Abstract
Introduction Malaria remains a major public health concern with substantial morbidity and mortality worldwide. In Malaysia, the emergence of Plasmodium knowlesi has led to a surge in zoonotic malaria cases and deaths in recent years. Signs of cerebral involvement have been observed in a non-comatose, fatal case of severe knowlesi infection, but the potential impact of this malaria species on the brain remains underexplored. To address this gap, we investigated circulating levels of brain injury, inflammation, and vascular biomarkers in a cohort of knowlesi-infected patients and controls.
Methods Archived plasma samples from 19 patients with confirmed symptomatic knowlesi infection and 19 healthy, age-matched controls from Peninsular Malaysia were analysed. A total of 52 plasma biomarkers of brain injury, inflammation, and vascular activation were measured using Luminex and SIMOA assays. Wilcoxon tests were used to examine group differences, and biomarker profiles were explored through hierarchical clustering heatmap analysis.
Results Bonferroni-corrected analyses revealed significantly elevated brain injury biomarker levels in knowlesi-infected patients, including S100B (p<0.0001), Tau (p=0.0007), UCH-L1 (p<0.0001), αSyn (p<0.0001), Park7 (p=0.0006), NRGN (p=0.0022), and TDP-43 (p=0.005). Compared to controls, levels were lower in the infected group for BDNF (p<0.0001), CaBD (p<0.0001), CNTN1 (p<0.0001), NCAM-1 (p<0.0001), GFAP (p=0.0013), and KLK6 (p=0.0126). Hierarchical clustering revealed distinct group profiles for circulating levels of brain injury and vascular activation biomarkers.
Conclusions Our findings highlight for the first time the impact of Plasmodium knowlesi infection on the brain, with distinct alterations in cerebral injury and endothelial activation biomarker profiles compared to healthy controls. Further studies are warranted to investigate the pathophysiology and clinical significance of these altered surrogate markers, through both neuroimaging and long-term neurocognitive assessments.
1 Introduction
Malaria is a life-threatening, vector-borne infection caused by parasites of the genus Plasmodium that led to an estimated 249 million cases worldwide and approximately 608,000 deaths in 2022 (1). Malaria is endemic in Southeast Asia, where it poses a significant public health challenge. In Malaysia, P. falciparum and P. vivax have historically been the species responsible for most malaria infections and deaths, but thanks to national efforts in eradicating the disease, no indigenous cases have been reported since 2018 (1). However, an emerging concern in the country is P. knowlesi (Pk), a zoonotic species that has become a significant contributor to malaria infections in humans locally (2,3), with a total of 19,625 Pk cases and 57 deaths reported in Malaysia since 2017, including 2,500 cases and 9 deaths in 2022 alone (1). Malaysia accounts for most Pk infections globally, and this species is currently the major cause of human malaria in the country (4–6).
One of the most severe manifestations of Plasmodium falciparum infection is cerebral malaria, a life-threatening neurological complication characterized by coma (7) and pathological hallmarks such as petechial hemorrhages and parasite sequestration in the brain vasculature (8,9). Neurocognitive sequelae are frequent in survivors and can persist long after the infection has been treated (10,11). While cerebral malaria is predominantly associated with P. falciparum (7), recent case reports have highlighted instances of cerebral involvement and severe neurological complications in patients infected with other Plasmodium species (12–18). A high proportion of severe Pk infections has been reported in Southeast Asia (5,19), as well as fatal cases (18–20). A recent study also showed that, similarly to P. falciparum, Pk-infected erythrocytes are able bind to endothelial cells (21). Post-mortem findings in one fatal case of severe Pk malaria revealed brain pathology features similar to those seen in fatal falciparum cerebral malaria, including Pk-infected erythrocytes sequestration in the microvasculature (18), suggesting that Pk malaria may also affect the brain. Remarkably, coma was not observed in this patient (18), which contrasts with the World Health Organization (WHO) definition of cerebral malaria caused by P. falciparum (7).
We recently reported that patients with severe and uncomplicated falciparum malaria —and therefore without coma— have a wide range of brain changes detected using magnetic resonance imaging (MRI) (22,23). Severe malaria patients from our cohort had elevated plasma levels of the neurotrophic factor S100 calcium-binding protein β (S100B), a biomarker associated with central nervous system insults (24,25), which correlated with brain MRI features typically found in cerebral malaria (22). Overall, our findings suggest that there is a frequent and unrecognized impact of malaria infection on the brain in falciparum malaria, irrespective of coma. These effects are likely aggravated by acute kidney injury (AKI), with potential long-lasting repercussions on quality of life and productivity in survivors (26).
Despite the similar pathologies reported between cerebral malaria and fatal Pk infection (18) and the high incidence of kidney dysfunction in severe Pk malaria (27), the potential impact of Pk infection on the human brain remains largely unexplored. To bridge this important knowledge gap, we quantified plasma biomarkers of brain injury, vascular activation, and inflammation in a cohort of Malaysian Pk-infected patients and healthy controls.
2 Methods
2.1 Participants and samples
This study leveraged archived plasma samples from patients aged 18 years and above with malaria symptoms attending either a government hospital or private clinic in Johor, Selangor, Pahang, Perak, and Trengganu states between December 2019 and January 2023 (Figure 1). Recruitment of community-matched, age-matched uninfected controls was conducted via active screening of individuals from communities in Johor, Selangor, Negeri Sembilan, and Kedah. Subjects were approached for participation if they had no fever, were aged 18 years and above, and were part of any of the considered high-risk groups. A high-risk group is defined as individuals working in proximity with forest and forest fringes, i.e., servicemen, farmers, hunters, and natural resource collectors (2,28).
Data collection procedures are described elsewhere (29). Briefly, samples were acquired using EDTA anti-coagulant tubes. Informed consent was obtained prior to sample collection. All participants were allowed sufficient time to consider their participation in the project. Individual data were collected and recorded anonymously. Plasmodium infection and speciation in each collected sample were confirmed by microscopic examination of Giemsa-stained blood smears and nested polymerase chain reaction (PCR) based on the 18S rRNA gene. Out of the 50 individuals sampled as part of the parent project, 38 had associated clinical information needed for this study, including parasitemia. The de-identified Malaysian samples, consisting of 19 infected patients and 19 healthy controls, were analysed.
2.2 Biomarker panel selection
A total of 52 plasma biomarkers were selected, including 21 markers of brain alterations or cerebral injury, 21 markers of infection and immune activation, and 10 vascular biomarkers. Our literature review-based selection followed the assumption that any plasma biomarker ever reported to be altered in any species of human malaria could potentially be altered in our cohort of Pk patients. For brain injury biomarkers, a second premise applied: since cerebral malaria can cause long-lasting cognitive disorders (10,11), any plasma marker reported to be increased or decreased in cognitively impaired subjects (e.g., neurodegenerative diseases, traumatic brain injury, etc) could also be altered in our patients.
To inform the biomarker panel selection, a systematic search of scientific literature on biomarker plasma levels in human patients was conducted in PubMed database, including all relevant publications from 2010 onwards. The results of this systematic search are comprehensively summarized in Table 1 (brain injury biomarkers), Supplementary Table 1 (infection and immune activation biomarkers) and Supplementary Table 2 (vascular biomarkers).
2.3 Measurement of plasma biomarker levels via Luminex and SIMOA assays
Thirty-eight biomarkers were analysed using the customizable Human Luminex® Discovery Assay provided by R&D Biotechne (catalogue ref.: LXSAHM): Alpha-Synuclein (αSyn), Amyloid-beta precursor protein (APP), Angiopoietin-1 (Ang-1), Angiopoietin-2 (Ang-2), Bone Morphogenetic Protein 9 (BMP-9), Calbindin D (CaBP), Chemokine Ligand 2 (CCL2), Chemokine Ligand 4 (CCL4), Chemokine Ligand 5 (CCL5), Chemokine Ligand 18 (CCL18), Contactin-1 (CNTN1), C-Reactive Protein (CRP), Enolase 2 / Neuron-specific Enolase (ENO2/NSE), Fetuin A, Granulocyte-Macrophage Colony-Stimulating Factor (GM-CSF), Interferon Gamma (IFN-γ), Intercellular Adhesion Molecule 1 (ICAM-1), Interleukin 1 alpha (IL-1 α), Interleukin 1 beta (IL-1 β), Interleukin 1RA (IL-1RA), Interleukin 2 (IL-2), Interleukin 4 (IL-4), Interleukin 6 (IL-6), Interleukin 8 (IL-8), Interleukin 10 (IL-10), Interleukin 17 (IL-17A), Lipocalin-2 (NGAL), Myeloperoxidase (MPO), Osteopontin (OPN), Parkinsonism associated deglycase (Park7), Platelet-Derived Growth Factor AA (PDGF-AA) and BB (PDGF-BB), Receptor for Advanced Glycation Endproducts (RAGE), Serine Proteinase Inhibitor!Zclade E1 (Serpin E1), Tumour Necrosis Factor alpha (TNF-α), Vascular Cell Adhesion Molecule (VCAM-1), Vascular Endothelial Growth Factor (VEGF), and von Willebrand Factor A2 domain (vWF-A2).
Additionally, 10 biomarkers were analysed using the Neuroscience 18-plex Human ProcartaPlex™ Panel assay provided by ThermoFisher Scientific Invitrogen™ (catalogue ref.: EPX180-15837-901): Amyloid beta (1–42) (Aβ(1-42)), Brain-Derived Neurotrophic Factor (BDNF), Kallikrein 6 (KLK6), Migration Inhibitory Factor (MIF), Nerve Growth Factor beta (NGF-β), Neural Cell Adhesion Molecule (NCAM-1), Neurogranin (NRGN), S100 calcium-binding protein β (S100β), TAR DNA-binding protein 43 (TDP-43), and Chitinase-3-like protein 1 (YKL-40).
Serum levels of biomarkers screened via the Human Luminex® Discovery Assay (#LXSAHM) and the Neuroscience 18-plex Human ProcartaPlex™ Panel (#EPX180-15837-901) were measured using a MAGPIX© bioanalyzer, according to manufacturer’s instructions. Singlicate measurements were taken of each sample. Sample concentrations were extrapolated from a standard curve, fitted using a 6-parameter logistic algorithm. Luminex technology is a multiplex assay platform that allows for the simultaneous measurement of multiple biomarkers in a single plasma sample and can detect molecule concentrations in the picogram range. The use of Luminex technology in the study of biomarkers of brain injury is widely extended (30) due to its high sensitivity and reproducibility.
Lastly, 4 brain injury biomarkers were analysed using the highly sensitive single molecule immunoassay (SIMOA), using the Human Neurology 4-Plex A assay provided by Quanterix (catalogue ref.: 102153): Neurofilament light (NfL), total Tau protein (Tau), Glial Fibrillary Acidic Protein (GFAP), and Ubiquitin Carboxyl-terminal Hydrolase L1 (UCH-L1). Levels of biomarkers screened via the SIMOA Human Neurology 4-Plex A assay (#102153) were measured using a SIMOA HD-X™ Analyzer (Quanterix), according to manufacturer’s instructions. Briefly, samples were thawed at 21°C, and centrifuged at 10,000 RCF for five minutes at the same temperature. On-board the instrument, samples were diluted 1:4 with sample diluent and bound to paramagnetic beads coated with a capture antibody specific for human GFAP, NfL, Tau, and UCH-L1. GFAP, NfL, Tau, and UCH-L1 bound beads were then incubated with a biotinylated GFAP, NfL, Tau, and UCH-L1 detection antibodies in turn conjugated to streptavidin-β-galactosidase complex that acts as a fluorescent tag. Subsequent hydrolysis reaction with a resorufin β-D-galactopyranoside substrate produces a fluorescent signal proportional to the concentration of GFAP, NfL, Tau, and UCH-L1 present. Singlicate measurements were taken of each sample. Sample concentrations were extrapolated from a standard curve, fitted using a 4-parameter logistic algorithm. SIMOA is an ultrasensitive digital immunoassay that allows for the detection and quantification of low-abundance plasma biomarkers with sub-femtomolar sensitivity, high specificity, and high reproducibility (31).
2.4 Serological assessment of malaria exposure via Luminex immunoassay
As past infections with Plasmodium spp. could also influence observed biomarker levels across participants, total IgG antibody responses of all Pk-infected patients and healthy controls were measured using a multiplex bead-based immunoassay built on Luminex® xMAP® technology to assess previous exposure to P. knowlesi as well as P. falciparum, P. vivax, P. malariae, and P. ovale. Sera were screened against a previously optimised panel of 14 blood-stage antigens representing varied markers of malaria exposure, listed in Supplementary Table 3. This incorporated 6 Pk antigens: PkAMA1 and PkMSP119, historical (long-term) exposure markers that can persist in blood for several years with repeated infections, PkSera3 Ag2 and PkSSP2/TRAP, which have previously been utilised as markers in seroprevalence studies on Malaysian populations (32–34), and Pk1 and Pk8, exploratory antigens that demonstrated high immunogenicity in preliminary data from assay screenings of pooled sera from Pk-infected Malaysian hyperimmune individuals (K. Tetteh, personal communication). For P. falciparum and P. vivax, markers of historical exposure, Pf / PvMSP119 and Pf / PvAMA1, and markers of recent (short-term) exposure known to persist in blood for up to 6-12 months following infection, PfEtramp5 Ag1 and PvRBP 2b, were included (35,36). PmMSP119 and PoMSP119 were used to assess historical exposure to P. malariae and P. ovale, respectively. Additionally, tetanus toxoid vaccine protein from Clostridium tetani and glutathione-S-transferase (GST) from Schistosoma japonicum were included as non-malaria internal assay controls (37).
The Luminex assay was performed as previously described (38). Briefly, based on previously identified antigen-specific optimal EC50 concentrations, each antigen was covalently coupled to a colour-coded MagPlex® bead region (MagPlex, Luminex Corp., Austin, TX) via N-Hydroxysuccinimide/1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (NHS/EDC) chemistry. Test sera were prepared at 1/400 in diluent buffer (1x Phosphate Buffer Saline (PBS); 0.05% Tween® 20; 0.5% bovine serum albumin; 0.02% sodium azide; 0.1% casein; 0.5% polyvinyl alcohol; 0.5% polyvinylpyrrolidone; 15.25ug/mL E. coli lysate) and incubated at 4°C overnight. Antigen-coupled beads were incubated with 50μL of diluted samples and incubated for 90 minutes at room temperature with shaking at 600 RPM before incubation with 50μL of 1/200 R-Phycoerythrin-conjugated AffiniPure F(ab’)₂ Fragment Goat Anti-Human IgG secondary antibody (JacksonImmunoResearch; 109-116-098) for 90 minutes under the same conditions (38).
Pooled sera from hyperimmune individuals in Malaysia (Pk+), Tanzania (CP3), and Peru (S1) were included as Pk, P. falciparum, and P. vivax positive controls, respectively, in 6-point 5-fold serial dilution curves (1/10 - 1/31250). Commercial WHO reference reagents for anti-P. falciparum (10/198) human serum and anti-P. vivax (19/198) human plasma were also added as positive controls at 1/400 and 1/4000. Public Health England (PHE) malaria naïve human sera (n=30) were included as negative controls at 1/400. Two wells of diluent buffer served as blank controls to allow subtraction of background signal. The plates were read using the MAGPIX© instrument (Luminex Corp., Austin, TX) with the raw data recorded as Median Fluorescent Intensity (MFI) values using an acquisition target of ≥30 beads per region per well. The data were background-adjusted and normalised as described by Wu et al (38).
2.5 Statistical analysis
Sociodemographic characteristics of the participants were reported as mean (±SD) for continuous data, or absolute frequencies (%) for categorical data. Continuous data was assessed for normality using Shapiro-Wilk tests. Comparisons between Pk-infected cases and healthy controls were made using t-tests or Wilcoxon tests for normally and non-normally distributed continuous data, respectively, and χ2 tests for categorical data.
Differences in levels of individual biomarkers between Pk-infected patients and controls were examined using two-tailed t-tests and Wilcoxon tests for normally and non-normally distributed data, respectively. Bonferroni correction for multiple comparisons was applied. Correlation matrixes were used to visualize the relationships between biomarker levels in both groups, with previous data scaling. Any potential correlations between biomarker levels and demographic and clinical parameters such as age and parasitemia were evaluated using Spearman’s rank correlation coefficient.
To elucidate potential profiles of biomarker levels that could distinguish between infected patients and controls, a hierarchical clustering heatmap analysis was performed, prior scaling of biomarker data. This provided a visual representation of the relationships between individuals based on the similarity of their biomarker profiles. Initially, the analysis was conducted on the entire set of biomarkers to comprehensively explore the clustering patterns within the dataset. Subsequently, the clustering analysis was applied separately to subsets of biomarkers representing brain injury, infection and immunity, and vascular biomarkers.
To assess malaria exposure, a hierarchical clustering heatmap was similarly performed to allow visual comparison of individual antibody responses to Pk-specific antigen profiles across study participants. Comparisons of group means of MFI responses between Pk-infected cases and healthy controls were conducted using the non-parametric Kruskal-Wallis test with Dunn’s correction for multiple comparisons applied.
Statistical analyses were performed in R (version 4.3.1) with RStudio software (version 2023.09.1-494). Data was visualized in R (version 4.3.1) with RStudio software (version 2023.09.1-494) and GraphPad Prism 10. P values of less than 0.05 (two-tailed) were considered statistically significant.
3 Results
3.1 Cohort characteristics
All participants in our sub-study were male (N=38, 100%). Both Pk-infected patients and community-matched, age-matched uninfected controls had an average age of 39 (±15) years. In the Pk-infected group, patients presented with median parasitemia 15,200.00 parasites/μL and 26,746.00 interquartile range.
3.2 Concentrations of individual markers between Pk-infected patients and controls
Biomarkers of brain injury
Bonferroni-corrected analyses revealed significantly higher plasma levels of the following brain injury biomarkers in the Pk-infected group, compared with uninfected controls: Tau (p=0.0007), UCH-L1 (p<0.0001), αSyn (p<0.0001), Park7 (p=0.0006), NRGN (p=0.0022), and TDP-43 (p=0.005). In contrast, levels of the following biomarkers were found to be significantly lower in the Pk-infected group when compared with uninfected controls: CaBD (p<0.0001), CNTN1 (p<0.0001), NCAM-1 (p<0.0001), BDNF (p<0.0001), GFAP (p=0.0013), and KLK6 (p=0.0126). Results obtained on CaBD and CNTN1 levels revealed clear cutoff values that allowed for subject classification based on infection status: All Pk-infected patients (19/19, 100%) presented with plasma CaBD levels below 1,400pg/mL or CNTN1 levels below 12,500pg/mL, whereas concentrations in all uninfected controls (19/19, 100%) were above 1,400pg/mL for CNTN1 and above 15,000pg/mL for CNTN1, respectively (Figure 2, Supplementary Table 4).
Since S100B and Aβ(1-42) plasma levels were below the range of detection in most uninfected subjects, we first compared the proportion of individuals with detectable levels of these biomarkers in each group using a Chi-square test with Yates’ correction. The percentage of participants with plasma levels within the range of detection was significantly higher in the Pk-infected group (19/19, 100%) than in the uninfected control group (4/19, 21.05%) (p<0.0001). Similarly for Aβ(1-42), the proportion of individuals with detectable plasma levels in the Pk-infected group (17/19, 89.47%) was significantly higher compared with their uninfected peers (7/19, 36.84%) (p= 0.0025).
To enable group comparisons for these two biomarkers, participants with values below the range of detection were assigned a numerical value corresponding to half the value of the lower threshold of detection. According to the manufacturer’s information, these values were 27.31 pg/mL for S100B and 0.22 pg/mL for Aβ(1-42). Achieved this, Wilcoxon tests revealed significantly higher plasma S100B levels in the Pk-infected group, which survived Bonferroni correction for multiple comparisons (p<0.0001), whereas no significant group differences were found for Aβ(1-42) plasma levels.
Biomarkers of infection and immune activation
Group differences in plasma levels were found after correction for multiple comparisons. IL-1RA (p<0.0001), IL-10 (p<0.0001), and MPO levels (p=0.0314) were significantly higher in the Pk-infected group when compared with their uninfected peers, whereas CCL4 (p<0.0001), CCL5 (p<0.0001), CRP (p=0.0280), and RAGE levels (p=0.0025) were significantly lower (Supplementary Figure 1, Supplementary Table 4).
Biomarkers of vascular activation and injury
Ang-2/Ang-1 ratios (p<0.0001) and VCAM-1 levels (p=0.0001) were significantly higher in Pk-infected patients compared with uninfected controls, while Ang-1 (p<0.0001), BMP-9 (p<0.0001), PDGF-AA and -BB (p<0.0001), and Serpin E1 levels (p<0.0001) were significantly lower (Supplementary Figure 2, Supplementary Table 4).
3.3 Hierarchical clustering of samples based on biomarker blood levels
Hierarchical clustering heatmap analyses revealed distinct group profiles for brain injury biomarkers (Figure 3A). Most infected individuals clustered together (17/19, 89.47%), indicating a cohesive pattern of elevated levels of Park7, S100B, αSyn, and TDP-43. Two infected individuals exhibited atypical clustering with the healthy controls, suggesting a subgroup with a distinct biomarker profile. In the control group (19/19, 100%), certain biomarkers, including BDNF, CaBD, CNTN1, and GFAP, displayed higher levels and clustered together, representing a baseline biomarker profile in healthy individuals. As mentioned, two infected individuals (2/19, 10.53%) clustered with the control group in this category, indicating potential overlap or similarity in the levels of these specific biomarkers between infected and control individuals.
Clustering of biomarkers associated with infection and immune activation did not reveal distinct separation between the two groups (Figure 3B). Higher levels of IL-1RA and MPO predominantly clustered in the infected group (17/19, 89.47%), and two control individuals also exhibited this pattern (2/19, 10.53%). One infected individual (1/19, 5.26%) displayed very high levels of IL-1β, GM-CSF, TNF-α, CCL4, and CCL2. Another infected individual (1/19, 5.26%) exhibited very high levels of OPN and IL-10. Additionally, one control individual showed high levels of IL-6 (1/19, 5.26%).
Lastly, clustering of vascular biomarkers revealed clear and distinctive group profiles (Figure 3C). In the control group (19/19, 100%), a cohesive cluster formed by high levels of PDGF AA, Ang-1, PDGF BB, Serpin E1, and BMP-9 was observed. Conversely, in the infected group (19/19, 100%), a distinct cluster of higher levels of VCAM-1 and the Ang-2/Ang-1 ratio was identified.
Correlations between biomarkers in each group can be observed in Supplementary Figure 3.
3.4 Correlation of biomarker levels with age and parasitemia
No significant correlations were identified between the levels of any of the examined biomarkers and the age of participants within each respective group, as illustrated in Supplementary Figure 4. Furthermore, within the Pk-infected group, no significant correlations were observed between biomarker levels and the percentage of parasitemia (Supplementary Figure 5).
3.5 Serological markers of previous malaria exposure
Overall, the majority of participants generated low antibody responses across Pk, P. vivax, P. falciparum, P. ovale, and P. malariae antigens with the exception of a few high responders (Supplementary Figure 6). Clustering of exposure markers in the hierarchical heatmap analysis did not reveal distinct separation between responses of Pk-infected patients and uninfected controls (Supplementary Figure 7). Among Pk antigens, uninfected controls exhibited higher responses against Pk8 (p<0.0001), PkSERA3 Ag2 (p=0.0406), and Pk1 (p=0.0250) than Pk-infected patients, but no differences were observed for PkAMA1 (p>0.9999), PkMSP119 (p=0.7778), or PkSSP2 (p=0.2592).
4 Discussion
Our study sheds light on the potential neurological implications of Pk infection in Malaysia. Through the investigation of circulating levels of brain injury, inflammation, and vascular biomarkers in Pk-infected patients compared to healthy controls, we uncovered compelling evidence suggesting a significant impact of Pk infection on brain health. In line with previous reports suggesting neurological involvement in severe malaria, our findings demonstrate increased levels of several biomarkers associated with cerebral injury in Pk-infected patients. Notably, elevated levels of S100B, Tau, UCH-L1, αSyn, Park7, NRGN, and TDP-43 signify potential brain damage in response to Pk infection in our cohort.
Alterations in circulating levels of S100B, Tau, and UCH-L1 have been documented in previous investigations of severe malaria. S100B, an abundant neurotrophic factor predominantly expressed in astrocytes, serves as a biomarker of blood-brain barrier permeability and central nervous system injury (39). Our group reported increased plasma levels of S100B in Indian patients with falciparum severe malaria compared to uncomplicated cases (40). Similarly, levels of microtubule-associated protein Tau, a neuropathological hallmark of Alzheimer’s Disease and a biomarker for brain injury (41), were found increased in the plasma of Ugandan children with severe falciparum malarial anemia or cerebral malaria compared to uninfected controls. In cerebral malaria cases, Tau levels correlated with mortality and persistent neurocognitive impairment in survivors (42). Lastly, plasma levels of the ubiquitin-protein hydrolase UCH-L1, a marker of neuronal damage (43) were significantly higher in Ugandan children with falciparum cerebral malaria and severe malarial anemia compared with asymptomatic community children. The elevated plasma levels were associated with surrogate markers of blood-brain barrier dysfunction and cognitive deficits over follow-up (44). Our findings of increased S100B, Tau, and UCH-L1 plasma levels during Pk infection align with these observations, suggesting potential cerebral involvement, neuronal injury, and blood-brain barrier disruption in knowlesi malaria pathogenesis.
In broader clinical contexts, elevated circulating biomarkers in our Pk-infected cohort have been associated with diverse conditions and are used as indicators of brain injury or dysfunction. For instance, patients with Parkinson’s Disease had significantly higher plasma levels of synaptic vesicle trafficking regulator αSyn compared to healthy controls (45), and elevated levels of post-synaptic protein NRGN were associated with mild to acute traumatic brain injury (46,47), as well as neonatal encephalopathy (48). The increased levels of αSyn and NRGN in our Pk-infected patients suggest potential neurological implications, prompting further investigation into the underlying mechanisms of knowlesi malaria pathogenesis and their clinical significance over time.
In our study, several biomarkers were significantly decreased in Pk-infected individuals compared to healthy controls, including GFAP, BDNF, CaBD, CNTN1, NCAM-1, and KLK6. The astrocytic intermediate filament protein GFAP, used in clinical settings to assess severity of traumatic brain injury (49), glymphatic function (50), and neurodegeneration (51,52), was not found to be elevated in Ugandan children with falciparum cerebral malaria and severe malarial anemia (44). In contrast, the observed differences in GFAP levels in our groups may suggest altered astroglial and/or glymphatic functions in infected participants, supporting our hypothesis of cerebral involvement in Pk malaria. Conversely, plasma levels of the neurotrophic factor BDNF, reported to be increased in early Alzheimer’s disease (53) and decreased in multiple sclerosis (54), were also lower in our study participants, potentially indicating dysregulated neuroprotective mechanisms during Pk infection. Furthermore, increased cerebrospinal fluid levels of CaBD, a protein involved in calcium homeostasis and neuronal signalling, were associated with risk of dementia in cognitively normal patients (55). CNTN1 and NCAM-1, two crucial factors for neural development and synaptic formation, were decreased in our Pk-infected patients. Downregulated CNTN1 serum levels predicted cognitive and motor declines in patients with Parkinson’s Disease (56), while lower plasma NCAM-1 levels correlated with cognitive impairment (57,58). Lastly, plasma levels of the neuroinflammation modulator KLK6 were significantly increased in patients with advanced Alzheimer’s Disease compared with healthy controls (59). Although challenging to interpret, these results suggest cerebral involvement during Pk infection and a potential risk of cognitive decline, indicating a complex interplay between neuroinflammatory and neuroprotective mechanisms that warrant further investigation.
Despite our expectations based on existing malaria literature (Table 1), we did not observe significant group differences in levels of some biomarkers, such as the iron-trafficking protein Lipocalin-2 (NGAL), a recognized biomarker of neuroinflammation (60). Elevated plasma NGAL levels were associated with cerebral malaria in adult patients from India and distinguished between fatal and nonfatal outcomes (61). Similarly, plasma NfL levels in Mozambican children with uncomplicated and severe falciparum malaria showed no baseline group differences, but a significant increase over time, particularly in severe malaria cases with neurological symptoms, suggesting NfL as a potential follow-up biomarker of brain injury in malaria (62). Additionally, the glycoprotein Fetuin A was previously reported as elevated in serum from Malaysian patients with Pk malaria compared to uninfected controls (63). Lastly, plasma levels of the neurotrophic factor CNTF were significantly reduced in Thai patients with cerebral malaria or severe malaria with renal failure, compared to severe malaria cases without cerebral or renal involvement (64). Our findings, diverging from anticipated outcomes, underscore the complexity of biomarker dynamics in infections with different Plasmodium species and emphasize the need for further exploration in the context of Pk malaria.
In our hierarchical clustering analysis, we observed distinct group profiles for brain injury biomarker levels, indicating differences between Pk-infected patients and healthy controls. Most infected individuals clustered together with elevated levels of specific biomarkers such as S100B, αSyn, Park7, and TDP-43, suggesting a cohesive pattern of cerebral injury. However, two infected individuals exhibited an atypical clustering with the healthy controls, hinting at potential subgroup variability within the infected cohort, or different stages of disease progression. Conversely, in the control group, several biomarkers including BDNF, GFAP, CaBD, and CNTN1 displayed higher levels and clustered together, representing a baseline biomarker profile in healthy individuals. As mentioned, two infected individuals clustered with the control group in this category, indicating potential overlap or milder infection.
Clustering of vascular biomarkers revealed clear and distinctive group profiles. Uninfected individuals formed a cohesive cluster with high levels of PDGF AA, Ang-1, PDGF BB, Serpin E1, and BMP-9, revealing a healthy, baseline vascular profile. Conversely, the infected group cluster had high levels of VCAM-1 and Ang-2/Ang-1 ratio, indicative of vascular involvement and endothelial activation during Pk infection. The elevated Ang-2/Ang-1 ratio observed in our infected group aligns with findings from previous studies on severe malaria caused by other Plasmodium species (65–70). Similarly, increased levels of VCAM-1 have been reported in both falciparum and vivax malaria infections (71,72). In contrast to brain injury and vascular biomarker profiles, hierarchical clustering of biomarkers associated with infection and immune activation did not reveal distinct separation between the two groups.
Lastly, total IgG antibody responses observed against Plasmodium spp. antigens, including Pk, were low not only among most uninfected controls but also Pk-infected patients, and did not form distinct clusters between the two groups. This suggests the cohort was largely malaria naïve, with low reactivity to long-term infection markers Pk/Pv/Pf/Pm/PoMSP119 and Pk/Pv/PfAMA1 observed at levels comparable to the PHE malaria naïve controls in all but two Pk-infected patients (73). As naturally acquired responses against these antigens develop cumulatively with repeated exposure, this indicates the occurrence of very few historical Plasmodium infections across participants (74). This could highlight a lack of clinical immunity, rendering Pk-infected patients more susceptible to potential malaria complications reflected in the brain and vascular biomarker profiles in this study, though this was not possible to investigate due to insufficient clinical information. The lack of elevated responses to Pk antigens SSP2, SERA3 Ag2, Pk1 and Pk8 may indicate there was inadequate time for Pk-infected patients to mount a response prior to the sampling period (32).
Among uninfected controls, serological analyses identified two individuals exhibiting notably higher responses: one control to Pk exploratory marker Pk8 and P. vivax short-term exposure marker PvRBP2b, and another also to PvRBP2b. Although this suggests previous exposures to Pk and P. vivax, this did not appear to explain any atypical deviations in plasma biomarker levels observed in a minority of individuals in this group.
Our study had several limitations. Firstly, the lack of detailed clinical information regarding malaria severity and the presence or absence of renal failure among the Pk-infected group restricts our ability to fully contextualize the observed alterations in brain injury biomarkers. While recent studies have implicated AKI in the pathogenesis of brain injury (40) and long-term neurocognitive sequelae (10,11), we were unable to directly investigate this relationship. Since AKI is a common complication of knowlesi malaria (27), it is likely that similar distant organ pathways are involved, resulting in an exacerbation of the brain changes (26). Secondly, our study did not include long-term follow-up assessments to determine whether the elevated biomarker levels observed in Pk-infected patients return to baseline or exhibit a steady increase over time. This would have been particularly relevant for NfL, which increases in the weeks following cerebral insult and could provide valuable insights into the persistence or resolution of brain injury following Pk infection (43,62). Lastly, a symptomatic, non-malaria control group would allow stronger ascertainment of associations between biomarkers and Pk infection.
In conclusion, our study represents the first comprehensive exploration of surrogate markers of cerebral involvement in Pk-infected patients from Malaysia. Through the investigation of circulating levels of brain injury, inflammation, and vascular biomarkers in Pk-infected patients compared to healthy controls, we uncovered compelling evidence suggesting a significant impact of Pk infection on brain and vascular health. The identification of distinct patterns in the clustering of biomarkers associated with cerebral injury and endothelial dysfunction highlights the complex interplay between these pathways during Pk infection. Our study lays the groundwork for further research to elucidate the pathophysiology of this disease, and its potential long-term effects on the brain. Additional investigations are urgently needed to assess structural and functional changes in the brain during knowlesi malaria, through both neuroimaging and longitudinal neurocognitive evaluations. Our findings not only contribute to our understanding of Pk malaria but may also have future implications for the diagnosis, management, and therapeutic targeting of neurological complications associated with this emerging infectious disease.
5 Ethics approval and consent
This study has been reviewed and approved by the Medical Research and Ethics Committee of the Ministry of Health in Malaysia (reference NMRR ID 22-02557-1KV) and by the Observational/Interventions Research Ethics Committee at the London School of Hygiene and Tropical Medicine, UK (reference 27902). The shipment of samples from Malaysia to the UK was performed in compliance with national and international regulations in both sites. The samples are registered with the Human Tissue Authority, in accordance with UK national guidelines.
Data Availability
All data produced in the present work are contained in the manuscript.
6 Conflict of Interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
7 Author Contributions
CBC: conceptualization, methodology, investigation, formal analysis and interpretation, visualization, writing – original draft, review & editing. ED: methodology, formal analysis and interpretation, visualization, writing – review & editing. NRN: conceptualization, methodology, investigation, writing – review & editing. MYL: conceptualization, methodology, investigation, writing – review & editing. KT: methodology, investigation, resources, and writing – review & editing. CD: methodology, investigation, resources, and writing – review & editing. SK: conceptualization, investigation, and writing – review & editing. YLL: conceptualization, methodology, investigation, resources, supervision, and writing – review & editing. SCW: conceptualization, methodology, investigation, resources, supervision, and writing – review & editing. All authors approved the final version.
8 Funding
SCW is supported by the Medical Research Council UK under award number MR/S009450/1, and the National Institute of Allergy and Infectious Diseases of the National Institutes of Health (NIH) under award number U19AI089676. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
9 Acknowledgments
We extend our gratitude to the local health officers and personnel from Mersing, Kota Tinggi, and Pejabat Kesihatan Dareah Kota Tinggi (Johor), Kuala Lipis and Kesihatan Betau (Pahang), Sungai Siput (Perak), Kuala Kubu Bharu (Selangor), and Setiu (Trengganu) for their invaluable assistance with sampling. We express our sincere appreciation to Dr Christian Bottomley, Ana Chopo-Pizarro, and Helena Brazal Monzó at LSHTM for their invaluable insights, which greatly enhanced the quality of this research.
10 References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.
- 67.
- 68.
- 69.
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.
- 76.
- 77.
- 78.
- 79.
- 80.
- 81.
- 82.
- 83.
- 84.
- 85.
- 86.
- 87.
- 88.
- 89.
- 90.
- 91.
- 92.
- 93.