symptomatology could be an approach worth considering in future studies.
Another point to be taken into account in interpreting the present results concerns the polymorphism of the 5-HT6 receptor. The human 5-HT6 receptor protein is coded by chromosome 1 and the specific gene of the receptor contains many trinucleotide polymorphisms [5]. Some previous studies, in Chinese and Korean populations, showed an association between this polymorphism and AD, and considered this allele to be a risk factor [40,41]. Another study, in a Caucasian population, found no significant differences between controls and AD patients [42], and considered this 5-HT6 receptor polymorphism as a silent mutation that does not affect the function of the protein. To date, an association between 5-HT6 polymorphism in AD patients and the binding properties of [18F]2FNQ1P is still an open question.
Although the present study is the first to visualize 5-HT6 receptors in AD patients with a radiopharmaceutical usable for neuroimaging in vivo, a few previous studies also showed a decrease in these receptors in AD and related disorders, using dedicated in-vitro exploration probes. Lorke et al.,in an immunohistochemistry study [43], reported decreased cellular expression of 5-HT6 receptors in the prefrontal cortex of AD patients in comparison with normal age-matched subjects. 5-HT6 receptors were expressed in pyramidal cells and stellate-shaped cells, and AD patients showed a significant 40% decrease in 5-HT6 receptor expression. Garcia-Alloza et alreportedsimilar results. They assessed the involvement of serotoninergic disturbance of 5-HT6 receptors in AD impairment. Binding assays with [125I]SB–258585, an in-vitro radiotracer, were performed on tissue samples from frontal and temporal cortex. Results showed a significant decrease in 5-HT6 receptor density in both frontal (56% reduction) and temporal (58% reduction) cortex in AD patients compared to controls [19]. While these results were consistent with 5-HT6 receptor involvement in AD, the authors did not use the clinicopathologic diagnostic classification used in the present study.
The present study also differs from a majority of previous studies in the choice of the brain region of interest: the caudate nucleus, well-known to be a region rich in 5-HT6 receptors [44,45]. It is known that 5-HT6 receptors are highly and homogeneously concentrated in the caudate nucleus, putamen and nucleus accumbens [46]. The first preclinical explorations of our radiopharmaceutical confirmed its preferential striatal binding [26,27,30], justifying the choice of this region for the present postmortem study. However, it must be recognized that the pathophysiological involvement of the caudate nucleus in AD is not yet well established. Brain atrophy and neuron loss occurs mainly in the frontal cortex, hippocampus and limbic areas [47]. De Jong et al. suggested a distinction in striatal pattern morphology in AD patients compared to subjects with memory complaints without objective cognitive impairment [48]. Their study showed that cognitive impairment is related to the degree of surface deformity, hypothesizing that associative and limbic cerebral networks are primarily affected in AD. These findings highlight the interest of tracing the progression of 5-HT6 receptors as striatal marker during neurodegeneration.
This is all the more interesting as the 5-HT6 receptor is also a potential therapeutic target in Alzheimer’s disease [49]. Recently, drug-candidates have been developed on the pharmacological hypothesis that 5-HT6 receptor blockade induces acetylcholine release and so improves cognition processes in AD [50]. So far, clinical studies were disappointing. Idalopirdine, a 5-HT6 receptor antagonist, did not improve cognition in patients with mild to moderate AD compared to placebo [21]; intepirdine, another selective 5-HT6 receptor antagonist, did not show efficacy in a phase II study [51]. However, these early studies do not signify the abandonment of therapeutic targeting of 5-HT6 receptors. Their main limitations were the lack of knowledge of 5-HT6 receptor status in groups of heterogeneous AD patients. The difficulty of staging patients may have led to underdosing in phase III compared to phase II [52]. It is therefore crucial to have better knowledge of the progression of 5-HT6 receptors during the course of Alzheimer’s disease, through PET imaging, to highlight potential early biomarkers, to refine the recruitment for future therapeutic trials and, finally, to contribute to the understanding of 5-HT6 drug-candidates by drug occupancy studies. The translational results we obtained with the radiopharmaceutical [18F]2FNQ1P therefore encourage us to implement a PET neuroimaging study in AD.