Skip Navigation
Johns Hopkins Bloomberg School of Public HealthCAAT

Technical Report No. 4

Cell Culture Systems And In Vitro Toxicity Testing

A Report of the CAAT Technical Workshop of June 13-15 1990


The workshop on Cell Culture Systems and In Vitro Toxicity Testing was held in conjunction with the 41st Annual Meeting of the Tissue Culture Association in Houston, Texas. This workshop was the second in the series of scientific workshops, organized by the Johns Hopkins Center for Alternatives to Animal Testing (CAAT), to address technical problems associated with the development and use of in vitro toxicity testing systems and focused on cell, tissue, and organ culture procedures used in in vitro toxicology.

The objective of the workshop was to identify the biological characteristics of the cultured cell and its environment that influence endpoint measurements in toxicity testing systems. The biological component of any cultured cell system is the key to the design of in vitro toxicity systems. Critical elements include: the cell themselves, the ways in which they are isolated and cultured, and the rationale for selecting one cell system over another for testing purposes. This report: (1) describes a paradigm of toxicity testing in which in vitro test systems can take a natural place in the toxicology arena; (2) surveys cell culture systems; (3) deals with perhaps one of the most important components of tissue culture, the medium and the substrate essential for cell growth; and (4) describes the requirements for standardization of tissue culture methodology.

CAAT has long recognized the need and responsibility to validate in vitro testing methodologies in order to move from current research involving the use of animals to modern research practices that could reduce, refine, and replace certain aspects of in vivo testing protocols. It is hoped that this technical report will provide guidance to investigators participating in both test development and test validation of new in vitro methodologies.

The technical report originated through a series of questions generated by the workshop co-chairpersons, Dr. June Bradlaw and Dr. Oliver Flint. Workshop presenters with expertise in cell culture and in vitro toxicity testing strategies were assigned specific questions and were asked to prepare working reports for discussion. It was clear that the questions generated fell under the headings of four discussion 'topics': (1) strategies for using in vitro tests; (2) the cultured cell, including (2a) cell lines, primary cultures and organ cultures, and (2b) human cells: use, advantages and limitations; (3) the in vitro environment; and (4) standardization in tissue culture: cell lines, primary cultures, media, sera, culture ware and terminology. Following the presentations by the invited participants, both the presenters and the observers attending the workshop spent several hours in working groups discussing one of the above topics. Each chair of a discussion group was instrumental in preparing the summary report presented in this technical report.


This series of workshops sponsored by CAAT is intended to identify technical problems associated with in vitro toxicity testing systems. Technical Report No. 1 (May 17-18, 1989) focused on certain analytical aspects of these systems. This report also identified that characterization of biological components of an in vitro system was essential to the successful development of new testing methodologies and their practical application and projected that this subject would be the subject of a separate workshop. The tissue culture environment, including the cells themselves, the substratum on which they grow and the medium from which they gain nutrition is usually given the least adequate description in any report of  in vitro toxicity. Yet, this environment is the single largest source of factors influencing the outcome of any experiment. In order to broaden the scope of this subject, the workshop was held in conjunction with the Tissue Culture Association Annual Meeting to avail itself of the scientific expertise of the association and its members in classical tissue and cell culture methodology. This collaboration allowed the infusion of fresh ideas and an evaluation of the basic concepts of in vitro toxicity strategies.


Jack Lipman1, Oliver Flint2, June Bradlaw3, and John Frazier4

1Investigative Toxicology, Hoffman-La Roche, Inc., 340 Kingsland Street, Nutley, NJ 07110-1199; 2Investigative Toxicology, Bristol-Myers Squibb, P.O. Box 4755, Syracuse, NY 13221; 3Division of Toxicological Studies, HFF-162, Food and Drug Administration, 200 C Street, S.W., Washington, DC 20204; and 4The Johns Hopkins Center for Alternatives to Animal Testing, 615 North Wolfe Street, Room 7033, Baltimore, MD 21205


One of the key issues in toxicology is to formulate a logical and flexible approach for evaluating test agent toxicity. The aim is to establish human risk before human exposure. Can in vitro tests make a contribution to this process?

Each test, in vivo or in vitro, can potentially supply some of the information we need to determine the profile of human risk. At present, in vitro tests only contribute in a small way to this 'risk assessment' process. However, we should not dismiss the possibility that in vitro data will eventually make a significant contribution to, and perhaps improve, our determination of human risk. We should, therefore, plan to evaluate the in vitro testing methodologies available today, and those yet to be described, by establishing an appropriate strategy for toxicity testing that involves the collection of both in vitro and in vivo data. The strategy described below is intended for the broad spectrum of chemical industry activity, including the pharmaceutical, household products, and cosmetics industries. This is only a framework for which application, in part or in full, will depend on the type of test agent and its intended use.

Our central strategy is to determine whether or not the test agent is safe for human exposure under the conditions of its proposed usage. When a test agent (for example, a drug) is found to be effective in a target area of interest, important aspects to its further development are: (1) to address its potential toxicity in the pre-project selection process, the discovery phase; (2) to investigate the mechanism of any test agent-induced injury; and (3) to aid in the modification of the parent compound structure or test agent formulation by toxicity screening. This approach, when used in conjunction with pharmacokinetic and metabolism studies (and Good Laboratory Practice) will allow the development of the best candidate with the lowest possible potential toxicity. Ideally, this process must occur in a rapid, reliable, and cost-effective manner. Moreover, it must use a logical and scientifically sound decision-making process for determining toxicity. Our objective should be to achieve some or all of the above goals by including in vitro tests to the process of test agent evaluation, and thus reduce the number of animals required for the testing program.

In vitro approaches have the potential to provide information about the function of specific organs and how they can be adversely affected by a toxic agent. Additionally, it is possible to study the effects of agents on different cell types from multiple species, including human. Currently it is estimated that at least 37 different assay systems exist for alternatives to the eye irritation tests, covering such areas as cytotoxicity, release of inflammatory mediators, and impairment of cellular functions (Goldberg, 1987). Clearly, a logical and sound scientific approach must be used in choosing the appropriate assay system for predicting toxicity.

A single universal assay does not exist that is reliable and accurate for all agents or capable of functioning as an alternative to the animal model. We must use batteries of tests capable of evaluating the impact of an agent on a wide variety of potential areas of toxicological concerns, including for example; phototoxicity, neurotoxicity, hepatoxicity, and renal toxicity.

The strategy must be based on a continuous decision-making process that will evaluate all relevant information about the agent. The caveat that demands our constant attention is the extrapolation from specific in vitro assay endpoints (that describe a specific toxicological endpoint) to the prediction of the potential toxicological impact in humans. Since in vitro toxicology is in its early stage of development, we must have a strategy that will allow us to learn from both our successes and our failures. This strategy will serve as a tool for refining our ability to predict toxicity.

The proposed strategy which we will call "The Toxicological Decision Paradigm", attempts to incorporate all of the following considerations:

  1. Each compound/product or class of compounds/products is treated uniquely.
  2. Each step in the decision-making process is documented in order to draw conclusions for use in a prospective and retrospective manner to further refine the process.
  3. A continuous dialogue with the product development team allows improvement of the molecular structure of the compound or product formulation.
  4. The in vitro screening process allows consideration of mechanistic questions and integrates this testing, wherever necessary or possible, with limited animal testing.
  5. This strategy is applicable to different industrial settings (e.g. pharmaceuticals, consumer products, cosmetics).

The Toxicological Decision Paradigm (Table I) is a five-part approach designed to integrate in vitro testing into the overall process of data collection, interpretation, and report generation aimed at determining human risk. It can only accomplish this by requiring written documentation explaining why each step (phase or test) was chosen and interpreted, and how the next step in the decision process will be chosen. This is an idealized approach to the development of new products. There is no doubt in our mind that a substantial evolution of methods, both in vivo and in vitro, is required before this ideal is widely accepted. Indeed, the ideal we achieve may be very different from that described here because methods will evolve and change. In describing this decision-free approach we do not imply a lack of flexibility. Furthermore, though we describe a linear stepwise path from in vitro to in vivo animal studies to human clinical trials, we expect much greater flexibility in practice. For example, it is inevitable that clinical findings may be reconsidered by returning to animal and in vitro studies. The results may affect the discovery process and lead to less toxic but equally effective compounds. In another example, the animal phase of testing may be eliminated altogether, as is currently the case with the Limulus amebocyte tests (Prior, 1990) for endotoxins, and as may be the case in the future with in vitro methods for identifying ocular toxins (Flint, 1990). Flexibility and adaptability to the problem must be the key.

Table I: Toxicology Decision Paradigm

Phase I: The compound and its family
  • Name, structure, formulation, or reformulation
  • Purpose (intended use)
  • Length of treatment (possible exposure)
  • Known or predicted biological activity
  • Literature search (published and internal reports)
  • Structure activity relationships (if any)
  • Summary:
    Anticipated dose
    target organs (if drug)
    toxic dose
  • Conclusion and recommendations

Phase II: In vitro tests

General considerations:

  • Test representative cell types
  • Test cells alone and cocultured with hepatocytes
  • Interspecies comparisons to include human cells
  • Compare general and organ specific endpoints

Phase IIA: Initial toxicity profile

Assay (examples only)

  • Neutral red bioassay
  • Rhodamine 123 uptake
  • LDH leakage
  • Total protein


  • Lysosomal integrity
  • Mitochondrial function
  • Cell membrane integrity
  • Cytotoxicity

Phase IIB: Organ-specific toxicity testing

Examples (will involve mostly primary cultures):

  • Cardiotoxicity
  • Dermal toxicity
  • Genotoxicity
  • Hematotoxicity
  • Hepatotoxicity
  • Immunotoxicity
  • Renal toxicity
  • Neurotoxicity
  • Ocular toxicity
  • Phototoxicity
  • Respiratory toxicity
  • Reproductive toxicity

Phase III: Limited animal testing

  • Optimization of study design (e.g., piggy backing)
  • Reducing animal numbers and dose size by use of biotelemetry
  • Drug metabolism and pharmacokinetics to determine relevance of in vitro to animal and human studies
  • Conclusion and recommendation on toxicity

Phase IV: Mechanistic studies

  • In vitro and in vivo. Some in vitro mechanistic studies in phase IIB also

Phase V: Recommendations and report

Finally, during validation of an in vitro test method and when it is applied to test agent evaluation, some limited animal testing will be required to confirm the in vitro test data. This will involve establishing that the target organ is affected in vivo as the in vitro data suggest. Pharmacokinetic data obtained in vivo will also be essential to determine that the target organ is exposed to the test agent at concentrations which were determined to be toxic in vitro. In vivo and in vitro data are, by definition, interrelated.

The Toxicology Decision Paradigm

Phase I: The Compound

A detailed review of all published literature and internal reports must be carefully studied for the test agent and related compounds or formulations. In addition, structure-activity relationship SAR data will be used if available. This information will be the basis for determining anticipated toxicity, target organs, and predicted toxic dose. If there are no available data from the published literature (we are dealing with an entirely new agent), then initial planning will be based on internal reports of purpose, pharmacokinetics, and metabolism studies when available. Clearly, this is the most important step, since careful review and evaluation of all data, in conjunction with available analytical skills, will be the key to the design and interpretation of all other phases.

Phase II: In Vitro Screening

Two phases of in vitro toxicity testing might be planned. First is a comparison of the effects of test agent on a variety of well-defined cell lines. All tests should be done with representative cell types (e.g., epithelial, endothelial, fibroblastic) as determined by Phase I. The tests should measure the direct effects of the agent on the cells and possibly during cocultivation with hepatocytes (to determine the effects of potential metabolites). These preliminary studies are intended to define a profile of test agent toxicity. Secondly, more detailed studies using primary cultures address the fundamental problems of potential human toxicity and toxic mechanism. Ideally, we would only study human cells, but rats and other species are routinely used in all drug safety assessment studies. Thus, interspecies comparisons are essential and in vitro toxicity studies comparing human with animal cells may be of central importance in determining the relevance of animal data for human toxicity. Appropriate interspecies comparisons of cellular toxicity are possible with primary cultures prepared from a variety of organs and provide a unique and, so far, under exploited potential application of tissue culture methodology. Liver and renal cultures, for example, are two specialized tests that represent common targets of toxicity and should be considered (see Table I for additional areas that should be evaluated) and could be done routinely in addition to other tests.

It is likely that, in the future, contract laboratories will perform the more basic primary screening with cell lines, thereby enabling the contracting laboratory to further refine specialized toxicity testing. It is essential that all results on Phases I and II be analyzed and conclusions drawn prior to undertaking the more specific tasks of Phases III and IV. Once the general toxicity of a compound is established, the screening laboratory must interact with chemistry, pharmacology, and metabolism groups to attempt to reduce any measureable toxicity by customizing the molecular structure of the compound or formulation of the product. This method of generating potentially nontoxic analogues by a rapid in vitro screening process must lead to an improvement in the rate at which drugs/formulations with minimal toxicity reach the market. An example is the selection of nonteratogenic antifungal agents by using an in vitro reproductive toxicity test (Flint and Boyle, 1990).

Phase III: Limited Animal Testing

After completion of the prior stages, this phase should only be used for those agents that are important lead candidates with anticipated or completely unknown toxicity. If the in vitro tests suggest that the compound or formulation is extremely toxic, is should first be evaluated by structure or formulation modification studies, and if toxicity remains, abandonment should be considered. There are exceptions to this approach, including the chemotherapeutic agents, where some toxicity may be acceptable. In vitro and in vivo studies with potential inhibitors may reduce these side effects in otherwise therapeutically effective compounds. The use of biotelemetry (implantation of radiotransmitters for measurement of physiological parameters in unrestrained animals) offers the potential to limit the number of animals (2-3/group) and permits lower test doses to determine toxicity. Biotelemetry is a part of a continuing effort by industrial toxicologists and regulatory agencies to optimize animal use for toxicity studies. Another approach has been to piggy back immunotoxicity or genetic toxicity studies onto more routine chronic and acute animal studies, rather than to use additional animals in separate studies.

Phase IV: Mechanistic Studies

Studies to evaluate mechanism should be done to augment Phase II for those compounds with unique and unexpected toxicity. In addition, these studies will serve as additional support for toxicities observed in later animal or clinical studies.

Phase V: Conclusion

As each test material is evaluated, all information is placed into a report. In addition, each step of the decision-making process will be documented to indicate the scientific rationale for its implementation. All of the quantitative data will be carefully reviewed and a conclusion on the anticipated toxicity will be reported. Specific recommendations about the compound or formulation should be incorporated to allow the next decision in the preclinical development process (to either abandon or continue testing the compound) to be made. It is not the purpose of the 'Toxicological Decision Paradigm' to make the final choice on the disposition of the test material, but to provide as much information as possible in a complete, rational, and sound manner as possible.

In conclusion, we are at the point in toxicology where we must use all of our skills and experience as scientists to advance our field. We must use a decision-oriented model involving in vitro toxicology.


  • Flint, O.P. (1990). In vitro alternatives to ocular toxicity testing: A report of a meeting organized by the Industrial In Vitro Toxicology Group. In Vitro Toxicology 3:281-292.
  • Flint, O.P. and Boyle F.T. (1990) Structure-teratogenicity relationships among antifungal triazoles. Handbook of Experimental Pharmacology, Vol. 96, Chemotherapy of Fungal Diseases (ed. J.F. Ryley). pp. 231-249. Verlag, Berlin, Heidelberg, New York.
  • Goldberg, A.M. (ed.) (1987). Alternative Methods in Toxicology Vol. 4. (pp. 31-34). Mary Ann Liebert, Inc., New York.
  • Prior, P.B. (1990). Clinical Applications of the Limulus Amebocyte Lysate Test. CRC Press, Boca Raton, Florida.

Section II: The Cultured Cell

Charlene McQueen1, Carol Green2, Daniel Acosta3, John Harbell4, James Klaunig5, James Resau6, Ellen Borenfreund7, Rajendra Mehta8, Robert Van Buskirk9, Bjorn Ekwall10, Oliver Flint11, June Bradlaw12, and John Frazier13

1Department of Pharmacology and Toxicology, College of Pharmacy, University of Arizona Health Sciences Center, Tuscon, AZ 85721; 2SRI International, Life Sciences Division 333 Ravenswood Avenue, Menlo Park, CA 94025; 3University of Texas, College of Pharmacy, Austin, TX 78712; 4Microbiological Associates Inc., Bethesda Laboratories, 5221 River Road, Bethesda, MD 20816; 5Medical College of Ohio, 3000 Arlington Avenue, Toledo, OH 43599; 6University of Maryland, 22 South Green Street, Baltimore, MD 21201; 7Rockefeller University, 1230 York Avenue, New York, NY 10021; 8IIT Research Institute, 10 West 35th Street, Chicago, IL 60616; 9State University of New York at Binghampton, Department of Biological Sciences, Binghampton, NY 13901; 10Department of Toxicology, University of Uppsala, Biomedical Center, Box 594, S-751 24 Uppsala, Sweden; 11Investigative Toxicology, Bristol-Myers Squibb, P.O. Box 4755, Syracuse, NY 13221; 12Division of Toxicological Studies, HFF-162, Food and Drug Administration, 200 C Street, S.W., Washington, DC 20204; and 13The Johns Hopkins Center for Alternatives to Animal Testing, 615 North Wolfe Street, Room 7033, Baltimore, MD 21205


Cells can be maintained as primary cultures of single cells, tissue slices or tissue fragments as well as in a continuous cell line. Regardless of whether primary cultures or cell lines are used, it is necessary that the origin of the cells be known. This includes species, strain, sex, age, and tissue. It is well recognized that variations in biotransformation and tissue-specific function can occur as a result of these variables. For example, differences in the pattern of amphetamine metabolites were observed among hepatocytes isolated from rat, rabbit, dog, squirrel monkey or human liver (Green et al, 1986). Differences in conjugation of 2-acetylaminofluorene have been noted in monolayer cultures of hepatocytes from male and female rats (McQueen et al, 1986).

A cell line is defined as arising "from a primary culture at the time of the first successful subculture. The culture consists of lineages of cells present in primary culture" (Schaeffer, 1990). Cell line characteristics are summarized in Table II. A system using a cell line should include the following information:

  1. origin of the cell line;
  2. population doubling;
  3. culture conditions including subculturing methodology;
  4. karyotype or DNA fingerprinting;
  5. morphologic and biochemical characteristics; and
  6. procedures for cryopreservation and thawing.

Table II: Some Characteristics of Cell Lines

  1. Cell lines are originally derived from a primary culture that has been successfully subcultured.
  2. Most cell lines can be stored at ultra-low temperatures.
  3. Cell lines may either have a finite life span or be a continuous cell line capable of an unlimited number of population doublings.
  4. Differentiated functions may be altered in cell lines depending on culture conditions.
  5. The metabolic capacity of the cells may decrease depending on culture conditions.

A primary culture is defined as one "started from cells, tissues or organs taken directly from the animal" (Schaeffer, 1990). Some ideal characteristics of primary cultures are summarized in Table III. In using primary cultures, some general areas that need to be described in reports include:

  1. pretreatment of animals;
  2. method of cell isolation;
  3. culture conditions; and
  4. cell type.

In the category of culture conditions, it is necessary to determine such parameters as suspension or monolayer cultures, medium components, serum source, components of matrix used as a substrate (if any), antibiotics (if any), atmosphere and temperature. The effect of these components is described in Section III. The cell type should be characterized by features such as morphology, cell function, cell-specific markers,and biochemical and immunologic features. For primary cultures of tissue slices or organ fragments, the thickness, culture method (stationary or fixed), and spatial orientation are important.

Table III: Characteristics of Primary Cultures

  1. Cultures must be isolated directly from the donor.
  2. The cultures generally have a finite life span.
  3. Cell morphology is similar to the intact tissue.
  4. Initially, metabolic capacity is similar to intact tissue but changes occur with increasing age of the cultures.
  5. The cells retain many tissue-specific functions. These can alter as the cultures age.

Each cell system has distinct advantages and disadvantages. In some, although not all cases, primary cultures mimic the intact tissue better than cell lines. One example of this phenomenon is the capacity of rat hepatocytes to activate carcinogens from a number of chemical classes to DNA-damaging products (Williams et al, 1989). In contrast, the capacity of adult rat liver (ARL) cell lines to activate these compounds is diminished and dependent upon the cell line tested (Tong et al, 1984). Primary cultures may contain specific receptors and under proper culture conditions may become organotypic (for organoids). Additionally, since primary organ or slice cultures do not require cell dissociation, they maintain the three-dimensional structure of the tissue and cell-to-cell contact. Limitations of primary cultures include the necessity to isolate tissue for each experiment, donor to donor variation, the longevity of the cultures, and loss of biotransformation and tissue-specific functions with time in culture. Additional limitations of organ and slice cultures include the duration of tissue viability and variation due to thickness of the tissue fragment.

Because the cells can be prepared in large numbers and cryopreserved, cell lines have a greater reproducibility and ease of use. A donor is only used at the initiation of the cell line and not at each experiment. Often, cells can be maintained under more defined culture conditions than primary cultures. Depending upon the cell type and culture conditions, there are cell lines that maintain tissue-specific function. The presence of dividing cells may be advantageous for some applications. One important limitation of cell lines is that the phenotype and genotype of the cultured cell changes with time in culture; loss of tissue-specific function, biotransformation capacity, and alterations in karyotype, morphology, and biochemical properties can occur. More recently, the use of immortalized (occasionally human) cell lines derived from primary cultures transfected with viral DNA (Hurko et al, 1986) or from mice made transgenic with the SV 40 large T antigen (Jat et al., 1991) had led to the discovery of conditionally immortal cell lines with apparently stable phenotypes and genotypes retaining much of the differentiated function of the parent cell.

Once the cell systems have been delineated, toxicity evaluations can be established. The choice of the most appropriate cell culture system will then depend upon the adverse effect being evaluated. When using any cell culture system, experimental details must be clearly defined, described, and documented. The availability of such information facilities interlaboratory comparisons and identification of potential toxicity.

Cell Lines

One of the most obvious advantages of using established cell lines is their ready availability, either from a tissue culture bank such as the ATCC (Hayet al, 1988) or obtained from another laboratory. Medium requirements and details concerning propagation are usually already known. Furthermore, for inter- as we as intralaboratory reproducibility of experiments, the use of a given cell line gives an assurance of working with the same starting material, though clonal selection with passage of a cell line may lead to deviation of phenotype and genotype in different laboratories. Freshly isolated cell preparations, on the other hand, can vary only in individual laboratories, but more importantly, in other laboratories which attempt to reproduce these experiments. Close attention to experimental protocol and Good Laboratory Practice (GLP) should minimize such differences in primary cell cultures. Readily available cell lines will also provide a convenient solution for those investigators who are not equipped for the time-consuming and often complex isolation procedures required to establish primary cultures.

Use of a tissue culture cell line originally isolated from a particular organ or tissue is not necessarily an indication that these cells are actually representative of the tissue of origin. This is particularly true for cultures that have been maintained and propagated in vitro for extended time periods. Fibroblast cell lines maintained as long term cultures are usually characteristic only of species of origin, rather than tissue of origin. Cytotoxicity studies in vitro indicate that sensitivity to xenobiotics is often, but not always, quite similar regardless of tissue or species of origin (Borenfreund and Borrero, 1984). Epithelial cells, on the other hand, are frequently able to maintain specialized functions such as organ-specific secretions or production of plasma proteins (Knowles et al, 1980), hormones (McAllister and Hornsby, 1987; Sato et al, 1970), lipids, enzymes (Thompson et al, 1966), cytochrome P-450 dependent mixed function oxygenases (Babich et al, 1988) and growth factors (Sirbasku, 1978). Such characteristics can sometimes be induced in vitro (Thompson et al, 1966; Sirbasku, 1978) if they are not expressed by the cell line in normal maintenance culture conditions. For a general review see McKeehan et al (1990), Grisham and Smite (1984), and Sato (1981). Some cell lines may require special incubation conditions. For example those derived from fish require low temperature (28-34° C) reflecting those characteristics of their in vivo environment. Functional cell lines can at times be best isolated from tumor tissues, as, for example, neurons from mouse neuroblastomas (Augusti-Tocco and Sato, 1969) and hepatocytes from hepatomas (Knowles et al, 1980).

The actual application of in vitro technology using a particular cell line will depend on the problem to be investigated. There has been great interest in this area of research as attested by the large number of publications. This is evidenced by journals such as In Vitro Toxicology (Mary Ann Liebert), Toxicology In Vitro (BIBRA), Alternatives to Laboratory Animals (ATLA published by FRAME, Nottingham, England), the Special Issue on In Vitro Toxicology to be found in Molecular Toxicology (1988, Vol. 1, pp.281-603), and the series of publications emerging from the Center for Alternatives to Animal Testing at Johns Hopkins University under the general title of Alternative Methods in Toxicology (Volumes 1 to 7 to date; Alan Goldberg, editor; Mary Ann Liebert, publisher). Unfortunately, indiscriminate use of certain assay procedures without prior standardization or understanding of the particular endpoints used can easily lead to 'muddled' data. This is especially true if original procedures are modified without careful examination to determine whether such modifications significantly change the validity of the method.

The following important considerations should be taken into account before undertaking any toxicity study with a cell line:

  1. Many cell lines have been available for extended time periods, having undergone many in vitro passages. Before use, cultures should be carefully examined for homogeneity of cell type, and if necessary, must be re-cloned. Cell lines are usually more resistant to toxic insult than freshly isolated short-term cultures.
  2. Investigators should be aware of population doubling time since this may play a role in sensitivity to a toxic agent (i.e., cells may have to undergo mitosis to reflect toxicity). Replication time may also affect the number of cells seeded in order to obtain a required density.
  3. Type of cell and culture conditions used at the time of exposure to the test agent should best serve the particular assay at hand (i.e., suspension versus adherent cultures; sparsely seeded versus dense cultures). For certain experimental procedures, sparsely seeded cultures maintained over several days might be most suitable. Other experiments might be more informative when cells are in exponential growth or are confluent. Types of seeding must be kept constant for all experiments within a given study.
  4. Information concerning normal or tumorigenic characteristics should be noted. This may also be reflected in chromosome number and ability to replicate in medium with low serum concentration. (Tumor cells can grow more readily in low serum medium.)
  5. Some cell lines maintain or secret tissue-specific proteins and enzymes which can be used for examination as toxicity markers. Specific biomarkers should be noted, such as special receptor sites or responses to specific growth factors.
  6. If defined medium for a particular line is not available, serum concentrations should be kept as low as possible without affecting cell viability. Serum can adsorb xenobiotics, and thus reduce the concentration of test agent available (Babich and Borenfreund, 1987; Borenfreund and Puerner, 1986).
  7. Temperature effects may yield important information which can be obtained from cultures adapted to lower temperature or cell lines. For example, it may be more appropriate to expose keratinocyte cell lines, such as the XB-2 (Rheinwald and Green, 1975; Duffy and Flint, 1987), to the test agent at the lower temperatures found at the skin surface.
  8. For assays involving test agents which are in need of metabolic activation to convert to a toxic intermediate, cell lines that are known to contain P-450 microsomal enzymes, or can be induced to make them, should be used (Babich et al, 1988). Alternatively, preparations of S-9 mixtures which contain the mixed functions oxygenase, can be added to the incubation medium (Borenfreund and Puerner, 1987).
  9. Cell lines must be clean of contaminating mycoplasma, bacteria or viruses and sterile techniques must be strictly adhered to.
  10. Frequent microscopic examination and monitoring of stock cultures is mandatory to detect changes in growth patterns and morphology.
  11. It is essential that all experiments are repeated several times, using the same environmental conditions to determine variability.
  12. Immortalized epithelial cells (nontumorigenic) derived from some organs, such as the lung and liver, are now available (due to transfection of viral components) for long term studies.
  13. Stable lymphocytic lines (mostly of neoplastic origin) which maintain the functions of macrophages and other cells of the hematopoietic system are available. Similar stable cell lines maintaining the function of cells from other organ systems are not easily available, with some exceptions (for example, the PC12 neuroblastoma cell line).

Primary Cultures: Advantages and Limitations

As described in the introduction, a primary culture is defined as one 'started from cells, tissues or organs taken directly from organisms' (Schaefer, 1990). The major advantages of primary cultures are the retention of: (1) the capacity for biotransformation; and (2) tissue-specific functions.

Many compounds require biotransformation to exert a toxic effect. The importance of such biotransformation capabilities in a cell culture system has been evident in short-term tests to evaluate carcinogenicity. In order to adequately evaluate compounds in bacterial mutagenicity tests such as the Ames assay, an exogenous source of metabolic enzymes, generally an Aroclor-induced rat liver S-9 preparation, must be provided (Maron and Ames, 1983). While mammalian cell lines may have some intrinsic capacity for biotransformation, it will vary. An assessment of several adult rat liver (ARL) cell lines showed qualitative and quantitative differences in the induction of mutants by compounds that require bioactivation (Tong et al, 1984). The genotoxic compounds 2-aminofluorene and 2-acetylaminofluorene were mutagenic for ARL 18 but not ARL 14. In contrast, in primary culture of rat hepatocytes, which have a greater capacity for biotransformation than ARL cell lines, both compounds induced DNA repair (Williams et al, 1989).

In many cases, the metabolic profile generated by a primary cell culture such as hepatocytes has greater similarity to in vivo than does the pattern seen with subcellular fractions used as an exogenous source for biotransformation. Hepatocytes possess the ability for conjugation reactions (McQueen and Williams, 1988) that may be absent in subcellular fractions. In fact, metabolic studies in hepatocytes have resulted in the identification of previously undetected in vivo products. Studies in hepatocytes lead to the discovery of ketoethinimate as a metabolite of ethinimate (Billings et al, 1977). Subsequent investigation revealed this metabolite was also formed in vivo (McQueen and Williams, 1988).

The second advantage of primary cultures is the retention of tissue specific functions. For example, primary cultures of rat myocardial cells consisting of synchronously beating cells can be prepared (Acosta and Ramos, 1984). When these cultures were exposed to tricyclic antidepressants that are cardiotoxic, both arrhythmias and cessation of beating were observed.

One limitation of primary cultures is the necessity to isolate cells for each experiment. Procedures to isolate cells require the disruption of the tissue, often with proteolytic enzymes. This may result in the loss or damage of specific membrane receptors, damage to the integrity of the membrane, and loss of cellular products. During the interval necessary to establish monolayer cultures, damage is often repaired. An example of this phenomenon is the loss of Ca++ during the isolation of hepatocytes. There is a 60% decrease in the concentration of Ca++ in freshly isolated cells; however, following 24 hours in culture, the level is restored to that of the intact liver (Paine, 1990).

Primary cultures have a limited life span and changes in metabolism and tissue specific functions will occur with time in culture. Many alterations can be postponed depending upon the culture conditions. It has been shown using hepatocytes that P-450 levels can be maintained with appropriate medium (Paine, 1990), and inducibility can be altered by the substrate upon which the cells are cultured (Guzelian et al, 1988).

In summary, primary cultures offer a cell system which can often provide a better approximation of the intact tissue than can cell lines. However, care must be taken to minimize the changes that occur with increasing time in culture.

Advantages and Limitations of Freshly Isolated Cell Suspensions Versus Monolayer Culture

If in vitro systems are to serve as reliable and predictable experimental models to study the toxicity of xenobiotics, they must retain differentiated functions and responses that are characteristic of the intact tissue in vivo. One target tissue that has received the most attention by investigators is the liver because of its role in the metabolism of xenobiotics to inactive, pharmacologically active, or toxic metabolites. For example, of the several in vitro hepatic models available to investigate the metabolism and toxicity of xenobiotics, the following preparations have been utilized the most extensively: perfused liver, liver cell lines, freshly isolated hepatocytes in suspension, and primary monolayer cultures of hepatocytes (Acosta and Sorensen, 1983; Acosta et al, 1985, 1987). Table IV summarizes the advantages and disadvantages of the various liver models.

Table IV: In Vitro Liver Systems Used for Cytotoxicity and Metabolism Studies

Perfused liverRetention of structural integrity; maintenance of cell-to-cell interrelationships.Viable for only a few hours; complex and costly perfusion apparatus; high interlaboratory variability; statistical sampling problems.
Liver cell linesIncreased viability period. Easier to maintain than primary cultures.Loss of differentiated liver functions; characteristics of transformed cells.
Freshly isolated hepatocytesEase of isolation; drug-metabolizing capacity similar to intact liver; ability to evaluate toxicity and metabolism of xenobiotic in the same system.Lack of cell-cell contact; viable for a few hours only; isolation procedure damages membranes; cofactors and enzymes leak out; impaired intermediary metabolism.
Primary hepatocyte cultureIncreased longevity (24 to 48 hours, longer with co-culture); recovery from trauma and damage from isolated procedure; retention of several differentiated liver functions; useful for chronic as well as acute in vitro studies; useful for metabolism studies.Significant loss of cytochrome P-450 during first 24 hours of culture; loss of several other differentiated functions upon subculturing cells (functional liver cell lines are difficult to establish).

To establish the functionality of in vitro liver systems and their similarity to the liver in vivo, several investigators have established a number of parameters that are considered characteristic of functional hepatocyte. These include sulfobromophthalein uptake, glutathione conjugation, L- and M-pyruvate kinase activity, cytochrome P-450 levels, O-demethylation and conjugation activity, metabolic activation of xenobiotics to toxic intermediates, total urea, maintenance of stable lactate-to-pyruvate ratios, and cytotoxic injury after exposure to well-known hepatotoxic agents. These activities can be monitored to ensure the continued functionality and uniformity of cell populations and to reveal the injurious effects of toxic chemicals as well. It is important to note that morphological changes such as blebbing (indicative of cell membrane and cytoskeletal disturbance) often precede changes detectable by these biochemical assays. Microscopic observation of cultures should always be done immediately after addition of test agent and at regular intervals thereafter.

To illustrate some potential problems of using freshly isolated hepatocytes as an experimental system in the study of the interactions of calcium and toxic chemicals, the following description of studies will be discussed.

Extracellular calcium has been shown to have paradoxical, beneficial, and detrimental effects on primary cultures or isolate suspensions of rat hepatocytes that have been exposed to cytotoxic agents. Schanne et al (1979) showed that 10 different membrane-active toxins caused more cell damage in the presence than in the absence of 3.6 mM calcium chloride as assessed by viability, plating efficiency, and dye hydrolysis. Carbon tetrachloride hepatotoxicity was reported to occur in the presence of 3.6 mM calcium chloride when the ionophore A23187 was used to disrupt intracellular calcium homeostasis; cultured hepatocytes showed marked swelling and blebbing, as well as increased LDH (lactate dehydrogenase) and GPT (glutamic-pyruvic transaminase) leakage (Chenary et al, 1981). Casini and Farber (1981) found that carbon tetrachloride decreased cellular viability with increasing levels of extracellular calcium from 0.3 to 3.6 mM.

Although these three studies from two laboratories showed detrimental effects of calcium on hepatocyte cultures less than two hours old, other laboratories have reported either beneficial effects or no effects of calcium on cultures or isolated cells of a similar age. Smith et al (1981) found that several cytotoxic agents caused less damage to cultured hepatocytes in the presence of 2.6 mM calcium than under calcium-free conditions. Edmondson and Bang (1981), furthermore, reported that freshly isolated hepatocytes incubated in calcium-free medium lacked microvilli and nuclear contents, rapidly lost the ability to exclude trypan blue and to retain LDH, and failed to accumulate Y-aminoisobutryic acid. Stacey and Klaassen (1982) found that the addition of 1 mM calcium chloride to the incubation medium did not change hepatocyte viability or leakage of potassium or AST (aspartate aminotransferase) from freshly isolated liver cells exposed to cadmium, copper, amphotericin B or lysolecithin. Omission of calcium from the medium, however, reversed the toxicity of the calcium ionophore A23187 in this same study.

Because the previously cited cytotoxicity studies used either freshly isolated hepatocytes or hepatocytes cultured for two hours or less, it is likely that even slight plasma membrane damage, caused by the rigorous isolation procedures usually employed, would allow altered hepatotoxin influx and subsequent damage more severe than would be observed if cells were allowed to recover following isolation.

In fact, freshly isolated parenchymal hepatocytes are reported to be much altered cells (metabolically, morphologically, and functionally) primarily because of the use of collagenase and mechanical agitation for dissociation of the intact liver into individual cells or clusters of cells. Freshly isolated cells have reduced ornithine decarboxylase activity, reduced protein synthesis with altered response to hormones, disaggregation of polysomes, lower albumin, transferrin and fibrinogen levels, and lower 5'-nucleotidase and alkaline phosphatase activities when compared to intact liver cells or to cells cultured for one or more days. Although freshly isolated rat hepatocytes maintain their overall shape and dye-exclusion capacities, these cells generally lose about 10% of their metabolic capacity per hour and thus are probably less suited for chemical challenge than are those cells allowed to recover from the trauma of isolation and to regain metabolic activities characteristic of the normal liver (Tanaka et al, 1978; Kato et al, 1979; Krebs  et al, 1979; Bissell and Guzelian, 1980; Just and Schimassek, 1980; Ichihara et al, 1980.).

Tissue Slice Cultures: Use, Advantages, and Disadvantages

Method of Slice Procurement

Isolated and cultured cells offer many advantages both in the study of mechanisms of cellular toxicity by chemical agents and for in vitro bioassays. These systems have the advantage of ease of use and reproducibility, but, frequently lack (or express in greatly reduced amounts) the differentiated functions of the in vivo tissue from which they were isolated. The culture of tissue explants (pieces or the entire organ taken without enzyme dissociation) has received extensive use in recent years as a means of maintaining differentiated tissue function in an in vitro model. While explant cultures have proven successful for a number of tissues, particularly the epithelial lining of organs such as skin, bronchus, esophagus, and bladder, others have met with less success. Cultures of solid parenchymal tissue such as kidney and liver have become possible only within recent years. The difficulty with these techniques appears to be due, in part, to the inability of nutrients and oxygen to reach the cells inside the solid tissue when they are placed in explant culture.

Recently, several investigators have developed a procedure using very thin liver and kidney slices that appears to minimize the ischemic effects seen in the explants. A tissue slicer was designed to provide controlled, reproducible sections of tissue that maintain the differentiated functions of the intact in vivo tissue while allowing for the diffusion of nutrients and oxygen into the cells within the tissue slice. In the past, the problem with liver slices has been the inability to produce slices thin enough to prevent ischemia of the cells in the center of slice. Also, cells along the cut surfaces of the prepared sliced were often destroyed due to compression and cutting injury. A simple machine has been developed that readily produces slices of tissue from organs that are of consistent thickness with minimal traumatic effects produced during the slice preparation (Krumdieck et al., 1980; Sipes et al, 1987). This precision cut slicer can produce slices from liver, heart, and kidney tissue thin enough (250 µm) to allow for the adequate diffusion of nutrients and gases into the cells in the center of the slice.

Uses of Tissue Slices in Toxicology Studies

Tissue slices are ideal systems to study comparative metabolism of chemical compounds between rodents and humans. Since the slice maintains the three-dimensional structure between the cells in an exact relationship as that found in vivo, the tissue slice method offers an ideal in vitro system available for in vivo to in vitro comparisons. Tissue slice cultures also offer an effective means by which metabolism, DNA damage, detoxification, and lethality of toxic chemicals can be compared between species. The use of slices can result in a considerable reduction in the number of animals needed for toxicity studies. From a rodent liver, for example, over a hundred liver slices can be prepared from each animal. This allows for the examination of a number of concentrations and exposure durations for each toxic compound. Slices can also be made from animals that have been pretreated in vivo with one compound and treated in vitro with another compound(s), thus allowing for the examination of synergistic, additive or antagonistic effects of chemicals. Slices have also been made of the liver from tumor-containing animals, allowing for the study of normal cell-tumor cell interaction.

Disadvantages of Slices in Toxicology Studies

There are a number of disadvantages to the use of tissue slices in in vitro toxicology evaluations, however. Perhaps of primary importance is that viability of the tissue slice in current culture techniques is limited to approximately 36 hours. This time constraint presently limits that use of the tissue slices to in vitro studies of 24 hours or less. Other minor disadvantages of the slice technique include the difficulty with morphological evaluation of the tissue while in culture and the variation, due to small changes of slice thickness, that may be observed in uptake and absorption of compounds from the culture medium.

Advantages of Slices in Toxicology .Studies

There are a number of advantages over primary and continuous cell culture techniques. Cells in slices are maintained in a differentiated state. In the tissue slice, cell-to-cell and cell-to-matrix relationships are supported in a manner similar to that seen in vivo. Differentiated function and morphology of the cells in the tissue are also sustained. Because enzymatic dissociation of cells from the tissue matrix is not performed, the functional and morphologic heterogeneity of the tissue is preserved in the same manner as that found in the intact organ. A final advantage is the ease with which slices can be obtained and placed in culture. For example, after a little practice, it is easy to produce 15-20 good quality slices from rodent liver per minute by using the Krumdieck tissue slicer (Krumdieck et al, 1980). These slices can then be placed into a rocking culture apparatus and cultured up to 24 hours. With practice, the time from acquisition of the liver tissue to placement of slices into culture is usually less than 15 minutes.


Tissue slices provide another useful  in vitro model in which the mechanisms and effects of toxic agents can be examined. The maintenance of the same three-dimensional structure and cellular heterogeneity in vitro to that seen in the organ in vivo provides an excellent system in which differentiated function and morphology of the tissue is maintained. Despite some shortcomings, the tissue slice model provides an excellent means for reducing the number of animals needed for comparative toxicity studies while furnishing a suitable and close adjunct to in vivo testing.

Human Cells

Although most toxicologists are concerned with solving problems that affect humans, most research is conducted in nonhuman species because techniques for studying responses in vivo with human subjects are obviously limited. The significant differences among species in cellular regulatory and metabolic processes, and in responses to specific cytotoxic perturbations, make it difficult to predict the susceptibility of human cells to toxic chemicals. The toxic response to a given challenge can also be an organ-specific phenomenon as well as species-specific. Thus, human cell in vitro models which may approximate the range of species-specific cellular targets, xenobiotic chemical metabolism and responses to toxic injury (e.g., repair, release of secondary mediators, etc.) expected of the intact target organ should be valuable tools in predicting toxicity to humans.

Species differences in toxic responses are regularly observed in vivo, and thus similar differences may be expected at the cellular level in vitro. Cells isolated from the same organ of several species may differ with respect to cellular target, xenobiotic transport and metabolism, repair capacity, and the production of secondary mediators. Direct comparative studies between human and nonhuman cells derived from the same tissues have been relatively limited, except in the area of xenobiotic metabolism by hepatocytes (see below). Those studies available which examined transport, repair or production of secondary mediators have shown differences between human and nonhuman cells (often rodent cells), providing some examples of differences to be expected. Steel and Morgan (1988) reported a comparison of human, rat, and rabbit nasal turbinate cells in response to 12-0-tetradecanoylphorbol-13-acetate (TPA). This compound was severely toxic to human cells and had no detectable effect on rabbit cells, but stimulated cell replication in rat cells. In the same study, cadmium toxicity and uptake were measured in human and rat nasal turbinate cells. While the rat cells bound less of the cadmium (at higher doses), they were more sensitive to this metal when compared to human cells.

Several studies have shown greater DNA repair capacity in normal human cells compared to normal rodent cells. For example, Grafstrom et al (1984) compared six different tissue types of human and rat origin for nonspecific DNA repair capacity, by measuring O6-methylguanine-DNA methyltransferase activity. In all six tissues, the human tissue showed activity at least four times greater than that observed in the corresponding rat tissue. Secondary mediators of inflammation play a very significant role in the overall toxic response of a tissue. Cell type differences in the mediators released, such as the arachidonate metabolites or interleukins, are well documented (Powanda, 1985; Churchill et al, 1989). Species differences in the specific products released and the methods that must be used for their identification and determination support the use of human cells for such studies (DeLeo et al, 1987).

In general, primary cultures of human cells metabolize and respond to chemicals qualitatively the same as cells from laboratory species (Harris, 1987; Butterworth et al, 1989). Species-related differences in the in vivo metabolism of xenobiotics are well known however, and have been correlated with their toxicity in certain species (Caldwell, 1980; Calabrese, 1983). In vitro comparative metabolism studies using hepatocytes from several species, both suspension and monolayer cultures, have shown that these cells reproduce species-related differences in metabolite profiles and chemical toxicity, validating them as models of the intact liver (Gee et al., 1984; Green et al, 1984, 1986; Chenery et al, 1987; LeBigot et al, 1987; Cole et al, 1988; Berthou et al, 1989). Cultured hepatocytes have been used primarily for toxicology assays of cytotoxicity, comparative metabolism, peroxisome proliferation, and genotoxicity such as unscheduled DNA synthesis (UDS). Kidney cells are also capable of metabolizing xenobiotics. The metabolism of 2-acetylaminofluorene (2AAF) and benzo(a)pyrene [B(a)P] were compared in human and rate cells (Rudo et al, 1987). Considerable differences in the metabolism of 2AAF, but not B(a)P, by the cells from these species, were observed.

In the past, access to human tissue or early passage cell lines has been much more limited than it was for tissues from laboratory species. Generally, only those organizations associated with clinical centers has a dependable supply of human tissue. However, in recent years a number of programs have been established to provide such tissue for research. For example, the Cooperative Human Tissue Network, The National Disease Research Exchange, Tissue Culture and Human Tissues Laboratory (University of Maryland), Center for Human Cell Biotechnology (The University of Texas Health Sciences Center at San Antonio), Human Liver Resource Facility (Carol E. Green, SRI International) and other such organizations can provide human tissues and cell lines.1 In addition, there are now commercial sources of early passage human cells or tissues (for example, the International Institute for the Advancement of Medicine) or providing toxicity testing services using cultured human tissue (for example, in vitro Technologies Inc., Technology Enterprise Center, University of Maryland). Established human cell lines and strains, which range from well characterized to simply banked, can also be obtained from the American Type Culture Collection, the Aging Cell Repository and the NIGMS Human Cell Genetic Mutant Cell Repository.

Concurrent with the availability of the human tissues has come the formulation of specific media for their culture (Ham, 1981). Some of these media allow for the retention of proliferative potential or induction of differentiation as required. Many of these media are now available from commercial sources and contain little or no serum. Techniques for the isolation and culture of many human cell types of interest in toxicology have been published. Representative examples of these techniques can be found in publications by Triffillis et al (1985), Boyce and Ham (1986), Green et al (1986), Shapiro et al (1989) and Resau et al (1990).

The principal limitations to the use of primary human cell cultures in toxicity and metabolism assays are practical in nature. The problem of rapid phenotypic changes, particularly a decline in xenobiotic-metabolizing enzyme levels, exists in human cell cultures as it does for those from other species. There is some evidence that cytochrome P-450 is more stable in human liver cells than in rodents (Blaauboer et al, 1985). The loss of metabolic capacity with time in culture has focused attention on the need to use primary or early passage cultures, depending on the cell type, for many in vitro studies. However, the use of human tissue, whether from the liver or another organ, does pose several unique problems. These include the following: (1) human tissue is frequently difficult to obtain in a viable state; (2) the heterogeneity of the human population can result in increased viability in the data; (3) human tissues should be handled as if contaminated with pathogens; (4) the cell isolation method must be optimized for human tissue; and (5) the timing of experiments with primary human cell lines or tissues depends on organ availability.

Human tissue for use in toxicology studies can be acquired from various types of clinical specimens and donor sources. Experience with these tissues suggests that some are more representative of the organ in vivo than others. For example, adult liver specimens can be obtained at autopsy, from diagnostic biopsies, as surgical waste from pathology specimens, and from organ donors. Autopsy liver is generally relatively easy to obtain, but it is usually difficult to control the quality of the specimens because it may be several hours after death before the tissue is available. A dramatic loss of microsomal protein and enzyme activities in autopsy liver tissue, as compared to fresh liver tissue, has been demonstrated (Powis et al, 1988). However, many other tissues can be obtained and their cells successfully cultured from autopsy specimens (Shapiro et al, 1989; Resau et al, 1990). Surgical waste specimens can be acquired and are a primary source of tissue available to commercial firms producing human cell lines from skin. For other organs obtained from surgical waste, the influence of major illness (e.g., primary or metastic tumors), surgery, and drug therapies on the tissue is uncertain. Organ donor specimens are usually acquired from individuals with no history of chronic disease. With the increased use of new organ preservation solutions for flushing organs prior to transplant, the tissues are viable for increased lengths of time and facilitate shipment. An additional advantage of this tissue source is the availability of organs. Even though the numbers of liver and other transplant programs has increased dramatically, excess organs are still available.

The varied genetic background of the human population and differences in the exposures of individuals to environmental factors result in greater variability in the data obtained with human cells compared to laboratory species. For example, the genotoxicity of isoniazid to human hepatocytes has been related to the ability of the liver cells to form the acetylated metabolite (Neis et al, 1986), an activity for which genetic polymorphism in humans is well documented (Peters et al, 1990). In addition, the tissue available for use in research may not reflect the population in general. Gender, racial, and age imbalances in the particular donor populations may significantly alter the pattern of responses. Economic, social, and religious influences, in a community, can select for a donor population at large. To better represent the 'human' response, those experiments which focus on metabolism should be repeated with as many donors as practical. Usually 3 to 6 are adequate, but when wide ranges of variability are found, a larger number of specimens may be needed to derive statistically significant values for comparison or validation. All available information on each donor should be obtained and applied to the interpretation of data acquired with human cell cultures. Ideally, standard characterization assays, e.g., a set of xenobiotic metabolism activities or other responses appropriate to the test system, should be repeated with each specimen. It is recommended that an epidemiological profile accompany the tissue to facilitate interpretation of the data. Such information should include age, sex, race, cause of death, postmortem interval, and significant therapy or drug history. This information is by no means complete but would facilitate comparison of one set of human in vitro data with another.

All human tissues should be handled with caution, as if it were contaminated with pathogens. In many cases, data on possible viral infection do not exist and there can be a reluctance on the part of clinical centers to release such data if they exist. In any event, it is often necessary to prepare cells from the donor before total clinical data are available. Fortunately, many human specimens, in particular excess transplant organs, are recovered from donors that have been screened for major viral diseases, such as hepatitis B and HIV. Standard safety precautions, i.e., use of biological safety cabinets, decontamination of glassware and bench surfaces with household bleach (sodium hypochlorite solution), and protective clothing, are required when working with human specimens ("Guidelines for Handling Human Tissues and Body Fluids in Research", The National Disease Research Interchange, 1987). In addition, laboratory personnel should be immunized against hepatitis B.

Even though specific media are available for the culture of many human cell types, the preparation of cell cultures from human tissue usually requires modifications of the isolation procedure used for rodent cells. One common finding is that human organs generally contain more connective tissues than those of rodent; thus human cells isolated by enzyme perfusion (for example, with collagenase) are more difficult to recover. For the preparation of both hepatocytes and renal proximal tubules from human organs, the collagenase concentration in the perfusate must be increased 2- to 4-fold compared to the concentration used for rat cell isolations, and the perfusion must be continued for a longer period of time. Additions to the perfusates such as streptokinase, to digest blood clots in the vasculature, or trypsin inhibitor, to inhibit the action of damaging proteases, can also be beneficial. The availability of the whole organ or a large section, makes it possible to perfuse a tissue such as the liver through the portal vein, adapting the techniques used with smaller species (Frabre et al, 1988) or through a vessel on the cut surface of the tissue (Reese and Byard, 1981). High yields of hepatocytes (up to 5 x 109 viable cells per 200 g of liver) can be obtained by using these methods. However, if only small pieces of tissue are available, such as specimens from surgical waste, a nonperfusion method must be used, resulting in a lower cell yield as compared to the whole organ perfusion technique (Rudo et al, 1987).

Another problem associated with the use of primary human cells is the need to be prepared to conduct an experiment with very little advance notice on an irregular schedule. It is therefore necessary to maintain at all times a reserve of glassware and reagents, experimental protocols designed in advance, and trained personnel who are on call. Although this is a completely practical problem, it should not be dismissed as insignificant. Human tissue is seldom available at a convenient time and the stress of long work hours can lead to experimental errors. Similarly, if the assays are intended to screen chemicals for their toxicity, those chemicals must be collected and held until this tissue is available. As a practical matter, screening of small numbers of samples becomes very expensive in terms of time and materials. The frequent availability of specimens can help, but the real solution to this problem is the development of culture conditions that maintain stable differentiated cells and cryopreservation methods for human cells such as hepatocytes that allow storage of excess cells.

Other problems unique to human cells relate to ownership, profit making and ethical and legal considerations. These topics were dealt with at length by the US Congress Office of Technology Assessment in a Special Report published in 1987. To that end, sources of tissues must have tissue collection and distribution policies and procedures in place which meet all current laws and are approved by an Institutional Review Board (or other appropriate governing board) for the use of human subjects. Reports and use of material can be done anonymously without source identifiers, but initial collection should require acquisition of the pathology report relevant to the specimen and, whenever possible, preparation of histological reference materials from the same specimen.

Although past and future litigation concerning the ownership of human cells could change current policies, care in collection of tissues with proper approvals should minimize potential problems. Integral to that effort should be the idea that preparing, culturing or shipping human tissues for research is a service activity which generates cost by the nature of the service. However, it does not imply the buying or selling of body parts and, therefore, places no defined value on the specific tissue.

Tissues which have been broken down into subcomponent cellular or molecular parts by a specific method can be considered a processed raw material. This material may be considered as a potentially protective property in its raw state or in any modified state. An example of such a modification would be transfer of a genetic element (gene transfection) into the cells. This issue is far from clear and may await legal decisions from pending cases. Thus, the distribution or sharing of human cellular materials generally does not necessarily imply transfer of ownership but, simply, use of the cells or tissue. This written agreement can be for a specified time period and can include the details of subsequent developments of cell strains or clones from the parental cells, when and to whom any unused materials or banks of cells should be returned, and how co-authorships and/or acknowledgements should be dealt with in publications. Such agreements should also deal with matters of compensation. For example, the establishment of the use of distribution fee from a tissue or cell bank is a routine procedure. When government or other monies support the overall banking operations, such fees are minimal. However, commercial and nongovernment supported laboratories may assess higher fees for initiating or providing primary or passage level human cells. This will result in a perceived 'profit' which generally would come from the services or materials provided with the tissues rather than from the tissues themselves.

Donors may or may not receive compensation for their tissues, depending on the policies of the banking or distribution source. These may be nominal because, as stated above, providing the expertise and service activities to prepare the raw, routine tissue materials for their intended use are the real value. An exception may arise when the donor has some rare disease or condition which might increase the value of the raw tissue material. Value increased through biotechnology applications is secondary to the real value of the source material, but legal arguments may hinge on whether or not the donor material had any known attributes worthy of special consideration prior to beginning the study which could feasibly lead to product(s) with commercial potential. Consequently, attempts should be made to document release of potential ownership claims and/or to provide compensation. However, in those cases, it is unnecessary and burdensome to consider anything but immediate compensation, since it is unlikely that the manpower for the paperwork and follow-up of the donors would be available. This would become even more impractical if it were required for all possible donor sources, including discarded tissues.

In summary, human tissue represents a very important alternative to animal testing. Use of freshly derived tissues or cells generates unique, but not insurmountable, problems when compared to the use of established cell lines. The availability of cells through banking and distribution facilities should aid the continued development and use of human cell models. However, more emphasis must be given to the development of useful, immortalized human cells from a variety of parenchymal cell types. Such cells and their derivatives can be validated as representative models, and then banked and distributed for routine testing without the need for repeated initiation of primary cultures. Such efforts should be supported through industry-government-university foundation collaborative ventures.

1The Need for and Management of Human Tissue for Research Purposes. Report of a Concept Review by Outside Consultants to NIH, November, 1987.


  • Acosta, D. and Sorensen, E.M.B. (1983). Role of calcium in cytotoxic injury of cultured hepatocytes. Annals of the New York Academy of Science 407: 78-92.
  • Acosta, D. and Ramos, K. (1984). Cardiotoxicity of tricyclic antidepressants in primary cultures of rat myocardial cells. Journal of Toxicology and Environmental Health 14: 137-143.
  • Acosta, D., Sorensen, E.M., Anufuco, D.C., Mitchell, D.B., Ramos, K., Santone, K.S. and Smith, M.A. (1985). An in vitro approach to the study of target organ toxicity of drugs and chemicals. In Vitro Cellular and Developmental Biology 21: 495-504.
  • Acosta, D., Mitcell, D.B. and Sorensen, E.M.B. (1987). The metabolism and toxicity of xenobiotics in a primary culture system of postnatal rat hepatocytes. The Isolated Hepatocytes (eds. Rauchman and Padella) pp. 189-214. Academic Press, New York.
  • Augusti-Tocco, G. and Sato, G. (1969). Establishment of functional clonal cell lines of neurons with mouse neuroblastoma. Proceedings of the National Academy of Science USA 64: 311-315.
  • Babich, H. and Borenfreund, E. (1987). Fathead minnow FHM cells for use in in vitro cytotoxicity assays of aquatic pollutants. Ecotoxicological and Environmental Safety 14: 78-87.
  • Babich, H., Sardana, M.K. and Borenfreund, E. (1988). Acute cytotoxicities of polynuclear aromatic hydrocarbons determined in vitro with the human liver tumor cell line HepG2. Cell Biol. Toxicol. 4: 285-309.
  • Babich, H., Martin-Aguacil, N. and Borenfreund, E. (1989). Comparison of the cytotoxicities of dermatotoxins to human keratinocytes and fibroblasts in vitro. Alternative Methods in Toxicology, Vol 7. (ed. Goldberg, A.). pp. 153-167. Mary Ann Liebert Inc., New York.
  • Berthou, F., Ratanasavanh, D., Riche, C., Picart, D., Voirin, T. and Guillouzo, A. (1989). Comparison of caffeine metabolism by slices, microsomes, and hepatocyte cultures from adult human liver. Xenobiotica 19: 401-417.
  • Billings, R.E., McMahon, R.E., Ashmore, J. and Wagle, S.R. (1977). The metabolism of drugs in isolated rat hepatocytes. A comparison with in vivo drug metabolism and drug metabolism in cellular fractions. Drug Metab. Dispos. 5: 518-526.
  • Bissell, D.M. and Guzelian, P.S. (1980). Phenotypic stability of adult rat hepatocytes in primary monolayer culture. Annals of the New York Academy of Science 349: 85-98.
  • Blaauboer, B.J., van Holsteijn, I., Van Graft, M. and Paine, A.J. (1985). The concentration of cytochrome P-450 in human hepatocyte culture. Biochemical Pharmacology 34: 2405-2408.
  • Borenfreund, E. and Borrero (1984). In vitro cytotoxicity assays. Potential alternatives to the Draize ocular irritancy test. Cell Biology and Toxicology 1: 55-65.
  • Borenfreund, E. and Puerner, J.A. (1986). Cytotoxicity of metals, metal-metal and metal-chelator combinations assayed in vitro. Toxicology 9: 121-134.
  • Borenfreund, E. and Puerner, J.A. (1987) Short-term quantitative in vitro cytotoxicity assay involving an S-9 activating system.  Cancer Lett. 34: 243-248.
  • Boyce, S.T. and Ham, R.G. (1986). Normal human epidermal keratinocytes. In Vitro Models for Cancer Research. Vol 3. (eds. Weber, M.M. and Sekey, L. pp. 245-274. CRC Press, Inc. Boca Raton, Florida.
  • Butterworth, B.E., Smith-Oliver, T., Earle, L., Loury, D.J., White, R.D., Doolittle, D.J., Working, P.K., Cattley, R.C., Jirtle, R., Michaloupoulos, G. and Strom, S. (1989). Use of primary cultures of human hepatocytes in toxicology studies. Cancer Research 49: 1075-1084.
  • Calabrese, E.J. (1983). Comparative Metabolism: the principal cause of differential susceptibility to toxic and carcinogenic agents. Chapter 5 Principals of Animal Extrapolation. pp. 203-282. John Wiley & Sons, New York.
  • Caldwell, J. (1980). Comparative aspects of detoxicification in mammals. Enzymatic Basis of Detoxication, Vol. I. (ed. Jakoby, W.J.). pp. 85-114). Academic Press, New York.
  • Casisni, A.F. and Farber, J.L. (1981). Dependence of the carbon tetrachloride-induced cell death of cultured hepatocytes on the extracellular calcium concentration. American Journal of Pathology 105: 138-148.
  • Chenery, R.M., George, M. and Krishna, G. (1981). The effect of ionophore A23187 and calcium on carbon tetrachloride induced toxicity in cultured rat hepatocytes. Toxicology and Applied Pharmacology 60: 241-252.
  • Chenery, R.J., Ayrton, A. Oldham, H.G., Standring, P., Norman, S.J., Seddon, T. and Kirby, R. (1987). Diazepam metabolism in cultured hepatocytes from rat, rabbit, dog, guinea pig, and man. Drug Metabolism and Disposition 15:312-317.
  • Churchill, L., Chilton, F.G., Resau, J.H., Bascom, R., Hubbard, W.C. and Proud, D. (1989). Cyclooxygenase metabolism of endogenous arachidonic acid by cultured human trachael epithelial cells. American Review of Respiratory Diseases 140: 449-459.
  • Cole, K.E., Jones, T.W., Lipsky, M.M., Trump, B.F. and Hsu, I.C. (1988). In vitro binding of aflatoxin B1 and 2-acetylaminofluorene to rat, mouse, and human hepatocyte DNA: the relationship of DNA binding to carcinogenicity. Carcinogenesis 9: 711-716.
  • De Leo, V.A., Harber, L.C., Kong, B.M. and De Salve, S.J. (1987). Surfactant-induced alteration of arachidonic acid metabolism of mammalian cells in culture. Proceedings of the Society for Experimental Biology and Medicine 184:477-482.
  • Duffy, P.A. and Flint, O.P. (1987). In vitro dermal irritancy tests. In Vitro Methods in Toxicology (eds. Atterwill, C.K. and Steele, C.E.) pp. 279-297. Cambridge University Press, Cambridge.
  • Edmonson, J.W. and Bang, N.O. (1981). Deleterious effects of calcium deprivation on freshly isolated hepatocytes. American Journal of Physiology 241: 63-68.
  • Fabre, G., Rahmani, R., Placidi, M., Combalbert, J., Covo, J., Cano, J-P., Coulange, C., Ducros, M. and Rampal, M. (1988). Characterization of midazolam metabolism using human hepatic microsomal fractions and hepatocytes in suspension obtained by perfusing whole human livers. Biochemical Pharmacology 37: 4389-4397.
  • Gee, S. Green, C.E. and Tyson, C.A. (1984). Comparative metabolism of tolbutamide by isolated hepatocytes from rat, rabbit, dog, and squirrel monkey. Drug Metabolism and Disposition 12: 174-178.
  • Grafstrom, R.C., Pegg, A.E., Trump, B.F. and Harris, C.H. (1984). O6-Alkylguanine-DNA alkyltransferase activity in normal human tissues and cells. Cancer Research 44: 2865-2867.
  • Green, C.E., Dabbs, J.E. and Tyson, C.A. (1984). Metabolism and cytotoxicity of acetaminophen in hepatocytes isolated from resistant and susceptible species. Toxicology and Applied Pharmacology 76: 139-149.
  • Green, C.E., LeValley, S.E. and Tyson, C.A. (1986). Comparison of amphetamine metabolism using isolated hepatocytes from five species including human. Journal of Pharmacology and Experimental Therapeutics 237: 931-936.
  • Grisham, J.W. and Smith, G.J. (1984). Predictive and mechanistic evaluation of toxic responses in mammalian cell culture systems. Pharmacological Reviews. 36: 1515-1715.
  • Guzelian, P.S., Li, D., Schuetz, E.G., Thomas, P., Levin, W., Mode, A. and Gustaffsson, J.A. (1988). Sex change in cytochrome P-450 phenotype by growth hormone treatment of adult rat hepatocytes maintained in a culture system on matrigel. Proceedings of the Natational Academy of Science USA 85: 9783-9787.
  • Ham, R.G. (1981). Survival and growth requirements of nontransformed cells. Handbook of Experimental Pharmacology, Vol. 57 (ed. Baserga, R.) pp. 13-88. Springer-Verlag, Berlin.
  • Harris, C.C. (1987). Human tissues and cells in carcinogenesis research. Cancer Research 47: 1-10.
  • Hay, R., Macy, M., Chen, T.R., McClintock, P. and Reid, Y. (1988). American Type Culture Collection Catalogue of Cell Lines and Hybridomas. 6th Edition American Type Culture Collection, Rockville, Maryland 20852.
  • Hurko, O. McKee, L. and Zuurveld, G.E.M. (1986). Transfection of human skeletal muscle cells with SV40 large T-antigen gene coupled to a metallothinein promoter. Annals of Neurology 20: 573-582.
  • Ichihara, A., Nakamura, T. Tanaka, K., Tomita, Y., Aoyama, K., Kato, S. and Shinno, H. (1980). Biochemical functions of adult rat hepatocytes in primary culture. Annals of the New York Academy of Science 349: 77-84.
  • Jat, P.S., Noble, M.D., Ataliotis, P., Tanaka, Y., Yannoutsos, N. Larsen, L. and Kioussis, D. (1991). Direct derivation of conditionally immortal cell lines from an H-2Kb-tsA58 transgenic mouse. Proceedings of the National Academy of Science USA 88: 5096-5100.
  • Just, W.W. and Schimassek, H. (1980). Lack of ornithine decarboxylase activity in isolated rat liver parenchymal cells. European Journal of Cellular Biology 22: 649-653.
  • Kato, S., Aoyama, K. Nakamura, T. and Ichihara, A. (1979). Biochemical studies of liver functions in primary cultured hepatocytes of adult rats. III. Changes of enzyme activities on cell membranes during culture. Journal of Biochemistry (Tokyo) 86: 1419-1425.
  • Knowles, B., Hower, C.C. and Aden, D.P. (1980). Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B antigens. Science 209:497-499.
  • Krebs, H., Lund, P. and Edwards, M. (1979). Features of isolated liver cells. Cell Populations, Methodological Surveys in Biochemistry, Vol. 8. (ed. Reid, E.) pp 1-6. Ellis Horwood Limiteed, Chichester, England.
  • Krumdieck, C.L., dos Santos, J.E. and Ho, K.J. (1980). A new instrument for the rapid preparation of tissue slices. Analytical Biochemistry 104: 118-123.
  • LeBigot, J.F., Begue, J.M., Kiechel, J.R. and Guillouzo, A. (1987). Species differences in metabolism of ketotifen in rat, rabbit, and man: demonstration of similar pathways in vivo and in cultured hepatocytes. Life Science 40: 883-890.
  • Maron, D.M. and Ames, B.N. (1983). Revised methods for the Salmonella mutagenicity test. Mutatation Research 113: 173-215.
  • McAllister, J.M. and Hornsby, P.J. (1987). Improved clonal and nonclonal growth of human, rat, and bovine adrenocortical cells in culture. In Vitro Cellular and Developmental Biology 23: 667-685.
  • McKeehan, W., Barnes, D., Reid, L., Stanbridge, E., Murakami, H. and Sato, G.H. (1990). Frontiers in mammalian cell culture. In Vitro Cell and Dev. Biol. 26: 9-23.
  • McQueen, C.A. and Williams, G.M. (1988). Conjugation of chemical carcinogens in cultured hepatocytes. Cellular and Molecular Aspects of Glucoronidation, Vol. 173 (eds. Siest, G., Magdalou, J. and Burchell, B.) pp. 121-127. Cooloque INSERM/John Livvey Eurotest Ltd.
  • McQueen, C.A. Miller, M.J. and Williams, G.M. (1986). Sex differences in the biotransformation of 2-acetylaminofluorene in cultured rat hepatocytes. Cell Biology and Toxicology 2: 271-281.
  • Guidelines for Handling Human Tissues and Body Fluids in Research. National Disease Research Interchange, Philadelphia, PA (1987).
  • Neis, J.M., Yap, S.H., Roeloefs, H.M.J., Van Gemert, P.J.L., Te Brommelstroet, B.W.J. and Henderson, P.Th (1986). Mutagenicity and DNA-excision repair induced by isoniazid after metabolic activation by isolated human and rat hepatocytes. Cancer Letters 30: 103-111.
  • Paine, A.J. (1990). The maintenance of cytochrome P-450 in rat hepatocyte culture: some applications of liver cell cultures to the study of drug metabolism, toxicity, and the induction of the P-450 system. Chemical and Biological Interactions 74: 1-31.
  • Peters, J.H., Gordon, G.R., Lin, E., Green, C.E. and Tyson, C.A. (1990). Polymorphic N-acetylation of sulfamethazine and benzidine by human liver: implications of cancer risk? Anticancer Research 10: 225-230.
  • Powanda, M.C. (1985). The role of interleukin-I in homeostasis. The Physiologic, Metabolic, and Immunologic Actions of Interleukin-I(eds. Kluger, M.J., Oppenheim, J.J. and Powanda, M.C.) Alan R. Liss, Inc., New York.
  • Powis, G.Y. Jardine, I., Van Dyke, R. Weinshilbourn, R. Moore, D., Wilke, T.O., Rhodes, W., Nelson, R., Benson, L. and Szumalansk, C. (1988). Foreign compound metabolism studies with human liver obtained as surgical waste. Drug Metabolism and Disposition 16: 582-589.
  • Reese, J.A. and Byard, J.L. (1981). Isolation and culture of audlt hepatocytes from liver biopsies. In Vitro 17: 935-940.
  • Reasu, H.J., Phelps, P.C., Zhu, S.M., Smoot, D., Lee, H.K., Cottrell, J.R., Hudson, E.A. and Elliget, K.A. (1990). Long-term culture of human esophageal explants and cells. Cytotechnology 3: 61-73.
  • Rheinwald, J.G. and Green, H. (1975). Formation of a keratinizing epithelium in culture by a cloned cell line derived from a teratoma. Cell 6: 317-330.
  • Rudo, K.M., Dauterman, W.C. and Langerbach, R. (1987) Human and rat kidney cell metabolism of 2-acetylaminofluorene and benzo(a)pyrene. Cancer Research 49: 1187-1192.
  • Rudo, K.M., Meyers, W.C., Dauterman, W.C. and Langenbach, R. (1987). Comparison of human and rat hepatocyte metabolism and mutagenic activation of 2-acetylaminofluorene. Cancer Research 47: 5861-5867.
  • Sato, G. (ed.) (1981). Functionally Differentiated Cell Lines. Alan R. Liss, New York.
  • Sato, G., Augusti-Tocco, G. and Psoner, M. (1970). Hormone secreting and hormone-responsive cell cultures. Record of Progress in Hormone Research 26: 539-546.
  • Schaeffer, W.I. (1990). Terminology associated with cell, tissue and organ culture, molecular biology and molecular genetics. Tissue Culture Association Terminology Committee. In Vitro Cellular and Developmental Biology 26: 91-101.
  • Schanne, F.A., Kane, A.B., Young, E.E. and Farber, J.L. (1979). Calcium dependence of toxic death: a final common pathway. Science 206: 700-702.
  • Shapiro, D.L., Finkelstein, J.N. and Van Diver, T. (1989). Isolation of alveolar epithelial cells from lung tissue obtained at autopsy. In Vitro Cellular and Developmental Biology 25: 1051-1054.
  • Sipes, I.G., Fisher, R.L., Smith, P.F., Stine, E.R., Gandolfi, A.J. and Brendel, K.A. (1987). Dynamic liver culture system: a tool for studying chemical biotransformation and toxicity. Archives of Toxicology. Supplements 11: 20-33.
  • Sirbasku, D.A. (1978). Estrogen induction of growth factors specific for hormone-responsive mammary, pituitary, and kidney tumor cells. Proceedings of the National Academy of Science USA 75: 3786-3790.
  • Smith, M.T., Thor, H. and Orrenius, S. (1981). Toxic injury to isolated hepatocytes is not dependent on extracellular calcium. Science 213: 1257-1259.
  • Stacey, N.H. and Klaassen, C.D.J. (1982). Lack of protection against chemically induced injury to isolated hepatocytes by omission of calcium from the incubation medium. Journal of Toxicology and Environmental Health. 9: 267-276.
  • Steele, V.E. and Morgan, D.L. (1988). Chemically induced toxicity in cultured respiratory tract epithelium from animals and humans. Toxicology of the Lung (eds. Gardner, D.E., Grapo, J.D. and Massaro, E.J.). pp.315-322. Raven Press, New York.
  • Tanaka, K., Sato, M., Tomita, Y. and Ichihara, A. (1978). Biochemical studies of liver functions in primary cultures of hepatocytes of adult rats. I. Hormonal effects on cell viability and protein synthesis. Journal of Biochemistry (Tokyo) 84: 937-946.
  • Thompson, E.B., Tomkins, G.M. and Curran, J.F. (1966). Induction of tyrosine alpha-ketoglutarate transaminase by steroid hormones in a newly established tissue culture cell line. Proceedings of the National Academy of Science USA 56: 296-303.
  • Tong, C., Telang, S. and Williams, G.M. (1984). Differences in responses of four adult rat liver epithelial cell lines to a spectrum of chemical mutagens. Mutation Research 130: 53-61.
  • Trifillis, A.L., Regec, A.L. and Trump, B.F. (1985). Isolation, culture, and characterization of human renal tubular cells. Journal of Urology 133: 324-329.
  • US Congress Office of Technology Assessment (1987). New Developments in Biotechnology I: Ownership of Human Tissues and Cells, Special Report No. OTA-BA-337, Corp. Service Codes 05874000 (Washington DC: US Government Printing Office). March.
  • Williams, G.M., Mori, H. and McQueen, C.A. (1989). Structure-activity relationships in the rat hepatocyte DNA-repair test for 300 chemicals. Mutatation Research 221: 263-286.


Richard Ham1, David Barnes2, Charlene McQueen3, Oliver Flint4, June Bradlaw5, and John Frazier6

1University of Colorado, MCD Biology Campus, Box 347, Boulder, CO 80309; 2Department of Biochemistry and Biophysics, Environmental Health Center, Oregon State University, Corvallis, OR 97331; 3Department of Pharmacology and Toxicology, College of Pharmacy, University of Arizona Health Sciences Center, Tucson AZ 85721; 4Investigative Toxicology, Bristol-Myers Squibb, P.O. Box 4755, Syracuse, NY 13221; 5Division of Toxicological Studies, HFF-162, Food and Drug Administration, 200 C Street, S.W., Washington, DC 20204; and 6The Johns Hopkins Center for Alternatives to Animal Testing, 615 North Wolfe Street, Room 7033, Baltimore, MD 21205


Selection of Test System

In vitro toxicity testing is best done with cells that respond as much like those in the intact human body as possible (Alberts et al, 1989). Cultured normal human cells of the type that would be most critically affected in vivo by the substance being tested would be preferred. Where this is not feasible, the next best choice may be a continuous cell line of human origin that has undergone minimal deviation from the behavior of comparable normal cells, or else the same cell type (primary culture usually preferable to cell line) from a species of laboratory animal that is also used for in vivo testing of related substances.

For extrapolation to effects on intact humans, the most effective approach may be to compare effects on a small number of living animals with effects on cultured cells from the same species, and then to compare species differences in cell culture responses. Such comparisons should be both the basis for making cell culture to whole animal correlations and an insight into species-specific differences in response to the test substances, both of which facilitate extrapolation to potential effects on intact humans.

Cell Type

Classical continuous lines of cultured cells (BALB/c 3T3 for example) have generally undergone extensive selection for ability to grow well in conventional media, and therefore, tend to exhibit similar growth requirements. Normal cells and most artificially transformed or tumor-derived continuous lines that continue to exhibit tissue-specific properties have not undergone a comparable level of adaptation. These tend to exhibit much greater individuality in their requirements for growth and differentiation in culture. Because of this, there are no generic culture media or culture conditions that can reliably be used for normal cells or for minimally altered continuous cell lines. Instead, it is necessary to optimize a wide range of variables specifically for growth and differentiation of each cell type studies. In many cases good, but not necessarily optimal, culture conditions for a particular cell type have been described in the literature. These may have to be modified for the particular needs of toxicity testing. For example, it may be important to use serum-free medium to avoid serum enzymes or serum protein binding.

Differentiated State

For many cell types, the culture conditions that yield optimal growth are quite different from those that result in optimal expression of differentiated properties. Because one of the major objectives of in vitro testing is to reduce the number of animals that must be used, it is important to expand cultured cell populations as much as is practical before using them for testing. Therefore, it will often be necessary to use one culture medium to expand the cellular population, followed by induction of appropriate differentiated states through use of a different medium.

It is also important to view differentiation as a multi-step process. Many cell types of interest, such as hematopoietic cells and epidermal keratinocytes, undergo a series of stepwise changes from proliferating stem cells through progressively altered immediate types, and ultimately to fully differentiated cellular phenotypes that are nonproliferative or have only limited capacity for further proliferation. In many cases, the stem cells or the early intermediates are more sensitive to toxic or irritating agents than the terminally differentiated cells. In contrast to bone marrow cell cultures, where cell division in the presence of specific growth factors may be encouraged, the differentiated phenotype of other cell types (hepatocytes, neuronal cells) may only be preserved in the absence of cell division. Thus, for any given toxicity assay, it is necessary to decide which stages of differentiation are the appropriate targets, and then to manipulate the culture conditions to ensure that an appropriate distribution of cellular phenotypes will be present.

Nature of Assay

It is also necessary to determine what types of effects are to be examined. In some cases, one is looking for generalized cytotoxicity, but more often than not the primary concern is over possible damage to a specific type of tissue resulting from a specific type of exposure. Thus, in addition to cell type and differentiated state, it is necessary to be concerned with other issues such as whether the cells are cycling, position in the cell cycle at the time of exposure, overall metabolic activity of the cell, whether the cells have formed tight junctions and the way that the test substance is presented.

For example, if one is concerned with mutagenic or carcinogenic effects, it will probably be necessary to have the cells in an active state of proliferation when they are exposed, and also to keep them proliferating long enough to detect induced changes. For substances that must be metabolically activated, it is necessary to use cell types with the enzymes needed for activation, or else to activate the substances by other means, such as supplying the enzymes exogenously.

Optimized In Vitro Environment

Thus, it is only after cell type, the differentiated state(s), and the nature of the assay have been decided upon that serious attempts can be made to identify an appropriately optimized culture environment for in vitro toxicity testing. For certain assays such as the keratinocyte neutral red assay, the conditions have already been partially optimized. However, for many other normal cell culture systems, the work of optimization has yet to be done. Moreover, assay systems will generally have to be formulated for each new cell type.

Essential Components in Nutrient Media for Growth of Normal Cells


Certain minimal requirements, depending on cell type, must be satisfied before cells will multiply in culture. Serum helps to support cell growth in a wide variety of different ways, all of which must be replaced or compensated for in order to obtain growth in a defined medium (Ham and McKeehan, 1978, 1979; Ham, 1981, 1984). Hormones and growth factors are only one part of a much larger picture, which includes the nutrient medium, the culture substrate, and the overall physiological environment, with holistic interactions, such as changes in any individual component of the system, can potentially alter responses to virtually any other component.

Nutrient Media

Serum replacement with defined mixtures of hormones and growth factors was highly successful for a wide range of continuous cell lines and strains that had retained only modest serum requirements (Barnes and Sato, 1980). However, to achieve comparable results with normal cells (primary cultures and untransformed, diploid cell lines), it has generally been necessary to begin by optimizing the nutrient medium in order to reduce the serum requirement to a level that makes it feasible to replace the remainder with more highly defined substances (Ham, 1981, 1982, 1984).

Individuality of Nutrient Requirement

Different types of normal cells usually require nutrient media that differ substantially from one another, with quantitative differences in relative concentrations of specific nutrients often being a major importance (Ham, 1984). As an example, the key to successful growth of human epidermal keratinocytes without a feeder layer was the development of nutrient medium MCDB 151 and its further refinement to the modified MCDB 153 formulation now used in Clonetics KGM growth medium.

Key Issues Related to Nutrient Media

A detailed analysis of nutrient media for diverse types of normal and minimally altered cells is clearly beyond the scope of this report. However, the following broad principles are important:

  1. Cell type specificity: For growth of normal cells in defined media, it is usually necessary to optimize the nutrient medium specifically for the cell type in question. Among other things, optimization is usually needed to reduce the amount of undefined supplementation that is needed to a level where it can reasonably be replaced with more highly defined substances.
  2. Selenium: It is critically important to have an adequate amount of the trace element selenium in serum-free media (generally added as selenite ion). Selenium is a key structural component of the enzyme glutathione peroxidase, which plays an essential role in the neutralization of metabolically generated peroxides, potentially including those generated in response to substances being tested for toxicity.
  3. Iron: It is difficult to provide adequate amounts of soluble iron in defined media. Ferric iron is highly insoluble at neutral pH, and ferrous iron tends to be oxidized rapidly to ferric. Cells appear to be able to use the oxidized form despite its insolubility, but care must be taken not to lose iron from the medium on a sterilizing filter. The best procedure is aseptic addition of a sterile stock solution of ferrous sulfate after the medium has been filtered.
  4. Transferrin: Transferrin is often regarded as an invariable requirement for defined medium growth. This is largely due to loss of iron on sterilizing filters (which appears to be a particularly major problem when nutrient media are prepared from dry powders). Transferrin is used to solubilize ferric iron and present it to cells in a receptor-dedicated fashion. However, transferrin preparations are generally not highly purified and this reduces the level of definition of the culture medium. The requirement for transferrin can be largely eliminated by addition of inorganic iron salts in a manner that makes the iron directly available to the cells. Transferrin also binds a number of other heavy metals, and tends to have a buffering effect on toxic excesses, as well as serving as a reservoir for needed trace elements. If transferrin is used, it is important to match the species closely enough so that the transferrin is compatible with the cellular receptors. Mammalian transferrins tend to be interchangeable, but avian and mammalian are not, and the presence of avian transferrin can tie up all iron and keep it away from mammalian cells.
  5. Trace elements: 'Standard' medium formulations tend to contain few, if any, deliberately added trace elements. It is therefore often necessary to add a trace element supplement for consistent defined medium growth.
  6. Biotin: Certain standard media, such as Eagle's MEM and Dulbecco's modified Eagle's medium lack biotin. Serum, even when dialyzed, contains bound biotin, as well as fatty acids, which are one of the major end products of biotin-catalyzed metabolism. Serum replacement in such media requires the addition of biotin.
  7. Nutrients in serum: Serum can also serve as a source for a variety of other nutrients that must be replaced for serum-free growth. One example is choline, which is at inadequate levels for serum-free growth in some standard media such as Dulbecco's. A variety of other nutrients, including many essential amino acids, can also become rate-limiting for growth in the absence of serum.

Hormones and Growth Factors

There is major individuality in cellular requirements for hormones and growth factors. This was already clearly evident for continuous lines at least 10 years ago (Barnes and Sato, 1980), and has been further confirmed by studies on various types of normal cells (Ham, 1984). The current literature contains frequent reports of the discovery of new growth factors, and there is substantial evidence that still more remain to be discovered. The approach to serum-free media, pioneered by Gordon Sato's laboratory, was based on replacement of serum with known hormones and growth factors, rather than identification of the actual growth factors that were present in the serum. Evidence is now accumulating that serum contains additional potent growth factors that have not yet been isolated and characterized. Also, the study of oncogenes and proto-oncogenes has identified a substantial number of apparent growth factor receptors for which the corresponding growth factors have not yet been identified. Thus, for some cell types, the use of defined media probably must await the discovery of additional growth factors. The increased availability and competitive pricing of growth factors produced with recombinant DNA technology is rapidly expanding the possibilities for the use of defined media in toxicity testing and other applications.

Serum-free cell culture has become widely possible because of recently developed purification procedures and general availability of peptide growth factors (Barnes and Sirbasku, 1987; Barnes et al, 1984) and cell adhesion proteins (Barnes, 1984). The number of hormones used in serum-free cell culture is large (Table V), representing the complexity of growth regulation in vivo. Hormones include peptide growth factors, prostaglandins, classical endocrine messengers such as pancreatic and pituitary peptides, vasopressin, steroid and thyroid hormones, peptide growth factors such as epidermal growth factor, neuroendocrine or gastrointestinal peptides such as bombesin and hypothalamic releasing factors like luteinizing hormone releasing hormone (Barnes, 1984; Barnes, 1985; Barnes and Sato, 1980). Almost all cells require insulin. Although some of the hormones are relatively cheap commercially, others, particularly the peptide growth factors, may be quite expensive. Recent progress in large scale production of recombinant products and peptide synthesis is slowly reducing the price for some of these.

Table V: Hormones

Pancreatic hormonesInsulin, somatostain, glucagon
Peptide growth factorsPlatelet-derived growth factor, insulin-like growth factors I/II, connective tissue-activating factor, epidermal growth factor, fibroblast growth factor, acidic/basic nerve growth factor, transforming growth factor beta
Steroid hormonesGlucocorticoids, estrogens, androgens, progestins
Products of arachidonic acid metabolismProstaglandins, thromboxanes, leukotrienes, prostacyclin
Pituitary hormonesAdrenocorticotropin, growth hormone, follicle-stimulating hormone thyrotropin, luteinizing hormone, prolactin and other related serum gonadotropins, thyroid and parathyroid hormones
Leukocyte productsInterferons, tumor necrosis factor, interleukins

The Advantages and Disadvantages of a Completely Defined Medium


Cell cultures maintained in basal nutrient culture media supplemented with bovine, fetal bovine or calf serum are well established conventional models. Although modern basal nutrient medium formulations allow the serum in the range of 5% to 20% to be used as a supplement, this biological fluid still represents a significant contribution of undefined components to the culture medium. Among the approaches taken to eliminate this problem were attempts to isolate growth-promoting factors from serum and the selection of variant cells capable of growing in little or no serum. Isolation of growth-promoting activities from serum was not successful on a widely applicable scale, largely because of difficulties in obtaining purified material for bulk use and because serum activities are synergistic among multiple components. Selection of cells in unsupplemented basal nutrient medium was of limited value because the cells were altered from the parents, producing growth factors mimicking serum. The successful solution to this problem was found in the development of better defined media utilizing supplements from multiple sources to replace serum.

The use of defined media provides technical advantages over serum-containing media at the basic and applied science levels, and also allows the growth of some cell types and the expression of some cell properties that cannot be duplicated by traditional culture procedures. Development of a serum-free medium formulation may seem complicated, considering the number of components that might be tested and the interactions among these components, but the medium optimization process can be approached in a systematic and logical manner (Bettger and Ham, 1982; Bottenstein et al., 1979; Ham and McKeehan, 1979). Commonly used basal medium formulations include Ham's F12, Dulbecco-modified Eagle's medium, RPMI 1640, MCDB media, and combinations of these media (Mather and Sato, 1979). To these are added hormones, binding proteins, attachment factors, extracellular enzymes, and supplemental nutrients (Barnes, 1987).

Serum-free medium allows precise control of the extracellular environment for experiments that are difficult or impossible in conventional serum-containing medium. This permits studies of hormones, drugs or toxins in the absence of serum components that may mimic, antagonize, inactivate or act synergistically with the factors of interest and allows identification of serum components that may affect biological responses, using the serum-free culture as an assay mode. An additional advantage is the selectivity of many serum-free media, allowing the growth of only a specific cell type and preventing overgrowth from other cell types. Furthermore, it is possible to design serum-free media for the selection of malignantly transformed cells or of mutants that are altered in response to toxins, carcinogens or hormones (Chiang et al, 1985; Taub et al, 1981). A major contribution of this approach is the culture of cell types that cannot be grown in conventional, serum-containing medium. Examples include normal rat thyroid, bovine parathyroid, human mammary epithelia, and diploid mouse embryo cells (Allegra and Lippman, 1978; Brandi et al, 1986; Hammond et al, 1984; Loo et al,1987).

The serum-free approach is widely applicable. A number of differentiated cell lines derived from the endocrine system have been propagated that retain tissue-specific responses to endocrine factors, and also exhibit the ability in serum-free culture to synthesize the appropriate endocrine hormones or other modulators of growth and differentiation. Other epithelia cell types of interest to toxicologists for which serum-free media have been developed include kidney, liver, bronchial epithelia, and cells of neural tissue origin. Kidney cultures in serum-free media exhibit characteristic renal transport properties (Palmoski et al, 1991). Serum-free formulations have been developed for fibroblasts and lymphoid cells from humans and other species allowing direct toxicological evaluations and individualized analyses.

Although serum-free media often provide a more reproducible environment for cells, there are also disadvantages. Often these media are more expensive than conventional media, principally because of the cost of hormones and growth factors. In addition, serum-free culture can be tedious and variable. Attention must be paid to length and temperature of storage of medium, supplements, and water quality; disposable tubes and pipettes should be used to avoid contamination from washing and loss of supplements at low protein concentration. Often cells are more sensitive to routine passaging or other manipulations in serum-free media than in conventional protocols. In addition, complicated relationships exist among the actions of the medium supplements. Synergisms are common and some factors may exert positive or negative effects, depending on the cell type or experimental protocol. Furthermore, some serum-free medium supplements probably are serving multiple functions. Such complications may make serum-free experiments difficult to interpret.


There is a critical need for more precise definitions of terms such as "defined" and "serum-free". These terms are currently used very loosely and inconsistently. In the broadest sense, "serum-free" implies only the absence of serum and says nothing about what else may be in the medium. Thus, a "serum-free" medium can contain crude serum fractions such as fetuin, or even totally undefined additives, such as chicken embryo extract or bovine pituitary extract. More precise definitions should be established and a uniform usage encouraged.

"Defined" is even more difficult to interpret. The term implies absolute knowledge of the overall chemical composition of the medium. This is an abstract ideal that does not exist in real life. All chemicals used in medium preparation contain some degree of impurity. In addition, at least some of the impurities (e.g., inorganic trace elements) are almost certainly essential requirements for cellular growth under absolutely defined conditions (Ham, 1981).

Thus, the term "completely defined medium" is meaningless and should be replaced with a more moderate and realistic term such as "highly defined medium".

Sensitivity and False Positives

Serum protein tends to have a "buffering" effect on a wide variety of substances, depending on the chemical properties of the test agent, due to reversible and irreversible protein binding. This is also true for large amounts of various proteins such as serum albumin or transferrin that are used in some "serum-free" media. "Defined" media are usually prepared with relatively low levels of growth factors and other essential proteins to minimize such effects. Cells grown or maintained in "defined" media, therefore, tend to exhibit toxic responses at lower threshold levels of test substances than cells in serum-containing media.

Depending on the circumstances, this can be viewed as either good or bad. An orally taken drug reaches the target organ after having achieved solution in the blood. Protein is an important component of this tissue. Thus, protein binding can be expected in vivo. Sensitivity of the assay is increased sharply in serum-free cultures, but so is the risk of false positives. In assays that are examining systemic toxicity of ingested or injected substances, the presence of serum in the culture medium may more closely approximate the protective effects of proteins in the circulating plasma. However, for tests of toxicity under conditions where the test substances come into direct contact with tissues such as the cornea or conjunctiva of the eye, the skin or the respiratory epithelium, a reasonably defined serum-free medium may be more appropriate.

Direct data on this issue are still rather limited. However, in studies of a group of test substances provided by the Soap and Detergent Association (SDA), in vitro testing with serum-free media provided a better correlation with Draize ocular irritancy scores than in vitro testing with serum-containing media (Shopsis and Eng, 1988; Shopsis, 1989). The impact of possible false-positive tests, which are more likely with defined media, versus false negatives, which are more likely with serum-containing media, varies with the primary objectives of the test program. False positives increase the margin of safety in circumstances where rejection of a few too many of the candidate substances is not a serious problem. On the other hand, in a search for a modified form of a useful substance whose toxicity or irritancy must be reduced by manipulation of molecular structure, they may result in rejection of potentially valuable candidate substances.

Technical Difficulty

A major disadvantage of typical defined medium systems is the added technical difficulty involved in their use. They require precise formulated media and skilled technicians who have learned all of the special quirks of such systems. Cells growing in the defined media are often pushed close to their biosynthetic and physiological limits, and tend to be extremely unforgiving of any defects in medium composition, as well as rough handling, harsh trypsinization, or any other deviations from established procedures. The use of serum-containing media, on the other hand, provides a substantial margin of safety in many of these areas. However, it also provides the cells with a margin of safety against test substances, as discussed above.

Extracellular Matrix (ECM) Components Used to Culture Specific Cell Types


Because this topic is so broad, it is necessary to take into consideration all aspects of the cellular substrate, beginning with the nature and quality of tissue culture plastic surfaces.

Variability of Cell Culture Plastic

There have been numerous accounts of inconsistencies in cell growth and test results being traceable to inconsistency of cell culture plastic surfaces. In some cases these could be traced to quality control problems that resulted in lot-to-lot variation from a single manufacturer, whereas in other cases there appeared to be systematic differences from one manufacturer to another. Procedures for the generation of cell culture plastic surfaces are proprietary; there are no generally accepted standards for a "good" surface other than customer satisfaction with the final product. Thus, there is no firm basis for predicting performance without testing the product for growth of the cell type in question.

Another current variable is the introduction by various manufacturers of modified culture surfaces that are claimed to be superior for various types of cells. Because the modifications are also proprietary, it is necessary for each end user to evaluate the usefulness of each type of surface for each cell type with very little information from the manufacturer other than advertising claims about the superiority of new products.

Porous Membranes

The use of porous membranes as substrates for cellular growth has been in the literature for a long time, but is now becoming more widespread, with convenient systems offered commercially by several different companies (Millipore, for example). Systems of this sort permit epithelial cells to form solid sheets and become fully polarized with basolateral feeding and apical secretion or stratification and terminal differentiation. The use of such substrates provides a means for testing of surface exposure to toxic agents that more closely reflect the arrangement of cells in the body. It also provides a means for testing whether test agents have disrupted tight junctions in a confluent cellular monolayer growing on the membrane (for example, radiolabeled insulin transport or electrical resistance determination).

Transmembrane Cocultures

In the intact body, individual cell types do not deal with toxic substances in isolation. Instead there is an organized community of cells that responds, sometimes in complex ways that involve communication from one cell type to another. The use of porous membranes with, for example, epithelial cells on one side and stromal cells on the other side provides an in vitro model that is far more like the in vivo situation. It is likely that many responses of epithelial cells in vivo are influenced significantly by stromal cells on the other side of the basal lamina. Thus, for example, stromal cells could metabolically activate a carcinogen that then has its primary effect on epithelial cells. A possible example of such an effect, which has never received an adequate follow-up study, is the apparent metabolic activation of DMBA by mammary stromal cells, which do not themselves become transformed, but may contribute significantly to mammary carcinogenesis by DMBA (Menon et al, 1987).

Cells can be cocultured in three possible ways. First, both cell types can be cultured on the same face of the membrane to allow feeding from underneath. An example would be keratinocytes layered onto dermal fibroblasts with the filter acting as an air-culture medium interface as a more appropriate system for studying skin toxicity. Secondly, each cell type might be cultured on opposite faces of the membrane, the pore size determining the degree of cellular contact. Such a system has been usefully applied to the study of cell contact and differentiation in embryo development, but has not yet been described as a system for toxicity testing. Finally, one cell type, hepatocytes for example, might be established in the culture dish and the other, cells from the putative target organ, on the filter in a separate culture dish. Then, at the appropriate time the filter insert could be transferred to the dish with hepatocytes in the presence of the test agent to expose the target cells to parent compound and metabolites. This can be done with permeable collagen membranes or with a variety of commercially available transmembrane culture units. Transmembrane cocultures are likely to become very important for toxicity testing. Membranes fused to plastic rings suitable for tissue culture are now available from Nunc, Wheaton Scientific (Anocell porous tissue culture insert), Millipore (Millipore-HA opaque membrane insert (Palmoski et al, 1991 and Millicell-CM clear membrane insert), and Costar (Transwell insert, either opaque or collagen-treated and transparent).

Extracellular Matrix Components

As with peptide growth factors, recent commercial advances have been made in the large scale production of extracellular matrix components. In addition, major advances have been made in the biochemistry and cell biology of this diverse group, allowing considerable improvement in serum-free medium formulations and uses (Barnes, 1984). Soon after serum-free formulations began to appear, experiments with the replacement of serum showed that other classes of serum components were important for cell growth in vitro in addition to hormones and nutrients. Cell adhesion proteins such as fibronectin that promote attachment and spreading of cells on plastic or glass are critical for the growth of some cell types (Barnes, 1985; Barnes and Sato, 1980). Commercially available mixtures of extracellular matrix components are available. A common one is a mixture of collagen, laminin, entactin, and proteoglycans isolated from the EHS mouse tumor (MatrigelTM). Although this material has the advantage of being cheap and more effective than the individual components in some cases, it has the disadvantage of being undefined and possibly irreproducible.

Although cell adhesion proteins that are components of basement membrane (collagens, laminin) are not present in serum, these extracellular matrix components elicit effects on cells in serum-free culture by mimicking the extracellular environment of the cell (Barnes, 1984; Barnes and Sato, 1980). The relationship between adhesion factors and other medium components is complex. In some situations hormones may elicit extracellular matrix production, or extracellular matrix may alter cellular response to a hormone. Mammary epithelia in vitro have been particularly informative in understanding the relationship between cell-substratum interactions and cell-hormone interactions (Imagawa et al, 1982; Salomon et al, 1981). Synthesis of collagen, which is active on these cells, is regulated by growth factors. Another component that, like collagen, has been used for many years in cell culture as an attachment factor is fetuin, a partially purified fraction of fetal calf serum. The activity of fetuin is probably due to several contaminants, including fibronectin and serum spreading factor, as well as peptide growth factors and binding proteins for hormones.

Artificial Approaches to Attachment

A number of attempts have been made to improve cellular attachment through synthetic means. For at least some cell types, coating of standard cell culture plastic with polylysine (which is commercially available from several distributors) can improve both attachment and growth. An adhesive substance derived from muscles is sold commercially under the trade name-Cell-Tak (Table VI lists the more commonly used cell adhesion products).

Table VI: Some Commercially Available Cell Adhesion Products for Tissue Culture

Cell-substrate adhesion productSupplier
Cell-TakBiopolymers Inc., 309 Farmington Avenue, CT 06032
Pronectin-FProtein Polymer Technologies, 10655 Sorrento Valley Road, San Diego, CA 92121
CR-FibronectinCollaborative Research, Inc., Two Oak Park, Bedford, MA 01730-9902
LamininCollaborative Research, Inc.
MatrigelCollaborative Research, Inc.
Collagen Type I (rat tail) - powderCollaborative Research, Inc.
Collagen Type IV (mouse) - powderCollaborative Research, Inc.
Poly D LysineCollaborative Research, Inc.
Collagen Type I (bovine) - solutionOrganogenesis, 83 Rogers Street, Cambridge, MA 02142
Collagen-glycosaminoglycan discsBiomat Corporation, 57 Raleigh Road, Belmont, MA 02178

Cell culture dishes with an altered surface charge distribution are sold under the trade name of Primaria (Falcon). There are claims that all of these products improve cellular adhesion and growth. However, the effects appear to be strongly influenced by cell type and by the overall composition of the culture medium.

Trypsin Inhibitors

There is a danger of confusion between the attachment-promoting effects of serum and matrix molecules and effects that are due to neutralization of trypsin. Most investigators who grow cells in serum-containing media are not very careful about avoiding the introduction of trypsin into the medium along with the cellular inoculum. Serum contains potent trypsin inhibitors that neutralize the residual trypsin and no harm is done. However, under serum-free conditions, the same amount of trypsin will block cellular attachment or cause the cells to release again after they have initially attached. Thus, serum-free growth can often be achieved without specific attachment factors through use of greatly reduced amounts of trypsin, through blocking of trypsin internalization by low temperature when feasible, and by careful washing of subcultured cells. In some cases, use of alternative trypsin inhibitors may be necessary, but they are often rather impure and tend to reduce the level of definition of "defined" media.

Preservation of the Differentiated Cell in Culture

Selection of Cell Types

Modern cell biology texts recognize the existence of more than 200 different cell types within the human body. Because of extensive individuality of requirement and the fact that adequate culture systems have been developed for only a few of these cell types, it becomes absolutely essential to begin by selecting the cell type(s) with which to be worked. In addition, one must recognize the multistep nature of differentiation and select the specific state(s) of differentiation that are the most desirable for use in a particular type of toxicity testing. It may also be necessary to specify a required level of maintenance of specific metabolic activities that are important to the test system. Only after this selection process has been completed, can development of a culture system designed to maintain all of the desired states be seriously undertaken. The basic principles of individuality of cellular requirements generally apply at least as rigorously to differentiation and maintenance of metabolism in culture as they do to growth.

Important Variables

Depending on the cell type, the differentiated state and the in vitro environment any of the following aspects of the culture system may be critical variables: overall composition of the nutrient medium, concentrations of specific regulatory ions (with special emphasis on calcium), hormones, presence or absence of growth factors, presence of specific inducers or activators, absence of specific inhibitors, basolateral feeding, ability to form tight junctions or gap junctions between adjacent cells, ability to reorganize into histotypic structures, presence of a flexible substrate, presence of specific extracellular matrix components, presence of substrates for inducible enzymes, oxygen tension, carbon dioxide tension, a suitable liquid-gaseous interface, appropriate pH, appropriate osmolarity, cross-talk with additional cell types, adequate removal of toxic waste products (etc.).

Thus, as with growth condition, it becomes necessary to tailor the culture system specifically for promotion of the differentiated and metabolic states that are needed for optimal toxicity testing. It is also important to recognize that a two step process may be needed in which the cells are first grown up to an appropriate population in an optimized growth medium, and then switched to a quite different medium that causes them to enter the desired states of differentiation and metabolism.

Primary Cultures of Hepatocytes: An Example

If cell culture systems are to have relevance as toxicity testing models, it is important that the chosen cellular systems retain metabolic features similar to the tissue in situ or in vivo. For example, the primary culture of rat hepatocytes has historically led to the loss of certain differentiated functions (e.g., cytochrome P-450 activity); the addition of certain medium components will help to reduce or minimize the loss of that important liver function. It is the function of detoxification and processing of xenobiotics that makes the liver particularly susceptible to toxicity. The endoplasmic reticulum of liver cells contains enzymes of the cytochrome P-450 monoxygenase system responsible for the conversion of chemicals to more polar forms that can be further conjugated to achieve sufficient hydrophilicity to make urinary and biliary excretion attainable. Therefore, there is little doubt that metabolism usually serves as a protective mechanism of physiologic significance against the toxic effect of xenobiotics on vital cellular functions. However, the toxicity of some chemicals results from their metabolic conversion to derivatives that can alter tissue macromolecules, a process that has come to be known as metabolic activation. In recent years, a significant body of information has accumulated on the role of metabolism in the activation of xenobiotics to toxic metabolites that may be carcinogenic, mutagenic or teratogenic, or that may be responsible for producing tissue injury and necrosis.

The relevance of primary culture systems, such as rat hepatocytes, must be established by rigorous comparison with the xenobiotic-metabolizing capability of the intact liver. In general, it has been shown qualitatively that various biotransformation reactions can be maintained in liver cell cultures but the actual level of activity is lower than in freshly isolated hepatocytes and whole liver. Numerous attempts have been made to prevent or diminish the dramatic loss of cytochrome P-450 activity when hepatocytes are grown in culture. Modifications such as culturing the cells on floating collagen membranes, growing the cells on specially treated culture dishes or supplementing the culture medium with special substrates or hormones, such as those described above, have partially resolved the problem.

Thus, a major objective of several investigators has been to develop a system of cultured hepatocytes that retains cytochrome P-450 levels over an extended period of time. Initial attempts were unsuccessful in maintaining cytochrome P-450 levels in postnatal hepatocyte cultures for several days in culture (Acosta et al, 1979; Acosta et al, 1985). A medium enriched with several hormones (insulin, hydrocortisone, thyroxine, and testosterone) did not prevent the decline in cytochrome P-450 when compared to those found in freshly isolated hepatocytes for seven days in culture (Michalopoulos et al, 1976; Acosta et al, 1979). By supplementing the culture medium with insulin, hydrocortisone, and nicotinamide and by removing cystine, the cultures of rat postnatal hepatocytes retained cytochrome P-450 levels, over several days, which were essentially the same (0.5 to 1.5 nmol per mg microsomal protein) as freshly isolated hepatocytes (Nelson and Acosta, 1982; J. Klimczk, personal communication). In addition, pretreatment of the cultured hepatocytes with phenobarbital resulted in a 2-fold increase of cytochrome P-450 when compared to untreated controls. Similar results were obtained in adult rat hepatocyte cultures grown in cystine-free medium by Paine et al (1982). Hence, the ability to maintain high levels of P-450 in culture allows one to study more confidently the metabolism of a xenobiotic and potential toxicity of the formed metabolites in a hepatocyte culture model (Fry and Bridges, 1979; Decad et al, 1977, Guzelian et al, 1977; Dickins and Peterson, 1980, Davila et al, 1991).

Another important aspect to consider in maintaining the cellular metabolism of liver cells in vitro is the prevention of overgrowth of the parenchymal hepatocytes with fibroblasts and other nonparenchymal liver cells (e.g., endothelial cells) (Leffert and Paul, 1972; Acosta et al, 1978; Wenzel et al, 1970; Smith et al, 1986; Gilbert and Migeon, 1975). The fibroblasts replicate rapidly in culture while the hepatocytes are nonreplicating or slowly replicating cells in culture. Thus, even if one is able to maintain the cellular metabolism of the hepatocytes, the presence of fibroblasts which can quickly overgrow and crowd out the hepatocytes will result in a culture system that is no longer reflective of the in vivo situation. The use of arginine-free, ornithine-enriched culture medium allows the selective growth of liver parenchymal cells and the inhibition of fibroblast overgrowth. Arginine is an essential amino acid for the growth and survival of cells in culture. The removal of arginine from the culture medium does not allow the fibroblasts to replicate in culture, while the hepatocytes have the enzymatic machinery to synthesize arinine from ornithine (a capability lacking in fibroblasts) and thus are able to grow and survive in culture. Similar manipulations of the culture medium have resulted in the selection of viable, functional parenchymal cells isolated from the heart, kidneys, central nervous system, and other tissues (Acosta et al, 1985; Belleman, 1980; Waymouth, 1978; Smith et al, 1986; Kisby and Acosta, 1987).


  • Acosta, D., Anufuro, D. and Smith, R.V. (1978). Primary monolayer cultures of postnatal rat liver cells with extended differentiation functions. In Vitro 14: 428-436.
  • Acosta, D., Anuforo, D.C., McMillin, R., Soine, W.H. and Smith, R.V. (1979). Comparison of cytochrome P-450 levels in adult rat liver, postnatal rat liver, and primary cultures of postnatal rat hepatocytes. Life Science 25:1413-1418.
  • Acosta, D., Sorensen, E.M., Anufuro, D.C., Mitchell, D.B., Ramos, K., Santone, K.S. and Smith, M.A. (1985). An in vitro approach to the study of target organ toxicity of drugs and chemicals. In Vitro Cellular and Developmental Biology 9: 485-504.
  • Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J.D. (1989). Molecular Biology of the Cell, Second Edition (Appendix to Chapter 17: Cells of the Adult Human Body: A Catalog, pp. 995-997). Garland Publishing Inc., New York.
  • Allegra, J.C. and Lippman, M.E. (1978). Growth of a human breast cancer cell line in serum-free hormone-supplemented medium. Cancer Research 38:3823-3829.
  • Barnes, D. (1984). Attachment factors in cell culture. In Use of Serum-Free and Hormone-Supplemented Media (ed. Mather, J.P.) pp. 915-237. Plenum Press, New York.
  • Barnes, D. (1985). Nutritional and hormonal requirements of cells in culture. World Review of Nutrition and Dietetics 45: 167-197.
  • Barnes, D.W. (1987). Serum-free animal cell culture. Biotechiques 5: 534-545.
  • Barnes, D.W. and Sato, G.H. (1980) Serum-free cell culture: A unifying approach. Cell 22: 649-655.
  • Barnes, D. and Sato, G. (1980). Methods for growth of cultured cells in serum-free media. Analytical Biochemistry 102: 255-270.
  • Barnes, D. Sirbasku, D. (eds.) (1987). Peptide Growth Factors: Methods in enzymology. Vols. 146-147. Academic Press, Inc., New York.
  • Barnes, D., Sirbasku, D. and Sato, G. (eds.) (1984). Cell Culture Methods for Molecular and Cell Biology. Vols. 1-4. Alan R. Liss, Inc., New York.
  • Belleman, O. (180). Primary monolayer culture of liver parenchymal cells and kidney cortical tubules as a useful new model for biochemical pharmacology and experimental toxicology. Archives of Toxicology 44: 63-84.
  • Bettger, W.J. and Ham, R.G. (1982). The nutrient requirements of mammalian cells grown in culture.
  • Bottenstein, J., Hayashi, I., Hutchings, S., Masui, H., Mather, J., McClure, D., Ohasa, S., Rizzino, A., Sato, G., Serrero, G., Wolfe, R. and Wu, R. (1979). The growth of cells in serum-free hormone-supplemented media. Methods in Enzymology, Vol. 45 (eds. Jakoby, W.B. and Pastan, I.H.) pp.94-109. Academic Press, New York.
  • Brandi, M., Fitzpatrick, L., Coon, H. and Aurbach, G. (1986). Bovine parathyroid cells: culture maintained for more than 140 population doublings. Proceedings of the National Academy of Science USA 83: 1709-1714.
  • Chiang, L., Siluntzer, J., Pipas, J. and Barnes, D. (1985). Selection of transformed cells in serum-free media. In Vitro Cell Developmental Biology 21: 707-712.
  • Davila, J.C., Davis, P.J. and Acosta, D. (1991). Changes in glutathione and cellular energy as potential mechanisms of papaverine induced hepatotoxicity culture. Biochemical and Biophysical Research Communications 78: 279-287.
  • Decad, G.M., Hsieh, D.P. and Byard, J.L. (1977). Maintenance of cytochrome P-250 and metabolism of aflatoxin B1 in primary hepatocyte culture. Biochemical Biophysical Research Communications 78: 279-287.
  • Dickins, M. and Peterson, R.E. (180). Effects of hormone supplemented medium on cytochrome P-450 content and monooxygenase activities of rat hepatocytes in primary culture. Biochemical Pharmacology 29: 1231-1238.
  • Fry, J.R. and Bridges, J.W. (1979). Use of primary hepatocyte cultures in biochemical toxicology. Review of Biochemical Toxicology 1: 201-247.
  • Gilbert, S.F. and Migeon, B.R. (1975). D-Valine as a selective agent for normal human and rodent cells in culture. Cell 5: 11-17.
  • Guzelian, P.S., Bissell, M. and Meyers, U.S. (1977). Drug metabolism in adult rat hepatocytes in primary monolayer culture.
  • Ham, R.G. (1981). Survival and growth requirements of nontransformed cells. Handbook of Experimental Pharmacology 57: 13-88.
  • Ham, R.G. (1981). Introduction: cell growth requirements - the challenge we face. The Growth Requirements of Vertebrate Cells In Vitro (eds. Waymouth, C., Ham, R.G. and Chapple, P.J.) pp. 1-15. Cambridge University Press, Cambridge.
  • Ham, R.G. (1982). Importance of the basal nutrient medium in the design of hormonally-defined media. Cold Spring Harbor Conferences on Cell Proliferation 9: 39-60.
  • Ham, R.G. (1984). Methods for formulation of basal nutrient media. Cell Culture Methods for Molecular and Cell Biology Vol. 1 (eds. Barnes, D., Sirbasku, D. and Sato, G.) pp. 3-21. Alan R. Liss, Inc., New York.
  • Ham, R.G. (1984). Selective media. Cell Separation: Methods and Selected Applications Vol. 3 (eds. Pretlow, T.G., II and Pretlow, T.P.) pp. 209-236. Academic Press, New York.
  • Ham, R.G. and McKeehan, W.L. (1978). Nutritional requirements for clonal growth of nontransformed cells. Nutritional Requirements of Cultured Cells (ed. Katsuta, H.) pp. 63-115. Japan Scientific Societies Press, Tokyo.
  • Ham, R.G. and McKeehan, W.L. (1979). Media and growth requirements. Methods of Enzymolology 58: 44-93.
  • Hammond, R., Ham, R. and Stampfer, M. (1984). Serum-free growth of human mammary epithelial cells: rapid clonal growth in defined medium and extended serial passage with pituitary extract. Proceedings of the National Academy of Science USA 81: 5434-5439.
  • Imagawa, W., Tomooka, Y. and Yandi, S. (1982). Serum-free growth of normal and tumor mouse mammary epithelial cells in primary cultures. Proceedings of the National Academy of Science USA 79: 4074-4077.
  • Kisby, G.E. and Acosta, D. (1987). Cytotoxic effects of aluminium in hippocamapla, cerebellar, and astrocyte cell cultures. In Vitro Toxicology 1: 85-102.
  • Leffert, H.L. and Paul, D.J. (1972). Studies on primary cultures of differentiated fetal liver cells. Cell Biology 52: 559-568.
  • Loo, D.T., Fuquay, J.I., Rawson, C.L. and Barnes, D.W. (1987). Extended Culture of Mouse Embryo Cells without Senescence: Inhibition by Serum. Science 236: 200-202.
  • Mather, J.P. and Sato, G.H. (1979). The growth of mouse melanoma cells in hormone-supplemented serum-free medium. Exp. Cell. Res. 120:191-200.
  • Mather, J.P. and Sato, G.H. (1979b). The use of hormone-supplemented serum-free media in primary cultures. Experimental Cellular Research 124: 215-221.
  • Menon, R.S., Van der Haegen, B.A., Bartley, J.C. and Ham, R.G. (1987). DMBA metabolism by cultured rat mammary stromal cells. In Vitro Cellular and Developmental Biology 12: 13A (Abstract No. 8).
  • Michalopoulos, G., Sattler, G.L. and Pitot, H.C. (1976). Maintenance of microsomal cytochromes b5 and P-450 in primary cultures of parenchumal liver cells on collagen membranes. Life Science 18: 1139-1144.
  • Nelson, K.F. and Acosta, D. (1982). Long-term maintenance and induction of cytochrome P-450 in primary cultures of rat hepatocytes. Biochemical Pharmacology 31: 2211-2214.
  • Paine, A.J., Hockin, L.J. and Allen, C.M. (1982). Long term maintenance and induction of cytochrome P-450 in rat liver cell culture. Biochemical Pharmacology 31: 1175-1178.
  • Palmoski, M.J., Masters, B.A., Ford, S.M. and Flint, O.P. (1991). Growth characteristics and measurement of confluency of rabbit primary proximal tubule kidney cells grown on Millicell-HA membrane filters. Journal Tissue Culture Methods (accepted for publication).
  • Salomon, D.S., Liotta, L. and Kidwell, W.R. (1981). Differential response to growth factors by rat mammary epithelium plated on different collagen substrata in serum free medium. Proceedings of the National Academy of Science USA 78: 382-386.
  • Shopsis, C. (1989). Validation study: Ocular irritancy prediction with the total cell protein, uridine, uptake, and neutral red assays applied to human epidermal keratinocytes and mouse 3T3 cells. Alternative Methods in Toxicology, Vol. 7 In Vitro Toxicology: New Directions (ed. Goldberg, A.M.) pp. 273-286. Mary Ann Liebert, Inc., New York.
  • Shopsis, C. and Eng, B. (1988). Poster A13: In vitro ocular irritancy prediction: Assays in serum-free media correlate better with in vitro data. Alternative Methods in Toxicology (ed. Goldberg, A.M.) pp. 253. Mary Ann Liebert, Inc.: New York.
  • Smith, M.A., Acosta, D. and Bruckner, J.V. (1986). Development of a primary culture system of rat kidney cortical cells to evaluate the nephrotoxicity of xenobiotics. Food Chemical Toxicology 24: 551-556.
  • Taub, M., Chuman, L., Rindler, M., Saier, M. and Sato, G. (1981). Alterations in growth requirements of kidney epithelial cells in defined medium associated with malignant transformation. Journal of Supramolecular Structure and Cellular Biochemistry 15: 63-72.
  • Waymouth, C. (1978). Studies on chemically defined media and nutritional requirements of cultures of epithelial cells. Nutritional Requirements of Cultured Cells (ed. Katsuta, H.) pp. 39-61. University Park Press: Baltimore.
  • Wenzel, D.G., Wheatley, J.W. and Byrd, G.D. (1970). Effects of nicotine on cultured rat heart cells. Toxicology and Applied Pharmacology 17: 774-785.


Robert Hay1, Warren Schaeffer2, Oliver Flint3, June Bradlaw4 and John Frazier5

1American Type Culture Collection, 12301 Parklawn Drive, Rockville, MD 20852; 2University of Vermont Medical College, Department of Medical Microbiology, Given Building, Burlington, VT 05405;3Investigative Toxicology, Bristol-Myers Squibb, P.O. Box 4755, Syracuse, NY 13221; 4Division of Toxicological Studies, HFF-162, Food and Drug Administration, 200 C Street, S.W., Washington , DC 20204; 5The Johns Hopkins Center for Alternatives to Animal Testing, 615 North Wolfe Street, Room 7033, Balitmore, MD 21205


Research today is becoming more and more interdisciplinary, and nowhere is this more obvious than within the fields of tissue and cell culture. As a result, standardization is of critical importance when it comes to replication of experimental models involving such components as media, sera, cultureware, cell lines or primary cultures described in the literature. Each of these components will be discussed. Misuse of terminology is also considered here because, potentially, it is a serious source of problems, especially when relevant areas of in vitro research gain prominence in the regulatory arena and become the basis for the definition of officially sanctioned protocols.

Media and Sera

Very little has been done to generate uniform standards for media and sera. Many laboratories, including the cell culture department at the American Tissue Culture Collection (ATCC), establish their own criteria for quality control and in-house standardization of media used (Hay et al, 1988). Very often a substantial variety of formulations may be needed.

Factors to be considered before choosing a particular medium formulation along with details concerning important supplements and environmental variables have been discussed thoroughly by Ham and McKeehan (1979). Because short wavelengths of light induce chemical modifications such as peroxides in cultural media (Ham and McKeehan, 1979; Hay, 1985) and because of the relative lability of other medium components, care must be taken to prepare and store media according to manufacturers' recommendations. In turn, manufacturers must become more responsible for the provision of accurate descriptions of the best methods for media preparation and storage.

At the ATCC, wherever possible, dehydrated media are now obtained commercially for reconstitution and final manufacturing. The appropriate amount of pretested serum, NaHCO3 and other supplements are added to the reconstituted fluid, and the solutions are filter-sterilized and stored at 4°C, pending completion of quality-control tests (Hay, 1985).

Selection of a Suitable Serum Lot

Methods for selecting serum lots to be used in culture media vary among laboratories depending upon the particular anticipated research and practical applications involved. Experience indicates that repeat tests for bacteria, fungi, mycoplasma, and viruses (Hay, 1985) are necessary for an institution such as the ATCC. The problem of BVD virus contamination in fetal bovine serum has been reviewed elsewhere (Hay, 1985). Checks for bacteriophage contamination are not applied routinely because the sera utilized have been collected aseptically.

Testing of Media and Sera Lots

The use of clonal analyses and quantitative assessment of growth at varying inoculation densities provides simple yet reliable indications of serum quality. Indices of both cell plating and growth promotion are reflected in the final result. One can now utilize automated colony counters to quantitate both plating efficiencies and colony growth areas. Under appropriate conditions these should give a reliable index of the properties of medium and serum used (Hay, 1985).


At the ATCC, quality control trials on plasticware are routine and lot-to-lot variation is observed. The assay method used is that described above for media and sera, namely clone-forming efficiencies as quantitated with an automated counter (Hay, 1985). This is an important consideration for the scientist responsible for in vitro studies, since the plating efficiency of the cell type being studied will be critical to the endpoint measured.

Cell Lines

Whenever cells are grown in culture there is a risk of contamination and overgrowth by cells of another individual or species. The scientific literature is replete with articles and reviews documenting both interspecies and intraspecies cell culture cross-contaminations (Hay, 1988; Nelson Rees et al, 1981). The loss of time and research funds resulting from the use of such lines is incalculable.

Bacterial and fungal contaminations present an additional problem but one which is often, though emphatically not always, readily identified. Mycoplasmal contamination is a significantly greater concern because it is frequently invisible. The metabolic and biological effects of these microorganisms in cell cultures can negate research or diagnostic findings entirely (Del Guidice and Gardella, 1984; McGarrity et al, 1982).

Other serious problems relating to quality control exist. However, by applying relatively simple but critical testing procedures or relying on service organizations to authenticate cultures, one can avoid the potential pitfalls associated with use of cell lines obtained or characterized casually. The following is an outline of steps which can be taken to increase the reliability of cell culture stocks maintained in a collection.

The Seed Stock Concept

Even for comparatively small cell banks, it is recommended that a seed stock be created for each specific line. The cell suspensions in such stocks are the most thoroughly characterized and are thus retained for generations of working or stock distribution as the bank holdings are utilized through the years.

Generally, starter cultures or ampoules are obtained from the originator, and progeny are propagated according to instructions to yield the first "token" freeze. Cultures derived from such token material are then subjected to critical characterizations. If these tests suggest that further efforts are warranted, the material is expanded to produce the seed and distribution stocks. The major characterization efforts are applied to cell populations in the seed stock of ampoules. Test results refer back to specific numbered stocks. The distribution stock consists of ampoules that are distributed on request to investigators. The reference seed stock, on the other hand, is retained to generate further distribution stocks as the initial stock becomes depleted.

Although this procedure has been developed to suit the needs of a large central repository, it is also applicable in smaller laboratories. Even where the number of cell lines and users may be limited, it is important to separate "seed stock" from "working or distribution stock". Problems associated with genetic instability, cell line selection, senescence or transformation may be minimized or avoided entirely by strict adherence to this principle.

While short-term preservation of cell lines in mechanical freezers (-75°C) is possible, storage in liquid nitrogen (-196°C) or its vapor (to -135°C) is much preferred. The use of liquid nitrogen refrigerators is advantageous not only because of the lower temperatures and, consequently, almost infinite storage times possible, but also due to the total absence of risk of mechanical failures and the prolonged holding times now available. Certainly for all the smallest cell line banking activity, storage in a liquid nitrogen refrigerator is essential.

Critical characterizations involve tests to assure the absence of microbial contamination (including mycoplasma and viruses), species verification, and, where possible, individual identification of the specific line. The latter can best be accomplished by cytogenetic analysis (Hay, 1985) or by DNA fingerprinting (Nelson-Rees et al, 1981). An extensive series of additional characterizations may be applied depending upon resources at hand and the nature of the cell bank being developed (Hay, 1985).

Primary Cultures

Species, Strain, Sex, and Age of Tissue Donor

Species, strain, sex and age differences in metabolism are well-known from in vivo studies. These differences are often reflected in primary cell cultures. For example, a comparison of the metabolism of amphetamine in hepatocytes from five species demonstrated differences that correlated with those seen in vivo (Green et al, 1986). Hepatocytes from male and female F-344 rats differ in the formation of conjugates of 2-acetyl-aminofluorene (McQueen et al, 1986). Sulfate conjugates predominate in hepatocytes from male rats, while glucoronides were found in greater quantities in hepatocytes from females (McQueen et al, 1986). Age-related changes in conjugation have also been seen in hepatocytes (McQueen and Williams, 1987). Finally, strain differences occur in the induction of P-450 and maintenance of associated enzyme activities in hepatocytes (Grant et al, 1986).

Method of Cell Isolation

The procedure for cell isolation should be clearly described, because the method of cell dissociation can affect the quality of the cells. Comparison of cellular function of hepatocytes prepared by the two-step perfusion of the liver originally described by Seglen (1973) and those isolated using ethyl-enediamine tetraacetic acid (EDTA) followed by Percoll density gradient centrifugation (Meredith, 1988) showed that cells prepared by the latter method have a higher glutathionine content and a more constant level of P-450 and y-cystathionase (Meredith, 1988). It should be noted that cells prepared with collagenase were not subjected to a density gradient centrifugation which itself might result in the selection of a sub-population of hepatocytes.

Identification of specific cell types

The specific cell type(s) that are found in the primary culture should be defined. This is necessary since differences in sensitivity to toxic compounds may vary from one cell type to another. This was shown in a study of the induction of DNA repair in rabbit lung cells in which Clara cells were shown to have a lower capacity to repair alkylated DNA than did Type II alveolar cells (Deilhaug et al, 1985).

Cell Culture Environment

Culture conditions are vital to maintaining tissue specific function and metabolic capacity, two of the major reasons for using primary cultures. In standardizing a primary culture system, the following parameters must be known: medium, serum, supplements, and substrate. One well-studied example is the effect of these conditions on P-450 levels in hepatocytes; this has recently been reviewed (Paine, 1990). Another example of the differences that can occur because of culture conditions can be seen when comparing suspension and monolayer cultures of hepatocytes. Uridine 5'-diphosphoglucuronic acid (UDPG) content in freshly isolated hepatocytes was reported as only 5% of the content in liver, while monolayer cultures, after 24 hours, had almost 60% (Croci and Williams, 1985). Thus, suspension cultures which are used immediately following cell isolation would have decreased levels of UDPG compared to monolayer cultures.

In summary, it is vital that the source of cells and isolation conditions as well as the nature of the cells and the culture conditions be described. Standard protocols can then be adopted which will help to minimize variation, facilitating interlaboratory comparisons and system validations.

Terminology, Historical Perspective, Scientific Communication

To communicate is to impart or give information, according to Webster's Dictionary. Scientists have not done particularly well communicating with either the lay community (Linder, 1989; Iglewski, 1989) or each other, and there are numerous examples of the results of the failure to do so.

Why is this so? Research is becoming more and more interdisciplinary and often differences in the terminology used frustrates attempts to communicate with others in areas peripheral to our own expertise. In some cases, one may not even know that the jargon he/she is using is being misinterpreted by others. That is, the same terms may, in fact, have a different meaning in other, even closely related areas. This has resulted from the fact that a particular field's jargon, techniques and body of information are being used widely and in diverse disciplines. One now must communicate in a more global sense during scientific presentations, and in publications and research proposals. Terminology, used by individuals who have not previously used it, is often misused and, therefore, confusion occurs because the writer and reader, who represent different areas of specialization, have been brought together by the common technology used in their work.

Here are some very common examples of misuse and abuse of cell and tissue culture terms which are important in the area of toxicology. Three of the most controversial terms in cell culture are "established cell line", "cell strain", and "cell line". In 1984 (Schaeffer, 1984) it was proposed by the Tissue Culture Association (TCA) Terminology Committee that use of the term "established" be eliminated when it is used as an adjective referring to cells in culture. In the past, this term was used to denote a cell line that had acquired immortality with respect to its ability to proliferate in culture. Exactly when this point occurred was never acquired by such a cell were likewise not defined. It was presumed that such a cell was not normal and probably had acquired tumorigenic properties. Difficulties with the term ensued with widespread use. Examples: "The culture described was established in 1975." "We established the culture in order to examine the effect of toxins on it." "The establishment of human, diploid, epithelial cells is difficult." "The established culture was used for viral propagation." Precisely what is meant by these uses of the term? Which use is consonant with the then prescribed definition? The misuse and overuse of the term "establish" has made it impossible to know what is being discussed when it is used to describe cell cultures; it really should no longer be used as an adjective because it has no clear meaning as such. The term has now been replaced by the term "continuous" which, when used in conjunction with cell, cell line or cell strain, denotes a culture which can be subcultured continuously (i.e., indefinitely).

"Cell strain" and "cell line", were originally defined (Fedoroff, 1966) as they are currently used, namely: a cell line is one which has been transferred at least once and a cell strain is one which has been derived from a primary culture or a cell line by cloning or some other means of selection. Subsequent subcloning then yields a "substrain" of the original strain. Relationship is, therefore, linearly described as it is in Webster's Dictionary where a strain is defined as "a group of presumed common ancestry" and "a line of individuals differentiated from the main species"; clearly a relevant definition for cell cultures even in 1966, for cloned cells were already known from the work of Earle et al (1954) and Puck and Marcus and Moorhead (1961) in describing their novel fibroblast cell culture, WI-38, stated that since a cell line refers "to those cells that have been grown in vitro for extended periods of time (years)..[and]..presumes potential 'immortality' of the cells...[and further since]...all mammalian cell lines examined to date vary from the diploid chromosome number, this fact alone should exclude the diploid cells (WI-38) from being termed "cell lines" and we have chosen to refer to them as 'cell strains'" (Hayflick and Moorhead, 1961). Tumorigenesis and anchorage independence were other possible characteristics attributed to cell lines by Hayflick and Moorhead. They used "cell line" to describe all other cell cultures than theirs and defined "strain" as cells with a finite life span. Thus, use of the terms "line" and "strain" as defined by Hayflick and Moorhead (1961) became entrenched. An example of the extent of the entrenchment occurred at a 1982 international meeting where one speaker discussing terminology standardization was prompted to say "I use Hayflick's definition, although I known it is wrong. Still, in my mind, when I refer to a cell strain I think of the cells with a finite life" (Litwin, 1982).

The current definitions take into account newer information available about cells in culture since the publication of Hayflick and Moorhead, divorce the definitions from any particular cell culture system and, by simply adding the adjectives "finite" or "continuous", clearly state what is known or presumed about the cultural longevity of the cell culture in question (Schaeffer, 1990). Hayflick and Moorhead, in their definitions, imply that a cell line, being immortal, is also tumorigenic. Today we recognize that a continuous cell line need not be tumorigenic and that, for some cultures, tumorigenesis is a characteristic acquired over time in culture (Aaronson and Todaro, 1968: Chen and Heidelberger, 1969; Schaeffer and Heintz, 1978; Todaro and Green, 1963). Anchorage dependence, another criterion for defining a cell strain by Hayflick and Moorhead, does not reliably predict normalcy or tumorigenic ability as there are false positives and negatives (Jari-walla et al, 1979; Iype et al, 1975; Morel-Chany et al, 1978; Schaeffer and Polifka, 1975). In contrast, the definitions of "cell line" and "cell strain" that are currently extant are not restricted by new knowledge. Hayflick and Moorhead's definitions do not account for cell lines which are not tumorigenic but which are immortal, cell cultures that are cloned or cell cultures which have recently been introduced to in vitro life. Would a recently introduced culture be considered a cell strain before it is known whether or not it will eventually become immortal? And what happens to this "strain" if it becomes immortal?

It is appropriate to be less ambiguous and call a cell in culture a "line" if it has been passaged even one time and a "strain" if it has undergone deliberate selection, whether by cloning or some other means; no other bias or restriction should be placed upon the definitions. If, at some future time, one determines the mortal, tumorigenic, etc., status of the culture, one can then affix the appropriate adjective(s) such as continuous, finite, immortal, tumorigenic, etc. but "line" and "strain" should remain simply defined. As a result of our increasing knowledge about the process of immortalization, it should be possible soon to define when a culture is immortal vis á vis continuous. Currently, the two words can be used interchangeably but only in the context of discussing the longevity of cell lines. We know that a continuous line is immortal but not all immortal lines are tumorigenic. Tumorigenesis is a property which is separable from, but preceded by, the process of immortalization.

"Transformation" is another example of a term which is becoming more and more ambiguous to use. In the area of prokaryotic genetics, "transformation" is defined as the process by which a cell gains genetic information from another cell without the intervention of a vector such as a virus, that is, via naked DNA. Today, the cavalier use of this word by other than microbiologists has extended the meaning to include various phenotypic changes in cells such as morphological, antigenic, proliferative, etc., with no exchange of DNA implied. For many, especially those working with cell cultures, transformation has been used to describe the process by which cells begin to express neoplastic or malignant properties. Because of the problems inherent with using "transformation" in its original genetic sense, eukaryotic molecular biologists substituted the term "transfection" when they began to describe the transfer and integration of foreign DNA into cells in culture. However, originally, "transfection" was coined by virologists to describe the infection of cells with DNA isolated from viruses, thus transferring the infection, bypassing receptors which would be required if intact viruses were used. Today transfection is widely used to describe the transfer of DNA, from any source, into cultured cells and its use is now widely entrenched. Thus, in eukaryotic systems, "transformation" must be relegated to define a phenotypic change in a cell, no matter what the cause, in order for there to be no misunderstandings.

Lastly, "transfection" is often defined as "plasmid-mediated DNA transfer" and it can be so defined, but, in reality, this is only one aspect of its definition in eukaryotic systems. If one were to restrict the definition in this manner it would be necessary to coin yet another term when the DNA is not integrated into a plasmid. Rather, defining "transfection" simply as "DNA transfer" allows the widest use of the term as long as the origin of the DNA is defined, such as integrated within a plasmid, as isolated from a cell or from a gene synthesizer, etc.

"Passage" and "passage number" are two terms that, although they are rather simply defined as subculturing and the number of times the cultures were subcultured, respectively, are not used by some researchers at all. Consider the number of times in the literature that one has seen the following: "The cells were used after their 53rd week in culture." "The cells were injected into nude mice following their 3rd month in culture." "The cells were found to be normal even after having been cultured continuously for 18 months." What do these authors mean? Did the cells sit in culture for the respective periods of time? Did any subculturing go on at all and if it did, at what intervals? Were the cells replicating? If one were to try to repeat experiments described as above, how would one go about doing so? These descriptions are so vague as to render the results useless to anyone else. These means of designating cultural age would not be so vague if the authors defined the intervals at which cells were subcultured during the period described, if, indeed, they were subcultured. To do otherwise makes it virtually impossible for one to interpret or repeat such experiments. The phase "population doubling level" would obviate the difficulties mentioned but it is a phrase most researchers seem to resist, although those involved in cytogerontological research and some involved with in vitro carcinogenesis find its use indispensable. At the very least, one should state the number of passages which have occurred, the interval at which these were done and the amount by which the cells were dilated before replanting occurred. Much can happen during the in vitro life span of cells in culture and unless one knows the relative cultural age of the various subcultures it is possible that characteristics such as enzymes systems required for activation of toxicological compounds could become noninducible with culture age. Were this to happen, it might be impossible for ongoing work in the laboratory to be repeated or for work in other laboratories which have received the cell cultures to confirm original findings.

Finally, it is important to discuss the description or naming of cells and media. With cells obtained from another laboratory, any deviations from the way in which the culture was originally maintained must be stated whenever the culture appears in a publication. Cultures can vary in their responses depending upon the way they are maintained and, thus, a biological parameter which the cultures were supposed to possess could be absent because of differences in cultural procedures in subsequent laboratories. The same holds true for cell cultures or culture media that are defined in one laboratory and then are widely used throughout the research community. It is important to realize that if a laboratory using a medium changes its formulation, in any way, it is no longer correct to call that medium by its original name. It should be called a modification of the original medium and any publication should be describe the modifications in detail. Likewise, a cell culture which was derived in a particular manner and given a unique name which becomes widely used throughout the research community is the only culture which should have this name. Another cell culture, derived in another laboratory, even if derived in precisely the same manner and from the same inbred species of animal may not be given the same name as the originally described cell line. It may not be the same culture in all important respects. One can correctly state that the cell line was derived in the same manner and from the same animal species. This avoids the chaos which can develop when one believes that he/she is obtaining the original named cell line when, in reality, it is a similarly derived copy.

Since our goal is effective communication and meaningful and rigorous conduct of science, we should be striving to make our communications and protocols as lucid as possible; we should make every effort to describe phenomena so that others can repeat and, hopefully, reproduce our findings. In oral communication, it is easier to explain an ambiguity. However, with written communication, where there is no immediate exchange of dialogue, it is more difficult. For a manuscript, a congressional appropriation, a grant proposal or a report to a special interest group, it is important not to be misunderstood or misinterpreted. Therefore, only precisely defined terms, which are universally accepted, should be used.


  • Aaronson, S.A. and Todaro, G.J. (1968). Basis for the acquisition of malignant potential by mouse cells cultivated in vitro. Science 162: 1024-1026.
  • Chen, T.T. and Heidelberger, C. (1969). In vitro malignant transformation of cells derived from mouse prostate in the presence of 3-methylcholanthrene. Journal of the National Cancer Institute 42: 915.
  • Croci, T. and Williams, G.M. (1985). Activities of several phase I and phase II xenobiotic biotransformation enzymes in cultured hepatocytes from male and female rats. Biochemical Pharmacology 234: 3029-3035.
  • Deihau, T., Myrnes, B., Aune, T., Krokan, H. and Haugen, A. (1985). Differential capacities for DNA repair in Clara cells, alverolar Type II cells and macrophages of rabbit lung. Carcinogenesis 6: 661-663.
  • Del Guidice, R.A. and Gardella, R.S. (1984). Mycoplasma infection of cell culture. In Vitro Monograph 5, Uses and Standardizations of Vertebrate Cell Cultures. pp.105-115. Tissue Culture Association, Gaithersburg, MD.
  • Earle, W.R., Bryant, J.C. and Schilling, E.L. (1954). Certain factors limiting the size of the tissue culture and the development of massive cultures. Annals of the New York Academy of Science 58: 1000-1011.
  • Fedoroff, S. (1966). Proposed usage of animal tissue culture terms. In Vitro 2: 155-159.
  • Grant, M.H., Smith, S.J. and Burke, M.D. (1986). Strain differences in the maintenance of cytochrome P-450 and mixed function oxidase activities in cultured rat hepatocytes. Biochemistry Journal 239: 785-788.
  • Green, C.E., LeValley, S.E. and Tyson, C.A. (1986). Comparison of amphetamine metabolism using isolated hepatocytes from species including human. Journal of Pharmacology and Experimental Therapeutics 237: 931-936.
  • Ham, R.G. and McKeehan, W.L. (1979). Media and growth requirements. Methods in Enzymology Vol. LVIII. Cell Culture (eds. Jakboby, W.B. and Pastan, I.H.) pp. 44-93. Academic Press, New York.
  • Hay, R.J. (ed.) (1985). ATCC Quality Control Methods for Cell Lines, 1st ed, American Tissue Culture Association: Rockville, MD.
  • Hay, R.J. (1988). The seed stock concept and qulaity control. Analytical Biochemistry 171: 225-237.
  • Hayflick, L. and Moorhead, P.S. (1961). The serial cultivation of human dipoid cell strains. Experimental Cellular Research 25: 585-621.
  • Iglewski, B. (1989). Communicating with the public. A new scientific imperative. ASM News 55 (6): 306.
  • Iype, P.T., Allen, T.D. and Pillinger, D.J. (1975). Certain aspects of chemical carcinogenesis in vitro using adult rat liver cells. Gene Expression and Carcinogenesis in Cultured Liver (eds. Gerschenson, L.E. and Thompson, E.B.) pp.425-440. Academic Press: New York.
  • Jariwall, R.J., Aurelian, L. and Ts'O, P.O.P. (1979). Temporal acquisition of phenotypic alterations and tumorigenicity in cells exposed to DNA of Herpes simplex virus Type 2. Proceedings of the Annual Meeting of the American Association of Cancer Research 20: 153.
  • Linder, R. (1989). Letter to the Editor. ASM News 55(1).
  • Litwin, J. (1982). Discussion Section. Development of Biological Standards Vol. 50 (eds. Griffith, J.B., Hennessen, W., Horodniceanu, F. and Spier, R.E.) pp. 329-334. S. Karger A.G., Basel.
  • McGarrity, G.I., Vanaman, V. and Sarama, J. (1982). Cytogenetic effects of mycoplasmal infection of cell cultures. A review. In Vitro 20: 1-18.
  • McQueen, C.A., Miller, M.J. and Williams, G.M. (1986). Sex differences in the biotransformation of 2-acetylaminofluorene in cultured rat hepatocytes. Cell Biology and Toxicology 2: 271-281.
  • McQueen, C.A. and Williams, G.M. (1987). Toxicology studies in cultured hepatocytes from various species. The Isolated Hepatocytes: Use in Toxicology and Xenobiotic Biotransformations (eds. Rauckman, E.J. and Padilla, G.M.) pp. 51-67. Academic Press: New York.
  • Meredith, M.J. (1988). Rat hepatocytes prepared without collagenase: Prolonged retention of differentiated characterisitcs in culture. Cellular Biology Toxicology 4: 405-425.
  • Morel-Chany, E., Guillouzo, C., Trincal, G. and Szajnert, M-F. (1978). "Spontaneous" neoplastic transformation in vitro of epithelial strains of rat liver. Cytology, growth and enzymatic activities. European Journal of Cancer 14: 1341-1352.
  • Nelson-Rees, W.A., Daniels, W.W. and Flandermeyer, R.R. (1981). Cross contamination in cell culture. Science 212: 446-452.
  • Paine, A.J. (1990). The maintenance of cytochrome P-450 in rat hepatocyte culture: Some applications of liver cell cultures to the study of drug metabolism, toxicity and the induction of the P-450 system. Chemico-Biological Interactions 74: 1-31.
  • Puck, T.T. and Marcus, P.I. (1955). A rapid method for viable cell titration and clone production with HeLa cells in tissue culture, the use of irradiated cells to supply conditioning factors. Proceedings of the National Academy of Science USA 41: 432-437.
  • Schaeffer, W.I. (1984). Usage of vertebrate, invertebrate and plant cell, tissue, and organ culture terminology. In Vitro 20: 19-24.
  • Schaeffer, W.I. (1990). Terminology associated with cell, tissue and organ culture, molecular genetics. Tissue Culture Association Terminology Committee. In Vitro Cellular and Developmental Biology 26: 91-101.
  • Schaeffer, W.I. and Polifka, M.D. (1975). A diploid rat liver cell culture. III. Characteristics of the heteroploid morphological variants which develop with time in culture. Experimental Cellular Research 95: 167-175.
  • Schaeffer, W.I. and Heintz, N.H. (1978). A diploid rat liver cell culture. III. Malignant transformation by aflatoxin-B1. In Vitro 14 (5): 418-427.
  • Seglen, P.O. (1973). Preparations of rat liver cells. III. Enzymatic requirements for tissue dispersion. Experimental Cellular Research 82:391-398.
  • Todaro, G.J. and Green, H. (1963). Quantitative studies of the growth of mouse embryo cells in culture and their development into established lines. Journal of Cellular Biology 17(2): 299-313.



Dr. Daniel Acosta
University of Texas
College of Pharmacy
Austin, TX 78712
Dr. David W. Barnes
Department of Biochemistry and Biophysics
Environmental Health Center
Oregon State University
Corvallis, OR 97331
Dr. Ellen Borenfreund
Rockefeller University
Box 2
1230 York Avenue
New York, NY 10021
Dr. June A. Bradlaw
Division of Toxicological Studies, HFF-162
Food and Drug Administration
200 C Street, S.W
Washington, DC 20204
Dr. Robert G. Van Buskirk
State University of New York at Binghampton
Dep. of Biological Sciences
Binghampton, NY 13901
Dr. Björn Ekwall
Department of Toxicology
University of Uppsala
Biomedical Center, Box 594
S-751 24 Uppsala
Dr. Oliver P. Flint
Bristol-Myers Squibb
Room 388, Building 32A
P.O. Box 4755
Syracuse, NY 13221-4755
Dr. John M. Frazier
Associate Director
The Johns Hopkins Center for Alternatives to Animal Testing
615 North Wolfe Street
Room 7033
Baltimore, MD 21205
Dr. Carol Green
SRI International
Life Sciences Division
333 Ravenswood Avenue
Menlo Park, CA 94025
Dr. John Harbell
Bioltechnology Srvs. Div. Division
Microbiological Associates Inc.
Bethesda Laboratories
5221 River Road
Bethesda, MD 20816
Dr. Richard G. Ham
University of Colorado
MCD Biology Campus, Box 347
Boulder, CO 80309
Dr. Robert Hay
American Type Culture Collection
12301 Parklawn Drive
Rockville, MD 20852
Dr. James E. Klaunig
Director, Toxicology Program
Medical College of Ohio
3000 Arlington Avenue
Toledo, OH 43699
Dr. Jack M. Lipman
Investigative Toxicology
Hoffman-LaRoche, Inc.
340 Kingsland Street
Nutley, NJ 07110-1199
Dr. Rajendra Mehta
IIT Research Institute
10 West 35th Street
Chicago, IL 60616
Dr. Charlene A. McQueen
Department of Pharmacology and Toxicology
College of Pharmacy
University of Arizona Health Sciences Center
Room 232
Tucson, AZ 85721
Dr. Marshall Palmosku
Dristol-Myers Squibb
P.O. Box 4755
Syracuse, NY 13221-4755
Dr. James H. Resau
University of Maryland
22 South Green Street
Baltimore, MD 21201
Dr. Warren I. Schaeffer
University of Vermont Medical College
Department of Medical Microbiology
Given Building
Burlington, VT 05405


Katherine Allen
SRI International
Life Sciences Division
333 Ravenswood Avenue
Menlo Park, CA 94025
Leon H. Bruner
Human and Environmental Safety Division
The Procter & Gamble Co.
P.O. Box 398707
Cincinnati, OH 45239-8707
Janice Demetrulias
Dial Corporation
Technical & Admin Center
15101 North Scottsdale Rd.
Scottsdale, AZ 85254
David A. Epstein
Supervisor - Biochemical Fluids Group
Cell Culture Research and Development
3175 Stanley Road
P.O. Box 88
Grand Island, NY 14072-0068
Eugene Elmore
NSI Technology Services, Inc.
P.O. Box 12313
Research Triangle Park, NC 27709
Robert Finch
US Army Research and Development Laboratory
Fort Dietrich
Frederick, MD 21701-5010
Robert Gay
Organogenesis, Inc
83 Rogers Street
Cambridge, MA 02142
Dianne K. Hammond
The University of Texas Medical Branch at Galveston
Room 1.101, Pharmacology Building
Galveston, TX 77550-2782
Rhonda Jones
Purdue Cancer Center Laboratory
R.H. Pharamacy Building
West Lafayette, IN 47097
Chris Kemp
Whittaker BioProducts
8830 Biggs Ford Road
Walkersville, MD 21793-0127
Mary Pat Moyer
University of Texas
Health Science Center at San Antonio
7703 Floyd Curl Drive
San Antonio, TX 78284
Mary L. Nicholson
Manager, Research and Development
Charles River Laboratories
268 Concord Road
Weston, MA 02193
Paul J. Price
Hana Biologics Inc.
850 Marina Village Parkway
Alameda, CA 94501-1034
Michael Rozen
Unilever Research
45 River Road
Edgewater, NJ 07020
Charles Shopsis
Department of Chemistry
Adelphi University
Garden City, NY 11530
Thomas J. Stephens
In Vitro Alternatives Inc.
Product Safety, Testing and Consulting
201 Renner Road
Richardson, TX 75080
Belina Veronesi
USEPA, Neurotoxicology Division
Maildrop 74B
Research Triangle Park, NC 27711
Kathleen Wallace
Whittaker Bioproducts
8830 Higgs Ford Road
Walkersville, MD 21793-127