Search

Objective This study examines sex differences in social inferencing deficits after traumatic brain injury (TBI), and examines the odds of males and females being impaired while controlling for potential confounders. Design Cross-sectional survey. Setting Outpatient.USA and a University in Canada. Participants One hundred five participants with TBI (60 males, 45 females) and 105 healthy controls (HC; 57 males, 48 females). Interventions Not applicable. Main Outcome Measures The Awareness of Social Inference Test (TASIT), which includes 1) Emotional Evaluation Test (EET), 2) Social Inference-Minimal (SI-M) test, and 3) Social Inference-Enriched (SI-E) test. Results Within the HC sample, males and females performed similarly on all three TASIT subtests. Within the TBI group, males had significantly lower scores than females on EET (P = 0.03), SI-M (P=0.01) and SI-E (P=0.04). Using impairment cutoffs derived from the HC sample, significantly more males with TBI (30%) were impaired on the EET than females(16.7%); impairment was similar between males and females on SI-M and SI-E. When adjusting for executive functioning and education, the odds of being impaired on the EET did not significantly differ for males and females (OR = 0.47; 95% CI: 0.16 - 1.40; P = 0.18). Conclusions While more males with TBI have emotion perception deficits than females, the difference appears to be driven by education and executive functioning. Research is needed in larger samples with more definitive norms to better understand social inferencing impairments in males and females with TBI, and translation to interpersonal behaviors.

IL1β is a central regulator of systemic inflammatory response in breast cancer, but the precise regulatory mechanisms that dictate the overproduction of IL1β are largely unsolved. Here, we show that IL1β secretion is increased by the coculture of human monocyte–like cells and triple-negative breast cancer (TNBC) cells. In addition, macrophages robustly produced IL1β when exposed to the conditioned media of TNBC cells. Consistent with these observations, macrophage depletion decreased serum IL1β and reduced breast cancer progression in an orthotopic breast cancer mouse model. Profiling the secretome of human breast cancer cells revealed that the CD44 antigen was the most differentially released protein in basal conditions of TNBC cells. Antibody-mediated neutralization of CD44 abrogated IL1β production in macrophages and inhibited the growth of primary tumors. These results suggest IL1β-mediated oncogenic signaling is triggered by breast cancer cell membrane–derived soluble CD44 (sCD44) antigen, and targeting sCD44 antigen may provide an alternative therapeutic strategy for breast cancer treatment by modulating inflammatory tumor microenvironment. Significance: A novel positive feedback loop between IL1β and CD44 promotes TNBC malignant progression.

Functional markers and their quantitative features (eg, maximum value, time to maximum, area under the curve [AUC], etc) are increasingly being used in clinical studies to diagnose diseases. It is thus of interest to assess the diagnostic utility of functional markers by assessing alignment between their quantitative features and an ordinal gold standard test that reflects the severity of disease. The concept of broad sense agreement (BSA) has recently been introduced for studying the relationship between continuous and ordinal measurements, and provides a promising tool to address such a question. Our strategy is to adopt a general class of summary functionals (SFs), each of which flexibly captures a different quantitative feature of a functional marker, and study its alignment according to an ordinal outcome via BSA. We further illustrate the proposed framework using three special classes of SFs (AUC-type, magnitude-specific, and time-specific) that are widely used in clinical settings. The proposed BSA estimator is proven to be consistent and asymptotically normal given a consistent estimator for the SF. We further provide an inferential framework for comparing a pair of candidate SFs in terms of their importance on the ordinal outcome. Our simulation results demonstrate satisfactory finite-sample performance of the proposed framework. We demonstrate the application of our methods using a renal study.

We propose a practical principal component analysis (PCA) framework that provides a nonparametric means of simultaneously reducing the dimensions of and modeling functional and vector (multivariate) data. We first introduce a Hilbert space that combines functional and vector objects as a single hybrid object. The framework, termed a PCA of hybrid functional and vector data (HFV-PCA), is then based on the eigen-decomposition of a covariance operator that captures simultaneous variations of functional and vector data in the new space. This approach leads to interpretable principal components that have the same structure as each observation and a single set of scores that serves well as a low-dimensional proxy for hybrid functional and vector data. To support practical application of HFV-PCA, the explicit relationship between the hybrid PC decomposition and the functional and vector PC decompositions is established, leading to a simple and robust estimation scheme where components of HFV-PCA are calculated using the components estimated from the existing functional and classical PCA methods. This estimation strategy allows flexible incorporation of sparse and irregular functional data as well as multivariate functional data. We derive the consistency results and asymptotic convergence rates for the proposed estimators. We demonstrate the efficacy of the method through simulations and analysis of renal imaging data.

Existing missing data methods for functional data mainly focus on reconstructing missing measurements along a single function-a univariate functional data setting. Motivated by a renal study, we focus on a bivariate functional data setting, where each sampling unit is a collection of two distinct component functions, one of which may be missing. Specifically, we propose a Bayesian multiple imputation approach based on a bivariate functional latent factor model that exploits the joint changing patterns of the component functions to allow accurate and stable imputation of one component given the other. We further extend the framework to address multilevel bivariate functional data with missing components by modeling and exploiting inter-component and intra-subject correlations. We develop a Gibbs sampling algorithm that simultaneously generates multiple imputations of missing component functions and posterior samples of model parameters. For multilevel bivariate functional data, a partially collapsed Gibbs sampler is implemented to improve computational efficiency. Our simulation study demonstrates that our methods outperform other competing methods for imputing missing components of bivariate functional data under various designs and missingness rates. The motivating renal study aims to investigate the distribution and pharmacokinetic properties of baseline and post-furosemide renogram curves that provide further insights into the underlying mechanism of renal obstruction, with post-furosemide renogram curves missing for some subjects. We apply the proposed methods to impute missing post-furosemide renogram curves and obtain more refined insights.

Research Objectives To test the hypothesis that carbamazepine (CBZ) reduces irritability/aggression among individuals > 6 months post-traumatic brain injury (TBI). Design Parallel-group, randomized, double-blind, placebo-controlled, forced-titration trial of CBZ versus placebo. Setting Outpatient. Participants 70 individuals with chronic TBI and irritability (CBZ n=35 versus placebo n=35). Interventions CBZ (titrated up to 400 mg twice daily) or placebo equivalent two times daily. Main Outcome Measures A composite measure of Neuropsychiatric Inventory Irritability and Aggression Domains (NPI-I/A). Global impression of change recorded from participant, observer, and study clinician. Results The CBZ group did not differ significantly from the placebo group (p=0.60 and 0.59 for NPI-I/A observer and participant ratings respectively). High placebo effects were observed with Minimal Clinically Important Difference in observer NPI-I/A 57% in CBZ group and 77% in placebo group (p=0.09). Findings were similar for participant ratings. 18 of 35 had therapeutic CBZ level > 4. Therapeutic sample analysis revealed similar high placebo response and non-significant differences except clinician ratings favoring CBZ. Non-serious adverse events occurred more frequently in the CBZ group with greater nervous system effects. Conclusions up to 400 mg two times daily was not superior to placebo at reducing irritability/aggression according observers and participants. Large placebo effects may have masked the detection of differences. Clinician rating metrics suggest benefit, and thus, CBZ should remain a treatment option for the experienced brain injury clinician; data are provided that may aid treatment decisions. Author(s) Disclosures Dr. Hammond serves on the Avanir Scientific Advisory Committee.

Signal transducer and activator of transcription 3 (STAT3) plays important roles in cancer-associated inflammation by controlling expression of proinflammatory cytokines and chemokines. Recent studies suggest that C/EBPβ (CCAAT-enhancer binding protein beta) and STAT3 synergistically stimulate cancer cell proliferation and epithelial-mesenchymal transition. C/EBPβ is a leucine-zipper transcription factor that regulates expression of a variety of inflammatory cytokines or chemokines, such as IL-8, G-CSF (granulocyte colony stimulating factor), and GM-CSF (granulocyte macrophage colony stimulating factor) which induce neutrophil infiltration and differentiation. However, molecular mechanisms by which STAT3 and C/EBPβ cooperatively interact had not been fully elucidated. In this study, we found that the level of C/EBPβ protein, but not that of its mRNA transcript, was decreased in the absence of STAT3 in H-Ras transformed human mammary epithelial (H-Ras MCF10A) cells. In addition, silencing STAT3 dramatically induced ubiquitination of C/EBPβ for proteasomal degradation. Furthermore, direct interaction between STAT3 and C/EBPβ was confirmed by immunoprecipitation and proximity ligation assays. Taken together, these results suggest that STAT3 stabilizes C/EBPβ, thereby promoting cancer-associated inflammation.

OBJECTIVES To compare construct and predictive validity, readability, and time-to-administer of 2 negative attribution measures in participants with traumatic brain injury (TBI). SETTING Two TBI rehabilitation hospitals. PARTICIPANTS Eighty-five adults with complicated mild to severe TBI. MAIN MEASURES Negative attributions (intent, hostility, and blame) and anger responses to hypothetical scenarios were measured with the Epps scenarios and the Ambiguous Intention Hostility Questionnaire (AIHQ). Trait aggression was measured with the Buss-Perry Aggression Questionnaire (BPAQ). RESULTS Associations between attributions and anger responses (ie, construct validity) within each measure were significant (Epps: r = 0.61-0.74; AIHQ: r = 0.39-0.71); however, associations were stronger for Epps (Ps < .001). Receiver operating characteristics (ROC) revealed attributions from both measures predicted BPAQ scores (area under the ROC curves = 0.6-0.8); predictive validity did not statistically differ between the 2 measures. Both had comparable readability (fifth- to sixth-grade levels), but Epps required longer administration times. CONCLUSION Negative attributions affect anger and aggression after TBI, making it important to identify suitable assessments for the TBI population. While psychometric properties of the AIHQ and Epps scenarios should be further explored, this study offers early support for the use of either instrument in persons with TBI. Advantages and disadvantages of the AIHQ and Epps scenarios are highlighted.

Background Across many settings, lack of virologic control remains common in people with HIV (PWH) due to late presentation and lack of retention in care. This contributes to neuronal damage and neurocognitive impairment, which remain prevalent. More evidence is needed to understand these outcomes in both PWH and people without HIV (PWOH). Methods We recruited PWH initiating antiretroviral therapy (ART) as well as PWOH at two sites in the United States. 108 adults were enrolled (56 PWOH and 52 PWH), most of whom had a second assessment at least 24 weeks later (193 total assessments). Tumor necrosis factor alpha (TNFa), monocyte chemotactic protein-1 (MCP-1), neopterin, soluble CD14, and neurofilament light chain protein (NFL) were measured in plasma and cerebrospinal fluid (CSF). Using multivariate models including Bayesian Model Averaging (BMA), we analyzed factors associated with global neuropsychological (NP) performance (NPT-9) and CSF NFL at baseline and over time. Results At baseline, higher CSF MCP-1 and plasma sCD14 were associated with worse NPT-9 in PWH, while CSF HIV RNA decrease was the only marker associated with improved NPT-9 over time. Among PWH, higher CSF neopterin was most closely associated with higher NFL. Among PWOH, higher CSF MCP-1 was most closely associated with higher NFL. Following ART initiation, decrease in CSF MCP-1 was most closely associated with NFL decrease. Conclusion Monocyte-associated CSF biomarkers are highly associated with neuronal damage in both PWH and PWOH. More research is needed to evaluate if therapies targeting monocyte-associated inflammation may ameliorate HIV-associated neurobehavioral diseases.

Background and Purpose: Previous data suggest patient demographics and clinical presentation are primary predictors of motor recovery poststroke, with minimal contributions of physical interventions. Other studies indicate consistent associations between the amount and intensity of stepping practice with locomotor outcomes. The goal of this study was to determine the relative contributions of these combined variables to locomotor outcomes poststroke across a range of patient demographics and baseline function. Methods: Data were pooled from 3 separate trials evaluating the efficacy of high-intensity training, low-intensity training, and conventional interventions. Demographics, clinical characteristics, and training activities from 144 participants >1-month poststroke were included in stepwise regression analyses to determine their relative contributions to locomotor outcomes. Subsequent latent profile analyses evaluated differences in classes of participants based on their responses to interventions. Results: Stepwise regressions indicate primary contributions of stepping activity on locomotor outcomes, with additional influences of age, duration poststroke, and baseline function. Latent profile analyses revealed 2 main classes of outcomes, with the largest gains in those who received high-intensity training and achieved the greatest amounts of stepping practice. Regression and latent profile analyses of only high-intensity training participants indicated age, baseline function, and training activities were primary determinants of locomotor gains. Participants with the smallest gains were older (≈60 years), presented with slower gait speeds ( Conclusions: Regression and cluster analyses reveal primary contributions of training interventions on mobility outcomes in patients >1-month poststroke. Age, duration poststroke, and baseline impairments were secondary predictors. Registration: URL: https://www.clinicaltrials.gov . Unique identifier: NCT02507466 and NCT01789853.

Cancer-associated fibroblasts (CAFs) constitute a major compartment of the tumor microenvironment. In the present study, we investigated the role for CAFs in breast cancer progression and underlying molecular mechanisms. Human breast cancer MDA-MB-231 cells treated with the CAF-conditioned media manifested a more proliferative phenotype, as evidenced by enhanced messenger RNA (mRNA) expression of Cyclin D1, c-Myc, and proliferating cell nuclear antigen. Analysis of data from The Cancer Genome Atlas revealed that fibroblast growth factor-2 (FGF2) expression was well correlated with the presence of CAFs. We noticed that the mRNA level of FGF2 in CAFs was higher than that in normal fibroblasts. FGF2 exerts its biological effects through interaction with FGF receptor 1 (FGFR1). In the breast cancer tissue array, 42% estrogen receptor-negative patients coexpressed FGF2 and FGFR1, whereas only 19% estrogen receptor-positive patients exhibited coexpression. CAF-stimulated MDA-MB-231 cell migration and invasiveness were abolished when FGF2-neutralizing antibody was added to the conditioned media of CAFs. In a xenograft mouse model, coinjection of MDA-MB-231 cells with activated fibroblasts expressing FGF2 dramatically enhanced tumor growth, and this was abrogated by silencing of FGFR1 in cancer cells. In addition, treatment of MDA-MB-231 cells with FGF2 enhanced expression of Cyclin D1, a key molecule involved in cell cycle progression. FGF2-induced cell migration and upregulation of Cyclin D1 were abolished by siRNA-mediated FGFR1 silencing. Taken together, the above findings suggest that CAFs promote growth, migration and invasion of MDA-MB-231 cells via the paracrine FGF2-FGFR1 loop in the breast tumor microenvironment.

The proviral integration site for Moloney murine leukemia virus (PIM) family of serine/threonine-specific kinases consist of three isoforms, that regulate proliferation, apoptosis, metabolism, invasion, and metastasis of cancer cells. Among these, abnormally elevated kinase activity of PIM-1 contributes to the progression of gastric cancer and predicts poor prognosis and a low survival rate in gastric cancer patients. In the present study, we found that resveratrol, one of the representative chemopreventive and anticarcinogenic phytochemicals, directly binds to PIM-1 and thereby inhibits its catalytic activity in human gastric cancer SNU-601 cells. This resulted in suppression of phosphorylation of the proapoptotic Bad, a known substrate of PIM-1. Resveratrol, by inactivating PIM-1, also inhibited anchorage-independent growth and proliferation of SNU-601 cells. To understand the molecular interaction between resveratrol and PIM-1, we conducted docking simulation and found that resveratrol directly binds to the PIM-1 at the ATP-binding pocket. In conclusion, the proapototic and anti-proliferative effects of resveratrol in gastric cancer cells are likely to be mediated through suppression of PIM-1 kinase activity, which may represent a novel mechanism underlying its chemopreventive and anticarcinogenic actions.

ObjectivesImmune non-responders (INR) have poor CD4 recovery and are associated with increased risk of serious events despite antiretroviral therapy (ART). A clinically relevant definition for INR is lacking.MethodsWe conducted a retrospective analysis of three large cohorts: Infectious Disease Clinic at the Atlanta Veterans Affairs Medical Center, the US Military HIV Natural History Study and Infectious Disease Program of the Grady Health System in Atlanta, Georgia. Two-stage modeling and joint model (JM) approaches were used to evaluate the association between CD4 (or CD4/CD8 ratio) slope within two years since ART initiation and a composite endpoint (AIDS, serious non-AIDS events and death) after two years of ART. We compared the predictive capacity of four CD4 count metrics (estimated CD4 slope, estimated CD4/CD8 ratio slope during two years following ART initiation and CD4 at 1 and 2 years following ART initiation) using Cox regression models.ResultsWe included 2,422 patients. Mean CD4 slope (±standard error) during two years of ART was 102 ± 2 cells/μl/year (95% confidence interval: 98-106 cells/μl/year), this increase was uniform among the three cohorts (p = 0.80). There were 267 composite events after two years on ART. Using the JM approach, a CD4 slope ≥100 cells/μL/year or CD4/CD8 ratio slope >0.1 higher rate per year were associated with lower composite endpoint rates (adjusted hazard ratio [HR] = 0.80, p = 0.04 and HR = 0.75 pConclusionsUsing a complex JM approach, CD4 slope and CD4/CD8 ratio slope the first two years after ART initiation were associated with lower rates of the composite outcome. Moreover, the uniformity observed in the mean CD4 slope regardless of the cohort suggests a common CD4 response pattern independent of age or CD4 nadir. Given the consistency observed with CD4 slope, availability and ease of interpretation, this study provides strong rationale for using CD4 gains

Signal transducer and activator of transcription 3 (STAT3) functions as a major molecular switch that plays an important role in the communication between cytokines and kinases. In this role, it regulates the transcription of genes involved in various biochemical processes, such as proliferation, migration, and metabolism of cancer cells. STAT3 undergoes diverse post-translational modifications, such as the oxidation of cysteine by oxidative stress, the acetylation of lysine, or the phosphorylation of serine/threonine. In particular, the redox modulation of critical cysteine residues present in the DNA-binding domain of STAT3 inhibits its DNA-binding activity, resulting in the inactivation of STAT3-mediated gene expression. Accumulating evidence supports that STAT3 is a key protein that acts as a mediator of metabolism and mitochondrial activity. In this review, we focus on the post-translational modifications of STAT3 by oxidative stress and how the modification of STAT3 regulates cell metabolism, particularly in the metabolic pathways in cancer cells.

Cancer stem cells (CSCs) are responsible for tumor initiation, metastasis and recurrence. Caveolin-1 (Cav-1) is a major protein of caveolae, which participates in various cellular functions, such as vesicle trafficking, cholesterol homeostasis, tumor progression, etc. In the present study, we investigated a role for Cav-1 in regulating the stemness of human breast cancer (MDA-MB-231) cells. Cav-1 expression was significantly lower in tumorspheres than in adherent cells. The silencing of Cav-1 enhanced stemness of MDA-MB-231 cells. Mechanistically, Cav-1 silencing was accompanied by enhanced expression of Bmi-1, which is a representative self-renewal regulator, and promoted epithelial-mesenchymal transition. In a CSC-like state, reduced Cav-1 depends on its destabilization through ubiquitin-proteasome degradation. We further found that Src-mediated phosphorylation of Cav-1 at the Tyr 14 residue is essential for its degradation. Taken together, these findings suggest that Cav-1 destabilization by Src may play a pivotal role in manifestation and maintenance of stemness in breast cancer cells.

The need to assess agreement exists in various clinical studies where quantifying inter-rater reliability is of great importance. Use of unscaled agreement indices, such as total deviation index and coverage probability (CP), is recommended for two main reasons: (i) they are intuitive in a sense that interpretations are tied to the original measurement unit; (ii) practitioners can readily determine whether the agreement is satisfactory by directly comparing the value of the index to a prespecified tolerable CP or absolute difference. However, the unscaled indices were only defined in the context of comparing two raters or multiple raters that assume homogeneity of variances across raters. In this paper, we introduce a set of overall indices based on the root mean square of pairwise differences that are unscaled and can be used to evaluate agreement among multiple raters that often exhibit heterogeneous measurement processes in practice. Furthermore, we propose another overall agreement index based on the root mean square of pairwise differences that is scaled and extends the concept of the recently proposed relative area under CP curve in the presence of multiple raters. We present the definitions of overall indices and propose inference procedures in which bootstrap methods are used for the estimation of standard errors. We assess the performance of the proposed approach and demonstrate its superiority over the existing methods when raters exhibit heterogeneous measurement processes using simulation studies. Finally, we demonstrate the application of our methods using a renal study.

Mammalian target of rapamycin (mTOR), a serine/threonine protein kinase, forms two different complexes, complex 1 and 2, and plays a key role in the regulation of Akt signaling-mediated cell proliferation and transformation. This study reveals aschantin, a natural compound abundantly found in Magnolia flos, as a novel mTOR kinase inhibitor. Aschantin directly targeted the active pocket of mTOR kinase domain by competing with adenosine triphosphate (ATP), but not PI3K and PDK1. Aschantin inhibited epidermal growth factor (EGF)-induced full activation of Akt by phosphorylation at Ser473/Thr308, resulting in inhibition of the mTORC2/Akt and Akt/mTORC1/p70S6K signaling pathways and activation of GSK3β by abrogation of Akt-mediated GSK3β phosphorylation at Ser9. The activated GSK3β inhibited cell proliferation by c-Jun phosphorylation at Ser243, which facilitated destabilization and degradation of c-Jun through the ubiquitination-mediated proteasomal degradation pathway. Notably, aschantin treatment decreased c-Jun stability through inhibition of the mTORC2-Akt signaling pathway, which suppressed EGF-induced anchorage-independent cell transformation in non-malignant JB6 Cl41 and HaCaT cells and colony growth of LNCaP and MIAPaCa-2 cancer cells in soft agar. Altogether, the results show that aschantin targets mTOR kinase and destabilizes c-Jun, which implicate aschantin as a potential chemopreventive or therapeutic agent.

15-Deoxy-Δ

15-Deoxy-Δ15d-PGJ15d-PGJ15d-PGJ

Background Epidermal autografting has been used to treat vitiligo. The pigmentation achieved at the recipient site can be variegated and incomplete compared with that of the surrounding normal skin, and sometimes remains that way for a fairly long time. Objective We investigated whether the clinical results from epidermal autografting are related to a change in the number of melanocytes. This was performed by counting the number of melanocytes in the epidermis obtained from biopsy and suction with and without psoralen plus UVA (PUVA) exposure of the donor sites before grafting. Methods The numbers of melanocytes in the epidermis were counted after staining with dopa. The epidermis from suction and biopsy was included. The biopsied specimen was treated with NaBr for dermo–epidermal separation before staining, whereas the epidermis obtained from suction was stained directly. Results The epidermis obtained from suction contained 40–60% of the number of melanocytes found in the biopsied epidermis. Melanocytes around the hair follicles seemed to be omitted. Treatment with PUVA 10–21 times caused the number of melanocytes to increase by 1.5–2 times the normal level with a promising clinical result. Conclusions The preparation of donor sites with PUVA before the treatment of vitiligo by epidermal autografting induced an increased number of melanocytes and improved the clinical result.

Background Magnolin is a natural compound abundantly found in Magnolia flos, which has been traditionally used in oriental medicine to treat headaches, nasal congestion and anti-inflammatory reactions. Our recent results have demonstrated that magnolin targets the active pockets of ERK1 and ERK2, which are important signaling molecules in cancer cell metastasis. The aim of this study is to evaluate the effects of magnolin on cell migration and to further explore the molecular mechanisms involved. Methods Magnolin-mediated signaling inhibition was confirmed by Western blotting using RSK2+/+ and RSK2−/− MEFs, A549 and NCI-H1975 lung cancer cells, and by NF-κB and Cox-2 promoter luciferase reporter assays. Inhibition of cell migration by magnolin was examined by wound healing and/or Boyden Chamber assays using JB6 Cl41 and A549 human lung cancer cells. The molecular mechanisms involved in cell migration and epithelial-to-mesenchymal transition were determined by zymography, Western blotting, real-time PCR and immunocytofluorescence. Results Magnolin inhibited NF-κB transactivation activity by suppressing the ERKs/RSK2 signaling pathway. Moreover, magnolin abrogated the increase in EGF-induced COX-2 protein levels and wound healing. In human lung cancer cells such as A549 and NCI-H1975, which harbor constitutive active Ras and EGFR mutants, respectively, magnolin suppressed wound healing and cell invasion as seen by a Boyden chamber assay. In addition, it was observed that magnolin inhibited MMP-2 and −9 gene expression and activity. The knockdown or knockout of RSK2 in A549 lung cancer cells or MEFs revealed that magnolin targeting ERKs/RSK2 signaling suppressed epithelial-to-mesenchymal transition by modulating EMT marker proteins such as N-cadherin, E-cadherin, Snail, Vimentin and MMPs. Conclusions These results demonstrate that magnolin inhibits cell migration and invasion by targeting the ERKs/RSK2 signaling pathway. Electronic supplementary material The online version of this article (doi:10.1186/s12885-015-1580-7) contains supplementary material, which is available to authorized users.