Tumor-derived cytokines impair myogenesis and alter the skeletal muscle immune microenvironment

Mutant RAS-driven Secretome Causes Skeletal Muscle Defects in Breast Cancer

Cancer-induced skeletal muscle defects differ in severity between individuals with the same cancer type. Cancer subtype-specific genomic aberrations are suggested to mediate these differences, but experimental validation studies are very limited. We utilized three different breast cancer patient-derived xenograft (PDX) models to correlate cancer subtype with skeletal muscle defects. PDXs were derived from brain metastasis of triple-negative breast cancer (TNBC), estrogen receptor-positive/progesterone receptor-positive (ER+/PR+) primary breast cancer from a BRCA2-mutation carrier, and pleural effusion from an ER+/PR- breast cancer. While impaired skeletal muscle function as measured through rotarod performance and reduced levels of circulating and/or skeletal muscle miR-486 were common across all three PDXs, only TNBC-derived PDX activated phospho-p38 in skeletal muscle. To further extend these results, we generated transformed variants of human primary breast epithelial cells from healthy donors using HRASG12V or PIK3CAH1047R mutant oncogenes. Mutations in RAS oncogene or its modulators are found in approximately 37% of metastatic breast cancers, which is often associated with skeletal muscle defects. Although cells transformed with both oncogenes generated adenocarcinomas in NSG mice, only HRASG12V-derived tumors caused skeletal muscle defects affecting rotarod performance, skeletal muscle contraction force, and miR-486, Pax7, pAKT, and p53 levels in skeletal muscle. Circulating levels of the chemokine CXCL1 were elevated only in animals with tumors containing HRASG12V mutation. Because RAS pathway aberrations are found in 19% of cancers, evaluating skeletal muscle defects in the context of genomic aberrations in cancers, particularly RAS pathway mutations, may accelerate development of therapeutic modalities to overcome cancer-induced systemic effects.

SIGNIFICANCE

Mutant RAS- and PIK3CA-driven breast cancers distinctly affect the function of skeletal muscle. Therefore, research and therapeutic targeting of cancer-induced systemic effects need to take aberrant cancer genome into consideration.

Trends and Future Research in Skeletal Muscle Tissue Engineering in the Past Decade (2012-2022)

To learn about advances in skeletal muscle tissue engineering (SMTE) in recent years, we used VOSviewer and Citespace software to quantitatively analyze and visualize relevant literature in the Web of Science database during the period 2012-2022. By mapping high-frequency keyword relationship networks, keyword time zones, and journal article cocitations, we clarified the areas of great interest, evolutionary paths, and developmental trends in research on SMTE. We conducted an in-depth analysis of highly cited and representative articles at various stages to summarize the mainstream research areas of great interest in SMTE and discussed the future development and challenges in this field, intending to provide a reference for the clinical treatment of skeletal muscle injury repair. We found that a collaborative network of authors has formed in this field; the journals publishing SMTE articles belong to the fields of biomaterials and tissue engineering, and open-access journals have played a key role in the promotion of the development of SMTE; and in the past decade, there has been rapid progress in SMTE research in terms of both depth and breadth. Impact statement Compared with the literature review method, bibliometrics can provide a comprehensive knowledge of a knowledge area based on a huge amount of literature. In this article, based on the Web of Science database, CiteSpace, and Vosviewer visualization tools were used to measure and analyze the literature reports in the field of skeletal muscle tissue engineering (SMTE). The research hotspots and cutting-edge information on SMTE were mined in terms of the number of publications, the number of citations, the keywords, the authors, and the publishing institutions to understand the current status of the research on SMTE in the world, to provide a reference for related researchers, engineering research in the field of SMTE, to comprehensively understand the current status of global research in the field of SMTE, and to provide a reference for related researchers.

The Molecular Aspects of Functional Activity of Macrophage-Activating Factor GcMAF

Group-specific component macrophage-activating factor (GcMAF) is the vitamin D3-binding protein (DBP) deglycosylated at Thr420. The protein is believed to exhibit a wide range of therapeutic properties associated with the activation of macrophagal immunity. An original method for GcMAF production, DBP conversion to GcMAF, and the analysis of the activating potency of GcMAF was developed in this study. Data unveiling the molecular causes of macrophage activation were obtained. GcMAF was found to interact with three CLEC10A derivatives having molecular weights of 29 kDa, 63 kDa, and 65 kDa. GcMAF interacts with high-molecular-weight derivatives via Ca2+-dependent receptor engagement. Binding to the 65 kDa or 63 kDa derivative determines the pro- and anti-inflammatory direction of cytokine mRNA expression: 65 kDa-pro-inflammatory (TNF-α, IL-1β) and 63 kDa-anti-inflammatory (TGF-β, IL-10). No Ca2+ ions are required for the interaction with the canonical 29 kDa CLEC10A. Both forms, DBP protein and GcMAF, bind to the 29 kDa CLEC10A. This interaction is characterized by the stochastic mRNA synthesis of the analyzed cytokines. Ex vivo experiments have demonstrated that when there is an excess of GcMAF ligand, CLEC10A forms aggregate, and the mRNA synthesis of analyzed cytokines is inhibited. A schematic diagram of the presumable mechanism of interaction between the CLEC10A derivatives and GcMAF is provided. The principles and elements of standardizing the GcMAF preparation are elaborated.

Unravelling the Role of Cancer Cell-Derived Extracellular Vesicles in Muscle Atrophy, Lipolysis, and Cancer-Associated Cachexia

Cancer-associated cachexia is a metabolic syndrome that causes significant reduction in whole-body weight due to excessive loss of muscle mass accompanied by loss of fat mass. Reduced food intake and several metabolic abnormalities, such as increased energy expenditure, excessive catabolism, and inflammation, are known to drive cachexia. It is well documented that cancer cells secrete EVs in abundance which can be easily taken up by the recipient cell. The cargo biomolecules carried by the EVs have the potential to alter the signalling pathways and function of the recipient cells. EV cargo includes proteins, nucleic acids, lipids, and metabolites. Tumour-secreted EVs have been found to alter the metabolic and biological functions of adipose and muscle tissue, which aids in the development of the cachexia phenotype. To date, no medical intervention or FDA-approved drug exists that can completely reverse cachexia. Therefore, understanding how cancer-derived EVs contribute to the onset and progression of cancer-associated cachexia may help with the identification of new biomarkers as well as provide access to novel treatment alternatives. The goal of this review article is to discuss the most recent research on cancer-derived EVs and their function in cellular crosstalk that promotes catabolism in muscle and adipose tissue during cancer-induced cachexia.

ROCK1 activates mitochondrial fission leading to oxidative stress and muscle atrophy

Chronic kidney disease (CKD) is often associated with protein-energy wasting (PEW), which is characterized by a reduction in muscle mass and strength. Although mitochondrial dysfunction and oxidative stress have been implicated to play a role in the pathogenesis of muscle wasting, the underlying mechanisms remain unclear. In this study, we used transcriptomics, metabolomics analyses and mouse gene manipulating approaches to investigate the effects of mitochondrial plasticity and oxidative stress on muscle wasting in mouse CKD models. Our results showed that the expression of oxidative stress response genes was increased, and that of oxidative phosphorylation genes was decreased in the muscles of mice with CKD. This was accompanied by reduced oxygen consumption rates, decreased levels of mitochondrial electron transport chain proteins, and increased cellular oxidative damage. Excessive mitochondrial fission was also observed, and we found that the activation of ROCK1 was responsible for this process. Inducible expression of muscle-specific constitutively active ROCK1(mROCK1ca)exacerbated mitochondrial fragmentation and muscle wasting in CKD mice. Conversely, ROCK1 depletion (ROCK1-/-) alleviated these phenomena. Mechanistically, ROCK1 activation promoted the recruitment of Drp1 to mitochondria, thereby facilitating fragmentation. Notably, the pharmacological inhibition of ROCK1 mitigated muscle wasting by suppressing mitochondrial fission and oxidative stress. Our findings demonstrate that ROCK1 participates in CKD-induced muscle wasting by promoting mitochondrial fission and oxidative stress, and pharmacological suppression of ROCK1 could be a therapeutic strategy for combating muscle wasting in CKD conditions.

Non-immune cell components in tumor microenvironment influencing lung cancer Immunotherapy

Lung cancer (LC) is one of the leading causes of cancer-related deaths worldwide, with a significant morbidity and mortality rate, endangering human life and health. The introduction of immunotherapies has significantly altered existing cancer treatment strategies and is expected to improve immune responses, objective response rates, and survival rates. However, a better understanding of the complex immunological networks of LC is required to improve immunotherapy efficacy further. Tumor-associated antigens (TAAs) and tumor-specific antigens (TSAs) are significantly expressed by LC cells, which activate dendritic cells, initiate antigen presentation, and activate lymphocytes to exert antitumor activity. However, as tumor cells combat the immune system, an immunosuppressive microenvironment forms, enabling the enactment of a series of immunological escape mechanisms, including the recruitment of immunosuppressive cells and induction of T cell exhaustion to decrease the antitumor immune response. In addition to the direct effect of LC cells on immune cell function, the secreting various cytokines, chemokines, and exosomes, changes in the intratumoral microbiome and the function of cancer-associated fibroblasts and endothelial cells contribute to LC cell immune escape. Accordingly, combining various immunotherapies with other therapies can elicit synergistic effects based on the complex immune network, improving immunotherapy efficacy through multi-target action on the tumor microenvironment (TME). Hence, this review provides guidance for understanding the complex immune network in the TME and designing novel and effective immunotherapy strategies for LC.

Zoledronic acid improves bone quality and muscle function in a high bone turnover state

SUMMARY

Zoledronic acid (ZA) prevents muscle weakness in mice with bone metastases; however, its role in muscle weakness in non-tumor-associated metabolic bone diseases and as an effective treatment modality for the prevention of muscle weakness associated with bone disorders, is unknown. We demonstrate the role of ZA-treatment on bone and muscle using a mouse model of accelerated bone remodeling, which represents the clinical manifestation of non-tumor associated metabolic bone disease. ZA increased bone mass and strength and rescued osteocyte lacunocanalicular organization. Short-term ZA treatment increased muscle mass, whereas prolonged, preventive treatment improved muscle mass and function. In these mice, muscle fiber-type shifted from oxidative to glycolytic and ZA restored normal muscle fiber distribution. By blocking TGFβ release from bone, ZA improved muscle function, promoted myoblast differentiation and stabilized Ryanodine Receptor-1 calcium channel. These data demonstrate the beneficial effects of ZA in maintaining bone health and preserving muscle mass and function in a model of metabolic bone disease.

Context and significance

TGFβ is a bone regulatory molecule which is stored in bone matrix, released during bone remodeling, and must be maintained at an optimal level for the good health of the bone. Excess TGFβ causes several bone disorders and skeletal muscle weakness. Reducing excess TGFβ release from bone using zoledronic acid in mice not only improved bone volume and strength but also increased muscle mass, and muscle function. Progressive muscle weakness coexists with bone disorders, decreasing quality of life and increasing morbidity and mortality. Currently, there is a critical need for treatments improving muscle mass and function in patients with debilitating weakness. Zoledronic acid's benefit extends beyond bone and could also be useful in treating muscle weakness associated with bone disorders.

Single cell analysis reveals satellite cell heterogeneity for proinflammatory chemokine expression

Background: The expression of proinflammatory signals at the site of muscle injury are essential for efficient tissue repair and their dysregulation can lead to inflammatory myopathies. Macrophages, neutrophils, and fibroadipogenic progenitor cells residing in the muscle are significant sources of proinflammatory cytokines and chemokines. However, the inducibility of the myogenic satellite cell population and their contribution to proinflammatory signaling is less understood.

Methods: Mouse satellite cells were isolated and exposed to lipopolysaccharide (LPS) to mimic sterile skeletal muscle injury and changes in the expression of proinflammatory genes was examined by RT-qPCR and single cell RNA sequencing. Expression patterns were validated in skeletal muscle injured with cardiotoxin by RT-qPCR and immunofluorescence.

Results: Satellite cells in culture were able to express Tnfa, Ccl2, and Il6, within 2 h of treatment with LPS. Single cell RNA-Seq revealed seven cell clusters representing the continuum from activation to differentiation. LPS treatment led to a heterogeneous pattern of induction of C-C and C-X-C chemokines (e.g., Ccl2, Ccl5, and Cxcl0) and cytokines (e.g., Tgfb1, Bmp2, Il18, and Il33) associated with innate immune cell recruitment and satellite cell proliferation. One cell cluster was enriched for expression of the antiviral interferon pathway genes under control conditions and LPS treatment. Activation of this pathway in satellite cells was also detectable at the site of cardiotoxin induced muscle injury.

Conclusion: These data demonstrate that satellite cells respond to inflammatory signals and secrete chemokines and cytokines. Further, we identified a previously unrecognized subset of satellite cells that may act as sensors for muscle infection or injury using the antiviral interferon pathway.

Proteomic analysis of brain metastatic lung adenocarcinoma reveals intertumoral heterogeneity and specific alterations associated with the timing of brain metastases

BACKGROUND

Brain metastases are associated with considerable negative effects on patients' outcome in lung adenocarcinoma (LADC). Here, we investigated the proteomic landscape of primary LADCs and their corresponding brain metastases.

MATERIALS AND METHODS

Proteomic profiling was conducted on 20 surgically resected primary and brain metastatic LADC samples via label-free shotgun proteomics. After sample processing, peptides were analyzed using an Ultimate 3000 pump coupled to a QExactive HF-X mass spectrometer. Raw data were searched using PD 2.4. Further data analyses were carried out using Perseus, RStudio and GraphPad Prism. Proteomic data were correlated with clinical and histopathological parameters and the timing of brain metastases. Mass spectrometry-based proteomic data are available via ProteomeXchange with identifier PXD027259.

RESULTS

Out of the 6821 proteins identified and quantified, 1496 proteins were differentially expressed between primary LADCs and corresponding brain metastases. Pathways associated with the immune system, cell-cell/matrix interactions and migration were predominantly activated in the primary tumors, whereas pathways related to metabolism, translation or vesicle formation were overrepresented in the metastatic tumors. When comparing fast- versus slow-progressing patients, we found 454 and 298 differentially expressed proteins in the primary tumors and brain metastases, respectively. Metabolic reprogramming and ribosomal activity were prominently up-regulated in the fast-progressing patients (versus slow-progressing individuals), whereas expression of cell-cell interaction- and immune system-related pathways was reduced in these patients and in those with multiple brain metastases.

CONCLUSIONS

This is the first comprehensive proteomic analysis of paired primary tumors and brain metastases of LADC patients. Our data suggest a malfunction of cellular attachment and an increase in ribosomal activity in LADC tissue, promoting brain metastasis. The current study provides insights into the biology of LADC brain metastases and, moreover, might contribute to the development of personalized follow-up strategies in LADC.

Impaired muscle stem cell function and abnormal myogenesis in acquired myopathies

Skeletal muscle possesses a high plasticity and a remarkable regenerative capacity that relies mainly on muscle stem cells (MuSCs). Molecular and cellular components of the MuSC niche, such as immune cells, play key roles to coordinate MuSC function and to orchestrate muscle regeneration. An abnormal infiltration of immune cells and/or imbalance of pro- and anti-inflammatory cytokines could lead to MuSC dysfunctions that could have long lasting effects on muscle function. Different genetic variants were shown to cause muscular dystrophies that intrinsically compromise MuSC function and/or disturb their microenvironment leading to impaired muscle regeneration that contributes to disease progression. Alternatively, many acquired myopathies caused by comorbidities (e.g., cardiopulmonary or kidney diseases), chronic inflammation/infection, or side effects of different drugs can also perturb MuSC function and their microenvironment. The goal of this review is to comprehensively summarize the current knowledge on acquired myopathies and their impact on MuSC function. We further describe potential therapeutic strategies to restore MuSC regenerative capacity.

Interleukin-11 (IL11) inhibits myogenic differentiation of C2C12 cells through activation of extracellular signal-regulated kinase (ERK)

Cancer-associated cachexia (CAC) is a multifactorial wasting syndrome characterized by loss of skeletal muscle. Interleukin-11 (IL11), one of the IL6 family cytokines, is highly expressed in various types of cancer including cancers frequently associated with cachexia. However, the impact of IL11 on muscle metabolism remains to be determined. Since one of the mechanisms of muscle wasting in cachexia is defective muscle regeneration due to impaired myogenic differentiation, we examined the effect of IL11 on the differentiation of C2C12 mouse myoblasts. Treatment of C2C12 cells with recombinant mouse IL11 resulted in decreased myotube formation. In addition, IL11 treatment reduced the protein and mRNA levels of myosin heavy chain (MHC), a marker of myogenic differentiation. Moreover, the levels of myogenic regulatory factors including myogenin and Mrf4 were significantly reduced by IL11 treatment. IL11 treatment increased the number of BrdU-positive cells and the level of phosphorylated retinoblastoma (Rb) protein, while the levels of p21^Waf1 and p27^Kip1 were reduced by IL11 treatment in differentiating C2C12 cells, suggesting that IL11 interferes with cell cycle exit during the early stages of myogenic differentiation. Consistent with this, IL11 treatment at the late stage of differentiation did not affect myotube formation and MHC expression. IL11 treatment resulted in an activation of ERK, STAT3, and AKT in differentiating C2C12 cells. However, only ERK inhibitors including PD98059 and U0126 were able to ameliorate the suppressive effect of IL11 on the expression of MHC and myogenin. Additionally, pretreatment with PD98059 and U0126 resulted in improved myotube formation and reduced BrdU staining in IL11-treated cells. Together, our results suggest that IL11 inhibits myogenic differentiation through delayed cell cycle exit in an ERK-dependent manner. To our knowledge, this study is the first to demonstrate an inhibitory role of IL11 in myogenic differentiation and identifies the previously unrecognized role of IL11 as a possible mediator of CAC.

Transforming L1000 profiles to RNA-seq-like profiles with deep learning

The L1000 technology, a cost-effective high-throughput transcriptomics technology, has been applied to profile a collection of human cell lines for their gene expression response to > 30,000 chemical and genetic perturbations. In total, there are currently over 3 million available L1000 profiles. Such a dataset is invaluable for the discovery of drug and target candidates and for inferring mechanisms of action for small molecules. The L1000 assay only measures the mRNA expression of 978 landmark genes while 11,350 additional genes are computationally reliably inferred. The lack of full genome coverage limits knowledge discovery for half of the human protein coding genes, and the potential for integration with other transcriptomics profiling data. Here we present a Deep Learning two-step model that transforms L1000 profiles to RNA-seq-like profiles. The input to the model are the measured 978 landmark genes while the output is a vector of 23,614 RNA-seq-like gene expression profiles. The model first transforms the landmark genes into RNA-seq-like 978 gene profiles using a modified CycleGAN model applied to unpaired data. The transformed 978 RNA-seq-like landmark genes are then extrapolated into the full genome space with a fully connected neural network model. The two-step model achieves 0.914 Pearson’s correlation coefficients and 1.167 root mean square errors when tested on a published paired L1000/RNA-seq dataset produced by the LINCS and GTEx programs. The processed RNA-seq-like profiles are made available for download, signature search, and gene centric reverse search with unique case studies.

Deciphering the role of predicted miRNAs of polyomaviruses in carcinogenesis

Human polyomaviruses are relatively common in the general population. Polyomaviruses maintain a persistent infection after initial infection in childhood, acting as an opportunistic pathogen in immunocompromised populations and their association has been linked to carcinogenesis. A comprehensive understanding of the underlying molecular mechanisms of carcinogenesis in consequence of polyomavirus infection remains elusive. However, the critical role of viral miRNAs and their potential targets in modifying the transcriptome profile of the host remains largely unknown. Polyomavirus-derived miRNAs have the potential to play a substantial role in carcinogenesis. Employing computational approaches, putative viral miRNAs along with their target genes have been predicted and possible roles of the targeted genes in many significant biological processes have been obtained. Polyomaviruses have been observed to target intracellular signal transduction pathways through miRNA-mediated epigenetic regulation, which may contribute to cancer development. In addition, BKPyV-infected human renal cell microarray data was coupled with predicted target genes and analysis of the downregulated genes indicated that viruses target multiple signaling pathways (e.g. MAPK signaling pathway, PI3K-Akt signaling pathway, PPAR signaling pathway) in the host as well as turning off several tumor suppression genes (e.g. FGGY, EPHX2, CACNA2D3, CDH16) through miRNA-induced mechanisms, assuring cell transformation. This study provides a conceptual framework for the underlying molecular mechanisms involved in the course of carcinogenesis upon polyomavirus infection.

The Importance of CXCL1 in Physiology and Noncancerous Diseases of Bone, Bone Marrow, Muscle and the Nervous System

This review describes the role of CXCL1, a chemokine crucial in inflammation as a chemoattractant for neutrophils, in physiology and in selected major non-cancer diseases. Due to the vast amount of available information, we focus on the role CXCL1 plays in the physiology of bones, bone marrow, muscle and the nervous system. For this reason, we describe its effects on hematopoietic stem cells, myoblasts, oligodendrocyte progenitors and osteoclast precursors. We also present the involvement of CXCL1 in diseases of selected tissues and organs including Alzheimer’s disease, epilepsy, herpes simplex virus type 1 (HSV-1) encephalitis, ischemic stroke, major depression, multiple sclerosis, neuromyelitis optica, neuropathic pain, osteoporosis, prion diseases, rheumatoid arthritis, tick-borne encephalitis (TBE), traumatic spinal cord injury and West Nile fever.

Human pancreatic tumour organoid-derived factors enhance myogenic differentiation

BACKGROUND

Most patients with pancreatic cancer develop cachexia, which is characterized by progressive muscle loss. The mechanisms underlying muscle loss in cancer cachexia remain elusive. Pancreatic tumour organoids are 3D cell culture models that retain key characteristics of the parent tumour. We aimed to investigate the effect of pancreatic tumour organoid-derived factors on processes that determine skeletal muscle mass, including the regulation of muscle protein turnover and myogenesis.

METHODS

Conditioned medium (CM) was collected from human pancreatic cancer cell lines (PK-45H, PANC-1, PK-1, and KLM-1), pancreatic tumour organoid cultures from a severely cachectic (PANCO-9a) and a non-cachectic patient (PANCO-12a), and a normal pancreas organoid culture. Differentiating C2C12 myoblasts and mature C2C12 myotubes were exposed to CM for 24 h or maintained in control medium. In myotubes, NF-kB activation was monitored using a NF-κB luciferase reporter construct, and mRNA expression of E3-ubiquitin ligases and REDD1 was analysed by RT-qPCR. C2C12 myoblast proliferation and differentiation were monitored by live cell imaging and myogenic markers and myosin heavy chain (MyHC) isoforms were assessed by RT-qPCR.

RESULTS

Whereas CM from PK-1 and KLM-1 cells significantly induced NF-κB activation in C2C12 myotubes (PK-1: 3.1-fold, P < 0.001; KLM-1: 2.1-fold, P = 0.01), Atrogin-1/MAFbx and MuRF1 mRNA were only minimally and inconsistently upregulated by the CM of pancreatic cancer cell lines. Similarly, E3-ubiquitin ligases and REDD1 mRNA expression in myotubes were not altered by exposure to pancreatic tumour organoid CM. Compared with the control condition, CM from both PANCO-9a and PANCO-12a tumour organoids increased proliferation of myoblasts, which was accompanied by significant downregulation of the satellite cell marker paired-box 7 (PAX7) (PANCO-9a: -2.1-fold, P < 0.001; PANCO-12a: -2.0-fold, P < 0.001) and myogenic factor 5 (MYF5) (PANCO-9a: -2.1-fold, P < 0.001; PANCO-12a: -1.8-fold, P < 0.001) after 48 h of differentiation. Live cell imaging revealed accelerated alignment and fusion of myoblasts exposed to CM from PANCO-9a and PANCO-12a, which was in line with significantly increased Myomaker mRNA expression levels (PANCO-9a: 2.4-fold, P = 0.001; PANCO-12a: 2.2-fold, P = 0.004). These morphological and transcriptional alterations were accompanied by increased expression of muscle differentiation markers such as MyHC-IIB (PANCO-9a: 2.5-fold, P = 0.04; PANCO-12a: 3.1-fold, P = 0.006). Although the impact of organoid CM on myogenesis was not associated with the cachexia phenotype of the donor patients, it was specific for tumour organoids, as CM of control pancreas organoids did not modulate myogenic fusion.

CONCLUSIONS

These data show that pancreatic tumour organoid-derived factors alter the kinetics of myogenesis, which may eventually contribute to impaired muscle mass maintenance in cancer cachexia.

Quantification of Muscle Stem Cell Differentiation Using Live-Cell Imaging and Eccentricity Measures

Culturing primary muscle stem cells ex vivo is a useful method for studying this cell population in controlled environments. Primary muscle stem cells respond to external stimuli differently than immortalized myoblasts (C2C12 cells), making ex vivo culture of muscle stem cells an important tool in understanding cell responses to stimuli. Primary muscle stem cells cultured ex vivo retain a majority of the characteristics they possess in vivo such as the abilities to differentiate into multinucleated structures, and self-renew a stem cell-like population. In this chapter, we describe methods for isolating primary muscle stem cells, controlled differentiation into myotubes, and quantification of differentiation using IncuCyte live cell imaging and analysis software.

Disrupted NOS2 metabolism drives myoblast response to wasting-associated cytokines

Skeletal muscle wasting drives negative clinical outcomes and is associated with a spectrum of pathologies including cancer. Cancer cachexia is a multi-factorial syndrome that encompasses skeletal muscle wasting and remains understudied, despite being a frequent and serious co-morbidity. Deviation from the homeostatic balance between breakdown and regeneration leads to muscle wasting disorders, such as cancer cachexia. Muscle stem cells (MuSCs) are the cellular compartment responsible for muscle regeneration, which makes MuSCs an intriguing target in the context of wasting muscle. Molecular studies investigating MuSCs and skeletal muscle wasting largely focus on transcriptional changes, but our group and others propose that metabolic changes are another layer of cellular regulation underlying MuSC dysfunction in cancer cachexia. In the present study, we combined gene expression and non-targeted metabolomic profiling of myoblasts exposed to wasting conditions (cancer cell conditioned media, CC-CM) to derive a more complete picture of the myoblast response to wasting factors. After mapping these features to annotated pathways, we found that more than half of the mapped pathways were amino acid-related, linking global amino acid metabolic disruption to conditioned media-induced myoblast defects. Notably, arginine metabolism was a highly enriched pathway in combined metabolomic and transcriptomic data. Arginine catabolism generates nitric oxide (NO), an important signaling molecule known to have negative effects on mature muscle. We hypothesize that tumor-derived disruptions in Nitric Oxide Synthase (NOS)2-regulated arginine catabolism impair differentiation of MuSCs. The work presented here further investigates the effect of NOS2 overactivity on myoblast proliferation and differentiation. We show that NOS2 inhibition is sufficient to rescue wasting phenotypes associated with inflammatory cytokines. Ultimately, this work provides new insights into MuSC biology and opens up potential therapeutic avenues for addressing disrupted MuSC dynamics in cancer cachexia.

Hormonally Regulated Myogenic miR-486 Influences Sex-specific Differences in Cancer-induced Skeletal Muscle Defects

Cancer-induced skeletal muscle defects show sex-specific differences in severity with men performing poorly compared to women. Hormones and sex chromosomal differences are suggested to mediate these differences, but the functional skeletal muscle markers to document these differences are unknown. We show that the myogenic microRNA miR-486 is a marker of sex-specific differences in cancer-induced skeletal muscle defects. Cancer-induced loss of circulating miR-486 was more severe in men with bladder, lung, and pancreatic cancers compared to women with the same cancer types. In a syngeneic model of pancreatic cancer, circulating and skeletal muscle loss of miR-486 was more severe in male mice compared to female mice. Estradiol (E2) and the clinically used selective estrogen receptor modulator toremifene increased miR-486 in undifferentiated and differentiated myoblast cell line C2C12 and E2-inducible expression correlated with direct binding of estrogen receptor alpha (ERα) to the regulatory region of the miR-486 gene. E2 and toremifene reduced the actions of cytokines such as myostatin, transforming growth factor β, and tumor necrosis factor α, which mediate cancer-induced skeletal muscle wasting. E2- and toremifene-treated C2C12 myoblast/myotube cells contained elevated levels of active protein kinase B (AKT) with a corresponding decrease in the levels of its negative regulator PTEN, which is a target of miR-486. We propose an ERα:E2-miR-486-AKT signaling axis, which reduces the deleterious effects of cancer-induced cytokines/chemokines on skeletal muscle mass and/or function.

Role of Cytokines and Chemokines in NSCLC Immune Navigation and Proliferation

With over a million deaths every year around the world, lung cancer is found to be the most recurrent cancer among all types. Nonsmall cell lung carcinoma (NSCLC) amounts to about 85% of the entire cases. The other 15% owes it to small cell lung carcinoma (SCLC). Despite decades of research, the prognosis for NSCLC patients is poorly understood with treatment options limited. First, this article emphasises on the part that tumour microenvironment (TME) and its constituents play in lung cancer progression. This review also highlights the inflammatory (pro- or anti-) roles of different cytokines (ILs, TGF-β, and TNF-α) and chemokine (CC, CXC, C, and CX3C) families in the lung TME, provoking tumour growth and subsequent metastasis. The write-up also pinpoints recent developments in the field of chemokine biology. Additionally, it covers the role of extracellular vesicles (EVs), as alternate carriers of cytokines and chemokines. This allows the cytokines/chemokines to modulate the EVs for their secretion, trafficking, and aid in cancer proliferation. In the end, this review also stresses on the role of these factors as prognostic biomarkers for lung immunotherapy, apart from focusing on inflammatory actions of these chemoattractants.

Skeletal Muscle and Childhood Cancer: Where are we now and where we go from here

Skeletal muscle (muscle) is essential for physical health and for metabolic integrity, with sarcopenia (progressive muscle mass loss and weakness), a pre-curser of aging and chronic disease. Loss of lean mass and muscle quality (force generation per unit of muscle) in the general population are associated with fatigue, weakness, and slowed walking speed, eventually interfering with the ability to maintain physical independence, and impacting participation in social roles and quality of life. Muscle mass and strength impairments are also documented during childhood cancer treatment, which often persist into adult survivorship, and contribute to an aging phenotype in this vulnerable population. Although several treatment exposures appear to confer increased risk for loss of mass and strength that persists after therapy, the pathophysiology responsible for poor muscle quantity and quality is not well understood in the childhood cancer survivor population. This is partly due to limited access to both pediatric and adult survivor muscle tissue samples, and to difficulties surrounding non-invasive investigative approaches for muscle assessment. Because muscle accounts for just under half of the body's mass, and is essential for movement, metabolism and metabolic health, understanding mechanisms of injury responsible for both initial and persistent dysfunction is important, and will provide a foundation for intervention. The purpose of this review is to provide an overview of the available evidence describing associations between childhood cancer, its treatment, and muscle outcomes, identifying gaps in current knowledge.

Increased myocellular lipid and IGFBP-3 expression in a pre-clinical model of pancreatic cancer-related skeletal muscle wasting

BACKGROUND

Skeletal muscle wasting (SMW) in cancer patients is associated with increased morbidity, mortality, treatment intolerance and discontinuation, and poor quality of life. This is particularly true for patients with pancreatic ductal adenocarcinoma (PDAC), as over 85% experience SMW, which is responsible for ~30% of patient deaths. While the established paradigm to explain SMW posits that muscle catabolism from systemic inflammation and nutritional deficiencies, the cause of death, and the cellular and molecular mechanisms responsible remain to be elucidated. To address this, we investigated the relationship between tumour burden and survival in the KCKO murine PDAC model.

METHODS

Female C57BL/6J mice 6-8 weeks of age underwent orthotopic injection with KCKO-luc tumour cells. Solid tumour was verified on Day 5, post-tumour inoculation. In vivo, longitudinal lean mass and tumour burden were assessed via dual-energy X-ray absorptiometry and IVIS imaging, respectively, and total body weight was assessed, weekly. Animals were sacrificed at a designated end point of 'failure to thrive'. After sacrifice, lower limb hind muscles were harvested for histology and RNA extraction.

RESULTS

We found a strong correlation between primary tumour size and survival (r2  = 0.83, P < 0.0001). A significant decrease in lower limb lean mass was first detected at Day 38 post-implantation vs. no tumour controls (NTCs) (P < 0.0001). SMW was confirmed by histology, which demonstrated a 38%, 32.7%, and 39.9% decrease in fibre size of extensor digitorum longus, soleus, and tibialis anterior muscles, respectively, in PDAC mice vs. NTC (P < 0.002). Histology also revealed a 67.6% increase in haematopoietic cells within the muscle of PDAC mice when compared with NTC. Bulk RNAseq on muscles from PDAC mice vs. NTC revealed significant increases in c/ebpβ/Δ, il-1, il-6, and tnf gene expression. Pathway analyses to identify potential upstream factors revealed increased adipogenic gene expression, including a four-fold increase in igfbp-3. Histomorphometry of Oil Red-O staining for fat content in tibialis anterior muscles demonstrated a 95.5% increase in positively stained fibres from PDAC mice vs. NTC.

CONCLUSIONS

Together, these findings support a novel model of PDAC-associated SMW and mortality in which systemic inflammation leads to inflammatory cell infiltration into skeletal muscle with up-regulated myocellular lipids.

A Novel Technique for Radiographic Diagnosis of Sarcopenia that Accurately Predicts Postoperative Complications in Lower Extremity Free Flap Patients

INTRODUCTION

 Sarcopenia is linked to poor outcomes throughout the surgical literature and can be assessed on preoperative imaging to potentially aid in risk stratification. This study examined the effects of sarcopenia on surgical morbidity following lower extremity (LE) reconstruction, and also compared two methods of assessment, one of which is novel ("ellipse method").

METHODS

 A retrospective cohort study of 50 patients receiving free flap-based reconstruction of the LE was performed. Bilateral psoas density and area were quantified at L4 through tracing ("traditional method") and encircling ("ellipse method") to calculate Hounsfield unit average calculation (HUAC). Logistic regression and receiving operator curve analysis for the primary outcome of any postoperative complication was used to determine HUAC cutoffs (≤ 20.7 vs. ≤ 20.6) for sarcopenia. Risk of complications associated with sarcopenia was evaluated using Fisher's exact tests.

RESULTS

 Twelve patients (24%) met criteria for sarcopenia via the traditional method and 16 (32%) via the ellipse method. By both methods, sarcopenic patients were older and more often female and diabetic. These patients also had higher American Society of Anesthesiologists scores and lower serum prealbumin levels. The ellipse method was found to be more accurate, sensitive, and specific than the traditional method in predicting postoperative morbidity (p = 0.009). Via the ellipse method, sarcopenic patients were at higher risk for any complication (p = 0.002) and were at a higher risk for a deep vein thrombus or pulmonary embolism via the traditional method (p = 0.047).

CONCLUSION

 Sarcopenia is associated with greater pre- and postoperative morbidity in LE reconstruction. The novel ellipse method is a simplified and accurate method of assessing sarcopenia that can be easily performed in the clinical setting.

Mitochondria-cytokine crosstalk following skeletal muscle injury and disuse: a mini-review

Skeletal muscle mitochondria are highly adaptable, highly dynamic organelles that maintain the functional integrity of the muscle fiber by providing ATP for contraction and cellular homeostasis (e.g., Na⁺/K⁺ ATPase). Emerging as early modulators of inflammation, mitochondria sense and respond to cellular stress. Mitochondria communicate with the environment, in part, by release of physical signals called mitochondrial-derived damage-associated molecular patterns (mito-DAMPs) and deviation from routine function (e.g., reduced ATP production, Ca²⁺ overload). When skeletal muscle is compromised, mitochondria contribute to an acute inflammatory response necessary for myofibril regeneration; however, exhaustive signaling associated with altered or reduced mitochondrial function can be detrimental to muscle outcomes. Here, we describe changes in mitochondrial content, structure, and function following skeletal muscle injury and disuse and highlight the influence of mitochondria-cytokine crosstalk on muscle regeneration and recovery. Although the appropriate therapeutic modulation following muscle stressors remains unknown, retrospective gene expression analysis reveals that interleukin-6 (IL-6), interleukin-1β (IL-1β), chemokine C-X-C motif ligand 1 (CXCL1), and monocyte chemoattractant protein 1 (MCP-1) are significantly upregulated following three unique muscle injuries. These cytokines modulate mitochondrial function and execute bona fide pleiotropic roles that can aid functional recovery of muscle, however, when aberrant, chronically disrupt healing partly by exacerbating mitochondrial dysfunction. Multidisciplinary efforts to delineate the opposing regulatory roles of inflammatory cytokines in the muscle mitochondrial environment are required to modulate regenerative behavior following skeletal muscle injury or disuse. Future therapeutic directions to consider include quenching or limited release of mito-DAMPs and cytokines present in cytosol or circulation.

Cancer-Associated Muscle Wasting-Candidate Mechanisms and Molecular Pathways

Excessive muscle loss is commonly observed in cancer patients and its association with poor prognosis has been well-established. Cancer-associated sarcopenia differs from age-related wasting in that it is not responsive to nutritional intervention and exercise. This is related to its unique pathogenesis, a result of diverse and interconnected mechanisms including inflammation, disordered metabolism, proteolysis and autophagy. There is a growing body of evidence that suggests that the tumor is the driver of muscle wasting by its elaboration of mediators that influence each of these pro-sarcopenic pathways. In this review, evidence for these tumor-derived factors and putative mechanisms for inducing muscle wasting will be reviewed. Potential targets for future research and therapeutic interventions will also be reviewed.

Clostridium perfringens α-toxin inhibits myogenic differentiation of C2C12 myoblasts

Clostridium perfringens type A is the causative agent of clostridial myonecrosis, and α-toxin has been reported to be responsible for the pathogenesis. Recently, it was reported that regeneration of skeletal muscle after C. perfringens-induced muscle disorders is delayed, but the detailed mechanisms have not been elucidated. Here, we tested whether α-toxin impairs the differentiation of C2C12 myoblasts, a useful cell line to study muscle growth, maturation, and regeneration in vitro. α-Toxin dose-dependently inhibited myotube formation in C2C12 cultures after induction of their differentiation by horse serum. Also, immunoblot analysis revealed that α-toxin dose-dependently decreases the expressions of two skeletal muscle differentiation markers, myogenic differentiation 1 (MyoD) and myogenin. These results demonstrate that α-toxin impairs the myogenic differentiation of C2C12 myoblasts. To reveal the mechanism behind α-toxin-mediated impairment of myogenic differentiation, we focused on ceramide production since α-toxin is known to promote the formation of ceramide by its sphingomyelinase activity. Immunofluorescent analysis revealed that ceramide production is accelerated by treatment with α-toxin. Furthermore, a synthetic cell-permeable ceramide analog, C₂-ceramide, inhibited myotube formation in C2C12 cells and decreased the expressions of MyoD and myogenin, suggesting that accelerated ceramide production is involved in the α-toxin-mediated blockage of myogenic differentiation. Together, our results illustrate that the impairment of myogenic differentiation by α-toxin might be crucial for the pathogenesis of C. perfringens to delay regeneration of severely damaged skeletal muscles.

The wasting-associated metabolite succinate disrupts myogenesis and impairs skeletal muscle regeneration

BACKGROUND

Muscle wasting is a debilitating co-morbidity affecting most advanced cancer patients. Alongside enhanced muscle catabolism, defects in muscle repair/regeneration contribute to cancer-associated wasting. Among the factors implicated in suppression of muscle regeneration are cytokines that interfere with myogenic signal transduction pathways. Less understood is how other cancer/wasting-associated cues, such as metabolites, contribute to muscle dysfunction. This study investigates how the metabolite succinate affects myogenesis and muscle regeneration.

METHODS

We leveraged an established ectopic metabolite treatment (cell permeable dimethyl-succinate) strategy to evaluate the ability of intracellular succinate elevation to 1) affect myoblast homeostasis (proliferation, apoptosis), 2) disrupt protein dynamics and induce wasting-associated atrophy, and 3) modulate in vitro myogenesis. In vivo succinate supplementation experiments (2% succinate, 1% sucrose vehicle) were used to corroborate and extend in vitro observations. Metabolic profiling and functional metabolic studies were then performed to investigate the impact of succinate elevation on mitochondria function.

RESULTS

We found that in vitro succinate supplementation elevated intracellular succinate about 2-fold, and did not have an impact on proliferation or apoptosis of C2C12 myoblasts. Elevated succinate had minor effects on protein homeostasis (~25% decrease in protein synthesis assessed by OPP staining), and no significant effect on myotube atrophy. Succinate elevation interfered with in vitro myoblast differentiation, characterized by significant decreases in late markers of myogenesis and fewer nuclei per myosin heavy chain positive structure (assessed by immunofluorescence staining). While mice orally administered succinate did not exhibit changes in overall body composition or whole muscle weights, these mice displayed smaller muscle myofiber diameters (~6% decrease in the mean of non-linear regression curves fit to the histograms of minimum feret diameter distribution), which was exacerbated when muscle regeneration was induced with barium chloride injury. Significant decreases in the mean of non-linear regression curves fit to the histograms of minimum feret diameter distributions were observed 7 days and 28 days post injury. Elevated numbers of myogenin positive cells (3-fold increase) supportive of the differentiation defects observed in vitro were observed 28 days post injury. Metabolic profiling and functional metabolic assessment of myoblasts revealed that succinate elevation caused both widespread metabolic changes and significantly lowered maximal cellular respiration (~35% decrease).

CONCLUSIONS

This study broadens the repertoire of wasting-associated factors that can directly modulate muscle progenitor cell function and strengthens the hypothesis that metabolic derangements are significant contributors to impaired muscle regeneration, an important aspect of cancer-associated muscle wasting.

Distinct cachexia profiles in response to human pancreatic tumours in mouse limb and respiratory muscle

BACKGROUND

Cancer cachexia is a life-threatening metabolic syndrome that causes significant loss of skeletal muscle mass and significantly increases mortality in cancer patients. Currently, there is an urgent need for better understanding of the molecular pathophysiology of this disease so that effective therapies can be developed. The majority of pre-clinical studies evaluating skeletal muscle's response to cancer have focused on one or two pre-clinical models, and almost all have focused specifically on limb muscles. In the current study, we reveal key differences in the histology and transcriptomic signatures of a limb muscle and a respiratory muscle in orthotopic pancreatic cancer patient-derived xenograft (PDX) mice.

METHODS

To create four cohorts of PDX mice evaluated in this study, tumours resected from four pancreatic ductal adenocarcinoma patients were portioned and attached to the pancreas of immunodeficient NSG mice.

RESULTS

Body weight, muscle mass, and fat mass were significantly decreased in each PDX line. Histological assessment of cryosections taken from the tibialis anterior (TA) and diaphragm (DIA) revealed differential effects of tumour burden on their morphology. Subsequent genome-wide microarray analysis on TA and DIA also revealed key differences between their transcriptomes in response to cancer. Genes up-regulated in the DIA were enriched for extracellular matrix protein-encoding genes and genes related to the inflammatory response, while down-regulated genes were enriched for mitochondria related protein-encoding genes. Conversely, the TA showed up-regulation of canonical atrophy-associated pathways such as ubiquitin-mediated protein degradation and apoptosis, and down-regulation of genes encoding extracellular matrix proteins.

CONCLUSIONS

These data suggest that distinct biological processes may account for wasting in different skeletal muscles in response to the same tumour burden. Further investigation into these differences will be critical for the future development of effective clinical strategies to counter cancer cachexia.

Organ System Crosstalk in Cardiometabolic Disease in the Age of Multimorbidity

The close association among cardiovascular, metabolic, and kidney diseases suggests a common pathological basis and significant interaction among these diseases. Metabolic syndrome and cardiorenal syndrome are two examples that exemplify the interlinked development of disease or dysfunction in two or more organs. Recent studies have been sorting out the mechanisms responsible for the crosstalk among the organs comprising the cardiovascular, metabolic, and renal systems, including heart-kidney and adipose-liver signaling, among many others. However, it is also becoming clear that this crosstalk is not limited to just pairs of organs, and in addition to organ-organ crosstalk, there are also organ-system and organ-body interactions. For instance, heart failure broadly impacts various organs and systems, including the kidney, liver, lung, and nervous system. Conversely, systemic dysregulation of metabolism, immunity, and nervous system activity greatly affects heart failure development and prognosis. This is particularly noteworthy, as more and more patients present with two or more coexisting chronic diseases or conditions (multimorbidity) due in part to the aging of society. Advances in treatment also contribute to the increase in multimorbidity, as exemplified by cardiovascular disease in cancer survivors. To understand the mechanisms underlying the increasing burden of multimorbidity, it is vital to elucidate the multilevel crosstalk and communication within the body at the levels of organ systems, tissues, and cells. In this article, we focus on chronic inflammation as a key common pathological basis of cardiovascular and metabolic diseases, and discuss emerging mechanisms that drive chronic inflammation in the context of multimorbidity.

Inflamma-miR-21 Negatively Regulates Myogenesis during Ageing

Ageing is associated with disrupted redox signalling and increased circulating inflammatory cytokines. Skeletal muscle homeostasis depends on the balance between muscle hypertrophy, atrophy and regeneration, however during ageing this balance is disrupted. The molecular pathways underlying the age-related decline in muscle regenerative potential remain elusive. microRNAs are conserved robust gene expression regulators in all tissues including skeletal muscle. Here, we studied satellite cells from adult and old mice to demonstrate that inhibition of miR-21 in satellite cells from old mice improves myogenesis. We determined that increased levels of proinflammatory cytokines, TNFα and IL6, as well as H₂O₂, increased miR-21 expression in primary myoblasts, which in turn resulted in their decreased viability and myogenic potential. Inhibition of miR-21 function rescued the decreased size of myotubes following TNFα or IL6 treatment. Moreover, we demonstrated that miR-21 could inhibit myogenesis in vitro via regulating IL6R, PTEN and FOXO3 signalling. In summary, upregulation of miR-21 in satellite cells and muscle during ageing may occur in response to elevated levels of TNFα and IL6, within satellite cells or myofibrillar environment contributing to skeletal muscle ageing and potentially a disease-related decline in potential for muscle regeneration.

A novel transplantable model of lung cancer-associated tissue loss and disrupted muscle regeneration

BACKGROUND

Cancer-associated muscle wasting (CAW), a symptom of cancer cachexia, is associated with approximately 20% of lung cancer deaths and remains poorly characterized on a mechanistic level. Current animal models for lung cancer-associated cachexia are limited in that they (1) primarily employ flank transplantation methods, (2) have short survival times not reflective of the patient condition, and (3) are typically performed in young mice not representative of mean patient age. This study investigates a new model for lung cancer-associated cachexia that can address these issues and also implicates muscle regeneration as a contributor to CAW.

METHODS

We used tail vein injection as a method to introduce tumor cells that seed primarily in the lungs of mice. Body composition of tumor-bearing mice was longitudinally tracked using NMR-based, echo magnetic resonance imaging (echoMRI). These data were combined with histological and molecular assessments of skeletal muscle to provide a complete analysis of muscle wasting.

RESULTS

In this new lung CAW model, we observed (1) progressive loss in whole body weight, (2) progressive loss of lean and fat mass, (3) a circulating cytokine/inflammatory profile similar to that seen in other models of CAW, (4) histological changes associated with muscle wasting, and (5) molecular changes in muscle that implicate suppression of muscle repair/regeneration. Finally, we show that survival can be extended without lessening CAW by titrating injected cell number.

CONCLUSIONS

Overall, this study describes a new model of CAW that could be useful for further studies of lung cancer-associated wasting and accompanying changes in the regenerative capacity of muscle. Additionally, this model addresses many recent concerns with existing models such as immunocompetence, tumor location, and survival time.

Systemic Actions of Breast Cancer Facilitate Functional Limitations

Breast cancer is a disease of a specific organ, but its effects are felt throughout the body. The systemic effects of breast cancer can lead to functional limitations in patients who suffer from muscle weakness, fatigue, pain, fibromyalgia, or many other dysfunctions, which hasten cancer-associated death. Mechanistic studies have identified quite a few molecular defects in skeletal muscles that are associated with functional limitations in breast cancer. These include circulating cytokines such as TNF-α, IL-1, IL-6, and TGF-β altering the levels or function of myogenic molecules including PAX7, MyoD, and microRNAs through transcriptional regulators such as NF-κB, STAT3, and SMADs. Molecular defects in breast cancer may also include reduced muscle mitochondrial content and increased extracellular matrix deposition leading to energy imbalance and skeletal muscle fibrosis. This review highlights recent evidence that breast cancer-associated molecular defects mechanistically contribute to functional limitations and further provides insights into therapeutic interventions in managing functional limitations, which in turn may help to improve quality of life in breast cancer patients.

PIMKL: Pathway-Induced Multiple Kernel Learning

Reliable identification of molecular biomarkers is essential for accurate patient stratification. While state-of-the-art machine learning approaches for sample classification continue to push boundaries in terms of performance, most of these methods are not able to integrate different data types and lack generalization power, limiting their application in a clinical setting. Furthermore, many methods behave as black boxes, and we have very little understanding about the mechanisms that lead to the prediction. While opaqueness concerning machine behavior might not be a problem in deterministic domains, in health care, providing explanations about the molecular factors and phenotypes that are driving the classification is crucial to build trust in the performance of the predictive system. We propose Pathway-Induced Multiple Kernel Learning (PIMKL), a methodology to reliably classify samples that can also help gain insights into the molecular mechanisms that underlie the classification. PIMKL exploits prior knowledge in the form of a molecular interaction network and annotated gene sets, by optimizing a mixture of pathway-induced kernels using a Multiple Kernel Learning (MKL) algorithm, an approach that has demonstrated excellent performance in different machine learning applications. After optimizing the combination of kernels to predict a specific phenotype, the model provides a stable molecular signature that can be interpreted in the light of the ingested prior knowledge and that can be used in transfer learning tasks.

The reliable classification of biomedical samples to predict phenotypic differences requires not only robust methods, but also interpretable approaches that can explain the reasons behind a prediction. A team led by María Rodríguez Martínez at IBM Research - Zürich has developed PIMKL, a methodology that exploits prior knowledge and enables the integration of multiple types of data with varying predictive power. Even when noisy datasets are simultaneously analyzed, PIMKL is able to discard uninformative data and achieve strong prediction power. Importantly, PIMKL produces a molecular signature that enables the interpretation of the results in terms of known biological functions. This signature can be transferred to other cohorts without loss of performance, demonstrating surprising robustness across cohorts. Interpretable algorithms can effectively help gain insights about disease mechanisms and build trust in a model.