ERYTHROSINE
First draft prepared by Dr J.C. Larsen,
Institute of Toxicology, National Food Agency of Denmark
1. EXPLANATION
Erythrosine was evaluated for acceptable daily intake for man
(ADI) by the Joint FAO/WHO Expert Committee on Food Additives at its
eighth, thirteenth, eighteenth, twenty-eighth, thirtieth and thirty-
third meetings (Annex 1, references 8, 19, 35, 66, 73 and 83).
Toxicological monographs or monograph addenda were published after
the thirteenth, eighteenth, twenty-eighth, thirtieth and thirty-
third meetings (Annex 1, references 20, 36, 67, 74 and 84). At its
eighteenth meeting the Committee allocated an ADI of 0-2.5 mg/kg
body weight. This ADI was reduced at the twenty-eighth meeting to
0-1.25 mg/kg body weight and made temporary following observations
that erythrosine produced effects on thyroid function in short term
studies in rats and that, in long-term studies, male rats receiving
4% erythrosine in the diet developed thyroid tumours. At the
thirtieth meeting the Committee reduced the temporary ADI to 0-0.6
mg/kg body weight, based on studies on the biochemical effects of
erythrosine on thyroid hormone metabolism and regulation, and
required further data from pharmacokinetic studies relating the
amount of absorption to the amount ingested, which would enable a
correlation to be established between blood/tissue levels of
erythrosine and effects on the thyroid. At the thirty-third meeting
the Committee further reduced the temporary ADI to 0-0.05 mg/kg body
weight, based on a no-observed-effect level with respect to thyroid
function in human beings ingesting 60 mg per person per day
(equivalent to 1 mg per kg body weight per day) for 14 days and
applying a safety factor of 20. The Committee again requested the
pharmacokinetic studies required by the previous Committee.
Since the previous evaluation, additional information have
become available and are summarized and discussed in the following
monograph addendum.
2. BIOLOGICAL DATA
2.1 Biochemical aspects
2.1.1 Absorption, distribution, and excretion
No new information.
2.1.2 Biotransformation
No new information.
2.1.3 Effects on enzymes and other biochemical parameters
See 2.2.6. Special studies on thyroid function.
2.2 Toxicological studies
2.2.1 Acute toxicity studies
No new information.
2.2.2 Short-term studies
See 2.2.6. Special studies on thyroid function.
2.2.3 Long-term/carcinogenicity studies
2.2.3.1 Mouse
No new information.
The long term feeding study reviewed at the thirtieth meeting
(Richter et al., 1981; see Annex 1, reference 74) has now been
published (Borzelleca & Hallagan, 1987).
2.2.3.2 Rat
No new information.
The results of the two long term feeding studies in rats after
in utero exposure to erythrosine that were reviewed at the thirtieth
meeting (Brewer et al., 1981; Brewer et al., 1982; see Annex 1,
reference 74) have now been published (Borzelleca et al., 1987).
In the statistical analyses thyroid follicular cell adenomas and
carcinomas are treated as separate tumour classes. The authors'
conclusion remains that erythrosine at a level of 4% in the diet for
128 weeks induces an increased incidence in thyroid follicular cell
adenomas in male rats (15/69 compared to 1/69 in controls). The
incidence of thyroid follicular cell carcinomas (3/69) was not
statistically significantly different from the control value (2/69).
In the females at the 4% level the incidence of thyroid follicular
cell adenomas (5/68) or carcinomas (0/68) were not different than
the controls (5/66 and 0/66, respectively). In female rats fed 0.1,
0.5, or 1% erythrosine in the diet, a numerical increase in adenomas
was observed (1/68, 3/67 and 5/768, respectively compared to 1/138
control females), but the increases were not statistically
significant. (The incidences of females with carcinomas were 0/68,
0/67 and 1/68 compared to 0/138). In the males at the 0.1, 0.5 and
1.0 levels the incidence of adenomas (0/67, 2/68, and 1/69 compared
to 0/139) and carcinomas (3/67, 1/68, and 3/69 compared to 0/139)
were not considered significantly different.
The microscopic findings in the thyroids from the above-
mentioned studies and the statistics used have been reviewed (FD&C
Red No. 3 Review Panel, 1987; Federal Register, 1990). Slight
discrepancies in the diagnoses of adenomas/carcinomas were reported.
When the combined incidence of adenomas and carcinomas was used in
the statistical evaluation the following results were obtained: As
might be expected an increased incidence of combined adenomas and
carcinomas was seen in the males fed 4% erythrosine in the diet
(18/68 compared to 2/68 in control males). A statistically
significant increase was also found for combined adenomas and
carcinomas in male rats fed 0.1, 0.5 or 1.0% erythrosine for 122
weeks (3+3/64, 7+1/66, 1+3/57, respectively, compared to 0+1/128 in
control male rats). In the female rats a significant increase in
tumour yield was only found in the 1.0% group (5+1/68 compared to
1+0/138 in controls).
2.2.4 Reproduction studies
No new information.
2.2.5 Special studies on genotoxicity
Erythrosine was tested for the induction of point mutations in
the Salmonella typhimurium plate incorporating assay using strains
TA1535, TA1537, TA1538, TA98, and TA100. No mutagenic effects were
observed. In a modified assay using the addition of flavin
mononucleotide to the activation mixture negative results were also
obtained (Cameron et al., 1987).
Erythrosine was non-mutagenic in the Ames test in strains
TA97a, TA98, TA100, TA102, and TA104 to a concentration of 2
mg/plate, with or without metabolic activation with rat liver S9 or
caecal-cell free extracts. The comutagens harman and norharman (+/-
S9) did not affect mutagenicity. A dose dependent suppression in
spontaneous reversion frequencies was observed. Toxicity
(phototoxicity) was observed in the repair-deficient strains (TA97a,
TA98 and TA100) but not in the repair-proficient strains (TA102 and
TA104). Erythrosine was antimutagenic to benzo(a)pyrene and
mitomycin C but not to 4-nitroquinoline-N-oxide and
methylmethanesulfonate (Lakdawalla & Netrawali, 1988a).
Erythrosine did not induce DNA repair in rat hepatocytes in
vitro at concentrations up to 1 mM, or in vivo after an oral
dose of 200 mg/kg body weight (Kronbrust & Barfknecht, 1985).
In the mouse lymphoma assay using L5178Y TK+/- cells
erythrosine was reported positive both with and without the addition
of S9. At concentrations exerting high toxicity the response was
similar to the positive control ethylmethanesulfonate (Cameron et
al., 1987). These results are in contrast to the results obtained
by Lin & Brusick (1986).
Erythrosine was reported to increase the yield of multigene
sporulation minus mutants of Bacillus subtilis excision repair-
proficient strain 168 when incubated in the presence of fluorescent
light. This effect was not seen in the excision repair-deficient
strain her-9 (exc). Erythrosine was highly toxic to both strains
(Lakdawalla & Netrawali, 1988b).
Erythrosine was tested for genotoxicity in V79 Chinese hamster
lung cells. Reduced colony size was seen at 200 µg/ml and more than
90% lethality was seen at 400 µg/ml. Erythrosine was non-mutagenic
to V79 cells at the HGPRT and Na+, K+, ATPase gene loci, and did
not increase the frequency of sister-chromatid exchanges with or
without rat hepatocyte activation. At 300 µg/ml erythrosine
produced an increase in micronucleus frequency in the absence of
hepatocytes. A dose related increase in the mitotic frequency was
observed due to an increase in the number of first mitosis. Thus
increased genotoxicity was observed only at concentrations well in
the range of cytoxicity (Rogers et al., 1988).
A re-evaluation of an earlier published mouse micronucleus test
(Lin & Brusick, 1986) revealed a positive response at the low dose
used (24 mg/kg body weight erythrosine given i.p.), but not at the
two higher doses (80 and 240 mg/kg body weight) (Brusick, 1989).
2.2.6 Special studies on thyroid function
Three groups of 160 male Sprague-Dawley rats were administered
erythrosine at dose levels of 0.0, 0.25 or 4.0% in the diet
(corresponding to 0.0, 147.1 or 2514.3 mg per kg body weight per
day). Physical observations and body weight and food consumption
measurements were performed on all animals pretest and at weekly
intervals during the treatment period. Necropsy was performed with
up to 20 animals per test group at days 0, 3, 7, 10, 14, 21, 30 and
60. Serum was prepared from blood samples taken from the abdominal
aorta at each sacrifice interval and analyzed by radioimmunoassays
for thyrotropin (TSH), thyroxine (T4), 3,5,3'-triiodothyronine
(T3) and 3,3'5'-triiodothyronine (rT3). Thyroid and pituitary
were weighed at each interval and organ/body weight ratios were
calculated. Gross postmortem examinations were conducted on the
thyroid and pituitary only. Three rats receiving 4.0% erythrosine
in the diet died spontaneously during the second week of the study.
The animals receiving 4% erythrosine in the diet lost weight during
the first week of the study and the mean body weights were
significantly lower than control values throughout the study (13% at
week one and 17% at week 8). Food consumption of the animals
receiving 4.0% erythrosine in the diet was significantly lower than
the control value at week one, but after week 2 it was comparable.
This probably reflected a palatability problem during the first two
weeks. The absolute pituitary weights of males receiving 4%
erythrosine were statistically significantly lower than control
values at days 7, 10, 14, 21 and 60. The differences were
considered to reflect the body weight differences between the high-
dose animals and the controls. The absolute thyroid/parathyroid
weights of the rats at the 4% level were generally lower than the
control values, but the differences were slight and may be due to
the body weight differences between these groups. The relative
weights of these organs were significantly greater at day 21;
otherwise relative weights were only slightly greater and not
significant. Thyroid/parathyroid absolute and relative weights of
the rats fed 0.25% erythrosine were significantly lower at day 60,
otherwise they were comparable to controls. Gross postmortem
examinations of thyroid and pituitaries did not show treatment
related changes (Kelly & Daly, 1988).
The analysis of serum hormone levels in these rats reveals the
following: There was a change (slight increase) in serum TSH levels
in the control rats during the 60 day experimental period. The
baseline (day 0) TSH level was significantly lower than the levels
on days 21, 30, and 60. In the 0.25% group serum TSH concentrations
were significantly increased over baseline (day 0) at days 14, 21,
30 and 60. When compared to the TSH levels in control animals a
significant increase was observed at days 21, 30 and 60. In the
4.0% group the TSH levels were significantly increased over the
baseline (day 0) level and the corresponding control levels at all
time points. When compared to the 0.25% group the serum TSH levels
in the high dose group were significantly greater at days 3, 7, 10,
and 14. Serum T4 concentrations were increased over baseline and
control values at days 10 and 14 in the 0.25% group, while in the
4.0% group the T4 concentrations were increased at all time points.
Furthermore, the high dose animals had significantly greater T4
concentrations than the low dose animals at days 7,10, 21, 30 and
60. Serum T3 concentrations in the low dose rats were comparable
to the control values except for a decrease at day 30. In the high
dose rats serum T3 concentrations were significantly lower than
baseline (day 0) and control values at all time points. In
addition, serum T3 concentrations were decreased compared to those
of the low dose animals on days 3, 10, 14, 21, 30, and 60. Serum
rT3 concentrations were increased above baseline (day 0) in the low
dose group at days 7, 10, 14, 21, 30 and 60; and increased above
control values at days 10, 14 and 21. A marked increase in serum
rT3 over controls and low dose animals was seen in the high dose
group at all time points.
The results indicate that the ingestion of a dietary
concentration of 4% erythrosine induces a rapid and sustained
increase in serum TSH, T4, and rT3 and a comparable decrease in
serum T3 concentrations, and that these changes are also induced,
but are less pronounced, after 0.25% in the diet. These findings
are consistent with an inhibition by erythrosine of the deiodination
in the 5'-position of T4 and rT3, resulting in a decreased
production of T3 from T4 and a decreased deiodination of rT3,
respectively (Braverman & DeVito, 1988).
Three groups of 80 male Sprague-Dawley rats were administered
erythrosine at dose levels of 0.0, 0.03, 0.06 and 4.0% in the diet
for a maximum of 60 days (corresponding to 0.0, 17.5, 35.8, and
2671.7 mg/kg body weight per day, respectively). Control animals
(100 males) received standard laboratory diets. Physical
observations, body weight and food consumption measurements were
performed on all animals pretest and at weekly intervals during the
study period. For the determination of baseline data, 20 control
animals were bled for radioimmunoassays of TSH, T4, T3, and rT3
and sacrificed on test day 0, prior to the initiation of dosing.
Additional necropsy intervals were staggered so that on days 7, 21,
30 and 60, an additional 20 animals per group at each interval were
bled for radioimmunoassay samples. Brain, pituitary and thyroid
were weighed and organ/body and organ/brain weight ratios were
calculated for all animals. Gross postmortem examinations were
performed on the thyroids, pituitary and brains of all animals. In
the animals receiving 4% erythrosine in the diet a substantial loss
of body weight and decreased food consumption during week 1 of the
study, probably due to poor palatability of the diet, resulted in
statistically significantly lower body weights of the animals
throughout the study period. The absolute and relative
thyroid/parathyroid weights of the animals receiving 4.0%
erythrosine were increased at days 21 and 30, and at day 60
(relative organ to body weight ratio). The absolute and relative
(organ to brain weight ratio) pituitary weights of animals at the
4.0% level were lower than control values at day 7. In the 0.03%
group absolute and relative thyroid/parathyroid weights were greater
than corresponding control values at day 21 and 30, but comparable
to control values at days 7 and 60. Thus, no consistent and dose
related changes in organ weight, absolute or relative, were found at
the lower doses. Gross postmortem examination of the thyroid,
pituitaries and brain did not reveal any treatment related effects
(Kelly & Daly, 1989).
In the 0.03% and 0.06% groups there were no significant changes
in serum TSH, T4, T3, and rT3, concentrations during the 60 day
treatment period. In the 4.0% group TSH concentrations were
significantly greater than the corresponding control values at days
21, 30, and 60. A 41% increase after 7 days was not statistically
significant compared to the control value. Serum TSH concentrations
were significantly greater than those of the 0.03 group at days 21,
30, and 60, and the 0.06% group at day 30. In the 4.0% group serum
T4 concentrations were slightly elevated above controls during the
treatment period. However, the increase was only statistically
significant on day 30. In the high dose animals serum T3
concentrations were significantly lower than controls at all time
points. Serum rT3 concentrations were markedly increased in the
high dose animals compared to controls or animals fed 0.03% and
0.06% erythrosine at all time points (Braverman & DeVito, 1989).
2.3 Observations in humans
No new information.
3. COMMENTS
The Committee considered additional studies on thyroid hormone
metabolism and regulation in male rats during 60-day feeding trials
with erythrosine. The studies showed a rapid onset in the expected
hormonal changes of a statistically significant rise in serum
levels of thyrotropin, thyroxine (T4), and 3,3,5'-triiodothyronine
(rT3), and a decrease in serum 3,5,3'-triiodothyronine (T3) after
ingestion of 40 mg/kg erythrosine in the diet. A no-observed-effect
level of 0.6 mg/kg erythrosine in the diet corresponding to 30 mg
per kg of body weight per day was obtained. The changes seen in
these studies are consistent with the hypothesis that erythrosine
inhibits the hepatic conversion of circulating T4 to T3, and the
resulting decrease in the concentration of T3 stimulates the serial
release of thyrotropin-releasing hormone from the hypothalamus and
then thyrotropin from the pituitary. The sustained increases in the
levels of thyrotropin produce hyperstimulation of the thyroid, which
may be associated with the tumorigenic effects noted below.
The Committee also reconsidered the carcinogenicity data from
two long-term feeding studies on erythrosine in which an increase in
the incidence of thyroid follicular-cell adenomas in male rats was
demonstrated at a level of 40 mg/kg of erythrosine in the diet.
When thyroid follicular-cell adenomas and carcinomas were combined
in the statistical analysis, significant (but not clearly dose-
related) increases in the incidence of thyroid tumours in male rats
given 1, 5, 10 and 40 mg/kg of erythrosine in the diet were found.
Effects in females were significant only at one dose level. The
Committee agreed that it was appropriate to combine thyroid
follicular-cell adenomas and carcinomas in the statistical analysis,
in view of evidence that adenomas are an earlier stage of carcinomas
in the thyroid.
The Committee reviewed additional data on the mutagenicity of
erythrosine, and, taking into account extensive data from other
mutagenicity studies, concluded that the compound is not genotoxic.
4. EVALUATION
While a no-effect-level could not be determined for the
tumorigenic effect of erythrosine in rats, the Committee considered
that the occurrence of thyroid tumours in rats was most likely
secondary to hormonal effects and concluded that it would be
possible to establish an ADI from the no-effect-level for effects on
thyroid function. In view of the differences in thyroid physiology
between humans and rats the Committee based its evaluation on the
previously reported no-observed-effect level derived from human
data. Therefore the Committee allocated an ADI of 0-0.1 mg/kg of
body weight for erythrosine, based on the no-effect-level at 60 mg
per person per day (equivalent to 1 mg per kg body weight per day)
and a safety factor of 10.
5. REFERENCES
BORZELLECA, J.F., & HALLAGAN, J.B. (1987). Lifetime
toxicity/carcinogenicity study of FD&C Red No. 3 in mice. Fd.
Chem. Toxicol., 25, 735-737.
BORZELLECA, J.F., CAPEN, C.C. & HALLAGAN, J.B. (1987). Lifetime
toxicity/carcinogenicity study of FD&C Red No. 3 (erythrosine) in
rats. Fd. Chem. Toxicol., 25, 723-733.
BRAVERMAN, L.E. & DEVITO, W.J. (1988). Effects of FD&C Red No. 3
(Tetraiodofluorescein) on serum thyroid hormone and TSH
concentrations in male Sprague-Dawley rats; a 60-day study.
Unpublished report dated December 7, 1988. Submitted to WHO by
Certified Color Manufacturers' Association, Washington, D.C., USA.
BRAVERMAN, L.E. & DeVITO, W.J. (1989). Effects of FD&C Red No. 3 on
serum TSH and serum thyroid hormone concentrations in male Sprague-
Dawley rats; Results of a 60-day study (B/d Project No. 88-3378).
Unpublished report dated July 26, 1989. Submitted to WHO by
Certified Color Manufacturers' Association, Washington, D.C., USA.
BRUSICK, D.J. (1989). Addendum to review the genotoxicity of FD&C
Red No. 3. Unpublished report. Submitted to WHO by Certified Color
Manufacturers' Association, Washington, D.C., USA.
CAMERON, T.P., HUGHES, T.J., KIRBY, P.E., FUNG, V.A. & DUNKEL, V.C.
(1987). Mutagenic action of 27 dyes and related chemicals in the
Salmonella/microsome and mouse lymphoma TK+/- assays. Mutat.
Res., 189, 223-261.
FD&C RED NO. 3 REVIEW PANEL (1987). An inquiry into the mechanism
of carcinogenic action of FD&C Red No. 3 and its significance for
risk assessment. Unpublished report. Submitted to WHO by Certified
Color Manufacturers' Association, Washington, D.C., USA.
FEDERAL REGISTER (1990). Termination of provisional listing of FD&C
Red No. 3 for use in cosmetics and externally applied drugs and of
lakes of FD&C Red NO. 3 for all uses. Department of Health and
Human Services, Food and Drug Administration, 21 CFR parts 81 and 82
(Docket Nos. 76C-0044 and 76N-0366), Thursday, February 1.
KELLY, C.M. & DALY, I.W. (1988). A 60-day study to investigate the
effects of FD&C Red No. 3 on thyroid economy in male Sprague-Dawley
rats. Bio/Dynamics Inc. Project No. 88-3378. Report dated December
7, 1988.
KELLY, C.M. & DALY, I.W. (1989). A 60-day study to investigate the
effects of FD&C Red No. 3 on thyroid economy in male Sprague-Dawley
rats. Bio/Dynamics Inc., Project No. 88-3378. Report dated August
2, 1989. Submitted to WHO by Certified Color Manufacturers'
Association, Washington, D.C., USA.
KORNBRUST, D. & BARFKNECHT, T. (1985). Testing of 24 food, drug,
cosmetic, and fabric dyes in the in vitro and the in vivo/in
vitro rat hepatocyte primary culture/DNA repair assays. Environ.
Mutat., 7, 101-120.
LAKDAWALLA, A.A. & NETRAWALI, M.S. (1988a). Mutagenicity,
comutagenicity, and antimutagenicity of erythrosine (FD&C Red 3), a
food dye, in the Ames/ Salmonella assay. Mutat. Res., 204, 131-
139.
LAKDAWALLA, A.A. & NETRAWALI, M.S. (1988b). Erythrosine, a
permitted food dye, is mutagenic in the Bacillus subtilis
multigene sporulation assay. Mutat. Res., 206, 171-176.
LIN, G.H.Y. & BRUSICK, D.J. (1986). Mutagenicity studies on FD&C
Red No. 3. Mutagenesis, 1, 253-259.
ROGERS, C.G., BOYES, B.G., MATULA, T.I, HEROUX-METCALF, C. &
CLAYSON, D.B. (1988). A case report: A multiple end-point approach
to evaluation of cytotoxicity and genotoxicity of erythrosine (FD&C
Red No. 3) in a V79 hepatocyte-mediated mutation assay. Mutat.
Res., 2-5, 415-423.