1. Introduction
Yellow fever virus (YFV) belongs to the Flavivirus genus and forms a
virion of approximately 50 nm that is composed by a capsid, a
premembrane/membrane (prM/M) protein and an envelope (E) protein, a
single-stranded positive-sense RNA and a lipid envelope (Lindenbach et
al., 2007). YFV is an arthropod-borne human pathogen transmitted by
mosquitoes that causes a “historically devastating disease” (Paules &
Fauci, 2017), for which there is no specific treatment available yet
(CDC, 2019).
Before the current live-attenuated egg-derived vaccine became available
in the 1930s, yellow fever (YF) was a major health threat, since
outbreaks used to kill approximately 10% of the population of affected
cities, such as occurred in Philadelphia (USA) in 1793 (Frierson, 2010;
Monath & Vasconcelos, 2015; Paules & Fauci, 2017). Despite the
availability of a highly protective vaccine, nowadays 29,000-60,000
people die annually from YF worldwide (Garske et al., 2014). Moreover,
recent outbreaks in Africa (2016) and Brazil (2017-2018) have shown an
inadequate preparedness for YF. In Africa in 2016, depletion of the WHO
YF vaccine stockpile along with worldwide vaccine shortage led WHO
authorities to start dose fractionation (1/5) as an emergency measure
for outbreak control. Despite being the largest world producer of YF
vaccine, Brazil also suffered from vaccine shortage in the latest
outbreak and started using the 1/5 fractional dose in January 2018 for
mass vaccination of the population. Since the African outbreak in 2016,
when there were several imported cases to China, representing the first
ever documented cases of yellow fever in Asia, concerns have arisen
about vaccine availability in case the virus starts to circulate in
other world regions where the mosquito vector is present, such as Asia
(Paules & Fauci, 2017; WHO, 2016; Wasserman et al., 2016).
Although the existing live-attenuated vaccine confers lifetime immunity
in up to 99% of vaccinees (Paules & Fauci, 2017), rare but serious
adverse effects have been reported (Seligman, 2014; Porudominsky &
Gotuzzo, 2018) and usually become more apparent during mass vaccination
campaigns in the setting of outbreaks. Thus, considering the risks of
YFV spread to world regions where the population is naïve but the
mosquito vector circulates, the recently experienced shortages of the
vaccine and the occurrence of rare, but fatal adverse effects
demonstrate that there is a need for a new yellow fever vaccine (Shearer
et al., 2018; Wong et al., 2019).
In this context, virus-like particles (VLPs) represent a promising
alternative, since they are 3D particles formed by structural viral
proteins produced in recombinant form, thus mimicking the virus, but
lacking its genetic material. Antigens are presented in a repetitive
mode, enhancing immunogenicity, so VLPs are considered a vaccine
platform that combines good safety and high efficacy (Fuenmayor et al.,
2017; Krol et al., 2019; Mohsen et al., 2017; Wang & Roden, 2013).
Vaccines must be affordable and their production needs to be scalable.
One way to develop cost-effective production processes is to develop
continuous perfusion cultivation processes, using cells stably
expressing the product. Perfusion technology usually results in high
product quality and high volumetric productivity. However, it is a more
complex technology in comparison to batch processes, since cells need to
be retained inside the bioreactor while medium is continuously renewed
and product is continuously harvested (Castilho, 2015). Finding an
appropriate cell retention device may be a challenge, considering that
the device needs to work continuously for long periods presenting high
efficiency, not causing cell damage and not retaining product in order
to allow its continuous recovery in the harvest stream (Bielser et al.,
2018; Carvalho & Castilho, 2017).
In the present work we generated a stable recombinant HEK293 cell line,
which constitutively secretes yellow fever VLPs comprised of the prM and
E proteins. In order to develop an intensified VLP production process,
we combined two strategies: the use of FACS to sort recombinant cells
for high producers and the use of the resulting enriched cell pools to
develop continuous perfusion processes based on two different cell
retention devices (ATF and cell settler). In order to better
characterize the VLP product, VLPs were purified by steric exclusion
chromatography and then analyzed by conventional biochemical techniques
and by transmission electron microscopy.