Genome-Wide Expression Profiling of Complex Regional Pain Syndrome

Complex regional pain syndrome (CRPS) is a chronic, progressive, and devastating pain syndrome characterized by spontaneous pain, hyperalgesia, allodynia, altered skin temperature, and motor dysfunction. Although previous gene expression profiling studies have been conducted in animal pain models, there genome-wide expression profiling in the whole blood of CRPS patients has not been reported yet. Here, we successfully identified certain pain-related genes through genome-wide expression profiling in the blood from CRPS patients. We found that 80 genes were differentially expressed between 4 CRPS patients (2 CRPS I and 2 CRPS II) and 5 controls (cut-off value: 1.5-fold change and p<0.05). Most of those genes were associated with signal transduction, developmental processes, cell structure and motility, and immunity and defense. The expression levels of major histocompatibility complex class I A subtype (HLA-A29.1), matrix metalloproteinase 9 (MMP9), alanine aminopeptidase N (ANPEP), l-histidine decarboxylase (HDC), granulocyte colony-stimulating factor 3 receptor (G-CSF3R), and signal transducer and activator of transcription 3 (STAT3) genes selected from the microarray were confirmed in 24 CRPS patients and 18 controls by quantitative reverse transcription-polymerase chain reaction (qRT-PCR). We focused on the MMP9 gene that, by qRT-PCR, showed a statistically significant difference in expression in CRPS patients compared to controls with the highest relative fold change (4.0±1.23 times and p = 1.4×10−4). The up-regulation of MMP9 gene in the blood may be related to the pain progression in CRPS patients. Our findings, which offer a valuable contribution to the understanding of the differential gene expression in CRPS may help in the understanding of the pathophysiology of CRPS pain progression.


Introduction
Complex regional pain syndrome (CRPS) is a chronic, progressive, and devastating pain syndrome that is characterized by spontaneous pain, hyperalgesia, allodynia, altered skin temperature, and motor dysfunction [1,2]. CRPS is generally classified into 2 types by the absence or presence of nerve injury. Patients with CRPS type I show no nerve injury, while type II patients exhibit nerve injury [3].
Due to the phenotypic complexity of CRPS, it is difficult to conduct a human based genome-wide association study in CRPS. Nonetheless, microarray tools have been commonly used to identify novel biomarkers that are known to contribute to pain pathways in animal pain models. Genome-wide expression analyses have been successfully performed only in animals. A different regulation of 86 genes after nerve injury was detected by a cDNA microarray analysis of spinal nerves from a rat model of neuropathic pain [4]. Furthermore, 124 co-regulated genes were identified in 3 neuropathic pain models (spared nerve injury, chronic construction injury, and spinal nerve ligation) by gene expression profiling of the rat dorsal root ganglion (DRG). Additionally, following a microarray-based screening study in large international pain cohort [5], a genetic association study was performed using single nucleotide polymorphisms (SNPs) of the potassium channel alpha subunit, KCNS1.
In addition to animal studies, recent studies focused on the identification of novel molecules or genetic loci related to neuropathic pain in humans suffering from CPRS. A genetic association study conducted in CRPS patients and controls provided a new CRPS susceptibility locus (D6S1014) in human leukocyte antigen (HLA) class I region [6]. Uçeyler et al. compared the cytokine expression (at the mRNA and protein level) in the serum between CRPS II or CRPS I patients and controls. The mRNA and protein levels of transforming growth factor (TGF)-b1 and interleukin (IL)-2 were higher and those of IL-4 and IL-10 were lower in CRPS patients than in controls [7]. Furthermore, the levels of tumor necrosis factor (TNF) receptor and IL-1b in cerebrospinal fluid and serum were found to be related to pain intensity in CRPS II patients [8]. So far, there are no reports of a genome-wide expression profiling analysis successfully conducted in CRPS patients.
In this study, we analyzed the gene expression levels in the whole blood of CRPS patients using a genome-wide expression profiling analysis and identified the molecules that were highly expressed in CPRS depending on type-I or -II. These different transcriptional profiles in CPRS may contribute to the understanding of the pathogenesis of CRPS progression.

Ethics Statement
All individuals enrolled in this study provided written informed consent for blood collection and use. The study protocol was approved by the Institutional Review Board (IRB) of Chungnam National University Hospital, St. Mary's Hospital, and Samsung Medical Center.

Patients and Pain Evaluation
A clinical diagnosis of CRPS was established using the 'Budapest criteria' published by the International Association for the Study of Pain (IASP) 2007 [9]. CRPS I and CRPS II were distinguished by the presentation of nerve injury as defined by the IASP [10]. The diagnosis is CRPS I if there is no nerve lesion, while the diagnosis is CRPS II if a nerve lesion is present. Inclusion criteria included CRPS patients who received the medication of CRPS or CRPS related depressive disorder such as pregabalin (or gabapentin), tricycllic antidepressant, oipiods, acetaminophen, selective serotonin reuptake inhibitors, serotonin norepinephrine reuptake inhibitors, and benzodiazepine derivatives. Exclusion criteria included CRPS patients who received the medication described in the inclusion criteria as a cause of other neurologic disorders or any other medications not mentioned in the inclusion criteria (Table 1). Blood samples of CRPS patients obtained while taking their medications. The healthy control group was free of infectious diseases and pain disorders, and had undergone no recent surgery at the sampling time. Mechanoallodynia was determined by the pain evoked by the Von Frey hair or brush application. A pinprick or cold ice was to the lesion to detect hyperalgesia. Thermography has been used for the evaluation of temperature asymmetry (1 . C) or skin color change. The microarray analysis was conducted in 5 controls and 2 CRPS I and 2 CRPS II patients. We selected 2 CRPS I and 2 CRPS II samples considering allodynia, hyperalgesia, spontaneous pain, temperature change, vasomotor change, atrophic change symptoms, and a high RIN (RNA integrity number) value. The mean ages of the CRPS patients and controls were 46.6610.1 and 44.764.5 y, respectively. CRPS patients (24) and 18 controls were used for quantitative real-time PCR (qRT-PCR) validation (Table 1).

Genome-wide Transcriptional Profiling
Whole blood samples were collected using PAXgene blood RNA tubes (PreAnalytiX, Hilden, Germany). Total RNA was extracted using the TRIzol Reagent (Ambion, CA, USA) and purified using RNeasy columns (Qiagen, Valencia, USA) according to the manufacturer's protocol. The concentration of the RNA was assessed using a NanoDrop spectrophotometer (NanoDrop Technologies, Inc., Wilmington, DE, USA). For quality control, RNA purity and integrity were evaluated by denaturing gel electrophoresis and the OD 260/280 ratio, and analyzed on an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, USA).

Raw Data Processing and Statistical Analysis
The raw data were processed using the software provided by the manufacturer (Illumina GenomeStudio version 2011.1, Gene Expression Module v1.9.0. We applied a filtering criterion for data analysis: a high signal value was required to obtain a detection p value ,0.05. The selected signal value of the probe was transformed using a logarithmic function and normalized using the quantile method. Statistical significance of the expression data was determined using independent t-test and fold change in which the null hypothesis was that no difference exists between the CRPS group and the control group. The false discovery rate (FDR ,0.05) was controlled by adjusting the pvalue using the Benjamini-Hochberg algorithm [11]. The data were further processed with 2 cut-off values, p-value ,0.05 and fold change .1.5. A hierarchical cluster analysis was performed using complete linkage and Euclidean distance as a measure of similarity. The significant probe list was classified into biological process and molecular function using the panther classification system (http://www.pantherdb.org). All data analysis and visualization of differentially expressed genes (DEG) was conducted using R 2.14.1 (www.r-project.org).

qRT-PCR Analysis
qRT-PCR was used to verify the differential expression that was initially detected by the array. Total RNA was isolated from whole HDC (Hs00157914_m1), STAT3 (Hs00374280_m1) and GAPDH (Hs99999905_m1). Amplification reactions were performed in triplicate with a StepOne Plus system (Applied Biosystems, CA, USA) using the following conditions: 10 min at 95uC, 40 cycles of 15 s at 95uC and 1 min at 60uC for the primer assay; and 2 min at 50uC, 10 min at 95uC, 40 cycles of 15 s at 95uC and 1 min at 60uC for the probe assay. The threshold cycle (Ct) of the GAPDH gene was used as a reference control to normalize the expression level of the target gene (DCt) to correct for experimental variation. The relative level of gene expression (DDCt) was calculated as DCt CRPS patient 2 DCt control , and the relative fold changes were determined by using the 2 2DDCt method [12]. Statistical analysis of the difference in gene expression (2 2DDCt values) levels between CRPS patients and controls was calculated by a nonparametric Mann-Whitney U test (SPSS ver 20.0). A p-value ,0.05 was considered to indicate statistical significance.

Validation of DEGs in CRPS by qRT-PCR
To validate the DEGs selected from gene expression profiling, we performed qRT-PCR with 24 CRPS (13 CRPS I and 11 CRPS II) patients and 18 controls. The expression levels of HLA-A29.1, MMP9, PTGS2, IL-8, MMP25, ANPEP, HDC, G-CSF3R, and STAT3 genes showed concordant results with the microarray data, while that of ARHGEF10 was not consistent with microarray results (Figure 3). The expression levels of 6 of those 10 genes (HLA-A29.1, MMP9, ANPEP, HDC, G-CSF3R, and STAT3) were significantly different between CRPS patients and   (Fig. 3). We also analyzed the gene expression levels in the subgroups CRPS I and CRPS II through a comparison of the 2 2DDCt value between CRPS I or CRPS II patients and controls. We found that the expression level of HLA-A29.1, MMP9, IL8, HDC, and ARHGEF10 showed a statistical difference between the CRPS I group and the control group (p = 0.011, 0.045, 0.045, 0.005, and 3.0610 24 , respectively). The relative fold changes of HLA-A29.1, MMP9, IL8, HDC, and ARHGEF10 in the CRPS I group compared to the control were 1.760.23, 1.960.51, 1.160.38, 1.760.31, and -1.360.17 (Fig. 4). We also observed that the expression level of HLA-A29.1, MMP9, ANPEP, HDC, G-CSF3R, and STAT3 significantly differed in the CRPS II patients, when compared to that in the controls (p = 0.020, 3.  (Fig. 4).

Discussion
Although the pathophysiology of neuropathic pain has not been completely understood, previous studies implicate that trauma induces activation of mast cells and macrophages and neutrophils are recruited to the injury region [23,24]. Studies have been demonstrated that CRPS are associated with inflammatory and neuroinflammatory mediators in blood of patients compared to controls [25,26]. Based on previous reports, we investigated the gene expression profiling in the blood of CRPS patients. In the present study, by genome-wide expression profiling followed by qRT-PCR validation, we found that HLA-A29.1, MMP9, ANPEP, HDC, G-CSF3R, and STAT3 genes were highly expressed in the blood of CRPS patients, compared to controls (Fig. 3). Though pain-related genes have been identified through microarray analyses in animals, no reports of successful genome-wide transcriptional profiling in CRPS have been published. This is the first successful genome-wide expression profiling analysis in the blood of CRPS patients. We observed fold change in HLA-DRB1 and HLA-DRB6 expression were the largest among the 80 genes that were up-or down-regulated in the microarray (14.9-fold and 3.1-fold, respectively) ( Table 2). However, when examined by qRT-PCR, we were not able to confirm this microarray finding. Additionally, the expression level of ARHGEF10 in qRT-PCR was inconsistent with that in microarray (Fig. 3).
In our subgroup analysis, the expression level of HLA-A29.1, MMP9, ANPEP, HDC, G-CSF3R, and STAT3 genes in both CRPS group and CRPS II subgroup was statistically different (as assessed by the 2 2DDCt value) compared to that of the control group. Fold changes in the expression of HLA-A29.  (Fig. 4). There are literature evidences on the involvements of HLA-A29.1, MMP9, IL8, ANPEP, HDC, G-CSF3R, STAT3, and ARHGEF10 genes in pain progression. A HLA polymorphism was associated with postherpetic neuralgia in a Japanese population [13]. MMP9 was up-regulated in dorsal root ganglion neurons of spinal nerve-ligated rats [23]. IL-8, a  proinflammatory cytokine, induced hyperalgesia in rats [16]. In peripheral inflamed tissues of injured rats, pain was reduced by inhibition of opioid degradation with ANPEP or neutral endopeptidase [18]. HDC is the enzyme that produces histamine and participates in central pain modulation; intrathecal administration of histamine evoked hyperalgesia in HDC knockout mice [19]. The expression of G-CSF increased in a mouse model of bone tumorinduced pain and G-CSF signaling via its receptor led to nerve remodeling and bone cancer pain [20]. STAT3 has been shown to play an important role in inducing astrocyte proliferation and tactile allodynia in a neuropathic pain rat model [21]. ARH-GEF10 was found to play an important role in myelination of peripheral nerves [22].
Based on previous studies and our results, we could assume that the direction of regulation of HLA-A29.1, MMP9, and HDC genes may be the same in both CRPS I and CRPS II, although the level of regulation in CRPS II was greater than that of CRPS I, that the up-regulation of IL8 and the down-regulation of ARHGEF10 gene may be related with the pain progression of CRPS I, and that the up-regulation of ANPEP, G-CSF3R, and STAT3 genes may be associated with the pathogenesis of CRPS II. Interestingly, the expression of MMP9 of validated genes was prominently upregulated in subgroups CRPS I (1.960.26 times and p = 0.045) and CRPS II patients (6.462.47 times and p = 3.4610 27 ) (Fig. 4). Thus, we particularly focused on the MMP9 gene expression.
There has been interesting evidence that supports the involvement of MMP9 in neuropathic pain. Matrix metalloproteinases are a family of endopeptidases that play an important role in neuroinflammation, developmental processes, and wound healing [27,28]. MMP9 is one of the major gelatinases. MMP9 was upregulated in rat DRG after a sciatic nerve crush that led to demyelination and its levels were regulated by TNF-a and IL-1b [29]. MMP9 was also up-regulated in the DRG neurons of spinal nerve-ligated rats and induced neuropathic pain by cleaving IL-1b in the dorsal root ganglion and spinal cord; MMP9-null mice showed a reduction of pain in the form of mechanical allodynia [30]. Elevated MMP9 levels were observed in the plasma of migraineurs, even during headache-free periods. [14].
There are some limitations to our study. First, the sample size was too small to have statistical power. Second, all CRPS patients who participated in this study took several pain medications, such as pregabalin, gabapentin, tricyclic antidepressants, and opioids. Thus, we cannot rule out that the medications had an effect on the gene expression. To adequately control for this possibility further studies would be required with a control group of medication only. Third, CRPS patients that participated in this study were heterogenous with respect to disease duration.
In conclusion, based on the genome-wide gene expression profiling in the blood of CRPS patients, we suggest that the upregulation of the MMP9 gene in the blood might be related to pain progression in CRPS, although further replication and functional studies conducted in large populations are required to define the role of this gene in CRPS. This study offers an early and fascinating assay of gene expression in peripheral leukocytes in CRPS patients, one which may lead to new mechanisms and therefore potentially new therapies.