Adequate immunotoxicity testing in drug development

Adequate immunotoxicity testing in drug development

Toxicology Letters 149 (2004) 115–122 Adequate immunotoxicity testing in drug development Danuta J. Herzyk∗ , Elizabeth R. Gore Department of Safety ...

100KB Sizes 7 Downloads 126 Views

Toxicology Letters 149 (2004) 115–122

Adequate immunotoxicity testing in drug development Danuta J. Herzyk∗ , Elizabeth R. Gore Department of Safety Assessment, GlaxoSmithKline Pharmaceuticals, King of Prussia, PA 19406, USA

Abstract Modulation of the immune system can lead to either immunostimulation or immunosuppression and can be either intended or unintended. While many effects on the immune system’s components can be found as a result of a drug treatment or chemical exposure, true immunotoxicity occurs when such treatment results in adverse effects or defects in the immune response. Regulatory expectations to evaluate potential adverse effects of pharmaceuticals warrants a need for reliable and readily standardized methods. Moreover, criteria to classify a drug as an “immunotoxicant” need to be established. Examples of studies using a modified approach to measure T-cell-dependent antibody responses (the rat KLH model) and interpretation of the results in the context of immunotoxicity evaluation are discussed in this paper. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Immunotoxicity; T-cell-dependent antibody test; KLH antibodies

1. Introduction The immune system is a tightly regulated and very complex network of various lymphoid and other cell types interacting by cell-to-cell contact and communicating via soluble mediators such as cytokines. Over the past two decades, there has been a growing awareness that pharmaceuticals can cause a variety of immunologically mediated adverse effects in patients. Immune-mediated disorders can be related to immunostimulation, including hay fever or the more serious life-threatening conditions such as anaphylaxis, autoimmune hemolytic anemia or systemic lupus erythaematosus. At the other end of the spectrum, consequences of immunosuppression lead to increased ∗ Corresponding author. Tel.: +1-610-270-7781; fax: +1-610-270-7504. E-mail address: [email protected] (D.J. Herzyk).

susceptibility to the development of lymphomas and infections, particularly among patients receiving long-term (intended) immunosuppressive therapy following organ transplantation. There is agreement that reliable and readily standardized immunotoxicity methods are needed to address potential adverse effects on a target organ that is as dynamic and complex as the immune system. The primary antibody response to a T-cell-dependent antigen is widely viewed as a comprehensive evaluation of immune function as it enables assessment of the various components of the immune system (e.g., macrophages, T-helper cells and B lymphocytes) involved in the antigen-specific antibody response. Any drug-induced alteration in antigen processing, presentation, synthesis and release of interleukins, cell proliferation, differentiation and/or secretion is thought likely to modify the response (Luster et al., 1992). However, while a primary antibody to sheep red blood cells (conventionally

0378-4274/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2003.12.026

116

D.J. Herzyk, E.R. Gore / Toxicology Letters 149 (2004) 115–122

referred to as the plaque forming cell, PFC) assay has been extensively evaluated by the National Toxicology Program (NTP) with a large number of chemicals and pesticides as part of environmental safety testing, the use of this assay for evaluation of pharmaceuticals in drug safety assessment has been limited. One of the alternative approaches to immunotoxicity testing in drug development is an immunization model using the highly immunogenic and stable T-cell-dependent antigen, keyhole limpet hemocyanin (KLH) that can be used in multiple species. This paper focuses on the application of the modified model in rats, the conventional rodent species used in toxicology. The utility of the model in detecting suppression of immune function was demonstrated with the well-established immunosuppressive drugs: cyclosporin A, azathioprine and prednisolone. The results from this evaluation and application of the model to identify immunotoxic potential of new drugs in development are discussed.

2. Materials and methods All experimental procedures were conducted in accordance with NIH guidelines and reviewed by the GlaxoSmithKline Animal Care and Use Committee. 2.1. Animals Adult male and female Crl:CD Br (CD) rats (12–16 weeks of age) were obtained from Charles River Laboratories (Raleigh, NC) and allowed 1 week to acclimate prior to study. 2.2. Reagents Keyhole limpet hemocyanin was supplied as a lyophile by Sigma Chemical Co. (St. Louis, MO). Cyclosporin (NDC 0078-0110-22), azathioprine (NDC 55390-600-20) and prednisolone (NDC 0451-220104) were supplied by Hanna’s Pharmaceutical Supply Co., Wilmington, DE. Pharmaceutical grade olive oil was used as the vehicle for cyclosporin A and USP grade saline for both azathioprine and prednisolone. Cyclosporin, azathioprine and prednisolone were administered at maximum tolerated doses indicated in

previously published reports (Dean and Remandet, 1997; Crevel et al., 1997; ICICIS Group Investigators, 1997; Matsuura et al., 2000). Two proprietary compounds in drug development at GSK were also tested. 2.3. Study design 2.3.1. Immunization with KLH by various routes Various immunization routes were compared to determine optimal induction of an anti-KLH antibody response. Rats were given a single injection of KLH without adjuvant by the footpad or intravenous (iv) route. Blood was collected from a tail vein prior to immunization and on days 3, 5, 7, and 10 post-immunization. A terminal blood sample from vena cava was collected 14 days post-immunization. Serum from blood was frozen at −80 ◦ C until analyzed for the presence of anti-KLH IgM- and IgG-specific antibodies. 2.3.2. Evaluation of test compounds in the rat KLH model The effect of cyclosporin (Cyc), azathioprine (Aza) and prednisolone (Pred) on KLH antibody response following single immunization with KLH by footpad route was assessed. Based on existing information about immunosuppressive activities of all the selected drugs, dosing duration prior to immunization with KLH was shortened from 14 to 7 days and the established dosing routes (oral for Cyc and Pred, intraperitoneal for Aza) were used. Two drugs (designated as compounds X and Y) of unknown immunotoxicity were tested in the rat primary antibody response model at two dose levels (a maximal tolerated dose and the non-observed adverse effect level, previously determined in routine 28-day toxicity study). Vehicle or drug was administered daily by oral gavage for 28 days with KLH immunization (by foot pad or intravenous injection) on day 14 of dosing, and subsequent measurement of serum anti-KLH antibodies. 2.4. Anti-KLH antibody ELISAs Quantitative measurement of rat anti-KLH IgM and IgG serum antibodies was performed by ELISA using mouse monoclonal IgM and IgG1 for preparation

D.J. Herzyk, E.R. Gore / Toxicology Letters 149 (2004) 115–122

of standard curves. The assay range for the IgM- and IgG-specific ELISAs was 32–2000 and 5–1000 ng/ml, respectively. Briefly, 96-well microtiter plates were coated with KLH (500 ng/well) overnight at 4 ◦ C. After a 1 h blocking step with 5% goat serum in PBS, standard and test sample dilutions (prepared in 5% goat serum/0.05% Tween-20 in PBS) were added to plates for 2 h. The detection antibody, goat anti-rat IgM (␮) or IgG (␥) conjugated to horse radish peroxidase (Jackson Immunoresearch, West Grove, PA) was then added for 1 h, followed by 30–45 min incubation with tetramethylbenzidine substrate. Plates were read at 450 nm single wavelength on a 96-well plate spectrophotometer. All steps were carried out at RT unless otherwise indicated, and plates were washed in between each step using a commercially available ELISA wash solution (Biosource International, Camarillo, CA). Dilution curves of serum samples of pre-determined anti-KLH antibody concentrations were generated and compared to the respective standard curve (i.e., IgM or IgG) to compare the binding of rat serum antibodies and mouse mAbs. The data, indicating the reliability of the quantitative measurement by the established ELISA methods, have been reported previously (Gore et al., 2004). 2.5. Statistical analysis The data for antibody determination were expressed as geometric mean concentrations and ranges. Prior to conducting statistical tests, raw data were converted to the log 10 scale due to large animalto-animal variability. Converted data were analyzed for group differences between the KLHimmunized control group and KLH-immunized, drug-treated groups using a repeated measure ANOVA test. Separate statistical analysis was performed for IgM and IgG antibody results at the respective time of peak response whereby groups were compared at each timepoint using a repeated measure analysis. Group comparisons were expressed as ratios of geometric means between control and treatment groups. Two-way ANOVA with factors immunization method, sex, and interaction between immunization method and sex was used in the analysis when both males and females were evaluated. The within sex animal-to-animal variability estimate for each immu-

117

nization method was used to estimate the coefficient of variation.

3. Results 3.1. Comparison of magnitude and kinetics of primary antibody response to KLH Rat primary antibody response to KLH was compared using immunization by footpad (fp) or intravenous (iv) route. Immunization of rats with 300 ␮g KLH by fp or at 300 ␮g/kg (equivalent of approximately 100 ␮g per rat) by iv injection resulted in a robust antibody response with 100% induction of IgM- and IgG-specific antibodies. The peak in anti-KLH IgM and IgG antibody production was observed 5 and 14 days post-immunization, respectively (Fig. 1). Within the various immunization groups, there were no significant differences between the males and females in regard to antibody response to KLH. Furthermore, variance was equal among the KLH treatment groups. 3.2. Effect of immunosuppressive drugs on primary antibody response to KLH Evaluation of the KLH primary antibody response model in the rat with the immunosuppressive drugs (Cyc, Aza and Pred) was performed using footpad immunization and then with Cyc using iv immunization. The kinetics of the IgM- and IgG-specific antibody response in the rat were evaluated by ELISA. As shown in Table 1, treatment with Cyc (20 mg/kg per day), Aza (25 mg/kg per day) or Pred (5 mg/kg per day) caused approximately 60% reduction in IgM antibody production at the time of peak response (5 days post-immunization) compared to the respective vehicle control groups. Similarly, inhibition of the IgG-specific antibody response (≥90%, 14 days post-immunization) was observed following treatment with each of the immunosuppressive drugs (Table 2). The effect of Cyc on the primary antibody response to KLH given by iv injection was also assessed to determine if similar susceptibility to immunosuppression could be demonstrated utilizing the iv route of immunization. Treatment with Cyc (20 mg/kg per day) caused statistically significant suppression of

118

D.J. Herzyk, E.R. Gore / Toxicology Letters 149 (2004) 115–122

Fig. 1. Comparison of primary antibody response in rats administered KLH by footpad (300 ␮g per rat) or intravenous (300 ␮g/kg) route. Serum anti-KLH IgM and IgG antibodies were measured prior to immunization (day 0) and at the indicated time points post-immunization by quantitative ELISA (n = 6 per sex per group). Geometric mean concentration for a given group (male and female combined) is indicated (—).

anti-KLH IgM and IgG antibody production in response to KLH iv immunization (P ≤ 0.001, Fig. 2).

immunotoxicity studies followed the standard 28-day rat toxicity studies.

3.3. Effects of novel drug candidates on primary antibody response to KLH

3.3.1. Effects of compound X KLH immunization by footpad induced a robust, albeit highly variable, anti-KLH IgM and IgG primary antibody response with 100% incidence in control and drug-treated groups. Anti-KLH IgM antibody concentrations in the control group and the 300 and 1500 mg/kg dosed groups ranged from 27 to 1948, 20

The established rat KLH model was utilized to evaluate immunotoxic potential of two compounds being developed at GSK as new drug candidates, designated here as compounds X and Y. For both compounds the

D.J. Herzyk, E.R. Gore / Toxicology Letters 149 (2004) 115–122

119

Table 1 Effects of cyclosporin, azathioprine and prednisolone on the primary IgM antibody response to KLH Groupa

IgM (␮g/ml) post-KLHb Day 3

Day 5

Day 7

Day 10

Day 14

0 mg/kg Cyc, p.o. 20 mg/kg Cyc, p.o.

6.9 (2–17) 6.3 (2–16)

33.2 (7–175) 14.4∗ (8–108)

32.1 (8–79) 11.6∗ (5–91)

16.8 (6–41) 10.6 (7–52)

11.2 (3–28) 7.9 (4–32)

0 mg/kg Aza, i.p. 25 mg/kg Aza, i.p.

14.2 (0–43) 9.1 (0–21)

121.5 (30–682) 50.9 (17–347)

91.6 (18–733) 40.2 (14–169)

57.5 (11–469) 25.7 (10–100)

55.9 (10–646) 26.8 (10–75)

0 mg/kg Pred, p.o. 5 mg/kg Pred, p.o.

15.6 (9–25) 13.7 (0–23)

195.7 (60–420) 81.5∗ (33–201)

146.2 (46–397) 65.4∗ (24–215)

82.2 (38–199) 52.5 (28–195)

65.9 (27–142) 46.2 (26–262)

a All groups (n = 12 males) received immunization with KLH (300 ␮g per rat, fp) on day 7 during dosing between days 1 and 20 by either oral (p.o.) or intraperitoneal (i.p.) routes of administration. b Antibody concentrations expressed as geometric means in ␮g/ml; concentration range in each group is included in parenthesis. ∗ Statistically significant (P < 0.05) decrease relative to time-matched control.

to 1575, and 19 to 808 ␮g/ml, respectively. Anti-KLH IgG antibody concentrations ranged from 44 to 1857, 59 to 1264, and 208 to 1753 ␮g/ml, respectively. The geometric mean concentrations of serum anti-KLH IgM antibodies in the control, low- and high-dosed groups (n = 6 per sex per group) are presented in Table 3. While these mean values show a decreasing trend, there was no statistically significant difference between control and drug-treated groups. The anti-KLH IgG antibody response differed significantly between males and females (P = 0.0065) with the females eliciting an overall higher response (Table 3). Geometric mean concentrations of serum anti-KLH IgG antibody response showed an increasing trend with drug treatment in males, but not in females. Statistical analysis of the anti-KLH IgG

antibody data indicate no significant difference between control and drug-treated groups for either sex. These results indicate that the induction of a primary antibody response to KLH in rats treated with compound X resulted in adequate detection of IgM-specific antibodies and effective class switch to the IgG isotpye, reflecting a complete immune response to T-cell-dependent antigen in the immunocompetent host. Therefore, compound X does not cause immunosuppression in rats. 3.3.2. Effects of compound Y KLH immunization by iv route elicited an anti-KLH IgM and IgG antibody response in all rats in both the control and drug-treated groups. Antibody responses in male and female rats were comparable. Anti-KLH

Table 2 Effects of cyclosporin, azathioprine and prednisolone on the primary IgG antibody response to KLH Groupa

IgG (␮g/ml) post-KLHb Day 3

Day 5

Day 7

Day 10

Day 14

0 mg/kg Cyc, p.o. 20 mg/kg Cyc, p.o.

8.3 (0–12) 3.4 (0–12)

5.3 (0–16) 3.4 (0–13)

24.8 (3–172) 3.5∗ (0– 16)

42.6 (5–253) 3.3∗ (0–17)

63.6 (10–304) 3.1∗ (0–20)

0 mg/kg Aza, i.p. 25 mg/kg Aza, i.p.

5.2 (0–6) 7.1 (0–7)

13.0 (5–23) 6.7 (0–13)

133.0 (42–733) 22.0∗ (0–70)

250.5 (95–593) 24.4∗ (3–81)

436.9 (200–1933) 39.9∗ (4–167)

0 mg/kg Pred, p.o. 5 mg/kg Pred, p.o.

4.4 (0–4) 4.4 (0–4)

10.1 (5–21) 5.1∗ (0–21)

143.7 (42–493) 13.3∗ (0–57)

254.2 (90–641) 16.6∗ (2–111)

417.1 (126–1124) 27.1∗ (6–208)

All groups (n = 12 males) received immunization with KLH (300 ␮g per rat, fp) on day 7 during dosing between days 1 and 20 by either oral (p.o.) or intraperitoneal (i.p.) routes of administration. b Antibody concentrations expressed as geometric means in ␮g/ml; concentration range in each group is included in parenthesis. ∗ Statistically significant (P < 0.05) decrease relative to time-matched control. a

120

D.J. Herzyk, E.R. Gore / Toxicology Letters 149 (2004) 115–122

group) are presented in Table 4. While the mean values for IgM antibodies are approximately 40% lower in rats treated with high dose (1000 mg/kg) of the compound, the ranges of antibody concentrations in all groups were comparable and there was no statistically significant difference between the control and the high dose-treated group. Similarly, anti-KLH IgG antibody production was not statistically different between controls and the drug-treated groups. These results indicate that the induction of a primary antibody response to KLH in rats treated with compound Y was not compromised. Therefore, compound Y does not cause immunosuppression in rats. Fig. 2. Comparison of effects of cyclosporin on rat anti-KLH antibody response to footpad (fp) or intravenous (iv) immunization with KLH. Percent inhibition of anti-KLH IgM and IgG antibody production was determined by comparing geometric mean concentrations from immunized control group with concurrent drug-treated group [1 − (GMC treated group/ GMC control group) × 100]. Studies were conducted with 12 or 6 male rats per group for fp or iv route, respectively.

IgM antibody concentrations in the control group and the 100 and 1000 mg/kg dosed groups (including both sexes) ranged from 10 to 200, 23 to 591 (with the exception of one rat’s response of 2841), and 8 to 191 ␮g/ml, respectively. Anti-KLH IgG antibody concentrations ranged from 4 to 306, 2 to 328, and 1 to 471 ␮g/ml, respectively. The geometric mean concentrations of serum anti-KLH IgM antibodies in the control, low and high-dosed groups (n = 10 per sex per

4. Discussion We have developed a modified model for the primary response to T-cell-dependent antigens in rats using the ELISA. In this test system, we propose a single immunization with keyhole limpet hemocyanin by intravenous injection on study day 14, while drug is given on days 1–28, to enable assessment of both the IgM (peak response day 19) and IgG-specific (peak response day 29) primary antibody response as measured by ELISA. A major advantage of this system is detection of antigen-specific antibodies in the serum, a hallmark of the humoral immune response, as effector functions of both IgM and IgG antibodies take place systemically and not at the site of production (e.g.,

Table 3 Effects of compound X on rat anti-KLH antibody response post-footpad immunizationa Dose (mg/kg per day) on days 1–28

Sex

IgMb Day 0

Sex Day 19

IgGb Day 0

Day 29

M F

NQ NQ

118.8 (27–689) 385.7 (43–1948)

Mc Fc

NQ NQ

136.5 (44–278) 480.5 (214–1857)

300

M F

NQ NQ

106.0 (20–471) 252.8 (100–1575)

M F

NQ NQ

204.6 (54–909) 499.6 (84–1264)

1500

M F

NQ NQ

143.8 (62–321) 71.8 (33–808)

M F

NQ NQ

376.2 (270–650) 501.4 (208–1753)

0

a Anti-KLH IgM and IgG antibody results taken from time of peak response: study days 19 and 29 (days 5 and 14 post-immunization, respectively); n = 6 per sex per group. b Antibody concentrations expressed as geometric means in ␮g/ml; concentration range in each group is included in parenthesis; NQ: not quantifiable. c Males and females significantly different (P = 0.0065).

D.J. Herzyk, E.R. Gore / Toxicology Letters 149 (2004) 115–122

121

Table 4 Effect of compound Y on rat anti-KLH antibody response post-intravenous immunizationa Dose (mg/kg per day) on days 1–28

Sex

IgMb

IgGb

Day 0

Day 19

Day 0

Day 29

0

M F

NQ NQ

60.3 (10–116) 96.4 (48–200)

NQ NQ

50.8 (4–306) 62.2 (9–223)

100

M F

NQ NQ

118.3 (23–591) 96.9 (31–2841)

NQ NQ

104.6 (35–328) 34.8 (2–162)

1000

M F

NQ NQ

36.9 (8–114) 46.0 (17–191)

NQ NQ

34.0 (1–471) 103.0 (22–337)

a Anti-KLH IgM and IgG antibody results taken from time of peak response: study days 19 and 29 (days 5 and 14 post-immunization, respectively); n = 10 per sex per group. b Antibody concentrations expressed as geometric means in ␮g/ml; concentration range in each group is included in parenthesis; NQ: not quantifiable.

spleen, bone marrow and lymphoid tissue) (Abbas et al., 2000). A caveat to the model, however, is the high degree of variability in the antibody response to KLH, albeit not unexpected for an outbred strain of rats such as the Sprague Dawley (Sovcikova et al., 2002). Moreover, similar variability is seen using the PFC assay (personal communications). The inherent, high variability in rat antibody production can be greatly challenging with regard to data interpretation, yet cannot be avoided as interindividual differences is a natural feature of the functioning immune system. The type of antigen, the route of immunization and the method of antibody analysis seem to be secondary to the prominent factor of “individuality” of one’s immune system. Despite the high animal-to-animal variability, our rat KLH model has demonstrated utility in predicting immunosuppressive potential based on its evaluation with three classic immunosuppressive drugs: cyclosporin, azathioprine and prednisolone. All three drugs, with diverse mechanistic properties, caused statistically significant immunosuppression of the KLH-induced primary antibody response in the presence of high interindividual differences. Again, this model seems to be very comparable to the model that uses sheep red blood cells (SRBC) as antigen, as consistent suppressions of anti-SRBC antibodies (in PFC assay or ELISA) by immunosuppressive positive controls are demonstrated in periodical testing to confirm validity of the model (personal communications). The advantages of the KLH immunization model over alternative approaches are emphasized by use of a stable, soluble antigen; ability to preserve samples for

batch testing; endpoint evaluation by a readily standardized and quantitative test system; and evaluation of both IgM and IgG primary antibody isotypes that allow for holistic assessment of the immune response. Rats dosed with compounds X or Y demonstrated ample production of both IgM- and IgG-specific antibodies relative to controls. The later finding is evidence of an effective isotype class switch, dependent on T-cells and cytokine production (Chin et al., 1995; Lederman et al., 1994). IgG-specific antibodies, like IgM, activate the classical pathway of complement as a mechanism of clearance, but also play a role in other effector functions including the following: opsonization of antigens for phagocytosis by macrophages and neutrophils; antibody-dependent cell-mediated cytotoxicity involving natural killer cells and macrophages; and feedback inhibition of B cell activation (Abbas et al., 2000). Our data show that IgG-specific antibodies, that have enhanced effector functions, are more sensitive to the effects of well-known immunosuppressive compounds compared to IgM-specific antibodies. This finding suggests that the IgG endpoint may be of greater value than IgM in the assessment of immunotoxic potential of development compounds. In this test system, compounds X and Y did not have adverse effects on immune function. The interpretation of the data was based on both the biological context and the statistical evaluation. The described modified rat KLH model is robust, practical and provides comprehensive evaluation of immune responsiveness, however, it does not im-

122

D.J. Herzyk, E.R. Gore / Toxicology Letters 149 (2004) 115–122

prove the high animal-to-animal and study-to-study variability seen with the existing T-cell-dependent antibody tests. Although consistent in detection of suppression with the intentional immunosuppressive drugs, the results, even if derived from the limited number of studies, suggest that a likelihood of detection of unexpected, but clinically relevant, immunosuppression of novel compounds by the immune function test used as a screening tool is quite low. Moreover, immunotoxicological evaluation of 27 new molecular entities over 10 years revealed a significant modification of immune parameters in only two cases, both of which had immunomodulatory activities (Dean et al., 1998), therefore, their effects on the immune function tests would not be unexpected. While evaluation of immunotoxicity potential of new drugs is important, the “prescribed” approaches focused on early screening of compounds may not be optimal. Consideration of tailored programs designed to address the most relevant effects on immune function assessment, based on overall knowledge of drug activities, should become more prominent in immunotoxicological testing of pharmaceuticals.

References Abbas, A.K., Lichtman, A.H., Pober, J.S., 2000. Effector mechanisms of humoral immunity. In: Abbas, A.K., Lichtman, A.H.,

Pober, J.S. (Eds.), Cellular and Molecular Immunology, 4th ed. Saunders, Philadelphia, pp. 309–334. Chin, L.T., Malmborg, A.C., Kristensson, K., 1995. Mimicking the humoral immune response in vitro results in antigen-specific isotype switching supported by specific autologous T helper cells-generation of human HIV-1-neutralizing IgG monoclonal antibodies from naive donors. Eur. J. Immunol. 25, 657–663. Crevel, R.W.R., Buckley, P., Robinson, J.A., Sanders, I.J., 1997. Immunotoxicological assessment of cyclosporin A by conventional pathological techniques and immune function testing in the rat. Hum. Exp. Toxicol. 16, 79–88. Dean, J.H., Remandet, B., 1997. Integration of immunotoxicology evaluation in drug development. Drug Inform. J. 31, 1347– 1356. Dean, J.H., Hinks, J.R., Remandet, B., 1998. Immunotoxicology assessment in the pharmaceutical industry. Toxicol. Lett. 102–103, 247–255. Gore, E.R., Gower, J.K., Kurali, E., Sui, J.-L., Bynum, J., Ennulat, D., Herzyk, D.J., 2004. Primary antibody response to Keyhole Limpet Hemocyanin in rat as a model for immunotoxicity evaluation. Toxicol (in press). ICICIS Group Investigators, 1997. Report validation study of assessment of direct immunotoxicity in the rat. Toxicology 125, 13–201. Lederman, S., Yellin, M.J., Cleary, A.M., 1994. T-BAM/CD40-L on helper T lymphocytes augments lymphokine induced B cell Ig isotype switch recombination and rescues B cells from programmed cell death. J. Immunol. 152, 2163–2171. Luster, M.I., Portier, C., Pait, D.G., 1992. Risk assessment in immunotoxicology: sensitivity and predictability of immune tests. Fundam. Appl. Toxicol. 18, 200–210. Matsuura, M., Imayoshi, T., Okumoto, T., 2000. Int. J. Immunopharmacol. 22, 323–331. Sovcikova, A., Tulinska, J., Kubova, J., 2002. Effect of cyclosporin A in Lewis rats in vivo and HeLa cells in vitro. J. Appl. Toxicol. 22, 153–160.