Whole Exome Sequencing in Children with Sickle Cell Anemia Who Developed Ischemic Stroke Prior to 4 Years of Age Reveals a Strong Association with Inherited Dyslipidemias

Ischemic stroke is a devastating complication of sickle cell anemia (SCA) and can affect children from a very young age. It is the commonest cause of stroke in the general pediatric population. The peak incidence is in the first decade of life and approaches 11% in the absence of stroke prevention programmes. The advent of transcranial doppler (TCD) screening as part of a primary stroke prevention programme has reduced the incidence of stroke by 90%. However, not all strokes are prevented by TCD, and TCD abnormalities are not specific, with up to 60% children receiving transfusions unnecessarily. Understanding the underlying pathophysiology of cerebrovasculopathy in SCA and recognition of important risk factors will lead to a more stringent identification of the at-risk population.

Analysis of siblings with stroke and abnormal TCDs offers good evidence that a proportion of this variability is conferred by inherited genetic variation. Identifying those specific variants has proved elusive, with many conflicting findings reported in the literature, no doubt reflecting the limited power of many studies to address the complex overlay of environmental and genetic influences. To attempt to minimise the impact of non-genetic confounders, our study looked specifically at children who had an overt ischemic stroke or developed abnormal TCD measurements (TAMMV >200cm/s) prior to their 4^th birthday, in the absence of potentially precipitating acute medical events.

We recruited 22 patients, 19 of whom had an ischemic stroke, and 3 who developed abnormal TCD measurements, prior to 4 yrs of age. Additionally, we recruited the dizygotic twin of one stroke case, who had SCA, but, importantly, no cerebrovascular complications, as confirmed by MRI/A aged 15 yrs. We extracted DNA from peripheral blood samples. Exome library was prepared using Agilent SureSelect XT V7, sequencing was performed using Illumina NovaSeq. Alignment, assembly, variant calling and annotation were based on a GATK Best Practices workflow. For each sample, around 15000 non-synonymous variants were identified. Across the 22 case samples, 58 variants in 56 genes were predicted to be pathogenic or likely pathogenic, as determined with InterVar. These variants were interpreted with respect to dbSNP documentation and biological role of the gene they affected. Variant segregation within the twin pair was also considered.

Two patients were homozygous for rs429358, a missense variant in the APOE gene that is pathogenic for familial hyperlipoproteinemia, type 3. Moreover, 3 further patients were compound heterozygous for this variant plus another APOE missense variant pathogenic for the same condition, rs7412. One of these 3 cases was the affected sibling of the twin pair and notably, the unaffected twin carried only heterozygous rs429358 and not rs7412. A further patient was heterozygous for rs121908043 in the LDL-R gene, which is pathogenic of Familial Hypercholesterolaemia, even in the heterozygous state. Finally, in an additional patient, we identified a potential compound heterozygote of a known pathogenic variant, rs118204068 with another missense variant rs11542065 in the LPL gene, to cause Hyperlipoproteinemia type 1.

We have used whole exome sequencing to analyse patients with sickle cell anemia and stroke at a very young age, an extreme phenotype, in whom we predicted genetic factors would be a significant cause of stroke. We found 7 of the 22 patients (32%) had variants diagnostic of inherited dysplipidaemias, including 5 with familial hyperlipoproteinemia type 3, 1 with hyperlipoproteinemia type 1 and one with familial hypercholesterolemia. Our analysis included one twin set and it is noteworthy that the unaffected twin sibling did not carry the necessary variants defining the inherited dyslipidaemia. These conditions are characterised by improper breakdown and accumulation of LDL-C, triglycerides and cholesterol, which has been associated with vascular dysfunction and hemolysis in SCA and are strongly associated with stroke and cardiovascular disease in the general population. This is potentially an important finding that warrants further investigation into the role of lipids and hyperlipidemia in the development of stroke and cerebral vasculopathy in the sickle cell population. In the long term, we suggest that measurement, and potentially, control of lipids may form a component of the primary stroke prevention programme.