Wild Brazilian yellow fever virus infection in Syrian hamsters (Mesocricetus auratus): Clinical and histopathological analyses Authors: Fernanda de Oliveira Bottino, Barbara Cristina Euzebio Pereira Dias de Oliveira, João Paulo Rodrigues dos Santos, Mariana Barata Viana Tiradentes, Yuli Rodrigues Maia de Souza, Tainah Silva Galdino de Paula, Hyago da Silva Medeiros Elido, Isabele Barbieri dos Santos, Ieda Pereira Ribeiro, Myrna Cristina Bonaldo, Marcelo Pelajo Machado, Pedro Paulo de Abreu Manso Published: 2024 ABSTRACT Highlights•Wild Brazilian YFV strains infect and cause histopathologic changes in the hamster.•The histopathological findings were compatible with YF pathology observed in humans.•The PR4408 strain showed increased virulence, despite causing a non-fatal infection.•Evaluating new YFV isolates help to understand the viscerotropic phenotype of these.Keywords: Yellow fever virus, Non-hamster-adapted strains, PR4408 strain, Hamster model Introduction Yellow fever (YF) is an infectious disease caused by the Orthoflavivirus flavi virus, also known as yellow fever virus (YFV), the prototype member of the Orthoflavivirus genus and Flaviviridae family (ICTV, 2022). All YFV strains belong to a single serotype, categorized into seven genotypes (five of African origin and two of South American origin, designated I and II) (Monath, 2001). The South American I genotype encompasses five lineages (1A to 1E) associated with outbreaks and epidemics in South America (de Souza et al., 2010; Vasconcelos et al., 2004). The YFV responsible for the spread of the disease in Brazil since 2008, however, belongs to sub-lineage 1E (de Souza et al., 2010; Nunes et al., 2012). Similarly to other arboviruses, the natural YFV infection begins with mosquito bites. Transmission takes place in three distinct cycles: the sylvatic cycle, where the YFV is transmitted in forest areas between non-human primates (NHPs) and mosquitoes belonging to the Haemagogus and Sabethes genera in South America (Hanley et al., 2013), and Aedes africanus in Africa (Germain et al., 1982); the urban cycle, in which the virus is transmitted urbanely between humans by the peridomestic mosquito Ae. aegypti (Monath, 2001); and the intermediate cycle, described only in the African savanna, near forests, where transmission occurs between NHPs, humans, and mosquitoes with semi-domestic habits, such as Ae. furcifer, Ae. bromeliae and Ae. luteocephalis (Germain et al., 1982). While most human YF cases are self-limited, it is important to note that clinical symptoms can range from mild, resembling a nonspecific influenza-like illness, to severe. Alarmingly, 5 to 10 % of all cases can progress to a fatal outcome (Monath, 2001). The severe forms of YF primarily affect the liver, leading to jaundice, as well as the kidney and myocardium, potentially causing hepatorenal syndrome, hemorrhage, and shock (Barrett and Teuwen, 2009; Pastorino et al., 2010). Despite being a vaccine-preventable disease, YF is still endemic in Africa, Latin America and Caribbean (LAC). Until 2018, 40 countries (27 in Africa and 13 in the Americas) were at the highest risk for YF, including Brazil. Moreover, the increasing risks of urban outbreaks and international spread and number of mosquitos, as well as a changing epidemiology, places YF as an emerging global threat (WHO, 2018). Significant epizootic YF waves were reported from 2008 to mid-2017. The 2016–2019 Brazilian outbreak, for example, reached the entire southeastern region. This spread towards the Brazilian south, an area with no vaccination recommendations (Brasil, 2018), and, in the last 70 years, is considered to be the most significant New World YF activity, with a total of 2251 human cases, 772 deaths, and 2590 confirmed NHP infections (Sacchetto et al., 2020; Silva et al., 2020). In February 2017, a set of nine specific amino acid markers was detected in circulating YFV (Gómez et al., 2018). Evidence indicates that YFV strains circulating in Brazil carry amino acid variations that can modulate its transmission and virulence in mammalian cells and immunodeficient mice (Furtado et al., 2022b). However, further studies are required to assess these polymorphism impacts on YF pathophysiology. In this sense, YFV infection assessments in a non-immunocompromised animal model are essential to better understand this. Non-human primates, such as rhesus monkeys, present a clinical infection pattern almost identical to humans (Stokes et al., 1928). However, high costs and the complex regulatory procedures make it difficult to carry out studies in this model (Tesh et al., 2001). Considering the need for an experimental YF animal model, the Syrian hamster (Mesocricetus auratus) was established as a YF model by Tesh et al. (2001) and McArthur et al. (2003), although demanding YFV-adapted strains. Therefore, this study aimed to evaluate Syrian hamster susceptibility to infection by non-hamster-adapted Brazilian YFV strains (PR4408 and RJ155) and an 17DD experimental vaccine strain. The main aim was to assess whether or not the increasing virulence previously observed in mammalian cells and mice could make these viruses capable of infecting the hamster model without the need for any adaptation. Material and methods Animals Twenty-four male Syrian hamsters (Mesocricetus auratus), 4 to 6 weeks old, were employed in this study. All animals were maintained in ventilated racks (Alesco, Brazil) equipped with a controlled ventilation system (10 to 20 air changes per hour) in the Center for Animal Experimentation (CEA) facilities at the Oswaldo Cruz Institute (IOC), Oswaldo Cruz Foundation (Fiocruz). The hamsters were housed in polypropylene cages (Alesco, Brazil) measuring 490 mm × 340 mm × 160 mm under a 12-hour photoperiod at 21 ± 2 °C, and a relative humidity of 40 % to 60 %. The animals had daily access to paper towels and hay as environmental enrichment material and ad libitum access to standard chow (Nuvilab, Brazil) and autoclaved water. Ethics statement All experimental protocols adopted in this study followed the Guide for the Care and Use of Laboratory Animals (Animals, 2011) and were carried out according to Brazilian law, licensed under number (L-023/2021) by the Oswaldo Cruz Institute Animal Use Ethics Committee (CEUA IOC-Fiocruz). Viruses Three YFV strains were employed herein: (i) YFV PR4408, isolated in Vero cells from the serum of a dying howler monkey (Alouatta sp.) from Caibaté, in the state of Rio Grande do Sul, Brazil, in 2008. This strain has undergone six previous passages in a mosquito cell (C6/36) culture (Furtado et al., 2022b); (ii) YFV RJ155, isolated in Vero cells from the serum and whole blood of a dead howler monkey (Alouatta guariba clamitans), from Casimiro de Abreu, in the state of Rio de Janeiro, Brazil, in 2019. This strain has undergone four previous passages in a mosquito cell (C6/36) culture (Abreu et al., 2019; Furtado et al., 2022b); and (iii) the YFV 17DD experimental vaccine (17DDexp), developed in chicken embryo skeletal muscle cells by our research group (Souza et al., 2019). All strains were sequenced by the Sanger method to confirm the absence of mutations following cell culture passages. Both the PR4408 and RJ155 strains belong to sub-lineage 1E isolates from Brazil. The inocula titers were standardized between 1.6 and 2.1 × 105 PFU/mL and the virus samples were diluted immediately prior to the assays and kept on ice. Experimental infection and sample collection Hamsters were divided into three groups of eight animals each, in which six animals were intraperitoneally (IP) inoculated with 0.1 mL of a different viral strain (PR4408, RJ155, or 17DDexp) and two were inoculated with 0.1 mL of sterile medium 199 (mock-infected animals). Each group was then categorized into two subgroups: (a) blood sample collection at 1- and 5-days post-infection (dpi); and (b) blood sample collection at 3 and 7 dpi. Following infection (day 0), animal blood was collected 1, 3, 5 and 7 dpi, alternating subgroups, and the liver, spleen, pancreas, and kidneys were collected at 5 and 7 dpi. Remaining animals were euthanized on day 7 post-infection. At 1 and 3 dpi, blood was collected through the gingival plexus, following intraperitoneal sedation with a 0.1 mL of a ketamine 10 % and xylazine 2 % combination (7:3 ratio). At 5 and 7 dpi, blood collection was performed during euthanasia, via cardiac puncture, following the administration of 3 mL of a ketamine 10 %, xylazine 2 % and morphine 0,04 % association (20:8:2 ratio). Animal death was confirmed by verifying cardiorespiratory arrest. The collected blood was placed in sterile DNAse/RNAse free tubes and centrifuged for 5 min at 9000 rpm (5415R centrifuge, Eppendorf), to obtain serum, which was stored in a freezer at -80οC. At 5 and 7 dpi, the liver, spleen, pancreas, and kidneys were collected and submitted to conventional histological processing. Small liver fragments were also placed in Trizol (Invitrogen, USA) and stored in a freezer at -80 °C. Clinical evaluation Animals were weighed daily and their body temperature was measured by an infrared thermometer (DM300), at a distance of about 0.5 cm from the xiphoid process. This thermometer ranges from 0° to 100 °C. In addition, animals were also clinically evaluated daily to determine the humanitarian endpoint, if needed. The following clinical parameters were analyzed: weight loss (less than 5 %, 5 to 10 %, 11 to 15 %, 16 to 20 % and above 20 %, adding up to, respectively, 0, 3, 6, 9 or 12 points), body temperature (normal temperature, hypothermia bellow 35.2 °C and hyperthermia above 38.9 °C), posture (normal body posture, slight locking, moderate locking and severe locking, adding up to, respectively, 0, 6, 9 or 12 points), breathing (normal breathing, tachypnea, and abdominal breathing, scoring, respectively, 0, 6 or 12 points), eye opening (above 75 %, between 50 and 75 %, between 25 and 50 % and below 25 %, adding, respectively, 0, 6, 9 or 12 points), mobility (active, with reduced mobility, ataxia/unsteady gait, reluctance or inability to move, scoring, respectively, 0, 3, 6, 9 or 12 points), and coat appearance (normal hair coat, discreet piloerection or noticeable piloerection, adding, 0, 6 or 12 points). Euthanasia was performed on animals that reached a total score equal to or greater than 36 points or that scored 12 points in any item alone at the end of the day. Weight, viral load and biochemical analysis graphs were constructed using the GraphPad Prism 8.0 software. Concerning weight loss monitoring, values ​​were normalized (100 %) on day 0. Biochemical analyses Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were determined at 5 and 7 dpi employing a Vitros 250 apparatus (Ortho Clinical Diagnostics Johnson & Johnson, USA) according to the dry chemistry methodology. RNA extraction Viral RNA was extracted from 140 μL of serum samples using the commercial QIAamp® Viral RNA Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions. RNA was also extracted from 100 mg of liver samples from infected and mock-infected animals by the Trizol (Invitrogen, USA) method. After mechanical dissociation and incubation with 0.2 mL of chloroform, 0.5 mL of isopropanol and 4 μL of polyacrylamide were added to the upper aqueous phases. Following centrifugation, the pellets were rehydrated with 1 mL of 75 % alcohol and centrifuged for 10 min at 10.000 xg at 4 °C. Alcohol was poured, and the RNA was resuspended in 40 μL of RNAse free water. Electrophoresis runs on 2 % agarose gels (Labtrade, Brazil) in Tris/Borate/EDTA (1X TBE), and bright green plus (NeoBio, Brazil) were performed to assess extracted RNA integrity. RT-qPCR For Real-time PCR preceded by reverse transcription (RT-qPCR), 5 μl of viral RNA extracted from serum samples were subjected to a reverse transcription step and amplified using the TaqMan Fast Virus One-Step Master Mix kit (Applied Biosystems, USA) in the Applied Biosystems StepOnePlus equipment. For each reaction, 300 nM forward primers (5′-GCTAATGAGGTGCATTGGTCTGC-3′, genome position 15–38), 600 nM reverse primer (5′-CTGCTAATCGCTCAACGAACG-3′, genome position 83–103) and 250 nM of the probe (5′ FAM-ATCGAGTTGCTAGGCAATAAACAC-3′ TAMRA, genome position 41–64) were used. Reverse transcription was performed at 50 °C for 5 min, followed by qPCR under the following conditions: 95 °C for 20 s, forty 95 °C amplification cycles for 15 s and 60 °C for 1 min. The genome copy numbers/mL were calculated by absolute quantification based on Cycle ​​(Ct) threshold values in relation to a previously characterized standard curve, obtained by a tenfold serial dilution (ranging from 10 to 106) (Abreu et al., 2019). All samples were run in triplicate and with the constitutive gene 18S from mammals as an internal reaction control. RT-PCR and electrophoresis For the conventional PCR preceded by reverse transcription (RT-PCR), 5 μL of RNA extracted from liver samples was Reversed Transcripted (Thermoscript RT-PCR kit - cat. 11,146,016, Life Technologies, USA). The amplification was then carried with ThermoScript RT-PCR System (Thermoscript – cat. 1,383,211, Life Technologies, USA), Platinum Taq DNA Polymerase (Thermoscript – cat. 10.966–018, Life Technologies, USA) and a pair of universal Flavivirus Yellow Fever primers (YF-1 – 5` GGTCTCCTCTAACCTCTAG 3` and YF-3 – 5` GAGTGGATGACCACGGAAGACATGC 3`) (Tanaka, 1993). The amplicons (675 base pairs) were submitted to electrophoresis on 1.5 % agarose gels (Labtrade, Brazil) in Tris/Borate/EDTA (1X TBE) and stained with bright green plus (NeoBio, Brazil). A 100 bp molecular mass standard was used as reference (Mid-Range DNA Ladder – Cellco Biotec, Brazil) and the amplified products were visualized by a photodocumenter. Histological analyses Liver, spleen, pancreas, and kidneys fragments were fixed in Carson's formalin-Millonig for 48 h at room temperature (Carson et al., 1973) and processed according to standard histological techniques for paraffin embedding. Sections (5 μm thick) were stained with hematoxylin-eosin (Mayer, 1903) and analyzed under an AxioHOME brightfield microscope (Carl Zeiss, Germany). Images were acquired with an Axiocam 712 color camera (Carl Zeiss, Germany). YFV antigen detection Paraffin-shaped liver and spleen sections (5 μm) from mock-infected and animals infected with the YFV PR4408 strain at 7 dpi were examined by indirect immunofluorescence using a mouse monoclonal antibody that recognizes the YFV antigen (ascitic fluid anti-ptn E YFV) (Instituto Evandro Chagas; Belém – PA) at a 1:100 dilution. A goat anti-mouse polyclonal antibody conjugated with Alexa Fluor 488 (IgG, 2mg/ml) [cat: A11001] (Thermo Fisher Scientific, USA) was used as a secondary antibody. Antigenic retrieval was carried out in a 0.01 M citrate buffer (Imuno Retriever 20x; Bio SB) pH 6.0 in TintoRetriever chamber (Bio SB), according to the manufacturer's instructions. Counter-staining with DAPI 1:5000 and a background staining with Evans Blue 1:10,000 were subsequently performed. Negative controls were performed by duplicating each sample and omitting primary antibody treatment, so that any reactions resulting from the secondary antibodies or reagents employed in the analyses could be adequately traced. Slides were mounted with ProLong Glass Antifade Mountant [cat: P36980] (Thermo Fisher Scientific, USA) and analyzed under a LSM 980 Confocal Laser Scanning Microscope (Carl Zeiss, Germany). Results Clinical observations Animals infected with RJ155 and 17DDexp strains presented an average weight gain of 3.1 % and 3.5 %, respectively, at the end of the experiment (day 7) compared to the first day of infection (day 0). The average weight loss of PR4408-infected animals up to euthanasia was 9.82 %. Mock-infected animals also presented weight loss (1.2 %) at day 7 (Fig. 1A). From day 2 post-infection, one PR4408 infected animal presented a loss of over 5 % of its body weight. For the other infected animals, this 5 % loss began at 5 dpi (n = 2) and 6 dpi (n = 3) and persisted until the end of the experiment (7 dpi) (Fig. 1B). No other alterations concerning the clinical parameters evaluated for the humanitarian endpoint were observed. In addition, none of the animals evolved to death or reached the humanitarian endpoint. Infected hamster weight loss. The above statistics depict the (A) average weight of infected and mock-infected animals and (B) individual weight of PR4408-infected animals. At 5 dpi, PR4408 infected animal serum samples were icteric (n = 2/3) (Fig. 2A), with the liver exhibiting a whitish appearance (n = 2/3). At 7 dpi, all PR4408 infected animals (n = 3) presented icteric sera samples (Fig. 2B), splenomegaly (characterized by an enlarged spleen of about 10 mm in length compared to uninfected animals) (Fig. 2C), and a whitish liver (Fig. 2D). These changes were not observed in the other infected and mock-infected animals (Fig. 2E-H). Necropsy findings. PR4408-infected animals exhibiting (A) icteric sera samples at 5 dpi, (B) icteric sera samples at 7 dpi, (C) splenomegaly (30 mm in length), and (D) livers with whitish appearance at 7 dpi. Mock-infected animals exhibiting (E) non-icteric sera at 5 dpi, (F) non-icteric sera at 7 dpi, (G) regular spleen (20 mm in length), and (H) regular liver at 7 dpi. Serum ALT (Fig. 3A) and AST (Fig. 3B) levels were significantly elevated only in hamsters infected with the PR4408 virus, suggesting more severe liver damage. Hamsters infected with the 17DDexp and RJ155 viruses showed little or no increase in these liver enzymes compared to the control group. Biochemical analysis. Alanine aminotransferase (ALT) and Aspartate aminotransferase (AST) levels in hamsters infected with the YFV 17DDexp, RJ155 and PR4408 strains and mock-infected animals at 5 and 7 dpi. Viremia Serum samples were obtained at 1-, 3-, 5- and 7-days post-infection. YFV viremia was detected at 3, 5, and 7 dpi in PR4408-infected animals, except for one animal that did not display detectable viral RNA at 5 dpi. This animal was also the only one that recovered from the weight loss during the experiment. Regarding RJ155, the YFV in infected animals was detected at 3 and 5 dpi. Excluding animals with detectable viremia, the mean maximum blood virus titers occurred on day 3 after infection in RJ155 infected animals (4,87 log10 copies/mL), and on day 5 after infection in PR4408 infected animals (6045 log10 copies/mL). Viremia remained undetectable in 17DDexp infected and mock-infected animals (Fig. 4). Dynamic of YFV RNA measured by RT-qPCR. YFV RNA from PR4408-, RJ155z, and 17DDexp-infected animals were quantified by Real-time PCR preceded by reverse transcription. The figure shows the viral load in Log10 copies/mL. Hepatic YFV detection An electrophoresis assay was performed to confirm the integrity of the extracted RNA (data not shown). YFV RNA was detected in the liver of two PR4408 infected animals (n = 2/3) at 5 dpi and in three other infected animals (n = 3/3) at 7 dpi. The animal that recovered from weight loss and presented no detectable viremia also exhibited no detectable YFV RNA in the liver. All RJ155-infected animals (n = 3/3) at 5 dpi were positive for YFV in liver, while no YFV RNA was detected at 7 dpi. Moreover, all 17DDexp were negative for viral RNA detection (Fig. 5). YFV RNA detection in infected hamster liver by RT-PCR. YF amplicons were assayed via agarose gel electrophoresis. The figure shows samples from PR4408-, RJ155-, and 17DDexp-infected animals. L: Ladder (100 bp). Co: Negative controls (mock-infected animals). Black arrows indicate the 675 bp amplicon obtained from YFV strain amplification. Hepatic histopathology Peri-portal inflammatory infiltrates were observed on day 5 after infection in YFV PR4408-infected animals, with a predominance of neutrophils and monocytes (n = 3/3) (Fig. 6A); and Councilman bodies (n = 2/3) (Fig. 6B). No pathological changes were observed in mock-infected animals (Fig. 6C). On day 7, all animals displayed an inflammatory process associated with Councilman bodies (Fig. 6D), necroapoptosis (Fig. 6E) and steatosis (Fig. 6F) (n = 3/3, but more significant in two animals). In addition, peri‑portal inflammatory infiltrates, karyorrhexis, sinusoidal congestion, foam cells, and hepatocytic regeneration (mitosis figures) were observed in one animal at 7 dpi. Hepatocytes and Kupffer cells were positive for the viral antigen according to an immunofluorescence analysis at 7 dpi (Fig. 7A). No pathological changes or YFV antigen-labeled cells (Fig. 7C) were noted in mock-infected animals in the same point of infection. Brightfield microscopy analysis of the hepatic histopathology of Syrian hamsters infected with YFV PR4408 strain. Infected animals at 5 dpi exhibiting (A) peri‑portal inflammatory infiltrate and (B) acidophilic bodies. Mock-infected animals at 5 dpi exhibiting (C) absence of histopathological changes. Infected animals at 7 dpi presenting (D) acidophilic bodies, (E) necroapoptosis, and (F) steatosis and acidophilic bodies. White arrows indicate neutrophils and monocytes and black arrows indicate acidophilic bodies. Hematoxilin and Eosin stain. YFV antigen detection in the liver and spleen of Syrian hamsters infected with YFV PR4408 strain. Infected animal at 7 dpi showing the viral antigen in (A) hepatocytes (yellow arrows) and Kupffer cells (red arrows), and (B) splenic cells. Mock-infected animal at 7 dpi with no viral antigen in (C) hepatocytes and Kupffer cells, and (D) splenic cells. YFV antigen detection in green, nuclei stained with DAPI in blue and erythroid cells and background in red. In general, no specific pathological abnormalities were observed in YFV RJ155-infected animals. However, rare Councilman bodies (n = 2/3) (Fig. 8A) and peri‑portal inflammatory infiltrates (n = 3/3) (Fig. 8B) were observed at 5 dpi, while no histopathological changes were observed in mock-infected animals (Fig. 8C). Conversely, rare peri‑portal inflammatory infiltrates (n = 3/3) (Fig. 8D) and Councilman bodies (n = 1/3) (Fig. 8E) were described at 7 dpi. No histopathological changes were detected in mock-infected animals (Fig. 8F). Brightfield microscopy analysis of the hepatic histopathology of Syrian hamsters infected with YFV RJ155 strain. Infected animals at 5 dpi exhibiting (A) acidophilic bodies; (B) peri‑portal inflammatory infiltrate. Mock-infected animals at 5 dpi exhibiting (C) no histopathological changes with microsteatosis areas. Infected animals at 7 dpi exhibiting (D) peri‑portal inflammatory infiltrate, (E) acidophilic bodies. Mock-infected animal at 7 dpi showing (F) absence of histopathological changes. Black arrows indicate acidophilic bodies. Hematoxilin and Eosin stain. A single 17DDexp infected animal exhibited the presence of Councilman bodies (n = 1/3) at 5 dpi, while another animal presented a mild congestion of focal sinusoids (n = 1/3) and occasional apoptotic bodies (n = 1/3) at 7 dpi. Apparently, a slight to mild degree of hepatic microsteatosis and rare inflammatory infiltrates may be normal for Syrian hamsters, as they were observed in some mock-infected animals. Splenic histopathology On day 5 after infection, animals infected with the YFV PR4408 strain presented histopathological spleen changes, consisting of cells with pyknotic nucleus and cellular apoptosis, especially in the white pulp (n = 2/3) (Fig. 9A), as well as white pulp disorganization, with intense cellular apoptosis (n = 1/3) (Fig. 9B), stromal cell death, red pulp hyperplasia and marginal zone loss (n = 1/3) (Fig. 9C). One infected animal presented no splenic pathology (n = 1/3). On day 7, all infected animals presented white pulp apoptosis, with white pulp edge loss (n = 3/3) (Fig. 9D), and an exacerbated macrophage response predominantly in the marginal zone of the lymphoid follicle (n = 3/3). In addition, leukophagocytosis (n = 2/3) (Fig. 9E) and erythrophagocytosis (n = 2/3) in the marginal zone of the spleen were also observed at 7 dpi. Splenic cells in the white pulp and marginal zone were positive for viral antigen detection by immunofluorescence analysis at 7 dpi (Fig. 7B). No histopathological changes were observed in mock-infected animals at 5 (Fig. 9F) or 7 dpi. Furthermore, no YFV antigen-labeled cells were observed in mock-infected animals at 7 dpi (Fig. 7D). Brightfield microscopy analysis of the splenic histopathology of Syrian hamsters infected with YFV PR4408 strain. Infected animals at 5 dpi exhibiting (A) pyknotic nucleus and cellular apoptosis with the presence of macrophages (black arrow) in the white pulp, (B) white pulp disorganization with intense cellular apoptosis, (C) intense cellular apoptosis in the white pulp and marginal zone loss. Infected animals at 7 dpi exhibiting (D) cellular apoptosis in the white pulp, with loss of the white pulp boundary, (E) leukophagocytosis (*) and activated macrophages (black arrow). Mock-infected animal at 5 dpi showing (F) no histopathological changes in the white pulp and red. Black circles indicate cellular apoptosis. Hematoxilin and Eosin stain. Five days after infection, all YFV RJ155-infected animals presented lymphoid hyperplasia, with loss of lymphoid follicle and marginal zone structure (n = 3/3) containing cells with less condensed chromatin, probably indicating the presence of young lymphocytes (Fig. 10A). Furthermore, rare cells in apoptosis (n = 3/3) were observed (Fig. 10B). At 7 dpi, a more heterogeneous population of lymphocytes was observed in two animal lymphoid follicles, containing both young and mature lymphocytes (Fig. 10C), and a heterogeneous marginal zone exhibiting signiicant cellularity compared to 5 dpi, without the presence of activated macrophages and rare points of apoptosis (n = 2/3) (Fig. 10D). Some white pulp disorganization was also observed. Only one animal exhibited young lymphocyte follicle hyperplasia and white pulp disorganization, with a slight cell rarefaction in the marginal zone, as well mitosis and activated macrophages (n = 1/3) (Fig. 10E). No histopathological alterations were observed in the spleens of mock-infected animals (Fig. 10F) and YFV 17DDexp-infected animals at 5 and 7 dpi. Brightfield microscopy analysis of the splenic histopathology of Syrian hamsters infected with YFV RJ155 strain. Infected animals at 5 dpi exhibiting (A) white pulp hyperplasia, with loss of lymphoid follicle and marginal zone structure, cells with less condensed chromatin, (B) rare cellular apoptosis. Infected animals at 7 dpi showing: (C) heterogeneous lymphocyte population, containing both young and mature lymphocytes, (D) rare points of apoptosis, and white pulp disorganization, (E) young lymphocyte follicle hyperplasia. Mock-infected animals at 5 dpi with (F) no histopathological changes. Black circles indicate cellular apoptosis. Hematoxilin and Eosin stain. Pancreatic histopathology No histopathological changes were observed in the pancreas of YFV RJ155- and YFV 17DDexp -infected animals, at 5 and 7 dpi. Some cells undergoing apoptosis were observed on one YFV-PR4408 infected animal at 5 dpi (n = 1/3) (Fig. 11A). At 7 dpi, cellular apoptosis was noted in both exocrine (n = 3/3) (Fig. 11B) and endocrine (n = 1/3) (Fig. 11C) components. No histopathological changes were detected in mock-infected animals at 5 dpi (Fig. 11D) and 7 dpi. Brightfield microscopy analysis of the pancreatic histopathology of Syrian hamsters infected with YFV PR4408 strain. Infected animal at 5 dpi exhibiting (A) cellular apoptosis in the exocrine portion. Infected animals at 7 dpi exhibiting cellular apoptosis in the (B) exocrine and (C) endocrine portions of the pancreas. Mock-infected animals at 5 dpi with (D) no histopathological changes. Black circles indicate cellular apoptosis. Hematoxilin and Eosin stain. Renal histopathology No histopathological changes were observed in the kidneys of all analyzed animals. Discussion The present study assessed Syrian hamster (Mesocricetus auratus) infection by two non-hamster-adapted Brazilian YFV isolates originated from dying or dead howler monkeys obtained during outbreaks in the states of Rio Grande do Sul in 2008 (PR4408) and Rio de Janeiro in 2019 (RJ155) (Abreu et al., 2019; Furtado et al., 2022b), comparing the results with a YFV experimental vaccine strain (17DDexp). Both non-hamster-adapted strains caused non-fatal infections in hamsters, with animals exhibiting weight loss and hepatic, splenic, and pancreatic histopathological changes compatible with YF pathology. Furthermore, infection by the YFV PR4408 strain led to persistent viremia from day 3 post-infection up to day 7. No viscerotropic model for YF in immunocompetent mice is currently available. In this animal, wild-type YFV strains exhibit little or no ability to evade IFN-I antiviral action (Meier et al., 2009). Thus, immunocompromised mice, such as A129 and STAT129, and chimeric models, have been employed to mimic severe YF human cases, as well as to study YFV pathogenesis and test vaccine and antiviral drugs (Furtado et al., 2022a; Meier et al., 2009; Strom et al., 2010). Syrian hamsters are already being used for vaccine and therapeutic candidate assessments employing hamster-adapted-YFV-strains, due to their intact immune systems. Previous studies (McArthur et al., 2020; Tesh et al., 2001) have demonstrated that these strains cause illness and mortality in immunocompetent animals. Wild-type strains can be adapted after serial passages, as described by Tesh et al. (2001) and McArthur et al. (2003) to obtain an immunocompetent hamster model that leads to severe illness and death. Animals infected with the Jimenez p.10 and Asibi p.7 strains manifested lethargy, loss of appetite, ruffled fur, and hunched postures, but only the Asibi p.7 strain caused death in 100 % of the infected hamsters (McArthur et al., 2003; Tesh et al., 2001). However, little information is available regarding experimental infection by YFV wild-type strains in this model. The two wild-type strains evaluated herein can be grouped according to the presence (RJ155) or absence (PR4408) of a 2016–2019 outbreak molecular signature consisting of nine non-synonymous substitutions (Bonaldo et al., 2017; Furtado et al., 2022b). This signature is associated with YFV 2016–2019 outbreak strains and is present in most sub-lineage 1E sequences described since 2016, in contrast with YFV 2000–2010 outbreak strains, which do not present these substitutions (Furtado et al., 2022b). In a recent study, Furtado et al. (2022) described that, despite the YFV 2016–2019 molecular signature, the PR4408 and RJ155 strains presented higher viral titers, exhibited larger and more diverse plaque sizes, and induced the highest percentages of cytotoxicity in Vero cells compared to other Brazilian isolates, such as GO05 (YFV 2000–2010) and ES-504 (YFV 2016–2019). In addition, the YFV PR4408 and RJ155 strains led to a lower average survival time in a neurovirulence test conducted on a mouse model and displayed higher viral titers, despite cell treatment with IFN-α and IFN-β. These results suggest that these two Brazilian isolates carry genomic markers that increase infectivity in mammal hosts (Furtado et al., 2022b), corroborated by the data reported herein. The capacity of the RJ155 and PR4408 strains to infect hamsters without any adaptation supports the findings reported by Furtado et al. (2022), which suggest that strains associated with the 2016–2019 outbreak are more virulent in mammals. In Tesh et al. (2001), who carried out a preliminary study with the YFV wild-type BeH 413,820, Jimenez, Asibi, and A709–4A2 strains, animals developed a transient but significant viremia that lasted 3 to 5 dpi. In addition, in a previous study, animals infected with the wild-type Asibi strain developed a modest viremia, which was not detected in 40 % of the animals (McArthur et al., 2003). In the present study, RJ155-infected animals showed detectable viral RNA in sera at 3 and 5 dpi, while viremia was detected in PR4408-infected animals from day 3 post-infection to at least day 7. Performing an extended experimental infection will aid in understanding whether the PR4408 strain induces a more prolonged viremia in infected animals or if the infection resolves spontaneously. When inoculated with adapted Jimenez p.10 and Asibi p.7 strains, hamsters developed detectable viremia between 24 h and 5- or 6 days post-infection. Both exhibited a viremia peak at 3 dpi, similar to the wild-type strains evaluated in this study (McArthur et al., 2003; Tesh et al., 2001). Viremia remained undetectable in 17DDexp-infected animals,. In contrast, the YFV 17D vaccine strain was isolated from the blood of hamsters immunosuppressed with cyclophosphamide (Mateo et al., 2007). Furthermore, RT-PCR detected YFV RNA in the liver of PR4408- and RJ155-infected animals, respectively, at 5 dpi, and 5- and 7-dpi. In previous studies, YFV antigens were detected in the brain and liver of animals inoculated by the intracerebral route with 17DD and GOI 4191 strains; and in the liver and spleen by the IP route with adapted Jimenez p.10 strain (Rodrigues et al., 2004; Tesh et al., 2001). All animals infected with the YFV PR4408 wild-type strain presented weight loss from day 2 post-infection, in addition to hepatic and splenic changes, such as color change and enlargement, respectively, as well as an icteric serum observed following necropsy. In contrast, animals infected with the YFV wild-type RJ155 strain did not present any clinical signs of the disease during the 7 days post IP infection, similar to the subclinical infection described by McArthur et al. (2003) in animals infected with the wild-type Asibi strain. Additionally, none of the hamsters infected with YFV 17DDexp strains in the present study showed signs of illness during 7 days post IP infection. Since it the immunocompetent Syrian hamster model requires YFV-adapted strains, and because 17DD is an attenuated vaccine strain, it is expected that infected animals will not manifest the disease, as observed herein. In a previous study, Syrian hamsters infected with the GOI4191 strain, isolated from a fatal adverse event following YFV vaccination, were compared with the 17DD vaccine strain by intracerebral, intrahepatic, and subcutaneous routes (Rodrigues et al., 2004). No clinical data, however, were provided. In humans, rhesus monkeys, and hamsters infected with the YFV, hepatocytes can be identified by their characteristic morphology, presenting acidophilic corpuscles known as Councilman and Rocha-Lima corpuscles, as well as microsteatosis, associated with a moderate periportal inflammatory infiltrate (Engelmann et al., 2014; Monath, 2001; Quaresma et al., 2013; Xiao et al., 2001). Herein, the most common finding in animals infected with the PR4408 and RJ155 strains was the presence of periportal inflammatory infiltrates and Councilman bodies at 5- and 7 dpi. The infection was eliminated in animals infected with the RJ155 strain at 7 dpi. No viremia or detectable viral RNA was observed in liver, and a decrease in histological changes was observed. However, animals infected with the PR4408 strain remained infected, also exhibiting liver steatosis and necrosis day 7 post-infection. The altered sera AST and ALT levels observed in animals infected with the PR4408 and RJ155 strains indicate hepatocyte death and corroborate with the alterations observed in the histopathological assessments. As with the histopathological alterations, the AST and ALT increases were more pronounced in animals infected with the PR4408 strain. In contrast, McArthur et al. (2003) described that hamsters infected with the YFV Asibi strain (Asibi p.0) showed no significant changes at 5 or 6 dpi when compared with mock-infected animals. However, histological liver, spleen, and pancreas changes were observed in animals infected with adapted strains. At 5 and 6 dpi, changes observed by Xiao et al. (2001) and Sbrana (2006) were similar to those observed in the present study, encompassing hepatic portal tract inflammatory infiltrates, microvesicular steatosis, and Councilman bodies (Sbrana et al., 2006; Xiao et al., 2001). Also, hamsters infected with the Asibi p.7 strain exhibited microvesicular steatosis, moderate to severe inflammation, as well as mild to moderate hepatic necrosis, at 5 and 6 dpi (McArthur et al., 2003). The presence of viral protein in hepatocytes and Kupffer cells in the liver and in white spleen pulp cells indicates that cell death was caused by viral infection. Hamster spleen pathology of PR4408 and RJ155 strain infections at 5 and 7 dpi were consistent with that noted by Xiao et al. (2001); McArthur (2003), and Sbrana (2006) following infection with adapted strains, comprising lymphoid hyperplasia, white pulp depletion, and an increased number of macrophages. Xiao et al. (2001) described spotty parenchymal necrosis (numerous single necrotic cells in the acini) in the pancreas, between 5 and 7 dpi. Herein, we observed cellular apoptosis in both exocrine and endocrine portions at 7 dpi only in animals infected with the PR4408 strain. Thus, pathophysiological aspects of hamster infection by this Brazilian strain require further investigation and will be the subject of future studies by our research group. The data reported here indicate that Syrian hamsters reproduce a non-fatal infection when experimentally infected with non-hamster-adapted Brazilian YFV strains but develop histopathological and clinical alterations compatible with yellow fever pathology. The observed histopathological changes were similar to those detected in human pathology and in a hamster model using adapted strains. Furthermore, the YFV PR4408 strain was found to be more virulent than the RJ155 strain, causing a persistent infection for at least up to day 7 post-infection. Funding statement This work was supported by INOVA Fiocruz Program (Fiocruz). PPAM receives the fellowship for Young Scientist of Our State Program – 2023 (FAPERJ – 2007.04381.9). MCB receives the CNPq fellowship for Productivity in Technological Development and Innovative Extension (308,164/2020–2) and Scientist of Our State Program – 2020 (FAPERJ). MPM receives the CNPq fellowship for Productivity in Research (317,510/2021–5), Scientist of Our State Program – 2023 (FAPERJ - E-26/200.528/2023). CRediT authorship contribution statement Fernanda de Oliveira Bottino: Writing – review & editing, Writing – original draft, Methodology, Investigation, Data curation, Conceptualization. Barbara Cristina Euzebio Pereira Dias de Oliveira: Writing – review & editing, Methodology, Investigation, Data curation, Conceptualization. João Paulo Rodrigues dos Santos: Methodology, Investigation. Mariana Barata Viana Tiradentes: Methodology, Investigation. Yuli Rodrigues Maia de Souza: Writing – review & editing. Tainah Silva Galdino de Paula: Methodology, Investigation. Hyago da Silva Medeiros Elido: Methodology, Investigation. Isabele Barbieri dos Santos: Methodology, Investigation. Ieda Pereira Ribeiro: Writing – review & editing, Methodology, Investigation, Data curation. Myrna Cristina Bonaldo: Writing – review & editing, Funding acquisition, Data curation, Conceptualization. Marcelo Pelajo Machado: Writing – review & editing, Funding acquisition, Data curation, Conceptualization. Pedro Paulo de Abreu Manso: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The authors thank Rita de Cássia dos Passos Ferraz da Silva, João Gabriel Regis Sobral (Institute of Science and Technology of Biomodels, Oswaldo Cruz Foundation, Rio de Janeiro, Brazil), Daniele Guimarães, Luzia Fátima Gonçalves Caputo, Luciana Silva Souza, Rafaela Almeida da Silva and Nathália Furtado (Laboratory of Experimental Medicine and Health, Oswaldo Cruz Institute, Oswaldo Cruz Foundation, Rio de Janeiro, Brazil) for their support with the analyses. Footnotes Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.virusres.2024.199505. Supplementary materials Data availability Data will be made available on request. References Associated Data Supplementary Materials Data Availability Statement Data will be made available on request.