Biosynthesis of 3-Iodothyronamine (T₁AM) Is Dependent on the Sodium-Iodide Symporter and Thyroperoxidase but Does Not Involve Extrathyroidal Metabolism of T₄

Abstract

3-Iodothyronamine (T₁AM) is an endogenous thyroid hormone derivative with unknown biosynthetic origins. Structural similarities have led to the hypothesis that T₁AM is an extrathyroidal metabolite of T₄. This study uses an isotope-labeled T₄ [heavy-T₄ (H-T₄)] that can be distinguished from endogenous T₄ by mass spectrometry, which allows metabolites to be identified based on the presence of this unique isotope signature. Endogenous T₁AM levels depend upon thyroid status and decrease upon induction of hypothyroidism. However, in hypothyroid mice replaced with H-T₄, the isotope-labeled H-T₃ metabolite is detected, but no isotope-labeled T₁AM is detected. These data suggest that T₁AM is not an extrathyroidal metabolite of T₄, yet is produced by a process that requires the same biosynthetic factors necessary for T₄ synthesis.

Thyroid hormone (TH) is synthesized in the thyroid gland predominantly as T₄. The first step in the biosynthesis of T₄ involves concentration of dietary iodide in the thyroid gland by the sodium-iodide symporter (NIS), which is followed by thyroperoxidase (TPO)-catalyzed iodination of tyrosine residues within the thyroglobulin protein. T₄ is a prohormone that is secreted from the thyroid gland and undergoes enzymatic outer ring deiodination to form T₃. T₃ is the active form of TH, and it binds to the TH nuclear receptors to regulate transcription of TH responsive genes (1). TH-mediated gene transcription affects a variety of physiological processes, including cardiac function, metabolism, bone turnover, body temperature, and cholesterol levels (2). Although most effects of TH are transcriptionally mediated, there are some rapid effects that are thought to be transcriptionally independent (3, 4).

In addition to outer ring deiodination, T₄ is deiodinated on the inner ring to form rT₃, a metabolite with unknown physiological significance. Inner ring deiodination of either T₄ or T₃ is thought to be a mechanism for deactivating TH (5), yet deiodination pathways sequentially produce diiodo- and monoiodothyronines (6, 7). Additional metabolic pathways of TH include conjugation with either a sulfate or a glucuronic acid (8). T₄ and T₃ can undergo oxidative deamination of the alanine side chain to form the thyroacetic acid derivatives, Tetrac and Triac, both of which have biological activity. Triac has thermogenic effects (9), whereas Tetrac inhibits T₄ actions initiated at the plasma membrane (10). This extensive metabolism of TH beyond direct activation and deactivation of T₃ indicates that TH metabolism is critical for regulating hormone action.

3-Iodothyronamine (T₁AM) is an endogenous derivative of TH present in rat, mouse, hamster, and human, with acute pharmacological effects (11–15). A single pharmacological dose of T₁AM results in hypothermia, bradycardia, a shift in the respiratory quotient, hyperglycemia, and hypoinsulinemia within minutes to a few hours (11, 13, 16). Structural similarities between T₄ and T₁AM, including the presence of iodine and the biaryl ether carbon skeletons (Fig. 1), suggest that T₁AM is a metabolite of T₄. T₁AM is sulfated and deiodinated in vitro and oxidatively deaminated in vivo (17–19), indicating that it is a substrate for the same metabolic pathways that act on TH. A hypothesized biosynthetic pathway for T₁AM involves deiodination and decarboxylation of TH. If decarboxylation of TH occurs, this would represent a previously uncharacterized branch of TH metabolism and signaling.

Fig. 1.

Structures of T₄ and T₁AM.

Several analytical methods have been developed to measure T₁AM levels, including an immunoassay and liquid chromatography-tandem mass spectrometry (LC-MS/MS) assays (12, 14, 20, 21). The advantage of LC-MS/MS over an immunoassay is that additional structural information is obtained. We have recently synthesized an isotope-labeled T₄, ¹³C₉-¹⁵N-T₄ (Hackenmueller, S. A., and T. S. Scanlan, Synthetic Communications, in press). ¹³C₉-¹⁵N-T₄ [heavy-T₄ (H-T₄)] is 10 atomic mass units heavier than unlabeled T₄, creating a unique mass signature that is detectable by mass spectrometry. This mass signature allows for H-T₄ and any metabolite arising from H-T₄ to be distinguishable from unlabeled T₄ and metabolites. In this study, we use the labeled H-T₄ to test the hypothesis that T₁AM is a metabolite of T₄.

Materials and Methods

Animals

Male C57Bl/6J mice, 8- to 10-wk-old, were housed four to five per cage in a climate-controlled room with a 12-h light, 12-h dark cycle. All animals had ad libitum access to food and water. Experimental protocols were in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Oregon Health and Science University Institutional Animal Care and Use Committee.

Chemicals

For use with animals, potassium perchlorate (KClO₄) was purchased from MP Biomedicals (Solon, OH). Methimazole (MMI) was purchased from Acros Organics (Pittsburgh, PA). T₄ was purchased from Sigma (St. Louis, MO), and H-T₄ was synthesized as described elsewhere (Hackenmueller, S. A., and T. S. Scanlan, Synthetic Communications, in press). Human serum deficient in T₄ and T₃ was purchased from Sigma.

Hormone manipulations

To determine the length of treatment necessary to induce hypothyroidism, mice were treated with 0.1% MMI and 0.2% KClO₄ in drinking water for 2, 4, or 6 wk (22). Blood was collected and allowed to clot on ice for at least 30 min before centrifugation at 6000 × g. Sera were removed and stored at −80 C until further use. Total serum T₄ was measured using a Coat-A-Count Total T₄ RIA kit (Siemens Healthcare Diagnostics, Deerfield, IL).

For T₄ replacement studies, mice were pretreated for 2 wk and subsequently maintained on MMI and KClO₄ in drinking water for the duration of hormone replacement. T₄ or H-T₄ was dissolved in 0.01 n NaOH in sterile saline (0.9% NaCl; Hospira, Lake Forest IL) and administered as once daily sc injections (23). T₄ was administered once daily for 2–5 wk to determine the time necessary to establish steady-state serum hormone levels. H-T₄ was injected once daily for 3 wk to reestablish euthyroid hormone levels. Serum was collected as described above, and tissues, including liver, heart, brain, and kidney, were frozen immediately on dry ice and stored at −80 C until used for hormone extraction. Total serum T₄ levels after hormone replacement were determined using the RIA kit described above.

T₁AM, T₃, and T₄ extraction from liver

T₁AM, T₃, and T₄ were extracted from mouse livers following a similar protocol to what has been previously described (21). Frozen livers were thawed, weighed, and homogenized in cold homogenization buffer [1 ml per 1 g of tissue, 154 mm NaCl, and 10 mm NaH₂PO₄ (pH 7.4)] with 30 passes in a Potter-Elvehjem tissue homogenizer attached to a Glas-Col motor drive. Homogenates were centrifuged at 4000 × g for 15 min. Supernatants were collected and transferred to clean 15-ml polypropylene tubes, and 2.5 pmol d₄-T₁AM, and 25 pmol each of ¹³C₆-T₃ and ¹³C₆-T₄ were added as internal standards. Additional 60 mg of NaCl were added per 1 ml of supernatant, and the samples were incubated at room temperature for 1 h. Acetone was added to each sample to precipitate proteins at a ratio of 2:1 acetone:supernatant volume, and samples were vortexed and left on ice for 30 min. Samples were centrifuged at 2000 × g for 15 min, and supernatants were evaporated in an Eppendorf Concentrator Plus 5301 evaporator (Eppendorf, Hamburg Germany) to remove acetone. Potassium acetate buffer [1 ml of 0.1 m (pH 4.0)] was added to each sample before solid phase extraction (SPE).

Bond-Elut Certify SPE cartridges (130 mg, 3 ml; Varian, Inc., Middleburg, The Netherlands) were preconditioned by gravity with 2 ml of methylene chloride/isopropanol (75:25 by volume), followed by 2 ml of methanol and 2 ml of potassium acetate buffer. Samples in potassium acetate buffer were added to the SPE cartridges, and the cartridges were washed sequentially with 3.5 ml of water, 1.6 ml of 0.1 m HCl, 7 ml of methanol, and 3.5 ml of methylene chloride/isopropanol (75:25). T₁AM was eluted with 2 ml of methylene chloride/isopropanol/ammonium hydroxide (70:26.5:3.5 by volume) in clean borosilicate glass tubes. Eluates were transferred to Eppendorf tubes and evaporated to dryness in the Eppendorf evaporator described previously. Dried residues were redissolved in 100 μl of 1:1 methanol:0.1 m HCl before injection for LC-MS/MS analysis.

LC-MS/MS analysis

Samples were analyzed by LC-MS/MS using previously established selected reaction monitoring method (21). A 200 series HPLC system (PerkinElmer, Boston, MA), constituted by a binary micropump system, a column oven, and an autosampler, was coupled to an API 4000 triple quadrupole mass spectrometer (Applied Biosystems-MDS Sciex, Concord, Ontario, Canada) with a Turbo-V IonSpray source operated in the positive mode. Samples were injected onto a Gemini C18 2 × 50 mm, 3-μm particle size column (Phenomenex, Torrance, CA). Buffer A (0.1% formic acid in water) and buffer B (0.1% formic acid in methanol/acetonitrile) (1:4) were used in the following gradient: 5–100% A from 0 to 5 min, 100% A from 5 to 6 min, and equilibrate 6 to 9 min. IonSpray voltage was 5.25 kV, gas source 1 was 70, gas source 2 was 55, turbo temperature was 650 C, and collision-activated dissociation gas pressure was 5.7 mPa. Declustering potential, collision energy, and collision exit potential (CXP) parameters for each selected reaction monitoring ion pair monitored are listed in Table 1.

Table 1.

MS/MS instrument parameters

Analyte	ID	Retention time (min)	Q1 (m/z)	Q3 (m/z)	CE	DP	CXP
T1AM	Known analyte	2.4	356	339	18	53	10
				212	28	53	18
				195	36	53	17
d4-T1AM	Internal standard	2.4	360	343	18	53	10
				216	28	53	18
				199	36	53	17
H-T1AM	Predicted metabolite	2.4	365	347	18	53	10
				220	28	53	18
				203	36	53	17
T3	Known analyte	3.4	652	606	29	103	17
				508	29	103	15
				479	47	103	14
13C6-T3	Internal standard	3.4	658	612	29	103	17
				514	29	103	15
				485	47	103	14
H-T3	Predicted metabolite	3.4	662	615	29	103	17
				517	29	103	15
				488	47	103	14
T4	Known analyte	3.6	778	732	33	106	20
				634	34	106	18
				605	51	106	16
13C6-T4	Internal standard	3.6	784	738	33	106	20
				640	34	106	18
				611	51	106	16
H-T4	Predicted metabolite	3.6	788	741	33	106	20
				643	34	106	18
				614	51	106	16

DP, Declustering potential; CE, collision energy; CEP, collision exit potential; ID, identity; Q, quadrupole.

Statistics

Data are presented as mean ± sem. Data were analyzed by one-way ANOVA to determine statistical significance.

Results

Hypothyroid and hormone replacement mouse model

Multiple reports in the literature use a combination of MMI and ClO₄⁻ to induce hypothyroidism over the course of 4–14 wk (22, 24, 25). Because the half-life of T₄ in mice is 13–18 h, we hypothesized that a period of treatment shorter than 4 wk may be sufficient to induce hypothyroidism (26). To test this, we conducted a time-course experiment with MMI and KClO₄ administered in drinking water and monitored biweekly thyroid status as indicated by total serum T₄. Wild-type euthyroid mice had serum T₄ levels of 2.52 ± 0.32 μg/dl before inducing hypothyroidism with MMI and KClO₄. After 2 wk of treatment with MMI and KClO₄, total serum T₄ levels were below the assay limit of detection (0.82 μg/dl). From these experiments, we concluded that 2 wk of MMI and KClO₄ treatment was sufficient to induce hypothyroidism. For future experiments involving hormone replacement, mice were pretreated with MMI and KClO₄ for 2 wk and maintained on MMI and KClO₄ in drinking water throughout the duration of the experiments to inhibit endogenous TH synthesis.

Next, a time-course experiment was conducted to determine the length of time necessary to reestablish steady-state serum T₄ in a hypothyroid mouse. Previous studies in the literature involving TH replacement in Pax8 knockout or Pax8/MCT8 double knockout mice use a dose of 20 ng/g daily, and thus, we used this as our initial replacement dose of T₄ (23, 27). At 2–5 wk of T₄ replacement, total serum T₄ levels were significantly decreased with respect to euthyroid mice (Fig. 2A) but were not significantly different from each other. Based on these data, T₄ serum levels reach steady-state within 2 wk, but the replacement dose of 20 ng/g was insufficient to reestablish euthyroid serum T₄ levels. A previous report in the literature found that 3 wk of hormone replacement were necessary for the thyroid axis to reach steady state, as measured by TSH (28), prompting us to use a 3-wk hormone replacement duration (27). Because a 20 ng/g T₄ replacement dose did not reestablish complete euthyroid serum T₄ levels, a dose-response study was conducted. Hypothyroid mice were hormone replaced for 3 wk with T₄ ranging in doses from 0–500 ng/g (Fig. 2B), which resulted in dose-dependent restoration of serum T₄. We selected 100 ng/g T₄ as a replacement dose for future studies, because this dose resulted in serum T₄ levels closest to those of the euthyroid control.

Fig. 2.

Establishment of a hypothyroid and TH replacement mouse model. Total serum T₄ levels were measured by RIA after treatment with 0.1% MMI and 0.2% KClO₄ and replacement with 20 ng/g T₄ for 2–5 wk (A) or replacement with 0–500 ng/g T₄ for 3 wk (B); n = 3–5 per group; *, P < 0.05; **, P < 0.01. ND, Not detectable.

H-T₄ replacement

In hypothyroid mice, replacement with 100 ng/g H-T₄ did not restore serum T₄ to euthyroid levels (data not shown). However, 200 ng/g H-T₄ resulted in euthyroid serum T₄ levels (2.79 ± 0.54 ng/dl) (Fig. 3A). Replacement with either 100 ng/g T₄ or 200 ng/g H-T₄ also normalized body weight after hypothyroid-induced weight loss (Fig. 3B). One interpretation of our results is that the T₄ antibody in the RIA kit binds less efficiently to isotope-labeled H-T₄ than to T₄, resulting in artificially lower T₄ serum concentrations in H-T₄-replaced mice. To determine whether the difference in replacement doses between T₄ and H-T₄ was due to isotope effects on antibody binding, the response of T₄ and H-T₄ in the RIA kit was investigated. T₄ and H-T₄ were spiked into T₄/T₃-deficient human serum and analyzed by RIA. There was no difference in detection of H-T₄ from T₄ by the T₄ antibody from 0 to 12.5 μg/dl (data not shown). This suggests that the difference in doses is not due to the presence of ¹³C isotopes but is likely due to a lack of bioequivalence between the T₄ and H-T₄ formulations, an issue that has been documented for different T₄ preparations (29).

Fig. 3.

H-T₄ replacement in a hypothyroid mouse model. A, Total serum T₄ levels as measured by RIA after 2 wk of treatment with 0.1% MMI and 0.2% KClO₄ and replacement with T₄ or H-T₄; n = 4–5 per group. B, Body weights of euthyroid, hypothyroid, T₄, and H-T₄-replaced mice. Hypothyroid, T₄ and H-T₄-replaced mice were treated with 0.1% MMI and 0.2% KClO₄ starting on d 1 and continuing through d 36. At d 15, replaced mice began receiving daily injections of either T₄ or H-T₄; n = 4–8 per group. ND, Not detectable.

Endogenous T₁AM levels in mouse liver

To test whether T₁AM biosynthetic origins are related to thyroid status, we measured endogenous T₁AM after induction of hypothyroidism. To date, the distribution of endogenous T₁AM in mouse tissues has not been reported. The distribution of T₁AM in rat tissues, where the greatest concentration of T₁AM was present in liver, was used as a guide (21). For these studies, multiple tissues were initially analyzed to determine in which tissues T₁AM is present, but ultimately, liver was used for analysis because this tissue contained the greatest concentration of T₁AM in the mouse tissues analyzed. Detection of T₁AM from other mouse tissues was not reproducible and thus not included in this analysis. Similar to the data reported for rat serum (21), mouse serum T₁AM is lower than liver T₁AM. T₄ and T₃ were monitored as experimental controls, because T₄ and T₃ levels are known to decrease in hypothyroidism. Endogenous T₁AM from liver homogenates was analyzed by LC-MS/MS, and the chromatographic trace corresponding to the mass to charge ratio (m/z) transition of 356.2 to 339.1 shows a decrease in T₁AM peak intensity after induction of hypothyroidism for 2 wk (Fig. 4). Ionization is sample dependent, and each sample is normalized by integrating peak areas with respect to the internal standards. As expected, T₄ and T₃ also decrease by over 90% upon induction of hypothyroidism. Due to sample-dependent variations in ionization efficiencies, to quantify T₄, T₃, and T₁AM, peak areas were integrated with respect to the corresponding internal standard (d₄-T₁AM, ¹³C₆-T₃, or ¹³C₆-T₄). We integrated the peak areas of T₁AM and d₄-T₁AM to determine the concentration of endogenous T₁AM and found that liver T₁AM concentration decreases time dependently with induction of hypothyroidism (Fig. 5).

Fig. 4.

Representative LC-MS/MS chromatogram for T₁AM, T₃, and T₄ from euthyroid and 2-wk hypothyroid mouse livers. The m/z transitions represented are: T₁AM, 356 to 339; d₄-T₁AM, 360 to 343; T₃, 652 to 606; ¹³C₆-T₃, 658 to 612; T₄, 778 to 732; and ¹³C₆-T₄, 784 to 738. Solid black line, Endogenous analyte; dashed line, internal standard. Representative traces of three samples analyzed for each group. cps, Counts per second.

Fig. 5.

T₁AM levels decrease in mouse liver as a function of time of 0.1% MMI and 0.2% KClO₄ treatment, as measured by LC-MS/MS; n = 3–5 per time point.

Metabolism of exogenous T₄ and H-T₄

We next investigated whether T₁AM is an extrathyroidal metabolite of T₄ by measuring T₁AM in hypothyroid mice that were hormone replaced with either T₄ or H-T₄. Homogenized livers were analyzed by LC-MS/MS for the presence of unlabeled and labeled (H-T₁AM) forms of T₁AM. T₄, T₃, and the corresponding heavy-labeled forms, H-T₄ and H-T₃, served as positive controls, because T₃ is a known metabolite of T₄. Representative chromatographic traces for T₄ and H-T₄-replaced mouse livers are shown in Fig. 6. The retention times of internal standards are used to additionally verify analytes, because the internal standards and corresponding unlabeled or heavy analytes coelute. In hypothyroid mice, replacement with T₄ results in increased peak intensities for T₄ and T₃ relative to hypothyroid levels but no increase in T₁AM peak intensity (Fig. 6). Replacement with H-T₄ results in the detection of chromatographic peaks corresponding to H-T₄ and H-T₃ but no detectable chromatographic peak corresponding to H-T₁AM.

Fig. 6.

Representative LC-MS/MS chromatogram for T₁AM, T₃, and T₄ from T₄ and H-T₄-replaced mouse livers. The m/z transitions represented are: T₁AM, 356 to 339; d₄-T₁AM, 360 to 343; H-T₁AM, 365 to 347; T₃, 652 to 606; ¹³C₆-T₃, 658 to 612; H-T₃, 662 to 615; T₄, 778 to 732; ¹³C₆-T₄, 784 to 738; and H-T₄, 788 to 741. Solid black line, Unlabeled analyte (endogenous analyte for T₁AM and H-T₄-replaced group, and metabolite of exogenous T₄ in T₄-replaced group); dashed line, internal standard; gray line, heavy metabolite from H-T₄. cps, Counts per second.

To quantify liver concentrations of unlabeled and labeled T₄, T₃, and T₁AM, we integrated peak areas with respect to the corresponding internal standard in euthyroid, 2-wk hypothyroid, T₄-replaced, and H-T₄-replaced mice (Fig. 7). Inducing hypothyroidism decreases endogenous T₄ and T₃ 92 and 93.6%, respectively (52.0 ± 16.0 to 4.1 ± 3.5 pmol/g for T₄ and 25.0 ± 5.5 to 1.6 ± 0.7 pmol/g for T₃). Replacement with T₄ increases liver T₄ to 91% and liver T₃ to 161.6% of euthyroid liver concentrations. Replacement with H-T₄ results in liver H-T₄ and H-T₃ of 46% (24.0 ± 8.2 pmol/g) and 114.4% (28.6 ± 6.8 pmol/g), respectively, whereas endogenous, unlabeled T₄ and T₃ remain decreased, at 4.2 ± 2.3 and 1.6 ± 0.3 pmol/g. Endogenous, unlabeled liver T₁AM concentration decreases with hypothyroidism, from 2.4 ± 0.4 to 0.8 ± 0.6 pmol/g, and remains decreased upon replacement with either T₄ (0.4 ± 0.1 pmol/g) or H-T₄ (0.2 ± 0.1 pmol/g). No H-T₁AM was detected from any mouse liver replaced with H-T₄. Although liver T₃ levels are increased after replacement with T₄, this is unlikely to impact the final results, because T₄ levels remain normal, and a lack of T₁AM formation is also observed upon replacement with H-T₄, which results in normal liver H-T₃ levels.

Fig. 7.

Liver T₄, T₃, T₁AM and H-T₄, H-T₃, and H-T₁AM levels in euthyroid, hypothyroid, and hypothyroid mice with T₄ or H-T₄ replacement, as measured by LC-MS/MS. Inducing hypothyroidism for 2 wk decreases endogenous T₄, T₃, and T₁AM. Replacement with T₄ results in normal T₄ and increased T₃ with respect to euthyroid but no increase in T₁AM with respect to hypothyroid. Replacement with H-T₄ results in measurable H-T₄ and H-T₃, no increase in endogenous T₄, T₃ or T₁AM, and no detectable H-T₁AM. H-T₄ and H-T₃ were only detected in mice replaced with H-T₄; n = 3 for euthyroid, hypothyroid, and H-T₄ replaced; n = 2 for T₄ replaced.

Discussion

T₁AM is a recently discovered endogenous compound that contains the unique chemical signature of a TH, namely a biaryl ether core structure containing an iodine substituent. Like the TH T₄ and T₃, T₁AM is found in various tissues and in circulation (21), and circulating T₁AM is largely bound to lipoprotein particles (30). Recent advances in the quantitative analysis of T₁AM using methods based on LC-MS/MS and T₁AM-specific immunoassay suggest that endogenous T₁AM is present in tissues (in the rat) and circulation (in humans) at levels that are comparable with those of total T₄ (14, 21). In several rat tissues, including heart, liver, kidney, skeletal muscle, and stomach, T₁AM is present at a greater concentration than T₃, the major active metabolite of T₄ (21). Although T₁AM has been detected previously in mice, the distribution and quantitative analysis of T₁AM in mice has not yet been studied with these improved methods.

Since its discovery, it has been proposed, if not assumed, that T₁AM is produced as a metabolite of T₄ by deiodination and decarboxylation, the hypothetical enzymatic steps required for conversion of T₄ to T₁AM. In support of this notion, plausible iodothyronamine deiodination pathways have been demonstrated in vitro (19). In addition, analysis of human serum samples showed higher levels of T₁AM in samples from thyroid cancer patients receiving elevated hormone replacement doses of T₄ to suppress TSH compared with samples from euthyroid control subjects [see Hoefig et al. (14)]. However, in this same study, no correlation was found between T₁AM levels and either T₄ or TSH levels, contrary to expectations were T₄ to be a biosynthetic precursor of T₁AM. Moreover, iodothyronine decarboxylation to iodothyronamines has not been demonstrated, and the well-studied aromatic amino acid decarboxylase, a somewhat nonspecific enzyme with respect to aromatic amino acid substrates, was recently shown to be unable to catalyze iodothyronine decarboxylation (31), suggesting that the enzymatic conversion of an iodothyronine to an iodothyronamine is not as straightforward as was originally thought. We have attempted here to address this question directly with a labeling experiment using a stable isotope-labeled form of T₄ (H-T₄) designed to afford a correspondingly labeled T₁AM if T₁AM were to arise from enzymatic processing of extrathyroidal T₄. Detecting the H-T₄ and its metabolites is done using LC-MS/MS, thus avoiding the use of radioactivity. Moreover, the labeled atoms of H-T₄ reside in the thyronine skeleton and are not lost upon deiodination, as would be the case with commercially available outer ring [¹²⁵I]-T₄. Of note, a study designed similarly to ours was recently attempted and failed, because the investigators were unable to extract and detect endogenous T₁AM by LC-MS/MS (32).

Our results show that endogenous liver T₁AM concentration decreases significantly in hypothyroid vs. euthyroid mice, indicating a role for a functioning thyroid gland in the biosynthesis of T₁AM. However, hypothyroid mice treated with H-T₄ did not generate detectable levels of H-T₁AM, whereas H-T₃ was unequivocally detected. Taken together, these results indicate that T₁AM biosynthesis can be inhibited by the antithyroid drugs MMI and KClO₄ that target thyroid gland function by inhibiting TPO and the NIS, respectively (33, 34), yet does not appear to involve extrathyroidal deiodination and decarboxylation of T₄, the previously favored and often hypothesized route of production of T₁AM. The NIS and TPO targets of these antithyroid drugs implicate iodide transport and thyronine biosynthetic machinery in T₁AM production that is generally regarded as being specific to the thyroid gland. These unexpected results lead to the conclusion that like T₄, T₁AM is either a direct product of the thyroid gland, or a product of another cell population that expresses NIS and TPO or homologs in functional form, or is a metabolite of another thyroid gland product that does not arise from extrathyroidal metabolism of T₄.

At least three hypothetical thyroid gland-dependent pathways for T₁AM biosynthesis can be considered, all of which are dependent upon thyroglobulin, the substrate for T₄ biosynthesis. The first of these three involves extrathyroidal coupling of a phenol to monoiodotyrosine, a known by-product of thyroglobulin proteolysis that occurs in the T₄ secretion process (35). The outer ring phenol could theoretically derive from tyrosine. However, the chemistry required for this putative coupling reaction is not presently known. The second possibility is that T₁AM is secreted directly from the thyroid gland as a direct cleavage product of thyroglobulin. This putative pathway would begin with a thyroglobulin-bound 3-iodothyronine residue arising from TPO-mediated cross-coupling of monoiodotyrosine and tyrosine residues within thyroglobulin. Release of T₁AM would then require an unknown proteolytic decarboxylation reaction or proteolysis followed by specific 3-iodothyronine decarboxylation occurring within the thyrocyte. The final possibility is that T₁AM is a metabolite of a non-T₄ thyroid gland product or a metabolite of T₄ that is specifically generated within the thyroid gland. One potential candidate for such a T₁AM precursor could be 3,5-diiodothyronine (3,5-T₂) (36). To date, there is no reported direct evidence for outer ring deiodination of T₃ in vivo to produce 3,5-T₂ (20, 37), leaving open the possibility that like T₁AM, 3,5-T₂ is not an extrathyroidal metabolite of T₄. If 3,5-T₂ is not a simple deiodination product of circulating T₄, then it most likely is produced from thyroglobulin-bound 3,5-T₂ arising from TPO-catalyzed cross-coupling of 3,5-diiodotyrosine and tyrosine residues of thyroglobulin, respectively. Alternatively, if free T₄ within the thyroid gland were converted to some other derivative via a reversible thyroid gland-specific biochemical process (e.g. N-methylation or acetylation), and this putative derivative was the direct biosynthetic precursor of T₁AM, then a lack of labeled T₁AM arising from labeled extrathyroidal T₄ might also be expected to occur.

Although often thought of as thyroid gland specific, the dependence of T₁AM synthesis on TPO and NIS does not preclude the involvement of an extrathyroidal de novo biosynthetic pathway. The expression of the NIS gene has been detected in numerous cultured skin cells (38), and both TPO and NIS mRNA expression has been detected in human skin samples (39). Additionally, a recent report documents that human fibrocytes express thyroglobulin and are able to take up and incorporate [¹²⁵I] into the protein (40). The expression of these proteins in a cell population outside the thyroid gland creates the possibility that T₁AM biosynthesis originates through a de novo pathway involving extrathyroidal iodination and cross-coupling of tyrosine residues.

Future studies are needed to distinguish between these possible routes or to discover alternative pathways from those proposed above. This work also points to a compelling need to reassess the unique chemistry that occurs within the thyroid gland. The powerful bioanalytical tools available today may reveal new chemical insights into the function of this important endocrine organ.