Emerging trends in Lassa fever: redefining the role of immunoglobulin M and inflammation in diagnosing acute infection

Abstract

Background: Lassa fever (LF) is a devastating hemorrhagic viral disease that is endemic to West Africa and responsible for thousands of human deaths each year. Analysis of humoral immune responses (IgM and IgG) by antibody-capture ELISA (Ab-capture ELISA) and Lassa virus (LASV) viremia by antigen-capture ELISA (Ag-capture ELISA) in suspected patients admitted to the Kenema Government Hospital (KGH) Lassa Fever Ward (LFW) in Sierra Leone over the past five years is reshaping our understanding of acute LF.

Results: Analyses in LF survivors indicated that LASV-specific IgM persists for months to years after initial infection. Furthermore, exposure to LASV appeared to be more prevalent in historically non-endemic areas of West Africa with significant percentages of reportedly healthy donors IgM and IgG positive in LASV-specific Ab-capture ELISA. We found that LF patients who were Ag positive were more likely to die than suspected cases who were only IgM positive. Analysis of metabolic and immunological parameters in Ag positive LF patients revealed a strong correlation between survival and low levels of IL-6, -8, -10, CD40L, BUN, ALP, ALT, and AST. Despite presenting to the hospital with fever and in some instances other symptoms consistent with LF, the profiles of Ag negative IgM positive individuals were similar to those of normal donors and nonfatal (NF) LF cases, suggesting that IgM status cannot necessarily be considered a diagnostic marker of acute LF in suspected cases living in endemic areas of West Africa.

Conclusion: Only LASV viremia assessed by Ag-capture immunoassay, nucleic acid detection or virus isolation should be used to diagnose acute LASV infection in West Africans. LASV-specific IgM serostatus cannot be considered a diagnostic marker of acute LF in suspected cases living in endemic areas of West Africa. By applying these criteria, we identified a dysregulated metabolic and pro-inflammatory response profile conferring a poor prognosis in acute LF. In addition to suggesting that the current diagnostic paradigm for acute LF should be reconsidered, these studies present new opportunities for therapeutic interventions based on potential prognostic markers in LF.

Figure 1

Lassa virus antigen levels in LF patients. Serum levels of LASV NP antigen in LF patients were quantitated by ELISA, using a sensitive caprine polyclonal antibody capture and detection sandwich method [11,12]. Follow-up convalescent patients primarily displayed undetectable levels of LASV antigen. Viral antigen levels between nonfatal (LF NF) and fatal LF cases (LF F) were significantly different at the time of admission or testing (N = 87, p < 0.01). The average time from onset of symptoms to admission was 8.0 ± 3.7 days for LF NF patients, and 10.4 ± 4.5 days for LF F patients. The difference between these times was not statistically significant (p = 0.07).

Figure 2

IgM and IgG responses for normal donors, LF and NL febrile subjects. Box plots of LASV-specific IgM (A) and IgG (B) levels determined by ELISA, are displayed as mean OD₄₅₀ with corrected cutoff values based on the 95^th percentile of established negative control sera. Each display shows the minimum non-outlying value, three quartiles, maximum non-outlying value, and outlying values. The comparison groups include U.S. normals (US N), Moyamba district normals (MOY NHS), Bombali district normals (BOM NHS), convalescent LF follow-up patients (between 8 and 108 weeks post discharge [LF FU]), nonfatal acute LF cases (LF NF), fatal LF cases (LF F), and non-Lassa febrile illness with LASV-specific IgM only (NL FI IgM+). IgM and IgG levels for patients in the LF FU sera group were significantly higher than those in for all of the other comparison groups, save for those in the BOM NHS and NL FI IgM+ cohorts. There were no significant differences between LF NF and LF F cases for both IgM and IgG responses. Bombali sera showed relatively high levels of LASV-specific IgM and IgG, but these levels did not significantly differ from those for LF FU patients, despite their undiagnosed recent LF or other febrile illnesses. Outliers are indicated with red asterisks (*). Significant p values for pairwise comparisons are displayed as * p < 0.05; ** p < 0.01; and *** p < 0.001.

Figure 3

Regression analysis for IgM and IgG responses against number of days post-discharge for convalescent LF patients. Corrected mean OD₄₅₀ values for LASV-specific IgM (A.) and IgG (B.) levels in LF convalescent patients did not reveal any dependence with time post-discharge for immunoglobulins responses. Hypothesis tests for the slope of each regression line revealed zero slopes for both profiles, suggesting that IgM and IgG responses for convalescent patients remained relatively constant after discharge. The fitted intercept for the regression line shown in (A.) was 0.46 (SE = 0.09), which showed prolonged elevation in IgM responses for convalescent LF patients. The fitted intercept of 1.14 (SE = 0.06) for (B.) was indicative of a prolonged mature humoral response (IgG) in convalescent LF patients.

Figure 4

Levels of relevant cytokines in LF pathogenesis. Serum levels of IL-8, Il-6, IL-10, and MIP-1β showed significant differences between LF NF and LF F patients on admission. For LF NF patients, IL-8 levels were significantly reduced when compared to LF F, normal, and LF FU subjects. Elevated levels of IL-6 and IL-10 were recorded in LF F patients, but were significantly lower in LF NF subjects. MIP-1β was significant reduced in LF NF only when compared to LF F patients. Interferon-γ was significantly higher for fatal LF patients than normal donors and follow-up controls, but IFN-γ levels did not differ between fatal and nonfatal LF patients. Similarly, IL12p70 levels were significantly elevated in LF F when compared to normal donors, but did not differ from the other comparison groups. CD40L was significantly reduced in LF NF when compared with normal controls, but not in LF F. Cytokine levels were relatively similar between LF NF and NL FI IgM+ groups, with the exception of MIP-1β, which was elevated in the later cohort. Outliers are denoted with red asterisks (*). Significant p values for pairwise comparisons are denoted as * p < 0.05; ** p < 0.01; and *** p < 0.001.

Figure 5

Relevant metabolic indicators in LF pathogenesis. Hepatic enzymes ALP, ALT, and AST were highly elevated in LF patients, particularly for LF F patients. AST levels were the best prognosticator in LF, with nearly all fatal cases showing extremely high levels. Dissolved tCO₂ levels were significantly lower than normal for LF F patients, whereas BUN levels were significantly higher than normal. Serum calcium levels (corrected for serum albumin levels) were elevated in LF patients regardless of eventual outcome, but no differences in elevation were observed between acute LF groups. Lower and upper outliers are denoted with pink and red asterisks (*), respectively. Significant p values are denoted as * p < 0.05; ** p < 0.01; *** p < 0.001.