Detection, characterization, and enrollment of donors of Ebola convalescent plasma in Sierra Leone

Abstract

Background: Passive therapy with convalescent plasma provides an early opportunity to intervene in Ebola virus disease (EVD). Methods for field screening and selection of potential donors and quantifying plasma antibody are needed.

Study design and methods: Recombinant Ebola virus glycoprotein (EBOV GP) was formatted into immunoglobulin G-capture, competitive, and double-antigen bridging enzyme immunoassays (EIAs). EVD survivors in Freetown, Sierra Leone, were recruited as potential plasma donors and assessed locally using sera alone and/or paired sera and oral fluids (ORFs). Uninfected controls comprised unexposed Gambians and communities in Western Area, Sierra Leone. Antibody neutralization in selected sera was measured retrospectively in a pseudotype virus assay.

Results: A total of 115 potential donors were considered for enrollment: 110 plasma samples were concordantly reactive in the three EIAs; three were concordantly unreactive and two were reactive in two of three EIAs (98.2% agreement; 95% confidence interval [CI], 93.9%-99.8%). In 88 donors with paired ORF and plasma, G-capture EIA reactivity correlated well in the two analytes (R² = 0.795). Plasma and ORF from 44 Gambians were unreactive. ORF samples from 338 of 339 unexposed Western Area community controls were unreactive (specificity, 99.7%; 95% CI, 98.4%-99.7%); ORF samples from 113 of 116 Kerry Town EVD survivors were reactive (sensitivity, 97.4%; 95% CI, 92.5%-99.5%). Strong reactivity in G-capture and/or competitive EIAs identified donors with high plasma EBOV GP antibody levels in the double-antigen bridging assay, correlating with high levels of neutralizing antibody.

Conclusions: In-field testing can qualify convalescent donors for providing high-titer antibody.

Figure 1

Recruitment process for volunteer convalescent donors seen first at the 34th Regimental Military Hospital Wilberforce Freetown (MH34) before referral to the Blood Bank, National Safe Blood Service, Connaught Hospital, Freetown. ¹34th Regiment Military Hospital, Wilberforce, Freetown, Sierra Leone. ²Sierra Leone Association of Ebola Survivors (SLAES). ³Compensation for cost of attendance of 40,000 Sierra Leone Leones (SLL, $8 USD). ⁴Anemia, HBsAg, anti‐HCV, anti‐HIV, and antibody to syphilis. ⁵HBV infection excluded three donors; compensation of 80,000SLL ($16 USD) for attendance. ⁶Current well‐being; vital signs including temperature, height, and weight; research samples including ORF sampling by Oracol device; blood drawn in tempus tubes, vacutainer PPT tubes, and EDTA tubes. ⁷Candidate donors returned to NSBS Blood Bank for first and subsequent apheresis and received compensation for each apheresis of 300,000 SLL ($60 USD).

Figure 2

Correlation between paired ORF and plasma reactivity, expressed as raw OD, from 10 convalescent donors tested in the G‐capture EIA at Connaught Blood Bank, Freetown (R² = 0.822 from linear regression). Linear regression line on logged titers is shown.

Figure 3

(A) Correlation of NOD in the competitive and the G‐capture EIAs of 37 donor samples field tested in Connaught Blood Bank, Freetown (R² = 0.625 from linear regression). (●) Samples from donors selected for further attendance. One sample is concordantly unreactive (●). (◻) Reactivities of the two WHO standards (15/262 top left; 15/220 bottom right). (B) Correlation between NOD reactivity of 88 paired ORF and plasma samples in the G‐capture EIA from donors taken at first attendance (R² = 0.795). One ORF sample had a NOD value less than 1.0 (●, 0.78). (○) Two dually reactive plasmas with anomalously low ORF NODs. Linear regression line on logged titers is shown. (C) Correlation between G‐capture and competitive EIAs, expressed as log NOD values, of 115 first‐attendance plasma samples (R² = 0.582). Three samples are concordantly unreactive in both EIAs (●). Two samples are discordantly unreactive (○), one is just below the cutoff in the G‐capture, and the other is just below cut‐off in the Competitive EIA (see Table 1). Linear regression line on logged titers is shown.

Figure 4

Anti‐EBOV GP levels in 115 seropositive convalescent donor plasmas, expressed as log₁₀ au/mL, measured in the DABA EIA. Results are shown for the entire cohort (All) superimposed with the first (15/220, lower of the two) and second (15/262, upper of the two) WHO EBOV standards (●) and for those in either of the top quadrants for the G‐capture (Capture UQ) or for the competitive EIA (Competitive UQ) and for those plasma samples reacting in both top quadrants of the G‐capture and the competitive EIAs (In both UQ). Horizontal bars represent geometric mean values anti‐EBOV GP in au/mL.

Figure 5

Anti‐EBOV GP levels in a selected panel of 25 convalescent donors. Plasma antibody measured by pseudotype neutralization (interpolated IC₅₀ neutralization titers) correlates with plasma reactivity in the G‐capture EIA (expressed as NODs, A, R² = 0.6794), with quantified plasma reactivity in the DABA EIA (expressed in au/mL, B, R²= 0.7633) and with paired ORF (21 ORF samples only) reactivity in G‐capture EIA (expressed as NODs, C, R² = 0.5700). (◻) IC₅₀ titers (A, B) for first (15/220 bottom left) and the second (15/252 bottom right) WHO standards at 162 and 192, respectively. Linear regression lines on logged titers are shown.