The Thyroid and Metabolic Rate

A person whose thyroid has been removed, and who takes only thyroxine (T4), still has the more active hormone, triiodothyronine (T3), in the blood. Where does it come from?

From the rest of the body. The thyroid secretes mostly T4, and tissues remove one iodine atom to make T3. This lesson follows the hormone from iodide to target cell, measures what it does, and uses feedback and a radioiodine scan to locate a fault. The first look at the axis is in endocrine system; paired hormone levels and assays are in endocrine axes and testing.

Iodide is trapped, then built into a protein

The thyroid is a gland of follicles: hollow balls of epithelial cells around a lumen of colloid, a protein gel made mostly of thyroglobulin. Each follicular cell faces blood on its basolateral side and colloid on its apical side. Synthesis runs from one face to the other.

1. Trapping. The sodium-iodide symporter (NIS) in the basolateral membrane moves iodide into the cell against its gradient, driven by the inward sodium gradient that the sodium-potassium pump maintains: secondary active transport (membrane transport). For rat NIS expressed in frog oocytes, Eskandari and colleagues (1997) measured two sodium ions per anion, one net positive charge in per cycle. NIS only traps iodide; it makes no hormone.
2. Iodination. Iodide crosses into the colloid (pendrin is thought to be one route). At the apical surface, thyroperoxidase (TPO) uses hydrogen peroxide from dual oxidase (DUOX) to oxidize iodide and attach it to tyrosines of thyroglobulin, making monoiodotyrosine (MIT) and diiodotyrosine (DIT) residues. Fixing iodine onto protein is organification.
3. Coupling. Still on thyroglobulin, the iodinated ring of a donor residue is joined onto an acceptor DIT. DIT plus DIT gives T4; MIT plus DIT gives T3. The hormone stays stored in the colloid as part of the protein.
4. Release. Under TSH stimulation the cell takes colloid back in by endocytosis, the vesicles fuse with lysosomes, proteases digest thyroglobulin, and T4 and T3 leave through basolateral transporters such as MCT8.
5. Recycling. Iodotyrosine dehalogenase strips iodide from the freed MIT and DIT for reuse.

Coupling is not random: structure, mutagenesis and hormone-production assays (Coscia and colleagues, Nature, 2020; Adaixo and colleagues, Nature Communications, 2022) place it at particular exposed, flexible, nearby tyrosine pairs of human thyroglobulin. The gland’s parafollicular C cells secrete calcitonin, taken up in calcium, phosphate and bone.

The gland makes mostly T4; tissues make most T3

Thyroid hormones dissolve poorly in water and travel bound to thyroxine-binding globulin (TBG), transthyretin and albumin. In human blood about 99.97 percent of T4 and 99.70 percent of T3 is protein-bound. The free fraction, the unbound part, is what enters cells and what the pituitary senses. The bound pool is a reservoir, one reason the half-life of T4 in euthyroid people (normal thyroid function) is about 6.2 days, against about 1 day for T3.

T4 binds the receptor much more weakly than T3 and is best treated as a prohormone. Three deiodinases, enzymes that each remove one iodine, decide its fate:

• D1 and D2 remove an outer-ring iodine from T4 and make T3. This activates. D1 also takes iodine from the inner ring, so it makes some reverse T3 as well and clears it.
• D3 removes an inner-ring iodine, turning T4 into inactive reverse T3 and T3 into 3,3’-diiodothyronine (3,3’-T2). This inactivates.

The ring decides: on or off. And the hormone does not simply diffuse into cells; it enters through transporters such as MCT8.

The classic experiment asked whether all T3 comes from the gland. Lewis Braverman, Sidney Ingbar and Kenneth Sterling (Journal of Clinical Investigation, 1970) studied 13 patients treated with T4, all but two athyreotic (without a functioning thyroid) or hypothyroid. If the gland were the only source, the athyreotic patients should have had no T3. They gave radiolabelled T4 and separated the iodine-containing compounds of serum by chromatography, looking for labelled T3. Two controls mattered. A specially purified lot of labelled T4 contained less than 0.1 percent contaminating T3 and gave the same result, so the T3 found could not simply have been injected. In two additional patients T4 was given intravenously, and stable and labelled T3 still appeared, so conversion was not confined to the gut during absorption. The result ruled in conversion outside the thyroid, in humans. It did not identify the enzyme or organ, or measure how much T3 is made this way; later work is commonly summarized as roughly 80 percent in healthy people with enough iodine, an approximate estimate.

So the T3 inside a cell need not match the T3 in plasma. The design problem tests that idea in mice.

Thyroid hormone raises the metabolic rate and drives the heart

Inside cells, T3 binds thyroid hormone receptors, nuclear receptors that regulate transcription. In mice whose receptors could not bind DNA (Hönes and colleagues, PNAS, 2017), heart rate and body-temperature regulation were preserved, while skeletal development (TRα) and feedback (TRβ) were not: some actions require the receptor on DNA and some do not.

The best-known action is on basal metabolic rate, energy use by an awake, resting, fasting person in the thermoneutral zone (thermoregulation). How does more hormone become more heat?

The modern experiment measured it in human muscle. Lebon and colleagues in Gerald Shulman’s laboratory (Journal of Clinical Investigation, 2001) studied seven healthy adults before and after three days of T3. Carbon-13 nuclear magnetic resonance (NMR) spectroscopy estimated tricarboxylic acid (TCA) cycle flux in skeletal muscle, phosphorus-31 saturation-transfer NMR estimated phosphate-to-ATP flux as a measure of ATP synthesis, and indirect calorimetry measured whole-body energy expenditure. If T3 acted only by raising ATP demand, TCA flux and ATP synthesis should rise together. Instead TCA flux rose by about 70 percent while measured ATP synthesis did not increase. Energy expenditure rose by about 15 percent and heart rate by about 30 percent. The authors’ reading is mitochondrial uncoupling: more fuel oxidized per ATP made, the difference released as heat (oxidative phosphorylation). Uncoupling is one contributor among several: pumping by the sodium-potassium pump, calcium cycling by SERCA with sarcolipin, and futile substrate cycles also turn ATP use into heat under thyroid hormone, and which matters most is still argued. The study does not identify the leaking protein or apply beyond muscle, and the fluxes are model-based estimates in seven people over three days.

Thyroid hormone raises heart rate and contractile force and lowers systemic vascular resistance (SVR), the resistance of the whole systemic circulation to flow, so cardiac output rises (cardiac output and venous return; resistance, compliance and local flow). In one 2024 study (Suonsyrjä and colleagues), SVR was 18 percent lower in 20 still-hyperthyroid patients, measured a median of six days after starting antithyroid medication, than in 60 matched euthyroid controls. Since mean arterial pressure is roughly cardiac output times SVR, mean pressure may change little; in that study aortic pressures were similar, radial pulse pressure was about 5 mmHg higher and finger blood pressure was lower.

Development depends on it too: in humans, too little thyroid hormone around birth impairs growth of the skeleton and the brain, and many countries screen newborns for congenital hypothyroidism. Entry into cells matters: Dumitrescu, Refetoff and colleagues (American Journal of Human Genetics, 2004) found MCT8 mutations in two unrelated families whose affected boys had global developmental delay.

One loop, several lesions: reading TSH, free T4 and T3

Thyrotropin-releasing hormone (TRH) from the hypothalamus reaches the anterior pituitary through portal vessels (hypothalamus and autonomic control) and stimulates thyrotropes to release thyroid-stimulating hormone (TSH). TSH acts through a G protein-coupled receptor on follicular cells (G proteins) to raise synthesis and release, and over weeks it enlarges the gland. T3 inhibits TRH and TSH release; much of the T3 acting in the pituitary is thought to be made there from T4 by D2, so the pituitary reads free T4 too.

TSH changes steeply with small changes in free hormone, so it is usually the most sensitive single readout of a problem in the gland itself. Read the pair. Hormone low and TSH high: the gland is failing and the pituitary is responding. Hormone low and TSH not high: this suggests a problem at the pituitary or hypothalamus.

State (untreated, in humans; overt disease unless noted)	TSH	Free T4	T3
Graves disease	Suppressed	High	High, sometimes rising before free T4
Toxic adenoma	Suppressed	High	High
Thyroiditis (early, thyrotoxic phase)	Suppressed	High	High or high-normal; raised less than T4
Primary hypothyroidism	High	Low	Can stay normal; may fall when severe
Secondary hypothyroidism (pituitary)	Low or inappropriately normal	Low	Often low or low-normal; not used to define it
Compensated iodine deficiency	Normal or high	Normal or low	Normal or high
TBG raised, at steady state	Normal	Normal	Free normal; total high

In subclinical disease TSH is abnormal while free T4 and T3 are still normal, and some thyrotoxic patients have a raised T3 with a normal free T4. In secondary hypothyroidism a normal TSH is itself abnormal, because it should be high when free T4 is low. In iodine deficiency T3 can be preserved: in 285 people from Nomane, an iodine-deficient region of New Guinea, Chopra and colleagues (1975) found a higher mean serum T3 than in an American comparison group, alongside a lower mean T4 and a raised mean TSH. That population was goitrous and under extra TSH drive, so it sits beyond the compensated row above.

State of the axis (untreated, in humans) Normal Graves disease (TSH-receptor antibody) Primary thyroid failure (primary hypothyroidism) Pituitary TSH failure (secondary hypothyroidism) Compensated iodine deficiency Thyroiditis (thyrotoxic phase) Toxic adenoma Show the radioiodine uptake scan TBG doubled (the worked problem; it assumes intact feedback, so it is offered for a normal axis)

TSH

Free T4

T3

In this sketch: the arrows are directions only, taken from the lesson’s table for untreated people, and carry no numbers. The only numbers are the lesson’s made-up worked problem (an adult at steady state with a total T4 of 80 nmol per L, 99.97 percent protein-bound, TBG carrying 75 percent of bound T4, unchanged carrier affinities). The dark scan panel, its brightness shades, the equal gland sizes and the outlined block are drawing choices made up for this sketch, and the scan shows a pattern, not a measurement.

The rule to keep: read TSH and free T4 as a pair. Low hormone with high TSH points to the gland; low hormone with TSH not high suggests the pituitary or hypothalamus; high hormone with suppressed TSH is thyrotoxicosis, and the uptake scan, read with the hormone pattern and never alone, tells high synthesis from leakage. In the worked problem, total T4 rose with the binding protein; free T4 is the level feedback defends.

Worked problem: a rise in binding protein

In this made-up example, an adult at steady state has a total plasma T4 of 80 nmol per L, 99.97 percent protein-bound. Estrogen, for example from an oral contraceptive, raises TBG; suppose TBG doubles. Assume intact feedback, unchanged carrier affinities, and TBG carrying 75 percent of bound T4 (a textbook approximation), transthyretin and albumin the other 25 percent.

1. Free T4 before. 80 nmol per L x 0.0003 = 0.024 nmol per L = 24 pmol per L.
2. The transient. The extra TBG takes up hormone, free T4 dips, TSH rises, and the gland releases more T4 until free T4 returns to the level feedback defends.
3. The new steady state. With free T4 back at 24 pmol per L, each carrier is as saturated as before, so bound hormone scales with carrier concentration. TBG-bound: 0.75 x 80 = 60 nmol per L, doubled to 120. Other carriers: 0.25 x 80 = 20 nmol per L, unchanged. Total = 120 + 20 = 140 nmol per L.
4. Answer. Total T4 rises 75 percent (140 / 80 = 1.75); free T4 and TSH end where they started.
5. How fast. Treat T4 as one well-mixed pool with first-order elimination and a 6.2-day half-life (euthyroid humans). After a step change in input, the fraction of the way to the new steady state after n half-lives is 1 - 2^(-n): 93.75 percent at n = 4 (24.8 days), 96.875 percent, about 97, at n = 5 (31.0 days). With no new input, 100 nmol of T4 falls to 25 nmol in 12.4 days.

free T4 = total T4 x free fraction; fraction of new steady state = 1 - 2^(-t / half-life)

T4-driven changes take about a month to settle, so an early measurement can mislead.

Radioiodine uptake asks the gland what it is doing

Radioiodine uptake asks a different question: give a tracer dose of radioactive iodine, commonly iodine-123, and measure hours later what fraction the gland has trapped; a scan shows where. Uptake reports how much tracer the gland still holds at the time of measurement, which depends on trapping by NIS, organification, the stable iodide supply and tracer loss. No single numerical normal range applies everywhere; it depends on tracer, timing, laboratory and local iodine intake. The patterns are what count:

• Graves disease: usually high uptake, even across both lobes; in a minority of cases it is normal or uneven.
• Toxic adenoma: one bright nodule, the rest faint because TSH is suppressed.
• Thyroiditis: low or absent uptake. Low uptake is not specific to it: thyroid hormone taken as a drug, and hormone made outside the gland, do the same.
• Iodine deficiency: high uptake, with little stable iodide competing with the tracer.
• A recent large iodine load, such as contrast dye: lower uptake, partly because stable iodide dilutes the tracer.

Test the mechanism with Graves disease. Autoantibodies bind and activate the TSH receptor; Faust and colleagues (Nature, 2022) solved structures of the receptor with such an antibody, showing how it can mimic the hormone. Predict: free T4 and T3 rise, T3 sometimes first; TSH falls to suppressed levels, yet stimulation continues because it never depended on TSH. The gland enlarges diffusely and uptake is usually high and diffuse. Downstream, thyrotoxicosis (excess thyroid hormone) raises metabolic rate, heat production, heart rate and cardiac output, and lowers SVR.

In the early thyrotoxic phase of thyroiditis, hormone is again high and TSH suppressed, but uptake is low: hormone is leaking from damaged follicles, not being made from new iodide. The scan is read with the hormone pattern, never alone.

Design problem

de Jesus and colleagues (Journal of Clinical Investigation, 2001) studied mice lacking type 2 deiodinase (Dio2-knockout). In the cold they became hypothermic because brown adipose tissue thermogenesis was impaired, yet plasma T3 was normal. A single injection of T3 reversed the hypothyroid-like abnormalities measured in their brown fat cells. What does this show about where active hormone is made, and what controls would you add? (Brown fat in adult humans is in skin and human heat.)

One way to do it

Plasma T3 was normal, yet the tissue behaved as if it lacked thyroid hormone. So brown fat relies on T3 made inside its own cells from T4 by D2, and plasma T3 does not fully report thyroid status in that tissue.

The T3 rescue supports a lack of T3 in the tissue rather than some function of D2 unrelated to making T3. A rescue confined to brown fat would test it more directly.

Controls: wild-type littermates in the same cold; both genotypes housed at thermoneutrality; plasma T4 and TSH as well as T3; T3 content of brown fat itself; a check that sympathetic drive to brown fat arrived normally, since cold-induced thermogenesis depends on noradrenaline; and a rescue dose too small to make the whole animal hyperthyroid.

Limits: this is mouse brown fat in the cold, not an adult human at rest, and a whole-body knockout lacks D2 everywhere, so not every phenotype belongs to brown fat.