|Home | About | Journals | Submit | Contact Us | Français|
Silent cerebral infarct (SCI) is the most commonly recognized cause of neurological injury in sickle cell anaemia (SCA). We tested the hypothesis that magnetic resonance angiography (MRA)-defined vasculopathy is associated with SCI. Furthermore, we examined genetic variations in glucose-6-phosphate dehydrogenase (G6PD) and HBA (α-globin) genes to determine their association with intracranial vasculopathy in children with SCA. Magnetic resonance imaging (MRI) of the brain and MRA of the cerebral vasculature were available in 516 paediatric patients with SCA, enrolled in the Silent Infarct Transfusion (SIT) Trial. All patients were screened for G6PD mutations and HBA deletions. SCI were present in 41.5%(214 of 516) of SIT Trial children. The frequency of intracranial vasculopathy with and without SCI was 15.9% and 6.3%, respectively (p<0.001). Using a multivariate logistic regression model, only the presence of a SCI was associated with increased odds of vasculopathy (p=0.0007, odds ratio (OR) 2.84; 95% Confidence Interval (CI)=1.55-5.21). Among male patients with SCA, G6PD status was associated with vasculopathy (p=0.04, OR 2.78; 95% CI=1.04-7.42), while no significant association was noted for HBA deletions. Intracranial vasculopathy was observed in a minority of children with SCA, and when present, was associated with G6PD status in males and SCI.
Silent cerebral infarct (SCI) is the most commonly recognized cause of neurological injury in sickle cell anaemia (SCA) (Miller, et al 2001) and occurs in 37 % of the paediatric SCA population prior to their 14th birthday (Bernaudin, et al 2011). There is limited data regarding primary and secondary prevention of SCI in the paediatric population. Although children with SCI have normal neurological examination (Glauser, et al 1995), they have significant cognitive impairment that affects academic performance(Armstrong, et al 1996, Kral, et al 2001, Schatz, et al 2001). Children with SCI are also at an increased risk for future strokes (Miller, et al 2001). Children with overt strokes receiving blood transfusion therapy, most of whom have moyamoya, have an ongoing risk of both overt strokes and SCI (Hurlet-Jensen, et al 1994).
G6PD deficiency, an X-linked genetic disorder, is the most common human enzyme defect (Beutler 1996) afflicting about 10% of African Americans and 10-25% of the populations where malaria is prevalent (Ruwende, et al 1995, Spolarics, et al 2001). Oxidative injury is thought to play a role in the development of vasculopathy seen in sickle cell disease (SCD). Sickling-induced oxidative stress occurs through formation of reactive oxygen species (ROS) that can be inactivated by reduced glutathione (GSH), due to activation of the pentose phosphate pathway that is mediated by glucose-6-phosphate dehydrogenase (G6PD) (Salvemini, et al 1999). The strength of the association between G6PD and SCD-related morbidity is controversial. Although G6PD deficiency has been reported to increase risk for stenosis as defined by magnetic resonance angiography (MRA) or abnormal cerebral velocity measured by Transcranial Doppler (TCD) in children with SCA (Bernaudin, et al 2011, Bernaudin, et al 2008), other studies reported no association between G6PD status and abnormal TCD evaluation (Miller, et al 2011, Rees, et al 2009).
HBA (α-globin gene) deletion has been associated with normal TCD measurement and with the absence of vasculopathy on MRA images in patients with SCA, suggesting that α-thalassaemia may play a protective role against cerebral vasculopathy (Belisario, et al 2010, Bernaudin, et al 2008, Hsu, et al 2003). However, in a small case control study, HBA deletion was found not to be significantly associated with reduced risk for specific neurological events (transient ischaemic attacks, ischaemic strokes or haemorrhagic strokes) (Filho, et al 2011). Taken together, studies evaluating the association between either G6PD or HBA status on cerebral vasculopathy have had mixed results, possibly due in part to the small sample size and in some cases the inherent biases associated with single institution studies.
The Silent Infarct Transfusion (SIT) Trial is a multicentre study which is investigating whether prophylactic blood transfusion in children with SCI will prevent and delay the progression of cerebral infarcts (Casella, et al 2010). The trial represents the largest cross-sectional study of children with SCA that have undergone both a magnetic resonance imaging (MRI) and MRA of the brain, with pre-defined central adjudication of infarcts and cerebral vasculopathy, respectively. As a planned secondary analysis, we tested the primary hypothesis that SCI is associated with MRA-defined intracranial vasculopathy. To clarify the roles of G6PD and HBA deletions, we also tested whether genetic variations associated with G6PD deficiency or α-thalassaemia would alter the prevalence of MRA vasculopathy.
The SIT Trial was approved by the Institutional Review Boards of all participating institutions and is registered at www.clinicaltrials.gov (NCT00072761) and www.ISRCTN.org (ISRCTN52713285). The details of the study design were described by Casella et al (2010). A subset of SCA patients enrolled in the Silent Cerebral Infarct Multicenter Transfusion trial studies was analysed (Casella, et al 2010). These patients were enrolled at 29 clinical sites across North America and Europe between 2004 and 2010 (age 5-14 years with haemoglobin genotypes of SS or Sβ0–thalassaemia). In this study, 516 children had both MRI and MRA studies, of whom 422 had DNA available for genetic analyses. The G6PD gene is X-linked, therefore, all G6PD analyses were further restricted to males only and thus the final sample for the genetic analyses included 208 males.
The neuroimaging protocol and definition of SCI were described by Vendt, et al (2009). Briefly, SCI is defined by an area of abnormal MRI signal intensity on fluid attenuated inversion recovery (FLAIR) T2-weighted images of the brain in a child with no prior history or physical findings of a focal neurological deficit as ascertained by physical examination. MRAs were not required for enrollment in the SIT Trial, but were commonly performed, and were collected as part of the trial, when available. Each MRA was assessed for intracranial vasculopathy by two investigators and discordant reads were reviewed by a senior neuroradiologist. For MRA studies, vessels examined were the terminal portions of the internal carotid arteries (cavernous and supraclinoid segments) andthe first segments of the anterior, middle, and posterior cerebral arteries bilaterally; each vessel was graded as normal, or at least moderate stenosis > 50% or occlusion, similar to Hulbert et al (2011) and as previously described (Lell, et al 2007) using vascular boundaries based on neuro-anatomical landmarks (Ropper, et al 2009). TCD velocities were obtained as part of standard care (Adams, et al 1998). TCD velocities were required for children with SCI prior to randomization; those without SCI were ineligible to continue in the SIT trial and TCDs were not required for these patients. Children with known elevated TCD measurements were excluded from enrolling in the SIT Trial, whereas those with abnormal MRAs were not excluded. Thus, we elected not to include the TCD evaluations in this analysis for assessment of vasculopathy, as the resulting subgroup was non-representative of the trial participants. Systolic blood pressure (SBP) was measured at the most recent wellness visit.
Genomic DNA was purified from Epstein-Barr virus-transformed lymphoblasts established from whole blood lymphocytes of the SCA patients enrolled in the SIT trial (Casella, et al 2010). Patients were screened for nine G6PD genetic variants associated with decreased enzyme function commonly found in African Americans, Caribbean, Mediterranean and European populations (Beutler 1994, Mason, et al 2007). These variants are spread across exons 3 to 9 of the G6PD gene. Genotype was determined using validated TaqMan® SNP Genotyping Assay following the manufacturer's instructions (Applied Biosystems, Foster City, CA). At least 10% of the samples were subjected to DNA sequencing to confirm the TaqMan® results using BigDye® Terminator v1.1 Cycle Sequencing Kits (Applied Biosystems). Additionally, direct DNA sequencing was performed to screen G6PD exons 10 and 11 that are clustered with point mutations that cause a severe class I phenotype. The TaqMan® results were analysed using the TaqMan® Genotyper v1.0 software (Applied Biosystems). The DNA sequencing data were analysed using SeqScape v2.5 software (Applied Biosystems).
Multiplex Ligation-Dependent Probe Amplification (MLPA) and two-tube multiplex polymerase chain reaction (PCR) assays modified from previous studies were used for detection of HBA deletions (de Mare, et al 2010, Liu, et al 2000). MLPA was carried out using the SALSA MLPA kit P140 HBA (MRC-Holland, Amsterdam, Netherlands) according to manufacturer's instructions. Products of the MLPA reaction were analysed on the 3730 Genetic Analyser (Applied Biosystems) using GeneScan™ 500 LIZ® Size Standard (Applied Biosystems). Data were analysed using GeneMapper® 4.0 (Applied Biosystems) and Coffalyser (MRC-Holland). PCR was carried out using two sets of primers. The first set of PCR primers (Set 1) were used to detect seven HBA mutations (-α3.7, -α4.2, --SEA, --FIL, --MED, -(α)20.5, and --THAI). The second set of PCR primers (Set 2) were used to discriminate -α3.7 trans-deletion. PCR was carried out in 25-μl reactions containing 200 μM of each dNTP, 1.5 mM MgCl2, 2.5 units HotStar Taq DNA polymerase with 1X Q-solution (Qiagen, Valencia, CA, USA), 200 ng of genomic DNA and the first or second set of primers at various concentrations as described previously (de Mare, et al 2010, Liu, et al 2000). Amplification was performed with an initial heat activation step of 15 min at 96°C, followed by 30 cycles of 98°C for 45 s, 60°C (Set 1 primers) or 65°C (Set 2 primers) for 90 s, 72°C for 150 s, and a 5 min final extension step at 72°C. PCR bands were separated and sized on 1% agarose gel.
Differences in the distribution of relevant demographic and clinical factors among children with and without vasculopathy were assessed using two-tailed chi-square tests (sex and silent cerebral infarct), t-tests (age, steady state haemoglobin, SBP, steady state white blood cell (WBC) count, baseline oxygen saturation), and negative binomial regression (acute chest syndrome episodes requiring hospitalization and pain events requiring hospitalization). The logit transformation of WBC count was used to approximate normality. Multivariate logistic regression models were constructed with all covariates to determine clinical and demographic factors associated with vasculopathy. Four hundred and eighty-seven of 516 children (94%) with valid data on all possible covariates were included in the multivariate models. The first multivariate model used all covariates potentially associated with vasculopathy entered in one block (forced). A subsequent model included only those covariates that were nominally significant predictors (p<0.20) from the first model. A final model included covariates that were statistically significant at p<0.05.
Bivariate associations between vasculopathy and G6PD mutations, and vasculopathy and HBA deletion status were evaluated using two-tailed chi-square tests. Multivariate logistic models including all genetic, demographic, and clinically relevant covariates (G6PD mutations, HBA deletion status, silent cerebral infarct, age, steady state haemoglobin, SBP, steady state WBC count, baseline oxygen saturation, acute chest syndrome rate, and pain event rate) were constructed to examine factors believed to be independently associated with vasculopathy. Of the 208 males with genetic information, 192 (92%) had non-missing information for all covariates and were thus included in the logistic regression models. All potential covariates were included in the initial models and only those with p-values < 0.05 were retained in the final models.
A total of 516 subjects had undergone both MRI and MRA studies. These 516 children had a mean age of 9.1 years (range 5-15 years) and 51.4% were male. SCI were detected in 41.5% (214 of 516) of the patients. MRA-defined intracranial vasculopathy was detected in 10.3% (53 of 516) of the patients. The frequency of SCI in children with and without vasculopathy was 64.2% and 38.9%, respectively (p<0.0004). Demographic and clinical features of participants and imaging studies are summarized in Table I.
Of the 34 children with vasculopathy and SCI, a total of 164 infarcts were identified, with an average of 4.8 infarcts per child. About 75% of SCI (123 of 164) were located in the frontal lobe. Infarcts located in the vascular distribution of the middle cerebral artery were the most common. These lesions were more prevalent in the vascular border-zone between the anterior cerebral artery (ACA) and the middle cerebral artery (MCA) territories [43% (70 of 164)], frontal lobe supplied by the MCA alone [36% (59 of 164)], and posterior border-zone [9% (15 of 164)], which is also supplied in part by the MCA. In addition, 10% (17 of 164) of the lesions were in the ACA territory, 1% (2 of 164) of the lesions were in the posterior cerebral distribution territory, and 1% (1 of 164) of lesions were in the territory of anterior and posterior border-zone.
Vasculopathy was unilateral in 53% (18 of 34) of the participants with SCI. We did not observe a relationship between the side of vasculopathy and the side of SCI (p=0.194). The side of vasculopathy was contralateral to the location of SCI in 6% (2 of 34) of the participants. The internal carotid artery was affected in 94% (32 of 34) patients, the middle cerebral artery was affected in 29% (10 of 34), and the anterior cerebral artery was affected in 15% (5 of 34) patients. Vasculopathy was not detected in the posterior cerebral artery.
We also studied the clinical and radiological features that are associated with vasculopathy. Using a multivariate logistic regression model, we initially included all plausible biological factors (n=9) (Table II). We subsequently included only the covariates with p-values less than 0.2, and only four factors remained in the multivariate model—SCI, age, SBP and baseline oxygen saturation. In the final model, only the presence of a SCI was associated with increased odds of vasculopathy (p=0.0007; odds ratio (OR) 2.84, 95% CI=1.55-5.21) (Table II).
Participants with available DNA were screened for nine non-synonymous G6PD genetic variants (rs78478128, rs1050828, rs1050829, rs5030872, rs5030868, rs137852330, rs137852328, rs76723693, and the G6PD Volendam variant that has not been registered in the dbSNP database) that are found to be prevalent in African Americans, European, Caribbean and Mediterranean populations (Beutler 1994, Mason, et al 2007). Additionally, exons 10 and 11 were screened by direct sequencing, as several point mutations have been reported in this region that interrupt G6PD dimerization and cause a severe class I phenotype (Cappellini and Fiorelli 2008). Two single-nucleotide mutations rs1050828 and rs1050829 were identified in the SCA patients with a prevalence of 17.8% (n=37) and 40.9% (n=85), respectively. Six additional single-nucleotide substitutions were found in exons 10 and 11; however, these synonymous polymorphisms resulted in no predicted amino acid change and were not considered in genetic association tests. Finally, a two-nucleotide deletion (g.20101delTC) was identified in intron 10. GeneSplicer analysis showed intron splicing was not affected by the deletion, and thus this variation was also excluded from further analyses (http://www.cbcb.umd.edu/software/GeneSplicer).
In the bivariate analyses of males with SCA, the association between genetic variations in G6PD and vasculopathy was significant (p=0.02), however,HBA deletions were not associated with cerebral vasculopathy as determined by MRA (Table III). In the final multivariate logistic regression model, only the presence of any G6PD variation (p=0.04; OR 2.78, 95% CI=1.04-7.42) was associated significantly with vasculopathy in male participants (Table IV, Model 2).
This study demonstrated that, when present, MRA-defined vasculopathy was associated with the presence of SCI, however, based on our observation that approximately 84% of the children with SCI did not have evidence of MRA-defined vasculopathy, the finding of MRA-defined vasculopathy alone is insufficient to explain the presence of SCI among children with SCA. Although vasculopathy clearly is associated with SCI, multiple other lines of evidence strongly suggest that perturbations in cerebral haemodynamics, other than vasculopathy, are involved in the pathogenesis of SCI, including 1) the location of SCIs predominantly in border-zone areas, an infarct prone region susceptible to abnormal cerebral haemodynamics (present study); 2) the association of SCI with acute anaemia events (Dowling, et al 2010); and 3) the association of SCI with relative low haemoglobin levels and high baseline SBPs in children with SCA (Debaun, et al 2012). .
A perhaps unexpected finding was a lack of correlation between the side of vasculopathy and the side of SCI. However, in adults without SCD, it was demonstrated that the percent stenosis in the carotid artery was not a reliable indicator of oxygen extraction fraction (OEF), a measure of severe cerebral haemodynamic compromise placing the patient at risk for cerebral infarcts (Powers et al 1987). This same group demonstrated a lack of correlation between angiographic studies of patients with carotid occlusion and increased OEF (Derdeyn et al 1999). Our findings in this study are supportive of this previous literature suggesting that angiography-defined stenosis or occlusion is not necessarily associated with cerebral haemodynamic compromise on the same side. Another possibility is the presence of small vessel vasculopathy not detectable by MRA or rheological properties intrinsic to sickle cell adhesiveness in the microvasculature may be contributing factors (Cheung, et al 2002, Cheung, et al 2010). Future studies using positron emission tomography or MRI assessment of OEF or its corrollary are needed to determine the basis for SCI in SCD.
We also found that G6PD genetic variation (rs1050828 or rs1050829) was associated with vasculopathy in male participants with SCA. G6PD enzyme activity is decreased by about 30% when either variation is present (Town, et al 1992). The G6PD haplotype (rs1050828/rs1050829) shows synergistic effects on enzyme thermal instability (Town, et al 1992), but only five patients with vasculopathy carried both mutations and, thus, could not be analysed separately (data not shown).
G6PD enzyme is a major source of cytosolic NADPH and serves to maintain redox balance and synthesis of nitric oxide (NO) (Spolarics, et al 2001). Excessive oxidative stress and/or diminished bioavailability of NO in endothelial cells due to G6PD deficiency may in part contribute to vasculopathy, especially in SCA patients when NO level is low due to massive haemolysis (Abboud and Musallam 2009, Donadee, et al 2011). Supporting this postulate is the observation that increased oxidative stress and decreased NO levels have been observed in G6PD-deficient endothelial cells (Leopold, et al 2007). Endothelial dysfunction is a hallmark of vasculopathy that is involved in the development of vasoocculsive, ischaemia-reperfusion injury, haemostatic activation, and stroke in individuals with SCD (Hoppe 2011, Kato, et al 2009, Morris 2008). Given G6PD-deficient foreskin fibroblasts exhibit faster cell senescence and slower cell growth and proliferation than G6PD normal cells (Ho, et al 2005), we postulate that G6PD deficiency could impair endothelium turnover and may contribute to SCI and cerebral vasculopathy. The SIT Trial cohort did not include direct measurements of haemolysis; hence, we could not pursue whether there was a relationship between the magnitude of haemolysis, G6PD mutations and vasculopathy.
On a molecular level, oxidative stress induces redox-sensitive transcription factors, such as NF-κB and AP-1 (Karin, et al 2001), which in turn increase expression of the adhesion molecules E-selectin (Kyriakis and Avruch 2001), vascular cell adhesion molecule-1 (Kyriakis and Avruch 2001, Tu, et al 2001) and intercellular adhesion molecule-1 (Wang, et al 2007), and promote inflammation or lesions via stimulating cytokine secretion (Chen, et al 1998, Seo, et al 2002, Zachwieja, et al 2005) or by recruiting adherent leucocytes to endothelial cells (Forstermann 2008, Huet, et al 2011). This pathological process may be more severe in G6PD-deficient endothelial cells that do not have the maximum capacity of NO (Kupatt, et al 1997, Leopold, et al 2007, Marui, et al 1993). More extensive studies are needed to better define the mechanism of the association between G6PD and vasculopathy.
We can only postulate why HBA deletions were not significantly associated with MRA-defined vasculopathy in the present study. In previous SCD studies, α-thalassaemia was associated with lower TCD measurements and, hence, reduced risk for stroke (Bernaudin, et al 2008, Flanagan, et al 2011, Hsu, et al 2003). The most likely explanation for the lack of association is the observation that the SIT Trial cohort does not represent the full spectrum of children with SCA and neurological complications, elevated TCD measurements and stroke, respectively. Potential participants who were screened for the SIT Trial were excluded if they had overt strokes, previously elevated TCD measurement, or epilepsy. Alternatively, the lack of an association between HBA deletions and MRA-defined vasculopathy may reflect a true negative result. Although this is the largest study to date of this type, additional studies are needed to substantiate our findings in other populations. Despite these limitations, the results of the current study are generalizable to the vast majority of children with SCA, as we did not pre-select children with MRA-defined vasculopathy to participate in the trial.
In summary, most children with SCA do not have evidence of MRA-defined intracranial vasculopathy, but when vasculopathy is present, there is an approximate 2 times greater odds of SCI. Based on the results of this study, the utility of MRA for risk stratification of children with SCI cannot be determined. Only an extremely large prospective longitudinal study can assess the increased risk of progressive cerebral infarcts, if any, for children with MRA-defined cerebral vasculopathy and SCI. Also, an association between cerebral vasculopathy in male patients with SCA and genetic variations in G6PD linked to reduced enzymatic function has been established. Future investigation is warranted to determine more precisely the underlying mechanisms of SCI in this high-risk paediatric population.
The authors would like to thank the SIT Trial study coordinators, the families and children with sickle cell anaemia who were participants and laboratory members in both the Division of Blood Disorders (CDC) and Division of Pediatric Hematology (JHU).
Funding: The SIT Trial is supported by the National Institute of Neurological Disorders and Stroke (NINDS), Rockville, Maryland (ClinicalTrials.gov identifier: U01 NS042804).
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Author contributions: MT, DF, AJ, MRP and RCM interpreted neuroradiology studies and participated in the data analysis. JFC, RCM, BAV and MJN participated in the protocol design. MRD designed the study, participated in the protocol design, acquired data, and performed data analysis. GY, WCH and CJB contributed to designing and interpreting the genetic experiments. EBC directed the selection and processing of the DNA used in the study. GY, MEP, ABP, JD and HAT performed the DNA genotyping and sequencing assays. SLB and MR performed statistical analysis. GY, MT, JFC, CJB, SLB and MRD prepared the manuscript.
The authors declare no competing financial interests.