Sustained response to pan-BRAF inhibitor Plixorafenib (FORE8394, PLX8394) in a young adult with neuro-degenerative Langerhans cell histiocytosis

Introduction

Langerhans cell histiocytosis (LCH) is an inflammatory myeloid neoplasia caused by MAP Kinase pathway mutations, most frequently BRAF^V600E arising in myeloid precursors.¹ Patients with BRAF^V600E mutations more frequently develop pituitary and brain abnormalities and have a higher risk of relapse to initial therapy than patients without this mutation. Approximately 20% of LCH patients will develop neurodegenerative syndrome (LCH-ND), with LCH-associated abnormal CNS imaging (LACI) consisting of MRI changes (T2/FLAIR signal in the cerebellum, pons or basal ganglia) and only some also with LCH-associated CNS symptoms (LACS) (e.g. ataxia, dysmetria, dysarthria, psychologic, and neurodevelopmental changes)² that can arise coincident with or years following systemic disease.³ Historically, LCH-ND was thought to be caused by abnormal CNS immune responses. However, our group has shown that the active demyelinating process is caused by brain infiltration by BRAF^V600E-mutated monocyte-derived CD11a+ perivascular macrophages. Autopsy of the brain from a LCH-ND patient identified >10% of cells in the brainstem and cerebellum were BRAF^V600E+.² Therefore LCH-ND should be considered a CNS-specific LCH lesion. We previously reported responses to LCH-directed chemotherapy (cytarabine⁴) and MAPK inhibitors (MAPKi)⁵ in LCH-ND patients. The potential of long-term responses to MAPK inhibition for patients with LCH-ND and importance of CNS bioavailability of MAPKi for durable responses is unknown.

Here we report a patient who initially presented with skin, bone, and pituitary LCH with new onset diabetes insipidus (DI). She subsequently developed brain and spinal cord lesions with dramatic neurologic deficits that failed to respond to multiple chemotherapeutic regimens. Ultimately, she enrolled into PLX120–03, a phase 1/2a clinical trial of a brain-penetrant BRAF inhibitor (plixorafenib; FORE8394, PLX8394). She experienced a dramatic and sustained response to plixorafenib, lasting >6.5 years with no significant toxicity.

Case Presentation

The patient presented at 6.5 years old with a hyperpigmented, post-auricular scaly rash. Three months later she developed a painful left temporal lesion; biopsy was consistent with LCH. Skeletal survey was otherwise negative as were blood, liver, and chemistry studies. Thus she was classified as multisystem LCH-CNS Risk (no hematologic risk organ involvement). She received vinblastine/prednisone treatment for one year according to the Histiocyte Society LCHIII protocol⁶ with rash and bony lesion resolution. Nine months off-therapy, she developed polyuria and polydipsia. She was diagnosed with DI and started on desmopressin. While her neurologic exam was normal, a brain MRI showed a slightly thickened pituitary stalk, absence of bright spot, and cerebellar T2 hyperintense signals--consistent LACI in the context of clinical CNS changes (DI with normal neurologic exam). A new parietal lesion appeared the following month and was treated with surgery only. Repeat MRI later that year showed stable LACI, and she remained asymptomatic for 46 months. At 12.7 years old she developed a new mandibular lesion and was treated with cytarabine until persistent fevers prompted a change to cladribine. Tremors and ataxia began as cladribine treatment ended and progressed over the next two months; then she deteriorated rapidly.

Peripheral blood BRAF^V600E was positive early in her course, and cell-free CSF BRAF^V600E was detectable by qPCR when she developed LCH-ND imaging and symptoms. In an institutional series of patients with LCH-ND, only 2/20 cases had detectable BRAF^V600E in the CSF--both with spinal cord involvement, including this patient.²

Over the next half-year, she received multiple regimens including intrathecal cytarabine/hydrocortisone/methotrexate, clofarabine, intravenous immunoglobulin, alemtuzumab, hydroxyurea, and dexamethasone. Imaging during this time showed worsening changes extending beyond the cerebellum into the cerebrum, brain stem, and spinal cord (Figure 1A-C). Collectively she experienced multiple therapy-related toxicities including a 90-day hospitalization for E. coli urosepsis complications. During this time her blood BRAF^V600E quickly became undetectable, but her CSF BRAF^V600E variant allele level increased from 0.11 to 0.25, possibly reflecting poor CSF penetration of the various agents (Table 1). After eight months of worsening neurologic function (unintelligible speech, poor balance, neurogenic bladder, emotional lability, impulsivity) and progressive brain MRI changes, she was started on vemurafenib (400 mg/m² or 480 mg twice daily [BID]) at 14 years old. Within one month the patient was stronger, with improved neurologic function (e.g., better speech clarity, head control, bowel/bladder control, standing unassisted, improved walking). CSF BRAF^V600E also became undetectable. However as there was concern for more seizures, her dosage was increased to 720 mg (550 mg/m²) BID. Brain MRI improved after three cycles. However after six cycles, progress had plateaued so rituximab was briefly added (as it has improved the psychological problems of some LCH-ND patients)⁷ but had no obvious benefits. After 13 months of vemurafenib, she developed grade 2 intolerable uveitis, so vemurafenib was held. (She had no other toxicities from vemurafenib, including no laboratory, skin, or EKG abnormalities.) While the uveitis improved over four weeks, the patient quickly weakened, so vemurafenib was resumed and gradually re-escalated without recurrence of any symptoms. However, after twenty months of vemurafenib therapy, the patient was found to have acute left hip osteonecrosis so vemurafenib was discontinued due to concern of an associated adverse event. At nine months off vemurafenib therapy while the brain MRI was stable, her ataxia rating score (using the International Cooperative Ataxia Rating Scale [ICARS] for pharmacological assessment of the cerebellar syndrome)⁸ rose to 45, so dabrafenib 100mg BID was added to maintenance hydroxyurea. While this had fewer side effects than vemurafenib (limited rash), none of the patient’s neuro/psychiatric symptoms improved. Thus dabrafenib was stopped after 2.5 months. When the PLX120–03 trial (NCT02428712) opened in pediatric patients, she was enrolled and assigned to the dose-escalation group 1B of single agent plixorafenib at the 250 mg/m² BID dose level.

Figure 1.

(A-B) Axial T2 FLAIR images of the brain obtained at 13.5 years old due to neurologic deterioration demonstrate abnormal T2 hyperintensity of the central grey matter (A, arrows) as well as the dentate and peridentate parenchyma with extension into the pontine brainstem (B, arrows). (D-E) Axial T2 FLAIR images of the brain at similar levels 6 months into plixorafenib treatment demonstrates marked improvement in signal abnormality (D-E, arrows). Sagittal T2 weighted image at 13.5 years old of the cervical and thoracic spine (C) demonstrates abnormal T2 hyperintensity and expansion of the cervical and thoracic cord consistent with longitudinally extensive myelitis. Sagittal T2 weighted image 6 months into plixorafenib treatment demonstrates normalization of the cervical (F) and thoracic cord (G).

Table 1.

Detailed BRAF testing summary. CSF samples indicated in italics.

Time (months)	Sample Type	Evidence Type	Result	VAF%*	Clinical Comments
0	Blood	BRAF QPCR	None Identified		4 years from LCH diagnosis, no active therapy
0.2	Blood	BRAF QPCR	BRAFV600E	0.02
14.5	Blood	BRAF QPCR	BRAFV600E	0.04	new mandibular lesion; baseline prior to starting monthly cytarabine
14.5	Blood	ddPCR	BRAFV600E	0.03
17.4	Blood	BRAF QPCR	BRAFV600E	0.02	Monitoring
21.6	Blood	BRAF QPCR	BRAFV600E	0.04	Monitoring after switching to cladribine and prednisone
25.4	CSF	BRAF QPCR	BRAFV600E	0.11	Tremor/balance problems, MRI with diffuse spinal cord involvement; started clofarabine, dexamethasone, and intrathecal chemotherapy
25.4	Blood	BRAF QPCR	BRAFV600E	0.016
26.4	Blood	BRAF QPCR	None Identified		Monitoring; stabilized balance and speech
27.9	Blood	BRAF QPCR	None Identified		Monitoring; gradual worsening balance and speech
29.9	Blood	BRAF QPCR	None Identified		2 months into switch to dexamethasone and alemtuzumab
31.25	Blood	BRAF QPCR	None Identified		Monitoring; some neurologic improvement
32.25	Blood	BRAF QPCR	None Identified		Detectable CSF BRAFV600E with negative blood testing
32.25	CSF	BRAF QPCR	BRAFV600E	0.25
32.6	Blood	BRAF QPCR	None Identified		Monitoring; started vemurafenib 2 weeks later
33.7	CSF	BRAF QPCR	None Identified
34.25	CSF	BRAF QPCR	None Identified		Monitoring; improved neurologic function
34.25	Blood	BRAF QPCR	None Identified
35.5	CSF	BRAF QPCR	None Identified		Monitoring
35.5	Blood	BRAF QPCR	None Identified
36.5	Blood	BRAF QPCR	None Identified		Monitoring; increasing vemurafenib to max dose
40.25	Blood	BRAF QPCR	None Identified		Neurologic improvements plateau + new double vision, baseline prior to adding rituximab
40.25	Blood	Clinical BRAFV600E Test	None Identified
42.4	Blood	BRAF QPCR	None Identified		Monitoring
45.5	Blood	BRAF QPCR	None Identified		Monitoring
46.5	Blood	BRAF QPCR	None Identified		Monitoring; overall marked improvements since starting vemurafenib
47.9	Blood	BRAF QPCR	None Identified		Monitoring
50.25	Blood	BRAF QPCR	None Identified		Monitoring
81.6	Blood	Clinical BRAFV600E Test	None Identified		Baseline prior to starting Plixorafenib
106.1	Blood	Clinical BRAFV600E Test	None Identified		Monitoring
114.1	Blood	Clinical BRAFV600E Test	None Identified		Monitoring
120.1	Blood	Clinical BRAFV600E Test	None Identified		Monitoring
122.3	Blood	Clinical BRAFV600E Test	None Identified		Monitoring; achieved ataxia score low of 28
127	Blood	Clinical BRAFV600E Test	None Identified		Monitoring
139.4	Blood	Clinical BRAFV600E Test	None Identified		Monitoring

^*limit of assay detection: 0.02%

Plixorafenib was initiated just before her 18^th birthday, starting at a dose of 300 mg in the morning and 375 mg in the evening. Her baseline ataxia score was 42 and improved to 37 after one month of therapy. The ataxia score continued to improve to a low of 28 at start of cycle 40, and then stabilized in the mid-30s the second half of the study, which was in part thought to be related to physical therapy availability (Figure 2). She briefly developed a chalazion in cycle 35 that resolved with supportive care. Otherwise the patient tolerated plixorafenib well, with mild, intermittent acneiform rash and dry skin (responded to topical creams) and mild dry eyes (responded to artificial tears). She had no neoplasms or laboratory abnormalities. All brain and spine MRIs showed stable disease on plixorafenib (Figure 1D-G), and peripheral blood BRAF^V600E PCRs all remained negative. Her diabetes insipidus, neurogenic bowel/bladder and movement disorders were stable. She continued with physical therapy throughout the study as tolerated and available. She graduated from high school, then community college during the study, and is currently employed as a bookkeeper. The patient has been on plixorafenib for >82 cycles (6.5 years).

Figure 2.

Patient’s ataxia score and functional status at baseline and while on plixorafenib.

Statement of Publication Consent and Human Investigation Protocol Approval

We obtained informed consent from the patient and her guardian for publication. Investigations were performed after protocol approval by a local Human Investigations Committee.

Discussion

We report a patient with profound LCH-ND experiencing prolonged positive clinical response based on improved ataxia score, clinical improvement and disease stabilization with single agent plixorafenib, following multiple therapies including BRAF inhibitors (BRAFi) with less CNS penetration. She had rapid improvement of clinical symptoms and ataxia score, with radiographic disease stabilization including the unusual involvement of her spinal cord. (See Supplemental Figure for entire treatment time course.)

Vemurafenib, dabrafenib, and plixorafenib are all BRAF targeted small molecules with slightly different mechanisms of action. Both vemurafenib and dabrafenib are kinase inhibitors designed to inhibit V600E mutant BRAF,^9–11 while plixorafenib inhibits BRAF with or without V600 mutations. Furthermore vemurafenib/dabrafenib resistance often result in BRAF dimerization and paradoxical downstream MAPK activation, while plixorafenib also inhibits dimerization and does not activate MAPK signaling.¹² Plixorafenib has a LogD of 2.86 (indicating acceptable lipophilicity per Lipinski’s rules), is a P-gp and BCRP inhibitor (personal communication, FORE Biotherapeutics), and patients with gliomas have demonstrated benefit,¹³ which support its CNS penetrance. Our patient remains on plixorafenib following study closure due to continued benefit, good tolerability, and continuous MAPK pathway suppression to prevent recurrent disease. Her sustained improvement over 6.5 years is the longest we have seen reported for patients with LCH-ND receiving single agent BRAFi.^14–17 As her lowest ataxia score was achieved after 40 cycles, this suggests that patients may have continued improvement in LCH-ND with plixorafenib, even after many months of treatment.

Identification of the BRAF^V600E mutation in tumor tissue and in peripheral blood mononuclear cells (PBMC) at diagnosis have prognostic implications including extent of disease, risk of front-line treatment failure, and risk of developing LCH-ND.^18,19 In a recent retrospective series, we found that >25% of patients with BRAF^V600E+ PBMC at diagnosis ultimately developed LACI or LACS.²⁰ Accordingly, percentage of BRAF^V600E+ PBMC tracks with clinical responses in patients treated with chemotherapy.^18,21,22 This appears to be less consistent in patients treated with MAPKi, where molecular responses are less predictable. In the literature, high-sensitivity CSF BRAF qPCR has only been reported in two patients (including this one).² We were able to confirm that her CSF BRAF became negative with vemurafenib treatment and stayed negative two months into treatment. As no further CSF BRAF testing was performed, her CSF BRAF could possibly have been negative prior to starting plixorafenib. BRAFi/MAPKi do not cure the disease; rather they just suppress it, and the disease progresses once the agent is stopped.^5,23,24 Furthermore, some patients are now developing LCH-ND after prolonged BRAFi/MAPKi therapy.²⁵ Fortunately, our patient has had no toxicities with plixorafenib and a durable partial response over years.

In summary, we believe this highlights the potential for CNS-penetrant BRAFi to have clinically meaningful, safe, and markedly prolonged response in patients with LCH-ND with no evidence of development of resistance to the single agent. Progressive LCH-ND represents a devastating condition with few examples of therapeutic promise.³ Prospective studies with plixorafenib and/or other CNS-penetrant MAPKi for LCH-ND are warranted. A phase 2 basket study of plixorafenib in participants ≥10 years of age with cancers harboring BRAF alterations is currently enrolling (NCT05503797).

Supplementary Material

PV Supplemental Figure 1

Supplemental Figure 1. Timeline of patient’s overall LCH course and ataxia rating scoring (top), and treatment and BRAF testing (bottom). Abbreviations: Alem: alemtuzumab; AraC: cytarabine; ARS: ataxia rating scale; BRAF: quantitative or digital droplet PCR (qPCR or dPCR) or clinical test for BRAF^V600E of peripheral blood (PB) or cerebrospinal fluid (CSF); CLD: cladribine; Clo: clofarabine; Dabraf: dabrafenib; Dex: dexamethasone; HC: hydrocortisone; HU: hydroxyurea; IT: intrathecal; IVIg: intravenous immunoglobulin; LACI: LCH-associated abnormal CNS imaging; MRI: magnetic resonance imaging; neg: negative; Plix: plixorafenib; Pred: prednisone; Ritux: rituximab; Rx: treatment; Sx: symptoms; VBL: vinblastine; Vem: vemurafenib.