D. Cell-to-cell transcriptional remodeling: a nuclear role for myeloperoxidase (MPO) in endothelial inflammation

Professor Argyris Papantonis

Rationale and aims

MPO is a known inflammatory mediator severely affecting vascular and myocardial function. MPO is secreted by activated neutrophils and monocytes, accumulates in the plasma, and is taken up by endothelial cells (ECs) in the microenvironment. This leads to the remodeling of EC gene expression both in vitro and in vivo. Yet, how this is achieved remains unknown. We offer preliminary evidence in support of the hypothesis that MPO function is a result of its translocation into cell nuclei where, given its predicted capacity to bind chromatin, it may modulate transcription. If this holds true, MPO will constitute a unique example of cell-to-cell crosstalk via a secreted enzyme directly acting to regulate gene expression. We propose probing the molecular function of MPO in ECs, using a combination of cutting-edge genomics and cell biology technologies, along with animal models. Such an approach should expand our current understanding of MPO’s impact on vascular inflammation and will help to establish MPO-directed specific pharmacological therapies in c

Current state of research and own preliminary work

Figure 1: Nuclear localization of MPO in ECs. A. HUVECs were incubated with MPO (red) for 2 h, which localized in cytoplasmic and nuclear foci; nuclei were visualized using an anti-histone antibody (green). B. Cultured rat heart ECs were incubated with MPO and nuclear (N), cytoplasmic (C) or mitochondrial lysates (M) were immuno-blotted. C. HUVECs incubated on ice or at 37°C with MPO were queried for MPO translocation into nuclei over time; nuclear (diamonds), membrane (circles), cytosolic (squares), and cytoskeletal fractions (triangles) are shown. D. MPO enzymatic activity was assessed spectrophotometrically in fractionated HUVECs.

We and others have identified MPO as associated with various cardiovascular disorders and causally linked to atherosclerosis, heart failure, and arrhythmia. ECs actively take up MPO via trans-cytosis (Baldus et al., JCI 2001) and MPO induces (amongst others) expression of the endothelin receptor type-B (Lau et al., JMCC 2014). MPO deficiency suppresses adenosine receptor A3 expression in ECs of a diabetes murine model (Nishat et al., JCP 2014). We performed proof-of-principle experiments validating MPO translocation into EC nuclei within 2 hours (Figure 1). Moreover, analysis of MPO’s primary and tertiary structure revealed an overall negative charge, focused on long arginine/lysine-rich stretches reminiscent of “AT-hook” DNA binding motifs. Finally, because we now know that AT-hook factors are difficult to capture on chromatin, we tested a dual-fixation protocol at suboptimal temperature, allowing us to obtain rich chromatin-binding profiles. Thus, all necessary preliminary data and tools are in place to support the spatio-temporal characterization

Experimental approach and work program

The key aim of this project is to investigate the role of MPO in EC nuclei at high spatiotemporal resolution.

First, to provide imaging of MPO translocation dynamics, we will use confocal microscopy and a high-resolution time-course (0-6 h in 30 min intervals) to follow the translocation of MPO in EC nuclei in vitro. This approach will define the time-point of maximal MPO titers used in all subsequent experiments (as well as entry, exit, and retention times). Next, we will use super-resolution imaging (a gSTED in-house platform) to colocalize nuclear MPO with hallmarks of active (H3K4me3, H3K27ac) or inactive chromatin (H3K9me2, H3K27me3), as well as with nuclear hallmarks (Lamin B1, nuclear pore components, nucleolar markers). This will provide a detailed subnuclear localization of MPO. Finally, we will use EC fractionation into cytoplasm, nucleoplasm, chromatin, and “transcription factories” to orthogonally verify the results from super-resolution colocalizations.

Second, we will assess MPO chromatin-binding and transcriptome remodeling. Here, we will use genomics to identify MPO chromatin binding preferences (ChIP-seq) and EC expression profile remodeling (RNA-seq) at the time-points specified above. We will apply our dual-fixation protocol on ECs treated with MPO to stabilize its (direct or indirect) chromatin association, before pulling-down MPO-bound complexes using a highly-specific monoclonal antibody (MAB3174, R&D). Massively parallel DNA sequencing of these complexes and computational analysis of the data in comparison to “background” data from untreated ECs will reveal its binding positions. These positions can then be further analyzed in comparison to available genome-wide histone modification and transcription factor binding data (from the Papantonis lab or the ENCODE project). In parallel, RNA from control and MPO-treated ECs will be sequenced to obtain complete transcriptomes of the two conditions, using “spiked-in” libraries (as per ERCC instructions), so as to normalize data. Curated up-/down-regulated gene lists will reveal the pathways and functions affected by MPO. Finally, integrating ChIP-seq and RNA-seq data will allow us to identify direct MPO targets.

Third, we will evaluate MPO and chromatin organization and provide in vivo validation. Here, we will dissect MPO function at the molecular level by pulling-down MPO from EC nuclear extracts and subjecting protein complexes to proteomics. In combination with label-free quantitative analysis this will return putative MPO interactors, which can be validated by knocking-down selected ones and checking the extent to which they affect MPO association with chromatin (by ChIP-qPCR) and gene expression (by RT-qPCR). Second, we will also apply targeted Chromosome Conformation Capture (3C) technology to uncover the interaction between MPO-bound sites and gene promoters in three-dimensional nuclear space. This approach relies on designing capture probes for all MPO binding sites genome-wide, and using them to subselect spatial chromatin interactions from untreated and MPO-treated ECs. Following massively parallel sequencing, we will be able to get bona fide physical connections between MPO-bound sites and the promoters of genes the enzyme regulates. Of note, this approach will also be invaluable if MPO binding sites are found to be mostly intergenic. Another advantage is that such 3C-based data will allow us to further filter the lists of direct MPO targets identified earlier on in this project.

Finally, in vitro results from ECs will be validated in vivo: For this purpose, we will test animal models of kidney dysfunction (Benzing), atherosclerosis (Brüning, Pasparakis), pulmonary hypertension (Rosenkranz) and mitochondrial dysfunction (Trifunovic) for enhanced MPO secretion in plasma and consequent expression of MPO-downstream targets in mesenteric tissue and cremaster muscle (comprising microcirculatory vessels) via immunostaining and immunoblotting. Microvascular function will be tested in vivo with and without pharmacological MPO inhibition in these models, given that enhanced expression of MPO targets is detectable (systemic vascular resistance, blood flow velocity, neutrophil and monocyte recruitment, endothelial permeability). Finally, patient specimens, such as explanted LIMA tissue from patients undergoing coronary artery bypass grafting, will be directly investigated at the single-cell level for changes in response to MPO.

Potential future therapeutic implications

The identification and mechanistic validation of direct MPO targets will add valuable information that will help us to establish pharmacological inhibition of the MPO regulatory pathway as a specific therapeutic strategy in cardiovascular disease.

Added value through collaborations within the CCRC

The collaboration formed here bridges clinical (Baldus) with basic research (Papantonis). The interdisciplinary approach is key to the success of this project, since the clinical model system will be investigated using cutting-edge genomics, thereby enabling – for the first time – the study of inter-cell transcriptome remodeling via the signal transducer itself. Once downstream MPO targets are identified, appropriate mouse models for atherosclerosis, PH, CKD and mitochondrial dysfunction will be provided by CCRC groups (Baldus, Pasparakis, Rosenkranz) to assess the pathophysiological relevance of MPO-regulated molecules (see above). Furthermore, this project will benefit from the core facilities for sequencing (Functional Genomics; P. Nürnberg) and proteomics (C. Frese / M. Krüger).