Transcriptomics-based prediction of response to immune checkpoint inhibitor in patients with malignant melanoma
Domenico Mallardo, Andrey Alexeyenko, […], Rolf Lewensohn, Giuseppe V Masucci, Paolo Antonio Ascierto

Most cancer patients fail to respond to immune checkpoint inhibitors, hence biomarkers needed in order to predict clinical benefit in individual patients. The goal of this study was to augment the prediction accuracy by identifying and testing novel candidate biomarkers that could envisage response in patients with metastatic melanoma. The analysis had two specific features: validation against previously published predicting biomarkers and characterization of patients’ transcriptomes at individual gene and pathway levels, where network enrichment analysis (NEA) integrated disparate genes into pathway scores. First, candidate transcription-based biomarkers were discovered in our cohorts via correlation to clinical benefit and then analyzed for significance by covariate adjustment. Secondly, the candidates performance was validated using a similar previously published NanoString-based gene dataset. In the anti-PD1 cohort, we identified genes which were informative on the clinical benefit regardless of the known determinants. In the anti-CTLA4 cohort, the individual gene analysis did not yield any significant and validated associations. However instead, we revealed a number of NEA-based correlates between disease progression and relevant pathways as well as a number of immune-related differentially expressed gene lists.

Proteomics and network analysis of cardiologic samples: paroxysmal versus persistent atrial fibrillation
Manolis Charitakis, Lars Karlsson, Andrey Alexeyenko, Carl-Johan Carlhäll Patients samples from two cohorts, one with paroxysmal atrial fibrillation (PAF) and the other with persistent atrial fibrillation (PEAF), were subjected to proteomics analysis using Olink INFLAMMATION panel that contains probes to 92 proteins. This revealed a number of differentially expressed (DE) proteins which we characterized with Network Enrichment Analysis. However, the 92-protein panel itself was “pre-enriched” in certain functional categories, like e.g. chemokines, interferon signaling etc., so that any list of DE proteins from the panel would be likely to exhibit enrichment in similar pathways.  Therefore, we complemented the NEA analysis with Differential Enrichment Analysis (DEA). This procedure gave statistical estimates of differential enrichment between PAF and PeAF with regard to specific pathways and placed the pathological conditions in relevant functional context.
Problem: State-of-the-art pathway enrichment techniques are inapplicable to results obtained on small, targeted panels that measure protein or gene expression as well as e.g. somatic point mutations. Solution: differential enrichment analysis.
By definition and the very approach, smaller diagnostic panels are designed “pre-enriched” in certain functional categories that correspond to their intended usage. As a result, any list of differentially expressed (DE) proteins generated by the panel would appear enriched in certain pathways, thus making impossible an unbiased pathway enrichment analysis. Therefore, we suggested to complement the latter with differential enrichment estimation: a random permutation procedure over sufficiently many (e.g. 10000) protein lists of same length as the actual DE list. The N lists are randomly picked from the full N-protein panel list (so called protein universe). Then we compare how often enrichment score values from the random lists exceeded the actual pathway value. This procedure estimates “differential” enrichment which enables an unbiased placement of the pathological conditions in a specific pathway context.
Protein profiling and network enrichment analysis in individuals before and after the onset of rheumatoid arthritis
Mikael Brink, Anders Lundquist, Andrey Alexeyenko, Kristina Lejon, Solbritt Rantapää-Dahlqvist In order to identify pathways related to the early development of rheumatoid arthritis (RA), we analyzed plasma samples from pre-symptomatic individuals (median predating time 4.1 years), early RA patients, and 74 matched controls. The levels of proteins related to autoimmunity were measured using 153 antibodies and a bead-based multiplex system FlexMap3D (Luminex). The data were analyzed using multifactorial linear regression model, random forest, and network enrichment analysis (NEA). There was a high agreement between the different statistical methods to identify the most significant proteins. The adipogenesis and interferon alpha response hallmarks differentiated pre-symptomatic individuals from controls. Between pre-symptomatic individuals and RA patients, three hallmarks were identified: apical junction, epithelial mesenchymal transition, and TGF-β signaling, including proteins suggestive of cell interaction, remodulation, and fibrosis. This confirmed the importance of interferon alpha signaling and lipids in the early phases of RA development. Network enrichment analysis provided a tool for a deeper understanding of molecules involved in the disease progression.
Protein expression in tonsillar and base of tongue cancer and in relation to human papillomavirus and clinical outcome

Torbjörn Ramqvist, Anders Näsman, Bo Franzén, Cinzia Bersani, Andrey Alexeyenko, Susanne Becker, Linnea Haeggblom, Aeneas Kolev, Tina Dalianis, Eva Munck-Wikland

Human papillomavirus (HPV) is a major etiological factor for tonsillar and the base of tongue cancer. HPV-positive and HPV-negative samples were analyzed for expression of 167 proteins using two Olink assays. Major differences in protein expression between cancer and normal tissue were identified, especially in chemo- and cytokines. A number of immunoregulatory proteins and chemokines were differently expressed in HPV-positive vs HPV-negative cancers. Several proteins were potentially related to clinical outcome for HPV-positive or HPV-negative tumors. For HPV-positive tumors, these were mostly related to angiogenesis and hypoxia. Differences in immune related proteins between HPV-positive and HPV-negative samples reflected the stronger activity of the immune defense in the former. Angiogenesis related proteins might serve as potential targets for therapy of HPV-positive cases.

Plasma RNA profiling unveils transcriptome signatures associated with resistance to osimertinib in NSCLC EGFR T790M positive patients
Andrey Alexeyenko, Odd-Terje Brustugun, […], Rolf Lewensohn, Per Hydbring & Simon Ekman Targeted therapy with tyrosine kinases inhibitors (TKIs) against Epidermal Growth Factor receptor (EGFR) is part of routine clinical practice for advanced non-small cell lung cancer (NSCLC) patients whose tumors harbor an activating mutation in the kinase domain of EGFR. Osimertinib is a third-generation EGFR TKI displaying potency to both activating kinase domain EGFR mutations and T790M resistance mutation. Here we aimed to analyze if exosomal RNAs, isolated from longitudinally sampled plasma of osimertinib-treated EGFR T790M NSCLC patients could provide biomarkers of acquired resistance to osimertinib and potentially unveil novel therapeutic opportunities to overcome this resistance. This revealed differential expression (>1.25-fold, FDR p<0.10) of 128 transcripts including those from protein coding genes, long non-coding RNAs, microRNAs, and pseudogenes. In order to summarize and generalize the observations, we applied network enrichment analysis (NEA) over a large pathway collection. An overall NEA revealed that the set of most differentially expressed genes was characterized by alterations in pathways related to EGFR and activity of other ERBB genes as well as to syndecan and glypican pathways. NEA applied to the 40 individual samples detected immune-related gene sets differentially activated between baseline and progression. These “overall” and “differential pathway activation” (DPA) signatures can reflect transition to tumor progression detectable via blood plasma sampling. Exosomal RNA sampled from plasma was thus informative on molecular phenotypes of emerging osimertinib resistance.

Bo Franzén, Andrey Alexeyenko, Masood Kamali‐Moghaddam, Thomas Hatschek, […] Giuseppe Masucci, Gert Auer, Ulf Landegren, Rolf Lewensohn

There are increasing demands for informative cancer biomarkers, accessible via minimally invasive procedures, both for initial diagnostics and to follow-up personalized cancer therapy. Fine-needle aspiration (FNA) biopsy provides ready access to relevant tissues; however, the minute sample amounts require sensitive multiplex molecular analysis to achieve clinical utility. We have applied proximity extension assays and NanoString technology for analyses of proteins and of RNA, respectively, in FNA samples. Using samples from patients with breast cancer (BC, n = 25) or benign lesions (n = 33), we demonstrate that these FNA-based molecular analyses (a) can offer high sensitivity and reproducibility, (b) may provide correct diagnosis in shorter time and at a lower cost than current practice, (c) correlate with results from routine analysis (i.e., benchmarking against immunohistochemistry tests for ER, PR, HER2, and Ki67), (d) could be divided into two main clusters, characterized by differences in expression levels of the estrogen receptor (ER) and the proliferation marker Ki67, and (e) may also help identify new markers related to immunotherapy. A specific 11-protein signature, including FGF binding protein 1, decorin, and furin, distinguished all cancer patient samples from all benign lesions in our main cohort and in smaller replication cohort. We also identified three signatures that could predict Ki67status, ER status, and tumor grade, respectively, based on a small subset of proteins dominated by chemokines. Expression profiles of CCL13 in benign lesions and BC have never previously been found to correlate with proliferation (P = 0.00095), suggesting a role in advanced BC. Due to the minimally traumatic sampling and clinically important molecular information for therapeutic decisions, this methodology is promising for future immunoscoring and monitoring of treatment efficacy in BC.

Efficient de novo assembly of large and complex genomes by massively parallel sequencing of Fosmid pools
Andrey Alexeyenko, Björn Nystedt, Francesco Vezzi, Ellen Sherwood, Rosa Ye, Bjarne Knudsen, Martin Simonsen, Benjamin Turner, Pieter de Jong, Cheng-Cang Wu & Joakim Lundeberg In order to sequence the genome of Norway spruce, which is of great size and complexity, we developed and applied a new technology based on the massive production, sequencing, and assembly of Fosmid pools (FP). The spruce chromosomes were sampled with ~40,000 bp Fosmid inserts to obtain around two-fold genome coverage, in parallel with traditional whole genome shotgun sequencing (WGS) of haploid and diploid genomes. Compared to the WGS results, the contiguity and quality of the FP assemblies were high, and they allowed us to fill WGS gaps resulting from repeats, low coverage, and allelic differences. The FP contig sets were further merged with WGS data using a novel software package GAM-NGS. By exploiting FP technology, the first published assembly of a conifer genome was sequenced entirely with massively parallel sequencing.