Signal Transducer and Activator of Transcription 5B Deficiency–associated Lung Disease

To the Editor:

Deficiency of STAT5B (signal transducer and activator of transcription 5B) causes a rare disease characterized by growth failure, immune dysfunction, and lung disease (1). Although lung disease is the cause of death in most STAT5B-deficient patients, its pathogenesis is unknown.

STAT5B belongs to a seven-member transcription factor family mediating responses to multiple cytokines, including GM-CSF (granulocyte–macrophage colony–stimulating factor), IL-2, IL-3, IL-5, IL-7, and IL-15 (2). STAT5 includes two members (STAT5A and STAT5B) that share 95% amino acid homology but are not functionally interchangeable. STAT5B deficiency causes growth failure by reducing growth hormone–stimulated IGF-1 (insulin-like growth factor-1) expression and immunodeficiency by disrupting cytokine signaling.

GM-CSF signals through heterodimeric cell-surface receptors and receptor-associated JAK2 (Janus kinase 2), resulting in activation of STAT5 and other regulators with pleiotropic effects on alveolar macrophages, including regulation of survival, proliferation, and differentiation (3, 4). Disruption of GM-CSF signaling causes pulmonary alveolar proteinosis (PAP), a syndrome of surfactant accumulation, lymphocytosis, increased pulmonary M-CSF (macrophage colony–stimulating factor), MCP-1 (monocyte chemoattractant protein-1), and GM-CSF (when caused by GM-CSF receptor dysfunction) resulting in respiratory failure, immunodeficiency, and in some patients serious infection, pulmonary fibrosis, and death (5).

Here, we report the characterization, course, and therapy of STAT5B deficiency–related lung disease. Some of the results of these studies have been previously reported in the form of an abstract (6).

Clinical Presentation

Patients 1 and 2 were nontwin sisters from Kuwait homozygous for a loss-of-function STAT5B mutation (c.1680delG) (7). Both had normal gestation, delivery, and birth weight (3.6 and 2.4 kg), but both developed severe postnatal growth failure, and by 1 year both had developed pneumonia requiring hospitalization, antibiotics, and initiation of albuterol and inhaled corticosteroid therapy; patient 1 also developed arthritis. Pneumonia recurred at ∼3-monthly intervals and was managed with antibiotics. However, by 6 years, both had developed severe pulmonary disease requiring ICU admission.

Diagnostic Evaluation

Both sisters received evaluation and therapy in the United States, approved by our institutional review boards. Radiological evaluation identified diffuse ground-glass opacification superimposed on interlobular septal thickening and bronchiectasis (Figures 1A–1C). BAL was turbid, and cytologic and ultrastructural examination revealed enlarged, foamy, oil red O–positive macrophages, extracellular debris, and surfactant globules (Figures 1D–1G). Lung histology identified intraalveolar eosinophilic material, acute inflammation, nonnecrotizing granulomas, interstitial lymphoid hyperplasia, and interstitial fibrosis (Figures 1H–1J). Immunohistochemistry demonstrated interstitial infiltration of both T and B lymphocytes (Figures 1K and 1L). BAL fluid contained increased dipalmitoyl phosphatidylcholine and cholesterol, with an increased cholesterol-to-phospholipid ratio (Figures 2A–2C), and increased concentrations of GM-CSF, M-CSF, and MCP-1 (Figures 2D–2F). Results demonstrated that STAT5B deficiency–associated lung disease comprised PAP (including lymphocytosis), bronchiectasis, and fibrosis.

Figure 1.

Radiographic appearance and pathological manifestations of STAT5B-associated lung disease. (A and B) Posterior–anterior (A) and lateral (B) chest radiographs showing diffuse lung disease (patient 1, age 10 yr). (C) Axial chest computed tomogram showing multifocal geographic ground-glass opacification and superimposed interlobular septal thickening and bronchiectasis (patient 2, age 15 yr). (D–F) BAL fluid showing lipid-laden macrophages and extracellular lipid (patient 2, age 16 yr; oil red O; original magnification, 600×) (D) and acellular globular, eosinophilic material (patient 1, age 18 yr; hematoxylin and eosin; original magnification, 600×) (E) corresponding to large aggregates of lamellar surfactant (patient 1, age 18 yr; uranyl acetate; electron photomicrograph; scale bar, 0.5 μm) (F). (G) Macrophage containing numerous intracytoplasmic lamellar bodies, multivesicular bodies, and phagolysosomes and adjacent extracellular lamellar material (patient 1, age 18 yr; uranyl acetate; electron photomicrograph; scale bar, 2 μm). (H) Early lung biopsy (patient 2, age 6 yr) showing predominantly interstitial lymphoid hyperplasia (hematoxylin and eosin; original magnification, 40×) with patchy pulmonary alveolar proteinosis (PAP) (lower inset; original magnification, 600×) and small peribronchovascular, nonnecrotizing granuloma containing a conchoidal body (upper inset; original magnification, 600×). (I) Extensive filling of alveoli with eosinophilic material indicating florid PAP (patient 2, age 13 yr; hematoxylin and eosin; original magnification, 200×). (J) Persistence of intraalveolar eosinophilic material and alveolar septal fibrosis 49 days after stem cell transplantation (patient 2, age 16 yr; hematoxylin and eosin; original magnification, 400×). (K) T lymphocytes distributed diffusely within alveolar septa (patient 1, age 6 yr; CD3 immunostain, 40×). (L) B lymphocytes confined to lymphoid aggregates within the lung parenchyma (patient 1, age 6 yr; CD20 immunostain, 40×). CD = cluster of differentiation; STAT5B = signal transducer and activator of transcription 5B.

Figure 2.

Evaluation of BAL, GM-CSF (granulocyte–macrophage colony–stimulating factor) signaling, STAT5 (signal transducer and activator of transcription 5) phosphorylation, and macrophage-mediated surfactant clearance in STAT5B-associated lung disease. (A–F) Evaluation of BAL from healthy control subjects (HC), our patient with STAT5B deficiency (patient 2, age 13 yr), and patients with autoimmune pulmonary alveolar proteinosis (aPAP), including measurement of diphosphatidylcholine (A), cholesterol (B), the ratio of cholesterol to phospholipid in surfactant lipids (C), GM-CSF (D), M-CSF (E), and MCP-1 (F). Each symbol represents one person (HC and aPAP) or a single measurement from independent BAL specimens (patient 2); bars represent the group mean, and T bars represent the SEM of the results shown. (G) GM-CSF signaling determined by the STAT5 phosphorylation index (STAT5-PI) test. Each symbol represents one individual (patient 2, n = 1; HC, n = 146; and patients with aPAP, n = 165); bars represent the group mean for the results shown. Heparinized blood was incubated with or without GM-CSF (10 ng/ml, 30 min), incubated with anti–human phosphorylated STAT5 (pSTAT5) antibody, and evaluated by flow cytometry. The STAT5-PI is calculated as the mean fluorescence intensity (MFI) of GM-CSF–stimulated minus unstimulated leukocytes divided by the MFI of unstimulated leukocytes and multiplied by 100. The STAT5-PI is ⩾216 units in healthy individuals (n = 77) and ⩽20 units in patients with aPAP (n = 80) (B. Carey and B. Trapnell, unpublished results). The dotted line represents the lower limit of the 95% confidence interval of the normal range (in healthy people), and the dashed line represents the upper limit of the 95% confidence interval for the abnormal range (in patients with pulmonary alveolar proteinosis). (H) GM-CSF signaling measured by the GM-CSF EC₅₀ test. Heparinized blood was incubated with the indicated concentrations of GM-CSF, and then pSTAT5 was measured by flow cytometry. For each curve, the horizontal dashed line (gray) denotes the 50% maximum concentration of GM-CSF signaling (measured as STAT5 phosphorylation), and the vertical dashed line represents the corresponding concentration of GM-CSF needed to achieve 50% of the maximum degree of signaling (i.e., the EC₅₀). Each curve represents three determinations for each individual (indicated). The EC₅₀ is ⩽20 ng/ml GM-CSF in healthy individuals (n = 77) and ⩾106 ng/ml in patients with aPAP (n = 80; B. Carey and B. Trapnell, unpublished results). (I) Western blot showing detection of total pSTAT5, STAT5B, and STAT5A (indicated) in iPSC-derived macrophages (prepared as reported [12]) with or without stimulation by GM-CSF (indicated) (4). (J and K) Macrophage surfactant clearance assay. iPSC-derived macrophages prepared from an HC or patient 2 were incubated without (control) or with human surfactant for 24 hours and washed (pulse) and then incubated for an additional 24 hours to permit clearance of surfactant lipids (chase), stained with oil red O, and evaluated microscopically (J) to determine the percentage of cells with staining. In (A–G), each symbol represents one specimen from a separate HC or patient with aPAP or a separate specimen from the same patient with STAT5B deficiency (patient 2). In H, symbols and error bars represent the mean and SD of three separate determinations from the same clinical specimen. In K, each symbol represents an independent measurement, and bars and error bars represent the mean and SD. ***P < 0.001. DPPC = dipalmitoylphosphatidylcholine; EC₅₀ = effective GM-CSF concentration sufficient to elicit a half-maximal signaling response; iPSC = induced pluripotent stem cell; MCP-1 = monocyte chemoattractant protein-1; M-CSF = macrophage colony–stimulating factor; ns = not significant.

Because the pattern of results suggested that GM-CSF signaling disruption could be driving the development of PAP, we evaluated GM-CSF signaling and macrophage function. Results of serum GM-CSF autoantibody testing (8) were normal (not shown), but STAT5 phosphorylation index testing indicated that GM-CSF signaling was severely reduced (Figure 2G). A GM-CSF EC₅₀ (effective concentration for 50% of maximal signaling) test indicated that GM-CSF signaling occurred at normal ligand concentrations but resulted in a severely reduced maximal phosphorylated STAT5 response (Figure 2H). Induced pluripotent stem cell–derived macrophages confirmed the reduction in GM-CSF–stimulated total STAT5 phosphorylation and identified STAT5A but not STAT5B in both patients (Figure 2I). Surfactant clearance by induced pluripotent stem cell–derived macrophages was also impaired (Figures 2J and 2K).

Clinical Course

Throughout childhood and adolescence, both patients experienced recurrent pneumonia caused by various bacterial, viral, and/or fungal pathogens and unremitting lymphoproliferative/fibrotic lung disease resulting in progressive mixed obstructive/restrictive airflow limitation and hypoxemia. In both, PAP responded to whole-lung lavage (WLL), recurrent pneumonia to periodic antibiotics (targeting bacterial, fungal, and viral pathogens), and immune dysregulation to immunosuppressants (including chronic oral and inhaled corticosteroids, immunoglobulin, abatacept, and sirolimus). Patient 1 stabilized after WLL, atorvastatin (to improve surfactant clearance), pirfenidone, and supplemental oxygen/nocturnal bilevel positive airway pressure therapy and remains alive at age 18 but was not a suitable candidate for lung or stem cell transplantation. In contrast, patient 2 was a suitable candidate and underwent fludarabine/melphalan/alemtuzumab conditioning and matched sibling–donor hematopoietic stem cell transplantation at age 15. Despite high-degree donor myeloid chimerism, she died of Aspergillus pneumonia at 11 months while continuing routine posttransplantation immunotherapy. Additional details of the clinical course were reported previously (9).

Discussion

We report the clinical, radiological, and pathological manifestations, natural history, and therapy for lung disease in two patients with STAT5B deficiency. Both presented in early childhood with recurrent pneumonia and PAP associated with GM-CSF signaling disruption, which progressed to include mixed obstructive/restrictive airflow limitation, pulmonary fibrosis, and eventually chronic hypoxemia. Our results identify STAT5B deficiency as a newly defined genetic cause of PAP.

Our observation that GM-CSF signaling was impaired in both patients with STAT5B deficiency has mechanistic implications. First, it explains the development of PAP, which results suggest is caused by disruption of GM-CSF–stimulated, alveolar macrophage–mediated surfactant clearance. Several lines of evidence support this mechanism. Disruption of GM-CSF signaling in alveolar macrophages is known to cause PAP in >90% of patients, including autoimmune PAP (caused by neutralizing GM-CSF autoantibodies) and hereditary PAP (caused by CSF2RA/B [colony stimulating factor 2 receptor subunit alpha/beta] mutations) (5). Normally, the GM-CSF receptor activates STAT5 (via JAK2-mediated phosphorylation) at the cytoplasmic tail of its β chain immediately after ligand binding (3). In both PAP and STAT5B deficiency, both GM-CSF signaling and STAT5 phosphorylation are disrupted and cause a strikingly similar pattern of abnormalities: alveolar macrophage dysfunction, surfactant accumulation, cytokine abnormalities (increased pulmonary GM-CSF, M-CSF, and MCP-1), pulmonary lymphocytosis, immunodeficiency, and fibrosis (5). The observation that STAT5A is phosphorylated normally in the absence of STAT5B but was unable to prevent the pathological manifestations of STAT5B deficiency supports the concept that STAT5A and STAT5B are nonredundant (2). Together, these findings suggest that STAT5B may mediate the regulatory effects of GM-CSF on surfactant clearance by alveolar macrophages and other GM-CSF–stimulated macrophage functions. Second, our finding of impaired GM-CSF signaling provides a mechanism (at least partly) explaining the development of pulmonary fibrosis, as patients with autoimmune and hereditary PAP also develop pulmonary fibrosis, and GM-CSF deficiency augments bleomycin-induced fibrosis by reducing expression of prostaglandin E₂, a potent antifibrotic eicosanoid (10). Third, it provides one mechanism (among others) explaining the increased infection risk, as GM-CSF signaling is vital to host defense mediated by macrophages, neutrophils, natural killer cells, and T-helper cell type 17 cells (3–5, 11).

The observations that pneumonia responded to antibiotics, lymphocytic infiltration to immune modulation, and PAP to WLL underscore the value of these therapeutic modalities in the management of STAT5B deficiency–related lung disease. Although allogeneic stem cell transplantation was initially successful, its ultimate failure because of sepsis after transplantation highlights the problem of immune dysregulation in STAT5B deficiency. Effective donor myeloid cell engraftment may have protected our patient from future episodes of PAP had there been sufficient time for replacement of the endogenous alveolar macrophage population. In the future, autologous stem cell transplantation of STAT5B gene–corrected cells or pulmonary transplantation of autologous STAT5B gene–corrected macrophages may be attractive alternatives (4).