Low IL‐23 levels in peripheral blood and bone marrow at diagnosis of acute leukemia in children increased with the elimination of leukemic burden

Abstract IL‐23 is an IL‐12 cytokine family member with pleiotropic functions that regulates tumour growth in various cancer types, exhibiting both anti‐tumorigenic and pro‐tumorigenic properties. Preclinical studies have shown a potential anti‐leukemic action on childhood B‐ALL cells. The study involved 65 children with acute leukemia [59 patients with acute lymphoblastic leukemia (ALL) and 6 patients with acute myeloid leukemia (AML)] and 27 healthy controls. Using an enzyme‐linked immunosorbent assay, we aimed to determine the IL‐23 levels in the peripheral blood (PB) and bone marrow (BM) of patients at diagnosis and at the end of the induction therapy (EIT). PB IL‐23 levels were lower in leukemia patients compared to the healthy controls. In all acute leukemia patients, IL‐23 levels were significantly lower at diagnosis both in PB (P = .015) and in BM (P = .037) compared to the PB and BM concentrations at the EIT. The same pattern was present in both subgroups of ALL and AML patients. The high leukemic burden at diagnosis was related with lower IL‐23 levels, which were increased with the disease remission. Considering the anti‐leukemic potential of this cytokine, the elevation of the IL‐23 concentration at the disease remission indicates a beneficial role of IL‐23 in paediatric acute leukemia.

immunomodulators. Among them, cytokines are of primary importance, regulating the functions of B and T lymphocytes as well as natural killer cells, and inducing lymphopoiesis. [7][8][9] It has been proposed that aberrant cytokine production supplements genetic abnormalities during leukemogenesis, providing leukemic cells with a survival and proliferation advantage, attributed to dysregulated cell cycle and impairment of the normal immune surveillance. 10 Both cytokine over-expression by leukemic and immune cells, as well as the inadequate cytokine expression, can promote leukemogenesis. [10][11][12] The aberrant cytokine secretion in leukemia patients is normalized in complete remission, suggesting an autonomous pattern of cytokine expression that is dependent on leukemic activity. 13 Moreover, the balance between pro-inflammatory and antiinflammatory cytokines plays an essential role on how the inflammatory cells are involved in tumour growth. 14,15 Malignant cells are able to evade the immune system surveillance by using a variety of methods, including production and secretion of soluble factors that suppress specific antitumour immune responses. There is evidence that the inhibition of antitumour immune responses, and thus, the promotion of tumour growth, is related to the shift from Th1 lymphocyte responses to Th2 responses in the tumour microenvironment. 16 Th1 cells produce cytokines, such as IL-12, IFNγ and TNFα, that are pro-inflammatory and promote cellmediated immunity to act versus tumour cells, while the Th2 cytokines exert anti-inflammatory and immunosuppressive properties. The Th1 and Th2 cytokines act antagonistically and the shift in favour of Th1 facilitates the anti-tumour immune responses. 14 In solid tumours, as well as in acute and chronic leukemias microenvironment, there is a shift towards the Th2 cytokine secretion by immune and tumour cells. 17,18 Indeed, in several cancer types, the IL-12 levels, that reflect the Th1 immune response, were significantly lower in patients with advanced disease in comparison with patients with moderate disease. 14,19 The IL-12 family members are mainly involved in the promotion and maintenance of Th1 cell differentiation and high IL-12 levels have been associated with less advanced malignancies and better prognosis. 19 Moreover, in some cancer types, the response to therapy was followed by an elevation of the initially low serum IL-12 levels. 14,20 IL-23, an IL-12 cytokine family member, exhibits structural similarities with IL-12 but substantially distinct biological role. 21,22 IL-23 is a heterodimeric cytokine that is comprised of the IL-12p40 subunit, common with IL-12, and the IL-23-specific p19 subunit. The IL-23 receptor is composed of the IL-12Rβ1 chain, that is shared with the IL-12 receptor, and the exclusive IL-23R chain. 21,23 IL-23 is produced by myeloid dendritic cells and by type 1 macrophages in response to microbial and host immune stimuli, such as Toll-like receptors and IFNs. IL-23 activates the same Jak/STAT signalling molecules as IL-12, namely Jak2, Tyk2 kinases and STAT1, −3, −4 and −5 transcription factors. In contrast to IL-12 that predominantly induces the activation of STAT4, the main transcription factor induced by IL-23 is STAT3. 24 IL-23 stimulates the proliferation of Th1 memory cells and promotes the proliferation of Th17 cells, a subset of T cells that expresses IL-17, which has a crucial role in the development of autoimmune inflammation. 23,25 Also, IL-23 modulates IgM and IgG secretion in plasma cells by upregulating the IgM secretion, while inhibiting the IgG secretion, suggesting that IL-23 is mainly involved in primary immune responses. 26 The potent anti-tumour activity of the IL-12 family cytokines has been put under investigation, and while the anti-tumour activity of IL-12 and IL-27 has been clearly established, the role of IL-23 appears to be controversial. 22 In this context, we hypothesized that IL-23 expression profile may be different in states of leukemic burden, and contribute to disease pathogenesis. Therefore, the aim of this study was to determine the levels of IL-23 in the peripheral blood (PB) and the BM of children with acute leukemia, reveal any potential correlations with clinicopathological features and detect any significant differences between diagnosis and the end of the induction therapy (EIT).

| Patients
Our study involved 65 children with acute leukemia, diagnosed and treated in the University Pediatric Hematology-Oncology Unit of the "Agia Sophia" Children's Hospital and the Pediatric Hematology-Oncology Unit of the "P&A Kyriakou" Children's Hospital in Athens, Greece, from January 2011 to May 2016, and 27 children admitted to the hospital for minor surgical procedures, serving as controls.
The patient cohort was comprised of 65 children (34 males and 31 females) aged from 11 months to 18 years old, with median age 4.9 years old (2.9-9.8) ( Table 1)

| PB and BM sample collection and IL-23 determination
Peripheral blood samples and BM aspirates were collected from children with acute leukemia before the initiation of treatment (day 0) and at the EIT (Day 33 for patients with ALL and Day 28 for patients with AML, respectively). In cases of Burkitt-cell acute leukemia, samples were collected at diagnosis and after the completion of COPADM1 scheme of induction therapy.
Bone marrow was collected into EDTA vacutainer tubes and PB in serum-separating tubes. PB serum and BM plasma were obtained by centrifugation, divided into aliquots, and stored at −80℃ until assayed. Repeated freeze-thaw cycles were avoided for all samples. The analytical sensitivity was 5.0 pg/mL and the assay range is 31.3-2000 pg/m. The inter-assay coefficient of variation was 4.5% and intra-assay, 6.7%. The absorbance value at 450-nm wavelength was read with a microplate reader (VersaMax Tunable, Molecular Devices).
Each measurement represents the average of triplicate wells.

| Statistical analysis
Continuous variables normally distributed are presented as means (standard deviation (SD)), whereas non-normally distributed variables are presented as median (25th, 75th percentile). The Kolmogorov-Smirnov goodness-of-fit test was used to assess the distribution of each variable. Comparisons between continuous variables were performed with Student's t test and paired t test or Mann-Whitney and Wilcoxon test depending on normality of distribution.
Statistical analyses were performed with the STATA software (version 13.0, College Station™). All P-values were two-sided and a cut-off value of P < .05 was set to denote statistical significance. Evaluation of demographic and clinical data of ALL and AML patients in correlation with pretreatment PB and BM IL-23 levels revealed no statistical significance between subgroups for age, gender, as well as for white blood cell count (WBC), lactate dehydrogenase (LDH) and the minimal residual disease (MRD) result at the EIT, as shown in Table 2.

| Lower IL-23 levels in PB of ALL patients at diagnosis compared to healthy controls
In all leukemia patients, PB IL-23 levels at diagnosis were lower compared to healthy controls [11.38 (6.14-18.61) vs 18

| Lower IL-23 levels in PB and BM at diagnosis compared to the EIT in ALL and AML patients
In all leukemia patients, PB IL-23 levels at diagnosis were signifi- The comparison between the IL-23 concentration at diagnosis and at the EIT in PB and BM is shown analytically in Table 3. that support haematopoiesis and modulate immune response via cytokine secretion. 50,51 In various hematopoietic disorders, BMMSCs presented altered cytokine expression and an impaired immunoregulatory function. 52 AML-derived BMMSCs were more immunosuppressive/anti-inflammatory and reduced the secretion of pro-inflammatory cytokines, contributing to disease evolution in AML patients. 53 In childhood ALL, the MSC niches in BM were

| D ISCUSS I ON
shown to contribute to asparaginase resistance. 54 It is possible that the leukemic cells themselves reprogramme BMMSCs to provide a niche that protects their growth and clonal evolution. 55 The role of  Note: IL-23 concentration is presented as median value (25th, 75th percentile) in cases of non-normal distribution and as mean value (SD) in cases of normal distribution. Bold values indicate significance.
Abbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; B-ALL, B-cell acute lymphoblastic leukemia; BM, bone marrow; EIT, end of the induction therapy; n, number of patients; PB, peripheral blood; T-ALL, T-cell acute lymphoblastic leukemia.
as hepatitis while anti-PD-1 treatment significantly increases IL-23 levels. 27,59 Understanding the way that IL-23 contributes to the control of different types of cancer, including acute leukemias of childhood, may lead to re-evaluation of the potential therapeutic use of this molecule.
IL-12, which is the prototype cytokine of this family, is considered as a strong candidate for immunotherapy-based interventions. In both preclinical and clinical trials, IL-12 administration has exerted significant anti-tumour and anti-metastatic activity. 35 A trial performed in low-grade non-Hodgkin lymphoma patients demonstrated that combined administration of IL-12 with rituximab resulted in longer response compared to stand-alone. 22 However, the therapeutic effect of IL-12 in clinical trials has been limited by systemic toxicity. 31 The use of IL-23 in cancer patients has not been yet investigated.
However, in preclinical trials, the IL-23 administration has showed direct anti-tumour activity in several cancer types, including pediatric B-ALL, exerting lower toxicity compared to IL-12 treatment, possibly because of the lower induction of IFNγ. 22 Collectively, these data suggest that IL-23 may be a good candidate treatment to be tested in phase 1 trial in pediatric acute leukemia patients.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data have been reported.