This study sought to investigate biological/clinicopathological characteristics of neuroblastoma, undifferentiated subtype (NBUD).
This study sought to investigate biological/clinicopathological characteristics of neuroblastoma, undifferentiated subtype (NBUD).
This study examined 157 NBUD cases filed at the Children's Oncology Group Neuroblastoma Pathology Reference Laboratory, and survival rates of the patients were analyzed with known prognostic factors. Immunostainings for MYCN and MYC protein were performed on 68 tumors.
NBUD cases had a poor prognosis (48.4% ± 5.0% 3-year event-free survival [EFS]; 56.5% ± 5.0% overall survival [OS]), and were often associated with high mitosis-karyorrhexis index (MKI, 65%), prominent nucleoli (PN, 83%), ≥ 18 months of age (75%), MYCN amplification (MYCN-A, 83%), diploid pattern (63%), and 1pLOH (loss of heterozygosity (72%). However, these prognostic indicators, except for MYCN status, had no significant impact on survival. Surprisingly, EFS for patients with MYCN-A tumors (53.4% ± 5.6%) was significantly better (P = .0248) than for patients with MYCN-nonamplified (MYCN-NA) tumors (31.7% ± 11.7%), with MYCN-NA and PN (+) tumors having the worst prognosis (9.3% ± 8.8%, P = .0045). Immunohistochemically, MYCN expression was found in 42 of 48 MYCN-A tumors. In contrast, MYC expression was almost exclusively found in the MYCN-NA tumors (9 of 20) especially when they had PN (8 of 11). Those patients with only MYC-positive tumors had the worst EFS (N = 8, 12.5% ± 11.7%) compared with only MYCN-positive (N = 39, 49.9% ± 17.7%) and both negative tumors (N = 15, 70.0% ± 17.1%) (P = .0029). High MKI was often found in only MYCN-positive (30 of 38) but rarely in only MYC-positive (2 of 8) tumors.
NBUD represents a unique subtype of neuroblastoma associated with a poor prognosis. In this subtype, MYC protein expression may be a new prognostic factor indicating more aggressive clinical behavior than MYCN amplification and subsequent MYCN protein expression. Cancer 2013;119:3718–3726. © 2013 American Cancer Society.
Peripheral neuroblastic tumors (pNTs) derive from the developing sympathetic nervous system, and are the most common extracranial tumors of childhood, with approximately 700 new patients each year in the United States, accounting for 15% of childhood cancer mortality.[1-7] According to the International Neuroblastoma Pathology Classification (INPC), pNTs are classified into 4 categories: neuroblastoma (Schwannian stroma-poor), ganglioneuroblastoma, intermixed (Schwannian stroma-rich), ganglioneuroma (Schwannian stroma-dominant), and ganglioneuroblastoma, nodular (composite, Schwannian stroma-dominant/stroma-rich and stroma-poor). Among the neuroblastoma category, 3 subtypes are recognized: undifferentiated, poorly differentiated, and differentiating.[3, 4]
MYCN (V-Myc myelocytomatosis viral-related oncogene, neuroblastoma derived) amplification is a hallmark of aggressive clinical behavior for patients with pNTs. In contrast to conventional neuroblastoma with a typical salt-and-pepper nucleus, MYCN-amplified tumors are often characterized by the presence of one or few prominent nucleoli.[8-10] Recently, in our report describing genotype-phenotype discordant neuroblastomas, we suggested that those prominent nucleoli could be the putative site of RNA synthesis/accumulation critical for protein expression.
In this study, we summarize clinicopathological characteristics of neuroblastoma, undifferentiated subtype (NBUD) with a special reference to prominent nucleolar formation. This is a unique and histologically defined subtype, of which little is known or has been previously published. In this report, we also present and discuss, for the first time, clinical and biological significance of MYC (C-myc) protein expression in human neuroblastoma cases.
A total of 3501 newly diagnosed cases of pNTs were enrolled in the COG (Children's Oncology Group) Neuroblastoma Studies between January 1, 2001, and August 31, 2010. Among these cases, 159 (4.5%) were identified as NBUD by the criteria of INPC[3, 4] at the COG Neuroblastoma Pathology Reference Laboratory (COG-NPRL), Department of Pathology and Laboratory Medicine, Children's Hospital Los Angeles, Los Angeles, California. Because 2 patients were not included in the COG biology and/or survival databases, 157 NBUD cases were analyzed in this study. Most of these patients were in the high-risk group: they had intensive chemotherapy, surgery, radiation, and in many cases, myeloablative therapy.[5-7, 12] Informed consent approved by the institutional review board was obtained for all patients at the time of enrollment in the COG biological or therapeutic study.
Hematoxylin-and-eosin–stained slides (Fig. 1) from the 157 NBUD cases filed at the COG-NPRL were re-reviewed (by L.W., R.S., and H.S.). Those slides (average of 3; range, 1-24 per case) were either from biopsy or surgery specimens obtained prior to chemotherapy/irradiation. Mitosis-karyorrhexis index (MKI: low, < 100/5000 cells; intermediate, 100-200/5000 cells; high, ≥ 200/5000 cells) class was assigned to each case. All NBUD cases were classified into the Unfavorable Histology Group according to the INPC regardless of their MKI class.[3, 4]
The INPC defines NBUD as a tumor without identifiable neuropil (thin neuritic processes), and supplemental techniques are usually required to establish the diagnosis. In this series of cases, the diagnosis was confirmed by immunohistochemistry and molecular tests performed by the contributing institutions, by the Department of Pathology and Laboratory Medicine at Children's Hospital Los Angeles, and/or by the COG Neuroblastoma Biology Reference Laboratory, Nationwide Children's Hospital, Columbus, Ohio. Tumor cells were usually positive for NSE (neuron-specific enolase), PGP9.5, NB84, and synaptophysin, often positive for chromogranin, and sporadically positive for TH (tyrosine hydroxylase, a marker for cells of neural crest origin). They were negative for CD99, WT1, muscle markers (desmin, myogenin, MyoD1, and so forth), and lymphoma markers (LCA, CD3, CD20, TdT, and so forth). Fluorescence in situ hybridization (FISH) test for detecting EWS (Ewing sarcoma) gene rearrangement, when performed, was always negative. NBUD tumors were frequently noted to have MYCN amplification. In rare cases, especially when TH was negative immunohistochemically, MYCN status along with clinical correlation including younger age at diagnosis and primary site of the tumor (adrenal gland or paraspinal sympathetic ganglion, for example), supported the diagnosis of NBUD.
In this study, we distinguished 2 subgroups of NBUD based on the nuclear morphology of neuroblastic cells: “prominent nucleolar” NBUD, where 10% or more of the neuroblastic cells had a nucleus containing prominent nucleoli, and “conventional” NBUD, where all or a majority of neuroblastic cells had a salt-and-pepper type nucleus (none or < 10% of the neuroblastic cells with prominent nucleoli). The prominent nucleoli were identified according to our previous report[11, 13]: They were either single or few, well-defined, discrete, eosinophilic, and enlarged nucleoli, and found in medium- or large-sized, often vesicular or open nuclei. In typical cases, those prominent nucleoli were easily recognized using low-power magnification. In some cases, however, examination under high-power magnification was required, with or without using an oil immersion lens. The proportion of neuroblastic cells containing prominent nucleoli varied from tumor to tumor. We determined an arbitrary 10% cutoff for distinguishing between the 2 subgroups, based on our previous experience reported elsewhere.
MYCN and MYC (c-myc) protein expression were detected immunohistochemically using formalin-fixed, paraffin-embedded sections from 68 NBUD tumors with available unstained sections filed in the COG Neuroblastoma Pathology Reference Laboratory. Among those tumors, 63 (MYCN-A in 45, MYCN-NA in 18) were stained for both MYCN and MYC proteins, 3 (all MYCN-A) for MYCN protein only, and 2 (both MYCN-NA) for MYC protein only. Immunostaining was performed with Leica BOND-MAX (Leica Microsystems Inc., Bannockburn, Ill) using steamer antigen retrieval for 30 minutes in Bond Epitope Retrieval Solution 2 (no. AR9640; Leica Biosystems Newcastle, Ltd., Newcastle Upon Tyne, UK). The sections were incubated with either anti-MYCN polyclonal rabbit antibody (Proteintech Group, Inc., Chicago, Ill) at a dilution of 1:50, or anti-human MYC monoclonal rabbit antibody (no. 1472-1; Epitomics, Inc., Burlingame, Calif) at a dilution of 1:50 in Bond Primary Antibody Diluent (no. AR9352; Vision BioSystems Inc., Norwell, Mass). Staining was visualized using Bond Polymer Refine Detection (no. DS9800; Leica Microsystems). The slides stained for MYC protein were counterstained with hematoxylin. No counterstaining was performed for the slides after MYCN protein staining. The staining results were evaluated as either “positive” or “negative.” Appropriate positive and negative controls were stained along with those NBUD tumors.
We examined the association of survival with multiple parameters, including age at diagnosis (< 18 months versus ≥ 18 months), INSS stage (not stage 4 versus stage 4), MKI (low versus intermediate versus high), the presence/absence of prominent nucleoli (“prominent nucleolar” tumors versus “conventional” tumors), MYCN status (nonamplified versus amplified), combinations of MYCN status with the presence/absence of prominent nucleoli, ploidy (hyperdiploid versus diploid), 1p status (no loss versus loss of heterozygosity [LOH]), and 11q (no loss versus LOH). Also, differences in outcome between patients with 1) MYCN protein–expressing tumors, 2) MYC protein–expressing tumors, and 3) tumors that were negative for both MYCN and MYC proteins were investigated. For event-free survival (EFS), the time to event was defined as the time from diagnosis until the first event or until the time of last contact if no event occurred. Patients who were alive without event (event-free) were censored at the time last known alive. For overall survival (OS), death was the only event considered. Survival analyses were performed using the methods of Kaplan and Meier with standard errors per the methods of Peto et al. Survival curves were compared using a log-rank test. P values < 0.05 were considered statistically significant.
To determine the independent prognostic strength for survival of factors found to be predictive in univariate analyses in the presence of standard prognostic factors age, INSS stage, MYCN status, ploidy, MKI, 1p status, and 11q status, Cox proportional hazards models with the Efron method of handling tied event times were fit. Backward selection was used to determine the most parsimonious model, with the least statistically significant term dropping out at each step.
A 2-sided Fisher's exact test was used to explore the following associations: 1) MKI with the combinations of MYCN status and presence/absence of prominent nucleoli; 2) MYCN status with MYCN and MYC protein expression, overall and within each of the prominent nucleoli groups; and 3) MKI with MYCN and MYC protein expression.
Overall 3-year EFS and OS rates of patients with NBUD were 48.4% ± 5.0% and 56.5% ± 5.0%, respectively. As listed in Table 1, NBUD tumors were frequently associated with poor prognostic factors; such as ≥ 18 months of age at diagnosis (117 of 157, 75%), stage 4 disease according to the INSS (128 of 157, 82%), MYCN amplification (124 of 150, 83%), diploid pattern (89 of 142, 63%), and 1pLOH (34 of 47, 72%). Whereas only 2 of 43 (5%) tested tumors had 11qLOH. It was noted, however, that all factors, except for the MYCN status, failed to show any prognostic significance for patients with NBUD. In contrast, the 3-year EFS for patients with MYCN-amplified NBUD tumors (53.4% ± 5.6%) was significantly better than for patients with MYCN-nonamplified tumors (31.7% ± 11.7%), with the MYCN-nonamplified plus prominent nucleoli (+) tumors faring worst of all (9.3% ± 8.8%; Fig. 2).
|Cohort||n||3-y EFS ± Standard Error (%)||EFS P Valuea||3-year OS ± Standard error (%)||OS P Valuea|
|<18 months||40||48.4 ± 11.0||.6273||58.5 ± 11.4||.9221|
|≥18 months||117||48.7 ± 5.6||56.2 ± 5.5|
|1,2,3,4s||29||60.5 ± 13.4||.5269||66.1 ± 12.8||.2996|
|4||128||46.2 ± 5.3||54.4 ± 5.4|
|Low||17||50.8 ± 17.8||.7175||53.2 ± 18.2||.7308|
|Intermediate||38||41.3 ± 8.8||47.7 ± 9.2|
|High||100||50.0 ± 6.2||60.1 ± 6.2|
|No||25||50.3 ± 13.4||.6483||52.1 ± 12.8||.7767|
|Yes||123||47.0 ± 5.6||56.7 ± 5.6|
|Nonamplified||26||31.7 ± 11.7||.0248||40.3 ± 13.9||.0524|
|Amplified||124||53.4 ± 5.6||59.6 ± 5.5|
|Amp/Nucl +||105||51.9 ± 6.0||.0045||59.1 ± 5.9||.1718|
|Amp/Nucl −||13||54.2 ± 18.3||53.3 ± 18.2|
|Non-Amp/Nucl +||13||9.3 ± 8.8||36.7 ± 29.2|
|Non-Amp/Nucl −||10||56.3 ± 21.5||53.3 ± 21.0|
|Hyperdiploid||53||54.3 ± 9.5||.7789||55.7 ± 9.6||.8840|
|Diploid||89||47.4 ± 6.3||56.9 ± 6.2|
|Non-LOH||13||30.2 ± 25.2||.1087||37.3 ± 29.5||.1153|
|LOH||34||62.2 ± 13.5||70.9 ± 12.7|
|Non- LOH||41||46.0 ± 15.1||.6487||57.1 ± 15.3||.9538|
|LOH||2||100.0 ± 0.0||100.0 ± 0.0|
A backward-selected Cox model in terms of EFS indicated that patients with nonamplified MYCN and prominent nucleoli (+) tumors were at more than 3 times greater risk of event than each of the other groups of patients (P = .0084). The order of removal of terms was 1p status, 11q status, MKI, ploidy, age, and INSS stage.
NBUD tumors often had a high MKI (100 of 155, 65%) and prominent nucleoli (123 of 148, 83%). The majority of patients with high MKI also had amplified MYCN (87 of 90, 97%) and prominent nucleoli (79 of 90, 88%). A statistically significant association was found between MKI and the combination of MYCN status and prominent nucleoli (P < .0001).
Among the 68 tumors immunohistochemically studied, 41 tumors were positive for MYCN (39 tested for both MYCN/MYC proteins; 2 tested for MYCN only), 9 tumors were positive for MYC (8 tested for both MYCN/MYC proteins; 1 tested for MYC only), 1 tumor was positive for both proteins, and 17 tumors were negative (15 tested for both MYCN/MYC proteins, 1 tested for MYCN only, and 1 tested for MYC only) (Fig. 3). MYCN protein expression was exclusively found in MYCN-amplified NBUD tumors regardless of the presence/absence of prominent nucleoli (Fig. 4). In contrast, MYC protein expression was found in the MYCN-nonamplified tumors especially when they had prominent nucleoli (8 of 11, 73%). One tumor expressing both MYCN and MYC protein had amplified MYCN. Diagnosis of neuroblastoma was confirmed for all of those MYC protein expressing cases by detecting TH positive cells, often sporadic, immunohistochemically in their tumor tissues. The results with statistical analyses are also summarized in Table 2.
|MYCN Status||P Valuea|
|MYCN (prominent nucleoli +)||<.0001|
|MYC (prominent nucleoli +)||<.0001|
|MYCN (prominent nucleoli −)||.0023|
|MYC (prominent nucleoli −)||1.0000|
As listed in Table 3, there were no differences in survival rates between MYCN-protein-positive tumors (N = 42) and MYCN-protein-negative tumors (N = 24). In contrast, there were significant differences in survival rates between MYC-protein–positive tumors (N = 10) and MYC-protein-negative tumors (N = 55). A backward-selected Cox model in terms of EFS indicated that patients with MYC-protein–positive tumors were 3.427 times more likely to experience an event than patients with MYC-protein–negative tumors (P = .0022). The order of removal of factors was 11q status, MYCN status, 1p status, INSS stage, MKI, ploidy, and age. In the model for OS, none of the factors were statistically significant.
|Cohort||n||3-y EFS ± Standard Error (%)||EFS P Valuea||3-y OS ± Standard Error (%)||OS P Valuea|
|Negative||24||46.0 ± 13.8||.5143||56.1 ± 15.2||.6236|
|Positive||42||50.5 ± 15.9||61.7 ± 14.4|
|Negative||55||54.4 ± 12.2||.0012||62.3 ± 11.5||.0491|
|Positive||10||12.5 ± 11.7||46.7 ± 34.1|
|Only MYCN positive||39||49.9 ± 17.7||.0029||61.1 ± 15.6||.1755|
|Only MYC positive||8||12.5 ± 11.7||50.0 ± 35.4|
|Both negative||15||70.0 ± 17.1||66.0 ± 17.2|
EFS rates were significantly different based on the protein expression of the tumors (Fig. 5; Table 3): Patients with only MYC-positive but MYCN-negative tumors had the worst 3-year EFS (N = 8, 12.5% ± 11.7%) compared to patients with only MYCN-positive tumors (N = 39, 49.9% ± 17.7%) and patients with both negative tumors (N = 15, 70.0% ± 17.1%) (P = .0029). One patient whose tumor expressed both MYCN and MYC proteins was excluded from these survival analyses. In a backward-selected Cox model for EFS, none of the factors were statistically significant.
There was a statistically significant association between MKI and MYCN protein expression and the combination of MYCN/MYC protein expression. The majority of patients with high MKI also had positive MYCN and, in addition, negative MYC.
This is the first report describing a clinicopathological summary of NBUD using the largest series of cases from the COG study. NBUD is a histologically defined entity according to the INPC, and all the tumors in this subtype are classified into the Unfavorable Histology Group. Our study clearly demonstrated that the majority of NBUD tumors had prominent nucleoli, and showed the feature of so-called “Large cell neuroblastoma.” These tumors were often associated with poor prognostic factors, such as older age (≥18 months old) at diagnosis, metastatic disease, MYCN amplification, a diploid pattern, 1pLOH, and a high MKI.[8, 9, 16] Patients with NBUD had a poor clinical outcome that was not significantly influenced by the presence/absence of those prognostic factors except for MYCN status and the combination of MYCN status with the presence/absence of prominent nucleoli.
Among these prognostic factors, MYCN amplification has been considered the most powerful indicator of clinically aggressive neuroblastoma.[8, 10] It is reported that MYCN-amplified tumors often show prominent nucleoli that could be a putative site for RNA synthesis and accumulation of excess amounts of MYCN protein leading to their aggressive behavior. In this series, the majority of NBUD tumors had prominent nucleoli and MYCN amplification as well. The results of immunohistochemical studies showed that 42 of 48 (88%) MYCN-amplified tumors expressed MYCN protein regardless of the presence or absence of prominent nucleoli.
Surprisingly, this study demonstrated that patients with MYCN-nonamplified NBUD tumors had significantly worse 3-year EFS than those with MYCN-amplified NBUD tumors. Further analysis demonstrated that MYCN-nonamplified tumors having prominent nucleoli had the worst prognosis in this subtype. These findings prompted us to look for other protein expression than MYCN, and led to our immunohistochemical study to systematically detect MYC protein expression in NBUD tumors.
The MYC (c-MYC) gene, along with the MYCN gene, belongs to the MYC gene family, and is well known to regulate tumor cell proliferation, differentiation inhibition, angiogenesis, and metastasis, in various cancers, such as lymphomas, breast cancers, and medulloblastomas.[17-25] In human neuroblastoma, however, little is known and reported about the role of MYC, whereas a significant body of literature describes basic, translational, and clinical data on MYCN. Recently, overexpression of either MYCN or MYC is reported to regulate tumorigenesis and tumor proliferation through BMI1 in human neuroblastoma. They both encode basic helix-loop-helix leucine zipper proteins that make heterodimers with their obligate partner protein, MAX. The MYC-MAX heterodimer binds to DNA consensus core binding sites, 5′-CACGTG-3′ or variants (E-boxes), which preferentially leads to transcriptional activation of target genes.
Amplified MYCN is almost consistently associated with high MYCN mRNA and protein levels, and increased MYCN expression is known to be involved in tumor initiation and progression in neuroblastoma. Approximately 20% of neuroblastoma tumors have amplified MYCN and are associated with a poor clinical outcome.[2, 13, 28-31] In contrast, neuroblastoma-derived cell lines that lack amplified MYCN are reported to express MYC rather than MYCN, and often at higher levels.[31, 32] From the experiments with neuroblastoma cell lines, it is documented that MYCN and MYC control their expression via autoregulatory loops and via repressing each other at defined promoter sites. Accordingly, in human neuroblastomas, there also seems to be an inverse correlation of MYCN and MYC expression.
Our immunohistochemical study clearly demonstrated that MYC-protein expression was almost exclusively found in MYCN-nonamplified tumors, especially when they had prominent nucleoli. Only one MYCN-amplified tumor was positive for MYC protein and the same tumor was also positive for MYCN protein. Furthermore, patients with MYC-protein–positive tumors had significantly low 3-year EFS rates compared to those with MYCN-protein–positive tumors and both negative tumors.
It should be noted here that MYCN and MYC protein expression may not always be associated with prominent nucleolar formation even though those nucleoli are considered as the site of RNA synthesis and accumulation. Six MYCN-amplified tumors showed MYCN protein expression without prominent nucleolar formation, and one MYCN-nonamplified tumor expressed MYC-protein with no detectable prominent nucleoli. Even one MYCN-amplified tumor having no prominent nucleoli expressed both MYCN and MYC protein. Further study should be required between molecular properties and their morphological manifestations in this group of tumors.
The MKI is proposed to define cellular activities of proliferation (mitosis) and apoptosis due to DNA instability (karyorrhexis) in neuroblastoma tumors. A high MKI was significantly associated with MYCN amplification and the presence of prominent nucleoli in this subtype of NBUD. It was also noted that MYCN-protein-positive NBUD tumors often (30 of 38, 79%) had a high MKI. In contrast, only 2 of 8 (25%) MYC-protein-positive NBUD tumors had a high MKI. This could suggest some difference(s) in the molecular/biological function between MYCN protein and MYC protein, even though both are the products of the same MYC family genes and seem to make strong indicators of a poor prognosis in patients with this disease.
In summary, NBUD represents a unique subtype of pNTs and is frequently associated with prominent nucleoli. Patients with NBUD had a poor clinical outcome. In this study, we presented the data suggesting that MYC-protein expression could make a new prognostic factor indicating more aggressive clinical behavior than MYCN amplification and subsequent MYCN-protein expression. We are now conducting further investigation into MYC-protein expression and its prognostic effects in other subtypes of neuroblastoma, especially in the subset of MYCN-NA tumors of Unfavorable Histology. Also necessary is a study on the mechanism(s) of MYC-protein expression in neuroblastoma tumors. Based on our limited experience using FISH, MYC expression in NBUD often seems to take place through mechanism(s) other than amplification (data not shown). However, we recently experienced a case of MYC-protein-expressing neuroblastoma having MYC gene amplification (Dr. Andrew Gifford, written personal communication on September 17, 2012).
No specific funding was disclosed.
Dr. Hogarty has received salary support for Children's Oncology Group (COG)-related activities, and travel to COG meetings has been reimbursed. He also has grants submitted and/or awarded for research from numerous granting agencies, received course instructor honorarium from Acadia University, has intellectual property on ARID1 (no patent filed, no monetary value anticipated), and is a committee member and is reimbursed for travel and honorarium from the American Board of Pediatrics. Dr. Gastier-Foster's institution has received, from NCI/SAIC, a contract to Biospecimen Core Resource for specimen processing, support from NCI/SAIC for travel to meetings for the study or other purposes, grants from NCI/SAIC, Children's Oncology Group, and SWOG, and travel/accommodations/meeting expenses from NCI, Children's Oncology Group, SWOG, American Board of Medical Genetics (unrelated to activities listed). Dr. London's institution has received grants from NIH/NCI–Children's Oncology Group Statistics and Data Center. Dr. Shimada has received National Institutes of Health grants U10 CA98413 and U10 CA98543. The other authors made no disclosure.