The presence of circulating antibody secreting cells and long-lived memory B cell responses to reticulocyte binding protein 1a in Plasmodium vivax patients

Cross-sectional and cohort studies were designed to explore the generation of long-term antibody and MBC responses to PvRBP1a in low malaria transmission areas of Southern Thailand (Ranong province). Plasma anti-PvRBP1a antibodies were measured in vivax malaria patients (n  = 68) seen between May 2014 and May 2019. Subjects positive to RBP1a in total IgG (n  = 41) were further analysed for anti-PvRBP1a in IgG subclasses. Of these 41 seropositive patients, 16 patients who were available for monitoring at all time-points in a 12-month cohort study (acute and recovery from infection for 3, 9 and 12 months) were recruited to monitor the longevity of total anti-PvRBP1a IgG responses. The characteristics of participants enrolled in seroprevalence study are shown in Table 1.

To explore the relationship of frequency of MBC subsets and plasmablasts/plasma cells with anti-PvRBP1a antibody levels during acute malaria illness, peripheral blood mononuclear cells (PBMCs) from acute patients (n  = 20) were isolated to perform flow cytometric analysis, and plasma was collected for detection of total IgG. Moreover, the presence of PvRBP1a-specific MBCs was investigated in acute patients (n  = 10) and recovered patients (1–3 months, n  = 7; 1–3 years, n  = 7). Healthy volunteers who lived in non-malaria endemic areas (Bangkok, Thailand) were recruited as healthy controls (HC, n  = 55).

The patients who had febrile illness and positive P. vivax parasitaemia by microscopic examination of both thick and thin blood smears, and nested PCR were recruited as acute vivax malaria infection. After receiving anti-malarial drug treatment, the patients were enrolled in post-infection phase of P. vivax infection. History of previous malaria infections in individual subjects were obtained from the records of the malaria clinic at Vector Borne Disease Unit. Subjects were scheduled for blood sample collection every 3 months for assessment of sub-patent malaria by nested PCR. Staff conducted weekly house to house visits from May 2014 to May 2019 to estimate the incidence of clinical malaria over the study period. Ethical approval was obtained from the Committee on Human Rights Related to Human Experimentation, Mahidol University, Thailand (MUIRB 2012/079.2408).

To express recombinant PvRBP1a proteins, the protocol followed that of a previous study [19]. In brief, overlapping fragments spanning the extracellular region of the RBP1a gene from the Sal1 allele (PVX_098585) were designed. The DNA sequences were codon optimised for Escherichia coli expression, commercially synthesized (Invitrogen, Carlsbad, CA, USA), and cloned into the pET21(a) +  expression vector with a C-terminal 6xHis-tag to facilitate purification. Protein expression was performed in E. coli BL21 (DE3) pLysE cells (Invitrogen). Bacterial cells were cultured in Luria–Bertani (LB) broth until reaching an OD₆₀₀ of 0.6–0.8; protein expression was induced with 1 mM IPTG (final concentration) for 3 h at 30 °C. Cells were harvested by centrifugation at 11,000×g for 20 min and lysed in 20 mM phosphate buffer, pH 7.4, containing 0.5 M NaCl and 20 mM imidazole. Expressed proteins were purified by affinity chromatography, using Ni  +  Sepharose 6 fast flow (GE Life Sciences, Pittsburgh, PA, USA) according to the manufacturer’s recommendations. Purity of the antigens was checked by analysis via Coomassie-stained sodium dodecyl sulfate polyacrylamide gel electrophoresis (see Additional file 1) and then dialyzed against phosphate-buffered saline (PBS) before storage at − 80 °C.

Anti-PvRBP1a antibody levels in plasma samples from patients during and after acute P. vivax infection were measured using an indirect ELISA. Briefly, 96-well microtiter plates were coated with 2 µg/mL recombinant PvRBP1a protein and incubated overnight at 4 °C. The plates were blocked with 5% skimmed milk in PBS-0.05% Tween for 2 h. The plasma samples at dilution 1:200 were then added in duplicate wells and incubated for 1 h. After washing, a 1:1000 dilution of goat anti-human IgG-alkaline phosphatase was added into the wells, and incubated for 1 h. Subsequently, ABTS substrate or 2,2-azino-bis (3-etylbenzthiazoline-6-sulfonic acid) (Sigma-Aldrich, St. Louis, MI, USA) was added to detect antigen–antibody reactivity. The Microplate Reader Synergy HTX (BIOTEK Instruments, USA) was used for reading absorbance at 405 nm. Two acute plasma samples with high titers were used as positive controls [19]. The OD values from duplicate wells per individual were averaged. A baseline OD was established using plasma samples from the 43 HCs.

To detect IgG subclass responses to RBP1a, a protocol similar to that for total IgG. In the step when each plasma sample was added at a dilution of 1:100, a modification used HRP-conjugated anti-human IgG1, IgG2, IgG3, and IgG4 antibodies at a dilution of 1:1000. Signal was developed with TMB (3, 3′5, 5′-tetramethylbenzidine) substrate (Merck KGaA, Sigma-Aldrich, USA). OD was read at 450 nm. Two seropositive samples for anti-PvRBP1a IgG response were used as positive controls. A baseline OD was established using plasma samples from 25 of the HCs.

The levels of total IgG or IgG subclasses specific to PvRBP1a were standardized to a reactivity index (RI), calculated by dividing OD values of tested samples by a cut-off value (mean  +  2SD) from HC samples. An RI  ≥  1 was considered to be positive for specific antibodies; those with an RI  <  1 were considered as non-responders (NRs). To distinguish levels of anti-PvRBP1a antibodies, a cut-off value was calculated as mean RI  +  0.5SD of all acute patients and used to classify high responder (HR; RI  >  mean  +  0.5SD) and low responder (LR; RI  <  mean  +  0.5SD but RI  ≥  1).

PBMCs collected during acute malaria were used for MBC subset and plasmablast/plasma cell phenotyping. Fluorochrome-conjugated, mouse anti-human monoclonal antibodies were used to stain 1 million PBMCs/100 μL FACS buffer. A cocktail consisting of the following mouse monoclonal antibodies was used: FITC-CD19, APC-CD21, APC/fire-CD27, AF700-CD38, PE-IgM and PE/Cy7-IgD (Biolegend, San Diego, CA, USA). After staining for 15 min, cells were washed with FACS buffer. Finally, cells were suspended in 250 μL FACS buffer. The analyses were done with a flow cytometer (FACSCanto II, Becton–Dickinson Immunocytometry Systems, San Jose, CA, USA). Data were processed using FlowJo software (Tree Star, San Carlos, CA, USA).

The presence of PvRBP1a-specific MBCs was determined by ELISPOT assay as performed in a previous study [25]. Briefly, 7 × 10⁶ PBMCs from P. vivax samples were prepared for cell stimulation along with negative controls. First, 1 × 10⁶ cells per mL were stimulated with a cocktail of polyclonal activator R848 (Invivogen) and recombinant human IL-2 (PrepoTech, NJ, USA) at 37 °C in a 5% CO₂ incubator for 72 h. The ELISPOT assay was performed by coating with 15 μg/mL of monoclonal antibodies of anti-human IgG (clones MT91/145; Mabtech), 5 μg/mL of recombinant PvRBP1a antigen, or 1 μg/mL of tetanus toxoid (TT) antigen (Merck Millipore, Darmstadt, Germany) onto ELISPOT plates (Merck Millipore). Cells were harvested and seeded in duplicate to yield 5 × 10⁴ cells per anti-human IgG-coated well, and 1 × 10⁶ cells per specific antigen-coated well. After overnight culture, 1 μg/mL of detection monoclonal antibody MT78/145 (Mabtech) was added and incubated for 2 h. Following thorough washing, the diluted (1:1000) streptavidin-HRP-conjugated, polyclonal goat anti-human IgG (Mabtech) and TMB substrate for ELISPOT (Mabtech) were added. The plates were rinsed with deionized water after distinct spots emerged. PvRBP1a-specific ASCs were expressed as spot-forming cells (SFCs) in the wells. The plates were analysed with a CTL ELISPOT reader. PBMCs from P. vivax-infected subjects that were cultured without polystimulators and incubated overnight with antigens were used as negative controls. A positive PvRBP1a-specific MBCs response was defined as detectable spots in duplicate wells with the total spots in the specific antigen-coated wells being at least twice the number of spots detected with the negative control samples. The circulating PvRBP1a-specific ASCs were enumerated from PBMC cultures without polyclonal activation and the frequencies of ASCs were counted as spot forming units/1 million PBMCs. The antigen-uncoated wells were used as negative controls for circulating PvRBP1a-specific ASC analysis.

To determine the seroprevalence of anti-PvRBP1a antibody responses in natural infection to the P. vivax parasite, a cross-sectional survey of antibody levels against this antigen was conducted in acutely infected patients. During symptomatic malaria, anti-PvRBP1a IgG levels were significantly higher than in HCs (P  < 0.0001) (Fig. 1a). Positivity to PvRBP1a was 60.3% (41 of 68) in the group of acute patients (Fig. 1a). Based on the level of anti-PvRBP1a antibody during acute infections, the individuals were classified into 1 of 3 responder groups [HR (RI  > 3.7), LR (RI 1 to 3.7), and NR (RI  < 1)]. Among these subjects, 16, 44 and 40% were high, low and non-responders, respectively (Fig. 1b).

Different IgG subclasses are involved in the protective humoral immunity to malaria [26, 27]. Thus, the relative anti-PvRBP1a IgG responses to each isotype in acute seropositive subjects were determined (Table 1). IgG1 was a significantly higher compared to IgG2, IgG3, IgG4 subclasses (Fig. 2). The seroprevalence of anti-PvRBP1a by IgG subclass was 82.93%, 48.78%, 58.54% and 41.46% for IgG1, IgG2, IgG3 and IgG4, respectively (Fig. 2). Thus, predominant subclass response to PvRBP1a antigen was IgG1.

To observe the kinetic of anti-PvRBP1a IgG responses after recovery from P. vivax infection, serial anti-PvRBP1a antibody levels were measured in 16 acute patients who were initially seropositive and agreed to additional blood collections (at 3, 9 and 12 months) (Table 1). At 3-month post-infection, total IgG anti-PvRBP1a levels were significantly reduced compared to the acute phase; responses continued to decline at the 9- and 12-month timepoints. There were 56.25% (9 of 16), 43.75% (7 of 16) and 37.5% (6 of 16) of the 16 subjects who maintained seropositive antibody responses to PvRBP1a at 3, 9 and 12 months, respectively (Fig. 3a). Further analysis of the response of these subjects in the kinetic study showed that most of the HR group (n  = 3, RI at acute  > 3.7) had persistent anti-PvRBP1a levels after parasite clearance through month 12, whereas antibody responses of the LR group (n  = 13, RI at acute 1–3.7) were significantly decreased at the 3-month timepoint (Fig. 3a, b).

As HR patients showed the maintenance of anti-PvRBP1a antibody for at least 12 months, this could be due to the ability of PvRBP1a to induce development of MBCs during acute P. vivax malaria which persist after parasite clearance. To assess which MBC subsets might be related to anti-PvRBP1a antibody in P. vivax patients, an independent set of P. vivax subjects were studied, quantitating both anti-PvRBP1a IgG level and frequency of each MBC subset. Based on the levels of anti-PvRBP1a antibodies, 4 and 16 subjects were found to be HRs and LRs, respectively (Fig. 4a). To relate the antibody levels with MBC subset responses, the proportions of activated MBCs (CD21⁻CD27⁺), atypical MBCs (CD21⁻CD27⁻), classical MBCs (CD21⁺CD27⁺), naïve B cells (CD21⁺CD27⁻) and plasmablasts/plasma cells (CD27^hiCD38^hi) from total CD19⁺ B cells were determined (Fig. 4b) [28, 29]. Both HR and LR subjects showed an expansion of activated and atypical MBCs. In the HR group, 8.09% and 25.44% had activated and atypical MBCs, respectively. Whereas, lower proportions of activated (6.26%) and atypical MBCs (17.82%) were observed in the LR group (Fig. 4c). Moreover, the phenotyping showed that 82% of activated MBCs had the phenotype of switched MBCs (IgM⁻IgD⁻ and IgM⁺ only), whereas atypical MBCs showed the phenotypes of both switched and unswitched (IgM⁺IgD⁺) MBCs (35.33% of switched, 45.15% of un-switched MBCs) (Fig. 4d).

To explore the presence of PvRBP1a-specific MBC responses, a cross-sectional study was conducted in acute (n  = 10) and recovered patients (1–3 months, n  = 7; 1–3 years, n  = 7). During acute malaria, 8 out of 10 patients had positive MBCs to PvRBP1a (SFCs ranging from 4 to 42). Post-infection, PvRBP1a-specific MBCs were detected in all subjects assessed 1–3-month post-infection (n  = 7; range, 4–70 SFCs) and 1–3-year post-infection (n  = 7; range, 10–46 SFCs) (Fig. 5a). As expected, nearly all subjects (95.83%) produced tetanus toxoid (TT)-specific MBCs (range, 13–150 SFCs) since TT commonly used as a routine vaccine and the high frequency of TT-specific MBCs was detected in vaccinated individuals (Fig. 5b). Analysis of the relationship between the responses of PvRBP1a-specific MBCs and antibodies within each patient showed that both individuals positive (80%) and negative (20%) for PvRBP1a-specific MBCs were seropositive for total IgG during acute infection (P value  = 1.00). For all positive specific MBCs at 1–3-month post-infection, 5 were seropositive and 2 were seronegative (P value  = 1.00). Interestingly, all seronegative subjects studied in 1–3-year post-infection had positive PvRBP1a-specific MBCs (P value  = 1.00) (Fig. 5c).

To analyse the generation and persistence of ASCs specific to PvRBP1a after natural P. vivax infection, PBMCs from patients were cultured without polyclonal activation, incubated overnight with PvRBP1a and assessed by ELISPOT assay. During acute infection, 90% of patients (9 of 10) produced PvRBP1a-specific ASCs with SFCs ranging from 4 to 18. Following recovery, these subjects had few circulating ASCs specific to PvRBP1a (range, 0–6 SFCs) whether in the 1–3-month or 1–3-year post-infection window (Fig. 6).