Jeffrey S. Morris

The consensus molecular subtypes of colorectal cancer

Colorectal cancer (CRC) is a frequently lethal disease with heterogeneous outcomes and drug responses. To resolve inconsistencies among the reported gene expression–based CRC classifications and facilitate clinical translation, we formed an international consortium dedicated to large-scale data sharing and analytics across expert groups. We show marked interconnectivity between six independent classification systems coalescing into four consensus molecular subtypes (CMSs) with distinguishing features: CMS1 (microsatellite instability immune, 14%), hypermutated, microsatellite unstable and strong immune activation; CMS2 (canonical, 37%), epithelial, marked WNT and MYC signaling activation; CMS3 (metabolic, 13%), epithelial and evident metabolic dysregulation; and CMS4 (mesenchymal, 23%), prominent transforming growth factor–β activation, stromal invasion and angiogenesis. Samples with mixed features (13%) possibly represent a transition phenotype or intratumoral heterogeneity. We consider the CMS groups the most robust classification system currently available for CRC—with clear biological interpretability—and the basis for future clinical stratification and subtype-based targeted interventions.

Reproducibility of SELDI-TOF protein patterns in serum: comparing datasets from different experiments

MOTIVATION There has been much interest in using patterns derived from surface-enhanced laser desorption and ionization (SELDI) protein mass spectra from serum to differentiate samples from patients both with and without disease. Such patterns have been used without identification of the underlying proteins responsible. However, there are questions as to the stability of this procedure over multiple experiments. RESULTS We compared SELDI proteomic spectra from serum from three experiments by the same group on separating ovarian cancer from normal tissue. These spectra are available on the web at http://clinicalproteomics.steem.com. In general, the results were not reproducible across experiments. Baseline correction prevents reproduction of the results for two of the experiments. In one experiment, there is evidence of a major shift in protocol mid-experiment which could bias the results. In another, structure in the noise regions of the spectra allows us to distinguish normal from cancer, suggesting that the normals and cancers were processed differently. Sets of features found to discriminate well in one experiment do not generalize to other experiments. Finally, the mass calibration in all three experiments appears suspect. Taken together, these and other concerns suggest that much of the structure uncovered in these experiments could be due to artifacts of sample processing, not to the underlying biology of cancer. We provide some guidelines for design and analysis in experiments like these to ensure better reproducible, biologically meaningfully results. AVAILABILITY The MATLAB and Perl code used in our analyses is available at http://bioinformatics.mdanderson.org

Phase II Trial of Infusional Fluorouracil, Irinotecan, and Bevacizumab for Metastatic Colorectal Cancer: Efficacy and Circulating Angiogenic Biomarkers Associated With Therapeutic Resistance

PURPOSE We investigated the efficacy of fluorouracil (FU), leucovorin, irinotecan, and bevacizumab (FOLFIRI + B) in a phase II trial in patients previously untreated for metastatic colorectal cancer (mCRC), and changes during treatment in plasma cytokines and angiogenic factors (CAFs) as potential markers of treatment response and therapeutic resistance. PATIENTS AND METHODS We conducted a phase II, two-institution trial of FOLFIRI + B. Each 14-day cycle consisted of bevacizumab (5 mg/kg), irinotecan (180 mg/m(2)), bolus FU (400 mg/m(2)), and leucovorin (400 mg/m(2)) followed by a 46-hour infusion of FU (2,400 mg/m(2)). Levels of 37 CAFs were assessed using multiplex-bead assays and enzyme-linked immunosorbent assay at baseline, during treatment, and at the time of progressive disease (PD). RESULTS Forty-three patients were enrolled. Median progression-free survival (PFS), the primary end point of the study, was 12.8 months. Median overall survival was 31.3 months, with a response rate of 65%. Elevated interleukin-8 at baseline was associated with a shorter PFS (11 v 15.1 months, P = .03). Before the radiographic development of PD, several CAFs associated with angiogenesis and myeloid recruitment increased compared to baseline, including basic fibroblast growth factor (P = .046), hepatocyte growth factor (P = .046), placental growth factor (P < .001), stromal-derived factor-1 (P = .04), and macrophage chemoattractant protein-3 (P < .001). CONCLUSION Efficacy and tolerability of FOLFIRI + B appeared favorable to historical controls in this single arm study. Before radiographic progression, there was a shift in balance of CAFs, with a rise in alternate pro-angiogenic cytokines and myeloid recruitment factors in subsets of patients that may represent mechanisms of resistance.

Phase II trial of the combination of bevacizumab and erlotinib in patients who have advanced hepatocellular carcinoma

PURPOSE The study objective was to determine the proportion of patients with hepatocellular carcinoma (HCC) treated with the combination of bevacizumab (B) and erlotinib (E) who were alive and progression free at 16 weeks (16-week progression-free survival [PFS16]) of continuous therapy. Secondary objectives included response rate, median PFS, survival, and toxicity. PATIENTS AND METHODS Patients who had advanced HCC that was not amenable to surgical or regional therapies, up to one prior systemic treatment; Childs-Pugh score A or B liver function; Eastern Cooperative Oncology Group performance status 0, 1, or 2 received B 10 mg/kg every 14 days and E 150 mg orally daily, continuously, for 28-day cycles. Tumor response was evaluated every 2 cycles by using Response Evaluation Criteria in Solid Tumors Group criteria. A total of 40 patients were treated. RESULTS The primary end point of PFS16 was 62.5%. Ten patients achieved a partial response for a confirmed overall response rate (intent-to-treat) of 25%. The median PFSevent was 39 weeks (95% CI, 26 to 45 weeks; 9.0 months), and the median overall survival was 68 weeks (95% CI, 48 to 78 weeks; 15.65 months). Grades 3 to 4 drug-related toxicity included fatigue (n = 8; 20%), hypertension (n = 6; 15%), diarrhea (n = 4; 10%) elevated transaminases (n = 4; 10%), gastrointestinal hemorrhage (n = 5; 12.5%), wound infection (n = 2; 5%) thrombocytopenia (n = 1; 2.5%), and proteinuria, hyperbilirubinemia, back pain, hyperkalemia, and anorexia (n = 1 each). CONCLUSION The combination of B + E in patients who had advanced HCC showed significant, clinically meaningful antitumor activity. B + E warrant additional evaluation in randomized controlled trials.

Feature extraction and quantification for mass spectrometry in biomedical applications using the mean spectrum

MOTIVATION Mass spectrometry yields complex functional data for which the features of scientific interest are peaks. A common two-step approach to analyzing these data involves first extracting and quantifying the peaks, then analyzing the resulting matrix of peak quantifications. Feature extraction and quantification involves a number of interrelated steps. It is important to perform these steps well, since subsequent analyses condition on these determinations. Also, it is difficult to compare the performance of competing methods for analyzing mass spectrometry data since the true expression levels of the proteins in the population are generally not known. RESULTS In this paper, we introduce a new method for feature extraction in mass spectrometry data that uses translation-invariant wavelet transforms and performs peak detection using the mean spectrum. We examine the methods performance through examples and simulation, and demonstrate the advantages of using the mean spectrum to detect peaks. We also describe a new physics-based computer model of mass spectrometry and demonstrate how one may design simulation studies based on this tool to systematically compare competing methods. AVAILABILITY MATLAB scripts to implement the methods described in this paper and R code for the virtual mass spectrometer are available at http://bioinformatics.mdanderson.org/software.html SUPPLEMENTARY INFORMATION http://bioinformatics.mdanderson.org/supplements.html.

Phase 2 study of erlotinib in patients with unresectable hepatocellular carcinoma

Growth factor overexpression, including epidermal growth factor receptor (EGFR) expression, is common in hepatocellular cancers. Erlotinib is a receptor tyrosine kinase inhibitor with specificity for EGFR. The primary objective of this study was to determine the proportion of hepatocellular carcinoma (HCC) patients treated with erlotinib who were alive and progression‐free (PFS) at 16 weeks of continuous treatment.

Quality Control and Peak Finding for Proteomics Data Collected from Nipple Aspirate Fluid by Surface-Enhanced Laser Desorption and Ionization

BACKGROUND Recently, researchers have been using mass spectroscopy to study cancer. For use of proteomics spectra in a clinical setting, stringent quality-control procedures will be needed. METHODS We pooled samples of nipple aspirate fluid from healthy breasts and breasts with cancer to prepare a control sample. Aliquots of the control sample were used on two spots on each of three IMAC ProteinChip arrays (Ciphergen Biosystems, Inc.) on 4 successive days to generate 24 SELDI spectra. In 36 subsequent experiments, the control sample was applied to two spots of each ProteinChip array, and the resulting spectra were analyzed to determine how closely they agreed with the original 24 spectra. RESULTS We describe novel algorithms that (a) locate peaks in unprocessed proteomics spectra and (b) iteratively combine peak detection with baseline correction. These algorithms detected approximately 200 peaks per spectrum, 68 of which are detected in all 24 original spectra. The peaks were highly correlated across samples. Moreover, we could explain 80% of the variance, using only six principal components. Using a criterion that rejects a chip if the Mahalanobis distance from both control spectra to the center of the six-dimensional principal component space exceeds the 95% confidence limit threshold, we rejected 5 of the 36 chips. CONCLUSIONS Mahalanobis distance in principal component space provides a method for assessing the reproducibility of proteomics spectra that is robust, effective, easily computed, and statistically sound.

The 15-lipoxygenase-1 product 13-S-hydroxyoctadecadienoic acid down-regulates PPAR-δ to induce apoptosis in colorectal cancer cells

Diminished apoptosis, a critical event in tumorigenesis, is linked to down-regulated 15-lipoxygenase-1 (15-LOX-1) expression in colorectal cancer cells. 13-S-hydroxyoctadecadienoic acid (13-S-HODE), which is the primary product of 15-LOX-1 metabolism of linoleic acid, restores apoptosis. Nonsteroidal antiinflammatory drugs (NSAIDs) transcriptionally up-regulate 15-LOX-1 expression to induce apoptosis. Peroxisome proliferator-activated receptors (PPARs) are nuclear receptors for linoleic and arachidonic acid metabolites. PPAR-δ promotes colonic tumorigenesis. NSAIDs suppress PPAR-δ activity in colon cancer cells. The mechanistic relationship between 15-LOX-1 and PPAR-δ was previously unknown. Our current study shows that (i) 13-S-HODE binds to PPAR-δ, decreases PPAR-δ activation, and down-regulates PPAR-δ expression in colorectal cancer cells; (ii) the induction of 15-LOX-1 expression is a critical step in NSAID down-regulation of PPAR-δ and the resultant induction of apoptosis; and (iii) PPAR-δ is an important signaling receptor for 13-S-HODE-induced apoptosis. The in vivo relevance of these mechanistic findings was demonstrated in our tumorigenesis studies in nude mouse xenograft models. Our findings indicate that the down-regulation of PPAR-δ by 15-LOX-1 through 13-S-HODE is an apoptotic signaling pathway that is activated by NSAIDs.

CpG Island Methylation in Colorectal Adenomas

Methylation of cytosines in CpG islands silences gene expression. CpG island methylator phenotype (CIMP) in colorectal cancers is characterized by abnormal methylation of multiple CpG islands including those in several tumor suppressor genes such as p16, hMLH1, and THBS1. CpG island methylation has not been well characterized in adenomas. We evaluated methylation status at p16, MINT2, and MINT31 loci, which are frequently methylated in colorectal carcinomas, in 108 colorectal adenomas from a prospective study of 50 patients without cancer. Methylation at one or more loci was present in 48% (52 of 108) of adenomas with 25% (19 of 76) CIMP-high (two or more methylated loci) and 32% (24 of 76) CIMP-low (one methylated locus). The p16 gene was methylated in 27% (19 of 71) of adenomas. Methylation status of different adenomas from the same patient was not correlated (odds ratio, 0.93; P = 0.77). Adenomas with tubulovillous or villous histology were frequently methylated: 73% (17 of 26) versus 41% (35 of 85) of tubular adenomas (odds ratio, 3.46; P = 0.02). High levels of microsatellite instability were more frequent in adenomas without methylation (13% versus 2%; odds ratio, 8.48; P = 0.05). Our results indicate that methylation plays an important role early in colorectal tumorigenesis. CpG island methylation is more common in adenomas with tubulovillous/villous histology, a characteristic associated with more frequent predisposition to invasive carcinoma. Methylation is distinct from microsatellite instability and develops in individual adenomas rather than resulting from a field defect in an individual patient.

Concordant CpG island methylation in hyperplastic polyposis.

The CpG island methylator phenotype (CIMP) is a newly described mechanism for carcinogenesis in colorectal carcinomas and adenomas characterized by methylation of multiple CpG islands. The causes of CIMP are unknown. We studied CIMP in hyperplastic polyps (HPs), with emphasis on patients with multiple HPs (5 to 10 HPs), large HPs (one HP >1 cm) or hyperplastic polyposis (>20 HPs). Methylation of p16, MINT1, MINT2, MINT31, and hMLH1 was analyzed by methylation-specific polymerase chain reaction in 102 HPs, 8 serrated adenomas, 19 tubular adenomas, and 9 adenocarcinomas from 17 patients, with multiple/large HPs or hyperplastic polyposis and in 16 sporadic HPs from 14 additional patients. Sporadic HPs were CIMP-negative (not methylated at any locus), but 43% of HPs from multiple/large HPs, or hyperplastic polyposis were CIMP-high (two or more methylated loci, P = 0.00001). Methylation among the four loci was correlated within HPs (odds ratio, 3.41; P = 0.002), and the methylation status of HPs within the same patient was also correlated (odds ratio, 5.92; P = 0.0001). CIMP-high HPs were present primarily in patients with a predominance of HPs in the right colon and/or serrated adenomas (P = 0.0009) and were associated with the absence of K-ras proto-oncogene mutations (odds ratio, 5.08; P = 0.03). Our findings of concordant CpG island methylation of HPs in multiple/large HPs or hyperplastic polyposis supports the concept that some patients have a hypermethylator phenotype characterized by methylation of multiple HPs and other colorectal lesions. The hypermethylator phenotype is related to patient-specific factors, such as carcinogenic exposure or genetic predisposition.