PU.1 regulation of human alveolar macrophage differentiation requires granulocyte-macrophage colony-stimulating factor

Tracey L. Bonfield, Baisakhi Raychaudhuri, Anagha Malur, Susamma Abraham, Bruce C. Trapnell, Mani S. Kavuru, Mary Jane Thomassen

Abstract

Granulocyte-macrophage colony-stimulating factor (GM-CSF) is critically implicated in lung homeostasis in the GM-CSF knockout mouse model. These animals develop an isolated lung lesion reminiscent of pulmonary alveolar proteinosis (PAP) seen in humans. The development of the adult form of human alveolar proteinosis is not due to the absence of a GM-CSF gene or receptor defect but to the development of an anti-GM-CSF autoimmunity. The role of GM-CSF in the development of PAP is unknown. Studies in the GM-CSF knockout mouse have shown that lack of PU.1 protein expression in alveolar macrophages is correlated with decreased maturation, differentiation, and surfactant catabolism. This study investigates PU.1 expression in vitro and in vivo in human PAP alveolar macrophages as well as the regulation of PU.1 by GM-CSF. We show for the first time that PU.1 mRNA expression in PAP bronchoalveolar lavage cells is deficient compared with healthy controls. PU.1-dependent terminal differentiation markers CD32 (FCγII), mannose receptor, and macrophage colony-stimulating factor receptor (M-CSFR) are decreased in PAP alveolar macrophages. In vitro studies demonstrate that exogenous GMCSF treatment upregulated PU.1 and M-CSFR gene expression in PAP alveolar macrophages. Finally, in vivo studies showed that PAP patients treated with GM-CSF therapy have higher levels of PU.1 and M-CSFR expression in alveolar macrophages compared with healthy control and PAP patients before GM-CSF therapy. These observations suggest that PU.1 is critical in the terminal differentiation of human alveolar macrophages.

  • PU-1
  • c-fms
  • alveolar proteinosis

granulocyte-macrophage colony-stimulating factor (GM-CSF) is critically implicated in lung homeostasis in the GM-CSF knockout mouse model. These animals developed an isolated lung lesion reminiscent of pulmonary alveolar proteinosis (PAP) seen in patients (3, 16). PAP is a rare disease (estimated 3.7 cases per million population) that results in the accumulation of surfactant protein and phospholipids in the terminal alveoli of the lung (12). The development of alveolar proteinosis, in adult patients with PAP, is not due to the absence of a GM-CSF gene or a receptor defect but to the development of an anti-GM-CSF autoimmunity (7, 17). All adult patients with PAP have circulating and localized levels of neutralizing anti-GM-CSF antibodies (2, 7, 17). Pilot clinical studies indicate that GM-CSF therapy maybe effective in resolving the lung disease in some patients (1, 6, 10, 11, 13). We have previously published that response to systemic GMCSF therapy depends on the level of circulating neutralizing anti-GM-CSF, with individuals with higher titers less likely to respond to the GM-CSF (1). Therefore, both the GM-CSF-deficient murine model of PAP and human PAP support the notion that GM-CSF is essential in keeping the lungs free of lipoproteinaceous material and in the functional absence of GM-CSF (either by gene knockout or neutralization by anti-GMCSF) that results in PAP. We believe that PAP provides a unique opportunity to study the role of GM-CSF in lung homeostasis and verify conclusions from the GM-CSF knockout mouse.

The exact mechanism of how GM-CSF is related to lung homeostasis and surfactant catabolism is unknown. Inefficient surfactant catabolism by alveolar macrophages and type 2 bronchoalveolar epithelial cells is related to GM-CSF deficiency and lack of maturation and/or differentiation of monocyte/macrophage lineage (19). Data by Shibata et al. (14) indicate that the GM-CSF-deficient state in the GM-CSF knockout mouse is associated with decreased expression of the transcription factor PU.1. In these studies, the absence of PU.1 protein expression in the alveolar macrophage correlated with decreased maturation, differentiation, and surfactant catabolism (14). Correction of the GMCSF defect in vitro resulted in increased expression of PU.1 and upregulation of PU.1-dependent markers CD32, mannose receptor (MR), and macrophage colony-stimulating factor receptor (M-CSFR) (14). This resulted in improved surfactant catabolism, adherence, and maturation. Whether PU.1 plays a pivotal role in healthy human alveolar macrophages in vivo or in PAP has not been previously described.

We therefore hypothesize that GM-CSF exerts its regulatory effects in the human lung through the regulation of PU.1 analogous to those observations in the GM-CSF knockout mouse model (14, 19). The purpose of the current study is to investigate PU.1 expression in human alveolar macrophages in vitro and in vivo in the context of human PAP. We further propose that decreased PU.1 expression downregulates markers of maturation including M-CSFR. We believe that these studies will provide important information on the role of GM-CSF in healthy lung homeostasis as well as PAP.

METHODS

Study population. This protocol was approved by the Cleveland Clinic Foundation Institutional Review Board, and written informed consent was obtained from all subjects. Healthy control individuals had no history of lung disease and were not on medication. The diagnosis of idiopathic PAP was established by histopathological examination of material from open lung or transbronchial biopsies and exclusion of secondary etiologies (1, 2, 6). All PAP patients were symptomatic with dyspnea, hypoxemia, and typical alveolar infiltrates on radiographs. Patients participated in a prospective clinical trial of recombinant human GM-CSF (Leukine; Berlex, Seattle, WA) as described previously (1, 6). Treatment consisted of 250 μg/day by subcutaneous administration with increased dosage every 2 wk and maximum daily dose of 18 μg · kg-1 · day-1 by 8 wk. Median duration of therapy was 25.5 wk.

Bronchoalveolar lavage fluid. Bronchoalveolar lavage (BAL) was performed as previously described (18). Differential cell counts were obtained from cytospins stained with a modified Wright's stain. Mean viability of lavage cells was >95% as determined by trypan blue dye exclusion. For culture, BAL cells were plated into six-well plates with or without 100 ng/ml of human GM-CSF for 48 h in RPMI 1640 medium supplemented with 5% human blood type AB serum (Gemini, Calabasas, CA), l-glutamine, and antibiotics. The BAL cell pellet was used to determine lavage cell differentials, cell culture for M-CSF, and real-time PCR for PU.1, M-CSF, and M-CSFR using primers from Applied Biomedical Incorporated (ABI, Foster City, CA). We used 300,000 BAL macrophages in 24-well plates for 24-h basal secretion of M-CSF. The median differentials from PAP patient BAL cells (n = 17) were as follows: alveolar macrophages, 87%; lymphocytes, 7%; and polymorphonuclear cells (PMN), 6%. In healthy controls (n = 12), BAL cells consisted of alveolar macrophages, 96%; lymphocytes, 3%; and PMN, 1%.

Cytokine assays. Supernatants from BAL cell cultures from PAP patients and healthy volunteers were assayed for M-CSF (R&D Systems, Minneapolis, MN) by ELISA with the sensitivity of 31.2-1,000 pg/ml for M-CSF. Samples below the sensitivity were reported as the lowest value, and samples above the sensitivity were diluted to obtain a value within the assay range. The coefficient of variation was <10%. All assays were carried out in duplicate.

RNA purification and analysis. Total RNA was extracted from BAL cells by RNeasy protocol (Qiagen, Valencia, CA). Expression of mRNA was determined by real-time RT-PCR using the ABI Prism 7000 Detection System (TaqMan; Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. RNA specimens were analyzed in duplicate with primer sets for a housekeeping gene (GAPDH) and PU.1, M-CSF, or M-CSFR (ABI). Threshold cycle values for genes of interest were normalized to GAPDH and used to calculate the relative quantity of mRNA expression in PAP (or GM-CSF-treated) samples relative to untreated or healthy control values. Data are expressed as fold difference in mRNA expression relative to control values.

Fluorescence-activated cell sorting analysis. Alveolar macrophages obtained from PAP patients (n = 4) and healthy controls (n = 6) were stained for CD14, human leukocyte antigen (HLA)-DR, and MR. Briefly, 3 × 105 BAL macrophages were stained with FITC-labeled CD14, HLA-DR, or phycoerythrin-labeled CD32 or MR (PharMingen, San Diego, CA) for 30 min in cold PBS. Cells were washed and fixed in 2% paraformaldehyde. Cells were analyzed on a BD FACscan Analyzer compared with isotype-specific background controls. Results are expressed as mean percent positive-staining alveolar macrophages for 10,000 events.

Electrophoretic mobility shift assay. For electrophoretic mobility shift assay, 10 μg of each of the whole cell extracts were incubated in binding buffer (8 mM HEPES, pH 7.0, 10% glycerol, 20 mM KCl, 4 mM MgCl2, and 1 mM sodium pyrophosphate) containing 1 μg of poly(dI-dC) and 40,000 cpm probe for 20 min at room temperature as previously described (9). Specificity of the probe was shown by incubating the extract with a 1,000-fold molar excess of cold oligonucleotide. Sequence of the PU.1 binding site is as follows: 5′GGGGGGGAAAAACAGGAAGGGGGGGG3′ and 5′CCCCCCCTTTTTGTCCTTCCCCCCCC3′.

Statistics. Data were analyzed by one-way analysis of variance and Student's t-test using Prism software (Graph-Pad, San Diego, CA). Significance was defined as P = 0.05.

RESULTS

PU.1 mRNA expression is deficient in BAL cells from PAP patients. BAL cells from nine PAP patients were evaluated for PU.1 expression by real-time PCR. The mRNA levels of PU.1 were modestly but significantly lower than those observed in seven healthy controls (Fig. 1, P = 0.003).

Fig. 1.

PU.1 mRNA expression is deficient in bronchoalveolar lavage (BAL) cells from pulmonary alveolar proteinosis (PAP) patients. BAL cells from PAP patients (n = 9) and healthy controls (HC, n = 7) were evaluated for PU.1 expression using real-time PCR. PAP patients had deficient PU.1 expression compared with HC (P = 0.003).

PAP alveolar macrophages have decreased expression of terminal differentiation markers CD32 and MR. GM-CSF knockout studies have suggested that PU.1 is essential for terminal differentiation (14, 19). So, to determine whether PU.1 expression may alter human alveolar macrophage maturation in PAP, we evaluated healthy control (n = 6) and PAP (n = 4) alveolar macrophages for CD32 (FCγIIR), MR, and CD14 as markers of terminal differentiation. We found that both CD32 (P = 0.007) and MR (P = 0.001) expression was decreased on PAP alveolar macrophages, whereas the surface expression of CD14 (P = 0.216) was not significantly different than controls (Fig. 2). HLA-DR staining of PAP alveolar macrophages, 83 ± 4%, was similar to healthy controls, 84 ± 6%.

Fig. 2.

PAP alveolar macrophages have decreased expression of terminal differentiation markers CD32 and mannose receptor (MR). Healthy control (n = 5) and PAP (n = 4) alveolar macrophages were evaluated for CD32 (FCγIIR), MR, and CD14 by FACS analysis. PAP CD32 (P = 0.007) and MR (P = 0.001) expression was significantly less than healthy control alveolar macrophages. CD14 expression was not significantly different than controls (P = 0.216).

BAL cells are not the primary source of the elevated levels of M-CSF in the BAL of PAP patients. BAL cells were obtained from healthy volunteers (n = 6) and PAP patients (n = 4) and evaluated for both M-CSF mRNA and protein secretion. BAL cell M-CSF mRNA expression was not different compared with healthy controls (P = 0.187). Similarly, PAP and healthy control BAL cell secretion of M-CSF protein (2,609 ± 219 pg/ml vs. 2,648 ± 633 pg/ml for PAP and healthy controls, respectively) was not different (P = 0.467, means ± SE).

M-CSFR expression is deficient in PAP BAL cells. PU.1 regulates c-fms (M-CSFR) and may affect BAL cell response to the elevated M-CSF in PAP BAL fluid (2). We examined PAP (n = 6) and healthy control (n = 6) BAL cells for expression of M-CSFR by real-time PCR and found that M-CSFR mRNA was deficient in PAP compared with healthy controls (Fig. 3, P = 0.003).

Fig. 3.

Macrophage colony-stimulating factor receptor (M-CSFR) expression is deficient in PAP BAL cells. PU.1 regulates c-fms (M-CSFR) and may affect BAL cell response to the elevated M-CSF in PAP BAL fluid (2). We examined PAP (n = 6) and healthy control (n = 6) BAL cells for expression of M-CSFR with real-time PCR. PAP M-CSFR expression was significantly less than healthy control BAL M-CSFR expression (P = 0.003).

GM-CSF increases PU.1 expression in vitro. Because PAP is an anti-GM-CSF autoimmune disease and GMCSF can regulate PU.1 expression in the GM-CSF knockout mouse, we evaluated the effect of GM-CSF on PAP and healthy control BAL cells in vitro. Cultured BAL cells were rested for 24 h and were then incubated in the presence of GM-CSF for an additional 48 h. Healthy control BAL cells, as anticipated, did not have a significant increase in PU.1 mRNA expression relative to cells cultured in the absence of GM-CSF [1.6 ± 0.1-fold, n = 2, P > 0.05 (not significant)]; PU.1 is constitutively expressed at high levels in control BAL (14). In contrast, PAP cells cultured with GM-CSF had a significant increase in PU.1 mRNA (2.3 ± 0.6-fold, n = 2, P = 0.02). Furthermore, GM-CSF treatment of healthy control and PAP alveolar macrophages resulted in an increase in M-CSFR mRNA (healthy control: 9.6 ± 1.6-fold, n = 2, P = 0.01; PAP: 2.4 ± 0.6-fold, n = 2, P = 0.03).

GM-CSF therapy in PAP patients increases PU.1 protein expression. We evaluated PAP patients in a preliminary pilot study of GM-CSF therapy for changes in PU.1 binding activity. Before GM-CSF therapy, PAP patients (Fig. 4) have deficient PU.1 binding activity compared with healthy controls. Patients on GM-CSF therapy showed levels of PU.1 activity similar to healthy controls. This suggests that the deficient PU.1 activity in PAP patients is corrected by the GMCSF therapy.

Fig. 4.

Granulocyte-macrophage colony-stimulating factor (GMCSF) therapy in PAP patients increases PU.1 protein expression. We evaluated PAP patients in a preliminary pilot study of GM-CSF therapy for changes in PU.1 binding activity by EMSA. Before GM-CSF therapy PAP patients (PAP 1 and 3) have deficient PU.1 binding activity compared with healthy controls (HC 1 and 2). Patients on GM-CSF therapy have levels of PU.1 binding activity similar to healthy controls (PAP 2 and 4). This suggests that the deficient PU.1 activity in PAP patients is corrected by the GM-CSF therapy.

GM-CSF therapy enhances PU.1 and M-CSFR expression in PAP patients. Both PU.1 (Fig. 5A) and M-CSFR (Fig. 5B) mRNA is decreased in PAP patients (n = 6) compared with healthy controls (n = 6). BAL cells from patients on therapy (n = 5) demonstrated increased PU.1 (P = 0.03) and M-CSFR (P = 0.01) expression compared with baseline.

Fig. 5.

GM-CSF therapy enhances PU.1 and M-CSFR expression in PAP patients. Baseline PAP patients (n = 6) and patients on GMCSF therapy (n = 5) were evaluated for expression of both PU.1 (A) and M-CSFR (B) and compared with healthy controls (n = 5) by real-time PCR. PU.1 mRNA is decreased in baseline PAP patients compared with healthy controls. BAL cells from patients on therapy had increased PU.1 (P = 0.03) and M-CSFR (P = 0.01) expression compared with baseline.

DISCUSSION

We have shown for the first time that, in human PAP, a GM-CSF-deficient autoimmune disease, PU.1 mRNA expression in alveolar macrophages is significantly decreased compared with healthy controls. We have also shown decreased protein expression in PAP BAL cells compared with healthy controls. Furthermore, PU.1-dependent terminal differentiation markers CD32 (FCγII), MR, and M-CSFR were also decreased. In vitro studies demonstrated that GM-CSF treatment upregulates PU.1 and M-CSFR gene expression in alveolar macrophages from PAP patients. Finally, PAP patients treated with GM-CSF therapy have higher levels of PU.1 and M-CSFR expression in alveolar macrophages ex vivo compared with healthy control and PAP patients before GM-CSF treatment. These observations suggest that PU.1 is critical in the terminal differentiation of alveolar macrophages.

GM-CSF promotes monocytic and granulocytic progenitor cell growth, differentiation, and activation (3, 15). GM-CSF activates PU.1, which in turn interacts with transcription binding sites in gene expression differentiation markers such as CD32, MR, and MCSFR. This results in a complex series of reactions within the myeloid progenitor, which ultimately leads to terminal differentiation and the upregulation of these maturation markers (8, 19). Together with data from the GM-CSF knockout mouse model (14), data presented here with the rare PAP lung disease indicate that GM-CSF acts through PU.1 as a regulator of human alveolar macrophages both in vitro and ex vivo.

M-CSF shares important and overlapping properties with GM-CSF related to monocyte/macrophage differentiation (4). M-CSF levels are elevated in the lungs of both the murine GM-CSF knockout model of PAP as well as patients with PAP. Data from our study indicate that M-CSFR expression, which is regulated by PU.1, is markedly reduced in PAP compared with healthy controls and can be upregulated by GM-CSF. The lack of M-CSFR expression may prevent and/or decrease alveolar macrophage response to the elevated BAL M-CSF, which may be generated by a compensatory response to the lack of biologically active GM-CSF in both human and mouse PAP (2, 14). The source of enhanced M-CSF levels in BAL remains unknown. M-CSF gene expression and protein secretion did not differ between BAL cells from healthy controls and PAP patients.

PU.1 knockout mice do not produce lymphoid or myeloid cells and die before or shortly after birth (5, 8), whereas GM-CSF knockout mice are deficient in PU.1 protein and survive until the development of an alveolar proteinosis syndrome in the lung (14, 19). In the GM-CSF knockout mouse model, GM-CSF expression is absent throughout development, whereas in human PAP, GM-CSF is present but is neutralized by circulating levels of anti-GM-CSF, which appears most commonly in the second or third decade of life (12). We have shown here that BAL cells from PAP patients demonstrate a modest but significant decrease in PU.1, consistent with the GM-CSF knockout mouse. These results suggest that GM-CSF is important in the regulation of PU.1 expression in alveolar macrophages in the lung. The implication is that GM-CSF is required for healthy lung homeostasis through the regulation of PU.1 and myeloid terminal differentiation.

In summary, we demonstrate decreased PU.1 gene expression and activity in PAP BAL cells. Furthermore, expression of the PU.1-regulated differentiation-associated cell surface markers MR and CD32, as well as M-CSFR gene expression, are decreased. The addition of exogenous GM-CSF to alveolar macrophages in vitro and in vivo enhances the expression of PU.1 and M-CSFR. These findings support the critical role of GM-CSF in human lung homeostasis.

DISCLOSURES

This work was supported by NIH Research Grant R01 HL-67676 funded by the National Heart, Lung, and Blood Institute and the National Institute of Allergy and Infectious Diseases.

We are also grateful for the generous support of Regina Taussig.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

References

View Abstract