Rationale: Granulocyte/macrophage colony–stimulating factor (GM-CSF) autoantibodies (GMAb) are strongly associated with idiopathic pulmonary alveolar proteinosis (PAP) and are believed to be important in its pathogenesis. However, levels of GMAb do not correlate with disease severity and GMAb are also present at low levels in healthy individuals.
Objectives: Our primary objective was to determine whether human GMAb would reproduce PAP in healthy primates. A secondary objective was to determine the concentration of GMAb resulting in loss of GM-CSF signaling in vivo (i.e., critical threshold).
Methods: Nonhuman primates (Macaca fascicularis) were injected with highly purified, PAP patient-derived GMAb in dose-ranging (2.2–50 mg) single and multiple administration studies, and after blocking antihuman immunoglobulin immune responses, in chronic administration studies maintaining serum levels greater than 40 μg/ml for up to 11 months.
Measurements and Main Results: GMAb blocked GM-CSF signaling causing (1) a milky-appearing bronchoalveolar lavage fluid containing increased surfactant lipids and proteins; (2) enlarged, foamy, surfactant-filled alveolar macrophages with reduced PU.1 and PPARγ mRNA, and reduced tumor necrosis factor-α secretion; (3) pulmonary leukocytosis; (4) increased serum surfactant protein-D; and (5) impaired neutrophil functions. GM-CSF signaling varied inversely with GMAb concentration below a critical threshold of 5 μg/ml, which was similar in lungs and blood and to the value observed in patients with PAP.
Conclusions: GMAb reproduced the molecular, cellular, and histopathologic features of PAP in healthy primates, demonstrating that GMAb directly cause PAP. These results have implications for therapy of PAP and help define the therapeutic window for potential use of GMAb to treat other disorders.
High levels of granulocyte/macrophage colony–stimulating factor autoantibodies (GMAb) are present in patients with the common (idiopathic) clinical form of pulmonary alveolar proteinosis (PAP) and are presumed to be central to disease pathogenesis. However, levels of GMAb do not correlate with disease severity, they are present in most or all healthy people at low levels, and a definitive role in the pathogenesis of PAP has not been established.
Highly purified, PAP patient-derived GMAb reproduced the biochemical, cellular, and histopathologic features of idiopathic PAP after transfer into healthy nonhuman primates.
Normally, the level of alveolar surfactant is tightly regulated by the balanced production of surfactant in alveolar epithelial cells and its clearance by recycling and catabolism in these cells and by catabolism in alveolar macrophages (8). The serendipitous discovery that PAP develops in mice deficient in granulocyte/macrophage colony–stimulating factor (GM-CSF) (9, 10) or its receptor (11) provided an important pathogenic clue. Surfactant production, uptake, and clearance by alveolar epithelial cells and uptake by alveolar macrophages were unaffected by the lack of GM-CSF, whereas catabolism in alveolar macrophages was reduced (9, 12, 13). The coordinate regulation of this and many other macrophage functions by GM-CSF by the “master” myeloid transcription factor PU.1 led to the conclusion that GM-CSF regulates the terminal differentiation of alveolar macrophages in the lungs of mice (3, 14–20). GM-CSF deficiency has not been reported in patients with PAP (21). However, the discovery that high levels of GM-CSF autoantibodies (GMAb) are strongly associated with the common clinical form known as “idiopathic” PAP provided a second critical pathogenic clue (22). These autoantibodies are polyclonal (23), comprise neutralizing and nonneutralizing IgG (24), primarily IgG1 and IgG2 (15, 23), and eliminate GM-CSF bioactivity in the serum and lungs of patients with idiopathic PAP (23, 25). Ex vivo, purified GMAb reproduced the impairment of GM-CSF–stimulated leukocyte CD11b levels (i.e., the CD11b stimulation index) and other neutrophil abnormalities observed in patients with PAP (25). Notwithstanding, serum GMAb levels in patients with PAP do not correlate with disease severity (5, 7). Further, GMAb are present in most or all healthy individuals (15), albeit at lower levels than in patients with PAP, and GMAb are routinely present and comprise the dominate anticytokine activity in pharmaceutical immunoglobulin prepared from healthy human individuals (26). Thus, a causal role for GMAb in the pathogenesis of PAP has not been definitively established.
Based on this background, we hypothesized that prolonged exposure to GMAb at levels sufficient completely to neutralize GM-CSF bioactivity in vivo (i.e., the critical threshold level) would cause PAP by blocking GM-CSF signaling to alveolar macrophages impairing their ability to clear surfactant (15). We injected healthy, nonhuman primates with highly purified, PAP patient-derived GMAb and monitored for the development of PAP for up to 11 months using serum, blood leukocyte, and lung lavage biomarkers of PAP and lung histopathology. These results, of which some were recently reported in brief (27–29), demonstrate that GMAb are central to the pathogenesis of idiopathic PAP.
Details of the experimental methods are provided in the online supplement and described briefly next.
GMAb were purified (25) from a patient with biopsy-proven PAP (30) and high levels of GMAb (3). GMAb were measured by ELISA (15), assessed for purity using gel electrophoresis (25), neutralization of GM-CSF using TF-1 cells (15), and endotoxin content using the Limulus Amebocyte assay (Cambrex, Walkersville, MD).
This study was conducted with institutional approval using four female primates (Macaca fascicularis, Covance, TX) in the Cincinnati Children's vivarium. The demographics, health screening, vaccination and treatment records, housing, feeding, care, and assessment of the primates are described in the online supplement. All were in good health and free of infection.
We tested the hypothesis that human GMAb directly cause the pathologic manifestations of PAP. Seven clinical protocols evaluated administration of highly purified GMAb to healthy primates, including three single- and two multiple-administration safety and dose-finding studies, and two chronic administration studies (Table 1). Results from each study informed subsequent study designs. Preliminary studies (protocols 1–5) established the conditions necessary for long-term studies (protocols 6 and 7) evaluating chronic exposure to human GMAb (passive immunization). Xenographic immune responses were blocked in protocols 6 and 7 by depleting B lymphocytes using anti-CD20 monoclonal antibody (Rituxan, Genentech, San Francisco, CA) and cyclophosphamide (Cytoxan, Baxter, Deerfield, IL). Control primates received this same treatment and phosphate-buffered saline without GMAb. GMAb were administered intravenously and serum levels were monitored to plan repeated administrations.
|Passively immunized subject||C18588||C18588||C18588||C18588||C18585||C27820||C27842|
|Start date, dd/mm/yy*||20/3/06||19/4/06||19/6/06||11/06/07||28/08/06||5/12/07||1/07/08|
|End date, dd/mm/yy*||10/4/06||27/4/06||30/6/06||18/7/07||17/10/6||17/11/08||17/11/08|
|Neutrophil function evaluated||Yes||Yes||No||Yes||Yes||Yes||Yes|
|Serum biomarkers evaluated||Yes||Yes||Yes||Yes||Yes||Yes||Yes|
|Chest high-resolution chest CT scan||No||No||No||No||Yes||Yes||No|
|Bronchoalveolar lavage, no.||No||No||No||Yes (1)||No||Yes (4)||Yes (2)|
|Number of administrations||1||1||1||8||3||13||5|
|Total amount, mg/subject||2.2||2.7||14.8||160||60||350||100|
|Doses administered, mg||2.2||2.7||14.8||20||20||50, 20||20|
|Peak serum concentration, μg/ml||—||29||162||137||210||702||243|
|Mean peak concentration, μg/ml||—||(29)||(162)||106 ± 9||204.5||283 ± 43||211 ± 12|
|Trough serum concentration, μg/ml||<1 (0.4)||<1 (0.02)||<1 (0.14)||<1 (0.5)||2.1||24||22|
|Mean trough concentration, μg/ml||—||—||—||9 ± 4.6||5.6||49 ± 3.4||46 ± 5.6|
|Initial half-life, d§||6.1||—||—||—||6.9||14.5||10|
|Subsequent half-life, d§||—||0.5||1.2||2 ± 0.08||2.3||18 ± 2.1||16.8 ± 2.7|
|Duration above serum level, d‖|
| 1 μg/ml||15||2||9||27||52||349||140|
Study-related procedures, including clinical assessments, phlebotomy, B-lymphocyte depletion, pharmacokinetic analysis, radiologic assessments, bronchoscopy and bronchoalveolar lavage (BAL), cytology including immunohistochemistry (20, 31–33), surfactant lipid and protein analysis (12, 34, 35), open lung biopsies, histopathologic and ultrastructural analysis of lung tissues, and alveolar macrophages are described in the online supplement. GMAb were measured by ELISA (15) in serum and lung epithelial lining fluid (ELF) (by the urea-dilution method ). GM-CSF neutralization was measured with the CD11b stimulation index (25). The development of PAP was monitored by measuring serum SP-D weekly. Weight, vital signs, behavior, activity, blood leukocyte counts, and chest radiographs were used to assess safety.
BAL was performed using three 10-ml aliquots per site in the right middle and lower lobes and lingula. The volume (21 ± 1.3 vs. 19.6 ± 1.4 ml) and efficiency (74.2 ± 2.6 vs. 68.9 ± 3.5%) of BAL fluid recovery was similar in GMAb-injected and control primates (n = 16 or 15 samples, respectively; P = 0.343, 0.242, respectively). Cell viability was routinely greater than 95%. Adherent alveolar macrophages were used to measure PU.1 and PPARγ mRNA levels by quantitative polymerase reaction amplification with Taq-Man primers (AB Biosystems, Carlsbad, CA) and endotoxin-stimulated tumor necrosis factor (TNF)-α release as described (16).
Numerical data were tested for normality using the Kolmogorov-Smirnov test and for equal variance using the Levene median test. Parametric data are presented as the mean ± SEM. Comparisons of parametric and nonparametric data used Student t test and Mann-Whitney rank-sum test as appropriate. P values less than 0.05 indicated statistical significance; P values of less than 0.01 and less than 0.001 are indicated with single and double asterisks. All experiments were done at least twice with similar results.
PAP patient-derived, GM-CSF affinity-purified GMAb had an electrophoretic pattern identical to that of purified human IgG as we previously reported (25), and blocked GM-CSF signaling in vitro as demonstrated by the inhibition of GM-CSF–dependent proliferation of TF-1 cells (Figure 1A). The amounts of GMAb required to inhibit the activity of GM-CSF by 50% (IC50) was 4.13 ± 0.273 mol of GMAb per mole of GM-CSF, similar to previously reported results (23, 25). GMAb blocked GM-CSF signaling in vivo as shown by the inhibition of GM-CSF receptor-mediated STAT5 phosphorylation in blood leukocytes (Figure 1B) and alveolar macrophages (Figure 1B). Disruption in GM-CSF signaling was reversible because the reduction in the CD11b stimulation index by GMAb was normalized by increasing the concentration of GM-CSF used (Figure 1C).
A series of five clinical studies were used to evaluate the dosing and administration of GMAb in nonhuman primates (Table 1). The serum level of GMAb was proportional to the dose administered (Figure 1D) and doses of 20 and 50 mg/primate resulted in high serum GMAb concentrations, comparable with those in patients with PAP (3). Following the initial dose of GMAb in naive primates, the serum half-life of GMAb (∼6–8 d) was shorter than expected, rapidly decreased further, and was short (∼1–2 d) after subsequent doses (see Figure 1E in the online supplement, Figure 2, and Table 1). In clinical studies 1 and 2, the CD11b stimulation index increased (see Figure 1E) rather than decreasing as expected  because of trace amounts of endotoxin in the GMAb used (see Figure E2). Endotoxin was removed from the GMAb used in clinical studies 3–7. Despite improvement in GMAb purity, knowledge of the GMAb dose–response curve, use of higher doses, reduction of the dosing interval, and evaluation in a naive primate, the dynamic reduction in serum GMAb half-life persisted and an adequate serum GMAb concentration could not be maintained (see Figure E1 and clinical studies 3–5), suggesting an antihuman immunoglobulin immune response against GMAb. Subsequent use of rituximab and cyclophosphamide to deplete B lymphocytes and block the antihuman GMAb immune response (Table 1 and clinical studies 6 and 7) resulted in a consistently prolonged half-life of GMAb in serum of 17.1 ± 1.6 days in 18 repeated administrations (see Figure 1E). These studies established the safety and conditions of passive immunization with GMAb required for prolonged blockade of GM-CSF signaling in healthy nonhuman primates.
Informed by preliminary studies, two chronic administration protocols were conducted: clinical studies 6 and 7 (Table 1 and Figure 2). In both, the dose of GMAb used was based on a previously reported minimum serum GMAb level of approximately 10.4–19 μg/ml present in patients with PAP with active lung disease (15). A target value of twice the upper limit of this range was chosen to ensure adequate GM-CSF blockade and was maintained by regular monitoring and readministering GMAb as required continuously to maintain serum levels greater than 40 μg/ml (Figure 2). A marked reduction in the CD11b stimulation index demonstrated continuous pharmacodynamic effects of GMAb on GM-CSF signaling in vivo (Figure 2). In both studies, chronic GMAb exposure resulted in a progressive, time-dependent increase in serum SP-D concentration (Figure 2). B-lymphocyte depletion was used in both studies and resulted in the expected rapid reduction in B-cell counts (Figure 2). An incidental finding was that GMAb blocked homeostatic reexpansion of B cells following pharmacologic depletion (Figure 2). Hematologic indices in GMAb-exposed and control primates, respectively, also included total leukocytes (7,646 ± 335, 8,515 ± 351; P = 0.76); neutrophils (3,338 ± 252, 3,241 ± 180; P = 0.757); monocytes (696 ± 44, 494 ± 22; P = 0.001); lymphocytes (3,228 ± 161, 4,304 ± 176; P < 0.001); and CD3+ cells (2,679 ± 133, 3,131 ± 145; P = 0.024) (×1000 per mm3; n = 67 determinations each).
Three single-administration acute exposure studies of escalating GMAb doses ranging from 2.2–14.8 mg per subject (studies 1–3) and two multiple administration studies (20 mg/dose) ranging from 3 to 8 doses/subject were well-tolerated and without any effects on appetite, growth, or behavior. Similarly, B-lymphocyte depletion and chronic exposure to repeated doses of GMAb (20 or 50 mg/dose) for up to 11 months (studies 6 and 7) had no effect on appetite, growth, or behavior. No adverse events were observed in any subject in any of the seven clinical studies, thus establishing the short-term safety of administering GMAb to healthy primates.
Examination of surgical lung biopsies from GMAb-exposed primates revealed a patchy distribution of histopathologic lesions comprised of well-preserved alveoli filled with alveolar macrophages and extracellular eosinophilic debris after exposure to GMAb for 3.5 or 11 months (Figures 3A–3C and Figure 3D, respectively) compared with controls, which were normal in appearance (not shown). Alveolar macrophages in GMAb-exposed primates were large and foamy in appearance (Figure 3B) and had increased oil red O staining of neutral lipids (Figure 3E) compared with controls (Figure 3F). The intraalveolar material in GMAb-exposed primates had increased SP-A, SP-B, and SP-D immunostaining compared with controls (Figures 3G, 3I, and 3K and Figures 3H, 3J, and 3L, respectively).
Cytologic examination of BAL cells from GMAb-exposed primates revealed that alveolar macrophages were foamy (Figures 4A and 4B) and morphometric analysis revealed they were 1.8-fold larger than alveolar macrophages from controls (Figure 4C). Alveolar macrophages from GMAb-exposed primates also had increased staining with periodic acid–Schiff reagent (Figures 4D–4F), oil red O (Figures 4G–4I), and abnormal immunostaining for SP-B (Figures 4J–4K). Enumeration BAL cells revealed increased numbers of alveolar macrophages, neutrophils, and lymphocytes in GMAb-exposed primates compared with controls (Figure 4L), similar to results in patients with PAP (23, 37) and GM-CSF–deficient mice (9, 10).
Ultrastructural analysis done after GMAb-exposure for 3.5 months revealed that the foamy appearance of alveolar macrophages in GMAb-exposed primates was caused by accumulation of lamellar bodies and lipid droplet inclusions (Figures 5A–5H). Compared with controls, the lipid droplets in alveolar macrophages from GMAb-exposed primates were 1.3-fold larger than those in controls (Figure 5K). The increased size of alveolar macrophages was proportional to the number of inclusions (Figure 5L).
Alveolar macrophages from GMAb-exposed primates contained reduced numbers of mRNA transcripts for the transcription factors PU.1 (Figure 6A) and PPARγ (Figure 6B), and reduced endotoxin-stimulated TNF-α release (Figure 6C) compared with controls, similar to results for patients with PAP (17, 38) and GM-CSF–deficient mice (20).
These abnormal histologic, cytologic, and molecular findings are similar to those of patients with PAP associated with GMAb (2, 3) or mutations in the GM-CSF receptor α chain gene (CSF2RA) (6), and to those of mice deficient in GM-CSF (9, 10) or its receptor (11).
Passive immunization disrupted surfactant homeostasis as demonstrated by analysis of BAL fluid (Figure 7). Compared with controls, GMAb exposure caused the BAL fluid to be milky-appearing (Figure 7A) and to have a greater than twofold increase in turbidity (Figure 7B), total phospholipids (Figure 7C), and saturated phosphatidylcholine (Figure 7D). Cholesterol levels in the BAL fluid were unchanged (14 ± 3, 11 ± 2 μg/ml, respectively; n = 10 each; P = 0.494) and triglycerides were not detected (assay sensitivity = 1 pmol/ml, not shown). GMAb did not change the relative fractional composition of surfactant phospholipids as measured by thin-layer chromatography (not shown). Compared with controls, GMAb-exposed primates had increased total protein (Figure 7E) and levels of SP-A, -B, -C, and -D in BAL fluid (Figure 7F). In summary, GMAb-exposure increased pulmonary surfactant pool size without significantly altering its composition.
Serum and leukocyte biomarkers of PAP were used to measure the critical threshold of GMAb in blood. GMAb reduced the CD11b stimulation index (Figure 8A) and phagocytic capacity of neutrophils in whole blood (Figure 8B), similar to results for patients with PAP (25). The CD11b stimulation index varied inversely with serum GMAb concentrations below 5 μg/ml and was zero at higher values (Figure 8C), similar to results for the CD11b stimulation index and endogenous GMAb levels in healthy individuals and patients with PAP (15). These results indicated the critical threshold of GMAb is 5 μg/ml in serum.
Biomarkers of PAP in serum and lung were used to estimate the critical threshold of GMAb in the lung ELF (measured by the urea-dilution method ). Repeated measurements of GMAb simultaneously in serum and lung ELF demonstrated that lung levels were 10.2 ± 2.1% that of serum levels in passively immunized primates (Figure 8D) and that levels were negligible in controls (not shown). Based on the minimum serum GMAb concentration of 40 μg/ml in immunized primates in Studies 6 and 7 and this value for the serum to lung ratio, the minimum GMAb concentration in lung ELF was calculated to be 3.8 to 4 μg/ml. Because both immunized primates in these studies developed PAP, the critical threshold of GMAb in lung ELF was estimated to be 4 μg/ml or less (assuming development of PAP requires levels of GMAb to remain continuously above the critical threshold).
Because serum SP-D is elevated in proportion to lung disease severity in patients with PAP (7) and was elevated in GMAb-exposed primates compared with controls (Figure 8D), we used this biomarker to monitor for the development of PAP. Serum SP-D concentration increased progressively over time in GMAb-exposed primates and remained unchanged and within the normal range in control primates in both chronic GMAb administration studies (Figure 2). The earliest time at which serum SP-D was clearly increased above the normal range was after 2 to 3 months after initiating GMAb exposure. From a plot of time versus GMAb concentration (Figure 8F), the critical threshold of GMAb in lung ELF was estimated to be between 4 and 6 μg/ml. This result is consistent with pharmacokinetic data from clinical study 7, which showed that the GMAb concentration in lung ELF was never greater than 15 μg/ml, greater than 10 μg/ml for only 7 days, and greater than 5 μg/ml for 80 of 140 days (Figure 8F).
In this study, we found that prolonged exposure to a high level of highly purified, PAP patient-derived GMAb reproduced the typical pathologic manifestations of PAP in healthy, nonhuman primates, including the disruption of GM-CSF signaling in vivo in blood leukocytes and alveolar macrophages; reduced PU.1 and PPARγ mRNA levels and impaired proinflammatory cytokine signaling in alveolar macrophages; surfactant accumulation in alveolar macrophages and alveoli; a progressive increase in serum SP-D concentration; pulmonary leukocytosis; and reduced neutrophil function. These results provide strong evidence that GMAb directly cause the pathogenesis of idiopathic PAP.
We propose a mechanism in which GMAb reduce GM-CSF bioactivity in vivo in rheostatic fashion at concentrations below a critical threshold level and completely block it at higher levels thereby impairing GM-CSF–dependent myeloid cell functions (Figure 9). In GMAb-injected healthy primates, GMAb impaired alveolar macrophage surfactant clearance, PU.1 and PPARγ expression, and TLR4 signaling. These observations are similar to results in patients with idiopathic PAP (15, 17, 38) or PAP caused by GM-CSF receptor dysfunction caused by CSF2RA mutations (6), and mice deficient in GM-CSF (20) or its receptor (11). In each, interruption of GM-CSF signaling causes PAP and similar morphologic changes and functional impairment in alveolar macrophages. Although the exact mechanism by which GMAb impair surfactant clearance by alveolar macrophages was not identified in our study, several lines of evidence suggest that the reduction of PU.1 in alveolar macrophages is critical to the pathogenesis of PAP. In mice, pulmonary GM-CSF is required for PU.1 expression (20) and surfactant catabolism (12) in alveolar macrophages. Importantly, forced expression of PU.1 in alveolar macrophages from GM-CSF–deficient mice restored surfactant catabolism (and other functions) (20). In humans, GM-CSF is also required for PU.1 expression in alveolar macrophages (17). PU.1 levels and activity are reduced in alveolar macrophages in patients with idiopathic PAP and are restored by GM-CSF therapy (17). We propose that pulmonary GM-CSF is required to stimulate the terminal differentiation of alveolar macrophages in humans based on observations that GM-CSF, by PU.1, stimulates the terminal differentiation of alveolar macrophages in mice (20), and a similar pattern of macrophage abnormalities is seen in patients with idiopathic PAP (14, 15, 17), patients with hereditary PAP (6, 39), GM-CSF–deficient mice (16, 18–20), GM-CSF receptor-deficient mice (11, 40), and GMAb-injected nonhuman primates (Figures 3–6). We further propose that GMAb cause PAP by blocking pulmonary GM-CSF signaling in vivo, thereby reducing PU.1 levels in alveolar macrophages impairing surfactant catabolism in alveolar macrophages. The presence of small alveolar macrophages engorged with only lamellar bodies and large cells containing only intracytoplasmic lipid droplets staining with oil red O (indicating neutral lipids), and the lack of any increase in neutral lipids in BAL fluid is consistent with the conclusion that alveolar macrophages internalize lamellar bodies that are broken down to neutral lipids, which cannot be further processed or properly excreted. Further studies are required to confirm these observations and determine the precise defects in surfactant lipid metabolism caused by the loss of GM-CSF stimulation.
Our observation that neutrophil functions were impaired in GMAb-injected primates is consistent with results for patients with PAP and GM-CSF–deficient mice, in which neutrophils have functional reductions in phagocytosis, respiratory burst, cellular adhesion, and bacterial killing (25). Further, purified GMAb rheostatically reduced GM-CSF–stimulated functions of neutrophils from healthy human individuals in ex vivo studies (25). In contrast to alveolar macrophages, however, neither PU.1 levels nor any markers of terminal differentiation were altered in neutrophils from human patients with PAP or GM-CSF–deficient mice, indicating that GM-CSF is not required for the terminal differentiation of blood neutrophils (25).
In summary, the striking similarity pattern of differential impairment of myeloid cell functions in mice, humans, and monkeys caused by interruption of GM-CSF signaling (9, 10, 15, 25, 41) suggests that GM-CSF may regulate myeloid cell functions in all three species by common mechanisms. These results and published reports (3, 22, 24, 25, 29, 42) suggest the term “autoimmune PAP” may now be used instead of “idiopathic PAP” in reference to the common clinical form of PAP associated with GMAb.
The conclusion that GMAb are central to the pathogenesis of PAP supports the value of an increased serum level of GMAb in establishing a diagnosis of PAP. This is further supported by a report that the both the sensitivity and specificity of the currently available blood test (15, 23) approach 100% (41). Notwithstanding, it will be important to establish universally accepted methods, calibration standards, and normal and abnormal ranges in certified laboratories to enable routine clinical use.
The progressive increase in serum SP-D concentration caused by exposure to GMAb suggests this simple, noninvasive blood test may be useful in monitoring lung disease severity in patients with PAP. In support of this are results from patients with autoimmune PAP (7), hereditary PAP (6), and mice with disruption of GM-CSF signaling (6), in each, the serum SP-D is increased. Because serum SP levels are increased in other lung disorders, an increased serum SP-D level that is nonspecific is of limited use in the diagnosis of PAP.
The observation that the critical threshold GMAb concentration associated with loss of GM-CSF signaling was similar in the blood and lung ELF suggests a clinical outcome measure (e.g., serum GMAb concentration) and target value (4–6 μg/ml). Although GMAb levels do not correlate with PAP disease severity, a high level of GMAb is diagnostic of PAP (41). Thus, our data predict that a decrease in serum GMAb concentration below the critical threshold may be associated with resolution of PAP. This may be useful for predicting clinical responses to therapies that target reduction in GMAb levels (i.e., plasmapheresis [30, 43] or rituximab-mediated depletion of antibody-producing B lymphocytes ). Further, these results predict that spontaneous resolution of PAP may be associated with a fall in GMAb levels. Because the critical threshold will vary with the IC50 and GMAb are polyclonal, it will be important to determine if this value varies significantly among patients with PAP or among patients with differing degrees of disease severity.
Our results inform the natural history of PAP at disease onset, which is poorly studied because patients with PAP usually present late. Our conclusion that GMAb reproduced early manifestations of PAP in nonhuman primates is supported by the presence of neutrophil and alveolar macrophage abnormalities; a mild (greater than twofold) increase in surfactant pool size (in contrast to greater than eightfold increases in patients with advanced PAP ); a small increase in serum SP-D; and the absence of radiologic, behavioral, or symptomatic manifestations. We propose the following model (and timing) of events in the initiation and progression of autoimmune PAP: an increase in GMAb (disease initiation) → loss of GM-CSF signaling (seconds) → altered GM-CSF transcriptome (minutes) → altered macrophage biochemistry (i.e., decreased surfactant catabolism) (hours) → surfactant accumulation (days) → radiologic manifestations (months to years) → symptoms (years). Several lines of evidence support this model. First, GMAb disrupted GM-CSF signaling in myeloid cells immediately and at similar concentrations in the blood and lungs. Second, purified GMAb instantly blocked the binding of GM-CSF to cognate receptors (22), GM-CSF signaling, and functions of normal neutrophils ex vivo (25), and disrupted GM-CSF–stimulated proliferation of monocytes in short-term culture (46). Third, the delay preceding an increase in serum SP-D levels suggests that disruption of surfactant homeostasis at the biochemical and cellular levels is delayed. Fourth, surfactant pool size seems to increase progressively over time in GM-CSF–deficient mice (9, 12). Fifth, because 31% of autoimmune patients with PAP with significant radiologic manifestations were asymptomatic, radiologic manifestations occur before the development of symptoms (7). Sixth, in two children with hereditary PAP caused by CSF2RA mutations resulting in loss of GM-CSF signaling from birth, one became symptomatic at 4 years old and the other was asymptomatic and had mild disease (i.e., mild radiologic changes and a moderately reduced diffusing capacity of carbon monoxide) at age 8 years (6). These two results indicate that factors in addition to loss of GM-CSF signaling are important in determining lung disease severity in PAP.
GMAb reduced TNF-α release by endotoxin-stimulated macrophages and basal neutrophil functions indicating that GM-CSF regulates the inflammatory responses of myeloid cells in vivo. This is supported by observations in GM-CSF–deficient mice (16, 47), GM-CSF receptor-deficient mice (11, 40), wild-type mice injected with antimurine GM-CSF antibodies (48), patients with autoimmune PAP (15), and patients with hereditary PAP (6). In each, disruption of GM-CSF signaling, regardless of the mechanism involved, similarly reduced macrophage-mediated proinflammatory cytokine signaling or neutrophil reactivity. Importantly, observations potentially implicating GM-CSF in the pathogenesis of rheumatoid arthritis (49) and multiple sclerosis (50) have led to interest in the potential use of neutralizing antihuman GM-CSF antibodies to treat serious inflammatory and autoimmune diseases (51). Our results suggest that PAP may potentially complicate such therapies depending on GM-CSF antibody affinity, dosing, and administration. The minimum concentration of GMAb in lung ELF associated with PAP was approximately 4 μg/ml, thus providing an estimate of the concentration required to disrupt surfactant homeostasis. The level required to block GM-CSF signaling in lung macrophages and blood leukocytes was similar, agrees well with prior ex vivo human studies (25), and is close to the endogenous levels of GMAb associated with active lung disease in patients with PAP (15). Complete blockade of GM-CSF signaling for as little as 2 to 3 months may result in initiation of PAP. Thus, our results help to define the therapeutic window for the use of GM-CSF autoantibodies to treat other diseases (15). The potential occurrence of iatrogenic PAP will likely be reversible because GM-CSF administration is therapeutic in patients with autoimmune PAP (52) and reverses the PAP phenotype in GM-CSF–deficient mice (53). Several biomarkers may be useful in monitoring for the development and resolution of PAP in this clinical setting, including the level of GMAb and SP-D in serum, and the CD11b stimulation index and STAT5 phosphorylation in blood leukocytes.
An interesting observation was the unexpected increase in the CD11b stimulation index in the early preliminary studies, which provided a clue about the trace endotoxin contamination of GMAb. The use of endotoxin-free GMAb in subsequent clinical studies resulted in the expected reduction in the CD11b stimulation index (15, 25). Notwithstanding, the rapid decline in serum half-life of GMAb indicated human immunoglobulins (i.e., GMAb) elicited a xenographic immune response in Macaca fascicularis. We confirmed this by conducted clinical studies without or with B-lymphocyte depletion using rituximab and cyclophosphamide. This drug combination had been shown to be safe and induce beneficial immunomodulatory effects in prior reports of adoptive immunotherapy (54, 55). Although considered a limitation of our study, B-lymphocyte depletion had no effect on GM-CSF signaling in blood neutrophils or alveolar macrophages (Figure 1B), surfactant clearance by alveolar macrophages (Figures 3–5), or surfactant homeostasis (Figure 7).
GMAb blocked the homeostatic expansion B lymphocytes following pharmacologically induced B lymphocytopenia. Because homeostatic lymphocyte expansion (56) and GM-CSF (51) are both important in autoimmune disorders, this observation suggests GMAb may be useful in immunotherapeutic strategies targeting B lymphocytes. Because immune globulin is effective as therapy of autoimmune and systemic inflammatory disorders (57) and it consistently contains GM-CSF autoantibodies as the dominant anticytokine activity (26), our observation suggests the mechanism of action of immune globulin therapy of autoimmune diseases may involve disruption of GM-CSF signaling. An effective dose of immune globulin (2 g/kg body weight) would result in serum GMAb levels of 0.5 to 1 μg/ml (15), which is within the range associated with negative regulation of myeloid cell functions in primates (Figure 8C) and healthy humans (15). Interestingly, in renal transplant candidates, immune globulin and rituximab reduced anti-HLA antibody levels, the wait-time before transplantation (from 144 ± 89 to 5 ± 6 mo), and increased the number of patients transplanted (58). Further studies to evaluate this potential therapeutic role of GMAb are needed.
This work was supported in part by grants from the National Heart, Lung and Blood Institute (HL0085453 to B.C.T. and HL085610 to J.A.W.), and from the National Center for Research Resources and the National Institutes of Health Office of Rare Diseases (RR019498 to B.C.T.). We are indebted to Diane Black, Claudia Chalk, Paula Blair, Dave Loudy, and Gail Macke for technical support; to Jasmine Hales and Jefferson Childs for help with care and management of the primates; and Drs. Christopher Karp, Division of Immunobiology, Cincinnati Children's Hospital Research Foundation, Lisa Young, Division of Pulmonary Medicine, Cincinnati Children's Hospital, and Frank McCormack, Division of Pulmonary, Critical Care and Sleep Medicine, University of Cincinnati Medical Center, for critical reading of the manuscript.
|1.||Rosen SG, Castleman B, Liebow AA. Pulmonary alveolar proteinosis. N Engl J Med 1958;258:1123–1142.|
|2.||Seymour JF, Presneill JJ. Pulmonary alveolar proteinosis: progress in the first 44 years. Am J Respir Crit Care Med 2002;166:215–235.|
|3.||Trapnell BC, Whitsett JA, Nakata K. Pulmonary alveolar proteinosis. N Engl J Med 2003;349:2527–2539.|
|4.||Milleron BJ, Costabel U, Teschler H, Ziesche R, Cadranel JL, Matthys H, Akoun GM. Bronchoalveolar lavage cell data in alveolar proteinosis. Am Rev Respir Dis 1991;144:1330–1332.|
|5.||Seymour JF, Doyle IR, Nakata K, Presneill JJ, Schoch OD, Hamano E, Uchida K, Fisher R, Dunn AR. Relationship of anti-GM-CSF antibody concentration, surfactant protein A and B levels, and serum LDH to pulmonary parameters and response to GM-CSF therapy in patients with idiopathic alveolar proteinosis. Thorax 2003;58:252–257.|
|6.||Suzuki T, Sakagami T, Rubin BK, Nogee LM, Wood RE, Zimmerman SL, Smolarek T, Dishop MK, Wert SE, Whitsett JA, et al. Familial pulmonary alveolar proteinosis caused by mutations in csf2ra. J Exp Med 2008;205:2703–2710.|
|7.||Inoue Y, Trapnell BC, Tazawa R, Arai T, Takada T, Hizawa N, Kasahara Y, Tatsumi K, Hojo M, Ichiwata T, et al. Characteristics of a large cohort of autoimmune pulmonary alveolar proteinosis patients in Japan. Am J Respir Crit Care Med 2008;177:752–762.|
|8.||Wright JR. Clearance and recycling of pulmonary surfactant. Am J Physiol 1990;259:L1–L12.|
|9.||Dranoff G, Crawford AD, Sadelain M, Ream B, Rashid A, Bronson RT, Dickersin GR, Bachurski CJ, Mark EL, Whitsett JA, et al. Involvement of granulocyte-macrophage colony-stimulating factor in pulmonary homeostasis. Science 1994;264:713–716.|
|10.||Stanley E, Lieschke GJ, Grail D, Metcalf D, Hodgson G, Gall JA, Maher DW, Cebon J, Sinickas V, Dunn AR. Granulocyte/macrophage colony-stimulating factor-deficient mice show no major perturbation of hematopoiesis but develop a characteristic pulmonary pathology. Proc Natl Acad Sci USA 1994;91:5592–5596.|
|11.||Robb L, Drinkwater CC, Metcalf D, Li R, Kontgen F, Nicola NA, Begley CG. Hematopoietic and lung abnormalities in mice with a null mutation of the common beta subunit of the receptors for granulocyte-macrophage colony-stimulating factor and interleukins 3 and 5. Proc Natl Acad Sci USA 1995;92:9565–9569.|
|12.||Ikegami M, Ueda T, Hull W, Whitsett JA, Mulligan RC, Dranoff G, Jobe AH. Surfactant metabolism in transgenic mice after granulocyte macrophage-colony stimulating factor ablation. Am J Physiol 1996;270:L650–L658.|
|13.||Yoshida M, Ikegami M, Reed JA, Chroneos ZC, Whitsett JA. GM-CSF regulates surfactant protein-a and lipid catabolism by alveolar macrophages. Am J Physiol 2001;280:L379–L386.|
|14.||Golde DW, Territo M, Finley TN, Cline MJ. Defective lung macrophages in pulmonary alveolar proteinosis. Ann Intern Med 1976;85:304–309.|
|15.||Uchida K, Nakata K, Suzuki T, Luisetti M, Watanabe M, Koch DE, Stevens CA, Beck DC, Denson LA, Carey BC, et al. Granulocyte/macrophage colony-stimulating factor autoantibodies and myeloid cell in healthy individuals. Blood 2009;113:2547–2556.|
|16.||Berclaz PY, Carey B, Fillipi MD, Wernke-Dollries K, Geraci N, Cush S, Richardson T, Kitzmiller J, O'Connor M, Hermoyian C, et al. GM-CSF regulates a PU.1-dependent transcriptional program determining the pulmonary response to LPS. Am J Respir Cell Mol Biol 2007;36:114–121.|
|17.||Bonfield TL, Raychaudhuri B, Malur A, Abraham S, Trapnell BC, Kavuru MS, Thomassen MJ. PU.1 regulation of human alveolar macrophage differentiation requires granulocyte-macrophage colony-stimulating factor. Am J Physiol 2003;285:L1132–L1136.|
|18.||Berclaz PY, Zsengeller Z, Shibata Y, Otake K, Strasbaugh S, Whitsett JA, Trapnell BC. Endocytic internalization of adenovirus, nonspecific phagocytosis, and cytoskeletal organization are coordinately regulated in alveolar macrophages by GM-CSF and PU.1. J Immunol 2002;169:6332–6342.|
|19.||Berclaz PY, Shibata Y, Whitsett JA, Trapnell BC. GM-CSF. via PU.1, regulates alveolar macrophage fcgamma r-mediated phagocytosis and the IL-18/IFN-gamma -mediated molecular connection between innate and adaptive immunity in the lung. Blood 2002;100:4193–4200.|
|20.||Shibata Y, Berclaz PY, Chroneos ZC, Yoshida M, Whitsett JA, Trapnell BC. GM-CSF regulates alveolar macrophage differentiation and innate immunity in the lung through PU.1. Immunity 2001;15:557–567.|
|21.||Bewig B, Wang XD, Kirsten D, Dalhoff K, Schafer H. GM-CSF and GM-CSF beta c receptor in adult patients with pulmonary alveolar proteinosis. Eur Respir J 2000;15:350–357.|
|22.||Kitamura T, Tanaka N, Watanabe J, Uchida K, Kanegasaki S, Yamada Y, Nakata K. Idiopathic pulmonary alveolar proteinosis as an autoimmune disease with neutralizing antibody against granulocyte/macrophage colony-stimulating factor. J Exp Med 1999;190:875–880.|
|23.||Uchida K, Nakata K, Trapnell BC, Terakawa T, Hamano E, Mikami A, Matsushita I, Seymour JF, Oh-Eda M, Ishige I, et al. High-affinity autoantibodies specifically eliminate granulocyte-macrophage colony-stimulating factor activity in the lungs of patients with idiopathic pulmonary alveolar proteinosis. Blood 2004;103:1089–1098.|
|24.||Bendtzen K, Svenson M, Hansen MB, Busch T, Bercker S, Kaisers U, Uchida K, Beck DC, Trapnell BC. GM-CSF autoantibodies in pulmonary alveolar proteinosis. N Engl J Med 2007;356:2001–2002.|
|25.||Uchida K, Beck DC, Yamamoto T, Berclaz PY, Abe S, Staudt MK, Carey BC, Filippi MD, Wert SE, Denson LA, et al. GM-CSF autoantibodies and neutrophil dysfunction in pulmonary alveolar proteinosis. N Engl J Med 2007;356:567–579.|
|26.||Svenson M, Hansen MB, Ross C, Diamant M, Rieneck K, Nielsen H, Bendtzen K. Antibody to granulocyte-macrophage colony-stimulating factor is a dominant anti-cytokine activity in human IgG preparations. Blood 1998;91:2054–2061.|
|27.||Beck DC, Uchida K, Luisetti M, Keller G, Trapnell BC. Anti-GM-CSF antibodies isolated from pulmonary alveolar proteinosis (PAP) patients reproduce myeloid cell dysfunction in non-human primates. Am J Respir Crit Care Med 2007;175:A491.|
|28.||Sakagami T, Beck D, Suzuki T, Uchida K, Wood RE, Carey BC, Wert SE, Ikegami M, Whitsett JA, Keller G, et al. Pulmonary alveolar proteinosis (PAP) reproduced in non-human primates. Am J Respir Crit Care Med 2009;179:A3030.|
|29.||Sakagami T, Suzuki T, Uchida K, Carey BC, Wood RE, Wert S, Whitsett J, Luisetti M, Trapnell B. Human GM-CSF autoantibodies and reproduction of pulmonary alveolar proteinosis. N Engl J Med 2009;361:2679–2681.|
|30.||Luisetti M, Rodi G, Perotti C, Campo I, Mariani F, Poszzi E, Trapnell BC. Plasmapheresis for treatment of pulmonary alveolar proteinosis. Eur Respir J 2009;33:1220–1222.|
|31.||Nogee LM, Wert SE, Proffit SA, Hull WM, Whitsett JA. Allelic heterogeneity in hereditary surfactant protein B (SP-B) deficiency. Am J Respir Crit Care Med 2000;161:973–981.|
|32.||Huffman JA, Hull WM, Dranoff G, Mulligan RC, Whitsett JA. Pulmonary epithelial cell expression of GM-CSF corrects the alveolar proteinosis in GM-CSF-deficient mice. J Clin Invest 1996;97:649–655.|
|33.||Schoch OD, Schanz U, Koller M, Nakata K, Seymour JF, Russi EW, Boehler A. BAL findings in a patient with pulmonary alveolar proteinosis successfully treated with GM-CSF. Thorax 2002;57:277–280.|
|34.||Mason RJ, Nellenbogen J, Clements JA. Isolation of disaturated phosphatidylcholine with osmium tetroxide. J Lipid Res 1976;17:281–284.|
|35.||Ikegami M, Whitsett JA, Jobe A, Ross G, Fisher J, Korfhagen T. Surfactant metabolism in SP-D gene-targeted mice. Am J Physiol 2000;279:L468–L476.|
|36.||Rennard SI, Basset G, Lecossier D, O'Donnell KM, Pinkston P, Martin PG, Crystal RG. Estimation of volume of epithelial lining fluid recovered by lavage using urea as marker of dilution. J Appl Physiol 1986;60:532–538.|
|37.||Caparros D, Steenhouwer F, Marquette CH, Carpentier F, Foulet A. Rev Mal Respir 1994;11:63–66. (Pulmonary alveolar proteinosis. A sequential analysis of the alveolar cell population after complete pulmonary lavage. Apropos of a case).|
|38.||Bonfield TL, Farver CF, Barna BP, Malur A, Abraham S, Raychaudhuri B, Kavuru MS, Thomassen MJ. Peroxisome proliferator-activated receptor-gamma is deficient in alveolar macrophages from patients with alveolar proteinosis. Am J Respir Cell Mol Biol 2003;29:677–682.|
|39.||Martinez-Moczygemba M, Doan ML, Elidemir O, Fan LL, Cheung SW, Lei JT, Moore JP, Tavana G, Lewis LR, Zhu Y, et al. Pulmonary alveolar proteinosis caused by deletion of the GM-CSFralpha gene in the x chromosome pseudoautosomal region 1. J Exp Med 2008;205:2711–2716.|
|40.||Scott CL, Hughes DA, Cary D, Nicola NA, Begley CG, Robb L. Functional analysis of mature hematopoietic cells from mice lacking the betac chain of the granulocyte-macrophage colony-stimulating factor receptor. Blood 1998;92:4119–4127.|
|41.||Presneill JJ, Nakata K, Inoue Y, Seymour JF. Pulmonary alveolar proteinosis. Clin Chest Med 2004;25:593–613 (viii).|
|42.||Bonfield TL, Russell D, Burgess S, Malur A, Kavuru MS, Thomassen MJ. Autoantibodies against granulocyte macrophage colony-stimulating factor are diagnostic for pulmonary alveolar proteinosis. Am J Respir Cell Mol Biol 2002;27:481–486.|
|43.||Kavuru MS, Bonfield TL, Thomassen MJ. Plasmapheresis, GM-CSF, and alveolar proteinosis. Am J Respir Crit Care Med 2003;167:1036–1037.|
|44.||Dalrymple H, Malur A, Gagnon G, Marshall I, Kavuru MS, Barna B, Thomassen MJ. Rituximab depletes B-lymphocytes and improves symptoms in pulmonary alveolar proteinosis. J Allergy Clin Immunol 2008;121:S225.|
|45.||Kuroki Y, Tsutahara S, Shijubo N, Takahashi H, Shiratori M, Hattori A, Honda Y, Abe S, Akino T. Elevated levels of lung surfactant protein a in sera from patients with idiopathic pulmonary fibrosis and pulmonary alveolar proteinosis. Am Rev Respir Dis 1993;147:723–729.|
|46.||Tanaka N, Watanabe J, Kitamura T, Yamada Y, Kanegasaki S, Nakata K. Lungs of patients with idiopathic pulmonary alveolar proteinosis express a factor which neutralizes granulocyte-macrophage colony stimulating factor. FEBS Lett 1999;442:246–250.|
|47.||Paine R III, Morris SB, Jin H, Wilcoxen SE, Phare SM, Moore BB, Coffey MJ, Toews GB. Impaired functional activity of alveolar macrophages from GM-CSF- deficient mice. Am J Physiol 2001;281:L1210–L1218.|
|48.||Bozinovski S, Jones J, Beavitt SJ, Cook AD, Hamilton JA, Anderson GP. Innate immune responses to LPS in mouse lung are suppressed and reversed by neutralization of GM-CSF via repression of TLR-4. Am J Physiol 2004;286:L877–L885.|
|49.||Campbell IK, Bendele A, Smith DA, Hamilton JA. Granulocyte-macrophage colony stimulating factor exacerbates collagen induced arthritis in mice. Ann Rheum Dis 1997;56:364–368.|
|50.||McQualter JL, Darwiche R, Ewing C, Onuki M, Kay TW, Hamilton JA, Reid HH, Bernard CC. Granulocyte macrophage colony-stimulating factor: a new putative therapeutic target in multiple sclerosis. J Exp Med 2001;194:873–882.|
|51.||Hamilton JA. Colony-stimulating factors in inflammation and autoimmunity. Natl Rev 2008;8:533–544.|
|52.||Seymour JF, Presneill JJ, Schoch OD, Downie GH, Moore PE, Doyle IR, Vincent JM, Nakata K, Kitamura T, Langton D, et al. Therapeutic efficacy of granulocyte-macrophage colony-stimulating factor in patients with idiopathic acquired alveolar proteinosis. Am J Respir Crit Care Med 2001;163:524–531.|
|53.||Reed JA, Ikegami M, Cianciolo ER, Lu W, Cho PS, Hull W, Jobe AH, Whitsett JA. Aerosolized GM-CSF ameliorates pulmonary alveolar proteinosis in GM-CSF- deficient mice. Am J Physiol 1999;276:L556–L563.|
|54.||Edwards JC, Szczepanski L, Szechinski J, Filipowicz-Sosnowska A, Emery P, Close DR, Stevens RM, Shaw T. Efficacy of B-cell-targeted therapy with rituximab in patients with rheumatoid arthritis. N Engl J Med 2004;350:2572–2581.|
|55.||Cambridge G, Leandro MJ, Edwards JC, Ehrenstein MR, Salden M, Bodman-Smith M, Webster AD. Serologic changes following B lymphocyte depletion therapy for rheumatoid arthritis. Arthritis Rheum 2003;48:2146–2154.|
|56.||King C, Ilic A, Koelsch K, Sarvetnick N. Homeostatic expansion of T cells during immune insufficiency generates autoimmunity. Cell 2004;117:265–277.|
|57.||Kazatchkine MD, Kaveri SV. Immunomodulation of autoimmune and inflammatory diseases with intravenous immune globulin. N Engl J Med 2001;345:747–755.|
|58.||Vo AA, Lukovsky M, Toyoda M, Wang J, Reinsmoen NL, Lai CH, Peng A, Villicana R, Jordan SC. Rituximab and intravenous immune globulin for desensitization during renal transplantation. N Engl J Med 2008;359:242–251.|