PLEASE NOTE THAT TO OBTAIN DETAILS OF THE PROTOCOL YOU SHOULD REGISTER

TO OBTAIN DETAILED PROTOCOL YOU SHOULD FIRST REGISTER AT ECVAM SIS

I
N
V
I
T
T
O
X

O
N
-
L
I
N
E

I
N
V
I
T
T
O
X

O
N
-
L
I
N
E

I
N
V
I
T
T
O
X

O
N
-
L
I
N
E

I
N
V
I
T
T
O
X

O
N
-
L
I
N
E

I
N
V
I
T
T
O
X

O
N
-
L
I
N
E

I
N
V
I
T
T
O
X

O
N
-
L
I
N
E

I
N
V
I
T
T
O
X

O
N
-
L
I
N
E

I
N
V
I
T
T
O
X

Protocol no. 66
IN VITRO PREDICTION OF THE MAXIMUM TOLERATED DOSE

The results of cytotoxicity tests in primary cultures of rat hepatocytes and in MDBK and McCoy cells can be used to predict the in vivo 4-wk maximum tolerated dose in rats and dogs. A correlation between in vitro cytotoxicity, as measured in this system, and LD50 values in rats and mice has also been established.

CONTACT

Dr. R. Shrivastava
RL-CERM Route de Marsat 63203 Riom Cedex France
Tel: 73 33 49 85 Fax: 73 33 13 97

RATIONALE

Chemical toxicity in vivo is ultimately a process that occurs at the cellular level and thus can be studied in vitro. The in vitro cytotoxicity of a test compound in cell lines of different sensitivities to toxic effects may be assumed to parallel the range of toxicity occurring in vivo in rats and dogs. The in vitro concentration of the test compound in mg/ml remains constant throughout the experiment and may be empirically expressed as an in vivo dosage in mg/kg/day. The minimum in vitro drug concentration which induces changes in cell morphology, LDH release or up to 50% cell mortality (CT50) is assumed to correspond to the drug dose in vivo which gives rise to initial or mild toxic signs, while the minimum in vitro drug concentration which elicits over 90% cell mortality (CT100) is assumed to correspond to the in vivo dose which gives rise to marked clinical signs. On the basis of these assumptions, the CT50 and CT100 values (mg/ml) in the less sensitive MDBK cells are transformed directly into mg/kg/day for the in vivo threshold and mild-to-moderate toxic doses, respectively, in rats, while the corresponding values in the more sensitive rat hepatocytes are used to obtain the in vivo values in dogs. McCoy cells, being of intermediate sensitivity, serve as an in vitro control.

BASIC PROCEDURE

Primary cultures of rat hepatocytes and confluent monolayer cultures of MDBK and McCoy cells are exposed to various concentrations of the test compound. The following parameters of cell growth and morphology are scored at 24 hr in hepatocytes and at 24, 48 and 72 hr in the cell lines: surface occupied by growing cells (cell lines only), changes in cell size and shape, presence of cytoplasmic vacuoles, cell detachment, dead and dying cells. Scoring is by a rating system ranging from + (mild) to +++ (100% cell death). At each time point studied, the cell lines are dispersed with trypsin-EDTA and cell viability is measured with Trypan Blue. The dye is added directly to hepatocytes in their culture plates. LDH release is measured in hepatocytes at 22 _ 2 hr. Cytotoxicity is expressed as CT50 and CT100 values (see above).

CRITICAL ASSESSMENT

Prediction of in vivo MTD It is recommended by the EEC countries that new pharmaceuticals be evaluated in a four-week maximum tolerated dose (MTD) study in a rodent and non-rodent species prior to a one-week test substance exposure to human beings in phase I clinical trials. The two species of choice are commonly the rat and the dog. Since the MTD study is the first test that is required, it is usually carried out with little knowledge of the likely toxicity of the test compounds. Therefore, the ability to obtain an approximate prediction of the in vivo toxic dose by means of a simple in vitro test could reduce the suffering caused during in vivo testing by permitting the selection of drug doses which would be better tolerated by the animals, and this might, in turn, also reduce the overall number of animals used in the MTD and other initial toxicity tests, by eliminating the testing of the most toxic doses. The hypothesis which forms the basis for the in vitro prediction of the in vivo MTD makes a number of assumptions and simplifications. The in vitro toxic concentration in mg/ml is converted directly into an in vivo toxic dose in mg/kg/day. This does not take into account the fact that in vitro drug concentrations remain at more or less the same level throughout the experiment and that cells in culture are directly exposed to the drug without any effects of absorption, diffusion, metabolism and excretion. Nevertheless, Shrivastava et al (1991) have found that even such a rough approximation appears to be valid for the majority of orally well-absorbed compounds that they examined. During the initial steps of establishing an in vitro-in vivo correlation, it is also possible to use a relative classification for predicted in vivo toxicity, i.e. the compound can be classified as very toxic, toxic, mildly toxic or non-toxic. The cytotoxicity parameters assessed in the systems described in this protocol are not mandatory. Other cytotoxicity tests may be used, provided that at least 2-3 indicators of cytotoxicity are determined. The cell types chosen for the test are intended to reflect the variation in sensitivity that occurs between species in vivo, and that represents the extremes of variability observed within a species. In vivo, dogs are usually more sensitive than rats to toxic effects. In vitro, sensitivity to cytotoxic effects depends not only on the species, but also on the tissue of origin. After examining a wide range of cell types, Shrivastava et al (1991) established that rat hepatocytes are generally more sensitive than MDBK cells, while McCoy cells have intermediate sensitivity. Thus, rat hepatocytes are assumed to represent a sensitivity in vitro that is comparable to the sensitivity of dogs in vivo, while the sensitivity of MDBK cells is assumed to be comparable to that observed in whole rats. McCoy cells are selected on the basis of their intermediate sensitivity to act as a control cell line in vitro. It is recommended that rat hepatocytes be used, because they are considered to be the best representative of a highly sensitive cell type. However, McCoy and MDBK cells may be substituted by other cells of comparable sensitivity. Toxicity in vivo may arise from the action of the original compound or from the formation of a toxic metabolite. The inclusion of rat hepatocytes within the present experimental system provides a means of testing whether toxicity relates to the parent drug or to its metabolites, provided that the metabolism of the parent drug takes place in the liver. The parent drug is first studied at various concentrations in cultures of rat hepatocytes and MDBK cells. In most cases, a concentration can be found at which the drug is toxic in hepatocytes but not in the MDBK cells. The supernatant from hepatocytes exposed for 24 hr to this concentration is transferred to fresh MDBK cells and studied for 72 hr. Any cytotoxicity arising in MDBK cells within this period will most probably be due to the toxic effects of a metabolite formed by the hepatocytes, since it has already been established that MDBK cells incubated for 72 hr with the parent drug did not show toxic effects. This approach assumes that the metabolite is excreted in sufficient quantity and is sufficiently stable to produce measurable effects in the MDBK cells. Other approaches that may be used to identify the effects of an unstable metabolite include the use of co-cultures of hepatocytes with target cells, and the use of a liver homogenate which will generate, but not detoxify, metabolites. Several factors which limit the relevance of in vitro testing to the in vivo situation must be borne in mind. The route of administration and the speed of injection will affect the plasma half-life of a drug and will therefore influence the expression of in vivo toxicity. In the case of the oral route, correction for the degree of oral absorption is necessary, since a poorly absorbed compound will be less toxic that one which is well absorbed. The length of the in vitro exposure period and the following observation period must be adapted to the types of cells being used in the system. For example, Shrivastava et al (1991) found that the best correlation with in vivo dog toxicity was obtained by studying in vitro hepatocyte toxicity at 20-24 hr. In relation to possible metabolic effects, it should be noted that primary rat hepatocytes begin to lose their metabolizing capacity after 24 hr in culture. On the other hand, in vivo toxicity in rats was best correlated with in vitro toxicity in MDBK cells at 48-72 hr. A study (Shrivastava et al, 1991) of 25 randomly selected compounds, for which in vivo MTD data in rats and dogs were available, found a better than 80% correlation between actual in vivo threshold and toxic doses in rats and dogs and in vivo values predicted from in vitro cytotoxicity results. However, in vitro predictions underestimated the in vivo toxicity of one compound in both rats and dogs, and of another compound in dogs. The first of these compounds was found to cause biochemical changes in vivo which were not accompanied by any other apparent clinical, haematological or histopathological abnormalities, while the toxicity of the second compound was judged to derive from an extension of its pharmacological effects. In conclusion, although in vitro prediction of in vivo toxicity using the systems described cannot fully replace in vivo testing, it can serve as a useful complement to in vivo MTD experiments. The systems are rapid, accurate, reproducible and more economic than whole animal tests. Results obtained in vitro can help in the prior selection of least toxic molecules out of a series of compounds with equivalent pharmacological activity, and they can serve in the selection of non-toxic or threshold toxic doses for subsequent in vivo testing, thus reducing the numbers of animals required and the degree of suffering imposed. Prediction of in vivo LD50 Positive correlations between CT50 and CT100 values for all three cell types and in vivo oral LD50 values for rats and mice have also been demonstrated and shown to permit the accurate prediction of LD50 values for at least 75% of 48 diverse chemicals (Shrivastava et al., 1992). A plot of log-transformed CT50 concentrations (in mM) against respective log oral LD50 values (in mM) gave a positive correlation of 0.80 for hepatocytes, 0.77 for McCoy cells and 0.83 for MDBK cells. When CT100 values were used, the respective correlations were 0.79, 0.76 and 0.80. The cytotoxicity parameters were assessed at 24 hr in hepatocytes, since these cells lose their physiological function after longer periods in culture, and at 72 hr in the cell lines to ensure a sufficient exposure period to evaluate any delayed toxic potential. On the whole, hepatocytes tended to be more sensitive than McCoy and MDBK cells. A significantly greater toxicity to hepatocytes than to the cell lines was demonstrated by nine of the 48 chemicals studied: phenol, 2,4-dichlorophenoxy acetic acid, potassium cyanide, warfarin, isoniazid, caffeine, hexachlorophene, propranolol and phenobarbital, while digoxin and 1,1,1-trichloroethane were less toxic to hepatocytes than to the other cells. A few chemicals showed lower toxicity in vitro than in vivo, especially those that did not dissolve well in the medium or formed a medium-solvent partition. Shrivastava et al. (1992) suggested that a better in vitro-in vivo correlation might be obtained with these compounds by the use of agitated suspension cultures which would allow direct contact between the test chemical and the cells. The degree of correlation that has been found between in vitro cytotoxicity and in vivo LD50 values gives grounds for hope that in vitro systems, such as the one described in this protocol, may eventually contribute toward replacing the LD50 test in animals, although the reservations expressed above concerning extrapolation from in vitro toxicity testing must be taken into account. The system described in this protocol requires hepatocytes from 10-20 rats in order to screen about 50 compounds in vitro. In vivo LD50 testing of the same number of compounds requires the use of about 1,000 rats or mice.

CHEMICALS TESTED

Over 2,500 compounds with different types of pharmacological activity have been tested in vitro, and the predictive capacity of in vitro results for in vivo toxicity has been verified for over 120 of these compounds.

TEST STATUS

In-house development.

REFERENCES

  1. Shrivastava, R., John, G. W., Rispat, G., Chevalier, A., and Massingham, R. (1991) Can the in vivo maximum tolerated dose be predicted using in vitro techniques? A working hypothesis. ATLA 19: 393-402.
  2. Shrivastava, R., Delomenie, C., Chevalier, A., John, G., Ekwall, B., Walum, E., and Massingham, R. (1992) Comparison of in vivo acute lethal potency and in vitro cytotoxicity of 48 chemicals. Cell Biol. Toxicol. 8: 157-170.

IP-66 February 1992