VIRAL TRANSMISSION BY BIOLOGICAL PRODUCTS: AN OVERVIEW 
Viral safety of biological products came to my attention many years ago when I worked within different technical groups of WHO. In the middle of the years 1980's as a virologist at the Pasteur Institute, I became more involved in the area of biological safety since colleagues from French and European regulatory systems asked me to help them.
Transmission of infectious agents by pharmaceutical has a long and rich history but I will try to mention here only a few significant examples in order to illustrate that every technical progress, that allowed a new generation of biologicals, was marked by tragic iatrogenic accidents. The term "iatrogen", that is, a disease induced by a medical procedure or a treatment, was used for the first time in the 1920's to designate the treatment of the neurological phase of syphilis by malaria initiated by the Austrian clinician Wagner von Jaureg. The plasmodium parasite was deliberately inoculated to the patients in order to provoke a high fever access capable of destroying the treponema of the syphilis. This treatment was called ""iatrogenic therapy" and Wagner von Jaureg received in 1927 the Nobel Prize.
Tissues and fluids of human and animal origin and plant extracts were probably the first medicines used by man. Tribal or religious practices led then to illnesses such as Kuru, a form of spongiform encephalopathy described by Gajdusek, which persisted in Papua New Guinea until recently.
Biologicals are a large category of medicinal products and their classification is shown in the slide 1. The risk of viral transmission by different category of biologicals is presented in the slide 2 and 2bis.
Louis Pasteur, developing the empirical observation of Jenner, created the basis of a new industry of vaccines, that used as starting material human and animal pathogenic micro-organisms. For this reason it became evident that vaccines brought, in association to their enormous benefic effect, a high risk of adverse reactions.
Severe accidents of transmission of infectious agents were reported from the early period of vaccination on. The first was linked to the Jennerian era of small pox immunization, when vaccinia material from human origin was used for vaccination by passages from arm to arm. The most spectacular accident occurred in 1861 in Rialta, Italy where 46 children and 20 nurses vaccinated with human material were contaminated with syphilis. It is very probable that during the 19th century the practice of small-pox vaccination with human vaccinia contributed significantly to the spread of syphilis.
Historically, the fortuitous spread of tuberculosis by BCG vaccine, became well known under the name of Lübeck (Germany) accident. Between December 1929 and 30 April 1930, 251 infants, in Lübeck, received, during the first 10 days of life, three doses of BCG vaccine administered by mouth. Out from the vaccinated children, 72 died of tuberculosis in the next 2-5 months, 135 developed a clinical tuberculosis but recovered and 44 became tuberculin-positive but remained well. In parallel, none of the 161 unvaccinated children, born during the same period, died of tuberculosis in the subsequent three years. This tragedy created a strong emotion of the scientific community and accounted for a number of international studies performed by well-known bacteriologists. Although it has been clear that the vaccine prepared in Lübeck in the laboratory of Dr. Georg Deycke, an experienced bacteriologist and a tuberculosis expert, contained virulent tuberculosis bacilli, it was not possible to determine the circumstances in which the BCG vaccine strain has been contaminated.
I refer in some details to this tragedy, because the Lübeck accident led to a trial, which, to the best of my knowledge was the first time a iatrogenic disease induced by a biological product, in this case a live attenuated bacterial vaccine, had legal consequences. The trial was long, since it took one year and a half, and the great British microbiologist Graham ? Wilson said that "The long-drawn-out spectacle of the trial resembled ancient Greek tragedy, played between the doctors and the fates pursuing its way relentlessly to its climax of horror and death".
Two defendants were found guilty of manslaughter by negligence in 68 cases and injury by negligence in 131 and a sentence imprisonment was passed and two other were acquitted. If the Lübeck accident was the first that ended with a trial, unfortunately it was not the last.
The list of iatrogenic accidents induced by vaccines is too long to be detailed here, but the place occupied by adverse reactions to viral vaccines was important and some of them are mentioned in slides 3 and 4.
If we examine chronologically the accidents induced vaccines we can observe that most of them occurred when a new generation of products became available. This is particularly true for viral vaccines, because the iatrogenic accidents usually paralleled a new cell substrate used for vaccine manufacturing. This was at the origin of the debate around the cell substrate safety used for the production of viral vaccines.
An extensive review of iatrogenic accidents generated by immuno-prophylactic products can be found in the book published by Sir Graham Wilson in 1966 ("The Hazard of Immunization").
The accidents provoked by vaccines have today only an historical interest. Thanks to the progress of scientific knowledge and therefore to a better risk-benefit analysis, and to the implementation of severe national and international requirements for manufacturing and control, the vaccines are today a safe category of biologicals.
The invention of a replacement therapy, pioneered in 1921 to 1922 by Banting and Best, was a major step forward in Twentieth Century medicine. These Canadian scientists identified the function of insulin in carbohydrate metabolism and demonstrated that the hormone extracted from animals was able to compensate for the pancreatic deficiency. As the same time, Landsteiner was identifying the blood groups, providing a scientific basis for transfusional medicine and a boosting blood transfusion in the 1920's. The beginning of the replacement therapy were followed, after the Second World War, by other important developments such as Cohn's alcoholic fractionation of plasma.
A look back to the 1980's shows that the AIDS pandemic surprised the medical world and before we were able to identify its causal agent, thousands of transfused patients and a large part of hemophiliac population was heavily contaminated. This tragedy will probably remain the major event that marked the public health, in the second half of our century.
After the transmission of HIV in the mid 1980's, governmental agencies in the USA and Europe, in cooperation with the industry, made a considerable effort to develop new regulations for the manufacturing and control of blood derivatives. Despite the remarkable progresses accomplished in the field of viral safety of blood and its derivatives, some iatrogenic transmission occurred until recently (Slide 5- data transmitted by PEI).
The story of blood viruses is not yet finished since new viruses are described, but their role as human pathogens is not yet well established. Nevertheless, the increased vigilance for the viral safety of plasma derivatives is justified, since they will continue to be used in the next years. At the present time, for conceptual, technical and commercial reasons, it is not possible to replace them with products obtained by other technologies, such as recombinant DNA.
Biologicals obtained recently from animal cell culture represent now a large category of products, including highly purified recombinant proteins or monoclonal antibodies. Although so far they were not involved in transmission of viruses (touch wood!) they are not exempted of risk.
These categories of biologicals are produced on large bio-reactors in media containing ingredients from animal origin (such calf sera) or from human origin (such as albumin). However, even in the absence of media ingredients from animal origin, viral contamination with a murine parvovirus has been reported.
The operator can be also the source of contamination as was the case for a rhinovirus present in a culture medium used for the production of a hepatitis A vaccine. Some recent data on the contamination in process, are presented in slide 6.
The transmission of CJD by pituitary growth hormone raised numerous and complex problems. When revisiting this tragic accident, we must consider what was rightly so stated by George Craig in the paper he published recently in the New-York Review of Books: "Our view of the past can never be a permanent one. Revision is a part of the historical process, made inevitable by the passage of the time and the change perspective that comes with it and the accumulation of new knowledge". In other terms, the knowledge we have today on CJD and its mode of transmission was not available in the early 1960's when the risk of hGH was evaluated and accepted in the treatment of hypophysary deficiency.
During the 1960's and 1970's important progresses were accomplished in the field of molecule separation and protein purification. These achievements made possible the preparation of a new generation of medicinal products, mainly toxoids and vaccines, cytokines and hormones. With hindsight, it is clear that the development of pituitary GH extracted from pituitary glands collected from cadavers in autopsy rooms, brought to the fore its clinical benefit without a rigorous assessment of the potential risk. In short, the power and the simplicity of the technology used to prepare the hormone in a form quite pure created the impression that the risk, if it existed at all, was extremely small.
Since the start of the 1920's when CJD was described in Germany, this disease remained for forty years a rare clinical curiosity, and archetypal degenerative disease of CNS occurring to elder persons. The situation changed in a spectacular manner when Gajdusek described, in 1957, Kuru disease and later, in 1965-66, when the team of Gajdusek transmitted the disease to chimpanzees. However, while the transmission of CJD to non-human primates by intra-cerebral inoculation was accepted until the appearance of the first identified cases of CJD transmitted by GH, the peripheral routes of transmission were matter of debate. We have not enough time to review in details this iatrogenic transmission and therefore I will comment only on the most significant factors that contributed to the transmission of CJD prions to patient treated with pituitary GH.
The frequency of CJD transmitted by hGH is presented on slide 7 (data of December 1998). From this table, it can be noted that in all situations, CJD was transmitted with a low frequency, a common circumstance for the iatrogenic transmission of the disease.
In slides 8, 9, 10 and 11 are presented some factors that contributed to the transmission of CJD.
The conditions in which the glands were collected from autopsy halls facilitated the cross contamination of hypophyses. Moreover, the presence of infectious material in the SNC of patients apparently healthy augmented the number of infected glands that were, probably, introduced into the pool used for manufacturing hGH.
One important notion to understand the low frequency of transmission is the existence of sub-infectious doses of TSE agents. Kimberline showed, in the case of scrapie, that the amount of infectious doses needed to transmit the disease by the peripheral route is much more higher than from the intra-cerebral inoculation. As mentioned in slide 10, at difference of conventional viruses, the repeated exposure to sub-infectious doses of a TSE agent cannot elicit an immunological response.
The low concentration of the infectious agent in the final product (Poisson distribution) is a situation encountered not only for CJD transmission by hGH, but also for BSE ("low exposure dose"- Kimberline).
It is well known that BSE affected mostly dairy herds. Farms specializing in the rearing of calves for meat production did not generally use the meat meal. Calves from dairy farms, fed with bone and meat, went on to develop the illness; both calves and their mothers were fed a diet including meat meal as a protein supplement. Overall, 85% of BSE cases involved the calves of dairy cows, reared elsewhere for meat production. The percentage of animals developing the illness within a herd increased with herd size, from 3.7% in herds of less that 50 head to 41.09% in specialist dairy farms with more that 200 head of cattle. The overall frequency of BSE in animals raised on dairy farms was 13.99% and more that 85% of all farms reported at least one case of BSE (data from Kimberline).
These figures confirm that the causal agent of BSE is not very virulent, a view that is now widely accepted. Based on the low frequency of the disease within herds, Kimberline developed the notion that the animals that developed the disease had been exposed to low concentrations of prions, which he described as "low dose exposure".
When reconstituting the history of CJD transmission by pituitary growth hormone, one notes that 1985 was the crucial year, when the correlation between the cases of CJD in young patients in the USA and the UK, and the fact that they were treated with pituitary growth hormone became evident. Why was this correlation established so late? This is a long story, but personally I think that the medical community was faced with a new problem and the example that illustrated this situation is given by Paul Brown in a paper published in 1988. A young 20-year-old patient, treated with hGH, developed CJD in 1984. Medical examination in several American University medical centers produced no firm diagnosis. Moreover, in the autumn of 1984, when the patient was still alive, a diagnosis of CJD was put forward at a pediatric neurology meeting but was rejected since the subject was too young.
However, even though there is no publication, before 1985, warning clearly about the danger of CJD transmission by pituitary growth hormone, some scientists, such as Montagnier in France or Wildy in UK evoked this risk in their reports addressed to different organizations. Nevertheless, the general opinion was so much in favor of the use of these new hormones and the safety report of the product was so good, that the decision to stop was, at that time, not justified.
It is interesting to see what happened finally in 1985 when the role of GH in the transmission of the disease was recognized. The USA and UK decided to discontinue the treatment: the same decision was taken by one of the pharmaceutical companies. In USA, Paul Brown expressed clearly doubts about the possibility of eliminating completely the CJD infectious agent. In France, based on the safety report of the hormone prepared by the non-profit organization, France Hypophyse, the treatment was continued until 1987 when the first cases of the disease appeared.
The collection of pituitary glands from cadavers is another factor that facilitated transmission, given the impossibility of selecting glands in the absence of a rapid and sensitive diagnostic test. The presence in the pool of glands contaminated with CJD prions from subjects who had died in the preclinical phase of the disease increased the probability of CJD transmission.
The analysis herein implicates a putative cross contamination during the production of the hormone as the major source of transmission. The origin of this contamination was the pituitary glands contaminated with CJD prions present in the pool of hypophyses used as a source material. The high resistance of the prions to decontamination procedures facilitated cross contamination at all of the steps involved in the manufacturing process. Such contamination must have been infrequent and its level differed between batches (Slide 12).
Under the light of current knowledge, it does not require a complex analysis to recognize that the risk associated with medicines originating from the human CNS is unacceptable. However, there was no reason to take this viewpoint before 1985 and the producers of pituitary hormone were ignorant of the large risk associated with their product.
As we have already seen, the corollary of the transmission of HIV and CJD was the particular attention given by state organizations and by the pharmaceutical industry to the viral safety of biological products and to the enforcement of national and international regulations. During the 1990's, in addition to the WHO recommendations and to the monographs of the European Pharmacopoeia, which have international relevance, Japan, Europe and the United States agreed to formulate very precise consensual norms within the framework of the International Harmonization Conference. Viral safety was considered having a particular priority in the formulation of these norms. Governmental bodies now cooperate directly with scientists in both academia and industry. Groups of experts are brought together to evaluate the risk of viral transmission of every product of human or animal origin. This policy was developed as a consequence of the BSE epidemic in the United Kingdom and the necessity of limiting the possible effects on public health. Despite the considerable efforts that have been made, we must remain vigilant because the lessons of the past show that we must be circumspect with all pharmaceutical products of human or animal origin.
 

Cell Banks and the Concept of Sterility and Purity
Robert J. Hay, American Type Culture Collection (ATCC)
10801 University Boulevard, Manassas, Virginia 20110, USAINTRODUCTION
National cell line resource banks have been created to serve industry as well as academic and governmental research institutions and agencies. Industrial scientists in many cases will require fewer lines but those generally need to be expanded and reauthenticated under GMP conditions prior to use.
This presentation summarizes steps taken by the American Type Culture Collection (ATCC), an international cell resource and supply center, to provide reference cell cultures for research, production, and testing purposes. Initial cultures are provided for propagation and final re-characterization by the recipient laboratory.
Most established cell lines have been characterized by the originator and collaborators well beyond the steps essential for quality control. Specific details include, for example, phase contrast and ultrastructural morphologies; detailed cytogenetic analyses; definition of proto-oncogene, oncogene, or oncogene product presence, nature, and location; detailed evaluation of intermediate-filament proteins; and demonstration of tissue-specific antigens or production of other specific products. These characterizations obviously increase the value of each line for research and for production work. However, cell banking organizations need not attempt to repeat all these tests before distributing the stock cultures. Decisions must be made to establish the most acceptable authentication steps, consistent with maintaining the lowest possible cost, to provide a high quality cell stock. Authentication can be considered the act of confirming or verifying the identity and critical features of a specific line, whereas characterization is the definition of the many traits of the cell line, some of which may be unique and also may serve later to identify or authenticate that line specifically. Essential steps for quality control will vary with the type of cell bank constructed; such minimal descriptive data frequently will be supplemented with a much broader characterization base for each particular cell line.
SEED/MASTER CELL BANK CONCEPT
The utility of any bank of cell lines depends upon the degree of characterization of the holdings performed by the originators, the banking agency, and by other scientists. Ready availability at reasonable cost both of the lines and such data, plus the capability to track distribution of the biologicals are additional critical considerations. Figure 1 outlines steps applied to characterize and authenticate cultures provided by the deriving investigator. Progeny from the ampule or flask culture initially supplied are utilized to produce the first or "token" freeze. Cultures derived from such token material are then tested for bacterial, fungal, and mycoplasmal contamination. The species of each cell line is verified. These quality control steps are the minimum that must be performed before eventual release of a line. If warranted, the material is expanded to produce the seed and distribution stocks. Additional major quality control and characterization efforts are applied to cell populations from seed stock ampules. Test results are derived for individual lots and, therefore, refer to specific stocks. The distribution stock consists of ampules that are distributed on request to investigators. The reference seed stock, however, is retained to generate further distribution stocks as the initial distribution stock becomes depleted. The degree of characterization applied to master cell banks or master working cell banks in production facilities is generally more rigorous, although the seed stock here, like the master cell bank, is used as a reservoir to replenish depleted distribution lots over the years. By adherence to this principle, one can avoid problems associated with genetic instability, cell line selection, senescence, or transformation.
MICROBIAL AND VIRAL CONTAMINATION
Microbial and fungal contamination of cell cultures often are overt and easily observed. Still, less apparent or masked infections occur undetected. The ATCC receives cell cultures, even for the Patent Depository, that contain cryptic bacteria, yeast, filamentous fungi, or mycoplasma.
a. Bacteria and Fungi. Microscopic examination is only sufficient for the detection of gross contaminations. Even some of these cannot be detected readily by simple observations. Therefore, an extensive series of culture tests is required to provide reasonable assurance that a cell line stock or medium is free of fungi and bacteria. Details are given in one of the reprints appended (Hay, 1998).
b. Mycoplasma. Contamination of cell cultures by mycoplasma can be a much more insidious problem. Although the presence of some mycoplasma species may be apparent because of the degenerative effects induced, other mycoplasmas metabolize and proliferate actively without producing any overt morphological change in the contaminated cell line. Thus, cell culture studies relating to metabolism, surface receptors, virus-host interactions, and so forth, are certainly suspect to interpretation, if not negated in interpretation entirely, when conducted with cell lines that harbor mycoplasma. The seriousness of these problems can be documented through published data from testing services and cell culture repositories.
Data from seven different testing laboratories on mycoplasma infection frequencies in cell lines examined is summarized as Table 1. The results indicate clearly that there is a significant problem internationally.
Protocols for test procedures are numerous. A sensitive Polymerase Chain Reaction (PCR)-based kit is now available for the detection and identification of the common species of mycoplasma and Acholeplasma laidlawii known to infect cell cultures. A photo of a representative gel showing amplicons generated using the kit is presented as Figure 2.
Four general recommendations can be offered to avoid mycoplasma infection. The implementation of an effective regimen to monitor cell lines for mycoplasma is one critical step. Quarantining all new untested lines and using mechanical pipetting aids are others. Most experts also strongly suggest that the use of antibiotics be eliminated when possible. Antibiotic-free systems permit overgrowth by bacteria and fungi to provide ready indication whenever a lapse in aseptic technique occurs. When the initial tissue is used, e.g., a human tumor sample, antibiotics may be employed, but after the primary population has grown out and been cryopreserved, reconstituted cells may be propagated further in antibiotic-free medium.
c. Viruses. Verification as to the absence of viruses in cell lines is recognized as a most significant problem. Industrial production for human and veterinary applications has especially stringent requirements in this regard. That virus may coexist as noncytopathic entities (e.g., with the c-type retroviruses) or in a latent form (e.g., papilloma viruses and some herpes viruses) compounds difficulties in detection. Judicious choices are necessary not only to select appropriate methods available for recognizing viruses associated with cell lines, but also to identify the offending species. The nature of the cell line resource, its users, the budget available, and the intended purposes for which the line will be needed all affect decisions on testing. More complete detail and protocols are provided in the article appended (Hay, 1998).
CELLULAR CROSS CONTAMINATION
Wherever cells are grown in culture, serious risk exists for the inadvertent addition and subsequent overgrowth of cells from another individual or species. One cannot rely on morphologic criteria alone to recognize specific cell lines. Data-documenting problems have been collected over the years by groups offering identification services for cell culture laboratories in the United States and elsewhere (Nelson Rees, et al., 1981; Hukku, et al., 1984, Hay, et al., 1992). Results suggest contamination frequencies of 16 to 35 percent or greater.
a. Species Verification. Species of origin can be determined for cell lines by a variety of immunological tests, by isoenzymology, and/or by cytogenetics. The indirect fluorescent antibody-staining technique is used in many laboratories to verify the species of a cell line (for details, see Hay, et al., 1992). Isozyme analyses performed on homogenates of cell lines from over 25 species have demonstrated clearly the utility for species verification by determining the mobilities of three isozyme systems–glucose-6-phosphate dehydrogenase, lactic acid dehydrogenase, and nucleoside phosphorylase. Using vertical starch gel electrophoresis, the species of origin of cell lines can be identified with a high degree of certainty. Alternatively, a standardized kit employing agarose gels and stabilized reagents may be obtained for this purpose (Innovative Chemistry, 1988).
Karyologic techniques have long been used informatively to monitor for interspecies contamination among cell lines. In many instances, the chromosomal constitutions are so dramatically different that even cursory microscopic observations are adequate. In others, for example, in comparisons among cell lines from closely related primates, careful evaluation of banded preparations is required. Cytogenetics has the advantage of detecting even very minor contaminants, on the order of 1 percent or less in some circumstances. However, it is a time-consuming procedure and interpretation may require a high degree of skill. Consult the "Atlas of Mammalian Chromosomes" (Hsu and Benirschke, 1967-1975) for examples of conventionally stained preparations from over 550 species. Detailed protocols are available elsewhere (Hay, et al., 1992).
b. Intraspecies Cross-Contamination. With the dramatic increase in numbers of cell lines being developed, especially from human tissues, the risk of intraspecies cross-contamination rises proportionately. The problem is especially acute in laboratories in which work is in progress with the many different cell lines of human and murine origin that are available today.

Methods for verifying cell line species employing enzyme mobility studies and cytogenetics have been mentioned. Using similar technology, one can also screen for intraspecies cellular cross-contamination (Hukku, et al., 1984; Hay, et al., 1992).
The application of PCR and recombinant DNA technology, cloned DNA probes, and small microsatellite loci (2-6 bp repetitive motif) to identify and quantitate allelic polymorphisms provides additional powerful means for cell line identification. These polymorphisms can be recognized as extremely useful markers, even if they are not expressed through transcription and translation to yield structural or enzymatically active proteins. For example, Edwards, et al. (1992), demonstrated the usefulness of short tandem repeat (STR) loci in differentiating humans at the DNA level. One significant advantage of STR loci over their minisatellite cousins is their small size. This allows multiplex PCR reactions to be developed in which many loci are simultaneously examined in a single reaction. The ATCC currently employs a commercially available multiplexed STR system for routinely screening new cell line accessions for authenticity, as well as validating any subsequent distribution of an authenticated cell line (Durkin and Reid, 1998; Sajantila, et al., 1992; Hay, et al., 2000).
When authenticating a new line, it is recommended that DNA be extracted from the cell line using a traditional liquid extraction method as this tends to minimize STR artifacts that may complicate allele assignment. For validation of subsequent passages of the cell line, more expedient DNA techniques may be used, which may produce ambiguous allele assignments. Generally, comparison with the authentic DNA fingerprint easily resolves these ambiguities. The STR system utilizes fluorescent labels and an automated collection device. Typical profiling results are presented in the electropherograms shown. Figure 3 represents the data generated when analyzing two cell lines derived from the same individual. Figure 4 compares the STR profiles generated from two unrelated cell lines.

NATIONAL CELL LINE RESOURCES
National cell banks have been established to provide reference lines for use by multiple investigators. Use of such cell lines assures improved research comparability both geographically and with time. Details on the more prominent, internationally-utilized cell line repositories are provided in Table 2. While there is some overlap among these organizations in terms of culture holdings, they attempt to augment each other's strengths and are in reasonably close contact with regard to problems and new methodologies. Cell line distributions range in number from about 65,000 annually (ATCC) and 16,000 holdings (CIMR), downwards.
CONCLUSIONS
In conclusion, the overall utility of any cell line resource depends on the degree of characterization of the holdings that has been performed by the originators, the banking agency, and other individuals within the scientific community. Documenting the verification of species and identity of each cell line, when possible, is considered essential. Freedom from bacterial, fungal, and mycoplasmal infection must be assured. However, from the cell banking perspective, applying all possible characterizations to every seed or master cell stock developed is neither essential nor practical. Recipients can apply specialized assays or have additional characterizations performed commercially if necessary. At ATCC, for example, screens for particular viruses have been applied when specific program support is available for such testing. Similarly, the definition of ultrastructural, tumorigenicity, and functional traits is performed given appropriate external support and adequate rationale. The central responsibility is to produce reference stocks, authenticated and well characterized for multiple purposes, and to return to those preparations over the years for development of working stocks for distribution or other specific applications. Each replacement distribution stock requires reauthentication prior to distribution to intended users.REFERENCES
DelGiudice, R.A. and Gardella, R.S. (1984). Mycoplasma infection of cell culture: Effects, incidence and detection. In "In Vitro Monograph 5: Uses and Standardization of Vertebrate Cell Cultures," pp.104-115. Tissue Culture Association, Gaithersburg, Maryland.
Durkin, A.S. and Reid, Y.A. (1998) Short Tandem Repeat loci utilized in human cell line identification. ATCC Quarterly Newsletter 18:1-7.
Edwards. A.; Hammond, H.A.; Jin, L.; Caskey, C.T.; Chakraborty, R. (1992). Genetic variation at five trimeric and tetrameric tandem repeat loci in four human population groups. Genomics 12: 241-253.
Gignac, S.M.; Brauer, S.; Hane, B.; Quentimeier, H.; Drexler, H.G. (1991). Elimination of mycoplasma from infected leukemia cell lines. Leukemia 5, 43-53.
Hay, R.J.; Caputo, J.; and Macy, M.L. (1992). "ATCC Quality Control Methods for Cell Lines," 2nd Ed. ATCC, Rockville, Maryland.
Hay, R.J.; Cleland, M.M.; Durkin, S.; and Reid, Y.A. (2000). Cell Line Preservation and Authentication. In "Animal Cell Culture," J.R.W. Masters (ed.), J.Wiley, Inc., New York, in press.
Hay, R.J.; Reid, Y.A.; Miranda, M.G. (1996) Advances in Methodologies for Metazoan Cell Line Authentication. In: R.A. Sampson, J.A. Stalpers, D vander Mei, and A.H. Stouthamer (eds.) "Culture Collections to Improve the Quality of Life" Central bureau voor Schimmelcultures, Baarn, The Netherlands, pp. 131-137.
Hay, R.J. (1998). Testing Cell Cultures for Microbial and Viral Contaminants, Cell Biology: A Laboratory Handbook 2:1) 43-62.
Hsu, T.C. and Benirchke, K. (1967-1975). "An Atlas of Mammalian Chromosomes." Springer-Verlag, New York/Heidelberg/Berlin.
Hukku, B., Halton, D.M., Mally, M., Peterson, W.D., Jr. (1984). Cell characterization by use of multiple genetic markets. In Acton, R.T., Lynn, J.D. (eds.): "Eukaryotic Cell Cultures." New York, Plenum Publishing Co., pp. 13-31.
Innovative Chemistry (1988). The Authentikit System. Handbook for Cell Authentication and Identification. 2nd ed. Marshfield, Massachusetts: 1988.
Lundin, D.J. and Lincoln, C.K. (1994). Mycoplasmal testing of cell cultures by a combination of direct culture and DNA-fluorochrome staining. In Vitro 30A, 111.
McGarrity, G.J. (1982). Detection of mycoplasmal infection of cell cultures. Adv. Cell Culture 2,9-131.
Nelson-Rees, W.A.; Daniels, D.W., and Flandermeyer, R.R. (1981). Cross-contamination of cell lines. Science 212, 446-452.
Sajantila, A.; Puomilahti, S.; Johnsson, V.; Ehnholm, C. (1992). Amplification of reproducible allele markers for amplified fragment length polymorphism analysis. Biotechniques 12(1):16-22.
Takeuchi, M.; Yoshida, T.; Satoh, M.; Kumo, H.; and Ohno, T. (1993). Survey of mycoplasmal contamination in animal cell lines collected by three cell banks in Japan. Bull. JFCC 9, 13-18.
 

FIGURE LEGENDS

Figure 1. Suggested scheme for the authentication of cell lines to be added to a cell line resource. Terminology and precise group of tests applied vary somewhat depending upon the goals of the client community.
Figure 2. Second-step PCR products from eight commonly encountered Mycoplasma species and Acholyplasma Laidlawii. Photo courtesy of Dr. Charles Buck, ATCC.
Figure 3. Cell lines from same individual. Comparison of the identical STR profiles for the EBV transformed lymphoblast line CRL-5957 and the tumor line CRL-5868 derived from the same female. Upper blue tracings represent the four STR loci D5S818, D13S317, D7S820, and D16S539. The lower black tracings represent the four STR loci vWA, TH01, TPOX, and CSF1PO, as well as the amelogenin locus used for gender identification.
Figure 4. Two unrelated cell lines. Comparison of unique STR profiles for the unrelated male cell line CRL-5963 and female cell line CRL-1855. Upper blue tracings represent the four STR loci D5S818, D13S317, D7S820, and D16S539. The lower black tracings represent the four STR loci vWA, TH01, and CSF1PO, as well as the amelogenin locus used for gender identification.Table 1
Mycoplasma Infection of Cell Lines
Reference Test Laboratory
Percent Positive

Total
McGarrity, 1982
?4.7

16,197
Del Guidice and Gardella, 1984
?11.4

34,697
Gignac, et al. 1991
?64.0

39
Takeuchi, et al., 1993
?21.0

2,332
Ludin and Lincoln, 1994
?10.2

1,000
Hay, et al., 1996
?16.0

5,362

Significance of Parvoviruses as Contaminants in Products of
Mammalian Origin
Günter Siegl
IKMI Institute for Clinical Microbiology and Immunology
Frohbergstrasse 3, CH-9001 St.Gallen, Switzerland
 In recent years the human parvovirus B19 has become notorious as an unwanted passenger in donated human blood as well as in products derived thereof. Comparable problems are frequently encountered with additional parvoviruses infecting man and/or animals. In this case, however, both source and vehicle of contamination are predominantly tissues, cells, or subcellular components purified from or produced within such substrates. Because animal parvoviruses are highly diverse, mostly species specific, and possess some exceptional biological and physicochemical properties which makes it difficult to reveal and to control their presence, the extent of the contamination problem is frequently underestimated.
The spectrum of parvoviruses either present in biological materials from the very beginning or introduced into cellular substrates in the course of production of pharmaceuticals is rather broad. Aside from parvovirus B19, which shall not be dealt with in this context, there are at least 5 types of adenovirus-associated-viruses (AAVs) which are able to infect man, his primate relatives as well as further mammalian and avian hosts. These viruses are of low tissue specificity, yet, replicate only in the presence of a co-infecting adeno-, herpes-, or papillomavirus helper. In the absence of the helpervirus, AAV can integrate its genome into the genome of an infected cell and, thereby, establish long-lasting and well masked persistence in cells, tissues, and organs. Fortunately, the disease potential of AAV seems to be low.
The great majority of parvoviruses identified as contaminants in biologicals belongs to the so called autonomously replicating parvoviruses of animals. They comprise the well known parvoviruses of rodents (rats, mice, hamsters), rabbits, pigs, cattle, carnivores (cats, dog, mink, etc.), as well as those of simian, avian (geese, duck, chicken) or even of to date unknown origin. Without control measures, these viruses are widely distributed in populations of their natural host. Infection of a susceptible host can lead either to a specific disease or to a complex syndrome comprising some few or several clinical manifestations such as diarrhoea, haemorrhagic enteritis, anaemia, panleukopenia, hepatitis, myocarditis, cerebellar ataxia, encephalopathy, gammaglobulinemia, osteolysis, abortion, stillbirth, malformations, etc. Under natural conditions, the viruses are highly species-specific. However, there are also examples in which originally species-specific parvoviruses gained access to, spread rapidly and caused acute disease in populations of animals considered to be fully refractory to such an infection. The genetic changes necessary for a switch in hostrange are minimal and, consequently, all parvoviruses must be considered to be able to cross species barriers.
Although the "autonomous" parvoviruses are independent of a co-infecting helpervirus, they nevertheless depend for productive replication on cellular helper functions provided by a susceptible cell in a distinct stage of differentiation and passing actively through the division cycle. Mitotically silent tissues or cells can be infected, yet, fail to support virus replication. Under these conditions a carrier-type persistence of the virus is frequently established which is terminated as soon as an increasing number of susceptible dividing cells (e.g. in growing and differentiating organs, regenerating tissues, or tissues and cells propagated in vitro) become available as potent substrate for viral replication. In the intact organism, parvovirus infection is also kept at bay by the immunesystem. The latter, however, is not able to eradicate infection. Rather, there is good evidence that parvoviruses frequently persist in an organism in the presence of a sometimes impressive humoral immunity. Infection then usually is reactivated as soon as the immunesystem is impaired (e.g. by immunosuppressive treatment) or absent as in the case of propagation of infected tissues or cells in vitro.
Contamination of biologicals by parvoviruses is favoured by the exceptional physicochemical characteristics of these agents. The envelopeless spherical particles measure only about 22nm in diameter. Their capsid is made up of 2-3 sequentially closely related proteins which are coded for in a single-stranded DNA genome composed of, on the average, only 5000 nucleotides. The virus particles are of outstanding resistance to elevated temperatures, dehydration, variation in pH, and treatment with organic solvents or detergents. Thus, it is usually rather difficult to remove parvoviruses from or to inactivate such contaminants in biologicals or pharmaceutical products.
A large body of evidence suggests that, under natural conditions, presence of antiviral antibody is strongly indicative for, if not equivalent to the presence of infectious parvovirus in the host organism. Data in support of this view are available for rats, mice, hamsters, cats, mink, dogs, pigs as well as for man. Consequently, seroprevalence data can be considered to yield a good estimate of the risk of an association of parvoviruses with animal tissues. In unprotected, unvaccinated populations of host animals the respective figures are usually in the range of between 40 and 80% but figures >95 % have also been reported. Tissues found to predominantly harbour the viruses are kidneys, spleen, pancreas, lung, lymphnodes, bone marrow, and testicles. In most of these tissues the concentration of viable virus is apparently low.
Many of the parvoviruses which are the subject of this paper have come to our notice as contaminants of tissue explants or primary cell cultures derived from organs and tissues of persistently infected animals. This is especially true for the rodent parvoviruses as well as for porcine parvovirus. What appears to be low level infection in some few cells within the original tissue evidently becomes amplified under in vitro conditions via the increasing number of mitotically active and, hence, virus-replication-competent cells in the cultures. Such virus replication is not necessarily accompanied by typical cytopathology. Therefore, it easily goes undetected unless appropriate measures of surveillance are instituted.
Contamination of cell cultures, specifically permanent cell lines, with parvoviruses can also result from unwanted, accidental infection in the laboratory. Biologicals like serum or enzymes used for propagation and passaging of cultures were identified repeatedly as source and vehicle of infection. However, widespread contamination within a laboratory may also be initiated by introduction of an already infected cell culture and its handling together with other cultures in the absence of stringent barrier precautions. A broad spectrum of cell lines of both human and animal origin and collected from various laboratories in Europe and the U.S. was found to be contaminated by parvoviruses of rodent, porcine, and human origin. Upon introduction into cultures outside their host cell spectrum the viruses have to adapt to growth in cells normally supporting virus replication at an extremely low level, a process that may extend over a prolonged period of time and a great number of passages. Consequently, detection of parvovirus contamination in permanent cell lines usually requires a high degree of vigilance as well as the availability of sufficiently sensitive diagnostic means.
In summary, parvoviruses pose a threat as contaminants in many ways: Unless special precautions are taken, they are easily introduced and spread readily in populations of their natural hosts. Due to their persistence in an infected organism even in the presence of measurable humoral immunity, they then can be present in organs, tissues, as well as other products derived directly from such animals. Besides, their small size, outstanding physicochemical stability and, last but not least, their tendency to establish persistent infection in cell cultures favour the spread of parvoviruses in laboratories or production units for enzymes, antibodies, antigens or vaccines. Surveillance of parvovirus contamination requires command of a well assorted set of specific and especially sensitive diagnostic instruments. Size and stability of the parvovirus particle also make it difficult to remove the viruses from or to inactivate them in biological products.

Enveloped Viruses: structure and resistance to physico-chemical treatment
James S Robertson
National Institute for Biological Standards and Control,
Potters Bar, UKVirus structure
Viruses can be classified taxonomically according to a variety of features. Virus structure, as visualised by electron microscopy, has been an important approach in this regard, but is not the only characteristic used for classification. Nowadays, the genetic structure and mode of replication contribute significantly to virus classification. Viruses exist in a variety of physical shapes, sizes and biochemical structure. A typical virus of vertebrates is roughly spherical in shape and between 20 and 200 nm in diameter (others may be pleomorphic and/or filamentous). Viruses have an inner core containing their genetic material (DNA or RNA), which is usually complexed with protein and often replicative enzymes. This inner core is contained within a protein shell and for many viruses - the enveloped viruses - this is additionally surrounded by a lipid membrane. The diagnostic test for the presence of an envelope is the effect of a lipid solvent, e.g., ether, on virus infectivity. Such envelopes are generally derived from a cellular membrane, typically the plasma membrane, during the maturation of the virus through the process of ‘budding’. Associated with the envelopes are viral-specific glycoproteins which have an important role in the recognition of, and entry into, the target host cell. In contrast, the non-enveloped viruses typically have their outer shell of protein arranged in an ordered and tightly packed pattern (e.g., icosahedral symmetry). The subunits of this protein shell also play an important role in initiation of virus infection.
 
Virus families
Viral species exist whose hosts include all types of living organisms from bacteria, through plants to higher mammalian species. As a rule, they are highly specific and their host range is quite narrow (the species barrier), even amongst viruses whose hosts are mammalian species. The narrow host range specificity of a virus derives mainly from its ability to infect only a specific cell type of (and even within) a species.
Virus infection involves recognition of highly specific receptors on the target host cell by the virus. For enveloped viruses, the receptor-binding molecule is a virus-encoded glycoprotein protruding from the lipid membrane. Thus, any physico-chemical damage to the viral envelope (and/or consequently to the viral receptor-binding molecule) will result in the loss of the ability of the virus to attach to and infect its target cell. This property of enveloped viruses has been used to advantage in inactivation and it is the relative ease with which a lipid envelope can be destroyed which makes these viruses much easier to inactivate than their robust protein-coated non-enveloped counterparts.
The following is a list of families of enveloped viruses and some of their more common members:
Arenaviridae (lymphocytic choriomeningitis virus)
Bunyaviridae (Haantan)
Coronaviridae (human coronavirus, murine hepatitis virus)
Filoviridae (Marburg, Ebola)
Flaviviridae (yellow fever, hepatitis C, bovine viral diarrhoea virus)
Orthomyxoviridae (influenza)
Paramyxoviridae (measles, mumps, canine distemper, parainfluenza, RSV)
Retroviridae (HIV, murine leukemia virus, avian leukosis virus)
Rhabdoviridae (rabies, vesicular stomatitis virus)
Togaviridae (Sindbis, rubella)
Hepadnaviridae (hepatitis B)
Herpesviridae (herpes, cytomegalovirus, Epstein-Barr virus)
Poxviridae (vaccinia, cowpox, orf)
The above list is not comprehensive but serves to illustrate the variety of viruses of vertebrates which exist. Even within a particular family, individual members can vary in their host range, disease and sensitivity to a particular inactivation process.
The latter three families of enveloped viruses have a more complex type of envelope compared with the others, which is not necessarily derived by a simple budding process from the cell plasma membrane. Nonetheless, their envelope plays a crucial role in the infectious process and, like all enveloped viruses, they can be inactivated by methods which destroy the lipid membrane.
For completeness, the following list illustrates the more common non-enveloped virus families:
Adenoviridae (adenovirus, from a variety of animal species)
Astroviridae (astrovirus, an enteric virus, from a variety of animal species)
Caliciviridae (Norwalk virus, hepatitis E)
Papovaviridae (SV40, human papilloma (wart) virus)
Parvoviridae (B19, minute virus of mice)
Picornaviridae (poliovirus, rhinovirus, hepatitis A)
Reoviridae (reovirus, rotavirus, blue tongue)
 
Virus inactivation
There is the potential for a biological medicinal product to be contaminated by an infectious agent by virtue of the starting materials involved and/or its method of manufacture. In considering methods which could be used during manufacture to inactivate a potential viral contaminant, one has to take into account the action of the inactivation process on the biological activity of the active substance itself. A variety of approaches to achieve this have been developed over the years. The most useful generic inactivation process in use today (and for many years), is solvent/detergent (S/D) treatment.
S/D has been used in the blood fractionation industry for approximately 15 years and is very effective at inactivating enveloped viruses due to the destruction of the lipid membrane by the combined action of the lipid solvent and the detergent. In addition, S/D treatment has been shown to be innocuous to the biological activity of the various products fractionated from blood. However, solvent/detergent treatment has no effect on non-enveloped viruses and it is perhaps fortunate that the human blood-borne viruses of major concern in the fractionation industry are the enveloped viruses HIV, HBV and HCV, all of which are effectively inactivated by S/D treatment. Incidents of HIV, HBV or HCV transmission by blood products have occurred, but only with products not subjected to S/D treatment
Plasma fractionators are strongly encouraged to ensure that an effective viral inactivation step is incorporated into the manufacturing process, not only for enveloped viruses but for non-enveloped viruses also. Indeed, this is also encouraged in the production of all biological medicinal products (where possible), regardless of their origin, including the production of biotech products. Such safeguards include, in addition to S/D, other physicochemical treatments such as heat (pasteurisation or dry), pH, other chemicals, gamma-irradiation and filtration. The greater inactivation capacity of these additional processes on enveloped viruses compared to non-enveloped viruses is again by virtue of the more sensitive nature of the lipid membrane than the tightly packed, ordered array of protein sub-units which form the shell of non-enveloped viruses. For processes which may not disturb the lipid membrane itself, viral inactivation may be due to denaturation of the viral glycoproteins protruding from the envelope and which are an essential part of the viral infectious cycle. These glycoproteins tend to be much more labile than the equivalent proteins forming the shell of the non-enveloped viruses.
Viral filtration methods have been developed in recent years and are an effective approach to removing potential contaminants. These methods are generally more effective against enveloped viruses than non-enveloped viruses simply because enveloped viruses tend to be larger than the average non-enveloped virus and, as a result, easier to remove by filtration. For example, the minimum diameter of an enveloped virus is of the order 40-50 nm (and most are much larger), a particle size which can be removed with reasonable effectiveness by specialised filters, whereas many of the non-enveloped viruses are as small as 20 nm (the parvoviridae) or 30 nm (the picornaviridae), which are much more difficult to remove by filtration.
In applying a specific virucidal step during the manufacture of a biological medicinal product, it is also essential that the relevant biological activity is not compromised. In general, the virucidal steps investigated have proved to be innocuous or within acceptable limits to the quality of the biological. However, one study investigating the presence of inhibitors of Factor VIII in recipients of a preparation which had been subjected to a double viral inactivation step, found altered biological activity of the FVIII. No modification of the FVIII occurred as a result of any single viral inactivation step and it was only the combination of the two steps, S/D and pasteurisation, which were detrimental. Thus, whilst it is highly recommended that manufacturers of biological medicinal products, where possible, incorporate virus inactivation or removal steps into the manufacturing process, care has to be taken that such steps do not compromise the efficacy or safety of the medicinal product.
 

FRED BROWN
PLUM ISLAND ANIMAL DISEASE CENTER
UNITED STATES DEPARTMENT OF AGRICULTURE
P.O. Box 848, GREENPORT,
NY 11944-0848, U.S.A.
Naked viruses: structure and resistance to physico-chemical
treatments

There are many naked viruses which cause disease (Table I). Forthe purpose of this presentation, however, I will focus on information
which we have obtained with the picornaviruses causing poliomyelitis,
foot-and-mouth disease and swine vesicular disease, which shows
clearly that, even within the same family, structural differences exist
which can influence their inactivation. This means that the results
obtained with one individual of a virus family may not necessarily
apply to other members of the family.
In considering reagents which inactivate viruses, it is imperative to
ensure that those used do not impair the properties of the
pharmaceutical product. Consequently there are lessons to be
learned from those who have been involved in making inactivated
vaccines. My presentation will concentrate on the application of
those lessons to the production of 'clean' pharmaceutical products.
?There are two major concerns in the preparation of inactivated
vaccines. The first is to ensure that the product is innocuous. The
second is to show that the vaccine is immunogenic, i.e. that the
relevant epitopes have not been damaged. To ensure that a virus is
non-infectious after treatment, it is essential to know that the DNA
or RNA has been inactivated. Surprisingly, to me at any rate, this
has not been done, even with some of the vaccines which have been
used extensively for many years.
Many reagents have been used to prepare inactivated vaccines
(Table II). Of these it is now widely appreciated that the time-
honoured formaldehyde is far from ideal. Not only is there concern
that the products are not innocuous, but there is increasing evidence
that important epitopes are altered by that reagent. The best known
example of the failure of formaldehyde to provide an innocuous
product is that of the Cutter incident in 1955 with the early
inactivated polio vaccine. But even as early as 1948, Moosbrugger
was questioning the procedure for the preparation of foot-and-mouth
disease vaccines. Molecular evidence in the 1980s vindicated his
doubts following outbreaks of the disease in Europe shortly after
vaccination with formaldehyde-inactivated products. However,
ensuring that the RNA of the virus is rendered non-infectious by
formaldehyde is virtually impossible because it cannot be extracted
from the virus particles following the inactivating procedure used to
prepare the vaccines. Moreover, evidence has been provided that
formaldehyde impairs the activity of important epitopes on both
poliovirus and foot-and-mouth disease virus.
?The search for alternative inactivating agents for vaccine
preparation is not a recent endeavour. In 1960 Lo Grippo showed
that ss-propiolactone inactivated several viruses. However, it is known
that this reagent reacts with roteins and there is recent evidence
that the human serum albumin, added to rabies virus before its
inactivation by ss-propiolactone, is altered sufficiently to cause
immunological problems with the vaccine.
At about the same time, Imperial ChemicalIndustries in the U.K.
described the use of aziridines for inactivating viruses. The N-acetyl
derivative was suggested by Weston Hurst to Ian Galloway at the
Animal Virus Research Institute, Pirbright, U.K., as an alternative to
formaldehyde for the inactivation of foot-and-mouth disease virus
(Fig. 1). This reagent was adopted by the Wellcome Foundation when
it entered this field in the early 1960s, although it was superseded by
the parent compound following Bahnemann's work in the 1970s.
The aziridines have been shown to inactivate a wide variety of
viruses (Table III). Moreover, from the limited amount of work which
has been done with poliovirus and foot-and-mouth disease virus, it
seems reasonable to conclude that the compounds do not affect the
antigenic or biological properties of the viral proteins. In addition,
they do not appear to affect a wide variety of proteins, including virus
neutralizing antibodies and factors in the sera used for growing tissue
culture cells (Table IV).
Unlike viruses which have been inactivated with formaldehyde, and
from which the RNA cannot be extracted, the nucleic acid can be
extracted from imine-inactivated poliovirus and foot-and-mouth
disease virus. These RNAs were non-infectious. Interestingly, their
physical integrity is retained as judged by their sedimentation rate
and electrophoretic mobility.
Further evidence that the inactivating reaction is restricted to the
RNA was obtained by comparing the rate of inactivation of poliovirus
and foot-and-mouth disease virus by the acetyl derivative. Whereas
foot-and-mouth disease virus was inactivated rapidly at 25oC, the rate
of inactivation of poliovirus was much slower (Fig. 2). We interpreted
this difference as being due to the tighter capsid of poliovirus, thus
slowing down the penetration of the imine. However, by lowering the
ionic strength of the medium from 100mM to 1mM, poliovirus is then
inactivated much more quickly (Fig. 3). Moreover, by allowing the AEI
to penetrate overnight at 2oC, the subsequent inactivation at 25oC
proceeds more rapidly (Table V).
This interpretation was confirmed by electron microscopy of the
poliovirus particles at 100mM and 1mM. In the latter conditions, the
particles have a much more open structure (Fig. 4). Intriguingly, the
particles retained their infectivity in 1mM solutions and returned to
their normal morphology when the ionic strength was increased to
100mM. Similar observations have been made with swine vesicular
disease virus, which has a structure similar to that of poliovirus.
We are now studying the effect of the imines on the RNA of the
three viruses. The sedimentation rate of the RNAs is unaltered and
several regions of the genome can be amplified by RT-PCR as
efficiently as the untreated RNAs. Nevertheless the RNAs are non-
infectious.
The imines thus appear to be ideal viral inactivating agents
because of their specificity in reacting with the nucleic acid of the
viruses without apparently affecting the activity of proteins.
 
 

Critical overview of methods currently used for viral inactivation: Limits and advantages

Philip MinorConcerns about the virological contamination of biological products originated with vaccines grown on primary cell cultures derived from animals harbouring viruses, where a large proportion of the cultures could be infected. In the mid 1980s the virological contamination of plasma products became a major issue with transmission of pathogenic agents notably with HIV, but also other blood borne viruses, and around the same time products of recombinant technology produced by large scale culture of transformed cells became available, leading to concerns about possible contamination from novel sources.
The strategy for minimising the risk of viral contamination has always been based on complementary approaches of screening, testing and virus removal or inactivation during production, each of which has short comings. In the case of removal of viral infectivity the concerns is that the virus which will challenge the production process may be more resistant than was anticipated. For example the early poliovaccines were formalin inactivated preparations based on the meticulous studies of Jonas Salk which suggested a substantial safety margin, yet the first vaccines licensed contained live poliovirus which caused disease in recipients and their contacts. The Cutter incident as it was termed was attributed to the presence of aggregates of virus in the production process which were resistant to formalin, in contrast to the laboratory studies which used monodisperse preparations. A second type of example concerns the presence of an unsuspected virus to challenge the process, such as the transmission of hepatitis A by factor VIII treated with solvent/detergent. There was no reason to think that solvent detergent treatment would be effective against HAV, which is non-enveloped, and the assumption was that the virus would not be present in the starting material.
The result of this type of occurrence has been that studies to evaluate production processes for the removal of virus infectivity have become increasingly sophisticated and wide ranging, and the approach to their interpretation has been conservative. Thus a study treating a step in a generic way is rarely acceptable, all steps needing to be examined in the context of a particular production process for a particular product.
Over the years the view has been taken that steps in a production process which may be effective can be evaluated independently: Desirable features of such steps are that they can be modelled on a laboratory scale with reasonable accuracy, that there should be reason to think that different steps have additive effects for instance if they remove infectivity by different mechanisms, and that the step itself should be robust. The term robust has not been defined, but it implies that the step can be viewed with confidence; among the features which have been proposed are that it should remove substantial amounts of virus, that it should be easily modelled, but also scalable, that it should be insensitive to changes in production parameters within set specification, and that it should be controllable at production scale, for example that adequate mixing of reagents should be demonstrable. Ideally the process step should also be effective irrespective of the strain of a particular virus used, and its presentation or state of aggregation, as well as working on all virus types. The point is clearly impossible to demonstrate, but it is interesting that viruses resistant to one type of process are often resistant to others. For instance retroviruses are sensitive to a variety of chemical and physical treatments to which parvoviruses are highly resistant.
In practice satisfactory inactivation procedures have either been introduced specifically to remove virus infectivity, in which case they must be shown to be compatible with product integrity and purity, or they are part of the production process, in which case they must be shown to be effective. Examples of production steps which have been used include extremes of pH and physical procedures such as ethanol fractionation or column separation. There may be difficulties in demonstrating the robustness of the physical steps in particular. Deliberately introduced inactivation steps include solvent/detergent treatment and gamma or UV irradiation, which must be shown not to affect the product. The steps must be evaluated for a particular process because of differences in the detail of an individual step, including the make up of the material in which the virus finds itself. While no procedure is yet regarded as so robust as to be generic in nature, some, such as solvent detergent treatment or nanofiltration, seem to be close to it.
 
Nanofiltration : Viral Elimination versus Industrial ConstraintsVirus retaining filters are an emerging concept of the last 10 decade. They have undergone considerable evolution in that time, which will continue on into the future. Manufacturers of nano- or virus-retaining (VR) filters claim that their products form efficient virus barriers and so enhance the viral safety of biopharmaceutical products. They act with a non-specific anti-viral action, providing particle size-dependent barriers. They should deal equally well with known or unknown viruses, whether infectious or not. At the same time they may exert relatively little influence on the biopharmaceutical product or its stability; but their type and use may need matching to the specific application. They can be simple to use and their efficiency may be accurately validated.
An ideal VR-filter system should remove all virus types to high efficiency. The filtration process should neither degrade nor remove the desired product. The system should have a high throughput and be easy to use. Validation of post use filter integrity should be simple and precise. In reality it can be difficult to reconcile all these demands.
VR-filtration is just one part of the concept of a Virus Barrier Strategy that has the aim of controlling the whole production process to ensure that viruses neither enter, nor survive the purification process. VR-filtration can assist in the removal of adventitious virus from raw material input streams. When used downstream of the fermentation or culture stage, there may be a desire to insert a VR-filter early into the process stream where it can immediately reduce the level of a known or suspected intrinsic viral loads such as results from retroviral contamination of fluids derived from rodent cell lines. This initial reduction in virus load will make further downstream virus clearance steps, such as ionic-exchange or size exclusion chromatography, more likely to reduce a potential virus load to below detectable limits, but will necessitate that large volumes will have to be processed through the filter. Alternatively, the "catch-all" nature of a virus-retaining filter can be best exploited late in the purification process to guard against adventitious contamination, and where process volumes are minimised.
VR-filtration technology has shown considerable progress in manufacturing reliability and has resulted in configurations that more are user-friendly, such as compact, in-line, pre-sterilised cartridge systems. There have been improvements in sanitisation technology, such as with filter assemblies which can be steamed-in-place. Dead-end cartridges now can produce viral clearances as respectable as many of the tangential-flow and hollow-fibre modules. Filter matrix and membrane considerations are at the heart of these advances. In common with other nanofilters , virus-retaining filter membranes are often of complex design. They may consist solely of intercepting voids of varying diameters, or they may have ultra-fine porous membranes bonded onto a supporting matrix. Each type of filter has its own individual characteristics, which need to be recognised.
In a depth-type VR-filter the product stream has no choice but to pass directly through the filter. The filters may therefore block up at the narrowest points in their channels. If they are constructed from microporous membranes bonded to a supporting matrix, the product can blind the surface and produce the so-called "gel layer". The resultant build up of pressure on an inadequate matrix structure may produce fistulae between adjacent pores, forming virus-permissive ducts.
The free flow of a product stream through an AV-filter of is part of a dynamically balanced equation. All macromolecules are in competition, and may assist or impede the free passage of each other. It is even possible that the load of virus added into a filter system during virus clearance validation can alter the filter characteristics to generate poor filtration performance.
In the case of tangential flow filtration, virus particles and product either flow across the membrane or pass through it. The passage of fluid across the membrane can have a scouring action at the membrane surface, and thus minimise blockage by a gel layer. However, the repeated passage of viruses across the filter increases the chance that they will seek out each virus-sized pore, even though the latter are a tiny minority in the pore size spectrum. This property generates the constant LRV phenomenon where the same proportion of virus passes through the filter, independent of the input load. The trans-membrane pressure governs the force imposed upon macromolecules to make the choice of going through or over the membrane, and must be judged and maintained accurately as the proportion of virus in the input volume rises with volume decrease.
Modern filters are constructed from a variety of materials that are designed to withstand a number of sanitisation regimes, such as heat or chemical treatments. These sterilisation protocols have to be appropriately validated and controlled. Sterile pre-tested modules, which can be just slotted into place, have a definite advantage.
Thus there are several factors to be taken into account before deciding on the most appropriate filter system for your product. We have tested all the major configurations and have settled upon different filters for different production scenarios. VR-filters can handle almost any biopharmaceutical material and we will show a selection of product recoveries from our own experiences. We shall also present some of the virus-clearance data that we have obtained. Depth filters tend to produce complete exclusion of viruses above a certain size limit; below this limit the filters tend to be very inefficient in retaining virus particles.
In tangential filtration, there appears to be a more linear relationship between virus diameter and the virus clearance. However the relative molecular size of the product can markedly influence this clearance, which can be turned to advantage, but which also counsels against using surrogate product streams when carrying out validation work
VR-filters can run into problems, such as with inappropriate protein levels, shear damage with cross-flow filters, and inconsistencies in filter performance due to product aggregation. All these need to be considered when testing and evaluating filters. After use, and before integrity testing, VR-filters should be cleaned using defined protocols. Some filters are capable of multiple-use, but our experience is that revalidation costs will outweigh any potential savings.
Between the various VR-filter types, several types of integrity tests are now used. A number of tests exploit the property of relatively low interfacial surface tension between one liquid and another. Pressures are set up to displace pores wetted with a primary liquid when a second is applied. This intrusion fluid flow can be directly related to the maximum pore diameter and thence the prospective virus retention capacity.
Alternatively filters may be challenged with a variety of solid or flexible micro particles such as colloidal gold, latex or virus particles. These latter have the advantage of their power of biological amplification. There are considerations that must be addressed when undertaking the virus challenge of filters during virus validation of VR-filters, and are usually tackled on a case-by-case basis for each biopharmaceutical and filter combination. Only the smaller viruses (<100nm) are sensible for use as calibrators for most VR-filters. Retroviruses have been employed as the chosen virus for many validation studies, because they are potential or real contaminants of several biotech products. However these viruses are not ideally suited to the investigation of the limits of VR-filter performance. Preparations are rarely pure and contain competitors for passage through filter pores. Better viruses to use are both rigid and icosahedral, such as Phi-X174 bacteriophage, picornaviruses and parvoviruses.
The quality of the virus used in virus-spiking experiment to validate filter systems is of paramount importance. The titre of virus to be used should depend on whether the virus is likely to be excluded by the filter system being evaluated (if so, use high spiking virus titre), or whether it will be only partially excluded (needing only a moderate level of virus spike). Pre-studies before embarking upon a validation study at a Contract House are always worthwhile, and bacteriophages are ideal viruses for such in-house investigations. The particulate distribution of the virus is important; virus aggregation should be minimal, so some purification methods for virus stocks should be avoided. We have also used bacteriophages in full-scale validation of VR-filter systems, on many occasions, enabling the detection of anomalous filter performance, and for our own calibration of the manufacturer’s post-use integrity test. For such assays it is important that closely defined protocols, that mirror manufacturing conditions, are employed.
The various pitfalls in the use and validation of VR-filters are becoming well known, but there needs to be an understanding of the precise interaction of the product stream with the VR-filter that is going to be used or validated. Undesired outcomes of validation experiments need to be anticipated so that alternative courses of action can be considered. This may involve the choice between the determination of the maximum possible virus clearance versus a desire to accomplish the complete filtration of an appropriate scaled-down volume.
During validation, the best moment to spike the product stream must be chosen. This can be done before prefiltration of the product, immediately before the VR-filter itself, or after a proportion of unspiked material has passed through the filter. The volumetric level of spike to be used can be as low as 0.1% but it can be as much as 5% by volume.
The reality of the use of VR-filters is that they involve the interplay of several balanced situations. For instance, the efficiency of virus removal is mediated by several factors, not all of them under operator control, or even considered in advance. The efficiency of virus removal will be virus-type dependent, involving virus morphology, virus purity, and virus stock preparation, as well as the physicochemical characteristics of a preparation of that virus type. Contingent upon the physical construction of virus filter systems and the ancillary hardware it is often not possible to recover 100% of the input product load, but recoveries in excess of 96% are often obtained. As with any filters systems, the throughput may be constrained by the product’s characteristics, and the only way round this problem is to increase the available filter surface area. This may result in having to use a number of individual filter modules in a parallel configuration in order to obtain the requisite surface area.
After filtration, each module will require individual integrity testing. Integrity testing of filters, and evaluating the results can be difficult. Validation at small scale can present additional problems, often related to the types of pumps which can be used at small scale, and the difficulty of matching pressure and flow characteristics with those which are known to obtain at full scale.
A thorough knowledge of the product to be filtered, of the filter characteristics, and of the necessary validation and control processes, should be coupled with a realistic expectation of the achievable power of antiviral filters. That said, VR-filters will provide an impressive increase in confidence about the viral safety of biopharmaceuticals.
Considering the human use, the initial viral risk linked to a raw material depends mainly on (1) the biological origin of this raw material (cell line, nature of the tissus or fluids, etc....), (2) the species barrier and (3) the route of administration in human.
The human risk may be reduced at three different levels of the bio-pharmaceutical preparation.
A first viral reduction may be obtained from raw material screening: human plasma screening for the absence of main human pathogens, selection of healthy animals through rigourous veterinary controls, and selection of cell lines and/or cell supernatant by using in vitro and in vivo testsing.
A second reduction is usually associated to the manufacturing process since it can be able to eliminate and/or inactivate viruses. In order to evaluate this process capacity, spiking experiments need to be carried out with the more significant manufacturing steps.
My presentation will be focused on this aspect.
A third reduction may result from virological controls applied to the final product or to the concentrated bulk.
 
If we exclude gene therapy products, the viral safety strategy consists in optimizing the two first reduction ways, i.e. the raw material screening and reduction through the manufacturing process. In a large number of cases, the cumulated reduction effect of the two ways is sufficient to reach an extremely-low residual viral risk which is finally acceptable for human use.
When a too high residual viral risk is obtained (resulting usually from an non-optimal reduction capacity of the manufacturing process), controls on final product have to be introduced.
Gene therapy vectors represent a particular case for two reasons: firstly manufacturing processes are not still optimized for viral contaminant reduction (early stage of development), and secondly the product itself is usually a non-replicative virus and this characteristic needs to be controled.
Then, the viral safety of vectors is mainly supported by the raw material screening (cell banks) and by controls on the final product.
The ability of a manufacturing process to reduce a viral load has to be evaluated through rigorous strategy. In a first time, a theoretical analysis will allow to identify all the putative reduction factors of the process. Some factors are able to kill (inactivate) viruses, as high temperature, acidic / basic pH, SD treatment, UV or b irradiation, etc....
Others factors, or more precisely treatments, create a partitioning (at least two independent phases or fractions) and then a putative viral elimination if viruses and the biological active component don’t segregate in the same fraction. Examples of partitioning are protein precipitation with solvents, filtration, chromatography, ultracentrifugation, etc...
In a large number of cases, manufacturing steps are more complexe, and two, three, or more reduction parameters are simultaneously effective.
From a such theoretical analysis, two values will be deduced for each step:
- putative reduction factors for viruses,
- complexity of the spiking experiment design.
Considering these data, manufacturing steps will be selected for experimental evaluation.

Problems, Hurdles and Triumphs with Live Virus Vaccines. Changing Public Policy That May Affect Vaccine Research
Maurice R. Hilleman
Merck Institute
West Point, PA 19486Modern era vaccinology that began about 1950 was fueled by Enders' renaissance of cell culture for propagating viruses for vaccines.
1.)Its first example was that of both the live and killed poliovirus vaccines of the early 1960s. Special problems relating to safety and immunizing potency were encountered. Killed vaccine was encumbered initially by incomplete inactivation of poliovirus, uncertainty of adequate potency of individual vaccine lots, and presence of a new indigenous polyomavirus contaminant (SV40) that derived from monkey kidney cells and that was incompletely inactivated in the vaccine. The SV40 problem was resolved by substituting a monkey species that was free of SV40 virus to provide renal tissue for cell culture. The live poliovirus vaccine, also troubled initially by SV40 contamination, still retains very low residual neurotropism for man but remains the paradigm for world eradication of the polioviruses. Interference between individual polioviruses in the trivalent vaccine and coincidental natural infections in vaccine recipients may compromise vaccine efficacy.• • •