Activated Pericytes and the Inhibition of Renal Vascular Stability: Obstacles for Kidney Repair

The hallmark feature of CKD and its progression to end stage is the presence and severity of interstitial fibrosis.^1,2 This long-recognized association is the prime driving force of research to understand the nature of renal scarification. Much of the research in this area can be grouped into one of two broad categories: efforts geared toward understanding the biochemical and pathophysiological factors that govern the degree of progression and efforts aimed at defining the cellular and molecular elements that comprise renal scars. With regard to the former, it is clear that a profibrotic environment is promoted by activity from a variety of factors with activity driving the production of extracellular matrix (ECM). Of these, TGFβ and other profibrotic cytokines have been widely studied, and these are enhanced by local tissue injury, epithelial cell cycle arrest, cell differentiation status, or hypoxia.^1,3

With regard to the latter, although multiple cell types produce ECM, myofibroblasts are considered a primary contributor to ECM production in interstitial fibrosis.³ Nevertheless, there has been considerable debate regarding the origin of these interstitial cells and whether they derive from proliferation, circulating fibrocytes, or transition from either epithelial or endothelial sources (EMT or EndoMT). Recently, compelling evidence indicates that myofibroblasts may derive from activated pericytes following detachment from the endothelial cells (ECs) for which they provide support.⁴ Although specific arguments for each of the potential sources are beyond the scope of this editorial, a comprehensive perspective on progressive fibrosis requires accommodation of theories regarding both the origin and acceleration of renal scarring.

Interestingly, a central common feature present in virtually all models leading to interstitial fibrosis is the reduction in peritubular capillary (PTC) density, which is thought to fuel hypoxia and accelerate the rate of fibrosis by directly influencing pathways of ECM production.¹ To date, there is very little known about the mechanism mediating PTC endothelial cell loss or the apparent inability to undergo successful repair. That both increased interstitial fibroblast deposition and reduced PTC density occur in the same setting raises the possibility that the two processes are mechanistically linked. One possibility, for which there is some evidence, is that injury promotes endothelial-mesenchymal transition, which simultaneously increases fibroblast number while decreasing vascular cell populations.^5,6 A second possibility exists in which pericyte differentiation into myofibroblasts reduces endothelial trophic activity resulting in capillary rarefaction. In our opinion, these possibilities are not mutually exclusive.

In the current issue of JASN, Schrimpf et al.⁷ address the hypothesis that PTC reduction after injury is the result of failed maintenance by pericytes following their activation and differentiation into myofibroblasts. The investigators isolated naive pericytes or activated pericytes following unilateral ureteral obstruction and, by microarray analysis, demonstrated that pericyte activation was characterized by increased production of the antiangiogenic factor ADAMTS-1, a disintegrin and metalloproteinase with thrombospondin motifs-1, and the downregulation of its inhibitor, tissue inhibitor of metalloproteinase-3 (TIMP-3). Human ECs form capillary networks in 3D collagen gels, which regress following exposure to the serine protease kallikrien. The authors cleverly used this phenomenon to demonstrate that coculture with pericytes stabilized EC capillary networks in response to the destructive stimulus, whereas activated pericytes fail to stabilize ECs. Interestingly, exogenous ADAMTS-1 completely blocked the ability of pericytes to stabilize capillaries, whereas TIMP-3 supplementation attenuated the destruction of capillary networks by kallikrein in the absence of pericytes. The authors also noted that Timp 3^−/−-null mice are predisposed to interstitial fibrosis and that pericytes from these animals appear to be activated and express high levels of ADAMTS-1. Moreover, these mice demonstrate a more severe capillary loss and fibrosis in response to a mild ischemia/reperfusion (I/R) injury than wild types.⁷

These elegant studies provide insight at the molecular and cellular level into the link between fibrosis and capillary loss after injury and, importantly, are compatible with existing dogma regarding progressive CKD. Clearly, peritubular capillary loss represents a primary component driving fibrosis according to the chronic hypoxia hypothesis,¹ but the basis for sustained vascular loss in the face of hypoxia is unclear. Why does the endothelium not repair itself? Several viewpoints exist including the loss of trophic factors, particularly vascular endothelial growth factor (VEGF) that maintain the endothelium of the renal microvasculature. VEGF is produced by epithelial cells, and its expression is lost in models of interstitial fibrosis.⁸ Germane to this view is the observation that blockade of VEGF results in regression of the renal microvasculature.⁹

Similarly, there are reports in models of fibrosis that demonstrate the generation of antiangiogenic factors including proteolytic fragments of the basement membrane.^10,11 Metalloproteinases, such as ADAMTS-1 and the gelatinases, are important in the generation of some of these angio-inhibitory factors and the activity of VEGF splice variants. These metalloproteinases may be produced by affected epithelial cells, endothelial cells,^10,11 or as demonstrated by the authors for ADAMTS-1, activated pericytes.

A third possibility posed by us is that renal endothelial cells fail to undergo a significant proliferative response.⁵ It is noteworthy that there are very few studies successfully demonstrating sustained cultures of rodent kidney-derived endothelial cells without the aid of viral or transgenically mediated transformation.¹² Although this may be considered a technical limitation, it likely reflects a failure of these cells to thrive in culture conditions because of a lack of intrinsic growth potential. This is in stark contrast to renal epithelial cells, which proliferate robustly in rodent injury models to repair damaged kidneys. However, the tubular epithelial proliferative capacity is not limitless. The proliferative potential of tubular cells derived from aging kidneys is reduced relative to young animals, coinciding with their reduced reparative ability.¹³ By extension, we suggest that the loss of PTC should be viewed not only in terms of the surrounding trophic environment, but also whether these cells, once stimulated, can be induced to proliferate. On this final point, we highlight that, although many strategies have been used to preserve the loss of PTCs in response to an injury (VEGF-121, angiopoeitin-1),⁸ we are not aware of any study demonstrating that these strategies can revascularize a kidney once rarefaction is established.

Like many intriguing studies, this provides a new perspective while also stimulating new questions. The current study provides strong evidence that activated pericytes lose their ability to provide adequate protection to endothelial cells. However, pericyte activation occurs following detachment from the endothelium, highlighting that a state of dynamic reciprocity exists between these two cell types. This scenario gives rise to a chicken-and-egg argument in which injury may be directed toward either cell type initially, with either leading to a similar outcome. Thus, we could envision that alterations in endothelial structure in response to injury contribute to the pericyte activation process, which subsequently reduces endothelial survival and regeneration.

In support of this, we highlight the connection between the activation of matrix metalloproteinases and both acute and chronic alterations in endothelial function. Matrix metalloproteinases (MMPs) are well known to play a role in angiogenesis, but their pathologic expression may result in vascular impairment following acute ischemic injury. The gelatinases, MMP2 and MMP9, are activated in kidney after I/R. The loss of endothelial barrier function of the PTCs is an early manifestion of renal I/R, occurring within 2 hours of reperfusion. Interestingly, the increased permeability of the renal capillaries can be attenuated by pharmacological blockade of the gelatinases.¹⁴ Whether these early alterations in PTC endothelial structure relate to long-term function and chronic instability of capillaries remains unclear. However, recent work suggests that PTC rarefaction is significantly attenuated in MMP-9–null mice following I/R injury,¹⁵ suggesting that similar molecular pathways that mediate both the early endothelial response and chronic changes in vascular stability.

Placed in context, it is tempting to speculate that rapidly activated alterations in endothelial structure and function, apparently linked to gelatinase activity, may initiate detachment from pericytes and set in motion the coordinated processes culminating in vascular dropout. Indeed, the authors have squarely placed reduced TIMP-3 activity in the crosshairs of the pericyte activation process. While TIMP-3 inhibits ADAMTS-1, it also is a known inhibitor of both MMP-2 and MMP-9 activity.¹⁶ The further investigation into these processes is likely to garner additional complexity into the process of pericyte activation and the regulation of capillary stability following injury.

Regardless of lingering questions relating to the proliferative capacity of renal endothelial cells or the further characterization of the metalloproteinase environment either mediating pericyte activation or resulting from pericyte activation, the implications raised by this paper are profoundly important. When one becomes interested in understanding limitations on the potential to revascularize damaged kidneys, consideration of efforts to stimulate endothelial growth (or endothelial replacement) may be futile if pericytes remain activated. Hopefully, the identification of this biologic barrier will help to provide a rational strategy to promote vascular repair and slow the development of renal fibrosis.

Disclosures

None.