PD-1 Inhibitor Immune-Related Adverse Events in Patients With Preexisting Endocrine Autoimmunity Free

Abstract

Context

Immune checkpoint inhibitors, including monoclonal antibodies directed against programmed cell death protein 1 (PD-1) and its ligand, have emerged as beneficial cancer immunotherapies. These therapies are known to cause immune-related side effects; however, their role in patients with a preexisting autoimmune disease is not clear.

Case Description

We describe two cases of anti-PD-1 immune-related adverse events. A 52-year-old male with longstanding type 1 diabetes (T1D), long-term stable kidney transplant, and hypothyroidism received two separate anti-PD-1 monoclonal antibodies for metastatic melanoma. The patient developed acute kidney graft rejection requiring hemodialysis and worsening of autoimmune hypothyroidism 3 weeks after starting treatment. He continued anti-PD-1 treatments and remained on hemodialysis and increased levothyroxine dosage. The second case is a 62-year-old male with no previous history of diabetes who received anti-PD-1 treatment and developed severe diabetic ketoacidosis (DKA) 5 days following the start of therapy. Further laboratory testing revealed high titer antibodies directed against glutamic acid decarboxylase. These antibodies, which were of the IgG isotype and involved in memory immune responses, were likely present before anti-PD-1 treatment. He also had human leukocyte antigen genes that confer T1D genetic risk. Despite normal pretreatment blood glucose levels and HbA1c, the patient requires permanent exogenous insulin treatment.

Conclusion

Patients with preexisting endocrine autoimmunity may have more frequent and severe immune-related side effects with anti-PD-1 treatment. Given the morbidity and mortality associated with solid organ transplant rejection and DKA, clinicians caring for patients receiving these state-of-the-art therapies need to be aware of the potential adverse events.

Immunotherapy, including immune checkpoint blockade, is becoming a highly effective treatment of multiple malignancies refractory to conventional chemotherapies. Programmed cell death protein 1 (PD-1) inhibitors inhibit the interaction between PD-1 and its ligand, thereby activating immune responses toward cancer cells. In a normal adaptive immune response, immune checkpoint inhibition is in place to ensure that immune cells do not harm the host when responding to a foreign antigen. Interfering with this mechanism can cause immune-related adverse events directed against self-tissues (1). Pembrolizumab and nivolumab are two of the most commonly used human monoclonal antibodies directed against PD-1 and are approved by the US Food and Drug Administration to be used in several stage IV cancers, including melanoma.

Clinical trials that led to the US Food and Drug Administration approval of these agents often excluded patients with a known autoimmune disease such as type 1 diabetes (T1D) or an organ transplant (1, 2). Thus, in clinical practice, the potential exists to encounter immune-related adverse events more often than registration trials. Currently, there are no guidelines for the treatment of patients with immune checkpoint inhibitors in the setting of preexisting autoimmunity or a prior organ transplant. We report cases of acute kidney transplant rejection and worsening of autoimmune hypothyroidism in a patient with longstanding T1D having previously stable kidney and thyroid function and a second case of autoimmune diabetes as a rapid consequence of anti-PD-1 treatment of metastatic melanoma.

Cases

Case 1

A 52-year-old white male with a 45-year history of T1D, diabetic nephropathy with a related living donor–kidney transplant 10 years ago, and hypothyroidism on a stable dose of levothyroxine for 15 years was diagnosed with stage IV melanoma. For T1D management, he was on continuous subcutaneous insulin infusion with an HbA1c of 8.8%. He did not require hemodialysis after the kidney transplant, and his immune suppressive regimen included prednisone 10 mg/d, mycophenolate 1080 mg/d, and tacrolimus 3 mg/d with a stable serum creatinine of 1.1 mg/dL. He was tapered off mycophenolate and tacrolimus, but continued prednisone, in preparation to begin PD-1 inhibitor therapy. A week later, he received one IV infusion of 200 mg pembrolizumab followed by 240 mg of IV nivolumab in 2 weeks. Four days after his last infusion, he presented with confusion, swelling of lower extremities, and decreased urination with laboratories consistent for an acute kidney injury (Table 1). The patient was admitted to the intensive care unit and received hemodialysis. Computed tomography imaging showed increased inflammation and edema of the transplanted kidney, and a biopsy was consistent with acute graft rejection. He has required ongoing hemodialysis over the ensuing 6 months.

For hypothyroidism, he was on a stable dose of 75 μg/d levothyroxine with a TSH value of 3.8 mIU/L before anti-PD-1 treatment. Upon presentation with acute kidney failure, his TSH was 62.7 mIU/L; his levothyroxine dose was increased to 175 μg/d and TSH gradually decreased to 6.8 mIU/L. In regard to his blood glucose management, insulin needs temporarily increased during hospitalization but returned to baseline thereafter. Evaluation of T1D autoantibodies was negative (Table 1), which is not unexpected given his longstanding disease. Autoantibodies for celiac disease (gluten sensitivity) and Addison disease (adrenal insufficiency) were not present. The patient continued nivolumab therapy, receiving eight additional infusions over 6 months, with a partial response to his tumor load.

Case 2

A 62-year-old white male with no previous history of diabetes presented with abdominal pain, confusion, nausea, and vomiting for 5 days after receiving a single IV infusion of nivolumab for metastatic melanoma. His laboratories upon hospital admission showed substantial hyperglycemia, a metabolic acidosis with an increased anion gap, and the presence of serum and urine ketones consistent with severe diabetic ketoacidosis (DKA) (Table 2). One week prior to starting anti-PD-1 therapy, he had normal blood glucose and HbA1c values. He was admitted to the intensive care unit for DKA management, discharged on multiple daily insulin injections, and continues exogenous insulin treatment 6 months later. We evaluated T1D-associated autoantibodies at the time of his DKA presentation, which revealed high levels of antibodies directed against glutamic acid decarboxylase (GAD65), which is a major autoantigen in T1D. The antibody measurements were done with highly sensitive and specific fluid-phase radioimmunoassay, which measure IgG antibody isotypes. Because IgG antibodies are involved in memory immune responses and the short interval from anti-PD-1 treatment to DKA presentation, these antibodies directed against GAD65 were likely present in the patient before treatment. Further phenotyping indicated that these antibodies were polyclonal and predominantly of the IgG₁ (56%) and IgG₂ (35%) subclasses. In addition to T1D-associated antibodies, HLA genes, especially the class II molecules DR and DQ, confer substantial genetic risk for prototypical childhood-onset T1D, and our patient was homozygous for the high-risk HLA-DR3/DQ2 (DRB*0301, DQA*0501, DQB*0201) haplotype. Autoantibodies for celiac and autoimmune Addison disease were negative. The patient continued nivolumab therapy, receiving 11 additional infusions over 6 months, with a good response to his tumor burden.

Discussion

Clinical trials for anti-PD-1 treatment have generally avoided enrolling patients with preexisting autoimmune disorders or those with organ transplants, and some trials even excluded patients with a strong family history of autoimmune disorders (1–3). Our two cases highlight the rapid-onset and severe autoimmune-related adverse events that can occur with anti-PD-1 therapy. Both cases had autoimmunity directed toward pancreatic islets, with the first case having established T1D, and the second case presenting with unrecognized autoantibodies directed toward GAD65. In a recent prospective analysis of patients treated with anti-PD-1 therapies having preexisting autoimmunity (n = 45), mainly vitiligo and rheumatologic diseases, immune-related adverse events were more frequent than in those without an autoimmune disease (3). However, this series of patients included only one with T1D, highlighting the need to follow more patients with endocrine autoimmunity treated with anti-PD-1 therapies.

Premarketing studies reported low rates of diabetes but a higher incidence of other endocrine side effects such as autoimmune thyroid disease (4). PD-1 inhibitor–induced autoimmune diabetes is unique, with a rapid onset of presentation and severe DKA (5, 6). Previous reported cases are indeed the immune-mediated form of diabetes as diabetes-specific T cells are present in the peripheral circulation, and some patients have T1D autoantibodies at presentation (6). The estimated frequency of immune checkpoint inhibitor–induced autoimmune diabetes is 1% to 3% at large medical centers (7). In March 2018, immune checkpoint inhibitor labels were updated to include T1D as a serious side effect. In our second case, we found strong positivity for GAD65 autoantibodies, >35 times the upper limit of normal. Based upon our presented case and review of the literature, we propose the hypothesis that a subset of patients developing PD-1 inhibitor–induced diabetes likely has preexisting islet autoantibodies, which may be an early form of latent autoimmune diabetes of adulthood. In our first case, we would not anticipate the presence of T1D-associated autoantibodies or worsening of disease because T1D was present for 45 years and there was likely very little to no residual pancreatic β cells remaining.

In clinical practice, screening strategies are needed to assess risk of developing anti-PD-1–induced diabetes and monitoring to prevent DKA. We would advocate for baseline screening of blood glucose level and HbA1c, T1D-associated autoantibodies, and HLA-DQ and HLA-DR genes because a recent literature review indicated that 14 of 21 patients with immune checkpoint inhibitor–induced diabetes had high T1D-risk HLA genotypes (8). However, more research is needed to identify factors that predispose risk for developing PD-1 inhibitor–induced autoimmune diabetes, including registries that assess islet autoantibodies and HLA genes in more patients. At-risk patients also need to be monitored during treatment to prevent DKA, including education about signs and symptoms of hyperglycemia and DKA along with monitoring blood glucoses. Very rapid-onset cases, such as case 2 presented here, may not be detected by intermittent blood glucose monitoring; however, continuous glucose monitoring systems may prove helpful in monitoring the timing and duration of blood glucose elevations with PD-1 inhibitor therapy. Besides metabolic measures, assays that evaluate activated lymphocytes involved in pancreatic islet destruction, such as self-antigen–specific T and B cells (9), need to be investigated before clinical onset. This line of research will aid in understanding the mechanisms of PD-1 inhibitor–induced diabetes development and allow for testing therapies to block this unwanted side effect.

Another remarkable finding in our first case was the rejection of a stable kidney transplant. Cardiac and renal allograft rejection has been reported in cases with PD-1 inhibitor therapy (10–12). Similar to our case, these previously reported cases of stable kidney transplant rejection have required immediate dialysis and result in permanent graft rejection (11, 12). However, high-dose immune suppression was able to effectively salvage a cardiac allograft following a single dose of nivolumab (10). Organ transplant rejection can occur in patients with and without baseline autoimmunity after receiving anti-PD-1 treatment.

Additionally, we observed an acute increase in TSH with PD-1 inhibitor therapy. In registration trials, new-onset hypothyroidism and thyroiditis were reported, with subsequent worsening of hypothyroidism, as indicated by an increase in TSH, have also been reported (13). Our patient in case 1 maintained a normal TSH without hypothyroid symptoms on a levothyroxine dose that was not a full replacement based upon body weight. Following anti-PD-1 treatment, the 20-fold increase in TSH and need for a full replacement dosage of levothyroxine likely represents the resultant destruction of his remaining thyroid tissue.

In conclusion, we report the rapid and severe onset of PD-1 inhibitor–induced immune-related adverse events in a patient with preexisting T1D and one with islet autoimmunity. Further research is needed to identify risk factors for the development of PD-1 inhibitor–induced autoimmunity to develop screening algorithms. Clinicians should be aware of the immune-related adverse events with PD-1 inhibitors and discuss the risks and benefits with patients before initiation of treatment, especially in high-risk patients such as those with an existing autoimmune disease or organ transplant.

Abbreviations:

Abbreviations:
 
• DKA

  diabetic ketoacidosis

 
• GAD65

  glutamic acid decarboxylase

 
• PD-1

  programmed cell death protein 1

 
• T1D

  type 1 diabetes

Acknowledgments

Financial Support: This work was supported by the National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases Grants DK108868, DK110845, and DK032083 (to A.W.M.).

Disclosure Summary: The authors have nothing to disclose.

References