Monocyte-endothelial interactions as a targetable node in clonal hematopoiesis-mediated cardiovascular disease

Reviewer #1 (Public review):

Summary:

Using single-cell RNA sequencing and bioinformatics approaches, the authors aimed to discover if and how cells carrying mutations common to clonal haematopoiesis were more adherent to endothelial cells.

Strengths:

(1) The authors used matched blood and adipose tissue samples from the same patients (with the exception of the control people) to conduct their analysis.

(2) The use of bioinformatics and in-silico approaches helped to fast-track their aims to test specific inhibitors in their model cell adhesion system.

Weaknesses:

(1) The analysis was done on pooled cells; it would have been interesting to know if the same adhesion gene signatures were observed across the donors.

(2) The adhesion assays were conducted under static conditions; shear flow adhesion experiments would have been better. Mixed cultures using cell trackers would have been even better.

(3) In the intervention studies, the authors should have directly targeted the monocytes (not the endothelial cells) and should have also included DNMT3A mutant/KO cells to show specificity to TET2 CHIP.

Reviewer #2 (Public review):

Summary:

The authors describe potential mechanisms underlying the changes in endothelial-monocyte interactions in patients with clonal hematopoiesis of indeterminate potential (CHIP), including reduced velocity and increased ligand interactions of CHIP-mutated monocytes. They use a combination of transcriptomics (some for the first time in these tissues in patients with CHIP), in silico analyses, and ex vivo approaches to outline the changes that occur in blood monocytes derived from patients with CHIP. These findings advance the current field, which has previously mostly used mice and/or has been focused on cancer outcomes. The authors identify distinct alterations in signaling downstream of DNTM3A or TET2 mutations, which further distinguish two major mutations that contribute to CHIP.

Strengths:

(1) Combinatorial transcriptomics was used to identify potential therapeutic targets, which is an important proof-of-concept for multiple fields.

(2) The authors identify distinct ligand interactions downstream of TET2 and DNMT3A mutations.

Weaknesses:

(1) The authors extrapolate findings in adipose tissue in diabetic patients to vascular disease (ostensibly in the carotid or cardiac arteries), citing the difficulty of using tissue-matched samples. Broad-reaching conclusions need to be backed up in the relevant systems, considering how different endothelial cells in various vascular beds react. Considering these data were obtained with n=3 patients being sufficient to identify these changes, it seems that this can be performed (perhaps in silico) in the correct tissue.

(2) The selection/exclusion criteria for the diabetes samples are not noted, and therefore, the relevant conclusions cannot be fully evaluated, nor is the source of adipose tissue stated.

Appraisal:

While authors describe how to as well as the technical feasibility of integrating a number of transcriptomic techniques, they do not seem to do so to produce highly compelling data or targets within this manuscript. The potential is there to drill down to mechanisms; however, the data gathered herein do not highlight novel targets. For example, CXCL2 and 3 are already shown to be differentially expressed in TET2 loss combined with LDL treatment in the macrophages of mice. Furthermore, these authors then show that in humans, the prototypical CXC chemokine, IL8 (which mice lack), is significantly higher in TET2-mutated patients (DOI: 10.1056/NEJMoa1701719). The authors should demonstrate the utility of their transcriptomics by identifying and testing novel targets and focusing on the proper disease states. This could easily be a deep dive into CHIP in adipose tissue in diabetic patients.