Neuroblastoma incidence, biology and outcome : With special emphasis on quantitative analysis of Tyrosine Hydroxylase mRNA in blood and bone marrow

Abstract: Neuroblastoma, the most common malignant disease of early childhood, accounts for 6% of childhood malignancies but is responsible for 9% of paediatric cancer related deaths. The tumour originates from cells in the neural crest and has a very heterogeneous clinical course. A minority of tumours may differentiate or regress spontaneously whereas the metastatic form is difficult to cure which is related to difficulties in treating minimal residual disease (MRD). This study aimed to characterise neuroblastoma in Sweden during 27 years as well as developing methods for detecting MRD and evaluate the clinical significance of detecting MRD in blood and bone marrow in children with neuroblastoma. By now 359 children with neuroblastoma below the age of 15 years have been included in the Swedish Childhood Cancer registry. The yearly age standardised incidence in Sweden was 1/100 000 children <15 years with no trend towards an increase or decrease. The median age at diagnosis was 18 months. The gender ratio was significant different with 1.16:1 male to female. The outcome for boys was slightly worse than for girls mainly due to more frequent metastatic disease, 1.41:1. Survival for all children with neuroblastoma improved from 57.8% (overall survival at 5 years for children diagnosed 1982-1990) to 74.1% (2000-2008). Survival for high-risk neuroblastoma increased significantly (p<0.001) from 8.6 % (1982-1990, n=35) to 17.0 % (1991-1999, n=53) and 54.5 % (2000-2008, n=50). In order to detect MRD in blood (PB) and bone marrow (BM) from children with neuroblastoma, we developed a quantitative real time RT-PCR (qRT-PCR) method with tyrosine hydroxylase (TH) mRNA as a target. PB and BM samples from 24 children with neuroblastoma of all different biological subsets and stages were analysed. Tyrosine hydroxylase mRNA analysis could significantly distinguish localised disease from metastatic disease and detect relapse before conventional methods. We used this sensitive method for detection of TH in pigment cells to understand whether this enzyme can take an active part in pigment formation or not. Our results indicated that there is no TH that can contribute to pigment formation in melanocytes or melanoma cells. In our fourth study we used qRT-PCR to compare the specificity, sensitivity and clinical usefulness of TH, dopa decarboxylase (DDC) and GD2 synthase (GD2S) mRNA for detection of minimal disease in PB and BM in neuroblastoma patients. In total 554 samples from 58 patients were analysed. TH and DDC mRNA could discriminate localised disease from metastatic disease both in PB and BM whereas GD2S could discriminate between localised and metastatic disease in PB but not in BM. Tyrosine hydroxylase and DDC higher than median at diagnosis predicted a worse outcome for 24 well-characterised uniformly treated high-risk neuroblastoma patients followed for a median of 63 months from diagnosis (survival probability 91% at 5 years for TH below median versus 33% for TH above median, p=0.009) These results suggest that TH and DDC transcript concentrations at diagnosis is a possiblenew way to stratify patients for different novel treatment strategies. In conclusion, high expression of TH and DDC both in PB and BM corresponds to metastatic neuroblastoma atdiagnosis, residual disease, and poor outcome. The concentration of TH and DDC can distinguish at diagnosis children who can be cured with current intensive therapy from those in need for novel therapeutic strategies as one way forward to further improved outcome for children with neuroblastoma.