The patient presented to our clinic at the age of 17 years, diagnosed with SCN of unknown cause during early childhood. Previous infections included a gangrene-induced sepsis with need for intensive care treatment. The patient had no dysmorphisms, no congenital skeletal defects, and did not display symptoms of organ failure. On first admission to our clinic, the patient was on G-CSF treatment with subcutaneous (s.c.) filgrastim at 50 µg/kg/d. Yet, very low neutrophil numbers were noted (around 200/µl). Genetic assessment at that time point only revealed a heterozygous somatic stop-mutation in the intra-cytoplasmatic region of the G-CSF-receptor gene (p.Q741*). At the age of 18 years, the patient became pregnant. After in-depth considerations of potential risks for the fetus, the patient was offered s.c. treatment with GM-CSF (sargramostim/Leukine^®). Given the previous long complicated course of disease without causal treatment options she agreed with the experimental application after thorough consideration in light of her ongoing pregnancy. At 12 weeks of gestation, GM-CSF was administered for 7 days, initially at 5 µg/kg/d (=165µg/m²/d) for 5 days. As no rise in neutrophils was detected, the dose was escalated to 10 µg/kg/d (=330µg/m²/d). At day 8, the neutrophils dropped below 100/µl and the patient presented clinically with gingivitis, ultimately requiring intravenous (i.v.) antibiotics. In the absence of a response, GM-CSF was stopped and G-CSF was restarted ( Figure 1A ). In spite of G-CSF therapy, the patient´s neutropenia persisted and she was admitted to the hospital repeatedly during her pregnancy with infectious complications such as severe gingivitis, fever of unknown origin, dyspnea, and two episodes of urosepsis. At 40 + 5 weeks of pregnancy, a healthy baby with normal blood cell and neutrophil counts was born. In spite of being on G-CSF, the patient continued to have infectious complications. Bone marrow assessment showed maturation arrest even under G-CSF treatment ( Figure 2A ). In the meantime, genetic testing revealed a homozygous missense mutation in JAGN1 (c.63G>T, p.Glu21Asp), which has previously been reported (4). At the age of 20 years, she underwent HSCT from a matched related sibling. At the last available follow-up two years after HSCT, she had full donor chimerism, normal neutrophil counts and no clinical issues ( Table 1 ).

A 4-month-old girl with pneumonia, elevated CRP (282 mg/L), anemia (8 g/dL) and severe neutropenia (200/µL) was referred to our hospital. Infection history was positive for neonatal infection at the age of 6 weeks and pneumonia at the age of 10 weeks, requiring inpatient treatment. In the bone marrow, there was a maturation arrest of granulopoiesis in line with a diagnosis of SCN ( Figure 2B ). The patient was immediately started on G-CSF (5µg/kg/d), which led to monocytosis without persistent effect on neutrophil numbers. The patient received posaconazole treatment due to oral thrush and elevated beta-D-glycan. At 5 months, a single dose IgG was given (IgG level 1.35 g/l) ( Table 2 ). Two compound heterozygous mutations in JAGN1 were detected (c.36del p.Asp13Thrfs*17 and c.63G>T, p.Glu21Asp). As the patient was highly symptomatic (recurrent febrile infections with CRP > 100 mg/l) and did not respond to G-CSF therapy, hematopoietic stem cell transplantation (HSCT) was recommended, based on case reports for JAGN1 deficiency. While preparing for HSCT, G-CSF therapy was continued without achieving a significant increase in the neutrophils. We offered daily GM-CSF treatment, which was accepted by the mother after informed consent, as she hoped HSCT may be dispensable. A three-day break was required due to organizational matters and was bridged by G-CSF (5 µg/kg/d s.c.). As an initial dose of 12 µg/kg/d (=212µg/m²/d) GM-CSF s.c. did not show any quantitative effect, the dose was subsequently increased to 20 µg/kg/d (350µg/m²/d) s.c. for another 6 days ( Figure 1B ). Total treatment time was 9 days. GM-CSF was discontinued and G-CSF re-started as the neutrophils fell below 100/µl and signs of infection were noted. At the age of 8 months the patient underwent HSCT from a mismatched unrelated donor (9/10). The patient is well 2 years after HSCT showing normal neutrophil counts and full donor chimerism on last assessment ( Table 1 ).

Both patients’ neutrophil count did not respond to G-CSF therapy resulting in a markedly increased infection history and long-lasting severe neutropenia under G-CSF. They both displayed persistent (patient 2) or recurrent (patient 1) monocytosis in peripheral blood while on G-CSF treatment ( Table 1 ). G-CSF therapy did not greatly affect eosinophil counts with mean values of around 220/µl in both patients. On treatment with GM-CSF, the neutrophil count ultimately dropped down below 100/µl in both patients, and GM-CSF was stopped as the patients subsequently developed signs of infection requiring antibiotic treatment. There was a rise in eosinophils under GM-CSF therapy, confirming general efficacy of GM-CSF. In both cases, monocytes were already elevated on G-CSF therapy, and did not show major variations on GM-CSF therapy ( Figures 1A, B ).

In contrast to the persistent severe neutropenia with a mean around 300/µl in both patients, no major other immunological deficiencies were evident ( Table 2 ). Both patients showed B-cells above the upper limit, but the adult had hypergammaglobulinemia while the infant had an IgG below the age-specific value. They both had normal IgA and IgM values and the infant displayed a normal B-cell differentiation. At the time of immunophenotyping the infant patient was recently treated for an infection and the lymphocyte panel showed an expansion of particularly CD8 effector and CD8 effector memory T-cells. CD4 T-cell counts were normal in both patients.

G-CSF ameliorates the phenotype in around 90% of patients with SCN by improving neutrophil survival (15, 19, 22). In contrast, JAGN1 patients have been described as low-responders (4, 6). In line with this, both our patients presented with a severe infectious phenotype and neutrophils below 500/µl in spite of G-CSF treatment. JAGN1-mutant neutrophils show increased sensitivity to apoptotic and necrotic stimuli and defective protein glycosylation as well as reduced granules and aberrant granule exocytosis (4). JAGN1 is located at the ER and mutated protein enhances ER stress and apoptosis. Recent data suggests that the induction of calcium- and calpain-dependent cell death in myeloid cells is a major contributor to the manifestation of JAGN1-mutant SCN as evidenced from experiments using HL-60 cells (5). As G-CSF displays part of its anti-apoptotic features via inhibition of calpain and calcium-influx (23), it, however, still remains unclear why many, albeit not all JAGN1 patients are often non-responsive to G-CSF treatment. One potential explanation may be the presence of deficient N-glycosylation of the G-CSF receptor in patients with JAGN1-associated neutropenia, which has been suggested to contribute to inefficient G-CSF-R mediated signaling and thus G-CSF unresponsiveness (4). Notably, several other genes leading to SCN have also be associated to ER dysfunction. Among those are mutations in G6PC3, which also lead to defects in glycosylation. Upon treatment with G-CSF, neutrophil numbers in patients with G6PC3 defects do, however, usually increase (24). Further research is therefore needed to address to which extent aberrant glycosylation patterns in JAGN1 SCN affect the treatment response to G-CSF and whether different JAGN1 mutations lead to diverse functional consequences on the molecular level.

Experiments by Wirnsberger et al. suggested a beneficial effect of GM-CSF on neutrophil function and Candida defense in a mouse model of hematopoietic JAGN1 deficiency as well as in human JAGN1-mutant bone marrow cells (13). In both our patients, however, there was neither an associated rise in neutrophil numbers nor clinical protection from infection under the otherwise well-tolerated GM-CSF therapy. The administered starting doses in our patients were well within recommended standard doses for various bone marrow stimulating applications of GM-CSF. Escalating the doses further did not lead to a beneficial effect. The appearance of infectious signs, concomitantly with a drop in neutrophils below 100/µl, ultimately resulted in discontinuation of GM-CSF therapy. Notably, G-CSF has also been ineffective in some but not all patients with inherited biallelic CSF3R mutations in the extracellular domain of the G-CSF receptor (25), and beneficial effects by GM-CSF have been reported (17, 26). Among the reported cases, 1 patient with compound heterozygous CSF3R germline mutation responded to GM-CSF (final dose 3 µg/kg twice a week) without adverse events over a 12-year period (17). On this dose, the patient’s neutrophil counts remained over > 1000/µl and no infections occurred. Another patient with homozygous germ line deletion in the CSF3R gene was also effectively treated with GM-CSF, and increased neutrophils from < 500/µl to > 1000/µl were already noticed after 2 injections with 5 µg/kg/d every other day and infections disappeared (27). Zhou et al. reported another case of biallelic CSF3R mutation, who was unresponsive to G-CSF, but maintained neutrophils around 1000/µl and remained infection free on a final dose of 3µg/kg/d once a week (16). This argues that the observation time and dose administrated in our patients should have been sufficient to expect a treatment response. Patients with CSF3R mutations frequently display full myeloid cell maturation in the bone marrow in spite of severe peripheral neutropenia (18). Notably, both our patients did show a characteristic maturation arrest in the bone marrow ( Figures 2A–C ), and this was also the case in most but not all patients originally reported (4). The divergent response to GM-CSF in our patients in contrast to the patients with CSF3R mutations and known reduced G-CSF-Receptor signaling underlines that neutrophil maturation and survival in JAGN1-mutant SCN is affected at different levels.

Differentiation from hematopoietic stem cells to mature neutrophils is driven predominantly by G-CSF and to a lesser extent by GM-CSF in combination with other cytokines (28). In healthy individuals, nicotinamide phosphoribosyltransferase (NAMPT) activation has been suggested to be essential for both GM-CSF and G-CSF triggered granulopoiesis (29), and inhibition of NAMPT decreased both G-CSF and GM-CSF-induced colonies from hematopoietic cells in vitro (30). In patients with SCN due to ELANE and HAX1 mutations, insufficient activation of NAMPT by GM-CSF, but marked elevation by G-CSF was observed and was suggested to account for the lacking treatment response to GM-CSF in these patients (30). Further functional studies, ideally using primary cells from JAGN1-mutant patients, are needed to assess whether aberrant NAMPT signaling could potentially play a role in the low response for both GM-CSF and G-CSF in JAGN1-mutant SCN.

Apart from its antiapoptotic properties, GM-CSF is known to enhance cytotoxic Candida albicans killing (28). JAGN1 was found to be required for MPO expression and Candida killing within NETs (neutrophil extracellular traps), and GM-CSF was able to restore the MPO expression in NETs and upregulated calprotectin expression in a model of HL-60 derived neutrophil-like cells with silenced JAGN1 expression (31). Pre-treatment with GM-CSF also restored fungal killing capacities in bone marrow cells from JAGN1 patients (13). While patients with neutropenia may be prone to fungal infections, the risk of severe affection by bacterial infection is much higher. Thus, the known pro-inflammatory properties of GM-CSF are likely insufficient in patients, where neutrophil counts remain below a certain threshold as observed in our patients. Notably, our adult patient, although severely neutropenic and highly susceptible to bacterial infections, did not suffer from any fungal infections, while the infant achieved good fungal control with posaconazole.

Experimental data from a mouse with JAGN1-deficient B-cells suggests a crucial role of JAGN1 for IgG production and humoral immune response (32). In human JAGN1-mutant SCN patients, hypogammaglobulinemia has been reported in some, but not in all patients (6, 11, 33). Our infant patient also had IgG below the age-specific normal values, while the young woman had elevated IgG-levels prior to HSCT. As the adult patient experienced several severe infections prior to HSCT we hypothesize that the hypergammaglobulinemia was caused by constant immune stimulation. In the infant, the reduced IgG value was not accompanied by aberrancies in B-cell differentiation. Specific antibody responses were not available. Although the reason behind the variable immunological phenotype remains to be determined, assessment of immunoglobulin levels and specific antibody responses in neutropenic patients of unknown origin appears reasonable from a clinical point of view.

JAGN1 is ubiquitously expressed, thus extra-hematopoietic manifestations are possible. Regarding the extended clinical phenotype there appears to be quite some variability and no strict genotype-phenotype association has been observed so far (4, 34). Extra-hematopoietic manifestation mentioned in literature are short stature, pyloric stenosis, scoliosis, convulsions, osteoporosis, pancreatic insufficiency, amelogenesis imperfecta, urogenital malformations, hypothyroidism and mild facial dysmorphism (11). In addition, one JAGN1-mutant SCN patient with bleeding abnormalities and recurrent intracranial hemorrhage has also been reported (33). Both our patients did not show systemic abnormalities and only very mild phenotypic variation such as a triangular face in patient 2 as described by others (6). While lack of apparent extra-hematopoietic manifestation may be a reflection of the young age in patient 2, normal findings in patient 1 in spite of a severe neutropenic and infectious phenotype suggests, that additional factors may play a role in the development of extra-hematopoietic features. The low number of reported patients with currently less than 25 hampers our understanding of potential genotype-phenotype correlations. Notably, many of the reported patients are derived of consanguineous families. Thus, there may be additional genes but also environmental factors likely involved in the systemic features, the degree of severity and the presence of additional immunological aberrations.