Abstract
Accumulating evidence suggests that the extent of brain injury and the clinical outcome after traumatic brain injury (TBI) are modulated, to some degree, by genetic variants. Aquaporin-4 (AQP4) is the predominant water channel in the central nervous system and plays a critical role in controlling the water content of brain cells and the development of brain edema after TBI. We sought to investigate the influence of the AQP4 gene region on patient outcome after TBI by genotyping tag single nucleotide polymorphisms (SNPs) along AQP4 gene. A total of 363 patients with TBI (19.6% female) were prospectively evaluated. Data including the Glasgow Coma Scale (GCS) scores at admission, the presence of intracranial hemorrhage, and the 6-month Glasgow Outcome Scale (GOS) scores were collected. Seven tag SNPs across the AQP4 gene were identified based on the HapMap data. Using logistic regression analyses, SNPs and haplotypes were tested for associations with 6-month GOS after adjusting for age, GCS score, and sex. Significant associations with TBI outcome were detected for rs3763043 (OR [95% confidence interval (CI)]: 5.15 [1.60–16.5], p=0.006, for recessive model), rs3875089 (OR [95% CI]: 0.18 [0.07–0.50] p=0.0009, for allele difference model), and a common haplotype of AQP4 tag SNPs (OR [95% CI]: 2.94, [1.34–6.36], p=0.0065). AQP4 tag SNPs were not found to influence the initial severity of TBI or the presence of intracranial hemorrhages. In conclusion, the present study provides evidence for possible involvement of genetic variations in AQP4 gene in the functional outcome of patients with TBI.
Key words: : AQP4, genetics, polymorphism, tag SNPs, traumatic brain injury
Introduction
Traumatic brain injury (TBI) is one of the leading causes of mortality and disability worldwide, especially among young persons in their productive years of life. Despite the advances in the management of TBI, there is great variability in the clinical outcome after TBI, with more than 40% of the hospitalized TBI survivors having some kind of long-term disability.1 The extent of brain damage after TBI is determined not only by the severity of the primary impact, but also by secondary brain changes such as cerebral edema, increased intracranial pressure, tissue hypoxia-ischemia, and disruption of the blood-brain barrier.2 Expression studies have shown that several genes are implicated in the pathophysiology of secondary brain damage3 and the extent of brain injury. Moreover, the widely observed clinical variability after TBI appears to be modulated, at least to some degree, by genetic variants.4,5
Aquaporin-4 (AQP4) is the predominant water channel in the central nervous system and plays a critical role in controlling the water content of brain cells. It is mainly expressed in astrocyte foot processes that surround blood capillaries or form the glial limiting membrane, and also in ependymal cells.6,7 AQP4 manifests opposing effects in the pathophysiology of cytotoxic and vasogenic edema. While protecting against vasogenic edema by increasing the rate of water clearance, it aggravates cytotoxic edema by inducing astrocytic cell swelling.6 Animal experimental studies have demonstrated that AQP4-null mice were found to have reduced cytotoxic cerebral edema; however, they showed increased vasogenic edema.8 Of note, both types of edema are present in the brain tissue after TBI and are the most significant predictors of TBI outcome.9 There is, therefore, a strong biological rationale for involvement of AQP4 gene in brain edema formation after TBI and consequently in the clinical outcome of patients with TBI.
The aim of the present study was to investigate possible associations between genetic variations of AQP4 gene and the patients' initial TBI severity, the presence of intracranial hemorrhage, and the long-term clinical outcome after TBI. Previous studies have described a number of nonsynonymous functional polymorphisms of AQP4 gene, which increase or reduce water permeability either by gain or loss of AQP4 protein function.10 These polymorphisms and the rest coding variants of AQP4 gene according to the HapMap Project data (http://hapmap.ncbi.nlm.nih.gov) are very rare, however, with observed frequencies less than 0.01, and therefore are highly unlikely to determine common diseases or complex traits.11 This probably was the reason why a previous study investigating possible influence of genetic variants of AQP4 exon 4 failed to demonstrate any significant influence on outcome of patients with TBI because no coding polymorphisms in exon 4 of AQP4 gene were detected.12
To overcome this problem, the present study was designed to examine the effect of both coding and noncoding AQP4 gene variability on the outcome after TBI. In an attempt to capture a substantial proportion of common sequence variations along AQP4 gene, the tag single nucleotide polymorphism (SNP) approach was used. On the basis of the HapMap linkage disequilibrium (LD) blocks, seven tag SNPs were identified across AQP4 gene, which were tested for association with TBI outcomes.
Methods
Patients
Patients with TBI were prospectively recruited from consecutive admissions in the neurosurgery department of the University Hospital of Larissa (a tertiary referral institution in central Greece) as previously described.13,14 Patients were included if they sustained isolated blunt head trauma, were of Greek origin, and 18 years of age or older. Patients with penetrating TBI, multiple organ injury, or anoxic brain damage were excluded. In total, 363 patients with TBI were enrolled in the study. The study protocol was reviewed and approved by our Institutional Review Board and a detailed written informed consent was obtained from each participant (or a close relative in the case of unresponsive or comatose patients).
Demographic and TBI related data regarding age and sex of the participants, severity of the initial injury, findings of the initial CT scan, and 6-month clinical outcome were also prospectively collected. The clinical severity of head injury at admission was measured by means of the Glascow Coma Scale (GCS).15 Patients were allocated into one of the three categories based on the GCS score: severe TBI (GCS=3–8), moderate (GCS=9–12), and mild TBI (GCS=13–15).16
Six months after TBI, most of the patients were reevaluated in the outpatient neurosurgical clinic. For 54 (14.8%) patients, information was received by telephone from the patients themselves or their caregivers. With respect to the functional recovery, patients were allocated to one of the categories of the Glascow Outcome Scale (GOS) by an examiner who was unaware of the initial GCS score at admission and the results of the performed genotyping. For the analyses, GOS was dichotomized into good (good recovery, moderate disability) and poor (severe disability, vegetative state, death) outcome to provide a binary primary outcome variable corresponding to functional independence or not, respectively.15,16
Isolation of DNA, SNP selection, and genotyping
Genomic DNA was extracted from peripheral blood leukocytes using a salting out method.
Based on the HapMap data for northern and western Europe (CEU) population (Release 27, Phase II+III, Feb09, on NCBI B36 assembly, dbSNP b126), tag SNPs across the entire region of AQP4 gene (13.7Kbp spanning from 22686005 to 22699714 in chromosome 18) were selected using the tagger algorithm (http://www.broadinstitute.org/mpg/tagger/) with a pairwise approach, an r2 cutoff of ≥0.8 and a minor allele frequency >0.05. A total of seven tag SNPs in three distinct gene regions were retrieved: in the 3′UTR region (rs335929, rs3763043), in intronic region between exons 4–5 (rs11661256, rs335931), and in intron 1–2 (rs3763040, rs4800773, rs3875089). The positions of tag SNPs along AQP4 gene are depicted in Supplementary Figure 1 (see online supplementary material at ftp.liebertpub.com). Detailed information on the AQP4 tag SNPs is presented in Table 1.
Table 1.
Genotyped AQP4 TAG SNPs Characteristics
| rs number | Chromosome position1 | Distance from gene start2 | Gene position | Function | Minor allele (frequency) | |
|---|---|---|---|---|---|---|
| 1. | rs335929 | 24435587 | 3585 | 3'UTR | Regulatory | C (0.164) |
| 2. | rs3763043 | 24435818 | 3816 | 3'UTR | Regulatory | T (0.358) |
| 3. | rs11661256 | 24438540 | 6538 | Intron 4–5 | No-coding | A (0.142) |
| 4. | rs335931 | 24439072 | 7070 | Intron 4–5 | No-coding | G (0.167) |
| 5. | rs3763040 | 24444374 | 12372 | Intron 1–2 | No-coding | A (0.241) |
| 6. | rs4800773 | 24444981 | 12979 | Intron 1–2 | No-coding | A (0.412) |
| 7. | rs3875089 | 24445433 | 13431 | Intron 1–2 | No-coding | C (0.150) |
Chromosome 18 positions of each SNP based on dbSNP human build 135 and human genome build 37.3.
Distance from gene start position (24432002)-reverse strand.
Minor allele frequencies for Caucasian population (HapMap Data Rel 27, Phase II+III, Feb09, on NCBI B36 assembly, dbSNP b126).
Tag SNPs genotyping was performed with TaqMan allele specific discrimination assays on an ABI PRISM® 7900 Sequence Detection System and analyzed with the SDS software (Applied Biosystems, Foster City, CA).
Statistical analysis
Hardy-Weinberg equilibrium (HWE) was assessed using the exact test. Power calculation was performed by means of the Quanto software (http://hydra.usc.edu/GxE, version 1.2.4) that computes sample size or power for genetic association studies.17 Based on the expectation maximization algorithm, haplotypes from the tag SNPs alleles were constructed using the SimHap18 and the SHEsis (http://analysis.bio-x.cn)19 platforms. Pair-wise D’ and r2 LD measures between SNPs were made using the JLIN program.20
Univariate and multivariate logistic regression models were used to evaluate possible associations between individual tag SNPs or constructed haplotypes and patients' initial TBI severity (mild [GCS score: 13–15] vs. severe [GCS score: 3–8] TBI), the presence or not of intracranial hemorrhage and the 6-month clinical outcome (GOS good vs. poor outcome at 6 months) after TBI. Models for TBI severity and presence of intracranial hemorrhage were adjusted for age and sex, whereas models for GOS outcome were also adjusted for GCS score at admission. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated assuming different modes of inheritance (dominant: [AA] vs. [Ab+bb], recessive: [AA+Ab] vs. [bb] and allele difference model [A] vs. [b]).21 Although conservative for tightly linked SNPs,22 the Bonferroni correction for multiple comparisons was applied, and the significance threshold was set to 0.05 divided by the number of the tested SNPs. Statistical analysis was performed with the SPSS software version 17.0 (SPSS Inc., Chicago, IL).
Results
The baseline demographic, clinical, and imaging characteristics of the 363 patients with TBI are presented in Table 2. Patient characteristics and GCS scores at admission in relation to the dichotomized 6 months GOS are presented in Table 3.
Table 2.
Demographic Characteristics of the TBI Cohort
| Characteristics (n=363) | Mean (SD, range) |
|---|---|
| Age (years) | 42.67 (21.8, 18–88) |
| Frequency | |
| Female (n=71) | 19.5% |
| GCS 13–15 (n=179) | 49.3% |
| Favorable outcome (n=170) | 95.0% |
| Unfavorable outcome (n=9) | 5.0% |
| GCS 9–12 (n=54) | 14.9% |
| Favorable outcome (n=36) | 66.7% |
| Unfavorable outcome (n=18) | 33.3% |
| GCS 3–8 (n=130) | 35.8% |
| Favorable outcome (n=78) | 60.0% |
| Unfavorable outcome (n=52) | 40.0% |
| Hemorrhagic event | |
| Any (n=268) | 73.8% |
| Hemorrhagic contusions (n=152) | 41.8% |
| SAH (n=178) | 49.0% |
| Epidural hematoma (n=48) | 13.2% |
| Subdural hematoma (n=95) | 26.2% |
| Intracerebral hematoma (n=30) | 8.3% |
| Intraventricular hemorrhage (n=47) | 13.0% |
SD, standard deviation; GCS, Glasgow Coma Scale; SAH, subarachnoid hemorrhage.
Table 3.
Patient Characteristics, Glasgow Coma Scale Score at Admission, and 6 months Glasgow Outcome Scale
| Whole TBI sample (n=363) | Favorable outcome (n=284) | Unfavorable outcome (n=79) | |
|---|---|---|---|
| Sex | |||
| Male | 292 (80.4) | 228 (80.3) | 64 (81.0) |
| Female | 71 (19.6) | 56 (19.7) | 15 (19.0) |
| Age (years) | |||
| Mean | 42.67 | 39.20 | 55.19 |
| SD | 21.80 | 20.15 | 23.05 |
| Range | 18–88 | 18–84 | 18–88 |
| GCS at admission | |||
| 3–8 | 130 (0.358) | 78 (0.275) | 52 (0.658) |
| 9–12 | 54 (0.149) | 36 (0.127) | 18 (0.228) |
| 13–15 | 179 (0.493) | 170 (0.598) | 9 (0.114) |
SD, standard deviation; GCS, Glasgow Coma Scale.
Allele and genotype frequencies of the study TBI population, as well as of the favorable and unfavorable 6-months GOS outcome groups are shown in Table 4. Among the seven genotyped tag SNPs, rs3763040 exhibited a departure from HWE (p=0.002) and therefore was excluded from further analyses. Accordingly, the p value significance level assuming the Bonferroni correction for multiple comparisons was set at 0.05/6=0.008. Power analysis showed that the sample size was sufficient to detect an OR of 1.67–2.15 with a statistical power of >80%, assuming a significance level of 0.05, the additive mode of inheritance and a minor allele frequency of 37.3% and 8.4% (the highest (in rs4800773) and the lowest (in rs11661256) in the sample. respectively).
Table 4.
Allelic and Genotype Frequencies for Single Nucleotide Polymorphisms in the Traumatic Brain Injury Sample and Those with Favorable or Unfavorable Outcome
| SNP | Genotypes/alleles | Whole TBI sample, n (%) | Favorable outcome, n (%) | Unfavorable outcome, n (%) |
|---|---|---|---|---|
| rs335929 | ||||
| Genotype | A/A | 205 (0.579) | 152 (0.553) | 53 (0.671) |
| A/C | 125 (0.353) | 104 (0.378) | 21 (0.266) | |
| C/C | 24 (0.068) | 19 (0.069) | 5 (0.063) | |
| Allele | A | 535 (0.756) | 408 (0.742) | 127 (0.804) |
| C | 173 (0.244) | 142 (0.258) | 31 (0.196) | |
| rs3763043 | ||||
| Genotype | C/C | 196 (0.552) | 166 (0.601) | 30 (0.380) |
| C/T | 131 (0.369) | 94 (0.341) | 37 (0.468) | |
| T/T | 28 (0.079) | 16 (0.058) | 12 (0.152) | |
| Allele | C | 523 (0.737) | 426 (0.772) | 97 (0.614) |
| T | 187 (0.263) | 126 (0.228) | 61 (0.386) | |
| rs11661256 | ||||
| Genotype | T/T | 297 (0.844) | 234 (0.848) | 63 (0.829) |
| T/A | 51 (0.145) | 38 (0.138) | 13 (0.171) | |
| A/A | 4 (0.011) | 4 (0.014) | 0 (0.000) | |
| Allele | T | 645 (0.916) | 506 (0.917) | 139 (0.914) |
| A | 59 (0.084) | 46 (0.083) | 13 (0.086) | |
| rs335931 | ||||
| Genotype | A/A | 207 (0.591) | 152 (0.561) | 55 (0.696) |
| A/G | 123 (0.351) | 102 (0.376) | 21 (0.266) | |
| G/G | 20 (0.057) | 17 (0.063) | 3 (0.038) | |
| Allele | A | 537 (0.767) | 406 (0.749) | 131 (0.829) |
| G | 163 (0.233) | 136 (0.251) | 27 (0.171) | |
| rs3763040 | ||||
| Genotype | G/G | 219 (0.607) | 170 (0.603) | 49 (0.620) |
| G/A | 110 (0.305) | 92 (0.326) | 18 (0.228) | |
| A/A | 32 (0.089) | 20 (0.071) | 12 (0.152) | |
| Allele | G | 548 (0.759) | 432 (0.766) | 116 (0.734) |
| A | 174 (0.241) | 132 (0.234) | 42 (0.266) | |
| rs4800773 | ||||
| Genotype | G/G | 137 (0.395) | 104 (0.388) | 33 (0.418) |
| G/A | 161 (0.464) | 130 (0.485) | 31 (0.392) | |
| A/A | 49 (0.141) | 34 (0.127) | 15 (0.190) | |
| Allele | G | 435 (0.627) | 338 (0.631) | 97 (0.614) |
| A | 259 (0.373) | 198 (0.369) | 61 (0.386) | |
| rs3875089 | ||||
| Genotype | T/T | 289 (0.810) | 226 (0.804) | 63 (0.829) |
| T/C | 61 (0.171) | 48 (0.171) | 13 (0.171) | |
| C/C | 7 (0.019) | 7 (0.025) | 0 (0.000) | |
| Allele | T | 639 (0.895) | 500 (0.890) | 139 (0.914) |
| C | 75 (0.105) | 62( 0.110) | 13 (0.086) | |
TBI, traumatic brain injury.
Univariate and multivariate single locus analysis with 6 months GOS outcome is presented in Table 5. In multivariate analysis, age (OR:1.09, 95% CI [1.06–1.12], p<0.001) and GCS score at admission (OR: 0.56, 95% CI [0.48–0.65], p<0.001) but not sex were significantly associated with GOS at 6 months. In addition, two tag SNPs were found to influence the outcome of TBI: rs3763043 and rs3875089. In the rs3763043 SNP, the TT genotype was significantly more prevalent in the poor outcome group of patients yielding an OR of 5.15 (95% CI [1.60–16.5], p=0.006) in the recessive model. The dominant and allele difference models also revealed significant associations with TBI outcome that survived the Bonferroni correction. In the rs3875089 SNP, the less common C allele was found to have a protective effect, because it was more frequent among the patients with favorable outcome. Both dominant model and allele difference showed strong associations (OR: 0.22, 95% CI [0.07–0.65], p=0.006 and OR:0.18, 95% CI [0.07–0.50], p=0.0009, respectively) that remained significant after adjustment for multiple comparisons.
Table 5.
Univariate and Multivariate Single Locus Association with 6 months Glasgow Outcome Score outcome after Traumatic Brain Injury
| Univariate analysis | Multivariate analysis | ||||
|---|---|---|---|---|---|
| Good vs. poor outcome | Good vs. poor outcome | ||||
| SNPs | Model | OR (95% CI) | p | OR (95% CI) | p |
| rs335929 | Dom | 0.60 (0.35–1.02) | 0.06 | 0.49( 0.25–1.00) | 0.051 |
| Rec | 0.91 (0.33–2.52) | 0.85 | 0.80 (0.21–3.16) | 0.76 | |
| Allele difference | 0.71 (0.46–1.09) | 0.12 | 0.61 (0.35–1.08) | 0.09 | |
| rs3763043 | Dom | 2.46 (1.47–4.12) | 0.0005 | 2.39 (1.25–4.58) | 0.008 |
| Rec | 2.91 (1.31–6.44) | 0.008 | 5.15 (1.60–16.5) | 0.006 | |
| Allele difference | 2.08 (1.42–3.04) | 0.0001 | 2.30 (1.38–3.83) | 0.001 | |
| rs11661256 | Dom | 1.14 (0.58–2.27) | 0.68 | 0.24 (0.07–0.75) | 0.014 |
| Rec | – | – | – | – | |
| Allele difference | 1.02 (0.54–1.92) | 0.93 | 0.24 (0.07–0.75) | 0.014 | |
| rs335931 | Dom | 0.55 (0.32–0.95) | 0.03 | 0.40 (0.19–0.82) | 0.012 |
| Rec | 0.58 (0.16–2.06) | 0.41 | 0.27 (0.05–1.44) | 0.126 | |
| Allele difference | 0.61 (0.39–0.97) | 0.039 | 0.44 (0.24–0.82) | 0.009 | |
| rs4800773 | Dom | 0.88 (0.53–1.47) | 0.63 | 0.96 (0.48–1.89) | 0.91 |
| Rec | 1.61 (0.82–3.14) | 0.16 | 2.18 (0.81–5.89) | 0.12 | |
| Allele difference | 1.07 (0.74–1.54) | 0.70 | 1.17 (0.72–1.92) | 0.51 | |
| rs3875089 | Dom | 0.84 (0.43–1.65) | 0.62 | 0.22 (0.07–0.65) | 0.006 |
| Rec | – | – | – | – | |
| Allele difference | 0.77 (0.42–1.41) | 0.39 | 0.18 (0.07–0.50) | 0.0009 | |
R, odds ratio; CI, confidence interval.
Pairwise LD between the studied AQP4 tag SNPs in the TBI sample (Supplementary Fig. 2; see online supplementary material at ftp.liebertpub.com), has revealed high disequilibrium (D’) values between most tag SNPs across the AQP4 gene. In the haplotype multilocus-based analysis of the tag SNPs (Table 6), six common haplotypes were predicted to be formed from the six tag SNPs (predicted frequency >0.05; Table 4). The ATTAGT common haplotype was predicted to be significantly overrepresented in the poor outcome group compared with the favorable outcome group (16.9% vs. 7.4%, OR: 2.94, 95% CI [1.34–6.36], p=0.0065). This haplotype combines the T allele of rs3763043 and the T allele of rs3875089 that were associated with poor outcome in the single locus analysis.
Table 6.
Predicted Haplotypes from the AQP4 TAG Single Nucleotide Polymorphisms in Good and Poor Outcome Traumatic Brain Injury Groups
| Univariate analysis | Multivariate analysis | ||||||
|---|---|---|---|---|---|---|---|
| Good vs. poor outcome | Good vs. poor outcome | ||||||
| Haplotypesa | % Good outcome | % Poor outcome | OR (95% CI) | p | OR (95% CI) | p | |
| I | ACTAGT | 26.0 | 21.3 | 0.99 (0.57–1.72) | 0.788 | 1.14 (0.55–2.34) | 0.71 |
| II | CCTGGT | 23.5 | 13.2 | 0.56 (0.30–1.04) | 0.061 | 0.57 (0.26–1.24) | 0.15 |
| III | ATTAAT | 13.6 | 13.7 | 1.05 (0.53–2.04) | 0.65 | 1.79 (0.67–4.60) | 0.24 |
| IV | ACTAAT | 10.4 | 13.8 | 0.78 (0.35–1.65) | 0.50 | 1.99 (0.79–4.87) | 0.14 |
| V | ACAAAC | 8.3 | 4.0 | 0.96 (0.39–2.29) | 0.75 | 0.43 (0.10–1.64) | 0.19 |
| VI | ATTAGT | 7.4 | 16.9 | 3.24 (1.74–5.96) | 0.00013 | 2.94 (1.34–6.36) | 0.0065 |
Haplotypes derived from polymorphisms rs335929 A/C, rs3763043 C/T, rs11661256 T/A, rs335931 A/G, rs4800773 G/A, and rs3875089 T/C.
OR, odds ratio; CI, confidence interval.
Univariate and multivariate single locus analyses with initial TBI severity and the presence of intracranial hemorrhage are presented in Supplementary Tables 1 and 2 (see online supplementary material at ftp.liebertpub.com). These analyses did not provide any significant effect of AQP4 SNPs, except for a marginal association of rs4800773 with hemorrhagic events (p=0.04), which, however, did not survive Bonferroni correction.
Discussion
In the present study, specific variants of the AQP4 gene, which codes for the predominant water channel in the central nervous system, were found to significantly influence the 6-month clinical outcome after TBI. No similar effects were found with respect to the initial TBI severity as measured by the GCS score at admission or the presence of intracranial hemorrhage.
It is known that the ongoing sequelae of damage to nervous tissue after the initial TBI is perpetuated by cerebral edema, increased intracranial pressure, tissue hypoxia-ischemia, and disruption of the blood-brain barrier (BBB)2. There are two major types of brain edema: cytotoxic and vasogenic.23 Cytotoxic edema occurs when brain cells are damaged and the Na–K ATPase fails to maintain transmembrane ion gradients,while the BBB remains intact. The cells swell from influx of fluid from the vascular compartment. Although all brain cell types swell in cytotoxic edema, the astrocytes contribute the most of the brain swelling.24 Vasogenic edema occurs when the BBB becomes disrupted (vascular leak) causing a net flux of water driven by hydrostatic pressure gradient from the blood to the brain. The excess filtrated water with plasma proteins and electrolytes expands the interstitial compartment.25 In TBI, both types of edema are present in the brain tissue, and they are developed in a biphasic pattern: vasogenic edema occurs in the first few post-traumatic hours as a result of rupture of the BBB, which is gradually restored within the following 7 days, whereas cytotoxic edema develops more slowly but persists for up to 2 weeks.9
AQP4 manifests opposing effects in the pathophysiology of the two types of brain edema by facilitating edema fluid formation in cytotoxic edema and by increasing the rate of edema fluid elimination in vasogenic brain edema. AQP4-null mice are less prone to cytotoxic brain edema6,8,26 whereas overexpression of AQP4 in transgenic mice accelerates cytotoxic brain swelling.27 On the contrary, AQP4 deletion aggravates vasogenic brain edema produced by a fluid leak.6,8 AQP4 gene, therefore, has a strong biological rationale for involvement in brain edema formation after TBI and consequently in the clinical outcome of patients with TBI.
Despite the considerable efforts made in recent years to identify any possible associations between AQP4 gene polymorphisms and various diseases including inflammatory demyelinating disease (multiple sclerosis and neuromyelitis optica),28,29 idiopathic intracranial hypertension,30 cerebral small vessel disease,31 and migraine,32 there has been little success in identifying regions of AQP4 gene with functional significance.33–35 A previous study in patients with TBI, examining possible involvement of coding variants in AQP4 exon 4, failed to demonstrate any significant influence on outcome of patients with TBI, because no coding polymorphisms in exon 4 of AQP4 gene were detected.12 AQP4 gene coding variants are known to influence water permeability of cell membrane.10 These variants are very rare, however, and therefore are unlikely to exert meaningful effects on common phenotypes.
Our study provides evidence for possible involvement of two genetic variations, rs3763043 and rs3875089, and a common haplotype of AQP4 gene in the clinical outcome of patients with TBI. Rs3763043 polymorphism is located in the 3′UTR site of the gene, a region without known regulatory function or transcription factor-binding sites. Interestingly, however, a previous study provided evidence for an association between the rs9951307 polymorphism of AQP4 gene and severe brain edema in patients with middle cerebral artery occlusion.33 Of note, the rs9951307 SNP is located outside the AQP4 gene, approximately 1.5 Kb downstream of the 3′ end point of the AQP4 gene and 5.3 Kb far from the rs3763043, which reached statistical significance in our study. According to HapMap LD data, these two polymorphisms are in complete LD (D’=1). It is therefore possible that polymorphisms within the 3′UTR may influence the function of the AQP4 gene.
The second SNP associated with TBI outcome in our study was rs3875089, which is located in intron 1–2. Rs3875089 is in complete LD (D’=1) with the AQP4 promoter region. Previous functional analysis of AQP4 promoter has revealed several transcriptional regulatory elements that influence AQP4 gene expression.36 In addition, intron 1–2 constitutes the promoter region of the shorter AQP4 protein isoform b (Supplementary Fig. 1; see online supplementary material at ftp.liebertpub.com). Indeed, intron 1–2 is a region with important regulatory properties, because many transcription-binding sites, CpG methylation sites, histone modification regions, and a CpG island are located in this gene region (information retrieved from UCSC genome browser (http://genome.ucsc.edu/cgi-bin/hgTracks?org=human). It is possible therefore, that intron 1–2 may also represent a region of the gene with functional significance and implications in diseases phenotypes, including TBI outcome.
In support of our findings, recent studies on patients with idiopathic demyelinating disorders of the central nervous system including neuromyelitis optica and in cases of sudden infant death syndrome reported significant associations in polymorphisms located in intron 1–234 and the promoter region.34,35
The precise mechanisms by which AQP-4 gene or its variants may influence the pathophysiology of TBI have not been fully elucidated. There is evidence that TBI induces biphasic changes in the expression of AQP4 in astrocytic processes with an early downregulation of AQP4, which facilitates an initial vasogenic edema, followed by a persistent upregulation of AQP4 expression that enhances formation of cytotoxic edema, and thus determines the extent of brain damage.9 Interestingly, an AQP4-deficient animal model showed remarkably improved neurological outcome, with reduced neuronal death and myelin breakdown after spinal cord compression injury.37 There was also a suggestion about the potential signaling role of AQP4 in terms of edema formation.38
The presence of brain edema after TBI and the ensuing changes in cerebral hemodynamics are critical predictors of TBI outcome.2 It is possible that alterations in expression of AQP4 gene because of genetic polymorphisms may influence the fluid and blood flow balance and consequently the extent of brain damage. Interestingly, similar correlations between genetic variants and TBI outcome through alterations in cerebral hemodynamics already exist in the literature. In particular, a common variant in NOS3 gene was reported to determine cerebral blood flow parameters during the acute phase of TBI and subsequently the overall clinical outcome.39 Besides the role of AQP4 in brain edema formation, AQP4 has also been implicated in cerebral response to brain injury by affecting astrocyte migration, neuroexcitation, neuronal regeneration, and neuroinflammation7 and through these mechanisms it may influence the degree of brain damage.
The use of certain SNPs as potential biomarkers to identify subgroups of patients with altered AQP4 expression or structure would be very promising in the future in the light of recent advances in understanding the role of AQP4 and the identification of potential aquaporin modulators.6 AQP4 modulators have been proposed to have therapeutic potential in the management of brain edema of various etiologies including brain trauma.6
Our study did not show any significant effects of AQP4 genetic variants on the initial TBI severity or the presence of intracranial hemorrhage. It is possible that secondary processes after TBI including brain edema may be induced after a certain period and thus may not have significant effects on the initial clinical presentation after TBI. In addition, AQP4 knockout animal models showed indications of altered BBB integrity but no evidence of intracranial hemorrhages,40 which is in accordance with our findings.
Our study has all the limitations that are applicable to genetic association studies.41 We cannot also exclude the possibility of selection bias because our study is based on hospital emergency department admissions and milder cases may be underrepresented. In addition, we used a single measure of physical disability—the five-point GOS. The use of additional scales, such as the Injury Severity Score or the Disability Rating Scale, would probably have increased the credibility of our findings. Further, to increase the power of the study, we used dichotomized GOS instead of the ordinal analysis. This approach, however, may in some instances alter the outcome sensitivity.42 Finally, the use of additional clinical covariates in the regression models such as the pupillary reactions might have increased the likelihood to disclose the net effects of SNPs on TBI outcome.
Conclusion
The present study provides evidence for possible involvement of genetic variations in the AQP4 gene in the functional outcome of patients with TBI. This association is of particular interest: AQP4 is regarded as a potential therapeutic target in patients with TBI for the prevention and treatment of brain edema because AQP4 inhibitors are expected to protect the brain in cytotoxic edema, whereas AQP4 activators or upregulators facilitate the clearance of vasogenic brain edema.6 Although the identification of selective AQP4 modulators still remains elusive,43 advances in this field are expected to have profound effects in the treatment of patients with TBI and may positively influence their overall outcome.
Supplementary Material
Acknowledgments
Georgios Tsivgoulis has been supported by the European Regional Development Fund - Project FNUSA-ICRC (No. CZ.1.05/1.1.00/02.0123)
The study was supported in part by a research grant of the Research Committee of the University of Thessaly, Greece (Code: 2845).
Author Disclosure Statement
No competing financial interests exist.
References
- 1.Selassie A.W., Zaloshnja E., Langlois J.A., Miller T., Jones P., and Steiner C. (2008). Incidence of long-term disability following traumatic brain injury hospitalization, United States, 2003. J. Head Trauma Rehabil. 23, 123–131 [DOI] [PubMed] [Google Scholar]
- 2.Golding E.M. (2002). Sequelae following traumatic brain injury. The cerebrovascular perspective. Brain Res. Brain Res. Rev. 38, 377–388 [DOI] [PubMed] [Google Scholar]
- 3.Lei P., Li Y., Chen X., Yang S., and Zhang J. (2009). Microarray based analysis of microRNA expression in rat cerebral cortex after traumatic brain injury. Brain Res, 1284, 191–201 [DOI] [PubMed] [Google Scholar]
- 4.Dardiotis E., Grigoriadis S., and Hadjigeorgiou G.M. (2012). Genetic factors influencing outcome from neurotrauma. Curr. Opin. Psychiatry 25, 231–238 [DOI] [PubMed] [Google Scholar]
- 5.Dardiotis E., Fountas K.N., Dardioti M., Xiromerisiou G., Kapsalaki E., Tasiou A., and Hadjigeorgiou G.M. (2010). Genetic association studies in patients with traumatic brain injury. Neurosurg. Focus 28, E9. [DOI] [PubMed] [Google Scholar]
- 6.Papadopoulos M.C., and Verkman A.S. (2007). Aquaporin-4 and brain edema. Pediatr. Nephrol. 22, 778–784 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Papadopoulos M.C., and Verkman A.S. (2013). Aquaporin water channels in the nervous system. Nat. Rev. Neurosci. 14, 265–277 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Verkman A.S., Binder D.K., Bloch O., Auguste K., and Papadopoulos M.C. (2006). Three distinct roles of aquaporin-4 in brain function revealed by knockout mice. Biochim. Biophys. Acta 1758, 1085–1093 [DOI] [PubMed] [Google Scholar]
- 9.Donkin J.J., and Vink R. (2010). Mechanisms of cerebral edema in traumatic brain injury: therapeutic developments. Curr. Opin. Neurol. 23, 293–299 [DOI] [PubMed] [Google Scholar]
- 10.Sorani M.D., Zador Z., Hurowitz E., Yan D., Giacomini K.M., and Manley G.T. (2008). Novel variants in human Aquaporin-4 reduce cellular water permeability. Hum. Mol. Genet. 17, 2379–2389 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Frazer K.A., Murray S.S., Schork N.J., and Topol E.J. (2009). Human genetic variation and its contribution to complex traits. Nat. Rev. Genet. 10, 241–251 [DOI] [PubMed] [Google Scholar]
- 12.Romeiro R.R., Romano-Silva M.A., De Marco L., Teixeira A.L., Jr, and Correa H. (2007). Can variation in aquaporin 4 gene be associated with different outcomes in traumatic brain edema? Neurosci. Lett. 426, 133–134 [DOI] [PubMed] [Google Scholar]
- 13.Hadjigeorgiou G.M., Paterakis K., Dardiotis E., Dardioti M., Aggelakis K., Tasiou A., Xiromerisiou G., Komnos A., Zintzaras E., Scarmeas N., Papadimitriou A., and Karantanas A. (2005). IL-1RN and IL-1B gene polymorphisms and cerebral hemorrhagic events after traumatic brain injury. Neurology 65, 1077–1082 [DOI] [PubMed] [Google Scholar]
- 14.Dardiotis E., Dardioti M., Hadjigeorgiou G.M., and Paterakis K. (2006). Re: Lack of association between the IL1A gene (-889) polymorphism and outcome after head injury. Tanriverdi T et al. Surg. Neurol. 2006;65:7–10. Sur Neuro 66, 334–335 [DOI] [PubMed] [Google Scholar]
- 15.Teasdale G., Murray G., Parker L., and Jennett B. (1979). Adding up the Glasgow Coma Score. Acta Neurochir. Suppl. (Wien) 28, 13–16 [DOI] [PubMed] [Google Scholar]
- 16.Teasdale G.M., Nicoll J.A., Murray G., and Fiddes M. (1997). Association of apolipoprotein E polymorphism with outcome after head injury. Lancet 350, 1069–1071 [DOI] [PubMed] [Google Scholar]
- 17.Gauderman W.J. (2002). Sample size requirements for association studies of gene-gene interaction. Am. J. Epidemiol. 155, 478–484 [DOI] [PubMed] [Google Scholar]
- 18.Carter K.W., McCaskie P.A., and Palmer L.J. (2008). SimHap GUI: an intuitive graphical user interface for genetic association analysis. BMC Bioinformatics 9, 557. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Li Z., Zhang Z., He Z., Tang W., Li T., Zeng Z., He L., and Shi Y. (2009). A partition-ligation-combination-subdivision EM algorithm for haplotype inference with multiallelic markers: update of the SHEsis (http://analysis.bio-x.cn). Cell Res. 19, 519–523 [DOI] [PubMed]
- 20.Carter K.W., McCaskie P.A., and Palmer L.J. (2006). JLIN: a java based linkage disequilibrium plotter. BMC Bioinformatics 7, 60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Dardiotis E., Jagiella J., Xiromerisiou G., Dardioti M., Vogiatzi C., Urbanik A., Paterakis K., Komnos A., Fountas K.N., Slowik A. and Hadjigeorgiou G.M. (2011). Angiotensin-converting enzyme tag single nucleotide polymorphisms in patients with intracerebral hemorrhage. Pharmacogenet. Genomics 21, 136–141 [DOI] [PubMed] [Google Scholar]
- 22.Balding D.J. (2006). A tutorial on statistical methods for population association studies. Nat. Rev. Genet. 7, 781–791 [DOI] [PubMed] [Google Scholar]
- 23.Klatzo I. (1994). Evolution of brain edema concepts. Acta Neurochir. Suppl. (Wien) 60, 3–6 [DOI] [PubMed] [Google Scholar]
- 24.Liang D., Bhatta S., Gerzanich V., and Simard J.M. (2007). Cytotoxic edema: mechanisms of pathological cell swelling. Neurosurg. Focus 22, E2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Papadopoulos M.C., Manley G.T., Krishna S., and Verkman A.S. (2004). Aquaporin-4 facilitates reabsorption of excess fluid in vasogenic brain edema. FASEB J 18, 1291–1293 [DOI] [PubMed] [Google Scholar]
- 26.Zador Z., Stiver S., Wang V., and Manley G.T. (2009). Role of aquaporin-4 in cerebral edema and stroke. Handb. Exp. Pharmacol. (1190), 159–170 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yang B., Zador Z., and Verkman A.S. (2008). Glial cell aquaporin-4 overexpression in transgenic mice accelerates cytotoxic brain swelling. J. Biol. Chem. 283, 15280–15286 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ban M., Walton A., Goris A., Gray J., Compston A., and Sawcer S. (2007). Polymorphisms in the neuromyelitis optica auto-antigen AQP4 and susceptibility to multiple sclerosis. J. Neurol. 254, 398–399 [DOI] [PubMed] [Google Scholar]
- 29.Park T.J., Kim J.H., Kim H.J., Bae J.S., Cheong H.S., Park B.L., and Shin H.D. (2013). Lack of association between AQP4 polymorphisms and risk of inflammatory demyelinating disease in a Korean population. Gene 536, 302–307 [DOI] [PubMed] [Google Scholar]
- 30.Kerty E., Heuser K., Indahl U.G., Berg P.R., Nakken S., Lien S., Omholt S.W., Ottersen O.P., and Nagelhus E.A. (2013). Is the brain water channel aquaporin-4 a pathogenetic factor in idiopathic intracranial hypertension? Results from a combined clinical and genetic study in a Norwegian cohort. Acta Ophthalmol. 91, 88–91 [DOI] [PubMed] [Google Scholar]
- 31.Li W., Hu B., Li G.L., Zhao X.Q., Xin B.Z., Lin J.X., Shen Y., Liang X.H., Liu G.F., Gao H.Q., Liao X.L., Liang Z.G., and Wang Y.J. (2012). Heterozygote genotypes at rs2222823 and rs2811712 SNP loci are associated with cerebral small vessel disease in Han Chinese population. CNS Neurosci. Ther. 18, 558–565 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Rubino E., Rainero I., Vaula G., Crasto F., Gravante E., Negro E., Brega F., Gallone S., and Pinessi L. (2009). Investigating the genetic role of aquaporin4 gene in migraine. J. Headache Pain 10, 111–114 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Kleffner I., Bungeroth M., Schiffbauer H., Schabitz W.R., Ringelstein E.B., and Kuhlenbaumer G. (2008). The role of aquaporin-4 polymorphisms in the development of brain edema after middle cerebral artery occlusion. Stroke 39, 1333–1335 [DOI] [PubMed] [Google Scholar]
- 34.Mai W., Hu X., Lu Z., Qiu W., Peng F., and Wang Y. (2013). Preliminary study on the association of AQP4 promoter polymorphism with anti-aquaporin-4 antibody positivity in southern Han Chinese patients with idiopathic demyelinating disorders of central nervous system. J. Neuroimmunol 255, 75–80 [DOI] [PubMed] [Google Scholar]
- 35.Opdal S.H., Vege A., Stray-Pedersen A., and Rognum T.O. (2010). Aquaporin-4 gene variation and sudden infant death syndrome. Pediatr. Res. 68, 48–51 [DOI] [PubMed] [Google Scholar]
- 36.Umenishi F., and Verkman A.S. (1998). Isolation and functional analysis of alternative promoters in the human aquaporin-4 water channel gene. Genomics 50, 373–377 [DOI] [PubMed] [Google Scholar]
- 37.Saadoun S., Bell B.A., Verkman A.S., and Papadopoulos M.C. (2008). Greatly improved neurological outcome after spinal cord compression injury in AQP4-deficient mice. Brain 131, 1087–1098 [DOI] [PubMed] [Google Scholar]
- 38.Thrane A.S., Rappold P.M., Fujita T., Torres A., Bekar L.K., Takano T., Peng W., Wang F., Rangroo Thrane V., Enger R., Haj-Yasein N.N., Skare O., Holen T., Klungland A., Ottersen O.P., Nedergaard M., and Nagelhus E.A. (2011). Critical role of aquaporin-4 (AQP4) in astrocytic Ca2+ signaling events elicited by cerebral edema. Proc. Natl. Acad. Sci. U. S. A. 108, 846–851 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Robertson C.S., Gopinath S.P., Valadka A.B., Van M., Swank P.R., and Goodman J.C. (2011). Variants of the endothelial nitric oxide gene and cerebral blood flow after severe traumatic brain injury. J. Neurotrauma 28, 727–737 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Zhou J., Kong H., Hua X., Xiao M., Ding J., and Hu G. (2008). Altered blood-brain barrier integrity in adult aquaporin-4 knockout mice. Neuroreport 19, 1–5 [DOI] [PubMed] [Google Scholar]
- 41.Bird T.D., Jarvik G.P., and Wood N.W. (2001). Genetic association studies: genes in search of diseases. Neurology 57, 1153–1154 [DOI] [PubMed] [Google Scholar]
- 42.Wilson J.T. (2001). Assessing outcome in head injury trials. Curr. Pharm. Des. 7, 1537–1552 [DOI] [PubMed] [Google Scholar]
- 43.Verkman A.S., Anderson M.O., and Papadopoulos M.C. (2014). Aquaporins: important but elusive drug targets. Nat. Rev. Drug Discov. 13, 259–277 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.