Fabio Parisi

Tumour response and secondary resectability of colorectal liver metastases following neoadjuvant chemotherapy with cetuximab: the CELIM randomised phase 2 trial.

BACKGROUND Neoadjuvant chemotherapy for unresectable colorectal liver metastases can downsize tumours for curative resection. We assessed the effectiveness of cetuximab combined with chemotherapy in this setting. METHODS Between Dec 2, 2004, and March 27, 2008, 114 patients were enrolled from 17 centres in Germany and Austria; three patients receiving FOLFOX6 alone were excluded from the analysis. Patients with non-resectable liver metastases (technically non-resectable or > or =5 metastases) were randomly assigned to receive cetuximab with either FOLFOX6 (oxaliplatin, fluorouracil, and folinic acid; group A) or FOLFIRI (irinotecan, fluorouracil, and folinic acid; group B). Randomisation was not blinded, and was stratified by technical resectability and number of metastases, use of PET staging, and EGFR expression status. They were assessed for response every 8 weeks by CT or MRI. A local multidisciplinary team reassessed resectability after 16 weeks, and then every 2 months up to 2 years. Patients with resectable disease were offered liver surgery within 4-6 weeks of the last treatment cycle. The primary endpoint was tumour response assessed by Response Evaluation Criteria In Solid Tumours (RECIST), analysed by modified intention to treat. A retrospective, blinded surgical review of patients with radiological images at both baseline and during treatment was done to assess objectively any changes in resectability. The study is registered with ClinicalTrials.gov, number NCT00153998. FINDINGS 56 patients were randomly assigned to group A and 55 to group B. One patient in each group were excluded from the analysis of the primary endpoint because they discontinued treatment before first full dose, one patient in group B was excluded because of early pulmonary embolism. A confirmed partial or complete response was noted in 36 (68%) of 53 patients in group A, and 30 (57%) of 53 patients in group B (difference 11%, 95% CI -8 to 30; odds ratio [OR] 1.62, 0.74-3.59; p=0.23). The most frequent grade 3 and 4 toxicities were skin toxicity (15 of 54 patients in group A, and 22 of 55 patients in group B), and neutropenia (13 of 54 patients in group A and 12 of 55 patients in group B). R0 resection was done in 20 (38%) of 53 patients in group A and 16 (30%) of 53 of patients in group B. In a retrospective analysis of response by KRAS status, a partial or complete response was noted in 47 (70%) of 67 patients with KRAS wild-type tumours versus 11 (41%) of 27 patients with KRAS-mutated tumours (OR 3.42, 1.35-8.66; p=0.0080). According to the retrospective review, resectability rates increased from 32% (22 of 68 patients) at baseline to 60% (41 of 68) after chemotherapy (p<0.0001). INTERPRETATION Chemotherapy with cetuximab yields high response rates compared with historical controls, and leads to significantly increased resectability. FUNDING Merck-Serono, Sanofi-Aventis, and Pfizer.

PCGF Homologs, CBX Proteins, and RYBP Define Functionally Distinct PRC1 Family Complexes

The heterogeneous nature of mammalian PRC1 complexes has hindered our understanding of their biological functions. Here, we present a comprehensive proteomic and genomic analysis that uncovered six major groups of PRC1 complexes, each containing a distinct PCGF subunit, a RING1A/B ubiquitin ligase, and a unique set of associated polypeptides. These PRC1 complexes differ in their genomic localization, and only a small subset colocalize with H3K27me3. Further biochemical dissection revealed that the six PCGF-RING1A/B combinations form multiple complexes through association with RYBP or its homolog YAF2, which prevents the incorporation of other canonical PRC1 subunits, such as CBX, PHC, and SCM. Although both RYBP/YAF2- and CBX/PHC/SCM-containing complexes compact chromatin, only RYBP stimulates the activity of RING1B toward H2AK119ub1, suggesting a central role in PRC1 function. Knockdown of RYBP in embryonic stem cells compromised their ability to form embryoid bodies, likely because of defects in cell proliferation and maintenance of H2AK119ub1 levels.

PLX4032, a selective BRAFV600E kinase inhibitor, activates the ERK pathway and enhances cell migration and proliferation of BRAFWT melanoma cells

BRAFV600E/K is a frequent mutationally active tumor‐specific kinase in melanomas that is currently targeted for therapy by the specific inhibitor PLX4032. Our studies with melanoma tumor cells that are BRAFV600E/K and BRAFWT showed that, paradoxically, while PLX4032 inhibited ERK1/2 in the highly sensitive BRAFV600E/K, it activated the pathway in the resistant BRAFWT cells, via RAF1 activation, regardless of the status of mutations in NRAS or PTEN. The persistently active ERK1/2 triggered downstream effectors in BRAFWT melanoma cells and induced changes in the expression of a wide‐spectrum of genes associated with cell cycle control. Furthermore, PLX4032 increased the rate of proliferation of growth factor‐dependent NRAS Q61L mutant primary melanoma cells, reduced cell adherence and increased mobility of cells from advanced lesions. The results suggest that the drug can confer an advantage to BRAFWT primary and metastatic tumor cells in vivo and provide markers for monitoring clinical responses.

Genome-wide remodeling of the epigenetic landscape during myogenic differentiation

We have examined changes in the chromatin landscape during muscle differentiation by mapping the genome-wide location of ten key histone marks and transcription factors in mouse myoblasts and terminally differentiated myotubes, providing an exceptionally rich dataset that has enabled discovery of key epigenetic changes underlying myogenesis. Using this compendium, we focused on a well-known repressive mark, histone H3 lysine 27 trimethylation, and identified novel regulatory elements flanking the myogenin gene that function as a key differentiation-dependent switch during myogenesis. Next, we examined the role of Polycomb-mediated H3K27 methylation in gene repression by systematically ablating components of both PRC1 and PRC2 complexes. Surprisingly, we found mechanistic differences between transient and permanent repression of muscle differentiation and lineage commitment genes and observed that the loss of PRC1 and PRC2 components produced opposing differentiation defects. These phenotypes illustrate striking differences as compared to embryonic stem cell differentiation and suggest that PRC1 and PRC2 do not operate sequentially in muscle cells. Our studies of PRC1 occupancy also suggested a “fail-safe” mechanism, whereby PRC1/Bmi1 concentrates at genes specifying nonmuscle lineages, helping to retain H3K27me3 in the face of declining Ezh2-mediated methyltransferase activity in differentiated cells.

Inhibition of Nonsense-Mediated RNA Decay by the Tumor Microenvironment Promotes Tumorigenesis

ABSTRACT While nonsense-mediated RNA decay (NMD) is an established mechanism to rapidly degrade select transcripts, the physiological regulation and biological significance of NMD are not well characterized. We previously demonstrated that NMD is inhibited in hypoxic cells. Here we show that the phosphorylation of the α subunit of eukaryotic initiation factor 2 (eIF2α) translation initiation factor by a variety of cellular stresses leads to the inhibition of NMD and that eIF2α phosphorylation and NMD inhibition occur in tumors. To explore the significance of this NMD regulation, we used an unbiased approach to identify approximately 750 NMD-targeted mRNAs and found that these mRNAs are overrepresented in stress response and tumor-promoting pathways. Consistent with these findings, the inhibition of NMD promotes cellular resistance to endoplasmic reticulum stress and encourages tumor formation. The transcriptional and translational regulations of gene expression by the microenvironment are established mechanisms by which tumor cells adapt to stress. These data indicate that NMD inhibition by the tumor microenvironment is also an important mechanism to dynamically regulate genes critical for the response to cellular stress and tumorigenesis.

Quantitative Assessment of Effect of Preanalytic Cold Ischemic Time on Protein Expression in Breast Cancer Tissues

BACKGROUND Companion diagnostic tests can depend on accurate measurement of protein expression in tissues. Preanalytic variables, especially cold ischemic time (time from tissue removal to fixation in formalin) can affect the measurement and may cause false-negative results. We examined 23 proteins, including four commonly used breast cancer biomarker proteins, to quantify their sensitivity to cold ischemia in breast cancer tissues. METHODS A series of 93 breast cancer specimens with known time-to-fixation represented in a tissue microarray and a second series of 25 matched pairs of core needle biopsies and breast cancer resections were used to evaluate changes in antigenicity as a function of cold ischemic time. Estrogen receptor (ER), progesterone receptor (PgR), HER2 or Ki67, and 19 other antigens were tested. Each antigen was measured using the AQUA method of quantitative immunofluorescence on at least one series. All statistical tests were two-sided. RESULTS We found no evidence for loss of antigenicity with time-to-fixation for ER, PgR, HER2, or Ki67 in a 4-hour time window. However, with a bootstrapping analysis, we observed a trend toward loss for ER and PgR, a statistically significant loss of antigenicity for phosphorylated tyrosine (P = .0048), and trends toward loss for other proteins. There was evidence of increased antigenicity in acetylated lysine, AKAP13 (P = .009), and HIF1A (P = .046), which are proteins known to be expressed in conditions of hypoxia. The loss of antigenicity for phosphorylated tyrosine and increase in expression of AKAP13, and HIF1A were confirmed in the biopsy/resection series. CONCLUSIONS Key breast cancer biomarkers show no evidence of loss of antigenicity, although this dataset assesses the relatively short time beyond the 1-hour limit in recent guidelines. Other proteins show changes in antigenicity in both directions. Future studies that extend the time range and normalize for heterogeneity will provide more comprehensive information on preanalytic variation due to cold ischemic time.

Regulation of Glutamine Carrier Proteins by RNF5 Determines Breast Cancer Response to ER Stress-Inducing Chemotherapies

Many tumor cells are fueled by altered metabolism and increased glutamine (Gln) dependence. We identify regulation of the L-glutamine carrier proteins SLC1A5 and SLC38A2 (SLC1A5/38A2) by the ubiquitin ligase RNF5. Paclitaxel-induced ER stress to breast cancer (BCa) cells promotes RNF5 association, ubiquitination, and degradation of SLC1A5/38A2. This decreases Gln uptake, levels of TCA cycle components, mTOR signaling, and proliferation while increasing autophagy and cell death. Rnf5-deficient MMTV-PyMT mammary tumors were less differentiated and showed elevated SLC1A5 expression. Whereas RNF5 depletion in MDA-MB-231 cells promoted tumorigenesis and abolished paclitaxel responsiveness, SLC1A5/38A2 knockdown elicited opposing effects. Inverse RNF5(hi)/SLC1A5/38A2(lo) expression was associated with positive prognosis in BCa. Thus, RNF5 control of Gln uptake underlies BCa response to chemotherapies.

TrAp: a tree approach for fingerprinting subclonal tumor composition

Revealing the clonal composition of a single tumor is essential for identifying cell subpopulations with metastatic potential in primary tumors or with resistance to therapies in metastatic tumors. Sequencing technologies provide only an overview of the aggregate of numerous cells. Computational approaches to de-mix a collective signal composed of the aberrations of a mixed cell population of a tumor sample into its individual components are not available. We propose an evolutionary framework for deconvolving data from a single genome-wide experiment to infer the composition, abundance and evolutionary paths of the underlying cell subpopulations of a tumor. We have developed an algorithm (TrAp) for solving this mixture problem. In silico analyses show that TrAp correctly deconvolves mixed subpopulations when the number of subpopulations and the measurement errors are moderate. We demonstrate the applicability of the method using tumor karyotypes and somatic hypermutation data sets. We applied TrAp to Exome-Seq experiment of a renal cell carcinoma tumor sample and compared the mutational profile of the inferred subpopulations to the mutational profiles of single cells of the same tumor. Finally, we deconvolve sequencing data from eight acute myeloid leukemia patients and three distinct metastases of one melanoma patient to exhibit the evolutionary relationships of their subpopulations.

Picking ChIP-seq peak detectors for analyzing chromatin modification experiments

Numerous algorithms have been developed to analyze ChIP-Seq data. However, the complexity of analyzing diverse patterns of ChIP-Seq signals, especially for epigenetic marks, still calls for the development of new algorithms and objective comparisons of existing methods. We developed Qeseq, an algorithm to detect regions of increased ChIP read density relative to background. Qeseq employs critical novel elements, such as iterative recalibration and neighbor joining of reads to identify enriched regions of any length. To objectively assess its performance relative to other 14 ChIP-Seq peak finders, we designed a novel protocol based on Validation Discriminant Analysis (VDA) to optimally select validation sites and generated two validation datasets, which are the most comprehensive to date for algorithmic benchmarking of key epigenetic marks. In addition, we systematically explored a total of 315 diverse parameter configurations from these algorithms and found that typically optimal parameters in one dataset do not generalize to other datasets. Nevertheless, default parameters show the most stable performance, suggesting that they should be used. This study also provides a reproducible and generalizable methodology for unbiased comparative analysis of high-throughput sequencing tools that can facilitate future algorithmic development.

Ranking and combining multiple predictors without labeled data

Significance A key challenge in a broad range of decision-making and classification problems is how to rank and combine the possibly conflicting suggestions of several advisers of unknown reliability. We provide mathematical insights of striking conceptual simplicity that explain mutual relationships between independent advisers. These insights enable the design of efficient, robust, and reliable methods to rank the advisers’ performances and construct improved predictions in the absence of ground truth. Furthermore, these methods are robust to the presence of small subgroups of malicious advisers (cartels) attempting to veer the combined decisions to their interest. In a broad range of classification and decision-making problems, one is given the advice or predictions of several classifiers, of unknown reliability, over multiple questions or queries. This scenario is different from the standard supervised setting, where each classifier’s accuracy can be assessed using available labeled data, and raises two questions: Given only the predictions of several classifiers over a large set of unlabeled test data, is it possible to (i) reliably rank them and (ii) construct a metaclassifier more accurate than most classifiers in the ensemble? Here we present a spectral approach to address these questions. First, assuming conditional independence between classifiers, we show that the off-diagonal entries of their covariance matrix correspond to a rank-one matrix. Moreover, the classifiers can be ranked using the leading eigenvector of this covariance matrix, because its entries are proportional to their balanced accuracies. Second, via a linear approximation to the maximum likelihood estimator, we derive the Spectral Meta-Learner (SML), an unsupervised ensemble classifier whose weights are equal to these eigenvector entries. On both simulated and real data, SML typically achieves a higher accuracy than most classifiers in the ensemble and can provide a better starting point than majority voting for estimating the maximum likelihood solution. Furthermore, SML is robust to the presence of small malicious groups of classifiers designed to veer the ensemble prediction away from the (unknown) ground truth.