Age distribution was comparable among individuals in the Clinical Cohort. The majority of the participants were Caucasian males (% males/% Caucasian): 78/76 for the entire Clinical Cohort, 84/74 for the FM-ALL Subcohort, and 84/77 for the FM-FMRP Subcohort. As shown in Table 1, the small subgroup of individuals with PM consisted of two males and three females.
Materials
FMR1 reference cell lines
The Reference Cohort consisted of three individuals in the normal FMR1 CGG range, five with PM, and three with FM, each of whom contributed a lymphoblastoid cell line. These lines were established at RUMC using standard techniques (i.e., peripheral blood mononuclear cell Epstein-Barr virus transformation). Transformed cell lines were expanded, cryopreserved, and provided as frozen stocks to the Asuragen team (Austin, TX) for further expansion. Cell lines were grown in suspension at 37°C in RPMI 1640 (Gibco, Thermo Fisher Scientific, Waltham, MA) supplemented with 15% fetal bovine serum, 2 mM L-glutamine, and penicillin and streptomycin. Viability and total cell counts were determined using a standard trypan blue stain procedure and a phase hemacytometer (VWR, Radnor, PA); bacterial contamination was assessed using a Universal Mycoplasma Detection Kit (ATCC, Manassas, VA). After being grown to confluence (~8 x 105 cells/mL), cells were harvested for DNA, RNA, and total protein isolation.
Whole blood collection and preservation
Whole blood was collected using: (i) EDTA vacutainer tubes, as the whole liquid blood for initial 20 subjects, (ii) FTA cards, preserved blood saturated spot cards for all (N = 42) subjects (GE Healthcare Life Sciences, Marlborough, MA), and/or (iii) 903 Specimen Collection Paper/Whatman protein saver cards for total 36/42 subjects (GE Healthcare Life Sciences). Both the whole liquid blood and FTA cards were analysed for the first 20/42 subjects. The approach demonstrated that FMR1 CGG repeat quantification and DNA methylation analysis were not affected by the method of the blood collection or storage; thus, the FTA cards were selected as a preferred sample type for the remainder 22/42 subjects of the study. FMRP analysis, however, was not supported by specimens preserved on FTA cards, which prompted use of the 903 protein saver cards as an additional specimen collection. Cards were spotted with ~125 µL of whole blood (EDTA tubes) and stored at room temperature. Specimens from each subject were shipped to Asuragen for subsequent molecular and protein testing (FMR1 and FMRP analysis).
Matched blood and buccal specimens
Matched whole blood and buccal cell specimens were collected from 42 patients from the Clinical Cohort. Buccal cell samples were collected using ORAcollect·DNA (DNA Genotek, OCR-100) following the manufacturer’s instructions, while blood samples were obtained using standard clinical procedures, both following informed consent and according to protocols approved by the Johns Hopkins Medicine’s IRB.
Genomic DNA and protein isolation and primary characterization
Genomic DNA was isolated from the whole blood specimens (Clinical Cohort, N = 42) and cell lines described above using the DNeasy Blood and Tissue Kit (Qiagen, Germantown, MD) according to the manufacturer’s instructions. The concentration of gDNA was determined using spectrophotometry (NanoDrop, Thermo Fisher Scientific). The level of intact or non-fragmented DNA was assessed by agarose gel electrophoresis (AGE, E-Gels Precast Agarose Gels; Thermo Fisher Scientific, Waltham, MA).
To isolate multiple analytes (DNA, RNA) simultaneously from the cell lines, we used the AllPrep DNA/RNA Mini Kit (Qiagen, Germantown, MD) following the manufacturer’s instructions. Nucleic acid concentration was determined by spectrophotometry (NanoDrop, Thermo Fisher Scientific). Additionally, RNA integrity was determined on the 2100 Bioanalyzer (Agilent, Santa Clara, CA) using standard procedures. The minimum RNA Integrity Number (RIN) for all samples was 9.6.
Total protein was isolated using Cell Extraction Buffer (CEB, Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. Briefly, cell pellets were washed by re-suspension in ice-cold phosphate-buffered saline (PBS) followed by centrifugation at 300 x g for 7 minutes at 4°C. After aspirating the PBS, cells were lysed using 1 mL CEB/108 cells and incubated on ice for 30 minutes with occasional vortexing followed by centrifugation at 300 x g for 7 minutes at 4°C to collect cell debris. Total protein concentration of the lysates was determined using a Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA).
Molecular assessments
CGG repeat genotyping and AGG interruption analysis
FMR1 genotypes from cell-line and blood sample gDNA were determined using triplet repeat-primed PCR, followed by fragment analysis of amplicons by capillary electrophoresis (CE) on an ABI 3500xl Genetic Analyzer (Thermo Fisher Scientific, Waltham, MA), and Sanger sequencing. PCR-based CGG repeat genotyping was performed with both two-primer and three-primer FMR1 PCR/CE [FMR1 gene-specific PCR/CE (GS-PCR/CE) and repeat-primed PCR/CE (RP-PCR/CE), AmplideX® PCR/CE FMR1 Kit (Asuragen, Austin, TX)] [20, 21]. Cell-line gDNA samples were also analysed using two-primer FMR1 PCR with products resolved by AGE (GS-PCR/AGE) [21]; this permitted sizing of repeat expansions larger than ~200 CGG that cannot be sized within the resolution limits of CE. Samples were annotated as size mosaics if they manifested peaks in two different size categories such as PM and FM with a peak signal intensity greater than 75 relative fluorescence units (RFU) by CE. The number and sequence context of interrupting AGG sequences in the repeat tract was determined using Xpansion Interpreter®, a PCR-based method (Asuragen, Austin, TX) [40, 41, 45, 48].
Methylation PCR analysis
The AmplideX mPCR FMR1 Kit (Asuragen, Austin, TX) was used to quantify relative allele-specific DNA methylation of the FMR1 gene [39, 49] on gDNA from blood and buccal samples or cell lines. Both X-inactivation and methylation of the expanded allele could be assessed using this method, revealing completely or partially methylated states [44]. A sample was defined as a methylation mosaic (MM) if it contained FM or PM fragment peak(s) above 75 RFU by CE that was/were <80% methylated in at least one sample type (buccal swab or whole blood).
Southern blot analysis
Southern blot analysis was performed utilizing gDNA isolated either from the whole blood samples or from the cell lines. DNA was digested with EcoRI and NruI and separated on an agarose gel. After DNA transfer, the membranes were hybridized with the FMR1-specific StB12.3 genomic probe, and imaged according to published procedures [50].
FMR1 mRNA analysis
FMR1 transcript expression was evaluated via quantitative real-time PCR (qPCR) for all cell lines samples. A total of 2.5 ug of purified total mRNA was converted to cDNA using a reverse transcription (RT) protocol. The RT reaction was incubated at 42°C for 45 minutes followed by 10 minutes at 93°C, and then rapidly cooled to 4°C. FMR1 mRNA specific primers (Integrated DNA Technologies, Coralville, IA) [Forward Primer: 5’- TATGCAGCATGTGATGCAACT-3’, Reverse Primer: 5’- TTGTGGCAGGTTTGTTGGGAT-3’] for use with the/a KAPA SYBR Fast qPCR kit (KAPA Biosystems, Wilmington, MA) were applied according to the manufacturer’s instructions. Each sample was run at two input concentrations of 31 ng/ul and 6.25 ng/ul, respectively. FMR1 mRNA expression was quantified relative to expression of the control housekeeping gene USP33 [42].
Quantitative FMRP analysis
FMRP levels were determined using a quantitative FMRP (qFMRP) assay developed by Lafauci et al [43], per the method described by Gustin et al. [42]. This assay utilizes a recombinant FMRP peptide (GST-SR7) as a standard for quantification of FMRP level. Protein (on 903 paper) and gDNA were co-extracted from 3 mm dried blood spot punches. DNA was quantified using the PicoGreen Quant-iT™ HS kit (Thermo Fisher Scientific, Waltham, MA). The amount of FMRP was normalized to the amount of gDNA, as a proxy for cell count, to reduce the variability among samples.
Clinical assessments
Table 3 summarizes the neurobehavioral assessments carried out on the FM-FMRP cohort (N = 31). They included: (i) presence (diagnosis) of ASD, social anxiety (SA), and/or unspecified anxiety, (ii) level of intellectual functioning/intellectual disability (ID), as determined by Full Scale Intelligence Quotient (FSIQ), (iii) level of adaptive functioning, (iv) severity of problem behaviors, (v) overall clinical severity, as determined by the Clinical Global Impression-Severity scale (CGI-S), and (vi) use of antipsychotics.
The following assessment tools were employed:
(i) Diagnostic and Statistical Manual-5th Edition criteria (DSM-5) [51], supplemented by available Autism Diagnostic Observation Schedule (ADOS) assessments, were used to diagnose ASD. Diagnosis of ASD was performed longitudinally for the Clinical Cohort by a clinician (DBB) with experience in idiopathic ASD and ASD in FXS [26]. Diagnoses of SA and unspecified anxiety were also performed using DSM-5 criteria [27, 32]. SA include a substantial social inhibition (shyness) accompanied by a broad range of fear of negative evaluation by others, which may be embarrassing, lead to rejection or offend others such as the expression of anger toward others. ‘Fragile-X handshake” and various forms of “escape” behaviors in familiar or especially unfamiliar situations are common as well. [See section (v) for profiling of severity/level of SA diagnosis].
(ii) FSIQ was determined by the Stanford-Binet Intelligence Scales-5th Edition (SB-5) for 15/31 individuals, the Wechsler Preschool and Primary Scale of Intelligence (WPPSI) for 1/31, the Wechsler Intelligence Scale for Children (Fourth and Fifth Edition, WISC-IV and WISC-V) for 7/31, The Differential Ability Scales (DAS) for 4/31 and the Mullen Scales of Early Learning for 4/31 [38]. To address the skewed effect of FSIQ-standard or other score testing, raw-score based z-score calculations from the IQ subtests were used by a senior neuropsychologist (EMM) to generate extended FSIQ values [53]. When there were several administrations of a test, the most current one was used for data analysis. Alternatively, if the scores were highly discrepant, estimation was made via interpolation of the two scores.
To determine level of ID, the extended FSIQ scores were also used instead of adaptive skill scores because the former, as scaled measures, better reflect the range of cognitive abilities. Subjects were assigned to one of four ID levels: normal range (FSIQ score ≥70), mild ID (FSIQ: 55-69), moderate (FSIQ: 35-54) and severe ID (FSIQ <35).
(iii) Adaptive functioning was assessed by using adaptive skill scales, which included the Vineland Adaptive Behavior Scales-Second Edition (VABS–II) for most participants, the Adaptive Behavior Assessment System Second and Third Editions (ABAS-2 and 3) for 10 individuals, and Scales of Independent Behavior-Revised (SIB-R) for 2 participants.
(iv) Problem behaviors were assessed by the Aberrant Behavior Checklist-Community Edition (ABC-C) adapted for FXS (ABC-CFX), which applies a subscale scoring algorithm developed specifically for the disorder and yields six subscales [52]: (i) Irritability, (ii) Lethargy/Social Withdrawal, (iii) Stereotypic Behavior, (iv) Hyperactivity, (v) Inappropriate Speech, and (vi) Social Avoidance [52, 54]. The ABC-CFX has been applied as a primary outcome measure in multiple observational and interventional studies in FXS (reviewed in [38]).
(v) The CGI-S score evaluates the overall impairment of a patient, using as reference the clinician's past experience with patients who have the same diagnosis. Possible ratings of the CGI-S are: 1-normal, not at all ill, 2-borderline ill, 3-mildly ill, 4-moderately ill, 5-markedly ill, 6-severely ill, and 7- extremely ill. The CGI approach was also separately applied to profile the severity/level of SA (CGI-SANX); based on CGI-SANX scores, two categories were defined: ≥5 (severe) and ≤4 (mild-moderate).
(vi) A patient’s use of antipsychotics was determined through health records and marked as ‘there is’ or ‘there is not’ (yes/no) use of this class of drugs.
Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics version 25 (IBM Corporation, Armonk, NY) or JMP version 14 (SAS Institute, Cary, NC). Descriptive statistics included frequency (percent) of nominal variables, and median, mean, standard deviation (SD), and range for continuous variables. Tests of normality and homogeneity of variances were also performed. Depending on data distribution, either parametric or non-parametric tests were applied. The Chi square test was used to test differences between nominal variables (frequencies). Pearson’s correlation coefficient was used as a measurement of the strength of the linear relationship between normally distributed variables. Welch’s t-test for unequal variances and equivalent non-parametric tests (Mann-Whitney) were performed to compare means of two samples. Welch's t-test maintains type I error rates close to nominal for unequal variances and for unequal sample sizes under normality. Significance was indicated by p £ 0.05 and high significance by p £ 0.01.