LRP5 regulates human body fat distribution by modulating adipose progenitor biology in a dose- and depot-specific fashion

Summary Common variants in WNT pathway genes have been associated with bone mass and fat distribution, the latter predicting diabetes and cardiovascular disease risk. Rare mutations in the WNT co-receptors LRP5 and LRP6 are similarly associated with bone and cardiometabolic disorders. We investigated the role of LRP5 in human adipose tissue. Subjects with gain-of-function LRP5 mutations and high bone mass had enhanced lower-body fat accumulation. Reciprocally, a low bone mineral density-associated common LRP5 allele correlated with increased abdominal adiposity. Ex vivo LRP5 expression was higher in abdominal versus gluteal adipocyte progenitors. Equivalent knockdown of LRP5 in both progenitor types dose-dependently impaired β-catenin signaling and led to distinct biological outcomes: diminished gluteal and enhanced abdominal adipogenesis. These data highlight how depot differences in WNT/β-catenin pathway activity modulate human fat distribution via effects on adipocyte progenitor biology. They also identify LRP5 as a potential pharmacologic target for the treatment of cardiometabolic disorders. © 2015 The Authors.

Mutations in known monogenic high bone mass loci only explain a small proportion of high bone mass cases

High bone mass (HBM) can be an incidental clinical finding; however, monogenic HBM disorders (eg, LRP5 or SOST mutations) are rare. We aimed to determine to what extent HBM is explained by mutations in known HBM genes. A total of 258 unrelated HBM cases were identified from a review of 335,115 DXA scans from 13 UK centers. Cases were assessed clinically and underwent sequencing of known anabolic HBM loci: LRP5 (exons 2, 3, 4), LRP4 (exons 25, 26), SOST (exons 1, 2, and the van Buchem's disease [VBD] 52-kb intronic deletion 3'). Family members were assessed for HBM segregation with identified variants. Three-dimensional protein models were constructed for identified variants. Two novel missense LRP5 HBM mutations ([c.518C>T; p.Thr173Met], [c.796C>T; p.Arg266Cys]) were identified, plus three previously reported missense LRP5 mutations ([c.593A>G; p.Asn198Ser], [c.724G>A; p.Ala242Thr], [c.266A>G; p.Gln89Arg]), associated with HBM in 11 adults from seven families. Individuals with LRP5 HBM ( approximately prevalence 5/100,000) displayed a variable phenotype of skeletal dysplasia with increased trabecular BMD and cortical thickness on HRpQCT, and gynoid fat mass accumulation on DXA, compared with both non-LRP5 HBM and controls. One mostly asymptomatic woman carried a novel heterozygous nonsense SOST mutation (c.530C>A; p.Ser177X) predicted to prematurely truncate sclerostin. Protein modeling suggests the severity of the LRP5-HBM phenotype corresponds to the degree of protein disruption and the consequent effect on SOST-LRP5 binding. We predict p.Asn198Ser and p.Ala242Thr directly disrupt SOST binding; both correspond to severe HBM phenotypes (BMD Z-scores +3.1 to +12.2, inability to float). Less disruptive structural alterations predicted from p.Arg266Cys, p.Thr173Met, and p.Gln89Arg were associated with less severe phenotypes (Z-scores +2.4 to +6.2, ability to float). In conclusion, although mutations in known HBM loci may be asymptomatic, they only account for a very small proportion ( approximately 3%) of HBM individuals, suggesting the great majority are explained by either unknown monogenic causes or polygenic inheritance.

R1SVM: A randomised nonlinear approach to large-scale anomaly detection

The problem of unsupervised anomaly detection arises in a wide variety of practical applications. While one-class support vector machines have demonstrated their effectiveness as an anomaly detection technique, their ability to model large datasets is limited due to their memory and time complexity for training. To address this issue for supervised learning of kernel machines, there has been growing interest in random projection methods as an alternative to the computationally expensive problems of kernel matrix construction and sup-port vector optimisation. In this paper we leverage the theory of nonlinear random projections and propose the Randomised One-class SVM (R1SVM), which is an efficient and scalable anomaly detection technique that can be trained on large-scale datasets. Our empirical analysis on several real-life and synthetic datasets shows that our randomised 1SVM algorithm achieves comparable or better accuracy to deep auto encoder and traditional kernelised approaches for anomaly detection, while being approximately 100 times faster in training and testing.

R1STM: One-class support tensor machine with randomised kernel

Identifying unusual or anomalous patterns in an underlying dataset is an important but challenging task in many applications. The focus of the unsupervised anomaly detection literature has mostly been on vectorised data. However, many applications are more naturally described using higher-order tensor representations. Approaches that vectorise tensorial data can destroy the structural information encoded in the high-dimensional space, and lead to the problem of the curse of dimensionality. In this paper we present the first unsupervised tensorial anomaly detection method, along with a randomised version of our method. Our anomaly detection method, the One-class Support Tensor Machine (1STM), is a generalisation of conventional one-class Support Vector Machines to higher-order spaces. 1STM preserves the multiway structure of tensor data, while achieving significant improvement in accuracy and efficiency over conventional vectorised methods. We then leverage the theory of nonlinear random projections to propose the Randomised 1STM (R1STM). Our empirical analysis on several real and synthetic datasets shows that our R1STM algorithm delivers comparable or better accuracy to a state-of-the-art deep learning method and traditional kernelised approaches for anomaly detection, while being approximately 100 times faster in training and testing.

Elliptical summary randomisation for sensor-based human activity recognition

Many conventional statistical machine learning al- gorithms generalise poorly if distribution bias ex- ists in the datasets. For example, distribution bias arises in the context of domain generalisation, where knowledge acquired from multiple source domains need to be used in a previously unseen target domains. We propose Elliptical Summary Randomisation (ESRand), an efficient domain generalisation approach that comprises of a randomised kernel and elliptical data summarisation. ESRand learns a domain interdependent projection to a la- tent subspace that minimises the existing biases to the data while maintaining the functional relationship between domains. In the latent subspace, ellipsoidal summaries replace the samples to enhance the generalisation by further removing bias and noise in the data. Moreover, the summarisation enables large-scale data processing by significantly reducing the size of the data. Through comprehensive analysis, we show that our subspace-based approach outperforms state-of-the-art results on several activity recognition benchmark datasets, while keeping the computational complexity significantly low.

The High Places [Review]

Sarah Holland-Batt reviews 'The High Places' by Fiona McFarlane

The effects of height and BMI on prostate cancer incidence and mortality: A Mendelian randomization study in 20,848 cases and 20,214 controls from the PRACTICAL consortium

Background Epidemiological studies suggest a potential role for obesity and determinants of adult stature in prostate cancer risk and mortality, but the relationships described in the literature are complex. To address uncertainty over the causal nature of previous observational findings, we investigated associations of height- and adiposity-related genetic variants with prostate cancer risk and mortality. Methods We conducted a case–control study based on 20,848 prostate cancers and 20,214 controls of European ancestry from 22 studies in the PRACTICAL consortium. We constructed genetic risk scores that summed each man’s number of height and BMI increasing alleles across multiple single nucleotide polymorphisms robustly associated with each phenotype from published genome-wide association studies. Results The genetic risk scores explained 6.31 and 1.46 % of the variability in height and BMI, respectively. There was only weak evidence that genetic variants previously associated with increased BMI were associated with a lower prostate cancer risk (odds ratio per standard deviation increase in BMI genetic score 0.98; 95 % CI 0.96, 1.00; p = 0.07). Genetic variants associated with increased height were not associated with prostate cancer incidence (OR 0.99; 95 % CI 0.97, 1.01; p = 0.23), but were associated with an increase (OR 1.13; 95 % CI 1.08, 1.20) in prostate cancer mortality among low-grade disease (p heterogeneity, low vs. high grade <0.001). Genetic variants associated with increased BMI were associated with an increase (OR 1.08; 95 % CI 1.03, 1.14) in all-cause mortality among men with low-grade disease (p heterogeneity = 0.03). Conclusions We found little evidence of a substantial effect of genetically elevated height or BMI on prostate cancer risk, suggesting that previously reported observational associations may reflect common environmental determinants of height or BMI and prostate cancer risk. Genetically elevated height and BMI were associated with increased mortality (prostate cancer-specific and all-cause, respectively) in men with low-grade disease, a potentially informative but novel finding that requires replication.

Polydimethylsiloxane (PDMS) modulates CD38 expression, absorbs retinoic acid and may perturb retinoid signalling

Polydimethylsiloxane (PDMS) is the most commonly used material in the manufacture of customized cell culture devices. While there is concern that uncured PDMS oligomers may leach into culture medium and/or hydrophobic molecules may be absorbed into PDMS structures, there is no consensus on how or if PDMS influences cell behaviour. We observed that human umbilical cord blood (CB)-derived CD34+ cells expanded in standard culture medium on PDMS exhibit reduced CD38 surface expression, relative to cells cultured on tissue culture polystyrene (TCP). All-trans retinoic acid (ATRA) induces CD38 expression, and we reasoned that this hydrophobic molecule might be absorbed by PDMS. Through a series of experiments we demonstrated that ATRA-mediated CD38 expression was attenuated when cultures were maintained on PDMS. Medium pre-incubated on PDMS for extended durations resulted in a time-dependant reduction of ATRA in the medium and increasingly attenuated CD38 expression. This indicated a time-dependent absorption of ATRA into the PDMS. To better understand how PDMS might generally influence cell behaviour, Ingenuity Pathway Analysis (IPA) was used to identify potential upstream regulators. This analysis was performed for differentially expressed genes in primary cells including CD34+ haematopoietic progenitor cells, mesenchymal stromal cells (MSC), and keratinocytes, and cell lines including prostate cancer epithelial cells (LNCaP), breast cancer epithelial cells (MCF-7), and myeloid leukaemia cells (KG1a). IPA predicted that the most likely common upstream regulator of perturbed pathways was ATRA. We demonstrate here that ATRA is absorbed by PDMS in a time-dependent manner and results in the concomitant reduced expression of CD38 on the cell surface of CB-derived CD34+ cells.