Search

Full information pertaining to supplementary figure legends.

Morphology and proliferation of knockdown cells.

Primer sequence.

Brain metastases occur in about 10% to 30% of breast cancer patients, which culminates in a poor prognosis. It is, therefore, critical to understand the molecular mechanisms underlying brain metastatic processes to identify relevant targets. We hypothesized that breast cancer cells must express brain-associated markers that would enable their invasion and survival in the brain microenvironment. We assessed a panel of brain-predominant markers and found an elevation of several neuronal markers (βIII-tubulin, Nestin, and AchE) in brain metastatic breast cancer cells. Among these neuronal predominant markers, in silico analysis revealed overexpression of βIII-tubulin (TUBB3) in breast cancer brain metastases (BCBM) and its expression was significantly associated with distant metastases. TUBB3 knockdown studies were conducted in breast cancer models (MDA-Br, GLIM2, and MDA-MB-468), which revealed significant reduction in their invasive capabilities. MDA-Br cells with suppressed TUBB3 also demonstrated loss of key signaling molecules such as β3 integrin, pFAK, and pSrc in vitro. Furthermore, TUBB3 knockdown in a brain metastatic breast cancer cell line compromised its metastatic ability in vivo, and significantly improved survival in a brain metastasis model. These results implicate a critical role of TUBB3 in conferring brain metastatic potential to breast cancer cells. Mol Cancer Ther; 14(5); 1152–61. ©2015 AACR.

Generation of TUBB3 knockdown cells.

(A) Sorted Tregs were placed the upper chamber of a 3.0 microm pore cell culture insert with or without 5 microM of the small molecule antagonist C 021 and chemokine-containing media was placed in the well. Migration index was determined after 4 hours through flow cytometric analysis. (B) A similar migration assay was performed as in A, except CCL21 and 10% FBS were used in the bottom chamber as a non-specific positive control. Data are representative of 3 independent experiments. Data are represented as mean {plus minus} SEM; *P < 0.05, by Student's t test.

(A-C) Vehicle control or C 021-treated mice were analyzed at 1 week post-GL261 injection for the accumulation of various T cell subsets and CD11b+Ly-6G+ MDSCs. (D-G) Quantification of flow cytometric analysis. (H-I) C 021 treatment improved the CD4/Foxp3 ratio and CD8/Foxp3 ratio. Data are representative of at least 2 independent experiments. Data are represented as mean {plus minus} SEM; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by Student's t test.

(A-G) Kaplan-Meier survival curves generated from TCGA data using GBM patients segregated by gene expression for a panel of putative Treg and MDSC recruiting chemokines. (H) GBM patient tissue array stained for CCL2. (I) Key showing CCL2 score corresponding to GBM patient cores. Scale bar = 7 mm. TCGA, The Cancer Genome Atlas. Survival curves were compared with the Log-rank test.

Antibody Array Key (RayBiotech Cat# AAM-CYT-6, related to Figure 3.

In many aggressive cancers, such as glioblastoma multiforme, progression is enabled by local immunosuppression driven by the accumulation of regulatory T cells (Treg) and myeloid-derived suppressor cells (MDSC). However, the mechanistic details of how Tregs and MDSCs are recruited in various tumors are not yet well understood. Here we report that macrophages and microglia within the glioma microenvironment produce CCL2, a chemokine that is critical for recruiting both CCR4+ Treg and CCR2+Ly-6C+ monocytic MDSCs in this disease setting. In murine gliomas, we established novel roles for tumor-derived CCL20 and osteoprotegerin in inducing CCL2 production from macrophages and microglia. Tumors grown in CCL2-deficient mice failed to maximally accrue Tregs and monocytic MDSCs. In mixed-bone marrow chimera assays, we found that CCR4-deficient Treg and CCR2-deficient monocytic MDSCs were defective in glioma accumulation. Furthermore, administration of a small-molecule antagonist of CCR4 improved median survival in the model. In clinical specimens of glioblastoma multiforme, elevated levels of CCL2 expression correlated with reduced overall survival of patients. Finally, we found that CD163-positive infiltrating macrophages were a major source of CCL2 in glioblastoma multiforme patients. Collectively, our findings show how glioma cells influence the tumor microenvironment to recruit potent effectors of immunosuppression that drive progression. Cancer Res; 76(19); 5671–82. ©2016 AACR.

Word document file containing supplementary materials and methods, supplementary figure legends, and supplementary table

CD45.1+ recipient mice were irradiated and reconstituted with CD45.1+ WT marrow and CD45.2+ Ccr4-/- marrow. After at least 6 weeks of reconstitution, mice were implanted with GL261 tumors and brain tumor-infiltrating cells were quantified. Data are representative of at least 2 independent experiments. Data are represented as mean {plus minus} SEM; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by Student's t test. n.s., not significant.

(A) GL261 cells were orthotopically implanted into the right hemisphere of syngeneic immunocompetent C57BL/6 mice. At 1 week post-intracranial (i.c.) injection whole brains were homogenized followed by Percoll gradient separation of leukocytes and non-leukocytes. qRT-PCR was performed for CCL2 in mice intracranially-injected with GL261 cells (n = 4) or PBS-injected controls (n = 3). (B) Detection of CCL2 protein in both brain hemispheres at 1 week-i.c. GL261 (n = 3). (C) Immunoreactivity for CCL2 was detected at 1 week post-GL261 injection. 20Ã- and 63Ã- magnification. Scale bar = 50 microm. (D) qRT-PCR for CCL1, CCL28, CCL2, CCL22, CCL17, and CCL20 from non-leukocytes (NL) and leukocytes (L) as in (A). Images in C are representative of 3 independent replicates. Data are representative of at least 2 independent experiments. Data are represented as mean {plus minus} SEM; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by one-way ANOVA with Tukey's multiple comparisons test (A, D) or Student's t test (B).

(A) CCR4 and CCR2 expression on CD4+Foxp3- conventional CD4 T cells, CD4+Foxp3+ Tregs, and CD8+ T cells across various tissues at 1 week post-GL261 implantation. (B) Gating strategy for quantification of myeloid cells in brains of PBS-injected or GL261-implanted mice at 1-week post-intracranial implantation. Data are representative of at least 4 independent experiments.

Epigenetic changes are among the most common alterations observed in cancer cells, yet the mechanism by which cancer cells acquire and maintain abnormal DNA methylation patterns is not understood. Cancer cells have an altered distribution of DNA methylation and express aberrant DNA methyltransferase 3B transcripts, which encode truncated proteins, some of which lack the COOH-terminal catalytic domain. To test if a truncated DNMT3B isoform disrupts DNA methylation in vivo, we constructed two lines of transgenic mice expressing DNMT3B7, a truncated DNMT3B isoform commonly found in cancer cells. DNMT3B7 transgenic mice exhibit altered embryonic development, including lymphopenia, craniofacial abnormalities, and cardiac defects, similar to Dnmt3b-deficient animals, but rarely develop cancer. However, when DNMT3B7 transgenic mice are bred with Eμ-Myc transgenic mice, which model aggressive B-cell lymphoma, DNMT3B7 expression increases the frequency of mediastinal lymphomas in Eμ-Myc animals. Eμ-Myc/DNMT3B7 mediastinal lymphomas have more chromosomal rearrangements, increased global DNA methylation levels, and more locus-specific perturbations in DNA methylation patterns compared with Eμ-Myc lymphomas. These data represent the first in vivo modeling of cancer-associated DNA methylation changes and suggest that truncated DNMT3B isoforms contribute to the redistribution of DNA methylation characterizing virtually every human tumor. Cancer Res; 70(14); 5840–50. ©2010 AACR.

Supplementary Methods, Figures 1-10, Tables 1-8 from DNMT3B7, a Truncated DNMT3B Isoform Expressed in Human Tumors, Disrupts Embryonic Development and Accelerates Lymphomagenesis

Objectives Tissue carryovers are contaminants of surgical pathology cases in which extraneous tissue is incorporated into tissue blocks. Carryovers occur most frequently at the grossing or embedding stations, but little is published about them. We sought to analyze their transmission during transit to the histology lab. Methods Cassettes of friable donor tissue were mixed with cassettes of spongy recipient tissue in formalin-filled containers and agitated by shipment via pneumatic tube. The tissue cassettes were processed, embedded as blocks, and cut as usual. Liquid samples were prepared from the submission containers as well as from workstation submission containers and histology tissue processor waste. Results A high rate of contamination (14.9%) was observed under these artificial conditions. Friable donor tissue, including urothelium and colorectal adenocarcinoma, were promiscuous contaminants, as were placental villi. Fluid from submission containers showed viable tumor cells and fragments, which were also present in workstation submission containers and in tissue processor waste fluid. Conclusions This study implicates liquid transport media as a possible avenue of contamination during submission and transportation of tissue cassettes for histologic processing. Attention should be given to the friability of submitted tissue and physical agitation of the cassettes in transit. Such contaminants may be present in the fluid in tissue submission bins and in tissue processor fluid.

Background: Cerebral microbleeds (CMBs) are manifestations of age-related microangiopathies associated with an increased risk of hemorrhagic stroke and cognitive decline. The pathophysiology of this age-related breakdown of the neurovascular unit remain unclear. Understanding CMB cellular architecture is needed for gene, pathway, and mechanism analyses, to pave the way for biomarker and therapeutic development. Herein, we aim to systematically describe histopathological changes in the CMB lesional milieu. Methods: Eight CMBs, along with contralateral non-lesional tissue, were identified by gross appearance at autopsy by a neuropathologist. Samples were first stained with H&E, then Prussian blue for non-heme iron, to localize the lesion. This was followed by multiplex fluorescence staining of β-amyloid and individual cell types including astrocytes (GFAP + ), macrophages (CD163 + /Iba1 + ), microglia (CD163 - /Iba1 + ), B-cells (CD20 + ), T-cells (CD3 + ), and endothelium (CD31 + ). The cell types within the lesional milieu, defined as 200μm from the periphery of the CMB, were quantified using QuPath. Astrocyte and microglia morphology were assessed via fluorescence imaging. Results: One CMB contained dilated capillaries, 2 contained lipohyalinized arterioles, and 4 contained dystrophic arterioles due to amyloid angiopathy. Astrocytes and macrophages showed higher normalized cell counts in all CMB tissues compared with controls ( p and p , respectively). Astrocytes in 7 of 8 CMB milieux showed reactive morphological changes, including larger cell bodies with peripheral nuclei. Microglia had higher normalized cell counts in 4 CMB tissues, with senescence-associated amoeboid, as well as ramified, reactive morphologies. The lesional milieu also showed higher counts of B-cells in 7 CMB tissues and T-cells in 3 CMB tissues when compared with controls. Conclusions: We provide the first description of the cellular architecture of the CMB and the associated lesional milieu. Despite heterogeneity of disrupted vasculature, there are common features of glial reactive changes and inflammation in CMBs. These results offer a basis for additional study of the CMB transcriptome and individual roles of respective cell types in CMB pathology.

High-grade astrocytoma with piloid features (HGAP) is a recently recognized glioma type whose classification is dependent on its global epigenetic signature. HGAP is characterized by alterations in the mitogen-activated protein kinase (MAPK) pathway, often co-occurring with CDKN2A/B homozygous deletion and/or ATRX mutation. Experience with HGAP is limited and to better understand this tumor type, we evaluated an expanded cohort of patients (n = 144) with these tumors, as defined by DNA methylation array testing, with a subset additionally evaluated by next-generation sequencing (NGS). Among evaluable cases, we confirmed the high prevalence CDKN2A/B homozygous deletion, and/or ATRX mutations/loss in this tumor type, along with a subset showing NF1 alterations. Five of 93 (5.4%) cases sequenced harbored TP53 mutations and RNA fusion analysis identified a single tumor containing an NTRK2 gene fusion, neither of which have been previously reported in HGAP. Clustering analysis revealed the presence of three distinct HGAP subtypes (or groups = g) based on whole-genome DNA methylation patterns, which we provisionally designated as gNF1 (n = 18), g1 (n = 72), and g2 (n = 54) (median ages 43.5 years, 47 years, and 32 years, respectively). Subtype gNF1 is notable for enrichment with patients with Neurofibromatosis Type 1 (33.3%, p = 0.0008), confinement to the posterior fossa, hypermethylation in the NF1 enhancer region, a trend towards decreased progression-free survival (p = 0.0579), RNA processing pathway dysregulation, and elevated non-neoplastic glia and neuron cell content (p < 0.0001 and p < 0.0001, respectively). Overall, our expanded cohort broadens the genetic, epigenetic, and clinical phenotype of HGAP and provides evidence for distinct epigenetic subtypes in this tumor type.

Introduction We studied a patient with a congenital myasthenic syndrome (CMS) caused by a dominant mutation in the synaptotagmin 2 gene (SYT2) and compared the clinical features of this patient with those of a previously described patient with a recessive mutation in the same gene. Methods We performed electrodiagnostic (EDX) studies, genetic studies, muscle biopsy, microelectrode recordings and electron microscopy (EM). Results Both patients presented with muscle weakness and bulbar deficits, which were worse in the recessive form. EDX studies showed presynaptic failure, which was more prominent in the recessive form. Microelectrode studies in the dominant form showed a marked reduction of the quantal content, which increased linearly with higher frequencies of nerve stimulation. The MEPP frequencies were normal at rest but increased markedly with higher frequencies of nerve stimulation. The EM demonstrated overdeveloped postsynaptic folding, and abundant endosomes, multivesicular bodies and degenerative lamellar bodies inside small nerve terminals. Discussion The recessive form of CMS caused by a SYT2 mutation showed far more severe clinical manifestations than the dominant form. The pathogenesis of the dominant form likely involves a dominant-negative effect due to disruption of the dual function of synaptotagmin as a Ca2+ -sensor and modulator of synaptic vesicle exocytosis. This article is protected by copyright. All rights reserved.

Autopsies of patients who have died from COVID‐19 have been crucial in delineating patterns of injury associated with SARS‐CoV‐2 infection. Despite their utility, comprehensive autopsy studies are somewhat lacking relative to the global burden of disease, and very few comprehensive studies contextualize the findings to other fatal viral infections. We developed a novel autopsy protocol in order to perform postmortem examinations on victims of COVID‐19 and herein describe detailed clinical information, gross findings, and histologic features observed in the first 16 complete COVID‐19 autopsies. We also critically evaluated the role of ancillary studies used to establish a diagnosis of COVID‐19 at autopsy, including immunohistochemistry (IHC), in situ hybridization (ISH), and electron microscopy (EM). IHC and ISH targeting SARS‐CoV‐2 were comparable in terms of the location and number of infected cells in lung tissue; however, nonspecific staining of bacteria was seen occasionally with IHC. EM was unrevealing in blindly sampled tissues. We then compared the clinical and histologic features present in this series to six archival cases of fatal seasonal influenza and six archival cases of pandemic influenza from the fourth wave of the ‘Spanish Flu’ in the winter of 1920. In addition to routine histology, the inflammatory infiltrates in the lungs of COVID‐19 and seasonal influenza victims were compared using quantitative IHC. Our results demonstrate that the clinical and histologic features of COVID‐19 are similar to those seen in fatal cases of influenza, and the two diseases tend to overlap histologically. There was no significant difference in the composition of the inflammatory infiltrate in COVID‐19 and influenza at sites of acute lung injury at the time of autopsy. Our study underscores the relatively nonspecific clinical features and pathologic changes shared between severe cases of COVID‐19 and influenza, while also providing important caveats to ancillary methods of viral detection.

Clear cell papillary cystadenoma of the epididymis is an uncommon benign neoplasm, usually seen in patients with von Hippel-Lindau disease. Morphologic and immunohistochemical examination aid in distinguishing clear cell papillary cystadenoma from malignant histologic mimics including low-grade mesothelial proliferations and metastatic clear cell renal cell carcinomas. Analogous lesions have been described in the female genital tract, often posing diagnostic challenges due to their low incidence. Here, we present the difficult diagnostic aspects of the first case of clear cell papillary cystadenoma involving the ovary, including the salient immunohistochemical, ultrastructural, and molecular characteristics.

Limited tissue in biopsies of malignant bone lesions can preclude definitive subclassification, especially when cellular or matrix elements are sparse, absent, or confounding. It is uncertain whether immunohistochemistry for SOX9 (marker of chondrogenesis) and SATB2 (marker of osteoblastic differentiation) may be discriminatory tools toward osteosarcoma and chondrosarcoma. This study interrogated the preresection biopsies of a cohort of osteosarcoma and chondrosarcoma with SATB2 and SOX9 in tandem, to assess their value as diagnostic adjuncts as well as their concordance with the final resection diagnoses. SATB2 was expressed more frequently in osteosarcoma (46/53, 86%) than in chondrosarcoma (9/18, 50%); SOX9 was expressed in high frequencies in both osteosarcoma (52/53, 98%) and chondrosarcoma (17/18, 94%), and SATB2 and SOX9 were coexpressed in both osteosarcoma (46/53, 89%) and chondrosarcoma (8/18, 44%). There exists significant overlap in the expression of SATB2 and SOX9 in osteosarcoma and chondrosarcoma. These markers are not expressed in a distribution that is unique enough for application toward this particular diagnostic differential.

Background: Desmoplastic small round cell tumors (DSRCT) are malignant neoplasms of young males arising most commonly in the abdominopelvic cavity, with a subset originating from extra-abdominal soft tissues. As either primary or metastatic lesions, they are rare in intraosseous sites. Case Presentation: We describe the fifth report of primary DSRCT of bone. A healthy 18-year old male presented with a blastic, 17 cm lesion within the left distal femur, suspicious for osteosarcoma or Ewing sarcoma. Subsequent biopsy revealed nests of small round blue cells infiltrating through a desmoplastic stroma. These cells were diffusely positive for epithelial markers, with paranuclear staining for desmin and focal reactivity with NSE. Break-apart FISH revealed a rearrangement in EWSR1, and RNA fusion panel confirmed WT1 as its partner in the pathognomonic t(11;22)(p13;q12) rearrangement. PET/CT showed widespread metastatic disease to visceral and bony sites. Conclusions: Due to their rarity as well as clinicopathologic and immunomorphologic overlap, primary intraosseous DSRCT can create diagnostic challenges with the more frequently encountered tumors of bone.

BackgroundTumor-infiltrating CD8+ T cells and neoantigens are predictors of a favorable prognosis and response to immunotherapy with checkpoint inhibitors in many types of adult cancer, but little is known about their role in pediatric malignancies. Here, we analyzed the prognostic strength of T cell-inflamed gene expression and neoantigen load in high-risk neuroblastoma. We also compared transcriptional programs in T cell-inflamed and non-T cell-inflamed high-risk neuroblastomas to investigate possible mechanisms of immune exclusion.MethodsA defined T cell-inflamed gene expression signature was used to categorize high-risk neuroblastomas in the Therapeutically Applicable Research to Generate Effective Treatments (TARGET) program (n=123), and the Gabriella Miller Kids First (GMKF) program (n=48) into T cell-inflamed, non-T cell-inflamed, and intermediate groups. Associations between the T cell-inflamed and non-T cell-inflamed group, MYCN amplification, and survival were analyzed by Cox proportional hazards models. Additional survival analysis was conducted after integrating neoantigen load predicted from somatic mutations. Pathways activated in non-T cell-inflamed relative to T cell-inflamed tumors were analyzed using causal network analysis.ResultsPatients with T cell-inflamed high-risk tumors showed improved overall survival compared with those with non-T cell-inflamed tumors (pMYCN amplification status, in both TARGET and GMKF cohorts. Higher neoantigen load was also associated with better event-free and overall survival (pMYCN, ASCL1, SOX11, and KMT2A transcriptional programs was inversely correlated with the T cell-inflamed signature in both cohorts.ConclusionsOur results indicate that tumors from children with high-risk neuroblastoma harboring a strong T cell-inflamed signature have a more favorable clinical outcome, and neoantigen load is a prognosis predictor, independent of T cell inflammation. Strategies to target SOX11 and other signaling pathways associated with non-T cell-inflamed tumors should be pursued as potential immune-potentiating interventions.

Objectives Large gene panel next-generation sequencing (NGS) is a powerful tool capable of generating predictive data on cancer prognosis and response to specific therapeutic interventions. The utility of large panel NGS data on tumor classification, however, may be underappreciated because of a workflow that often circumvents the surgical pathologist. We sought to describe cases in which NGS data lead to an unanticipated change in tumor classification and to discuss current workflow practices of NGS testing that limit its use as a diagnostic adjunct. Methods We performed a retrospective review to identify cases in which NGS testing uncovered data that led to a revision of the initial pathologic diagnosis that an outside or in-house pathologist had made. Results Nine cases are presented in which NGS data provided insights that led to a revision of the original pathologic diagnosis. Distinctive molecular signatures, mutational signatures, fusions, or identification of viral sequencing provided the critical evidence on which these tumors were reclassified. Conclusions The current workflow of NGS testing should always include the surgical pathologist as an active partner to ensure that the molecular results are fully reflected in the final diagnosis. In some instances, active participation by the surgical pathologist may require amendment of previously issued pathology reports.

Background Differential diagnosis of sellar masses includes adenoma, meningioma, craniopharyngioma, and metastasis. Primary paraganglioma is seldom considered. We present a case of this unique pathology, review the relevant literature, and compile a compendium of immunohistochemical characteristics for use as a resource. Case Description A 73-year-old woman presented to the hospital with visual changes in her left hemifield. Noncontrast head computed tomography demonstrated a large sellar mass with suprasellar extension and displacement of the optic chiasm (diameter of 3.1 cm). Magnetic resonance imaging was unobtainable owing to an incompatible pacemaker. Computed tomography characterization was most consistent with a macroadenoma. Given the acute visual decline, surgical decompression via an endonasal transsphenoidal route was performed without complication. A diagnosis of paraganglioma was made based on histopathology. Following resection, the patient's visual field deficit improved. Computed tomography body imaging was negative for a metastatic origin. Conclusions Paraganglioma is a rare but potential differential diagnosis to consider when evaluating sellar masses.

Background and Purpose— Previously, murine models Krit1 +/− Msh2 −/ − and Ccm2 +/ − Trp53 −/ − showed a reduction or no effect on cerebral cavernous malformation (CCM) burden and favorable effects on lesional hemorrhage by the robust Rock (Rho-associated protein kinase) inhibitor fasudil and by simvastatin (a weak pleiotropic inhibitor of Rock). Herein, we concurrently investigated treatment of the more aggressive Pdcd10/Ccm3 model with fasudil, simvastatin, and higher dose atorvastatin to determined effectiveness of Rock inhibition. Methods— The murine models, Pdcd10 +/ − Trp53 −/ − and Pdcd10 +/ − Msh2 −/ − , were contemporaneously treated from weaning to 5 months of age with fasudil (100 mg/kg per day in drinking water, n=9), simvastatin (40 mg/kg per day in chow, n=11), atorvastatin (80 mg/kg per day in chow, n=10), or with placebo (n=16). We assessed CCM volume in mouse brains by microcomputed tomography. Lesion burden was calculated as lesion volume normalized to total brain volume. We analyzed chronic hemorrhage in CCM lesions by quantitative intensity of Perls staining in brain sections. Results— The Pdcd10 +/ − Trp53 −/ − /Msh2 −/ − models showed a mean CCM lesion burden per mouse reduction from 0.0091 in placebos to 0.0042 ( P =0.027) by fasudil, and to 0.0047 ( P =0.025) by atorvastatin treatment, but was not changed significantly by simvastatin. Hemorrhage intensity per brain was commensurately decreased by Rock inhibition. Conclusions— These results support the exploration of proof of concept effect of high-dose atorvastatin on human CCM disease for potential therapeutic testing.

The 2016 WHO classification of brain tumors represents a major step towards the integration of molecular data into pathologic diagnoses. Our institution has included massively parallel sequencing technology in the diagnostic work-up of all gliomas since January 2016. The utilized platform successfully identifies copy number variations, individual gene mutations, small insertions and deletions, and selected gene fusions. Herein, we retrospectively review the first 51 glial tumor samples run for clinical purposes using the UCM-OncoPlus platform, a 1213 gene targeted hybrid-capture next generation sequencing (NGS) panel. NGS paired with histomorphology and clinical data allowed for reliable, comprehensive, and cost-effective classification of all the analyzed gliomas (51/51) with minimal tissue required and without the need for additional testing. In addition to detecting all diagnostically relevant mutations according to the 2016 WHO system, our data suggest a large NGS-based platform may improve the accuracy of classifying gliomas beyond the 2016 WHO system, to provide truly personalized diagnostics. Furthermore, this methodology assists in classifying histologically challenging or clinically unusual cases. And, finally, the versatile nature of this testing methodology allows for near effortless expansion as new therapeutic targets and prognostic markers are discovered.