Predicting likelihood of in vivo chemotherapy response in canine lymphoma using ex vivo drug sensitivity and immunophenotyping data in a machine learning model.

We report a precision medicine platform that evaluates the probability of chemotherapy drug efficacy for canine lymphoma by combining ex vivo chemosensitivity and immunophenotyping assays with computational modelling. We isolated live cancer cells from fresh fine needle aspirates of affected lymph nodes and collected post-treatment clinical responses in 261 canine lymphoma patients scheduled to receive at least 1 of 5 common chemotherapy agents (doxorubicin, vincristine, cyclophosphamide, lomustine and rabacfosadine). We used flow cytometry analysis for immunophenotyping and ex vivo chemosensitivity testing. For each drug, 70% of treated patients were randomly selected to train a random forest model to predict the probability of positive Veterinary Cooperative Oncology Group (VCOG) clinical response based on input variables including antigen expression profiles and treatment sensitivity readouts for each patient's cancer cells. The remaining 30% of patients were used to test model performance. Most models showed a test set ROC-AUC > 0.65, and all models had overall ROC-AUC > 0.95. Predicted response scores significantly distinguished (P < .001) positive responses from negative responses in B-cell and T-cell disease and newly diagnosed and relapsed patients. Patient groups with predicted response scores >50% showed a statistically significant reduction (log-rank P < .05) in time to complete response when compared to the groups with scores <50%. The computational models developed in this study enabled the conversion of ex vivo cell-based chemosensitivity assay results into a predicted probability of in vivo therapeutic efficacy, which may help improve treatment outcomes of individual canine lymphoma patients by providing predictive estimates of positive treatment response.

Down-regulation of EZH2 expression in myelodysplastic syndromes.

EZH2 genetic mutations are common in myelodysplastic syndrome (MDS), which implies that this gene has a pathophysiological role in the disease. To further characterize molecular alterations of EZH2, and their potential prognostic impact in MDS, we assessed EZH2 RNA expression in primary bone marrow CD34+ cells from 78 patients. We found that 47% of patients have reduced EZH2 expression compared to normal controls. Further analyses revealed that EZH2 is significantly underexpressed in patients bearing chromosome 7 or 7q deletions (7-alt) when compared to controls, diploid patients, and patients with other cytogenetic alterations (p<0.05). In survival analysis, we found a non-significant trend toward overall survival (OS) being better among patients with EZH2 underexpression (median OS 55 vs. 36 months; p=0.71). Importantly, this trend became significant when the analysis was restricted to the subset of cases without alterations in chromosome 7 (62 vs. 36 months; p=0.033). Furthermore, our previous work has identified a spectrum of innate immune genes in MDS CD34+ cells that are deregulated via abnormal promoter histone methylation. Because EZH2 is a key regulator of histone methylation, we assessed the relationship between deregulation of these genes and EZH2 underexpression. We observed that the mRNA levels of 11 immune genes were higher in the EZH2 underexpression group and that immune gene expression was significantly higher in patients with concomitant EZH2 underexpression and KDM6B (also known as JMJD3, an H3K27 demethylase) overexpression. Taken together, these data indicate that EZH2 underexpression may have unique impact on the molecular pathogenesis and prognosis in MDS and be an important marker for patients without chromosome 7 alteration.

Detectable FLT3-ITD or RAS mutation at the time of transformation from MDS to AML predicts for very poor outcomes.

BACKGROUND: The molecular events that drive the transformation from myelodysplastic syndromes (MDS) to acute myeloid leukemia (AML) have yet to be fully characterized. We hypothesized that detection of these mutations at the time of transformation from MDS to AML may lead to poorer outcomes. METHODS: We analyzed 102 MDS patients who were admitted to our institution between 2004 and 2013, had wild-type (wt) FLT3-ITD and RAS at diagnosis, progressed to AML, and had serial mutation testing at both the MDS and AML stages. RESULTS: We detected FLT3-ITD and/or RAS mutations in twenty-seven (26%) patients at the time of transformation to AML. Twenty-two patients (81%) had RAS mutations and five (19%) had FLT3-ITD mutations. The median survival after leukemia transformation in patients who had detectable RAS and/or FLT3-ITD mutations was 2.4 months compared to 7.5 months in patients who retained wt RAS and FLT3-ITD (hazard ratio [HR]: 3.08, 95% confidence interval [CI]: 1.9-5.0, p<0.0001). In multivariate analysis, FLT3-ITD and RAS mutations had independent prognostic significance for poor outcome.

Discontinuation of hypomethylating agent therapy in patients with myelodysplastic syndromes or acute myelogenous leukemia in complete remission or partial response: retrospective analysis of survival after long-term follow-up.

Hypomethylating agents (HMA), such as 5-azacitidine or decitabine, are currently used to treat patients with myelodysplastic syndromes (MDS) or acute myeloid leukemia (AML) until treatment failure. However, the outcomes for patients who discontinue therapy after achieving partial response (PR) or complete remission (CR) but before treatment failure have not been reported. We present a series of 16 patients with higher-risk MDS (n=5; 31%) or AML (n=11; 69%) who achieved PR (n=1) or CR (n=15) and stopped HMA therapy while in response in the context of clinical trials. They received a median of 12 courses (range 1-24) and achieved response after a median of 1 course of therapy (1-4). Loss of response after discontinuation of therapy was rapid, with a median progression-free survival of 4 months (95% CI: 2-6). Median overall survival (OS) from the time of therapy discontinuation was 15 months (95% CI: 6-24). Patients who received 12 cycles of therapy or more had significantly better OS (median: 20 months [95% CI: 12-27]) than those who received fewer than 12 cycles (median: 4 months [95% CI: 1-8]) (p=0.043). Poor-risk cytogenetics were also associated with lower 1-year OS (33% versus 69%; p=0.046). According to these results and considering the poor prognosis after HMA failure, HMA interruption should be avoided once a sustained response has been achieved.

A phase I study of oral ARRY-614, a p38 MAPK/Tie2 dual inhibitor, in patients with low or intermediate-1 risk myelodysplastic syndromes.

PURPOSE: Data suggest that activity of p38 MAPK and Tie2 kinases is dysregulated in myelodysplastic syndromes (MDS) and may be targets for novel therapies. A phase I study of ARRY-614, an oral dual inhibitor of p38 MAPK and Tie2, was conducted in patients with low or intermediate-1 International Prognostic Scoring System risk MDS to evaluate safety, pharmacokinetics (PK), pharmacodynamics (PD), and preliminary responses by International Working Group 2006 criteria. EXPERIMENTAL DESIGN: Forty-five patients received ARRY-614 either once daily or twice daily in dose escalation (400, 600, 900, or 1,200 mg once daily; 200 or 300 mg twice daily) or expansion cohorts. RESULTS: The 300 mg twice daily schedule was not tolerated, and an MTD was not reached for once daily dosing. Treatment-related adverse events were primarily grade 1-2, with the most common being rash, diarrhea, dry skin, fatigue and anorexia. Interpatient PK variability was high, although exposure was sufficient to achieve reduction in p38 MAPK activation in bone marrow and in the levels of circulating biomarkers. Disease responses were observed in 14 of 44 (32%) evaluable patients, 13 (93%) of whom had previously been treated with a hypomethylating agent. Responses were observed in all lineages, with 5 patients experiencing bilineage responses. Three of 25 red blood cell transfusion-dependent (TD) patients achieved transfusion independence (TI) and 5 of 7 platelet TD patients achieved TI. CONCLUSIONS: ARRY-614 was well tolerated and has sufficient activity to warrant further evaluation in this patient population. We recommend 1,200 mg once daily as the optimal dose for further study.

Relative packing groups in template-based structure prediction: cooperative effects of true positive constraints.

Most current template-based structure prediction methods concentrate on finding the correct backbone conformation and then packing sidechains within that backbone. Our packing-based method derives distance constraints from conserved relative packing groups (RPGs). In our refinement approach, the RPGs provide a level of resolution that restrains global topology while allowing conformational sampling. In this study, we test our template-based structure prediction method using 51 prediction units from CASP7 experiments. RPG-based constraints are able to substantially improve approximately two-thirds of starting templates. Upon deeper investigation, we find that true positive spatial constraints, especially those non-local in sequence, derived from the RPGs were important to building nearer native models. Surprisingly, the fraction of incorrect or false positive constraints does not strongly influence the quality of the final candidate. This result indicates that our RPG-based true positive constraints sample the self-consistent, cooperative interactions of the native structure. The lack of such reinforcing cooperativity explains the weaker effect of false positive constraints. Generally, these findings are encouraging indications that RPGs will improve template-based structure prediction.

Assessing side-chain perturbations of the protein backbone: a knowledge-based classification of residue Ramachandran space.

Grouping the 20 residues is a classic strategy to discover ordered patterns and insights about the fundamental nature of proteins, their structure, and how they fold. Usually, this categorization is based on the biophysical and/or structural properties of a residue's side-chain group. We extend this approach to understand the effects of side chains on backbone conformation and to perform a knowledge-based classification of amino acids by comparing their backbone phi, psi distributions in different types of secondary structure. At this finer, more specific resolution, torsion angle data are often sparse and discontinuous (especially for nonhelical classes) even though a comprehensive set of protein structures is used. To ensure the precision of Ramachandran plot comparisons, we applied a rigorous Bayesian density estimation method that produces continuous estimates of the backbone phi, psi distributions. Based on this statistical modeling, a robust hierarchical clustering was performed using a divergence score to measure the similarity between plots. There were seven general groups based on the clusters from the complete Ramachandran data: nonpolar/beta-branched (Ile and Val), AsX (Asn and Asp), long (Met, Gln, Arg, Glu, Lys, and Leu), aromatic (Phe, Tyr, His, and Cys), small (Ala and Ser), bulky (Thr and Trp), and, lastly, the singletons of Gly and Pro. At the level of secondary structure (helix, sheet, turn, and coil), these groups remain somewhat consistent, although there are a few significant variations. Besides the expected uniqueness of the Gly and Pro distributions, the nonpolar/beta-branched and AsX clusters were very consistent across all types of secondary structure. Effectively, this consistency across the secondary structure classes implies that side-chain steric effects strongly influence a residue's backbone torsion angle conformation. These results help to explain the plasticity of amino acid substitutions on protein structure and should help in protein design and structure evaluation.