|Home | About | Journals | Submit | Contact Us | Français|
Converging evidence suggests a possible link between thyroid state and Alzheimer’s disease (AD), including a higher probability of dementia in individuals with higher TSH levels and a two-fold risk of AD in patients with hypothyroidism. Thyroid hormones modulate factors associated with AD, including amyloid precursor protein expression in the brain, suggesting a possible role for thyroid hormone in AD pathology. The present study is the first to directly evaluate brain thyroid hormone levels in AD. Triiodothyronine (T3) and thyroxine (T4) levels were measured with radioimmunoassay (RIA) in post-mortem samples of prefrontal cortex of patients with pathologically confirmed AD, including Braak stage I–II (n=8), Braak stage V–VI (n=8), and controls without any primary neurological disease (n=8). T4 levels did not differ between groups. T3 levels were significantly lower in Braak stage V–VI brains relative to controls, but there was no statistically significant difference between T3 levels in Braak stage I–II versus controls. Results suggest that the conversion of T4 to T3 may be affected in advanced AD, perhaps due to alterations in deiodinase activity. Reduced conversion of T4 to T3 in AD may be associated with both AD pathology and the clinical presentation of dementia.
It is well established that hypothyroidism is associated with mental status changes, including secondary or reversible dementia , though the relationship between thyroid disease and risk for Alzheimer’s disease (AD) is unclear [2–5]. Some studies have demonstrated a relationship between subclinical thyroid disease (i.e., elevated or reduced thyroid stimulating hormone [TSH] levels with normal T3 and T4) and cognitive impairment and Alzheimer’s disease (AD) . There is a three- to four-fold higher probability of dementia among individuals with elevated serum thyroid stimulating hormone (TSH), but without overt hypothyroidism , a high prevalence (41%) of autoimmune thyroid disease among familial AD kindreds , and a significantly higher rate of antibodies to thyroid peroxidase (TPO) in AD patients compared to controls . There is also a relationship between mood state and thyroid status in euthyroid patients with AD .
Despite the link between HPT abnormalities and dementia, the exact nature of the relationship is unclear. Clinical studies of patients diagnosed with AD have identified HPT abnormalities in this population, including increased risk for AD associated with both lower and higher serum TSH levels. Individuals with reduced TSH levels outside of the normal range showed a greater than threefold increased risk of dementia at a two-year follow-up compared to euthyroid individuals . Similarly, individuals with lower TSH values within the reference range have an increased risk for AD after controlling for the effects of potentially confounding medical comorbidities, such as diabetes mellitus and blood pressure . In addition, patients with AD show significantly lower T3 levels and a blunted TSH response to thyrotropin releasing hormone (TRH) . TRH, a neuropeptide that regulates anterior pituitary release of TSH, is depleted in the hippocampus of post-mortem AD brains compared to controls . Finally, there may be an important association between thyroid state and risk for future cognitive decline and dementia. Volpato and colleagues (2002) found that thyroxine (T4) concentrations within the normal range were significantly associated with an increased risk for future cognitive decline in a large, community-based sample of older, physically impaired, non-demented women . Compared with women in the highest tertile of T4 levels, those in the lowest tertile had a twofold risk of future cognitive decline. The interrelationship between levels of thyroid hormone and AD is not without controversy, however, as higher serum total and free T4 levels were associated with AD pathology at autopsy  and higher serum concentrations of T4 were associated with worse cognition in AD . Similarly, higher serum T4 levels were associated with a better treatment response to donepezil in individuals with AD .
The association between the HPT axis and AD is plausible given that laboratory studies implicate a strong relationship between thyroid state and factors associated with the pathogenesis of AD, including β-amyloid (Aβ) deposition and neuronal apoptosis. For example, T3 plays a regulatory role of Aβ , a major component of senile plaques found in AD and central to the pathogenesis of AD . Thyroid response elements (TREs) have been found on the amyloid precursor protein (APP) gene, a precursor of Aβ, and T3 negatively regulates APP gene expression by directly repressing the APP promoter . Cell treatment with T3 also alters the splicing of APP and secretion of its various isoforms , leading to complex alterations in the expression of Aβ. There also appears to be a complex relationship between transthyretin (a serum transport protein), thyroid hormones, and AD pathogenesis. Transthyretin prevents the deposition of Aβ protein fibrils by creating soluble Aβ complexes . In addition, transthyretin is a transport protein for T4 and is the only carrier protein for T4 in cerebrospinal fluid (CSF). Transthyretin is reduced in the CSF of AD patients compared to age-matched controls , suggesting not only an etiologic role of transthyretin in the deposition of Aβ in AD, but also a possible reduction in T4 transport into the brain in AD patients. Furthermore, recent findings indicate a relationship between apolipoprotein (ApoE), transthyretin, and actin, which suggests a possible metabolic role of ApoE genetics and transthyretin in cytoskeleton biochemistry that may be associated with AD pathogenesis .
Despite the evidence linking the HPT axis and AD, there has been little work directly examining TH concentrations in the brains of AD patients. Circulating thyroid hormones are produced in the thyroid gland, primarily in the form of T4, and are transported into the brain by transthyretin . T4 is then converted to its bioactive form, T3, by 5’-iodothyronine deiodinases. T4 conversion to T3 is regulated in the CNS by type II deiodinase, and type III deiodinase is responsible for the degradation of T4 to reverse T3 (rT3; the inactive isomer of T3). The majority of T3 within the brain is produced locally by intracerebral conversion of T4 to T3 (75–90%), allowing the brain to maintain T3 concentrations within very narrow limits .
Given the interaction of thyroid hormones with several of the key components of AD pathology and their role in cognition and mood, it is possible that an association exists between reduced thyroid hormone concentrations on a tissue level and some of the clinical features of AD. Although a causal relationship between changes in the HPT axis is unlikely, evidence from in vivo and in vitro studies support a complex relationship between thyroid hormones and AD that may have some relevance for adjunctive treatment and management of AD. The aim of the present study was to measure both T4 and T3 concentrations in the prefrontal cortex of post-mortem brains of patients diagnosed with Braak I–II or Braak V–VI AD compared to age- matched controls with no pathological evidence of disease. If deiodinase activity is disrupted in AD, T3 should be lower in brains of AD patients, along with somewhat higher levels of T4. In contrast, a disturbance in T4 transport activity should result in both decreased T3 and T4.
All brains were obtained at autopsy with a post-mortem interval of 2 to 24 hours and stored in the Brain Collection of the Division of Neuropathology at Rhode Island Hospital. All participants provided informed consent for brain donation prior to their death. AD brains were pathologically confirmed and classified according to the Braak neuropathological staging of Alzheimer-related changes, where stages I through VI represent increasing levels neuropathological changes . Control brains were obtained from age-matched hospital patients with no history or evidence of neurological and neurodegenerative disease. A clinical synopsis of each case is provided in Table 1. There were no significant group differences in age or gender. The prefrontal cortex (Brodmann area 10) was sampled for three groups - Braak stage I–II (n=8), Braak stage V–VI (n=8), and control brains (n=8). The prefrontal cortex was selected for two primary reasons. This region contains one of the highest concentrations of thyroid hormone receptors in the CNS , and neuropsychological and neuroimaging studies suggest primarily frontal-subcortical system dysfunction in patients with thyroid disease [1, 29].
The original protocol for extraction of TH from brain tissue was adopted from Pinna et al. (1999) as follows . Frozen tissue samples were weighed and cut into approximately 50 mg sections. These were kept on ice for the remainder of the procedure. Tissue samples were homogenized in 2 ml of 100% methanol. An additional sample of control tissue was simultaneously homogenized in 1.5 ml of 100% methanol and 0.5 ml of buffer containing radiolabeled T3. This homogenate was labeled the “recovery sample” and it underwent the same procedures as all the other samples. The homogenates were collected and each tube used for homogenization was washed with another 2 ml of 100% methanol, which was then added to the homogenates. These mixtures were centrifuged at 3000 r.p.m. for 15 min at 4°C. The supernatants were collected, the pellets washed with another 3 ml of 100% methanol and then centrifuged at 3000 r.p.m. for 15 min at 4°C. The supernatants were again collected and added with the previously collected supernatants to poly-prep ion-exchange chromatographic columns prefilled with AG 1-X8 resin (BioRad, Hercules, California) kept at room temperature. The hormones were purified by washing the columns with the following solutions in the following order: 2 ml acetate buffer pH 7, 2 ml 100% ethyl alcohol, 2 ml acetate buffer pH 7, 2 ml acetate buffer pH 7, 2 ml 100% ethyl alcohol, 2 ml acetate buffer pH 7, 2 ml acetate buffer pH 4, 2 ml acetate buffer pH 3, 2 ml 1% acetic acid, 2 ml 35% acetic acid, 0.5 ml 70% acetic acid. The hormones were eluted with 2 ml of 70% acetic acid. At this point the radioactivity of the recovery sample was measured using a gamma counter, and it was compared to the radioactivity of 500 µl of buffer containing radiolabled T3.
TH concentration in the eluates was measured using T4 and T3 RIA kits (ICN Diagnostics, Costa Mesa, CA) in accordance with manufacturer's instructions. All standards and samples were run in duplicate. For T4, 25 µl of standards or samples were added to antibody-coated (Ab-coated) tubes along with 1 ml of buffer containing radiolabeled T4. The solution was mixed gently and the tubes incubated at room temperature for 1 hr. At the end of this time, the liquid was decanted from all tubes, which were then allowed to drain in an inverted position for at least 1 min. The radioactivity bound in the Ab-coated tubes was counted for 1 min with a gamma counter.
For T3 measurements, 100 µl of standards or samples were added to Ab-coated tubes along with 1 ml of buffer containing radiolabeled T3. This solution was mixed and the tubes incubated at 37°C for 1 hr. After this time, the liquid was decanted from all tubes and each tube washed with 2 ml of distilled water. The tubes were allowed to drain in an inverted position for at least 1 min. The radioactivity bound in the coated tubes was then counted for 1 min with a gamma counter.
Several steps were taken to further adapt the above protocols to account for current laboratory conditions, pH of the extract, measurement of TH in tissue, and low tissue concentrations of T4 and T3. First, an appropriate volume of NaOH was added to the eluates to bring their pH to 7. Second, a purification step was added where the eluates were dried in a speed vacuum pump overnight and then suspended in a PBS buffer. The RIA protocol was then followed without the need to adjust the pH with NaOH. Third, in order to further concentrate the hormones and maximize the yield, 30 mg of tissue was homogenized in 100 µl of 5% TCA, centrifuged, the supernatant mixed with Tris buffer and TH assayed without purification. Additionally, the Ab and bound hormones had to be detached from the RIA tubes and placed in test tubes suitable for measurement with the gamma counter present in the current laboratory. This was accomplished with 0.1 M citric acid containing 0.1% SDS and vigorous vortexing for 1 min. Finally, stripped human serum (free of hormone; ICN Diagnostics, Costa Mesa, CA) had to be added to the RIA mixture in order to provide the protein matrix necessary for hormone binding to the Ab.
The final determinations of TH levels in brain tissue were thus made as follows. Frozen tissue samples were weighed and cut into 100 mg pieces. They were homogenized in 200 µl of 5% TCA and centrifuged at 14,000 r.p.m. for 10 mins. The supernatant was collected and mixed 2:1 with 4M Tris buffer. TH concentrations were measured in duplicate using RIA kits. For T4, the sample was mixed 1:1 with stripped human serum. 25 µl of this new solution or 25 µl of standards were added to Ab-coated tubes along with 1 ml of buffer containing radiolabeled T4, and this incubated at room temperature for 1 hr. After decanting and draining, the tubes were filled with 1 ml of 0.1 M citric acid containing 0.1% SDS. Each tube was vortexed for 1 min and the liquid then transferred to smaller test tubes, which were placed in a gamma counter and the radioactivity counted for 1 min.
The final T3 measurements were obtained by mixing the 100 µl of the sample, 150 µl of stripped human serum and 1 ml of buffer containing radiolabeled T3 in Ab-coated tubes. These were incubated for 1 hr at 37°C. After decanting, washing with distilled water, and draining, the Ab-coated tubes were filled with 1 ml of 0.1 M citric acid containing 0.1% SDS. Each tube was vortexed for 1 min and the liquid then transferred to smaller test tubes, which were placed in a gamma counter and the radioactivity counted for 1 min.
The amount of radioactivity remaining in the RIA tubes, corresponding to the amount of radiolabeled TH bound to the Ab, was detected with the gamma counter for all standards and samples. The maximum amount of binding was obtained from a hormone-free standard. The level of binding for all other standards and samples was calculated as a percentage of this maximum binding. A standard curve was then plotted for known concentrations of T4 and T3 and the corresponding percent of radiolabeled tracer binding (Figs. 1a and 1b). Two samples were run independently and plotted on the standard curve, and the standard curves indicate that the assay kits used in the study were appropriate and stable. The level of T4 and T3 in all of the samples was determined from the standard curve.
Group differences were examined by one-way analysis of variance (ANOVA). Significant main effects and interactions were followed by Tukey post-hoc comparisons. An alpha level of p< 0.05 was used throughout.
The mean radioactivity level for each group was measured according to level of T3/T4 in the samples and adjusted for tissue weight. There were no significant group differences [F(2,23) =3.1, p >.05] in T4 level (ng/mg tissue). Results are presented in Fig. (2). In contrast, there was a main effect of group [F(2,23) = 8, p < .05] for T3. Follow-up Tukey’s t-tests for independent samples indicated significantly lower tissue T3 content in the Braak V–VI group compared to controls [t(14)=2.82, p<.05]; there were no differences in T3 levels between Braak I–II and controls (see Fig. 3).
Results revealed comparable concentrations of T4 in the prefrontal cortex of early (Braak stage I–II) and advanced (Braak stage V–VI) AD brains compared to controls. T3 concentrations were significantly lower in Braak V–VI brains compared to controls, though there was no statistically significant difference between Braak I–II brains and controls.
The mechanism underlying reduced T3, but not T4, concentrations in prefrontal cortex of AD brains is unclear, but data from previous studies examining in vivo thyroid hormone concentrations, metabolism, and transport in AD patients suggest that disruption of at least two possible mechanisms could explain these findings, including: (1) reduced transport of thyroid hormone into the brain due to alterations in transthyretin (a transporter peptide for T4 into the brain) activity; or (2) a disruption in deiodinase which would affect the conversion of T4 to T3 or the degradation of T4. Transthyretin, a transport protein for T4, is reduced in the CSF of AD patients compared to age-matched controls [23, 31]. Alternatively, the transport of T4 may be upregulated in advanced AD through a feedback mechanism designed to compensate for diminished bioactivity of thyroid hormone. Given that there were no group differences in tissue T4 content, however, an overall reduction in thyroid hormone due to reduced T4 transport is not supported.
Alternatively, lower T3 tissue content could be explained by alterations in deiodinase activity. T4 conversion to T3 is regulated by type II deiodinase, and type III deiodinase is responsible for the degradation of T4 to rT3 (the inactive isomer of T3). Alterations in type II deiodinase in AD brains would produce the lowered T3 levels observed in the current study. Interestingly, the possibility of reduced conversion of T4 to T3 in the CNS is supported by a recent study which demonstrated that AD patients who were responders to donepezil presented with higher total T4 and free T4 levels before treatment and showed a significant reduction in these levels after treatment . It is possible that reductions in serum T4 in that study were due to enhanced conversion of T4 to T3 by donepezil. Conversely, an upregulation of type III deiodinase, responsible for degradation of T4 to the inactive rT3, would also lead to a decrease in T3. In fact, patients with AD have been shown to have significantly increased rT3 levels and an increased rT3 to T4 ratio in the CSF, suggesting an abnormal intracerebral thyroid hormone metabolism and possibly brain hypothyroidism in AD . Because rT3 was not measured in this study, future studies are necessary to further address these competing hypotheses.
Results from the current study provide preliminary evidence suggesting that T3, but not T4, is reduced in the prefrontal cortex of post-mortem brains of advanced AD patients. Results should be viewed as preliminary, however, given some methodological limitations. First, tissue samples were restricted to the prefrontal cortex, so the findings from this study may not generalize to other brain structures affected in AD. This brain region was selected because it contains one of the highest concentrations of thyroid hormone receptors in the CNS , and neuropsychological and neuroimaging studies suggest primarily frontal-subcortical system dysfunction in patients with thyroid disease [1, 29]. However, future studies should sample from other structures affected early on in AD, such as the hippocampi. Second, it is possible that the lower T3 levels in the advanced AD brains compared to control brains identified in this study is merely an artifact of disease severity. Serum T3 levels are commonly reduced with an associated increase in rT3 concentrations in patients with prolonged, nonthyroidal illness. Tissue concentrations of thyroid hormone are also reduced in most organs . For example, Arem and colleagues (1993) demonstrated reduced T3, but not T4, in the cerebral cortex of patients who had died from nonthyroidal illness compared to sudden death from trauma . However, inspection of the causes of death of subjects in the present study indicates no systematic bias of nonthyroidal illness versus sudden death between groups, suggesting that illness severity may not fully explain the current findings. Relatedly, serum thyroid hormone concentrations were not available, so the relationship between the brain thyroid hormone levels and serum thyroid hormone levels is unclear. Future studies should measure rT3 concentrations and deiodinase expression and activity in the brain to more directly address this issue, and those studies are currently underway by our group. Finally, we were unable to control for any additional confounds, such as pharmacological agents, as such opiates, administered close to the time of death that could affect the HPT-axis . These potential confounds should be carefully controlled for in a future, larger study.
Results from the current study provide some evidence for an association between AD and the HPT axis. These data should be viewed as preliminary given that the procedures for thyroid hormone measurement in the human brain were adapted from techniques with demonstrated validity for the rat brain and have not been widely utilized . The methodology used in the current study requires replication in future studies to validate the sensitivity of thyroid hormone measurement in human brain tissue. In summary, results provide preliminary evidence to suggest that the brains of patients with advanced AD may be in a state of tissue hypothyroidism that may be associated with some of the clinical features of AD. Additional work is necessary to determine the mechanisms underlying reduced tissue T3 content in AD brains and the clinical correlates of these changes. Clinical investigation of the possible efficacy of adjunctive T3 treatment in AD patients may be warranted.
This research was supported, in part, by NIH grant P30-AG13846 (Boston University Alzheimer’s Disease Core Center).