Assessing and Addressing The Risks Associated With Sf-Rhabdovirus, An Adventitious Agent In The Baculovirus-Insect Cell System

Introduction

Since 1983, the baculovirus-insect cell system (BICS) has been used to produce thousands of different recombinant proteins for diverse areas of biomedical research.¹ Since 2009, the BICS has also been used to manufacture biologics for human or veterinary medicine.² Four BICS products are currently approved for human use and several others are in various stages of human clinical trials.³ Thus, the BICS is now recognized as a bona fide biologics manufacturing platform.

The insect cell lines most commonly used in the BICS are derived from two different moths, Trichoplusia ni (Tn) and Spodoptera frugiperda (Sf). Cell lines derived from the latter are used to produce most of the licensed biologics currently being manufactured in the BICS. Thus, it was disconcerting, to say the least, when scientists at the FDA’s Center for Biologics Research and Evaluation (CBER) found all Sf cell lines tested are contaminated with an adventitious viral agent, now known as Sf-rhabdovirus.⁴

Other investigators have independently detected⁵ and confirmed⁶ this novel rhabdovirus as a contaminant of various Sf lines, including one used for commercial manufacturing. These results indicate Sfrhabdovirus is a common, if not universal contaminant of Sf cells. The discovery of this adventitious agent revealed a new potential risk associated with the BICS, which could have an adverse impact on its growing acceptance as a commercial biologics-manufacturing platform. Now, experts in the field must meet the challenges associated with the need to assess and address this problem.

Assessment

In what might be considered as an initial risk assessment, phylogenetic analyses were performed to examine the relationship between Sfrhabdovirus and previously known viruses. The only relationship identified was high similarity between short sequences in the Sfrhabdovirus polymerase (L) gene and plant rhabdoviral L genes⁴ . In fact, this finding, together with the characteristic morphology and density of Sf-rhabdovirus particles, was used to classify the newly identified adventitious agent as a member of the Rhabdoviridae⁴ .

Viral host range studies also were performed to assess the risk associated with this agent. The results showed Sf-rhabdovirus is not only an adventitious agent of Sf cells, but is also found in cell lines derived from the related moths, Heliothis subflexa (Hs)5 and Bombyx mori (Bm; Maghodia, A., Geisler, C., and Jarvis, D. unpublished). Additional studies showed Sf-rhabdovirus is infectious for insect cells. These experiments involved inoculating cell lines derived from the moths, Tn and Spodoptera exigua (Se), and the fruitfl y, Drosophila melanogaster (Dm), none of which have any detectable Sf-rhabdovirus, with cellfree media from Sf cells.^4,5 Importantly, the results of analogous experiments suggested Sf-rhabdovirus is not infectious for human or monkey cell lines^4,5 and other investigators have independently confirmed and extended these results (A. Maghodia, C. Geisler, and D. Jarvis, unpublished). Thus, available data suggest Sf-rhabdovirus has a limited host range and most likely poses little, if any risk for humans or other mammals.

However, this interpretation is weakened by the fact that there are at least four distinct strains of Sf-rhabdovirus, whose host ranges have not yet been thoroughly assessed. The four strains identified to date have X genes with either full length coding sequences or 320 bp deletions and L genes with either full-length coding sequences or 6 bp deletions (Table 1).^4,5 The 320 bp deletion in the X gene eliminates most of the coding sequence and downstream untranslated region. The 6 bp deletion in the L gene eliminates two amino acids from the gene product. A limited comparison of the infectivity of individual Sfrhabdovirus strains revealed no major differences in their infectivity for Sf cells (A. Maghodia, C. Geisler, and D. Jarvis, unpublished). This is not surprising, considering rhabdoviral X genes encode non-conserved, non-essential, “accessory” proteins⁷ and the small, in-frame L gene deletion eliminates just two amino acids from an unconserved region of the polymerase. Further studies are needed to assess the host range of individual Sf-rhabdovirus strains and various combinations thereof, some of which co-exist in Sf cells. The most relevant assessment will be to determine their infectivity for mammalian cells. However, any new information on Sf-rhabdovirus as a relatively newly identified infectious agent will be of general interest to the virology community.

Solutions

Irrespective of the results and interpretations of risk assessment experiments, experts in this field will have to move beyond assessment and ultimately undertake efforts to directly address the risks associated with Sf-rhabdovirus. One approach could be to take steps to effectively clear this viral contaminant and document its removal from every biologic produced in the BICS with Sf cells as the host. Another approach could be to isolate Sf cell lines with no detectable Sf-rhabdovirus for use as improved hosts for biologics production in the BICS. Both approaches would enhance product safety, but only the latter would enhance platform safety.

Currently, there are no published examples of effective, clearly documented efforts to clear Sf-rhabdovirus from any commercial product manufactured in the BICS with Sf cells as the host. However, the successful application of viral clearance as one way to address the Sf-rhabdovirus contamination problem was presented at a recent conference focused on BICS technology (Sixth Annual Meeting of the International Society for BioProcess Technology, Washington, DC, 2016).

A poster at that same conference described efforts to create an Sf-rhabdovirus-negative (Sf-RVN) Sf cell line, which were recently published.⁶ The remainder of this article will focus on select features of this new cell line relevant to its application as an alternative host for biologics manufacturing in the BICS.

The definitive feature of Sf-RVN cells is they have no detectable Sfrhabdovirus. This can be demonstrated by using the ultrasensitive L-gene specifi c nested RT-PCR assay initially developed by the CBER group that discovered Sf-rhabdovirus.⁴ In our hands, this assay has a detection limit of ~24 copies of Sf-rhabdovirus DNA in extracts from 1 X 10⁶ Sf cells, which equates to ~2.4 x 10^-5 copies per cell. However, these calculations underestimate the true limit of detection because they do not include the amplification obtained by reverse transcribing the Sf-rhabdoviral RNA, which is difficult to measure. Using this assay to monitor Sf-RVN cells during routine maintenance in our lab, we found no evidence of Sf-rhabdovirus contamination during 120 serial passages over the course of about 10 months in culture (Fig. 1A). Total RNA from Sf-rhabdovirus-contaminated Sf (Sf9) cells was used for positive controls and total RNA from Dm (S2) cells and reactions with no template (H2 O) were used for negative controls in these experiments. In addition, RT-PCRs with primers specific for the endogenous Sf ribosomal protein L3 (RPL3) gene were used to validate the total RNA preparations and RT-PCR method (Fig. 1B). To control for the possibility that the Sf-RVN cells were contaminated with an Sf-rhabdovirus variant lacking one or more of the L-specific primer binding sites, additional RT-PCRs were performed with Sf-rhabdovirus N-, M-, P-, and G-specific primer pairs. The results showed all primer pairs produced strong signals with total RNA from Sf9, but not Sf-RVN cells (Fig. 1C-1D). RT-PCRs with any of these primer pairs also routinely produce strong amplification products with total RNA from pellets produced by ultracentrifugation of Sf9, but not Sf-RVN cell-free media6 (A. Maghodia, C. Geisler, and D. Jarvis, unpublished).

Productivity is another critically important feature of any insect cell line used as the host component of the BICS. One might reasonably expect Sf-RVN cells to have higher productivity due to elimination of the presumed metabolic load associated with Sf-rhabdovirus contamination. To examine this possibility, we compared recombinant protein production levels obtained using Sf9 and Sf-RVN cells with Escherichia coli ß-galactosidase, human secreted alkaline phosphatase (hSEAP), and human erythropoietin (hEPO) as models. We infected both cell lines in parallel with Sf-rhabdovirus-free baculovirus stocks encoding these proteins, then measured the amounts of each protein produced intracellularly (ß-galactosidase) or produced and secreted (hSEAP and hEPO) at various times after infection. The average immunoblotting and laser scanning densitometry results obtained in three biological replicates revealed no significant differences in productivity (Fig. 2). Thus, somewhat surprisingly, Sf-RVN cells do not appear to provide higher productivity, as compared to Sf9 cells. This suggests either there is little or no metabolic load associated with persistent Sf-rhabdovirus infection of Sf cells or eliminating that load has little or no impact on their productivity. Clearly, these interpretations are preliminary, as they are based on results obtained with a small set of just three model proteins. It will be important and interesting to follow up with additional work designed to compare the productivity of Sf-RVN and Sf9 cells using a larger and more diverse set of recombinant proteins.

The last feature of Sf-RVN cells to be addressed herein is their utility as hosts for the production of infectious baculovirus vectors. This is a particularly relevant capability for biomanufacturing, as that process requires not only scaling-up insect cells, but also scaling-up baculovirus vector stocks needed to infect them. Thus, we compared the amounts of infectious virus recovered after infecting Sf-RVN or Sf9 cells with baculoviral vectors at low multiplicity, which is the condition typically used to produce scaled-up working stocks. The results showed Sf-RVN cells produced higher titer stocks of both baculovirus vectors tested, as compared to Sf9 cells (Fig. 3).

The new data presented in this article support and extend the original conclusion that Sf-RVN cells have no Sf-rhabdovirus contamination.⁶ Even an extremely low level of contamination would have led to a detectable nested RT-PCR signal over 120 passages, as evidenced by the positive controls for that assay and the demonstrated ability of various Sf-rhabdovirus strains to replicate in Sf cells⁴ (A. Maghodia, C. Geisler, and D. Jarvis, unpublished). Thus, we have shown it is, indeed, possible to produce Sf-rhabdovirus-negative cell lines, which can be used as alternative hosts to address the Sf-rhabdovirus contamination problem and enhance the safety profile of the BICS.

Summary

The purpose of this article was to discuss the recent discovery of a new adventitious viral agent, Sf-rhabdovirus, in the BICS, and the need for experts in the field to assess and address this discovery. While initial assessments suggest Sf-rhabdovirus is unlikely to be a high risk factor for humans or other mammals, further investigation is needed. In addition, whether or not it is considered to be high-risk, regulatory agencies will likely contend the mere presence of this adventitious agent poses some level of risk that must be addressed. Two potential ways to address this problem include viral clearance, which will have to be effectively invoked for every commercial product destined for human use, or creation of an Sf-rhabdovirus-free cell line(s), which would circumvent the need for Sf-rhabdoviral clearance and enhance the overall safety of the platform. Our group has demonstrated successful completion of the latter approach.

However, we believe Sf-rhabdovirus will not be the last potential adventitious viral agent detected in the BICS. This opinion is based on the results of a recent comprehensive search of the Sf genome and transcriptome for sequences related to Sf-rhabdovirus³ and a deep-sequencing project designed to elucidate the genome and transcriptome of Sf-RVN cells (Geisler, C. and Jarvis, D, unpublished). Surprisingly, these projects revealed previously unidentifi ed intact open reading frames related to Sf-rhabdovirus N and P genes, as well as partial open reading frames related to Sf-rhabdovirus G and L genes. Importantly, these sequences were found in both the Sf genome and transcriptome and, based on extensive literature precedents, we concluded these are endogenous viral elements (EVEs) resulting from genomic integration of partial viral genetic material. We also concluded these Sf-rhabdovirus-like EVEs cannot produce infectious virus particles because they are disseminated across 4 genomic loci, the G and L sequences do not comprise complete open reading frames, and there are no M gene-related sequences.

The provocative CBER study resulting in the original discovery of Sf-rhabdovirus⁴ clearly underscored the value and power of deep sequencing as an approach for assessing the risks associated with various biologics production platforms. However, we believe our follow-up studies underscore an equally important need to elucidate and compare both genomic and transcribed sequences to produce clear conclusions. In our case, this refined approach enabled identification of transcribed EVE’s and allowed us to avoid the potentially incorrect conclusion that our deep sequencing results revealed contamination with yet another replication competent, adventitious viral agent.

References

1. Smith, G.E., Summers, M.D.Fraser, M.J. Production of human beta interferon in insect cells infected with a baculovirus expression vector. Mol. Cell. Biol. 1983;3(12):2156-2165.
2. Felberbaum, R.S. The baculovirus expression vector system: A commercial manufacturing platform for viral vaccines and gene therapy vectors. Biotechnol. J. 2015;10(5):702-714.
3. Geisler, C.Jarvis, D.L. Rhabdovirus-like endogenous viral elements in the genome of Spodoptera frugiperda insect cells are actively transcribed: Implications for adventitious virus detection. Biologicals. 2016;44(4):219-225.
4. Ma, H., Galvin, T.A., Glasner, D.R., Shaheduzzaman, S.Khan, A.S. Identifi cation of a novel rhabdovirus in Spodoptera frugiperda cell lines. J. Virol. 2014;88(12):6576-6585.
5. Haynes, J. in WO2015051255 A1. (ed. U.S.P.a.T. Offi ce) (2015).
6. Maghodia, A.B., Geisler, C.Jarvis, D.L. Characterization of an Sf-rhabdovirus-negative Spodoptera frugiperda cell line as an alternative host for recombinant protein production in the baculovirus-insect cell system. Prot. Expr. Purif. 2016;122(45-55.
7. Wagner, R.R. The Rhabdoviruses. (Springer Sciences + Business Media, New York; 1987).