Simple, sensitive and reliable in vivo assays to evaluate the estrogenic activity of endocrine disruptors

Abstract

Purpose

We compared three in vivo assays, determining changes of body weight, and uterotropic and vaginal cytology assays, for the evaluation of estrogenic activity of an estrogen disrupting compound, Pueraria mirifica (PM), in comparison with 17β‐estradiol (E).

Methods

Female rats were ovariectomized and gavaged with distilled water, 0.01, 0.1, 1, 10 and 20 mg/kg BW/day of E and 100 and 1,000 mg/kg BW/day of PM for 14 days. Body weights were measured weekly, and vaginal epithelium cells were monitored daily. The uterus was dissected at the end of the treatment period, weighed and examined for histomorphometry.

Results

There were a decrease in body weight and an increase in uterine weight, uterine, endometrium and myometrium areas, uterine gland numbers, and percent of cornified cell which were dependent on doses of E and PM treatments.

Conclusions

Of the three assays proposed, although all are reliable and had critical read‐out, measurements of body and uterine weights is likely convenient and simple, but the uterotropic assay needs to kill the animals. Vaginal cytology assay appears most promising for sensitivity and shortening the duration of the assay. Compared to those of E, the estrogenic activity of PM at concentrations of 100 and 1,000 mg/kg BW was in the range of 14 to >20 mg/kg BW.

Introduction

Estrogen disrupting compounds (EDCs) encompass a variety of chemical classes, including drugs, pesticides, industrial by‐products, pollutants, and naturally produced botanical chemicals. Since these compounds are fat soluble, it is likely they accumulate from the environment in the fatty tissue of animals or humans exposed to them, and consequently generate outbursts on health effects. Generally, EDCs occur in a very low dose in the environment and are difficult to detect [1, 2].

There are two models in determining the estrogenic activity, in vitro and in vivo assays. In in vitro assays, various kinds of estrogen responsive cancer derived cell lines, e.g., MCF‐7 human breast cancer cells [3, 4], HeLa cervical cancer cells [5], HepG2 hepatocarcinoma cells [6], CV‐1 monkey kidney cells [7] or yeast cells [8] have widely been used. This assay is based on the estrogen receptor (ER) binding assay, such as the relative binding affinity assays, and an ER transcriptional activation assay, including cell proliferation assays. Although the in vitro assay for estrogenic activity in EDCs is rapid and convenient, and quite applicable for commercial scale assays, it has some disadvantages. For example, using the in vitro MCF‐7 proliferation assay to determine estrogenic activity levels may not represent the estrogenic response in animals including humans due to the lack of bioavailability and biotransformation (absorption, distribution, metabolism, and excretion or ADME) of the chemicals in whole organisms, and the differences in ER types and expression levels between MCF‐7 cells and those in the different tissues of the target organism. Furthermore, the evaluation of the estrogenic activity of EDCs using different estrogen responsive cell lines yielded significantly different results for each chemical between the different cell lines [4, 8].

To avoid these pitfalls, in vivo methods are often used. The most popularly used in vivo assays to assess the estrogenic activity of EDCs are uterotropic and vaginal cornification assays [9, 10, 11, 12, 13, 14, 15, 16]. This is based on the principle that the growth phase of the uterus and vagina in the natural estrous cycle is under the control of estrogens. Uterine growth during the natural estrous cycle is rapid and easily measurable within 2 days. When the endogenous estrogens are not available, either because the animal is immature or because it has been ovariectomized (OVX), then the growth of the uterus becomes sensitive to external sources of estrogen. When the animals are exposed to those chemicals, the uterus of immature or OVX animals can increase in size and weight due to the imbibitions of fluid and cell proliferation stimulated by estrogens. Therefore, the end point of this uterotropic assay is uterine weight, dry or wet weight. The vaginal cytology assay or Allen–Doisy test is used to track changes in the morphology of desquamated vaginal epithelium cells and provides convenient means of assaying changes in estrogen levels. The vaginal epithelium cell is responsive to sex steroids, particularly estrogens, and rising levels of estrogens cause the vaginal epithelium to become “cornified” cells.

Thus, this study aims to search for the inexpensive, simple, sensitive and reliable in vivo assays, e.g., uterotropic and vaginal cytology assays, which can be used to determine the estrogenic activity of EDCs. Moreover, the recent publications indicated that body weights of OVX rats were significantly changed after the EDCs treatment [17, 18, 19, 20], thus, this study also assessed if changes of the body weights can be used as one of the in vivo assays of estrogenic activity. The 17β‐estradiol was used as a standard chemical. The Pueraria mirifica (PM) plant was selected as a representative EDC for this study, because the estrogenic activity of this plant has been widely tested [4, 5, 15, 21, 22, 23].

Materials and methods

Animals

Adult female Wistar rats, 7 weeks old, were obtained from the National Laboratory Animal Center, Mahidol University, Nakhon Pathom, Thailand. They were housed in stainless steel cages with sawdust bedding at five animals/cage in a room with controlled lighting (lights on 06:00–18:00 h) and temperature (25 ± 1 °C) at the Primate Research Unit, Department of Biology, Faculty of Science, Chulalongkorn University. The animals were fed with soybean fee‐rat diet (Pokaphan Animal Feed Co., Ltd., Thailand) and water ad libitum. All animals were acclimatized to the surroundings for 1 week before the onset of the study. The experimental protocol was approved by the Animal Ethical Committee in accordance with the guide for the care and use of laboratory animals prepared by Chulalongkorn University, Protocol Review No. 0823010.

The preparation of 17β‐estradiol

The powder of 17β‐estradiol (E, Fluka, China) was weighed and dissolved in absolute ethanol. After the powder was completely dissolved, the distilled water was added and the solution was allowed standing at room temperature to evaporate the ethanol. The final concentrations of 0.01, 0.1, 1, 10 and 20 mg/kg BW/day were prepared and kept in the dark bottle at 4 °C until used.

The preparation of Pueraria mirifica suspensions

The tuberous roots of PM were purchased from Dr. Sompoch Tubcharoen, Kasetsart University Kampang Sean Campus, Thailand. It was authenticated as the PM by comparing with the voucher specimen numbers BCU010250 and BCU011045, from Professor Kasin Suvatabhandhu Herbarium, Department of Botany, Faculty of Science, Chulalongkorn University. The PM roots used throughout this study were the same lot. The roots were sliced and dried at 70–80 °C, pulverized in a mortar, filtrated through a 100 μm mesh, and the powders were kept in dark bottles. The PM suspension at concentrations of 100 and 1,000 mg/kg BW/day in 1 ml distilled was prepared for this study. These concentrations were completely verified for their estrogenic activity on reproductive organs [10, 11, 12, 13, 24, 25]. However, their estrogenic activity in comparison to the 17β‐estradiol has never been estimated.

Experimental design

At the age of 8 weeks, the rats were bilaterally OVX and divided into three treatment groups; control, E and PM. The experimental schedule was divided into two periods; pre‐treatment and treatment (14 days for each period). During the treatment period, the rats were gavaged daily with 1.0 ml of distilled water (DW) for the control group, 0.01, 0.1, 1, 10 and 20 mg/kg BW/day of E (E‐0.01, E‐0.1, E‐1, E‐10 and E‐20, respectively) for the E group, and 100 and 1,000 mg/kg BW/day of PM (PM‐100 and PM‐1000, respectively) for the PM group. The gavage administration was selected for this study as to mimic the usual route that human intakes EDCs in their daily life. Eight rats were used in each group. The treatment was performed at 09:00–11:00 h. At the end of the 28‐day study period, each group of rats was euthanized under ether, and the uterus was dissected and weighed [17].

Measurement of body weight

The body weights were measured and recorded weekly. The body weight changes during the experimental period were calculated by the following equation

$$[\frac{{\text{body weight at D}}_{\mathrm{x}}-{\text{body weight at D}}_0}{{\text{body weight at D}}_0}]\times 100.$$

Uterotropic assay

After weighing, the uterus was then fixed in 10 % (w/v) neutral buffered formalin solution and manipulated for histomorphometric analysis [17]. The middle region of the fomalin‐fixed uterus was selected and embedded in paraffin and sectioned in 5 μm thickness. All sections were stained with hematoxylin and eosin following standard procedure [26]. The uterine, myometrium and endometrium areas were determined by the Image‐Pro Express program (Media Cybernetics, Inc., USA). The uterine glands were examined and counted under the Olympus compound light microscope using 40× magnification. Five sections of uterus in each rat (in total, 40 sections/group) were randomly selected and determined.

Vaginal cytology assay

Vaginal epithelial cells were monitored daily at 07:00–09:00 h. The vaginal cells observed under the Olympus compound light microscope were classified into three types; leukocyte cell (L), nucleated cell (O) and cornified cell (Co). A total of 100 cells were counted for calculating the percentage of cornified cells (% Co) as described previously [17]. The appearance of cornified cells was used as an indicator of estrogenic activity.

Statistical analysis

All data are presented as mean ± standard error of the mean (SEM). Analysis of variance (ANOVA) was used to determine the differences of means of all the results. The observed significance was then confirmed by using the least significant difference (LSD) test, with the level of significance was set at p <0.05.

Results

Body weight

After ovariectomy, the body weights of rats rapidly and highly significantly increased (p < 0.01) within a week and kept a linear pattern of increase for the control (DW) group (Fig. 1a). However, feeding with high doses of E (E‐10 and E‐20) for only 1 week (D₂₁ of the study period), the body weights were kept constant and tended to be lower than that of the control (DW) group (p = 0.057 and 0.123 for E‐10 and E‐20 groups, respectively). The significant differences of body weights between the control and the E‐10 and E‐20 groups (p < 0.05) could be detected at the second week of treatment (D₂₈ of the study period). On the contrary, the body weights of the lower dose treatment groups of E (E‐0.01, E‐0.1 and E‐1 groups) were not significantly different from the control group throughout the study period.

Figure 1.

Body weights of ovariectomized rats fed with distilled water (DW), 17β‐estradiol at concentrations of 0.01, 0.1, 1, 10 and 20 mg/kg BW/day (E‐0.01, E‐0.1, E‐1, E‐10 and E‐20, respectively) (a) and Pueraria mirifica at concentrations of 100 and 1,000 mg/kg BW/day (PM‐100 and PM‐1000) (b) for 14 days. * and ** indicate p < 0.05 and 0.001 compared to the DW group

Feeding of PM, rats showed the broadly similar patterns of changes of body weights to those of the high doses of E treatment (Fig. 1b). However, the significantly lower body weights from the control group could be detected since the first week of treatment period (D₂₁ of the study period).

Uterotropic assay

Patterns of changes in the uterine parameters (uterine wet weight, uterine, myometrium and endometrium areas, and numbers of uterine glands) of the rats fed with E were similar to those of the body weights, in that the increase was divided into two groups: low and high doses, and no differences between group members (Table 1). There were no significant differences of the uterine wet weights, uterine, myometrium and endometrium areas, and number of uterine glands of the low doses of E treatment (E‐0.01, E‐0.1 and E‐1 groups) when compared to the control group and also between groups. Those values of the rats treated with the higher doses of E (E‐10 and E‐20 groups) were significantly higher than those of the control group (0.001 ≤ p ≤ 0.05), and no significant differences between E‐10 and E‐20 groups.

Table 1.

Uterine wet weight, uterine, myometrium and endometrium areas, and number of uterine glands (mean ± SEM) in ovariectomized rats treated with distilled water (DW group), 0.01, 0.1, 1, 10 and 20 mg/kg BW/day of 17β‐estradiol (E‐0.01, E‐0.1, E‐1, E‐10 and E‐20 groups), and 100 and 1,000 mg/kg BW/day of Pueraria mirifica (PM‐100 and PM‐1000) for 14 days

Treatment	Uterine weight (g)	Uterine area (×105 μm2)	Myometrium area (×105 μm2)	Endometrium area (×105 μm2)	Number of uterine gland
DW	0.120 ± 0.004	13.28 ± 0.88	7.68 ± 0.66	5.30 ± 0.31	21.54 ± 1.99
E‐0.01	0.150 ± 0.016	16.82 ± 2.21	9.70 ± 1.28	6.67 ± 0.83	21.82 ± 1.91
E‐0.1	0.150 ± 0.028	19.02 ± 6.10	10.16 ± 2.90	7.20 ± 1.84	21.91 ± 3.17
E‐1	0.141 ± 0.014	16.15 ± 1.39	9.07 ± 0.67	6.71 ± 0.72	17.43 ± 1.87
E‐10	0.281 ± 0.027c	40.48 ± 2.96c	22.18 ± 1.35c	16.87 ± 1.41c	35.86 ± 4.82b
E‐20	0.297 ± 0.018c	42.41 ± 5.49c	23.67 ± 3.27c	17.37 ± 2.02c	31.22 ± 5.59a
PM‐100	0.433 ± 0.007d	49.35 ± 2.19c	27.71 ± 1.35d	19.89 ± 0.96c	22.26 ± 1.42
PM‐1000	0.451 ± 0.020d	60.24 ± 1.41d	34.16 ± 0.77d	22.77 ± 1.34d	30.06 ± 2.96

a, b, c and d indicate p < 0.05, 0.01, 0.001 and 0.0005 compared to the control group

All of the uterine parameters were significantly increased in rats treated with PM‐100 and PM‐1000 (p < 0.005) when compared to the control group (Table 1), except for the number of uterine glands that was a non‐significant difference in both the PM‐100 (p = 0.867) and PM‐1000 group (p = 0.068).

The uterus morphology (size, thickness and uterine gland) was similar between the E and PM treatments (Fig. 2).

Figure 2.

Uterine histology of ovariectomized rats fed with distilled water (DW), 17β‐estradiol at concentrations of 0.01, 0.1, 1, 10 and 20 mg/kg BW/day (E‐0.01, E‐0.1, E‐1, E‐10 and E‐20, respectively) and Pueraria mirifica at concentrations of 100 and 1,000 mg/kg BW/day (PM‐100 and PM‐1000) for 14 days. L lumen, En endometrium, My myometrium, G uterine gland, respectively

Vaginal cytology assay

After ovariectomy or during the pre‐treatment period, the majority of vaginal epithelium cells were L‐type. The %Co was ranging by 14.6–51.7 % (or 29.15 ± 0.72 %, mean ± SEM) during the pre‐treatment period in the control, five E and two PM treatment groups (Fig. 3a, b). It was kept lower than 30 % during the treatment period for the control group. After 2 days of all E treatments (or D₁₆ of the study period), the %Co was dose‐dependently increased (r ² = 0.880) (Fig. 3a, dotted box). This is different from the decrease in body weight and increase in uterine parameters in that we could not observe changes in the lower dose treatment. Generally, responses of vaginal epithelium cells to the E treatment could be divided into two groups; lesser and greater. The greater response was found in the higher doses of E treatments (E‐10 and E‐20 groups) when the significant increase in cornified cells, compared to the control group, could be observed as early as the second day of the treatments (D₁₆ of the study period) (p < 0.01) and ranging by 67.7–97.0 % from D₁₆ onwards (Fig. 3a). The lesser response was found in the lower doses of E treatment (E‐0.01, E‐0.1 and E‐1 groups) when the significant increase in cornified cells could be observed after 7 days of treatment (or D₂₁ of the study period) (p < 0.05). Thus, the %Co of all E treatment groups was significantly higher than the control group on D₂₁–D₂₈ of the study period (or during 7–14 days of E treatment, p < 0.05 and 0.01). However, the correlations of the increase of %Co during D₂₁–D₂₈ in relation to doses of E were lower than that of the D₁₆ (r ² = 0.530–0.828 for D_21–28 and 0.880 for D₁₆, respectively).

Figure 3.

Percent of cornified cells (%Co) of ovariectomized rats fed with distilled water (DW), 17β‐estradiol at concentrations of 0.01, 0.1, 1, 10 and 20 mg/kg BW/day (E‐0.01, E‐0.1, E‐1, E‐10 and E‐20, respectively) (a) and Pueraria mirifica at concentrations of 100 and 1,000 mg/kg BW/day (PM‐100 and PM‐1000) (b) for 14 days. * and ^† indicate p < 0.05 and 0.01 compared to the DW group. As the significant differences between the DW group and all low doses of E (E‐0.01, E‐0.1 and E‐1) groups or between the DW group and all high doses of E (E‐10 and E‐20) groups are similar, the * and ^† present only in one group. Dotted boxes indicate the second day of treatment when the %Co was first higher than the DW group

Feeding of PM‐100 and PM‐1000 showed a significant increase in %Co as early as the second day of treatment (or D₁₆ of study period), similar to those of the higher doses of E treatment (Fig. 3b), and the %Co of these groups were highly significantly increased throughout the treatment period when compared to the control group (p < 0.001).

Calculation of estrogenic activity of Pueraria mirifica in comparison to the 17β‐estradiol

Standard curves of estrogenic activity of E were drawn in terms of changes of body and uterine weights at the end of the study period (D₂₈ or 14 days after treatment) (Fig. 4a, b) and %Co on D₁₆ of the study period (or two days after feeding) (Fig. 4c). The increase in uterine weight was selected as a representative of uterotropic assay, because (1) changes of all uterine parameters are in the same patterns, and (2) increase in uterine weight was clearly seen and easily detected. The standard curve between E concentrations and % Co on D₁₆ of the study period was selected because it showed the highest correlation value. Comparing to the E, the estrogenic activity of PM‐100 and PM‐1000 in terms of body weight and uterine weight was >20 mg/kg BW, whilst it was 14.32 and 17.20 mg/kg BW for PM‐100 and PM‐1000 based on the increase in %Co.

Figure 4.

Standard curves of estrogenic activity of 17β‐estradiol in terms of body weight changes (a), uterine weights (b) and % Co (c)

Discussion

After ovariectomy for two weeks, all eight groups of rats dramatically increased in body weight. This might be explained by the effect of estrogen deficiency on hypothalamus and adipose cells. Hypothalamus, particularly at the paraventricular and arcuate nuclei and lateral hypothalamic area which are involved in control of food intake, and adipose tissue express both ERα and ERβ [27, 28, 29], and estrogens and EDCs can bind to either ER and exhibit anti‐lipogenic (or anti‐obesity) effects on the hypothalamus and adipose tissue [19, 30, 31, 32]. Thus, the ovariectomy induces an increase in food intake and body weight gain in rats [19, 20, 32, 33]. Gene expression analyses indicated that adipose tissue is the center of action for ER‐β‐selective ligands, and the reduction in body weight after estrogen treatment is likely due to increased energy expenditure via the peroxisome proliferator‐activated receptor γ antagonistic actions [18]. Other proposed mechanisms are that ovariectomy led to increase in plasma levels of leptin and adiponectin and expression level of hypothalamic phosphorylated adenosine monophosphate‐activated protein kinase‐α (pAMPKα) [19, 32], and decrease in gherin expression in fundus of stomach of rats [20], and, thus, replacement of both E and PM could significantly reverse these effects. Although much of the scientific literature has reported about the effects of estrogen deficiency and estrogen replacement on changes of body weight of rats, to the best of our knowledge, no one uses this simple read‐out parameter as one of the in vivo assays to evaluate the estrogenic activity of estrogens or EDCs. From this study, we found that determination of changes of body weights of rats can be a reliable method to detect the estrogenic activity of chemicals, especially for EDCs. It is also inexpensive, straight forward, simple, and the experimenters do not need any skills on animals weighing.

The uterotropic assay is designed to detect estrogenic activity based on the weight‐evidence analysis which is recommended by the US Environmental Protection Agency (EPA) and Organization for Economic Cooperation and Development (OECD) [9]. It is well known that the uterus is one of the main target organs of estrogen and the uterine proliferation is required the estrogen stimulation. Therefore, the uterotropic assay was established to determine the estrogenic activity of synthetic estrogens, xenoestrogens and phytoestrogens [9, 10, 11, 15, 16]. Since both types of ER (ERα and ERβ) can be seen at the uterine glands, and endometrium and myometrium layers of uterus [34], the increases in the number of uterine glands, and uterine, myometrium and endometrium areas after E and PM treatments in this study were consequentially in the same line with the increase in uterine weights. Interestingly, the window of the maximal action of E on increasing the uterine parameters seemed to be at 10 mg/kg BW, and the higher dose (E‐20) did not show the greater response. On the contrary, the increase in uterine parameters stimulated by PM was higher than those of E‐10 and E‐20 treatment groups. This can be explained by the fact that the phytoestrogens in PM have a higher binding affinity on ERβ, whilst the estrogen has a higher binding affinity on ERα [35, 36], and the ratio of ERβ/ERα was high at the uterus [34]. As mentioned above, the uterotropic assay is the preferable method for the estrogenic activity evaluation on EDCs, because uterine weight is also a simple read‐out parameter and weighing the uterus is not difficult. However, compared to the measurement of body weights of rats, we need to kill the animal at the end of the study period for the uterotropic assay, and because of that we could not follow up changes during the time of treatment. Additionally, we need to have experience in uterine histomorphometry. Interestingly, up to two weeks of the low doses of E treatments (E‐0.01, E‐0.1 and E‐1), changes in the body weight of the rats and all uterine parameters could not be detected. This should be due to the fact that an amount of E was likely to be below the threshold for efficacy of the whole body and uterus of rats.

The vaginal cytology assay can be performed in either immature or ovariectomized rodents [37, 38, 39]. In this study we used the ovariectomized rat as a model, because ovariectomy can up‐regulate the expression of estrogen receptors, especially for ERα [40, 41], and it should improve the sensitivity of the assay. Recent publications indicate that the vaginal cytology assay could detect very weak estrogenic activity of puerarin phytoestrogen [17]. Thus, comparing to the measurement of body weight and the uterotropic assay, the vaginal cytology assay is more sensitive which could detect the estrogenic activity of high doses of E (E‐10 and E‐20) within 2 days of treatment (D₁₆ of the study period) and also those of the low doses (E‐0.01, E‐0.1 and E‐1) within a week of treatment. Besides the sensitivity, the vaginal cytology assay can follow up changes of the treatment which could not be done by uterotropic assay; and we do not need to kill the animals. However, observation and classification of vaginal cell‐types under the compound light microscope every day was a tedious work and we need to practice.

Of the three alternatives of the in vivo assays proposed, although all are inexpensive, reliable and critical read‐out, determination of body weights changes is likely convenient and simple, and vaginal cytology assay appears most promising for sensitivity and shortening the duration of the assay. Thus, the researchers should consider what kind of chemicals they want to study, and how much of the concentration of the chemicals, time, complexity, skill and experience is required before they make a decision to use each an in vivo assay.

Phytoestrogens, classified as one of EDCs, are plant compounds with estrogen‐like biological activity which can bind to ERs and express estrogen‐like properties which are similar to endogenous estrogen. Four different families of phenolic compounds produced by plants are considered as phytoestrogens: isoflavonoids, stilbenes, lignans and coumestans [42]. There are more than 300 plant species which contain estrogenic compounds or phytoestrogens, but only a few of these are consumed by animals or humans. Phytoestrogens are widely studied and used as an estrogen replacement therapy in postmenopausal women for the purposes of reducing cardiovascular disease (cardioprotective) [43], maintaining bone mineral density [44, 45], increasing long term and short‐term memory [46] and inducing the low risk of breast, endometrial, ovarian, prostate and colon cancers [47]. However, the plants exhibited variations in the amount of phytoestrogens according to their genetics, location of crop, time of harvest, crop conditions, infection with fungal diseases, and processing in preparation of the raw material [23, 48]. Therefore, the estrogenic activities of phytoestrogens in each plant should be evaluated for the synthetic estrogen and standardized between lots of plants before applying for the further purposes of use in animals or humans.

It is well‐known that the plants in genus Pueraria are largely distributed in Asia, and PM is a Thai endemic herb which is now commercially cultivated and widely used in Thailand, Japan, Korea, China and USA [23]. Its tuberous roots contain various kinds of phytoestrogens, mainly isoflavonoids [49]. Thus, PM is selected as a representative of EDC of this study. Although the estrogenic activity of phytoestrogens in PM has been tested by the in vitro MCF7 proliferation assay [5], and the in vivo vaginal cytology and uterotropic assays [11, 12], the comparison of its activity in equivalent to that of the endogenous estrogen (17β‐estradiol) for the safety of use in animals and humans has never been done. Comparing to the estrogenic activity of E in terms of body and uterine weights and %Co, the estrogenic activity of PM‐100 and PM‐1000 was out of or at the upper range of the standard curve which is 14 to >20 mg/kg BW. Thus, this confirms that PM is a phytoestrogen‐rich plant exhibiting a high estrogenic activity [23] which should be applicable to ameliorate the menopausal symptoms in postmenopausal women. However, the users need to be aware of the suitable and non‐toxic doses before use.

Acknowledgments

The authors thank Dr. Robert Butcher, Faculty of Science, Chulalongkorn University, for proofreading of the manuscript and Dr. Jirarach Kitana for providing the Image‐Pro Express program. This study was supported by the Thai Government Stimulus Package 2 under the Project for establishment of a comprehensive center for innovative food, health products and agriculture (AS613A to S. Malaivijitnond), Ratchadaphisek Somphot Endowment Fund (Grant No. AG001B to S. Malaivijitnond), and by Chulalongkorn University (RES 560530191‐AS, Aging Cluster to S. Malaivijitnond).

Conflict of interest

The authors declare that they have no conflict of interest.