DeePaN: deep patient graph convolutional network integrating clinico-genomic evidence to stratify lung cancers for immunotherapy.

Immuno-oncology (IO) therapies have transformed the therapeutic landscape of non-small cell lung cancer (NSCLC). However, patient responses to IO are variable and influenced by a heterogeneous combination of health, immune, and tumor factors. There is a pressing need to discover the distinct NSCLC subgroups that influence response. We have developed a deep patient graph convolutional network, we call "DeePaN", to discover NSCLC complexity across data modalities impacting IO benefit. DeePaN employs high-dimensional data derived from both real-world evidence (RWE)-based electronic health records (EHRs) and genomics across 1937 IO-treated NSCLC patients. DeePaN demonstrated effectiveness to stratify patients into subgroups with significantly different (P-value of 2.2 × 10-11) overall median survival of 20.35 months and 9.42 months post-IO therapy. Significant differences in IO outcome were not seen from multiple non-graph-based unsupervised methods. Furthermore, we demonstrate that patient stratification from DeePaN has the potential to augment the emerging IO biomarker of tumor mutation burden (TMB). Characterization of the subgroups discovered by DeePaN indicates potential to inform IO therapeutic insight, including the enrichment of mutated KRAS and high blood monocyte count in the IO beneficial and IO non-beneficial subgroups, respectively. Our work has proven the concept that graph-based AI is feasible and can effectively integrate high-dimensional genomic and EHR data to meaningfully stratify cancer patients on distinct clinical outcomes, with potential to inform precision oncology.

Small molecule AZD4635 inhibitor of A2AR signaling rescues immune cell function including CD103+ dendritic cells enhancing anti-tumor immunity.

Accumulation of extracellular adenosine within the microenvironment is a strategy exploited by tumors to escape detection by the immune system. Adenosine signaling through the adenosine 2A receptor (A2AR) on immune cells elicits a range of immunosuppressive effects which promote tumor growth and limit the efficacy of immune checkpoint inhibitors. Preclinical data with A2AR inhibitors have demonstrated tumor regressions in mouse models by rescuing T cell function; however, the mechanism and role on other immune cells has not been fully elucidated. METHODS: We report here the development of a small molecule A2AR inhibitor including characterization of binding and inhibition of A2AR function with varying amounts of a stable version of adenosine. Functional activity was tested in both mouse and human T cells and dendritic cells (DCs) in in vitro assays to understand the intrinsic role on each cell type. The role of adenosine and A2AR inhibition was tested in DC differentiation assays as well as co-culture assays to access the cross-priming function of DCs. Syngeneic models were used to assess tumor growth alone and in combination with alphaprogrammed death-ligand 1 (αPD-L1). Immunophenotyping by flow cytometry was performed to examine global immune cell changes upon A2AR inhibition. RESULTS: We provide the first report of AZD4635, a novel small molecule A2AR antagonist which inhibits downstream signaling and increases T cell function as well as a novel mechanism of enhancing antigen presentation by CD103+ DCs. The role of antigen presentation by DCs, particularly CD103+ DCs, is critical to drive antitumor immunity providing rational to combine a priming agent AZD4635 with check point blockade. We find adenosine impairs the maturation and antigen presentation function of CD103+ DCs. We show in multiple syngeneic mouse tumor models that treatment of AZD4635 alone and in combination with αPD-L1 led to decreased tumor volume correlating with enhanced CD103+ function and T cell response. We extend these studies into human DCs to show that adenosine promotes a tolerogenic phenotype that can be reversed with AZD4635 restoring antigen-specific T cell activation. Our results support the novel role of adenosine signaling as an intrinsic negative regulator of CD103+ DCs maturation and priming. We show that potent inhibition of A2AR with AZD4635 reduces tumor burden and enhances antitumor immunity. This unique mechanism of action in CD103+ DCs may contribute to clinical responses as AZD4635 is being evaluated in clinical trials with IMFINZI (durvalumab, αPD-L1) in patients with solid malignancies. CONCLUSION: We provide evidence implicating suppression of adaptive and innate immunity by adenosine as a mechanism for immune evasion by tumors. Inhibition of adenosine signaling through selective small molecule inhibition of A2AR using AZD4635 restores T cell function via an internal mechanism as well as tumor antigen cross-presentation by CD103+ DCs resulting in antitumor immunity.

Adenosine Signaling Is Prognostic for Cancer Outcome and Has Predictive Utility for Immunotherapeutic Response.

PURPOSE: There are several agents in early clinical trials targeting components of the adenosine pathway including A2AR and CD73. The identification of cancers with a significant adenosine drive is critical to understand the potential for these molecules. However, it is challenging to measure tumor adenosine levels at scale, thus novel, clinically tractable biomarkers are needed. EXPERIMENTAL DESIGN: We generated a gene expression signature for the adenosine signaling using regulatory networks derived from the literature and validated this in patients. We applied the signature to large cohorts of disease from The Cancer Genome Atlas (TCGA) and cohorts of immune checkpoint inhibitor-treated patients. RESULTS: The signature captures baseline adenosine levels in vivo (r 2 = 0.92, P = 0.018), is reduced after small-molecule inhibition of A2AR in mice (r 2 = -0.62, P = 0.001) and humans (reduction in 5 of 7 patients, 70%), and is abrogated after A2AR knockout. Analysis of TCGA confirms a negative association between adenosine and overall survival (OS, HR = 0.6, P < 2.2e-16) as well as progression-free survival (PFS, HR = 0.77, P = 0.0000006). Further, adenosine signaling is associated with reduced OS (HR = 0.47, P < 2.2e-16) and PFS (HR = 0.65, P = 0.0000002) in CD8+ T-cell-infiltrated tumors. Mutation of TGFß superfamily members is associated with enhanced adenosine signaling and worse OS (HR = 0.43, P < 2.2e-16). Finally, adenosine signaling is associated with reduced efficacy of anti-PD1 therapy in published cohorts (HR = 0.29, P = 0.00012). CONCLUSIONS: These data support the adenosine pathway as a mediator of a successful antitumor immune response, demonstrate the prognostic potential of the signature for immunotherapy, and inform patient selection strategies for adenosine pathway modulators currently in development.

Acalabrutinib monotherapy in patients with chronic lymphocytic leukemia who are intolerant to ibrutinib.

The Bruton tyrosine kinase (BTK) inhibitor ibrutinib improves patient outcomes in chronic lymphocytic leukemia (CLL); however, some patients experience adverse events (AEs) leading to discontinuation. Acalabrutinib is a potent, covalent BTK inhibitor with greater selectivity than ibrutinib. We evaluated the safety and efficacy of 100 mg of acalabrutinib twice daily or 200 mg once daily in patients with CLL who discontinued ibrutinib because of intolerance as determined by the investigators. Among 33 treated patients (61% men; median age, 64 years; range, 50-82 years), median duration of prior ibrutinib treatment was 11.6 months (range, 1-62 months); median time from ibrutinib discontinuation to acalabrutinib start was 47 days (range, 3-331 days). After a median of 19.0 months (range, 0.2-30.6 months), 23 patients remained on acalabrutinib; 10 had discontinued (progressive disease, n = 4; AEs, n = 3). No acalabrutinib dose reductions occurred. During acalabrutinib treatment, the most frequent AEs included diarrhea (58%), headache (39%), and cough (33%). Grade 3/4 AEs occurred in 58%, most commonly neutropenia (12%) and thrombocytopenia (9%). Of 61 ibrutinib-related AEs associated with intolerance, 72% did not recur and 13% recurred at a lower grade with acalabrutinib. Overall response rate was 76%, including 1 complete and 19 partial responses and 5 partial responses with lymphocytosis. Among 25 responders, median duration of response was not reached. Median progression-free survival (PFS) was not reached; 1-year PFS was 83.4% (95% confidence interval, 64.5%-92.7%). Acalabrutinib was well tolerated with a high response rate in patients who were previously intolerant to ibrutinib. This trial was registered at www.clinicaltrials.gov as #NCT02029443.

Duplicated Enhancer Region Increases Expression of CTSB and Segregates with Keratolytic Winter Erythema in South African and Norwegian Families.

Keratolytic winter erythema (KWE) is a rare autosomal-dominant skin disorder characterized by recurrent episodes of palmoplantar erythema and epidermal peeling. KWE was previously mapped to 8p23.1-p22 (KWE critical region) in South African families. Using targeted resequencing of the KWE critical region in five South African families and SNP array and whole-genome sequencing in two Norwegian families, we identified two overlapping tandem duplications of 7.67 kb (South Africans) and 15.93 kb (Norwegians). The duplications segregated with the disease and were located upstream of CTSB, a gene encoding cathepsin B, a cysteine protease involved in keratinocyte homeostasis. Included in the 2.62 kb overlapping region of these duplications is an enhancer element that is active in epidermal keratinocytes. The activity of this enhancer correlated with CTSB expression in normal differentiating keratinocytes and other cell lines, but not with FDFT1 or NEIL2 expression. Gene expression (qPCR) analysis and immunohistochemistry of the palmar epidermis demonstrated significantly increased expression of CTSB, as well as stronger staining of cathepsin B in the stratum granulosum of affected individuals than in that of control individuals. Analysis of higher-order chromatin structure data and RNA polymerase II ChIA-PET data from MCF-7 cells did not suggest remote effects of the enhancer. In conclusion, KWE in South African and Norwegian families is caused by tandem duplications in a non-coding genomic region containing an active enhancer element for CTSB, resulting in upregulation of this gene in affected individuals.

Mutations in MAPKBP1 Cause Juvenile or Late-Onset Cilia-Independent Nephronophthisis.

Nephronophthisis (NPH), an autosomal-recessive tubulointerstitial nephritis, is the most common cause of hereditary end-stage renal disease in the first three decades of life. Since most NPH gene products (NPHP) function at the primary cilium, NPH is classified as a ciliopathy. We identified mutations in a candidate gene in eight individuals from five families presenting late-onset NPH with massive renal fibrosis. This gene encodes MAPKBP1, a poorly characterized scaffolding protein for JNK signaling. Immunofluorescence analyses showed that MAPKBP1 is not present at the primary cilium and that fibroblasts from affected individuals did not display ciliogenesis defects, indicating that MAPKBP1 may represent a new family of NPHP not involved in cilia-associated functions. Instead, MAPKBP1 is recruited to mitotic spindle poles (MSPs) during the early phases of mitosis where it colocalizes with its paralog WDR62, which plays a key role at MSP. Detected mutations compromise recruitment of MAPKBP1 to the MSP and/or its interaction with JNK2 or WDR62. Additionally, we show increased DNA damage response signaling in fibroblasts from affected individuals and upon knockdown of Mapkbp1 in murine cell lines, a phenotype previously associated with NPH. In conclusion, we identified mutations in MAPKBP1 as a genetic cause of juvenile or late-onset and cilia-independent NPH.

An Activating Janus Kinase-3 Mutation Is Associated with Cytotoxic T Lymphocyte Antigen-4-Dependent Immune Dysregulation Syndrome.

Heterozygous mutations in the cytotoxic T lymphocyte antigen-4 (CTLA-4) are associated with lymphadenopathy, autoimmunity, immune dysregulation, and hypogammaglobulinemia in about 70% of the carriers. So far, the incomplete penetrance of CTLA-4 haploinsufficiency has been attributed to unknown genetic modifiers, epigenetic changes, or environmental effects. We sought to identify potential genetic modifiers in a family with differential clinical penetrance of CTLA-4 haploinsufficiency. Here, we report on a rare heterozygous gain-of-function mutation in Janus kinase-3 (JAK3) (p.R840C), which is associated with the clinical manifestation of CTLA-4 haploinsufficiency in a patient carrying a novel loss-of-function mutation in CTLA-4 (p.Y139C). While the asymptomatic parents carry either the CTLA-4 mutation or the JAK3 variant, their son has inherited both heterozygous mutations and suffers from hypogammaglobulinemia combined with autoimmunity and lymphoid hyperplasia. Although the patient's lymph node and spleen contained many hyperplastic germinal centers with follicular helper T (TFH) cells and immunoglobulin (Ig) G-positive B cells, plasma cell, and memory B cell development was impaired. CXCR5+PD-1+TIGIT+ TFH cells contributed to a large part of circulating T cells, but they produced only very low amounts of interleukin (IL)-4, IL-10, and IL-21 required for the development of memory B cells and plasma cells. We, therefore, suggest that the combination of the loss-of-function mutation in CTLA-4 with the gain-of-function mutation in JAK3 directs the differentiation of CD4 T cells into dysfunctional TFH cells supporting the development of lymphadenopathy, hypogammaglobulinemia, and immunodeficiency. Thus, the combination of rare genetic heterozygous variants that remain clinically unnoticed individually may lead to T cell hyperactivity, impaired memory B cell, and plasma cell development resulting finally in combined immunodeficiency.

Mutations in TRAF3IP1/IFT54 reveal a new role for IFT proteins in microtubule stabilization.

Ciliopathies are a large group of clinically and genetically heterogeneous disorders caused by defects in primary cilia. Here we identified mutations in TRAF3IP1 (TNF Receptor-Associated Factor Interacting Protein 1) in eight patients from five families with nephronophthisis (NPH) and retinal degeneration, two of the most common manifestations of ciliopathies. TRAF3IP1 encodes IFT54, a subunit of the IFT-B complex required for ciliogenesis. The identified mutations result in mild ciliary defects in patients but also reveal an unexpected role of IFT54 as a negative regulator of microtubule stability via MAP4 (microtubule-associated protein 4). Microtubule defects are associated with altered epithelialization/polarity in renal cells and with pronephric cysts and microphthalmia in zebrafish embryos. Our findings highlight the regulation of cytoplasmic microtubule dynamics as a role of the IFT54 protein beyond the cilium, contributing to the development of NPH-related ciliopathies.

Whole-exome sequencing reveals overlap between macrophage activation syndrome in systemic juvenile idiopathic arthritis and familial hemophagocytic lymphohistiocytosis.

OBJECTIVE: Macrophage activation syndrome (MAS), a life-threatening complication of systemic juvenile idiopathic arthritis (JIA), resembles familial hemophagocytic lymphohistiocytosis (HLH), a constellation of autosomal-recessive immune disorders resulting from deficiency in cytolytic pathway proteins. We undertook this study to test our hypothesis that MAS predisposition in systemic JIA could be attributed to rare gene sequence variants affecting the cytotolytic pathway. METHODS: Whole-exome sequencing was used in 14 patients with systemic JIA and MAS and in their parents to identify protein-altering single-nucleotide polymorphisms/indels in known HLH-associated genes. To discover new candidate genes, the entire whole-exome sequencing data were filtered to identify protein-altering, rare recessive homozygous, compound heterozygous, and de novo variants with the potential to affect the cytolytic pathway. RESULTS: Heterozygous protein-altering rare variants in the known genes (LYST,MUNC13-4, and STXBP2) were found in 5 of 14 patients with systemic JIA and MAS (35.7%). This was in contrast to only 4 variants in 4 of 29 patients with systemic JIA without MAS (13.8%). Homozygosity and compound heterozygosity analysis applied to the entire whole-exome sequencing data in systemic JIA/MAS revealed 3 recessive pairs in 3 genes and compound heterozygotes in 73 genes. We also identified 20 heterozygous rare protein-altering variants that occurred in at least 2 patients. Many of the identified genes encoded proteins with a role in actin and microtubule reorganization and vesicle-mediated transport. "Cellular assembly and organization" was the top cellular function category based on Ingenuity Pathways Analysis (P < 3.10 × 10(-5) ). CONCLUSION: Whole-exome sequencing performed in patients with systemic JIA and MAS identified rare protein-altering variants in known HLH-associated genes as well as in new candidate genes.

LMX1B mutations cause hereditary FSGS without extrarenal involvement.

LMX1B encodes a homeodomain-containing transcription factor that is essential during development. Mutations in LMX1B cause nail-patella syndrome, characterized by dysplasia of the patellae, nails, and elbows and FSGS with specific ultrastructural lesions of the glomerular basement membrane (GBM). By linkage analysis and exome sequencing, we unexpectedly identified an LMX1B mutation segregating with disease in a pedigree of five patients with autosomal dominant FSGS but without either extrarenal features or ultrastructural abnormalities of the GBM suggestive of nail-patella-like renal disease. Subsequently, we screened 73 additional unrelated families with FSGS and found mutations involving the same amino acid (R246) in 2 families. An LMX1B in silico homology model suggested that the mutated residue plays an important role in strengthening the interaction between the LMX1B homeodomain and DNA; both identified mutations would be expected to diminish such interactions. In summary, these results suggest that isolated FSGS could result from mutations in genes that are also involved in syndromic forms of FSGS. This highlights the need to include these genes in all diagnostic approaches to FSGS that involve next-generation sequencing.

Construction of functional linkage gene networks by data integration.

Networks of functional associations between genes have recently been successfully used for gene function and disease-related research. A typical approach for constructing such functional linkage gene networks (FLNs) is based on the integration of diverse high-throughput functional genomics datasets. Data integration is a nontrivial task due to the heterogeneous nature of the different data sources and their variable accuracy and completeness. The presence of correlations between data sources also adds another layer of complexity to the integration process. In this chapter we discuss an approach for constructing a human FLN from data integration and a subsequent application of the FLN to novel disease gene discovery. Similar approaches can be applied to nonhuman species and other discovery tasks.

Phenotypic connections in surprising places.

Surprising correlations between human disease phenotypes are emerging. Recent work now reveals startling phenotype connections between species, which could provide new disease models.

Genome-wide prioritization of disease genes and identification of disease-disease associations from an integrated human functional linkage network.

We integrate 16 genomic features to construct an evidence-weighted functional-linkage network comprising 21,657 human genes. The functional-linkage network is used to prioritize candidate genes for 110 diseases, and to reliably disclose hidden associations between disease pairs having dissimilar phenotypes, such as hypercholesterolemia and Alzheimer's disease. Many of these disease-disease associations are supported by epidemiology, but with no previous genetic basis. Such associations can drive novel hypotheses on molecular mechanisms of diseases and therapies.

Computational reconstruction of protein-protein interaction networks: algorithms and issues.

Accurate mapping of protein-protein interaction networks in model organisms is a crucial first step toward subsequent quantitative study of the organization and evolution of biological systems. Data quality of experimental interactome maps can be assessed and improved by integrating multiple sources of evidence using machine learning methods. Here we describe the commonly used algorithms for predicting protein-protein interaction by genome data integration, and discuss several important yet often overlooked issues in computational reconstruction of protein-protein interaction networks.

High-precision high-coverage functional inference from integrated data sources.

BACKGROUND: Information obtained from diverse data sources can be combined in a principled manner using various machine learning methods to increase the reliability and range of knowledge about protein function. The result is a weighted functional linkage network (FLN) in which linked neighbors share at least one function with high probability. Precision is, however, low. Aiming to provide precise functional annotation for as many proteins as possible, we explore and propose a two-step framework for functional annotation (1) construction of a high-coverage and reliable FLN via machine learning techniques (2) development of a decision rule for the constructed FLN to optimize functional annotation. RESULTS: We first apply this framework to Saccharomyces cerevisiae. In the first step, we demonstrate that four commonly used machine learning methods, Linear SVM, Linear Discriminant Analysis, Naïve Bayes, and Neural Network, all combine heterogeneous data to produce reliable and high-coverage FLNs, in which the linkage weight more accurately estimates functional coupling of linked proteins than use individual data sources alone. In the second step, empirical tuning of an adjustable decision rule on the constructed FLN reveals that basing annotation on maximum edge weight results in the most precise annotation at high coverages. In particular at low coverage all rules evaluated perform comparably. At coverage above approximately 50%, however, they diverge rapidly. At full coverage, the maximum weight decision rule still has a precision of approximately 70%, whereas for other methods, precision ranges from a high of slightly more than 30%, down to 3%. In addition, a scoring scheme to estimate the precisions of individual predictions is also provided. Finally, tests of the robustness of the framework indicate that our framework can be successfully applied to less studied organisms. CONCLUSION: We provide a general two-step function-annotation framework, and show that high coverage, high precision annotations can be achieved by constructing a high-coverage and reliable FLN via data integration followed by applying a maximum weight decision rule.

VisANT 3.0: new modules for pathway visualization, editing, prediction and construction.

With the integration of the KEGG and Predictome databases as well as two search engines for coexpressed genes/proteins using data sets obtained from the Stanford Microarray Database (SMD) and Gene Expression Omnibus (GEO) database, VisANT 3.0 supports exploratory pathway analysis, which includes multi-scale visualization of multiple pathways, editing and annotating pathways using a KEGG compatible visual notation and visualization of expression data in the context of pathways. Expression levels are represented either by color intensity or by nodes with an embedded expression profile. Multiple experiments can be navigated or animated. Known KEGG pathways can be enriched by querying either coexpressed components of known pathway members or proteins with known physical interactions. Predicted pathways for genes/proteins with unknown functions can be inferred from coexpression or physical interaction data. Pathways produced in VisANT can be saved as computer-readable XML format (VisML), graphic images or high-resolution Scalable Vector Graphics (SVG). Pathways in the format of VisML can be securely shared within an interested group or published online using a simple Web link. VisANT is freely available at http://visant.bu.edu.

Rub1p processing by Yuh1p is required for wild-type levels of Rub1p conjugation to Cdc53p.

In Saccharomyces cerevisiae, Rub1p, like ubiquitin, is conjugated to proteins. Before protein conjugation, the carboxyl-terminal asparagine residue of Rub1p is removed. Rub1p conjugation is dependent on the carboxyl-terminal processing enzyme Yuh1p, whereas Rub1p lacking the asparagine residue is conjugated without Yuh1p. Thus, Yuh1p is the major processing enzyme for Rub1p.