Pericyte Bridges in Homeostasis and Hyperglycemia: Reconsidering Pericyte Dropout and Microvascular Structures

Diabetic retinopathy threatens the vision of a third of diabetic patients. Progression of the disease is attributed to the dropout of pericytes, a cell type that enwraps and stabilizes the microvasculature. In tandem with this presumptive pericyte dropout, there is enriched formation of structures assumed to be remnants of collapsed or regressed vessels, previously classified as acellular capillaries, string vessels, and basement membrane bridges. Instead of endothelial cells, we show that pericytes colocalize with basement membrane bridges, and both bridging structures are enriched by cell-specific knockout of KLF4 and reversibly enriched with elevation of Ang-2, PDGF-BB, and blood sugar. Our data suggests that pericyte counts from retinal digests have misclassified pericyte bridges as endothelial structures and have exaggerated the role of pericyte loss in DR progression. In vivo imaging of corneal limbal vessels demonstrates pericyte migration off-vessel, implicating pericyte movement in formation of pericyte bridges and pathogenesis of diabetic retinopathy.


Introduction
In diabetes, the malfunction of pericytes, cells that enwrap capillaries and considered to be key effectors in microvascular remodeling 1 , is thought to contribute to endothelial dysfunction and microvessel dropout across tissues including kidney 2 , skeletal muscle 3 , liver 2 , brain 2 , heart 4 , and retina 2 . Diabetic retinopathy is the primary condition responsible for working age blindness in the developed world 5 . Its associated pathologies, including aberrant vessel remodeling, pathological angiogenesis, progressive fibrosis, and retinal detachment 5 , have flawed treatment options. Laser photocoagulation treatment sacrifices peripheral and night vision for temporarily slowing pathological angiogenesis 6 . Anti-VEGF treatment can require monthly ocular injections resulting in mixed clinical outcomes: including lack of clinically significant visual improvement 7 , poor patience compliance 7 , and induction of retinal ganglion and neuronal cell death 8 . Better methods are needed for modulating microvascular cell behaviors to improve clinical outcomes for diabetic retinopathy patients.
Pericytes are considered an effector cell for microvascular remodeling and enwrap capillaries, maintaining close physical contact via cell soma and extended cellular processes within the vascular basement membrane 1 . However, some studies have observed pericyte-like cells bridging across two or more adjacent capillaries in both homeostatic 9 and pathological conditions such as diabetes, where their frequency is elevated 10 . These studies have framed the existence of pericyte-like bridges as evidence of pericyte detachment, where it is assumed that a fully-attached pericyte migrates (or begins to migrate) away from the capillary on which it resides and extends cell processes or its entire cell soma to form a bridge from one capillary to another [10][11][12][13] . Previous research has indicated that these bridging cells can colocalize with basement membrane structures that also span across, or bridge, adjacent capillaries 9,14 . These stand-alone (i.e. cell-free) basement membrane structures have been referred to in the literature as acellular capillaries 15 , intervascular bridges 9 , basal lamina and collagen-IV (Col-IV) sleeves 16 , and string vessels 15 . They also appear more frequently in pathological settings than in homeostasis, and some have presumed these basement membrane bridges to be residual structures left by collapsed and regressed capillaries (for review, please see 15 ). Together, these observations and assertions point to unanswered questions about the origin, functional significance, longevity, and reversibility of these cellular and acellular cross-capillary bridges, including whether they are formed by migrating pericytes that deposit basement membrane or from remnants of regressed vessels 9 ; and, given their increased prevalence in diabetic conditions, whether their frequencies are reversibly tunable by exogenous interventions, and, if so, on what time frame.
To begin to answer these questions, we first evaluated the phenotypic identity of these cell bridges using immunolabeling of a pericyte-specific marker and confirmed with a lineagetracing genetic reporter mouse driven by the same promoter, which enabled pericytes and their progeny to be fluorescently tagged with a high degree of specificity 17 . Given that all prior studies of these bridging cells have relied exclusively on non-specific proteins like NG2 and PDGFRβ 18 (which can be expressed by other cell types 19 , such as macrophages and microglia), we regarded this as a critical first step. Next, we tested the hypothesis that pericyte bridges in the retina constitute a subset of pericytes that can be reversibly enriched with multiple interventions that have relevance to diabetic retinopathy and pericyte dynamics, including: pathological elevation of blood glucose, delivery of exogenous pericyte chemokines that are upregulated in diabetes 20,21 and targeted in clinical trials 22,23 , and pericyte specific knockout of a gene that has been associated with cell migration 24 . Since the establishment and maintenance of basement membrane is an integral role for pericytes 25 , we also sought to examine the frequency of basement membrane bridges and their colocalization with pericyte bridges. Characterizing colocalization of basement membrane bridges with pericytes and endothelial cells could give insight to their origin and function within the microvasculature. The existence of pericyte bridges in healthy conditions suggest an unknown and potentially important role for them in the maintenance of homeostasis, while the enriched density of pericyte bridges in response to hyperglycemia could represent a putative reversible vascular remodeling event in diabetes that could have protective effects if appropriately modulated.  29 . Mice were sacrificed and immunostained using previously developed techniques 30,31 (See Supplementary Materials 3 and Supplementary Tables 1-2 for antibodies used).

Quantifying Microvascular Structure and Pericyte Phenotype
Cell counts of pericyte association state with the vasculature 32 was quantified using Fiji's Cell Counter plugin 33 in a blinded fashion. Vessel structure was analyzed with software written in MATLAB using previously developed methods 34 (Supplementary Materials 4).

Pericyte Marker NG2 Colocalizes with Basement Membrane Bridges in Homeostasis
Previous research shows that basement membrane bridges are found in healthy homeostatic conditions in the retina 9 , suggesting some form of ongoing remodeling of the basement membrane. We hypothesize that in the homeostatic retina, endothelial cells are not associated with these structures, while pericytes are. Immunostaining of retinal digests, an assay used to isolate the vasculature via enzymatic digestion where only endothelial cells, pericytes, and basement membrane bridges remain 10 , reveals that these thin basement membrane structures are Col-IV+, but negative for pan-endothelial marker CD31 35  A range of NG2+ cell morphologies were found, which we classify as either an attached pericyte with cell soma and all cell processes associated with a vessel (Fig. 2D, E) or a pericyte bridge with a cell soma or process extending partially off-vessel ( Fig. 2F-H). Additionally, we divided attached pericytes into a subgroup with those that are connected by an off-vessel basement membrane bridge (basement membrane bridged pericyte) that lack colabeling with pericyte and endothelial cell markers, giving an impression of an empty basement membrane track (Fig. 2E).

Pericyte Bridges Express the Smooth Muscle Cell and Pericyte Specific Marker Myh11
Previously, the Myosin Heavy Chain 11 (Myh11) promotor has been used in an inducible lineage tracing reporter mouse model to track the lineage of smooth muscle cells 37 . Myh11 lineage cells also colabel with the majority of NG2 or PDGFRβ expressing pericytes 17  suggesting that this morphology represents a fundamental pericyte phenotype found across vascularized tissues.

Short-Term and Long-Term Hyperglycemia Elevates While Insulin Treatment Reduces Pericyte Bridge Density in STZ-induced Diabetes
Previous data from retinal digests suggests that basement and pericyte bridges are

Knockout of KLF4 Results in Elevated Pericyte Bridge Density
In vascular smooth muscle cells, KLF4 can negatively regulate cell migration 24,40 , and its expression in Myh11 lineaged cells has been shown to alter smooth muscle cell phenotype 26  within KLF4-KO retinas (Fig. 6D, p=1.61E-3). YFP+ cell density was not altered in the knockout mice (Fig. 6E, p=0.699). Laminin bridges and the subset colabeled with YFP follow similar trends as pericyte bridge density across study groups (Fig. 6F-G). There was no evidence of angiogenesis as measured by vessel length density (Fig. 6H, p=0.451), no change to branchpoints per vessel length (Fig. 6I, p=0.395), enriched vessel segment tortuosity in the knockout (Fig. 6J, p=7.27E-3), and no change to vessel diameter (Fig. 6K, p=0.264).
Representative images are shown at 14 weeks of age (Fig 6L-M), 6 weeks after tamoxifen treatment is completed.

Pericytes are Capable of Migration and Process Extension Revealed Through In Vivo Timelapse
While population level analysis demonstrated in the previous results suggest that pericyte bridges could be formed by active cell movement, there is a lack of evidence directly demonstrating that pericytes can migrate or extend off-vessel processes. As a requirement for pericyte remodeling, we investigate whether pericytes are capable of dynamic process extension and migration off-vessel, and explored this through in vivo time lapse imaging of corneal limbal vessels, a vascular bed bordering the sclera and cornea noted for its utility in live imaging 41 . To provide an angiogenic response mimicking that in diabetic retinopathy, silver nitrate burns were applied to the Myh11-RFP mouse cornea 41  Our data show that enrichment of pericyte bridges on the order of a few days in response to hyperglycemia occurs long before the vascular regression that has previously been observed at 6 months post STZ-induction 43 . This rapid change in pericyte bridges was recapitulated by injection of recombinant Ang2 and PDGF-BB, both of which are cytokines upregulated in diabetes 20,21 . Elevation of pericyte bridges in Akita mice 11 provides further evidence that high blood sugar could be a short-term stimulus for enrichment of this pericyte phenotype.
For the first time, we show that the enrichment of pericyte bridges is a reversible phenomenon with restoration to basal levels following insulin treatment in STZ-induced hyperglycemia, both on a short-term time scale of days to a long-term timescale of months.
While we assert that insulin exerts its primary effect through the normalization of blood glucose levels, we cannot rule out direct effects of insulin on pericytes or on endothelial cells in communication with pericytes. The dynamic capacity of pericyte bridges is reinforced by our observations that at long-term timepoints after exogenous delivery of Ang2 and PDGF-BB, pericyte bridges return to basal levels. Given that the management of blood sugar levels is known to have a strong influence on clinical outcomes in diabetic patients 44 , it is possible that the enrichment of pericyte bridges, as an early and reversible process, could serve as a target for early therapeutic interventions to slow or perhaps prevent diabetic vasculopathies.
Across all stimuli investigated, we showed there was no evidence of altered capillary density with appreciable effect size over the time-course of our studies, suggesting that the capillary network remains static during alterations in pericyte bridge density. Previous literature has hypothesized that pericyte bridges are formed either through the active movement of pericytes 10 , or as a result of vessel regression leaving behind an attached pericyte 9 . While our studies were not designed to rule out either of these hypotheses, our data is more consistent with pericyte movement being responsible for the pericyte bridge phenotype. Our time-lapse movie demonstrates that pericytes are capable of off-vessel migration and process extension over the timescale of minutes, far more rapid than pericyte movement demonstrated previously in vivo 45 .
If vessels were regressing to create the enriched pericyte bridge density observed 4 days after injection of PDGF-BB and Ang2, an equal degree of angiogenesis would have to occur to maintain capillary density, yet without exception there were no signs of capillary sprouts or regression across all timepoints and stimuli. Our studies in STZ-induced diabetes also suggest that vessel regression is not implicated in the altered numbers of pericyte bridges because there were no changes in vascular density two weeks after STZ injection when pericyte bridge density was significantly elevated. Additionally, knockout of KLF4 in pericytes, a gene implicated in regulating cell migration 24,26  Our results also reveal shortcomings in the classical retinal digest assay that has been cited as support for pericyte drop-out in diabetic conditions 20 , and they suggest that pericyte loss as a causative factor in diabetic vasculopathies needs to be reexamined, as it has already in the Akita mouse model 11 . Our analysis of immunostained retinas revealed that up to 50% of all pericyte somas were associated with a basement membrane bridge (combination of pericyte bridges and basement membrane bridged pericytes), which would have been miscounted as endothelial somas 47,48 in the retinal digest assay and potentially accounts for the approximately 30% 20 loss in pericyte density observed with this assay in diabetes and Ang2 stimulation. We show that pericytes only occupy a subset of basement membrane bridges, but this cell-specific colocalization is not captured with the histological staining used in retinal digests, limiting the assay's usefulness for quantifying pericyte bridges and total pericyte cell count.
The existence of pericyte bridges in homeostatic conditions that are known to gradually decrease with age 9 suggests that this pericyte subpopulation serves a homeostatic function. The enrichment of pericyte bridges during diabetic conditions could alter pericyte coverage of the network 49 and suggests that they could play a contributing or protective role in vasculopathy.
Elevated pericyte bridges may prime the vessel network for remodeling, and basement membrane bridges are hypothesized to offer a preferential route for the rapid growth of new blood vessels 50 . Earlier events in disease progression are thought to be superior avenues for interventions since preventative treatments avoid the need to regenerate damaged tissue.
Modulation of pericyte bridges in hyperglycemia could provide a novel cell behavior as a potential therapeutic target.

Data and Resource Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request. The mouse strains generated during and/or analyzed during the current study are either commercially available or available through the labs that generated them.