- Email: [email protected]
Variance of body and organ weights in 28-day studies in mice
Regulatory Toxicology and Pharmacology 108 (2019) 104472
Contents lists available at ScienceDirect
Regulatory Toxicology and Pharmacology journal homepage: www.elsevier.com/locate/yrtph
Variance of body and organ weights in 28-day studies in mice ∗
T
Heike Antje Marxfeld , Karin Küttler, Martina Dammann, Sibylle Gröters, Bennard van Ravenzwaay Experimental Toxicology and Ecotoxicology, BASF SE, 67056, Ludwigshafen, Germany
A R T I C LE I N FO
A B S T R A C T
Keywords: Organ weight Mouse Immune system Toxicity Endocrine disruptor effects
The OECD guideline 407 outlines the conduct of 28-day studies in rodents to detect systemic toxicity with focus on endocrine and immunotoxic effects. It was validated with the rat as preferred model species. Justification is required for other rodent species, as an increased variability is expected compared to the rat. We investigated the variability of organ weights in the mouse and compared this to data published for the rat in the validation report of test guideline 407. Furthermore, the influence of the immunotoxic model substance cyclophosphamide on spleen and thymus weights in the mouse in immunotoxicity studies (duration 28 days) is reported and discussed, an immunotoxic model substance was not included in the validation report. Historical control data were compiled for mouse studies performed according to OECD 407 and for immunotoxicity studies between 2008 and 2013 at BASF SE. For absolute weights, the coefficient of variation was determined for each study group and compared with the rat. Adrenal glands, ovaries and to lesser degree testes and prostate showed higher coefficients of variation in the mouse (most pronounced in adrenal glands in male animals: rat 5%–17%, CD1 mouse 20%–51%). Effects of cyclophosphamide were best detected measuring the thymus weight.
1. Introduction The OECD guideline 407 gives guidance on the conduct of 28-day studies in rodents for the detection of systemic toxicity with special focus on endocrine and immunotoxic effects. This guideline was validated using the rat as the test species (OECD Guidelines for Testing of Chemicals; Method No. 407, 2008). However, the 407 guideline also gives the experimenter the choice to use other rodents like the mouse. According to this guideline, an increased variability is to be expected using the mouse which is due to technical challenges dissecting these small animals (OECD Guidelines for Testing of Chemicals; Method No. 407, 2008, point 11 “Selection of animal species” page 2–3). The aim of the present investigation was to determine the variability in organ weights in mice to provide guidance for the assessment of potential changes and set them in context with the rat. Furthermore, we report and discuss the influence of the immunotoxic model substance cyclophosphamide on spleen and thymus weights in the mouse in immunotoxicity studies where it serves as a recommended positive control. An immunotoxic model substance was not included in the validation report of the updated test guideline 407. To achieve this, we compiled data from 28-day studies in the mouse performed according to OECD 407 and for spleen and thymus weights
from immunotoxicity studies (according to EPA OPPTS 870.7800) at BASF SE from 2008 to 2013. The variance of terminal body weight and those organ weights mentioned in OECD 407 were investigated. Mandatory weights are those of liver, kidneys, adrenal glands, testes, epididymides, prostate, seminal vesicles with coagulating glands, thymus, spleen, brain, and heart of all animals. Optionally, the weight of ovaries, uterus, and thyroid glands can be determined. Spleen and thymus weights are the only weights required to be taken in immunotoxicity studies. 2. Materials and methods All studies were performed in an AAALAC-approved laboratory at BASF SE, Ludwigshafen, Germany, in accordance with the German Animal Welfare Act and the effective European Council Directive 2010/ 63/EU. All necropsies and organ weight measurements were conducted inhouse by highly experienced (> 15 years) and specialized technicians. Data were recorded in the BASF SE pathology data capture system and exported to SAS for the calculations of coefficients of variation (CVs) in this paper.
∗
Corresponding author. BASF SE, RB/TD - Z470, 67056, Ludwigshafen, Germany. E-mail addresses: [email protected] (H.A. Marxfeld), [email protected] (K. Küttler), [email protected] (M. Dammann), [email protected] (S. Gröters), [email protected] (B. van Ravenzwaay). https://doi.org/10.1016/j.yrtph.2019.104472 Received 20 March 2019; Received in revised form 13 August 2019; Accepted 3 September 2019 Available online 05 September 2019 0273-2300/ © 2019 Elsevier Inc. All rights reserved.
Regulatory Toxicology and Pharmacology 108 (2019) 104472
H.A. Marxfeld, et al.
OECD, 2006a). These CVs were additionally calculated from the data in Annex 5 (OECD, 2006b) to get an idea which variability might be expected in these organs. The number of studies on which this calculation is based varied from 5 to 21 (Calculations are provided in the supplementary data). Different rat strains were used in the validation studies, this information was, however, not detailed in the validation report. The ranges of CVs are defined in the validation report as “low” variability 8–11, mid 12–21 and high > 24 (p. 197, OECD, 2006a). Very low variability (CV < 8) was not included as a class in the validation report. A power analysis was performed to determine which difference is expected to be detected for a given CV. For this analysis a significance level of 5% and a power of 75% was assumed. The power calculation for the two-sided Wilcoxon test was performed using the software NQUERY (Nquery Technical Solutions). For the sake of this power calculation, the underlying distribution was assumed to be normal. The power calculation using the two-sided Dunnett test was based on the following assumptions:
Table 1 Animal data. Supplier
Strain
Age at beginning of study
Abbreviation
Number of control mice/sex
Number of studies
Janvier Janvier CRL CRL
C57BL/j Rj C57BL/j Rj CD-1 CD-1
5–7wks 7–8wks 5–7wks 7–8wks
J5-7 J7-8 CRL5-7 CRL7-8
15 23 30 5
3 4 4 1
2.1. Mice: 28-day studies according to test guideline OECD 407 Historical control data were compiled for 12 different 28-day studies (details are provided in Table 1) performed between 2008 and 2013 according to OECD 407 with mice from Janvier C57BL/j Rj (J) and Charles River CD-1 (CRL) including terminal body weight and the following organs: Adrenal glands, brain, epididymides, heart, kidneys, liver, ovaries, prostate, seminal vesicles (with coagulating glands), spleen, testes, thymus, uterus. Thyroid gland weights, which are optional in the OECD 407 guideline were not measured.
• a design of 3 dose groups and one control having 5, 8 or 10 animals per group, respectively • on the probability of a significant effect of the highest of three dose groups compared to the control
2.2. Mice: immunotoxicity studies (duration of 28 days) The calculation was performed using SAS functions probmc for the determination of the critical value of the Dunnett test and probt for the calculation of the power (SAS Institute Inc). The normal distribution of data of the mouse studies was checked using the Shapiro-Wilk test (Shapiro and Wilk, 1965).
Additionally, data were compiled for 8 immunotoxicity studies performed between 2008 and 2013 according to EPA Health Effects Test Guidelines OPPTS 870.7800 with mice from Janvier C57BL/j Rj (J). 48 control females and 16 control males were investigated as well as cyclophosphamide treated positive control groups with the same number of animals. They were all 5–7 weeks old at the beginning of the studies. Per study, 8 animals/group were used as this is the minimal number permitted in the guideline. The following parameters were included: Number of animals; terminal body weight, mean absolute and relative weights of spleen and thymus; and CVs. The number of animals tested should yield sufficient statistical power to detect a twenty percent change based upon the inter animal variation which may be encountered in these assays.
4. Results 4.1. 28-Day studies According to OECD 407, at the commencement of the study the body weight variation of animals used should be minimal and not exceed ± 20% of the mean weight of each sex. All studies fulfill these requirements.
2.3. Rats
4.2. Organ weights
Organ weight data for the rat was taken from the report of the validation of the updated test guideline 407 and Annex 5 of this document, referred to in the text as “validation report” or “Annex 5”.
A comparison of coefficients of variation for absolute organ weights for male and female mice of different strains and rats is presented in Tables 2 and 3. Scatterplots showing the distribution of coefficients of variations for male and female mice are shown in Figs. 1 and 2. In male mice, adrenal glands (especially in CRL CD1 mice), show a noticeable higher variability in both strains of mice compared to the
2.4. Statistics 2.4.1. Mouse 28-day studies according to guideline 407
Table 2 Comparison of minimum – maximum ranges of CVs of absolute weights in male rats and mice.
To judge the variation of the data the coefficient of variation (CV: standard deviation divided by the mean in percent) was determined for each control study group of mice in 28-day studies. The minimum – maximum ranges of the group means were compared to the rat data. 2.4.2. Mouse immunotoxicity studies The CV was determined for absolute and relative spleen and thymus weights of control and treated animals. Furthermore, the percent deviation of positive control weights from control weights was calculated. 2.4.3. Rat The validation report lists CVs only for absolute weights of selected organs; brain, kidney, heart, spleen and thymus are missing (p. 197,
Organ
Mouse C57BL/Rj CV(%) Based on 7 studies
Mouse CRL CD1 CV(%) Based on 5 studies
Adrenal glands Brain Epididymides Heart Kidney Liver Prostate Seminal vesicles Spleen Testes Thymus
6–35 1–4 5–20 2–18 4–13 4–11 9–31 4–18 3–20 a(10–20) 4–21 4–21 a(9–17)
20–51 2–13 4–10 6–15 3–17 6–11 14–26 9–29 10–21 7–15 18–29
a
2
5–17 2–4 5–13 4–13 3–19 8–15 11–22 8–24 6–25 5–12 8–29
Ranges in Immunotoxicity studies in brackets (2 studies).
Rat CV (%)
Regulatory Toxicology and Pharmacology 108 (2019) 104472
H.A. Marxfeld, et al.
rat. Testes (C57BL/Rj) and prostate (C57BL/Rj) show a slightly higher variability in mice. In female mice, ovaries and thymus showed a noticeable higher variability (especially in CRL CD1 mice), compared to the rat.
Table 3 Comparison of minimum – maximum ranges of CVs of absolute weights in female rats and mice. Organ
Mouse C57BL/Rj CV(%) Based on 7 studies
Mouse CRL CD1 CV(%) Based on 5 studies
Rat CV(%)
Adrenal glands Brain Heart Kidney Liver Ovary Spleen Thymus Uterus
6–17 2–5 4–12 3–9 4–13 6–31 7–17 a(9–15) 5–28 a(10–25) 12–45
6–28 4–6 1–11 6–11 11–19 13–51 17–23 16–40 22–32
7–20 2–4 6–13 2–13 6–17 10–33 6–25 9–27 8–42
a
4.3. Immunotoxicity studies (duration 28 days) Organ weights of spleen and thymus are requested in the EPA guideline OPPTS 870.7800 (EPA, 1998). The following investigations were designed to examine the suitability of this parameter to detect an immunotoxic effect. According to this guideline, at least eight animals should be included in each dose and control group. The number of animals tested should yield sufficient statistical power to detect a twenty percent change based upon the inter animal variation which may be encountered in these assays. A power analysis was performed which based on the two-sided Dunnett test taking a significance level of 5% and a power of 75%. Similarily to the validation document of OECD407, page 188, the highest dose was presumed to have changed versus the control with the low and intermediate doses similar to the control. In addition to the Dunnett test the Wilcoxon test was taken, because after checking with the Shapiro-Wilk test the normal distribution was not always given in our data (data not shown). Based on the assumption of the performed power analyses, the following conclusions can be drawn: If the aim is to detect a 20% difference from the control.
Ranges in Immunotoxicity studies in brackets (6 studies).
• For a group size of 5 the CV should be lower than 7% or no more than 10% using the Wilcoxon or the Dunnett test, respectively. • For a group size of 8 the CV should be lower than or equal to 12% using the Wilcoxon test or Dunnett test. • For a group size of 10 the CV should be lower than or equal to 14% using the Wilcoxon test or Dunnett test.Table 4
In Tables 5 and 6, the CVs of absolute and relative spleen or thymus weights of each study and the deviation of the positive control (cyclophosphamide) from the control are shown. The statistically significant difference with a significance level of 5% is denoted with one star (*) and with a significance level of 1% with two stars (**) using the two-sided Wilcoxon-test. Critical CVs above 12% for the group size of 8 are marked in bold. The ranges of CVs in immunotoxicity studies were within the range for thymus and spleen in 28-day studies according to OECD 407 (shown in brackets in Tables 2 and 3).
Fig. 1. Scatterplot of coefficients of variation in male mice. Abbreviation used in Fig. 1: Adrenal: Adrenal glands, sem ves: seminal vesicles.
4.4. Spleen () The magnitude of the weight change in the spleen induced by cyclophosphamide was mostly lower than 20%, only in two studies the change in absolute weights exceeded 20%, while none of the relative weight changes was higher than 20%. Based on the power calculation, the detection of a 20% change can only be expected if the CVs of both control and positive control group are lower than or equal to 12%. This is only given in 3/8 studies for absolute spleen weights and in 6/8 studies for relative spleen weights. The percent deviation of positive control from control was higher using absolute weights than relative weights. Therefore, a significant effect could be detected in 6/8 studies with absolute weights and in 3/8 studies with relative weights. In study 5, one animal in the treated group showed a markedly higher spleen weight (0.134 g) than both the other animals in this group (range from 0.04 g–0.088 g) as well as the concurrent controls (range from 0.057 g–0.075 g). In the absence of histopathology data, no further conclusions can be drawn.Table 5
Fig. 2. Scatterplot of coefficients of variation in female mice. Abbreviation used in Fig. 2: Adrenal: Adrenal glands.
3
Regulatory Toxicology and Pharmacology 108 (2019) 104472
H.A. Marxfeld, et al.
Table 4 Analysis of power. CV
6% 7% 8% 10% 12% 14% 16% 18% 20% 25% 30%
Percent deviation (n = 5)
Percent deviation (n = 8)
Percent deviation (n = 10)
Wilcoxon
Dunnett
Wilcoxon
Dunnett
Wilcoxon
Dunnett
18% 20.5% 23% 29% 35% 41% 47% 53% 59% 73% 88%
12% 15% 17% 21% 25% 29% 33% 37% 42% 52% 62%
10% 12% 13% 17% 20% 23% 27% 30% 33% 42% 50%
10% 11% 13% 16% 19% 22% 25% 29% 32% 40% 48%
8% 10% 11% 14% 17% 20% 22% 25% 28% 35% 42%
8% 10% 11% 14% 17% 20% 22% 25% 28% 35% 42%
CVs as in the rat with the notable exception of higher CVs for adrenal glands, ovaries, and to a lesser degree testes and prostate, which might implicate that the mouse is less suitable to detect endocrine related effects than the rat with standard group sizes of 5 animals per test group. Also in the rat, only clear antiestrogenic (Ethinyl Estradiol, Tamoxifen, CGS, 18320B) and antiandrogenic (flutamide) substances were successfully detected, while weaker antiestrogenic (Genistein, Nonylphenol) and antiandrogenic (1,1-dichloro-2,2-bis-(p-chlorophenyl)ethylene) substances did not show convincing weight effects according to the validation report (OECD, 2006b) and were only detected on histopathological examination. Although the focus of the immunotoxicity study according to OPPTS 870.7800 is the T-cell dependent antibody response, thymus and spleen weights are also required to be taken and evaluated in this study design. A decrease of weight interpreted as an immunotoxic effects of cyclophosphamide was successfully detected with the evaluation of thymus weights rather than spleen weights. The number of 8 animals per group was sufficient in 7/8 studies. One study did not show an effect of cyclophosphamide, due to high variation in thymus weights in the treated group caused by one animal. Histopathological assessment would be beneficial in the evaluation of questionable cases like this even if it is not required by the guideline for Immunotoxicity studies (EPA, 1998). Weight determination of the spleen seems to be a less valuable parameter in this study type as the effects of the positive control substance were lower (often less than 20% change) than the sensitivity aimed for in this study design. Concluding, in this study design only immunotoxic substances with equal or greater potency than cyclophosphamide can be detected using the thymus weights.
4.5. Thymus () The magnitude of the weight change in the thymus induced by cyclophosphamide was higher than 30% in 6/8 studies, which was detected as a significant effect in these studies. The CVs of both control and positive control group were lower than or equal to 12% in only 2/8 studies for absolute and relative thymus weights, which implies that a smaller effect would not have been detected. A significant effect could be detected in 7/8 studies with absolute weights and in 6/8 studies with relative weights. In study 8, which did not show an effect with cyclophosphamide, one animal in the treated group showed a markedly higher thymus weight (0.154 g) than both the other animals in this group (range from 0.014 g–0.041 g) as well as the concurrent controls (range from 0.028 g–0.051 g). If this animal is excluded from calculations, the group mean weights decrease and the effect is significant. In the absence of histopathology data, no further conclusions can be drawn.Table 6 5. Discussion This investigation was designed to assess the suitability of the mouse compared to the rat in studies conducted in accordance with Test Guideline 407 (OECD, 2008) focusing on organ weight assessment. Furthermore, we report the effects of cyclophosphamide on spleen and thymus weights in the mouse to expand the data in the validation report for test guideline 407 (OECD, 2006b) with an immunotoxic model substance, as only endocrine active substances were reported (OECD, 2006a,b). Generally, the coefficient of variation of organ weights determines the magnitude of an effect which can be detected statistically with a given group size. Most organ weights in control mice showed similar Table 5 CVs of mean absolute and relative spleen weights in immunotoxicity studies. Study number
1
2
3
4
5a
6
7
8
Number of animals
8
8
8
8
8
8
8
8
Sex
m
f
m
f
f
f
f
f
20 19 −22∗
11 16 −14∗
10 11 −25**
10 6 −4
12 45 0
9 7 −14**
9 16 −17**
15 16 −18∗
15 24 −12
12 12 −7
11 10 −17**
8 4 −1
12 45 +10
8 7 −9∗
9 12 −12∗
12 11 −9
Absolute weights Control group CV (%) Positive control group CV (%) Deviation of positive control from control (%) Relative weights Control group CV (%) Positive control group CV (%) Deviation of positive control from control (%)
m = male, f = female, *: p ≤ 0.05, **: p ≤ 0.01. CVs above 12% are marked in bold. a One animal is an outlier. 4
Regulatory Toxicology and Pharmacology 108 (2019) 104472
H.A. Marxfeld, et al.
Table 6 CVs of mean absolute and relative thymus weights and detected effects in immunotoxicity studies. Study number
1
2
3
4
5
6
7
8a
Number of animals
8
8
8
8
8
8
8
8
Sex
m
f
m
f
f
f
f
f
17 41 −39**
11 21 −17*
9 19 −50**
11 12 −38**
10 12 −33**
10 14 −35**
25 20 −34**
19 101 +8
20 55 −29*
13 21 −10
10 17 −44**
11 12 −36**
8 12 −30**
10 14 −32**
24 18 −31**
18 107 +19
Absolute weights Control group CV (%) Positive control group CV (%) Deviation of positive control from control (%) Relative weights Control group CV (%) Positive control group CV (%) Deviation of positive control from control (%)
m = male, f = female, *: p ≤ 0.05, **: p ≤ 0.01. CVs above 12% are marked in bold. a One animal is an outlier cf text.
6. Conclusion
Appendix A. Supplementary data
Adrenal glands and ovaries and to lesser degree testes and prostate showed higher coefficients of variation in the mouse in studies performed according to OECD 407, all other parameters were comparable between the two species. Therefore, the detection of endocrine effects in the mouse might require larger group sizes than in the rat. In immunotoxicity studies (EPA, 1998), the group size of 8 animals per group is only sufficient to detect the effects of immunotoxic substances with equal or greater potency than cyclophosphamide measuring the weight of the thymus.
Supplementary data to this article can be found online at https:// doi.org/10.1016/j.yrtph.2019.104472. References EPA, 1998. U.S. EPA Health effects test guidelines OPPTS 870.7800; immunotoxicity (aug 1998). https://www.regulations.gov/document?D=EPA-HQ-OPPT-2009-01560049, Accessed date: 12 July 2019. Nquery Technical Solutions, Miami, Florida 33176, USA. OECD, 2008. Guidelines for testing of Chemicals; method No. 407: repeated dose 28-day oral toxicity study in rodents; adopted 03 oct 2008. https://www.oecd-ilibrary.org/ environment/test-no-407-repeated-dose-28-day-oral-toxicity-study-in-rodents_ 9789264070684-en, Accessed date: 12 July 2019. OECD, 2006a. Environmental Health and Safety Publications Series on Testing and Assessment No. 59: Report of the Validation of the Updated Test Guideline 407 Repeat Dose 28-Day Oral Toxicity Study in Laboratory Rats. ENV/JM/MONO. vol. 2006. pp. 26. OECD, 2006b. Environmental Health and Safety Publications Series on Testing and Assessment No. 59: Annexes to the Report of the Validation of the Updated Test Guideline 407 Repeat Dose 28-Day Oral Toxicity Study in Laboratory Rats. ENV/JM/ MONO. vol. 2006 26/ANN. SAS Institute Inc, Cary, North Carolina 27511, USA+. Shapiro, S.S., Wilk, M.B., 1965. An analysis of variance test for normality (complete samples). Biometrika 52, 591–611.
Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflicts of interest No conflict of interest is declared by the authors.
5
Contents lists available at ScienceDirect
Regulatory Toxicology and Pharmacology journal homepage: www.elsevier.com/locate/yrtph
Variance of body and organ weights in 28-day studies in mice ∗
T
Heike Antje Marxfeld , Karin Küttler, Martina Dammann, Sibylle Gröters, Bennard van Ravenzwaay Experimental Toxicology and Ecotoxicology, BASF SE, 67056, Ludwigshafen, Germany
A R T I C LE I N FO
A B S T R A C T
Keywords: Organ weight Mouse Immune system Toxicity Endocrine disruptor effects
The OECD guideline 407 outlines the conduct of 28-day studies in rodents to detect systemic toxicity with focus on endocrine and immunotoxic effects. It was validated with the rat as preferred model species. Justification is required for other rodent species, as an increased variability is expected compared to the rat. We investigated the variability of organ weights in the mouse and compared this to data published for the rat in the validation report of test guideline 407. Furthermore, the influence of the immunotoxic model substance cyclophosphamide on spleen and thymus weights in the mouse in immunotoxicity studies (duration 28 days) is reported and discussed, an immunotoxic model substance was not included in the validation report. Historical control data were compiled for mouse studies performed according to OECD 407 and for immunotoxicity studies between 2008 and 2013 at BASF SE. For absolute weights, the coefficient of variation was determined for each study group and compared with the rat. Adrenal glands, ovaries and to lesser degree testes and prostate showed higher coefficients of variation in the mouse (most pronounced in adrenal glands in male animals: rat 5%–17%, CD1 mouse 20%–51%). Effects of cyclophosphamide were best detected measuring the thymus weight.
1. Introduction The OECD guideline 407 gives guidance on the conduct of 28-day studies in rodents for the detection of systemic toxicity with special focus on endocrine and immunotoxic effects. This guideline was validated using the rat as the test species (OECD Guidelines for Testing of Chemicals; Method No. 407, 2008). However, the 407 guideline also gives the experimenter the choice to use other rodents like the mouse. According to this guideline, an increased variability is to be expected using the mouse which is due to technical challenges dissecting these small animals (OECD Guidelines for Testing of Chemicals; Method No. 407, 2008, point 11 “Selection of animal species” page 2–3). The aim of the present investigation was to determine the variability in organ weights in mice to provide guidance for the assessment of potential changes and set them in context with the rat. Furthermore, we report and discuss the influence of the immunotoxic model substance cyclophosphamide on spleen and thymus weights in the mouse in immunotoxicity studies where it serves as a recommended positive control. An immunotoxic model substance was not included in the validation report of the updated test guideline 407. To achieve this, we compiled data from 28-day studies in the mouse performed according to OECD 407 and for spleen and thymus weights
from immunotoxicity studies (according to EPA OPPTS 870.7800) at BASF SE from 2008 to 2013. The variance of terminal body weight and those organ weights mentioned in OECD 407 were investigated. Mandatory weights are those of liver, kidneys, adrenal glands, testes, epididymides, prostate, seminal vesicles with coagulating glands, thymus, spleen, brain, and heart of all animals. Optionally, the weight of ovaries, uterus, and thyroid glands can be determined. Spleen and thymus weights are the only weights required to be taken in immunotoxicity studies. 2. Materials and methods All studies were performed in an AAALAC-approved laboratory at BASF SE, Ludwigshafen, Germany, in accordance with the German Animal Welfare Act and the effective European Council Directive 2010/ 63/EU. All necropsies and organ weight measurements were conducted inhouse by highly experienced (> 15 years) and specialized technicians. Data were recorded in the BASF SE pathology data capture system and exported to SAS for the calculations of coefficients of variation (CVs) in this paper.
∗
Corresponding author. BASF SE, RB/TD - Z470, 67056, Ludwigshafen, Germany. E-mail addresses: [email protected] (H.A. Marxfeld), [email protected] (K. Küttler), [email protected] (M. Dammann), [email protected] (S. Gröters), [email protected] (B. van Ravenzwaay). https://doi.org/10.1016/j.yrtph.2019.104472 Received 20 March 2019; Received in revised form 13 August 2019; Accepted 3 September 2019 Available online 05 September 2019 0273-2300/ © 2019 Elsevier Inc. All rights reserved.
Regulatory Toxicology and Pharmacology 108 (2019) 104472
H.A. Marxfeld, et al.
OECD, 2006a). These CVs were additionally calculated from the data in Annex 5 (OECD, 2006b) to get an idea which variability might be expected in these organs. The number of studies on which this calculation is based varied from 5 to 21 (Calculations are provided in the supplementary data). Different rat strains were used in the validation studies, this information was, however, not detailed in the validation report. The ranges of CVs are defined in the validation report as “low” variability 8–11, mid 12–21 and high > 24 (p. 197, OECD, 2006a). Very low variability (CV < 8) was not included as a class in the validation report. A power analysis was performed to determine which difference is expected to be detected for a given CV. For this analysis a significance level of 5% and a power of 75% was assumed. The power calculation for the two-sided Wilcoxon test was performed using the software NQUERY (Nquery Technical Solutions). For the sake of this power calculation, the underlying distribution was assumed to be normal. The power calculation using the two-sided Dunnett test was based on the following assumptions:
Table 1 Animal data. Supplier
Strain
Age at beginning of study
Abbreviation
Number of control mice/sex
Number of studies
Janvier Janvier CRL CRL
C57BL/j Rj C57BL/j Rj CD-1 CD-1
5–7wks 7–8wks 5–7wks 7–8wks
J5-7 J7-8 CRL5-7 CRL7-8
15 23 30 5
3 4 4 1
2.1. Mice: 28-day studies according to test guideline OECD 407 Historical control data were compiled for 12 different 28-day studies (details are provided in Table 1) performed between 2008 and 2013 according to OECD 407 with mice from Janvier C57BL/j Rj (J) and Charles River CD-1 (CRL) including terminal body weight and the following organs: Adrenal glands, brain, epididymides, heart, kidneys, liver, ovaries, prostate, seminal vesicles (with coagulating glands), spleen, testes, thymus, uterus. Thyroid gland weights, which are optional in the OECD 407 guideline were not measured.
• a design of 3 dose groups and one control having 5, 8 or 10 animals per group, respectively • on the probability of a significant effect of the highest of three dose groups compared to the control
2.2. Mice: immunotoxicity studies (duration of 28 days) The calculation was performed using SAS functions probmc for the determination of the critical value of the Dunnett test and probt for the calculation of the power (SAS Institute Inc). The normal distribution of data of the mouse studies was checked using the Shapiro-Wilk test (Shapiro and Wilk, 1965).
Additionally, data were compiled for 8 immunotoxicity studies performed between 2008 and 2013 according to EPA Health Effects Test Guidelines OPPTS 870.7800 with mice from Janvier C57BL/j Rj (J). 48 control females and 16 control males were investigated as well as cyclophosphamide treated positive control groups with the same number of animals. They were all 5–7 weeks old at the beginning of the studies. Per study, 8 animals/group were used as this is the minimal number permitted in the guideline. The following parameters were included: Number of animals; terminal body weight, mean absolute and relative weights of spleen and thymus; and CVs. The number of animals tested should yield sufficient statistical power to detect a twenty percent change based upon the inter animal variation which may be encountered in these assays.
4. Results 4.1. 28-Day studies According to OECD 407, at the commencement of the study the body weight variation of animals used should be minimal and not exceed ± 20% of the mean weight of each sex. All studies fulfill these requirements.
2.3. Rats
4.2. Organ weights
Organ weight data for the rat was taken from the report of the validation of the updated test guideline 407 and Annex 5 of this document, referred to in the text as “validation report” or “Annex 5”.
A comparison of coefficients of variation for absolute organ weights for male and female mice of different strains and rats is presented in Tables 2 and 3. Scatterplots showing the distribution of coefficients of variations for male and female mice are shown in Figs. 1 and 2. In male mice, adrenal glands (especially in CRL CD1 mice), show a noticeable higher variability in both strains of mice compared to the
2.4. Statistics 2.4.1. Mouse 28-day studies according to guideline 407
Table 2 Comparison of minimum – maximum ranges of CVs of absolute weights in male rats and mice.
To judge the variation of the data the coefficient of variation (CV: standard deviation divided by the mean in percent) was determined for each control study group of mice in 28-day studies. The minimum – maximum ranges of the group means were compared to the rat data. 2.4.2. Mouse immunotoxicity studies The CV was determined for absolute and relative spleen and thymus weights of control and treated animals. Furthermore, the percent deviation of positive control weights from control weights was calculated. 2.4.3. Rat The validation report lists CVs only for absolute weights of selected organs; brain, kidney, heart, spleen and thymus are missing (p. 197,
Organ
Mouse C57BL/Rj CV(%) Based on 7 studies
Mouse CRL CD1 CV(%) Based on 5 studies
Adrenal glands Brain Epididymides Heart Kidney Liver Prostate Seminal vesicles Spleen Testes Thymus
6–35 1–4 5–20 2–18 4–13 4–11 9–31 4–18 3–20 a(10–20) 4–21 4–21 a(9–17)
20–51 2–13 4–10 6–15 3–17 6–11 14–26 9–29 10–21 7–15 18–29
a
2
5–17 2–4 5–13 4–13 3–19 8–15 11–22 8–24 6–25 5–12 8–29
Ranges in Immunotoxicity studies in brackets (2 studies).
Rat CV (%)
Regulatory Toxicology and Pharmacology 108 (2019) 104472
H.A. Marxfeld, et al.
rat. Testes (C57BL/Rj) and prostate (C57BL/Rj) show a slightly higher variability in mice. In female mice, ovaries and thymus showed a noticeable higher variability (especially in CRL CD1 mice), compared to the rat.
Table 3 Comparison of minimum – maximum ranges of CVs of absolute weights in female rats and mice. Organ
Mouse C57BL/Rj CV(%) Based on 7 studies
Mouse CRL CD1 CV(%) Based on 5 studies
Rat CV(%)
Adrenal glands Brain Heart Kidney Liver Ovary Spleen Thymus Uterus
6–17 2–5 4–12 3–9 4–13 6–31 7–17 a(9–15) 5–28 a(10–25) 12–45
6–28 4–6 1–11 6–11 11–19 13–51 17–23 16–40 22–32
7–20 2–4 6–13 2–13 6–17 10–33 6–25 9–27 8–42
a
4.3. Immunotoxicity studies (duration 28 days) Organ weights of spleen and thymus are requested in the EPA guideline OPPTS 870.7800 (EPA, 1998). The following investigations were designed to examine the suitability of this parameter to detect an immunotoxic effect. According to this guideline, at least eight animals should be included in each dose and control group. The number of animals tested should yield sufficient statistical power to detect a twenty percent change based upon the inter animal variation which may be encountered in these assays. A power analysis was performed which based on the two-sided Dunnett test taking a significance level of 5% and a power of 75%. Similarily to the validation document of OECD407, page 188, the highest dose was presumed to have changed versus the control with the low and intermediate doses similar to the control. In addition to the Dunnett test the Wilcoxon test was taken, because after checking with the Shapiro-Wilk test the normal distribution was not always given in our data (data not shown). Based on the assumption of the performed power analyses, the following conclusions can be drawn: If the aim is to detect a 20% difference from the control.
Ranges in Immunotoxicity studies in brackets (6 studies).
• For a group size of 5 the CV should be lower than 7% or no more than 10% using the Wilcoxon or the Dunnett test, respectively. • For a group size of 8 the CV should be lower than or equal to 12% using the Wilcoxon test or Dunnett test. • For a group size of 10 the CV should be lower than or equal to 14% using the Wilcoxon test or Dunnett test.Table 4
In Tables 5 and 6, the CVs of absolute and relative spleen or thymus weights of each study and the deviation of the positive control (cyclophosphamide) from the control are shown. The statistically significant difference with a significance level of 5% is denoted with one star (*) and with a significance level of 1% with two stars (**) using the two-sided Wilcoxon-test. Critical CVs above 12% for the group size of 8 are marked in bold. The ranges of CVs in immunotoxicity studies were within the range for thymus and spleen in 28-day studies according to OECD 407 (shown in brackets in Tables 2 and 3).
Fig. 1. Scatterplot of coefficients of variation in male mice. Abbreviation used in Fig. 1: Adrenal: Adrenal glands, sem ves: seminal vesicles.
4.4. Spleen () The magnitude of the weight change in the spleen induced by cyclophosphamide was mostly lower than 20%, only in two studies the change in absolute weights exceeded 20%, while none of the relative weight changes was higher than 20%. Based on the power calculation, the detection of a 20% change can only be expected if the CVs of both control and positive control group are lower than or equal to 12%. This is only given in 3/8 studies for absolute spleen weights and in 6/8 studies for relative spleen weights. The percent deviation of positive control from control was higher using absolute weights than relative weights. Therefore, a significant effect could be detected in 6/8 studies with absolute weights and in 3/8 studies with relative weights. In study 5, one animal in the treated group showed a markedly higher spleen weight (0.134 g) than both the other animals in this group (range from 0.04 g–0.088 g) as well as the concurrent controls (range from 0.057 g–0.075 g). In the absence of histopathology data, no further conclusions can be drawn.Table 5
Fig. 2. Scatterplot of coefficients of variation in female mice. Abbreviation used in Fig. 2: Adrenal: Adrenal glands.
3
Regulatory Toxicology and Pharmacology 108 (2019) 104472
H.A. Marxfeld, et al.
Table 4 Analysis of power. CV
6% 7% 8% 10% 12% 14% 16% 18% 20% 25% 30%
Percent deviation (n = 5)
Percent deviation (n = 8)
Percent deviation (n = 10)
Wilcoxon
Dunnett
Wilcoxon
Dunnett
Wilcoxon
Dunnett
18% 20.5% 23% 29% 35% 41% 47% 53% 59% 73% 88%
12% 15% 17% 21% 25% 29% 33% 37% 42% 52% 62%
10% 12% 13% 17% 20% 23% 27% 30% 33% 42% 50%
10% 11% 13% 16% 19% 22% 25% 29% 32% 40% 48%
8% 10% 11% 14% 17% 20% 22% 25% 28% 35% 42%
8% 10% 11% 14% 17% 20% 22% 25% 28% 35% 42%
CVs as in the rat with the notable exception of higher CVs for adrenal glands, ovaries, and to a lesser degree testes and prostate, which might implicate that the mouse is less suitable to detect endocrine related effects than the rat with standard group sizes of 5 animals per test group. Also in the rat, only clear antiestrogenic (Ethinyl Estradiol, Tamoxifen, CGS, 18320B) and antiandrogenic (flutamide) substances were successfully detected, while weaker antiestrogenic (Genistein, Nonylphenol) and antiandrogenic (1,1-dichloro-2,2-bis-(p-chlorophenyl)ethylene) substances did not show convincing weight effects according to the validation report (OECD, 2006b) and were only detected on histopathological examination. Although the focus of the immunotoxicity study according to OPPTS 870.7800 is the T-cell dependent antibody response, thymus and spleen weights are also required to be taken and evaluated in this study design. A decrease of weight interpreted as an immunotoxic effects of cyclophosphamide was successfully detected with the evaluation of thymus weights rather than spleen weights. The number of 8 animals per group was sufficient in 7/8 studies. One study did not show an effect of cyclophosphamide, due to high variation in thymus weights in the treated group caused by one animal. Histopathological assessment would be beneficial in the evaluation of questionable cases like this even if it is not required by the guideline for Immunotoxicity studies (EPA, 1998). Weight determination of the spleen seems to be a less valuable parameter in this study type as the effects of the positive control substance were lower (often less than 20% change) than the sensitivity aimed for in this study design. Concluding, in this study design only immunotoxic substances with equal or greater potency than cyclophosphamide can be detected using the thymus weights.
4.5. Thymus () The magnitude of the weight change in the thymus induced by cyclophosphamide was higher than 30% in 6/8 studies, which was detected as a significant effect in these studies. The CVs of both control and positive control group were lower than or equal to 12% in only 2/8 studies for absolute and relative thymus weights, which implies that a smaller effect would not have been detected. A significant effect could be detected in 7/8 studies with absolute weights and in 6/8 studies with relative weights. In study 8, which did not show an effect with cyclophosphamide, one animal in the treated group showed a markedly higher thymus weight (0.154 g) than both the other animals in this group (range from 0.014 g–0.041 g) as well as the concurrent controls (range from 0.028 g–0.051 g). If this animal is excluded from calculations, the group mean weights decrease and the effect is significant. In the absence of histopathology data, no further conclusions can be drawn.Table 6 5. Discussion This investigation was designed to assess the suitability of the mouse compared to the rat in studies conducted in accordance with Test Guideline 407 (OECD, 2008) focusing on organ weight assessment. Furthermore, we report the effects of cyclophosphamide on spleen and thymus weights in the mouse to expand the data in the validation report for test guideline 407 (OECD, 2006b) with an immunotoxic model substance, as only endocrine active substances were reported (OECD, 2006a,b). Generally, the coefficient of variation of organ weights determines the magnitude of an effect which can be detected statistically with a given group size. Most organ weights in control mice showed similar Table 5 CVs of mean absolute and relative spleen weights in immunotoxicity studies. Study number
1
2
3
4
5a
6
7
8
Number of animals
8
8
8
8
8
8
8
8
Sex
m
f
m
f
f
f
f
f
20 19 −22∗
11 16 −14∗
10 11 −25**
10 6 −4
12 45 0
9 7 −14**
9 16 −17**
15 16 −18∗
15 24 −12
12 12 −7
11 10 −17**
8 4 −1
12 45 +10
8 7 −9∗
9 12 −12∗
12 11 −9
Absolute weights Control group CV (%) Positive control group CV (%) Deviation of positive control from control (%) Relative weights Control group CV (%) Positive control group CV (%) Deviation of positive control from control (%)
m = male, f = female, *: p ≤ 0.05, **: p ≤ 0.01. CVs above 12% are marked in bold. a One animal is an outlier. 4
Regulatory Toxicology and Pharmacology 108 (2019) 104472
H.A. Marxfeld, et al.
Table 6 CVs of mean absolute and relative thymus weights and detected effects in immunotoxicity studies. Study number
1
2
3
4
5
6
7
8a
Number of animals
8
8
8
8
8
8
8
8
Sex
m
f
m
f
f
f
f
f
17 41 −39**
11 21 −17*
9 19 −50**
11 12 −38**
10 12 −33**
10 14 −35**
25 20 −34**
19 101 +8
20 55 −29*
13 21 −10
10 17 −44**
11 12 −36**
8 12 −30**
10 14 −32**
24 18 −31**
18 107 +19
Absolute weights Control group CV (%) Positive control group CV (%) Deviation of positive control from control (%) Relative weights Control group CV (%) Positive control group CV (%) Deviation of positive control from control (%)
m = male, f = female, *: p ≤ 0.05, **: p ≤ 0.01. CVs above 12% are marked in bold. a One animal is an outlier cf text.
6. Conclusion
Appendix A. Supplementary data
Adrenal glands and ovaries and to lesser degree testes and prostate showed higher coefficients of variation in the mouse in studies performed according to OECD 407, all other parameters were comparable between the two species. Therefore, the detection of endocrine effects in the mouse might require larger group sizes than in the rat. In immunotoxicity studies (EPA, 1998), the group size of 8 animals per group is only sufficient to detect the effects of immunotoxic substances with equal or greater potency than cyclophosphamide measuring the weight of the thymus.
Supplementary data to this article can be found online at https:// doi.org/10.1016/j.yrtph.2019.104472. References EPA, 1998. U.S. EPA Health effects test guidelines OPPTS 870.7800; immunotoxicity (aug 1998). https://www.regulations.gov/document?D=EPA-HQ-OPPT-2009-01560049, Accessed date: 12 July 2019. Nquery Technical Solutions, Miami, Florida 33176, USA. OECD, 2008. Guidelines for testing of Chemicals; method No. 407: repeated dose 28-day oral toxicity study in rodents; adopted 03 oct 2008. https://www.oecd-ilibrary.org/ environment/test-no-407-repeated-dose-28-day-oral-toxicity-study-in-rodents_ 9789264070684-en, Accessed date: 12 July 2019. OECD, 2006a. Environmental Health and Safety Publications Series on Testing and Assessment No. 59: Report of the Validation of the Updated Test Guideline 407 Repeat Dose 28-Day Oral Toxicity Study in Laboratory Rats. ENV/JM/MONO. vol. 2006. pp. 26. OECD, 2006b. Environmental Health and Safety Publications Series on Testing and Assessment No. 59: Annexes to the Report of the Validation of the Updated Test Guideline 407 Repeat Dose 28-Day Oral Toxicity Study in Laboratory Rats. ENV/JM/ MONO. vol. 2006 26/ANN. SAS Institute Inc, Cary, North Carolina 27511, USA+. Shapiro, S.S., Wilk, M.B., 1965. An analysis of variance test for normality (complete samples). Biometrika 52, 591–611.
Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflicts of interest No conflict of interest is declared by the authors.
5