{"id":624325,"date":"2026-10-08T12:02:20","date_gmt":"2026-10-08T12:02:20","guid":{"rendered":"https:\/\/www.newsbeep.com\/nz\/624325\/"},"modified":"2026-10-08T12:02:20","modified_gmt":"2026-10-08T12:02:20","slug":"filovirus-reactive-antibodies-in-wildlife-hunting-communities-suggest-ongoing-zoonotic-spillover-in-forested-guinea","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/nz\/624325\/","title":{"rendered":"Filovirus reactive antibodies in wildlife hunting communities suggest ongoing zoonotic spillover in forested Guinea"},"content":{"rendered":"

Study sites and study populations<\/p>\n

Study sites were purposively sampled to include villages that were affected (n\u2009=\u200916) and unaffected (n\u2009=\u200928) by the 2013\u20132016 West African Ebola epidemic, as reported in national databases from the Macenta Prefectural Health Direction office. Official designations were confirmed by key informant interviews with village leaders and healthcare workers. Within villages, families known to practice wildlife hunting were approached, and the senior hunter and their spouse (or next closest relative) were invited to participate. Participants <18 years of age, pregnant, with comorbidities, and visitors to the region were excluded. Demographic data, including age, sex, occupation, and data on wildlife hunting practices, were collected by researchers via questionnaires administered in French and Toma (a regional indigenous language).<\/p>\n

A total of 511 participants were recruited from 40 villages in 2017 (as previously published46<\/a>), and 648 participants were recruited from 43 villages in 2023 (Table\u20091<\/a>). Of the 2017 study population, 52% (268\/511) were resampled as part of the follow-up 6 years later. Wildlife hunters (all male) constituted 56% of the study population, with the remaining 45% reporting close household contact with a hunter. Farming was the predominant occupation in 2017 (93%, 476\/511) and 2023 (86%, 556\/648), however some healthcare workers were also enrolled (3% and 12% in 2017 and 2023 respectively).<\/p>\n

Table 1 Demographic characteristics of participants in the Guinean wildlife hunting community, 2017 and 2023<\/p>\n

Serum samples (n\u2009=\u200963) collected in 2020 from healthy blood donors, aged 17\u201380 years, from the United Kingdom were utilized as a presumed negative, non-endemic comparative study population47<\/a>.<\/p>\n

Serum samples (n\u2009=\u20093) were collected between 2015 and 2017 from Guinean PCR-confirmed Ebola virus disease (EVD) survivors 3\u201314 months post-infection and pooled for use as a positive control for Ebola virus. Serum samples (n\u2009=\u200915) were also collected from urban settings (Coyah and Gu\u00e9ck\u00e9dou, Guinea) in 2015\u20132017 from individuals with no known epidemiological contact with EVD, no known case history, and who were negative by EBOV-GP ELISA41<\/a>.<\/p>\n

Wildlife hunting and bat exposure in Forested Guinea<\/p>\n

In 2023, a detailed questionnaire on wildlife hunting and other activities was administered to all participants in Toma and French. Consumption of wildlife or \u201cbushmeat\u201d was ubiquitous, with 90% of respondents (581\/648) consuming wildlife meat, and 94% (583\/619) helping handle, wash or prepare wildlife carcasses prior to cooking. Wildlife was consumed a median of 7 [IQR: 4\u201315] days per month, and amongst hunters, hunting was practiced a median of 6 [IQR: 3\u201310] days per month. Wildlife hunting generally was not considered to be seasonal by the majority of respondents (93%, 576\/619), with the notable exception of bat hunting. Bat exposure and hunting were reported to be highly seasonal, with 87% (292\/334) of respondents seeing them or hunting them (92%, 122\/133) at the start of the dry season. Agouti was the most commonly reported food source, followed sequentially by rodent, antelope, boar, monkey, pangolin and bat. Bat hunting was found to be a relatively specialized practice with only 32% (207\/648) of individuals either hunting bats or living with a bat hunter. Bats are typically hunted in the forest, in trees, or in the village, and rarely in caves.<\/p>\n

Bat contact was more frequent and was typically reported as 2-3 days per week (46%, 296\/648), taking place around the village (45%, 259\/574) or in the house (28%,158\/574) or the forest (14%, 80\/574) between 12\u20136\u2009am and in the afternoon. Forest activities included hunting or collecting wood, plants or food. Cave usage was uncommon (25%, 139\/546) and comprised short visits only, with an average time of 13\u2009min [min\u2013max; 0\u201360] spent inside per visit. Caves were typically used by both hunters and non-hunters. Guano harvesting was rarely reported (0.3%, 2\/648).<\/p>\n

Filovirus antigen reactivity in Guinean wildlife hunting communities by multiplex microsphere-based immunoassay screening<\/p>\n

Population-level differences in the distributions of IgG antibodies against multiple zoonotic filoviruses were detected between the Forested Guinea wildlife hunting communities (2017 and 2023) and the Ebola-negative Coyah and Gu\u00e9ck\u00e9dou (n\u2009=\u200918) and United Kingdom (UK) control study populations (n\u2009=\u200963) (Fig.\u20091<\/a>). Initial seropositivity cut-offs were based on the presumed unexposed non-endemic study population (UK blood donors) by calculating five standard deviations above the background-subtracted (MFI-bg) mean. As seen in Fig.\u20091<\/a>, a high proportion of the wildlife hunting population had responses above the seropositivity threshold for not only EBOV and closely related orthoebolaviruses, such as BDBV and TAFV, but also more distant filoviruses, including BOMV, LLOV, RESTV and MLAV.<\/p>\n

Fig. 1: Probability density plots of anti-filovirus glycoprotein IgG distributions among Guinean wildlife hunting communities in 2017 (n\u2009=\u2009511) and 2023 (n\u2009=\u2009648), Guinean Coyah controls (n\u2009=\u200918) and the United Kingdom (UK) blood donor study populations (n\u2009=\u200963).\"Fig.<\/p>\n

IgG responses were measured by multiplex microsphere-based immunoassay (MMIA) in background-subtracted median fluorescence intensity (MFI-bg) on the Luminex platform. Dashed vertical lines mark seropositivity cutoffs based on the presumed unexposed non-endemic study population (UK blood donor) by calculating five standard deviations above the background-subtracted (MFI-bg) mean. Abbreviations: BDBV Bundibugyo virus; BOMV Bombali virus; EBOV Ebola virus; IgG immunoglobulin G; LLOV Lloviu virus; MLAV M\u011bngl\u00e0 virus; RAVV Ravn virus; RESTV Reston virus; SUDV Sudan virus; TAFV Ta\u00ef Forest virus; UK United Kingdom.<\/p>\n

Classification of immune responses by K-means clustering reveals EBOV and non-EBOV filovirus reactivity in wildlife hunting communities, 2017 and 2023<\/p>\n

Given the heterotypic cross-reactivity seen between antigenically related viruses, we utilized an unsupervised machine learning algorithm (K-means clustering) to classify the IgG binding antibody responses of individuals against multiple filoviruses into \u201cclusters\u201d, or unique exposure profiles. The K-means clustering algorithm revealed a stable underlying data structure and was used to group individuals with similar serological responses together into k\u2009=\u200915 clusters (Fig.\u20092<\/a>). This approach to determining seropositivity allowed us to avoid the use of a serological threshold for each virus, thus minimizing the misclassification of cross-reactive individuals as true positives and identifying groups of individuals with common exposures across multiple viruses.<\/p>\n

Fig. 2: The K-means clustering algorithm partitioned anti-filovirus glycoprotein (GP) antibody responses from 2017 and 2023 into k\u2009=\u200915 clusters with qualitatively different serological responses.\"Fig.<\/p>\n

The mean reactivity for each filovirus antigen by cluster, and the number of samples in each cluster, is shown. IgG responses were measured by multiplex microsphere-based immunoassay (MMIA) in median fluorescence intensity (MFI) units on the Luminex platform. Several sets of clusters showed similar ordinal patterns of antibody reactivity across filovirus GPs. Specifically, clusters 2, 12, and 15 shared an ordinal pattern in which average MFI values were highest for TAFV, followed by BDBV and EBOV. Clusters 3, 6, 7, and 9 exhibited a pattern in which average MFI values were highest for EBOV, followed by BDBV and BOMV. Clusters 4 and 5 exhibited a pattern in which average MFI values were highest for BDBV, followed by EBOV. Individual responses of wildlife hunters within these clusters are presented in the Supplementary to show the intra-cluster variation as a qualitative check on whether these clusters are distinct. Abbreviations: BDBV Bundibugyo virus; BOMV Bombali virus; EBOV Ebola virus; GP glycoprotein; IgG immunoglobulin G; LLOV Lloviu virus; MLAV M\u011bngl\u00e0 virus; RAVV Ravn virus; RESTV Reston virus; SUDV Sudan virus; TAFV Ta\u00ef Forest virus.<\/p>\n

Three clusters with a total of 57 individuals (57\/1159, 4.9%, Clusters 6, 7, 9) were characterized by high anti-EBOV binding, and heterotypic cross-reactivity against related filoviruses, which decreased with phylogenetic distance (Fig.\u20093<\/a>) and were considered consistent with prior EBOV exposure, based on their similarity to binding patterns observed in PCR-confirmed EBOV survivors 12 and 24 months post-infection48<\/a>. Interestingly, clusters were also identified with more species-specific antibody binding against singular filoviruses, often in the absence of anti-EBOV antibodies. This included clusters of individuals with antibody binding against BDBV (n\u2009=\u200949, 4.2%, Cluster 4), TAFV (n\u2009=\u200913, 1.1%, Cluster 15), RESTV (n\u2009=\u20094, 0.3%, Cluster 8), MLAV (n\u2009=\u20098, 0.7%, Cluster 13), LLOV (n\u2009=\u20098, 0.7%, Cluster 14) and one individual with binding against BOMV (n\u2009=\u20091, 0.1%) (Fig.\u20093<\/a>). Individuals within these clusters were subsequently prioritized for functional assessment and further differentiation by parallel pseudotyped virus neutralization assays (PVNA).<\/p>\n

Fig. 3: Anti-filovirus glycoprotein (GP) antibody responses of individuals clustered by K-means clustering (k\u2009=\u200915), Guinean wildlife hunters, 2017 and 2023 (n\u2009=\u20091159).\"Fig.<\/p>\n

IgG responses measured by multiplex microsphere-based immunoassay (MMIA) in median fluorescence units (MFI) on the Luminex platform. Abbreviations: BDBV Bundibugyo virus; BOMV Bombali virus; EBOV Ebola virus; GP glycoprotein; IgG immunoglobulin G; LLOV Lloviu virus; MLAV M\u011bngl\u00e0 virus; RAVV Ravn virus; RESTV Reston virus; SUDV Sudan virus; TAFV Ta\u00ef Forest virus.<\/p>\n

Filovirus neutralizing activity in Guinean wildlife hunting communities<\/p>\n

Individuals with antibody profiles consistent with exposure to BDBV, TAFV, BOMV, RESTV and LLOV were then assessed for their ability to neutralize the filovirus in parallel with Ebola virus-Makona PV. Neutralizing ability was tested by a lentivirus-backbone filovirus glycoprotein pseudotyped virus (PV) neutralization assay. A pool of PCR-confirmed EBOV-Makona survivors from 2014 was included for comparison. Negative controls included an EBOV-negative Guinean pool from Coyah and Gu\u00e9ck\u00e9dou (n\u2009=\u20093) and a United Kingdom blood donor.<\/p>\n

Of those tested from the BDBV cluster, 5\/10 individuals had lower NT50 estimates against BDBV than EBOV (though 95% confidence intervals (CI) overlapped), and neutralized BDBV more strongly than the EBOV survivor pool, and all came from two villages which reported no cases of EBOV (Fig.\u20094<\/a>). Two of these individuals were household contacts (JAB135, JAB136), also from an EBOV-unaffected village. Persistent responses were shown by those resampled in 2023 (JAB297, JAB376). The remaining samples (including the EBOV survivor pool) showed stronger neutralization of EBOV, with some neutralization of BDBV (Fig.\u20094<\/a>).<\/p>\n

Fig. 4: Neutralizing ability of individual sera samples from the Guinean wildlife hunting community against.\"Fig.<\/p>\n

a Bundibugyo (BDBV) and Ebola (EBOV) viruses. b Ta\u00ef Forest virus (TAFV) and EBOV viruses, as tested in a lentiviral backbone pseudotyped virus neutralization assay. Sera were identified by K-means clustering as showing the highest specific binding antibody responses against a BDBV b TAFV. NT50 values (displayed as points, unless NT50\u2009=\u20090) and 95% confidence intervals (displayed as whiskers, unless >0) were calculated using nonlinear regression in GraphPad Prism 10.0.3. Neutralization assays were performed in technical triplicate for each independent serum sample. The EBOV survivor pool consists of sera from n\u2009=\u20093 PCR-confirmed EBOV survivors 3\u201314 months post-infection. Negative controls included sera from an EBOV-negative Guinean pool (n\u2009=\u20093) and a UK blood donor: neither showed neutralization. Additional demographic information is available in the Supplementary Information. Abbreviations: BDBV Bundibugyo virus; CI confidence interval; EBOV Ebola virus; IgG immunoglobulin G; NT50 50% neutralization titer; PCR polymerase chain reaction.<\/p>\n

Of the individuals tested from the TAFV cluster, 5\/8 individuals from 2023 had lower NT50 estimates against TAFV than EBOV (with some overlap of 95% CIs) (Fig.\u20094<\/a>). Of the LLOV cluster, 1\/9 individuals had stronger neutralizing ability against LLOV than EBOV. All samples from the BOMV and RESTV cluster neutralized EBOV more strongly than BOMV or RESTV, though all showed some neutralizing activity against both. In all neutralization assays, the PCR-confirmed EBOV survivor pool showed strong neutralization responses against EBOV-Makona (NT50 90, 95% CI: 57\u2013142), and also demonstrated a lower level of cross-neutralization of BDBV, RESTV, TAFV and LLOV.<\/p>\n

Durable EBOV antibody responses, 6 years post- 2013-16 West African EBOV epidemic<\/p>\n

For individuals with detectable EBOV antibodies in 2017, a comparison of 2017\u20132023 demonstrated the population-level durability of anti-EBOV antibodies across all villages, including in villages that were not reported as affected by the 2013\u201316 EBOV epidemic. Antibody decay rates were modeled as an exponential decrease from time of infection, with a regression model fitted to IgG antibody data from each individual above a seropositivity threshold for EBOV in 2017 who was resampled in 2023 (n\u2009=\u200911). Based on the population average of these models, the EBOV IgG antibody half-life was estimated as 7.5 years (Fig.\u20095<\/a>), though the possibility of re-infection and subsequent antibody boosting could not be excluded.<\/p>\n

Fig. 5: Population-level mean (regression line) with 95% confidence interval (shaded) and individual-level antibody decay models for Ebola virus (EBOV).\"Fig.<\/p>\n

IgG responses were measured by multiplex microsphere-based immunoassay (MMIA) in background-subtracted median fluorescence units (MFI). Antibody decay rates were modeled as an exponential decrease from time of infection, with a regression model fitted to IgG antibody data from each individual above a seropositivity threshold for EBOV in 2017 who was resampled in 2023 (n\u2009=\u200911). Abbreviations: IgG immunoglobulin G.<\/p>\n

Village-level hotspots identified for BDBV in locations unaffected by the 2013\u20132016 West African EBOV epidemic<\/p>\n

Mean antibody responses were calculated for each filovirus antigen by village, and revealed marked village-level heterogeneity (Fig.\u20096<\/a>). For example, Village 4 in 2017 showed significantly increased mean BDBV responses compared to other villages (P\u2009<\u20090.0001). Some of the highest BDBV-neutralizing individuals identified were located in Village 4 within the same household, and further, demonstrated superior binding and neutralizing activity against BDBV than EBOV in the MMIA and lentiviral pseudotyped neutralization assays (Fig.\u20097<\/a>). Village 4 was confirmed by both official Macenta Ministry of Health designations and key informant interviews of healthcare workers and village leaders to be unaffected by the 2013\u201316 West African Ebola epidemic.<\/p>\n

Fig. 6: Mean anti-filovirus IgG responses amongst Guinean wildlife hunting communities in 2017 and 2023, stratified by village.\"Fig.<\/p>\n

IgG responses were measured by multiplex microsphere-based immunoassay (MMIA) in background-subtracted median fluorescence intensity (MFI). Villages (n\u2009=\u200940 in 2017, n\u2009=\u200943 in 2023) were determined to be affected or unaffected by the 2013\u201316 West African Ebola virus epidemic based on official Macenta Ministry of Health designations followed by key informant interviews of healthcare workers and village leaders. Villages are numbered from 1\u201343 by geographic proximity. Abbreviations: BDBV Bundibugyo virus; BOMV Bombali virus; EBOV Ebola virus; IgG immunoglobulin G; LLOV Lloviu virus; MLAV M\u011bngl\u00e0 virus; RAVV Ravn virus; RESTV Reston virus; SUDV Sudan virus; TAFV Ta\u00ef Forest virus.<\/p>\n

Fig. 7: Individual IgG responses to Ebola (EBOV) and Bundibugyo (BDBV) virus amongst Guinean wildlife hunting communities from Village 4 in 2017 and 2023.\"Fig.<\/p>\n

IgG responses were measured by multiplex microsphere-based immunoassay (MMIA) as background-subtracted median fluorescence intensity (MFI). Individuals with a lower estimated NT50 for BDBV than EBOV in parallel lentiviral pseudotyped virus neutralization tests are denoted with a red dot. Abbreviations: IgG immunoglobulin G; NT50 50% neutralization titer.<\/p>\n

Wildlife hunting and zoonotic filovirus exposure<\/p>\n

Individual-level risk factors were investigated for the 2023 study population in univariable and multivariable models using questionnaire data. Sex was significantly associated with BDBV cluster seropositivity, with the odds of being in the BDBV cluster 2.49 times higher amongst males (95% CI: 1.00\u20136.16, P\u2009=\u20090.048) compared to females (Supplementary Table\u20098<\/a>). Nonhuman primate consumption was associated with higher odds of BDBV cluster seropositivity in both univariable (OR\u2009=\u20093.49, 95% CI: 1.25\u20139.74, P\u2009=\u20090.017) and multivariable models (aOR\u2009=\u20095.03, 95% CI: 1.70\u201314.90, P\u2009=\u20090.004) (Supplementary Table\u20098<\/a>).<\/p>\n

Influence of land cover patterns on exposure to zoonotic filoviruses in Guinean wildlife hunting communities<\/p>\n

We extracted metrics of land cover (trees, crops, built area) and forest fragmentation (perimeter area ratio, shape-index and fractal dimension) from satellite-derived land-cover maps for 2017 and 2023 (Environmental Systems Research Institute, European Space Agency Sentinel-2)49<\/a>. We extracted metrics at 500\u2009m, 1\u2009km, 2\u2009km, 5\u2009km and 10\u2009km distances from the village centroids, which were examined for significance across varying spatial scales using a data-driven approach.<\/p>\n

Mixed-effect multivariable logistic regression models were then used to examine the association between environmental drivers and risk for each zoonotic virus (based on their K-means-based antibody exposure profiles) and identified significant associations between landscape patterns and serological filovirus exposure. Models also examined individual risk factors reported in the questionnaire, including age, sex, occupation and active wildlife hunting; these did not significantly improve the model.<\/p>\n

Significant environmental variables were identified at various spatial scales for all antigens, with forest coverage and configuration (shape index, perimeter area ratios, fractal dimension index) found to be associated with EBOV, BDBV, TAFV, LLOV and MLAV reactivity (Table\u20092<\/a>). Differing landscape patterns were identified for each filovirus antigen, with EBOV exposure being negatively associated with tree coverage. Forest fragmentation (within a 500\u2009m radius of the village) and living in built areas were positively associated with BDBV seroreactivity. Reactivity to other filovirus antigens (TAFV, LLOV) was positively associated with proximity to intact forest (5\u201310\u2009km radius of the village).<\/p>\n

Table 2 Significant associations between viral antigen exposure risk and environmental factors<\/p>\n","protected":false},"excerpt":{"rendered":"Study sites and study populations Study sites were purposively sampled to include villages that were affected (n\u2009=\u200916) and…\n","protected":false},"author":2,"featured_media":624326,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10],"tags":[8433,156767,108750,134,1928,1929,111,139,69,147,306853],"class_list":["post-624325","post","type-post","status-publish","format-standard","has-post-thumbnail","category-health","tag-ebola-virus","tag-ecological-epidemiology","tag-forest-ecology","tag-health","tag-humanities-and-social-sciences","tag-multidisciplinary","tag-new-zealand","tag-newzealand","tag-nz","tag-science","tag-viral-epidemiology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts\/624325","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/comments?post=624325"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/posts\/624325\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/media\/624326"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/media?parent=624325"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/categories?post=624325"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/nz\/wp-json\/wp\/v2\/tags?post=624325"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}