Massachusetts General Hospital Logo

Neuroendocrine and Pituitary
Tumor Clinical Center (NEPTCC) Bulletin

Winter 2019/2020 | Volume 25, Issue 1

Hypophysitis Associated With Immune Checkpoint Inhibitors

-ALEX FAJE, MD

Background
Immune checkpoint inhibitors (CPIs) have transformed the landscape of oncology since the approval of ipilimumab by the US Food and Drug Administration in 2011. Eighteen cancer types are currently approved for treatment with 7 CPIs (ipilimumab, nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, and avelumab). A recent study estimated that almost half of all patients with cancer in the United States are eligible for treatment with CPIs, a proportion far outweighing those eligible for genome-targeted therapies (1-2).

Ipilimumab is an IgG1-based monoclonal antibody. Canonically, its effects are mediated through the inhibition of cytotoxic T-lymphocyte antigen-4 (CTLA-4) on activated T cells. CTLA-4 competitively binds B7 to block costimulation and thereby downregulates T cell activation and proliferation, acting as a brake on the immune response. Nivolumab, pembrolizumab, and cemiplimab are IgG4-based monoclonal antibodies targeting programmed cell death 1 (PD-1). Atezolizumab, durvalumab, and avelumab are IgG1-based monoclonal antibodies that inhibit programmed death-ligand 1 (PD-L1). PD-1 is expressed by activated T and B lymphocytes and monocytes; PD-L1 is found on antigen-presenting cells and many other cell types. Binding of PD-1 to PD-L1 inhibits the immune response of these cells primarily at the level of the tumor microenvironment (3-4). Tumoral expression of PD-L1 and/or PD-L2 can act as a mechanism of evasion from immune surveillance.

Immune-related adverse events (irAE) occur in some patients as sequelae from CPI treatment and can involve any organ system. Potential endocrine irAEs include hypophysitis, primary hypothyroidism with or without destructive thyroiditis and temporary thyrotoxicosis, hyperthyroidism due to Graves disease, diabetes mellitus, primary adrenal insufficiency (AI), and hypoparathyroidism. Primary hypophysitis is a rare disorder with an estimated annual incidence of 1 in 7-9 million (5), and it accounts for less than one percent of pituitary surgery cases (6-10). Hypophysitis secondary to CPIs is a recently recognized clinical entity and has likely become the most frequent etiology for new hypophysitis cases.

Epidemiology
The risk of hypophysitis after treatment with ipilimumab is significant. Pooling the largest endocrinology-focused studies yields an overall risk of approximately 10% following treatment with ipilimumab monotherapy (145/1,394-1,438) (11-18). The frequency of hypophysitis following anti-PD-1 treatment is much lower and appears to be less than 1% (17/3,522) (13). Whether combination therapy with ipilimumab plus anti-PD-1 agents increases the risk of hypophysitis has not yet been fully determined. No endocrinology study has addressed this question yet. Data derived from prospective oncology studies may be significantly limited by multiple inadequacies of the endocrinology categories in the Common Terminology Criteria for Adverse Events, many of which are imprecise and overlapping. Hypophysitis appears to be rare following treatment with agents targeting PD-L1, but available data are limited. No endocrinology study or review of oncology studies has analyzed this population. Male gender appeared to be a risk factor for ipilimumab-associated hypophysitis in one study (14) but not others (12, 15-17). It is unclear if age is a risk factor for hypophysitis after treatment with ipilimumab. The influence of age and gender has not been examined for anti-PD-1 and PD-L1-associated hypophysitis.

Presentation and Diagnosis
The clinical presentation of hypophysitis appears to be distinct in patients receiving ipilimumab compared to those receiving anti-PD-1 therapy. Hypophysitis develops relatively soon after treatment initiation with ipilimumab, typically after 2 or 3 cycles of therapy. Hypophysitis may be diagnosed over a much broader period of time in patients treated with anti-PD-1 agents, in some cases nearly a year after starting therapy (Figure 1). Headache is a frequent presenting symptom in ipilimumab-associated hypophysitis but uncommonly occurs in anti-PD-1 hypophysitis patients. Concordant with that difference, radiographic pituitary enlargement occurs in nearly all ipilimumab hypophysitis patients but is observed in only a minority of anti-PD-1 hypophysitis cases. Hyponatremia is common in both groups and may be a sign of hypothyroidism and/or AI. Generalized fatigue, loss of appetite, and myalgias/arthralgias are the primary symptoms in anti-PD-1 hypophysitis patients. Isolated central AI is typical in hypophysitis associated with anti-PD-1 agents; multiple anterior pituitary hormone deficiencies occur more commonly in hypophysitis associated with ipilimumab. Diabetes insipidus is extremely rare in both groups. Combination therapy (ipilimumab plus anti-PD-1 treatment) hypophysitis patients share characteristics of both monotherapy groups but appear to align much more closely with ipilimumab-associated hypophysitis (13,19).


Figure 1. Time (weeks) from the onset of therapy until the diagnosis of hypophysitis for patients treated with anti-PD-1 monotherapy (Panel A), ipilimumab monotherapy (Panel B) and ipilimumab plus concurrent nivolumab (Panel C). Reproduced with permission from Faje et al., European Journal of Endocrinology, 2019 (13).

No case of CPI-associated hypophysitis has been confirmed by surgical biopsy. An autopsy specimen has been described from a single case of tremelimumab-associated hypophysitis (tremelimumab is an IgG2-based anti-CTLA-4 antibody) (20). Clinical diagnoses of CPI-associated hypophysitis are based on the presence of new hypopituitarism and/or new reversible radiographic pituitary enlargement following treatment with CPIs in the absence of an alternative etiology. A serial decline in TSH values may precede the onset of ipilimumab-associated hypophysitis (19,21), and radiographic changes may also occur before the onset of clinical symptoms in these patients (14). Pituitary gland enlargement is generally mild, rarely compresses the optic apparatus, and may only be appreciated in some cases after comparison with prior imaging (19). Gland enlargement exists for only a brief time in some patients and may be missed if imaging is delayed (Figure 2) (22). Pituitary function test results may be influenced by factors such as recent or current exogenous glucocorticoid use, chronic illness and associated factors such as nutritional status, and/or the presence of an additional endocrine irAE in some patients. Due to the nonspecific character of their presenting symptoms, the frequent lack of localizing features, and typically negative imaging findings, diagnosis may be particularly challenging in anti-PD-1 hypophysitis cases; awareness is key.

Treatment and Outcomes
The precise therapeutic role of pharmacologic dosages of glucocorticoids in primary hypophysitis remains somewhat ill-defined. Early treatment strategies by many groups for ipilimumab-associated hypophysitis also incorporated supraphysiologic dosages of glucocorticoids. Multiple studies have reported that high dosages of glucocorticoids utilized for the treatment of irAEs do not appear to inhibit the antitumor efficacy of CPIs (23-30). Notably, some studies suggest that the development of an irAE may be associated with higher tumor response rates (24-25,28-30). Ipilimumab-associated hypophysitis likewise may be correlated with improved survival in patients with melanoma (22). The presence of an irAE may therefore represent a confounding factor in analyses assessing the impact of pharmacologic glucocorticoids on CPI antitumor efficacy. Ipilimumab-associated hypophysitis patients treated with physiologic or near-physiologic dosages of glucocorticoids appear to have improved overall survival, progression free survival, and time to treatment failure compared to those receiving high dosages. Additionally, higher dosages of glucocorticoids do not hasten the resolution of pituitary gland enlargement or help recover pituitary function in these patients (22). Limited additional studies have suggested similar potential negative antitumor effects from pharmacologic dosages of glucocorticoids in patients treated with CPIs (31-33). Given the apparent lack of benefit and questionable negative effects from higher dosages, physiologic glucocorticoid replacement appears to be appropriate for most patients with ipilimumab-associated hypophysitis and is also a reasonable management strategy for cases of hypophysitis associated with anti-PD-1 or PD-L1 therapy.

Central AI appears to be permanent in nearly all patients with CPI-associated hypophysitis. Other hormonal axes may recover in some patients (13,19,22). Treatment of central hypothyroidism with thyroid hormone replacement is appropriate at the time of diagnosis. Thyroid hormone weaning/withdrawal may be attempted at a future date in suitable patients. Gonadal hormone replacement can be deferred initially if appropriate in order to assess for potential axis recovery. Growth hormone replacement is contraindicated in patients with CPI-associated hypophysitis due to the underlying malignancy.

Mechanism of CPI-associated Hypophysitis and Potential Therapeutic Application
Ipilimumab-associated hypophysitis may occur secondary to direct drug binding and targeting of the pituitary gland. CTLA-4 is expressed by anterior pituitary cells in mice and humans. CTLA-4 expression levels in humans appear to be highly variable (20,34-35). Experiments in vitro with a melanoma cell line and ex vivo with T regulatory cells have demonstrated that ipilimumab can activate antibody-dependent cellular cytotoxicity (36-37). Mice injected with anti-CTLA-4 antibody show activation of the classical complement pathway and complement deposition onto anterior pituitary cells. In the previously noted autopsy case of tremelimumab-associated hypophysitis, the patient had a high expression level of CTLA-4 in the anterior pituitary, evidence of IgG2 binding, complement deposition, abundant macrophage infiltration as well as T and B cells consistent with type II and IV hypersensitivity reactions, and a lack of posterior pituitary gland involvement (consistent with the lack of diabetes insipidus observed in CPI-associated hypophysitis patients) (20,34). Coherent with the hypothesis that drug binding to anterior pituitary cells is important for the development of ipilimumab-associated hypophysitis, 2 studies of 25 patients with germline CTLA-4 mutations reported the presence of multiple severe autoimmune diseases but no case of hypophysitis/hypopituitarism (38-39). It is unknown whether PD-1 is expressed by pituitary cells. Regardless, anti-PD-1 medications are IgG4-based antibodies, which have a poor ability to bind complement and activate Fc receptor subtypes (40-41). The anti-PD-L1 medications atezolizumab and durvalumab cannot effectively bind Fc receptors or complement due to modified Fc domains.


Figure 2. Images from a patient with ipilimumab-associated hypophysitis. Radiographic pituitary gland enlargement developed over a 4-day period and resolved equally quickly followed by a gradual evolution to a nearly empty sella. Panels A) Prior to treatment with ipilimumab, B) 4 days prior to the diagnosis of hypophysitis, C) hypophysitis showing gland enlargement, D) 4 days after diagnosis, E) 19 weeks post-diagnosis. Adapted from Faje et al., Cancer 2018 (22).

The possibility that ipilimumab may directly target pituitary tissue suggests that the off-target effect of hypophysitis could potentially be harnessed for therapeutic treatment of selected aggressive pituitary tumors and carcinomas (19). In a recent case report, combination therapy with ipilimumab plus nivolumab was successfully utilized in the treatment of an ACTH-secreting pituitary carcinoma (42). Tumor mutational burden is correlated with response rates for ipilimumab and anti-PD-1 therapy within individual cancer types and across different malignancies (43-46). The authors of the case report hypothesized that a higher tumor mutation load following prior treatment with temozolomide was a significant factor for the tumor’s response to CPI treatment (42). Whether mutational burden or alternatively CTLA-4 expression is related to CPI treatment response in these tumors deserves further investigation.

Key Points

  • Checkpoint inhibitors (CPIs) occupy an expanding role in the treatment of a broad range of malignancies, and consequently CPI-associated hypophysitis has become a more frequently encountered irAE.
  • Hypophysitis is a more common side effect following treatment with ipilimumab compared to other CPIs.
  • The time of onset and clinical presentation of hypophysitis differs between ipilimumab and other CPIs. Hypophysitis develops relatively soon after ipilimumab treatment initiation, and radiographic pituitary gland enlargement, headache, and multiple anterior pituitary hormone deficiencies are observed more commonly in ipilimumab-associated hypophysitis. These clinical differences may result from distinct underlying mechanisms of toxicity.
  • Supraphysiologic dosages of glucocorticoids do not appear to be advantageous in the treatment of CPI-associated hypophysitis.
  • CPIs may have an application in the treatment of selected aggressive pituitary adenomas and carcinomas.

 

References:

  1. Haslam A, Prasad V. JAMA Netw Open. 2019; 2(5):e192535.
  2. Marquart J, et al. JAMA Oncol. 2018; 4:1093-8.
  3. Blank CU, Enk A. Int Immunol. 2015; 27:3-10.
  4. Mahoney KM, et al. Clin Ther. 2015; 37:764-82.
  5. Howlett TA, et al. Clin Endocrinol. 2010; 73:18-21.
  6. Buxton N, Robertson I. Br J Neurosurg. 2001; 15:242-5, discussion 245-6.
  7. Honegger J, et al. Neurosurg. 1997; 40:713-22; discussion 722-13.
  8. Imber BS, et al. Pituitary. 2015; 18:630-41.
  9. Leung GK, et al. J Neurosurg. 2004; 101:262-71.
  10. Sautner D, et al. Acta Neuropathol. 1995; 90:637-44.
  11. Albarel F, et al. Eur J Endocrinol. 2015; 172:195-204.
  12. Brilli L, et al. Endocrine. 2017; 58:535-41.
  13. Faje A, et al. Eur J Endocrinol. 2019; Epub
  14. Faje AT, et al. J Clin Endocrinol Metab. 2014; 99:4078-85.
  15. Min L, et al. Clin Cancer Res. 2015; 21:749-55.
  16. Ryder M, et al. Endocr Relat Cancer. 2014; 21:371-81.
  17. Snyders T, et al. Pituitary. 2019; 22:488-96.
  18. Eatrides J, et al. AACR advances in melanoma: from biology to therapy, abstract 2014.
  19. Faje A. Pituitary. 2016; 19:82-92.
  20. Caturegli P, et al. Am J Pathol. 2016; 186:3225-35.
  21. De Sousa SMC, et al. Pituitary. 2018; 21:274-82.
  22. Faje AT, et al. Cancer. 2018; 124:3706-14.
  23. Ascierto PA, et al. J Translational Med. 2014; 12:116.
  24. Beck KE, et al. J Clin Oncol. 2006; 24:2283-9.
  25. Downey SG, et al. Clin Cancer Res. 2007; 13:6681-8.
  26. Horvat TZ, et al. J Clin Oncol. 2015; 33:3193-8.
  27. Johnson DB, et al. Cancer Immunol Res. 2015; 3:464-9
  28. Ku GY, et al. Cancer. 2010; 116:1767-75.
  29. Weber JS, et al. J Clin Oncol. 2017; 35:785-92.
  30. Weber JS, et al. J Clin Oncol. 2008; 26:5950-6.
  31. Arbour KC, et al. J Clin Oncol. 2018; 36:2872-8.
  32. Chasset F, et al. Eur J Dermatol. 2015; 25:36-44.
  33. Margolin K, et al. Lancet Oncol. 2012; 13:459-65.
  34. Iwama S, et al. Science Trans Med. 2014; 6:230ra245.
  35. Faje A,et al. ENDO 2015, abstract.
  36. Laurent S, et al. J Trans Med. 2013; 11:108.
  37. Romano E, et al. Proc Natl Acad Sci U S A. 2015;112:6140-5.
  38. Kuehn HS, et al. Science. 2014; 345:1623-7.
  39. Schubert D, et al. Nat Med. 2014; 20:1410-6.
  40. Davies AM, Sutton BJ. Immunol Rev. 2015; 268:139-59.
  41. Vidarsson G, et al. Frontiers Immunol. 2014; 5:520.
  42. Lin AL, et al. J Clin Endocrinol Metab. 2018; 103:3925-30.
  43. Snyder A, et al. New Engl J Med. 2014; 371:2189-99.
  44. Van Allen EM, et al. Science. 2015; 350:207-11.
  45. Rizvi NA, TN, Chan TA. Science. 2015; 348:124-8.
  46. Yarchoan M, et al. New Engl J Med. 2017; 377:2500-1.