Rosetting correlates with total immunoglobulin M but not IgG in patients infected with P. vivax

Immunoglobulins are among the plasma components that influence rosetting in P. falciparum¹¹. Therefore, total immunoglobulin M (IgM) and G (IgG) were quantified in plasma from 38 P. vivax patients. IgM levels ranged from 1059 to 4628 μg/mL (mean=3070 μg/mL), and IgG varied from 5498 to 12,669 μg/mL (mean=90,608 μg/mL). Rosetting was positively correlated with IgM levels (Pearson correlation coefficient r=0.34, p<0.05) (Fig. 2A), but not with IgG levels (Pearson correlation coefficient r= − 0.05, p=0.74) (Fig. 2B). The ability of IgM and IgG in P. vivax-infected patients to bind to noninfected RBCs was next tested. However, neither the binding of IgM or IgG to noninfected RBC correlated with rosetting rates in P. vivax patients (Supplementary Fig. ^S3).

Open in a separate window

Figure 2

Total IgM but not IgG correlated with rosetting in vivax malaria. (A) Correlation of total IgM (Pearson correlation coefficient r=0.34, p<0.05, n=38) and (B) total IgG (Pearson correlation coefficient r= − 0.05, p=0.74, n=38) from malaria infected patients and rosetting rate of P. vivax isolates. CI: confidence interval.

Naturally acquired antibodies to merozoite antigens and rosetting in vivax malaria

Because rosetting was correlated with total IgM, the relationship of specific P. vivax-related antibodies with rosetting was investigated. The presence of antibodies against two merozoite proteins, PvAMA-1 (Fig. 3A, n=38) and PvMSP-1 (Fig. 3B, n=38), was investigated. Most individuals had IgG1 antibodies to both proteins (57.9%, n=22). The presence of anti-PvAMA-1 IgG1 antibodies was detected in 57.9% (n=22) of individuals; 21.1% (n=8) had IgG3; 10.5% (n=4) had IgG4 and 7.9% had IgG2 (n=3). The mean of rosetting was 46.97% in P. vivax isolates from individuals that had IgG1 antibodies to PvAMA-1 and 39.62% in P. vivax isolates from individuals with no detectable IgG1 antibodies. The mean of rosetting in P. vivax isolates from individuals with IgG3 antibodies to PvAMA-1 was 50.23% while in the group with no detectable IgG3 antibodies was 42.18%. Regarding the presence of antibodies to PvMSP-1, most samples had IgG1 and IgG3 antibodies, whereas 84.21% (n=32) of samples were detected with IgG1; 86.84% (n=33) with IgG3; 31.6% (n=12) with IgG2 and 73.68% (n=28) with IgG4. Infected individuals with detectable IgG1 antibodies to PvMSP-1 had a mean of rosetting of 44.22% while individuals whithout IgG1 antibodies had P. vivax rosettes with a mean of 42.04%. The mean of P. vivax rosetting was 45.74% in individuals with IgG3 against PvMSP-1 and 31.62% in individuals with no detectable antibodies.

Open in a separate window

Figure 3

Naturally acquired IgG and IgM antibodies to merozoite antigens and its relationship with rosetting. (A) Boxplot showing median and interquartile ranges of scaled reactivity index for IgG subclasses (IgG1, IgG2, IgG3 and IgG4) and IgM towards PvAMA-1 and (B) PvMSP-1 recombinant proteins. Open circles denote outliers observations.

Parasitemia, IL-6 and IL-10 are increased in patients with high frequency of P. vivax rosettes

Other parameter evaluated among studied isolates was parasitemia. When rosetting rate was compared with parasitemia a positive correlation was found (Spearman’s correlation coefficient rho=0.40, p<0.05, n=37) (Fig. 4A).

Open in a separate window

Figure 4

Parasitemia and plasmatic levels of IL-6 and IL-10 correlated with rosetting in vivax malaria. (A) Positive correlation between peripheral parasitemia and rosetting (Spearman’s correlation coefficient rho=0.40, p=<0.05, n=37). (B) Correlation of P. vivax rosetting and log IL-6 (Spearman’s correlation coefficient rho=0.46, p<0.05, n=33). (C) log IL-10 (Spearman’s correlation coefficient rho=0. 46, p<0.05 and n=33) and (D) IFN-γ (Student’s t-test, p=0.48 and n=38). CI: confidence interval.

Parasitemia has being associated with increased levels of specific cytokines in vivax malaria¹². Considering that rosetting levels were associated with parasitemia the role of plasmatic cytokines in rosetting was investigated. Plasmatic cytokine IFN-γ, IL-6 and IL-10 levels were elevated in infected patients compared to noninfected individuals. Concentrations of IL-6, IL-10 and IFN-γ varied from 0.1 to 10,489.0 ρg/mL, 0 to 9416.75 ρg/mL and 0 to 26.8 ρg/mL, respectively. Five and four individuals had atypical values for IL-6 and IL-10, respectively, and were considered outliers, and one individual had no detectable value for IL-10 and therefore disregarded in further analysis. The logarithm of IL-6 and IL-10 were positively correlated with rosetting rates (IL-6, Spearman’s correlation coefficient rho=0.46, p<0.05 and n=33; IL-10, Spearman’s correlation coefficient rho=0.46, p< 0.05 and n=33)  (Fig. 4B,C) while IFN-γ level was not associated with rosetting (Student’s t-test, p=0.48 and n=38) (Fig. 4D). In addition, the logarithm of IL-6 and IL-10 were also positively correlated (Spearman’s correlation coefficient rho=0.54, p<0.05, n=30) (Supplementary Fig. ^S4A). However, no association were found between cytokines and parasitemia (Supplementary Fig. ^S4B–D).

Genes related to phagocytosis pathway are differentially expressed in peripheral blood mononuclear cells of patients infected with P. vivax forming rosettes

Transcriptome analysis was performed on peripheral leukocytes from P. vivax infected patients with very low or moderate rosetting (Table ​(Table1).1). Five genes that showed a log2 fold change higher than 2 were identified (Table ​(Table2,2, Fig. 5). Three out of five genes were in the Fc gamma receptor (FCGR)-dependent phagocytosis pathway. Immunoglobulin kappa constant (IGKC) and immunoglobulin heavy constant gamma 1 (IGHG1) were upregulated and actin-related protein 2/3 complex subunit 2 (ARPC2) was downregulated in individuals with moderate rosetting compared to patients with low rosetting.

Open in a separate window

Figure 5

RNAseq analysis of PBMCs from patients infected with P. vivax. Volcano plot showing the range of the log2(fold change) relative to the − log2(q-values) of mapped genes from PBMCs of P. vivax infected patients between moderate and low rosetting isolates differential expression data analysis. Identified genes with a p-value<0.05, q-value<0.3 and log2 (fold change)>2 cut-offs obtained from RNAseq differential gene expression analysis are put in evidence (dark red dots).

Blood sample collection

Blood samples were obtained from malaria patients presenting at the Hospital of the Fundação de Medicina Tropical Dr, Heitor Vieira Dourado (Manaus, Amazonas State, Brazil). Prior to sample collection, patients granted informed consent. Blood samples were collected using BD Vacutainer tubes with sodium citrate anticoagulant. A thin blood smear was prepared from each sample to determine species of malaria parasites involved, parasitemia and the predominant erythrocytic stage of the parasite. Blood cell count was performed immediately after blood collection using a Sysmex KX21N (Sysmex Corporation-Roche, Japan). After collection of peripheral blood, patients received treatment following national guidelines with chloroquine plus primaquine.

Parasite isolation and enrichment

Once microscopic diagnosis of uncomplicated vivax malaria was made and before the treatment was initiated, 8 mL of blood was collected into citrate-coated Vacutainer tubes (BD). Blood was immediately processed to obtain enriched Pv-iRBCs. Immediately after collection, the RBCs containing trophozoites and schizonts were separated from the younger forms on a 45% Percoll (GE Healthcare) gradient as previously described⁵.

Rosetting assay

IRBCs (20 µL) at 2.5–5% parasitemia and 2.5–5% hematocrit were initially washed three times with McCoy’s 5A medium (400 g, 5 min., at room temperature). After that, iRBC were incubated for 40 min at 37 °C in rosetting medium (McCoy’s 5A medium supplemented with 20% of autologous plasma). Duplicated samples were stained with 45 µg/ml acridine orange and examined by direct light and fluorescence microscopy (Nikon Eclipse 50i, filter 96311 B-2E/C). Rosetting was assessed by counting 200 iRBCs, in duplicate. Slides were counted at diagonal vision, to balance for a possible irregular distribution of rosettes in the slide. A rosette was determined by the binding of two or more uninfected erythrocytes to an iRBC (Supplementary Fig. ^S5). To assess the involvement of plasma factors in the rosette formation, the plasma in rosetting medium was substituted for 0,5% of Albumax II. For heat-inactivated plasma, samples were heat inactivated for 30 min at 56 °C. Nonimmune heterologous plasma evaluated was from a donor without history of malaria and no detectable antibodies to PvAMA-1 or PvMSP-1 (reactivity index<1).

Measurements of plasma cytokines and IgG and IgMlevels

The levels of IL-2, IL-4, IL-6, IL-10 TNF-α and IFN-γ were quantified in cryopreserved plasma samples using the Cytometric Bead Array system (CBA; BD Biosciences) following the manufacturer’s instructions. Samples were analyzed using a BD FACSCalibur (BD Biosciences, USA). Standard curves were derived from cytokine standards for each cytokine analyzed. Total IgG and IgM were measured from plasma using an ELISA quantitation kit (Bethyl Laboratories) following the manufacturer’s recommendations.

IgG and IgM red blood cell binding

Briefly, erythrocytes (A+) at 0.8% hematocrit in RPMI medium were incubated with plasma from P. vivax-infected patients at 1:10 dilution for 1 h. After three washing steps with RPMI, erythrocytes were incubated for 30 min with anti-IgG or anti-IgM Alexa Fluor 488 conjugate (ThermoFisher Scientific). Erythrocytes were washed thrice and suspended in PBS solution, and 100.000 events were acquired using a BD FACSCanto II (BD Biosciences, USA). Flow cytometry results were analyzed using FlowJo software.

Naturally acquired IgG antibody subclasses of antibodies to PvAMA-1 and PvMSP-1₁₉ merozoite antigens

Specific IgG subclasses of antibodies (IgG1, IgG2, IgG3 and IgG4) to PvAMA-1 and PvMSP-1₁₉ were detected in plasma by enzyme-linked immunosorbent assays (ELISAs). PvAMA-1 and PvMSP-1₁₉ recombinant proteins were expressed and purified as previously described^36,37. ELISAs were conducted as previously described³⁸. The results for IgG subclasses were expressed as reactivity indices (RIs), which were calculated by dividing the mean optical density (OD) values of tested samples by the mean OD values plus three standard deviations of 20 non-exposed individuals living in nonendemic areas of malaria. RI values>1 were considered positive.

RNAseq of peripheral blood mononuclear cells

mRNA from PBMCs isolated from total blood of patients infected with P. vivax with high (n=2) or low (n=2) rosetting was analyzed by RNAseq. For this specific approach, moderate rosetting isolates were considered as 20% rosettes, and low rosetting isolates were considered as less than 10% rosettes. RNA extractions were performed a RNeasy Micro Kit (Qiagen) according to the manufacturer’s instructions. Before conducting Whole Transcriptome Shotgun Sequence (WTSS), RNA quality was accessed using an Agilent 2100 Bioanalyzer as well as Agilent RNA 6000 Pico Kit reagents and chips, and it was analyzed using 2100 Expert software according to the recommendations of Agilent Technologies. A SMART-Seq V4 Ultra Low Input RNA Kit was used for sequencing via Clontech’s patented Switching Mechanism at 5′ End of RNA Template (SMART) technology. cDNA quality, quantity and size range were evaluated using the BioAnalyzer platform from Agilent Technologies, Inc. using the Agilent High Sensitivity DNA Kit (cDNA, 5 to 500 pg/µL within a size range of 50 to 7000 bp) following the manufacturer’s instructions. Prior to generating the final library for Illumina sequencing, the Covaris AFA system was used for controlled cDNA shearing. cDNA output was then converted into sequencing templates suitable for cluster generation and high-throughput sequencing through the Low Input Library Prep v2 (Clontech Laboratories, Inc.; Takara Bio Company). Library quantification procedures were performed by qPCR using the Library Quantification Kit (Clontech Laboratories, Inc.; Takara Bio Company) and Agilent's High Sensitivity DNA Kit (Agilent Technologies, Inc.), and they were successfully completed before proceeding to the pool setup at a final concentration of 2 nM for direct sequencing. The library was sequenced on a HiSeq 2500 sequencer on Rapid Run mode with the HiSeq Rapid Cluster Kit v2 (100×100) Paired End and HiSeq Rapid SBS Kit v2 (200 cycles). The generated libraries were cluster amplified and sequenced on the Illumina platform using standard Illumina reagents and protocols for multiplexed libraries following the manufacturer’s loading recommendations. On average, approximately 474.553 paired end reads were obtained per sample from the 4 sample libraries sequenced. The RNAseq raw reads were checked for quality by running Fast Quality Control (FastQC—https://www.bioinformatics.babraham.ac.uk/projectY/fastqc/). The reads were then subjected to a RNAseq alignment v1.1.1 workflow from BaseSpace platform for cloud-based genomics analysis and storage, and they were integrated with Illumina sequencers. This workflow allowed trimming of Illumina adaptors, read mapping using TopHat 2 (Bowtie 2) aligner towards the Homo sapiens UCSC hg38 (RefSeq & Gencode gene annotations)³⁹, and FPKM estimation of reference genes and transcripts using Cufflinks 2. Final assembly and analysis of differentially expressed reference transcripts were performed with Cuffdiff 2 within the Cufflinks Assembly & DE pipeline v2.1.0. Differential gene expression between moderate and low rosetting groups was identified after a pairwise Wilcoxon test was used to compare the transcriptional profiles with the following cutoffs: p-value<0.05, q-value<0.5 and a log2 fold change>2.

Phagocytosis assay with P. vivax

THP-1 cells were grown on 8-well culture slides. Maturation was induced by incubation with 60 ng/mL phorbol 12-myristate 13-acetate (PMA) (CalbiochemH, San Diego, CA) for 24 h at 37 °C. The supernatant was then removed, and cells were washed twice with cell culture medium. In parallel, parasites were kept at room temperature in static conditions (intact rosettes) or agitated using a vortex for 30 min after complete disruption of rosettes using a needle (disrupted rosettes), in the presence of autologous plasma. P. vivax isolates (1×10⁶) at 5–8% parasitemia were added to each well of THP-1 cells. After 30 min of incubation in 5% CO₂ atmosphere at 37 °C, THP-1 cells were harvested and washed three times in cell culture medium, and the slides were fixed in methanol and stained with panoptic stain. At least 200 THP-1 cells were counted in each well, and all samples were tested in duplicated. THP-1 cell culture was checked for Mycoplasma contamination weekly.

Twenty mL of total blood were used to separate PBMC using Ficoll-Paque-PLUS following the fabricant instructions. After isolation, cells were counted at optical microscopy and 1×10⁶ cells were added to a Labtek slide and kept for 4 h at 37 °C. Non-adherent cells were removed and phagocytosis assay was performed. Immediately before phagocytosis assay, parasites collect from the same patients were kept at room temperature in static conditions (intact rosettes) or agitated using a vortex for 30 min after complete disruption of rosettes using a needle (disrupted rosettes), in the presence of autologous plasma. P. vivax isolates (2×10⁶) at 5–8% parasitemia were added to each slide of PBMC. After 1h30min of incubation at 37 °C, PBMCs were harvested and washed three times in cell culture medium, and the slides were fixed in methanol and stained with panoptic stain. At least 200 PBMCs were counted in each well, and all samples were tested in duplicated.

We want to express our gratitude to the people who agreed to participate in this study, and to the field team in the Fundação de Medicina Tropical Dr. Heitor Vieira Dourado (FMT-HVD) in Manaus. The authors also thank the Program for Technological Development in Tools for Health-PDTIS FIOCRUZ (Microscopy Facility and Flow Cytometry Facility) at the Carlos Chagas Institute, Fiocruz-Paraná, Brazil.