CNS demyelination and enhanced myelin-reactive responses after ipilimumab treatment

Ipilimumab is a monoclonal antibody that prolongs survival in patients with metastatic melanoma.¹ It targets the coinhibitory receptor cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4). CTLA-4 signaling induces a state of T-cell unresponsiveness, which facilitates tumor escape from immune surveillance. Blockade of CTLA-4 is believed to shift the immune status from T-cell exhaustion to a functional antitumor response. Anti-CTLA-4 therapy is associated with immune-related adverse events in 64% of patients. Autoimmunity involving the nervous system has a low incidence and manifests predominantly as peripheral inflammatory neuropathy.² We report new-onset inflammatory CNS demyelination in an ipilimumab-treated melanoma patient (figure), confirmed by histology and associated with enhanced responses of myelin-reactive CD4+ T cells.

Figure. Inflammatory CNS demyelination and enhanced myelin-reactive T-cell responses in a patient with melanoma treated with ipilimumab (IP).

Three-tesla MRI scans obtained 2 weeks (scan A) and 6 weeks (scan B) after completion of IP therapy. (A) Axial fluid-attenuated inversion recovery (FLAIR) at the level of the periatrial white matter of the lateral ventricles demonstrate an abnormal focus of increased T2 signal in the left posterior cingulate gyrus (arrow) with corresponding enhancement on T1 postgadolinium imaging (arrowhead). In addition, several scattered foci of abnormal FLAIR signal are seen in the periventricular and subcortical white matter consistent with small vessel ischemic disease. (B) FLAIR images demonstrate increased T2 signal changes seen in the intracanalicular segment of the left optic nerve, the left orbital gyrus of the inferior frontal lobe, and splenium of the corpus callosum (arrows). Changes related to stereotactic radiosurgery are present in the left posterior parietal lobe (arrowhead) with evidence of faint residual enhancement in this region on postcontrast sequences (not shown). The green dot indicates the site of the biopsy taken 1 week after this MRI. (C–F) Representative section of the corpus callosum biopsy. (C) Low-power magnification with a large area of demyelinated white matter, indicated by absence of myelin basic protein (MBP), and an abundance of infiltrating CD68+ macrophages (MBP, brown; CD68, purple). Scale bar 2 mm. (D) Detail from (C) shows MBP+ myelin debris within phagocytosing macrophages (arrowheads) as well as streaks of myelin still connected to axons (arrow). Scale bar 40 μm. (E) Immunolabeling for CD3 shows a cluster of tissue-infiltrating lymphocytes. Scale bar 150 μm. (F) High-power magnification demonstrates demyelinated but mostly preserved axons (arrowheads) separated by foamy macrophages (arrows). Bielschowsky stain. Scale bar 75 μm. C–E counterstained with hematoxylin (blue). (G, H) Phenotypic analysis of myelin-reactive CD4+ T cells from the IP-treated patient, patients with multiple sclerosis (MS), and healthy controls (HC). CCR6+ memory T-cell libraries were generated and cultured with irradiated autologous monocytes with or without a myelin peptide pool (MBP_85–99, myelin oligodendrocyte glycoprotein [MOG]_97–109, MOG_222–241, proteolipid protein [PLP]_30–49, PLP_129–148, and PLP_180–199) and with Candida albicans as control as described.⁷ (G) Heat map details functional responses of multiple T-cell libraries from a healthy control, a patient with MS, and the IP-treated patient, all age-, sex- and race-matched. Proliferation measured by ³H-thymidine incorporation on day 5 and culture supernatants examined on day 7 by ELISA for interferon (IFN)–γ, interleukin (IL)–17, granulocyte-macrophage colony-stimulating factor (GM-CSF), and IL-10. Data were z score–normalized for each parameter. A total of 96 T-cell libraries were tested for each individual. Each bar per column represents one T-cell library. Each row represents the different measurements that were performed on an individual library. (H) Percentage of proliferating and cytokine production of myelin-reactive T-cell libraries from the IP-treated patient compared to 14 matched patients with MS and healthy controls. Data are shown as mean ± SEM. CPM = counts per minute; N.D. = not detectable. *p < 0.05, **p < 0.01, unpaired Mann-Whitney test.

Methods.

A 76-year-old woman was diagnosed with stage III melanoma in 1997 for which she underwent surgery and received adjuvant vaccination. The patient did well until May 2014, when she was found to have kidney, spleen, and peritoneal masses. A kidney biopsy confirmed metastatic melanoma with BRAF^V600K mutation. The patient underwent 4 cycles of ipilimumab immunotherapy from June to August 2014. Restaging 2 weeks after completion of immunotherapy demonstrated partial responses but also a new enhancing lesion in the left posterior cingulate gyrus (figure, A) that was assumed to represent metastatic disease. Treatment was switched to dabrafenib and trametinib in mid-September and gamma knife stereotactic radiosurgery (SRS) was administered to the left posterior cingulate gyrus 2 weeks later (margin dose 20 Gy). Two weeks after SRS, the patient was readmitted with fatigue, memory loss, and vision changes and repeat MRI demonstrated new signal abnormalities in the left optic nerve, left inferior frontal lobe, and splenium of the corpus callosum, which extended into the parietal lobe and bordered on the SRS-treated lesion that by now was no longer enhancing (figure, B). A 5-day course of high-dose Solu-Medrol did not improve clinical symptoms. A biopsy of the left splenial lesion was performed, 1 month after radiosurgery and 2 months after the last ipilimumab dose, revealing acute/subacute inflammatory demyelination without evidence of tumor cells (figure, C–F). The patient was treated with cyclophosphamide and discharged to rehabilitation. She entered hospice care in January 2015, 3 months after admission, and died 1 month later.

Results.

We compared the functional profiles of the patient's myelin-reactive T cells to those of myelin-reactive T cells from 14 age-/sex-matched patients with multiple sclerosis (MS) and healthy controls by generating T-cell libraries from peripheral blood of all participants and measuring their responses to myelin peptides. Proliferation rates and proinflammatory cytokine production of myelin-reactive CD4+ T cells from the ipilimumab-treated patient were higher than that of myelin-reactive CD4+ T cells from healthy subjects and similar to those from patients with MS in response to myelin antigens (figure, G and H). Conversely, fewer T cells from the ipilimumab-treated patient and patients with MS than from healthy controls produced the anti-inflammatory cytokine IL-10, without differences in T-cell responses to Candida albicans (figure, G), confirming the specificity of the T-cell response to myelin peptides. Thus, myelin-reactive T cells from the ipilimumab-treated patient exhibited an enhanced autoimmune response similar to that observed in patients with MS and distinct from healthy controls.

Discussion.

Treatment with immune checkpoint inhibitors is currently revolutionizing cancer therapy.³ CTLA-4 is the best-characterized inhibitory pathway for T-cell activation and has been shown experimentally to play a key role in limiting T-cell reactivity in MS and its animal models.⁴ Moreover, a recent report of a patient with stable relapsing-remitting MS who had a severe relapse after ipilimumab treatment for metastatic melanoma suggests that blocking CTLA-4 signaling can induce MS exacerbations.⁵

The presentation of this patient is complicated by administration of SRS, which can induce inflammatory demyelination in susceptible individuals in the vicinity of the radiation field, typically 9–18 months after SRS.⁶ The white matter lesions reported here occurred 2 weeks after radiosurgery at multiple sites, including optic nerve and left frontal lobe, too early to be induced by SRS.

Finally, functional profiling of myelin-reactive CD4+ T cells revealed an inflammatory T_H1/T_H17 phenotype that resembled that of myelin-reactive T cells in patients with MS. This phenotype has been shown to be pathogenic in an animal model of MS, suggesting that demyelination in this patient was induced by an ipilimumab-associated pathogenic T-cell response against myelin. It would be interesting to examine whether this pathogenic phenotype is absent in ipilimumab-treated patients who did not develop inflammatory demyelination.

As the use of immune checkpoint inhibitors for cancer treatment increases, neurologists must be aware that immunotherapy can induce de novo inflammatory demyelination and worsen existing MS. Presence of new brain lesions after ipilimumab treatment requires careful evaluation to distinguish among brain metastases, SRS-induced encephalopathy, and inflammatory demyelination.