Skip Navigation
Johns Hopkins Bloomberg School of Public HealthCAAT

Technical Report No. 1

Technical Problems Associated with In Vitro Toxicity Testing Systems

A Report of the CAAT Technical Workshop of May 17-18, 1989

Editors: John M. Frazier and June A. Bradlaw


The idea of conducting a workshop on the topic of technical problems associated with in vitro toxicity testing systems arose from two converging problem areas. First, the initiation of several projects to validate in vitro toxicity testing methodologies brought to the forefront technical problems related to interactions between the physical-chemical properties of test chemicals and the biological components of the in vitro test system. These problems have resulted in difficulties in the interpretation of the data generated in these studies. Secondly, the Johns Hopkins Center for Alternatives to Animal Testing receives a large number of research grant proposals each year. Many of these proposals are directed toward developing new in vitro toxicity testing methodologies. In many cases, the proposed methodologies represent unique approaches to toxicity testing; however, basic technological requirements for test development are not addressed. It is hoped that this technical report will provide guidance to investigators participating in both test development and test validation. Attention to the technical issues discussed in this report should facilitate the development of new in vitro methodologies and enhance their transfer from the research laboratory to the regulatory arena.

The process by which this technical report was developed involved several stages. Initially, an outline of the issues involved was developed here at CAAT with input from the center's advisory board. With the project broadly defined, sponsors were solicited and a set of specific questions was promulgated in conjunction with the identified sponsors. Concurrently, workshop participants with expertise in cell culture and product chemistry were contacted. Each participant was responsible for several specific questions and was asked to prepare a working document. Draft materials were collected and the workshop convened on May 17-18, 1989 at the Johns Hopkins University. After two days of critical evaluations, each participant prepared revised working documents which were compiled into the draft report. The draft report was edited by Dr. June A. Bradlaw (Co-Chairperson, Division of Toxicological Studies, HFF 162, FDA, 200 C Street, SW, Washington, DC 20204) and myself to obtain a second working draft. This draft was submitted to two outside scientific reviewers (Dr. Daniel Acosta of the Department of Pharmacology and Toxicology, University of Texas at Austin, College of Pharmacy, Austin, TX 78712, and Dr. Wallace McKeehan of W. Alton Jones Cell Science Center, Old Barn Road, Lake Placid, NY 12946) as well as two members of the CAAT Advisory Board (Dr. Peter Ward of University of Michigan Medical School, 1315 Catherine Road, Box 45, Room M5240, Medical Science I, Ann Arbor, MI 48109 and Dr. Kurt Stenn, Professor of Dermatology and Pathology, Yale University, School of Medicine, 333 Cedar Street, New Haven, CT 06510). After receiving the critiques of the reviewers, the final report was prepared and submitted to the publisher. Although the process used to develop and review this report involved many research scientists from diverse backgrounds and areas of expertise, it has worked remarkably well and the final product has been published in a timely fashion. Mainly, this is due to the competence and dedication of the participating individuals whom I would like to thank publicly for their contributions.

The Johns Hopkins Center for Alternatives to Animal Testing depends on the financial support of public and private institutions to conduct its activities. This project would not have been possible without the sponsorship of the U.S. EPA, Mary Kay Cosmetics, Inc., The Procter and Gamble Company, and Hoffmann-LaRoche, Inc. The foresight and cooperation of these institutions is gratefully acknowledged.

Finally, I would like to think Marilyn Principe, administrative assistant to CAAT, and Robbie Betts, my secretary, for their assistance in organizing the workshop and in the preparation of this technical report.

John M. Frazier, Ph.D.
June, 1989


In vitro toxicity testing is a technically complicated activity and many pitfalls and sources or error exist which can lead to erroneous conclusions. A critical factor in the successful development of new testing methodologies and their practical application is proper and adequate training of individuals participating in these activities. To this end, it is recommended that personnel involved in all aspects in in vitro toxicity testing be adequately trained in cell cultures procedures and basic toxicology.

The workshop considered a series of technical questions which relate to factors that confound the use and interpretation of data derived from in vitro systems for toxicity testing. It is recommended that when developing a new in vitro toxicity testing system, certain issues must be addressed and adequate documentation of the test system performance provided. Table 1 lists specific activities which should be undertaken to characterize any in vitro testing system as well as precautions to be taken when introducing test chemicals into test systems. The following comments amplify on the issues identified in Table I.

  1. The characteristics and behavior of cells in culture and their effects on biochemical function and endpoint measurements in toxicity testing systems is an important factor in in vitro toxicity testing and will be the subject of future technical workshops. However, it is important to emphasize here that the biological component of the in vitro testing system should be fully characterized. If the biological component is a cell line, then it is important to provide documentation regarding the source, passage number, and quality control checks performed for cell parity. If, on the other hand, primary cells are used, it is necessary to determine and report the purity of the putative cell preparation, including method of isolation and culture, as well as representative cellular characteristics.
  2. Since alterations in basic physical/chemical characteristics of the media can influence the behavior of the biological component as well as the endpoint measurement of the test system, it is necessary to evaluate the effects of such alterations on test performance and experimental data. In the case of in vitro toxicity testing, the two factors which are generally most relevant are changes in media pH and osmolarity upon addition of test chemicals. To determine test limitations with respect to these two parameters, it is recommended that effects of both parameters are documented prior to use of the test system for toxicity evaluations. This can be provided by determining and reporting cell and culture condition and sensitivity of endpoint measurement to pH in the range of 6.5 to 8.0 and to osmolarity in the range of nominal for the medium to 350 mOsmol/L in the absence of test chemicals. Effects on the test system over these ranges of parameters should be reported.
  3. Foreign agents, such as solvents or antimicrobial agents, should be avoided whenever possible. However, many test chemicals may require solvents in order to prepare stock solutions; therefore, it is recommended that concentration-response curves for the more common solvents (ethanol and DMSO) be reported. In addition, if any antimicrobial or antifungal agents are used, then concentration-response curves for these agents should be reported over a ten-fold range including the concentration used in the test system.
  4. When chemicals are added to cell culture media for testing purposes, it is essential to obtain some preliminary evaluation of their compatibility with media components. Several simple observations can be made at the highest concentration of test chemical: change in pH, final osmolality, development of turbidity both immediately and at a time interval equivalent to the test exposure period, changes in color of medium, and finally, the formation of precipitates. These observations should be reported since they will be useful in interpreting results. Any indications of reactions between test chemicals and medium should be noted and reported.

Table I: Recommended Activities to Characterize In Vitro Toxicity Testing Systems

  1. Characterization of biological components of in vitro test system.
  2. Documentation of the effects of the following parameters on both the status of cells and the endpoint measurement:
    1. pH -- effect of varying pH between 6.5 and 8.0
    2. osmolality -- effect of varying osmolarity from that of standard media to 350 mOsmol.
  3. Documentation of concentration-response curves for:
    1. potential solvents to be used -- at minimum, data should be provided for ethanol and DMSO
    2. antimicrobial/antifungal agents present in test media.
  4. In the course of in vitro toxicity testing, conduct preliminary evaluations to document the reaction of the test chemical with culture medium. The following information should be reported for the highest test concentration employed:
    1. pH
    2. osmolarity
    3. turbidity at appropriate time intervals
    4. color change
    5. precipitates

If possible, analytically determine if there is any decrease in test chemical concentration in the medium during a time interval equal to the exposure period. This study should be conducted with a cell-free, control culture. Chemical losses should be reported.

Although these recommendations do not cover all possible problems which might arise, attention to these issues will significantly reduce the occurrence of some of the more common artifacts experienced when using in vitro systems for toxicity testing. It is recommended that researchers developing new in vitro tests investigate each of these issues in a rigorous manner and provide relevant information in the published literature for evaluation by interested parties.


In recent years, in vitro systems have been explored as possible tools in the evaluation of chemical safety (Berky and Sherrod, 1978; Dawson, 1972; Gourley, 1976; Klausner, 1987; Williams et al, 1983). In vitro cell culture systems used for toxicity testing are a unique combination of a living biological component (cells or tissues), fluid phase (culture media with appropriate nutrients, growth factor,s and hormones) and solid support (glass, plastic, and/or extracellular matrix components). This combination of components is designed to provide a test system in which the biological component can be maintained in a defined environment which supports the growth of cell lines and/or maintenance of the particular cell types in a well-characterized, differentiated state. The physical, chemical, and biological factors necessary to attain this condition have been determined for many cell lines, but are less well understood for primary cultures of human, primate, and rodent cells. Significant research efforts are devoted to the problem of maintaining differentiated, primary cultures of mammalian cells (Acton and Lynn, 1977; Christolfalo and Rothblatt, 1973; Freshney, 1983; Hay, 1986; Jakoby and Pastan, 1979; Kruse and Patterson, 1973; McKeehan, et al, 1989; Paul, 1975). Developments in this area have an important impact on in vitro toxicity testing but will not be discussed in detail in this report.

Summary Section I: Confounding Factors Which Affect Cells in Culture

In addition to the fundamental problem of the maintenance of well-characterized cells in culture for toxicity testing, there are a spectrum of technical problems which result from the interaction of the test chemical with the components of the cell culture environment. These effects confound the interpretation of the intrinsic toxicological effects of the test chemical and must be taken into account when interpreting in vitro toxicity data. The following factors must be considered when evaluating the toxicity of a test chemical in an in vitro system:

  1. Effect on cellular response in test system due to changes in pH of culture medium resulting from acidity/alkalinity of test chemical
  2. Effect of alterations on the osmolarity of the culture medium at high concentrations of the test chemical
  3. Effect of alterations in the composition of the culture medium due to reactions between medium components and the test chemical, e.g., protein denaturation, alteration of essential nutrient availability, etc.
  4. Confounding effects on cell response of organic solvents used to dissolve test chemicals, such as DMSO and alcohols
  5. Interactive effects of antimicrobial agents and test chemicals on cell response
  6. Complications due to evaporation of volatile test agents
  7. Interference with the endpoint measurement e.g., presence of colored test chemicals, interaction with assay enzymes
  8. Influence of protein binding of test chemicals on the concentration-response curve
  9. Effect of test incubation time on availability of essential nutrients
  10. Deterioration of culture purity due to contamination and overgrown with bacteria, fungi and mycoplasma, and
  11. Physical effects of insoluble particulates on cell response in culture.

Each of these factors can potentially influence the observed toxicity of a test chemical in an in vitro testing system and thus, must be taken into consideration.

Summary Section II: Confounding Factors Related to Dispersal of Test Material in Culture Medium

A significant technology problem faced by in vitro, as well as in vivo toxicity testing, is how to deal with aqueous insoluble materials. Many chemicals are hydrophobic or exist in a physical state which is not readily compatible with tissue culture medium and cells in culture. Greases, resins and mineral solids are just a few examples of such materials. Serious questions to be addressed are how to physically introduce the test material in to the tissue culture system, how to define dosages and lastly, how to account for mechanical effects as distinct from the intrinsic toxicity of the test substance. Several approaches have been pursued to introduce problematic test chemicals into in vitro test systems: physical dispersion, extraction and physical separation. Each of these approaches has its own merits as well as limitations. Specific issues which should be addressed concerning incompatible materials are:

  1. What physical/chemical properties of test chemicals can be used to define incompatible materials?
  2. What solvents can be used to dissolve these materials and introduce them into tissue culture media
  3. What physical techniques can be used to disperse test chemicals in tissue culture media?
  4. What extraction techniques can be used to prepare test material for dosing of cells in culture?
  5. What physical separation techniques (e.g., agar overlay) can be used to dose cells in culture?
  6. How can the effective concentration of test chemical to which cells are exposed be quantified?

These questions have plagued in vitro toxicologists for years and definitive answers are not yet forthcoming. However, familiarity with these issues will encourage investigators to accurately report these problems when encountered and describe procedures used to deal with them.

The basic problem of incompatible test materials will arise for most industries that require toxicity testing of commercial or industrial products, many of which are complex mixtures. This problem area can benefit from:

  1. a critical evaluation of the factors involved which are relevant to in vitro toxicity evaluations;
  2. a standardization of testing procedures; and
  3. establishment of guidelines to deal with chemicals from diverse industrial interests.


A. pH

The addition of a test chemical to the culture medium of in vitro toxicity testing systems may significantly alter the pH of the medium and thus modify biological responses. The potential effect of media pH on in vitro toxicity evaluation will depend, to a large extent, on the acidity/alkalinity of the test chemical, i.e., its intrinsic buffering capacity, and the buffering capacity of the cell culture medium. A test chemical with either a high or low pH, but with little buffering capacity, when added to a cell culture system with a highly buffered medium, is not likely to alter pH or elicit toxicity artifacts. On the other hand, a test chemical with moderate buffering capacity when added to an in vitro system with low buffering capacity can potentially produce an effect which is only related to the influence of pH. Whether this pH effect is relevant to in vivo systemic toxicity or merely an artifact of the in vitro system must be determined in order to interpret the toxicological significance of the observed effect.

The pH range for optimal cell growth and metabolism must be determined in the development of culture conditions for cells in vitro. An indicator, such as phenol red, is usually included in the culture medium to allow aseptic determination of pH. This is particularly important for biocarbonate buffered media wherein rapid pH changes may occur. Outside the optimal pH range, cell growth/proliferation is significantly impeded, metabolic rate is altered, and cell death may occur.

The basic question to be answered is whether or not the addition of the test chemical to the in vitro test system will alter the pH of the cell culture medium sufficiently to affect cell performance as determined by the endpoint measured. The essential information needed to make this evaluation can be obtained empirically, by adding the test chemical to culture medium in the absence of cells and measuring the pH change. If the change in pH of the medium is less than 0.2 pH units from the optimum pH for culturing the particular cells used in the test and the remaining buffering capacity of the medium is adequate to compensate for the production of metabolic acids, then the test chemical will have limited effect on the toxicological evaluation. An alternative approach makes use of the pH and acidity/alkalinity of the test capacity of the cell culture medium. With these data, it is possible to calculate the pH change resulting from the addition of known concentrations of the test chemical to the culture medium. In either case, the pH changes which can be tolerated by a given in vitro test system without affecting test performance must be evaluated in control experiments.

If significant pH effects occur when test chemicals are added to the in vitro system, then two approaches can be pursued. First, the pH of the test chemical can be adjusted to an acceptable value. This approach is adequate as long as the pH adjustment does not affect the physical/chemical properties of the test chemical. Close scrutiny is necessary if the pH of the test chemical is adjusted as this may alter the solubility of the test chemical. Solubility effects may also affect the outcome of testing the pure chemical upon addition to the neutral culture medium. There will often be some salt formation of pure chemicals in culture media. For this reason, the preparation and testing of sodium or chloride salts of test chemicals is recommended. The second approach is to determine the effect of pH on the performance of the test system and assume that the effect due to changes in pH is additive with the intrinsic toxicity of the test chemical. In this case, the pH effect can be subtracted and the "intrinsic toxicity" determined; however, a pH control (i.e., cells cultured at a pH equal to the medium plus test chemical but in the absence of test chemical) must be included in the test protocol to perform this calculation. The assumption that the intrinsic toxicity and pH effects are independent would not hold true for test chemicals which have a pKa near 7.4. In this case, small changes in pH would significantly alter the ration of ionized to non-ionized forms of the compound and, therefore, affect its availability to cells.

Additional precautions are necessary when working with bicarbonate buffered systems. If the system is open to the air, then CO2 will be lost and pH changes follow. Thus, whenever a test chemical is diluted in bicarbonate-based media, the samples should be monitored for pH changes and time delays between dilution and testing may become a critical variable.

B. Osmolality

Osmolality is the property of a solution responsible for the osmotic force of water across the semi-permeable cell membrane. This property is quantitated by the osmolarity which is defined as the total concentration of non-water particles (e.g. ions, salts, and proteins) in the medium (Osml = mole/L). Osmolarity of a solution and the associated osmotic pressure generated in cells can be responsible for changes in cell shape, volume, function, and viability. Failure to control for dissolved salts, ions, and particles in the media can result in artifactually increased cell death and injury.

Cell structure and function can be influenced by the osmolality of the cell culture medium. This effect can potentially become important for relatively innocuous test chemicals where high concentrations are required to produce measurable responses. In normal conditions, cells are in contract with extracellular fluids with an osmolarity of 281 mOsmols/L. High concentrations of test chemicals can significantly increase the osmolarity of cell culture media and thus potentially influence test responses. Water flows across selectively permeable cell membranes in order to stabilize the osmotic pressures generated by dissolved salts, ions, and particles. This pressure can cause cells to shrink or well and artifactually invalidate morphologic characterizations of cell injury. Cell viability assays like trypan blue dye exclusion assays are sensitive to osmotic forces. Marked and sudden changes in cellular morphology are characteristic of an osmotic effect.

As with pH, the magnitude of the change in cell culture conditions can be determined either empirically or by theoretical calculations. The effect of these changes on test system performance can be experimentally determined and the influence taken into account, assuming an additive effect. Osmolality of a solution is experimentally determined by measuring the reduction of the freezing point or increase in the boiling point/vapor point of water with an instrument known as an osmometer. Osmometers are easy to use, relatively inexpensive, and simple to calibrate and maintain.

An important question is: How can the change in osmolarity effect be accounted for in the toxicity evaluation? First, it is important to determine the appropriate osmolarity for normal maintenance of cells in culture (Jakoby and Pastan, 1979; e.g. mOsm/L for mouse cells; 290 mOsm/L for human cells). Osmolarity should be determined each time tissue culture media is prepared in order to: (1) determine if the nominal osmolarity is correct (260-320 mOsm/L; Freshney, 1983) and (2) to serve as a starting reference point to use in the assessment of potential effects of osmolarity change due to test chemical. Osmolality should be determined at all appropriate experimental doses of test chemical. However, if the osmolarity is not altered at the highest does concentration, lower concentrations need not be determined. In the event that the toxicity at any experimental concentration is associated with significantly elevated osmolarity, a control assay using normal media with an equivalent osmolarity obtained by addition of non-toxic agents must run to rule out toxicity that is, in reality, only an osmotic effect. The osmotic control media should be prepared using a biologically inactive or non-metabolic compound, such as mannitol, rather than Na, Ca, or K salts.

Examples of the effects of osmotic factors on morphology and function of cells and tissues in vitro have been demonstrated in rat-embryo culture (Clode, et al, 1987), tubulin proteins (Croom-Brown, 1986), and bronchia (Luk and Dulfano, 1983).

Osmolarity must be determined throughout the experimental procedure. Relative changes in osmolarity can occur as a result of evaporation of the aqueous phase of the media which occurs in less-humid incubators at 37° C. As water evaporates, salts are left behind. If fresh media is added and the cycle repeated, increasing amounts of salts/ions accumulate in the dishes. This increases the osmotic forces and may affect cells or tissue explants. This effect can be controlled by complete replacement of media at periodic intervals. Use of polar and non-polar solvents to dissolve test chemicals can also generate osmotic forces in culture media and these effects need to be determined.

C. Reactions Between Test Chemicals and Culture Medium

Another serious technical problem in in vitro toxicological testing exists when test chemicals react with medium components producing adducts or modifications which alter the ability of the cells to maintain normal metabolism. Cell status is thus altered independent of a direct toxic effect of the test chemical on the cells. Due to the complex nature of cell culture media, a vast array of toxicant interactions may occur including protein denaturation, oxidation of reactive thiols and lipids, chelation of metals, etc. Although a complex issue, consideration of the many possible medium-toxicant interactions is essential to the correct interpretation of in vitro test data.

The number of possible medium-toxicant interactions is defined by not only the reactive nature of the chemical toxicant but also by the reactive functional groups of the various medium components (Table II). Although there is a vast array of potential interactions, the medium components most essential to cell growth and most likely to be altered by toxicants will be discussed.

Table II: Possible Medium and Toxicant Interactions

Media ComponentReactive Functional GroupChemical Class of ToxicantCell Effect of Media Component
Free amino acids-NH2
Acid chlorides
Organic alcohols
Alkyl halides
Heavy metals
Synthesis of proteins
buffering capacity
Energy sources
Chelator of metals
Vitamins/coenzymes-SHHeavy metalsEnzyme activity
Cellular metabolism
Metals in salt solutionsMe++
Sulfhydryl compounds
Cell morphology
Enzyme activity
Cell adhesion
Growth rate
Serum proteinsMacromolecular structure
(Denaturation/polar absorption)
As with free amino acidsCell adhesion
Availability of:

growth factors
fatty acid
Serum lipids-C=C-C=C-?,?-unsaturated carbonyl compounds
Membrane fluidity

  1. Free amino acids in the medium, which are essential for cell protein synthesis, act as both hydrogen donors and acceptors and possess carboxyl, amino, and sulfhydryl groups as well, which are at risk for reaction. Toxicants which contain metals will chelate amino acids such as cysteine while test chemicals with strong oxidizing potential cause oxidation of such sulfhydryl groups. Carboxyl groups, as in glutamate and aspartate, can be esterified by acid halides and alcohols. Such esterification could interfere with energy production as well as protein synthesis. Due to the highly charged nature of some amino acid side chains, they are always at risk for adsorption onto or hydrogen bonding with polar toxicants thus lowering the free amino acid concentration of the medium.
  2. Metals in the salt solution of a medium are mainly the cations Na+, K+, Ca++, and Mg++ which are balanced by the anions chloride and phosphate. These ions maintain osmotic and ionic equilibrium. An additional nineteen trace metals are present in serum. A balance must be maintained in the medium between Ca++ and phosphate or precipitation will occur. Toxicants with potential chelating properties, e.g. polyglycols, diamines, sulfhydryl reagents, can complex metal ions which would alter the ionic equilibrium of the medium and eliminate essential trace elements needed for enzymatic reactions.
  3. Vitamins/coenzymes are present in limited concentrations within the cell. Hence, irreversible alterations of these compounds in the media may deplete the cellular complement of these substances. Lead, cadmium,and mercury ions are known to bind to the disulfide bonds present in thiamine pyrophosphate, lipoic acid, and coenzyme A.
  4. In addition to the salt solution components, standard cell culture medium requires supplementation with proteins usually supplied as whole or dialyzed serum or serum components. The primary function of the added serum proteins is not to supply amino acids for growth but to be a source of factors for the attachment of cells, regulators of cellular function and as carriers for micronutrients, growth factors, and lipids.
  5. The proteins of serum can be denatured by changes in their polar outer coat due to "salting out" effects by high concentrations of highly charged toxicants or changes in pH due to addition of acidic compounds such as halogenated acids. Adsorption of highly charged or polar toxicants on the surface of proteins due to hydrogen bond formation may inhibit the ability of proteins to act as carriers for essential nutrients. The ions of heavy metals (particularly silver, mercury, cadmium, and lead) readily form mercaptides with cysteinyl residues. Protein lability to such mercaptide formation varies with protein type.
  6. The composition of the lipids and fatty acids of serum are critical precursors for maintenance of fluidity and integrity of the cell membrane. The propensity of polyenoic fatty acids to undergo oxidation of their unsaturated bonds by free-radical mechanisms is an important concern in medium stabilization. Halogenated hydrocarbons and ethanol are both known to cause oxidation of fatty acids. Organic solvents and amphoteric detergents destroy lipid integrity. Heavy metal ions of mercury and cadmium complex with the phospholipid bases of ethanoamine and choline expanding the surface area of the membrane and presumably decreasing its fluidity and flexibility.

Although serum proteins and lipids are the only two classes of serum components discussed here, there are approximately ten additional classes of components in serum with varying concentrations dependent on the source of the serum and the age and nutrition of the animal. Each of these has the potential to interact with toxicants. To overcome the problem of serum variability and thus the potential for unknown toxicant interactions, many researchers adapt cultures to defined or serum-free medium which adds to the minimal basal medium only those components necessary for cell growth and maintenance of differentiated characteristics. Thus, knowledge of the concentrations of each component in a defined medium allows nutrient losses of any nature to be measured by assays for the functional groups thought to be affected by toxicant exposure.

While serum-free medium may eventually minimize some of the problems of serum-supplemented media, defined media require additions of essential proteins, such as growth factors, at ng and µg levels. At these concentrations, essential factors may be highly susceptible to reactions and even inactivation by test chemicals.

Due to variability in serum protein content, the measurement of loss of total protein due to toxicant action is possible only in serum-free formulations with defined and reproducible protein content. Total protein is most often measured by the Biuret, Bradford or Lowry assay after centrifugation of precipitated proteins due to toxicant denaturization. However, the problem with protein-toxicant interactions is not usually denaturation to the point of visible precipitation but instead a partial protein alteration which inhibits its usual functions. Assays for this type of protein modification should be limited to those proteins in the medium which are essential for the growth and phenotypic expressions of the particular cell type, e.g. transferrin, growth factors, etc.

Pretesting of the medium with toxicant in the absence of cells will potentially indicate toxicant-medium interactions without complications of cellular metabolism. If the medium alterations by toxicant are known, compensations can be made during incubation conditions, such as:

  1. additions of the modified component, e.g. additional transferrin;
  2. protection of the modified component, e.g. using sulfhydryl reducing agents; and
  3. elimination of the component from the medium if it is not necessary to the growth of the cells, e.g. dialysis of unnecessary salts.

In summary, toxicant-medium interactions are a function not only of the toxicant chemistry but also the composition of the medium. If the effects on the medium result in alterations of components which are not essential to the growth of the cell, then the medium can be modified to a simpler formulation excluding the modifiable component, e.g. serum-free medium. Alterations of medium components which are essential for cell growth and nutrition by toxicants can be compensated for by additions to the medium or protection of the modified component. Prior incubation of medium without cells with the test chemical and subsequent monitoring of selected factors essential to cell growth may provide critical information on medium-toxicant interactions. All changes in medium formulation for a given cell line or cell type must always be tested for effect on cell growth and phenotypic expression.

D. Organic Solvent Effects

Many test chemicals and formulations are not aqueous soluble. These solubility problems often create complications when polar solvents are employed (alcohols, DMSO) to solubilize the test agent before addition to the in vitro test system. Two possible effects may result from the dual exposure to both the test agent and the solvent. First, the solvent may be toxic. Thus, the toxicity of the test chemical must be evaluated in the context of interactive toxicity. The basic toxicity of the solvent must be carefully evaluated in the in vitro test system and maximum no effect levels determined. This concentration of solvent can be used as a criterion for allowable solvent levels in the test system. The second possibility is that the solvent is not directly toxic, but may modulate cellular functions by either altering cell differentiation and/or intermediate cellular metabolism. These effects could potentially alter the response of the cells to chemical toxins, and thus, confound toxicological evaluation. Either of these effects must be carefully evaluated in any in vitro test system.

A special case is the use of a volatile organic solvent (ether, chloroform, etc.) for the addition of a measured amount of the test chemical to the culture dish before addition of media. In this case, the test chemical is dissolved in the solvent and known quantities added to the culture dish. The organic solvent is allowed to evaporate naturally or with the aid of a vacuum atmosphere. Following addition of cells and culture media, the test chemical will partition into the medium up to the maximum of the solubility limit or partition coefficient for the chemical. Alternatively, diffusion-limited dissolution into a serum-based culture medium may be utilized to deliver the test chemical which has been "solvent-coated" on the surface of an appropriate container. The need for analytical verification of test material concentration should be considered for these special cases. Appropriate controls for the effects of any residual solvent should be included with these assays.

E. Effects of Antimicrobial Agents

Antimicrobial agents are added to media used in some cell culture laboratories to reduce contamination by bacteria and fungi. While this practice is advantageous when tissues or substances of questionable sterility must be utilized in primary cultures, systems free of all antibiotics are preferred.

Antimicrobial drugs can be toxic to some or all cells. They may also interact synergistically to mask or enhance activities of test compounds. Furthermore, the indiscriminate use of antibiotics in cell culture fluids permits repeated lapses in aseptic technique. Consequently, the system may become contaminated with resistant organisms, especially mycoplasma (DelGiudice and Gardella, 1984). Many strains of mycoplasma can coexist with cultured cells in the presence of antibiotics without causing any overt cytopathic effect. However, such infections have been shown to affect many quantifiable parameters, thus compromising or negating interpretations of experimental data entirely. This and related problems are discussed further in Section J.

The most commonly used antimicrobial agents include penicillin (at 10-1000 units/mL), streptomycin (50-500 µg/mL), and gentamicin (5-50 µg/mL). Even these compounds, which have been added routinely for years, may have adverse effects depending upon the cell type, other test compounds in the system, and the form and concentration of antibiotics used. For example, Huegin et al., (1986) reported that five antibiotics containing the lactam structure, including 6-amino penicillinic acid, suppressed generation of virus-specific cytotoxic T-cells and their proliferation in vitro. Similar findings were obtained with the erythroleukemic line K-562. Walker et al, (1988) found that streptomycin potentiated the nephrotoxicity of cyclosporin A on epithelia of the procine and canine kidney cell lines LLC-PK1 and MDCK. The pig kidney line is also affected by gentamicin, since Holohan et al, (1988) found that LLC-PK1 cells accumulated higher levels of intracellular Ca2+ rapidly and for a prolonged period after exposure.

One of the most widely used and effective antifungal agents, fungizone or amphotericin B, is inhibitory or even profoundly cytotoxic in some systems. Krause and Juliano (1988) noted that acute (2 hour) exposure of LLC-PK1 cells to amphotericin B inhibited protein synthesis, caused protein loss and eventual cell detachment with an IC50 of about 30 µg/mL. Sodium-stimulated glucose transport was inhibited at much lower doses (1.5 µg/mL). The methyl ester of amphotericin B is less toxic to cultured cells than the parent compound (Fisher et al, 1976).

Broad spectrum tetracyclines, which are often needed to eliminate mycoplasma infections from cell lines, can also have markedly inhibitory effects of the host cells. Borup-Christensen et al, (1988) found that minocycline inhibited not only proliferation rates of some human-human hybridomas but also IgM secretion. The effects were reversible with lower concentrations of the antibiotic, but cytostatic action has been observed at higher doses with some tumor cells in vitro (Inaba and Nagashima, 1986).

Ideally, antibiotic-free systems are preferred. In primary culture systems where a microbial flora is present, penicillin (100 U/mL) and streptomycin (50 µg/mL) can be employed with caution. If possible, antibodies should be removed from the primary cultures after the initial 5-6 days.

Test compounds may be contaminated with microorganisms and cannot be sterilized. In such cases, antimicrobial agents become essential components of the test system. Therefore, preliminary trials with appropriate antibiotics (penicillin, streptomycin, amphotericin B or others) may be necessary to establish the degree of effect on cell proliferation or endpoint parameters being evaluated. A concentration-response curve should be constructed. Subsequent test runs should also include several antibiotic concentrations to permit determination of possible synergistic actions.

Some preparations being tested may contain preservatives which are essential for stabilizing the product as marketed. If positive effects are noted during initial testing of the complete preparation, the manufacturer may wish to repeat the procedure on the isolated preservative and on the preservative-free preparation to identify the source of toxicity.

The use of sterile procedures in cell culture activities is essential in order to prevent contamination with bacteria, fungi, and mycoplasma. Even with the best technique and equipment, the possibility of contamination still exists. Therefore, it is necessary to include a quality control program in the standard operating procedures of the laboratory in order to verify the absence of contamination.

F. Volatile Test Chemicals

For many aqueous soluble chemicals, their vapor pressure is low and little material is off-gassed during experimental studies. Hence, nominal test concentrations remain effective during culture/exposure periods. However, many of the less water-soluble chemicals are highly volatile, resulting in a significant loss of test chemical during a test. This effect has two major consequences. The first is a constant loss of the test chemical from the culture system and hence the regulation of exposure conditions is less than adequate. Nominal test concentrations can differ significantly from actual exposure concentrations. This variance will produce anomalies in the concentration-response curves and result in reduced reproducibility of toxicity when comparisons are made between test systems. The second problem is cross contamination of test concentration groups. Test material volatizing from high concentration groups may dissolve in low concentration or control test groups, thus affecting baseline data. This problem must be prevented when testing volatile test chemicals in cell culture systems.

G. Interferences

Occasionally, a test chemical will possess physical and/or chemical properties which will result in an interference with the toxicity test endpoint measurement. In many cases, this is a chromatic effect. If the endpoint assay requires a spectrophometric measurement at a particular wavelength, then the potential for interference exists. In this case, the test chemical must be completely removed before the assay is performed unless some procedure to correct for its presence is possible. Spectral interference is an example of the most common type of interference. Other possibilities exist. Test chemicals may precipitate and/or co-precipitate with assay reagents and affect endpoint assay performance. The only way to avoid such unforeseen complications is to maintain constant surveillance at each stage of the test protocol to identify abnormal reactions. Technicians should be familiar with the details of assay performance so that they are aware of potential problematic conditions.

H. Protein Binding

Protein binding refers to the association of the test chemical with proteins. Direct reaction of the test chemical with proteins present in the media has already been addressed (Section C). As considered here, binding does not involve formation of covalent bonds between the test chemical and media proteins. Protein binding affinities for the test chemical may artifactually lower or reduce the rate or amount of cell injury and toxicity if protein binding leads to increased cellular uptake by either increasing solubility of the test chemical in culture media or by directly facilitating cellular uptake.

A particularly confounding factor is the reduced effective concentration of test material that results from the binding of test chemicals to either proteins in the culture medium and/or the culture system substrate (glass, plastic, membranes, biological substrates). This causes a distortion in the concentration-response curve which usually is manifested as a shift of the curve to higher concentrations. Thus, EC50s might vary solely on the basis of availability of free toxicant and artifactually influence the data. If the problem involves binding to plasma proteins present in the culture medium due to the presence of serum, this can be corrected by excluding serum during the period of exposure to test chemicals. However, if the binding is to the culture vessel itself, this problem may be undetected. Such an effect could result in false negatives and must be evaluated.

Toxicant binding to media proteins is a dynamic process. Eventually an equilibrium will be established between the free and bound species, but time factors might compromise the data because equilibrium may not be attained during the course of the experiment. This problem can be resolved by eliminating proteins in the medium during experimental exposure time, thus reducing non-specific binding problems during the toxicity test. If the toxicity assay is of short duration, consider the possibility of dosage in serum-free medium or in the lowest serum medium possible. If serum is necessary, one might have to determine when equilibrium has occurred before conducting the toxicity part of the test.

In vivo, non-specific binding of chemicals has been reported for blood cells and serum albumins. Protein binding of surfactant type physical/chemical components also have been characterized. These examples can serve as a guide or indicator for the magnitude and duration of the protein-binding problems.

When using serum containing media for in vitro toxicity studies, one needs to consider the amount of serum and the type of serum. Serum is used at concentrations of up to 20% in the medium; however, the minimum concentration of serum adequate to maintain test system performance should always be predetermined for each new batch of serum. The serum types are usually fetal or adult, bovine, equine or human. One must determine the albumin components of the serum by type and known their importance when designing test protocols in order to control for non-specific protein binding. Heat inactivation, dialysis, or centrifugation during the preparation of serum for culture will also impact on protein binding by altering the competitive binding with other serum components.

Concentration-response curves can be corrected once the actual amount of protein binding is determined either theoretically or experimentally. The "contact" level of test chemical which interacts with the cells can be utilized to express the EC50.

I. Availability of Essential Nutrients

Cells in culture require various essential factors for normal growth and maintenance of differentiation. To provide adequate supplies of these factors, under normal circumstances, the culture medium is replaced at regular intervals. If this schedule is interrupted by the experimental protocol (e.g. elimination of serum from the culture medium during exposure to the test chemical or prolonged periods without media changes), then the possibility exists for depletion of these essential factors. It is important that any such effects are evaluated and taken into account in order to prevent inaccurate interpretation of effects of the test chemical.

The basic consideration in designing the length of exposure to a toxicant and composition of the medium in a testing procedure is the type of cell used as a model system. Cell types vary widely in their nutritional needs for carbohydrates as energy stores, in oxygenation, in dependence on external growth factors, and in requirements for specific factors, such as hormones and trophic factors, to express characteristic features of the cell type. Such variations in nutritional needs are demonstrated by the array of various media formulations. Another consideration for some cells is the requirement for synthesis or availability of extracellular matrix (ECM) components involved in cell-cell and substratum adhesions. Such molecules as collagens, glycosaminoglycans, proteoglycans, and glycoproteins are a complex arrangement of interactive protein and polysaccharide molecules with specific precursor requirements. In testing toxicants, the ideal nutritional state for the cell is one in which the necessary components for growth and phenotypic expression are maintained at the optimal level in the original medium throughout the duration of the test. Considerations of the interaction of the gaseous phase with the media components must also be considered in relation to O2 delivery and pH buffering.

Glucose is the most common factor depleted in a toxic situation, and its depletion is often associated with an energy stress for the cell. The question of the necessary concentration of glucose for adequate cell nutrition over an extended period of time again can only be answered in relation to the specific needs of the particular cell line or cell type. The cellular rate of glucose metabolism will determine the carbohydrate needs of the cells during the testing procedure. Hepatic ascites, malignant, and virus-transformed cells generally have high glucose needs. These cells produce large amounts of lactic acid which change medium pH rapidly (a confluent culture may need new medium every 12 hours). In contrast, many fibroblastic lines grow and metabolize slowly necessitating new medium only every 48-72 hours. During a 24 hour toxicant exposure, the ascites line would be highly stressed toward the end of test period, while the fibroblastic line would be metabolized normally. Cells with high metabolic rates have been used for long-term or chronic testing by: (1) maintaining the cell cultures at very low cell density; (2) periodically adding small amounts of salt solution containing glucose (this does, however, change the concentration of the toxicant); (3) supplying a constant flow of fresh medium containing the toxicant over cells, e.g. capillary tube and high volume, mass cultures that require steady state conditions; and (4) substituting other more slowly metabolized sugars such as fructose or galactose for glucose. If additional glucose is added to the test medium, additional insulin may also be required in the medium to maintain the correct simple carbohydrate balance for the cells.

Another critical factor to consider in the relationship between testing time for toxicant exposure and the nutrition of the test culture systems is the organization of the cells in the culture used for the model system, i.e. explant, tissue slice, monolayer, suspension culture, co-culture or reaggregated cells. The differences in the nutritional needs of each of these systems is dependent on the ability or inability of medium to reach all cells and may dictate whether long-term or short-term exposures are feasible. Medium generally does not penetrate into inner cells of explants and large reaggregated cell bodies often resulting in necrosis due to hypoxia and starvation. Thus, slice preparations are routinely tested for less than 12-24 hours, while monolayers cultures can be incubated for days.

Exogenous specific factors needed for expression of phenotypic markers of certain cell types are another class of nutrients to be considered in toxicity testing, e.g. neuronal growth fact for neuronal cultures or hormones for mammary cells. Again, the question of what is an adequate concentration of any component during a test period is defined by the cell type and the rate of turnover of the specific factor. Rates of turnover of specific factors can often be found in the literature. If the turnover rate of a hormonal peptide is 12 hours and the toxicity test is 24 hours, only 25% of the factor remains at the end of the experiment. If factor turnover data is not available, it can be empirically determined by monitoring the temporal correlation between phenotypic expression of the desired characteristic and feeding of the component. Media formulations may have large excesses of certain factors so that even with metabolic turnover the medium concentration of a specific factor may greatly exceed the required effective level. The concentrations of salt components of the media are often published in manufacturer catalogs. The range of concentrations of components in various types of serum are identified in product analyses statements from serum suppliers.

For cells requiring serum, complications often arise in designing toxicity-testing procedures. Inclusion of serum in the test medium during incubation often lowers the effective concentration of the toxicant by its binding to proteins and lipids leading to false concentration-response data. However, the removal of serum eliminates a vast array of nutrients, growth factors, and metals in the serum which are necessary for normal cell metabolism. The ability of the cell in a test situation to serum deprivation again depends on the cell type. Neuronal cells differentiate in the absence of serum, but they also become less adherent and convert to a differentiated metabolism and produce neurotransmitters. For the 3T3-L1 cell line, removal of serum would inhibit the formation of adipose-like differentiated cells. In general, however, for most cell lines, for short periods (2-6 hours) serum can be removed from a complex salt medium with only minor effects on the cells. It is estimated there is a 2 hour time lag when a cell readjusts to new media. Cells at particularly high risk to serum deprivation will often start lifting from the plate within 6 hours. For such cells, serum substitutes or a lower concentration of serum can be added to test medium or a completely defined medium can be used. It is suggested that prior to toxicant testing, the cells should be observed and assayed for phenotypic markers using the medium conditions and time intervals to be used in the protocol to determine if major metabolic or morphological changes occur under these conditions. The phenotypic markers to be assayed are determined by the cell type in the test system, e.g. glycogen for liver cells, neurotransmitters for neuronal cells, etc.

For test substances which are of a nutrient nature, such as vitamins, amino acid analogs, and hormones, these may possibly serve as both test chemical and nutrient; however, this would depend on the capabilities of the cells to metabolize these compounds and the relative concentrations of the added test formulation to analogous compounds already present in the medium. For example, in the cosmetic industry, amino acids are often added to formulations which must be tested. If the medium already contains micromolar concentrations of lysine, addition of nanomolar levels of lysine to the medium in the formulation tested would have no consequence on cell nutrition.

Developments in studying complex tissue physiology by defined cell culture approaches are revealing the true complexity of interactions among nutrients, hormones, and extracellular matrix. Response to xenobiotics is likely to be a part of this complexity. The total in vitro environment (nutrients, hormones, matrix) employed as well as cell type should be rigidly standardized to achieve quantitative reproducibility of any cellular response. For example, the presence or absence of a single hormone or growth factor might completely alter how a cell would metabolically react to a xenobiotic at all levels from metabolism to damage repair.

In summary, the nutritional requirements of a cell must be met within the period of toxicant exposure to ensure proper interpretation of test results. Some factors which are critical in the nutritional maintenance of cells are an adequate energy source (usually maintained with glucose), proper exposure to oxygen, and in maintenance of trophic, hormonal or other factors necessary for phenotypic expression. For cells dependent on serum-supplemented medium, serum can be removed from test incubation medium for short time periods. Ideally, when culture medium must be modified, test conditions, cell growth, morphology, and phenotypic characteristics should be studied in the modified test medium prior to toxicant testing.

J. Contamination of Test Culture System

Even for small cell culture testing facilities, it is recommended that seed stocks be created for each specific line to be utilized. A starter culture can be obtained from a reputable source (e.g, American Type Culture Collection, Coriell Institute for Medical Research, European Collection of Animal Cell Cultures) and a suitable lot of cells produced (by growth in complete absence of antibiotics) and cryopreserved. Extensive quality control steps can then be applied to progeny from these stocks. Working stocks for toxicity screening can then be generated for use within 5-10 passages from seed stocks. By returning to the seed stock preparation periodically for production of new working stocks, one can be assured of comparability and uniformity within the cell system even over an extended time period.

For an adequate quality control program, one must consider at minimum appropriate precautions and tests to ensure both that the correct cell type and species are in use, and that the cell lines are free of microbial contaminants including mycoplasma. Numerous occasions have resulted in cross-contaminated cells when cell lines were exchanged among cooperating laboratories. Examples of cross-cell contamination with other species have been detailed, documented, and published elsewhere (Hay, 1988). Similarly, the problem of intraspecies cross-contamination among cultured human cell lines has been recognized for more than 20 years and detailed reviews are available on the subject (Nelson-Rees et al, 1981). The loss of time and research funds as a result of these problems is incalculable. The inclusion of both appropriate steps to reduce the likelihood of cellular cross-contamination in the laboratory and of tests to verify cell line identity (e.g., isoenzymology, fluorescent antibody staining, cytogenetics, and DNA fingerprinting) is critical. Commercial testing services to authenticate cell lines are readily available.

Although bacterial and fungal contaminations represent an added concern, in most instances they are overt and easily detected and, therefore, of less serious consequence than the more insidious contaminations by mycoplasma. The presence of these microorganisms in cultured cell lines often negates research findings entirely and has been emphasized over the years (DelGiudice and Gardella, 1984). Still, the difficulties of detection and prevalence of contaminated cultures in the research community show clearly that an especially high degree of diligence in quality control in this area is essential. One key to avoiding mycoplasma contamination is the use of media without antibiotics. Additional critical characterizations, such as production of cell-specific product, response to hormones or growth factors, and presence of characteristic inclusion or filaments, can be included for the seed stock as required.

Quality control methodologies are described elsewhere in detail (Chen, 1977; Cour et al, 1979; Macy, M. 1978, 1979; Hay, 1985, 1986, 1988; Gilbert et al, 1989).

For actual toxicity test runs, depending upon their length and nature, it may be prudent to include determination and quantitation of the degree of microbial contamination both at the beginning and at the end of the trial.

K. Effect of Insoluble Particles

Insoluble powders, such as talc, or insoluble particulates, such as polyethylene beads, which are used in commercial products, may have a direct, physical effect on cells in culture. This physical effect may elicit a phagocytic response or may physically injure the cell. Physical actions could influence test results by damaging the cell membrane thereby yielding spurious and unexpected results, most likely false positives. A physical effect could damage the cell wall thus activating the cell and stimulating cell proliferation. Insoluble particulates and powders may float or settle in the media. It is doubtful that they will disperse evenly in the media, thus heterogenous effects may be observed.

Since most insoluble particulates are larger (>20 m) than the cell, inert, and in lower concentrations than the cells in culture, the effect of physical actions in tissue cell cultures will be minimal. However, physical actions that may affect cells in culture include: (1) stimulation of phagocytosis by particulates less than 10 m; (2) availability of a large number of particulates whose surface chemistry may alter phenotypic expression of cells; and (3) cell damage resulting from test system handling techniques (stirring, addition of materials to test system) and shear forces that may impact particulates on cells (spinner vs. monolayer cultures).

Although adverse physical effects on cells in culture are unlikely to occur for most test materials, it is important that any potential interference or physical action caused by insoluble particulates to cells in culture be controlled by carefully observing for the presence and effects of insoluble particulates in in vitro systems. This may be accomplished by evaluating the test material with and without the insoluble particulate. Once dispersed, the test material may be filtered or centrifuged as a simple means of removing any insoluble particulate from a test material. The particulate-free fraction of the test material may then be used as a control to determine if particulate is contributing to observed responses. If particulates are of concern, then test parameters should be reported in terms of particulate size, distribution, whether cells are stimulated by the material (e.g., phagocytic activity), and potential culture shear forces which may contribute to positive responses.

When considering the effect of insoluble particulates for certain testing objectives, e.g. ocular irritation testing, it may be not necessary to subject the particulate to in vitro or in vivo assays, since the ability of these materials to produce physical actions on the eye is readily predicted.


A. Physical/Chemical Properties of Test Chemicals and Compatibility with Tissue Culture Medium

An incompatible material is defined here as insoluble in the tissue culture medium. Therefore, the basic question becomes how does one predict the solubility of a test material in tissue culture medium. An evaluation of the molecular structure is useful, because extensive information is available concerning the relationship between hydrophilicity, molecular structure, and solubility. Most bioactive molecules are amphiphilic, thus they contain both polar and nonpolar structural elements. All ionic amphiphilic compounds, as well as those nonionic compounds which display a high proportion of hydrophilic groups in comparison to lipophilic (hydrocarbon) groups, are potentially soluble. It is important to recognize that solubility has a qualitative dimension and is defined by the physical phase of the test material which coexists with a saturated solution of the test material in equilibrium with culture medium at a given temperature. This phase structure may be either a gas, a liquid, or a crystal. Since solubility is determined by the difference in the thermodynamic free energies between the test material in the saturated solution and the coexisting phase, the solubility will be low when the coexisting phase is highly stable.

It is clear that the inference of solubility from measurements of partition coefficients, such as between water and octanol, may be misleading. The partition coefficient provides information concerning the relative solubility of the test material between the two phase (aqueous vs. organic), but does not define absolute solubility of the test chemical in the aqueous phase. Another factor which is indicative of solubility is melting point. High melting points for the fry crystals of a test compound often signify low solubility.

Truly soluble materials will dissolve isothermally, i.e., simply by stirring with the solvent at a particular temperature, provided adequate time is allowed for dissolution to occur. Heating and cooling mixtures to accelerate solubility may cause problems, as in the case of a test compound which dissolves at high temperatures and precipitates on cooling. This procedure may result in the formation of colloidal dispersions of ill-defined structure.

In the case of poorly soluble materials no universally applicable test for solubility exists. Often a poorly soluble compound is actually soluble enough to act on cells via the soluble species. Isothermal dialysis experiments, properly designed to allow for analysis of transport kinetics, are the least ambiguous means to determine the solubility of these compounds. Such experiments are the least subject to interference due to colloidal phenomena, because colloidal particles do not generally diffuse through swollen membranes. Similarly, gels such as agarose or gelatin constitute effective barriers to the passage of colloidal particles, and only permit the diffusion of molecular species.

The effective dispersion of poorly soluble compounds to form clear mixtures, as by sonication, does not change their solubility. The formation of clear mixtures simply signifies small particle size or low particle densities, and does not necessarily signify that the compound is soluble. Clarity is an optical property of a mixture and is determined by the refractive indices of the particle and the medium, and the dimensions of the particle present. Solubility is determined by thermodynamics of molecular mixing.

B. Solvents Used to Deliver Incompatible Test Materials to the Tissue Culture Medium

The physical/chemical properties of anhydrous liquids, solids, and powders pose a potential obstacle to testing such materials in aqueous culture media. Guidelines are needed to assist the in vitro practitioner in finding ways to dissolve or otherwise make these materials compatible with the culture medium. The main issue is how to physically introduce the test material into the tissue culture system.

Depending upon the density of the incompatible test material, these materials will either settle or float in culture media. As an example, a thin film of relatively innocuous mineral oil on the surface of the culture medium can interfere with fast exchange across the surface of an aqueous system. Low oxygen tension or changes in pH may affect cells in culture, yielding a false positive result. One must also consider potential interactive toxicological effects of delivery systems and test materials which may lead to unanticipated interference with the response of the in vitro system. Furthermore, water dispersible powders may absorb the aqueous media and swell, thus posing another potential interference with the culture systems response. Finally, it must be recognized that there may be situations where incompatible test materials cannot be tested in specific in vitro assay systems because of the inability to cosolubilize or otherwise deliver them to the test system.

Two proposed methods of introducing otherwise incompatible materials into a test system are:

  1. solubilization of the test material in a suitable cosolvent; and
  2. transportation of the material to the cell via carrier molecules.

The in vitro practitioner and test material chemist may use these guidelines in determining the optimum method for introducing a test material into the culture system.

In general, water soluble materials need only be diluted to the appropriate concentration prior to addition to the test system medium. Whenever possible, the final dilution of a test material should be made using the same medium that is used in the in vitro system.

Water insoluble test materials usually require solubilization in a compatible solvent prior to testing them in the in vitro assay. An ideal solvent is one that permits the addition of an otherwise insoluble test material to the in vitro system without interfering with the assay. When solubility information is not available for a particular test material, dilution of the test material in a variety of solvents by trial and error is often used. However, the selection of a solvent can be greatly facilitated when Solubility Parameter (SP) or Hydrophilic-Lipophilic-Balance (HLB) values are available for the test material (Barton, 1983; ICI Americas, Inc., 1987; Laughlin, 1978 and 1981; Vaugha, 1985 and 1988). The Solubility Parameter is a numerical system for ranking solvents and other materials according to their relative solubility properties. Most test materials, which are likely to be soluble in ordinary solvents, will have a SP value between 5, for oily or oil soluble materials, and 23.4, which is the SP value for water. Solubilization of a particular test material is accomplished by selecting a solvent of comparable SP value. Example: nitrocellulose (SP = 11.25) and isopropanol (SP = 11.24). Solubility properties of solvent combinations can likewise be obtained, by direct proportion, using the SP value. For cases when a solvent has been found to be unsuitable for a particular test system, an alternative solvent of similar SP value can be selected directly from the literature, without resorting to extensive trial and error solubility tests.

SP values are available in the literature for most solvents and many materials. In addition, the SP value can be obtained by direct conversion from the HLB value. The HLB system, like the SP system, is a numerical system for ranking materials according to their solubility properties. It is particularly useful for finding an appropriate cosolvent for raw materials. The HLB value for a number of commercial raw materials is listed in McCutcheon's handbook (MC Publishing Co., 1989). Several raw material manufacturers supply this data in their specification sheets. High HLB emulsifiers, especially those with HLB >15, are excellent cosolubilizers of oils. By mixing the surfactant with the oil at ratios of 4:1 or 5:1, this mixture may be readily added to water. This approach produces a true solution of the test material, not an emulsion.

Temperature effects are not taken into account by either the SP or HLB systems. Additional solubility data related to cloud point or Phase Inversion Temperature (HLB temperature) may be required for some materials when this effect becomes important, (BASF Wyandotte Corp.; Rosen, 1978).

The Octanol/Water Partition Coefficient is widely used as a tool for predicting biological accumulations and effects of organic chemicals. However, the usefulness of this procedure for testing materials may be limited due to a lack of reliable methods for determination or estimation (Mailhot and Peters, 1988).

Specific examples of solvents to consider in solubilizing water insoluble test materials are listed in Table III. The selection of these solvents is based on their relatively safe history of use in household and personal care products. They have excellent potential as cosolubilizers with in vitro systems. The effect of these solvents on specific in vitro systems must be verified in each circumstance (see Section I.D).

Table III: Potential Solubilizers for Water Insoluble Test Materials

1. AlcoholsSP Value
Methyl alcohol14.33
Ethyl alcohol12.55
T-butyl alcohol10.28
Hexylene glycol12.32
Butylene glycol13.20
Propylene glycol14.00
Polyethylene glycols (PEG)Variable
3. Emulsifiers
Polysorbate 20 (Tween 20 -ICI)9.16
Block copolymers (Pluronics - BASF)Variable
4. Miscellaneous
5. Cocolvents for nonaqueous media
Dibutyl phthalate9.88
Diethyl phthalate10.00

1Glycols with ethyl or higher chains are potentially toxic to cells. Toxicity varies inversely with chain length.

Powders that swell in aqueous systems, such as magnesium aluminum silicate and carbomers, can be made thinner by water dilution or by the addition of chelating agents or glycols to the water phase before adding the powdered test material. However, care should be taken when using chelating agents so that essential elemental nutrients are available cells in culture.

Another method of delivering the test materials to the cell is via carrier molecules. These carriers include intralipid (Baxter), liposomes, cyclodextran, and mucopolysaccharides. In these cases, the test material is not truly solubilized in the culture medium, but is incorporated (i.e., engulfed, absorbed, etc.) instead into a carrier system which is transported into the cell.

When dealing with an unknown material, i.e., one for which no solubility information is available, it may be necessary to use a trial and error approach. First, it is necessary to evaluate the compatibility of the test material with a variety of solubilizers and solubilizer combinations. Once a suitable solvent system is obtained, it is important to determine the minimum amount of solubilizer necessary to cosolubilize the test material in the culture medium. Finally, when actual testing is conducted, it is essential to determine the compatibility of the solvent with the tissue culture media using appropriate controls as discussed in Section I. D.

C. Physical Dispersion Techniques

The nominal concentration of a test chemical in culture medium is the amount of the test chemical per unit volume as prepared in the laboratory. If a test material is less soluble than the nominal test concentration, then the dosing solution will typically be a colloidal dispersion of the compound with the aqueous phase saturated with the test chemical. If testing is to be carried out on colloidal dispersions in a reproducible and controlled manner, two important questions which must be addressed are:

  1. How does one create such dispersions?
  2. How stable are they with time?

Neither of these questions can be answered in the generic sense; the answer will depend on the individual case. Dispersion methods for producing colloidal structure (such as sonication or high shear mixing) are useful for dispersing fluids (liquids or liquid crystals) in the aqueous phase, but will not disperse crystalline compounds. Colloidal precipitation, obtained by dilution of a solution of the test chemical in a water-miscible solvent with water, is a commonly used method, but care in executing this method is important. The colloidal structure of the resulting dispersion is very sensitive to the precise conditions of the dilution process, and the product is often highly polydisperse. As long as these problems are recognized, however, such methods may be used. In principle, the method used is not important as long as the colloidal structure is measured and controlled.

The second issue when dispersion techniques are used is the stability of the colloidal structure. There are two mechanisms by which colloidal structure changes: flocculation/coalescence and Ostwald ripening. In ripening, small unstable particles dissolve and the material in them reprecipitates on the more stable large particles. Ripening is a linear function of solubility: the greater the solubility, the faster the ripening. Flocculation is the binding together of dispersed particles, which in turn results from action of thermal energy. Collision frequency is increased by elevated temperature and by increased particle number concentration. The rate of flocculation is strongly retarded by the development of a charge at the surface of the particle. The magnitude of this surface charge is strongly influenced by the valence of counterions, by the concentrations of electrolytes or salts present, and by surfactants. Divalent counterions and high salt concentrations significantly reduce the surface charge and accelerate flocculation. Furthermore, the kinetics of flocculation are retarded by the adsorption of polymers, such as albumin. Coalescence is the fusion of flocculated particles, and its rate depends largely on the physical state of the particle; only fluid particles readily coalesce.

For further information, the most authoritative reference to the field of colloid science is the two-volume series edited by Kruyt (1952). A recent concise treatment of this discipline is found in Everett (1988).

In addition to the purely physical aspects of colloidal structure, compounds that are poorly soluble may act on cells as particles, as well as free molecules, and in such cases, the colloidal structure of the dispersions, as well as the molar composition, must be considered. Further, it must be recognized that particles may interact with cells in two distinctive ways: phagocytosis and fusion. In phagocytosis, the particle is engulfed intact by the cell. In fusion, contact of the particle with the cell leads to partitioning of the contents of the particle into the cell membrane. The integrity of the particle is lost during this process.

If the test compound acts on cells strictly as a molecular species, then the molecular concentration as a solution is an important parameter. Separation of the cells from the test solution by a swollen hydrophilic membrane (dialysis membrane), or by a swollen gel layer (agar overlay), will restrict the response of the cells to the response resulting from an interaction with the molecular species which can diffuse through the barrier.

If the mode of action occurs via particles, such as in phagocytosis, then the particle number concentration (PNC) is important (see Section F). In such cases, the characteristics of the system are important and care should be taken to control the composition of the colloidal dispersion during the test and to determinate the dimensions of the colloidal structure of the dispersions. For most dispersion, quasielastic light scattering (QELS) data is a good method for defining colloidal structure.

D. Extraction Techniques

Extraction is an alternative approach to processing aqueous insoluble materials for in vitro testing. The test material is placed either in water to obtain an aqueous extract or in an organic solvent compatible with water, e.g., alcohol, and allowed to remain in contact for a defined interval of time with or without mixing. At the end of the extraction period, any soluble material is removed by filtration, centrifugation or phase separation and the soluble phase is used for testing in the in vitro system. For a complex mixture composed of several ingredients, components of the mixture may be differentially extracted depending on their solubility in the extraction solution. The main limitation of this approach is the lack of quantitation of the actual dosage to cells in culture of each constituent in the original test material. Therefore, reproducibility of results is difficult unless the extraction procedure is carefully standardized as to extraction time, mechanical agitation and temperature, among other potential modulating factors. In some respects, the extraction procedure can be designed to more closely resemble conditions of exposure in vivo. Development and validation of standardized extraction procedures are important and require additional research efforts.

E. Physical Separation Techniques (e.g., Agar Overlay)

The third technique in the preparation of test chemicals for in vitro toxicity testing is the physical separation of the test chemical from the cell culture system which allows the extractable constituents to diffuse into the cell culture system. This approach uses a physical separation by means of a mechanical barrier through which the extractable substances can diffuse. An example of this technique is the agar overlay test system used to evaluate the potential cytotoxicity of solid materials used in medical devices and implants. Cells are grown in a monolayer on culture plates until they establish a desired degree of confluency. A layer of low temperature melting agar is poured onto the monolayer and allowed to set. The test material is then placed on the surface of the agar and the system is incubated for a defined exposure period. At the end of the test, the cell layers is examined for alterations (either morphological or biochemical). The strength of the toxic effect is classified by the spatial extent of the damage indicated by whether altered cells are only located immediately beneath the test material or whether the zone of effect extends beyond the margins of the test material. This method allows for insoluble materials to be tested as well as soluble substances which are applied as a solution absorbed into filter paper disks. The major drawback of this approach is, as with other techniques described above, the lack of quantitation of the concentration delivered to the cultured cells. This problem is complicated by the uncertainty in specific diffusion of components through the agar layer and depends on the composition and thickness of this barrier. In all these methods, quantitation and/or standardization of the concentration is difficult.

F. Quantitation of Dosage to Cells in Culture

When concentration-response relationships are experimentally determined in in vitro systems, the question of documentation of the test chemical concentration is a critical consideration. The reproducibility of concentration-response relationships is highly dependent on the ability both to reproducibly prepare stock solutions of test chemicals and to dose test cultures accurately. The actual concentration of the test chemical to which the cells are responding can differ from nominal concentrations as a result of several factors.

The potential factor which can alter the concentration of test chemicals in vitro are: (1) insolubility, (2) adsorption to culture vessel surfaces, (3) binding to media components, (4) spontaneous reactions with media components, surfaces and/or photochemical reactions, (5) volatilization and (6) uptake and metabolism by cells. Each of these factors is discussed below:

  1. Insolubility. This issue is discussed in more detail later in this section. Experimental protocols should be developed which distinguish between responses of the cell culture system to the test chemical in the soluble phase as compared to physical reactions to the test chemical in solid phase. If the test chemical itself is immiscible or is dissolved in a solvent which is immiscible with the aqueous phase of the culture medium, a two-phase system will result. If the density of the test chemical or the solvent in which it is dissolved is less dense than the medium, the test chemical will have little contact with cells. If the test chemical or the solvent in which it is dissolved is heavier than the medium, then cells in monolayer culture will be in direct contact with high concentrations of the test chemical. Quantitation of exposure concentrations becomes extremely difficult in this case because cells are exposed in a non-uniform manner.
  2. Adsorption to culture vessel surface. Many chemicals will adsorb on plastic/glass surfaces depending on the specific nature of the surface. Loss of test chemical to vessel surfaces will reduce the actual concentrations to which cells in culture are exposed. If a radio-labelled form of the test chemical is available, then adsorption can be readily documented. If a direct analytical procedure is available, then it can be used to measure loss of test chemical from media in the absence of cells. These can be used to correct for adsorption effects.
  3. Binding to media components. Plasma proteins are noted for their ability to bind xenobiotics (see Section I.H). In this respect, albumin has the broadest spectrum for binding. If serum or albumin is present in the culture medium, then the potential for binding must be considered. One way to avoid this problem is to eliminate serum and albumin from the culture medium during the exposure to the test chemical. However, this option has its own drawbacks. Dramatic changes in culture medium, such as withdrawal of serum, can have detrimental effects on the state of cell differentiation and, thus, can alter cellular responses to toxins (see Section I.I). Such alterations would complicate interpretation of test results. A more involved option to solving this problem is to investigate xenobiotic binding to serum proteins by any one of several techniques available, e.g., equilibrium dialysis. Careful studies can provide accurate measurements of affinity binding constants which can be used to estimate the free toxicant concentrations.
  4. Spontaneous reactions. Reactions with culture media components, as opposed to binding, and/or photochemical reactions, may convert the test chemical into various degradative products which may be more or less toxic than the original test chemical (see Section I.C). The concentration of the test chemical and its metabolites may change with time. If these reactions can occur in vivo, then such reactions may be relevant to the toxicological evaluation. If, on the other hand, the reactions observed are unique to the in vitro culture system, then these modifications of the original test chemical may produce toxicological artifacts. Such reactions can be detected utilizing analytical techniques which are sensitive to the initial test chemical and degradation products. Blank test vessels containing media but no cells can be assayed to determine the presences of degradative products. If pseudo-first-order reaction kinetics apply, then a degradation rate constant can be determined, and available concentrations of test chemical and its metabolites over the exposure period can be computed.
  5. Volatilization. For volatile test chemicals (see Section I.F), the available test chemical will decrease with time if precautions to prevent this effect are not taken. One solution to this problem is to expose cells to the test chemical in a sealed vessel. In this case, the concentration of test chemical available to cells in culture media will not be the nominal concentration, but will be determined by the partial pressure of the test chemical and the relative volumes of the medium and the headspace. The actual concentrations in the medium can be determined by analytical means using sampling techniques which prevent loss of volatile components or by theoretical calculations using the partial pressures and volumes.
  6. Cellular uptake and metabolism. The concentration of the test chemical in the medium can be influenced by the cellular component of the system in a time-dependent manner. Even in the absence of all the effects described above (1-5), the concentration of toxicant may decrease below the nominal concentration due to partitioning between the aqueous phase and the cellular phase. In addition, if cells metabolize the toxicant, then the concentration will decrease with time. Metabolism is non-linear with concentration and partitioning can be non-linear if it involves intracellular protein binding. Thus, the effect of these processes can result in concentration-dependent distortions of the concentration-time profile of the test chemical in the medium. In many cases, the ratio of medium to cell volume is great and metabolism is slow, thus effects are negligible in short-term test protocols. However, with strongly hydrophobic chemicals or high rates of cellular metabolism, these factors can be important and concentration corrections must be made.

It is always important to document the procedures for preparation of stock solutions and dosing of test cultures. If solvents must be used to dissolve test chemicals or to solubilize them in culture media, these procedures must be carefully followed to attain reproducible dilutions of test chemicals. If a chemical is highly reactive or susceptible to photoreactions, then care must be taken to prevent spontaneous reactions. The time between preparation of working solutions from certified, reference solutions to dosing of test cultures must be standardized. Photoreactive test chemicals must be protected from exposure to light and the toxicity testing must be conducted in appropriate lighting.

The requirement for documentation of test chemical concentration in the in vitro culture system leads to certain analytical difficulties. If in vitro test systems are to be used for rapid screening purposes, it is counterproductive to become burdened with extensive measurements of test chemical concentrations. However, if the in vitro test system is being used for regulatory purposes and/or mechanistic studies, it may be important to determine actual test chemical concentrations during the course of the in vitro toxicity study. Blanks must be used to determine the time course of test chemical concentration in the absence of cells to account for spontaneous reactions and absorption/adsorption to culture vessels. In regulatory/mechanistic studies, the time-integrated area-under-the-concentration-curve should be considered as a possible measurement for expressing exposure.

When the true solubility of a test material is less than the nominal concentration, the response of cells in culture to the material may be highly variable depending on environmental conditions. Whether soluble or insoluble, an important variable for quantitating concentration is the particle number concentration (PNC) of the test material. The total number of particles relative to the total number of cells will influence the probability of interaction of the test compound with the cells. Further, unless the number of particles present in the liquid phase is very large in comparison to the number of cells present, non-uniform interactions with the culture may result. Finally, transport kinetics by diffusion through the aqueous phase surrounding the cells will also be influenced by PNC.

When the test compound is fully soluble at the nominal concentration used, the PNC is equal to the molarity times Avogadro's number. When the compound is not fully soluble, then the PNC will be far less than the predicted molar composition. For complex mixtures whose components are partially soluble at the test concentration, under saturating conditions the actual molecular concentration of each component in solution is equal to the solubility of the compound, not to its nominal concentration. The solution concentration is, in this case, independent of the gross composition of test material; therefore, the response of the cells may become independent of dose. This will only be true if the dispersed particles of undissolved material do not themselves interact directly with the cells.

For some test materials, cells may respond directly to particles (see Section I.K). The carcinogenicity of asbestos particles is an example; asbestos may be dispersed as needle-like particles, but has no solubility whatsoever in water. Under these circumstances, the PNC is dictated by two factors:

  1. the composition of the test compound; and
  2. the colloidal structure of the dispersion utilized.

The acquisition of reproducible test data for materials whose solubility is less than the nominal dose concentration hinges critically on achieving control over both the concentration of the test material and the colloidal structure of the disperse phase. The parameters describing colloidal structure are the mean particle size and the size distribution. Both of these parameters may now be determined with relative ease using modern instrumentation. The most useful method for size analysis uses a QELS spectrometer equipped with computer data analysis. If both of these parameters are known, the numerical magnitude of the PNC can then be ascertained. The total concentration in solution then equals the solubility plus the PNC of dispersed particles (i.e., PNC = [total concentration - solubility]/[mean particle volume of dispersed particles times the density]). This calculation is based on basic principles of colloid science and has been used successfully (Laughlin, et. al 1983). Control over colloidal structure both enhance the reproducibility of the data and increases the intrinsic activity of the compound.


This has been a brief discussion of the potential technical problems which can arise when using in vitro cell culture systems to evaluate chemical toxicity. Obviously, all of these factors will not have an effect on each and every chemical tested. However, the conscientious investigator must be aware that these problems exist and must carefully monitor the performance of the test system to ensure the quality and accuracy of test results.


  1. Acton, R.T., and Lynn, J.D. (Eds.) Cell Culture and its Application, Academic Press, New York (1977).
  2. Barton, A.F. Handbook of Solubility Parameters and Other Cohesion Parameters, CRC Press, Boca Raton (1983).
  3. BASF Wyandotte Corporation, Data Sheets for Pluronics and Cloud Point, NJ 07054.
  4. Berky, J. and Sherrod P.C., (Eds.) In Vitro Toxicity Testing, 1975-76: Short Term In Vitro Testing for Carcinogenesis, Mutagenesis, and Toxicity. The Franklin Press, PA (1978).
  5. Borup-Christensen, P., Erb, K. and Jensen, J.C. Curing human hybridomas infected with Mycoplasma hyorhinis. J. Immunol. Methods 110: 237-40 (1988).
  6. Chen, T.R. In situ detection of mycoplasma contamination in cell culture by fluorescent Hoechst 33258 strain. Exp. Cell Res. 104: 255-62 (1977).
  7. Christolfalo, V.J. and Rothblatt, G. (Eds.) Nutrition and Metabolism of Animal Cells in Culture, Academic Press, NY (1973).
  8. Clode, A.M., Pratten, M.K., and Beck, F. A stage dependent effect of ethanol on 9.5 day rat embryos grown in culture and the role played by the concomitant rise in osmolality. Teratology 35: 395-403 (1987).
  9. Cour, I., Maxwell, G. and Hay, R.J. Tests for bacterial and fungal contaminants in cell cultures as applied at the ATCC. TCA Manual 5: 1157-1160 (1979).
  10. Croom-Brown, H, Correia, J.J., and Williams, R.C., Jr. The effects of elevated pH and high salt concentrations on tubulin. Arch. Biochem. Biophys. 249(2) 397-406 (1986).
  11. Dawson, M. Cellular Pharmacology, Charles C. Thomas, Publisher, IL (1972).
  12. DelGiudice, R.A. and Gardella, R.S. Mycoplasma infection of cell cultures: Effects, incidence, and detection. TCA Monogram 5: 104-115 (1984).
  13. Everett, D.H. Basic Principles of Colloid Science. Royal Society of Chemistry Paperbacks, Herts, England (1988).
  14. Fisher, P.B., Goldstein, M.I., Bryson, V. and Schaffner, C.P. Reduced toxicity of amphotericin B metyl ester (AME) vs. amphotericin B and fungizone in tissue culture. In Vitro 12: 133-40 (1976).
  15. Freshney, R.I. (Ed.) Culture of Animal Cells. A Manual of Basic Technique (First Edition). Alan R. Liss, Inc., NY (1983).
  16. Gilbert, D.A., Reid., Y.A., Hay, R.J. and O'Brien, S.J. Application of DNA fingerprints for cell line individualization. J. Human Gent. 1989 (Submitted for publication).
  17. Gourley, D.R.H. Interactions of Drugs with Cells. Charles C. Thomas, Publisher, Springfield, IL (1976).
  18. Hay, R.J. (Ed.) ATCC Quality Control Methods for Cell Lines, ATCC Publication, Rockville, MD., First edition (1985).
  19. Hay, R.J. The seed stock concept and quality control for cell lines. Anal. Biochem. 171:225-237 (1988).
  20. Hay, R.J. Preservation and characterization of cell lines. Animal Cell Culture: A Practical Approach, R.I. Freshney (Ed.), IRL Press, Washington, DC. pp. 71-112 (1986).
  21. Holohan, P.D., Sokol, P.P., Ross, C.R., Coulson, R., Trimble, M.E., Laska, D.A. and Williams, P.D. Gentamicin-induced increases cytosolic calcium in pig kidney cells (LLC-PK1). J. Pharmacol. Exp. Ther. 247: 349-354.
  22. Huegen, A.W., Cerney, A., Zinkernagel, R.M. and Neftel, K.A. Suppressive effects of B-lactam-antibiotics or in vitro generation of cytotoxic T-cells. Int. J. Immunopharmacol. 8: 723-729 (1986).
  23. ICI Americas, Inc. The HLB System: A Time-saving Guide to Emulsifier Selection, ICI Americas, Inc., Wilmington, DE (1987).
  24. Inaba, M., Nagashima, K. Growth-inhibitory activity of minocycline on various tumor cell lines in vitro. Gan to Kaguku Rycho, 13: 2337-41 (1986).
  25. Jakoby, W.B. and Pastan, I.H. (Ed.): Cell Culture (Volume LVIII), Methods in Enzymology (Colowick, S.P. and Kaplan, N.O., Editors-in-Chief). Academic Press, NY (1979).
  26. Klausner, A. Tissue culture for improved toxicology, Bio/Technology, 5: 779-786 (1987).
  27. Krause, H.J. and Juliano, R.L. Interactions of liposome-incorporated amphotericin B with kidney epithelial cell cultures. Mol. Pharmacol. 34: 286-297 (1988).
  28. Kruse, P.F. Jr., and Patterson, Jr., M.K. (Eds.) Tissue Culture, Methods and Applications, Academic Press, NY (1973).
  29. Kruyt, H.R. Colloid Science: Irreversible Colloids Vol. I. 1952. Elsevier Publishing Company, NY (1952).
  30. Kruyt, H.R. Colloid Science: Reversible Colloids Vol. II. 1952. Elsevier Publishing Company, NY (1952).
  31. Laughlin, R.G. Solution and structural requirements of surfactant hydrophilic groups. Advances in Liquid Crytals, Vol. 3 (G.H. Brown, Ed.) (1978), pp. 41-98.
  32. Laughlin, R.G., Munyon, R.L., Ries, S.K. and Wert, V.F. Growth enhancement of plants by femtomole doses of colloidally dispersed triacontanol. Science, 219, 1219-1221 (1983).
  33. Luk, C.K., and Dulfano, M.J. Effect of pH, viscosity and ionic-strength changes on ciliary beating frequency of human bronchial explants. Clin. Sci. 64: 449-451 (1983).
  34. Macy, M.L. Identification of cell line species by isoenzyme analysis. TCA Manual. 4: 833-836 (1978).
  35. Macy, M.L. Tests for mycoplasmal contamination of cultured cells as applied at the ATCC. TCA Manual. 5: 1151-1156 (1979).
  36. Mailhot, H., and Peters, R.H. Empirical relationships between the 1-octonol water partition-coefficient and 9 physicochemical properties. Environ. Sci. & Technol. 22: 1479-1488.
  37. MC Publishing Co. McCutcheon's Emulsifiers and Detergents, North American Edition, McCutcheon Division, Geln Rock, NY (1987).
  38. McKeehan, W.L., Barnes, D., Reid, L., Stanbridge, E., Murakami, H. and Sato, G. Mammalian cell culture. Science 1989 (submitted for publication).
  39. Nelson-Rees, W., Daniels, W.W. and Flandermeyer, R.R. Cross-contamination of cells in culture. Science 212: 446-452 (1981).
  40. Paul, J. Cell and Tissue Culure, 4th Edition, Williams and Wilkins, MD (1970).
  41. Rosen, M.J., Surfactants and Interfacial Phenomena, John Wiley and Sons, Inc., New York (1978).
  42. Vaughan, C.D. Using solubility parameters in cosmetic formulations. J. Soc. Cos. Chem., 36: 319-333 (1985).
  43. Vaughan, C.D. Solubility: Effects in product, package, penetration, and presentation. Cosmetics & Toiletries, 103: 47-69 (1988).
  44. Walker, R.J., Lazzaro, V.A., Duggin, G.G., Horvath, J.S. and Tiller, D.J. Synergistic toxicity of cyclosporin A and streptomycin in renal epithelial cell cultures. Res. Commun. Chem. Pathol. Pharmacol. 62:447-60 (1988).
  45. Williams, G.M., Dunkel, V.C., and Ray, V.A. (Eds.) Cellular Systems for Toxicity Testing, Annals N.Y. Acad. Sci. 407: 1-484 (1983).


Dr. John M. Frazier
Associate Director
The Johns Hopkins Center for Alternatives to Animal Testing
615 N. Wolfe Street
Baltimore, MD 21205
Dr. Mary Delong
Division of Toxicological Sciences
Johns Hopkins University
615 N. Wolfe Street
Baltimore, MD 21205
Dr. June Bradlaw
Division of Toxicological Sciences
200 C. Street, SW
Washington, DC 20204
Dr. Robert Hay
American Type Culture Collection
12301 Parklawn Drive
Rockville, MD 20852
Dr. Daniel Acosta
Outside Reviewer
Department of Pharmacology and Toxicology
University of Texas at Austin
College of Pharmacology
Austin, TX 78712
Dr. Robert G. Laughlin
Procter and Gamble Co.
Miami Valley Laboratories
P.O. Box 398707
Cincinnati, OH 45239-8707
Dr. Wallace McKeehan
Outside Reviewer
W. Alton Jones Cell Science Center
Old Barn Road
Lake Placid, NY 12946
Dr. Sharon J. Northup
Baxter Health Care Corporation
Baxter Technology Park
Round Lake, IL 60073
Dr. John Bunch
Mary Kay Cosmetics, Inc.
1330 Regal Row
P.O. Box 47310
Dallas, TX 752417
Dr. James H. Resau
University of Maryland
22 South Green Street
Baltimore, MD 21201
Ms. Melanie Smith
Skin Care Product Development
Mary Kay Cosmetics, Inc.
1330 Regal Row
P.O. Box 47310
Dallas, TX 75247