Integrated Bioinformatic Analysis of the Expression and Prognosis of Caveolae-Related Genes in Human Breast Cancer

Benchmarking feature selection and feature extraction methods to improve the performances of machine-learning algorithms for patient classification using metabolomics biomedical data

Objective

Classification tasks are an open challenge in the field of biomedicine. While several machine-learning techniques exist to accomplish this objective, several peculiarities associated with biomedical data, especially when it comes to omics measurements, prevent their use or good performance achievements. Omics approaches aim to understand a complex biological system through systematic analysis of its content at the molecular level. On the other hand, omics data are heterogeneous, sparse and affected by the classical "curse of dimensionality" problem, i.e. having much fewer observation, samples (n) than omics features (p). Furthermore, a major problem with multi-omics data is the imbalance either at the class or feature level. The objective of this work is to study whether feature extraction and/or feature selection techniques can improve the performances of classification machine-learning algorithms on omics measurements.

Methods

Among all omics, metabolomics has emerged as a powerful tool in cancer research, facilitating a deeper understanding of the complex metabolic landscape associated with tumorigenesis and tumor progression. Thus, we selected three publicly available metabolomics datasets, and we applied several feature extraction techniques both linear and non-linear, coupled or not with feature selection methods, and evaluated the performances regarding patient classification in the different configurations for the three datasets.

Results

We provide general workflow and guidelines on when to use those techniques depending on the characteristics of the data available. To further test the extension of our approach to other omics data, we have included a transcriptomics and a proteomics data. Overall, for all datasets, we showed that applying supervised feature selection improves the performances of feature extraction methods for classification purposes. Scripts used to perform all analyses are available at: https://github.com/Plant-Net/Metabolomic_project/.

Expression Profiles of Dopamine-Related Genes and miRNAs Regulating Their Expression in Breast Cancer

This study aimed to assess the expression profile of messenger RNA (mRNA) and microRNA (miRNA) related to the dopaminergic system in five types of breast cancer in Polish women. Patients with five breast cancer subtypes were included in the study: luminal A (n = 130), luminal B (n = 196, including HER2-, n = 100; HER2+, n = 96), HER2+ (n = 36), and TNBC (n = 43); they underwent surgery, during which tumor tissue was removed along with a margin of healthy tissue (control material). The molecular analysis included a microarray profile of mRNAs and miRNAs associated with the dopaminergic system, a real-time polymerase chain reaction preceded by reverse transcription for selected genes, and determinations of their concentration using enzyme-linked immunosorbent assay (ELISA). The conducted statistical analysis showed that five mRNAs statistically significantly differentiated breast cancer sections regardless of subtype compared to control samples; these were dopamine receptor 2 (DRD2), dopamine receptor 3 (DRD3), dopamine receptor 25 (DRD5), transforming growth factor beta 2 (TGF-β-2), and caveolin 2 (CAV2). The predicted analysis showed that hsa-miR-141-3p can regulate the expression of DRD2 and TGF-β-2, whereas hsa-miR-4441 is potentially engaged in the expression regulation of DRD3 and DRD5. In addition, the expression pattern of DRD5 mRNA can also be regulated by has-miR-16-5p. The overexpression of DRD2 and DRD3, with concomitant silencing of DRD5 expression, confirms the presence of dopaminergic abnormalities in breast cancer patients. Moreover, these abnormalities may be the result of miR-141-3P, miR-16-5p, and miR-4441 activity, regulating proliferation or metastasis.

A mechanosensitive caveolae-invadosome interplay drives matrix remodelling for cancer cell invasion

Invadosomes and caveolae are mechanosensitive structures that are implicated in metastasis. Here, we describe a unique juxtaposition of caveola clusters and matrix degradative invadosomes at contact sites between the plasma membrane of cancer cells and constricting fibrils both in 2D and 3D type I collagen matrix environments. Preferential association between caveolae and straight segments of the fibrils, and between invadosomes and bent segments of the fibrils, was observed along with matrix remodelling. Caveola recruitment precedes and is required for invadosome formation and activity. Reciprocally, invadosome disruption results in the accumulation of fibril-associated caveolae. Moreover, caveolae and the collagen receptor β1 integrin co-localize at contact sites with the fibrils, and integrins control caveola recruitment to fibrils. In turn, caveolae mediate the clearance of β1 integrin and collagen uptake in an invadosome-dependent and collagen-cleavage-dependent mechanism. Our data reveal a reciprocal interplay between caveolae and invadosomes that coordinates adhesion to and proteolytic remodelling of confining fibrils to support tumour cell dissemination.

Amelioration of Cholesterol Rich diet-induced Impaired Cognition in AD Transgenic Mice by an LXR Agonist TO901317 Is Associated with the Activation of the LXR-β-RXR-α-ABCA1 Transmembrane Transport System and Improving the Composition of Lipid Raft

BACKGROUND

It has been reported that LXR agonist can inhibit Aβ generation and alleviate Aβ-induced various adverse reactions in vivo and in vitro experiments, but the mechanisms have not been clarified. The study aimed to observe the effect of LXR agonist TO901317 on the cognitive function of AD transgenic mice fed with cholesterol-rich diet (CRD), and to explore the possible mechanism. Methods: 32 male 6-month-old double transgenic AD mice were enrolled and randomly divided into 4 groups: control (normal diet) group, CRD treatment group, TO901317 treatment group and GSK2033 treatment group. After 3 month, Morris water maze was for the changes of spatial exploration and memory ability; ELISA was for detecting the production of Aβ42 in the brain; the concentration of total cholesterol (TC), low density lipoprotein (LDL) and high density lipoprotein (HDL) in serum were detected by cholesterol enzyme colorimetry; Finally, the expression of LXR-β, RXR-α, ABCA1, caveolin-1, BACE1 and APP at protein level in the brains was measured by Western blotting.

RESULTS

Compared with the control group, the learning, memory ability and spatial exploration ability of the mice were more significantly serious in the CRD group (P<0.05); The contents of TC and LDL in the serum and the production of Aβ42 in the brains were significantly increased (P<0.05), but HDL was remarkably decreased (P<0.05); The protein levels of LXR-β, RXR-α and ABCA1 were also significantly decreased (P<0.05); The expression of caveolin-1, APP and BACE1 were evidently increased (P<0.05). However, after treatment with TO901317, the impaired learning and memory and spatial exploration ability of the mice were significantly improved (P<0.05); The contents of TC and LDL in serum and the production of Aβ42 in the brains were significantly decreased (P<0.05), but HLD was increased (P<0.05); The protein levels of LXR-β, RXR-α, ABCA1were all significantly increased (P<0.05), while, the expression of caveolin-1, APP and BACE1 were all significantly decreased (P<0.05). All the changes were reversed by GSK2033 (P<0.05).

CONCLUSIONS

TO901317 attenuated the more serious impairment of spatial exploration, learning and memory in transgenic AD mice induced by CRD, and the mechanism may be that TO901317 could activate the LXR-β/RXR-α/ABCA1 transmembrane transport system, promote the cholesterol efflux, and decreased caveolin-1, APP and BACE1, further reduce Aβ42 in the brains.