The clinical and prognostic significance of FOXN3 downregulation in acute myeloid leukaemia

Abstract Introduction The expression of forkhead box N3 (FOXN3), also known as checkpoint suppressor 1 (CHES1), is reduced in many types of tumours. However, the clinical significance of FOXN3 and its potential role in acute myeloid leukaemia (AML) remain largely unknown. Methods A total of 117 de novo AML patients newly diagnosed between December 2015 and January 2018 were included in this study. The expression of FOXN3 and its clinical significance were analysed in these AML patients. Results The expression of FOXN3 was significantly downregulated in AML. In addition, lower FOXN3 expression was associated with older age and higher white blood cell counts. Moreover, a close correlation was observed between lower FOXN3 expression and a lower complete remission (CR) rate and shorter overall survival (OS), which was further analysed by multivariate analysis. Conclusion These data suggest that FOXN3 is a novel biomarker in AML and that lower FOXN3 expression predicts poor chemotherapy response and prognosis in AML.


| Patients
The present study enrolled 117 newly diagnosed AML patients between

| Therapy and follow-up
The follow-up information of 96 AML patients was available. The treatment followed protocol described as our previous study and response assessment was based on Chinese expert consensus on the treatment of AML (2011). 20

| RNA isolation and RT-qPCR
Ficoll-Paque™PLUS (GE Healthcare) was used to extract mononuclear cells from BM. Total RNA was isolated utilizing TRIzol reagent (Invitrogen), and cDNA was prepared from 1 µg of RNA using the PrimeScript™RT Reagent Kit with gDNA Eraser (TaKaRa). For the detection of FOXN3 expression levels in the BM of patients and normal controls, real-time quantitative PCR (RT-qPCR) was conducted using a TaqMan Gene Expression Assay on an ABI 7500 Real-Time PCR system (Applied Biosystems) as previously described, 20 and ABL was used as a control gene. The primers and TaqMan-based probes were as follows: FOXN3 forward 5′-TGCCAATCACTCCCATTGGG-3′, reverse   5′-CCGCATCCGGCAGCTGG-3′ and probe Fam-TGCCATTCCTCAT   GGCCGCTGTCA-Tam; and ABL forward 5′-TGGAGATAACACTCTAAGC   ATAACTAAAGGT-3′, reverse 5′-GATGTAGTTGCTTGGGACCCA-3′ and probe Fam-CCATTTTTGGTTTGGGCTTCACACCATT-Tam. For the detection of PIM2 and E2F5 expression levels in the BM of patients, RT-qPCR was conducted using SYBR Green technology as previously described. 21 The primers used are shown in Table S1.

| Immunocytochemistry
The cytospin smears of BM cells from AML and normal control samples were fixed in paraformaldehyde (4%, 5 minutes). The specimen was then incubated with a peroxidase-blocking enzyme and normal goat serum (10 minutes each), followed by incubation with rabbit antihuman FOXN3 protein antibodies (HPA059209, SIGMA) at 37°C for 30 minutes. Biotin-labelled goat antirabbit IgG was used as a secondary antibody, and the protein was detected using the streptavidin-peroxidase (SP) complex developed with DAB according to the manufacturer's instructions. Subsequently, the specimens were counter-stained with haematoxylin. Finally, the reactivity of the antibody was made visible with the Vector brown SP substrate.

| Gene expression data set
Forkhead box N3 expression was compared between haematopoietic stem cells (HSCs) and AML from the Bloodpool data set (probe number: 222494) using the online BloodSpot database (www.blood spot.eu). 24

| Statistical analysis
Statistical analysis was performed by GraphPad Prism 7.0a software and SPSS 15.1 software. Differences between groups were compared using the Mann-Whitney test or one-way analysis of variance (ANOVA) among multiple groups. Pearson chi-square analysis/Fisher's exact test was conducted to compare the differences of categorical variables.
Survival analysis was used to analyse the impact of FOXN3 on relapsefree survival (RFS) and overall survival (OS), and the differences were compared by a log-rank test. Univariate and multivariate analyses were performed using the Cox promotional hazards regression model.

| FOXN3 expression was abnormally downregulated in AML
Forkhead box N3 mRNA expression was detected in a total of 117 AML patients and 25 controls by RT-qPCR. As shown in Figure 1, the F I G U R E 1 FOXN3 expression is significantly downregulated in AML. A, Expression levels of FOXN3 mRNA were detected in AML patients and controls by RT-qPCR. ****P < .0001. B and C, Representative images showing decreased expression levels of FOXN3 protein in control (B) and AML patients (C) using immunocytochemical staining [Colour figure can be viewed at wileyonlinelibrary.com]  Figure 1A).
Moreover, the lower expression of FOXN3 protein was confirmed in 15 AML patients with decreased levels of FOXN3 mRNA by SP immunocytochemical staining ( Figure 1B,C). This downregulated expression of FOXN3 in AML was validated by analysing the online Bloodpool data set (www.blood spot.eu), revealing that FOXN3 expression was significantly lower in AML than in CD34+ HSCs.

| Lower FOXN3 expression correlated with older age and higher WBC
The 117 AML patients were divided into two groups according to whether their FOXN3 expression levels were below (lower expression group) or above (higher expression group) the median level of FOXN3 expression. The comparison of clinical features between the two groups showed that a lower expression of FOXN3 was correlated with older age and higher white blood cell counts (P = .033 and .032, respectively, Table 1). There were no significant differences observed in haemoglobin, platelet counts, BM blasts, FAB subtypes, cytogenetic subgroups and prognostic risk stratification 25 (Table 1).

| Lower FOXN3 expression is associated with poor chemotherapy response and shorter survival in AML
Follow-up data were collected from 96 AML patients, and the clinical information is summarized in Table S2. A total of 69 (72%) patients achieved CR after induction chemotherapy and 27 (28%) patients experienced induction chemotherapy failure. AML patients with lower FOXN3 expression showed a significantly lower CR rate than those with higher FOXN3 expression group (P = .012, Figure 3A). (P = .028, Figure 3B). Among these 34 patients, 16 were allocated to the lower FOXN3 expression group and the other 18 to the higher FOXN3 expression group. In the lower FOXN3 expression group, all 16 patients showed significantly higher FOXN3 mRNA expression at CR than when newly diagnosed (P = .0007, Figure   S1A), whereas no significant difference was observed in the higher FOXN3 expression group ( Figure S1B).
Survival analysis was performed to compare RFS and OS between the FOXN3 lower expression and higher expression groups.
The results indicated that patients with lower FOXN3 expression presented a significantly shorter OS time than those with higher FOXN3 expression ( Figure 3C). Although it was lack of significant difference in RFS between the two groups ( Figure 3D), the RFS of the higher FOXN3 expression group was significantly longer than that of the lower FOXN3 expression group when older patients were excluded ( Figure 3E). Univariate and multivariate analyses were further performed to reveal the prognostic significance of FOXN3 expression in AML according to ELN recommendations and previous studies 6,26,27 (Table 2). The multivariate analysis showed that the expression of FOXN3 was an independent prognostic factor correlated with OS (HR = 0.269, P = .003).

| Association of FOXN3 expression with PIM2 and E2F5 expression
To investigate the target genes of FOXN3 as a transcriptional suppressor in AML, the mRNA expression of PIM2 and E2F5 in the BM of 32 patients was detected by RT-qPCR. Unfortunately, there was no negative correlation between FOXN3 and PIM2 or E2F5 ( Figure   S2).

| D ISCUSS I ON
Forkhead box N3 belongs to the FOXN gene family and was first discovered as a suppressor of DNA damage-activated checkpoint mutations in yeast. 28 In recent years, studies on FOXN3 have suggested that it may play dual roles in different tumours. Acting as a tumour suppressor gene, FOXN3 is reduced in several types of tumours, such as HCC, colon cancer and osteosarcoma, 9,10,19 but it is upregulated in ovarian cancer and breast cancer, where it acts as an oncogene. 11,29 One explanation for the diverse effects of FOXN3 may be due to the specific cellular and tissue environment. 30,31 In our previous studies, the forced expression of FOXN3 inhibited cell proliferation, the induction of apoptosis, and cell cycle arrest. 21 These results indicated that FOXN3 may participate in the malignant transformation of leukaemia cells. In this study, the significant downregulation of FOXN3 was validated, and FOXN3 was also indicated to be an independent prognostic marker of AML. These findings affirmed the tumour suppressive role of FOXN3 in AML.
As a tumour suppressor, the impact of FOXN3 on prognosis has been demonstrated in solid tumours. Patients with high FOXN3 expression had longer OS and RFS times than those with low FOXN3 expression in HCC, osteosarcoma and breast cancer patients. 9,12,19 The results of this study suggested that downregulation of FOXN3 correlates with poor OS and RFS of non-APL.
The level of FOXN3 is not only an independent prognostic factor but also serves as a biomarker of treatment response in non-APL.
However, it seems that the prognostic effect of FOXN3 expression In this study, the abnormal expression of FOXN3 showed a tendency but not significant difference on mutations of NPM1, DNMT3A, TP53 and ASXL1, suggesting the potential correlation of FOXN3 with molecular aberrations. By searching the publicly available data containing a larger cohort of patients, 33 the correlation of lower FOXN3 expression with higher incidence of NPM1 mutation could be confirmed, whereas there was no significant distinction on mutations of DNMT3A, TP53 and ASXL1 between AML with different FOXN3 levels ( Figure S4).
As a transcription regulator, FOXN3 has been reported to inhibit the expression of some tumour oncogenes, such as PIM2 and E2F5. 9

ACK N OWLED G EM ENTS
This work was supported by the National Natural Science Foundation of China (81600117).

CO N FLI C T O F I NTE R E S T
The authors have no competing interest.