We All Know We Need Them,We Hope They Are Coming,But When?

No cell or tissue is the same when it comes to responding to thyroid hormone (TH). The action of TH requires the coordinated function of many different types of signaling proteins, and cells vary greatly in the expression levels and types of these proteins. For example, TH molecules require transmembrane transporters to gain access to the intracellular milieu. Once inside, TH molecules can be modified by deiodinases, enzymes that can both activate or terminate TH actions. Finally, transducing the presence of TH into a cellular response requires TH receptors (TRs) and an array of co-modulators. Therefore, TH signaling is unique for each cell type (tissue or organ), depending on circulating TH levels and on the exclusive blend of transporters, deiodinases, and TRs present in each cell (1).

Although fascinating, this level of complexity might create challenges in disease states, for example, in carriers of genetic mutations that disrupt the function of one or more TH signaling proteins, or when the thyroid gland itself is ill or has been removed, rendering patients hypothyroid. In both instances, treatment involves restoring TH signaling using serum thyrotropin (TSH) and TH levels as guidance, but—in many cases—success is difficult to achieve.

A success story is the utilization of serum TSH in the diagnosis of hypothyroidism. Through most of the 20th century, the diagnosis of hypothyroidism was clinical, based on signs and symptoms resulting from low levels of circulating TH. While diagnosis of overt hypothyroidism was straightforward in the hands of experienced clinicians, mild forms of thyroid deficiency were harder to identify, even when the rate of metabolism basal metabolic rate (BMR) and blood biochemistry was utilized. The availability of a clinical immunoassay to measure circulating levels of TSH changed all that, and serum TSH became a useful biomarker of TH action, the gold standard to diagnose most forms of hypothyroidism in iodine-sufficient areas (2).

Unsurprisingly, serum TSH also became the go-to-test to assess adequacy of TH replacement. However much less successfully. There are mounting questions about the value of serum TSH in faithfully assessing TH signaling in all tissues during therapy with levothyroxine (LT4). Despite normalization of serum TSH, many patients may exhibit residual symptoms that resemble tissue-specific hypothyroidism, for example, impaired cognition, sluggishness, or difficulty maintaining body weight (3). Likewise, patients with syndrome of resistance to TH may exhibit liver hypothyroidism alongside cardiac thyrotoxicosis and osteoporosis (4). How can we assess TH actions in individual tissues? In the current issue of Thyroid, Knock et al. (5) described how they utilized untargeted proteomics analyses of human and mouse plasma to identify 16 proteins produced in liver, spleen, or bone that respond to TH and might function as novel biomarkers of TH action.

The thyroid gland secretes a mixture of thyroxine (T4) and 3,5,3′-triiodothyronine (T3) at a ∼11:1 ratio. Whereas T4 is exclusively produced by the thyroid gland, most T3 in the circulation is produced outside of the thyroid gland. Hence, it makes sense that the hypothalamus–pituitary–thyroid (HPT) feedback mechanism evolved to be exquisitely sensitive to circulating levels of T4. This is due to the type 2 deiodinase (D2) that is expressed in the hypothalamic median eminence and pituitary gland, locally transforming T4 to T3. Thanks to its unique regulation, D2 faithfully informs the TRH and TSH secreting structures about the level of circulating T4 (6). As thyroidal secretion decreases during the initial phases of hypothyroidism, even minimal reductions in circulating levels of T4 are sufficient to elevate TSH secretion. This occurs while serum T3 levels remain within normal range, preserved by a series of thyroidal and extrathyroidal homeostatic adaptations that increase the fractional T3 production. These coordinated actions of the HPT axis explain the extraordinary value of serum TSH in the diagnosis of subclinical or clinical hypothyroidism (2).

In the 1970s, a strong case for LT4 monotherapy was made once humans were found capable of converting T4 to T3, “restoring” circulating T3 levels in LT4-treated athyrotic patients. In addition, knowledge of the mechanisms governing the HPT feedback mechanisms and the availability of clinical immunoassays to measure serum TSH led to the dogma that normalization of serum TSH would signal the ideal replacement dose of LT4 (7). This was a departure from 90 years of dose adjustments guided by “freedom from symptoms with the minimum dose that will accomplish it” (8). Indeed, clinical studies identified advantages in using LT4 alone as replacement therapy, and serum TSH measurements as an ideal tool for monitoring LT4 dosage and restoration of euthyroidism (9).

The utilization of serum TSH to assess monotherapy with LT4 assumes that the extrathyroidal T3 production compensates for the lack of thyroidal secretion of T3. Unfortunately, this does not seem to be the case (10). Because of differences in the way D2 is regulated in the hypothalamus vs the rest of the body, therapy with LT4 normalizes serum TSH before extrathyroidal T3 production can be fully restored, potentially disconnecting serum TSH from peripheral thyroid status (6).

One of the first studies to illustrate this critical point was conducted in 10 hypothyroid patients (aged 20–67 years; 7 females) treated consecutively with increasing doses of liothyronine (LT3; 10, 20, 25, and 50 mcg/day, single dose), each dosage given during at least 4 weeks; subsequently, patients were switched to LT4 (100–150 mcg/day) for an additional 6 weeks (11). Treatment with 20–25 mcg/day LT3 normalized some TH biomarkers, that is, cardiac systolic time intervals and serum creatine phosphokinase activity. However, the integrated serum T3 level, as well as BMR, total serum cholesterol, and TSH levels, remained abnormal in these patients. As the dose of LT3 was increased to 50 mcg/day, integrated serum T3 levels normalized, bringing all TH biomarkers to within normal reference range except for the high serum TSH and total cholesterol that remained >200 mg/dL (11). Remarkably, switching to therapy with LT4 normalized serum TSH but integrated serum T3, and BMR dropped below the normal reference range. Subsequent studies echoed these findings in a number of clinical settings. LT4-treated hypothyroid patients with normal serum TSH were found to have serum T3 levels ∼10% lower despite serum T4 levels that are ∼10% higher (12). When compared with normal individuals with similar TSH levels, LT4-treated patients weigh ∼5 kg more, exhibit a slightly lower BMR, and have higher serum cholesterol and LDL levels, even as they are more likely to be on statin medications (13). A corollary of these studies is that serum TSH levels do not reflect TH signaling in all tissues during therapy with LT4.

A less frequent but dramatic condition in which TH signaling may be inconsistent among different tissues is the syndrome of resistance to TH (4). In patients with this syndrome, TH signaling is selectively impaired, leading to markedly different thyroid status among tissues. Universally restoring TH signaling in such patients can be challenging, if at all possible. This is explained by the existence of two major forms of TRs, TRα and TRβ; tissues express different combinations of these receptors. For example, TRα predominates in the brain, heart, gastrointestinal, and musculoskeletal systems, whereas TRβ is the main TR form expressed in the pituitary gland, liver, kidney, retina, and inner ear. Therefore, TH signaling in patients that exhibit resistance to TRα or TRβ could be markedly different among tissues. The clinical presentation of these patients reflects a phenotypic mosaic depending on which receptor is affected and penetrance of the mutation. Therefore, in both LT4-treated patients and in patients with syndromes of resistance to TH, the existence of additional tissue-specific biomarkers could assist fine tuning TH signaling and restoring universal euthyroidism.

The usefulness of assessing TH on a tissue-specific manner has been demonstrated experimentally in a number of animal models. For example, important tissue-specific differences in TH signaling were identified in a mouse model that carries a luciferase-based reporter gene that is responsive to TH (14). In the clinical setting, however, we need a noninvasive approach to study TH signaling. Hence, TH biomarkers have been studied in the past decades, guided by the progressive understanding of TH actions. For example, circulating metabolites (serum total cholesterol) and proteins (low-density lipoproteins levels, osteocalcin, creatine kinase, ferritin, testosterone-binding globulin, tissue plasminogen activator, angiotensin converting enzyme, and glucose-6-phosphate dehydrogenase) that fluctuate between states of hypo- and hyperthyroidism have been identified (15). However, none of these markers reached widespread clinical utilization because of insufficient sensitivity and/or specificity.

With this in mind, Knock et al. (5) set out to identify enhanced tissue-specific TH biomarkers that could potentially be used in the clinical setting. They first looked at two published studies in which healthy young volunteers were treated with TH for up to eight weeks. The plasma of these individuals was studied by mass spectrometry, an unbiased approach that allows for the quantification of thousands of circulating proteins and metabolites that can ultimately be linked to specific cellular pathways. After matching the results with plasma obtained from thyrotoxic mice, they identified 16 unique plasma proteins that respond to TH excess. A subsequent in silico analysis identified liver, spleen, and bone as the putatively dominant tissues secreting these proteins. CD5L, produced by proinflammatory M1 macrophages in the liver, seems to be the most promising candidate of all, providing insight into TH action in hepatocyte–macrophage crosstalk. Serum CD5L levels are doubled in hyperthyroid vs hypothyroid patients, and correlated positively with free T3 levels and negatively with TSH levels.

As with all the other TH biomarkers identified so far, only future studies will determine the usefulness of this batch of biomarkers, and whether they faithfully reflect TH signaling on a tissue-specific manner. Subsequent studies will also determine whether these markers exhibit sufficient sensitivity to detect subtle changes in thyroid status and specificity to be used clinically. Nonetheless, the data presented by Knock et al. (5) are exciting and should encourage further research in this important field. That TH signaling exhibits such a high degree of tissue specificity is remarkable. The development of clinically relevant TH biomarkers is an important first step toward understanding and possibly modifying TH signaling in health and disease states.