Gut microbiota influences Plasmodium falciparum malaria susceptibility

The gut microbiota has recently been associated with susceptibility/resistance to malaria in animal models and humans, yet the impact of the gut microbiota on the risk of a malaria attack remains to be assessed. This study aims at assessing the influence of the gut microbiota on malaria attacks and Plasmodium parasitæmia in children living in a malaria-endemic area in Mali. Three hundred healthy children were included in a 16-months cohort study in Bandiagara. Their gut bacteria and fungi community structures were characterised via 16S and ITS metabarcoding from stool samples collected at inclusion. Clinician team monitored the occurrence of malaria attacks. Asymptomatic carriage of Plasmodium was assessed by qPCR. Over the 16-month period, 107 (36%) children experienced at least one occurrence of malaria attacks, and 82 (27%) at least one asymptomatic Plasmodium parasitæmia episode. A higher gut bacteria richness was independently associated with susceptibility to asymptomatic parasitæmia episodes and malaria attacks; while the Shannon H diversity and Chao-1 richness index of gut fungi community structure was relatively homogeneous in children who were and were not infected with P. falciparum. Using a linear discriminant effect size analysis of operational taxonomic units assigned to the species level, 17 bacteria, including Clostridiaceae, Eubacteriaceae, Senegalimassilia sp., Atopobiaceae and Lachnosipraceae, and seven fungi, including Dioszegia fristigensis, Ogataea polymorpha and Cutaneotrichosporon cyanovorans, were associated with susceptibility; whereas eight bacteria, including, Bifidobacterium spp., Weissela confusa and Peptostreptococcacea, and 3 fungi, Malassezia sp., Niesslia exosporoides, and Didymocrea leucaenae, were associated with resistance to malaria. Moreover, 15 bacteria, including Coproccus eutactus, Terrisporobacter petrolearius, Klebsiella pneumoniae and Ruminococcaceae, and 13 fungi, including Wallemia mellicola, were associated with susceptibility, whereas 19 bacteria, including Bifidobacterium spp., Bacteroides fragilis, Peptostreptococcacea, and Lactobacillus ruminis, and three fungi, including Cryptococcus neoformans, were associated with resistance to asymptomatic Plasmodium parasitæmia episodes. Further studies are needed to confirm these findings that point the way towards strategies aiming to reduce the risk of malaria by modulating gut microbiota components in atrisk populations.


Introduction
Malaria is caused by a protozoan parasite, Plasmodium spp., which are transmitted to humans through the infesting bite of a female Anopheles mosquito. The highest malarial burden is clustered in Africa and the area south of the Sahara makes up 98% of the 228 million cases worldwide each year. Plasmodium falciparum malaria is the most life-threating parasitic disease worldwide, despite the implementation of multiple control strategies. It causes 405,000 deaths or more each year and children under the age of five are the most vulnerable population, making up 67% (272,000) of these deaths [1].
Malaria parasite carriage can progress as an asymptomatic infection, as uncomplicated malaria with symptoms as fever, headache, and chills, or as severe malaria including high parasitaemia, severe anaemia, respiratory distress, and cerebral malaria (altered consciousness and seizures) [2,3]. However, factors have been described to be associated with malarial resistance, including ethnic and genetic factors, such as sickle cell traits (AS, C), Duffy negative blood group, HLA group, and polymorphisms in immune response genes [3][4][5][6][7][8]. In areas of intense malaria transmission, Young children are more susceptible to malaria due to the loss of maternal antibody protection from 9 months. They present a high malaria morbidity and mortality. Older children, as a result of repeated malaria attacks, acquire a non-sterilising premunition that reduces the risk of severe malaria and associated case fatality [9,10]. The immune response to microbiota has been associated with severe malaria infection. Lactobacillus and Bifidobacterium species in the gut have shown to play a protective role against Plasmodium infection by reducing the parasite load and attenuating the severity of malaria in mice [2].
Severe malaria infections alter the functional capacity of the microbiota, improving bacterial motility and amino acid metabolism in mice with a high parasite load compared to a mild infection [17]. Severe malaria infection can be modulated by the gut microbiota in genetically diverse mice and pregnant animals. The abundance of Akkermansia muciniphila, Allobaculum, Lactobacillus and S24-7 has also shown to be negatively correlated with parasite load [18].
Also, severe malaria infection in pregnant mice, which is a function of the composition of the intestinal microbiota, significantly influences foetal and postnatal outcomes [18]. Notably, the bacterial of the microbiota was associated with the risk of asymptomatic Plasmodium parasitaemia occurrence and not of malaria attack [19]. However, no studies have yet established that the bacterial microbiota contributes towards protecting against malaria in humans, and the influence of the fungal microbiota on malaria has not yet been investigated.
This study aimed to assess whether both bacteria and fungi communities in the gut were associated to the susceptibility/resistance against asymptomatic Plasmodium infection and malaria attacks in children living in a malaria-endemic area, Bandiagara (Mali).

Children's baseline characteristics
Three hundred (300) children were included in the cohort study; the median age of the cohort was 8 IQR [7; 8], respectively, from which the ages were stratified. Stools were able to collect from 296/300 children. Microbiota analysis of the faeces of the 296 children was performed.

Gut bacteria 16S metabarcoding
From the 296 stool samples, 24,913,960 sequences were generated for taxonomic assignment.
Sequences not classified at the species level and the other classes were indicated "unassigned" for taxonomy. Some classified sequences were also labelled as IHU_Bacteria (bacteria from the IHU database) with distinct numbers. The analysis of Operational Taxonomic Units (OTUs) in the children's gut bacterial communities showed a relatively evenly distributed frequency of the major phyla, Firmicutes (12%), Bacteroides (12%), Actinobacteria (12%) and Proteobacteria (12%), and a relative higher abundance of Firmicutes (47%) compared to other phyla, including Bacteroides (8%), Actinobacteria (8%) and Proteobacteria (6%) ( Figure 1A, relative frequency (C) and abundance (D) of the major gut fungal phyla, characterised via 16S or ITS metabarcoding.

Gut fungi ITS metabarcoding
The ITS1 region analysis yielded 647,816 reads and 532 single OTUs; the ITS2 region Figure 1C, 1D). Because gut fungi are less complex and relatively less known than bacterial communities, we further analysed the distribution of the fungal taxa according to the children's age groups. All fungi phyla were detected in children under the age of five; each phylum frequency was relatively higher in the 5 to 10-year-old group ( Figure 2A). Fungal phyla abundance relatively peaked in children under the age of 5 to 10-year-old and then decreased with age ( Figure 2B).   were independently statistically significantly associated with the risk of developing at least one asymptomatic Plasmodium parasitaemia episode (Table S2).
Both age and OTU richness were significantly associated with both malaria attack and asymptomatic Plasmodium parasitaemia in the univariate analysis. The survival analysis showed that being over the age of four was associated with an increased risk of both malaria attack (p= 0.015) and asymptomatic Plasmodium parasitaemia (p<10 -3 ) ( Figure S1A, S1B). The median OTU richness was 1,106 (112-2,551) and the median Chao-1 index was 1,303 (428-3,979). To estimate the risk of malarial infection, the cohort of 258 children were divided in two groups either lesser than or equal to and above the median of the richness indices (observed OTU richness and Chao-1). Cox regression analysis showed that a relatively low OTU richness was associated with a significantly lower risk of malaria attack than with high OTU richness in p=0.0009) ( Figure S2A, S2B). To assess whether age could be a confounding factor, the same analysis was conducted in children aged 0-4 years old or aged 5 years old and above. Children showed a lower risk of both malaria attack and asymptomatic Plasmodium parasitaemia episodes than children with an OTU richness above the median value (p=0.03) ( Figure S3).
This effect was even enhanced in children aged 5-15 years, where those with an OTU richness below the median value (1109; IQR [680 -1307]) showed a lower risk of both malaria attack and asymptomatic Plasmodium parasitaemia episodes than those with an OTU richness above the median value (p< 0.0001) ( Figure S4).

Malaria in the Dogon and Fulani ethnic groups
Regarding these two sympatric ethnic groups in Bandiagara, more Dogon (n=200) than Fulani (n=17) children were included in this study. Because of the relatively small number of Fulani enrolled into the study, the impact of the gut bacterial and fungal communities in these ethnic groups was not investigated further.

Gut bacterial community structure associated with malarial risk
The association of age, sex, eukaryotic enteric pathogens, and gut bacterial and fungal community structures with either the risk of a malaria attack or asymptomatic Plasmodium parasitaemia, were assessed using an unconditional logistic regression analysis. Univariate analysis showed that age and gut bacteria richness indices ( The gut bacterial community structure of children who developed at least one malaria attack were compared with those who did not, and that of those who developed at least one asymptomatic Plasmodium parasitaemia episode was compared with those who did not. At the phylum level, the relative abundance of the main bacterial phyla were Firmicutes (53.95%), Proteobacteria (53.84%), Actinobacteria (52.9%), and Bacteroides (49.54%) in the children who developed at least one malaria attack compared to those who did not ( Figure   Bacterial OTU richness (Chao-1 index and observed OTU richness) and diversity in Shannon H and Simpson D indices were evaluated according to malaria status and age (Table S3). The diversity indices were homogeneously distributed in each of the groups of children (Table S3).
In contrast, the Chao-1 index (median 1,457 [1,561]) in the children who developed at least one malaria attack was significantly higher than in those who did not (1,330 [1,257-1,402]) (p=0.036). Also, the median bacterial OTU richness observed (1,233 [1,152-1,313]) in the children who developed at least one malarial attack was statistically significantly higher (p=0.001) than that observed in those who did not (1,076 [1,022-1,129]) ( Figure 4; Table S4). We found that the median bacterial OTU richness observed in the children who developed at least one asymptomatic Plasmodium parasitaemia episode (1,217 [1,130-1,305]) was higher (p=0.02) than in those who did not (1,099 [1,046-, 1,151]) ( Figure 5; Table S5).  Bifidobacterium sp., Bacteroides fragilis, and Lactobacillus ruminis were more abundant in those who did not ( Figures 7A-D).
Regarding the effect of age on bacterial community structure; children from 0-4 years old who developed at least one asymptomatic episode of Plasmodium parasitaemia, Romboutsia timonensis and Ruminococcus bromii were more abundant among 107 species, whereas in children who did not 5 species, Collinsella aerofaciens and Bifidobacterium longum subsp.
longum were more abundant ( Figure S9); in 5-15 years old group who developed at least one asymptomatic episode of Plasmodium parasitaemia, 6 species, including pneumoniae subsp.
pneumoniae and IHU_PS_96_Ruminococcus_395, were more abundant, whereas 10 species, including Bacteroides fragilis and IHU_PS_96_Eubacterium_436, were more abundant in children who did not ( Figure S10). The alpha diversity indices did not statistically significantly differ between the children who 8 developed at least one malaria attack (Table S4), asymptomatic Plasmodium parasitaemia 9 episode (Table S5) and those who did not ( Figure S11).

10
The beta diversity of the fungal community between children with at least one malaria attack 11 (n = 105) and children tested negative (n = 191) was analysed by Principal Coordinated  The relative mean of phyla abundance in children who developed at least one malaria attack 24 was Ascomycota (46%), Basidiomycota (62%) compared to children with no malaria attack ( Figure S12). In children who developed at least one asymptomatic Plasmodium parasitaemia 26 episode, the relative mean of phyla abundance was Ascomycota (52%), Basidiomycota (50%) 27 against children without an asymptomatic Plasmodium parasitaemia episode ( Figure S13).

28
At the phyla and class level, no abundance of fungi was significant between children who 29 developed at least one attack of malaria and those who did not. five and nine years [25]. It is notable that the children included in our study were not exposed 84 to seasonal malaria chemoprevention, which has been associated with a significant reduction of 85 malaria incidence in children under the age of five [26,27]. However, they did use both long-86 lasting insecticidal nets (LLINs) and indoor residual spraying of insecticide (IRS), which might 87 explain the observed higher malaria burden in older children [28]. Indeed, age was an 88 independent factor that was strongly associated with both malaria attack and asymptomatic showed that the risks of both malaria attack and asymptomatic Plasmodium parasitaemia 92 significantly increased with the increase in bacteria OTU richness (Tables 2, S2) To the best of our knowledge, our study is the first to investigate the association of the gut 118 fungal community with the risk of malaria. We found that the gut fungal community structure 119 was relatively homogenous between children who were susceptible to malaria and those who 120 were resistant to it. This finding contrasts with previous reports of a higher abundance and   Table 1. Baseline characteristics of the children included in the cohort study, by age group. Table 2. Logistic regression analysis of the association between age and gut bacterial and fungal community structure with the risk of malaria attack. Table 3. PCR primer and probe sequences used in this study.                     Table S1. Malaria outcomes by age group within 16 months of follow-up. Table S2. Logistic regression analysis of the association of age and gut bacterial and fungal community structure with the risk of at least one asymptomatic Plasmodium parasitaemia episode. Table S3. Gut bacterial and fungal community structures according to the children's age groups. Table S4. Gut bacterial and fungal community structure in children who experienced, or did not experience, at least one malaria attack within 16 months of follow-up. Table S5. Gut bacterial and fungal community structure in children who developed, or did not develop, at least one asymptomatic Plasmodium parasitaemia episode within 16 months of follow-up.