Viral DNA contamination is responsible for Epstein-Barr virus detection in cytotoxic T lymphocytes stimulated in vitro with Epstein-Barr virus B-lymphoblastoid cell line

Abstract

Epstein-Barr virus (EBV)-transformed B-lymphoblastoid cell lines (LCLs) are used to prepare human EBV-specific T lymphocytes (EBV-CTL) in vitro. Within an LCL, up to 5-7% the cells release infectious EBV, and this has fostered safety concerns for therapeutic applications because of the exposure of T cells to EBV. The release of infectious viruses can be prevented by ganciclovir, but this drug may seriously affect LCL growth. In the wake of these concerns, the present work was designed to compile safety data on EBV-CTL preparation for the purpose of submission to a regulatory agency. We showed that further to supernatant exclusion, the number of EBV genome copies (EBVc) associated with the EBV-CTL always made up a constant proportion of the EBVc number detected in the culture supernatant. In addition, such was the case whether infectious virus could be produced by the LCL or not, suggesting that the EBV signal detected was due to a DNA contamination rather than an infection. Furthermore, we demonstrated that the number of EBVc associated with the EBV-CTL was highly sensitive to DNAse treatment, and finally that EBVc could no longer be detected after the EBV-CTL had been amplified in the absence of LCL. Consequently, during in vitro EBV-CTL preparation, either T cells cannot be infected or they die rapidly after EBV infection.

Fig. 1

GCV (15 μM) inhibits LCL amplification. LCL⁰ and LCL^G cultures from Do8, Do430, and Do698 were initiated at 0.5 × 10⁶ cells/ml, and cell concentration was adjusted to 0.5 × 10⁶ cells/ml every 2–3 days. Amplification factors were calculated between each dilution, and values show the mean amplification from day 0. Growth inhibition in the presence of GCV was highly significant (p = 0.0001)

Fig. 2

LCL treatment with GCV (15 μM) does not affect EBV-CTL amplification. Values show the mean amplification factors calculated for EBV-CTL of Do8, Do430, and Do698 selected either against LCL⁰ (CTL⁰) or LCL^G (CTL^G). The difference between the two groups was not significant (p = 0.5417)

Fig. 3

Effect of 15-day GCV (15 μM) treatment a on the number of EBV genome copies detected in 1 ml LCL supernatant of Do8 and b in 10⁵ LCL cells for Do8, Do430, and Do698. After GCV inhibition of the lytic cycle in the few lytically infected LCL cells, the mean EBVc per cell decreased to the basic level: 27, 24, and 27 for Do8, Do430, and Do698 LCL, respectively

Fig. 4

CTL (CTL⁰ and CTL^G) and culture supernatant (supCTL⁰ and supCTL^G) associated EBVc during EBV-CTL selection for Do8, Do430, and Do698. a Results are presented on an arithmetic scale to appreciate the evolution of EBVc in 1 ml of culture supernatant in relation to culture conditions; b results are presented on a log scale to appreciate the estimation of the number of EBVc associated with 100,000 EBV-CTL. Together, these results show that culture supernatant and EBV-CTL-associated EBVc evolved according to addition of LCL and culture dilution during the procedure of EBV-CTL selection

Fig. 5

EBV-PCR signal detected among EBV-CTL decreases after DNAse treatment. For 11 different cultures of EBV-CTL, DNAse treatment decreased the number of EBV genome copies per 10⁵ cells from a mean of 901 to 111 (87.6%)