Abstract
Seasonal malaria chemoprevention (SMC) is a novel strategy to reduce malaria infections in children. Infection with Plasmodium falciparum results in immune dysfunction characterized by elevated expression of markers associated with exhaustion, such as PD1 and LAG3, and regulatory CD4+FOXP3+ T cells.
In the current study, the impact of seasonal malaria chemoprevention on malaria-induced immune dysfunction, as measured by markers associated with exhaustion and regulatory T cells, was explored by flow cytometry.
Children that received seasonal malaria chemoprevention had fewer malaria episodes and showed significantly lower fold changes in CD4+PD1+ and CD4+PD1+LAG3+ compared to those that did not receive SMC. Seasonal malaria chemoprevention had no observable effect on fold changes in CD8 T cells expressing PD1 or CD160. However, children receiving SMC showed greater increases in CD4+FOXP3+ T regulatory cells compared to children not receiving SMC.
These results provide important insights into the dynamics of malaria-induced changes in the CD4 T-cell compartment of the immune system and suggest that the reduction of infections due to seasonal malaria chemoprevention may also prevent immune dysfunction.
NCT02504918.
Malaria caused by Plasmodium falciparum continues to threaten global health with devastating consequences. Recent statistics from the World Health Organization (WHO) estimate that there were over 219 million malaria infections, resulting in 435 000 deaths, in 2017 [1], indicating a stall in progress in malaria reduction over the last 3 years. Over 90% of all malaria-related deaths occurred in sub-Saharan Africa.
In the absence of an efficacious and long-lasting malaria vaccine, other strategies to reduce malaria morbidity and mortality have been introduced. These include distribution of bednets, indoor residual spraying, and targeted interventions in the most vulnerable groups such as intermittent preventative treatment for pregnant mothers and seasonal malaria chemoprevention (SMC) for children under the age of 5 years [2]. SMC has been shown to effectively reduce infections as well as clinical malaria cases in the Sahel region of Africa [3], where the malaria transmission season coincides with the commencement of the rainy season. The WHO-recommended SMC regimen consists of monthly dosing with amodiaquine and sulfadoxine-pyrimethamine for at least 4 months during the transmission season [2]. In 2017, 15.7 million children in 12 countries in the African Sahel received SMC, although an additional 13.6 million children could have benefited if additional funding had been available. Mali was one of the first countries in the Sahel region to implement SMC in 2012, which has now been gradually rolled out to cover the entire country.
Sterilizing immunity against P. falciparum infection does not develop even against repeated infections [4]; instead, there appears to be gradual acquisition of clinical immunity. In malaria-endemic regions, children under the age of 5 years experience the highest burden of the most severe forms of the disease (manifested as severe malarial anemia and cerebral malaria), while the prevalence of asymptomatic P. falciparum infections is greater in older children and adults [5, 6]. While the markers of clinical immunity remain poorly defined, they have been associated with reduced inflammatory responses and acquisition of strain-specific immunity to blood-stage P. falciparum antigens [7–10].
Malaria infection leads to increases in CD4 T cells expressing the immune exhaustion markers such as programmed cell death protein-1 (PD1) and lymphocyte activation gene-3 (LAG3) on T and B cells [11–13]. In animal models, blockade of PD1, as well as LAG3, results in clearance of plasmodium parasites [11, 14], suggesting that these markers hinder an effective immune response. Similarly, malaria infections have been associated with altered T regulatory cells (Tregs). While there are many different types of Tregs, forkhead box P3 (FOXP3)-expressing Tregs have been the most widely studied in the context of malaria [15, 16]. Treg responses are variable in the literature; some studies have shown increases in Treg numbers after infections, others have highlighted a trend of reduced risk of symptomatic malaria in children living in a high transmission area [15–17]. In animal models, expansion of Tregs has been shown to limit the early immune effector response to malaria, thereby assisting the pathogen to establish infection [18]. Together, these studies suggest that malaria infection leads to generalized immune dysfunction that may have consequences for subsequent responses to malaria, other diseases, and vaccines.
The effect of SMC on clinical immunity to malaria (and on the immune system in general) has not been extensively studied. One study in Mali showed decreases in antibodies to blood-stage antigens in children receiving SMC [19], while another study in Senegal showed no difference in antibody titers to P. falciparum antigens between children receiving SMC versus controls [17].
In the current study, the impact of SMC treatment over a whole malaria transmission season on the T-cell compartment was assessed. More specifically, Tregs and immune exhaustion markers were compared in children residing in neighboring villages in Mali: Beneko (SMC+) where SMC was being implemented and Ferekoroba (SMC−) where SMC had yet to be introduced. Malaria infections were dramatically reduced in children enrolled from Beneko compared to Ferekoroba. Consequently, the fold change in CD4+PD1+ or CD4+PD1+LAG3+ was significantly lower in the SMC+ group than the SMC− group over the course of transmission season, whereas Tregs were increased in the SMC+ group. These results highlight the beneficial impact of SMC in reducing malaria-induced immune dysfunction.
METHODS
Ethics
The study protocol was approved by Ethics committees in Mali (FMPOS) and the United States (ClinicalTrials.gov identifier: NCT02504918). The National Institute of Allergy and Infectious Diseases Institutional Review Board/Ethics Committee approved the study protocol. Written informed consent was obtained from the parents/guardians of the children as required.
Study Site
The study was conducted in 2 neighboring villages, Beneko and Ferekoroba, in the health district center of Ouelessebougou in Mali. In 2008, the incidence rate of clinical malaria in children under the age of 5 years was 1.99 episodes/child/year in Ouelessebougou and the incidence of severe malaria (as defined by WHO criteria [20]) was between 1% and 2% during the transmission season [3]. During the study period, SMC was being implemented in Beneko but not in Ferekoroba (where SMC was rolled out in 2016 after completion of the study).
Study Population
Children between the ages of 12 and 59 months from Beneko or Ferekoroba were included in the study. Children were clinically evaluated and those with significant anemia, unable to give blood, with chronic or debilitating conditions, immunosuppression, or chronic infection, or on co-trimoxazole were excluded from the study. Characteristics of study subjects are shown in Table 1.
Characteristics of Study Subjects
| Characteristics | SMC+ | SMC− |
|---|---|---|
| Sex, n (%) | ||
| Male | 27 (54.0) | 25 (50.0) |
| Female | 23 (46.0) | 25 (50.0) |
| Age, mo, mean (SD) | 33.2 (11.1) | 33.4 (13.5) |
| Hb type, n (%) | ||
| AA | 40 (80.0) | 34 (68.0) |
| AC | 8 (16.0) | 5 (10.0) |
| AS | 1 (2.0) | 9 (18.0) |
| CC | 0 (0.0) | 1 (2.0) |
| SC | 1 (2.0) | 1 (2.0) |
| Blood type, n (%) | ||
| A | 21 (42.0) | 5 (10.0) |
| AB | 1 (2.0) | 2 (4.0) |
| B | 15 (30.0) | 15 (30.0) |
| O | 13 (26.0) | 28 (56.0) |
| Infections >28 days apart, No. | 63 | 97 |
| Gametocytemia, mean (SD) | 10.18 (13.88) | 31.14 (82.62) |
| Parasitemia, mean (SD) | 1030.0 (1263) | 928.3 (1499) |
| Characteristics | SMC+ | SMC− |
|---|---|---|
| Sex, n (%) | ||
| Male | 27 (54.0) | 25 (50.0) |
| Female | 23 (46.0) | 25 (50.0) |
| Age, mo, mean (SD) | 33.2 (11.1) | 33.4 (13.5) |
| Hb type, n (%) | ||
| AA | 40 (80.0) | 34 (68.0) |
| AC | 8 (16.0) | 5 (10.0) |
| AS | 1 (2.0) | 9 (18.0) |
| CC | 0 (0.0) | 1 (2.0) |
| SC | 1 (2.0) | 1 (2.0) |
| Blood type, n (%) | ||
| A | 21 (42.0) | 5 (10.0) |
| AB | 1 (2.0) | 2 (4.0) |
| B | 15 (30.0) | 15 (30.0) |
| O | 13 (26.0) | 28 (56.0) |
| Infections >28 days apart, No. | 63 | 97 |
| Gametocytemia, mean (SD) | 10.18 (13.88) | 31.14 (82.62) |
| Parasitemia, mean (SD) | 1030.0 (1263) | 928.3 (1499) |
Unit for Gametocytemia: /500 WBC; Unit for Parasitemia: /300 WBC.
Abbreviations: Hb, hemoglobin; SMC, seasonal malaria chemoprevention.
Characteristics of Study Subjects
| Characteristics | SMC+ | SMC− |
|---|---|---|
| Sex, n (%) | ||
| Male | 27 (54.0) | 25 (50.0) |
| Female | 23 (46.0) | 25 (50.0) |
| Age, mo, mean (SD) | 33.2 (11.1) | 33.4 (13.5) |
| Hb type, n (%) | ||
| AA | 40 (80.0) | 34 (68.0) |
| AC | 8 (16.0) | 5 (10.0) |
| AS | 1 (2.0) | 9 (18.0) |
| CC | 0 (0.0) | 1 (2.0) |
| SC | 1 (2.0) | 1 (2.0) |
| Blood type, n (%) | ||
| A | 21 (42.0) | 5 (10.0) |
| AB | 1 (2.0) | 2 (4.0) |
| B | 15 (30.0) | 15 (30.0) |
| O | 13 (26.0) | 28 (56.0) |
| Infections >28 days apart, No. | 63 | 97 |
| Gametocytemia, mean (SD) | 10.18 (13.88) | 31.14 (82.62) |
| Parasitemia, mean (SD) | 1030.0 (1263) | 928.3 (1499) |
| Characteristics | SMC+ | SMC− |
|---|---|---|
| Sex, n (%) | ||
| Male | 27 (54.0) | 25 (50.0) |
| Female | 23 (46.0) | 25 (50.0) |
| Age, mo, mean (SD) | 33.2 (11.1) | 33.4 (13.5) |
| Hb type, n (%) | ||
| AA | 40 (80.0) | 34 (68.0) |
| AC | 8 (16.0) | 5 (10.0) |
| AS | 1 (2.0) | 9 (18.0) |
| CC | 0 (0.0) | 1 (2.0) |
| SC | 1 (2.0) | 1 (2.0) |
| Blood type, n (%) | ||
| A | 21 (42.0) | 5 (10.0) |
| AB | 1 (2.0) | 2 (4.0) |
| B | 15 (30.0) | 15 (30.0) |
| O | 13 (26.0) | 28 (56.0) |
| Infections >28 days apart, No. | 63 | 97 |
| Gametocytemia, mean (SD) | 10.18 (13.88) | 31.14 (82.62) |
| Parasitemia, mean (SD) | 1030.0 (1263) | 928.3 (1499) |
Unit for Gametocytemia: /500 WBC; Unit for Parasitemia: /300 WBC.
Abbreviations: Hb, hemoglobin; SMC, seasonal malaria chemoprevention.
Study Visits, Administration of SMC, and Sample Collection
The study was conducted over a 5-month period starting in August 2015 (month 0) until December 2015 (month 4). Children recruited to the study were clinically examined at each study visit. Blood smears were prepared and used to determine malaria infections. Sulfadoxine-pyrimethamine plus amodiaquine were administered monthly from August to November 2015 as per the guidelines of the WHO [2]. Children with symptomatic malaria were treated with artemether-lumefantrine and resumed SMC treatment at their next monthly visit. The participants were also seen on unscheduled visits at any point during the study.
During each scheduled and unscheduled visit, a 5 mL venous blood sample was collected from study participants. An aliquot of this sample was used for ex vivo flow cytometry, preparation of blood smears, and for hemoglobin (Hb) measurements. The remaining blood was used for peripheral blood mononuclear cell isolation.
Laboratory Evaluation
Malaria parasitemia and gametocytemia was determined in Giemsa-stained blood smears and reported against 300 and 500 white blood cells (WBCs), respectively, by expert microscopists. When prompt blood smear microscopy was not available, such as during SMC delivery, a rapid diagnostic test was used for malaria diagnosis in order to allow prompt treatment. All diagnoses were later confirmed by blood smear microscopy, and only blood smear-positive results were considered as patent parasitemia for analysis purposes.
Flow Cytometry
Whole blood (150 µL) was used to stain for surface markers using conjugated monoclonal antibodies against anti-CD160-PE, anti-CD4-PE-CY5, anti-PD1-APC, anti-CD3-Alexa-700, anti-CD8- APC-CY7, and anti-LAG3-Pacific Blue. Cells were incubated for 20 minutes, washed with phosphate-buffered saline, fixed, permeabilized, and then stained with anti-FOXP3-FITC. All samples were acquired on a BD LSR II flow cytometer and analyzed using FlowJo software (version 10.1; Figure 1). The data from flow cytometry were expressed as fold change from the baseline (day 0).
Gating strategy used to identify CD4+PD1+, CD4+PD1+LAG3+, CD4+FOXP3+, CD8+PD1+, and CD8+CD160+ populations within the CD3+ gate. Abbreviations: FSC, forward scatter; FOXP3, forkhead box P3; LAG3, lymphocyte activation gene-3; PD1, programmed cell death protein-1; SSC, side scatter.
Statistical Analysis
Χ2 and time to parasitemia analyses (log-rank test) were performed using GraphPad Prism version 7.00 for Windows.
T-cell data were expressed as fold change from baseline and assessed for extreme outlying observations, defined as observations in excess of 3 standard deviations above the mean for that T-cell subset. None of the observations for CD4+PD1+, CD4+PD1+ LAG3+, or CD4+FOXP3+ were in excess of this cleaning cutoff.
Fold change of each variable was compared between the groups at each individual time point and analyzed using the nonparametric Kruskal-Wallis test. All P values were 2-sided and P values <.05 were considered to be statistically significant.
Generalized estimating equations (GEE) were used to model changes in T-cell subsets between the SMC+ and SMC− groups over the whole season. Univariate analysis was conducted for the following T-cell subsets: CD4+PD1+, CD4+PD1+LAG3+, CD4+FOXP3+, CD8+PD1+, CD8+CD160+, CD8+CD160+PD1+, and CD8+CD160−PD1+ (Supplementary Table 1). Multivariate models were conducted for CD4+PD1+, CD4+PD1+LAG3+, and CD4+FOXP3+ with predictors for age, sex, Hb type, and blood group (Table 2). These analyses were done using SAS version 9.4.
Generalized Estimating Equation Output for Multivariate Models
| T cell | Parameters | β Estimate | SE | 95% CI | P |
|---|---|---|---|---|---|
| Total CD4+PD1+ QIC = 563.36 | Treatment | −0.2671 | 0.0681 | (−0.4006 to −0.1336) | <.0001 |
| Age | −0.0016 | 0.0023 | (−0.0060 to 0.0028) | .4743 | |
| Sex (Female) | −0.0453 | 0.0603 | (−0.1635 to 0.0728) | .4523 | |
| Hb (AA) | 0.0968 | 0.069 | (−0.0385 to 0.2321) | .1609 | |
| Blood type (O) | −0.0959 | 0.0654 | (−0.2240 to 0.0323) | .1425 | |
| CD4+PD1+LAG3+ QIC = 551.4030 | Treatment | −1.4725 | 0.7264 | (−2.8962 to 0.0489) | .0426 |
| Age | −0.0207 | 0.0183 | (−0.0565 to 0.0151) | .2565 | |
| Sex (Female) | −0.0501 | 0.5739 | (−1.1751 to 1.0748) | .9304 | |
| Hb (AA) | 0.718 | 0.603 | (−0.4639 to 1.8999) | .2338 | |
| Blood type (O) | −0.8097 | 0.6298 | (−2.0441 to 0.4247) | .1986 | |
| Total CD4+FOXP3+ QIC = 535.2059 | Treatment | 0.2628 | 0.0828 | (0.1005 to 0.4252) | .0015 |
| Age | −0.003 | 0.0029 | (−0.0086 to 0.0027) | .3075 | |
| Sex (Female) | 0.0226 | 0.0747 | (−0.0086 to 0.1689) | .7626 | |
| Hb (AA) | 0.0966 | 0.0894 | (−0.0787 to 0.2719) | .2804 | |
| Blood type (O) | −0.0394 | 0.0834 | (−0.2027 to 0.1240) | .6367 |
| T cell | Parameters | β Estimate | SE | 95% CI | P |
|---|---|---|---|---|---|
| Total CD4+PD1+ QIC = 563.36 | Treatment | −0.2671 | 0.0681 | (−0.4006 to −0.1336) | <.0001 |
| Age | −0.0016 | 0.0023 | (−0.0060 to 0.0028) | .4743 | |
| Sex (Female) | −0.0453 | 0.0603 | (−0.1635 to 0.0728) | .4523 | |
| Hb (AA) | 0.0968 | 0.069 | (−0.0385 to 0.2321) | .1609 | |
| Blood type (O) | −0.0959 | 0.0654 | (−0.2240 to 0.0323) | .1425 | |
| CD4+PD1+LAG3+ QIC = 551.4030 | Treatment | −1.4725 | 0.7264 | (−2.8962 to 0.0489) | .0426 |
| Age | −0.0207 | 0.0183 | (−0.0565 to 0.0151) | .2565 | |
| Sex (Female) | −0.0501 | 0.5739 | (−1.1751 to 1.0748) | .9304 | |
| Hb (AA) | 0.718 | 0.603 | (−0.4639 to 1.8999) | .2338 | |
| Blood type (O) | −0.8097 | 0.6298 | (−2.0441 to 0.4247) | .1986 | |
| Total CD4+FOXP3+ QIC = 535.2059 | Treatment | 0.2628 | 0.0828 | (0.1005 to 0.4252) | .0015 |
| Age | −0.003 | 0.0029 | (−0.0086 to 0.0027) | .3075 | |
| Sex (Female) | 0.0226 | 0.0747 | (−0.0086 to 0.1689) | .7626 | |
| Hb (AA) | 0.0966 | 0.0894 | (−0.0787 to 0.2719) | .2804 | |
| Blood type (O) | −0.0394 | 0.0834 | (−0.2027 to 0.1240) | .6367 |
Abbreviations: CI, confidence interval; Hb, hemoglobin; QIC, Quasilikelihood Information Criteria; SE, standard error.
Generalized Estimating Equation Output for Multivariate Models
| T cell | Parameters | β Estimate | SE | 95% CI | P |
|---|---|---|---|---|---|
| Total CD4+PD1+ QIC = 563.36 | Treatment | −0.2671 | 0.0681 | (−0.4006 to −0.1336) | <.0001 |
| Age | −0.0016 | 0.0023 | (−0.0060 to 0.0028) | .4743 | |
| Sex (Female) | −0.0453 | 0.0603 | (−0.1635 to 0.0728) | .4523 | |
| Hb (AA) | 0.0968 | 0.069 | (−0.0385 to 0.2321) | .1609 | |
| Blood type (O) | −0.0959 | 0.0654 | (−0.2240 to 0.0323) | .1425 | |
| CD4+PD1+LAG3+ QIC = 551.4030 | Treatment | −1.4725 | 0.7264 | (−2.8962 to 0.0489) | .0426 |
| Age | −0.0207 | 0.0183 | (−0.0565 to 0.0151) | .2565 | |
| Sex (Female) | −0.0501 | 0.5739 | (−1.1751 to 1.0748) | .9304 | |
| Hb (AA) | 0.718 | 0.603 | (−0.4639 to 1.8999) | .2338 | |
| Blood type (O) | −0.8097 | 0.6298 | (−2.0441 to 0.4247) | .1986 | |
| Total CD4+FOXP3+ QIC = 535.2059 | Treatment | 0.2628 | 0.0828 | (0.1005 to 0.4252) | .0015 |
| Age | −0.003 | 0.0029 | (−0.0086 to 0.0027) | .3075 | |
| Sex (Female) | 0.0226 | 0.0747 | (−0.0086 to 0.1689) | .7626 | |
| Hb (AA) | 0.0966 | 0.0894 | (−0.0787 to 0.2719) | .2804 | |
| Blood type (O) | −0.0394 | 0.0834 | (−0.2027 to 0.1240) | .6367 |
| T cell | Parameters | β Estimate | SE | 95% CI | P |
|---|---|---|---|---|---|
| Total CD4+PD1+ QIC = 563.36 | Treatment | −0.2671 | 0.0681 | (−0.4006 to −0.1336) | <.0001 |
| Age | −0.0016 | 0.0023 | (−0.0060 to 0.0028) | .4743 | |
| Sex (Female) | −0.0453 | 0.0603 | (−0.1635 to 0.0728) | .4523 | |
| Hb (AA) | 0.0968 | 0.069 | (−0.0385 to 0.2321) | .1609 | |
| Blood type (O) | −0.0959 | 0.0654 | (−0.2240 to 0.0323) | .1425 | |
| CD4+PD1+LAG3+ QIC = 551.4030 | Treatment | −1.4725 | 0.7264 | (−2.8962 to 0.0489) | .0426 |
| Age | −0.0207 | 0.0183 | (−0.0565 to 0.0151) | .2565 | |
| Sex (Female) | −0.0501 | 0.5739 | (−1.1751 to 1.0748) | .9304 | |
| Hb (AA) | 0.718 | 0.603 | (−0.4639 to 1.8999) | .2338 | |
| Blood type (O) | −0.8097 | 0.6298 | (−2.0441 to 0.4247) | .1986 | |
| Total CD4+FOXP3+ QIC = 535.2059 | Treatment | 0.2628 | 0.0828 | (0.1005 to 0.4252) | .0015 |
| Age | −0.003 | 0.0029 | (−0.0086 to 0.0027) | .3075 | |
| Sex (Female) | 0.0226 | 0.0747 | (−0.0086 to 0.1689) | .7626 | |
| Hb (AA) | 0.0966 | 0.0894 | (−0.0787 to 0.2719) | .2804 | |
| Blood type (O) | −0.0394 | 0.0834 | (−0.2027 to 0.1240) | .6367 |
Abbreviations: CI, confidence interval; Hb, hemoglobin; QIC, Quasilikelihood Information Criteria; SE, standard error.
RESULTS
Baseline Characteristics and Retention of Study Subjects
Fifty children from Beneko (SMC+) and 50 children from Ferekorobo (SMC−) villages were enrolled in the study that started in August 2015 (month 0), and baseline measurements were taken before SMC administration in Beneko. There was no difference in sex, age, or Hb variants in children recruited to the study between the SMC+ and SMC− groups. There was a greater number of blood group A children in the SMC+ versus SMC− group, while the SMC− group had a greater number of blood group O children (Table 1). Data from 3 volunteers in the SMC− group who only visited the clinic twice during the study period were excluded from the longitudinal analysis (Supplementary Figure 1). SMC treatment was well tolerated in the study population with only 7 reported instances of vomiting and only 1 of these was graded as a mild adverse event, from the 150 doses administered over the study period. Only 1 dose was considered unadministered as the child vomited just after receiving SMC as well as the replacement dose.
Comparison of Infections and Clinical Malaria in the SMC+ and SMC− Groups
Malaria transmission in Mali is highly seasonal and its onset coincides with the rainy season. Over this season, the frequency of malaria infections was significantly lower in children in the SMC+ group compared to the SMC− group (63 versus 97, X2P < .001; Table 1 and Figure 2A). Similarly, the proportion of clinical malaria was significantly lower in the children in the SMC+ compared to the SMC− group (Figure 2B). Furthermore, the SMC+ group had a significantly greater time to first infection than the SMC− group (89 vs 31 days median time to infection, P < .0001 log-rank test; Figure 3A). Time to clinical malaria was also significantly increased in the SMC+ children (P < .0001 log-rank test; Figure 3B).
Proportion of malaria infections (A) and clinical malaria episodes (B) during the follow-up months in the children enrolled in the SMC+ and SMC− groups. Differences between the groups were evaluated using the X2 test. Abbreviation: SMC, seasonal malaria chemoprevention.
Comparison of (A) time to first positive blood smear and (B) time to first clinical malaria infection in children in the SMC+)and control group, SMC−. Data were analyzed using the log-rank test. Abbreviation: SMC, seasonal malaria chemoprevention.
Reduction in CD4+PD1+ and CD4+PD1+LAG3+ T Cells in Children Receiving SMC
Ex vivo whole-blood staining was used to enumerate the percentage of CD4 and CD8 T cells expressing PD1, LAG3, FOXP3, and CD160 in all children and at each visit during the study period. There was no significant difference observed in CD8 T cells expressing PD1 at any time point (Supplementary Figure 2 and Supplementary Table 1) between the SMC+ and SMC− groups, while LAG3 and FOXP3 expression was only seen in CD4 T cells. Hence, the subsequent analysis focused on the CD4 T-cell compartment.
In the SMC− group, fold change from baseline in CD4+PD1+ T cells increased 1 month postenrollment followed by a gradual decrease in fold change. In the SMC+ group, CD4+PD1+ T-cells level remained similar 1 month postenrollment followed by a sharp decrease with a fold change of 0.5 after 2 doses of SMC.
Differences between SMC+ and SMC− groups achieved significance in month 2 and 3 (P < .0001 and .0001, respectively, Kruskal-Wallis test). Overall, during the study period the change in CD4+PD1+ T cells was 0.24-fold lower in the SMC+ group compared to the SMC− group (confidence interval [CI], 0.11–0.37; P = .0005; Figure 4A and Supplementary Table 1). A multivariate analysis was performed using fold change in CD4+PD1+ T-cell levels relative to baseline as the dependent variable and adjusted for age, sex, Hb type, and blood group. The point estimate (β coefficient) was −0.27 (CI, −0.40 to 0.13; P < .0001) when comparing the SMC+ to the SMC− group over the course of the study (Table 2).
Mean fold change from baseline of (A) CD4+PD1+ and (B) CD4+PD1+LAG3+ during the study period in the SMC+ and SMC− groups. Data are represented as mean fold change and error bars denote the 95% confidence interval. Generalized estimating equations were used to analyze differences between the groups. Abbreviations: LAG3, lymphocyte activation gene-3; PD1, programmed cell death protein-1; SMC, seasonal malaria chemoprevention.
Fold change from baseline in CD4+PD1+LAG3+ T cells increased in both SMC+ and SMC− groups 1 month postenrollment but to a greater degree in the SMC− group with a mean fold change of 4.0 and 1.8 in SMC− and SMC+ groups, respectively. Although throughout the season fold change in CD4+PD1+LAG3+ T cells was higher in the SMC− compared to the SMC+ group, differences between the 2 groups achieved significance only in month 2 (P = .0389 Kruskal-Wallis test; Figure 4B). A multivariate analysis was performed using fold change in CD4+PD1+LAG3+ T-cell levels relative to baseline as the dependent variable and adjusted for age, sex, Hb type, and blood group. The point estimate was –1.47 (CI, −2.9 to 0.05; P < .04) when comparing the SMC+ to the SMC− group over the course of the study (Table 2).
Increase in CD4+FOXP3+T Cells in the SMC+ Versus the SMC− Group
Tregs, defined as CD4 T cells that coexpress the transcription factor FOXP3, can be induced during P. falciparum infections [16, 21]. In the SMC+ group, Tregs increased from baseline levels at each monthly interval, peaking at a mean fold increase of 1.53 by the end of the malaria transmission season in month 4. In contrast, in the SMC− group, Tregs decreased from baseline levels in months 1 and 2, spiked in month 3, and were 0.6-fold lower at the end of the malaria transmission season in month 4. Differences between SMC+ and SMC− groups achieved significance in months 2 and 4 (P < .0003 and <.0001, respectively, Kruskal-Wallis test; Figure 5). In a univariate analysis using fold change of CD4+FOXP3+ T cells from baseline, the point estimate was 0.37-fold (CI, 0.17–0.57; P = .0004) when comparing the SMC+ and SMC− groups (Supplementary Table 1). In a multivariate GEE analysis using fold change of CD4+FOXP3+ T cells from baseline, that adjusted for age, sex, Hb type, and blood group, the point estimate was 0.26 (CI, 0.1–0.43; P = .0015) when comparing the SMC+ to the SMC− group (Table 2).
Mean fold change from baseline of CD4+FOXP3+ during the study period in the SMC+ and SMC− groups. Error bars represent the 95% confidence intervals. Generalized estimating equations were used to analyze differences between the groups. Abbreviations: FOXP3, forkhead box P3; SMC, seasonal malaria chemoprevention
DISCUSSION
This study capitalized on the fact that SMC was rolled out at different times in 2 neighboring villages, in the health district center of Ouelessebougou in Mali, to compare the impact of P. falciparum infection on immune exhaustion and T regulatory markers. SMC administration in children resulted in a 37% reduction of malaria infections and almost halved the number of clinical cases. These results are consistent with previous studies that similarly showed a dramatic reduction in malaria cases in children receiving SMC [3, 22–25]. While these results were encouraging, almost half of children receiving SMC were still infected despite receiving prophylaxis, highlighting the need for multipronged approaches to achieve the goal of malaria eradication.
Malaria infections have been linked with immune dysfunction that results in greater susceptibility to bacterial infections [26] and loss of vaccine-induced immunity [27]. Malaria-induced upregulation of PD1 on CD4 T cells is well established [11, 12, 15]. The design of the current study allowed us to measure the impact of SMC on malaria-induced upregulation of PD1 and LAG3 on CD4 and CD8 T cells. Over the course of the study, children receiving SMC had lower increases from baseline of CD4+PD1+ and CD4+PD1+LAG3+ T cells compared to those in the control group. There were no differences observed in CD8+PD1+ T cells between the 2 groups. These findings suggest that malaria infections had a more pronounced effect on CD4 T cells than CD8 T cells. A potential caveat to the interpretation of this finding is the fact that γδ T cells can also express CD8 and PD1 [28–30] but they were not evaluated during the study due to technical constraints. Recent studies have shown that malaria exposure results in expansion of CD4+PD1+ T cells that coexpress another inhibitory ligand, cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), associated with high interleukin-10 (IL-10) production [31, 32]. Furthermore, malaria-induced PD1 upregulation has been linked with alterations in the memory B-cell compartment [12]. While the expression of CTLA-4, IL-10, or memory B cells was not measured in this study, it could be hypothesized that children receiving SMC would have lower levels of circulating IL-10 and preserved levels of memory B cells compared to the control group.
The frequency of Tregs, defined as CD4 T cells that coexpress FOXP3, was evaluated here. The suppressive activity of Tregs during a malaria infection may prevent hyperimmune activation but could result in a failure to generate immune memory that protects from future infections [18]. Furthermore, the dynamics of Tregs during malaria infection have been reported to be higher, lower, or unchanged [16, 33]. In the current study, children receiving SMC had a gradual increase in Tregs over the course of the season, while in the control group Tregs declined over the first 2 months. The sudden increase in Treg levels in the fourth month of the study was unexpected and the cause unknown. Unlike the CD4+PD1+ T-cell subset that was almost identical in the 2 groups at the end of the study, Treg levels were markedly higher in the SMC+ group compared to the SMC− group at the end of the malaria transmission season. These results suggest that Tregs are induced in P. falciparum-exposed children but not in children that develop active infections. In animal models, Tregs generated during malaria infections have been shown to impede both early and long-term responses [18]. These findings also raise the concern about whether children receiving SMC may have delayed acquisition of clinical immunity against malaria. These questions will be important to address as SMC coverage is expanded across the Sahel region.
While the current study highlights some important changes in T cells, there were some limitations. The study relied on phenotypic characterization of markers associated with immune exhaustion and regulation on CD4 and CD8 T cells. The current findings can serve as a primer for further functional characterization of specific immune subsets on specific P. falciparum immunity in these children. Future studies have been designed to evaluate these same markers throughout the dry season and the following transmission season in children as well as in adults living in Mali.
Overall, results show that introduction of SMC led to a reduction in CD4+PD1+ and an increase in CD4+FOXP3+ Tregs over the course of a malaria transmission season. These findings underscore the need to reduce malaria transmission and potential immune dysfunction in children living in endemic countries.
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Presented in part: American Society of Tropical Medicine and Hygiene 65th Annual Meeting, 13–17 November 2016, Atlanta, Georgia.
Acknowledgment. The authors thank J. Patrick Gorres for assistance in writing and manuscript preparation.
Financial support. This work was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health (ZIA AI001147-09) as well as the Bill and Melinda Gates Foundation (OPP1134361).
Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
