Neutrophil-specific Granule Deficiency Results from a Novel Mutation with Loss of Function of the Transcription Factor CCAAT/Enhancer Binding Protein ε

Material from a previously described (5, 6) male patient lacking neutrophil-specific granules was studied. Research was conducted with informed consent under the guidelines of a National Institutes of Health (NIH) Internal Review Board– approved protocol, no. 92-I-99. The patient died from complications of pneumonia at age 20.

Peripheral blood neutrophils were isolated as described (28), cryopreserved with dimethylformamide (Sigma Chemical Co.), and maintained at −140°C. Cell proteins were extracted as described (29). DNA extraction from cryopreserved fibroblasts proceeded as described (30). RNA was extracted from patient bone marrow aspirate using RNAzol reagent (Teltest) as per manufacturer's protocol. Normal human bone marrow RNA was purchased from Clontech.

PCR reaction was performed using Platinum taq DNA polymerase (Life Technologies) per manufacturer's instructions and cycled as follows: 96°C for 12 min, followed by a three-step cycle—94°C for 30 s, 60°C for 30 s, and 72°C for 2 min—for 35–40 cycles. PCR products were gel purified and recovered using Gene Clean (Bio101). Products were sequenced with an ABI Prism Dye Terminator Cycle Sequencing Ready Reaction Kit (Perkin-Elmer Corp.). Primers were chosen from a published sequence available from EMBL/GenBank/DDBJ under accession no. U48865. Primer sets (upstream, downstream): B, 5′-AGC GGC CAT GCA AAA GGA AAG ACA, 5′-TCC ACC TAC CCC CAA GAG AAA GTT (bp 667–1186); C, 5′-CCC ACG GGA CCT ACT ACG A, 5′-GGG CTG GCC TGC TCT TAC (bp 1818–2343); F, 5′-CTC CCC GGC TGG CCC CTT ACA C, 5′-GCC AAC AGT CCC AAC ACC CAG TCA (bp 3133–3615); G, 5′-GGA GGT GGG GCT ACA AAA GAA ACT, 5′-TCA GGG AGG GGC AGG ACA (bp 1143–1553); H, 5′-ACA GGA GTG GGT GAC AGA GGA GAC, 5′-GGG CCG AAG GTA TGT GGA GGG TAG (bp 1563–2104); I, 5′-CCA TGC CCC CTC CTC TTG TTT CTC, 5′-ACT GCC TTC TTG CCC TTG TGT AA (bp 2594–3171); K, 5′-AAC TTT CTC TTG GGG GTA GGT GGA, 5′-TCG TAG TAG GTC CCG TGG (bp 1163– 1837). Homozygosity was determined by hybridization of the PCR product fragment C to a [³²P]γdATP-endlabeled internal oligonucleotide (H. downstream; sequence above). Labeled oligo was mixed with hybridization buffer (75 mM NaCl, 5 mM EDTA) in a ratio of 1:10, and 10 μl was added to 30 μl PCR product. Hybridization was cycled in a thermal cycler: 95°C for 5 min and 55°C for 10 min. Reactions were immediately placed on ice. Products were resolved on 4–20% Tris/borate/EDTA polyacrylamide gel (Novex) at 250 V.

10 μg of total RNA isolated from the patient's bone marrow and 0.25, 0.5, and 1 μg of control polyadenylated (pA)-mRNA was electrophoretically separated, blotted, hybridized, and washed as described (27). The membrane was stripped by boiling and stored at −20°C.

Protein quantitation was performed using a BCA Protein Assay kit (Pierce Chemical Co.) according to the manufacturer's instructions. 10–100 μg protein extracts were electrophoretically separated, transferred to nitrocellulose, and incubated with primary antibody as described (29). Primary antibody was generated in rabbits by Research Genetics, Inc., using a synthetic peptide encoded in exon 2 of C/EBPε, downstream of the SGD deletion (DPRAVAVKEEPRGPEGSR). The membranes were washed, incubated with anti–rabbit horseradish peroxidase conjugate antibody (ECL Western blot kit; Amersham Pharmacia Biotech, Inc.), and developed according to the manufacturer's instructions. Membranes were stripped and reblotted with anti–mouse human β actin antibody (Boehringer Mannheim) to control for protein loading.

The patient's mutation was introduced into the pCMV-C/EBPε³² expression vector using a Stratagene QuikChange site-directed mutagenesis kit per manufacturer's instructions using a complementary oligonucleotide (PAGE purified; purchased from Genosys Biotech) containing the deletion (5′-CCA CTA CTT GCC GCC CTC GGC CCT TTG CCT ACC). Presence of the mutation and maintenance of the vector sequences was verified by sequencing and restriction enzyme digestion, respectively.

HeLa cells were maintained in DMEM (BioWhittaker) supplemented with 10% heat-inactivated FBS (Life Technologies, Inc.) and penicillin/streptomycin at 37°C and 5% CO₂. Cells were plated in 6-well plates and transfected within 24 h, at 30–50% confluency. Transfections, using the Mammalian Transfection System (Stratagene), were performed using 5 μg reporter plasmid (G-CSF receptor promoter-luc); 1, 2, or 5 μg inducer plasmid (pCMV-C/EBPα, pCMV-C/EBPε³² isoform, pCMV-C/EBPε¹⁴ isoform, or pCMV-C/EBPε³²-SGD, described above); and 0.5 μg pCMVβ, as described (23, 31). The DNA content of transfections was normalized, and transfection was performed according to the manufacturer's instructions, with 300 μl transfection solution applied to the cells. Samples were harvested 24 h after transfection.

Luciferase and β-galactosidase activities were measured using a Dual-Light kit (Tropix, Inc.), according to the manufacturer's protocol, on a Turner 20/20 Luminometer. Samples were read for 15 s after a 3-s delay.

Sequencing of PCR products from genomic DNA detected a 5-bp deletion, TGACC, in exon 2 of the patient's C/EBPε sequence. Fig. 1 A shows sequence data from one normal control (top sequence) and the SGD patient (bottom sequence). The mutation predicts a frameshift and a premature termination of the encoded C/EBPε³² isoform (Fig. 1 B). The missense code after the frameshift results in the loss of the critical DNA binding domain and leucine zipper region required for C/EBP dimerization and function. C/EBPε transcripts encoding the shorter 27- and 14-kD isoforms are predicted to be unaffected, based upon the splice donor and acceptor and translational start sites (23).

Homozygosity of the deletion was determined by PCR amplification of the affected region and resolution of the DNA fragments on a 4–20% polyacrylamide gel (Fig. 1 C). DNA from one normal control and the SGD patient were mixed before amplification and electrophoresis (lane 3), showing bands from both affected and normal alleles. In comparison, PCR products from the SGD patient (lane 4) and normal controls (lanes 1 and 2) show only one fragment, indicating homozygosity for their respective alleles.

RNA blot analysis of the SGD patient's bone marrow total RNA showed decreased amounts of C/EBPε transcripts in comparison with control human bone marrow pA-RNA (Fig. 2 A). Hybridization with a [³²P]dCTP- labeled actin probe (provided by L. Perera, National Cancer Institute, NIH) showed that 10 μg of SGD patient bone marrow total RNA was equivalent to 1 μg of normal bone marrow pA-mRNA and verified the stability and quality of the patient's RNA preparation. Specific loss of C/EBPε transcripts in the SGD patient is likely due to mRNA instability secondary to the frameshift and the premature termination codon, as seen in other similar gene mutations (32, 33). Residual C/EBPε message is likely comprised by C/EBPε¹⁴ and C/EBPε²⁷ transcripts, which are unaffected by the 5-bp deletion and similar in size to the C/EBPε³² transcript. Transcripts of C/EBPα were present in normal amounts. C/EBPα has a more proximal role in the myelopoietic pathway and specifically induces expression of C/EBPε (31, 34, 35). As expected, message for lactoferrin was not detected in the SGD patient's bone marrow RNA.

As predicted from the C/EBPε transcript maps (Fig. 1 B), immunoblotting detected C/EBPε²⁷ and C/EBPε¹⁴ isoforms, but not C/EBPε³², in neutrophils from the SGD patient (Fig. 2 B). All three isoforms were seen in the normal control. The antibody used is specific for a peptide sequence immediately downstream of the 5-bp mutation and should not bind the C/EBPε³²-SGD protein.

Transient transfection assays in HeLa cells, using the G-CSF receptor promoter driving the luciferase gene (31), compared the transactivation potentials of the inducer genes C/EBPα, C/EBPε³², C/EBPε¹⁴, and C/EBPε³²-SGD (Fig. 3). C/EBPε³² has been shown to transactivate the G-CSF receptor promoter, whereas the C/EBPε¹⁴ isoform lacks transactivating function (23). Transient transfection of these plasmid constructs showed a significant loss of transactivation with the C/EBPε³²-SGD isoform (P = 0.02, Mann-Whitney U test). The demonstrated in vitro data, as well as the in vivo SGD phenotype, mark the full length, 32-kD isoform as the major transactivator encoded in the C/EBPε locus.

The temporal link between granule protein production and myeloid lineage differentiation is well described: primary granule proteins are synthesized in myeloblasts and promyelocytes, secondary granules are produced in myelocytes and metamyelocytes, and tertiary granule proteins are generated in band and segmented neutrophils (36).

Previous work suggested that C/EBPε functions at the terminal stages of myeloid differentiation (23, 26). However, the total absence of patient neutrophil secondary granules and the selective loss of primary granule defensins marks an early myelopoietic block at the promyelocyte transition (Fig. 4). Further evidence for this conclusion comes from in vitro differentiation experiments using C/EBPε-deficient stem cells, which do not proceed beyond the promyelocyte stage (26). Other functional defects seen in mouse and human C/EBPε-deficient neutrophils, such as loss of tertiary granule gelatinase (27) and abnormalities in chemotaxis and cytokine expression (6, 27), may occur secondary to the block at the promyelocyte or later stage.

Functional analysis of the previously developed C/EBPε knockout mouse model (26, 27) was critical for the interpretation of the C/EBPε mutation in SGD. The apparent multiplicity of C/EBPε target genes at different cell stages suggests that C/EBPε transactivates a set of early cell stage– specific genes, inducing normal promyelocyte differentiation and granule development. Additional evidence supporting these conclusions comes from recent observations suggesting that C/EBPε is induced by and transduces the G-CSF signal in neutrophils early in myelopoiesis (37). Absence of secondary granules, defensins, eosinophil cationic protein, eosinophil-derived neurotoxin (12), and platelet α granule high-molecular-mass von Willebrand factor (14) in SGD demonstrates a critical role for C/EBPε in the development of granules and their contents in multiple myeloid lineages.

We are grateful to Dr. Helene Rosenberg for providing SGD patient bone marrow RNA and Dr. Mitchell Horwitz for providing normal peripheral blood CD34⁺ selected cells and expertise.