An African swine fever vaccine-like variant with multiple gene deletions caused reproductive failure in a Vietnamese breeding herd

African swine fever (ASF) is a highly contagious and deadly viral disease primarily affecting all production stages in domestic pigs¹. The disease can profoundly impact animal well-being and productivity, often resulting in mortality rates as high as 100% and considerable economic repercussions for farmers and the agricultural sectors². ASF manifests in various clinical and pathological forms, from hyperacute to chronic, depending on several factors: host characteristics, the degree of viral virulence, dose, and infection route³. The initial ASF was discovered over a century ago in Kenya⁴; after re-emerging with a highly lethal strain in the Caucasus, it quickly spread throughout neighboring countries. Soon after its introduction into China (August 2018), ASF has continued posing a dramatic threat to Southeast Asian pork producers⁵. Shortly thereafter, the Vietnamese authorities announced the first ASF outbreaks in small-scale farms in February 2019⁶.

African swine fever virus (ASFV) is a giant, double-stranded enclosed DNA virus, approximately 170,000 to 194,000 base pairs (bp) in length, encoding numerous highly immunogenic proteins that exclusively cause ASF⁷. ASFV has exhibited substantial genetic diversity, with prior classification into at least 24 genotypes and eight serotypes relied on the B646L and EP402R genes, which translates into the capsid p72 protein and viral hemagglutinin CD2-like protein (CD2v), respectively⁸. Even so, a recent study has proposed consolidating ASFV into six distinct genotypes, as the previous classification method could not accurately capture the virus diversity due to its genomic complexity⁹. Genotype II, associated with highly lethal acute infections, has been the predominant strain in Eurasia since 2007. Emerging variants, encompassing avirulent genotype II, genotype I, and recombinants of both, have been reported in China and Vietnam^10,11,12,13. This genetic heterogeneity greatly influences disease severity, control measures, and vaccine development¹⁴.

In the absence of vaccinations or viable treatments, the mainstays of ASF control were prompt laboratory identification and removal of exposed and afflicted animals. A modified control approach was implemented by Vietnamese authorities, which mandated the immediate slaughter of any healthy pigs within the outbreak zone and allowed the selective culling of only diseased or infected pigs (official letter 5169/BNN-TY). Since 2022, Vietnam has successfully commercialized modified live-attenuated vaccines (LAVs) against ASF using working seed: ASFV-G-∆I177L, ASFV-G-∆MGF, and ASFV-G-∆I177L/∆LVR strains. These vaccines were developed through gene editing techniques or cell adaption and demonstrated safety, efficacy, and protection in laboratory and field trials^15,16,17. For the ASFV-G-∆MGF vaccine, a Diep’s macrophage cells (DMAC) line was selected and grown from swine embryonic macrophage for commercialized purposes¹⁸. A recent vaccine (HANVET ASF^®) was launched on the market, with a local ASF-HV21 isolate serving as a backbone (www.vietnamagriculture.nongnghiep.vn, accessed on 18 August 2024). Vaccination is recommended for piglets aged at least four weeks, while immunization of breeding stock is not advised¹⁹.

Since the ASF invaded Vietnam, the swine industry has faced a substantial economic burden. Comprehensive vaccination has been implemented to mitigate the spread of ASF, with nearly 4 million distributed doses nationwide until June 2024²⁰. These actions, coupled with unauthorized vaccine experiments, led to the improper release of master seeds, followed by the variation and evolution of the ASFV genetics²¹. For instance, retrospective surveillance determined a field-attenuated isolate (ASFV-GUS-Vietnam), characterized by alleviated clinical pathology and resembling ASFV-G-∆MGF’s genetic patterns, discovered in September 2021 during the LAVs trial stages. Despite its protective ability against ASF-challenged animals, the isolate exhibited risks such as horizontal transmission and death in experimental 6- to 7-week-old pigs²². Hence, it was speculated that immunized pigs might disseminate vaccine-like variants after vaccination and cause some adverse effects, highlighting the complexity of ASF management.

Accordingly, ASFV biological characteristics, improper culling practices, and widespread vaccine implementation may impact the viral genetic pool among the swine population. Nevertheless, these novel attenuated variants that infected animals asymptomatically were often neglected, serving as viral reservoirs and hindering disease control and eradication efforts. This paper explored and described a novel ASF vaccine-like variant that caused atypical reproductive problems in the non-ASF-vaccinated Vietnamese breeding herd.

Study design and sample collection

Following the ASF natural index cases in a swine breeding herd (capacity of domestic 2,400 Landrace/Yorkshire sows) in Southern Vietnam, a ten-week longitudinal study (September to November 2023) was conducted in the same herd. This closed-door, farrowing-to-wean farm had no prior history of ASF vaccination or reported cases. To understand the farm’s design and structure, refer to the layout and schematic perspective illustrated in Supplementary Fig. S1. The farm veterinarian recorded daily clinical signs, as well as the order, number, and location of infected/dead/cull pigs to aid in transmission analysis.

Sampling was carried out during the termination of the herd from October 24th to November 17th, 2023. A total of twenty multiparous, alive sows were randomly selected: twelve pregnant sows exhibited signs of reproductive failure (RF), such as abortions, stillbirths, and mummies; and eight lactating sows with multiple ulcerative dermatitis (UD) lesions. Anticoagulant whole blood was taken from suffering animals prior to their humane euthanasia using lethal doses of Succinyl chloride. Necropsy procedures were then performed to collect internal organs from the sows and fetuses aseptically (Supplementary Fig. S2). Additionally, whole blood samples from twenty-five healthy pigs, pooled in five specimens and samples employed to detect ASFV during the early outbreak stage were included (labeled “04Sep23_WB”). Fresh tissue samples were ground and homogenized at a ratio of 1:10 w/v in sterile phosphate-buffered saline, centrifuged for 5 min at 6000 g, and the supernatant was cryopreserved at – 80 °C until extraction. A separate portion of fresh tissues was fixed in 10% formalin buffer for further histopathological examination.

The research and animal sampling collection protocols in Vietnam complied with the institutional animal welfare guidelines and regulations approved by the Ministry of Agriculture and Rural Development (MARD) Vietnam (TCVN 8402:2010). This study was also reviewed and approved by the Chulalongkorn University Institutional Biosafety Committee (IBC, No. 2431021, Bangkok, Thailand). All studies complied with the ARRIVE guideline.

Data collection and key performance indicators (KPIs) calculation

Husbandry management procedure, biosecurity programs, pathogenic epidemiology, and reproductive performance were retrospectively obtained from the farm veterinarian’s weekly report. Data were categorized into two distinct stages: a “pre-outbreak” was from the first week of the productive year 2023 until a week before the outbreak (34th week), and ten-week subsequent monitoring was a “during the outbreak”. A dictionary of key performance indicators (KPIs) was created using industry-standard terminology and formulas. Mainly, farrowing rate, total born piglets per litter, born-alive piglets per litter, average birth weight, weaned piglets per litter, average weaning weight, and pre-weaning mortality rate were evaluated per each productive week. Details on KPIs calculation are provided in Supplementary Table S1.

DNA/RNA molecular extraction

In the laboratory, whole blood or tissue homogenates were extracted for viral DNA/RNA using the IndiSpin Pathogen Kit (Indical Bioscience GmbH, Leipzig, Germany), and complementary DNA (cDNA) was synthesized with the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific Inc., Waltham, MA, USA) following manufacturer’s instruction. Nucleic acid quantity and quality were assessed using Epoch2NS^® Microplate Spectrophotometer Reader and Gen5™ Microplate Software (Agilent Technologies Inc., Santa Clara, CA, USA) by measuring the A260/280 ratio. Extracted DNA was stored at – 80 °C until assays. A commercial ASFV-G-∆MGF vaccine (AVAC ASF LIVE^®, AVAC Viet Nam JSCo., Hung Yen, Vietnam) and a field Vietnamese ASFV virus, D/ASF/P1/Vietnam/2019²³, isolated from genotype-II-acute infected pigs’ tissue, served as positive controls for β-glucuronidases (GUS) gene and wild-type ASFV, respectively. The vaccine (Batch no. 0823) was purchased from a veterinary medicine store in November 2023 and extracted following the abovementioned instructions.

Molecular detection and genetic characterization by PCRs

Supplementary Figure S2 outlines the laboratory workflow for identifying an ASFV-G-∆MGF vaccine-like variant. An OIE-recommended qualitative (q) polymerase chain reaction (PCR) assay was utilized for ASFV DNA detection in whole blood and tissue specimens²⁴. The qPCR cycling condition was optimized for a fast ramp speed of the Luna^® Universal Probe qPCR Master Mix (New England Biolabs-NEB, Ipswich, MA, USA). Conventional (c) PCR was applied with a reaction mixture containing 12.5 µL of Q5^® High-Fidelity 2X Master Mix (NEB, Ipswich, MA, USA), 0.5 µM of each primer, 3.0 µL of DNA template, and nuclease-free water up to a total volume of 25 µL. The thermocycling profiles for qPCR and cPCR are presented in Supplementary Table S2. A specific and bright band from gel electrophoresis was cut and purified in preparation for DNA sequencing.

To identify the vaccine-like variants, the qPCR-positive samples with Ct-value below 35 were amplified by using PCR-specific targeted primers in numerous regions: (i) 478 bp within the 3’-end of the B646L gene as described previously⁸; (ii) the missing multiple family genes (MGF) 360/505 within the left variable region (LVR); (iii) the insertion of a β-GUS marker gene¹⁶, which is regularly used in vaccine development²⁵. By combining an MGF 505-2R primer²² with a newly designed MGF 360–12 L primer to detect the ASFV-G-∆MGF vaccine-like deletions. Two novel primer sets, 1R-GUS and GUS-3R, were custom-designed to amplify the full-length β-GUS gene according to the reference sequence (MasterSeed, PP529961). Escherichia coli (E. coli) DNA was included to control for false positives from contamination during sample collection and laboratory work.

These samples were subjected to detect the infection of other pathogens relevant to reproductive problems, such as classical swine fever virus, porcine reproductive and respiratory syndrome virus (PRRSV), porcine circovirus (PCV) type 2–3, porcine parvovirus (PPV), and suid herpesvirus type 1 as described previously protocols. Details of all primers used in this study are listed in Supplementary Table S3.

Sanger sequence, whole genome sequence, and phylogenetic analysis

Using the Sanger sequencing technique, the multi-target genes were sequenced from positive-cPCR samples by service (U2Bio, Bangkok, Thailand). The derived nucleotide sequences were assembled and aligned using BioEdit software v7.2.5 (Ibis Biosciences, Carlsbad, CA, USA) with the MUSCLE algorithm. Final consensus sequences were compared to other available NCBI database sequences using BLASTn. The phylogenetic tree of the B646L gene was generated using the MEGA X software package v10.2.6. The maximum likelihood with the best-fit nucleotide substitution Tamura-3 parameter model with gamma distribution (T92 + G) and 1000 bootstrap replicates were applied for ASFV genotyping.

Three udder tissue extractions from UD pigs with a Ct-value below 20 were chosen to attempt whole genome sequences (WGS, Table 1). A tiled amplicon-based sequencing Oxford Nanopore Platform was carried out following the procedure previously described²⁶ to enrich ASFV DNA before Nanopore sequencing. The brief sample preparation and WGS procedure are in the context of the Supplementary File. Nanopore reads were processed by trimming adapters, demultiplexing barcodes using Porechop (v0.5.0; https://github.com/rrwick/Porechop, accessed on 21 July 2024), and filtering short reads less than 50 bp with a quality threshold of 10 using BBDuk v38.84. To remove the host genome and duplicate reads in trimmed and filtered reads, following microbes’ identification and characterization using the v0.7 bioinformatics pipeline (Chan Zuckerberg ID, https://www.cizd.org, accessed on 10 August 2024). The initial consensus genomes were assembled using a mapping-to-reference algorithm with Minimap2 (v2.24)²⁷ in Geneious Prime (v2024.0.7)²⁸, followed by manual alignment-based corrections in MAFFT (v.7.490)²⁹ to complete the final sequence. WGS was sequenced and assembled by the Virology Unit, Institut Pasteur du Cambodge (Phnom Penh, Cambodia). ASFV sequencing database of Georgia 2007/1 (FR682468), VN/HY/2019 (MT872723), and MasterSeed strain (PP529961) were reference genomes.

Pathological examination

The farm’s veterinarian noted abnormal gross lesions during the investigation period. Six animals showing Ct-value below 40 in all organs from qPCR (three RF sows: #33700, #30256, #03300 and three UD sows: #32242, #32005, #34550) underwent further pathological examination (Table 1). The formalin-fixed paraffin-embedded tissues were sectioned at 4 μm thickness and stained with hematoxylin and eosin for histological assessment.

Immunohistochemistry stain

As in the previous protocol description, the immunohistochemistry (IHC) technique was employed to detect ASFV antigens in infected tissues, with minor optimization³⁰. A 1:200 dilution of a commercial monoclonal antibody targeting p72 capsid protein (clone 1BC11, Ingensa Inc., Madrid, Spain) was used as the primary antibody. The Histofine Simple Stain MAX PO (Multi; Nichirei Bioscience, Tokyo, Japan) was used as the secondary antibody. Negative controls involved treating slides with 1% bovine serum albumin in Tris-buffered saline instead of the primary antibody. Positive IHC results displayed the brown-yellow or brown-red staining in target cells, whilst no indicator color was visualized in the negative controls.

Statistical analysis

The distribution of KPIs data was verified by the Shapiro-Wilk Normal Distribution test. Descriptive analysis calculated each parameter’s mean and standard deviation within each surveyed period. With the non-parametric data, the Mann-Whitney U test was used to compare reproductive performance between pre- and during-outbreak periods. All statistical analyses were handled using Graph Prism 9 (GraphPad Software Inc., San Diego, CA, USA) with a significant level of P < 0.05.

Clinical signs observation and transmission features

In July 2023, replacement gilts introduced to the breeding house suspiciously manifested fever and anorexia. Retrospective data revealed a high replacement rate (up to 60% of sow heads) and gilts sourced internally from non-vaccinated herds. The farm’s ineffective biosecurity measures were documented from personal investigation, including traders’ frequent entry of cull pig transport vehicles. Meanwhile, several ASF outbreaks were reported within a 5–10 km radius of the investigated farm. During two months of acclimatization, gilts recorded atypical symptoms such as respiratory distress and prolonged loss of appetite. However, antibiotics, anti-inflammatory, and supplement treatments resulted in recovery within 3–4 days. As the pathogenic monitoring program, before the gilts entered the productive herd, laboratory services’ ASFV-free results were validated in ten pooled samples (5 gilts/sample) of roughly 250 replacement gilts.

After onset in gilts, sporadic occurrences in pregnant sows in other houses showed similar signs. Until early September 2023, four dead pigs were noticed: one lactating sow and three pregnant sows. Two out of three pregnant sows had only abortion, whereas long-term anorexia, nasal discharge, and intermittent fever in others. All four dead animals were sampled for ASFV diagnosis by laboratory service. The initial PCR result from the 04Sep23_WB sample showed ASFV-positive without the MGF 360–12 L gene, a portion of the deleted region in the current commercial vaccine, using multiplex PCR previously described³¹. Within the observed period, before the total depopulation of all herd, the farm losses of over 5% (111/2160) fatality, and the cull rate was 80% (Fig. 1A). The incidence rate of RF was up to 18% of the total pregnant sows, with aborted manifestation in the late-term gestation (Fig. 1B). Almost 20% of sows obtained anorectic, and approximately 25% of total lactating sows exhibited moderate-to-severe ulcerative dermatitis in the skin and udders (Figs. 1C and 4A). In addition, other symptoms such as cutaneous hyperemia, arthritis, lethargy, wavering stand, and bleeding nasal discharges also appeared at a low rate (< 1%).

The swine population in the herd and significant observed clinical signs in ASFV-infected sows. (A) A total of production animals was recorded each week during the study period. (B) Reproductive failure in pregnant sows in the third stage of gestation with aborted mummifies. (C) Moderate-to-severe multiple dermatologic ulcerations in lactating sows (arrowhead).

Full size image

A gestational house capacity for 536 cages was monitored for six weeks to simulate spatial and temporal transmission features. In the first week, three sudden fetal sows and a symptomized sow were adjacent (Supplementary Fig. S3A). Following this, new-onset cases are probably disseminated in almost any cages in the house, surrounding old cases (Supplementary Fig. S3B). Notably, a subset of pigs close to the symptomatic animals in the front weeks was removed to intercept the viral spreading (Supplementary Figs. S3C and D). After a period, over 50% of sows in the house were estimated ASFV-positive by clinical-pathological signs or the point-of-care lateral flow tests (Supplementary Figs. S3E and F). This control and management followed the selective culling procedure.

KPIs evaluation

A gradual drop in sow headcounts due to ASFV infection, removal strategy, and the difference in duration of the two phases resulted in a non-normally distributed sample data (Fig. 1A). A Mann–Whitney U non-parametric test compared the herd’s production performance prior to the outbreak and within the studied period (Table 2). All KPIs significantly declined during the ASF outbreak (P < 0.0001), except weaning weights, which had an analyzed value of P = 0.0031. The farrowing rate dropped by roughly 28%, and the total born was less than 1.59 piglets, primarily due to a reduction of 2.73 piglets born alive per litter. Additionally, the pre-weaning mortality rate markedly increased by nearly 30% compared to the pre-outbreak stage.

ASFV and co-infected pathogens detection

Seven and six individual samples per female were collected from twelve RF and eight UD pigs, respectively (Supplementary Table S4). All pigs tested positive for the p72-target qPCR assays in at least one testing sample. In RF sows, only 8/12 (66.67%) sows were shown Ct-value (range: 22.41–36.87) from their samples (21/71, 29.58%). Meanwhile, the remaining four sows had a DNA molecular positivity in the fetuses’ pooled internal organs. In UD sows, ASFV nucleotides were presented in widespan samples (41/48, 84.42%).

Among numerous sample types examined, whole blood samples, typically utilized for ASFV monitoring, were exposed in 8/20 (40%) with relatively high Ct-values (33.81–38.02). Udder tissues harbored viral DNA in almost all UD sows, and three suffered RF sows with udder ulcerations due to the fostering of neonatal piglets from the other sows. Remarkably, three UD sows carried high viral loads in their udder specimens (Ct-value: 18.41, 18.99, and 19.50 for pigs #32242, #32005, and #34550, respectively, Table 1). Separate organs in the reproductive tract and pooled internal organs of sows were found to be ASFV-positive with moderate-to-high frequency. Additionally, ten aborted fetal samples from 12 RF sows affirmed ASFV infection (Ct-value: 22.84–36.61, Table 1 and Supplementary Table S4). In the healthy pig without clinical signs examined, one out of five pooled samples was found positive with ASFV DNA (Ct-value: 35.31, Supplementary Table S4).

PCV3 was only detected as a co-infected pathogen in four out of twelve RF sows’ whole blood: #33,700 (34.05), #03392 (36.68), #03820 (37.01), and #33,042 (37.93). No DNA molecule detection of other reproductive-relevant agents in all sows’ and fetuses’ tissues (data not shown).

An ASF vaccine-like variant identification

A preliminary multiplex qPCR report detected ASFV and identified a partial deletion within the MGF 360–12 L gene³¹; two cPCR-based primer pairs verified this data. Results showed that the absence of MGF 360–12 L and MGF 505-2R differentiated from the field isolates containing these virulent-related genes (Fig. 2A and B). Besides, a 478 bp insertion of the internal β-GUS gene was found in both the outbreak-confirmed and late-stage samples. A lower band observed in gel electrophoresis was interpreted as non-specific amplification when using the previously published primers for marker gene monitoring¹⁶ (Fig. 2C). Subsequent sequencing clarified this as a fragment of the host chromosome. PCR reactions with new 1R-GUS and GUS-3R primers produced the complete inserted β-GUS gene sequence without cross-amplification with E. coli DNA (Fig. 2D and E). These comparable molecular outcomes across samples collected at different time points imply that a specific ASFV-G-∆MGF vaccine-like variant was circulating in this herd.

Molecular detection of ASFV-G-∆MGF vaccine-like variant among field samples, E. coli, and vaccine DNA by conventional PCR. (A) MGF 360–12 L gene and (B) MGF 505-2R gene amplification. (C) A part of the inner β-GUS marker gene detection with a non-specific amplification. (D) Amplicons length 1305 bp from 1R-GUS primer and (E) 1449 bp from GUS-3R primer. The PCR products were run gel electrophoresis in 1% agarose and measured with a 1500 bp DNA ladder.

Full size image

Phylogenetic analysis

Genotyping was analyzed according to a partial C-terminus sequence of the B646L gene, which encodes for the major capsid p72 protein. The investigated samples belong to genotype II, consistent with the predominant strains in Eurasia (Georgia 2007/1), and highly resembling the MasterSeed strain used to develop the ASFV-G-∆MGF engineering attenuated vaccine (Supplementary Fig. S4).

Only one amplicon-based sequencing from pig #32,005’s udder tissue resulted in a significant number of sequence reads mapped to the ASFV genome, achieving 81% genome coverage and a median depth of 297x. The WGS from this sample was assembled, annotated, and deposited in the NCBI GenBank database under accession number PQ629526, titled “VN/sows_32005/23”. As anticipated, after proofreading, the WGS authenticated that VN/sows_32005/23 missed six MGF genes in the LVR of the Georgia 2007/1 genome: specifically, three MGF 505-1R, -2R, -3R, and three MGF 360–12 L, -13 L, -14 L. Instead of this region, a 2,348 bp of β-GUS marker gene was inserted (Fig. 3). These observed genetic modifications were utterly consistent with those reported for the live-attenuated recombinant strain, a vaccine candidate described in the previous study¹⁶.

Schematic diagram of indel mutations among ASF strains: vaccine-like (VN/sows_32005/23), Eurasian isolate (Georgia 2007/1), the first Vietnamese outbreak isolates (VN/HY/2019), and vaccine strain (MasterSeed). Reference strains are downloaded from the NCBI GenBank database. Gene deletions, insertions, and partial deletions were visualized as dotted lines, yellow boxes, and red arrows, respectively. Reference genes were depicted as dark blue arrows with open reading frame annotation above.

Full size image

Further pathogenetic analysis using Georgia 2007/1, MasterSeed, and the early Vietnamese outbreak isolates revealed three major deleted fragments in LVR and a central conserved region containing several pathogenic-associated and unknown-functional genes (Fig. 3). The first fragment completely lacked genes related to MGF 360 and 110 (from 5,545 to 8,655). The second fragment deleted the entire sequences of A224L, ASFV G ACD 00600, A104R, A240L, A118R, A151R, and the MGF 360-15R gene, alongside partial truncations of two other genes, MGF 505-10R and A238L (from 46,535 to 57,578). The last fragment included part of the EP1242L gene and the entire EP84R, EP424R, EP152R, and EP153R genes (from 70,559 to 74,688, with position located in the Georgia 2007/1 strain).

Interestingly, in the third missing fragment, VN/sows_32005/23 showed a 334 bp absence of the 5’-end of the EP402R gene, responding to viral hemadsorption, and five inserted nucleotides “AATAT” at position 469 of the gene (Supplementary Fig. S5). Likewise, the MGF 505-5R gene experienced a loss of 1,047 nucleotides, leading to the omission of 349 amino acids. The novel deleted genes and sequence arrangement of the EP402R and MGF 505-5R genes were distinctive, as they displayed no significant homology to other sequences previously cataloged in the GenBank repository.

Regarding mutations, VN/sows_32005/23 dispersed multiple single nucleotide polymorphisms (SNPs) across the entire viral genome. A few substitutions in our sequence closely resembled the MasterSeed virus and ASFV-GUS-Vietnam isolates, disparate from the Georgia 2007/1 parental strain. Furthermore, the VN/sows_32005/23 isolate featured additional unique substitutions markedly divergent from reference genomes. The remaining genetic variations comprised silent mutations that did not alter the encoded amino acid sequence or were situated within the non-coding regions (Table 3).

Pathological and immunostaining findings

Grossly, all pathologically examined sows revealed variable-sized, multifocal to confluent skin ulcerations primarily located on the udders and other dependent body regions. These ulcerative lesions ranged from discrete, shallow erosions to deeper, more extensive skin necrosis and sloughing areas (Figs. 1C and 4A). The two affected sows displayed visible skin erythema, indicating an inflammatory response. Additionally, all the infected sows suffered from abortion, experiencing pregnancy loss and reproductive complications (Fig. 1B). Interestingly, no characteristic macroscopic lesions of ASFV genotype-II-acute infection were noted in the reproductive organs.

Pathological and immunohistochemical findings in ASF-infected sows. (A) A variable-sized ulceration on the udder of lactating sows at 1–2 weeks postpartum. (B) Massive necrosuppurative inflammation of the skin that extended from the epidermis to the reticular dermis. (C) Vasculitis (white arrowhead) with thrombus (T) at the interstitium of the mammary gland. (D) The lymphoid necrosis (N) in the germinal center of the lymph node. ASFV-antigen localization was detected in the macrophage-like cells in infected lesions (arrowhead, insets).

Full size image

Microscopically, ulcerative dermatitis is characterized by extensive and widespread necrosis of the epidermal layer, extending deep into the reticular dermis. This condition is accompanied by a marked infiltration and accumulation of neutrophilic inflammatory cells throughout the affected skin tissue (Fig. 4B). On occasion, neutrophils and lymphocytes were present in the glandular lumen of the mammary gland. Intriguingly, massive inflammatory cell infiltration and significant edema were noticed within the interstitial spaces of the mammary gland. Vasculitis is defined by inflammation of the blood vessel walls with fibrinoid degeneration, and the formation of thrombosis, or blood clots, within the vessels was observed (Fig. 4C). The germinal centers of the lymphoid follicles exhibited extensive necrosis, with marked destruction and loss of the typical follicular architecture (Fig. 4D).

Immunoreactivity to the ASFV antigen was demonstrated in the cytoplasm of macrophage-like cells in the skin lesions (Fig. 4A), the interstitium of the mammary glands (Fig. 4B), the lymph nodes (Fig. 4C), and the ovaries (Table 4). These findings first described viral distribution in udders and reproductive organs in affected sows. The viral antigens’ presence within these macrophages indicates their role as a critical target and reservoir for the virus, contributing to the persistent nature and dissemination of the ASFV infection.

ASF acutely impairs reproductive performance, resulting in pig losses and mortality across affected herds. Effective disease management and eradication require efficient vaccines and rigorous biosecurity protocols. Following the ASF’s endorsement in Southeast Asia, vaccine research and production have accelerated. Genetically engineered LAVs are promising vaccine candidates³². Many commercial ASF-modified LAVs are now available in Vietnam. Although their safety and efficacy have been assessed, some LAVs have been associated with elevated viral replication and shedding in laboratory and field trials^33,34. To identify the ability of vaccine seeds to spill over after nationwide vaccination, an ASF-negative breeding herd in an endemic area was monitored for ten weeks. By analyzing ASFV-specific genes and β-GUS reporter gene in atypically symptomatic sows, a variant analogous to the ASFV-G-∆MGF vaccine was identified in Southern Vietnam.

Genomic analysis revealed that the VN/sows_32005/23 variant had lost numerous genes relevant to viral manipulation and immune evasion, in addition to replacing specific marker genes for MGF genes (Fig. 3). The ASF pathogenesis impedes apoptosis within infected cells during the early phases, thereby promoting viral replication and propagation³⁵. The deleted MGF 360–505 genes and the A224L gene alleviated the infected macrophages, primarily targeted cells of ASFV, undergoing programmed cell death, ultimately attenuating viral replication^36,37. Other deleted genes, A104R, A151R, EP152R, and A238L, have been implicated in ASFV virulence and host immune modulation^38,39,40,41. The EP402R gene (CD2v protein) facilitates ASFV dissemination in blood circulation and causes viremia by binding to host erythrocytes⁴². Various research studies on recombinant strains with single or multiple deletions of EP402R and other harmful genes, such as the EP153R gene, were shown to drop a viral load in the bloodstream and attenuate infected pigs⁴³. The VN/sows_32005/23 variant’s high Ct-value and genetic modifications suggested that it may have impaired CD2v protein production, similar to previously published attenuated strains (Supplementary Fig. S5)¹¹. Unfortunately, this study did not perform hemadsorption inhibition assays because the ASF-live virus was not isolated from samples. This ASF genetic variant appeared to originate from a genetically attenuated vaccine characterized by replacing six virulent-related MGF 360/505 genes with the β-GUS marker gene, potentially due to many factors during its infection, for instance, under selective pressure exerted by the host’s immune defenses. Genetic mutations can facilitate the virus’s adaptation to the host’s cellular environment, evading immune recognition and suppressing immune responses, thereby prolonging the infection, as shown in this study⁴⁴.

Typical clinical signs of ASFV infection in affected pigs include intermittent fever, loss of appetite, respiratory distress, and reproductive disorders. The ASFV-G-∆MGF vaccine-like variant was the sole pathogen detected in all RF sows and their mummified fetuses, apart from only four RF sows co-infected with PCV3, but no PCV3 detection in fetuses. This suggests that PCV3 may not significantly contribute to these animals’ reproductive problems. Notably, the high symptomatic incidence of RF was obtained in multiparous sows and those in late-term gestation (Fig. 1B). These findings aligned with previous reports of varying infection types in female pigs and increased abortion rates speculated in chronic form due to animals’ prolonged exposure to the virus^2,3,7,35. Moreover, 25% of lactating sows in this study presented multiple udder ulcerations and dermatological lesions (Figs. 1C and 4A), primarily relevant to low-virulent infections as previously elucidated³.

From the histopathological results, extensive fibrin-necrotizing vasculitis with inflammatory cell infiltration caused by vascular damage and fibrinoid degeneration were shown in skin lesions and lymphoid organs in examined sows (Fig. 4). These pathological changes were often compatible with immune complex-induced lesions, mainly aberrating cytokine-mediated responses triggered by infection and monocyte/macrophage activations rather than direct virus-induced damage. ASFV-infected macrophages released potent proinflammatory cytokines such as interleukin (IL)-1, IL-6, and tumor necrosis factor-α (TNF-α), subsequently leading to endothelial cell lysis and predisposing to systemic vasculitis³⁵. Other experiments also demonstrated the enhanced interferon type I activations, stimulating the apoptosis-associated cytokine release of the IL-1 group (pro-apoptosis) and IL-18 (potent IFN-γ induction) from classically activated macrophages and monocytes in ASFV-attenuated infections (NH/P68 and BA71V strains) compared to highly pathogenic infection^45,46.

Although the fatality rate in the studied farm was lower than that of lethal ASF strains, the selective culling strategy may have influenced this data (roughly 80% cull rate, Fig. 1A). In a previous study, two out of three 100-days-gestational sows died on the 5th post-inoculated day by a Georgia 2007/1 isolate after showing severe clinical signs⁴⁷. Another study infected nursery pigs with the ASFV-GUS-Vietnam strain, which had six absent MGF 360/505 gens, and the β-GUS gene presence showed the fatality occurred at 2–3 weeks post-challenge. Horizontal transmission of this strain was also observed in the “contact pigs” and the recipients developed a mild fever on day 14 of the trial and died a week later²². While this variant’s direct contact transmission resembles those of strains, external factors such as removed pigs’ handling, human movements, husbandry management, and biosecurity practices likely played a role (Supplementary Fig. S3).

The ASFV molecules were detected in almost all internal organs of sampled pigs, including the reproductive tract of RF sows, entire fetal tissues, and the udder tissue of UD farrowing sows (Table 1 and Supplementary Table S4). Genetic characteristics in positive sows and fetuses were homologous (data not shown), predicting that the ASFV-G-∆MGF vaccine-like variant could transmit vertically from sows to their progeny in utero. Empirical research on reproductive issues in livestock breeding has consistently shown that viral etiologies, such as PCVs, PRRSV, and PPV, can be a potential transplacental transmission⁴⁸. In Lohse’s experiment, the vertical transmission was limited when the fetus carried a lower viral load than sows, which contrasted with the outcomes of this study. The prior research utilized an ASFV lethal strain administered in a short timeframe (5 days) and at a high dosage, likely constraining the virus manipulation within the fetal tissues⁴⁷. Meanwhile, this study observed natural field conditions when the exact ASFV exposure time point remains unclear. Most of the aborted mummies yielded a high viral load, coupled with the undetectable virus in sows’ blood and tissues; this suggested that survived sows contacted and carried the virus long before the disease onset. The immune systems of pregnant sows in the last trimester possibly mounted to eradicate or control it persistently in low concentrations, but only after the virus crossed the placenta barrier and infected fetuses with immature immunity⁴⁹.

Immunostaining analysis identified the presence of ASFV antigens in macrophage-like cells within various internal organs of infected sows despite the absence of typical macroscopic lesions, especially in reproductive organs (Fig. 4; Table 4). Regrettably, IHC in fetal tissues was not completed in this study. However, in the previous promulgation, the ASFV-G-∆MGF vaccine strain was cultured in specific DMAC cell lines derived from pig embryonic macrophages¹⁸, which differs from other common cell lines for ASF-modified LAVs development. With this information and qPCR results, it is questioned whether the novel ASFV-G-∆MGF vaccine-like variant is more susceptible to infecting reproductive organs and replicating in fetal myeloid cells as the behavior of the vaccine strain. This heightened tropism possibly leads to increased reproductive failure and embryonic mortality in affected sows, which has not been previously reported in the ASF highly fatal infection form.

Regarding healthy/asymptomatic pigs, viral DNA was detected in pooled whole blood samples, suggesting their potential role as reservoirs for viral dissemination or re-outbreaks (Supplementary Table S4). An early study reported that survival pigs in endemic areas might exhibit no viremia or viral excretion after 12 weeks of infection⁵⁰. Hence, further investigation is needed to determine the viral load dynamics and potential mechanisms for virus clearance or control in these surviving pigs. Understanding these factors could provide valuable insights for developing effective treatment strategies against ASF.

The KPIs are the farmers’ most concerning patterns to reflect a livestock herd’s economic reality. During ASFV circulating in this studied herd, KPIs were enormously reduced compared to the “pre-outbreak” stages (Table 2). Convalescent gilts or sows used for breeders may experience long-term impacts of up to 14 months⁵¹. According to retrospective data, the farm in this study was negative for ASF and had never been vaccinated. The outbreak occurred following intensive gilt replacement. In the meantime, various outbreaks occurred in nearby small-scale farms. It was hypothesized that the ASFV-G-∆MGF vaccine-like variant was perhaps introduced into the herd during this period. Vehicles used by traders to transport livestock between multiple farms and abattoirs, as well as inadequate biosecurity measures, speculatively facilitate viral entry. The origin of replacement gilts and semen employed for artificial insemination was a concern following the identification of ASFV in the reproductive organs of mature boars⁵². Before entering, an appropriate acclimatization protocol for replacement gilts, serological profiles, and regular pathogen monitoring programs are highly recommended for adequate implementation.

A part of this study points out that various human activities may contribute to the emergence of genetic diversity in ASF variants, negatively affecting disease control and mitigation efforts. Incomplete culling of the infected animals can leave behind infection reservoirs, letting the virus be able to circulate and mutate⁵³. While ASF LAVs are approved in Vietnam, effective vaccination programs for different epidemiological scenarios remain uncertain. Improper handling of vaccinated pigs and animal movement from ASF-vaccinated areas to the non-vaccinated zone could increase the risk of virus transmission into the naïve populations. Zoning should be implemented in the ASF-free regions to prohibit the introduction of ASF-vaccinated animals together with attenuated viruses into these zones. Also, counterfeit and substandard vaccines may promote ASF replication and the selection of resistant strains, as seen in mainland China and emphasized by the World Organization for Animal Health (WOAH)^54,55,56. To address these issues, comprehensive culling, rigorous surveillance and monitoring of vaccinated herds, regular on-farm inspections to assess stricter biosecurity compliance (restricted access to the farm, throughout cleaning and disinfection of vehicles, personnel, and facilities, all-in/all-out production system), and more vigorous enforcement of regulations against unauthorized vaccines are recommended^57,58.

Farmers remained skeptical about the vaccine despite extensive research, underscoring the need to balance scientific rigor and practical considerations. Animal studies and clinical settings have varied results due to multiple factors. Animal experiments provide a highly controlled setting to precisely measure vaccines’ action mechanisms, immune responses, and side effects, leveraging a restricted number of subjects within the constrained timeframe of the study. Contrarily, clinical trials test vaccines in field conditions to assess their effectiveness and safety in protecting animals, using a larger number of animals to ensure robust statistics while prioritizing animal welfare⁵⁹. Routinely approached in the practical farm scenarios, however, may yield unpredictable outcomes since many variables may be influenced, as previously discussed^{14,20,21,22,33}. Furthermore, vaccine safety in practice may face challenges from stress factors affecting pig health, interactions with pathogens or the microbiota present in the vaccinated pig.

The confined nature of this single-farm study constrains the generalizability of its findings. Further research on a larger scale is necessary to validate the observed disease transmission patterns and clinical manifestations. In addition, the long-term implications of infection, such as the potential for carrier animals and the persistence of the virus in the environment, warrant investigation. Regarding viral virulence and immune activation, the effectiveness of the VN/sows_32005/23 strain’s genetic variations in pathogenicity should be further verified through both in vivo and in vitro experiments.

A novel ASFV-G-∆MGF vaccine-like strain, the “VN/sows_32005/23” virus, was identified in an ASF-free breeding herd after a high gilt replacement rate and improper biosecurity measures. The outbreak caused by this variant proved an atypical reproductive failure in the pregnant sows and placental transmission ability, resulting in significant losses within the breeding population and viral spreading into farrow to finish pig populations, respectively. This alarming discovery highlights the potential risks associated with the use of genetically engineered ASF vaccines. Even though vaccines have been launched on the market, stringent biosecurity and husbandry management still play critical keys in fighting diseases. Last but not least, farmers and veterinary practitioners must exercise prudence and adopt robust safety protocols when replenishing livestock herds or using ASF-modified LAVs in ASF-free regions, followed by comprehensive surveillance and strong oversight, particularly among livestock producers.