Abstract
This study reports the development of a highly sensitive and reproducible RIA for the measurement of 3,5-diiodothyronine (3,5-T2) in human serum and tissue. The RIA employs 3-bromo-5-[125I]iodo-L-thyronine (3-Br-5-[125I]T1) as tracer, which was synthesized carrier free by an interhalogen exchange from 3,5-dibromo-L-thyronine (3,5-Br2T0). The detection limits were 1.0 fmol/g and 0.8 pmol/L in human brain tissue and serum, respectively. T3, diiodothyroacetic acid, and 3-monoiodothyronine cross-reacted with a 3,5-T2 antibody to the extent of 0.06%, 0.13%, and 0.65%, respectively.
Serum concentrations of 3,5-T2 were measured in 62 healthy controls and 4 groups of patients with nonthyroidal illness, i.e. patients with sepsis (n = 24), liver diseases (n = 23), head and/or brain injury (n = 15), and brain tumors (n = 21). The mean serum level of 3,5-T2 in the healthy subjects was 16.2 ± 6.4 pmol/L. Concentrations of 3,5-T2 were significantly elevated in patients with sepsis (46.7 ± 48.8 pmol/L; P < 0.01), liver diseases (24.8 ± 14.9 pmol/L; P < 0.01), head and/or brain injury (24.1 ± 11.3 pmol/L; P < 0.05), and brain tumors (21.6 ± 4.8 pmol/L; P < 0.01). In all 4 patient groups, serum levels of T3 were significantly reduced, confirming the existence of a low T3 syndrome in these diseases. Serum concentrations of 3,5-T2 were significantly elevated in patients with hyperthyroidism (n = 9) and were reduced in patients with hypothyroidism (n = 8). The levels of T4, T3, and 3,5-T2 were measured in normal human tissue samples from the pituitary gland and various brain regions and in brain tumors. In normal brain tissue, the concentrations of 3,5-T2 ranged between 70–150 fmol/g, and the ratio of T3 to 3,5-T2 was approximately 20:1. In brain tumors, however, T3 levels were markedly lower, resulting in a ratio of T3 to 3,5-T2 of approximately 1:1.
Recent findings suggest a physiological, thyromimetic role of 3,5-T2, possibly stimulating mitochondrial respiratory chain activity. Should this prove to be correct, then the increased availability of 3,5-T2 in nonthyroidal illness may be one factor involved in maintaining clinical euthyroidism in patients with reduced serum levels of T3 during nonthyroidal illness.
IT IS NOW generally accepted that the major physiological effects of thyroid hormones are exerted by the action of T3 at its specific nuclear receptor sites (1, 2). However, several studies indicate that other iodothyronines whose effects are probably mediated at different cellular loci may also have some functional significance. T4, for example, has been found to promote polymerization of actin cytoskeleton astrocytes in culture (3). In recent years several studies have described the effects of 3,5-diiodo-l-thyronine (3,5-T2) and/or 3,3′-diiodo-l-thyronine (3,3′-T2). Horst et al. (4) reported that 3,5-T2 stimulates oxygen consumption in isolated perfused livers from hypothyroid rats. The same effect has been demonstrated in mononuclear human blood cells (5). Other groups have found that both 3,5-T2 and 3,3′-T2 stimulate the mitochondrial respiratory rate and cytochrome oxidase activity more rapidly than T3 (6–9) and that 3,5-T2 has specific binding sites at hepatic mitochondria (10). Finally, 3,5-T2 has also been reported to suppress TSH secretion and serum concentrations of T4 in the rat (11). This rise in the number of studies reporting possible physiological effects of diiodothyronines contrasts with the as yet limited information available on the concentrations of these hormones in human blood and tissues as well as on their regulation in different physiological and pathophysiological states. We know of seven studies that have measured 3,5-T2 in human serum and found concentrations of between about 4–10 ng/dL (76–190 pmol/L) (1, 12–18). However, the results of these early studies on 3,5-T2 measurement were contradictory. Although some studies found 3,5-T2 concentrations to be elevated in hyperthyroidism (12, 14, 15) and reduced in hypothyroidism (14), others surprisingly reported completely normal concentrations in both thyroid diseases (13, 18) or in hypothyroidism alone (15).
The metabolic impact of the “low T3 state” in nonthyroidal illness is not yet understood. Patients with different somatic disorders often have subnormal serum concentrations of thyroid hormones over prolonged periods, but still do not exhibit clinical features of hypothyroidism (19–21). It would, therefore, seem justified to ask whether some biologically active thyroid hormone metabolite is being formed that is not being measured. In light of the above results and hypotheses regarding a possible physiological function of 3,5-T2, the question arises as to whether the production of 3,5-T2 is enhanced in nonthyroidal illness. We know of only two studies on 3,5-T2 in nonthyroidal illness, i.e. liver cirrhosis (13, 15). Both reported subnormal concentrations. However, it would seem all the more important to conduct more research into 3,5-T2 in nonthyroidal illness in as much as it has to date proven rather difficult to produce labeled 3,5-T2 with a specific radioactivity high enough to permit accurate measurement of low serum levels of 3,5-T2. We, therefore, developed a new 3,5-T2 RIA that employs 3-bromo-5-[125I]-L-thyronine (3-Br-5-[125I]T1) as tracer. The 3-Br-5-[125I]T1 tracer was obtained by a newly developed method in which one atom of bromine in the 3,5-dibromo-L-thyronine (3,5-Br2T0) substrate is replaced by one atom of 125I. The lower limits of detection for 3,5-T2 with this new RIA technique are 1.0 fmol/g and 0.8 pmol/L 3,5-T2 in tissue and serum, respectively. These threshold limits are considerably lower than those in all reports published to date (12–18). With this newly developed method we measured serum concentrations of 3,5-T2 in four different nonthyroidal illnesses as well as in healthy tissue from the human central nervous system and brain tumors.
Subjects and Methods
Patients and controls
Iodothyronines were measured in the following patients and control samples.
Patients with sepsis. Twenty-four patients admitted to the intensive care unit, either postoperatively or after trauma, whose clinical course was complicated by sepsis were studied. Ten of these patients had peritonitis, and 14 had pneumonia. Ten were women, and 14 were men; their mean age was 56.2 ± 19.0 yr (range, 16–89 yr). The diagnosis of sepsis was established according to criteria previously reported (22).
Patients with liver diseases. Serum T3 and T4 levels were measured in 23 patients with severe liver diseases before liver transplantation. Ten were women, and 13 were men; their mean age was 50.4 ± 30.2 yr (range, 23–82 yr). The patients were suffering from various underlying liver diseases, such as chronic hepatitis, Morbus Wilson disease, hemochromatosis, and alcoholic cirrhosis. They were part of a much larger sample of patients who were scheduled for liver transplantation and in whom serum concentrations of T3 and T4 were routinely measured. These 23 patients were purposely selected from the larger sample because they had particularly low serum levels of T3.
Patients with head and/or brain injury. Fifteen patients admitted to an intensive care unit with a diagnosis of closed head injury were also studied. Six of these patients were women, and nine were men; their mean age was 36.2 ± 13.8 yr (range, 28–44 yr). Blood samples were drawn from all patients during the first 2 days after the injury, when they were still in an unconscious state.
The critically ill patients admitted to the intensive care unit (sepsis, and head and/or brain injury groups) were treated with appropriate antibiotics and various drugs to stabilize vital functions. When necessary in patients with sepsis and head injuries, hemodynamic stabilization was achieved by controlled volume load, dobutamine, and norepinephrine. Four patients with sepsis were also given dopamine. None of these patients was receiving glucocorticoids. Patients with known thyroid disorders or those receiving thyroid hormone supplementation were not included in the study.
All patients received 30–35 Cal/kg BW daily either enterally or parenterally (70% in carbohydrate form and 30% in lipid form). Amino acids (1.5 g/kg BW) were also given.
Patients with brain tumors. Twenty-one patients hospitalized for brain surgery at the Department of Neurosurgery were also investigated. Twelve were women, and nine were men; their mean age was 50.1 ± 12.5 yr (range, 22–70 yr). Six of these patients had glioblastoma, five had meningioma, four had astrocytoma, and five were scheduled for removal of brain metastases (their primary tumors were at a different site). An adenoma was excised from the pituitary gland of one patient. The general condition of all of these patients was relatively good. None of them was suffering from a severe concomitant disease. Blood samples were drawn immediately before the operation, i.e. during full anesthesia. The anesthetics propofol and fentanyl as well as a muscle relaxant were given.
Patients with thyroid disorders. Nine patients with hyperthyroidism (six women and three men; mean age, 36.3 ± 4.5 yr; range, 28–44 yr) were also studied. All of these patients were suffering from Graves’ disease. Diagnoses were made according to laboratory parameters (elevated thyroid hormone concentrations and suppressed TSH secretion) and clinical symptoms. Furthermore, eight patients with hypothyroidism (six women and two men; mean age, 43.5 ± 6.3 yr; range, 34–58 yr) were included. Six of them had Hashimoto’s thyroiditis, and two had atrophic immune thyroiditis. All patients were studied before treatment for the underlying thyroid disorder was instituted.
Healthy controls. Sixty-two healthy subjects were studied for comparison. Thirty-two of them were women, and 30 men; their mean age was 47.3 ± 18 yr (range, 22–89 yr). As an age dependence of 3,5-T2 was reported in a previous study (16), we took care to include sufficiently large percentages of all relevant age groups in the control group. Ten subjects were between 20–30 yr of age, 10 were between 30–40 yr, 9 were between 40–50 yr, 11 were between 50–60 yr, 9 were between 60–70 yr, 9 were between 70–80 yr, and 4 were over 80 yr. The subjects less than 60 yr of age were employees of various departments of the Klinikum Benjamin Franklin who were personally known to the authors. Controls older than 60 yr were healthy retired volunteers who occasionally worked for a drug research institute and whose good and stable state of health had been documented in the records of this institute for a period of several years. None of the controls had any apparent clinical illness at the time of the investigation or any history or current signs of thyroid disorder, or was currently taking thyroid hormones or any other medication known to affect serum concentrations of thyroid hormones, such as oral contraceptives or β-adrenergic blockers (23, 24).
All blood samples from the patients and healthy controls were always drawn in the morning between 0700–1000 h. As blood samples were drawn from all patients routinely for diagnostic purposes, this opportunity was used to obtain 5 mL of additional blood to determine thyroid hormone levels for research purposes.
Samples of normal human tissue were collected by the Department of Neurosurgery and Neuropathology of the Klinikum Benjamin Franklin. They had been excised during neurosurgical operations (n = 5) for therapeutic purposes or at autopsy (n = 5) for histopathological investigation. Samples from the anterior pituitary and from various regions of the brain (cortex, hippocampus, pons, and cerebellum) were obtained from donors at autopsy, which was performed between 36–72 h after death. Between death and autopsy the bodies were stored at 4–6 C. Four of the donors were male, and one was female; their mean age was 58.4 ± 12.5 yr (range, 36–70 yr). They had died of different disorders, not primarily affecting the brain, such as lymphoma, myocardial infarction, diabetes mellitus, primary pulmonary hypertension, and dilatative cardiomyopathy.
Two tissue samples from the tempo-cortical lobe were obtained during temporal lobe resection in patients with temporal lobe epilepsy. Pieces of tumor tissue were obtained during surgery from the patients from whom serum samples had been collected (see above). Tissues obtained at operation were immediately frozen in liquid nitrogen and stored at −70 C until assay. The human tissue samples were taken from the tissue collections in which deiodinase activities and concentrations of T3 and T4 had been measured previously (25). All parts of the study were approved by the ethical committee of the Klinikum Benjamin Franklin and (in the case of patients with liver disease) that of the Klinikum Rudolf Virchow. In all cases, except the patients with sepsis and those with head injuries, informed consent was obtained for taking an extra 5-mL blood sample for research samples.
Hormone determinations
Reagents. Tetraiodothyroacetic acid (Tetrac), triiodothyroacetic acid (Triac), diiodothyroacetic acid (Diac), T4, T3, rT3, 3,5-T2, 3,3′-T2, 3′,5′-T2, and thyronine of the highest available purity as well as 3,5-T2 conjugate to BSA for antibody production were purchased from Henning Berlin (Berlin, Germany). Mercury-[(o-carboxyphenyl)thio]ethyl sodium salt (merthiolate), l-cysteine, and BSA were purchased from Sigma Chemical Co. (St. Louis, MO). The tracer 3-Br-5-[125I]T1 was provided by R. Thoma, Formula GmbH (Berlin, Germany). It was synthesized as carrier-free labeled product with a specific radioactivity of 74 megabecquerels/nmol by an interhalogen exchange from 3,5-Br2T0, separated from the reaction products, and purified by high performance liquid chromatography. Phosphate buffer (0.04 mol/L; pH 8.0) containing 243 mg/L merthiolate and 2 g/L BSA served as the experimental buffer. The iodoamino acids were dissolved in 0.1 mol/L sodium hydroxide and diluted to final assay concentrations using this buffer. The 3-Br-5-[125I]T1 tracer was dissolved in experimental buffer containing 100 mg/L l-cysteine. The stop solution formed by the experimental buffer, 30% (wt/vol) polyethylene glycol, and 1.3 mg/mL bovine γ-globulin was pipetted (1 mL/tube) to precipitate the antibody-bound radioactivity in the 3,5-T2 RIA.
Preparation of 3,5-T2 binding antiserum. Antiserum to 3,5-T2 was produced in three rabbits immunized by serial injections of 3,5-T2 conjugated to BSA in complete Freund’s adjuvant as earlier previously (26). All immunized rabbits produced antiserum satisfactory for RIA of 3,5-T2. The antiserum selected for the experiments was obtained after three injections of the immunogen, 8 weeks after starting immunization. It was used in a final dilution of 1:250,000, which bound about 40% of tracer in an incubation volume of 250μ L.
Preparation of test samples. For the serum 3,5-T2 measurements, serum was extracted with 2 vol dehydrated alcohol, evaporated to dryness, and taken up in the experimental buffer. Tissue concentrations of 3,5-T2 were determined after extraction of the tissue samples, as previously described (27). In brief, tissue samples were homogenized in 100% methanol containing 1 mmol/L 6-n-propyl-2-thiouracil, extracted in chloroform-methanol, and back-extracted into an aqueous phase, which was then purified through Bio-Rad AG 1 × 2 resin columns (Bio-Rad Laboratories, Richmond, CA). The iodothyronines were eluted with 70% acetic acid, evaporated to dryness, and taken up in the experimental buffer. Extracts from 200 μL original serum or tissue were processed individually and assayed together within the same run. Each sample was determined in triplicate. The results were corrected on the basis of individual recovery data obtained after the addition of tracer (∼1000 cpm/tube) during the initial extraction process. This amount of tracer did not affect the RIA measurements. The extraction procedure yielded a mean 3,5-T2 recovery of 81.4 ± 4.6% (range, 75.6–87.2%).
RIA procedure for 3,5-T2. The RIA of 3,5-T2 in serum and extracted tissue was carried out in 1.0 × 5.5-cm plastic tubes, adding various reagents as follows: 1) experimental buffer to give a final volume of 250 μL/tube, 2) 50 μL 3,5-T2 standard at concentrations ranging from 0.48–20 fmol/tube (for the serum measurements, standards were diluted in experimental buffer containing the same volume of ethanol-extracted hormone-free serum as the unknown samples), and 3) 100 μL tracer solution containing around 6000 cpm 3-Br-5-[125I]T1. After 24-h incubation at room temperature, the antibody-bound iodothyronine portion was precipitated by adding 1 mL polyethylene glycol stop solution and centrifuged. The supernatant was discarded, and precipitated bound radioactivity was counted.
Determinations of other iodothyronines and TSH. To confirm the low T3 syndrome in the four patient groups with nonthyroidal illness, serum concentrations of T3, T4, and rT3 were measured in the serum samples of these patients and in healthy controls. To confirm hyper- and hypothyroidism, respectively, serum concentrations of T4, free T4 (fT4), and TSH were measured in these two groups of patients. T3, T4, fT4, and TSH were determined in duplicate using the DYNOtest radioassay kits (B.R.A.H.M.S Diagnostica, Berlin, Germany); rT3 was determined using a kit obtained from Serono Diagnostica (Freiburg, Germany). Interassay coefficients of variation for all of these hormones had previously been determined by our laboratory at three different concentrations (28, 29). For each assay, the hormone determinations were performed on the serum samples drawn from the patient groups together with those from some of the healthy controls. We took care to ensure that the respective control group whose samples were investigated together with those of a patient group in a single assay were age and sex matched.
Data analysis
The data are presented as the mean ± sd, and P < 0.05 was considered significant. The mean ± 2 sd were regarded as within the normal range. Linear regression analysis was conducted to evaluate the effects of age and sex on 3,5-T2 concentrations in the healthy subjects. Comparison of hormone concentrations of the patients and controls was performed with the aid of the Mann-Whitney U test.
Results
Assay validation
Figure 1 represents the standard curve showing the displacement of the 3-Br-5-[125I]T1 tracer and the specific antibody to 3,5-T2 effected by increasing concentrations of nonradioactive 3,5-T2. The statistics reveal significant inhibition of 3-Br-5-[125I]T1 binding after the addition of 3,5-T2 concentrations as low as 0.16 fmol/tube. This sensitivity threshold allowed detection of 1.0 fmol/g and 0.8 pmol/L 3,5-T2 in tissue and serum, respectively.
![Standard curve for 3,5-T2. Increasing concentrations of nonradioactive 3,5-T2 were added to displace binding of the 3-Br-5-[125I]T1 tracer and its specific antibody.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/jcem/82/5/10.1210_jcem.82.5.3939/2/m_eg0573939001.jpeg?Expires=1501500118&Signature=R~fwk-mE9cZC0RkVAJpYkBBZ4w5OXt6YvOnaqWqsCNNVXUWBBNMMzXMTam70MaOP53~gGuEq~n7FUbHkYgfgCLA0~CNcjv55d8rb7mowH2Q1SrkBpDYlb5izHitx5Mdm0WTDtfO~RXl6p4lpTs0cEYVaDThNvvY40hyWizXgro0I~W2d-ql~O-aHDzKLd3EAPA9~hzg3lcJ1H58c8O8MXH8vPcjFLprVFSTwRqEZ1kwTRWpDGubutFbEmCVBlazlJe7jcQsy1O5~6wZMKuzZ3Bg19jr6zRtCzju~cU~DP2KfPGyECPXzJTBR2oNoaHuGolQ-xSjr6XbnR86oCbSYHA__&Key-Pair-Id=APKAIUCZBIA4LVPAVW3Q)
Standard curve for 3,5-T2. Increasing concentrations of nonradioactive 3,5-T2 were added to displace binding of the 3-Br-5-[125I]T1 tracer and its specific antibody.
Standard curve for 3,5-T2. Increasing concentrations of nonradioactive 3,5-T2 were added to displace binding of the 3-Br-5-[125I]T1 tracer and its specific antibody.
Dilutions of serum extracts from three 3,5-T2-enriched sera and two sera from hyperthyroid patients were essentially parallel to the standard curve (not shown).
The inter- and intraassay coefficients of variation (CVs) for tissue samples were determined in more than four tests. The intraassay CVs ranged between 6.6–7.8, and the interassay CVs ranged between 7.7–8.2. The measurements were performed using two different samples, both of which caused 50% inhibition of 3,5-T2 antibody binding.
Table 1 shows the abilities of various iodothyronines and iodotyrosines, tested in four or five different concentrations, to displace the binding of 3,5-T2 to its specific antibody. Cross-reaction of the antibody for 3,5-T2 with almost all compounds tested was minimal. In particular, Diac and 3-T1 cross-reacted with the antibody for 3,5-T2 0.13% and 0.65%, respectively. Particular attention was paid to examination of the cross-reaction with T3. After the addition of several concentrations of nonradioactive T3 to the serum of four normal patients, the serum was extracted and subsequently processed. Using this procedure, T3 cross-reacted to the extent of 0.06%.
Relative cross-reactivities of several iodothyronines with 3,5-T2 antibody
| Compounds | Relative reactivity (%) |
|---|---|
| 3,5-T2 | 100 |
| 3,5-Br2T0 | <0.14 |
| T3 | 0.06 |
| T4 | 0.001 |
| rT3 | 0.001 |
| Tetrac | <0.0001 |
| Triac | 0.0006 |
| Diac | 0.13 |
| 3′,5′-T2 | 0.001 |
| 3,3′-T2 | 0.005 |
| 3-T1 | 0.65 |
| 3′-T1 | 0.0005 |
| DIT | <0.0001 |
| MIT | <0.0001 |
| Compounds | Relative reactivity (%) |
|---|---|
| 3,5-T2 | 100 |
| 3,5-Br2T0 | <0.14 |
| T3 | 0.06 |
| T4 | 0.001 |
| rT3 | 0.001 |
| Tetrac | <0.0001 |
| Triac | 0.0006 |
| Diac | 0.13 |
| 3′,5′-T2 | 0.001 |
| 3,3′-T2 | 0.005 |
| 3-T1 | 0.65 |
| 3′-T1 | 0.0005 |
| DIT | <0.0001 |
| MIT | <0.0001 |
Tetrac, Tetraiodothyroacetic acid; Triac, triiodothyroacetic acid; Diac, diiodothyroacetic acid; DIT, diiodotyrosine; MIT, monoiodotyrosine.
Table 2 shows the effects of several drugs used for the treatment of intensive care patients on the binding of 3-Br-5-[125I]T1 to its specific antibody. Four of the 12 drugs tested showed no detectable effects. The highest cross-reactivities were measured for imipinem, vancomycin, and norepinephrine (1.3 × 10−7, 1.2 × 10−7, and 7.4 × 10−7, respectively). The cross-reactivities of all other drugs were between 1–3 orders of magnitude lower.
Cross-reactivity of various drugs with antiserum to 3,5-T2
| Drugs | Cross-reactivity (by wt) |
|---|---|
| Imipinem | 1.3 × 10−7 |
| Cefotaxime | 1 × 10−10 |
| Vancomycine | 1.2 × 10−7 |
| Dopamine | 3.7 × 10−8 |
| Norepinephrine | 7.4 × 10−7 |
| Dobutamine | 1 × 10−10 |
| Fentanyl | 1 × 10−10 |
| Furosemide | 1.8 × 10−8 |
| Propofol | 1.2 × 10−8 |
| Norcuron | 1 × 10−10 |
| Heparin | a |
| Drugs | Cross-reactivity (by wt) |
|---|---|
| Imipinem | 1.3 × 10−7 |
| Cefotaxime | 1 × 10−10 |
| Vancomycine | 1.2 × 10−7 |
| Dopamine | 3.7 × 10−8 |
| Norepinephrine | 7.4 × 10−7 |
| Dobutamine | 1 × 10−10 |
| Fentanyl | 1 × 10−10 |
| Furosemide | 1.8 × 10−8 |
| Propofol | 1.2 × 10−8 |
| Norcuron | 1 × 10−10 |
| Heparin | a |
Heparin (0.25–100 IU/assay tube) did not affect 3-Br-5-[125I]T1 binding to antiserum.
The mean (±sd) recoveries from 2.5, 5, and 10 pmol/L nonlabeled 3,5-T2 added to pooled sera obtained from euthyroid patients were 98.5 ± 7.3%, 96 ± 9.6%, and 101 ± 8.1%, respectively.
Serum concentrations of 3,5-T2 in patients and controls
The mean concentration of 3,5-T2 in the group of healthy controls (n = 62) was 16.2 ± 6.4 pmol/L. Linear regression analysis conducted to investigate the effects of age and sex on the variable hormone revealed no significant age effect. Therefore, in all further calculations the serum levels of 3,5-T2 in all healthy controls were compared to those in the respective group of the patient sample.
Figure 2 shows the full results for the serum concentrations of 3,5-T2 in the controls and various patient groups. The serum levels of 3,5-T2 were significantly enhanced in patients with hyperthyroidism (P = 0.009), sepsis (P = 0.004), liver diseases (P = 0.004), head injury (P = 0.015), and brain tumors (P = 0.003). Four of the 9 patients with hyperthyroidism, 8 of the 24 patients with sepsis, 6 of the 22 patients with liver diseases, and 4 of the 15 patients with head injury had serum 3,5-T2 concentrations above the normal range. In hypothyroid patients, serum 3,5-T2 levels were significantly reduced (P = 0.001).
Serum concentrations of 3,5-T2 in healthy controls and patients with thyroidal and nonthyroidal illnesses. Values are the mean ± sd. The hatched area indicates the normal range (3.4–29 pmol/L).
Serum concentrations of 3,5-T2 in healthy controls and patients with thyroidal and nonthyroidal illnesses. Values are the mean ± sd. The hatched area indicates the normal range (3.4–29 pmol/L).
The T4, fT4, T3, rT3, and TSH concentrations in the six patient groups and the healthy control group are listed in Table 3. All four patient groups with nonthyroidal illnesses had significantly reduced serum levels of T3. The lowest serum T3 levels were seen in patients with brain tumors, some of whose T3 and T4 concentrations were no longer measurable. The T3 measurements in the patients with brain tumors were, therefore, repeated in another assay, but the results showed serum concentrations as low as those determined in the first assay. Pearson’s coefficients of correlation calculated between the serum concentrations of 3,5-T2 and T3 separately for the four patient groups were not significant (r = 0.23 for patients with sepsis, r = −0.22 for liver disease, r = 0.11 for brain injury, and r = −0.34 for brain tumors; all not significant). The coefficients of correlation between T3 and rT3 were also not significant (data not shown).
Serum concentrations of different iodothyronines and TSH in patients with thyroid disorders and various nonthyroidal illnesses
| T4 (nmol/L) | fT4 (pmol/L) | T3 (nmol/L) | rT3 (nmol/L) | TSH (mU/L) | |
|---|---|---|---|---|---|
| Hyperthyroidism (n = 9) | 178 ± 321 | 58.7 ± 29.51 | NM | NM | <0.011 |
| Hypothyroidism (n = 8) | 44 ± 261 | 3.8 ± 3.31 | NM | NM | 37 ± 341 |
| Sepsis (n = 24) | 83 ± 30 | 15.1 ± 6.7 | 0.74 ± 0.371 | 2.13 ± 1.131 | 0.91 ± 1.12 |
| Liver diseases (n = 23) | 54 ± 51 | NM | 0.51 ± 0.181 | 1.73 ± 1.461 | 1.50 ± 1.40 |
| Head trauma (n = 15) | 90 ± 21 | 14.2 ± 4.7 | 1.35 ± 0.912 | 1.00 ± 0.652 | 1.40 ± 2.13 |
| Brain tumors (n = 21) | 6.7 ± 9.81 | NM | 0.18 ± 0.221 | 2.14 ± 1.031 | 0.02 ± 0.021 |
| Healthy controls (n = 62) | 94 ± 15 | 12.5 ± 2.0 | 2.01 ± 0.32 | 0.54 ± 0.13 | 1.21 ± 0.60 |
| Normal range | 64–124 | 8.5–16.5 | 1.37–2.65 | 0.28–0.80 | 0.10–2.41 |
| T4 (nmol/L) | fT4 (pmol/L) | T3 (nmol/L) | rT3 (nmol/L) | TSH (mU/L) | |
|---|---|---|---|---|---|
| Hyperthyroidism (n = 9) | 178 ± 321 | 58.7 ± 29.51 | NM | NM | <0.011 |
| Hypothyroidism (n = 8) | 44 ± 261 | 3.8 ± 3.31 | NM | NM | 37 ± 341 |
| Sepsis (n = 24) | 83 ± 30 | 15.1 ± 6.7 | 0.74 ± 0.371 | 2.13 ± 1.131 | 0.91 ± 1.12 |
| Liver diseases (n = 23) | 54 ± 51 | NM | 0.51 ± 0.181 | 1.73 ± 1.461 | 1.50 ± 1.40 |
| Head trauma (n = 15) | 90 ± 21 | 14.2 ± 4.7 | 1.35 ± 0.912 | 1.00 ± 0.652 | 1.40 ± 2.13 |
| Brain tumors (n = 21) | 6.7 ± 9.81 | NM | 0.18 ± 0.221 | 2.14 ± 1.031 | 0.02 ± 0.021 |
| Healthy controls (n = 62) | 94 ± 15 | 12.5 ± 2.0 | 2.01 ± 0.32 | 0.54 ± 0.13 | 1.21 ± 0.60 |
| Normal range | 64–124 | 8.5–16.5 | 1.37–2.65 | 0.28–0.80 | 0.10–2.41 |
NM, Not measured.
P < 0.01.
P < 0.05.
Tissue concentrations of 3,5-T2 in brain tumors and controls
Tissue concentrations of 3,5-T2 in different parts of the brain and pituitary glands obtained from five human donors at autopsy as well as in the temporal and occipital cortexes obtained during neurosurgery are shown in Fig. 3B. 3,5-T2 was detectable in all tissue samples; the concentrations ranged between 70–150 fmol/g and were markedly similar in the different regions of the brain and pituitary gland. The concentrations of T3 and T4, also measured in these tissue samples, ranged between 1–3 pmol/g (Fig. 3A). The mean tissue concentration ratios of T3 to 3,5-T2 in different brain regions ranged from 14:1 to 23:1 (Fig. 3C).
T4, T3 (A), and 3,5-T2 (B) concentrations and molar ratios of T3/3,5-T2 (C) in samples from various areas of the adult human brain and pituitary obtained either at autopsy or intraoperatively (*). T4, T3 (D), and 3,5-T2 (E) concentrations and molar ratios of T3/3,5-T2 (F) in tissue samples from different forms of human brain tumors, metastases, and a pituitary adenoma were determined. Crb, Cerebellum; Cf, cortex frontalis; Str, striatum; Pit, pituitary; Cx, cortex cerebralis; Met, metastases; Ast, astrocytomas; Glb, glioblastomas; Men, meningiomas; Pia, pituitary adenoma.
T4, T3 (A), and 3,5-T2 (B) concentrations and molar ratios of T3/3,5-T2 (C) in samples from various areas of the adult human brain and pituitary obtained either at autopsy or intraoperatively (*). T4, T3 (D), and 3,5-T2 (E) concentrations and molar ratios of T3/3,5-T2 (F) in tissue samples from different forms of human brain tumors, metastases, and a pituitary adenoma were determined. Crb, Cerebellum; Cf, cortex frontalis; Str, striatum; Pit, pituitary; Cx, cortex cerebralis; Met, metastases; Ast, astrocytomas; Glb, glioblastomas; Men, meningiomas; Pia, pituitary adenoma.
The concentrations of 3,5-T2 in the various brain tumors are shown in Fig. 3E. The mean 3,5-T2 concentrations over all 21 samples was 336 ± 223 fmol/g. The highest mean concentration was found in the metastases (427 ± 269 fmol/g), followed by the glioblastomas (403 ± 264 fmol/g) and the astrocytomas (283 ± 149 fmol/g). The lowest values were measured in the meningiomas (197 ± 148 fmol/g). The level of 3,5-T2 in the pituitary adenoma was 304 fmol/g. Thus, these values were considerably higher than the 3,5-T2 concentrations measured in the different nontumoral regions of the human brain (Fig. 3B).
The mean tissue concentration of T3 in all 21 tumors was 345 ± 203 fmol/g (detailed data are shown in Fig. 3D). The ratios of T3 to 3,5-T2 in the tumor tissues were remarkable lower than those in nontumoral tissues, ranging from 0.9:1 in metastases to 1.5:1 in astrocytomas (Fig. 3F).
Discussion
The main finding of the present study is the elevation of serum concentrations of 3,5-T2 in four patient groups with widely differing nonthyroidal illnesses. To our knowledge, serum levels of 3,5-T2 have to date been measured in nonthyroidal illness in only two studies, both of which reported subnormal levels of this hormone in patients with liver cirrhosis (13, 15). However, these studies also reported normal concentrations of 3,5-T2 in hyperthyroidism (13) and hypothyroidism (13, 15). Here we report serum 3,5-T2 levels (16.2 ± 6.4 pmol/L) between 5- and 12-fold lower than those previously reported (76 and 190 pmol/L) (1, 12–18). The possibility cannot therefore be excluded that accurate and reliable measurement of serum levels of 3,5-T2 was not possible with some of the RIAs available in the early 1980s.
Most of the patients were receiving a variety of different medications for treatment of their underlying disorders. Many of these medications, such as heparin, norepinephrine, and dopamine, have previously been reported to affect serum thyroid hormone levels (23, 24, 30). However, these drugs seem to affect serum levels of thyroid hormones in very different ways (23, 24, 30). It would, therefore, seem highly unlikely that the elevated serum levels of 3,5-T2 found in all four diagnostic groups were predominantly a drug-induced phenomenon. We also failed to find any relevant effects of the medication on our 3,5-T2 assay procedure. It would thus seem probable that our results were related to the low T3 syndrome seen in the four diagnostic groups investigated in our and many other studies (31–40) as well as in many other patients with nonthyroidal illnesses regardless of diagnosis and medication (19–21). With regard to the extraordinarily low serum levels of T3 and T4 in patients with brain tumors, the question arises of whether the anesthetics and muscle relaxant employed were responsible. However, T4 has a half-life of about 1 week. The fall in serum levels of T4 to less than one tenth of the values measured in healthy controls cannot therefore have taken place during anesthesia, but must have developed over a period of several weeks. In this connection it is interesting that the patients with brain tumors were certainly in a much better general state of health than those with sepsis or head trauma. It is, therefore, conceivable that the mental stress to which these patients had been subjected during the days and weeks before the operation may have been one factor contributing to the low T3 syndrome as well as the high serum levels of 3,5-T2.
As briefly outlined in the introduction above, it is as yet unclear why patients with nonthyroidal illnesses with markedly reduced serum levels of T3, sometimes over prolonged periods of time, still appear clinically euthyroid (19–21). Experiments investigating this issue have reasonably focused on the hypothesis that a change in intracellular thyroid hormone metabolism and/or function (e.g. enhanced T3 nuclear binding capacity) may compensate for the low serum levels of T3, thus maintaining the cell in a euthyroid condition. However, the few studies that have investigated this issue have yielded inconsistent results (20, 41–44). Alternatively, it may be considered that clinical euthyroidism is maintained by an as yet unknown active thyroid hormone metabolite other than T3. As mentioned in the introduction above, several groups have, independently of each other, shown that 3,5-T2 may have effects on mitochondria, such as an increase in oxygen consumption, respiratory rate, and cytochrome oxidase activity (4–10). However, most of these studies employed pharmacological doses of 3,5-T2. Whether the slight elevations in serum concentrations of 3,5-T2 seen in our patients with nonthyroidal illnesses are somehow involved in the maintenance of clinical euthyroidism in these patients remains to be investigated. The question is complicated by the fact that a direct comparison of the thyromimetic potencies of T3 and 3,5-T2 is not yet feasible, as the two hormones seem to act at different receptors located at different intracellular loci, T3 at its nuclear receptors and 3,5-T2, if at all, at putative mitochondrial receptors. It may be argued that, although significant, the increases in serum concentrations of 3,5-T2 measured in our study were rather small and therefore of questionable physiological relevance. If one considers, however, that the most likely origin of the 3,5-T2 measured in the serum is intracellular conversion from T3 in peripheral tissues, then it may well be that the increase in the intracellular production and function of this hormone is much greater than evidenced by measurement of the fraction that is transported into the blood circulation. We found it particularly striking, that serum 3,5-T2 concentrations were elevated in patients with brain tumors, whose T3 and T4 concentrations were hardly measurable at all.
Several causes have been discussed as the origin of the low T3 syndrome in nonthyroidal illness (21); an inhibition of 5′-deiodinase type I activity is generally regarded as the most likely explanation (19–21). To our knowledge, however, direct evidence of an inhibition of the conversion of T4 to T3 in nonthyroidal illness has yet to be furnished. Even studies in which the increases in serum concentrations of T3 in patients with nonthyroidal illness after the administration of T4 were slightly lower than those determined in healthy controls (37) cannot exclude the possibility that this may have been due to enhanced metabolization of T3, rather than to a fall in the production of T3. Interestingly, an enhanced metabolization of T3 to T3 sulfate may indeed occur in nonthyroidal illness, as recently reported by two study groups (45, 46). If the elevated serum levels of 3,5-T2 found in our patients with nonthyroidal illness are also taken into account, the possibility that the low levels of T3 seen in nonthyroidal illnesses may be at least partly due to an increase in the metabolization of T3 to its various metabolites should be considered.
Horst et al. (11) reported that 3,5-T2 may suppress serum TSH in euthyroid rats. Our finding that 3,5-T2 was elevated in nonthyroidal illness and the detection of 3,5-T2 in human pituitary glands postmortem could, therefore, be of relevance for the as yet unexplained finding that TSH concentrations remain normal (or are even decreased) in these patients despite marked and prolonged reductions in their serum levels of T3 and sometimes also of T4.
Another open question concerns the presently unknown origin of 3,5-T2. The results of several early studies all showed that this hormone is probably produced by peripheral deiodination from circulating T3, whereas production controlled by the thyroid gland is extremely unlikely (13, 14). We do not know of any study published to date that has investigated which type of deiodinase catalyzes the reaction by which T3 is converted to 3,5-T2. We have, however, been able to measure 3,5-T2 in different regions of the human central nervous system and the pituitary gland. Although exact comparison between concentration units of serum (picomoles per L) and brain tissue (femtomoles per g) is not possible, a rough estimate reveals that the brain concentrations of 3,5-T2 are many times higher than the serum concentrations of this hormone. This would, again, argue in favor of a local production of the hormone in the human central nervous system.
It is noteworthy that in normal human brain, T3 concentrations are almost as high as T4 concentrations. In contrast, in human serum T4 levels are approximately 50 times higher than T3 concentrations. Therefore, the question arises of whether the relevance of thyroid hormone function in the brain might have been somewhat underestimated in the past.
It is not yet possible to give an interpretation of the origin and possible functional relevance of the surprisingly high concentrations of 3,5-T2 in brain tumors. Future studies should clarify whether high levels of 3,5-T2 are specific to tumor tissue or whether they also occur in other diseased tissue, e.g. in the livers of cirrhotic patients, etc. It also remains to be determined whether these high levels of 3,5-T2 are local accumulations of the hormone originating from the serum. If 3,5-T2 does, in fact, enhance energy production at the mitochondria (see above), then the investigation of a possible role of enhanced production of this hormone in tumor tissue would seem particularly interesting as, to our knowledge, the exact causes of the high energy consumption (cachexy) in cancer have not yet been fully elucidated.
