The extent of postpartum cardiac reverse remodeling is reflected in urine proteome

Abstract

The association of postpartum cardiac reverse remodeling (RR) with urinary proteome, particularly in pregnant women with cardiovascular (CV) risk factors who show long-term increased risk of cardiovascular disease and mortality is unknown. We aim to profile the urinary proteome in pregnant women with/without CV risk factors to identify proteins associated with postpartum RR. Our study included a prospective cohort of 32 healthy and 27 obese and/or hypertensive and/or diabetic pregnant women who underwent transthoracic echocardiography, pulse-wave-velocity, and urine collection at the 3rd trimester and 6 months postpartum. Shotgun HPLC-MS/MS profiled proteins. Generalized linear mixed-effects models were used to identify associations between urinary proteins and left ventricle mass (LVM), a surrogate of RR. An increase in arterial stiffness was documented from 3rd trimester to 6 months after delivery, being significantly elevated in women with CV risk factors. In addition, the presence of at least one CV risk factor was associated with worse LVM RR. We identified 6 and 11 proteins associated with high and low LVM regression, respectively. These proteins were functionally linked with insulin-like growth factor (IGF) transport and uptake regulation by IGF binding-proteins, platelet activation, signaling and aggregation and the immune system's activity. The concentration of IGF-1 in urine samples was associated with low LVM regression after delivery. Urinary proteome showed a predicting potential for identifying pregnant women with incomplete postpartum RR.

Keywords: Cardiovascular reverse remodeling; Cardiovascular risk factor; Hemodynamic overload; Pregnancy; Proteome; Urine.

Figure 1

(A) Volcano Plot from multivariable generalized linear mixed-effects models (adjusted to cardiovascular risk factors, age at recruitment moment, parity, weight gain until 3rd trimester, and evaluation moments) showing the distribution of the proteins according to their effect (beta coefficient) and significance level (− log₁₀(p-value)) regarding LVM regression. Proteins are identified with the respective gene name. Red labels are non-significant associations whereas blue labels signify significant associations. (B) Urine proteins showing a significant association with LVM, according to the GLMM. (C) STRING Protein–Protein interaction network of the 17 most relevant proteins associated with postpartum LVM, according to a GLMM. Each node represents a protein, identified with the respective gene name. The most relevant biological processes found through gene ontology enrichment analysis are depicted through node colors. Each type of protein interaction is represented by a different edge color (cyan: known interaction from curated databases, pink: experimentally determined; yellow: text mining; black: co-expression; green: predicted interaction from gene neighborhood).