Here, we compared the biological properties of five CPXV strains that are representative of four different genetic clusters () 
. Based on viral kinetic analysis, a comparable growth in cells for all CPXVs was demonstrated. However, virulence in mice underlined marked differences between the CPXVs. Depending on the disease outcome, CPXVs could be allocated in three groups characterized by no death, 20% or a 100% mortality rate. Cohorts infected i.n. with CPXV-BR and VACV-WR depicted 100% lethality, with high levels of replicating virus in the lungs and spreading to various organs, which is consistent with published observations 
. CPXV-BR was originally isolated in 1937 from a milker and maintained by serial passage in rabbit skin 
, and VACV-WR is historically derived from NYCBH by several passages in mouse brain 
. While we can hypothesize that the virulence of VACV-WR might be explained by its adaptation in mouse brain, the pathogenicity of CPXV-BR in mice might rely on its genetic background (i.e. virulence factors, or immunomodulatory proteins).
Strikingly, two CPXV isolates derived from an elephant (CPXV-GER1980-EP4) and a cat (CPXV-AUS1999-867) induced a 20% mortality rate in contrast to two strains isolated from humans that did not cause mortality. A slight (CPXV-FIN2000-MAN) or no weight loss (CPXV-GER1991-3) was observed. In line with this, Huemer and colleagues demonstrated that a human-derived CPXV was impaired in pathogenicity following intravenous or i.n. infection of BALB/c mice, at virus doses ranging from 103
PFU per animal 
The variability in virulence among the CPXVs reported here correlates with the observations of Baxby 
. Eighteen CPXVs, including CPXV-BR, were evaluated by infecting mice intracerebrally (105
PFU per animal) and were classified in four groups based on mortality rates (0%, 15–25%, 55–65% and 90–100% mortality, CPXV-BR being part of the latter group). The authors noted that the variability was unlikely to be explained by a different passage history of the viruses 
, although this was not assessed here. Additionally, it has to be mentioned that our in vivo
observations relied on mice inoculated with a virus dose of 104
PFU per animal, and this dose was chosen because it is 100% lethal for mice i.n. exposed to either CPXV-BR or VACV-WR reference strains (). It cannot be excluded that CPXV isolates might be as pathogenic as the reference strains if animals were inoculated with a 10- or 100-fold increased virus dose, but only further investigations will give an answer to that question.
Our results suggested that the virus ability to replicate in the lungs may account for disease outcome, as observed with CPXV-BR, CPXV-GER1980-EP4 and CPXV-AUS1999-867, that had virus titers ≥4 logs PFU/g tissue. Also, elevated concentrations of the pro-inflammatory cytokine IL-6 seemed to correlate with CPXV disease outcome. This result is in line with increased secretion of IL-6 reported in the sera of cowpox-challenged macaques 
. The role of IL-6 induction in cowpox pathogenesis remains still unclear but a recent article evidenced that IL-6, IL-8 and CXCL1, produced upon in vitro
CPXV-BR infection, could be responsible for chemotaxis of monocytes in vitro
. Additionally, macrophages and monocytes are suspected to be involved in CPXV spread through the host 
. A potential role of IL-6 in CPXV pathogenesis might also be supported by the fact that, during inflammation, IL-6 has been shown to suppress neutrophils recruitment at sites of acute inflammation, making ways for the influx of monocytes as the inflammatory response is sustained 
. No evidence of systemic TNF-α induction was seen which might not be surprising as OPVs, and CPXV in particular, utilize many ways to counteract TNF-α production through the expression of NFκB inhibitors or of cytokine response modifiers (CrmB, C, D and E) (for review see 
). Undetectable levels of TNF-α were also reported in vitro
, following monkeypox virus infection 
, as well as in sera of macaques lethally infected with variola virus 
. In our hands, both lung virus titers and IL-6 sera levels emphasized differences between CPXV strains. Nevertheless, the absence of spreading of CPXV-AUS1999-867 to kidneys, MLNs and ovaries, and the avirulence of CPXV-GER1991-3 in NMRI mice still require additional studies that might point towards nonfunctional gene(s) needed for efficient virus spreading (CPXV-AUS1999-867), or for immunomodulation or host range (CPXV-GER1991-3).
The discrepancies observed here raised the question whether CPXV-BR should be considered a representative CPXV, as this strain has an unknown passage history in various laboratories since its isolation 80 years ago. It could be that a reduced virulence or an avirulent phenotype in NMRI mice may be a typical feature of CPXV clinical isolates, explaining their ability to survive in rodent populations. The virus has to rely on its host and should therefore not be lethal, but CPXV infections may still have a direct impact on the dynamics of their reservoir hosts. In line with that, infection of bank voles and wood mice with a CPXV strain isolated from a cat have been shown to impair fecundity although any overt disease and mortality were recorded 
. Also, CPXV infection of field voles reduced significantly their survival rates 
Antivirals in clinic could be valuable to manage cowpox infections in both humans and animals. We therefore examined the antiviral activities of promising anti-OPV compounds against CPXVs. Our in vitro
evaluation evidenced that compounds targeting viral DNA replication, i.e. cidofovir, (S
)-HPMP-5-azaC and CMX001, inhibited similarly CPXV clinical isolates and the CPXV, VACV and CMLV reference strains. Comparable findings were obtained with the viral egress ST-246, although the EC50
value against CPXV-BR appeared 5- to 10-fold higher than that of CPXV isolates, VACV and CML1. These results are in agreement with our previous studies demonstrating that the way of propagation of CPXV-BR in HEL cells had an impact on ST-246 antiviral activity 
. ST-246 inhibits the wrapping of virions and, as a consequence, inhibits the production of enveloped viruses. However, we have shown that CPXV-BR intracellular mature viruses can still be released in the cell supernatant following ST-246 treatment, and this viral form could most probably account for CPXV-BR EC50
. Our in vitro
findings together with numerous in vivo
antiviral studies performed with laboratory strains 
strongly support that these compounds will likely be active in vivo
against genetically diverse CPXV isolates.
In this study, we demonstrate the importance of both genomic and biological testing in characterizing emerging isolates. The extrapolation of these results for the clinicians and veterinarians should be made in function of the immune status of the patient and the virulence of the strain which may give variable clinical pictures. Antiviral treatments (i.e. cidofovir, (S)-HPMP-5-azaC, CMX001 and ST-246), specific for managing OPV-related infections, exist and, depending on the clinical case, they could be administered locally or systemically.