The Thyroid Pathologist Meets Therapeutic Pharmacology

Abstract

The effects of many pharmacological agents on thyroid function are well known. Direct influences on measurements of thyroid function tests are also described. However, certain classes of drugs produce morphological changes in the gland. This review focuses on the significance of the following drug classes for the thyroid pathologist: iodine, antithyroid drugs, psychotropic drugs, antibiotics, cardiotropic drugs, antidiabetic drugs, and immunomodulatory agents. Radioactive iodine initially induces mild histologic changes; however, the long-term effects include marked follicular atrophy, fibrosis, and nuclear atypia—changes that vary depending on the pre-therapy condition of the gland. Some psychotropic drugs have been associated with a spectrum of inflammatory changes throughout the gland. The tetracycline class of antibiotics, namely minocycline, can lead to a grossly black thyroid gland with pigment seen in both colloid and follicular epithelial cells while variably present within thyroid nodules. The surgical pathologist most commonly sees an amiodarone-affected gland removed for hyperthyroidism, and the histologic findings again depend on the pre-therapy condition of the gland. While GLP-1 receptor agonists carry an FDA black box warning for patients with a personal or family history of multiple endocrine neoplasia or medullary thyroid carcinoma, the C cell hyperplasia originally noted in rats has not borne out in human studies. Finally, thyroiditis and hypothyroidism are well known complications of checkpoint inhibitor therapy, and rare cases of severe thyroiditis requiring urgent thyroidectomy have been reported with unique histologic findings. In this review, we describe the histologic findings for these drugs and more, in many cases including their functional consequences.

Introduction

Many chemical compounds and drugs affect thyroid function. The vast majority of them interfere with production, release, or function of thyroid hormones. However, others act on peripheral receptors of these hormones or may interfere with the laboratory testing of thyroid function. A relatively small number of agents directly affect the thyroid histology, and it is this latter group of compounds that will be the focus of this review. These agents include iodine, drugs that suppress thyroid hormone production (such as methimazole and propylthiouracil), certain psychotropic agents, antibiotics, cardiotropic drugs, antidiabetic agents, and immune regulatory compounds used as therapy of malignancies. This paper will review and illustrate the effects of these drugs on the morphology of the thyroid and discuss, when known, the functional consequences of these reactions.

Iodine

Iodine is an integral part of the biosynthesis and function of thyroid hormones. Deficiency of iodine has been shown by numerous worldwide historical studies to induce the development of goiter as well as hypothyroidism [1]. Histologically, these glands show follicles distended with colloid and often degenerative changes (including fibrosis and calcifications), which reflect the longstanding nature of the condition.

The common past practice of “preparing” the Graves gland for surgery included the administration of potassium iodide (also known as Lugol solution) for 2–3 weeks prior to surgery. It is believed that this induced colloid storage, decreased papillary hyperplastic changes, and also decreased the vascularity of the gland. This idea was introduced in the early part of the twentieth century, and the histologic evidence of these findings was very weak due to the rarity of thyroid surgeries at this time (because of the significant mortality due to hemorrhage from the very vascular gland and possible postoperative severe hyperthyroidism). The advent of more advanced anesthetic, operative, and postoperative techniques has made use of potassium iodide in thyroid surgery obsolete [2].

When iodide is given in excess, the normal gland can result in clinical hypothyroidism. A study from Japan described and illustrated the changes of iodide-induced hypothyroidism [3]. They indicated five reversible changes that reversed with iodine restriction: hyperplastic follicles, relative lack of colloid in follicular lumens, clear cell change, variation in cell size, and dilated stromal vessels [3].

Radioactive Iodine

Radioiodine (RaI) is used in the treatment of hyperthyroidism. The dose of RaI is relatively low (< 30 mCi), and it may acutely result in mild inflammatory changes and follicular destruction as well as vascular dilatation. However, its long-term effects are histologically striking: severe follicular atrophy, minimal colloid, stromal fibrosis, focal oncocytic metaplasia, and random marked nuclear atypia (Fig. 1) [4]. Several papers have described the difficulty, especially in cytological preparations, in distinguishing the radiation-induced atypia from malignancy. Due to the similarities between these radiation-induced findings and those of malignancy, it is critical to obtain a clinical history of prior radiation therapy in order not to mistake the changes for carcinoma [5–9].

Fig. 1.

Atrophic thyroid from patient with history of radioiodine treatment for Graves disease 15 years before and had developed a new nodule (not shown). The background thyroid shows extensive fibrosis follicular atrophy and clusters of stromal lymphocytes (a, H&E, original 20 × magnification) and high-power magnification of the follicles shows oncocytic cytoplasm and enlarged, atypical nuclei (b, H&E, original 400 × magnification)

In addition to the cytological features in these treated glands, the morphological effects of radioiodine depend upon the initial histology of the thyroid. For example, in the setting of diffuse toxic goiter, the blood supply throughout the gland is evenly distributed, and thus the morphologic changes of the RaI are diffuse. However, in the setting of toxic nodular goiter, the blood distribution is centered around the toxic nodules; thus, the gland will demonstrate unequal RaI effects.

Overall, the alterations found from RaI result in hypothyroidism, which can be severe, and thus careful clinical monitoring of patients after RaI treatment is mandatory to maintain a functionally euthyroid state.

Antithyroid Drugs

Propylthiouracil (PTU) and methimazole are the mainstays of medical treatment for autoimmune hyperthyroidism. Muzukami et al. found that PTU histologically causes clear and vesicular change in the follicular epithelial cells [3]. One case of marked cellular pleomorphism after methimazole therapy has been reported, yet it is unclear if the histological change was a result of the epithelial atypia that can occur in longstanding hyperthyroid glands or if it was a direct effect of the therapeutic drug [10].

Psychotropic Agents

Lithium

Lithium is used in the treatment of various mood disorders, and similarly to other psychopharmaceutical drugs, its side effects have been extensively identified. This drug leads to decreased thyroid hormone synthesis in the thyroid while peripherally decreasing thyroxine deiodination [11]. These effects clinically present most commonly as hypothyroidism, though few reports of hyperthyroidism have been made [12].

Histologically, the effects of lithium are diffuse. Atrophic and occasionally disrupted follicles with various lymphoid follicles are the most common histologic finding (Fig. 2). The lymphoid aggregates consist of both primary and secondary follicles, and the lymphocytes may be interspersed throughout the follicles. Histiocytes and palpation thyroiditis may also be identified, and depending on the severity of the disease, bands of fibrosis may be seen [13]. These histologic findings are present in cases of hypothyroidism and hyperthyroidism alike, though the extent of each varies depending on each case.

Fig. 2.

Lithium effects on the thyroid can manifest histologically as chronic lymphocytic thyroiditis (H&E, original 100 × magnification)

Phenytoin

Phenytoin is utilized to prevent tonic–clonic and focal seizures, arrhythmias, and neuropathic pain. Within the brain, the drug is known to block voltage-gated sodium channels in the motor cortex. It can also inhibit binding of thyroxine to serum binding proteins thus increasing thyroxine clearance through urine [14, 15]. Clinically, this results in a hypothyroid picture, which is most commonly treated with thyroid replacement therapy [16].

Although this drug has no direct effect on the thyroid gland, there are histologic findings that have been reported in rat studies. El-Bermawy described that rat thyroid demonstrated evidence of follicular destruction with intact follicles demonstrating scant colloid [17]. Review of the English literature reveals no articles detailing the thyroidal histology of a patient taking phenytoin; however, it may be reasonable to presume thus far that findings may be similar to those of lithium or other hypothyroid-inducing drugs.

Antibiotics

Minocycline

Minocycline is one of the tetracycline class of antibiotics, which are bacteriostatic drugs that function by binding the 30S ribosomal unit to inhibit protein synthesis. Minocycline is available in topical or oral forms and is commonly used in the treatment of acne and urinary tract infections, among others. While the drug can cause hypersensitivity reactions, autoimmune conditions, and other, more severe adverse outcomes, animal trials of the drug described the development of a “black thyroid” that occurred after use in monkeys, dogs, and rats (but not mice) in a range of doses and persisted for 1 year after cessation [18].

Within years of its release for clinical use, the first autopsy report described a grossly black thyroid in a man who had taken minocycline for one year. The microscopic findings on hematoxylin–eosin stained slides confirmed the presence of black pigment aggregates both within follicles and the surrounding follicular cells, and the authors noted hyperplastic changes and pyknosis in some of the affected follicular cells [19]. Similar findings have been observed in thyroidectomy specimens (Fig. 3), and in both benign and malignant thyroid nodules, pigmentation has been reported in all combinations: unpigmented nodule and pigmented thyroid parenchyma, pigmented nodule and pigmented thyroid parenchyma, and, rarely, pigmented nodule and unpigmented parenchyma [20–27].

Fig. 3.

Minocycline use can result in the accumulation of black pigment, which can be observed grossly as a black thyroid as seen in this autopsy (a). Histologically, black pigment is seen predominantly accumulating within the colloid but also within thyroid follicular cells. In this example, increased pigment accumulation is present within a hyperplastic nodule compared to the surrounding parenchyma (b, H&E, original 40 × magnification). On high power, the pigment is fine and granular, forming loose but distinct aggregates in colloid. Within follicular cells, an apical distribution can be seen (c, H&E, original 200 × magnification). In this example, essentially no pigment is seen within an encapsulated papillary thyroid carcinoma while robust pigmentation is present in the surrounding parenchyma (d, H&E, original 100 × magnification)

Additional studies have described the nature of the pigment. In addition to central pigment accumulation within follicles, light and electron microscopy demonstrated that pigment initially developed in the apical aspect of follicular cells in lysozyme-like granules [28]. These granules are typically membrane-bound with electron-dense contents, variably described as melanin-like or lipofuscin-like [21, 29, 30]. Special stains have shown that the pigment bleaches with potassium permanganate, is not birefringent or autofluorescent, and is positive on Masson-Fontana and Schmorl stains while generally negative on Ziehl–Neelsen, Prussian blue, and PAS stains [21, 30–32].

Physiologic studies on minocycline-treated animals have shown no difference in TSH-stimulated T3 release compared to untreated controls, although a significant decrease in T4 release was noted [28]. In general, minocycline had been considered to have little impact on human thyroid function; however, several reports have brought this belief under scrutiny, particularly in the pediatric population. Transient abnormalities of thyroid function associated with minocycline use have been reported in children involving autoimmune thyroid disease [33, 34]. More recently, persistent, non-immune thyroid dysfunction has been reported in children [35, 36]. Histologic reports of drug-induced thyroiditis with clinical hyperthyroidism associated with chronic minocycline use have provided additional evidence to support a risk for thyroid dysfunction associated with the drug [37]. Additionally, immunohistochemical studies have shown decreased thyroglobulin and ubiquitin staining in the thyroid follicular cells of pigmented thyroids [32]. Finally, some groups have associated black thyroids with a higher rates of malignancy in the surgical setting [38].

Tetracycline and Doxycycline

Tetracycline and doxycycline function in a manner similar to minocycline. While tissue pigmentation has long been a concern with tetracycline administration, particularly in children, pigmentation of the thyroid has typically not been as profound as is seen with minocycline. Animal studies have shown a dark brown discoloration; however, the pigment noted microscopically was brownish yellow in contrast to the black pigment seen in minocycline [39]. Only one case of a pigmented thyroid in the setting of doxycycline therapy has been reported [40]. Notably, a similar phenomenon of a dark or black thyroid with dark brown pigment has been observed in the setting of lithium and antidepressant use, though reports are rare [41].

Cardiotropic Drugs

Amiodarone

Amiodarone is used to treat certain cardiac arrhythmias. It contains iodine moieties as part of its structure and iodine accounts for 37% of its molecular weight. Due to its characteristics (including its iodine content), it resembles T4 [42]. The drug is very lipophilic and is concentrated in adipose tissue, muscle, and thyroid. Interestingly, it has been found to induce ultrastructural changes in several organs, including the lung, thyroid, and liver, which correlates with its extensive side effect profile [43–45]. Furthermore, Smyrk et al. demonstrated that the drug is taken up by thyroid follicular cells by direct measurement of thyroid tissue from a patient taking the drug [42].

Amiodarone can lead to either hypothyroidism or hyperthyroidism. The time course for the development of thyroidal disease is unpredictable: Some patients show thyroid dysfunction within weeks of starting amiodarone and others may take up to six months. Hypothyroidism may result from the drug’s interference with thyroid hormone production due to the iodine within the drug; when intrathyroidal iodine concentrations reach a critically high level, iodine transport and thyroid hormone synthesis are transiently inhibited, which can lead to hypothyroidism [46–48]. However, patients in this setting most commonly are treated with thyroid hormone supplementation rather than undergoing thyroidectomy.

While surgical pathologists rarely (if ever) see an amiodarone-affected thyroid for hypothyroidism, it is fairly common for the gland to be seen in the setting of amiodarone-associated hyperthyroidism. In certain cases of severe hyperthyroidism caused by amiodarone, the thyroid gland must be removed as withdrawal of the drug will induce potentially dangerous alterations to cardiac status [42].

The literature classifies amiodarone hyperthyroidism as type 1 and type 2. Which category the gland falls into depends on the disease state of the gland prior to initiating amiodarone therapy (i.e. a nodular or a normal gland) [42]. Type 1 amiodarone-associated thyrotoxicosis occurs in the setting of a nodular gland (with adenomatous nodular hyperplasia) or with autoimmune disease (Hashimoto disease). As these patients do not effectively autoregulate iodine, adding more iodine through amiodarone results in excess thyroid hormone synthesis and thyrotoxicosis.

On the other hand, type 2 amiodarone-associated thyrotoxicosis occurs in the setting of a normal gland, where the drug affects and destroys the follicular epithelium (Fig. 4). These surgical specimens are often grossly unremarkable, but the histology shows a variable histiocytic reaction (sometimes with multinucleated cells) around rare or multiple groups of follicles, follicular disruption, loss of follicular epithelial cells, breaking of the follicular basement membrane, and even identifiable colloid released into the bloodstream. The release of this colloid results in increased thyroid hormone in the blood, which results in the hyperthyroid state. Lymphocytic infiltrates are not found [46, 47]. This phenomenon is similar to that seen in subacute thyroiditis or in rapidly growing thyroidal tumors where follicles are destroyed and thyroid hormone is released (known as the mechanicodestructive type of mechanism of hyperthyroidism) [49].

Fig. 4.

This thyroid is from a 78-year-old patient with severe hyperthyroidism who had been treated with amiodarone for cardiac arrythmias. Part a shows a follicle with prominent histiocytic reaction to colloid (H&E, 100 × original magnification). Part b is higher power to show follicular epithelium replaced by histiocytes, which also are present in the colloid (H&E, 400 × original magnification)

Anti-diabetic Drugs

GLP-1 Receptor Agonists

Numerous drugs are available to treat type 2 diabetes mellitus; however, one class which should be of concern to the thyroid surgical pathologist are the GLP-1 receptor agonists. These agents bind to the GLP-1 receptors causing a cascade effect: increasing the production and release of insulin, preventing glucagon release, and preventing hepatic glucose production [50]. It is extremely successful in reducing hyperglycemia in patients, while preventing hypoglycemia may be noted [50]. In humans, GLP-1 receptors are located on pancreatic β-cells, arterial smooth muscle cells in the kidneys and lungs, and in Brunner’s glands in the duodenum [51].

Semaglutide (commercially sold as Ozempic^® and Wegovy^®) is one of the most popularly used GLP-1 receptor agonists. This drug, as well as others within this class, is of particular interest to the thyroid surgical pathologist due to findings in the initial mouse studies performed prior to clinical trials. These early mouse trials demonstrated focal C cell hyperplasia, increased prevalence of thyroidal C cell adenomas and carcinomas, and elevated calcitonin levels after 1 dose of semaglutide for male and female rats [52]. In male rats, calcitonin was elevated 6 weeks after treatment. Subsequent studies by the same group interestingly reported that none of these findings occurred during their monkey studies [52].

Given the serious prognoses implied with these findings, the results were concerning, yet the variation amongst rodent and primate studies was found to depend predominantly on the differences of rodent versus primate physiology. A prior study demonstrated that rodent thyroidal C cells express the GLP-1 receptor [53], and other studies demonstrated that primate C cells (both monkey and human) do not express it [53, 54]. These findings are ones that support the findings for the original semaglutide studies submitted for FDA approval.

When the human clinical trials for semaglutide were performed, there was no evidence of elevated calcitonin, C-cell hyperplasia, or increased risk of medullary carcinoma, yet concern over the possibility of these outcomes grew tremendously due to the possible serious consequences. Therefore, the US Food and Drug Administration (FDA) placed a warning for clinicians to not prescribe this drug to patients with a personal or family history of medullary thyroid carcinoma or patients with multiple endocrine neoplasia 2. This same warning occurs with numerous other GLP-1 receptor agonists such as dulaglutide (sold as Trulicity^®), liraglutide (sold as Victoza^®), and other forms of semaglutide (such as Rybelsus^®, Wegovy^®).

Given this warning, much is left to know about how this drug may affect thyroid histology. Fuchs et al. demonstrated that attempting to identify C cell hyperplasia by quantifying the number of C cells in a thyroid is impossible given the widely variable number of C cells in human thyroids with no association with medullary carcinoma [55]. Additionally, no formal studies have been performed identifying the quantity of C cells or the presence of medullary carcinoma in patients on any GLP-1 receptor agonists.

Immunologic Agents

Pembrolizumab

The emergence of immunomodulatory therapy has produced dramatic results in the treatment of many cancer types, most well-known in melanoma and non-small cell lung cancer. One target of such therapy is the interaction between programmed death ligand 1 (PD-L1) and its receptor (PD-1). Physiologic activation of this signaling pathway commonly occurs between PD-L1-expressing macrophages and PD-1-expressing T cells, resulting in an inhibitory effect on T cells. Hijacking of this interaction has been demonstrated in cancer, in which upregulation of PD-L1 by tumor cells can allow them to evade the immune responses that would otherwise eliminate the aberrant cells. Pharmacologic inhibition of this interaction with monoclonal anti-PD-1 or PD-L1 antibodies can overcome this immune evasion and is the mechanism of immunologic agents such as pembrolizumab (anti-PD-1) and atezolizumab (anti-PD-L1).

Thyroiditis and hypothyroidism are well-known adverse events caused by pembrolizumab and are included on the drug’s FDA package insert. Patients with preexisting thyroiditis appear to be at increased risk for adverse events. Indeed, studies in papillary thyroid carcinomas have demonstrated increased expression of PD-L1 in tumors arising in a background of Hashimoto thyroiditis in contrast to near absent expression in tumors arising in thyroids without thyroiditis. These differential levels of PD-L1 expression persisted in metastases. Furthermore, PD-L1 expression was also noted in the non-neoplastic follicular epithelium in cases with Hashimoto thyroiditis, potentially predisposing those cells for injury with the initiation of PD-1 blockade [56]. Clinical studies have confirmed that thyroid dysfunction is common after PD-1 inhibitor administration, and patients with a higher baseline TSH level may be at increased risk for thyroid-related adverse events [57].

Case reports of patients with pembrolizumab-induced thyroiditis have shown subtle gross findings with firm, nodular, brown thyroid parenchyma. Morphologic findings consisted of diffuse follicular injury and extensive replacement of follicular cells by lymphocytic infiltrates composed of B and T cells accompanied by a proliferation of histiocytes [58, 59]. PD-L1 expression in these cases was essentially confined to the histiocytes (Fig. 5). In one of these cases, the patient experienced thyroid storm and refractory thyrotoxicosis necessitating thyroidectomy [58].

Fig. 5.

Histologically, pembrolizumab-associated thyroiditis shows diffuse involvement of the gland with a vaguely nodular appearance on low power (a, H&E, original 40 × magnification). At higher magnification, a robust histiocytic infiltrate fills most follicles, replacing most of the follicular cells with only rare intact follicle remaining (b, H&E, original 200 × magnification). The intervening inflammatory infiltrate is predominantly B and T lymphocytes. PD-L1 staining predominantly highlights the membranes of histiocytes (c, original 200 × magnification), which are confirmed by a CD68 immunohistochemical stain (d, original 200 × magnification)

SARS-CoV-2 Vaccines

With the COVID-19 pandemic, reports emerged of patients developing thyroid dysfunction in the setting of SARS-CoV-2 infection, including the initial onset or relapse of subacute thyroiditis and Graves disease with most patients returning to normal within months after infection [60–63]. Autopsy studies have similarly demonstrated the presence of SARS-CoV-2 antigens and nucleic acids in the thyroid follicular cells of affected patients [64, 65]. The underlying mechanism for these findings has been supported by in vitro studies, which have demonstrated expression of the SARS-CoV-2 receptor (ACE-2 receptor) and co-receptor (TMPRSS2) on thyroid follicular cells [65, 66]. With the rapid development and wide implementation of effective SARS-CoV-2 vaccines, including mRNA-based agents, some have reported associated subacute thyroiditis and Graves disease [67, 68]. However, epidemiological studies have not demonstrated increased rates of thyroid disease among vaccinated individuals, suggesting that thyroid disease in vaccinated patients is more likely coincidental [69].

Conclusion

This review has provided an overview of the histologic associations with various, commonly used therapeutic agents often prescribed for non-thyroidal diseases. It is important to recognize that when the pathologist is faced with a thyroid nodule as the clinical problem, there are sometimes changes in the background “normal” thyroid that should be recognized to understand the entire pathological picture. In this review, we have compiled some of the common microscopic reactions that are produced by various pharmaceutical agents.

Author Contribution

Dr. LiVolsi conceptualized the manuscript. Drs. Sande, Tondi Resta, and LiVolsi led the literature review, manuscript preparation, and figure production for the antibiotics/immunologic drugs, psychotropic/anti-diabetic drugs, and iodine/antithyroid/cardiotropic drugs sections, respectively. The manuscript was reviewed and revised by all authors, and there was joint agreement and approval of the final manuscript for submission.

Availability of Data and Materials

Not applicable for this review article.

Declarations

Ethics Approval

Not applicable for this review article.

Conflict of Interest

The authors declare no competing interests.