Reduced Expression of Cerebral Metabotropic Glutamate Receptor Subtype 5 in Men with Fragile X Syndrome

Glutamatergic receptor expression is mostly unknown in adults with fragile X syndrome (FXS). Favorable behavioral effects of negative allosteric modulators (NAMs) of the metabotropic glutamate receptor subtype 5 (mGluR5) in fmr1 knockout (KO) mouse models have not been confirmed in humans with FXS. Measurement of cerebral mGluR5 expression in humans with FXS exposed to NAMs might help in that effort. We used positron emission tomography (PET) to measure the mGluR5 density as a proxy of mGluR5 expression in cortical and subcortical brain regions to confirm target engagement of NAMs for mGluR5s. The density and the distribution of mGluR5 were measured in two independent samples of men with FXS (N = 9) and typical development (TD) (N = 8). We showed the feasibility of this complex study including MRI and PET, meaning that this challenging protocol can be accomplished in men with FXS with an adequate preparation. Analysis of variance of estimated mGluR5 expression showed that mGluR5 expression was significantly reduced in cortical and subcortical regions of men with FXS in contrast to age-matched men with TD.


Background
Fragile X syndrome (FXS) is caused by expansion full mutation (≥200 CGGs) of the fragile X mental retardation 1 (FMR1) gene leading to epigenetic silencing of the gene, resulting in reduction of its product: fragile X mental retardation protein (FMRP) [1]. FXS is the leading single-gene cause of inherited intellectual disability (ID) and autism spectrum disorder (ASD) [2,3]. Indeed, studies of humans with FXS have consistently demonstrated a wide range of global neurobehavioral impairments [4][5][6][7][8][9]. This is not surprising, as FMRP controls translation around 4% of mRNA in human brains. To illustrate, FMRP binds brain mRNAs, inhibits synthesis of a myriad of proteins [10], and increases the dosages of FMRP target proteins (over 600 to date) of relevance to ASD [11]. The FMRP expression in the brain is the ultimate factor determining the severity of the neurobehavioral phenotype [12]. The absence of adequate FMRP results in overactive glutamatergic signaling of group 1 metabotropic (mGluR 1 and mGluR 5 ) pathways, and consequently overactive downstream signaling cascades, such as the mammalian target of rapamycin (mTOR) and mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK). The overactive downstream signaling leads to excessive protein synthesis in an fmr1 knockout (KO) mouse model [13]. Namely, abnormal mGluR 5 -modulated long-term depression (LTD) in the hippocampus in the fmr1 KO model led to "the mGluR 5 theory" of neuronal dysfunction in FXS [14]. Indeed, the abnormal signaling in the absence of FMRP is associated with aberrant synaptic plasticity and immature dendritic spine morphology. The abnormal excitation-inhibition that leads to an excessive de novo protein synthesis also occurs in humans with FXS [15][16][17][18]. Targeted treatment studies using mGluR 5 negative allosteric modulators (NAMs) then unfolded in both the fmr1 KO mouse model and in humans with FXS [3,12]. Yet, mGluR 5 expression in animal studies and in autopsy studies of humans with FXS produced inconsistent results. Moreover, mGluR 5 expression in vivo has not been measured in humans with FXS.
Although a necropsy study pooling human brains with FXS and premutation of the FMR1 gene (PM,  showed increased mGluR 5 s and marginal protein overexpression [19], these studies do not exist in the living human brain. Since the limited necropsy findings may represent the changes in agonal and post-mortem periods, in vivo measurement of the expression of mGluR 5 s is needed, which may bring an initial insight into failed clinical trials that used investigational agents acting on mGluR 5 in humans FXS. Novel, specific mGluR 5 ligands that allow quantitative measurement of the density and distribution of mGluR 5 s in the brain, such as 3-[ 18  Indeed, quantification of mGluR 5 expression in the living human brain of men with FXS is needed to help understand results of past mGluR 5 trials in humans with FXS, and to help provide information for successful clinical trial designs. For example, an alteration of expression of mGluR 5 s, such as internalization of membrane mGluR 5 s, may be one possible explanation for the negligible therapeutic effect of NAMs in "failed" clinical trials of humans with FXS [20]. Since proteins and receptors occupy different locations on the membranes, the living brain may show protein overexpression and reduction of receptors due to receptor internalization or other alterations. Thus, the use of [ 18 F]FPEB may serve as an effective tool to confirm target engagement of NAMs for mGluR 5 s.

Measurement of mGluR 5 s in the Living Human Brain
While several techniques exist to estimate the concentration of glutamate in the living brain, including magnetic resonance imaging (MRI) and brain biopsy, positron emission tomography (PET) uniquely provides the optimal means to measure mGluR 5 s. For these reasons, radiotracers that bind to mGluR 5 in the living brain and can be visualized with PET are promising tools to quantify the density and the distribution of mGluR 5 s in humans with FXS.
Brain Sci. 2020, 10, x FOR PEER REVIEW  3 of 18 uniquely provides the optimal means to measure mGluR5s. For these reasons, radiotracers that bind to mGluR5 in the living brain and can be visualized with PET are promising tools to quantify the density and the distribution of mGluR5s in humans with FXS.

Recruiting Sites
The study is approved by Johns Hopkins Medicine Institutional Review Board IRB 169 249. The protocols for the study of humans with FXS were approved by the Institutional Review Boards of the Institute for Neurodegenerative Disorders (IND) in New Haven, Connecticut [52] and Johns Hopkins University (JHU) in Baltimore, Maryland [53,54]. Since exposure to radioactivity in PET constitutes greater than minimal risk, this pilot study was restricted to adults. Written informed consent was obtained from each participant at both locations.

Inclusion Criteria
Inclusion criteria for all subjects were age 18-60 years and a diagnosis of FXS based on FMR1 DNA gene testing by PCR/Southern Blot, supplemented by clinical neurobehavioral profiling [52].

Exclusion Criteria
Exclusion criteria were clinically significant abnormal laboratory values and/or clinically significant unstable serious medical, neurological, or psychiatric illnesses [52].

Institute for Neurodegenerative Disorders (IND)
Participants with FXS had completed genetic and other evaluations before traveling to the IND with a caregiver. One day after arrival to the IND, they underwent a screening assessment to confirm the inclusion and exclusion criteria, neuropsychological evaluation, mock scanner training, and PET scans. Participants with TD were recruited from community residents.
Seven men with FXS (mean age 25 ± 5, range 23-34 years) recruited from Rush University Medical Center, Chicago, Illinois, and three age-matched men with typical development (TD) (mean aged 32 ± 4, range 27-39 years) participated in the protocol. Clinical and demographic data [55] confirmed that all participants met the criteria to receive the adult dose of 185 megabecquerels (MBqs) (5 millicuries (mCis)) of [ 18 F]FPEB.

Johns Hopkins University (JHU)
Four men with FXS (mean age 28 ± 9, range 19-41 years) were recruited from the Kennedy Krieger Institute, Baltimore, Maryland, and Rush University Medical Center, including referrals from the Fragile X Online Registry With Accessible Research Database (FORWARD) of the National Fragile X Foundation (NFXF) funded by the Centers for Disease Control and Prevention (CDC), Atlanta, Georgia. The results of two of the four men with FXS (mean age 25.5 ± 2.1, range 24-27) who completed PET scans were reported in this article. Findings were compared and contrasted with five age-matched historical control men with TD who had already completed similar protocols (mean age 29.6 ± 6.02, range 24-39 years) [25,30]. Clinical and demographic data confirmed that all participants met the criteria to receive the adult dose of 185 MBqs (5 mCis) of [ 18 F]FPEB [55].

Institute for Neurodegenerative Disorders (IND)
Assessments of participants with FXS (600 to 1600 CGGs) included mean FMRP of 0.047 ± 0.04 ng/microgram total protein (reference mean FMRP of 0.87 for healthy normal controls with TD), reading level under first grade level, and scores for the Dementia Screening Questionnaire for Individuals with Intellectual Disabilities (DSQIID) [56] ranging from 0 to 2 [55].

Positron Emission Tomography (PET)
IND With the head stabilized by a gauze strip taped across the forehead and a rounded head holder, each participant received an intravenous bolus injection of 185 MBqs (5 mCis) of [ 18 F]FPEB [30] at 1 PM, followed by scans on an ECAT EXACT HR+ PET attaining an axial resolution of approaching y = 4-5 mm [61], with consecutive 6 × 300 s frames performed for 90 to 120 min after the injection time.
Statistical parametric mapping (SPM) [33] was applied to PET frames to obtain regional time (radioactivity) curves (TACs). The ratio of uptake in the volumes of interest (VOIs) to the uptake in the whole cerebellum, a reference region with minimal [ 18 F]FPEB uptake [28,29], was calculated.
JHU MRI was performed an hour before PET. Each participant with FXS underwent training using a mock scanner [62][63][64]. Each participant had a custom fitted face mask made by nuclear medical technologists to hold the head in the same position throughout the scan [65,66]. After receiving intravenous bolus injections of 185 MBq (5 mCis) of [ 18 F]FPEB (30), participants underwent PET scans on a high resolution research tomograph (HRRT), attaining an axial resolution approaching 2.3 to 2.5 mm [67,68] at 1 PM for 90 min.
VOIs were obtained automatically of cortical regions with Freesurfer 6.0 [69,70] and of subcortical regions with the subcortical segmentation tools of the software library of the Oxford Centre for fMRI of the Brain [71][72][73]. VOIs were transferred from MRI to PET space according to MRI-to-PET co-registration parameters obtained with the co-registration module [74,75] of statistical parametric mapping (SPM) [33] and applied to PET frames to obtain regional TACs. With the cerebellar white matter as the reference VOI [28,29], regional BP ND s [32] were obtained by reference tissue graphical analysis (RTGA) [76,77].

Comparisons and Contrasts of Cohorts from the IND and JHU
In order to directly compare and contrast data from both cohorts including nine men with FXS (mean age 27.21 ± 4.17, range 22.3-33.6) and eight historical control age-matched men with TD who had already completed similar protocols (mean age 30.63 ± 5.58, range 24-39 years) [25,30,52,55], we approximated the data with several estimates by means of multiple assumptions: (1) consistency over time of both standard uptake volume ratios (SUVRs) and distribution volume ratios (DVRs), since JHU PET scans spanned 0 to 90 min after radiotracer injections, while IND PET scans spanned 90-120 min; (2) approximately equivalent anatomical brain regions, as MRI-based segmentation was utilized for VOI analysis at JHU, but an atlas-based approach was applied to the IND data; and (3) approximately equivalent analyses, although the resolution of the scans from the IND was approximately twice the resolution of scans from JHU. Using the measurements of SUVR from the IND dataset, we derived estimates of binding potentials as DVR-1 [78], which were pooled with the comparable BP ND estimates from the JHU data.

IND
The density of mGluR5s was comparable in the men with FXS and the men with TD ( Figure 2)  VOIs were obtained automatically of cortical regions with Freesurfer 6.0 [69,70] and of subcortical regions with the subcortical segmentation tools of the software library of the Oxford Centre for fMRI of the Brain [71][72][73]. VOIs were transferred from MRI to PET space according to MRIto-PET co-registration parameters obtained with the co-registration module [74,75] of statistical parametric mapping (SPM) [33] and applied to PET frames to obtain regional TACs. With the cerebellar white matter as the reference VOI [28,29], regional BPNDs [32] were obtained by reference tissue graphical analysis (RTGA) [76,77].

Comparisons and Contrasts of Cohorts from the IND and JHU
In order to directly compare and contrast data from both cohorts including nine men with FXS (mean age 27.21 ± 4.17, range 22.3-33.6) and eight historical control age-matched men with TD who had already completed similar protocols (mean age 30.63 ± 5.58, range 24-39 years) [25,30,52,55], we approximated the data with several estimates by means of multiple assumptions: (1) consistency over time of both standard uptake volume ratios (SUVRs) and distribution volume ratios (DVRs), since JHU PET scans spanned 0 to 90 min after radiotracer injections, while IND PET scans spanned 90-120 min; (2) approximately equivalent anatomical brain regions, as MRI-based segmentation was utilized for VOI analysis at JHU, but an atlas-based approach was applied to the IND data; and (3) approximately equivalent analyses, although the resolution of the scans from the IND was approximately twice the resolution of scans from JHU. Using the measurements of SUVR from the IND dataset, we derived estimates of binding potentials as DVR-1 [78], which were pooled with the comparable BPND estimates from the JHU data.

IND
The density of mGluR5s was comparable in the men with FXS and the men with TD ( Figure 2) [55].

JHU
Participant JHUFXS1 withdrew before scans due to a family emergency. Participant JHUFXS2 completed both MRI and PET scans in one day without mock scanner training. Due to scheduling problems, MRI and PET scans were conducted on Participant JHUFXS3 on separate days a week apart without mock scanner training. Despite the administration of 2.0 mg of lorazepam before each scan, he could not complete either scan due to anxiety and agitation.
Participant JHUFXS4 had already completed MRI scans of 30 and 60 min on separate days at another institution. For this prior investigation, a psychologist met with him online regularly for weeks before the scans to practice holding still despite the noise. He had never had a PET scan. His mother began practicing relaxation and holding still while listening to MRI sounds for weeks before the session at JHU. His mother and an investigator accompanied him into the MRI chamber throughout the MRI scan. His mother sat at the operator's booth throughout the PET scan to praise him for holding still during the PET scan.
The non-displaceable binding potentials (BP ND s) [32] of [ 18 F]FPEB by reference tissue graphical analysis (RTGA) [76,77] in each VOI of two men with FXS were below the BP ND s of five age-matched men with TD [25,30,55] (Figure 3). mGluR 5 expression was lower in the men with FXS than the men with TD ( Figure 3). Due to scheduling problems, MRI and PET scans were conducted on Participant JHUFXS3 on separate days a week apart without mock scanner training. Despite the administration of 2.0 mg of lorazepam before each scan, he could not complete either scan due to anxiety and agitation.
Participant JHUFXS4 had already completed MRI scans of 30 and 60 min on separate days at another institution. For this prior investigation, a psychologist met with him online regularly for weeks before the scans to practice holding still despite the noise. He had never had a PET scan. His mother began practicing relaxation and holding still while listening to MRI sounds for weeks before the session at JHU. His mother and an investigator accompanied him into the MRI chamber throughout the MRI scan. His mother sat at the operator's booth throughout the PET scan to praise him for holding still during the PET scan.
The non-displaceable binding potentials (BPNDs) [32] of [ 18 F]FPEB by reference tissue graphical analysis (RTGA) [76,77] in each VOI of two men with FXS were below the BPNDs of five age-matched men with TD [25,30,55] (Figure 3). mGluR5 expression was lower in the men with FXS than the men with TD ( Figure 3).

IND and JHU
Combined (IND and JHU) estimates of mGluR 5 were significantly reduced in all eight volumes of interest (anterior cingulate, caudate, occipital, parietal, posterior cingulate, putamen, temporal, and thalamus) in the men with FXS (N = 9) in contrast to the age-matched men with TD ( Figure 4, Table 2).
Although the axial resolution of the IND scans was approximately twice that of the JHU scans, the combined results are striking.
Brain Sci. 2020, 10, x FOR PEER REVIEW 8 of 18 Table 2). Although the axial resolution of the IND scans was approximately twice that of the JHU scans, the combined results are striking.    Furthermore, a two-way analysis of variance on an initial pooled dataset showed that both independent variables of institution (IND and JHU, df = 1, F = 34.3, p < 0.0001) as well as diagnosis (FXS and TD, df = 1, F = 38.7, p < 0.0001) had non-random effects on regional estimates of BP [80] ( Table 2).

Discussion
We showed the feasibility and safety of administering MRI and PET in two independent pilot samples of men with FXS. We applied PET to quantitatively measure the density of mGluR 5 s in cortical and subcortical brain regions of these men with FXS following exposure to [ 18 F]FPEB), which is a first study to our knowledge. We found that mGluR 5 density was significantly reduced in the cingulate, cortex, striatum, and thalamus in men with FXS in contrast to age-matched men with TD. The tracer is a novel, specific mGluR 5 ligand to measure the density and distribution of mGluR 5 s in the brains of humans, which constitutes an effective tool to confirm target engagement of NAMs for mGluR 5 s. The feasibility of this complex protocol requires a multidisciplinary effort that includes mock scanner training and practice sessions taught with behavioral psychology.

Feasibility of a Complex Protocol of MRI and PET Scans in Participants with FXS Adults
A primary goal of this study was to determine the feasibility and safety of a complex protocol that included MRI and PET scans on men with FXS. We showed that this challenging protocol can be accomplished with mock scanner training and practice sessions taught with behavioral psychology [62][63][64] and trained parents. Additionally, an investigator and a parent routinely accompanied participants into the MRI chamber to assist with the process during the entire MRI series. Since state-of-the-art PET scanners provide three-dimensional image reconstruction, face masks may no longer be required to stabilize heads. Scans may be accomplished with gauze for optimal comfort.
We recommend several modifications to facilitate the completion of the MRI and PET scans on individuals with FXS. Mock scanner training beginning online for weeks before the actual scans provides the means to train participants and parents to relax quietly without moving while loud noises like a jackhammer are played [62][63][64]. Behavioral psychologists can meet with participants and parents repeatedly online to utilize training sessions for holding still while MRI soundtracks are played through recordings. The sessions can begin with short practices of 15 s. Gradually, the duration of the session can be increased to 30 or 60 min to train participants to calmly endure the challenges of the noise and stillness. Additionally, behavioral psychologists can provide the example of providing positive feedback to the participants. In other words, praising the participant for holding still during the practice session is a valuable positive reinforcement for desired behavior. On the other hand, criticizing the participant for moving may increase anxiety and lead to agitation and uncooperative behavior. Therefore, parents can be taught to reward the desired behavior.
Another approach to facilitate successful completion of scans includes the shortening of the duration of PET scans and the use of gauze instead of a rigid face mask. Additionally, performing PET and MRI scans on two separate days allows participants to recover from the stress of one scan before undergoing the next. The use of PET/MRI machines would simplify the protocol to accomplish both PET and MRI in a single session [81].

Adolescents and Children
Since PET involves greater than minimal risk due to radiation exposure, the safety and efficacy must be shown in adults before exposing vulnerable populations. For this reason, the current protocol was administered only to adults with FXS. After safety and efficacy are established in adults, then the procedure will be sequentially administered to adolescents, followed by children. The procedure may be modified for children to reduce the duration of scans. The procedure of the IND to conduct a 30 min scan 90 to 120 min after radiotracer injection with gauze to stabilize the head will shorten the stress of remaining on the scanner table. Another modification will be the utilization of PET/MRI scanners to conduct both PET and MRI scans in a single session instead of separate sessions for PET and MRI scans [81]. Mock scanner training by experienced behavioral psychologists [62][63][64] will be crucial to prepare children and adolescents for scans. Additionally, the participation of parents for each step is key to the accomplishment of this challenging protocol.

mGluR 5 Measurement in Men with FXS
Another goal of this investigation was to find out if the study protocol can quantify mGluR 5 expression in the brains of adult males with FXS. The data from our study show that the PET ligand binds mGluR 5 s in the brains of men with FXS, and that the expression of these receptors is decreased. This finding could be mediated by excessive upstream signaling resulting in reduced expression of mGluR 5 s. Internalization of the mGluR 5 s [20,82] throughout the brain induced by the radiotracer, the scanner, or other aspects of the environment of PET scans may explain the reductions in mGluR 5 s in our participants with FXS.
A preliminary attempt to perform an analysis on a combined dataset of both FXS/TD data from the IND and JHU showed that the source of the data was a non-random factor that influenced the outcome variable. We shall strive to reduce this possible confounding influence to improve the effect size in future analyses. As a future direction, we are developing other means of analyzing larger datasets from multiple institutions in a comparable manner so that the data can be pooled after removing the confounding factors of methodological differences.

Avoiding Effects of Diurnal Variations of mGluR 5 s
We administered PET scans to participants with FXS at the same time of day (1 PM) to minimize effects of diurnal variations of mGluR 5 s. Participants with TD received radiotracer injections 32 ± 120 min (range −135 to +163) from 1 PM [55], resulting in a confounding influence of diurnal variation. Large alterations in radiotracer uptake on the same individuals during the same day suggest that there may be considerable diurnal variation in mGluR 5 s, with increased uptake later in the day [21,[83][84][85]. Participants with FXS may experience greater anxiety with scans than participants with TD. Anxiety may increase cortisol values and result in diurnal variations. Thus, we assume that our participants with FXS likely exhibited the maximal radiotracer uptake at the time of their scans. Measurement of cortisol levels and administration of PET scans at the same time of day to all participants minimizes the effects diurnal variations of mGluR 5 s.

Limitations and Future Studies
There is a need for comprehensive protocols uniformly administered to all cohorts. The use of different protocols for PET at the collaborating institutions [52,86] confounds comparisons and contrasts of the results. Future investigations at multiple centers will benefit from the use of identical protocols and analyses for PET and MRI conducted contemporaneously. Analysis of results by a single center will facilitate the uniformity of the findings. Despite different protocols, the uniformity of the finding of reduced mGluR 5 expression in multiple brain regions independent of protocol strengthens this study's key finding.
Administration of the full neuropsychological battery to contemporaneous cohorts at all participating centers will provide the foundation to apply statistical analyses. Normalization of cognitive test scores for participants with FXS will remove a "floor effect" [58]. Future studies will benefit from examining participants with FXS exhibiting a spectrum of ID and ASD and comparison groups without FXS with levels of ID matched to the participants with FXS.
The current pilot study is limited by the incomplete FMR1 gene and epigenetic (methylation) parameter identification, and incomplete size mosaicism and quantification of FMRP. Future studies will be enhanced by including these measures and whole exome sequencing (WES) [87] on all participants to test the hypothesis that the parameters are correlated [12].
Since increased protein synthesis has been demonstrated in fibroblasts of individuals with FXS and fmr1 KO mice [88], measurement of protein synthesis, particularly in the mTOR and ERK signaling cascades, would be a valuable parameter to correlate with mGluR 5 density and distribution in future investigations of FXS in humans. However, the absence of increased protein synthesis in young men with FXS sedated with dexmedetomidine for PET with L[1-11 C]leucine suggests that humans with FXS may not demonstrate the increased protein synthesis seen in animal models [89].

Multimodal Imaging
Multimodal imaging can enhance future investigations by linking PET, electroencephalography [90], event-related brain potential (ERP) [91], resting state functional magnetic resonance imaging (rs-fMRI), diffusion tensor imaging (DTI), and movement measurement [92], along with quantitative measurements of FMRP and fmr1 [3]. Newly developed PET/MRI scanners [81] may produce visualization of the density and distribution of mGluR 5 s that is superior to images obtained from HRRT co-registered with MRI. PET/MRI units are appealing for future investigations because a single session would be required. PET/MRI provides both functional (PET) and structural (MRI) findings in one encounter. Thus, PET/MRI instrumentation and many other multimodal techniques may be utilized when available for subsequent investigations.

Conclusions
We showed the feasibility and safety of applying PET as a tool to quantify mGluR 5 receptor expression in the brains of humans with FXS.
We showed that the proposed protocol of MR and PET scans in one day is feasible in individuals with FXS who have received mock scanner training by an experienced behavioral psychology team.
Most importantly, we found for the first time that quantified mGluR 5 expression using [ 18 F]FPEB is reduced in the living human brain of men with FXS in contrast to healthy normal age-and sex-matched controls with TD.
Larger studies with additional molecular biomarkers [93] are needed to expand on the feasibility finding of this protocol to evaluate the receptor expression of mGluR 5 s using [ 18

Acknowledgments:
The authors thank the patients and families for their participation and dedication to these studies; they are the inspiration for our efforts at improving treatments. The authors thank the FORWARD Database and Registry of the National Fragile X Foundation (NFXF) funded by the Centers for Disease Control and Prevention (CDC), Atlanta, Georgia, for referral of participants. The authors thank the teams of the Institute of Neurodegenerative Disorders, the Positron Emission Tomography (PET) Radiotracer Service Center, and the Research Magnetic Resonance Imaging (MRI) Service Center of the Johns Hopkins University School of Medicine for conducting the scans. The authors thank Hiroto Kuwabara for PET analysis. The authors acknowledge Rohan Panaparambil and Mathew Shneyderman for their guidance to revise the paper. The authors thank Brian Hwang for contributing to the graphical abstract. Earlier versions of this article were presented at the 2020 Annual Meeting, Society of Nuclear Medicine and Molecular Imaging, 11-14 July 2020 [86], and the World Molecular Imaging Congress 2020.

Conflicts of Interest:
The authors declare no financial conflicts of interest. The funders had no role in the design of the study, in the collection, analyses, or interpretation of the data, in the writing of the manuscript, or in the decision to publish the results.