Malaria is still one of the most important infectious diseases, causing 249 million cases and 608,000 deaths in 2022. Nearly 95% of malaria cases and deaths occurred in sub-Saharan Africa; 76% of total deaths were reported in under-5 children (WHO, 2022; Venkatesan, 2024). In 2022, more than 50% of all deaths occurred in four African countries including Nigeria (31%), the Democratic Republic of the Congo (12%), Niger (6%), and Tanzania (4%) (Venkatesan, 2024). In countries such as Ghana, Kenya, and Malawi where the first malaria vaccine RTS,S/AS01 has been implemented, reductions in severe malaria cases and a 13% decrease in early childhood deaths have been reported (Venkatesan, 2024).

P. falciparum infection is associated with various clinical outcomes ranging from asymptomatic parasitaemia and uncomplicated malaria to severe malaria (SM) (Miller et al., 2002). The major complications of SM include cerebral malaria, severe anaemia, respiratory distress, pulmonary oedema and acute renal failure (Miller et al., 2002). In malaria-endemic areas, only a subset of cases progress to severe malaria and death (Mackinnon et al., 2005). The variations in malaria clinical outcomes are partly attributed to host genetic factors with estimated narrow-sense heritability of 20%–25% (Mackinnon et al., 2005; Damena and Chimusa, 2020).

Identifying genetic variations associated with clinical presentation during malaria infection may contribute to a better understanding of molecular mechanisms associated with host-pathogen interactions, which influence susceptibility to and protection against the disease (Achidi et al., 2008; Alamad et al., 2024). A classic example is the observation that African populations lacking Duffy blood group antigen were protected against P. vivax infection, because Duffy antigen is required for P. vivax to invade erythrocytes (Miller et al., 1976). This lack of Duffy antigen expression on erythrocyte surface is now known to be caused by a regulatory SNP in the Duffy blood group, within the chemokine receptor (DARC) gene that is near fixation in sub-Saharan Africa but absent in non-African populations (Hamblin and Di, 2000).

In continuing efforts to better understand P. falciparum pathogenesis, several genome-wide association studies (GWASs) have been conducted in diverse malaria-endemic populations, primarily by the MalariaGEN consortium (Timmann et al., 2012; Band et al., 2015; Ravenhall et al., 2018; MalariaGEN, 2019). GWASs have replicated the known protective loci including sickle cell (HBB) and ABO blood group, and identified new variants in ATP2B4 and Glycophorin regions. However, the cumulative heritability attributable to these loci constitute about 10% (MalariaGEN, 2019), suggesting additional genetic variations that influence malaria disease severity remained to be discovered.

Although GWASs have elucidated the genetic basis of susceptibility or resistance to SM, the method suffers from limitations including weak performance in genetically diverse African populations, lack of translation of associated loci into suitable biological hypotheses, and the well-known problem of missing heritability (Visscher et al., 2017; Damena et al., 2019). To address some of these challenges, we recently applied various computational methods to SM GWAS summary statistics datasets (N = 17,000) (MalariaGEN, 2019) and predicted plausible candidate genes (N = 182) with their respective biological pathways (Damena et al., 2021). However, these genes were mainly prioritized by in silico functional analysis based on GWAS-summary statistics and were not supported by clinical observations in affected communities. Here, we investigate the association of human candidate gene polymorphisms with the likelihood of having a severe malaria episode in a cohort of Malian children that were followed from birth for up to 5 years.

In this study, we systematically selected plausible novel candidate genes from our previous study (Damena et al., 2021) and performed candidate gene analysis in a cohort of Malian children. As African populations have high genetic diversity, severe malaria susceptibility variants have been shown to differ in allele frequencies, LD structure and effect sizes in sub-populations within endemic areas (Teo et al., 2010; Esoh et al., 2023). For instance, HbS, (rs334) is present at different frequencies in sub-Saharan Africa. It is prevalent at a frequency of ∼15% in northern Angola, Gabon and in southwestern Nigeria and east of Lake Victoria (Piel et al., 2013a). Frequencies of 7.5%–12.5% were predicted in large areas in West Africa extending from southern Senegal to northern Liberia, and from southern Ghana to northern Zambia. While HbC allele is largely absent in the South and Horn of Africa (Piel et al., 2013a), it is common in some parts of West Africa such as Mali, Burkina Faso, Ghana, Togo and Benin, though absent in other West African countries such as Cameroon and Chad, and East Africa (Piel et al., 2013b). This underscores the need for population-specific studies to capture severe malaria susceptibility variants that might be unique to an individual population.

In the current study, we identified a novel association between an intron SNP, rs13340578 in CSMD1 gene and increased odds of severe malaria. The SNP is distant from the previously reported SNPs located in the same gene in Tanzanian populations (Ravenhall et al., 2018). Allelic heterogeneity in malaria-endemic regions has been well described for genes encoding red blood cell proteins such as β globin that influence susceptibility to SM (Bauduer, 2013) The genomic region containing rs13340578 is composed of multiple SNPs with regulatory features. This might suggest the association is driven by linked SNPs or group of SNPs which might alter the expression of the CSMD1 gene. The MAF of these SNPs is generally higher in African populations, particularly in Mali. This may suggest the greater impact of these SNPs on health and susceptibility to diseases in the study population. The common-disease common-variant (CDCV) hypothesis depicts that common traits are most likely influenced by common variants with small to modest effects on diseases that may have escaped selection pressure (Risch and Merikangas, 1996). The differences in allele frequency in different populations is due to various demographic and evolutionary events in different parts of the world at different time points in history (Choudhury et al., 2014).

CSMD1 is a multiple domain gene consisting of 71 exons and spans a 2 MB DNA region on the short arm of chromosome 8 (8p23.2) (Sun et al., 2001). The gene is composed of 14 N-terminal CUB domains that are separated from each other by a short consensus repeat (SCR) followed by 15 tandem SCR domains, a transmembrane domain, and a short cytoplasmic tail (Kraus et al., 2006). CSMD1 is predominantly expressed in epithelial tissues and the central nervous system (CNS) (Uhlén et al., 1979). The gene is known to be an important regulator of complement activation and inflammation in the CNS (Kraus et al., 2006; Gialeli et al., 2018).

Complement is a system of plasma proteins that constitute a major component of the innate immune systems. Activation of complement leads to proteolytic cascades, which results in opsonization and lysis of the pathogen as well as in the generation of the classical inflammatory response through the production of potent proinflammatory molecules (Dunkelberger and Song, 2010; West et al., 2024). Complement can be activated by the classical, alternative, and mannose-binding lectin (MBL) pathways as described elsewhere (Rathnayake et al., 2021). All three complement pathways were reported to be activated during malaria infection by recognition of parasite and parasite-driven proteins in the host (Rathnayake et al., 2021). However, the malaria parasite has been known to escape host complement attack as a survival strategy (Rathnayake et al., 2021).

Apart from its protective role, activation of the complement pathway can also cause excess inflammation and extensive damage to self-tissues (Dunkelberger and Song, 2010). To prevent such damages, complement activation is tightly regulated by soluble and membrane-bound complement regulatory proteins (CRPs) at different points. CSMD1 inhibits the classical and lectin pathways of complement by promoting enzymatic cleavage of the activated C4b and C3b. Fragments one and two of CSMD1 bind C4 and C3 and facilitate their degradation by Factor I (Kraus et al., 2006). Thus, downregulation of this gene leads to elevated complement activation and associated pathology. Several studies have shown that variants and mutations in CSMD1 genes are linked to different pathological conditions including susceptibility to neurodegenerative diseases, psychiatric disorders, infertility, and cancer (Liu et al., 2019). In schizophrenia sufferers, decreased CSMD1 gene expression and its proteins predicted psychosis (Abd El Gayed et al., 2021) while increased complement (C4) expression predicted worse clinical outcomes after first psychosis (Mondelli et al., 2020).

The findings in the current study suggest that increased malaria disease severity might be due to dysregulation of complement. We propose the following model in which functional SNPs in the Sushi domain of CSMD1 gene might downregulate gene expression in the brain, which in turn can lead to elevated complement levels. The resulting hypercomplementemia may mediate excess inflammatory reactions and formation of membrane attack complex. Previous studies implicated complement component, C5a, as a driver of cerebral malaria pathogenesis (Patel et al., 2008) and foetal growth restriction due to placental malaria (Conroy et al., 2013). Further studies are needed to validate this hypothesis.

SM is a complex disease with different clinical presentations including cerebral malaria, severe malarial anaemia, and others which may arise from distinct pathophysiological processes. Sub-phenotype analyses were not conducted in this study due to lack of adequate power. Detecting an association with a particular SM phenotype may highlight whether complement pathway is involved with a specific SM syndrome. Additionally, haplotype association analysis was not investigated for the identified putative regulatory SNPs in the significant locus. Furthermore, causal variants and the underlying molecular mechanisms have not been elucidated. In conclusion, CSMD1 variants that modify regulation of complement may contribute to malaria disease severity. Further studies are needed to identify causal variants in this locus and the molecular mechanisms that mediate SM.