Introduction

Severe congenital neutropenia (SCN), originally described by Rolf Kostmann in 1950 [1], is a primary immunodeficiency syndrome characterized by early onset of severe bacterial infections. Due to a paucity of peripheral neutrophil granulocytes (<500/µl) patients are prone to invasive infections, typically in the absence of pus formation. “Infantile agranulocytosis”, as Rolf Kostmann named this hereditary syndrome, has been known for more than half a century, yet the underlying genetic mutations have remained unknown for many decades.

Modern genetic research in the field of congenital neutropenia has not only allowed a rational approach to classifying SCN but has also shed light on molecular mechanisms controlling the development and function of neutrophil granulocytes.

From a clinical point of view, two major groups of inherited immunodeficiency diseases associated with congenital neutropenia can be distinguished, first, severe congenital neutropenia without additional organ manifestations, and second, severe congenital neutropenia associated with syndromal features [2].

Here, we provide an overview of recent progress in identifying genetic defects causing congenital neutropenia.

  1. (a)

    Congenital neutropenia without syndromal features

Severe congenital neutropenia

In contrast to cyclic neutropenia, which characterized by oscillating neutrophil counts, SCN classically shows continuously low neutrophil counts. The typical morphological finding is a “myeloid maturation arrest” at the level of promyelocytes, seen in conventional bone marrow smears.

Using genetic linkage studies, Horwitz and colleagues defined mutations in the gene encoding neutrophil elastase (ELA2) as cause of autosomal dominant cyclic neutropenia [3]. Subsequently, sporadic and autosomal dominant cases of SCN have also been associated with ELA2 mutations (reviewed in [4]), suggesting that the feature of cycling hematopoiesis may not be a strong discriminatory factor defining subgroups of neutropenia disorders. Even though certain ELA2 mutations appear to be seen more frequently in patients with cyclic neutropenia, it remains unclear, why some ELA2 variants are associated with cyclic neutropenia while others are associated with SCN (Table 1).

Table 1 Genetic causes of congenital neutropenia
Full size table

ELA2 mutations are either transmitted in an autosomal dominant inheritance pattern or occur sporadically. Recently, Boxer and co-workers published compelling data supporting autosomal dominant transmission of ELA2 mutations as cause for SCN. Five unrelated children, all affected by severe congenital neutropenia, were found to be descendents of a single sperm donor carrying an ELA2 mutation, whose semen was used to impregnate five healthy mothers [5]. A few patients with mosaicism have also been reported [6, 7].

In Caucasian patients, ELA2 mutations appear to represent the most common genetic variant causing SCN, accounting for approximately 50–60% of SCN [8, 9]. Neutrophil elastase is a serine protease exclusively expressed in neutrophils and monocytes and displays pleiotropic functions. It has direct antimicrobial activity and also integrates inflammatory responses by proteolytic modification of chemokines, cytokines, and cell surface receptors (reviewed in [10]).

It is still unclear, how heterozygous mutations in ELA2 cause SCN. Benson suggested that aberrant intracellular trafficking of mutated neutrophil elastase may explain non-physiological enzymatic activity ultimately leading to neutropenia [11]. Alternatively, mutated elastase may be misprocessed and degraded via the activation of the so-called unfolded protein response [12]. For a comprehensive review, the reader is referred to an excellent summary on this topic [4].

Few SCN patients have been described who have mutations in WAS [13], the gene encoding the Wiskott–Aldrich-Syndrome protein. While patients with Wiskott–Aldrich syndrome have loss of function mutations in WAS [14], patients with congenital neutropenia may have gain of function mutations. Physiologically, inadequate activation of WASP is controlled by a conformational change leading to autoinhibition. Certain mutations in WAS prevent this autoinhibitory conformation and thus lead to excessive actin polymerisation. This altered state of the actin cytoskeleton also implies aberrant mitosis and may lead to chromosomal instability and myelodysplastic features [15].

It remains controversial whether the phenotype of myeloid maturation arrest may be caused by increased cell death of mature neutrophils and/or aberrant myeloid differentiation. Since differentiation is controlled by transcription factors, a search for dysfunctional transcription factors involved in SCN has been initiated by several groups. GFI1 is a transcriptional repressor and splice-control factor of the zinc-finger family of transcription factors and has been implicated in the differentiation of various hematopoietic cell lineages, including neutrophils [16, 17]. Two families with heterozygous mutations in GFI1 and congenital neutropenia have been described [18]. GFI1 mutations act in a dominant-negative way and cause dysregulation of several target genes such as C/EBPepsilon [19], ELA2 [18], and the monopoietic cytokine Csf1 [20].

Further evidence for aberrant transcriptional control of neutrophil development in SCN patients stems from an analysis of the Wnt signaling pathway. Wnt proteins are involved in mitogenic stimulation, cell fate specification, and differentiation. In unstimulated cells the transcriptional regulator β-catenin is degraded via a complex interplay of factors, such as disheveled, glycogen synthase kinase-3β, Axin, and adenomatous polyposis coli. In the so-called canonical Wnt pathway, Wnt proteins bind to their receptor, the frizzled/low density lipoprotein receptor-related protein complex, and inhibit the degradation of β-catenin. β-catenin accumulates in the cytoplasm and nucleus where it regulates transcription through interaction with the lymphoid enhancer factor (LEF-) 1 (reviewed in [21]). Skokowa and co-workers provided evidence that in myeloid progenitor cells from Kostmann-type SCN patients LEF-1 as well as LEF-1 target genes, e.g. cyclin D1, c-myc. ELA2, surviving, and C/EBPα are transcriptionally downregulated [22]. Further mechanistic studies will have to show how decreased LEF-1 may inhibit neutrophil differentiation.

While ELA2 and GFI1 may account for sporadic or autosomal dominant variants of SCN, the underlying molecular genetic defect of classical, autosomal recessive Kostmann-type SCN remained enigmatic more than 50 years after the first description (reviewed in [23]). Using linkage analysis and a candidate gene sequencing approach, we have identified homozygous mutations in the antiapoptotic molecule HAX1 in patients with autosomal recessive SCN [24]. HAX1 is known to interfere with signal transduction and cytoskeletal control [25–27]. Furthermore, HAX1 is a critical protein for maintaining the inner mitochondrial membrane potential and protects myeloid cells from apoptosis [24]. Deficiency of this key protein destabilizes the inner mitochondrial membrane potential and leads to SCN via increased apoptosis of myeloid progenitors. At least two transcripts of HAX1 have been identified. Interestingly, some patients with HAX1 deficiency suffer from neurological problems, ranging from mild cognitive problems to severe developmental delay and epilepsy [28–31]. These differences in phenotype can be attributed to the type of mutation in HAX1, which has two distinct splice variants. In case the mutation affects only isoform A, patients suffer from congenital neutropenia but not from neurological problems. In case the mutation affects both isoform A and isoform B, all patients show signs and symptoms of neurological impairment, indicating a strict genotype–phenotype correlation [32, 33].

Interestingly, not all patients with autosomal recessive SCN have mutations in HAX1. Ongoing studies in our laboratories suggest that there are several additional genes involved in the pathogenesis of autosomal recessive SCN (unpublished). Despite the phenotypical similarities in patients with HAX1 deficiency and ELA2 mutations [34], no common mechanism has been characterized explaining why both of the defects lead to isolated congenital neutropenia.

  1. (b)

    Congenital neutropenia associated with syndromal features

A subset of patients with congenital neutropenia has additional features such as lymphoid cell dysfunction, pigmentation defects, inner ear hearing loss, or cardiac/urogenital anomalies. Elucidation of the underlying genetic and molecular defects has provided fascinating insights into the basic function of mitochondria, lysosomes, and the endoplasmic retriculum.

Reticular dysgenesis

Reticular dysgenesis is an autosomal recessive form of early differentiation arrest in the myeloid lineage associated with severe combined immunodeficiency due to impaired lymphoid development [35]. Furthermore, affected patients suffer from sensorineural hearing loss [36]. Congenital neutropenia in reticular dysgenesis is non-responsive to granulocyte colony stimulating factor (G-CSF) [37]. Using a genome linkage analysis and candidate gene sequencing approach, mutations in adenylate kinase 2 (AK2) have been identified by two independent groups [38, 39]. Like HAX1, AK2 is localized in the mitochondrial intermembrane space and may be important in mitochondrial energy metabolism and control of apoptosis via FADD and caspase 10 [40].

Congenital neutropenia with hypopigmentation

While Chédiak-Higashi syndrome (CHS) [41, 42] and Griscelli syndrome type 2 (GS2) [43] may only show transient episodes of neutropenia, Hermansky–Pudlak syndrome type 2 (HPS2) and p14 deficiency are characterized by constantly reduced peripheral neutrophil counts [44, 45]. All four disorders can be differentiated based on their clinical phenotype [46]. In contrast to SCN with myeloid maturation arrest, CHS, GS2, HPS2, and p14 deficiency show presence of mature neutrophils in the bone marrow.

Hermansky–Pudlak syndrome was originally described as oculocutaneous albinism and platelet storage disease with increased bleeding tendency [41, 42]. Currently, HPS comprises a group of at least eight distinct genetic defects in humans (reviewed in [47]). All defective genes play a role in intracellular protein trafficking. As a result of the mutations, dysfunction of lysosome-related organelles (LROs) is seen in cells from HPS patients.

Among all variants of HPS, only HPS2 is characterized by congenital neutropenia [48–50]. The first patients, two descendents of a consanguineous pedigree, were described by Kotzot and colleagues in 1994 [51]. Up to date, less than ten patients have been described in the literature [44, 48–55]. HPS2 is caused by mutations in the AP3B1 gene which encodes the β3A subunit of the heterotetrameric adaptor protein (AP-) 3 complex. β3A deficiency leads to an almost complete abolishment of the remaining AP-3 subunits, i.e. δ, μ3A and σ3A. The AP-3 complex mediates cargo protein selection into transport vesicles and their trafficking to lysosomes [53, 56]. In the absence of AP-3, proteins destined to traffic to lysosomes therefore accumulate at the plasma membrane. It remains unknown why patients with AP-3 deficiency are neutropenic. So far, no functional defect could be identified in AP-3-deficient neutrophils [44]. Interestingly, AP-3-deficient dogs show cyclic neutropenia and AP-3-deficient mice appear to have no neutrophil defect (reviewed in [47]). However, AP-3 deficiency is associated with defects of NK and NKT cells [44, 55]. Lethal hemophagocytosis has recently been reported in one patient [52], suggesting that defective cytotoxicity may be clinically relevant in AP-3 deficiency.

A previously unrecognized immunodeficiency syndrome associating growth delay, partial albinism, lymphoid deficiency, and congenital neutropenia has recently been described in four members of a large pedigree [45]. Using a combination approach of genome wide linkage analysis and transcriptional profiling, we identified a homozygous mutation in the gene encoding the endosomal adaptor molecule p14. A point mutation in the 3’-UTR leads to decreased p14 mRNA. As a consequence of decreased p14 protein levels, neutrophils show an altered ultrastructure of azurophilic granules and decreased microbicidal activity in phagosomes. Furthermore, p14-deficient cells have a defect in cytokine-receptor-mediated ERK phosphorylation and show a marked dislocalisation of late endosomes. p14 is ubiquitously expressed and located at the outer membrane of late endosomes [57, 58]. The molecule acts as a scaffold protein by interacting with MEK partner (MP) 1 to direct the mitogen-activated protein kinase pathway. Its function is crucial for subcellular compartmentalization of signals [59]. Recently, a murine p14-deficient mouse model has been generated [60]. Complete knock-out of the p14 gene causes early intrauterine lethality. Epidermis-specific knock-out of p14 in mice results in a severely reduced number of epidermal skin layers, perturbed terminal skin differentiation, aberrant epidermal growth factor receptor degradation, and defective cellular proliferation. Affected mice die shortly after birth [60]. These studies support the idea that p14 is critical for cellular proliferation and differentiation. p14-deficient patients may only survive because of residual p14 protein levels (hypomorphic phenotype). Of note, some patients with a phenotype of p14 deficiency are known who do not show any mutation in p14, suggesting that this disorder is heterogeneous.

Congenital neutropenia and glycogen metabolism

Neutrophils are exquisitely sensitive to glucose deprivation. Two human genetic disorders have been identified in which neutrophil granulocytes show a predisposition to accelerated apoptosis due to disturbances of energy metabolism in the endoplasmic reticulum. Patients with glycogenosis type 1b, a variant of glycogenosis type 1, is not only characterized by glycogen storage, hypoglycaemia, and lactic acidosis, but also show a paucity of mature neutrophil granulocytes. This disease is caused by mutations in the ubiquitously expressed glucose-6-phosphate translocase (G6PT, encoded by the gene SLC37A4), a transporter mediating translocation of G6P into the endoplasmic reticulum. G6PT-deficient neutrophils show a significantly impaired respiratory-burst response. Therefore, GSD1b patients are not only at risk for severe bacterial infections due to neutropenia but may also develop clinical symptoms secondary to poor neutrophil function, such as inflammatory bowel disease [61].

Recently, we identified a second congenital neutropenia syndrome associated with defective glucose metabolism. In contrast to GSD1 and GSD1b, affected patients do not show signs of glycogen storage or symptoms related to hypoglycemia. Instead, there is an association of congenital neutropenia, various developmental defects of the cardiovascular and/or urogenital system and increased visibility of superficial veins [62]. There is a high degree of variation in clinical presentation and the determinants of the wide clinical spectrum remain unknown at the present time. Cardiac anomalies may include atrial septal defect, cor triatriatum or pulmonary vein stenosis, and urogenital problems comprise urachal fistulation, renal dysgenesis, or undescended testes. Myeloid progenitor cells are characterized by increased signs of the “unfolded protein response” or ER-stress, as evidenced by ultrastructural anomalies and increased expression of the chaperone BiP [62]. The defective protein, G6PC3, is a paralog of G6PC1, the classical glucose-6-phosphate, mutated in GSD1. In contrast to G6PC1, G6PC3 is expressed ubiquitously, suggesting that G6PC3 may in fact be the phylogenetically more ancient variant of glucose-6-phosphatase. G6PC2 is expressed in the pancreatic islets and has recently been implicated in diabetes mellitus [63, 64]. In the absence of G6PC3, myeloid cells are prone to undergo apoptosis, and we speculate that aberrant apoptosis due to increased ER-stress may also be the cause for defective organ development.

WHIM syndrome (warts, hypogammaglobulinemia, immunodeficiency, myelokathexis)

Myelokathexis is defined as dysgranulopoiesis (hypersegmented neutrophils, increased apoptosis) with increased bone marrow cellularity. First described by Zuelzer in a sporadic case [65], myelokathexis was subsequently found to occur as an autosomal dominant trait in association with hypogammaglobulinemia and dermal/genital papilloma virus infections [66]. Severe infections are rare since neutrophils can be adequately released to the periphery in case of infections.

Hernandez and colleagues [67] identified heterozygous mutations in the chemokine receptor gene CXCR4 in WHIM syndrome patients. Mutant CXCR4 shows significantly increased calcium flux in response to its ligand stromal-derived factor-1α (SDF-1α) consistent with dysregulated signaling of the mutant receptor. Likewise, SDF-1α stimulation of neutrophils with a truncating CXCR4 mutation, but not with non-mutated CXCR4, blocked the constitutive apoptosis seen in neutrophils from WHIM patients [68]. Kawai and colleagues [69] investigated the role of CXCR4 mutations in myelokathexis. They transduced wild type and mutated CXCR4 in human CD34+ peripheral blood-mobilized stem cells (PBSCs) and analyzed the properties of these cells in cell culture and after transplantation into nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice, an elegant xenotransplantation system to unequivocally assess the engraftment and expansion of the transplanted human cells. NOD/SCID mice are devoid of B and T cells and have a defect in Natural Killer cell function. Thus, they do not reject the xenograft. Interestingly, mutant CXCR4 transduced PBSCs behaved like the wild-type CXCR4 transduced PBSCs in the cell culture system but showed a better engraftment with a significantly enhanced apoptosis and decreased release of mature neutrophils from the BM only after transplantation into NOD/SCID mice suggesting that the CXCR4 mutation alone is insufficient to explain the increased apoptosis of BM stem cells in WHIM patients.

Interestingly, not all patients with myelokathexis have mutations in CXCR4 [70]. In particular, the WHIM phenotype can also be transmitted as an autosomal recessive trait, suggesting that other genes are involved in this disorder.

Natural history, treatment, and complications

Before recombinant human granulocyte colony stimulating factor became available, most patients with SCN succumbed to severe and recurrent bacterial infections in infancy [70–72]. In the G-CSF era, morbidity and mortality from infectious complications has greatly decreased and the quality of life has markedly improved. Today, most patients with congenital neutropenia survive into adulthood.

G-CSF has pleiotropic effects by orchestrating a series of molecular events in favor of neutrophil differentiation. In addition to its effects on differentiation, G-CSF has antiapoptotic effects. A very recent study has shown that G-CSF acts through Nampt/PEBF to increase intracellular NAD1, which in combination with SIRT1 binds and activates C/EBP-beta and leads to the so-called emergency neutrophil granulopoies [73].

Long-term studies, facilitated by the international congenital neutropenia registries in Seattle, USA, and Hannover, Germany, revealed long-term complications. There is a small but significant risk of death secondary to infectious complications [74]. In spite of effective G-CSF treatment, there is a risk (0.9% per year) of SCN patients to die as a consequence of infectious complications [74]. Thus, the reconstitution of normal neutrophil counts may not suffice to re-establish antibacterial immunity. ELA2-mutated patients show abnormal expression of neutrophil elastase and other granule-associated proteins such as myeloperoxidase, lactoferrin, cathepsin-G, and human-neutrophil-peptide [75]. These abnormalities may contribute to altered neutrophil function. More importantly, many SCN patients develop clonal hematopoietic disorders such as myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML). After 12 consecutive years of treatment, the risk is 8% per year [74]. Long-term studies have not shown a plateau effect, so there is a certain likelihood that the overall risk of developing leukemia continues to rise. The molecular etiology of leukemogenesis remains unknown. Many SCN patients develop somatic mutations in the CSF3R gene. Germeshausen and colleagues followed up on CSF3R mutations in 148 SCN patients 23 of whom developed signs of malignant transformation [76]. Eighteen of these individuals had acquired CSF3R mutations. Additionally, in 50 patients analyzed they could show an accumulation of different CSF3R mutations over time.

CSF3R mutations appear in early hematopoietic progenitor cells and confer a survival advantage in myeloid cells [77]. This observation has also been seen in murine studies. Liu et al. [78] used a murine model in which HSC expressed a truncating mutation of the G-CSF-R. This study demonstrated that administration of G-CSF permitted the outgrowth of G-CSF-Rmut stem cells in competitive repopulation experiments. This selective advantage was associated with and dependent on G-CSF induced increased activation of STAT5. However, no autonomous G-CSF-independent growth was observed and none of the mice developed AML or MDS.

In spite of successful supportive treatment using recombinant human G-CSF, definitive and curative treatment of SCN is currently only possible by means of allogeneic bone marrow transplantation. Preclinical studies using transplantation of genetically corrected hematopoietic stem cells are under way, but no clinical protocol has been opened up to date.

Conclusions and perspectives

Suspicion of congenital neutropenia should prompt a complete diagnostic workup including molecular diagnosis. The identification of underlying genetic defects now allows a molecular diagnosis in many patients, a prerequisite for a better assessment of the natural evolution and therapy-related risks. Furthermore, genetic diagnosis is the first step towards development of gene-based therapies. However, in many patients with congenital neutropenia underlying mutations are still unknown. Therefore, sustained scientific efforts and multi-national collaborations are needed to elucidate the genetic basis of all remaining variants of SCN.