Introduction

Cerebral malaria (CM) is a neurological syndrome caused by Plasmodium falciparum, and its mechanism of pathogenesis is tightly linked to the adhesion and sequestration of parasitised erythrocytes in the brain microvasculature [1]. Vascular-related pathological events such as brain endothelial activation and dysfunction [2], infiltration of activated CD8+ T cells into the brain microvasculature [3] and disruption of the blood–brain barrier (BBB) [4] are thought to occur secondary to the sequestration of infected erythrocytes. However, since these pathological mechanisms are vascular-related, the exact molecular processes occurring in the brain tissues that cause neuronal dysfunction in CM are poorly understood. Although most CM patients fully recover, 15–20% die, and greater than 20% of the survivors develop long-term post-discharge neurodevelopmental sequelae, including hemiplegia, aphasia, cortical blindness and ataxia [5], indicating a similarity between CM and neurodegeneration. Here, we explore available literature and hypothesise that CM might share a common pathological pathway with Alzheimer’s disease (AD) despite the former being largely remediable.

APOE Modulates Amyloidosis in AD in a Gene Dose-Dependent Manner

AD is a neurological disorder caused by age-associated changes in the brain, alongside environmental and genetic factors. Some age-associated changes that could lead to AD include brain atrophy [6], vascular damage [7, 8], neuroinflammation [9], production of free radicals leading to excessive peroxidation of lipids [10] and dysregulation of energy metabolism in brain cells leading to energy deficits and neuronal dysfunction [11]. Some AD cases are caused by the accumulation and deposition of amyloid-β (Aβ) peptides in the brain, a process termed amyloidosis [12], followed by hyperphosphorylation of tau to form toxic neurofibrillary tangles, a process termed tauopathy [13]. However, some studies also report that tauopathy can occur independently of amyloidosis. The Aβ peptides are derived from the proteolysis of the neuronal amyloid precursor protein (APP) [12]. A cholesterol-transporting protein called apolipoprotein E (APOE), secreted primarily by astrocytes, plays a central role in AD by modulating either the transcription of APP [14] or the clearance of Aβ [15]. APOE and APP levels are increased in the cerebral spinal fluid (CSF) of AD patients [16,17,18] and in the brain tissues of transgenic mouse models of Aβ amyloidosis [19,20,21,22]. In addition, accumulating studies in mouse models of AD suggest that the absence of APOE, either genetically or pharmacologically, dramatically decreases Aβ amyloidosis in a gene dose-dependent manner [19,20,21,22,23,24]. APOE is expressed in three isoforms, APOE2, APOE3 and APOE4, at a single gene locus, with APOE4 being the leading genetic risk factor for AD [25] and the one associated with the highest levels of Aβ synthesis and deposition [14, 15]. Although the APOE4 variant is widely believed to contribute to neurodegeneration by promoting the accumulation of APP and Aβ [14, 15, 19,20,21,22,23,24], APOE4 has also been associated with vascular dysfunction and BBB leakage [7, 8].

Amyloidosis-Related CSF Proteins are Associated with CM

The role of APOE in CM pathogenesis has been considered [26,27,28]. An observational study reported that children (< 5 years of age) with APOE4 isoforms have a higher risk of developing and dying from CM compared to those with other isoforms [28]. Additionally, a recent study observed that APOE−/− mice did not develop experimental CM (ECM), even at 70–80% peripheral parasitemia, had lower parasite sequestration in the brain, reduced disruption of the BBB and decreased infiltration of T cells into the brain [27]. Another mouse study used immunochemistry to show that Aβ accumulates in the brain tissues of ECM-sensitive mice infected with P. berghei but not in ECM-resistant mice [29]. The APP protein was upregulated in the brain sections of patients who died from CM relative to those with no clinical cerebral pathology [30]. Also, increased CSF levels of tau were associated with long-term neurological and cognitive deficits in CM patients [31], and anti-tau immunotherapy prevented parasite-induced cognitive impairment and was associated with significantly reduced neuroinflammation and vascular congestion in ECM [32]. While primarily searching for biomarkers of acute bacterial meningitis (ABM) relative to CM, Njunge and colleagues used mass spectrometry to compare the CSF protein profiles of paediatric admissions with either of the two infections [33]. They observed that APOE was among the most upregulated CSF proteins in CM [33]. We reanalysed Njunge et al. proteomic data [33] and observed that in addition to APOE, the CSF levels of APP and other AD-related proteins such as NPTX1, PRNP, NCAM1, SPARC, AGT and IGF2 were also significantly elevated in CM compared to ABM (Fig. 1A, B; Supplementary File Table 1). We also noticed that APOE was significantly positively correlated to APP in CM (R = 0.5, p = 0.02) but not in ABM (R = 0.18, p = 0.32) (Fig. 1C), which was concordant with the gene dose-dependent amyloidogenic effect of APOE observed previously in AD [19,20,21,22,23,24]. When we performed disease ontology enrichment analysis [34], we found that genes upregulated in CM were strongly associated with AD-related terms such as “amyloidosis”, “Alzheimer’s disease” and “tauopathies” (Fig. 1D). Next, we obtained a list of 162 CSF proteins reported in at least two studies to be significantly altered in AD patients compared to healthy controls [35] and overlapped them with proteins that we found significantly enriched in CSF from CM patients compared to ABM (Supplementary File Table 2). This analysis showed that 68 of the 168 AD-altered CSF proteins, including APOE and APP, overlapped with those enriched in CM compared to ABM, while only 12 overlapped with those enriched in ABM (Fig. 1E, Supplementary File Table 2). These observations and data suggest that AD-linked proteins are strongly associated with CM.

Fig. 1
figure 1

The cerebral spinal fluid (CSF) levels of amyloid-associated proteins are co-increased in cerebral malaria. A Volcano plot showing that the CSF protein levels of apolipoprotein E (APOE) and amyloid precursor protein (APP) are elevated in cerebral malaria (CM) compared to acute bacterial meningitis (ABM) at log2Fold-Change > 2 and false discovery rate (FDR)-adjusted p-value < 0.05. Spectrum searches were performed using Maxquant version 2.0.3, and moderated t.test in limma R package was used to perform the statistical analysis. The colours depict the following: blue, proteins downregulated in CM; grey, not significantly altered proteins; and red, proteins upregulated in CM. Consistent with what was observed previously [33], lipocalin 2 (LCN2), myeloperoxidase (MPO) and lactotransferrin (LTF) have lower CSF protein levels in CM compared to ABM. B Boxplots comparing the CSF protein abundance of APOE and APP between CM (n = 22) and ABM (n = 33). The centre lines show the medians; limits show the median ± interquartile range; whiskers show values 1.5 times above or below the 75th and 25th percentiles, respectively; and each point represents a sample. C Scatterplot showing that CSF protein levels of APOE (x-axis) and APP (y-axis) were significantly positively correlated in CM but not in ABM. The Spearman’s rank correlation test was used to calculate the correlation statistics. D Enrichment analysis [34] showing the top four disease ontology terms associated with CM compared to ABM. Colour depicts − log10 (p-value) while numbers show the enrichment score. E Overlap of proteins enriched in CM and ABM [33] with AD-altered proteins from a previous meta-analysis study [35]

Full size image

Concluding Remarks

We have hypothesised and provided partial evidence that druggable AD signatures are augmented in CM. This information might benefit the search for treatment and management of both AD and CM. For instance, amyloidosis-targeted and anti-APOE therapies could be explored to prevent post-infection neurodevelopmental sequelae observed in some children after CM recovery [5] (Fig. 2). Furthermore, the general reversibility of CM might provide insights into how to reverse or prevent the progression of AD. However, additional scientific data will be required to support the amyloidosis hypothesis of CM pathogenesis. First, a multicentre large CSF proteomic study would be required to confirm the co-upregulation of APOE and APP in CM patients relative to age-matched patients with non-plasmodial encephalopathies and healthy community controls. In addition to proteomics, we propose an integrated analysis of the CSF transcriptomes, including mRNA, lncRNA and miRNA [36], to interrogate the similarities between CM and AD further. In vivo amyloid imaging using positron emission tomography could confirm or exclude the accumulation of Aβ and tau in antemortem cases of CM. A large human study to confirm whether the APOE4 isoform is associated with CM would also be highly informative. In summary, we hypothesise that amyloidosis might be a common pathophysiological process underlying the neurological disorders of CM and AD, most likely initiated by parasite-induced co-upregulation of APOE and APP in the cerebral tissues.

Fig. 2
figure 2

The APOE-APP hypothesis of cerebral malaria. Parasite sequestration in the brain capillaries can trigger the upregulation of APOE and APP, accumulation of amyloid-beta peptides and phosphorylation of neuronal tau. The APOE4 isoform can cause alteration of the blood–brain barrier (BBB) and neuroinflammation. Similarly, brain hypoperfusion may trigger dysregulated vasodilation and breakage of BBB, leading to leakage of vascular material into the brain parenchyma, resulting in further neuroinflammation and dysfunction of neurons. Therapeutic targets that block APOE or prevent the build-up of beta-amyloid peptides and phosphorylated tau could be explored as interventions for post-developmental sequelae observed in cerebral malaria survivors

Full size image