Downregulation of Genes Involved in Metabolism and Oxidative Stress in the Peripheral Leukocytes of Huntington's Disease Patients

Background Huntington's disease (HD) is caused by expanded CAG repeats encoding a polyglutamine tract in the huntingtin (HTT) protein. A number of differentially-expressed protein molecules have been identified in striatum of HD animal models. Here we examined if the expression changes could be visualized in the peripheral leukocytes of HD patients and pre-symptomatic HD (PreHD) carriers. Methods and findings The expression levels of 17 candidate genes that differentially expressed in striatum between transgenic HD and wild-type mice in literature were measured in the peripheral leukocytes of 4 PreHD carriers, 16 HD patients and 20 healthy controls. Four genes majorly involved in metabolism and oxidative stress response, including AHCY1, ACO2, OXCT1 and CAP1, demonstrated consistent downregulation in peripheral leukocytes of both PreHD carriers and HD patients, while UCP2 was only down-regulated in HD patients. Conclusion These results provide potential peripheral biomarkers to indicate disease onset in preclinical stage, and to monitor the efficacy of early treatment. Further studies of a large series of preHD carriers and symptomatic HD patients will be warranted to verify the findings and examine if these markers correlate with clinical features.


Introduction
Huntington's disease (HD), characterized by choreiform movements, cognitive impairment and psychiatric symptoms, is an autosomal-dominant neurodegenerative disorder [1]. The major pathological presentation of HD is a regional degeneration of neurons in the striatum and cortex, which leads to movement disorders and dementia [2]. The genetic cause of HD is a long polyglutamine tract encoded by an expanded CAG trinucleotide repeats in the exon 1 of Huntingtin (HTT) [1]. The normal HTT gene has 35 or fewer repeats in the N-terminal region, whereas the appearance of neurological symptoms is associated with 36 or more CAG repeats in the HTT gene [1]. This polyglutamine expansion leads to a conformational change in the HTT and subsequently causes intracellular aggregates and neuronal degeneration [2,3].
While extensive therapeutic strategies were developed, none of them proved to be effective in halting the disease progression. One of the major drawbacks of the clinical trial in treatments is lack of a useful biomarker that can be used as the primary end-point to test the efficacy. Thus a sensitive and practical biomarker is an unmet need for the validation of new therapeutic strategies. Given that HTT is expressed ubiquitously, and a central nervous system sample from HD patients is practically difficult to access, a biomarker in peripheral tissue, especially from blood, should be more feasible as an indicator for the disease status. Although the pathology of HD is mainly in the striatum, a few studies have identified substantial biochemical deficits in peripheral tissues [4,5,6,7,8,9,10,11]. These peripheral changes include increased stress-induced lymphoblast apoptosis and vacuolation [9,10], upregulation of H2A histone family, member (H2AFY) in leukocytes [6], increased A(2A) receptor binding sites in blood platelets [8], pro-catabolic serum metabolite profiles [11], increased IL-6 and clusterin in plasma [5], reduced 24Shydroxycholesterol in plasma [7], and increased oxidative stress and mitochondrial dysfunction in peripheral blood [4]. Furthermore, parallel peripheral and neuronal dysfunctions have been shown [12,13,14,15,16]. Increased oxidative damage to DNA was found in the urine, plasma and striatal microdialysates of HD mice [13]. Lower Akt activation status [14], reduced cAMP and ATP/ ADP ratio [15], abnormal immune activation [12], and reduced creatinine kinase-BB have been demonstrated in both peripheral blood and brain of HD [16].
Transcriptional dysregulation is one of mandatory pathogenic mechanisms in HD [17,18]. Differential expression profiles in human peripheral tissues using oligonucleotide microarrays have been reported, whereas these two reports showed inconsistent results [19,20]. Recently, one study has also shown that the similar disruption pattern of RNA transcription in the brain could be visualized in peripheral blood [21]. Studies have shown a number of proteins that were up-or down-regulated specifically in the striatum of transgenic HD mouse models [22,23]. Since these results provide candidate molecules as potential biomarkers and the mRNA expression in peripheral blood may parallel the striatal changes in HD, we thus set out to examine if the expression of these molecules are altered in peripheral leukocytes of preHD and HD patients.

Results
To gain new insights into the pathological mechanisms of neuronal degeneration in HD patients and identify potential biomarkers, we chose 17 candidate genes that expressed differentially in striatum between transgenic HD and wild-type mice in literature (Table 1) [22,23]. The expression levels of these genes were measured on human leukocytes of HD patients. Due to its paucity of gene expression in human leukocytes [24,25,26], UCP1 was replaced by its homologue UCP2. Amongst 14 genes downregulated in the striatum of HD mice, 5 genes, including AHCY, ACO2, OXCT1, UCP2, CAP1, were in good agreement with the results in previous animal studies (Table 1). By contrast, all genes up-regulated in the striatum of HD mice displayed the similar expression preferences in HD patients.
Since the expression levels of AHCY, ACO2, OXCT1, UCP2 and CAP1 were lower in the striatum of HD mouse models and the human leukocytes when compared with their wild-type and the controls respectively, they serve as potential biomarkers to indicate the disease stage or progression not only for animal studies but also for clinical trials. To test this hypothesis, we compared the expression levels of peripheral leukocytes between HD patients, preHD carriers and controls. The results showed that the mRNA expression levels of AHCY (0.006460.00067 vs 0.007460.00076 vs 0.012160.00137, P = 0.003), ACO2 (0.2060.014 vs 0.2060.034 vs 0.5660.087, P,0.001), OXCT1 (0.004360.00037 vs 0.003660.00029 vs 0.005960.00054, P = 0.039) and CAP1 (0.09460.0095 vs 0.05960.0052 vs 0.13560.0109, P = 0.004) were significantly reduced in both HD patients and preHD carriers when compared with the controls by ANCOVA with adjustment of age and gender ( Table 2). UCP2 mRNA expression level of leukocyte was significantly reduced in HD patients (0.04760.0018, P = 0.011) compared with preHD carriers (0.06560.0070) and the controls (0.05860.0029). With the estimated standard deviations in each marker, at the level of 0.05, present sample sizes achieve a power of 100%, 76%, 88%, 93%, and 99.5% for AHCY, OXCT1, UCP, CAP, and ACO2, respectively, to detect differences in the mean of each marker between case and control groups. To understand the effect of medications and severity, we calculated the correlation between gene expression levels and age, medications, or UHDRS by covariate-adjusted generalized linear model. All factors did not present significant correlation with gene expression levels (data not shown).

Discussion
The application of a biomarker identified in animal models lies in the reproducibility in human samples. Although a number of animal studies have shown a long list of candidate biomarkers in HD mice or flies, the reappearance of these results in human system is uncertain. Here we validated that the expression changes of candidate biomarkers obtained from the striatum of HD mouse models are also present in human peripheral leukocytes of HD patients and preHD carriers. Of 17 genes quantified by RT-PCR,  The AHCY gene encodes S-adenosylhomocysteine hydrolase, which catalyzes the hydrolysis of S-adenosylhomocysteine (SAH) to adenosine and homocysteine [27]. It is known that SAH, strongly inhibits many S-adenosylmethionine (SAM)-dependent methyltransferases [27] that is crucial for the DNA, RNA and histone methylations. Therefore, S-adenosylhomocysteine hydrolase is important for maintaining the proper methylation potential in the cell. Moreover, SAH hydrolysis is the only source of homocysteine in mammals [28]. Thus reduced AHCY expression would obviously affect a wide variety of cellular processes. AHCY knockout mice have generally confirmed this by demonstrating early embryo lethality [29]. The patients with AHCY mutation showed white-matter atrophy and delayed myelination [27], further suggesting the downregulation of AHCY in HD patients may contribute to the CNS degeneration. Considering the downregulation of AHCY may result in the accumulation of its upstream substrates generated from methionine, restricted methionine intake could be a logical approach for the improvement of functional deterioration in HD patients and preHD carriers. Aconitase 2 encoded by ACO2 converts citrate to isocitrate in the tricarboxylic acid cycle and is involved in the ATP generation [30]. Decreased aconitase activity was also found previously in the striatum of R6/2 HD mice at late stage and postmortem brains of HD patients [30,31,32]. Aconitase 2 has been shown to be susceptible to increased oxidative stress that would lead to inactivation of aconitase 2 activity [30]. Since mitochondrial abnormalities and increased oxidative stress have long been suggested to play an important role in neurodegeneration of HD, the decreased peripheral leukocytes ACO2 expression level in preHD carriers could be an early peripheral event for HD.
OXCT1 encodes a mitochondrial matrix enzyme that functions as a homodimer to catalyze acetoacetate activation to acetoacetyl-CoA, playing a critical role in ketone body utilization [33]. Ketone bodies make an important contribution to brain energy production and biosynthetic processes [34]. In addition, several observations support the notion that ketone bodies exert neuroprotective effects [34]. The implication of OXCT1 in generation of ketone bodies, both in brain and peripheral tissues, suggests its important role in energy deficit in HD. Several lines of evidence have suggested metabolic deficits in HD (see review in Naia et al. [35]). Reduced ATP production has been shown in brain of presymptomatic and symptomatic HD mice [15,36]. In preHD carriers and symptomatic HD patients, there is strong evidence of hypometabolism in the brain, especially in the basal ganglia [37,38,39]. A systemic metabolic defect associated with early weight loss has been noted in HD patients as well [40]. Our study for the first time demonstrated down-regulated OXCT1 in preHD carriers and HD patients, suggesting an impairment of proper energetic supply from ketone bodies in HD. Supplement of ketogenic diet may be beneficial for the symptomatic improvement in HD patients.
The mitochondrial protein called uncoupling protein 2 encoded by UCP2 plays an important role in generating heat and burning calories by creating a pathway that allows dissipation of the proton electrochemical gradient across the inner mitochondrial mem-  brane [41]. UCP2 is widely expressed in adult human tissues, including brain and macrophages [41]. In addition to the generation of ATP, UCP2 has been identified as an inducible neuroprotective protein by decreasing the concentrations of reactive oxygen species inside mitochondria [42]. In the mouse overexpressing human UCP2, brain damage was diminished after experimental stroke and traumatic brain injury, and neurologic recovery was enhanced [43]. In cultured cortical neurons, UCP2 overexpression reduced cell death and inhibited caspase-3 activation induced by oxygen and glucose deprivation [43]. In this study, the downregulation of UCP2 is specific in HD patients, but not seen in PreHD patients, suggesting its potential role in association with the disease progression. Given that increased oxidative stress, impaired metabolism, mitochondrial dysfunction, and their interplay contribute to neuronal dysfunction in HD, the downregulation of UCP2 in the peripheral leukocytes may be used as a potential biomarker to monitor the disease deterioration and the treatment response.
CAP1 encodes a multifunctional protein with several structural domains involved in actin binding, adenylate cyclase association, and oligomerization [44]. Upon treatment with a number of agents that induce apoptosis, CAP1 rapidly translocates to the mitochondria [45]. This translocation is a proapoptotic event because CAP-knockdown cells are resistant to induction of apoptosis [45]. Therefore we proposed that the downregulation of CAP1 in peripheral leukocytes may have a protective effect through anti-apoptosis.
Systemic gene or protein expression analyses have been proven to provide useful candidate biomarkers, which may indicate disease status and are important for testing potential therapeutic strategies [46]. Although combining data sets obtained from independent research groups may identity biomarkers in a more reliable manner, a huge discrepancy existed amongst the studies of HD peripheral blood using microarray. For example, the global gene expression in lymphocytes from HD patients failed to show similar changes that were observed in freshly isolated blood samples [19,20]. Although some of the HD-related changes in the expression of 12 genes identified by Borovecki et al. were confirmed in one independent study [47], the levels of serum mRNA markers (ROCK1 and ANXA1) were unaffected by HD in another study [48]. The study of gene expression in peripheral lymphocytes by Runne and colleagues also failed to show consistent results found in brains of HD patients [19]. Similarly, the available studies in literature did not uncover the potential biomarkers revealed in our study. These conflicts may be originated from the differences in sample preparations, disease status, technologies, and individual biological status. Thus standard methods will be necessary to appropriately attenuate the variability in confounding factors in such studies. Furthermore, it will be promising to combine genomic and proteomic approaches with neuroimaging to successfully detect biomarkers related to disease progression in HD. Nevertheless, the downregulated genes identified in this study highlight that treatments aimed at correcting metabolic deficits and decreasing oxidative damage may be beneficial to HD patients. In addition, these molecules may serve as the potential good indicators for evaluating efficacy of preclinical and clinical treatments aimed at improving metabolic dysfunction and oxidative stress. Given that the numbers of our patients and preHD carriers are small, a larger, multi-center, longitudinal study regarding the correlation of these gene expressions with clinical and neuroimaging features will determine the role of their applications as biomarkers for HD, and might shed light on the development of novel treatments.

Ethics statement
This study was performed under a protocol approved by the Institutional Review Boards of Chang Gung Memorial Hospital and all examinations were performed after obtaining written informed consents.

Study population
Twenty subjects with HD, including 4 preHD carriers and 16 symptomatic HD patients, and 20 healthy controls were recruited in this study. Unified Huntington's Disease Rating Scale (UHDRS) were recorded for each patient [49]. The scale ranges (normal to most severe) of UHDRS include total motor score (0 to 124), independence score (100 to 10), and total functional capacity (13 to 0). None of the patients or the controls had systemic infection, autoimmune diseases, malignancies, or chronic renal, cardiac, or liver dysfunction. results were generated using the 59-nuclease assay (TaqMan) if the probe is commercially available and the ABI 7900HT Sequence Detection System (Applied Biosystems). Primers for ACO2 designed by using Primer Express software, Version 2.0, Applied Biosystems were used for amplification of Aco2 cDNA. Primers were designed to amplify transcripts across an exon junction to avoid genomic DNA contamination. Each reaction included cDNA from 100 ng of RNA, 900 nM of each primer and 100 nM of probes or primers and Universal PCR Master Mix (Applied Biosystems). Probe sequence information is indicated in Table 3. PCR parameters were 50uC for 2 min, 95uC for 10 min, 40 cycles of 95uC for 15 sec, 60uC for 1 min. Each sample was assessed in duplicate. Cycle threshold (CT, the fractional cycle number where the fluorescent signal reaches detection threshold) in each reaction was set in the linear range. Relative expression values were normalized to b-ACTIN. Relative gene expressions were calculated using the 2 DCT method, DCT = CT (b-ACTIN) -CT (target gene).The CTs of b-ACTIN across different samples ranged between 20 and 22. For each set of values, data were expressed as means 6 standard error (SE). Differences between groups were evaluated by Student's t-test, ANOVA, or ANCOVA (adjusted by age and gender) with post-hoc Bonferroni test where appropriate. Correlations of UHDRS (motor scale, independence scale and functional capacity), size of expanded polyglutamine or disease duration with levels of mRNA were analyzed by covariateadjusted generalized linear model (adjusted by age and gender). All P-values were two-tailed. The values of P,0.05 were considered significant.

Author Contributions
Conceived and designed the experiments: CC. Performed the experiments: WL. Analyzed the data: KC YC WL CC. Contributed reagents/materials/ analysis tools: KC YC YR CC. Wrote the paper: KC CC.