A rare subtype of hypereosinophilic syndrome successfully managed with anti-IL5–based therapy
History and presentation
A 54-year-old Caucasian man was hospitalized for 3 weeks in Germany in October 2012 due to herpes zoster in addition to Chlamydia pneumoniae infection, which was successfully treated with intravenous (IV) acyclovir. He experienced post-infectious fatigue, but chest CT scan was normal.
He also had a history of gout in 2001, myalgia and cramps of unknown origin in 2012, and catheter ablation of the left-sided concealed accessory pathway for tachycardia in 2012 and 2015.
After moving office in 2014, the patient started to experience bloating, flatulence and abdominal discomfort. He also developed anterior uveitis of unknown cause, cough productive of purulent sputum, and shortness of breath. Results of upper and lower endoscopy were normal. Various strains of mould, including Aspergillus, were found in his new office.
Investigations and treatment
High-resolution CT scan in October 2014 in UK showed cylindrical bronchiectasis in the upper and lower lobes of the lungs with minimal wall thickening suggestive of bronchiectasis with asthmatic element. Bronchoscopy showed endobronchial oedema and narrowing of orifices in the mid and lower lobes, with mucus exudation. Bronchial washings showed no infection, and cytology was negative.
In 2015, the patient complained of headache, short-term memory loss, episodes of confusion, loss of visual focus, photosensitivity, fatigue, pain on the right side of the neck and lower back, and dry and swollen nose. Wheezing was heard in his chest. Brain MRI was normal. He was prescribed budesonide, which provided only a slight improvement of the respiratory symptoms.
The patient was found to have mild lymphopenia during a follow-up visit in April 2015, with CD3+ T cell count of 516 cells/µL (normal range, 840–3,060 cells/µL), CD4+ T cell count of 319 cells/µL (normal range, 490–1,740 cells/µL), CD8+ T cell count of 173 cells/µL (normal range, 180–1,170 cells/µL), CD19+ B cell count of 69 cells/µL (normal range, 110–660 cells/µL), and absolute lymphocyte count of 811 cells/µL (normal range, 850–3,900 cells/µL). His immunoglobulin G1 (IgG1) level was also slightly low (320 mg/L; normal range, 382–929 mg/L). IV immunoglobulin (IVIg) 30 g was therefore administered in June 2015.
CT coronary angiography in July 2015 revealed mild stenosis (20–30 percent) at the proximal left circumflex and right coronary arteries. No antineutrophil cytoplasmic antibodies or antinuclear antibodies were detected. Eosinophil count was 800 cells/µL (10.7 percent).
In September 2015, the patient was hospitalized for bilateral pneumonia and respiratory failure. C-reactive protein was elevated. CT scan showed consolidation in the right and left upper lobes with enlarged and probably reactive mediastinal lymph nodes. (Figure 1) Bronchoalveolar lavage (BAL) was negative for fungi, tuberculosis and polymerase chain reaction (PCR) respiratory pathogens but showed abundant eosinophils. CT scan showed bilateral maxillary and ethmoid sinusitis.
In October 2015, the patient’s eosinophil count had risen to 19,700 cells/uL (67.8 percent). His pneumonia improved with azithromycin and levofloxacin, although he continued to expectorate mucus plugs. Sputum showed presence of Candida only. Extensive tests for Aspergillus, bacteria, fungi, parasites and viruses were negative. Testing for 1,3-beta-glucan (marker for invasive fungal infections) was also negative. Allergic bronchopulmonary aspergillosis (ABPA) was excluded as Aspergillus-specific IgE and IgG were negative. He was given a course of anidulafungin, followed by an empirical course of albendazole, praziquantel and ivermectin. No improvement was noted in his respiratory symptoms.
Bone marrow biopsy showed changes compatible with hypereosinophilic syndrome (HES). He was therefore started on high-dose IV hydrocortisone (200 mg TDS), which resulted in immediate decrease in peripheral blood eosinophil count. Tests for several molecular variants of HES (ie, FIP1L1-PDGFRA fusion, PDGFRB rearrangement, CBFB rearrangement, FGFR1 rearrangement), BCR-ABL1 fusion (for chronic myelogenous leukaemia, acute lymphoblastic leukaemia and acute myelogenous leukaemia), and KIT D816V mutation (for systemic mastocytosis) were negative.
The patient’s peak flows (PFs) improved on high-dose steroids, but he continued to expectorate mucus plugs. His prednisolone dose was reduced to 30 mg/day. Blood and sputum samples showed PCR evidence of Malassezia restricta yeast infection. He was started on a course of itraconazole, but no significant improvement was observed.
The patient was treated with intermittent antibiotics along with increased steroid doses as required, as well as montelukast 10 mg daily plus nebulized budesonide 0.5 mg BID and ipratropium bromide/salbutamol solution 2.5 mL BID. However, he was hospitalized in April 2016 and October 2016 due to chest exacerbations.
In January 2017, he was admitted to a specialist respiratory hospital in Germany where Pseudomonas aeruginosa was detected from BAL. He was treated initially with steroids and subsequently with nebulized colistimethate sodium upon returning to Hong Kong. Bronchoscopy performed in Germany revealed intraepithelial laryngeal neoplasia, which was subsequently treated in Hong Kong.
In June 2017, the patient underwent functional endoscopic sinus surgery together with surgical correction of a deviated nasal septum. His lymphopenia had largely resolved, with only a mild decrease in CD8+ T cells (217 cells/µL).
He was readmitted in November 2017 for respiratory exacerbation, and in June 2018 for sudden-onset left otitis media resulting in a burst eardrum within a few hours, which was managed conservatively and healed by itself without intervention.
In July 2018, he was readmitted for suspected respiratory infection, with various pathogens inconsistently identified on repeat testing. Skin prick test was positive for mixed moulds. A very marked non-immediate reaction was noted hours after skin prick testing for Candida albicans, but the clinical significance of this observation was unclear. Total IgE level was raised slightly, at 106 KU/L (normal range, <87 KU/L). He was started on monthly IVIg in August 2018 in light of his low IgG1 level.
Although he was readmitted again in September 2018 for suspected respiratory infection, he reported fewer exacerbations following the initiation of IVIg treatment. In May 2019, IgG1 level was normalized, at 733 mg/L (normal range, 382–929 mg/L). He has been on IVIg 40 g every 4 weeks since then.
Apart from IVIg and prednisolone, the patient has also been receiving antiasthmatic treatment since 2015. During this period, there have been fluctuations in peripheral eosinophil count (10–700 cells/µL), significant diurnal variations in PFs, as well as frequent expectoration of light green gelatinous sputum and blockage of sinuses and ears.
The initial diagnosis was either Churg-Strauss syndrome or HES. Many of the molecular changes in HES had already been excluded, but there was also no evidence of vasculitis or granulomata to support the diagnosis of Churg-Strauss syndrome. Nevertheless, eosinophil is the key pathobiologic effector cell type in both disorders.
The patient’s peripheral eosinophil count was mostly within normal range in 2018, but this might be attributed to the use of steroids, which was expected to reduce eosinophil level. His fractional expired nitric oxide (FENO) level was 45 ppb (normal range, <25 ppb), suggesting persistent eosinophilic airway inflammation.
The patient continued to experience upper and lower airway respiratory symptoms in 2019, which were possibly attributable to airway eosinophilia. He was therefore started on a course of benralizumab (30 mg subcutaneously). About 2 weeks later, his PFs started to improve. (Figure 2) In addition, sputum production decreased and peripheral eosinophil was undetectable. The patient also felt more physically energetic.
Meanwhile, our team identified a rare, recently described genetic abnormality associated with HES, namely PCM1-JAK2 fusion, in the patient’s peripheral blood sample. (Figure 3)
The patient will undergo another bone marrow examination imminently to guide further management. Meanwhile, benralizumab treatment is continued.
While it was difficult to obtain precise details of the patient’s health history and laboratory tests, the root cause of his disease seemed to be pulmonary eosinophilia and asthma with possible multisystem involvement and mild immunodeficiency, which could have resulted in opportunistic infections.
Peripheral hypereosinophilia is defined as absolute eosinophil count >1,500 cells/µL and/or pathological confirmation of tissue hypereosinophilia. HES is defined as hypereosinophilia with eosinophil-mediated organ damage, after excluding other potential causes for the damage.1 HES is rare, with an incidence rate of 0.36–6.3 per 100,000.2
HES can be subclassified as primary, secondary or idiopathic. Primary HES is caused by an underlying stem-cell, myeloid or eosinophilic neoplasm, and is clonal. Secondary HES is polyclonal and is typically caused by parasitic infestations, certain solid tumours, and T cell lymphomas.1
There are also some specific syndromes of unknown pathogenesis, such as Churg-Strauss syndrome (also known as eosinophilic granulomatosis with polyangiitis) and hypereosinophilia of undetermined significance. Differentiating between Churg-Strauss syndrome and HES can be difficult due to substantial overlap.3 (Table)
Hypereosinophilia can also be caused by certain immunodeficiencies, such as hyper-IgE syndrome, a rare primary immunodeficiency disease characterized by eczema, recurrent deep-seated Staphylococcal skin abscesses, recurrent lung infections, eosinophilia and high serum IgE levels.4 While our patient had a mild reduction in IgG1, he did not have the clinical features consistent with the diagnosis of hyper-IgE syndrome.
Our patient’s diagnosis falls into the category of “primary/neoplastic HES” based on the 2012 consensus classification proposed for eosinophilic conditions.1 The 2016 WHO classification listed a special category of myeloid and lymphoid disorders with eosinophilia that included PDGFRA, PDGRB, FGFR1, and PCM1-JAK2 rearrangement subtypes.5
The PCM1-JAK2 fusion subtype is a myeloid neoplasm with chromosomal translocation t(8;9), which results in the fusion of PCM1 and the cytoplasmic tyrosine kinase JAK2 genes.6 The PCM1-JAK2 fusion protein is a drug target for ruxolitinib and other newer JAK2 inhibitors.7-9 Patients with PCM1-JAK2 fusion may present with or progress to acute myeloid or lymphoblastic leukaemia.6
Anti–interleukin-5 (anti-IL5) agents such as benralizumab and mepolizumab, originally developed for treatment of eosinophilic asthma, were shown to be effective in HES in recent trials.10,11 In a randomized, double-blind phase II trial in 20 patients with PDGFRA-negative HES, for example, 90 percent of patients in the benralizumab group achieved ≥50 percent reduction in absolute eosinophil count at week 12, compared with 30 percent in the placebo group.10 In the subsequent open-label phase, which allowed tapering of background therapy as tolerated, clinical and haematologic responses were observed in 17 of 19 patients (89 percent), and were sustained for 48 weeks in 14 of 19 patients (74 percent).10 However, response rates in HES differ depending on the clinical subtype, suggesting that the mechanism driving eosinophilia plays an important role in therapeutic response.10,11 Of interest, a recent study showed that benralizumab may be ineffective against the JAK2 variant of HES.10
Mepolizumab and benralizumab have different mechanisms of action (MoA). Mepolizumab targets human IL-5, preventing its interaction with the alpha-chain of the IL-5 receptor on eosinophils and basophils.12 Benralizumab is directed against the alpha-chain of the IL-5 receptor, thereby blocking the binding of IL-5. In addition, simultaneous binding of benralizumab to FcγRIIIa on natural killer (NK) cells triggers antibody-dependent cell-mediated cytotoxicity (ADCC), resulting in amplified eosinophil apoptosis and reduced eosinophilic inflammation.13 This unique MoA of benralizumab is not observed with other anti-IL5 drugs, such as mepolizumab and reslizumab.14
Both mepolizumab and benralizumab are given subcutaneously using prefilled syringes, and their costs are similar in Hong Kong. Their efficacy is similar, and adverse events (AEs) are infrequent and mild. The most frequent AEs are headache (19 percent), injection site reactions (8 percent), back pain (5 percent) and fatigue (5 percent) for mepolizumab, and headache (8.6 percent), pharyngitis (4 percent), arthralgia (3.9 percent), cough (3.3 percent) and injection site reactions (2.2 percent) for benralizumab.12,13
In the present case, benralizumab was selected as the treatment of choice over mepolizumab because its dosing schedule is more convenient to our patient (every 4 weeks for 8 weeks, then every 8 weeks thereafter for benralizumab 30 mg; every 4 weeks for mepolizumab 100 mg). It is worth noting that benralizumab’s efficacy requires participation of NK cells. Mepolizumab may be more preferable for patients with low NK cell counts.