Cell-Specific Targeting of the Endothelium in the Cardiorenal Syndrome

Chronic kidney disease (CKD) is a global health burden that affects 850 million individuals, roughly 13% of the global population and exceeds that of diabetes [1, 2]. The dilemma regarding the management of CKD is the inordinate use of resources for the care of patients with CKD that progress to end stage and the care for accompanying comorbidities [3, 4]. Obesity and diabetes contribute significantly to the morbidity and mortality associated with the pre-dialysis population (e.g., CKD stage 1 through 4) [3‒5]. More specifically, cardiovascular diseases (CVDs) account for ∼40–50% of mortality in CKD and have a strong, continuous, graded relationship between decreases in estimated glomerular filtration rate and increasing incidents of CVD [6]. CVD risk factors in CKD range from high blood pressure, hyperglycemia, dyslipidemia, and tobacco use to non-traditional risk factors, such as vascular calcification and inflammation. High blood pressure remains the leading modifiable risk factor that can improve CVD outcomes [6]. Strategies to control blood pressure and those that target the renin-angiotensin-aldosterone system (RAAS) improve diabetic kidney disease outcomes to a greater extent than other anti-hypertensive regimens [7]. However, despite the widespread use of RAAS inhibition there is still a considerable “residual risk” for disease progression and enhanced risk for CVD. In this context, available therapies only provide a 20–30% risk reduction in CKD [4, 8] and there has been an active search for alternative treatment strategies from dual RAAS blockade to complementary glycemic control therapies such as sodium-glucose co-transporter-2 inhibition and glucagon-like peptide-1 agonism.

It is increasingly recognized that obesity not only is associated with an increased risk for CKD and CVD, but the presence of obesity enhances treatment resistance and residual risk for progression of disease. In the kidney, it is known that obesity alters intrarenal/intraglomerular hemodynamics by raising renal plasma flow and altering tubuloglomerular feedback mechanisms [5, 9‒11]. Obesity or the use of our Western diet has further shown to enhance the hyperfiltration mechanisms that lead to sodium retention, increased flow, and extracellular expansion [10]. However, from these observations it is difficult to distinguish between the confounding impact of associated hyperinsulinemia and insulin resistance, which have similar effects on vascular responses. Classically, insulin acts on the phosphatidylinositol 3-kinase (PI3K/Akt) pathway to facilitate glucose uptake and utilization [5]. However, impairments in insulin-dependent metabolic signaling contribute to reductions in bioavailable nitric oxide (NO), inflammation, oxidant stress, and growth and fibrotic pathways that trigger end-organ heart and kidney injuries.

Recent data indicate that in the kidney the glomerular endothelium, which is highly fenestrated and covered by a glycocalyx, is integral to the sieving properties of the glomerular filtration barrier and in the maintenance of the structure and function of podocytes that surround the glomerular capillaries. For instance, enhanced glomerular endothelium-related NO production increases blood flow and can alter the integrity of the filtration barrier, enhance filtration rate, and subsequent tubuloglomerular feedback mechanisms. Thereby, endothelial responses play a significant role in regulating kidney blood pressure and volume responses especially as it relates to the RAAS. In this regard, enhanced systemic and tissue RAAS in obesity occurs despite the salt and volume excess, which should suppress juxtaglomerular renin secretion. There is sufficient experimental evidence to support the concept that visceral adipose tissue enhances renin release, angiotensinogen, circulating aldosterone, and its mineralocorticoid receptor (MR) [9, 12‒14]. MR actions in heart failure and kidney diseases have been well defined clinically [14‒18]. Recent work has highlighted a role for MR on bioavailable NO and endothelial function as a link between heart and kidney diseases. Specifically, downstream actions of MR activation on the endothelial sodium channel (EnNaC) contribute to increased actin polymerization and cross-linking, reduced endothelial NO synthase (eNOS) activity and NO production leading to endothelial cell stiffening, impaired endothelial-dependent vessel relaxation, and target organ injury.

EnNaC as a molecular target: ENaC is known to regulate Na⁺ transport in kidney epithelial cells yet is also expressed in non-epithelial cells including endothelial cells (EnNaC) [4, 19]. ENaC consists of 3 subunits, α, β, and γ are members of the ENaC/degenerin family of ion channels, generally, all three subunits are essential for proper channel function and are expressed in vascular endothelial cells [19]. The distribution in kidney epithelial cells is pertinent to our understanding of ENaC regulation by flow-induced shear stress and mechanical forces. Work to date supports that western diet consumption increases aldosterone, which increases αEnNaC protein expression and its translocation to the cell membrane [20‒22], an effect mediated by MR activation in the endothelium [19]. EnNaC activation then results in polymerization of G- to F-actin [19] and reductions in bioavailable NO occur via negative regulation of eNOS during flow-mediated vasodilation [19, 22]. In this context, a number of studies support that EnNaC contributes to shear sensing in conduit and resistance arteries and regulates vascular endothelial stiffness [22, 23]. While vascular stiffening is associated with aging, it is independently associated with CKD and CVD [23] and is increasingly recognized as a therapeutic target for intervention.

Generally, endothelial stiffening is characterized by reduced NO bioavailability, increased expression of pro-inflammatory factors, and reduced endothelial-dependent vasodilation [19, 23]. This stiffness results from endothelial insults, which are dictated by a number of forces including epithelial cell myofibroblast activation and extracellular matrix stiffening [19, 20, 23]. Recent work supports that impaired endothelial NO production leads to extracellular matrix remodeling that contributes to endothelial stiffening and alterations in the endothelial glycocalyx that increase vessel wall permeability [21]. This relationship is dictated by applied tension and strain from derangements in cell-cell adhesions and the extracellular matrix via actin-myosin cytoskeletal rearrangements. Enhanced expression of endothelial cell adhesion molecules ICAM-1 (intercellular adhesion molecule-1) and VCAM-1 (vascular cell adhesion molecule-1) contribute to leukocyte/macrophage adhesion and infiltration that dictate endothelial stiffening [21, 22]. In this regard, consumption of a Western diet induces abnormal expression of tight junction proteins, including claudin-5 and activated endothelial cell MR, which in turn increases expression of VCAM-1, ICAM-1, and IL-6 that dictate glycocalyx degradation and stiffness.

Glycocalyx degradation EnNac and transglutaminases as a molecular target: In a well-developed glycocalyx, membrane expression of EnNaC is low, access of Na⁺ into the endothelial cells is limited, and NO release and vasodilation are maintained. In obesity, with higher vascular MR activity, membrane abundance of EnNaC is increased together with a deranged glycocalyx, facilitating Na⁺ entry into renal microcirculation endothelial cells and triggering the polymerization of G-actin to F-actin [21, 22]. This polymerization reduces caveolar eNOS release and bioavailable NO, which results in activation of TG2, collagen cross-linking, and endothelial stiffness [20‒22, 24]. In this regard, TG2 cross-links extracellular matrix proteins making them more resistant to breakdown and promote extracellular matrix remodeling. TG2 is expressed in the vasculature and kidney epithelial cells [19‒22, 24], where it is largely confined to the cytosol. Outside the cytosol, TG2 catalyzes a transamidation reaction [19, 25] enhancing cross-linking of extracellular matrix proteins. Recent work indicates that targeting ENaC and TG2 improves tight-junction protein expression, MMP-9 activation, and endothelium-dependent arterial relaxation suggesting their role in glycocalyx remodeling and increased endothelial permeability [22, 26]. Therefore, nanomechanics of endothelial stiffening in the microcirculation of the heart and kidney provide a rich environment for cell-specific molecular targeting.

The utility of targeting the endothelium has been limited by vectors that are non-selective and interact with numerous cell types that are permissive and allow transduction that limit their effectiveness in vascular-based disease such as in the heart and kidney. Gene therapy is an approach aiming to deliver or modify gene(s) in cells in order to alter their function and to gain a therapeutic effect. Currently, more than 20 gene therapy products have been approved by FDA or EMEA as gene medicines. Adenovirus, adeno-associated virus, or Lentivirus vectors are the most used gene therapy vehicles to deliver transgenes into the target tissue due to their efficacy and relative ease of large-scale, GMP-grade manufacturing. The vector of choice depends largely on the disease pathology, target tissue, the chosen delivery route, and desired duration of the transgene expression. Adenoviral vectors allow short-term transgene expression (weeks), and besides their current development or use as preventive vaccines, such as against COVID-19 [24], adenoviral vectors have been mostly developed for cancer therapy [27] and to treat ischemic diseases [28, 29]. In contrast to adenoviral vectors, lenitviral vectors and adeno-associated viral vectors allow more pro-longed transgene expression (months to years) and are more suitable for diseases requiring long-term therapeutic effects that can impact treatment resistance and disease progression.

All of these vector types have been tested for targeting the heart and kidney in animal models as gene therapy vectors. However, so far the main challenges have been a low efficacy and off-target expression of the transgene in the endothelium (reviewed in [30, 31]). In addition, due to the complex architecture of tissues like the heart and kidney and depending on the used delivery route and vector type/serotype variable results have been obtained. There are several choices for gene therapy vector delivery into a tissue such as the endothelium of the kidney, including systemic administration, anterograde delivery into the renal artery, retrograde administration, or intraparenchymal/subcapsular injection. Systemic delivery has been the most inefficient partly due to the vector sequestration into the liver and spleen. This could be circumvented by shielding the vectors with molecules, such as polyethylene glycol [32]. Also, certain adeno-associated viral serotypes have shown tropism to specific organs after systemic delivery. However, most of these studies have not been able to induce high transgene expression in the kidneys [31, 33, 34]. Of interest, the limited transduction of viral vectors following intravenous injection is also limited to the glomeruli [34, 35] and in the mouse kidney to the endothelial and mesangial cells of the glomerulus [35].

Direct delivery of gene therapy vectors to the kidneys has been a more promising delivery route to achieve high transduction efficiency. In a recent comparative study of three vector types, retrograde ureteral injections of the vectors to kidneys in mice were shown to result in a robust transduction of renal tubular cells [33] confirming previous observations [36‒38]. Higher transgene expression in off-target tissues was observed with adeno-associated viral than adenoviral or lenitviral vectors likely due to adeno-associated viral smaller capsid size [33]. Subcapsular injections were able to reduce the off-target expression, and AAV6.2 and AAV8 serotypes were shown to have the highest transduction efficiency in the kidneys [33]. Adenoviral vectors, however, only demonstrated transduction in the needle track and in addition showed damage to the renal parenchyma due to repeated injections [39]. To reduce the potential tissue damage, ultrasound-guided intraparenchymal injections of vectors have been tested and found to be feasible [40].

While most delivery methods that have been successful in the kidney, as an example, have resulted in transduction of the tubular epithelial cells, there are fewer reports of endothelial transduction. In context of endothelium targeting, slow infusion of the renal artery with adenoviral vector was shown to lead to a high transduction efficiency of the glomeruli in pigs [41] and in particular to the glomerular endothelium in rats [42]. As the glomerular barrier has been suggested to prevent virus vectors from reaching the kidney parenchyma, this could also be considered as a benefit to restrict vector entry into non-endothelial cells. Overall, improved retention of the vectors within the kidney and their cell-specific endothelial targeting are needed for the development of gene therapy for molecular targets such as EnNaC and TG2.

Targeting transduction would be ideal to redirect the virus vectors to a specific cell type, which would reduce off-target tissue expression. In addition, a combination of improved vector efficiency and cell-targeted delivery is likely to reduce immune responses to the vector or transgene due to lower doses required and limited biodistribution. One strategy to enable cellular targeting is to modify the surface molecules that interact with cellular receptors or molecules that facilitate viral entry. Modification of these viral proteins and thus a change in virus tropism has been achieved by introducing point mutations, using adapter molecules, inserting peptides or cysteine-moieties enabling coupling of the maleimide-activated ligands or pseudo-typing [42‒48]. Different serotypes of adeno-associated viral vectors exist naturally due to variations of their capsid proteins; however, these serotypes demonstrate a preference for a tissue and not specific cellular tropism. Much work on capsid engineering has been done, including using high-throughput, unbiased methods, such as directed evolution, to improve the specificity of adeno-associated vector to endothelial cells [43, 47, 49‒51]. However, this strategy is not feasible for all viral vectors; for example, for lentiviral vectors a change in the viral surface protein can negatively affect viral production [52]. We have used transferrin as a ligand to retarget adenoviral vectors across the endothelial monolayer [53], and α_v integrin binding adenoviral vector was used to target kidney endothelium [54].

To circumvent limited viral tropism, other means of targeting transgene expression specifically to endothelial cells are necessary. Traditionally, one such method is the use of endothelial-specific promoters. An example is the FLT-1 promoter, which has been described in the context of Ad vectors to robustly and selectively drive the expression of a transgene in endothelial cells in contrast to ICAM-2 and vWF promoters [55]. Likewise, VE-cadherin promoter was used in an LV to drive the expression of the transgene in endothelial cells in the rat kidneys [56]. An additional advantage of using cell-type-specific promoters is the tendency for reduced liver expression, which most vectors sequester to [57]. However, cell-type-specific promoters have a tendency for lower transgene expression or “leakiness” in other cell types and this has limited their use. Other means of affecting endothelial targeting is via insertion of transcriptional regulatory elements to regulate the transgene expression. We have recently demonstrated such a strategy, which relied on the use of enhancers derived from endothelial-specific super enhancers [58]. Viral vectors containing this enhancer specifically activated reporter gene in endothelial cells and not in the other cell types. In addition, since this enhancer contained a hypoxia-inducible factor (HIF)-1a binding site, then reporter transgene was further increased in low oxygen condition. Hypoxia is a feature of some CVD and kidney diseases, making this vector an attractive option for endothelial-targeted viral vector-mediated therapy. Likewise, oxidative stress-regulated vectors may be useful for these diseases. A lentiviral vector containing an antioxidant response element induced transgene expression in cells with elevated nuclear factor E2 [59], which is a key transcription factor that facilitates intracellular redox balance and is activated in ischemia. Thus, in choosing an appropriate vector to target the endothelium in tissues like the kidney, many aspects need to be carefully considered including the choice of vector type, serotype, and design of the expression cassette. Although kidney injury has not been strongly on the gene therapy radar, encouragingly a number of groups have reported effective therapy mediated by gene therapy vectors in the kidney [30, 41, 60]. Furthermore, gene therapies targeting the kidney are likely to increase with advances in vector design and specificity that will overcome the wide cellular heterogeneity and architectural constraints of the kidney.

Recent advances on techniques, such as single-cell/-nuclei sequencing (scRNA-/snRNA-seq) and spatial transcriptomics, will pave the way for new therapies for endothelium-specific targeting of the heart and kidney. In sc/snRNA sequencing, cells or nuclei are isolated from the target tissue, allowing comparison of RNA transcripts within different cell types. Recent work has already shown the complexity and heterogeneity of kidney endothelial cells in adult mice, in which 24 endothelial cell subtypes were found, including 5 subtypes of glomerular endothelial cells alone [61]. Validation of these findings at the protein level is important in order to find the best targeting strategies for heart and kidney disease. As gene expression profiles of cells also change in different disease states, and most scRNA-seq studies have so far been performed in healthy animals, this should be taken into account when planning the endothelium targeting strategy. Gene expression profiles between human and mice likely also differ due to various factors such as different timelines of the disease development, genetics, age, and flow conditions. Besides sc/snRNA-seq, spatial transcriptomics, allowing detection of mRNAs spatially in the tissue context and comparison of different areas of the diseased tissue, will be an important novel tool to understand CKD pathology and to find the best targeting strategies of the endothelium.

The vascular endothelium should be thought of as a gatekeeper or sentinel to monitor and regulate not only vessel wall function but also target organ function like the heart and kidney. Within this framework, injury to the endothelium promotes pro-thrombotic-proinflammatory/oxidant pathways that impair vessel wall relaxation and stiffening as well as target organ injury. Current treatment options available for patients with CKD and CVD are limited to pharmacological compounds and achieve only a modest risk reduction in clinical trials, suggesting the need for adjunct and alternative therapies and targets. In this regard, recent advances in techniques, such as sc/snRNA sequencing and spatial transcriptomics, will pave the way for new therapies for targeting of CVD and CKD. The development of endothelial cell-specific targeting of virus vectors will enable their use for the treatment of heart and kidney diseases with gene therapy.

The authors have no conflicts of interest to declare.

This review was supported by the Research Council of Finland Flagship GeneCellNano program (Seppo Ylä-Herttuala, Nihay Laham-Karam, and Johanna P. Laakkonen) and research fellow (328835; Johanna P. Laakkonen) and the EU-funded ERC-advanced grant HeartGenes (884382; Seppo Ylä-Herttuala) and EU Horizon grant CardioReGenix (GA825670; Seppo Ylä-Herttuala). The funders had no role in the design, data collection, data analysis, and reporting of this study.

Nihay Laham-Karam, Johanna P. Laakkonen, and Adam Whaley-Connell conceptualized and planned the review. All authors (Nihay Laham-Karam, Johanna P. Laakkonen, Seppo Ylä-Herttuala, Annayya Aroor, Guanghong Jia, and Adam Whaley-Connell) wrote and/or edited the review.

Nihay Laham-Karam and Johanna P. Laakkonen should be considered as first authors.