Issues associated with residual cell-substrate DNA in viral vaccines

The presence of some residual cellular DNA derived from the production-cell substrate in viral vaccines is inevitable. Whether this DNA represents a safety concern, particularly if the cell substrate is derived from a tumor or is tumorigenic, is unknown. DNA has two biological activities that need to be considered. First, DNA can be oncogenic; second, DNA can be infectious. As part of our studies to assess the risk of residual cell-substrate DNA in viral vaccines, we have established assays that can quantify the biological activities of DNA. From data obtained using these assays, we have estimated the risk of an oncogenic or an infectious event from DNA. Because these estimates were derived from the most sensitive assays identified so far, they likely represent worst-case estimates. In addition, methods that inactivate the biological activities of DNA can be assessed and estimations of risk reduction by these treatments can be made. In this paper, we discuss our approaches to address potential safety issues associated with residual cellular DNA from neoplastic cell substrates in viral vaccines, summarize the development of assays to quantify the oncogenic and infectivity activities of DNA, and discuss methods to reduce the biological activities of DNA.     

Safety of Viral DNA in Biological Products

Data from studies of the infectivity of DNA injected directly into laboratory animals are used to estimate the potential infectivity risk of residual DNA in biological products. The potential for some novel products to contain infectious quantities of residual cellular DNA is discussed, and further study of this subject is suggested.     

Quantitative determination of the infectivity of the proviral DNA of a retrovirus in vitro: Evaluation of methods for DNA inactivation

All viral vaccines contain contaminating residual DNA derived from the production cell substrate. The potential risk of this DNA, particularly when derived from tumorigenic cells, has been debated for over 40 years. While the major risk has been considered to be the oncogenicity of the DNA, another potential risk is that a genome of an infectious virus is present in this DNA. Such a genome might generate an infectious agent that could establish an infection in vaccine recipients. To determine the quantity of a retroviral provirus in cellular DNA that can establish a productive infection in vitro, we developed a transfection/co-culture system capable of recovering infectious virus from 1 pg of cloned HIV DNA and from 2 μg of cellular DNA from HIV-infected cells. We demonstrate that infectivity can be reduced to below detectable levels either by lowering the median size of the DNA to 350 base pairs or by treatment with β-propiolactone. From the amount of reduction of infectivity, we calculate that clearance values in excess of 10⁷ are attainable with respect to the infectivity associated with residual cell-substrate DNA. Thus, the potential risk associated with DNA can be substantially reduced by degradation or by chemical inactivation.     

Localization of SV40 genes within supercoiled loop domains

Recent studies indicate that eukaryotic DNA is organized into supercoiled loop domains. These loops appear to be anchored at their bases to an insoluble nuclear skeleton or matrix. Most of the DNA in the loops can be released from the matrix by nuclease digestion; the residual DNA remaining with the nuclear matrix represents sequences at the base of the loops, and possibly other sequences which are intimately associated with the nuclear matrix for other reasons. Using a quantitative application of the Southern blotting technique, we have found this residual DNA from SV40 infected 3T3 cells to be enriched in SV40 sequences, indicating that they reside near matrix-DNA attachment points. An enrichment of 3-7 fold relative to total cellular DNA, was found in each of three different lines of SV40 infected 3T3 cells. Control experiments with globin genes showed no such enrichment in this residual matrix DNA. This sequence specificity suggests that the spatial organization of DNA sequences within loops may be related to the functionality of these sequences within the cell.     

Measuring synthesis rates of muscle creatine kinase and myosin with stable isotopes and mass spectrometry

The biochemical measure of success in assisted cartilage repair is normally judged by repair tissue cell density, mRNA and protein expression, and accumulation of extracellular matrix molecules. Existing methods to solubilize cartilage matrix proteoglycans and cellular DNA for quantification, such as papain digestion, often destroy one or more species of the above-named parameters, in order to render others measurable. We have therefore developed a methodology to measure specific levels of mRNA, protein, DNA, glycosaminoglycan, and collagen content on single pulverized 10-mg samples of cartilage, or tissue-engineered cartilage, using successive extractions in concentrated guanidine hydrochloride (GuCl) and guanidine thiocyanate (GITC) solutions. Conditions were developed to solubilize most cellular proteins, DNA, proteoglycans, and some matrix proteins with an initial GuCl extraction step. A subsequent extraction with GITC was essential to solubilize the majority of the cellular RNA. Guanidine-insoluble material was rendered soluble by papain digestion, to enable quantification of collagen, residual glycosaminoglycan, and residual unextracted DNA in individual samples. In general, total collagen, GAG, and DNA content measured in multivalent-extracted samples was similar to that obtained with samples digested directly with papain. Moreover, we were able to reliably detect, in these same multivalent extracts, expressed mRNA as well as specific cellular and extracellular matrix proteins. This multivalent assay could be applied to a variety of cells cultured in biopolymers and to tissues from which biochemical components may be otherwise difficult to extract.     

Development of a purification process for adenovirus: controlling virus aggregation to improve the clearance of host cell DNA

The clearance of host cell DNA is a critical goal for purification process development for recombinant Ad5 (rAd5) based vaccines and gene therapy products. We have evaluated the clearance of DNA by a rAd5 purification process utilizing nuclease digestion, ultrafiltration, and anion exchange (AEX) chromatography and found residual host cell DNA to consistently reach a limiting value of about 100 pg/10(11) rAd5 particles. Characterization of the purified rAd5 product using serial AEX chromatography, hydroxyapatite chromatography, or nuclease treatment with and without particle disruption showed that the residual DNA was associated with virus particles. Using a variety of additional physical characterization methods, a population of rAd5 virus in an aggregated state was detected. Aggregation was eliminated using nonionic detergents to attenuate hydrophobic interactions and sodium chloride to attenuate electrostatic interactions. After implementation of these modifications, the process was able to consistently reduce host cell DNA to levels at or below 5 pg/10(11) rAd5 particles, suggesting that molecular interactions between cellular DNA and rAd5 are important determinants of process DNA clearance capability and that the co-purifying DNA was not encapsidated.     

Effects of large and small T antigens on DNA synthesis and cell division in simian virus 40-transformed BALB/c 3T3 cells.

The roles of the large T and small t antigens of simian virus 40 in cellular DNA synthesis and cell division were analyzed in BALB/c 3T3 mouse cells transformed by wild-type, temperature-sensitive A (tsA), or tsA-deletion (tsA/dl) double mutants. Assessment of DNA replication and cell cycle distribution by radioautography of [3H]thymidine-labeled nuclei and by flow microfluorimetry indicate that tsA transformants do not synthesize DNA or divide at the restrictive temperature to the same extent as they do at the permissive temperature or as wild-type transformants do at the restrictive temperature. This confirms earlier studies suggesting that large T induces DNA synthesis and mitosis in transformed cells. Inhibition of replication in tsA transformants at the restrictive temperature, however, is not complete. Some residual cell division does occur but is in large part offset by cell detachment and death. This failure to revert completely to the parental 3T3 phenotype, as indicated by residual cell cycling at the restrictive temperature, was also observed in cells transformed by tsA/dl double mutants which, in addition to producing a ts large T, make no small t protein. Small t, therefore, does not appear to be responsible for the residual cell cycling and plays no demonstrable role in the induction of DNA synthesis or cell division in stably transformed BALB/c 3T3 cells. Comparison of cell cycling in tsA and tsA/dl transformants, normal 3T3 cells, and a transformation revertant suggests that the failure of tsA transformants to revert completely may be due to leakiness of the tsA mutation as well as to a permanent cellular alteration induced during viral transformation. Finally, analysis of cells transformed by tsA/dl double mutants indicates that small t is not required for full expression of growth properties characteristic of transformed cells.     

Degradation of DNA in Haemophilus Influenzae Cells after X-Ray Irradiation: I. Experimental Results

Sequential measurements of DNA in Haemophilus influenzae cells after X-ray irradiation show rapid initial degradation of DNA followed by a plateau after about 40 min at normal growth conditions. Both the initial rate and final amount of degradation increase with radiation exposure. Degradation is somewhat greater in stationary-phase than in log-phase cells, but colony-forming ability (CFA) is independent of cell stage. Distributions of single-strand lengths of DNA in unirradiated or irradiated cases, as measured by alkaline sucrose gradient techniques, are neither monodispersive nor random, and possible causes for nonrandomness are discussed. The energy dissipated in the DNA is estimated as 40-50 eV per single-strand break for log-phase cells. The fractions of initial DNA remaining in heavily irradiated cells after long incubation are much greater than either the residual CFA or the number of DNA strands free of breaks. Hence, we conclude that cellular degradation of DNA, after exposure to ionizing radiation, cannot be explained quantitatively or qualitatively by simple correlations to these measures of cellular damage, but rather requires a more complex theory.     

Purification, potency and immunogenicity analysis of Vero cell culture-derived rabies vaccine: a comparative study of single-step column chromatography and zonal centrifuge purification

Continuous Vero cell lines are more suitable for large-scale production of rabies vaccine. The purification of Vero cell-derived rabies vaccine is critical because of the residual cellular DNA and serum proteins. The perfection of techniques using column chromatography with different matrix material, gel filtration and zonal centrifugation is of paramount importance for the optimal purification of rabies vaccine, leaving minimal residual cellular DNA, below the permissible level of 100 pg per dose and serum protein content of 1 ppm. In this study the potency, immunogenicity and safety of Vero cell-derived rabies vaccines were compared following purification by densely or loosely packed DEAE-sepharose CL-6B columns with different bed heights or by zonal centrifugation. The optimal virus recovery and maximum removal of substrate DNA and serum proteins were achieved only when the sepharose CL-6B column bed height was maintained at a thickness of 2-2.5 cm. The rabies virus material was purified by layering over the matrix without applying pressure. DEAE-sepharose CL-6B column purification using a simplified, cost effective technique as described in this study enhances the antigen yield by 50% in comparison with zonal purification.     

Temperature sensitive initiation of DNA synthesis in a mutant of Salmonella typhimurium

Summary A mutant (dna-1) of Salmonella typhimurium defective in DNA synthesis is described. DNA synthesis is stopped in this mutant at 42° after a residual synthesis amounting to about 50 to 60% of the total cellular DNA in minimal medium and about 120 to 200% in a medium enriched with amino acids. Reshift back to permissive temperature after the inhibition of DNA synthesis at 42° allows for recovery of DNA synthesis after a lag of about 30 min. Protein synthesis is required during that lag for the recovery of DNA synthesis at permissive temperature. The density transfer experiments indicate that in the mutant dna-1 chromosome termini are replicated normally at 42° while the initiation of new rounds of replication is inhibited although the mutation is probably leaky at this temperature. The mutant is hypersensitive to sodium deoxycholate at 42° which suggests alteration of the membrane structure.     

A Mouse Strain Defective in Both T Cells and NK Cells Has Enhanced Sensitivity to Tumor Induction by Plasmid DNA Expressing Both Activated H-Ras and c-Myc

As part of safety studies to evaluate the risk of residual cellular DNA in vaccines manufactured in tumorigenic cells, we have been developing in vivo assays to detect and quantify the oncogenic activity of DNA. We generated a plasmid expressing both an activated human H-ras gene and murine c-myc gene and showed that 1 µg of this plasmid, pMSV-T24-H-ras/MSV-c-myc, was capable of inducing tumors in newborn NIH Swiss mice. However, to be able to detect the oncogenicity of dominant activated oncogenes in cellular DNA, a more sensitive system was needed. In this paper, we demonstrate that the newborn CD3 epsilon transgenic mouse, which is defective in both T-cell and NK-cell functions, can detect the oncogenic activity of 25 ng of the circular form of pMSV-T24-H-ras/MSV-c-myc. When this plasmid was inoculated as linear DNA, amounts of DNA as low as 800 pg were capable of inducing tumors. Animals were found that had multiple tumors, and these tumors were independent and likely clonal. These results demonstrate that the newborn CD3 epsilon mouse is highly sensitive for the detection of oncogenic activity of DNA. To determine whether it can detect the oncogenic activity of cellular DNA derived from four human tumor-cell lines (HeLa, A549, HT-1080, and CEM), DNA (100 µg) was inoculated into newborn CD3 epsilon mice both in the presence of 1 µg of linear pMSV-T24-H-ras/MSV-c-myc as positive control and in its absence. While tumors were induced in 100% of mice with the positive-control plasmid, no tumors were induced in mice receiving any of the tumor DNAs alone. These results demonstrate that detection of oncogenes in cellular DNA derived from four human tumor-derived cell lines in this mouse system was not possible; the results also show the importance of including a positive-control plasmid to detect inhibitory effects of the cellular DNA.     

Influence of environmental exposure to PAHs on the susceptibility of lymphocytes to DNA-damage induction and on their repair capacity

The influence of occupational exposure to environmental carcinogenic polycyclic aromatic hydrocarbons (c-PAHs) on DNA damage detected in lymphocytes of exposed people (city policemen) was studied. The cellular susceptibility to the induction of the DNA damage and the repair capacity of exposed donors are presented in comparison with matched controls. Monitoring was performed and blood samples (164 donors) were collected in Prague, Czech Republic, during the winter and summer seasons. The single-cell gel electrophoresis (SCGE) assay with an internal standard was applied to evaluate the DNA damage. A challenging dose of 2Gy of X-rays was used to study cellular capacities. In the results of studies of the DNA damage induced in vivo or as an immediate response to the challenging treatment no significant difference was found between exposed and unexposed subgroups. The percentage of non-repaired X-ray-induced DNA damage (residual damage, RD) overall in both seasons was significantly higher in lymphocytes of policemen exposed to c-PAHs than in matched controls (RD(T-DNA), %DNA in the comet tail: winter 36.4+/-22.1 versus 22.7+/-10.8, p < 0.001; summer 47.7+/-22.9 versus 34.7+/-15.2, p < 0.001). The results suggest that occupational exposure to environmental c-PAHs significantly reduces the cellular capacity to repair the DNA damage induced by a challenging treatment. A significant decrease of repair efficiency in donors occupationally exposed to environmental c-PAHs was also observed when subgroups were stratified according to smoking history. In conclusion, our results suggest that environmental exposure to c-PAHs affects the cellular repair processes and can lead to harmful effects hazardous to human health.     

Rapid inactivation and proteasome-mediated degradation of OGG1 contribute to the synergistic effect of hyperthermia on genotoxic treatments

Inhibition of DNA repair has been proposed as a mechanism underlying heat-induced sensitization of tumour cells to some anticancer treatments. Base excision repair (BER) constitutes the main pathway for the repair of DNA lesions induced by oxidizing or alkylating agents. Here, we report that mild hyperthermia, without toxic consequences per se, affects cellular DNA glycosylase activities, thus impairing BER. Exposure of cells to mild hyperthermia leads to a rapid and selective inactivation of OGG1 (8-oxoguanine DNA glycosylase) associated with the relocalisation of the protein into a detergent-resistant cellular fraction. Following its inactivation, OGG1 is ubiquitinated and directed to proteasome-mediated degradation, through a CHIP (C-terminus of HSC70-interacting protein) E3 ligase-mediated process. Moreover, the residual OGG1 accumulates in the perinuclear region leading to further depletion from the nucleus. As a consequence, HeLa cells subjected to hyperthermia and exposed to a genotoxic treatment have a reduced capacity to repair OGG1 cognate base lesions and an enhanced cell growth defect. The partial alleviation of this response by OGG1 overexpression indicates that heat-induced glycosylase inactivation contributes to the synergistic effect of hyperthermia on genotoxic treatments. Taken together, our results suggest that OGG1 inhibition contributes to heat-induced chemosensitisation of cells and could lay the basis for new anticancer therapeutic protocols that include hyperthermia.     

Oncogenicity of DNA in vivo: Tumor induction with expression plasmids for activated H-ras and c-myc

All vaccines and other biological products contain contaminating residual DNA derived from the production cell substrate. Whether this residual cell-substrate DNA can induce tumors in vaccine recipients and thus represent a risk factor has been debated for over 50 years without resolution. As a first step in resolving this issue, we have generated expression plasmids for the activated human H-ras oncogene and for the murine c-myc proto-oncogene. Their oncogenic activity was confirmed in vitro using the focus-formation transformation assay. Two strains of adult and newborn immune-competent mice were inoculated with different amounts of either plasmid alone or with a combination of the H-ras and c-myc plasmids. Tumors developed only in mice inoculated with both plasmids and only at the highest amount of DNA (12.5 microg of each plasmid). The NIH Swiss mouse was more sensitive than the C57BL/6 mouse, and newborn animals were more sensitive than adults. Cell lines were established from the tumors. PCR and Southern hybridization analyses demonstrated that both inoculated oncogenes were present in all of the tumor-derived cell lines and that the cells in the tumors were clonal. Western analysis demonstrated that both oncoproteins were expressed in these cell lines. These results demonstrate that cellular oncogenes can induce tumors following subcutaneous inoculation. Such information provides a possible way of evaluating and estimating the theoretical oncogenic risk posed by residual cell-substrate DNA in vaccines.     

Rapid clearance of intranasally administered DNA from rat tissues

The cold-adapted (ca) live attenuated influenza vaccine (LAIV) strains are manufactured in embryonated hens' eggs. Recently, a clonal isolate from Madin Darby Canine Kidney (MDCK) cells was derived and characterized to assess its utility as a potential cell substrate for the manufacturing of LAIV [1]. Since MDCK cells are a transformed continuous cell line [2], and low levels of residual cellular components (DNA and protein) are found in the intermediates and final filled vaccine, we sought to characterize the uptake and clearance of MDCK DNA from tissues in order to assess theoretical risks associated with manufacturing LAIV in MDCK cell culture.In order to address this concern, MDCK DNA uptake and clearance studies were performed in Sprague Dawley rats. DNA extracted from MDCK Master Cell Bank (MCB) cells was administered via an intranasal (IN) or intramuscular (IM) route. Tissue distribution and clearance of MDCK DNA were then examined in fourteen selected tissue types at selected time points post-administration using a quantitative PCR assay specific for canine (SINE) DNA.Results from these studies demonstrate that the uptake and clearance of MDCK DNA from tissues vary depending on the route of administration. When DNA was administered intranasally, as compared to intramuscularly, detectable DNA levels were lower at all time points. Thus, the intranasal route of vaccine administration appears to reduce potential risk associated with residual host cell DNA that may be present in cell culture produced final vaccine products.     

Purification process monitoring in monoclonal antibody preparation: contamination with viruses, DNA and peptide growth factors.

Administration in vivo of monoclonal antibodies to humans is challenged by considerations regarding their safety. Contamination with viruses, potentially oncogenic nucleic acids and biologically active components like growth factors and hormones forms a serious point of concern in this respect. We have investigated the potential risk of viral contamination by measuring the reduction of 12 different viruses (after spiking) in the standard downstream purification process of ascitic fluid. Depending on the type of virus added and the purification step employed, the reduction of infectious virus particles varies considerably. The overall reduction ranges from about 10(3), observed for a member of the family of Papovaviridae, to more than 10(12) for members of the families of Herpesviridae and Orthomyxoviridae. Using hybridization analysis with a mouse (genomic) DNA probe, we show that the amount of residual DNA in ascitic fluids may also vary considerably, ranging from 75 ng/ml to 1 microgram/ml. In crude preparations produced in cell culture, much lower DNA concentrations are found (0.3 ng/ml). When standard downstream purification procedures are applied to ascitic fluid, a significant reduction of residual DNA levels is observed in the purified monoclonal antibody preparations and in intermediate fractions. The overall reduction factors vary from about 10(3) to 10(4), which is also confirmed by spiking experiments with either purified DNA or crude chromatin-like DNA. Using in-vitro cellular assays, we further show that peptide growth factors like PDGF and TGF beta are present in considerable amounts in ascitic fluids. The observed biological activities, however, are completely eliminated during the purification steps applied.     

Estimation of residual DNA in Feulgen reaction. A new correction of DNA determination per nucleus.

For the determination of the residual DNA amount after acid hydrolysis of Feulgen's method, a high salt-fluorochrome assay for DNA (5 microM Hoechst 33258 with 1 M NaCl) was effectively applied. At an optimal time length of acid hydrolysis for Feulgen reaction, the ratio of the residual DNA of non-hydrolysis to total DNA is 10% or more in hepatocyte or lymphocyte nuclei. A lot of residual DNA seems not to be negligible in Feulgen's method. A more accurate determination of DNA can be made by correcting the loss ratio of the residual DNA value to Feulgen DNA value. Thus, the combination assay of Feulgen's method with the present fluorometry is enough to measure separately both the amounts of Feulgen DNA and its residual DNA and successfully determines more accurately the total DNA per nucleus by summing both the amounts. The residual DNA, a resistant portion of the chromatin DNA against acid hydrolysis, is a possible constituent as the physiological component of nuclear structures.     

Low-dose hyper-radiosensitivity is not caused by a failure to recognize DNA double-strand breaks

One of the earliest cellular responses to radiation-induced DNA damage is the phosphorylation of the histone variant H2AX (gamma-H2AX). gamma-H2AX facilitates the local concentration and focus formation of numerous repair-related proteins within the vicinity of DNA DSBs. Previously, we have shown that low-dose hyper-radiosensitivity (HRS), the excessive sensitivity of mammalian cells to very low doses of ionizing radiation, is a response specific to G(2)-phase cells and is attributed to evasion of an ATM-dependent G(2)-phase cell cycle checkpoint. To further define the mechanism of low-dose hyper-radiosensitivity, we investigated the relationship between the recognition of radiation-induced DNA double-strand breaks as defined by gamma-H2AX staining and the incidence of HRS in three pairs of isogenic cell lines with known differences in radiosensitivity and DNA repair functionality (disparate RAS, ATM or DNA-PKcs status). Marked differences between the six cell lines in cell survival were observed after high-dose exposures (>1 Gy) reflective of the DNA repair capabilities of the individual six cell lines. In contrast, the absence of functional ATM or DNA-PK activity did not affect cell survival outcome below 0.2 Gy, supporting the concept that HRS is a measure of radiation sensitivity in the absence of fully functional repair. No relationship was evident between the initial numbers of DNA DSBs scored immediately after either low- or high-dose radiation exposure with cell survival for any of the cell lines, indicating that the prevalence of HRS is not related to recognition of DNA DSBs. However, residual DNA DSB damage as indicated by the persistence of gamma-H2AX foci 4 h after exposure was significantly correlated with cell survival after exposure to 2 Gy. This observation suggests that the persistence of gamma-H2AX foci could be adopted as a surrogate assay of cellular radiosensitivity to predict clinical radiation responsiveness.     

Simian virus 40 associates with nuclear superstructures at early times of infection.

The association of infecting simian virus 40 with insoluble nuclear structures was assayed by disrupting infected nuclei and assaying insoluble fractions for virus. Three methods were used which lyse nuclei but maintain the insolubility of residual nuclear structures: sonication, high-salt-Triton-EDTA extraction, and low-salt-lithium diiodosalicylate extraction. After each type of nuclear extraction, infecting simian virus 40 remained associated with the residual nuclear structures. This association depended strictly on natural viral infections and on the use of buffers containing moderate amounts of salt and Mg2+ for the isolation of infected nuclei. These viral interactions exhibited behavior similar to host cell DNA interactions studied by analogous assays. Both viral DNA and coat proteins were found associated with the host cell nuclear superstructure. We concluding that at early times after infection the viral templates mimic the state of the host cell chromatin by attaching to the cellular nuclear matrix.