November 15, 2014 (Vol. 34, No. 20)

Amy Eckert scientist Charles River Laboratories

Novel Approach for Biological Cell Bank Characterization

Cell bank characterization is required for assuring the safety of pharmaceutical biologics. Typical cell bank characterization safety tests, such as in vitro adventitous virus assays, detect infectious viral contaminants in the cell lysate or supernatant.

However, in any particular sample, and at any stage of the manufacturing process, a virus may be present but not actively replicating and, therefore, not detected by conventional infectivity assays. Performing safety tests that only detect replicating virus can be seen as a potential risk when the goal should be to detect latent or endogenous viruses, as well.

With this in mind, an approach to consider for detecting latent or endogenous viruses in cell substrates is the use of assays that employ specific chemical compounds to induce virus expression or replication. Following induction, latent and endogenous viruses can then be detected through subsequent assays which include transmission electron microscopy (TEM) testing for visible virus particles, and quantitative fluorescent polymerse chain reaction (QF-PCR) testing for either retroviral reverse transcriptase activity or viral DNA sequences.

Our initial research and development for the induction study aimed to balance toxicity of the compounds with a successful induction of latent or endogenous viruses. Cell confluence observations were used to monitor compound toxicity and determine optimal induction compound concentrations. Four compounds were tested, singly and in combination, on Madin-Darby canine kidney (MDCK) cells, an adherent cell line:

  • sodium butyrate (NaB)
  • 12-O-tetradecanoylphorbol-13-acetate (TPA)
  • 5-azacytidine (AzaC)
  • 5-iodo-2-deoxyuridine (IdU)

NaB and TPA were used to induce DNA viruses, and AzaC and IdU were used to induce retroviruses.

Initial Experiments

The first experiments tested six different concentrations of each compound on the MDCK cells with three different exposure times—8, 24, and 48 hours. After the exposure to the agents, the cells were washed and fed with growth medium. Confluence and resulting toxicity were monitored for up to five days.

The 8-hour exposure showed minimal effect on the cell monolayer with nearly 100% confluence and no effect on the cell morphology, while the 48-hour exposure decreased final confluence, and for a few compounds, complete cell death occurred at higher compound concentrations. Induction compound concentrations were selected that showed a 25–50% decrease in confluence after treatment following a five-day recovery period compared to nontreated controls.

Based on these results, a 24-hour treatment, which showed only moderate effects on the cells, was selected as an optimal exposure time. To compensate for added toxicity from combining two induction compounds, concentrations for each were decreased by 50%.

Additional studies compared the effect of the induction compound concentrations on MDCK cells grown in medium supplemented with and without fetal bovine serum (FBS). The absence of FBS appeared to make the cells more sensitive to the effects of the compounds; this was demonstrated by increased toxicity and MDCK cell death. Induction compound concentrations were reduced for subsequent testing in serum-free medium.

Compound Concentrations

The induction compound concentrations were also optimized on two positive control cell lines containing endogenous viruses, BC-3 and K-BALB. BC-3, human lymphoblasts containing human herpes virus (HHV-8), is a suspension cell line selected to monitor DNA virus induction. K-BALB, a BALB/c 3T3 cell line transformed by Kirsten murine sarcoma virus, is an adherent cell line selected to monitor retrovirus induction.

K-BALB cells reacted similarly to the induction compounds as sample MDCK cells (Figure 1). Suspension BC-3 cells, however, were more sensitive to the induction compounds, especially TPA (data not shown), and concentrations were subsequently reduced to avoid complete cell death.

After cell confluence tests were used to determine optimal induction compound concentrations, two different PCR tests were designed to detect the presence of induced viruses in the MDCK cell line.

Multiplexed QF-PCR assays were developed and qualified for DNA virus sequence detection. Canine sequences were targeted to identify viruses in the sample MDCK cell line, while human viruses were targeted to identify human viruses that may grow in the canine cell line and potentially be present in the final pharmaceutical product.

Figure 1. Cell growth after treatment of K-BALB cells with various concentrations of IdU. Cells were treated with 0 to 0.25 mg per mL of IdU for 24 hours and then monitored for four days for monolayer confluence.

Targeted Multiple Strains

The following broad-range QF-PCR assays targeted multiple strains of both human and canine viruses:

  • herpesviruses (eight human strains: HHV-1 to HHV-8; two canine strains: CHV-1 and CHV-2)
  • adenoviruses (seven human strains; five canine strains)
  • polyomaviruses (human strains: JC, BK, Merkel cell, KI, WU, HPyV-6, HPyV-7, and HPyV-9; no known canine viruses described in the NCBI database)
  • papillomaviruses (two human strains: HPV-16 and HPV-18; eight canine strains)

In this assay, target sequences in a background of genomic DNA were diluted as copy number standards for quantification of spiked and nonspiked sample preparations. For example, the broad range detection assay for the selected canine and human herpesviruses consisted of 6 multiplexed master mixes to evaluate a total of 10 targeted DNA sequences.

The assays included negative and positive extraction controls, PCR inhibition controls, and an endogenous gene control. Each assay achieved one-copy sensitivity and a 50-copy limit of detection (positive amplification in at least 95% of the reactions).

The second PCR assay, a PCR-based reverse transcriptase assay (PBRT, PERT), was performed to detect broad-range retrovirus activity in the induced cell supernatant samples. In the presence of retrovirus-generated reverse transcriptase enzyme, cDNA is generated from an RNA template (MS2) and detected in the real-time fluorescent PCR assay.

Samples, as well as negative and positive (SMRV, MuLV) internal controls, were tested at multiple concentrations. The PBRT assay was used to monitor the optimal time necessary to detect virus expression levels. To note when PBRT activity peaked then decreased, induced viral products from the positive control cells were tested at 24 to 96 hours after induction.

The PBRT activity levels in the K-BALB positive control peaked on day 2 or day 3, depending on the specific compound used, and decreased on day 4 (Figure 2). Based on this study, samples were harvested on days 2, 3 and 4 following induction to detect levels of PBRT activity in the uninduced and induced MDCK and K-BALB cell lines.

Figure 2. PBRT assay for reverse transcriptase activity in K-BALB cells. Cells were either untreated (NC) or treated with AzaC or IdU, singly or in combination, for 24 hours, washed to remove the compounds, and then assayed for reverse transcriptase (RT) on days 2, 3, and 4. The increase in activity is indicted by the values above the bars and is relative to the uninduced control.

In addition to PCR, cell pellets were tested following induction by TEM to detect virus particles within cells. Retrovirus-like particles were observed in the IdU-induced K-BALB cells. A-type retrovirus-like particles were identified as intracisternal particles within the endoplasmic reticulum, and C-type retrovirus-like particles were identified extracellularly and budding from the cell membrane (Figure 3).

The final results of the induction study did not detect any human or canine viruses in uninduced or chemically induced MDCK cells by QF-PCR or TEM. In addition, no detectable retroviruses were observed in uninduced or chemically induced MDCK cells by the PBRT assay or TEM.

Viruses were detected in the corresponding positive controls using TEM, as well as multiplexed, broad-range DNA QF-PCR and PBRT assays. Multiple tests are crucial to provide an increased level of confidence that each cell bank and pharmaceutical product is safe and free of adventitious agents. Characterization of this MDCK cell bank included routine safety testing for infectious viruses, as well as custom induction studies to detect latent or endogenous viruses. The induction study used multiple compounds to induce viral expression or replication. It was critical to optimize induction conditions for sample and positive control cell lines considering length of induction, medium conditions, and time of induction detection. An induction study can provide added assurance that products of pharmaceutical biologics are free of viral contaminants.

Figure 3. TEM detection of retrovirus-like particles in IdU-treated K-BALB cells. (A) Intracisternal A-type particles in the endoplasmic reticulum. (B) C-type particles observed extracellularly. (C) C-type particles observed budding from the cell membrane.

Amy Eckert ( is a scientist in the virology department, Biologics Testing Solutions, Charles River Laboratories.