4.6
2.2
| Citation: |
Ahmad Rusdan Handoyo Utomo, Yusnita Yusnita, Siti Maulidya Sari, Octaviani Indrasari Ranakusuma, Sunu Bagaskara, Wening Sari, Yulia Suciati, Anggi Puspa Nur Hidayati, Silviatun Nihayah, Catur Anggono Putro, Neni Nurainy. Buccal DNA global methylation and cognitive performance in stunted children under five years of age[J]. The Journal of Biomedical Research. DOI: 10.7555/JBR.37.20230295
|
Online publication: The manuscript has been published online and can be indexed in databases like PubMed and WoS. It will be available in the next/future issue of the journal, depending on the journal's publication schedule.
The prevalence of stunting in Indonesian children under five years of age is approximately 20%. Chronic maternal malnutrition contributes to the risk of stunting by affecting global DNA methylation. In the present study, we aimed to evaluate the levels of 5-methyl-cytosine (5mC) as a surrogate marker of global DNA methylation in buccal swabs and its potential association with the risk of stunting and cognitive performance. The levels of 5mC were measured using an enzyme-linked immunosorbent assay. The Wechsler Preschool and Primary Scale of Intelligence was used to measure cognitive function. Buccal swab DNA samples and anthropometric data were collected from a total of 231 children aged zero to five years. In this cross-sectional cohort, the prevalence of stunting was 37% in 138 children aged zero to two years and 30% in 93 children aged over two years. The univariable analysis revealed that the levels of 5mC in buccal swab DNA were significantly lower in severely stunted children (median, 2.84; interquartile range [IQR], 2.39–4.62) and children aged less than two years (median, 2.81; IQR, 2.53–4.62) than those in normal children (median, 3.75; IQR, 2.80–4.74; P-value, 0.028) and children aged over four years (median, 4.01; IQR, 3.39–4.87; P-value < 0.001), respectively. We also found that the average cognitive scores tended to be low in boys and stunted children, although the differences were not statistically significant. Furthermore, the levels of 5mC found in buccal swab and mouthwash DNA were not associated with cognitive scores.
Stunting, defined by the World Health Organization as having a height-for-age Z score (HAZ) less than –2 standard deviations[1], is a public health burden, because it reduces quality of life in adulthood and increases the risk of non-communicable diseases[2] and low cognitive performance[3]. A national nutrition survey conducted by the Ministry of Health Directorate of the Republic of Indonesia in 2017 reported a stunting prevalence of approximately 20.1% among children under two years of age[4]. Regional clusters in Indonesia have a higher prevalence of stunting than the national average. For example, in the Pandeglang district in western Java, the prevalence of stunting is about 30%[5].
Chronic malnutrition is an established risk factor for stunting, particularly maternal malnutrition during the first
Stunting may lead to a poor neural development, affecting cognitive performance during school years and later in adulthood[2]. More importantly, persistent early-onset stunting leads to low cognitive scores[3]. Poor cognitive function may result from defects in neural development, as observed in animal models of malnutrition[12].
Efforts to prevent stunting have faced many challenges, including low compliance among participants[13–14], the lack of relevant biomarkers for specific micronutrient deficiencies[15], and non-standardized anthropometric measurements[16], which may affect the accuracy of stunting prevalence estimates. Recent studies have shown that buccal cells and neurons share similar global DNA methylation patterns[17]. Therefore, buccal cells, which are readily accessible in children, may serve as surrogate specimens for cognitive studies.
In the present study, we aimed to investigate the association between 5mC levels in buccal DNA, collected by a minimally invasive method, and stunting or cognitive performance.
The present cross-sectional study enrolled 231 children (aged zero to five years) from local community health facilities in three villages (Kadumaneuh, Kadubelang, and Medong) in Pandeglang district, Banten province, Indonesia. The prevalence of stunting in Pandeglang, especially in Kadumaneuh, has been estimated at approximately 30% according to previous studies[5,18]. The sample size needed for the analysis was not calculated, because no references in the literature had described global DNA methylation in buccal specimens in children with stunting. However, cross-sectional studies using buccal DNA as a source of global DNA methylation in non-stunted children recruited between 48 and 73 children[19–20]. Considering the budget, we conducted a consecutive sampling of children who attended community healthcare facilities in the villages on September 10, 2022, from 9 a.m. to 12 p.m., to collect their buccal cells and anthropometric measurements. There were four healthcare facilities in each village, amounting to a total of 12 recruitment sites. Cognitive measurements were performed on children aged four years or older within one month after recruitment. A diagram of subject recruitment, inclusion criteria, and analyses is shown in Fig. 1.
Buccal DNA samples were collected from all 231 children, consisting of 138 children aged two years or younger and 93 children older than two years. In addition, children older than two years were invited to participate in DNA mouthwash collection. Mouthwash DNA was successfully collected from 80 children. Since stunting may also affect cognitive function, a subset of these 80 children, i.e., children aged four years or older (n = 55), participated in and completed the cognitive evaluation. After a detailed explanation of the study, we obtained the informed consent from their parents or guardians, and the present study was conducted following the Helsinki Declaration. This study was approved by the YARSI Ethics Committee No. 238/KEP-LPUY/VIII/2022.
The height/length of the children was measured directly through physical examination, which was recorded in centimeters according to the standards of the Ministry of Health of the Republic of Indonesia, as previously described[5]. Briefly, the measurement of body length for children under two years old was conducted in a supine position using the GEA WB-C Stature Gauge (Changzhou Wujin Weighing Apparatus Factory, Changzhou, China). For children over two years old, height was measured in a standing position using the GEA SH2A Stature Meter (Changzhou Wujin Weighing Apparatus Factory). Length and height measurements were recorded to the nearest 0.1 cm. All measurements were performed by trained medical students with expertise in anthropometric assessment. The height-for-age Z score (HAZ) for children was calculated using the WHO growth standard. Children with −2 ≤ HAZ ≤ 2 were considered normal, −3 ≤ HAZ < −2 were considered stunted, and HAZ < −3 were considered severely stunted.
Buccal swab DNA was collected from all participants using commercial buccal DNA isolation kits (BioSaliva Kit, Biofarma, Bandung, Indonesia). Parents were asked not to give meals to their children 30 min before sample collection. To collect the buccal cells, the trained medical students rubbed a sterile swab head against the inside cheeks of the child for 20 rounds on each cheek. In addition, children aged four to five years were also asked to take about 2.5 mL of mouthwash liquid provided by the kit solution and spit it into the preservative solution[21]. The samples were stored in a sterile tube at room temperature and transported to the YARSI University molecular testing laboratory. DNA was extracted using an automated DNA extraction kit (Maxwell® RSC Blood DNA Kit, Cat. AS1400, Promega, Madison, WI, USA) with magnetic bead technology (Maxwell® RSC Instrument, Promega). The DNA concentration was quantified using spectrophotometry (Tecan Infinite M200 Pro, Switzerland).
Global methylation of child buccal swab DNA samples was analyzed using an indirect ELISA-based commercial kit (5-mC DNA ELISA Kit; Cat. D5325; Zymo Research, Irvine, CA, USA), as previously described. Briefly, 100 ng of input DNA was bound to the wells, and methylated DNA was detected using antibodies to 5mC and then quantified colorimetrically by reading absorbance at 405 nm using an ELISA plate reader (Tecan Infinite M200 Pro, Switzerland). The standard curve was generated using the absorbance values of seven standards made by mixing positive and negative controls provided by the kit. The final standard methylation concentrations were 0%, 5%, 10%, 25%, 50%, 75%, and 100%, respectively. The 5mC levels for unknown DNA samples were calculated as follows: %5mC = e(Absorbance−Intercept)/Slope.
Child cognitive function was evaluated using the Wechsler Preschool and Primary Scale of Intelligence (WPPSI), which is suitable for children aged four to 7.3 years as described previously[22]. The trained psychologist assistants conducted the assessment, which was closely supervised by a licensed psychologist. The assessment was administered at each child's home to minimize unfamiliar surroundings that may distract children's attention during the test. The administration of the test took 60–70 min. The test was paused when the child looked tired or bored. During the break, the assistant offered the child a snack or drink and talked about a toy or game the child favored. The break usually lasted five minutes before the assistant asked the participant to continue the test. When the child refused to continue, the assistant stopped the assessment and returned the next day.
Cognitive function was measured using the Indonesian version of the WPPSI published by the Faculty of Psychology, Universitas Indonesia[22]. The WPPSI consists of two scales: the Verbal Scale and the Performance Scale. The Verbal Scale was measured by five subtests: information, vocabulary, comprehension, similarities, and arithmetic. The Performance Scale was measured by another five subtests: animal house, picture completion, mazes, geometric design, and block design. The raw score of each subtest was standardized according to age group before summing the scores for each scale and converting them into verbal intelligence quotient (VIQ) and performance intelligence quotient (PIQ). The standard scores of the Verbal and Performance Scales were summed and converted into full-scale intelligence quotient (FIQ)[22].
We used the Chi-square test to compare the proportions of normal, stunted, and severely stunted children in different independent variables, such as sex, age, and villages. Because the values of 5mC levels were not normally distributed, we compared the median of 5mC levels in multiple groups of HAZ status (i.e., normal, stunted, and severely stunted children) and age (zero to 24 months, 24 to 48 months, and older than 48 months) using the Kruskal-Wallis rank test. When the Kruskal-Wallis test showed statistical significance, post hoc Dunn-Bonferroni tests were calculated to reveal differences among different subgroups (HAZ status and age groups). The Mann-Whitney test was used to compare the median values of 5mC levels between boys and girls. Binary logistic regression was used to test numerical data (5mC levels) as a predictor of binary stunting outcomes in different age groups. To determine the relationship between global DNA methylation and cognitive function, we used a linear regression model as described previously. A two-sided P-value < 0.05 was considered statistically significant. All statistical analyses were performed using Statplus for Mac statistical software (AnalystSoft Inc., Walnut, CA, USA), GraphPad Prism version 10.0.0 for Windows (GraphPad Software, Boston, MA, USA), and DATAtab: Online Statistics Calculator (DATAtab e.U. Graz, Austria).
Children ranged in age from two to 73 months, with a mean age of 28.4 months (SD = 20.1 months) and a median age of 19 months (IQR, 12–49 months). As shown in Table 1, the prevalence of stunting (−3 ≤ HAZ < −2) and severe stunting (HAZ < −3) in this cohort was 21% and 13% out of 231 subjects, respectively. Moreover, the prevalence of stunting was 37% in children aged zero to two years and 30% in children older than two years. Approximately 57% of the subjects were male, and 60% were aged zero to two years. Of the three villages, Kadumaneuh had the highest prevalence of severe stunting. The average concentration of buccal DNA and the corresponding 5mC levels were 28.3 ng/μL and 4.06%, respectively.
| Variables | n (%) |
| Age (months) | |
| 0–24 | 138 (60) |
| 24–48 | 14 (6) |
| ≥ 48 | 79 (34) |
| Sex | |
| Male | 131 (57) |
| Female | 100 (43) |
| Outcomes | |
| Normal (−2 ≤ HAZ ≤ 2) | 152 (66) |
| Stunting (−3 ≤ HAZ < −2) | 49 (21) |
| Severe stunting (HAZ < −3) | 30 (13) |
| Villages (normal, stunting, and severe stunting) | |
| Kadubelang (N = 73) | 47 (65), 17 (23), 9 (12) |
| Kadumaneuh (N = 72) | 43 (60), 14 (19), 15 (21) |
| Medong (N = 86) | 61 (71), 19 (22), 6 (7) |
| Abbreviation: HAZ, height-for-age Z score. | |
DownLoad:
CSV
As shown in Table 2, stunting outcomes were not significantly associated with age, sex, or the villages in which the children lived (P-values > 0.05).
| Variables | Normal [n (%)] |
Stunting [n (%)] |
Severe stunting [n (%)] |
P-value |
| Age (months) | ||||
| 0–24 | 87 (57) | 27 (55) | 24 (80) | 0.204 |
| 24–48 | 9 (6) | 4 (8) | 1 (33) | |
| ≥ 48 | 56 (37) | 18 (37) | 5 (17) | |
| Sex | ||||
| Male | 90 (59) | 29 (59) | 12 (40) | 0.140 |
| Female | 62 (41) | 20 (41) | 18 (60) | |
| Villages | ||||
| Kadubelang | 47 (31) | 16 (32) | 9 (30) | 0.776 |
| Kadumaneuh | 44 (29) | 14 (29) | 15 (50) | |
| Medong | 61 (40) | 19 (39) | 6 (20) | |
| The height-for-age Z score (HAZ) for children was calculated using the WHO growth standard. Children with −2 ≤ HAZ ≤ 2 were considered normal, −3 ≤ HAZ < −2 were considered stunted, and HAZ < −3 were considered severely stunted. P-value was calculated by the Chi-square analysis. | ||||
DownLoad:
CSV
As shown in Table 3, the levels of 5mC were significantly associated with age, sex, and HAZ status. Specifically, normal children had significantly higher 5mC levels than severely stunted children, and the youngest age group (zero to 24 months) had the lowest levels of 5mC, compared with older age groups.
| Variables | 5mC levels in % [median (IQR)] | P |
| Age (months) | <0.001a | |
| 0–24 | 2.81 (2.53–4.62) | |
| 24–48 | 4.09 (3.76–4.56) | |
| > 48 | 4.01 (3.39–4.87) | |
| Post hoc analysis | ||
| 0–24 months vs. 24–48 months | 0.038b | |
| 0–24 months vs. > 48 months | < 0.001b | |
| 24–48 months vs. > 48 months | 1b | |
| Sex | 0.0187c | |
| Male | 3.83 (2.79–4.83) | |
| Female | 3.25 (2.60–4.48) | |
| Outcomes | 0.031a | |
| Normal (−2 ≤ HAZ ≤ 2) | 3.75 (2.80–4.74) | |
| Stunting (−3 ≤ HAZ < −2) | 3.64 (2.70–4.57) | |
| Severe stunting (HAZ < −3) | 2.84 (2.39–4.62) | |
| Post hoc analysis | ||
| Normal vs. stunting | 1b | |
| Normal vs. severe stunting | 0.028b | |
| Stunting vs. severe stunting | 0.132b | |
| aThe Kruskal-Wallis test was employed to compare the median of multiple groups.bDunn-Bonferroni was employed as a post hoc analysis.cThe Mann-Whitney test was employed to compare the median of the two groups.Bold font indicates P-values < 0.05.Abbreviations: HAZ, height-for-age Z score; IQR, interquartile range. | ||
DownLoad:
CSV
The distribution of 5mC levels by age and HAZ status is shown in Fig. 2. Because 5mC levels were not normally distributed, the Spearman correlation analysis was used to examine the correlations between 5mC levels and age in different HAZ statuses. As shown in Table 4, a significant correlation was found between age and 5mC levels in stunted children (−3 ≤ HAZ < −2), but not in normal (−2 ≤ HAZ ≤ 2) or severely stunted (HAZ < −3) children.
| Outcomes | n | r | 95% CI | P |
| Normal (−2 ≤ HAZ ≤ 2) | 152 | − |
||
| Stunting (−3 ≤ HAZ < −2) | 49 | |||
| Severe stunting (HAZ < −3) | 30 | − |
||
| Spearman correlation analysis was performed. Bold font indicates P-values < 0.05. Abbreviations: HAZ, height-for-age Z score; CI, confidence interval. | ||||
DownLoad:
CSV
Although 5mC levels were lower in severely stunted children, we tested whether this association remained after adjusting for age and sex in this cohort (n = 231). The results of multivariate logistic regression analysis showed that the outcome of stunting (including both stunted and severely stunted children with an HAZ < −2) was not significantly associated with age (coefficient, 0.004; 95% confidence interval [CI], −0.010–0.018), 5mC levels (coefficient, 0.099; 95% CI, −0.039–0.278), and sex (coefficient, 0.252; 95% CI, −0.304–0.808).
Mouthwash DNA was another source of non-invasive DNA collection, which did not require swabbing tools. Out of 93 children aged more than two years, mouthwash DNA was successfully extracted from 80 children. Therefore, there were 80 children with paired DNA collected from buccal swabs and mouthwash. The 5mC levels in swab DNA (median, 4.01; IQR, 3.40–4.70) were significantly higher than those in mouthwash DNA (median, 3.69; IQR, 3.27–4.20) (P = 0.018 by Wilcoxon matched-pairs signed rank test). There was no significant difference in 5mC levels in mouthwash DNA among children of different sex and HAZ status (Table 5).
| Variables | n | 5mC levels [median (IQR)] |
P |
| Sex | 0.59a | ||
| Male | 47 | 3.69 (3.26–4.03) | |
| Female | 33 | 3.68 (3.38–4.23) | |
| HAZ | 0.63b | ||
| −2 ≤ HAZ ≤ 2 | 51 | 3.68 (3.25–4.27) | |
| −3 ≤ HAZ < −2 | 18 | 3.65 (3.39–4.09) | |
| HAZ < −3 | 11 | 3.96 (3.34–4.20) | |
| aMann-Whitney was employed to compare median values of 5mC levels in males and females.bKruskal-Wallis was employed to compare median values of 5mC levels in three different HAZ statuses.Abbreviations: HAZ, height-for-age Z score; IQR, interquartile range. | |||
DownLoad:
CSV
Of the 80 children, 55 aged four years and over completed the cognitive assessment. For the three cognitive scores (i.e., PIQ, VIQ, and FIQ), the differences in performance attributed to sex and HAZ were not significant (Table 6). Of note, children who lived in Kadumaneuh village had significantly lower VIQ and FIQ scores than those living in the other two villages.
| Variables | n | PIQ [median (IQR)] |
P | VIQ [median (IQR)] |
FIQ [median (IQR)] |
||
| Sex | 0.983a | 0.936a | 0.996a | ||||
| Male | 32 | 99.5 (86–111) | 100 (95–116) | 101 (94–110) | |||
| Female | 23 | 100 (92–108) | 104 (89–116) | 101 (90–113) | |||
| HAZ | 0.797a | 0.573a | 0.825a | ||||
| −2 ≤ HAZ ≤ 2 | 38 | 100 (85–108) | 101 (94–117) | 101 (94–111) | |||
| HAZ < −2 | 17 | 99 (92–113) | 97 (89–115) | 99 (91–112) | |||
| Villages | 0.293b | 0.0065b | 0.019b | ||||
| Medong | 22 | 102 (97–111) | 109 (99–121) | 111 (97–115) | |||
| Kadumaneuh | 18 | 97 (80–105) | 93 (85–100) | 96 (79–103) | |||
| Kadubelang | 15 | 100 (92–108) | 105 (98–113) | 102 (99–106) | |||
| aP-value by the Mann-Whitney test.bP-value by the Kruskal-Wallis ANOVA.Abbreviations: HAZ, height-for-age Z score; PIQ, performance intelligence quotient; VIQ, verbal intelligence quotient; FIQ, full-scale intelligence quotient; IQR, interquartile range. | |||||||
DownLoad:
CSV
To assess different variables that might influence performance IQ scores, we performed multivariable regression analysis, and the results are shown in Table 7. None of the variables had a significant association with cognitive performance.
| Variables | PIQ | VIQ | FIQ | ||||||||
| Coefficient (95% CI) | P | R2 | Coefficient (95% CI) | P | R2 | Coefficient (95% CI) | P | R2 | |||
| Sex (Female) | −0.0625 (−9.14–9.01) | 0.98 | 0.10 | 3.05 (−6.20–12.31) | 0.50 | 0.10 | 2.18 (−6.64–11.01) | 0.62 | 0.10 | ||
| Villages (Kadumaneuh) | −6.96 (−18.16–4.23) | 0.21 | 0.36 | −11.39 (−22.81–0.03) | 0.050 | 0.36 | −9.81 (−20.71–1.08) | 0.07 | 0.36 | ||
| Stunting outcomes (stunted) | −6.88 (−16.92–3.15) | 0.17 | 0.16 | −6.21 (−16.46–4.02) | 0.23 | 0.16 | −7.11 (−16.88–2.64) | 0.14 | 0.16 | ||
| 5mC levels | |||||||||||
| Buccal swab DNAa | 1.33 (−2.02–4.68) | 0.42 | 0.24 | −0.21 (−3.63–3.21) | 0.90 | 0.24 | 0.67 (−2.59–3.93) | 0.68 | 0.24 | ||
| Mouthwash DNAb | 9.34 (−14.45–33.13) | 0.43 | 0.07 | 7.18 (−17.08–31.46) | 0.55 | 0.075 | 9.47 (−13.66–32.61) | 0.41 | 0.07 | ||
| aValues of 5mC levels in buccal swab DNA were left untransformed.bValues of 5mC levels in mouthwash DNA were natural logarithm transformed to obtain a normal distribution.Abbreviations: PIQ, performance intelligence quotient; VIQ, verbal intelligence quotient; FIQ, full-scale intelligence quotient; CI, confidence interval. | |||||||||||
DownLoad:
CSV
In the present study, we successfully collected a sufficiently high concentration of DNA from buccal swabs and mouthwash for suitable epigenetic analysis. Both methods were non-invasive for children. We demonstrated that 5mC levels were higher in boys and older children than in girls and younger children, consistent with previous studies[23]. The exact mechanism underlying sex-specific global DNA methylation remains unclear but may involve the epigenetic regulation of steroid hormones[24].
Iqbal et al[25] investigated global DNA methylation in children aged two to three years using blood DNA samples and the ELISA method, reporting slightly higher 5mC levels in stunted children than in normal children, albeit the difference was not statistically significant. In the present study, where most children were under two years old, the 5mC levels in buccal DNA were significantly lower in severely stunted children than in normal children. This discrepancy may be influenced by the natural increase in global methylation during early childhood[26].
Although the present study does not support the use of 5mC levels as a predictor of stunting, previous research demonstrates that the rapid increase in genome methylation during early normal childhood[26] may be impaired in children with growth problems[27]. Therefore, global methylation levels may vary with age[26]. Older children with stunting may also exhibit higher levels of methylated DNA, because of increased susceptible to bacterial infections and poor oral health, such as caries[28]. Bacterial DNA also contains 5mC, which complicates analysis using tools like ELISA as these methods cannot distinguish between bacterial and human DNA. It is estimated that 0.8% of all cytosines in E. coli DNA are methylated, compared with about 4% of all cytosine nucleotides in the human genome[29].
In the present study, we showed that the 5mC levels in buccal swab DNA were higher in boys, consistent with findings from other studies using peripheral blood DNA[25] and buccal swab DNA. However, this pattern was not observed in mouthwash DNA. The discrepancy between buccal swab and mouthwash DNA may be due to bacterial DNA methylation, which the ELISA method cannot distinguish from human DNA[30].
Stunting is a public health concern because of its potential impact on cognitive function, with long-term consequences in adulthood[31–32]. However, in the present study, the differences in cognitive function between normal and stunted children were not statistically significant. Notably, children living in the village of Kadumaneuh had the lowest cognition scores, although multivariable regression analysis did not confirm this difference as statistically significant. Furthermore, mean scores (> 90) of PIQ, VIQ, and FIQ in stunted children were still considered normal within the general population. These findings do not support previous studies showing that stunted children have lower IQ[33–34]. Koshy et al[35] also found that stunted children had significantly lower VIQ than non-stunted children at two, five, and nine years of age. However, these studies did not use the same measures. Aurora et al[34] used the Colored Progressive Matrices (CPM), while Koshy et al[35] used the Malin's Intelligence Scale of Indian Children (MISIC), an adaptation of the Wechsler Intelligence Scale for Children (WISC), both assessing children aged nine years. Venables et al[33] examined the intelligence of three-year-old children using the WPPSI subtest and found direct correlations with VIQ and PIQ. Meanwhile, Mohd Nasir et al[36] used the CPM to measure general cognitive function in children aged four- to six-year-old and found that height-for-age contributed to cognitive performance only after controlling for sociodemographic background and parental nutrition knowledge.
A meta-analysis showed that cognitive ability was highly dependent on social education and a stimulating environment[37]. Although nutritional supplements are important, the risk of low cognitive function may be minimized by providing children with social support. However, a prospective study would be ideal to follow cognitive performance in school at a later age in this cohort and assess the association with global DNA methylation or specific genes affected by methylation using the NextGen sequencing.
Using the linear regression model, one study showed a decreasing tendency (though insignificant) of all cognition scores for each increasing unit of 5mC levels in children aged four years[38]. Interestingly, we did not observed this decreasing trend for any of the cognitive parameters. The cause of this discrepancy is currently being investigated, including factors such as the small size of our cohort, the specific cognitive measurement tools used, and the exposure of our cohort to a stimulating environment.
A limitation of the present study was that we did not assess dietary intake in our cohort. Additionally, we did not analyze breastfeeding history, which may influence DNA methylation[39]. Stunting primarily results from chronic malnutrition during the first
Regarding cognitive assessment in stunted children, challenges arise when assessing preschool children, whose attention spans are often limited. Although test administrators were trained to build rapport and carefully administer the tasks, some obstacles arose during the assessment. The high dropout rate of the participants reflected these challenges, with children either refusing to continue the test on the following day or the caregivers tooking them to family events preventing completing the second day of assessment. Despite these challenges, the assessment of cognitive function is important for early identification of cognitive delays so that appropriate interventions may be implemented at an earlier stage.
In conclusion, we demonstrated that both buccal and mouthwash samples were suitable materials for a follow-up study to clarify the precise molecular mechanisms of epigenetic alterations affecting cognitive function in stunted children. These stunted children in this cohort may be followed up for further cognitive studies when they reach primary school age.
The authors would like to thank the participation of mothers and the support of our local community healthcare cadres, research assistants, and YARSI student volunteers in Indonesia. Any opinions, findings, conclusions, or recommendations expressed here are those of the authors alone.
The present study was supported by the Ministry of Education, Culture, Research, and Technology of the Republic of Indonesia through the Kedaireka 2022 Matching Fund Programme (Grant No. 282/E1/KS.060.02/2022).
CLC number: R179, Document code: A
The authors reported no conflict of interests.
| [1] |
Leroy JL, Frongillo EA. Perspective: what does stunting really mean? A critical review of the evidence[J]. Adv Nutr, 2019, 10(2): 196–204. doi: 10.1093/advances/nmy101
|
| [2] |
De Sanctis V, Soliman A, Alaaraj N, et al. Early and long-term consequences of nutritional stunting: from childhood to adulthood[J]. Acta Biomed, 2021, 92(1): e2021168. doi: 10.23750/abm.v92i1.11346
|
| [3] |
Alam MA, Richard SA, Fahim SM, et al. Impact of early-onset persistent stunting on cognitive development at 5 years of age: results from a multi-country cohort study[J]. PLoS One, 2020, 15(1): e0227839. doi: 10.1371/journal.pone.0227839
|
| [4] |
Laksono AD, Wulandari RD, Amaliah N, et al. Stunting among children under two years in Indonesia: does maternal education matter?[J]. PLoS One, 2022, 17(7): e0271509. doi: 10.1371/journal.pone.0271509
|
| [5] |
Wuryanti S, Marsiati H. Relationship between maternal education and knowledge on coverage of exclusive breastfeeding in ten stunting locus villages in Pandeglang district, Banten[J]. Indones J Med Sci Public Health, 2021, 2(1): 40–46. doi: 10.11594/ijmp.02.01.05
|
| [6] |
Black RE, Victora CG, Walker SP, et al. Maternal and child undernutrition and overweight in low-income and middle-income countries[J]. Lancet, 2013, 382(9890): 427–451. doi: 10.1016/S0140-6736(13)60937-X
|
| [7] |
Messerschmidt DM, Knowles BB, Solter D. DNA methylation dynamics during epigenetic reprogramming in the germline and preimplantation embryos[J]. Genes Dev, 2014, 28(8): 812–828. doi: 10.1101/gad.234294.113
|
| [8] |
Geraghty AA, Lindsay KL, Alberdi G, et al. Nutrition during pregnancy impacts offspring's epigenetic status-evidence from human and animal studies[J]. Nutr Metab Insights, 2016, 8(S1): 41–47. https://pubmed.ncbi.nlm.nih.gov/26917970/
|
| [9] |
Pauwels S, Ghosh M, Duca RC, et al. Maternal intake of methyl-group donors affects DNA methylation of metabolic genes in infants[J]. Clin Epigenetics, 2017, 9: 16. doi: 10.1186/s13148-017-0321-y
|
| [10] |
Dominguez-Salas P, Moore SE, Baker MS, et al. Maternal nutrition at conception modulates DNA methylation of human metastable epialleles[J]. Nat Commun, 2014, 5: 3746. doi: 10.1038/ncomms4746
|
| [11] |
Schulze KV, Swaminathan S, Howell S, et al. Edematous severe acute malnutrition is characterized by hypomethylation of DNA[J]. Nat Commun, 2019, 10(1): 5791. doi: 10.1038/s41467-019-13433-6
|
| [12] |
Laus MF, Vales LDMF, Costa TMB, et al. Early postnatal protein-calorie malnutrition and cognition: a review of human and animal studies[J]. Int J Environ Res Public Health, 2011, 8(2): 590–612. doi: 10.3390/ijerph8020590
|
| [13] |
Saffari A, Shrestha S, Issarapu P, et al. Effect of maternal preconceptional and pregnancy micronutrient interventions on children’s DNA methylation: findings from the EMPHASIS study[J]. Am J Clin Nutr, 2020, 112(4): 1099–1113. doi: 10.1093/ajcn/nqaa193
|
| [14] |
Hossain M, Choudhury N, Abdullah KAB, et al. Evidence-based approaches to childhood stunting in low and middle income countries: a systematic review[J]. Arch Dis Child, 2017, 102(10): 903–909. doi: 10.1136/archdischild-2016-311050
|
| [15] |
Chandak GR, Silver MJ, Saffari A, et al. Protocol for the EMPHASIS study; epigenetic mechanisms linking maternal pre-conceptional nutrition and children’s health in India and Sub-Saharan Africa[J]. BMC Nutr, 2017, 3(1): 81. doi: 10.1186/s40795-017-0200-0
|
| [16] |
Chanyarungrojn PA, Lelijveld N, Crampin A, et al. Tools for assessing child and adolescent stunting: lookup tables, growth charts and a novel appropriate-technology “MEIRU” wallchart - a diagnostic accuracy study[J]. PLoS Global Public Health, 2023, 3(7): e0001592. doi: 10.1371/journal.pgph.0001592
|
| [17] |
Nishitani S, Isozaki M, Yao A, et al. Cross-tissue correlations of genome-wide DNA methylation in Japanese live human brain and blood, saliva, and buccal epithelial tissues[J]. Transl Psychiatry, 2023, 13(1): 72. doi: 10.1038/s41398-023-02370-0
|
| [18] |
Mardhiyah D, Desiar Y, Sary F, et al. Maternal factors and stunting among children age 0-24 months in Banten province Indonesia[J]. Malaysian J Med Health Sci, 2023, 16(Suppl 10): 55–58. https://medic.upm.edu.my/upload/dokumen/2020111315140310_MJMHS_0657.pdf
|
| [19] |
Ducreux B, Frappier J, Bruno C, et al. Genome-wide analysis of DNA methylation in buccal cells of children conceived through IVF and ICSI[J]. Genes, 2021, 12(12): 1912. doi: 10.3390/genes12121912
|
| [20] |
Henriksson P, Lentini A, Altmäe S, et al. DNA methylation in infants with low and high body fatness[J]. BMC Genomics, 2020, 21(1): 769. doi: 10.1186/s12864-020-07169-7
|
| [21] |
Utama R, Hapsari R, Puspitasari I, et al. Self-collected gargle specimen as a patient-friendly sample collection method for COVID-19 diagnosis in a population context[J]. Sci Rep, 2022, 12(1): 3706. doi: 10.1038/s41598-022-07690-7
|
| [22] |
Tanjung ICD, Sekartini R, Gunardi H, et al. Stimulation and cognitive function in short-stature preschoolers[J]. Paediatr Indonesiana, 2021, 61(2): 74–81. doi: 10.14238/pi61.2.2021.74-81
|
| [23] |
Taylor RM, Smith R, Collins CE, et al. Methyl-donor and cofactor nutrient intakes in the first 2–3 years and global DNA methylation at age 4: a prospective cohort study[J]. Nutrients, 2018, 10(3): 273. doi: 10.3390/nu10030273
|
| [24] |
Arathimos R, Sharp GC, Granell R, et al. Associations of sex hormone-binding globulin and testosterone with genome-wide DNA methylation[J]. BMC Genet, 2018, 19(1): 113. doi: 10.1186/s12863-018-0703-y
|
| [25] |
Iqbal MS, Rahman S, Haque MA, et al. Lower intakes of protein, carbohydrate, and energy are associated with increased global DNA methylation in 2- to 3-year-old urban slum children in Bangladesh[J]. Matern Child Nutr, 2019, 15(3): e12815. doi: 10.1111/mcn.12815
|
| [26] |
McEwen LM, O’Donnell KJ, McGill MG, et al. The PedBE clock accurately estimates DNA methylation age in pediatric buccal cells[J]. Proc Natl Acad Sci U S A, 2020, 117(38): 23329–23335. doi: 10.1073/pnas.1820843116
|
| [27] |
Mayer F, Becker J, Reinauer C, et al. Altered DNA methylation at age-associated CpG sites in children with growth disorders: impact on age estimation?[J]. Int J Legal Med, 2022, 136(4): 987–996. doi: 10.1007/s00414-022-02826-w
|
| [28] |
Sadida ZJ, Indriyanti R, Setiawan AS. Does growth stunting correlate with oral health in children?: A systematic review[J]. Eur J Dent, 2022, 16(1): 32–40. doi: 10.1055/s-0041-1731887
|
| [29] |
Langie SAS, Moisse M, Declerck K, et al. Salivary DNA methylation profiling: aspects to consider for biomarker identification[J]. Basic Clin Pharmacol Toxicol, 2017, 121(S3): 93–101. doi: 10.1111/bcpt.12721
|
| [30] |
Nishitani S, Parets SE, Haas BW, et al. DNA methylation analysis from saliva samples for epidemiological studies[J]. Epigenetics, 2018, 13(4): 352–362. doi: 10.1080/15592294.2018.1461295
|
| [31] |
Leroy JL, Ruel M, Habicht JP, et al. Linear growth deficit continues to accumulate beyond the first 1000 days in low- and middle-income Countries: Global evidence from 51 national surveys[J]. J Nutr, 2014, 144(9): 1460–1466. doi: 10.3945/jn.114.191981
|
| [32] |
Ekholuenetale M, Barrow A, Ekholuenetale CE, et al. Impact of stunting on early childhood cognitive development in benin: evidence from demographic and health survey[J]. Egypt Pediatr Assoc Gaz, 2020, 68(1): 31. doi: 10.1186/s43054-020-00043-x
|
| [33] |
Venables PH, Raine A. The impact of malnutrition on intelligence at 3 and 11 years of age: the mediating role of temperament[J]. Dev Psychol, 2016, 52(2): 205–220. doi: 10.1037/dev0000046
|
| [34] |
Aurora WID, Sitorus RJ, Flora R. Effect of stunting on intelligence quotient (IQ) of school-age children[C]//Proceedings of the 3rd Green Development International Conference (GDIC 2020). Dordrecht: Atlantis Press, 2021: 176–180.
|
| [35] |
Koshy B, Srinivasan M, Gopalakrishnan S, et al. Are early childhood stunting and catch-up growth associated with school age cognition? -Evidence from an Indian birth cohort[J]. PLoS One, 2022, 17(3): e0264010. doi: 10.1371/journal.pone.0264010
|
| [36] |
Mohd Nasir MT, Norimah AK, Hazizi AS, et al. Child feeding practices, food habits, anthropometric indicators and cognitive performance among preschoolers in Peninsular Malaysia[J]. Appetite, 2012, 58(2): 525–530. doi: 10.1016/j.appet.2012.01.007
|
| [37] |
Prado EL, Larson LM, Cox K, et al. Do effects of early life interventions on linear growth correspond to effects on neurobehavioural development? A systematic review and meta-analysis[J]. Lancet Glob Health, 2019, 7(10): e1398–e1413. doi: 10.1016/S2214-109X(19)30361-4
|
| [38] |
Taylor RM, Smith R, Collins CE, et al. Global DNA methylation and cognitive and behavioral outcomes at 4 years of age: A cross-sectional study[J]. Brain Behav, 2020, 10(4): e01579. doi: 10.1002/brb3.1579
|
| [39] |
Hartwig FP, De Mola CL, Davies NM, et al. Breastfeeding effects on DNA methylation in the offspring: A systematic literature review[J]. PLoS One, 2017, 12(3): e0173070. doi: 10.1371/journal.pone.0173070
|
| [1] | Renqi Li, Qiuting Zeng, Muhuo Ji, Yue Zhang, Mingjie Mao, Shanwu Feng, Manlin Duan, Zhiqiang Zhou. Oxytocin ameliorates cognitive impairments by attenuating excitation/inhibition imbalance of neurotransmitters acting on parvalbumin interneurons in a mouse model of sepsis-associated encephalopathy[J]. The Journal of Biomedical Research. DOI: 10.7555/JBR.37.20230318 |
| [2] | Zhenxiang Gong, Li Ba, Jiahui Tang, Yuan Yang, Zehui Li, Mao Liu, Chun Yang, Fengfei Ding, Min Zhang. Gut microbiota links with cognitive impairment in amyotrophic lateral sclerosis: A multi-omics study[J]. The Journal of Biomedical Research, 2023, 37(2): 125-137. DOI: 10.7555/JBR.36.20220198 |
| [3] | Hamed Amini Amirkolaee, Hamid Amini Amirkolaee. Medical image translation using an edge-guided generative adversarial network with global-to-local feature fusion[J]. The Journal of Biomedical Research, 2022, 36(6): 409-422. DOI: 10.7555/JBR.36.20220037 |
| [4] | Asra Mariyam, Lin Xiaofang, Ul Haq Ijaz, Pasha Asfia Banu, Shao Yingying, Qin Rui, Hu Manli, Ding Ye, Wang Zhixu. Malnutrition associated factors on children under 5 years old in Lhaviyani Atoll, Maldives[J]. The Journal of Biomedical Research, 2020, 34(4): 301-308. DOI: 10.7555/JBR.33.20180141 |
| [5] | Tao Chun'ai, Gan Yongxin, Su Weidong, Li Zhutian, Tang Xiaolan. Effectiveness of hospital disinfection and experience learnt from 11 years of surveillance[J]. The Journal of Biomedical Research, 2019, 33(6): 408-413. DOI: 10.7555/JBR.33.20180118 |
| [6] | Huanqiang Wang, Congying Yang, Siyuan Wang, Tian Wang, Jingling Han, Kai Wei, Fucun Liu, Jida Xu, Xianzhen Peng, Jianming Wang. Cell-free plasma hypermethylated CASZ1, CDH13 and ING2 are promising biomarkers of esophageal cancer[J]. The Journal of Biomedical Research, 2018, 32(6): 424-433. DOI: 10.7555/JBR.32.20170065 |
| [7] | David Li, Hong Liu. Cognitive function assessment should be included in preoperative evaluation[J]. The Journal of Biomedical Research, 2018, 32(3): 161-163. DOI: 10.7555/JBR.32.20180008 |
| [8] | Nuan Wang, Xianming Chen, Deqin Geng, Hongli Huang, Hao Zhou. Ginkgo biloba leaf extract improves the cognitive abilities of rats with D-galactose induced dementia[J]. The Journal of Biomedical Research, 2013, 27(1): 29-36. DOI: 10.7555/JBR.27.20120047 |
| [9] | Aixia Zhang, Brian Cao. Generation and characterization of an anti-GP73 monoclonal antibody for immunoblotting and sandwich ELISA[J]. The Journal of Biomedical Research, 2012, 26(6): 467-473. DOI: 10.7555/JBR.26.20120057 |
| [10] | Xinjian Liu, Xiaomin Feng, Qi Tang, Zhongcan Wang, Zhenning Qiu, Yuhua Li, Changjun Wang, Zhenqing Feng, Jin Zhu, Xiaohong Guan. Characterization and potential diagnostic application of monoclonal antibodies specific to rabies virus[J]. The Journal of Biomedical Research, 2010, 24(5): 395-403. DOI: 10.1016/S1674-8301(10)60053-X |
Figures(2) / Tables(7)
Supported by: Beijing Renhe Information Technology Co., Ltd. E-mail:
info@rhhz.net 
| Variables | n (%) |
| Age (months) | |
| 0–24 | 138 (60) |
| 24–48 | 14 (6) |
| ≥ 48 | 79 (34) |
| Sex | |
| Male | 131 (57) |
| Female | 100 (43) |
| Outcomes | |
| Normal (−2 ≤ HAZ ≤ 2) | 152 (66) |
| Stunting (−3 ≤ HAZ < −2) | 49 (21) |
| Severe stunting (HAZ < −3) | 30 (13) |
| Villages (normal, stunting, and severe stunting) | |
| Kadubelang (N = 73) | 47 (65), 17 (23), 9 (12) |
| Kadumaneuh (N = 72) | 43 (60), 14 (19), 15 (21) |
| Medong (N = 86) | 61 (71), 19 (22), 6 (7) |
| Abbreviation: HAZ, height-for-age Z score. | |
DownLoad:
CSV
| Variables | Normal [n (%)] |
Stunting [n (%)] |
Severe stunting [n (%)] |
P-value |
| Age (months) | ||||
| 0–24 | 87 (57) | 27 (55) | 24 (80) | 0.204 |
| 24–48 | 9 (6) | 4 (8) | 1 (33) | |
| ≥ 48 | 56 (37) | 18 (37) | 5 (17) | |
| Sex | ||||
| Male | 90 (59) | 29 (59) | 12 (40) | 0.140 |
| Female | 62 (41) | 20 (41) | 18 (60) | |
| Villages | ||||
| Kadubelang | 47 (31) | 16 (32) | 9 (30) | 0.776 |
| Kadumaneuh | 44 (29) | 14 (29) | 15 (50) | |
| Medong | 61 (40) | 19 (39) | 6 (20) | |
| The height-for-age Z score (HAZ) for children was calculated using the WHO growth standard. Children with −2 ≤ HAZ ≤ 2 were considered normal, −3 ≤ HAZ < −2 were considered stunted, and HAZ < −3 were considered severely stunted. P-value was calculated by the Chi-square analysis. | ||||
DownLoad:
CSV
| Variables | 5mC levels in % [median (IQR)] | P |
| Age (months) | <0.001a | |
| 0–24 | 2.81 (2.53–4.62) | |
| 24–48 | 4.09 (3.76–4.56) | |
| > 48 | 4.01 (3.39–4.87) | |
| Post hoc analysis | ||
| 0–24 months vs. 24–48 months | 0.038b | |
| 0–24 months vs. > 48 months | < 0.001b | |
| 24–48 months vs. > 48 months | 1b | |
| Sex | 0.0187c | |
| Male | 3.83 (2.79–4.83) | |
| Female | 3.25 (2.60–4.48) | |
| Outcomes | 0.031a | |
| Normal (−2 ≤ HAZ ≤ 2) | 3.75 (2.80–4.74) | |
| Stunting (−3 ≤ HAZ < −2) | 3.64 (2.70–4.57) | |
| Severe stunting (HAZ < −3) | 2.84 (2.39–4.62) | |
| Post hoc analysis | ||
| Normal vs. stunting | 1b | |
| Normal vs. severe stunting | 0.028b | |
| Stunting vs. severe stunting | 0.132b | |
| aThe Kruskal-Wallis test was employed to compare the median of multiple groups.bDunn-Bonferroni was employed as a post hoc analysis.cThe Mann-Whitney test was employed to compare the median of the two groups.Bold font indicates P-values < 0.05.Abbreviations: HAZ, height-for-age Z score; IQR, interquartile range. | ||
DownLoad:
CSV
| Outcomes | n | r | 95% CI | P |
| Normal (−2 ≤ HAZ ≤ 2) | 152 | − |
||
| Stunting (−3 ≤ HAZ < −2) | 49 | |||
| Severe stunting (HAZ < −3) | 30 | − |
||
| Spearman correlation analysis was performed. Bold font indicates P-values < 0.05. Abbreviations: HAZ, height-for-age Z score; CI, confidence interval. | ||||
DownLoad:
CSV
| Variables | n | 5mC levels [median (IQR)] |
P |
| Sex | 0.59a | ||
| Male | 47 | 3.69 (3.26–4.03) | |
| Female | 33 | 3.68 (3.38–4.23) | |
| HAZ | 0.63b | ||
| −2 ≤ HAZ ≤ 2 | 51 | 3.68 (3.25–4.27) | |
| −3 ≤ HAZ < −2 | 18 | 3.65 (3.39–4.09) | |
| HAZ < −3 | 11 | 3.96 (3.34–4.20) | |
| aMann-Whitney was employed to compare median values of 5mC levels in males and females.bKruskal-Wallis was employed to compare median values of 5mC levels in three different HAZ statuses.Abbreviations: HAZ, height-for-age Z score; IQR, interquartile range. | |||
DownLoad:
CSV
| Variables | n | PIQ [median (IQR)] |
P | VIQ [median (IQR)] |
FIQ [median (IQR)] |
||
| Sex | 0.983a | 0.936a | 0.996a | ||||
| Male | 32 | 99.5 (86–111) | 100 (95–116) | 101 (94–110) | |||
| Female | 23 | 100 (92–108) | 104 (89–116) | 101 (90–113) | |||
| HAZ | 0.797a | 0.573a | 0.825a | ||||
| −2 ≤ HAZ ≤ 2 | 38 | 100 (85–108) | 101 (94–117) | 101 (94–111) | |||
| HAZ < −2 | 17 | 99 (92–113) | 97 (89–115) | 99 (91–112) | |||
| Villages | 0.293b | 0.0065b | 0.019b | ||||
| Medong | 22 | 102 (97–111) | 109 (99–121) | 111 (97–115) | |||
| Kadumaneuh | 18 | 97 (80–105) | 93 (85–100) | 96 (79–103) | |||
| Kadubelang | 15 | 100 (92–108) | 105 (98–113) | 102 (99–106) | |||
| aP-value by the Mann-Whitney test.bP-value by the Kruskal-Wallis ANOVA.Abbreviations: HAZ, height-for-age Z score; PIQ, performance intelligence quotient; VIQ, verbal intelligence quotient; FIQ, full-scale intelligence quotient; IQR, interquartile range. | |||||||
DownLoad:
CSV
| Variables | PIQ | VIQ | FIQ | ||||||||
| Coefficient (95% CI) | P | R2 | Coefficient (95% CI) | P | R2 | Coefficient (95% CI) | P | R2 | |||
| Sex (Female) | −0.0625 (−9.14–9.01) | 0.98 | 0.10 | 3.05 (−6.20–12.31) | 0.50 | 0.10 | 2.18 (−6.64–11.01) | 0.62 | 0.10 | ||
| Villages (Kadumaneuh) | −6.96 (−18.16–4.23) | 0.21 | 0.36 | −11.39 (−22.81–0.03) | 0.050 | 0.36 | −9.81 (−20.71–1.08) | 0.07 | 0.36 | ||
| Stunting outcomes (stunted) | −6.88 (−16.92–3.15) | 0.17 | 0.16 | −6.21 (−16.46–4.02) | 0.23 | 0.16 | −7.11 (−16.88–2.64) | 0.14 | 0.16 | ||
| 5mC levels | |||||||||||
| Buccal swab DNAa | 1.33 (−2.02–4.68) | 0.42 | 0.24 | −0.21 (−3.63–3.21) | 0.90 | 0.24 | 0.67 (−2.59–3.93) | 0.68 | 0.24 | ||
| Mouthwash DNAb | 9.34 (−14.45–33.13) | 0.43 | 0.07 | 7.18 (−17.08–31.46) | 0.55 | 0.075 | 9.47 (−13.66–32.61) | 0.41 | 0.07 | ||
| aValues of 5mC levels in buccal swab DNA were left untransformed.bValues of 5mC levels in mouthwash DNA were natural logarithm transformed to obtain a normal distribution.Abbreviations: PIQ, performance intelligence quotient; VIQ, verbal intelligence quotient; FIQ, full-scale intelligence quotient; CI, confidence interval. | |||||||||||
DownLoad:
CSV
Authors and Reviewers
