Emerging Insights into Glomerular Vascular Pole and Microcirculation

Abstract

The glomerular vascular pole is the gate for the afferent and efferent arterioles and mesangial cells and a frequent location of peripolar cells with an unclear function. It has been studied in definitive detail for >30 years, and functionally interrogated in the context of signal transduction from the macula densa to the mesangial cells and afferent arteriolar smooth muscle cells from 10 to 20 years ago. Two recent discoveries shed additional light on the vascular pole, with possibly far-reaching implications. One, which uses novel serial section electron microscopy, reveals a shorter capillary pathway between the basins of the afferent and efferent arterioles. Such a pathway, when patent, may short-circuit the multitude of capillaries in the glomerular tuft. Notably, this shorter capillary route is enclosed within the glomerular mesangium. The second study used anti-Thy1.1–induced mesangiolysis and intravital microscopy to unequivocally establish in vivo the long-suspected contractile function of mesangial cells, which have the ability to change the geometry and curvature of glomerular capillaries. These studies led me to hypothesize the existence of a glomerular perfusion rheostat, in which the shorter path periodically fluctuates between being more and less patent. This action reduces or increases blood flow through the entire glomerular capillary tuft. A corollary is that the GFR is a net product of balance between the states of capillary perfusion, and that deviations from the balanced state would increase or decrease GFR. Taken together, these studies may pave the way to a more profound understanding of glomerular microcirculation under basal conditions and in progression of glomerulopathies.

Interest in the morphologic and functional richness of the glomerular vascular pole has recently grown, although the potential clinical significance of this anatomic site has been recognized for >20 years. Time-sequence analysis of the dynamics of renal pathology revealed that segmental sclerosis commences in the vascular pole of fawn-hooded hypertensive rats,¹ associated with loss of negative charge on the surface of endothelial cells.² The researchers attributed proteinuria in the rats to increased mean glomerular capillary pressure to 60–65 mm Hg³ due to preexisting defective myogenic constriction.⁴ A similar hilar starting point of glomerulosclerosis was documented in rats with reduced renal mass, especially when fed a high-cholesterol diet,⁴ or in patients with thrombotic microangiopathy.⁵ These pathophysiologic studies culminated in a definitive morphologic description of the vascular pole⁶ >30 years ago. This portion of the glomerulus—the hilum—has witnessed a recent rebirth of interest coupled with some enthralling findings, discussed in this brief overview.

Glomerular Hilum

The site of the entry of the afferent arteriole and exit of the efferent arteriole—the glomerular hilum or a vascular pole—is a bottleneck with potential valves. Elger et al.⁶ examined this site with great precision using electron microscopy. The glomerular hilum is crowded, indeed: it also houses contractile syncytial mesangial cells traversing the extraglomerular space to the glomerulus, rare peripolar (perihilar) cells, and transitional visceral-to-parietal epithelial cells. Mesangial cells envelop the exit of the efferent arteriole. Extraglomerular mesangial cells occupy the narrow spaces among the macula densa, afferent arteriolar smooth muscle cells, renin-secreting juxtaglomerular granular cells, and occasional peripolar cells in the hilum. The peripolar cells express α-smooth muscle actin but not renin or cytokeratin.⁷ Extraglomerular mesangium is contiguous with the intraglomerular mesangial stalk, anchoring to the basement membranes of capillary tufts.⁸

Contractile Apparatus of Glomeruli

Remarkably, several mesangial cells comprising this extra- to intraglomerular chain are syncytial, coupled via gap junctions upregulated by cAMP and downregulated by activation of protein kinase C. The cells have voltage-gated calcium channels capable of propagating calcium waves in communication-competent cell cultures.^9,10 These findings appear to complement the earlier observation of contractile activity of explanted human glomeruli,¹¹ later attributed to contractility of mesangial cells.

Stereologic analysis of glomeruli in volume-expanded and volume-depleted rats showed that glomerular tuft volume and capillary lumen volume were 28% and 32% smaller, respectively, in the volume-depleted condition.¹² The above changes were accompanied by increased surface density of glomerular basement membrane (GBM) in the perimesangial area, suggestive of mesangial contraction. In vivo relationships may be more complex than fixed sections because contractions of glomeruli may cycle whereas those of mesangial cells may oscillate.^11,13,14

Endothelial cells in the efferent arteriole bulge into the lumen, additionally impeding blood flow, as intravital videomicroscopy and electron microscopy¹⁵ have revealed. The idea of a shunt pathway (Trueta or Oxford shunt), suggested to connect afferent and efferent arterioles and redistribute blood flow away from glomerular capillaries, has been definitively rejected in recent electron microscopy reconstruction from serial sections and graph analysis of mouse glomeruli. These investigations demonstrate a “no cross zone” between afferent and efferent arterioles with no detectable “shunt” vessels.¹⁶ Nonetheless, these studies show (Figure 1) within each glomerular tuft (which has 75–250 nodes or brunching points) a shorter pathway of eight to 12 nodes between the afferent and efferent arterioles.¹⁶ This pathway is not in the hilum per se, but it is the closest transverse vessel beneath it.

Figure 1.

Demonstration of the shortest path between the basins of the afferent and efferent arterioles. (A) The number of nodes (bifurcations) in the shortest path as opposed to other vessels in capillary tufts (gray). (B) The shortest path (delineated in red; in some cases double paths) at the vascular pole of examined glomeruli. In all cases, the afferent arteriole is on the left side (green) and the efferent arteriole on the right side (red). Reprinted from ref. 16, with permission from Dr. M. Terasaki and Scientific Reports publisher.

According to Terasaki et al.,¹⁶ this shorter circuit represents a developmental primordium of the glomerular tuft, which undergoes eventual vasculogenic or angiogenic sprouting, giving rise to the rich network of glomerular capillaries. Shorter path patency could regulate the proportion of blood flowing toward the deeper bifurcating glomerular capillaries to undergo filtration in the glomerular capillary tuft. This action may separate erythrocytes, preferentially short-circuiting through this pathway in accord with the Zweifach–Fung bifurcation effect. In another study using serial resin section reconstruction of human glomeruli, Neal et al.¹⁷ described afferent and efferent arteriolar chambers with multiple outlets/inlets, which they modeled as a plenum manifold. This model does not conflict with that of Terasaki et al.¹⁶ because individual plenum manifold outlets could be subjected to regulation. This shorter path with relatively few bifurcations would be predicted to have lower resistance to flow compared with the rest of the glomerular capillary network with many more nodes. Thus, it should exhibit higher blood flow velocities.

According to the Zweifach–Fung bifurcation effect,¹⁸ single-file erythrocytes preferentially separate into the high-velocity bifurcation, resulting in the increased hematocrit and viscosity in such a branch—the Fahreus effect.^19–21 Preferential flow in a branch eventually increases flow resistance and deceleration of erythrocyte flow, thus negating and reversing the driver for the Zweifach–Fung bifurcation effect (Figure 2). Furthermore, efferent arteriolar endothelial bulging¹⁵ adds resistance to flow with elevated pressure in the upstream feeding capillaries. The computational and experimental data imply that the flow and velocity of erythrocytes in capillary networks are heterogeneous and that bifurcations function as self-sustained intrinsic oscillators, on the basis of erythrocyte separation and equalizing perfusion pressure without requirement for any external driving force.²² The similar Zweifach–Fung blood flow should exist in all bifurcations of glomerular capillaries, but exactly how the cells flow in individual branches in complex networks is unclear.²³ The complexity may arise from the fact that even minute changes in the curvature or the angle of bifurcation in daughter capillaries would produce a “butterfly effect” of changing resistance to flow and feeding pressure in the upstream capillaries.

Figure 2.

Auto-oscillatory flow pattern at bifurcations. According to the Zweifach–Fung bifurcation effect, inequality of flows at bifurcations, such as in glomerular capillaries, red blood cell (RBC) separation is driven by the differences in the flow rate. The branch with the preferential RBC flow experiences higher wall shear stress and more efficient mechanotransduction to generate NO. Eventually, increasing resistance in this branch leads to a shift of the preferential RBC flow to the second branch with similar consequences. This process is self-repetitive and alternating, not affecting the height of endothelial glycocalyx. However, when the reduced flow and shear stress become prolonged, the height of glycocalyx becomes compromised.

Could the contractile activity of mesangial cells and α-smooth muscle actin–positive peripolar cells serve as a rheostat, or a traffic controller, that regulates the degree of deformation (that is, more nuanced than the “all-or-nothing” mode) and achieve states of graded opening/closure of the shorter pathway? Different variations of this question have long been vexing, as Steinhausen et al.²⁴ describe. In all examined glomeruli, the shorter path crossed the intraglomerular mesangial stalk,¹⁶ raising the possibility that the intrinsic pulsatile contractility of mesangial cells could regulate the extent of open/closed states of the shorter path. Could the oscillatory pattern of chloride ion concentration ([Cl⁻]) at the luminal side of the macula densa trigger commensurate contraction-relaxation cycles of mesangial cells, the oscillating pattern of deformation of glomerular capillaries, and thus glomerular perfusion? The latter could act via changing the curvature of glomerular capillaries or by periodic near opening/closing of the shorter pathway.

Contractility of mesangial cells and effects of angiotensin II have been contentious subjects, as Steinhausen et al.²⁴ and Schlondorff²⁵ review in detail, both emphasizing the lack of in vivo data. Ziegler et al.²⁶ have addressed this paucity of in vivo confirmation using anti-Thy1.1–induced mesangial cell ablation. After mesangial depletion, single-nephron GFR (SNGFR) was reduced almost three-fold, even as the diameter of glomerular capillaries increased, whereas the glomerular sieving coefficient for albumin increased, indicating elevated permeability of the glomerular capillary barrier for macromolecules. Intravital multiphoton videomicroscopy revealed that angiotensin II infusion in control rats resulted not only in the reduction of SNGFR, but also in a marked rotational deformation of capillary loops of glomerular tufts (Figure 3), thus directly confirming in vivo prediction of mesangial tensile forces acting on glomerular capillaries.⁶ Both phenomena—reduction in SNGFR and change in capillary architecture—were absent in rats subjected to anti-Thy1.1–induced mesangiolysis. These studies are remarkable for several reasons: they have confirmed (1) the role of mesangial cells in regulating the SNGFR; (2) the ability of mesangial cells to affect the geometry of glomerular capillaries; and (3) demonstrated mesangial involvement in regulating the permeability of glomerular capillaries for proteins. Given that diverse kidney diseases include mesangiolysis, such as diabetic nephropathy, thrombotic microangiopathy, hemolytic-uremic syndrome, anti-GBM nephritis, and the aging kidney, these studies reinforce the idea that mesangial cells are critically involved in a range of pathologic conditions. Their ability to gently “pull” and release the glomerular tuft, which can deform the curvature of capillaries, may be essential in controlling glomerular perfusion and permeability, as Elger et al.⁶ earlier conjectured. Notably, tubuloglomerular feedback plays a considerable role in consecutive cycles of mesangial relaxation and contraction.

Figure 3.

Movement of glomerular capillaries within the Bowman’s space of rat kidneys: controls (white bars) and after anti-Thy1.1 (blue bars) mesangiolysis. (A) A glomerular tuft during perfusion of saline. (B) The glomerular tuft during perfusion of angiotensin II depicting a rotating movement. (C) Restoration of the capillary tuft after cessation of angiotensin II infusion. (D) Summary of intraglomerular movement of capillaries in response to the infusion of angiotensin II—note that movements were curtailed by mesangiolysis. (E) Summary of intraglomerular movement of capillaries in response to NE—note that movements are significantly nonexistent. White arrows show sites of deformation. Reprinted from ref. 26, with permission from Acta Physiologica.

Tubuloglomerular Feedback

The continuation of the tubule conveying the urine out, the loop of Henle, turns back toward its own glomerulus. In a pioneering investigation to identify the initiating signal for tubuloglomerular feedback, Schnermann et al.²⁷ perfused the loop of Henle with 17 different salt solutions. The most pronounced feedback responses were with monovalent Cl⁻ or bromine ion salts and mitigation of responses in Cl⁻-free perfusates or upon addition of furosemide. They concluded that Cl transport, probably in macula densa cells, initiates tubuloglomerular feedback. Macula densa cell volume changes in accordance with the chemical composition of the tubular fluid: luminal osmolarity of 180 mOsm results in swelling abrogated by furosemide.²⁸ Measurements of cytosolic Cl concentration in macula densa cells showed that furosemide reduced the cytosolic Cl concentration two-fold, similar to the effect of reduction in luminal sodium chloride (NaCl), whereas basolateral Cl channel inhibition almost tripled the cytosolic Cl concentration.²⁹

These studies established the existing apical Na, potassium, 2Cl cotransporter and basolateral Cl channel as regulators of macula densa cell volume. In turn, macula densa cell swelling is a cognate activator of basolateral maxi-anion channel.³⁰ This ubiquitous channel, with a radius of 1.3 nm and unitary conductance of 300–400 pS, partners with the channel-forming 12-membrane-spanning solute carrier organic anion transporter family member 2A1 (SLCO2A1) protein, which functions as a PGE1, PGE2, PGD2, and PGF2A channel in the basal state and as an ATP channel upon cell swelling, ischemia, or hypoxia, activating downstream purinergic signaling.³¹ The channel’s function could be critical for signal transduction from the tubule Cl sensing by the macula densa cells and to the extraglomerular mesangium and afferent arteriole. Specifically, one would envision the reciprocal interplay between relaxing PGE2 and constricting ATP in regulating the tensile state of extraglomerular mesangial cells, which is transmitted to intraglomerular mesangium. Activation of a single cell in this syncytium propagates in a way that activates distant cells in the adjacent afferent arteriole and inside the glomerulus, where mesangial tension could affect the geometry of capillaries and the patency of the short-circuit pathway.²⁶

Schnermann³² has emphasized the dual regulatory role of the macula densa Cl sensor: constriction of the afferent arteriole during high macula densa transport activity and renin release during low transport activity. In support of this prediction, reducing NaCl concentration in cultures of macula densa cells activated cyclooxygenase-2 and PGE2 release by promoting phosphorylation of p38 and extracellular signal–regulated 1/2 kinases.³³ In addition, reporter fura-2–loaded HEK293 cells expressing prostanoid receptor EP1 and positioned at the basolateral surface of isolated perfused macula densa cells detected purinergic signaling. Reduction in luminal NaCl concentration led to a calcium signal, but EP1 receptor antagonist, cyclooxygenase-2 inhibitor, or furosemide³⁴ inhibited it.

Examples of oscillatory microvascular perfusion abound. Applying dynamic Bayesian inference to laser Doppler recordings of microvascular flow has identified low-frequency, around 0.1 Hz, oscillations.^35–38 Discovery of similar phenomena in other microcirculatory beds would support the existence of similar oscillations in the glomerular microcirculation. Indeed, Marsh and colleagues’³⁹ analyses of renal microcirculation reveal that tubuloglomerular feedback and myogenic contractions orchestrate regulation of SNGFR and blood flow and generate self- sustained oscillations.

Contraction/Relaxation of Mesangial Cells May Change the Curvature of Glomerular Capillaries

Sakai and Kriz⁴⁰ have thoroughly investigated connectivity of mesangial cells to basement membrane of the glomerular tuft. The investigators applied a new staining technique of unosmicated specimens, better revealing cellular outlines in their electron microscopy survey of mesangial, epithelial, and endothelial cells and the GBM, and demonstrated extensive direct or microfibrillar connections between them. They concluded that mesangial cells and GBM form a “biomechanical unit capable of developing wall tension in glomerular capillaries and of changing the geometry of glomerular capillaries following mesangial contraction or relaxation.” The forces of mesangial pull/release cycles may be insufficient to open or close glomerular capillaries, but experimental data show these forces are sufficient to change the geometry or curvature of capillaries.^7,25,26,41 Although the forces syncytial mesangial cells produce should be equally distributed between different capillary beds and, morphologically, the shorter path vessel is indistinguishable from other capillaries in the glomerulus, important hemodynamic differences may exist. The forces acting on the shorter circuit are at a near-right angle, as opposed to the tangential vectors applied to peripheral capillaries. This could be one of the reasons to presume that the shorter path is the most sensitive to deformation (Figure 4).

Figure 4.

A schematic illustration of a hypothetical relationship between mesangial cell contraction-relaxation status (modeled with two dented clock wheels) and the patency of the shorter path between inflow and outflow glomerular circulation (shown in red as in Figure 1). The tensile state of mesangial cells could establish a rheostat, regulating the flow through the shorter path between afferent and efferent arteriolar basins. The humoral agents promoting mesangial contraction and relaxation are summarized at the top portion of the schema (on the basis of refs. 25 and 41), but require caution as they were derived in vitro. It is possible that the increased blood flow through the capillary tuft counterbalances contractile pressure, whereas shorter circuit–redirected flow in the relaxed state diminishes perfusion of distal capillaries. These states of contracted and relaxed mesangial cells are probably alternating. ANP, atrial natriuretic factor; AVP, arginine vasopressin; cGMP, cyclic guanosine monophosphate; ET-1, endothelin-1; PAF, platelet activating factor; PDGF, platelet-derived growth factor.

One consequence of the changes in capillary curvature is the concomitant change in blood flow and wall shear stress—important determinants of the height of endothelial glycocalyx.^42–45 Pries et al.⁴⁵ demonstrated that removal of glycocalyx via selective infusion of heparinase into a branch of the mesenteric microvascular network leads to the selective decrease in flow resistance in that branch. Although the physical finding is indisputable, the biologic consequences of ensuing unhindered tethering and adhesion of leukocytes and platelets⁴⁶ may, in the long run, reverse the effect of glycocalyx removal on resistance to flow. Notably, the loss of glycocalyx is common in a broad range of cardiovascular, renal, metabolic, and infectious diseases.^47–51

Possible Implications for CKD Progression

Taking the idea of oscillating mesangial cell contraction-relaxation and a glomerular perfusion rheostat aided by the shorter circuit pathway a step further, it follows that there may be a dynamic balance between the near-open/closed time of this circuit and the blood flow rate through the glomerular tuft. This is a rheostatic adjustment of blood flow, not an “all-or-nothing” effect, that has an intrinsic oscillating capacity described by Zweifach–Fung bifurcation and Fahreus viscosity effects. This cumulative balance determines the measured SNGFR and the GFR. When the balance shifts in favor of the increased duration of the near-open state of the shorter circuit, both the SNGFR and the GFR are reduced (like in the anti-Thy1.1 mesangiolysis model), whereas renal blood flow remains little changed. A consequence of shorter-circuit circulation will be reduced NaCl delivery to the macula densa cells, afferent arteriolar dilation, and restoration of balance between the near-open/closed time of this circuit. If the net-sum reduction in glomerular deep capillary perfusion occurs for a protracted period, it should affect the integrity of endothelial glycocalyx and repress endothelial nitric oxide (NO) synthase activity and NO production. Among other actions of NO, a decrease in its bioavailability promotes TGF-β signaling. Saura et al.⁵² demonstrated that the endothelial NO–protein kinase G pathway is involved in proteasomal degradation of Smad-2, thus curtailing effects of TGF-β on fibrogenesis and endothelial-mesenchymal transition.⁵³ Furthermore, NO and carbon monoxide are capable of antagonizing TGF-β signaling by facilitating internalization of its receptor TGF-beta receptor, ALK5–activin receptor-like kinase 5 (TβR/ALK5).⁵⁴ Chronic inhibition of NO synthases promotes endothelial-mesenchymal transition.⁵⁵ In addition, reduced bioavailability of NO is linked to the loss of endothelial glycocalyx.⁵⁶ All of the above culminate in endothelial-mesenchymal transition, fibrogenesis, and capillary obliteration and dropout.^53–58 Gradual accumulation of these abnormalities would promote progression of CKD.

Potential Limitations of this Hypothesis

Alterations in glomerular microcirculation may contribute to the progression of kidney diseases. The data presented here are limited to the superficial cortical glomeruli, which are more amenable to investigation. The contribution of juxtamedullary glomeruli and their microcirculation is less clear, although Cowley⁵⁹ proposed their role in pressure natriuresis and BP regulation. Because renal medullary microcirculation is relatively insensitive to vasoconstrictive effects of angiotensin II,⁶⁰ it follows that juxtamedullary glomeruli could be most vulnerable to increased pressure transmission and resulting barotrauma.^61,62 In this context, (1) the existence of the shorter path, and (2) the contractility of mesangial cells in juxtamedullary glomeruli need further studies. These are especially important in view of preferential vulnerability of juxtamedullary glomeruli in FSGS. Finally, the physiologic and pharmacologic profiles of the mesangial contraction/relaxation oscillatory cycle are incompletely understood, as is the coordination of this rhythm with that of BP oscillation. These and many other questions need to be resolved.

In summary, the contractility of glomerular mesangial cells and their ability to change capillary geometry have been experimentally established in vivo. The shorter capillary path, enclosed within the glomerular mesangium between the afferent and efferent arteriolar basins, has been identified. Taken together with knowledge of the oscillating Cl⁻ concentration at the level of the macula densa, periodic activation of the extraglomerular mesangium transmitted inside glomeruli, and the auto-oscillatory blood flow at bifurcations, these new findings support the idea that the shorter path is periodically becoming more and less patent, reducing or increasing blood flow through the entire glomerular capillary tuft. The GFR is a net product of balance between the states of capillary perfusion, and any deviations from the balanced state would increase or decrease GFR. The opening state of the shorter pathway prevailing may functionally or structurally obliterate a portion of glomerular capillaries and progressive loss of filtration.

The Proof Is in the Pudding

The above hypothetic corollary of two recent studies calls for experimental evidence; I envision a few possibilities. Intravital videomicroscopy, similar to what Ziegler et al.²⁶ used, focusing on the periportal area, may discern contraction/relaxation cycles of glomerular capillaries. Velocitometry of red blood cells may reveal the degree of homogeneous flow in distinct capillaries. Other areas of study could involve examining the oscillations in flow velocity and their harmonics using Fourier analysis; conditions leading to prolongation of the underperfused state of the glomerular capillary tuft; and matrix deposition, TGF-β activation, and density of patent capillaries. Finally, mathematic modeling of intricate relations between tubuloglomerular feedback, mesangial contractility, and capillary patency, all advancing earlier modeling,⁶³ may shed light on their complex interaction.

Disclosures

None.

Funding

Studies in the author’s laboratory are supported, in part, by National Institutes of Health grant HL144528 and New York Community Trust funds for research and education.

Author Contributions

M.S. Goligorsky conceptualized the study, wrote the original draft, and reviewed and edited the manuscript.