Carriers of the premutation (PM) are characterized by a moderate (55 to 200) expansion of the cytosine-guanine-guanine (CGG) nucleotide repeats in the first exon and promoter of the X-linked FMR1 gene (Verkerk et al., 1991; Bagni et al., 2012). Originally, PM carriers were assumed to be free of any apparent phenotypic traits. However, over the last decade, a growing number of neuropsychiatric manifestations (including depression, anxiety, and insomnia), visuospatial deficits, and immune dysregulation have been reported to occur at a greater frequency among adult PM carriers than in the general population (Hagerman and Hagerman, 2013). Generally, PM carriers show lower performance in neuropsychological testing including full-scale intellectual quotient (FSIQ) and working memory (WM) subtests on the Wechsler Adult Intelligence Scale (Lozano et al., 2016). In the case of children with the PM, they are often diagnosed with ADHD, autism, anxiety, and other psychopathologies (Napoli et al., 2018). The PM has also been associated with conditions beyond those involving the CNS, such as hypertension, hypothyroidism, high blood glucose, as well as a higher incidence of thyroid, prostate and other cancers (Lozano et al., 2016). With age, both female and male carriers of the PM are at a higher risk for developing the late-onset (usually appearing after age 50) neurodegenerative disorder fragile X-associated tremor/ataxia syndrome [FXTAS; OMIM:300623; (Hagerman et al., 2001, 2004; Hagerman and Hagerman, 2013)]. FXTAS-affected carriers may exhibit intention tremor and gait ataxia, accompanied by cerebral atrophy, white matter disease, parkinsonism, neuropathy, autonomic dysfunction, and cognitive deficits (Berry-Kravis et al., 2007a; Bourgeois et al., 2007; Lozano et al., 2016; Claeys et al., 2020).

As it is the case for many X-linked disorders, women have a lower absolute risk of developing FXTAS symptoms compared to men (Hagerman and Hagerman, 2016). However, women carrying FMR1 PM allele have a higher risk of developing premature or primary ovarian insufficiency (POI) [16% (Schwartz et al., 1994; Allingham-Hawkins et al., 1999; Rifé et al., 2004; Terracciano et al., 2004; Schuettler et al., 2011; Elizur et al., 2014; Pouresmaeili and Fazeli, 2014)] as compared to full mutation females (>200 CGG repeats), who carry the same risk for POI as the general population (1%). About 20 to 30% of female carriers experiencing irregular periods or amenorrhea due to ovarian insufficiency prior to age 40 are diagnosed with fragile X-associated primary ovarian insufficiency [FXPOI; (Allingham-Hawkins et al., 1999; Allen et al., 2007)]. Even PM carriers without signs of ovarian dysfunction have an earlier (on average by 5 years) age at menopause compared with non-carriers (Patsalis et al., 1999; Sullivan et al., 2005; Besterman et al., 2014). Women with alleles between 35–44 CGG repeats seem to present diminished ovarian function but regular menses and occult primary ovarian insufficiency (Streuli et al., 2009; Karimov et al., 2011; Pastore et al., 2012); however, other studies found no association between FMR1 intermediate alleles and POI (Bennett et al., 2010; Murray et al., 2014; Voorhuis et al., 2014). Other medical and psychological issues reported in females are hypothyroidism, hypertension, endocrine dysfunctions, chronic pain, fibromyalgia, autoimmune diseases, neuropathies, migraines, dementia, and psychiatric conditions, such as anxiety and depression (Allen et al., 2007, 2020; Bailey et al., 2008; Hunter et al., 2010; Winarni et al., 2012; Wheeler et al., 2014a,b; Lozano et al., 2016; Movaghar et al., 2019). Collectively included under the term FXAND [fragile X–associated neuropsychiatric disorders; (Hagerman et al., 2018)], such emotional and neuropsychiatric disorders, have been shown to be more common in female carriers compared to non-carriers. In a recent work by Dr. S. Sherman’s group, which investigated the association between the PM diagnosis and CGG repeat expansion in female carriers, the most common symptoms reported were anxiety and depression, migraine, headaches, and sleep problems (Allen et al., 2020).

Our team was the first to report mitochondrial dysfunction as a common feature in biological samples from PM carriers (Napoli et al., 2013, 2016a,b, 2018; Song et al., 2016) as well as in murine models of the PM (Napoli et al., 2016a). This decreased mitochondrial bioenergetics is present in PM carriers with and without FXTAS and even in some pediatric carriers (Napoli et al., 2018). However, to our knowledge, no study has to date characterized the metabolic footprint of the PM and related clinical and cognitive features in female carriers of the PM.

To bridge this gap in knowledge, we performed a comprehensive biochemical assessment (including metabolomics and proteomics profiling, and bioenergetics) in peripheral blood mononuclear cells (PBMC) and plasma samples obtained from 24- to 70-year-old female carriers. To elucidate peripheral bioenergetics markers that may function as surrogates for CNS function, we utilized a multivariable correlation matrix to identify correlations between mitochondrial outcomes and cognitive parameters (FSIQ), executive function (BDS-2), anxiety, tremor, and FXTAS (and its severity) and FXPOI diagnoses. As such, this study is ideally positioned to perform comprehensive deep metabolic and mitochondrial phenotyping by gathering complementary outcomes on genomics, proteomics, metabolomics, mitochondrial physiology and clinical information to systematically develop a metabolic profile from established female carriers by taking advantage of a substantial repository of patient samples. This multi-faceted approach, as opposed to a simple model based on the statistical differences of few (and sometimes unconnected) metabolites or proteins between diagnostic groups. The profile generated will, for the first time, allow researchers to fully assess the larger biological impact of the premutation on metabolic status of female carriers, thereby providing unprecedented insight into the biological consequences of metabolic deficits and mitochondrial dysfunction, aiding the discovery of disease mechanisms. Importantly, due to the depth of the phenotyping across multiple readouts, the integrated metabolic profiles generated will have the detail required to cluster patients according to their clinical pathology when more patients’ data will be available. This necessary step, in turn, will lead to truly novel and testable hypotheses regarding individualized pathogenesis and treatment.

Mounting evidence has shown that female carriers of the PM are at higher risk for developing several health-related issues compared to non-carrier females (Bailey et al., 2008; Hunter et al., 2010; Winarni et al., 2012; Wheeler et al., 2014b; Movaghar et al., 2019). Awareness of these risks and correlation of the clinical signs with the biochemical footprint of carriers could help to identify critical biomarkers in early diagnosis and likely prognosis. If ascertained, these associations could lead to an integrated approach between clinical specialties and basic science, which could be extremely beneficial in the management of symptoms and challenges that female PM carriers experience in their day-to-day life.

The brain’s high mitochondrial energy consumption makes neurons highly vulnerable to impaired glucose metabolism (Hyder et al., 2013). Then, it is not surprising that a decline in mitochondrial activity has been associated with memory loss and, particularly, with age-dependent cognitive impairment (Freeman and Young, 2000; Liu et al., 2002). Moreover, most mtDNA diseases are associated with brain disorders because adequate neuronal development (Atamna et al., 2002; Zini et al., 2002) and structure (Johnson and Byerly, 1993; Bristow et al., 2002; Liu et al., 2002) and axonal and synaptic activity (Freeman and Young, 2000; Zenisek and Matthews, 2000) all involve mitochondrial genes (Wallace, 1999; Schon, 2000). Mitochondrial dysfunction has been reported not only in neurodegeneration (Friedland et al., 1983; Minoshima et al., 1997; Chiaravalloti et al., 2015) but also in pre-symptomatic, genetically-susceptible individuals (Small et al., 1995; Reiman et al., 1996; Mosconi et al., 2006; Ossenkoppele et al., 2013). In line with these studies and our previous reports on the PM (Napoli et al., 2016a, 2018), our study strongly indicates that global impairment of bioenergetics, and likely the subsequent energy depletion, is one of the earliest functional changes prior to the onset of overt clinical symptoms. This is supported by the relatively milder mitochondrial dysfunction in PM carriers without FXTAS and FXPOI which is enhanced with the diagnosis and progression of FXTAS. While the decline in cognitive/intellectual (FSIQ) and executive (BDS-2) function was correlated with longer CGGs, intellectual decline (FSIQ) was significantly correlated with an overall mitochondrial deficit. This is highly suggestive that mitochondrial deficits may influence neuronal dysregulation and, over the years, degenerative mechanisms such as described for AD (Rice et al., 2014).

The metabolomics findings are particularly relevant in the context of several psychiatric and neurological symptoms experienced by some female carriers as oleamide has a critical role in mood and sleep disorders as well as depression due to its interaction with serotonergic and GABAergic neurotransmission (Mendelson and Basile, 2001). Similarly, tryptophan is not only an essential amino acid for protein synthesis but also a precursor of several biological mediators involved in stress response, antioxidant system, behavioral response, and immune system (Firk and Markus, 2009; Hoglund et al., 2017). Decreased tryptophan levels have been linked to lower serotonin levels, likely setting the basis for the establishment of neuropsychological symptoms including depression, anxiety, irritability, attention deficits, and insomnia (Jenkins et al., 2016). The higher cholesterol levels are of interest in the context of neurodegeneration, as mitochondrial cholesterol loading has recently emerged as a key player in the pathology of neurological disorders such as AD and Niemann-Pick Type C (NPC) disease (Elustondo et al., 2017; Torres et al., 2019).

In line with these findings, the lower fatty acid beta-oxidation resulting mainly from lower ABAD and ECHS1, deserves a separate discussion. Lower levels of ABAD may disrupt the ABAD-beta-amyloid interaction in mitochondria and suppresses apoptosis by increasing the levels of beta-amyloid as reported for AD patients and transgenic mouse models (Lustbader et al., 2004). Deficiency in ECHS1 results in metabolic acidosis with a combined respiratory chain deficiency (Sakai et al., 2015). Consistent with these findings, higher levels of lactic acid and a trend towards higher 2-hydroxyvaleric acid, proline and organic acids, such as adipic, azelaic and glutaric, were observed in plasma from female carriers (Figures 2C, 3) similar to the cases reported by others (Fitzsimons et al., 2018). Then, the significant lower expression of ABAD and ECHS1 in female carriers makes them as interesting candidates to investigate further in terms of its regulation in the context of beta-amyloid and proteostasis. Furthermore, since the OXPHOS decline in ECHS1 deficiency has been attributed to the accumulation of inhibitory fatty acid intermediates and to the disruption of ETC Complex biogenesis and/or stability (Burgin and McKenzie, 2020), it would not be unlikely that these mechanisms are contributing to the OXPHOS deficiencies observed in the PM.

Proteomics analysis not only confirmed and extended the functional and metabolomics results, but also shed light into the proteostasis status of the PM. The biological consequence of higher mitochondrial ROS was identified with the higher levels of aconitate likely the result of inactivation of aconitase, an enzyme highly sensitive to oxidative stress damage. The increase in oxidative stress may contribute to proteotoxicity and to the generation of misfolded proteins that accumulate upon ER stress, as a result, increase in the detoxification demands. However, an uncoordinated proteomic response in the ER was identified in the PM. Only one chaperone, HSPA5, was higher in PM vs. non-carriers, and a mixed response was observed for subunits of the proteasome. These results suggest that proteasomal degradation, a fundamental mechanism for degrading the misfolded proteins that accumulate in response to the metabolic challenge, is somehow not coordinated to adequately cope with the proteotoxicity. Indeed, it has been a long-standing hypothesis that protein aggregates in diseased brain impair the protein degradation function of the 26S proteasome (Taylor et al., 2002; Ciechanover and Brundin, 2003; Valera et al., 2005), then it would be reasonable to hypothesize that as the UPR is not fully coordinated in female carriers, any proteasomal response expected to ensue in response to a metabolic, environmental or pharmacological challenge might be adversely affected.

Our findings are particularly relevant for female carriers, as mitochondria have a critical role in oocyte developmental competence and function (Dumollard et al., 2007; Wang et al., 2009), and in the development and function of the reproductive system fertility (Reynier et al., 2001; El Shourbagy et al., 2006; Santos et al., 2006; Cagnone et al., 2016; Demain et al., 2017) with an irrefutable implication in primary ovarian insufficiency (POI) (Conca Dioguardi et al., 2016; Tiosano et al., 2019). The energy deficit along with the lower levels of ECHS1 in carriers, one of the genes identified as critical for oocyte developmental competence (Biase, 2017), set the basis for future research to identify more clearly the role of this protein in POI and FXPOI.

In conclusion, through a combined multi-omics approach in PBMC and plasma of PM females in association with thorough clinical and bioenergetics assessments, we found impaired metabolic pathways which can result from the direct action of toxic intermediates derived from the PM genetic background (either accumulation of mRNA or proteotoxicity of RAN-derived protein products). In this regard, we are adding the novel observation that the accumulation of aberrant metabolites resulting from deficiencies in critical metabolic steps may add to the altered interaction between fatty acid oxidation and the electron transport chain contributing to the overall OXPHOS decline as it has been described for ECHS1 deficiency (Burgin and McKenzie, 2020; Figure 3). In turn, either mechanism elicits a suboptimal activation of UPR and ERAD responses, setting a challenging scenario to withstand other, more severe forms of stress. Along these lines, the development of neurodegeneration or other clinical symptoms in older carriers could be linked to a lifetime accumulation of cellular damage, aggravated by the aging process.

RH has received funding from the Azrieli Foundation, Zynerba, Ovid and Neuren for treatment trials in fragile X syndrome, and consulted with Zynerba and Fulcrum regarding treatment for fragile X syndrome. FT has received funding from the Azrieli Foundation and Zynerba for studies on fragile X syndrome and has consulted with Zynerba. The other authors have no financial disclosures relevant to this article.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by IRB Ethics Committee at UC Davis Medical Center. The patients/participants provided their written informed consent to participate in this study.

EN processed all samples, carried out all polarographic and spectrophotometric measurements, performed correspondent statistical analyses, wrote the manuscript, and revised and approved the final version as submitted. YM collected and provided demographic, clinical, and molecular data, revised the manuscript, and approved the final version as submitted. AS carried out the neuropsychological testing and also revised the manuscript and approved the final as submitted. FT provided CGG repeats and XAR, revised the manuscript, and approved the final version as submitted. RH carried out clinical assessment of the women enrolled in this study and wrote clinical findings, revised the manuscript, and approved the final manuscript as submitted. CG conceptualized and designed the study, analyzed the omics data, wrote the manuscript, and approved the final manuscript as submitted.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.