2024 年 71 巻 1 号 p. 23-29
Immune checkpoint inhibitors (ICIs) can cause immune-related adverse events (irAEs) in several organs including endocrine glands. Among endocrine irAEs, thyroid and pituitary irAEs are frequently observed, followed by primary adrenal insufficiency, insulin-dependent diabetes mellitus, and hypoparathyroidism. These conditions could lead to life-threatening consequences, such as adrenal crisis and diabetic ketoacidosis. On the other hand, several types of irAEs including thyroid and pituitary irAEs are reported to be associated with better overall survival. Therefore, it is important to understand and manage endocrine irAEs, which differ depending on the ICI regimen used. In this review, we describe the clinical features, potential biomarkers, management strategies, and possible mechanisms of thyroid and pituitary irAEs.
Immune checkpoint inhibitors (ICIs) are widely used for the treatment of cancer, and show promising effects in several types of malignancies. On the other hand, immune-related adverse events (irAEs) induced by ICIs can develop in several organs including endocrine glands [1, 2]. Endocrine irAEs comprise thyroid dysfunction, pituitary dysfunction, primary adrenal insufficiency, insulin-dependent diabetes mellitus, and hypoparathyroidism [3-5]. To clarify the accurate clinical characteristics of endocrine irAEs, we have been conducting prospective studies analyzing irAEs in patients treated with ICIs since November 2, 2015. In this review, we introduce the findings of our studies regarding thyroid and pituitary irAEs.
The thyroid gland is known to be a susceptible target for autoimmune diseases [6], and thyroid dysfunction is often observed during ICI treatment. The clinical presentation of thyroid irAEs mainly manifest in two types: destructive thyroiditis (DT) and hypothyroidism without preceding thyrotoxicosis. Hyperthyroidism (Graves’ disease) has also been reported as a thyroid irAE but is very rare [7]. The incidence of thyroid irAEs varies among types of ICIs. In a systematic review, the incidence of thyroid irAEs was significantly higher in anti-programmed cell death-1 antibody (PD-1-Ab) monotherapy or a combination of PD-1-Ab with anti-cytotoxic T-lymphocyte antigen-4 antibody (PD-1/CTLA-4-Abs) compared to CTLA-4-Ab monotherapy [8]. In our real-world prospective study, the incidence of thyroid irAEs was significantly higher in PD-1/CTLA-4-Abs (37.0%) compared to PD-1-Ab monotherapy (9.9%) [9] and the incidence of thyroid irAEs induced by anti-programmed death ligand 1 antibody (PD-L1-Ab) was 10.1% [10], which was similar to that of PD-1-Ab.
Thyroid irAEs are shown to develop relatively early, two to six weeks after the start of administration of ICIs [5]. It is thus recommended to measure thyroid hormone levels once a month during the first six months after the start of administration of ICIs [11]. When thyrotoxicosis is observed, anti-thyroid stimulation hormone receptor antibody (TRAb) measurement and thyroid ultrasonography should be performed to distinguish DT from Graves’ disease. In cases of DT, antithyroid drugs should not be used and thyroid function can be followed up with or without administration of β blockers. Most patients with DT subsequently develop hypothyroidism. In our prospective study, 71.4% (25/35) of patients who developed overt DT after PD-1-Ab or PD-1/CTLA-4-Abs developed subsequent hypothyroidism [75% (21/28) in PD-1-Ab and 57.1% (4/7) in PD-1/CTLA-4-Abs, respectively] [9]. Progression from DT to hypothyroidism often takes a period of weeks, requiring careful follow-up of thyroid hormone levels. In case of hypothyroidism, administration of levothyroxine and adjustment of the dose based on the serum thyroid stimulating hormone (TSH) level are needed. ICIs can be administered again when the general condition of the patient becomes stable after appropriate treatments.
There are several studies reporting that the development of thyroid irAEs is associated with prolonged overall survival (OS) in patients with non-small cell lung cancer (NSCLC) [12-19]. On the other hand, it is also demonstrated that thyroid irAEs are not associated with prolonged OS in patients with malignant melanoma (MM) [18, 19]. Therefore, thyroid irAEs could be a potential biomarker to predict better outcome of ICIs in patients with NSCLC but not in patients with MM. One possible explanation is that there may be shared antigens between thyroid and NSCLC tissues but not between thyroid and MM tissues. Another explanation is that there may be common human leukocyte antigen (HLA) types which restrict the T cells from responding to both thyroid and NSCLC antigens but not to both thyroid and MM antigens. Further investigation is needed to clarify the association between thyroid irAEs and OS in malignancies other than NSCLC and MM.
To identify biomarkers of thyroid irAEs, we analyzed the association of anti-thyroid-antibodies (ATAs) at baseline and thyroid irAEs induced by nivolumab (PD-1-Ab). Our data showed that the incidence of thyroid irAEs was significantly higher in patients who had positive ATAs [anti-thyroglobulin-antibodies (TgAbs) or anti-thyroid peroxidase-antibodies (TPOAbs)] at baseline than those who were negative for ATAs [3/6 (50.0%) vs. 1/60 (1.7%), p < 0.001] [20]. Subsequent analysis revealed that patients who were positive for TgAb were at higher risk than patients who were positive for TPOAb alone [14]. Moreover, we revealed that ATAs can also be a biomarker of thyroid irAEs induced by PD-1/CTLA-4-Abs [positive for ATAs 6/10 (60.0%) vs. negative for ATAs 4/17 (23.5%), p < 0.05] (Fig. 1) [9]. Notably, the incidence of thyroid irAEs induced by PD-1/CTLA-4-Abs in the patients negative for ATAs at baseline was not statistically different from that induced by PD-1-Ab in the patients positive for ATAs (p = 0.399) (Fig. 1), suggesting that the risk of thyroid irAEs induced by PD-1/CTLA-4-Abs is high. In addition, another prospective study of ours showed that positive TgAb at baseline could be a biomarker for thyroid irAEs induced by PD-L1-Ab [odds ratio 11.927 (95% CI, 2.526–56.316)] [10]. Therefore, patients who are positive for ATAs at baseline or patients who are treated with PD-1/CTLA-4-Abs require close monitoring of thyroid function. The clinical guidelines of the Japan Endocrine Society point out that it is beneficial to identify patients at high risk such as those positive for ATAs in advance [5]. In addition to ATAs, irregular echo pattern in thyroid gland at baseline [7], elevated TSH level at baseline [10, 21, 22], prior treatment with tyrosine kinase inhibitors (TKIs) [23], prior treatment with ramucirumab or TKIs [10], increased uptake of 18fluorodeoxyglucose in the thyroid [18], higher body mass index [24], and HLA-DPA1*01:03 and DPB1*02:01 [25] are shown to be associated with the development of thyroid irAEs.
The risk of thyroid irAEs is high after PD-1/CTLA-4-Abs even in patients negative for ATAs at baseline
Cumulative incidence of thyroid irAEs in patients treated with PD-1-Ab or PD-1/CTLA-4-Abs. Blue and red lines indicate Kaplan-Meier curves of patients treated with PD-1-Ab and PD-1/CTLA-4-Abs. Solid and dashed lines indicate patients with positive (Ab+) and negative (Ab–) for ATAs at baseline. The cumulative incidence of thyroid irAEs was significantly higher in the PD-1-Ab (Ab+) than PD-1-Ab (Ab–) group [28/87 (32.2%) vs. 13/329 (4.0%), log-rank test, p < 0.001] and in the PD-1/CTLA-4-Abs (Ab+) than PD-1/CTLA-4-Abs (Ab–) group [6/10 (60.0%) vs. 4/17 (23.5%), log-rank test, p < 0.05]. Notably, the risk of thyroid irAEs in the PD-1/CTLA-4-Abs (Ab–) group was not statistically different from that for the PD-1-Ab (Ab+) group (p = 0.399).
Abbreviations: irAE, immune-related adverse event; PD-1-Ab, anti-programmed cell death-1 antibody; PD-1/CTLA-4-Abs, a combination of PD-1-Ab with anti-cytotoxic T-lymphocyte antigen-4 antibody (CTLA-4-Ab); Ab, anti-thyroid-antibodies; Combi, a combination of PD-1-Ab with CTLA-4-Ab.
To elucidate the mechanism of thyroid irAEs, a mouse model was reported from our laboratory in which DT was induced by PD-1-Ab injections after thyroglobulin immunization [26]. In this model, it was shown that PD-1-Ab administration increased the frequencies of central memory and effector memory CD4 T cells in the thyroid and cervical lymph nodes and that granzyme B was expressed on CD4 T cells infiltrating the thyroid, suggesting that cytotoxic CD4 T cells could directly damage thyrocytes. The development of DT was completely prevented by the prior depletion of CD4 T cells and was partially prevented by the depletion of CD8 T cells, suggesting an essential role of CD4 T cells in the pathogenesis. Furthermore, flow cytometric analyses of peripheral blood mononuclear cells revealed that the frequencies of central memory and effector memory CD4 T cells expressing the cytotoxic marker CD27 were higher in patients with DT induced by PD-1-Ab than in those without [26]. These findings suggest that cytotoxic memory CD4 T cells, which are activated by PD-1-Ab, play a critical role in the pathogenesis of DT induced by PD-1-Ab in humans.
Autoimmune hypophysitis is a chronic inflammatory disease that is characterized by infiltration of lymphocytes in the pituitary gland [27, 28]. Although autoimmune hypophysitis itself is a very rare disease, many cases with similar clinical features during ICIs administration have been reported, and pituitary dysfunction as an irAE has thus attracted attention.
The clinical characteristics of pituitary irAEs differ among causative ICIs. In previous studies, it was reported that the incidence of pituitary irAEs induced by CTLA-4-Ab and PD-1-Ab was 4–10% and 0–1.2%, respectively [29-32]. However, in our prospective study including 174 patients treated with ICIs, the incidence of pituitary irAEs induced by ipilimumab (CTLA-4-Ab) and PD-1-Ab was 24.0% (6/25) and 6.0% (10/167), respectively [19], which was much higher than in the previous studies and suggests that pituitary irAEs might have been overlooked in previous retrospective studies. In addition, it was shown that pituitary irAEs have two different disease types: combined pituitary hormone deficiencies accompanied by pituitary enlargement and isolated adrenocorticotropic hormone (ACTH) deficiency (IAD) without pituitary enlargement. It was also shown that ACTH was almost always impaired in pituitary irAEs and that 1) IAD can be induced by any classes of ICIs, and 2) combined pituitary hormone deficiencies with pituitary enlargement is likely to be induced when CTLA-4-Ab is used as a monotherapy or a combination therapy with PD-1-Ab [19, 29, 33, 34]. This suggests there were different underlying mechanisms between pituitary irAEs induced by CTLA-4-Ab and those induced by PD-1-Ab/PD-L1-Ab.
It should be noted that symptoms of pituitary irAEs, such as fatigue or appetite loss, may be overlooked in patients with advanced malignancies. In our prospective study, the frequency of hyponatremia was significantly higher in patients treated with ICIs who developed pituitary irAEs than in those who did not [6/16 (37.5%) vs. 10/120 (8.3%), p < 0.01] [19]. Therefore, we should measure pituitary hormones when hyponatremia is observed in patients treated with ICIs so that ACTH deficiency will not be overlooked. In addition to hyponatremia, decreases in TSH level [35], elevated FT3 level [36], and eosinophilia [37] are reported to be a potential indicator for pituitary irAEs. Since pituitary irAEs can develop even after discontinuation of ICIs [38], it may be useful to regularly monitor the pituitary hormones and electrolytes, as well as to evaluate the subjective symptoms.
As a treatment for ACTH deficiency induced by ICIs, the guidelines of the Japan Endocrine Society recommend administration of physiological doses of hydrocortisone (10–20 mg/day) [5], since there is no evidence that high doses of glucocorticoids improve the pituitary hormone deficiencies or the survival outcomes [39, 40]. However, high-dose glucocorticoids can be administered in cases with marked enlargement of the pituitary gland accompanied by headaches and visual disturbance. Similarly, high-dose steroids are also recommended in the presence of severe compressive symptoms or adrenal crisis in the clinical guidelines of the American Society of Clinical Oncology [41] and the European Society for Medical Oncology [42]. If pituitary hormones other than ACTH are impaired, individual replacement therapy should be performed. However, in case of both ACTH and TSH deficiency, hydrocortisone must be administered before starting thyroid hormone replacement so that adrenal insufficiency will not be exacerbated. In our prospective study in which all patients with pituitary irAE were treated with physiological doses of hydrocortisone, the development of pituitary irAEs was associated with better OS in patients with NSCLC and MM [19]. Thus, it was suggested that pituitary irAEs could be a potential biomarker to predict better outcome of ICIs if appropriately treated.
It is possible that autoantibodies and HLAs may be useful as biomarkers of pituitary irAEs. Anti-pituitary antibodies (APAs) are autoantibodies against human pituitary cells and are present in some patients with autoimmune hypophysitis [43]. HLA-DQ8 and DR53 were reportedly present in patients with lymphocytic hypophysitis, although this is only one cohort study analyzing 15 patients, most of who were Caucasian [44]. To clarify the biomarkers of pituitary irAEs, we investigated APAs and HLAs in 22 patients with pituitary irAEs and 40 patients without pituitary irAEs [45]. We used indirect immunofluorescence of human pituitary gland substrates to assess the presence of serum APAs, as described previously [43, 46, 47]. The prevalence of APAs at baseline was significantly higher in patients with IAD than in control patients [64.7% (11/17) vs. 2.5% (1/40), p < 0.05]. Although APAs were negative at baseline in all patients who developed combined pituitary hormone deficiency with pituitary enlargement, they had become positive before the onset of pituitary irAEs in 3 of 4 (75.0%) patients several weeks after ipilimumab administration (Fig. 2). Furthermore, HLA analysis revealed that the prevalence of HLA-Cw12, DR15, DQ7, and DPw9 was significantly higher in patients who developed IAD, whereas that of HLA-Cw12 and DR15, but not DQ7 or DPw9, was significantly higher in patients who developed combined pituitary hormone deficiency with pituitary enlargement than in controls. HLA-DPw9 was also reported to be associated with IAD induced by PD-1-Ab but not with idiopathic IAD in another study [48]. These results suggest that positive APAs at baseline and after treatment, together with susceptible HLA alleles, could be predictive biomarkers to identify patients with a high risk of pituitary irAEs. It would be desirable to establish an enzyme-linked immunosorbent assay (ELISA) system to measure APAs quantitatively that can be clinically implemented in the future.
Positive APAs are biomarkers of pituitary irAEs
Schema of the time course of APAs after ICI administration in patients who developed pituitary irAEs. APAs are positive at baseline in patients who developed IAD. On the other hand, they are negative at baseline but become positive several weeks before the onset of pituitary irAEs in patients who developed combined pituitary hormone deficiency with pituitary enlargement.
Abbreviations: ICIs, immune checkpoint inhibitors; irAEs, immune-related adverse events; APAs, anti-pituitary antibodies; IAD, isolated adrenocorticotropic hormone (ACTH) deficiency; CPHD, combined pituitary hormone deficiencies.
In a mouse model of hypophysitis induced by CTLA-4-Ab [49], it was shown that CTLA-4 was expressed in pituitary cells, particularly TSH-secreting cells and prolactin-secreting cells, and that CTLA-4-Ab directly acted on this CTLA-4 resulting in complement activation in the pituitary. In other words, it was clarified that CTLA-4-Ab administration induces localized inflammation mediated by type Ⅱ hypersensitivity in the pituitary gland. It is postulated that this inflammation leads to subsequent autoimmune response to the pituitary gland in the mice systemically injected with CTLA-4-Ab, leading to the development of pituitary hormone deficiency and pituitary enlargement. This hypothesis is also supported by the fact that APAs became positive after ipilimumab administration in all 7 patients who developed pituitary irAEs induced by ipilimumab [49].
On the other hand, the pathogenesis of pituitary irAEs induced by PD-1-Ab is unclear. A study analyzing an autopsy case of esophageal cancer where the subject developed ACTH deficiency after nivolumab administration reported that infiltration with CD8-dominant T cells, CD68 macrophages and CD20 B cells were observed in the pituitary gland and that IgG and C4d were deposited on the pituitary cells, whereas IgG4, a subclass of nivolumab, was not observed [50]. These findings suggest that type Ⅳ (Ⅳa + Ⅳc) hypersensitivity and secondary type Ⅱ hypersensitivity are involved in the pathogenesis of pituitary irAEs induced by PD-1-Ab. Another study reported that anti-corticotroph antibody was detected in two out of 20 patients who developed ACTH deficiency induced by ICIs, sera from the two patients recognized proopiomelanocortin (POMC), and that tumors from the two patients ectopically expressed ACTH, suggesting that autoimmunity against corticotrophs as a form of paraneoplastic syndrome could be involved, at least in part, in the pathogenesis of IAD induced by ICIs [51].
Endocrine irAEs are expected to increase further due to the expansion of use of ICIs in the near future. In particular, ACTH deficiency as a pituitary irAE could be fatal if overlooked, so close monitoring should be required after administration of ICIs. As introduced in this review, there are some biomarkers to predict the onset of endocrine irAEs. Moreover, the development of endocrine irAEs could be a potential biomarker of better outcome of ICIs. Therefore, this is expected to spur the development of novel tailor-made cancer immunotherapy based on the risk of irAEs and refine the predicted treatment efficacy for each patient.
TK wrote the manuscript. TK, SI and HA were involved in revising the manuscript.
None.
S.I. received personal fees from Ono Pharmaceutical Co. Ltd., Bristol-Myers Squibb, Chugai Pharmaceutical Co., Ltd., MSD K.K., AstraZeneca Co., Ltd., and Merck Biopharma Co., Ltd. outside of this study. H.A. received grants from Ono Pharmaceutical Co., Ltd., MSD K.K., and Chugai Pharmaceutical Co., Ltd., and personal fees from Ono Pharmaceutical Co., Ltd., Bristol-Myers Squibb, and MSD K.K. outside of this study. T.K. has nothing to disclose.
S.I. is a member of Endocrine Journal’s Editorial Board.
