Neutrophil dysfunction triggers inflammatory bowel disease in G6PC3 deficiency

The glucose‐6‐phosphatase catalytic subunit 3 (G6PC3) encodes a ubiquitously expressed enzyme that regulates cytoplasmic glucose availability. Loss‐of‐function biallelic G6PC3 mutations cause severe congenital neutropenia and a diverse spectrum of extra‐hematological manifestations, among which inflammatory bowel disease (IBD) has been anecdotally reported. Neutrophil function and clinical response to granulocyte colony‐stimulating factor (G‐CSF) and hematopoietic stem cell transplantation (HSCT) were investigated in 4 children with G6PC3 deficiency‐associated IBD. G6PC3 deficiency was associated with early‐onset IBD refractory to treatment with steroids and infliximab. The symptoms of IBD progressed despite G‐CSF treatment. In vitro studies on the patients’ blood showed that neutrophils displayed higher levels of activation markers (CD11b, CD66b, and CD14), excessive IL‐8 and reactive oxygen species, and increased apoptosis and secondary necrosis. Secondary necrosis was exaggerated after stimulation with Escherichia coli and could be partially rescued with supplemental exogenous glucose. HSCT led to normalization of neutrophil function and remission of gastrointestinal symptoms. We conclude that neutrophils in G6PC3 deficiency release pro‐inflammatory mediators when exposed to gut bacteria, associated with intestinal inflammation, despite treatment with G‐CSF. HSCT is an effective therapeutic option in patients with G6PC3 deficiency‐associated IBD refractory to immune suppressants.


G6PC3 deficiency is caused by biallelic loss-of-function variants in
G6PC3, which encodes glucose-6-phosphatase-(G6Pase-), a ubiquitously expressed enzyme that regulates cytoplasmic glucose availability through hydrolysis of glucose-6-phosphate during glycolysis and glycogenolysis. 12 As well as SCN, G6PC3 is associated with a diverse spectrum of extra-hematological manifestations, including cardiac and urogenital abnormalities. 12 G6PC3 deficiency results in neutropenia because neutrophils are highly dependent on glycolysis as an energy source. 13,14 Studies in humans and mice with G6PC3 deficiency have demonstrated neutrophil dysfunction characterized by accelerated apoptosis. 10,[15][16][17] However, IBD is not a feature of other SCNs that also demonstrate accelerated neutrophil apoptosis. 18 This suggests a mechanism independent of apoptosis underpins IBD in G6PC3 deficiency.
Treatment with G-CSF normalizes peripheral neutrophil count and reduces the risk of neutropenic infections, but it is not known if it promotes G6PC3 deficiency-associated IBD. 12,18 Here, we describe 4 children with G6PC3 deficiency-associated IBD and neutrophil dysfunction characterized by excessive production of pro-inflammatory mediators and increased necrosis. Hematopoietic stem cell transplant (HSCT) was performed in 3 children and led to normalization of neutrophil function and remission of IBD, and thus should be considered as a therapeutic option for refractory G6PC3 deficiencyassociated IBD.

Participants
The research protocol was approved by the NHS Research Ethics

Bacteria
Escherichia coli American Type Culture Collection (ATCC) 25922 (gift from Professor Ian Roberts, University of Manchester) were grown in lysogeny broth medium (Thermo Fisher) at 37 • C and harvested in midlog phase. gating determined using the fluorescence-minus-one principle.

Statistical analysis
Statistical analysis was undertaken using Prism (Version 7.0, GraphPad Software   Table 1). Homozygous variants in G6PC3 were identified in all 4 patients (Supplemental Fig.   1). In 2 patients, mutations were truncating (P1: c.282delA and P2: c.882_903dup) and predicted to be more deleterious and were associated with a more severe pattern of extra-hematological manifestations (Supplemental Table 1 Table 2). 19 P3 and P4 did not require endoscopic evaluation. The colitis was poorly responsive to prednisolone in patients P1-P3 and azathioprine, sulfasalazine, and infliximab in patients P1 and P2.
Both amelioration 5,7 and deterioration 2,8 of enteritis has been described following the commencement of G-CSF therapy in patients with G6PC3 deficiency. In all our patients, severity of the diarrhea increased following the initiation of G-CSF therapy, except P1 who had commenced treatment in the neonatal period. Absolute neutrophil count (ANC) was maintained on a daily G-CSF dose of 3-10 mcg/kg/day in all patients except P3 who received weekly pegylated G-CSF injections at 120 mcg/kg/week (Supplemental Table 1). In P3, ANC trough was 1.7 × 10 9 /L and peak was 9.7 × 10 9 /L 2 days after pegylated G-CSF injection, which coincided with a significant increase in stool frequency (Fig. 1A).
HSCT can cure neutropenia and has been recently reported to ameliorate IBD in an adult with G6PC3 deficiency. 11, 20 We first performed HSCT in P4 primarily due to poorly tolerated local and systemic side effects of G-CSF injections, and the wishes of the patient and her family, as well as her enteritis (Table 1). We then performed HSCT on P2 and P3 for treatment-refractory enteritis (   Table 2).
In addition, all are now attending school or university 2-3 years post-HSCT, with the only residual medical problems being bronchiectasis requiring prophylactic antibiotics (P2 and P4), and gonadal dysfunction requiring hormone replacement (P2 and P3; Supplemental Table 2).
Given that G-CSF injections were associated with increased stool frequency and HSCT was associated with remission of gastrointestinal symptoms, we considered whether G6PC3 deficiency-associated IBD may be driven by phagocyte dysfunction. As previous studies have described accelerated neutrophil cell death from patients with G6PC3 deficiency, 10,15,16 we hypothesized that this might drive intestinal inflammation, particularly in the gut where neutrophils are exposed to commensal microbes. We therefore assessed neutrophil cell death following incubation with live Escherichia coli. Neutrophils from patients with G6PC3 deficiency underwent significantly more apoptosis and secondary necrosis, particularly when incubated with E. coli, compared with equivalents from healthy control volunteers ( Fig. 2A-C). The pattern of neutrophil death in 1 patient (P4) who had undergone HSCT was in keeping with the healthy controls.
Accelerated neutrophil death in G6PC3 deficiency has been attributed to ER and oxidative stress driven by a lack of cytoplasmic G6P. 21 It is therefore possible that increasing the exogenous glucose to supra-physiological levels might rescue neutrophils from death.
We therefore compared cell death of neutrophils from a patient with G6PC3 deficiency (P2) and a healthy control, following culture with increasing concentrations of D-glucose (Fig. 2D). In the complete absence of exogenous glucose, the median proportion of neutrophils undergoing secondary necrosis was similar in the patient with G6PC3 deficiency (14%) and the healthy control (16%; Fig. 2D). Increasing the exogenous glucose concentrations to 1 mM or more led to almost complete rescue of neutrophils of the healthy donor from necrosis (2-3% necrotic neutrophils), while only partial rescue (7-8% necrotic neutrophils) in the patient with G6PC3 deficiency (Fig. 2D). This Significance determined by Student's t-test with Welch's correction for unpaired data where "**" and "***" denote P < 0.01 and < 0.001, respectively (A-C), and 1-way ANOVA with Bonferroni's correction for multiple comparisons where *, **, ***, and **** denote P < 0.05, < 0.01, < 0.001, and < 0.0001, respectively (D) defective glucose metabolism (e.g., HAX1 deficiency) do not suffer from IBD. 23 Increased necrosis implies that neutrophils in patients with G6PC3 deficiency have an altered inflammatory activation status, or a lower proportion that are immature. 24,25 We tested this theory by measuring expression of neutrophil surface markers. 26,27 There was significantly higher expression of CD14, CD66b, and CD11b on freshly isolated live neutrophils from patients with G6PC3 deficiency compared with healthy controls (Supplemental Fig. 2A). There was however no significant difference in the expression of CD64, CD16, TREM-1, and CD62L (Supplemental Fig. 2A) nor the proportion of CD11b high /CD16 low/immature neutrophils (Supplemental Fig. 2B). These observations suggest that neutrophils from patients with G6PC3 deficiency are mature, but constitutively activated.
Constitutively activated neutrophils in patients with G6PC3 deficiency could contribute to IBD through excessive release of soluble inflammatory mediators. We therefore measured soluble inflammatory mediators and found significantly higher IL-8 release, but not TNF-, IL-6, and IL-1 , from the unstimulated neutrophils of patients with G6PC3 deficiency, compared with healthy controls (Fig. 3A).
Colonic IL-8 levels correlate with IBD severity, as excessive IL-8 drives tissue inflammation by its action as a potent neutrophil chemoattract and activation ligand. 28,29 We also found significantly lower release of neutrophil elastase and MMP-9 in E. coli-stimulated neutrophils from G6PC3 patients compared with healthy controls (Fig. 3A).
Dysregulated reactive oxygen species (ROS) production by neutrophils is associated with mucosal inflammation, as observed in colonic resection specimens of IBD patients in which ROS is increased, and IBD caused by CGD in which ROS is reduced. 30,31 We therefore Fluorescence minus one sample denoted by unfilled histogram. Significance determined by Student's t-test with Welch's correction for unpaired data where "*" denoted P < 0.05, and where appropriate control of False Discovery Rate (FDR) by the Benjamini-Hochberg method was applied in which "*" denotes FDR < 0.05 had a significantly lower DHR index compared with healthy controls (Fig. 3B). Where cellular yield was sufficient, stimulation with N-formyl-methionyl-leucyl-phenylalanine (fMLP) was also performed and similar results were observed (Fig. 3C). In contrast, unstimulated neutrophils from patients with G6PC3 deficiency demonstrated a significantly higher mean fluorescence intensity of rhodamine-123 compared with healthy controls (Fig. 3D). The pattern of inflammatory mediator production of the patient with G6PC3 deficiency who had undergone successful HSCT 2 years earlier (P4) closely aligned with the healthy controls. Collectively these observations suggest that constitutively activated neutrophils in G6PC3 deficiency produce excessive pro-inflammatory mediators in steady state conditions but have a reduced capacity to do so upon stimulation.
Given previous reports of mononuclear phagocyte dysfunction in G6PC3 deficiency, we considered whether monocyte dysfunction might also contribute to intestinal inflammation. 9,32 We therefore stimulated freshly isolated peripheral monocytes with LPS and assessed cytokine production and cell death. Under both media and LPS conditions, there were no significant differences in innate cytokine production by monocytes from patients with G6PC3 deficiency compared with healthy controls (Supplemental Figure 3).
There were also no significant differences in the RNA expression of genes involved in regulating apoptosis or LDH production from the monocytes of patients compared with healthy controls (Supplemental Figure 3). Therefore, defective function of peripheral phagocytes in G6PC3 deficiency appears restricted to neutrophils.
For the first time, we show that neutrophils from patients with G6PC3 deficiency exhibit elevated constitutive expression of inflammatory cell surface markers (CD11b, CD66b, and CD14), cytokines (IL-8), and ROS reflecting an intrinsically higher inflammatory activation state of neutrophils in G6PC3 deficiency, possibly because neutrophils are particularly dependent on glycolysis as an energy source. 13,14 Reduced release of neutrophil elastase and MMP-9 from G6PC3 deficient neutrophils suggests a defect in degranulation, which could be tested in future studies. It is unlikely that the observed differences were caused by G-CSF treatment rather than G6PC3 deficiency, as previous studies have demonstrated that G-CSF treatment is associated with reduced apoptosis, normal ROS production and changes in surface marker expression that return to normal after a few days treatment. [33][34][35] An important limitation of our study was the small number of participants and therefore it would be useful if our findings were confirmed in larger independent studies. We were unable to recruit more patients with G6PC3 due to the rarity of the condition, and the necessity to perform functional neutrophil studies on freshly drawn blood. Nonetheless, our data highlight that early-onset IBD can be a major complication of G6PC3 deficiency associated with significant morbidity. Gastrointestinal symptoms were more severe in patients with genotypes predicted to be more deleterious. Three patients successfully underwent HSCT, associated with complete remission of their IBD. Post-HSCT, neutrophil responses closely aligned with those of healthy controls. We therefore conclude that G6Pase-(and therefore cytoplasmic glucose availability in neutrophils) has an important role in mucosal immunity and increased neutrophil inflammatory output may drive enteritis in patients with G6PC3 deficiency. Future studies should seek to test whether G6PC3 deficiency-associated IBD can be ameliorated by rescuing neutrophils from death, by lowering 1,5-anhydroglucitol-6-phosphate levels with therapeutic inhibitors of the kidney glucose transporter SGLT2. 22 In the interim, we propose that clinicians check for IBD in patients with G6PC3 deficiency and consider HSCT when IBD is difficult to control.