Dejnirattisai2010 - Anti-prM Antibodies Enhance Dengue ADE

Full citation: Dejnirattisai W, Jumnainsong A, Onsirisakul N, Fitton P, Vasanawathana S, Limpitikul W, Puttikhunt C, Edwards C, Duangchinda T, Supasa S, Sae-Jang K, Malasit P, Mongkolsapaya J, Screaton GR. “Cross-Reacting Antibodies Enhance Dengue Virus Infection in Humans.” Science 2010; 328(5979):745–748.

Raw file: [[raw/dejnirattisai2010.pdf]]

Note: A published erratum (July 2010) corrected an overstatement of priority in the original paper — the claim to be “the first” to describe cross-reactive ADE was revised. The scientific findings reported here are unaffected by the erratum.

Summary

This paper characterises the anti-prM (pre-membrane protein) component of the human anti-dengue antibody response using human monoclonal antibodies (hmAbs) isolated from B cell lines of seven DENV-infected donors in Thailand. The authors find that anti-prM antibodies constitute approximately 60% of the structural antibody response (more than anti-E antibodies), are fully cross-reactive across all four dengue serotypes, neutralise virus poorly due to incomplete prM cleavage, and mediate exceptionally potent ADE in primary monocytes and dendritic cells — up to 10^5-fold enhancement.

The mechanistic explanation centres on incomplete furin-mediated prM cleavage during virion maturation, which produces a mixed population of prM-containing (immature/partially immature) and M-containing (mature) particles. Anti-prM antibodies bind the prM-containing particles but cannot neutralise them because the prM-coated particles are not infective in the absence of FcR-mediated uptake. Once opsonised by anti-prM antibodies, these particles enter monocytes/DCs via FcγR, replicate intracellularly, and release progeny virus.

The findings have major implications for dengue vaccine design: all current vaccine candidates use native prM sequences, meaning they will prime potent anti-prM antibody responses with the same ADE-promoting profile characterised here.

Study Design

• Type: Original experimental research — human monoclonal antibody isolation + in vitro functional assays
• Sample size: 7 DENV-infected donors; 3,020 B cell lines screened; 301 anti-dengue positive; 26 anti-prM hmAbs + 26 anti-E hmAbs selected for full characterisation
• Setting: Donors recruited from two Thai hospitals (Khon Kaen Hospital and Songkhla Hospital, Thailand); B cell lines and assays conducted at Wellcome Centre, Oxford/MOPH Bangkok
• Population: DENV-infected adults with confirmed dengue; serotype data not fully specified for all donors; secondary infections included

Key Findings

• Anti-prM antibodies are the dominant structural antibody class: ~60% of isolated anti-DENV hmAbs were anti-prM; only ~35% were anti-E; ~5% bound other structural antigens. This was unexpected — prior literature had focused almost exclusively on anti-E antibodies as the relevant immunological targets.
• Complete cross-reactivity across all four serotypes: All 26 anti-prM hmAbs bound DENV-1, -2, -3, and -4 with high avidity by ELISA. Anti-E hmAbs showed more serotype-restricted binding, consistent with their role in type-specific neutralisation.
• Poor neutralisation: Anti-prM hmAbs neutralised virus poorly — maximum neutralisation plateau of 10–60% in PRNT, with most anti-prM antibodies achieving <50% neutralisation even at saturating concentrations. Complete neutralisation (>90%) was not achieved by any anti-prM hmAb against any serotype. The incomplete prM cleavage model explains this: anti-prM antibodies bind only the prM-containing (non-infective) fraction; the mature virion fraction remains unbound and infective.
• Potent ADE in primary monocytes and dendritic cells: Anti-prM hmAbs mediated ADE in primary human monocytes and monocyte-derived DCs at enhancement levels up to 10^5-fold over antibody-free controls — substantially higher than typical anti-E ADE (which is also substantial but generally in the 10-1,000× range). Enhancement was FcγR-dependent (blocked by anti-FcγR antibodies).
• Anti-JEV serum shows minimal cross-reactivity with DENV prM: Sera from JEV-immune donors did not cross-react substantially with DENV prM, distinguishing the anti-prM response as dengue-specific rather than a broad flavivirus cross-reactive phenomenon.
• Vaccine implication: CYD-TDV, DENVax (TAK-003), and TV003/TV005 all incorporate native prM sequences; all will prime anti-prM responses with the same ADE-promoting characteristics documented here. The authors call for heterologous prM design or prM deletion in future vaccine constructs.
• Erratum (July 2010): The original version claimed this was the first description of anti-prM ADE; the correction acknowledges prior in vitro work on anti-prM enhancement.

Methods Used

• PRNT (neutralisation assay for anti-prM and anti-E hmAbs against all four serotypes)
• IgM-IgG Serology ELISA (anti-prM and anti-E binding by ELISA; cross-reactivity panel across serotypes; anti-JEV cross-reactivity controls)
• B cell line culture and EBV-immortalisation (B cell hmAb isolation from DENV-infected donors)
• ADE assay in primary human monocytes and monocyte-derived dendritic cells (flow cytometry-based infection measurement)
• Antibody competition assays (blocking anti-FcγR to confirm receptor dependence of ADE)

Entities Mentioned

• prM Protein
• E Protein
• DENV-1
• DENV-2
• DENV-3
• DENV-4

Concepts Addressed

• Antibody-Dependent Enhancement
• Cross-Reactive Antibodies
• Dengue Vaccine Candidates
• Secondary Dengue Infection

Relevance & Notes

This paper is one of the most important mechanistic contributions to ADE pathogenesis in this wiki. It reframes the immunological landscape by identifying anti-prM antibodies — not anti-E — as the numerically dominant and functionally most dangerous component of the structural antibody response. Prior work had almost exclusively focused on anti-E antibodies as the ADE-mediating population.

The finding is directly relevant to the ADE-in-secondary-infection model (see Antibody-Dependent Enhancement): in secondary infection, anti-prM antibodies from the primary exposure would be present at high titres and cross-reactive against the new serotype, primed for potent FcγR-mediated enhancement. The 10^5-fold ADE magnitude observed in primary monocytes/DCs exceeds typical anti-E enhancement and may explain why certain secondary infections produce rapid, overwhelming viral replication.

The vaccine implication is sobering and unresolved: as of the paper’s publication, and for years afterward, all leading vaccine candidates (CYD-TDV, TAK-003, TV003) retain native prM sequences. Whether any vaccine has moved toward heterologous prM design is not yet documented in this wiki.

The paper’s findings also strengthen the FcγRIIa genotype literature: if anti-prM-mediated ADE is the dominant enhancement pathway, then FcγRIIa genotype effects on ADE risk (see FcγRIIa Receptor, Garcia2010 - Asymptomatic Dengue FcγRIIa Polymorphism) should be interpreted in the context of anti-prM (not just anti-E) opsonised particles.

Questions Raised

• Have any current vaccine candidates (post-2010) adopted heterologous or modified prM sequences to reduce anti-prM ADE risk?
• Does anti-prM ADE in vivo operate primarily in monocytes and DCs, or also in other FcγR-bearing cells (neutrophils, NK cells)?
• What is the kinetic relationship between anti-prM and anti-E titres across the inter-infection interval? Does anti-prM wane faster or slower? (The Bos2025 longitudinal data does not distinguish anti-prM from total XR E/prM responses.)
• Does the 10^5-fold monocyte ADE of anti-prM translate linearly to in vivo viraemia in secondary infection, or are there homeostatic constraints?
• Why are anti-prM antibodies so numerous (~60%) relative to anti-E, when E protein is far more abundant on the mature virion surface? Does this reflect preferential prM immunogenicity during the natural course of infection, or selection bias in the assay?