Apolipoprotein E is a genetic risk factor for fetal iodine deficiency disorder in China

Article metrics


Fetal iodine deficiency disorder (FIDD) is the principal form of endemic cretinism, and the most common cause of preventable mental deficiency in the world. However not everyone at risk develops FIDD and familial aggregation is common. This suggests that genetic factors may also be involved. The Apolipoprotein E (APOE) gene encodes for a lipoprotein that possesses a thyroid hormone binding domain, and APOE genotype may affect the efficiency with which thyroid hormone influences neuronal cell growth during the first and second trimesters of fetal development. We have compared ApoE genotypes in 91 FIDD cases with 154 local control subjects, recruited from three iodine deficiency areas in central China. We have also genotyped 42 FIDD family cases and 158 normal individuals from the families of local controls, and 375 population controls from Shanghai. APOE ε4 genotypes were significantly enriched in FIDD probands from each of the three iodine deficiency areas; the ε4 allele frequency was 16% vs 6% in controls. The same effect was also observed when we compared FIDD family cases with controls and control families. Our data suggest that in iodine-deficient areas, the APOE ε4 allele is a genetic risk factor for FIDD. The phenomenon may affect population selection and contribute to the low frequency of the ε4 allele in Chinese compared to Caucasian populations.


Iodine deficiency is the commonest cause of preventable mental retardation.1 One hundred thousand children a year are born with features of frank cretinism,2 and many times more are born with lesser mental and neurological deficits attributable to iodine deficiency.3 The neurological form of cretinism, also known as fetal iodine deficiency disorder (FIDD), predominates in China and is characterized by mental retardation and abnormalities of speech, hearing impairment, and disturbance of posture and gait.3 It results from inadequate amounts of thyroid hormone being available to the developing fetal brain in the second and probably some of the first trimester.4, 5 The damage is permanent and cannot be reversed by postnatal thyroid hormone supplementation. This contrasts with the myxedematous forms of cretinism whose severity depends upon the duration and magnitude of postnatal hypothyroidism, and is treatable to a large extent by postnatal thyroid hormone therapy.6

Clinicopathological studies of FIDD are limited in number. Macroscopically the brain may be smaller than normal but is otherwise unremarkable.7 Microscopic examination shows reduced numbers, irregular arrangement and degeneration of neurons in several areas of the brain. The neurons have fewer dendrite branches, and the number and pattern of their spines is deranged.8 While fetal iodine deficiency during early development is the major agent responsible for neurological cretinism/FIDD, it does not explain why only some of those at risk develop FIDD. Familial clustering is reported, which suggests that genetic factors may interact with an iodine-deficient physical environment to determine overall risk of FIDD.2 APOE is the structural gene for apolipoprotein E (apoE), an apolipoprotein involved in lipid transport and metabolism. It is synthesized in the liver and by the astrocyte cells of the CNS. The three major isoforms (E2, E3, E4), are encoded by alleles ε2, ε3, ε4 and modulate the biological functions of apoE, in part by altering the binding of the different lipoprotein lipid classes.9 Individuals carrying the APOE ε4 allele have higher plasma and neuronal levels of cholesterol than individuals with ε2 or ε3 and the ε4 allele is strongly linked with Alzheimer's disease in both familial and sporadic forms.9, 10, 11 ApoE also has an isoform-specific effect on neuronal growth with E3 stimulating and E4 impairing neuronal elongation and neurite outgrowth in dorsal root ganglia.12 The abnormalities of the neuronal dendrites and their synaptic spines reported in neurological cretinism bear a striking resemblance to those reported in Alzheimer's disease especially among APOE ε4 carriers.13 Similar changes are reported in APOE knockout mice, and also when the human ApoE ε4 transgene is introduced into APOE null mice.14 ApoE has a thyroid hormone (T4) binding domain, and may be involved in transport of T4 into cells.3, 15 In addition 3,5,3′-Triiodothyronine (T3) upregulates transcription of APOE and also acts post-transcriptionally.16 In the light of this circumstantial evidence we decided to evaluate APOE as a candidate susceptibility locus for FIDD.

Materials and methods

We began by investigating APOE genotypes in FIDD cases from Ning Qiang county, an iodine-deficient area in Shaanxi province, China. When we found a significant increase in ε4 allele frequency, we proceeded to examine FIDD cases from two other geographically separate iodine deficiency areas, Zha Shui and Fang counties. Ning Qiang, and Zha Shui counties are both in Shaanxi province but over one hundred kilometers apart, and were selected by the health department of the provincial government, for a longitudinal social, developmental and epidemiological survey of FIDD in the Qin Ba mountain region. Fang county is in the neighboring province of Hubei. All three are areas of subsistence farming situated in mountainous regions (average elevation 750–1500 meters). Soil erosion is widespread and soil and water iodine levels are low. We recorded iodine levels in water of 1.178 μg L−1 in Ning Qiang vs >5 μg L−1 in non-iodine deficient areas. There is no selenium deficiency. FIDD is prevalent and affects 5–8% of the adult population. Endemic goiter has been observed but is now relatively infrequent. We encountered four men in Fang county who had large visually obvious goiters, otherwise goiters were only detectable by palpation. For several years the provincial health authorities have conducted iodination programs, and this together with improved communications has resulted in a marked decline in the frequency of new cases of endemic cretinism in all three areas.

Our group, which included two authors (FZ and JF) who were familiar with the local dialects, visited each of these areas for periods of 3–4 weeks. We explained the nature of our study to leading members of the local population in each of the villages we surveyed. We then asked permission to approach children and adults whom they thought might have FIDD. We indicated that mental retardation was a key feature we were looking for, as well as other abnormalities associated with endemic cretinism including speech difficulties, hearing problems, gait disturbance and small stature. We also requested to be introduced to non-mentally retarded local inhabitants who might volunteer as control subjects. With this cooperation we screened over 400 children aged 4–14 years at risk, as well as a high proportion of potentially suitable adults. Where circumstances allowed, the families of probands also participated. We performed a three-stage screening procedure.

Screen for social adaptability or mental handicap

Participants were screened using Chinese versions of the Denver Developmental Screening Test (DDST) (0–3 years).17 Subjects aged 4–14 years were screened using the Adaptive Scale of Infant and Children revised by Qi-Hua Zuo et al, Beijing Medical University.18 Those suspected of mental retardation aged over 14 years were evaluated using the Rating Scale of Mental Handicap for Adults, as revised by Yao-Xian Gong et al, Hunan Medical University.19 Using these scales each person was given a social adaptive score or mental handicap score. Those with no disability on these scales invariably also have an IQ within the normal range, and for the purpose of the study, were therefore classified as normal controls.

IQ testing

Children 4–5 years old were tested with the Chinese-Wechsler Young Children Scale of Intelligence, C-WYCSI.20 Those 6–16 years old were tested with the Chinese-Wechsler Intelligence Scale for Children, C-WISC.21 For those over 16 years old we used the Wechsler Adults Intelligence Scale – Revision & Chinese, WAIS – RC.22 We selected as a cut-off for mental retardation (MR) an IQ of less than 70. We defined IQs in the range 70–55 accompanied by social disability (SD) scores of 8 or less as modest MR, IQs 55–40 and SD scores of 7 or less as moderate MR, and IQs less than 40 and SD scores of 6 or less as severe MR.

Neurological examination

It was usually not possible or appropriate to perform a formal IQ test in cases of frank cretinism, and we had to depend upon clinical diagnosis. A neurological examination, conducted by a physician, included tests of hearing, vision, voice and speech, reflexes, posture and gait. Most cases we clinically classified as cretins had moderate to severe mental retardation with varying degrees of deaf-mutism and a characteristic gait. We encountered no cases of microcephaly.

We excluded from the study cases of MR if causes other than iodine deficiency were suspected. These included infection, trauma and birth complications. The remainder of MR cases with IQ below 70 we classified as FIDD. There was no evidence in all three areas that other genetic causes of MR or deaf mutism were overrepresented. Cases affected by trachoma, which was common in Fang county were excluded. Many FIDD probands came from multiplex families. Wherever possible we also recruited other family members with MR thought to be due to FIDD. Where a child below 14 years old met the criteria for FIDD he/she was called the proband. Otherwise we called the proband the first adult we identified in each family who fulfilled criteria for FIDD. Controls came from the same iodine-deficient areas and were selected from families with no history of MR. They resembled FIDD cases in age, sex, and origin except that generally they had received more education and their living conditions were slightly better. If permission was granted a blood sample was removed for routine hematology, serology and DNA analysis. We did not include individuals who we identified with an IQ 70–80. At the start of the study it was decided that we could not with confidence classify these cases for genetic analysis, and blood for DNA analysis was not obtained. No significant biochemical differences were observed in serum TSH, T3 and T4 values in FIDD compared to local controls or values reported for other Chinese populations.

DNA extraction and ApoE genotyping

Blood samples were stored at −20°C. Genomic DNA was extracted from blood using a modified phenol/chloroform method. PCR was performed on the DNA thermal cycler (Perkin-Elmer 480 or 9600) with 50 μl of total volume containing 50 mM KCl, 10 mM Tris-HCl (pH 8.0), 0.001% gelatin (w/v), 1.5 mM MgCl2, 10% DMSO, 200 mM dNTPs, 8 pM each primer, 100 ng DNA and 2 U Taq polymerase (Sangon, Ottawa, Canada). ApoE genotyping was performed as described22 except PCR amplification used primers 5′-GGC ACG GCT GTC CAA GGA GCT-3′ (forward) and 5′-GAT GGC GCT GAG GCC GCG CT-3′ (reverse). Because we used a slightly different reverse primer than described,22 the length of the PCR product (262 bp) was 4 bp longer. Otherwise the PCR and digestion (Hha1) conditions were identical.23 PCR products were separated on 8% PAGE gels and visualized by silver staining. All genotyping was performed twice, and genotypes assigned by two raters familiar with apoE genotyping but blind to diagnosis. Any differences were resolved by re-genotyping samples.

Statistical methods

Allele frequencies were calculated using a Microsoft Excel spreadsheet. Hardy–Weinberg equilibrium was tested by a χ2 test. The Woolf method was used to combine data from the different areas in a stratified analysis and to obtain P-values, odds ratios (ORs) and 95% confidence intervals.24 This method of meta-analysis is particularly good for combining studies where the overall population allele frequencies may be different and is easy to implement via any spreadsheet package. Its results compare well with other methods of meta-analysis, such as Mantel–Haenzel, although they can differ slightly particularly when the effect size is small and the number of studies to be analyzed large.


Table 1 presents combined demographic data for the samples collected in each area and gives number collected, sex ratio and mean age for the MR and control cohorts.

Table 1 Number of samples collected in each area, sex ratio and mean age

Of a total of one hundred and thirty-two FIDD cases, forty had pronounced features of endemic cretinism. As well as moderate to severe mental retardation, 34 had profound deaf mutism, two profound hearing impairment, and four some hearing impairment but profound speech disorder and disturbance of gait. These cases were distributed throughout the three areas. The remaining 92 cases had mental retardation together with varying degrees of hearing and speech difficulties and disturbance of gait. None of the cases we included could read more than a few simple characters and words of Chinese.

Table 2 presents combined disability and IQ data for the FIDD cases. Cut-off points for the degrees of mental retardation are described in the text. We found a similar proportion and range of disability and IQ scores in FIDD cases in each of the three areas. Table 3 gives significance levels, odds ratios and confidence intervals for FIDD probands vs local controls for each of the three geographical regions.

Table 2 Disability and IQ data for FIDD cases in three iodine-deficient areas
Table 3 Genotypes and allele frequencies of FIDD and controls from three iodine-deficient areas

Table 4 compares ApoE genotypes and allele frequencies for FIDD probands with other FIDD cases from the same families. Control probands are also compared with other family controls. We found that the highly significant enrichment of ApoE ε4 observed in FIDD probands was also observed when family FIDD cases were compared to control family cases. The same significant differences were present when FIDD cases were compared to the Shanghai population, although ε4 frequency was slightly higher in this latter group than the other controls.

Table 4 Odds ratios, confidence intervals and P-values for FIDD vs controls

Table 5 gives ApoE genotypes and allele frequencies for FIDD and control probands in each of the three areas. Also presented are genotypes and allele frequencies for 375 population controls under the age of 65 years from Shanghai.

Table 5 Genotype and allele frequencies for all FIDD probands, FIDD affected relatives, controls and relatives and non-iodine deficient controls (Shanghai)


Our results show that in each of the iodine-deficient areas ApoE ε4 genotypes and allele frequencies are significantly enriched in FIDD cases compared to matched controls, who were screened and selected from the same background for the absence of mental retardation. Indeed the pattern is remarkably consistent across each of the three areas. In each case ε4 allele frequency in FIDD cases is at least twice that observed in the local mentally unimpaired controls. When the results from the three areas are combined the differences are highly significant (P = 0.00032). We observed a similar pattern when we compared ε4 frequencies in 42 FIDD relatives of FIDD probands with frequencies in unimpaired relatives of control probands (P = 0.00002). Indeed the ε4 allele frequency was higher in the FIDD family cases than in the FIDD probands themselves (23.81% vs 16.48%).

We initially attempted to establish segregation ratios for FIDD in the families. Unfortunately this did not prove possible on account of difficulties in classifying members of the FIDD families. Many had mild degrees of mental retardation (IQ 70–80), and there were relatively few with IQs in the range we encountered in the normal families, whom we could with confidence classify as being unaffected. We also had too few ε4 homozygotes to assess if copy number affected the severity of the phenotype. Segregation analysis of 50 families with FIDD in Ecuador2 estimated a P of 0.245, suggesting autosomal recessive inheritance; and a study in Greece yielded a heritability estimate of 32% for endemic goiter.24 Unfortunately neither study included assessment of intellectual performance. These questions concerning mode of inheritance therefore remain unanswered. Nevertheless we are confident that the FIDD cases we included from the families of FIDD probands are a representative sample of familial FIDD. Their enrichment for the ε4 allele is striking and is strong confirmatory evidence that the highly significant ε4 association we encountered in the FIDD probands was not a chance consequence of population stratification.

The APOE ε4 allele frequency of 9% in the coastal Shanghai population controls was higher than the ε4 allele frequency of 6% in the intellectually normal controls from the iodine-deficient areas. We are probably underestimating the real ε4 frequency in the iodine-deficient areas, because a substantial proportion of ε4 alleles are contained in the FIDD cases. We estimate that 5–8% of the population in these areas have some degree of FIDD. Using these figures the real population frequency for ε4 may be around 7.5% and so is similar to that reported in an unselected control group from neighboring Sichuan province (Tao Li, personal communication).

Our findings have several important implications. First our observations need to be confirmed by an independent group. A formal segregation analysis on a larger series of FIDD families including members with mild mental deficiency, should also be performed to determine the range of the clinical phenotype associated with one or two copies of the ε4 allele. It will also be important to determine whether families with FIDD cases associated with the ε4 allele differ in outcome or clinical response to treatment, and whether they should be identified and especially targeted during iodination programs.

Why should the developing brain of an APOE ε4-positive fetus be more vulnerable to the effects of iodine deficiency? Although iodine deficiency causes a reduction in the absolute amount of thyroid hormone available to bind at apoE, the thyroid hormone-binding domain, located at the N terminal of apoE, is unlikely to be directly affected by mutations at position 112 (apoE4) or 158 (apoE2).15 It is possible however that uptake of ApoE bound thyroid hormone by cells is facilitated by interacting with ApoE receptors. ApoE isoforms may have differing receptor-binding affinities and this could influence the relative amount of thyroid hormone delivered to the cells. However since T3 directly influences transcription of APOE, it seems equally likely that the reduced ability of ε4 compared to ε3 carriers to catalyse neuronal growth and branching of dendrites is further compromised by a change in transcription driven by reduced T3 levels.16 Whatever the explanation a threshold leading to intellectual impairment is breached more easily with profound implications for neuronal development during key periods of fetal life.

Further genotyping with other polymorphisms at the APOE locus is required, since it is possible that another allele in linkage disequilibrium with APOE ε4 is functionally important in conferring risk from iodine deficiency. Expression studies of the effects of different polymorphisms on thyroid hormone binding to ApoE and subsequent events are also necessary. Autopsies of human fetal brain from iodine-deficient areas are desirable but difficult to organize. It will be interesting to determine whether fetal iodine deficiency has a disproportionate effect upon the behavioral and neurological phenotype in strains of APOE null mice,13 or mice where a human ε4 transgene has been introduced.14 Neuropathological studies on these mice may also demonstrate increased abnormalities of neuronal differentiation and dendrite branching resembling endemic cretinism.11 APOE ε4 genotype and hypothyroidism are both associated with raised cholesterol. Are these processes additive and what are the consequences? Finally iodine deficiency is widespread in all the mainland provinces of China.4 Indeed the earliest description of goiter was reported in Shaanxi province in 2700 BC by Emperor Sheng Nung who recommended treatment with seaweed.25 Perhaps 350 million individuals are currently at risk in China.26 This raises the intriguing possibility that the increased risk of endemic cretinism/FIDD in ε4 carriers confers a reproductive disadvantage, so that in iodine-deficient regions, over the course of history, ε4 is selected against. This could in part account for the low frequency of the ε4 allele in the Chinese population compared to Caucasians. Comparative studies of APOE haplotypes in different ethnic Chinese populations may help to answer some of these interesting questions.


  1. 1

    WHO . National strategies for overcoming micronutrient malnutrition. 89th Session EB 89/27 WHO, Geneva, Switzerland 1992

  2. 2

    Held KR, Cruz ME, Moncayo F . Clinical pattern and the genetics of fetal iodine deficiency disorder (endemic cretinism): results of a field study in Highland Ecuador Am J Med Genet 1990; 35: 85–90

  3. 3

    Ma T, Guo J, Wang F . The epidemiology of iodine-deficiency diseases in China Am J Clin Nutr 1993; 57 Suppl 2: 264S–266S

  4. 4

    Benvenga S, Cahnmann H, Robbins J . Characterization of thyroid-hormone binding to apolipoprotein-e — localization of the binding-site in the exon 3-coded domain Endocrinology 1993; 133: 1300–1305

  5. 5

    Cao XY, Jiang XM, Dou ZH . Timing of vulnerability of the brain to iodine deficiency in endemic cretinism N Engl J Med 1994; 331: 1739–1744

  6. 6

    Boyages SC, Halpern JP . Endemic cretinism: toward a unifying hypothesis Thyroid 1993; 3: 59–69

  7. 7

    DeLong GR . Effects of nutrition on brain development in humans Am J Clin Nutr 1993; 57 Suppl 2: 286S–290S

  8. 8

    Yan Y, Guan C, Leng L, Li J . Quantitative histology study on brain nervous cells of neurological endemic cretins. In: DeLong GR, Robins J, Condiffe PG (eds) Iodine and the Brain Plenum: New York 1989; 359

  9. 9

    Strittmatter WJ, Roses AD . Apolipoprotein E and Alzheimer's disease Annu Rev Neurosci 1996; 19: 53–77

  10. 10

    Saunders AM, Schmader K, Breitner JCS . Apolipoprotein E ε4 allele distributions in late-onset Alzheimer's disease and in other amyloid-forming diseases Lancet 1993; 342: 710–711

  11. 11

    Corder EH, Saunders AM, Strittmatter DE . Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families Science 1993; 261: 921–923

  12. 12

    Nathan BP, Bellosta S, Sanan DA . Differential effects of apolipoproteins E3 and E4 on neuronal growth in vitro Science 1994; 264: 850–852

  13. 13

    Arendt T, Schilder C, Bruckner MK . Plastic neuronal remodelling in patients with Alzheimer's disease carrying apolipoprotein epsilon 4 allele J Neurosci 1997; 17: 516–529

  14. 14

    Raber J, Wong D, Buttini M . Isoform-specific effects of human apolipoprotein E on brain function revealed in APOE knockout mice: increased susceptibility of females PNAS USA 1998; 95: 10914–10919

  15. 15

    Benvenga S . Restriction isotyping of apolipoprotein-E Ann Intern Med 1995; 122: 68–69

  16. 16

    Vandenbrouck Y, Janvier B, Loriette C . The modulation of apolipoprotein E gene expression by 3,3′-5-triiodothyronine in HepG2 cells occurs at transcriptional and post-transcriptional levels Eur J Biochem 1994; 224: 463–471

  17. 17

    Frankenburg WK . The Denver Developmental Screening Test Dev Med Child Neurol 1969; 11: 260–262

  18. 18

    Zuo QH, Zhang ZhX, Lei ZhW . Adaptive Scale of Infant and Children, Beijing, China Medical University of China 1988

  19. 19

    Gong YX . Rating Scale of Mental Handicap for Adults, Hunan, China Hunan Medical College 1985

  20. 20

    Gong YX, Dai XY . Chinese-Wechsler Young Children Scale of Intelligence (C-WYCSI) Hunan China Map Press 1992

  21. 21

    Gong YX, Cai TSh . Wechsler Intelligence Scale for Children, Chinese revision (C-WISC) Map Press: Hunan, China 1993

  22. 22

    Gong YX . Manual for the Wechsler Adult Intelligence Scale: Chinese revision Hunan Medical College: Hunan, China 1989

  23. 23

    Hixon JE, Vernier DT . Restriction isotype of human apolipoprotein E polymorphism by gene amplification and cleavage with Hhal J Lipid Res 1990; 31: 545–548

  24. 24

    Woolf B . On estimating the relation between blood groups and disease Ann Hum Genet 1995; 19: 251–253

  25. 25

    Malamos B, Koutras DA, Kostamis GA . Endemic goiter in Greece: a study of 379 twin pairs J Med Genet 1967; 4: 16

  26. 26

    Merke F . History and Ichnography of Endemic Goiter and Cretinism MTP Press: Lancaster 1984

Download references


We thank Professor Ruilin Li, Qizheng Chen for clinical diagnosis and everyone in the iodine-deficient regions for their hospitality and cooperation. This work was supported by the Chinese Academy of Sciences, the government of Shaanxi province, the 973 Project, the National Natural Science Foundation of China, and the Technology Commission of Shanghai Municipality.

Author information

Correspondence to L He.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Wang, H., Zhang, F., Gao, J. et al. Apolipoprotein E is a genetic risk factor for fetal iodine deficiency disorder in China. Mol Psychiatry 5, 363–368 (2000) doi:10.1038/sj.mp.4000735

Download citation


  • fetal iodine deficiency disorder
  • endemic cretinism
  • APOE genotype
  • apolipoprotein
  • China
  • mental retardation

Further reading