Genetic heterogeneity and pathogenic potential of historical Crimean-Congo hemorrhagic fever virus isolates in China

Abstract

The Crimean-Congo hemorrhagic fever virus (CCHFV) poses a significant public health threat. In China, CCHFV has been circulating for decades, yet the genomic diversity and pathogenic potential of the circulating strains remain poorly characterized, hindering risk assessment and countermeasure development. In this study, we recovered 24 historical CCHFV strains isolated between 1966 and 2004 from humans, ticks and jerboas in Xinjiang Uyghur Autonomous Region of China. Whole-genome sequencing was performed, followed by comprehensive analyses of their phylogenetic relationships, in vitro infectivity and in vivo pathogenicity. Phylogenetic analyses revealed high genetic heterogeneity, identifying seven genotypes for the L segment, nine for the M segment (including a novel Asia 4 genotype), and nine for the S segment. Amino acid mutation analysis revealed that the mucin-like domain (MLD) of the glycoprotein (GP) exhibited the highest mutation rate, contributing substantially to sequence diversity. In vitro, Asia 2 (75024) and Asia 3 (79121M18) strains exhibited robust replication in monkey-, hamster-, and human-derived cell lines. In C57BL/6 mice, all four representative strains induced viral replication and specific antibody responses (IgM and IgG), causing mild to moderate pathological damage in the liver, spleen, and kidneys. In IFNAR^−/− mice, virulence varied markedly among representative strains: Asia 2 and Asia 3 strains were highly lethal (LD₅₀ < 1 TCID₅₀), Asia 1 was moderately virulent (LD₅₀ = 142.5 TCID₅₀), and Asia 4 exhibited atypical, non-dose-dependent mortality. Collectively, our work reports a novel Asia 4 genotype and suggests strain- and lineage-associated differences in virulence for CCHFV in China, providing critical insights for surveillance and targeted countermeasure development.

Highlights

• •Historical CCHFV strains in China reveal a novel Asia 4 genotype.
• •The glycoprotein's mucin-like domain shows the highest mutation rate.
• •Virulence in IFNAR^−/− mice is genotype-dependent, with Asia 2/3 being highly lethal.
• •Representative strains induce antibody responses and organ pathology in immunocompetent mice.

Introduction

Crimean-Congo hemorrhagic fever (CCHF) is a widespread and severe febrile illness caused by the Crimean-Congo hemorrhagic fever virus (CCHFV). It poses a significant global public health threat because of its high case-fatality rate of 10%–40%, the absence of approved vaccines or specific antiviral therapies, and its broad geographical distribution (Frank et al., 2024; Kaushal et al., 2025). CCHFV is endemic to numerous countries in Africa, Asia, the Middle East, and Southeastern Europe (Fereidouni et al., 2023, 2025). The disease is characterized by the rapid onset of high fever and myalgia, and in severe cases, progressive hemorrhagic manifestations that can lead to hepatic failure and multi-organ dysfunction (Frank et al., 2024). CCHFV is maintained in a tick-vertebrate host cycle, primarily involving Hyalomma ticks, with sporadic spillover to humans through tick bites or contact with infected animal tissues (Bhowmick et al., 2022). The World Health Organization (WHO) has designated CCHFV as a priority pathogen that requires urgent research and development (Ozdarendeli, 2023). This situation underscores the critical need to deepen our understanding of CCHFV biology and pathogenesis to facilitate the development of effective countermeasures (Freitas et al., 2022).

CCHFV is a member of the genus Orthonairovirus in the family Nairoviridae. It is an enveloped, trisegmented, negative-sense RNA virus with a genome comprising small (S), medium (M), and large (L) segments (Iglesias-Rivas et al., 2025). The genetic evolution of CCHFV is closely linked to its geographical distribution, and the high potential for mutation and segment reassortment of its RNA genome has fostered substantial genetic diversity. Globally, at least nine genotypes have been identified: Europe 1–3, Africa 1–3, and Asia 1–3 (D'Addiego et al., 2024; Guo et al., 2017). Genetic differences among genotypes may influence viral phenotypes, including their pathogenicity. Thus, the genotypic variation could be directly translate into substantial differences in disease severity and fatal outcomes. CCHFV displays significant genotypic diversity, however, whether this diversity drives variations in infectivity and pathogenicity remains unclear. Although independent investigations of individual genotypes have yielded clues to potential virulence disparities (Rao et al., 2026), systematic research dedicated to cross-comparative analysis of pathogenicity across distinct genotypes remains scarce.

First reported in Xinjiang, China, in 1965 (Guo et al., 2017), CCHF has caused over 330 confirmed human cases in the following decades (Teng et al., 2022). Although no confirmed human CCHF cases have been reported in Xinjiang in the past two decades, the virus has been consistently detected and isolated from local tick populations (Guo et al., 2017; Zhang et al., 2018). The minimum infection rate of CCHFV in ticks is 6.85% (Moming et al., 2018), indicating that the potential risk of CCHFV transmission in Xinjiang remains substantial. Recent epidemiological modeling suggests that areas in China at high risk for CCHF epidemics are far more extensive than previously recognized, exposing approximately 135 million people to potential infection (Teng et al., 2022). Although genetic differences among CCHFV strains isolated from ticks have been identified (Moming et al., 2018), it remains unknown whether these subtypes differ significantly in virulence and pathogenic potential. Moreover, the scarcity of viral isolates from early outbreaks because of technical limitations has hindered the comprehensive reconstruction of the virus's evolutionary properties in China, resulting in a limited understanding of the virulence and pathogenicity of Chinese CCHFV strains.

This study aimed to systematically characterize the genetic diversity and pathogenic potential of CCHFV strains in China. We recovered and performed whole-genome sequencing of 24 historical CCHFV isolates spanning several decades. Subsequent phylogenetic analyses were conducted to elucidate the evolutionary relationships. Furthermore, we selected representative strains from the identified genotypes to compare their infectivity and replication efficiency in a panel of cell lines and assess their pathogenicity in mouse models. The findings of this study provide crucial insights into the genetic landscape and biological characteristics of CCHFV in China, which will facilitate the development of more effective strategies for disease surveillance, risk assessment, and prevention.

Results

CCHFV isolates from China are genetically heterogeneous, encompassing several genotypes

Twenty-four CCHFV strains were isolated from diverse host origins during 1966–2004 using the method of suckling mouse inoculation (Tang et al., 2023), and lyophilized brain homogenates from infected suckling mice of the 24 strains were preserved in National Virus Resource Center (NVRC) of the Wuhan Institute of Virology, Chinese Academy of Sciences. Of these, strains 66019, BA68038, BA69034, A7001, A73017, 75024, A76048, HF7802, BA8004, and BA84051 were isolated from humans; strains Ha75001, 84007, BA84021, BA84027, BA04032, and YL04041 were isolated from ticks; strain 79121M18 was isolated from jerboas; and the host sources of the remaining seven strains were unknown (Table 1). These strains were resuscitated by suckling mouse inoculation, followed by cell culture, and genomic sequences were obtained. Phylogenetic analyses based on the sequences of the L, M, and S segments revealed that whereas all the previously reported Chinese CCHFV strains were classified mainly into Asia 1–3, the inclusion of the 24 strains demonstrated the presence of an additional genotype containing Chinese CCHFV strains. Among the 24 strains, 11 (BA68038, BA69034, A73017, 75024, Ha75001, A76048, BA7708, HF7802, 7903, BA7902, and 88017) were classified as the Asia 2 genotype across all three segments. Four strains (BA84051, 84007, BA84027, and 88168) were classified as Asia 1 for the L and M segments, while their S segments belonged to Asia 2. Three strains (BA68045, BA84002, and BA84021) were assigned to Asia 2 for the L and S segments but to Asia 4 for the M segment. Two strains (BA04032 and YL04041) were categorized as Asia 3 for the L and S segments and Asia 4 for the M segment. Further, two strains (66019 and BA8004) belonged to Asia 2 for the L and S segments but Asia 1 for the M segment. Strain A7001 was classified as Asia 1 for the L and M segments and Asia 3 for the S segment; only strain 79121M18 belonged to the Asia 3 genotype across all segments (Fig. 1A–C). Phylogenetic analysis of CCHFV strains from China revealed a predominantly tick origin for the Asia 3 and Asia 4 genotypes. All strains in the L, M, and S segments were tick-derived, except for strain 7001 in the M and S segments.

Table 1.

Summary of CCHFV isolates used in this study.

Strains	Year	Source	L	M	S	Storage number	GeneBank accession numbers (S, M, L)
66019	1966	Human	Asia 2	Asia 1	Asia 2	CSTR:16533.06.IVCAS 6.7393	OR047145, OR047165, OR047185
BA68038	1968	Human	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7397	OR047134, OR047154, OR047174
BA68045	1968	Unknown	Asia 2	Asia 4	Asia 2	CSTR:16533.06.IVCAS 6.7396	OR047133, OR047153, OR047173
BA69034	1969	Human	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7394	OR047132, OR047152, OR047172
A7001	1970	Human	Asia 1	Asia 1	Asia 3	CSTR:16533.06.IVCAS 6.7417	OR047141, OR047161, OR047181
A73017	1973	Human	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7379	OR047140, OR047160, OR047180
75024	1975	Human	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7414	OR047144, OR047164, OR047184
Ha75001	1975	Tick	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7420	OR047129, OR047149, OR047169
A76048	1976	Human	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7416	OR047139, OR047159, OR047179
BA7708	1977	Unknown	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7415	OR047137, OR047157, OR047177
HF7802	1978	Human	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7399	OR047128, OR047148, OR047168
7903	1979	Unknown	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS6.7430	OR047146, OR047166, OR047186
BA7902	1979	Unknown	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7400	OR047136, OR047156, OR047176
BA8004	1980	Human	Asia 2	Asia 1	Asia 2	CSTR:16533.06.IVCAS 6.7418	OR047135, OR047155, OR047175
BA84002	1984	Unknown	Asia 2	Asia 4	Asia 2	CSTR:16533.06.IVCAS 6.7413	OR047131, OR047151, OR047171
BA84051	1984	Human	Asia 1	Asia 1	Asia 2	CSTR:16533.06.IVCAS 6.7401	OR047130, OR047150, OR047170
84007	1984	Tick	Asia 1	Asia 1	Asia 2	CSTR:16533.06.IVCAS 6.7410	OR823870, OR823871, OR823872
BA84021	1984	Tick	Asia 2	Asia 4	Asia 2	CSTR:16533.06.IVCAS 6.7408	OR823873, OR823874, OR823875
BA84027	1984	Tick	Asia 1	Asia 1	Asia 2	CSTR:16533.06.IVCAS 6.7411	OR823876, OR823877, OR823878
88017	1988	Unknown	Asia 2	Asia 2	Asia 2	CSTR:16533.06.IVCAS 6.7424	OR047143, OR047163, OR047183
88168	1988	Human	Asia 1	Asia 1	Asia 2	CSTR:16533.06.IVCAS 6.7427	OR047142, OR047162, OR047182
BA04032	2004	Tick	Asia 3	Asia 4	Asia 3	CSTR:16533.06.IVCAS 6.7368	OR047138, OR047158, OR047178
YL04041	2004	Tick	Asia 3	Asia 4	Asia 3	CSTR:16533.06.IVCAS 6.7382	OR047127, OR047147, OR047167
79121M18a	1979	Jerboa	Asia 3	Asia 3	Asia 3	CSTR:16533.06.IVCAS 6.7391	GU477492, GU477493, GU477494

^aThe genome sequence has been previously reported.

Fig. 1.

Phylogenetic analysis and genetic characterization of CCHFV strains in China. Maximum-likelihood phylogenetic trees were constructed based on the complete coding sequences of the L segment (A), M segment (B), and S segment (C). The strains are presented as GenBank accession number_strain name_isolated country_isolated year. All Chinese CCHFV strains are represented in blue font, and the CCHFV strains in this study are highlighted with a red solid circle. D Amino acid mutation frequencies across RdRp, GP, and NP from Chinese CCHFV strains.

To further explore the phylogenetic relationships among Chinese CCHFV strains isolated from 1966 to 2016 in Xinjiang Province, nucleotide (nt) and amino acid (aa) identities were calculated for all previously reported Chinese strains available in GenBank and strain IbAr10200. The aa identities among the Chinese strains for RNA-dependent RNA polymerase (RdRp), glycoprotein (GP), and NP were 94.9%–100%, 72.6%–99.9%, and 94.8%–100%, respectively, whereas those between IbAr10200 and Chinese strains were 95.1%–96.7%, 73.6%–91.9%, and 94.8%–97.7%, respectively (Supplementary Fig. 1). The nt identities among Chinese strains for the L, M, and S segments were 87.0%–100%, 69.7%–99.9%, and 87.6%–99.9%, respectively, whereas those between IbAr10200 and Chinese strains were 86.1%–88.0%, 70.3%–89.2%, and 85.6%–89.0%, respectively (Supplementary Fig. 1). In the L segment, CCHFV isolates from the 1960s–1970s exhibited over 98% nucleotide identity, except for strain A7001. However, the homology between isolates from the 20th century onward and earlier isolates was relatively low, with an nt identity of 87.0%–93.3%. CCHFV strains 66019 and UCCR4459, along with BA68038 and C68031, exhibited 100% aa identity in their RdRp (Supplementary Fig. S1A). The M segment displayed more complex mutations and variations, such as isolates from 1970s, showing over 98% nt and aa identity, whereas strains including BA68045, BA84002, BA84021, BA04032, YL04041, YL04057, YL16070, and YL16204 showed less than 80% nt or aa identity with other strains (Supplementary Fig. S1B). In the S segment, strains isolated before the 21st century exhibited lower identity than those isolated later. Strains from the 1960s to the 1980s had over 97% nt and aa identity, except for strains 79121M18 and A7001, which showed less than 95% nt identity with the strains from the 20th century. In addition, multiple strains displayed 100% aa identity in NP (Supplementary Fig. S1C).

As shown by the analysis of aa mutations in the RdRp, GP, and NP strains of CCHFV in China, there are several stable distributions of aa site mutations in RdRp and NP (Kaushal and Baranwal, 2023), and the overall mutation rate was relatively low, which may be a reason for maintaining a relatively high sequence identity. However, GP from different strains exhibit high-frequency mutations at multiple aa sites, resulting in low sequence identity of the M segment. The region that may cause significant differences in sequence homology in GP is the mucin-like domain (MLD). Compared to GP38, Gn, and Gc (Mcfadden et al., 2024), the MLD in GP displayed the highest mutation rates at single aa sites (Fig. 1D). We found multiple aa mutations in RdRp and GP in the Asia 4 strain BA04032 compared with those in Asia 1–3. These mutations included E126D, K218R, A596T, I1472V, K1709R, D2130S, G2142S, and I2186V in RdRp (Supplementary Fig. S2A) and V323A, S433T, D449N, I533V, S762N, I865M, V869L, K946R, G950S, R964K, E1232D, and M1500I in GP (Supplementary Fig. S2B). In addition, strains 88168 and 75024, derived from human sources, exhibited several amino acid differences in NP compared to strain 79121M18 from jerboas and BA04032 from ticks, including E44D, N125D, A146T, D262N, N275S, and M353L (Supplementary Fig. S2C). These mutations may contribute to the observed variations in infection capacity and virulence across strains from different host sources or genotypes. The mucin domain at the N-terminus of the GP protein is associated with CCHFV pathogenicity. Compared to GP38, Gn, and Gc, the mucin domains (aa positions 1–252 of GP) of the CCHFV strains representing the four genotypes exhibited different N-glycosylation modifications, whereas no differences were observed in N-glycosylation modifications in other regions (Supplementary Fig. S3). This variation in N-glycosylation may contribute to the different pathogenicities of various genotypes (Mcfadden et al., 2024). The CCHFV strains 75024 and 79121M18 exhibited more intense N-glycosylation modification sites in the mucin domain, which may enhance pathogenicity.

In vitro replication of representative CCHFV strains differs across cell lines

Chinese strains of CCHFV exhibit four Asian lineages based on M-segment phylogeny: Asia 1–4. We investigated the infection characteristics of four strains, including strain 88168 (source: human) from Asia 1, strain 75024 (source: human) from Asia 2, strain 79121M18 (source: jerboa) from Asia 3, and strain BA04032 (source: ticks) from Asia 4, each representing a different M-segment lineage and derived from different origins. IFA results showed that CCHFV infection levels in Vero, BHK-21, DH82, SW-13, SH-SY5Y, and HEK-293 cells were generally higher than those in MDOK, MDBK, MDCK, U87MG, Huvec, and HepG2 cells (Fig. 2A), indicating that permissiveness to CCHFV varies across cell lines.

Fig. 2.

In vitro infectivity and replication kinetics of four representative CCHFV genotypes in various cell lines. A Infectivity of CCHFV strains (88168-Asia 1, 7502 -Asia 2, 79121M18-Asia 3, BA04032-Asia 4) in 13 cell lines at 3 dpi, as determined by IFA detecting the viral nucleoprotein (NP); scale bars, 100 μm. B Viral RNA load (copies/mL) in the culture supernatant was determined by qRT-PCR at 3 dpi following infection of the indicated cell lines with four CCHFV strains (MOI = 0.1). Data are presented as mean ± SD from the biological replicates. Statistical significance was determined using one-way ANOVA (∗P < 0.05, ∗∗P < 0.01).

Consistent with the high infection rates observed by IFA, quantification of viral RNA revealed high permissiveness in Vero, BHK-21, SW-13, and HEK-293 cells, which yielded viral RNA approaching 10⁸ copies/mL, whereas DH82, MDCK, and MDBK cells were less susceptible, with viral RNA below 10⁴ copies/mL (Fig. 2B). In Vero cells, the viral loads of 88168, 75024, and 79121M18 strains reached ∼10⁸–10⁹ copies/mL, whereas that of BA04032 was slightly lower (1.7 × 10⁷ copies/mL), although the difference between 75024 and 79121M18 in Vero cells was not significant. In BHK-21 cells, 88168, 75024, and 79121M18 also achieved viral loads of ∼10⁸ copies/mL, with no significant difference among them; however, BA04032 had a lower load (1.1 × 10⁷ copies/mL). Replication was notably restricted in DH82 and MDCK cells; all four strains only reached ∼10⁴ copies/mL, with no significant differences among strains, indicating that these cell types are poorly permissive. In MDBK cells, strain 75024 showed a significantly higher viral RNA load of 6.7 × 10³ copies/mL, whereas other strains were barely detectable (∼10 copies/mL). This suggests a strain-specific adaptation, with the 75024 strain exhibiting a unique ability to replicate in the bovine-derived MDBK cell lines. MDOK cells supported moderate replication of all four strains (10⁶ copies/mL), with no significant differences among them. In SW-13 cells, 88168 and 75024 reached viral loads of ∼10⁹ copies/mL, which was significantly higher than that of 79121M18 and BA04032 (∼10⁸ copies/mL). In U87MG cells, viral replication was reduced across all strains (∼10⁶–10⁷ copies/mL). Huvec cells supported similar replication of 88168, 75024, and 79121M18 (∼10⁸ copies/mL), with BA04032 having a significantly lower load (5.1 × 10⁷ copies/mL). In SH-SY5Y cells, 88168 (∼10⁷ copies/mL) replicated better than 75024 and 79121M18 (∼10⁶ copies/mL), while BA04032 had a significantly lower load (2.1 × 10⁵ copies/mL). HEK-293 and HepG2 cells showed relatively uniform replication across all four strains (∼ 10⁸ copies/mL), with no significant differences, suggesting that these human-derived cell lines are broadly permissive. This highlights that Vero, BHK-21, SW-13, Huvec, HEK-293, and HepG2 cells are permissive to multiple Asian viral strains, although subtle differences in tropism for BA04032 may be present. The differential replication profiles across cell lines underscore the strain- and cell type-specific interactions that govern the viral tropism. Strains 88168 and 75024 exhibited a broad tropism. In contrast, BA04032 exhibited a more restricted tropism, consistently showing lower replication across most cell types, which may reflect differences in viral entry or replication machinery.

Infection dynamics, immune responses, and tissue pathology of CCHFV strains in C57BL/6 mice

To evaluate the pathogenicity of the CCHFV strains in vivo, C57BL/6 mice were inoculated intraperitoneally with the four CCHFV strains. Although no mortality was occurred, yet all infected mice exhibited detectable viral loads and obvious organ lesions. As expected, IgM and IgG responses were observed in these mice. IgM responses were rapid and transient; 75024 and 79121M18 showed peaked IgM responses on day 4, whereas 88168 and BA04032 showed peaked IgM responses on day 6, after which the levels declined (Fig. 3A). Strain 75024 induced the strongest IgM response, followed by strains 88168 and 79121M18 at moderate levels, and BA04032 at the weakest. All viral strains elicited robust IgG responses, with OD values steadily increasing from days 4–12 (Fig. 3B).

Fig. 3.

In vivo pathogenicity and immune responses of CCHFV strains in C57BL/6 mice. A, B Kinetics of anti-CCHFV IgM (A) and IgG (B) antibody responses in mouse serum following intraperitoneal infection with 1 × 10⁵ TCID₅₀ of the indicated CCHFV strains. C–F Viral RNA loads (log10 copies/mL) in homogenates of six organs (brain, heart, lung, liver, spleen, and kidney) from mice infected with strains 88168 (C), 75024 (D), 79121M18 (E), and BA04032 (F) at 3 and 12 dpi. Data are presented as the mean ± SD (n = 3 mice per time point per group). G Representative hematoxylin-eosin-stained sections of mouse organs (liver, spleen, and kidney) at 3 and 12 dpi. Scale bars, 100 μm.

Following infection with each of the four CCHFV strains, viral RNA was broadly detected in six organs (brain, heart, lung, liver, spleen, and kidney) at 3 dpi but declined significantly by 12 dpi. Notably, delayed clearance was observed in the hearts of 79121M18-infected mice and lungs of BA04032-infected mice at levels similar to those observed at 3 dpi (Fig. 3C–F). Each CCHFV strain exhibited differential organ tropism, with significantly higher viral RNA levels detected in the liver, spleen, and kidney, and lower levels in the lungs at 3 dpi. This demonstrates elevated CCHFV tropism within murine liver, spleen, and kidney tissues, but not lung tissues. In addition, the RNA load in the brain infected with strain 79121M18 was lower than that of the other strains, indicating weaker mouse brain tropism (Fig. 3E).

Subsequently, we found that infection with the four strains resulted in severe lesions in the liver, spleen, and kidneys of mice at 3 dpi (Fig. 3G), whereas no obvious pathology was detected in the brain, heart, or lungs, as assessed by hematoxylin-eosin (H&E) staining (data not shown). Infection with CCHFV strains 75024, 79121M18, and 88168 induced mild hepatic pathology, characterized by hepatocellular edema, pale and loosely stained cytoplasm, disorganized architecture, slight hemorrhage, and limited hepatic steatosis. In the spleen, infection with human-derived strains 75024 and 88168 for 3 days pi led to white pulp reduction and lymphocyte necrosis in both white and red pulps, as evidenced by nuclear condensation, fragmentation, or lysis, and mild granulocyte infiltration in the red pulp. All strains caused only minor renal injury, manifesting as edema and swelling of occasional tubular epithelial cells, pale cytoplasm, and vacuolar degeneration with small cytoplasmic vacuoles. At 12 dpi, the lesions caused by viral infection in these mice were cleared, and the mice recovered (Fig. 3G).

The IFNAR^−/− mice model reveals differences in lethality among different genotypes of CCHFV

To better characterize the pathogenicity of the four CCHFV strains, IFNAR^−/− mice were challenged with serially diluted doses and monitored for body weight changes and survival rates, and the LD₅₀ value of each strain was determined. Infection with the CCHFV strain 88168 at doses of 10–10000 TCID₅₀ resulted in 33.3%–83.3% mortality and a 2.6%–25.2% reduction in body weight compared to the mock group at 8 dpi. In contrast, all mice survived the infection with this strain at a dose of 1 TCID₅₀ and exhibited no loss in body weight (Fig. 4A and B), resulting in an LD₅₀ value of 142.5 TCID_50. Infection with the strain 75024 resulted in severe body weight loss, exceeding 20% across all groups (Fig. 4C), with 100% mortality in all groups challenged with doses ranging from 1 to 1000 TCID₅₀ (Fig. 4D). All mice in these groups succumbed between 4 and 8 dpi, resulting in an LD₅₀ value of less than one TCID₅₀ for strain 75024. Infection with strain 79121M18 caused body weight loss in all groups (Fig. 4E), resulting in 66.7%–100% mortality (Fig. 4F). Of these mice, those infected with doses of 10–10000 TCID₅₀ died within a narrow window of 3–5 dpi. The observed lethality profile indicated an LD₅₀ of less than 1 TCID₅₀ for this strain. Infection with strain BA04032 induced dose-dependent weight loss, with the 100 and 1000 TCID₅₀ groups showing the most severe reduction (approximately 30%; Fig. 4G). Mortality was observed across all challenge doses from 1 to 1000 TCID₅₀, commencing between 3- and 4-days post-infection (dpi) and continuing until 8 dpi (Fig. 4H). However, the mortality pattern was not dose-dependent, which precluded the determination of a definitive LD₅₀. This contrasts with the abovementioned CCHFV strains representing Asia 1–3 genotypes, which cause dose-dependent lethality in IFNAR^−/− mice.

Fig. 4.

Lethality and virulence assessment of CCHFV genotypes in IFNAR^−/− mice. Mice infected with the Asia 1 strain 88168 exhibited dose-dependent weight loss (A), with an LD₅₀ of 142.5 TCID₅₀ (B). In stark contrast, the Asia 2 strain 75024 induced severe, rapid weight loss exceeding 20% across all challenged groups (C) and resulted in 100% mortality, even at the lowest dose of 1 TCID₅₀, indicating an LD₅₀ < 1 TCID₅₀ (D). Similarly, the Asia 3 strain 79121M18 caused significant weight reduction (E) and high mortality (66.7%–100%) with an LD₅₀ below 1 TCID₅₀ (F). The Asia 4 strain BA04032 demonstrated a unique pattern, inducing substantial weight loss that was most severe at intermediate doses (G), whereas mortality occurred across all challenge doses without a clear dose-response relationship (H). Mice were euthanized after losing 25% of their initial body weight. Data are presented as mean ± SD.

Discussion

CCHFV poses a significant threat to public health because of its ability to cause severe hemorrhagic symptoms. It has a broad geographical distribution and is maintained through a transmission cycle involving ticks and various animal hosts. CCHFV infections are prevalent in Africa, Europe, and Asia. Phylogenetic analyses have revealed a distinct geographical distribution of the virus across continents. Approximately nine genotypes have been documented, including Asia 1–3, Europe 1–3, and Africa 1–3, with at least three identified from each region (Guo et al., 2017; Zhang et al., 2018). In this study, we identified three CCHFV genotypes (Asia 1, Asia 2, and Asia 3) based on the L and S gene segments. Phylogenetic analysis of the M segment revealed a novel genotype, Asia 4, defined by five Chinese CCHFV strains (BA68045, BA84002, BA84021, BA04032, and YL04041), which formed a clade distinct from all previously recognized Asian genotypes. Although this genotype is closely related to the Asia 3 genotype, it branches separately and can be further divided into two subclades. Of the five strains belonging to genotype Asia 4 based on their M segments, the L and S segments of strains BA68045, BA84002, and BA84021 were classified under the Asian 2 genotype, and the other two strains, BA04032 and YL04041, belonged to the Asia 3 genotype (Fig. 1A–C, Table 1). The genotypic differences observed among segments of the same virus could be attributed to reassortment and migration events between CCHFV strains of different genotypes (Zhou et al., 2013). Inconsistency of genotypes in segments was also observed in the Chinese CCHFV strains 88168, BA84027, 84007, and BA84051, which contained L and M segments belonging to the Asia 1 genotype and S segments belonging to Asia 2. Conversely, strains 66019 and BA8004 had L and S segments corresponding to Asia 2, but their M segments were aligned with the Asia 1 genotype (Table 1).

Since 2003, no confirmed human CCHF cases have been reported in the Xinjiang region of China, and the virus has only been detected in ticks (Moming et al., 2018; Zhang et al., 2018). Given the current situation, the possibility of a CCHFV outbreak in China remains (Teng et al., 2022). However, there is a notable lack of research on the infection characteristics and virulence of the diverse CCHFV genotypes found in China. In this study, four strains, each representing a different CCHFV M-segment lineage and of different origins—88168 (human-derived, M-segment Asia 1 with an Asia 2 S segment), 75024 (human-derived, Asia 2), 79121M18 (jerboa-derived, Asia 3), and BA04032 (tick-derived, M-segment Asia 4 with Asia 3 L and S segments) were characterized in vitro and in vivo. Because 88168 and BA04032 are reassortants, the observed phenotypic variations likely reflect a combination of M-segment lineage specific traits and reassortment effects. The human-derived strains 88168 (Asia 1) and 75024 (Asia 2), as well as the Jerboa-derived strain 79121M18, were replicated in various cell lines, including Vero (monkey kidney), BHK-21 (hamster kidney), MDCK (canine kidney), SW-13 (human-derived adrenocortical carcinoma), U87MG (glioma), Huvec (endothelial), SH-SY5Y (neuroblastoma), HEK-293 (kidney), and HepG2 (liver) cell lines (Fig. 2A). The results of IFA and viral RNA copy number determination indicated that the tick-derived strain BA04032 had significantly (P < 0.05) weaker infection and replication capabilities in Vero, BHK-21, SW-13, Huvec, and SH-SY5Y cell lines than other strains (Fig. 2B). Therefore, the tick-derived strain exhibited a lower capacity to infect mammalian cells than the human- or animal-derived CCHFV strains, indicating a lower ability of the strain from ticks to infect and replicate in mammalian hosts. Our observation that Vero, BHK-21, SW-13, and HEK-293 cells support robust CCHFV replication is consistent with previous reports demonstrating high permissiveness to CCHFV infection (Dai et al., 2021). The restricted replication of BA04032 in mammalian cells, particularly its significantly lower titers in Vero, BHK-21, and SW-13 cells compared to Asia 2 and Asia 3 strains, represents a novel observation that may reflect the tick-adapted nature of this Asia 4 reassortant strain.

Infection of adult immunocompetent mice with CCHFV has demonstrated limited viral replication and little to no signs of disease (Hawman et al., 2021). Investigating the pathogenicity of CCHFV in mice has necessitated the use of immunocompromised models, such as those with genetic deficiencies in type I interferon (IFN) or STAT-1 signaling (Bente et al., 2010; Zivcec et al., 2013). The virulence of CCHFV varies among different genotypes in immunocompromised mice. Bente et al. conducted virulence assays on the CCHFV strain IbAr 10200 of the Africa 2 genotype and found that its LD₅₀ in STAT-1 knockout mice was four plaque-forming units (PFU) (Bente et al., 2010). Bereczky et al. demonstrated that infection of adult IFNAR^−/− mice with CCHFV caused rapid onset of symptoms and death in animals infected with as few as 10 focus-forming units (FFU) of the virus (Bereczky et al., 2010), whereas another study reported an LD₅₀ of 0.05 TCID₅₀ (Zivcec et al., 2013). The LD₅₀ of the Europe 1 strain Turkey Kelkit06 against IFNAR^−/− mice is 2.5 PFU (Canakoglu et al., 2015), and that of the strain Ank2 is 10 TCID₅₀ (Aligholipour et al., 2019). In this study, we determined that the LD₅₀ of strain 88168 from Asia 1 was 142.5 TCID₅₀, whereas strains 75024 from Asia 2 and 79121M18 from Asia 3 had LD₅₀ values of less than 1 TCID₅₀. The lethality of the Asia 4 genotype strain BA04032 in mice was non-dose-dependent, warranting further investigation into the underlying mechanisms. The differences in lethal doses among CCHFV strains of genotypes Africa 2, Europe 1, and Asia 1–4 also indicate variations in their virulence. However, because strains 88168 and BA04032 analyzed here are reassortants (Table 1), these differences may reflect the combined influence of M-segment lineage–specific traits and reassortment event.

One limitation of this study was that all 24 CCHFV isolates were recovered from lyophilized brain homogenates that underwent multiple serial passages in suckling Kunming mice. However, the exact passage history, intervals between passages, and associated laboratory adaptation events have not been documented. Consequently, we cannot quantify how repeated intracerebral inoculation and brain-to-brain transfer may have altered the replication kinetics, cell tropism, or virulence relative to the original clinical or tick-derived isolate. This precludes a definitive assessment of whether the differences observed in vitro and in vivo are inherent to each lineage.

Conclusions

In this study, we resuscitated 24 historical Chinese CCHFV strains (1966–2004) using suckling mouse inoculation and obtained their whole-genome sequences. Phylogenetic analyses of CCHFV gene segments revealed high genetic heterogeneity, with seven genotypes for L (Asia 1–3, Africa 2–3, Europe 1–2), nine for M (including novel Asia 4), and nine for S (Asia 1–3, Africa 1–3, Europe 1–3). In vitro, the Asian 2 and 3 strains replicated robustly in multiple cell lines, whereas the Asian 4 tick-derived strain (BA04032) showed weaker infectivity, reflecting the genotype and host origin effects. In vivo, all genotypes induced specific antibodies and mild-to-moderate liver, spleen, and kidney damage in C57BL/6 mice. In IFNAR^−/− mice, virulence was genotype-associated: Asia 2 and 3 were highly lethal (LD₅₀ < 1 TCID₅₀), Asia 1 was moderately virulent (LD₅₀ = 142.5 TCID₅₀), and Asia 4 had atypical non-dose-dependent mortality. These findings confirm the presence of a novel Asia 4 genotype and establish a correlation between the CCHFV genotype and virulence in China. These findings fill the gaps in historical isolate characterization, provide critical data for surveillance and risk assessment, and lay the foundation for targeted vaccines and therapeutics.

Materials and methods

Viruses

The 24 historical CCHFV strains used in this study were originally collected between 1966 and 2004 from Xinjiang Uygur Autonomous Region, China. Host origins included CCHF patients, ticks, and one jerboa (Table 1). Virus isolation was performed at the time of collection by intracerebral inoculation of 2-day-old suckling Kunming mice with clarified tissue homogenates (human serum, tick pools, or rodent organ samples). Brains of moribund or euthanized suckling mice were harvested, homogenized, lyophilized, and stored at the National Virus Resource Center (NVRC) of the Wuhan Institute of Virology, Chinese Academy of Sciences, until use in the present study.

Cell lines, experimental mice, and antibodies

The following cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA), namely African green monkey kidney cells (Vero, ATCC®CCL-81™), baby hamster kidney cells (BHK-21, ATCC®CCL-10™), canine macrophage cells (DH82, ATCC®CRL-10389™), human adrenocortical carcinoma cells (SW-13, ATCC®CCL-105™), human embryonic kidney cells (HEK-293, ATCC®CRL1573™), human glioma cells (U87MG, ATCC®HTB-14™), human liver cells (HepG2, ATCC®HB-8065™), human neuroblastoma cells (SH-SY5Y, ATCC®CRL-2266™), human umbilical vein endothelial cells (HUVEC, ATCC®CRL4053™), and the Madin-Darby canine kidney cells (MDCK, ATCC®CCL-34™); these cell lines were grown in Dulbecco's modified Eagle's medium (DMEM; Sigma, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (Gibco, Grand Island, NY, USA), whereas bovine kidney cells (MDBK, ATCC®CCL-22™) and Madin-Darby ovine kidney cells (MDOK, ATCC®CRL-1633™) were grown in Minimum Essential Medium (MEM; Sigma) supplemented with 10% fetal bovine serum (FBS). All cell lines were incubated at 37 °C in a 5% CO₂ atmosphere. The 2-day-old suckling KM, C57BL/6, and IFNAR^−/− C57BL/6 mice (lacking the type I IFN receptor) were obtained from the Animal Center of Hubei Provincial Centers for Disease Control and Prevention and raised in a specific pathogen-free environment. A monoclonal anti-CCHFV nucleoprotein (NP) antibody, mAb 43E5 (obtained from the Xinjiang Centers for Disease Control and Prevention), was used for indirect immunofluorescence assay (IFA) as described previously (Guo et al., 2017).

Inoculation of suckling mice

Freeze-dried powders of the CCHFV strains were dissolved in phosphate-buffered saline (PBS; pH 7.4) and centrifuged (4000×g, 5 min) to remove tissue debris. Suckling Kunming mice were inoculated via intracranial and subcutaneous routes with 0.03 mL of supernatant containing 100 U/mL penicillin and 100 μg/mL streptomycin, as described previously (Guo et al., 2017; Tang et al., 2023). In total, 24 groups of suckling mice were inoculated and monitored daily until day 12. Diseased mice were euthanized, and their brains were harvested for further analysis.

RNA extraction and metagenomic sequencing

Mouse brains were homogenized in 1 mL of PBS using a tissue cell destroyer (D1000, Novastar). The homogenates were centrifuged at 4000×g for 5 min at 4 °C to remove tissue debris. Total RNA was extracted from 0.1 mL of the clarified supernatant, as previously described (Guo et al., 2017). A total of 24 RNA samples were subjected to metagenomic sequencing to identify viral genomic sequences (Zhang et al., 2018). RNA libraries were constructed following the TruSeq Stranded Total RNA Sample Preparation Guide (Illumina). Ribosomal RNA (rRNA) was depleted using the VAHTS™ Total RNA-seq (H/M/R) Library Prep Kit for Illumina (Vazyme Biotech, China) according to the manufacturer's instructions. All libraries were subsequently sequenced on the MiSeq platform (Illumina). The resulting raw reads were quality-controlled using FastQC and de novo assembled into contigs with Trinity using default parameters. Assembled contig sequences were queried against the viral nucleotide database using BLASTn to identify CCHFV-related sequences. Putative CCHFV contigs were further screened against the whole-genome database using BLASTn to eliminate false-positive hits. Contigs related to CCHFV were extracted and assembled using SeqMan to obtain complete genomes of the CCHFV strains. The raw RNA sequencing data were submitted to the China National GeneBank (CNGB) under accession numbers CNR0809854–CNR0809873 and CNR1060510–CNR1060512.

Phylogenetic analyses

The genomic sequences of the L, M, and S segments of the 24 CCHFV strains obtained through metagenomic sequencing were deposited in GenBank under accession numbers OR047127–OR047186 and OR823870–OR823878 (Table 1). Phylogenetic analyses were performed on the complete coding sequences of the 79 L, 96 M, and 96 S segment sequences collected from diverse CCHFV isolates from humans, ticks, and jerboa, respectively. Sequence alignments were performed using ClustalW, and maximum-likelihood phylogenetic trees were constructed using MEGA 10.0 (Kumar et al., 2018) software, using the maximum composite likelihood model. Bootstrap analysis was conducted with 1000 replicates for tree validation purposes.

Virus end-point dilution assay

Viral titers in mouse brain homogenate supernatants were determined using an end-point dilution assay, as previously described. Vero cells were cultured in 96-well plates and infected with tenfold serial dilutions of the supernatant. Five days post-infection (dpi), the cells were subjected to an IFA to measure viral titers (TCID₅₀/mL) using mAb 43E5. Fluorescence microscopy (ECLIPSE TE2000-S; Nikon) was used to visualize the results, and viral titer endpoints were calculated using the Reed–Muench method.

Growth curve analyses

The viral growth properties of four CCHFV strains (88168, 75024, 79121M18, and BA04032), representing genotypes Asia 1–4, were assessed in various cell lines using growth curve analysis. Cells were seeded in 12-well plates and infected with CCHFV at a multiplicity of infection (MOI) of 0.1 in a medium containing 2% FBS. After 1 h of viral adsorption at 37 °C, the cells were supplemented with fresh medium. At 3 dpi, the cells were fixed and analyzed using IFA to detect CCHFV NP expression. In addition, 100 μL of supernatant from each well was collected to assess the viral load via real-time reverse transcription polymerase chain reaction (qRT-PCR). Total RNA was extracted from the culture supernatants of all 13 cell lines infected with CCHFV. qRT-PCR was performed using a One Step PrimeScript™ RT-PCR Kit (Takara), with primers and RNA standards prepared according to a previous study (Dai et al., 2021).

Animal experiments

The infection characteristics and median lethal doses (LD₅₀) of different CCHFV strains were examined in C57BL/6 and IFNAR^−/− mice. Six-to-eight-week-old C57BL/6 mice (n = 6 per group) were intraperitoneally injected with 150 μL of CCHFV strains (88168, 75024, 79121M18, and BA04032), containing 1 × 10⁵ TCID₅₀. Blood samples were collected every two days, and CCHFV–specific IgG and IgM antibodies were measured using a CCHFV IgG and IgM enzyme-linked immunosorbent assay (ELISA) detection kit (Hamburg, Germany), following the manufacturer's instructions. At 3 and 12 dpi, the animals were dissected, and organ samples from the brain, heart, lungs, liver, spleen, and kidneys of three mice per group were fixed in 4% paraformaldehyde for 24 h. Hematoxylin-eosin (H&E) staining was performed to observe pathological changes in the tissues. For the LD₅₀ experiment, female IFNAR^−/− mice (n = 5 per group for BA04032, n = 6 per group for the other three strains) were infected with CCHFV strains diluted in tenfold increments from 10,000 TCID₅₀ to 1 TCID₅₀. Body weight changes and survival were monitored daily for 10 days, and the LD₅₀ was calculated using logistic regression.

Statistical analysis

Statistical analyses were performed using GraphPad Prism version 8.3.0 (GraphPad Software, San Diego, CA, USA). The significance for all comparisons was accepted at P < 0.05 (∗P < 0.05, ∗∗P < 0.01).

Data availability

The metagenomic sequencing data are available in China National GeneBank (CNGB) under the accession numbers CNR0809854−CNR0809873 and CNR1060510−CNR1060512. The genomic sequences of the L, M, and S segments of 24 CCHFV strains are deposited in GenBank under accession numbers OR047127−OR047186 and OR823870−OR823878.

Ethical statement

Animal experiments were approved by the Ethics Committee of Wuhan Institute of Virology, Chinese Academy of Sciences (approval number WIVA332024017). All cell and animal experiments related to CCHFV infection were performed in biosafety level 3 laboratories (BSL-3) and animal biosafety level 3 laboratories (ABSL-3).

Author contributions

Abulimiti Moming: Investigation, methodology, software, data curation, visualization, writing-original draft, writing-review and editing. Yuan Bai: Investigation, data curation, and formal analysis. Qiong Zhu, Jiayin Jin, Yaohui Fang and Shouwei Huang: investigation and methodology. Hongfeng Chen, Haixia Yang, Qiaoli Wu, and Shuang Tang: Resources. Manli Wang: Conceptualization, and resources. Zhihong Hu: Conceptualization and Project administration. Yujiang Zhang: Resources, investigation, data curation, and supervision. Fei Deng: conceptualization, writing-review and editing, project administration, resources, supervision, and funding acquisition. Shu Shen: conceptualization, project administration, resources, writing-review and editing, and supervision.

Conflict of interest

The authors declare no conflict of interest. Prof. Fei Deng is an editorial board member for Virologica Sinica and was not involved in the editorial review or the decision to publish this article.

Appendix A. Supplementary data

The following are the Supplementary data to this article.

Supplementary Figure S1.

Homology analysis was performed to compare the nucleic acid and aa sequences of the CCHFV strains, specifically the L (A), M (B), and S (C) segments. This analysis included 23 strains from our study, as well as 9 L segments, 15 M segments, and 13 S segments from previously reported Chinese CCHFV strains.

Supplementary Figure S2A-1.

Supplementary Figure S2A-2.

Supplementary Figure S2B.

Supplementary Figure S2C.

Multiple protein sequence alignments of RdRp (A), GP (B), and NP (C) from CCHFV strains 88,168, 75,024, 79121M19, and BA04032. The protein lengths of RdRp, GP, and NP were 3945, 1689, and 482 aa, respectively.

Supplementary Figure S3.

Prediction results of N-glycosylation sites of glycoprotein GP from 88,168, 75,024, 79121M19, and BA04032. Four genotypes of CCHFV were predicted (https://services.healthtech.dtu.dk/service.php?NetNGlyc-1.0) for N-glycation sites in GP, and it was found that the mucin domain was the region with significant differences in N-glycation in different genotypes in Asia.