PHOX2B expression in bone marrow and peripheral blood is associated with metastasis and prognosis in patients with neuroblastoma

Background Paired-like homeobox 2B (PHOX2B) is specifically expressed in the nervous system including neuroblastoma cells, but little is known about the clinical significance of the expression of PHOX2B in bone marrow (BM) and peripheral blood (PB) samples of newly diagnosed neuroblastoma patients. Methods The expression of PHOX2B in 276 paired BM and PB samples of neuroblastoma patients at diagnosis was tested by quantitative reverse transcriptase polymerase chain reaction (RT-PCR). Then the relationship between PHOX2B level and clinical characteristics including metastasis and prognosis was explored by receiver operating characteristic (ROC) analysis and Kaplan-Meier method. Results We identified the combined expression of PHOX2B in both BM and PB provided a high diagnostic accuracy for metastasis of neuroblastoma patients (AUC = 0.920) with the sensitivity and specificity of 86.7% and 92.9%, respectively. At last, 246 patients were enrolled for prognostic analysis. The median follow-up time was 22 months. Positive expressions of PHOX2B in BM and PB at diagnosis were associated with worse EFS and OS in neuroblastoma patients (P < 0.05). What’s worse, 19.7% (31/157) and 6.4% (10/157) patients with positive expression of PHOX2B in BM and PB samples in low/intermediate-risk group also had shorter EFS and poor OS (P < 0.05). CONCLUSIONS The expressions of PHOX2B in BM and PB were high in patients with unfavorable clinical characteristics. PHOX2B could be an appropriate biomarker for predicting metastasis and prognosis in patients with neuroblastoma. BM PHOX2B in 98.7% BM cytology-positive and 98.9% BM immunocytology-positive patients, in 23.2% BM cytology-negative and 17.4% BM immunocytology-negative patients on the high sensitivity method of RT-PCR. The positive of PHOX2B in BM and PB were also higher in patients with unfavorable clinically relevant factors in the present

PHOX2B expression in bone marrow and peripheral blood is associated with metastasis and prognosis in patients with neuroblastoma Beijing Children's Hospital Background Neuroblastoma, originating from the peripheral sympathetic nervous system, accounts for 12% of deaths associated with cancer in children [1]. Although patients with low-and intermediate-risk neuroblastoma have favorable outcomes, the prognosis of high-risk patients is still dismal with 5-year overall survival (OS) less than 50% [2]. Nearly one-half of neuroblastoma patients are classified into the high-risk group with metastasis at the initial diagnosis [3]. Positron emission tomographycomputed tomography (PET-CT), computed tomography (CT), magnetic resonance imaging (MRI) and iodine-123 metaiodobenzylguanidine ( 123 I-MIBG) scanning are widely used in detecting neuroblastoma patients with distant metastasis. However, repeated CT or 123 I-MIBG scans during the long-term treatment expose children to radiation and also unable to detect small primary lesions and metastases accurately. Therefore, new methods should be tested for predicting metastasis in neuroblastoma patients.
Liquid biopsy, including circulating tumor cells (CTCs), cell-free tumor DNAs (ctDNAs), as well as RNAs, proteins, and exosomes, has been introduced as a new method for the diagnosis of cancer metastasis in the past decades [4,5]. Since the detection of neuroblastoma cells in the blood of patients with disseminated disease by Moss et al in the 1990s [6], some studies have been devoted to the development and improvement of the methods for liquid biopsy in neuroblastoma patients.
Paired-like homeobox 2B (PHOX2B) gene, located on chromosome 4p13 [7], is specifically expressed in the nervous system including neuroblastoma cells, but not in normal bone marrow or peripheral blood cells of patients [8,9]. Previously, some studies showed PHOX2B could be used as a biomarker for minimal residual disease (MRD) and predicting prognosis in neuroblastoma patients [9][10][11][12]. In our study, we evaluated the diagnostic significance of PHOX2B expression in bone marrow (BM) and peripheral blood (PB) for disseminated diseases and patients' outcomes.

Patients and samples
Paired BM and PB samples were collected from 276 newly diagnosed neuroblastoma patients, who were enrolled in our center from June 1st, 2017 to December 31th, 2018. The study was approved by the Ethical Committee of our Hospital. Informed written consent was obtained from each participant according to the Declaration of Helsinki. The International Neuroblastoma Risk Group Staging System (INRGSS) was used for stratifying patients with different clinical stages at the time of diagnosis before any treatment [13]. We defined cases as patients with metastasis by the evaluation of PET-CT, bone scans, CT, MRI, bone marrow aspirates and trephine biopsies at diagnosis [1]. Each patient had two sites bone marrow aspirates of the sternum and ilium at diagnosis. The patients were treated in different risk groups according to prognostic category [2,14], and the chemotherapy regimen was guided by BCH-2007-NB [15,16].

Real-time Polymerase Chain Reaction
The Trizol Reagent was used to extract total RNA from 2 mL BM and PB samples, and then the RNA was reverse transcribed into cDNA [17]. RT-PCR was performed in a ViiA™ 7 Dx System (Applied

Statistical analysis
The primary analytic endpoint was event-free survival (EFS) and overall survival (OS). Time to event was defined as the time from diagnosis until the time of the first occurrence of relapse, progression, death, or until time of the last contact if none of these occurred. Kaplan-Meier method was performed on survival analysis and the log-rank test was used to compare the survival curves between PHOX2B positive and negative subgroups. Hazard ratio (HR) and 95% confidence interval (95% CI) were estimated from Cox proportional hazard models.
Statistical analyses were performed using SPSS19.0 and Graphpad Prism 5.0 software. The significance of the differences in clinical features between PHOX2B positive and negative patients was determined by the χ 2 test. The efficacy of PHOX2B expression to diagnose metastasis was assessed by receiver operating characteristic (ROC) analysis. The P-value lower than 0.05 was considered as statistically significant.

PHOX2B mRNA expression in BM and PB of patients and Patients' characteristics
A total of 276 patients, aged 1 month to 138 months (median, 35 months) with newly diagnosed neuroblastoma, were enrolled in this study. The clinical and biological characteristics were shown in  To evaluate the ability of the PHOX2B mRNA level for predicting metastasis of neuroblastoma patients at diagnosis, ROC curves were utilized. As shown in Fig. 1A  expression in patients with different risk subgroups. In the high-risk group, the patients with PHOX2B expression either in BM or PB had poorer EFS than those negative ones (P < 0.05, Fig. 2C and 2D).
Although there was no statistical difference (P > 0.05, Fig. 2C and 2D), the positive BM and PB patients had a poor OS trend. In the Low-/Intermediate-Risk group, 29.0% (9/31) patients with PHOX2B expression in BM had an event occurrence, while only 1.6% (2/126) occurred in patients without PHOX2B expression (P < 0.05). Likely, 30.0% (3/10) patients with PHOX2B expression in PB had event occurrence, while only 5.6% (8/147) occurred in patients without PHOX2B expression (P < 0.05). Unexpectedly, in Low-/Intermediate-risk group with favorable outcomes, the patients with positive expression of PHOX2B in BM and PB at diagnosis showed worse EFS and OS than the negative patients (P < 0.05, Fig. 2E -2F, Fig. 3E -3F). The univariate survival analysis showed that age ≥ 18 months, high-risk group, MYCN amplification, chromosome 1p loss, BM and PB positive of PHOX2B, LDH ≥ 587 U/L, SF ≥ 92 ng/mL, and NSE ≥ 100 ng/mL were associated with OS in patients with neuroblastoma (P < 0.05). Further analysis using multivariate COX proportional hazards regression model demonstrated that MYCN amplification and the positive expression of PHOX2B in BM were the independent risk factors for the prognosis of patients with neuroblastoma (P < 0.05, Table 3).

Discussion
The clinical importance of PHOX2B gene expression examination for predicting metastasis and prognostic evaluation in neuroblastoma patients were evaluated in our study. PHOX2B was first been shown as a highly sensitive and specific biomarker for neuroblastoma in 2008 [8,18], and it was confirmed not expressed in BM and PB samples of children in the control group without neuroblastoma [8]. But in neuroblastoma patient samples, PHOX2B was detected in 32% cytologynegative and 14% immunocytology-negative BM samples, and in more than 90% of cytology-positive and immunocytology-positive BM samples [8]. Our study revealed PHOX2B was detected in 98.7% BM cytology-positive and 98.9% BM immunocytology-positive patients, and also in 23.2% BM cytologynegative and 17.4% BM immunocytology-negative patients based on the high sensitivity method of RT-PCR. The positive rates of PHOX2B in BM and PB were also higher in patients with unfavorable clinically relevant factors in the present study.
Today, besides CT and MRI, 123 I-MIBG and PET-CT have been widely used as more sensitive and specific imaging of neuroblastoma especially in soft tissues and bone sites [1,19]. However, about 10% of neuroblastomas do not take up radiolabelled MIBG [20]. Even the combined use of 123 I-MIBG and PET-CT can not detect all tumor lesions in neuroblastoma patients [19,21,22]. So, liquid biopsy like CTCs, ctDNA, ctRNA, and even exosomes which well used in diagnosing metastasis in solid tumors need to be further studied in neuroblastoma when imaging data are ambiguous [23][24][25][26].
Neuroblastoma patients with ≥ 3 CTCs per 4 ml of PB were showed with an increased likelihood of having metastasis with the sensitivity and specificity of 88.89% and 78.59%, respectively [27]. The univariate logistic regression results also showed that CTC and NSE levels were associated with metastasis [27]. In another study, the combined 9 circulating miRNAs in newly diagnosed neuroblastoma patients' serum strongly associated with metastatic stage 4 disease, disease burden, and treatment response. Our data found that combined the expression of PHOX2B in both BM and PB for predicting metastasis of neuroblastoma at diagnosis with the sensitivity and specificity high to 86.7% and 92.9%, respectively. That means PHOX2B is a very useful biomarker for predicting metastasis of neuroblastoma patients.
PHOX2B was also revealed to be related to the survival of neuroblastoma patients. The stage 4 neuroblastoma patients with negative expression of PHOX2B in BM at 3 months after diagnosis and after completion of induction chemotherapy had a favorable outcome [28]. However, the expression of PHOX2B in BM at diagnosis did not correlate to survival in patients with stage 4 neuroblastoma [28]. In another research, 19% of high-risk neuroblastoma patients at diagnosis with both high expression of PHOX2B and tyrosine hydroxylase (TH) in PB were identified as ultrahigh-risk, with 5- year EFS and OS rates of unacceptable 0% [29]. In our study, the patients with negative PHOX2B in both BM and PB at diagnosis had a more favorable prognosis than the positive patients in the highrisk group. Our data were consistent with the above studies.  Figure 1 The