TH-MYCN tumors, but not tumor-derived cell lines, are adrenergic lineage, GD2+, and responsive to anti-GD2 antibody therapy.

Neuroblastoma is a commonly lethal solid tumor of childhood and intensive chemoradiotherapy treatment cures ~50% of children with high-risk disease. The addition of immunotherapy using dinutuximab, a monoclonal antibody directed against the GD2 disialoganglioside expressed on neuroblasts, improves survival when incorporated into front-line therapy and shows robust activity in regressing relapsed disease when combined with chemotherapy. Still, many children succumb to neuroblastoma progression despite receiving dinutuximab-based immunotherapy, and efforts to counteract the immune suppressive signals responsible are warranted. Animal models of human cancers provide useful platforms to study immunotherapies. TH-MYCN transgenic mice are immunocompetent and develop neuroblastomas at autochthonous sites due to enforced MYCN expression in developing neural crest tissues. However, GD2-directed immunotherapy in this model has been underutilized due to the prevailing notion that TH-MYCN neuroblasts express insufficient GD2 to be targeted. We demonstrate that neuroblasts in TH-MYCN-driven tumors express GD2 at levels comparable to human neuroblastomas but rapidly lose GD2 expression when explanted ex vivo to establish tumor cell lines. This occurs in association with a transition from an adrenergic to mesenchymal differentiation state. Importantly, not only is GD2 expression retained on tumors in situ, treatment with a murine anti-GD2 antibody, 14G2a, markedly extends survival in such mice, including durable complete responses. Tumors in 14G2a-treated mice have fewer macrophage and myeloid-derived suppressor cells in their tumor microenvironment. Our findings support the utility of this model to inform immunotherapy approaches for neuroblastoma and potential opportunities to investigate drivers of adrenergic to mesenchymal fate decisions.

Augmented expression of MYC and/or MYCN protein defines highly aggressive MYC-driven neuroblastoma: a Children's Oncology Group study.

BACKGROUND: MYCN amplification with subsequent MYCN protein overexpression is a powerful indicator of poor prognosis of neuroblastoma patients. Little is known regarding the prognostic significance of the homologous MYC protein expression in neuroblastoma. METHODS: Immunostaining for MYCN and MYC protein was performed on 357 undifferentiated/poorly differentiated neuroblastomas. Results were analysed with other prognostic markers. RESULTS: Sixty-seven (19%) tumours were MYCN(+), 38 (11%) were MYC(+), and one(0.3%) had both proteins(+). MYCN(+) tumours and MYC(+) tumours were more likely diagnosed in children>18months with stage4-disease. MYCN(+) tumours were associated with amplified MYCN, Unfavourable Histology (UH), and High-MKI (Mitosis-Karyorrhexis Index). MYC(+) tumours were also frequently UH but not associated with MYCN amplification, and more likely to have low-/intermediate-MKI. Favourable Histology patients without MYC/MYCN expressions exhibited the best survival (N=167, 89.7±5.5% 3-year EFS, 97.0±3.2% 3-year OS), followed by UH patients without MYC/MYCN expressions (N=84, 63.1±13.6% 3-year EFS, 83.5±9.4% 3-year OS). MYCN(+)patients and MYC(+)patients had similar and significantly low (P<0.0001) survivals (46.2±12.0% 3-year EFS, 63.2±12.1% 3-year OS and 43.4±23.1% 3-year EFS, 63.5±19.2% 3-year OS, respectively). Notably, the prognostic impact imparted by MYC expression was independent from other markers. CONCLUSIONS: In this series, ∼30% of neuroblastomas had augmented MYCN or MYC expression with dismal survivals. Prospective study of MYC/MYCN protein expression signature as a new biomarker for high-risk neuroblastomas should be conducted.

Multiple components of the spliceosome regulate Mcl1 activity in neuroblastoma.

Cancer treatments induce cell stress to trigger apoptosis in tumor cells. Many cancers repress these apoptotic signals through alterations in the Bcl2 proteins that regulate this process. Therapeutics that target these specific survival biases are in development, and drugs that inhibit Bcl2 activities have shown clinical activity for some cancers. Mcl1 is a survival factor for which no effective antagonists have been developed, so it remains a principal mediator of therapy resistance, including to Bcl2 inhibitors. We used a synthetic-lethal screening strategy to identify genes that regulate Mcl1 survival activity using the pediatric tumor neuroblastoma (NB) as a model, as a large subset are functionally verified to be Mcl1 dependent and Bcl2 inhibitor resistant. A targeted siRNA screen identified genes whose knockdown restores sensitivity of Mcl1-dependent NBs to ABT-737, a small molecule inhibitor of Bcl2, BclXL and BclW. Three target genes that shifted the ABT-737 IC50 >1 log were identified and validated: PSMD14, UBL5 and PRPF8. The latter two are members of a recently characterized subcomplex of the spliceosome that along with SART1 is responsible for non-canonical 5'-splice sequence recognition in yeast. We showed that SART1 knockdown similarly sensitized Mcl1-dependent NB to ABT-737 and that triple knockdown of UBL5/PRPF8/SART1 phenocopied direct MCL1 knockdown, whereas having no effect on Bcl2-dependent NBs. Both genetic spliceosome knockdown or treatment with SF3b-interacting spliceosome inhibitors like spliceostatin A led to preferential pro-apoptotic Mcl1-S splicing and reduced translation and abundance of Mcl1 protein. In contrast, BN82865, which inhibits the second transesterification step in terminal spliceosome processing, did not have this effect. These findings demonstrate a prominent role for the spliceosome in mediating Mcl1 activity and suggest that drugs that target either the specific UBL5/PRPF8/SART1 subcomplex or SF3b functions may have a role as cancer therapeutics by attenuating the Mcl1 survival bias present in numerous cancers.

BH3 response profiles from neuroblastoma mitochondria predict activity of small molecule Bcl-2 family antagonists.

Bcl-2 family proteins regulate mitochondrial apoptosis downstream of diverse stressors. Cancer cells frequently deregulate Bcl-2 proteins leading to chemoresistance. We have optimized a platform for solid tumors in which Bcl-2 family resistance patterns are inferred. Functional mitochondria were isolated from neuroblastoma (NB) cell lines, exposed to distinct BH3-domain peptides, and assayed for cytochrome c release. Such BH3 profiles revealed three patterns of cytochrome c response. A subset had a dominant NoxaBH3 response implying Mcl1 dependence. These cells were more sensitive to small molecules that antagonize Mcl1 (AT-101) than those that antagonize Bcl-2, Bcl-xL and Bcl-w (ABT-737). A second subset had a dominant BikBH3 response, implying a Bcl-xL/-w dependence, and was exquisitely sensitive to ABT-737 (IC(50) <200 nM). Finally, most NB cell lines derived at relapse were relatively resistant to pro-death BH3 peptides and Bcl-2 antagonists. Our findings define heterogeneity for apoptosis resistance in NB, help triage emerging Bcl-2 antagonists for clinical use, and provide a platform for studies to characterize post-therapy resistance mechanisms for NB and other solid tumors.

N-myc augments death and attenuates protective effects of Bcl-2 in trophically stressed neuroblastoma cells.

N-myc has proapoptotic functions, yet it acts as an oncogene in neuroblastoma. Thus, antiapoptotic mechanisms have to be operative in neuroblastoma cells that antagonize the proapoptotic effects of N-myc. We conditionally activated N-myc in SH-EP neuroblastoma cells subjected to the trophic stress of serum or nutrient deprivation while changing the expression of Bcl-2, survivin and FLIP(L), antiapoptotic molecules often overexpressed in poor prognosis neuroblastomas. Bcl-2 protected SH-EP cells from death during nutritional deprivation by activating energetically advantageous oxidative phosphorylation. N-myc overrode the metabolic protection provided by Bcl-2-induced oxidative phosphorylation by reestablishing the glycolytic phenotype and attenuated the antiapoptotic effect of Bcl-2 during metabolic stress. Survivin partially antagonized the growth suppressive function of N-myc in SH-EP neuroblastoma cells during serum deprivation whereas FLIP(L) did not. These findings advance our understanding of the functions of N-myc in neuroblastoma cells.

Deregulated Wnt/beta-catenin program in high-risk neuroblastomas without MYCN amplification.

Neuroblastoma (NB) is a frequently lethal tumor of childhood. MYCN amplification accounts for the aggressive phenotype in a subset while the majority have no consistently identified molecular aberration but frequently express MYC at high levels. We hypothesized that activated Wnt/beta-catenin (CTNNB1) signaling might account for this as MYC is a beta-catenin transcriptional target and multiple embryonal and neural crest malignancies have oncogenic alterations in this pathway. NB cell lines without MYCN amplification express higher levels of MYC and beta-catenin (with aberrant nuclear localization) than MYCN-amplified cell lines. Evidence for aberrant beta-catenin-TCF transcriptional activity was demonstrated using expression profiles from 73 primary NBs. Findings included increased WNT ligands (WNT1, WNT6, WNT7A, WNT10B), DVL1 and TCF7 expression in high-risk NBs without MYCN amplification, consistent with canonical beta-catenin signaling. More directly, Patterns of Gene Expression and Gene Set Enrichment Analyses demonstrated beta-catenin target genes (for example, MYC, PPARD, NRCAM, CD44, TCF7) as coordinately upregulated in high-risk NBs without MYCN amplification in comparison to high-risk MYCN-amplified or intermediate-risk NBs, supporting pathway activation in this subset. Thus, high-risk NBs without MYCN amplification may deregulate MYC and other oncogenic genes via altered beta-catenin signaling providing a potential candidate pathway for therapeutic inhibition.

BH3 peptidomimetics potently activate apoptosis and demonstrate single agent efficacy in neuroblastoma.

The major impediment to cure for many malignancies is the development of therapy resistance with resultant tumor progression. Genetic alterations leading to subversion of inherent apoptosis pathways are common themes in therapy resistance. Bcl-2 family proteins play a critical role in regulating mitochondrial apoptosis that governs chemotherapeutic effects, and defective engagement of these pathways contributes to treatment failure. We have studied the efficacy of BH3 peptidomimetics consisting of the minimal death, or BH3, domains of the proapoptotic BH3-only proteins Bid and Bad to induce apoptosis using neuroblastoma (NB) as a model system. We demonstrate that BH3 peptides, modified with an arginine homopolymer for membrane transduction (called r8-BidBH3 and r8-BadBH3, respectively), potently induce apoptosis in NB cells, including those with MYCN amplification. Cell death is caspase 9 dependent, consistent with a requirement for the intrinsic mitochondrial pathway. Substitutions at highly conserved residues within the r8-BidBH3 peptide abolish apoptotic efficacy supporting activity through specific BH domain interactions. Concomitant exposure to r8-BadBH3 and r8-BidBH3 at sublethal monotherapy doses revealed potent synergy consistent with a competitive displacement model, whereby BH3 peptides displace sequestered BH3 proteins to induce cell death. Further, BH3 peptides demonstrate antitumor efficacy in a xenograft model of NB in the absence of additional genotoxic or trophic stressors. These data provide proof of principle that targeted re-engagement of apoptosis pathways may be of therapeutic utility, and BH3-like compounds are attractive lead agents to re-establish therapy-induced apoptosis in refractory malignancies.

MYC messenger RNA expression predicts survival outcome in childhood primitive neuroectodermal tumor/medulloblastoma.

PURPOSE AND EXPERIMENTAL DESIGN: Cerebellar primitive neuroectodermal tumors/medulloblastomas (PNET/MB) are the most common malignant brain tumors in childhood. To identify PNET/MB biological prognostic factors that define a patient group with a sufficiently good prognosis to permit a reduction in treatment intensity, we determined the expression levels of MYC mRNA in fresh frozen tumor samples from 26 PNET/MB patients using semiquantitative reverse transcription-PCR. RESULTS: MYC mRNA expression levels in primary PNET/MB showed a wide range with a 22-fold difference between the highest and lowest values and did not correlate with MYC gene amplification. MYC mRNA expression was an independent significant prognostic factor for progression-free survival outcome and was more predictive than standard clinical factors. The combination of low MYC mRNA expression and high TrkC mRNA expression identified a good outcome group of PNET/MB patients (n = 7) with 100% progression-free survival after a median follow-up time of 55 months (range, 15-91 months). Three of these seven good outcome patients survived without radiotherapy. CONCLUSIONS: Low MYC mRNA expression is a powerful independent predictor of favorable clinical outcome in PNET/MB. Assessment of MYC mRNA levels is feasible and may be incorporated in prospective PNET/MB clinical trials to aid in treatment planning for patients with PNET/MB on confirmation of our results in larger studies.

Allelic deletion at chromosome bands 11q14-23 is common in neuroblastoma.

BACKGROUND: Neuroblastoma tumorigenesis may involve the differential inactivation of multiple tumor suppressor genes. Recent data have suggested that a neuroblastoma suppressor gene may be located on the long arm of chromosome 11 (11q). PROCEDURE: We therefore analyzed 295 primary neuroblastomas from a representative group of patients for loss of heterozygosity (LOH) at 25 polymorphic markers spanning 11q. RESULTS: LOH was observed in 129 primary neuroblastomas (44%), and a common region of LOH mapped to 11q14-23. No correlation was found between 11q LOH and adverse prognostic variables, but a strong inverse relationship between 11q LOH and MYCN amplification (P < 0.001) was observed. There was no difference in overall survival when patients were stratified by 11q LOH status. However, 11q LOH was associated with a decreased overall survival probability when patients whose tumors had a single copy of MYCN were analyzed separately (P = 0.008). CONCLUSION: These data support the hypothesis that a tumor suppressor gene mapping within 11q14-23 is frequently inactivated during the malignant evolution of neuroblastoma.

Comprehensive analysis of chromosome 1p deletions in neuroblastoma.

BACKGROUND: Chromosome 1p deletions are common in advanced neuroblastomas, but the biological and clinical implications of this clonal rearrangement remain controversial. Previous studies of chromosome 1p loss of heterozygosity (LOH) have been limited by analyses of relatively small number of tumors derived from heterogeneously assessed and treated patient populations. Therefore, a strictly representative cohort of 288 Children's Cancer Group neuroblastoma patients treated on the most recent phase III therapeutic trials was identified. PROCEDURE: Primary tumors from these patients were analyzed for LOH at precisely mapped and highly informative 1p polymorphic loci located from 1p32 to 1p36.3 by multiplex PCR. RESULTS: Ninety-three primary tumor specimens (32%) had LOH at multiple 1p36 marker loci. All 1p deletions overlapped the previously determined smallest region of overlap (SRO). One tumor had a small terminal deletion completely within 1p36.3, allowing for further refinement of the 1p36 SRO. We found no evidence to support an additional, nonoverlapping region of LOH within 1p32-36. We confirmed the strong correlation of 1p36 LOH with MYCN amplification (P < 0.001), advanced disease stage (P < 0.001), and decreased both 3-year event-free survival and overall survival probabilities (P< 0.001). When stratified for MYCN amplification status or entered into a multivariate analysis, 1p36 LOH remained predictive for decreased event-free survival, but not overall survival probability. CONCLUSIONS: These data support the hypothesis that inactivation of a tumor suppressor gene within 1p36.3 is associated with an increased risk for disease relapse.

Detailed molecular analysis of 1p36 in neuroblastoma.

BACKGROUND: Several lines of evidence es tablish that chromosome band 1p36 is frequently deleted in neuroblastoma primary tumors and cell lines, suggesting that a tumor suppressor gene within this region is involved in the development of this tumor. PROCEDURE: We analyzed the status of 1p36 in primary neuroblastomas and cell lines to define the region of consistent rearrangement. RESULTS: Loss of heterozygosity (LOH) studies of primary neuro blastomas identified allelic loss in 135 of 503 tumors (27%), with the smallest region of overlap (SRO) defined distal to D15214 (1p36.3). No homozygous deletions were detected at 120 loci mapping to 1p36.1-p36.3 in a panel of 46 neuroblastoma cell lines. A recently identified patient with neuroblastoma was found to have a constitutional deletion within 1p36.2-p36.3, and this deletion, when combined with the LOH results, defined a smaller SRO of one megabase within 1p36.3. We constructed a comprehensive integrated map of chromosome 1 containing 11,000 markers and large-insert clones, a high-resolution radiation hybrid (RH) map of 1p36, and a P1-artificial chromosome (PAC) contig spanning the SRO, to further characterize the region of interest. Over 768 kb (75%) of the SRO has been sequenced to completion. Further analysis of distal 1p identified 113 transcripts localizing to 1p36, 21 of which were mapped within the SRO. CONCLUSION: This analysis will identify suitable positional candidate transcripts for mutational screening and subsequent identification of the 1p36.3 neuroblastoma suppressor gene.

Analysis of genomic imprinting at 1p35-36 in neuroblastoma.

BACKGROUND: Deletion of the distal short arm of chromosome 1 occurs in 25-35% of primary neuroblastomas, and a putative tumor suppressor gene has been mapped to a consensus region of deletion at 1p36.2-36.3. Indirect evidence suggests the presence of an imprinted neuroblastoma suppressor gene within this region, as well as an additional nonimprinted, proximal suppressor gene, inactivation of which correlates with MYCN amplification. PROCEDURE: To test this hypothesis, we performed 1p loss of heterozygosity (LOH) studies on a series of neuroblastomas for which parental DNA had been collected. PCR-formatted polymorphic markers were used to determine the size of the 1p deletion and the parental origin of the deleted 1p homologue. RESULTS: Twenty-six neuroblastomas with 1p LOH were evaluated. Twenty-four had MYCN amplification, and of these, 15 demonstrated loss of the paternally inherited 1p. Two neuroblastomas with a single copy of MYCN were evaluated and both had deletion of the paternally inherited 1p, with one case exhibiting a small terminal deletion. In addition, we have reviewed 49 previously reported neuroblastomas where 1p LOH data and the parental origin of the deleted lp homologue were available. CONCLUSIONS: Analyzed together, these 75 neuroblastomas demonstrate random deletion of parental 1p homologues (P = 0.30). Further, tumors with smaller deletions (breakpoints distal to D1S201 or D1S7) showed a random loss of the parental 1p homologues (P = 0.59), contrary to the expected preferential maternal 1p deletion if an imprinted suppressor gene mapped to this region. However, 19 tumors with 1p LOH and single copy MYCN had deletion of the maternal 1p homologue preferentially (P = 0.02), which does not exclude the possibility that loss of an imprinted suppressor gene plays a role in this subset.

Homozygous deletion of CDKN2A (p16INK4a/p14ARF) but not within 1p36 or at other tumor suppressor loci in neuroblastoma.

Loss of heterozygosity of several specific genomic regions is frequently observed in neuroblastoma tumors and cell lines, but homozygous deletion (HD) is rare, and no neuroblastoma tumor suppressor gene (TSG) has yet been identified. We performed a systematic search for HD, indicative of a disrupted TSG, in a panel of 46 neuroblastoma cell lines. An initial search focused on a well-characterized consensus region of hemizygous deletion at 1p36.3, which occurs in 35% of primary neuroblastomas. Each cell line was screened with 162 1p36 markers, for a resolution of 13 kb within the consensus 1p36.3 deletion region and 350 kb throughout the remainder of 1p36. No HDs were detected. This approach was expanded to survey 21 known TSGs, specifically targeting intragenic regions frequently inactivated in other malignancies. HD was detected only at the CDKN2A (p16INK4a/p14ARF) gene at 9p21 and was observed in 4 of 46 cell lines. The observed region of HD included all exons of both CDKN2A and the closely linked CDKN2B (p15INK4b) gene for cell lines LA-N-6 and CHLA-174, all exons of CDKN2A but none of CDKN2B for CHLA-179, and only 104 bp within CDKN2A exon 2 for CHLA-101. All four deletions are predicted to inactivate the coding regions of both p16INK4a and p14ARF. HD was observed in corresponding primary tumor samples for CHLA-101 and CHLA-174 but was not present in constitutional samples. These results suggest that for neuroblastoma, large HDs do not occur within 1p36, most known TSGs are not homozygously deleted, and biallelic inactivation of CDKN2A may contribute to tumorigenicity in a subset of cases.

Identification of a 1-megabase consensus region of deletion at 1p36.3 in primary neuroblastomas.

BACKGROUND: Deletion of the distal short arm of chromosome 1 occurs frequently in neuroblastoma. In addition, neuroblastoma has been described in children with constitutional deletions within 1p36, supporting the existence of one or more neuroblastoma suppressor genes within this region. PROCEDURE: We have pursued a 1p36 tumor suppressor gene identification strategy that has included deletion mapping of 566 primary neuroblastomas and 46 neuroblastoma-derived cell lines, and have determined the parental origin of the deleted 1p homologue in 44 cases to determine whether there is evidence for genomic imprinting within this region. RESULTS AND CONCLUSIONS: We have identified a 1-Mb consensus region of deletion within 1p36.3 defined by primary tumor deletions, constructed a physical map of the region that is being sequenced to completion, and have identified and prioritized candidate genes within this region for further analyses.

Localization of a hereditary neuroblastoma predisposition gene to 16p12-p13.

BACKGROUND: Hereditary predisposition to develop neuroblastoma segregates as an autosomal dominant Mendelian trait. PROCEDURE: We have performed linkage analysis on 10 families with neuroblastoma to localize a hereditary neuroblastoma predisposition gene (HNB1). RESULTS: A single genomic interval at chromosome bands 16p12-p13 was consistent with linkage (lod = 3.46), and identification of informative recombinants defined a 25.9-cM critical region between D16S748 and D16S3068. Loss of heterozygosity was identified in 5/12 familial (42%) and 55/259 nonfamilial (21%) neuroblastomas at multiple 16p polymorphic loci. A 12.8-cM smallest region of overlap of deletions was identified within the interval defined by linkage analysis (tel-D16S764-D16S412-cen). CONCLUSIONS: Taken together, these data suggest that HNB1 is located at 16p12-p13 and that inactivation of this gene may contribute to the pathogenesis of nonfamilial neuroblastomas.

Deletion of 11q23 is a frequent event in the evolution of MYCN single-copy high-risk neuroblastomas.

BACKGROUND: Deletions of the long arm of chromosome 11 are frequently identified in human neuroblastomas. PROCEDURE: We screened 394 primary neuroblastomas and 52 tumor-derived cell lines with a panel of 11q and 11p polymorphic markers to determine the frequency of chromosome 11 allelic deletion, and to differentiate partial deletions of chromosome 11q (unb[11q] LOH) from whole chromosome loss. RESULTS: Allelic deletion occurred most frequently at cytogenetic band 11q23 and was detected in 161 primary neuroblastomas (41%) and 18 cell lines (35%). Eighty-seven tumors (22%) had unb[11q] LOH with a heterogeneous distribution of deletion breakpoints. Unb[11q] LOH was highly correlated with age > 1 year at diagnosis (P = 0.008), stage 4 disease (P = 0.001), unfavorable Shimada histopathology (P < 0.001), and assignment to a high-risk therapeutic protocol (P < 0.001), and was inversely correlated with MYCN amplification (P = 0.018). Patients whose tumors showed unb[11q] LOH were less likely to survive (P < 0.001), but there was only a trend towards an independent prognostic influence in multivariate analyses. CONCLUSIONS: Thus, structural rearrangements resulting in unb[11q] LOH commonly occur during the malignant evolution of high-risk neuroblastomas with single-copy MYCN.

BIN1 inhibits colony formation and induces apoptosis in neuroblastoma cell lines with MYCN amplification.

BACKGROUND: MYCN amplification and overexpression occurs in 25% of neuroblastomas and independently predicts for poor prognosis disease, an effect thought to be mediated by its role as a transcriptional activator of growth promoting genes. However, in many mammalian cells, deregulated expression of MYC family genes (including MYCN) induces apoptosis. We hypothesized that BIN1, a MYC interacting protein capable of inducing apoptosis, may be an important regulator of MYCN in neuroblastoma. RESULTS: BIN1 expression was found to be reduced in MYCN-amplified cell lines. Further, forced expression of BIN1 markedly reduced colony formation in MYCN-amplified, but not single-copy, cell lines. This effect appeared to be caused by an increase in apoptosis, and was augmented by serum deprivation and concurrent cytotoxic drug therapy in cell culture CONCLUSION: BIN1 inactivation may be necessary for MYCN overexpression to lead to cellular proliferation rather than programmed cell death in neuroblastomas with MYCN amplification.

Loss of heterozygosity at 1p36 independently predicts for disease progression but not decreased overall survival probability in neuroblastoma patients: a Children's Cancer Group study.

PURPOSE: To determine the independent prognostic significance of 1p36 loss of heterozygosity (LOH) in a representative group of neuroblastoma patients. PATIENTS AND METHODS: Diagnostic tumor specimens from 238 patients registered onto the most recent Children's Cancer Group phase III clinical trials were assayed for LOH with 13 microsatellite polymorphic markers spanning chromosome band 1p36. Allelic status at 1p36 was correlated with other prognostic variables and disease outcome. RESULTS: LOH at 1p36 was detected in 83 (35%) of 238 neuroblastomas. There was a correlation of 1p36 LOH with age at diagnosis greater than 1 year (P = .026), metastatic disease (P<.001), elevated serum ferritin level (P<.001), unfavorable histopathology (P<.001), and MYCN oncogene amplification (P<.001). LOH at 1p36 was associated with decreased event-free survival (EFS) and overall survival (OS) probabilities (P<.0001). For the 180 cases with single-copy MYCN, 1p36 LOH status was highly correlated with decreased EFS (P = .0002) but not OS (P = .1212). Entering 1p36 LOH into a multivariate regression model suggested a trend toward an independent association with decreased EFS (P = .0558) but not with decreased OS (P = .3687). Furthermore, allelic status at 1p36 was the only prognostic variable that was significantly associated with decreased EFS in low-risk neuroblastoma patients (P = .0148). CONCLUSION: LOH at 1p36 is independently associated with decreased EFS, but not OS, in neuroblastoma patients. Determination of 1p36 allelic status may be useful for predicting which neuroblastoma patients with otherwise favorable clinical and biologic features are more likely to have disease progression.

Longitudinal evaluation of cardiopulmonary performance during exercise after bone marrow transplantation in children.

OBJECTIVE: Abnormalities in cardiopulmonary performance during exercise have been reported in children after bone marrow transplantation (BMT). We sought to study changes in exercise performance over time in pediatric BMT survivors. STUDY DESIGN: We retrospectively reviewed the results of serial cardiopulmonary exercise tests performed by patients who had undergone BMT at our institution. Four measurements of cardiopulmonary function are reported: maximum cardiac index (MCI), maximal oxygen consumption (Max VO(2)), oxygen consumption at ventilatory threshold (VO(2) at VT), and maximum work (Max Work) performed. A linear mixed-effects model was fitted to assess changes in these parameters over time. RESULTS: Thirty-three patients performed 96 cardiopulmonary exercise tests. MCI and VO(2) at VT were depressed at initial testing and did not change over time. Max VO(2) increased by 4% per year to 69% predicted, and Max Work increased to 77% predicted at 6 years after BMT. CONCLUSIONS: In spite of an impaired cardiovascular response to exercise as indicated by the persistently low MCI, aerobic and physical working capacity increase. Improved Max VO(2) suggests that oxygen extraction at the musculoskeletal level becomes more efficient with recovery from BMT. This may represent a compensatory response to an impaired ability to increase cardiac output.

Allelic deletion at 11q23 is common in MYCN single copy neuroblastomas.

Deletions of the long arm of chromosome 11 (11q) have been noted in primary neuroblastomas, but a comprehensive analysis has not been performed. Therefore, we analysed 331 neuroblastomas (295 sporadic, 15 familial and 21 tumor-derived cell lines) to determine the prevalence of 11q allelic deletions, to map the location of a putative tumor suppressor gene and to perform clinical correlative studies. Assays for loss of heterozygosity (LOH) were performed at 24 microsatellite loci spanning 11q. LOH was observed at multiple 11q loci in 129/295 (44%) sporadic neuroblastomas, 5/15 (33%) familial neuroblastomas, and 5/21 (24%) neuroblastoma cell lines. A single region of 2.1 cM within 11q23.3, flanked by markers D11S1340 and D11S1299, was deleted in all specimens with 11q LOH. Allelic loss at 11q23 was inversely related to MYCN amplification (P<0.001). Within the subset of cases with a single copy of MYCN, 11q LOH was associated with advanced stage disease (P=0.008), unfavorable histopathology (P=0.042), and decreased overall survival probability (P=0.008). However, 11q LOH was not independently prognostic in multivariate analyses. These data support the hypothesis that a tumor suppressor gene mapping within 11q23.3 is commonly inactivated during the malignant evolution of a large subset of neuroblastomas, especially those with unamplified MYCN.