Three-Dimensional Perfusion Imaging of Microcirculatory Networks Within Axially Vascularized Artificial Tissue

Abstract

Evaluating the complex, three-dimensional (3D) architecture of de novo angiogenesis in artificially engineered tissue remains a significant challenge, as conventional methods like 2D histology and microimaging techniques are limited. For axial vascularization techniques, a reproducible method for complete visualization of the microcirculatory system is needed. We present an integrated workflow for high-resolution 3D visualization of neovascularization within arteriovenous (AV) loop-based tissue constructs in a rat model. An intravascular perfusion with a cationic near-infrared fluorescent dye, MHI148-polyethylenimine, was used to 3D label the patent vasculature. Following perfusion-fixation and explantation, the construct was rendered optically transparent using an ethyl cinnamate–based clearing protocol. The fluorescent signal was then imaged using confocal and light-sheet fluorescence microscopy at 7 and 28 days postimplantation. Our workflow successfully achieved high-contrast, 3D visualization of the microvascular network, allowing for whole-mount and segmental analysis of the vascular tree. At day 7, imaging delineated solely the AV loop axis while by day 28, a dense and complex, interconnected capillary plexus from the central axis demonstrated a progressive neovascularization. Downstream processing compatibility was confirmed through successful rehydration and 3D nuclear counterstaining. This workflow offers a powerful and reproducible method for detailed structural assessment of microvascular networks in large engineered constructs, overcoming key limitations of existing techniques.

Impact Statement

Visualizing the complete 3D microvascular network in engineered tissue is a major challenge, as current methods are limited and often destructive. We introduce a novel workflow combining intravascular fluorescent perfusion, optical tissue clearing, and 3D microscopy in an AV loop model. This provides unprecedented high-resolution, whole-mount imaging of the entire patent vascular tree. Crucially, our aqueous-based protocol preserves endothelial integrity, allowing for subsequent rehydration and molecular analysis—a major advantage over traditional casting techniques. This reproducible tool accelerates the study of neovascularization and the development of clinically relevant engineered tissues.

Keywords

optical clearing 3D microscopy vascular imaging tissue engineering axially vascularized artificial tissue

Introduction

The reconstruction of extensive tissue defects remains a critical challenge in regenerative medicine, particularly when addressing the complex requirements of soft and functional tissues. Engineered tissue constructs must not only mimic native tissue architecture but also integrate a fully functional microvascular network to sustain metabolic demands. The ability to ensure sufficient vascularization is pivotal, as nutrient and oxygen diffusion alone cannot support the viability of thick tissue grafts. This necessitates a high-resolution, three-dimensional (3D) visualization approach to understand and optimize the microvascular integration of tissue-engineered constructs.

Among the most effective vascularization strategies is the arteriovenous (AV) loop model, which provides an axial vascularization mechanism that enables immediate perfusion and microcirculatory development within engineered constructs. Unlike extrinsic angiogenesis strategies relying on peripheral vessel sprouting, the AV loop model facilitates de novo angiogenesis, independent of external growth factor supplementation.¹ By embedding a venous graft between an arterial–venous shunt within an isolating chamber filled with a biocompatible scaffold, vascular shear stress within the graft together with a hypoxia gradient in the chamber promotes endothelial sprouting and microvessel formation, a process likely linked to Connexin43 transactivation.^2,3

Current strategies for 3D vascular visualization largely rely on two distinct approaches: casting techniques and fluorescence-based tracing. While vascular corrosion casting using high-viscosity resins (e.g., Microfil®) combined with microcomputed tomography (µCT) remains a standard for structural analysis, it is inherently destructive for the endothelium lining of the vascular tree. Conversely, fluorescence-based vascular imaging offers compatibility with optical tissue clearing (OTC) and advanced microscopy. In vivo administration of fluorescent dextrans or albumin conjugates is widely used for dynamic flow assessment; however, these tracers are nonfixable and prone to extravasation in permeable neovasculature, leading to signal “blurring” and washout during processing.^4,5 To address structural mapping, perfusion with fluorophore-conjugated lectins (e.g., Lycopersicon esculentum) is considered the gold standard for labeling the endothelial glycocalyx. Yet, lectin perfusion is limited by significant cost, variable binding affinity across different vascular beds, and inconsistent staining of immature neovessels.^6–8 As passive fluid tracers (dextrans) wash out and specific receptor ligands (lectins) may face saturation, there is a growing need for a robust, cost-effective, and “fixable” coating of the vascular lumen compatible with solvent-based optical clearing protocols, which not only facilitates high-resolution, volumetric visualization of microvascular structures in TE constructs but also preserves vessel wall integrity, ensuring optimal conditions for further downstream analysis.

Recently, we demonstrated that intravascular perfusion of an aqueous and cationic near-infrared (NIR) MHI148-polyethylenimine (PEI) solution in polylactic acid (PLA)-based AV loop constructs is ideally suited for intraluminal vessel visualization. Compared with conventional CD31 immunostainings, we achieved superior signal intensity in higher-order branches and capillaries making it a powerful tool for quantitative vascular analysis in TE constructs.⁹

Building on these findings, we recognized their potential for an advanced methodology that not only enhances vascular visualization but also addresses the limitations of conventional 3D angiogenesis assessment techniques. We developed an imaging protocol for AV loop tissue engineered (TE) constructs—an in vivo model for intrinsically vascularized soft tissue—integrating aqueous fluorescence-based perfusion imaging, OTC, and high-resolution fluorescence microscopy. Combining the PLA scaffold and intravascular perfusion with fluorescent dye MHI148-PEI, we achieved superior dye penetration with high contrast and signal-to-noise ratio, making it an ideal technique for 3D vascular imaging in TE constructs. These techniques may ultimately enable spatial visualization and identification of key molecular and cellular regulators within specific subdomains of TE constructs and help to understand the dynamics and mechanisms underlying vessel formation in axially vascularized systems. Leveraging these techniques may ultimately contribute to TE construct development in regenerative medicine and promote translational and clinical use of engineered TE constructs.

Materials and Methods

Microsurgical procedure and AV loop model

All experiments were performed using female Lewis rats (8–12 weeks of age, mean body weight: 240 g) obtained from Charles River Laboratories (Sulzfeld, Germany). All animal procedures complied with the German Animal Welfare Act and were approved by the Institutional Animal Care and Use Committee (Landesuntersuchungsamt Rheinland-Pfalz; approval number AZ G 18-7-032). For analgesia, buprenorphine (0.05 mg/kg body weight; Bayer, Leverkusen, Germany) was administered subcutaneously. Prior to surgery, 4 international units of heparin (Rotexmedica GmbH, Trittau, Germany) were injected via the tail vein, and 7.5 mg/kg of enrofloxacin (Bayer) was administered orally.

Anesthesia was induced with 5% isoflurane (Baxter, Vienna, Austria) in oxygen (2 L/min flow) and maintained at 1.9% isoflurane with 0.6 L/min oxygen. All microsurgical procedures were performed under 16× magnification using a surgical microscope (OPMI® Technoskop pico, Carl Zeiss, Jena, Germany) by the same surgeon. After depilation, skin disinfection, and sterile draping, a bilateral midventral incision was made to expose the femoral neurovascular bundles. A 20 mm segment of the femoral vein was harvested and interposed between the contralateral femoral artery and vein via microsurgical end-to-end anastomoses using 11/0 nylon sutures (Ethilon, Ethicon, Norderstedt, Germany). Following completion of anastomoses, 10 IU of heparin were administered intravenously, and hemostasis was ensured using bipolar forceps (KLS Martin, Freiburg, Germany).

The isolation chamber was assembled with two layers of 2 mm-thick lactocapromer-terpolymer (SupraSDRM, PolyMedics, Denkendorf, Germany) enclosing the AV loop and was implanted subcutaneously. The chamber was fixed onto the underlying adductor fascia using 6-0 polypropylene sutures (Prolene 6/0, Ethicon), and the skin was closed with interrupted vertical mattress sutures using 4-0 Vicryl (Ethicon). Postoperatively, animals received daily buprenorphine (0.05 mg/kg) until postoperative day 3, and one additional dose of 10 IU heparin on day 1. Animals were housed under a 12 h light/dark cycle with ad libitum access to standard chow (Sniff) and water. Rats were sacrificed under deep isoflurane anesthesia (5%) by exsanguination and subsequently reperfused with the near-infrared dye MHI148-PEI.

Intravascular perfusion with MHI148-PEI

A 3–4 cm skin incision was made along the inguinal ligament on the limb carrying the AV loop. Blunt dissection was performed along the fascia of the external abdominal oblique muscle using microsurgical forceps until the vascular pedicle, including the superficial femoral artery and great saphenous vein, was visualized. AV loop patency was confirmed by observing arterial pulsation and venous distensibility. Subsequently, the abdominal midline was incised, and jejunal and ileal loops were displaced cranially after mesenteric mobilization. The abdominal aorta and vena cava were exposed in the retroperitoneum. The aorta was cannulated above the renal arteries using a 24-gauge catheter, and 300 mL of prewarmed (37°C) isotonic saline containing heparin (100 IU/mL) was perfused after venotomy of the inferior vena cava. Following this, 30 mL of aqueous MHI148-PEI solution (1.5 mg/mL in distilled water) was administered intra-aortically at a flow rate of 1 mL/min in an antegrade manner. To enhance dye binding, the vascular pedicle was clamped with a microsurgical clamp for 15 min. Perfusion fixation was then performed with 50 mL of 4% paraformaldehyde (PFA; pH 7.0; 4°C). Samples were stored at 4°C in 4% PFA, protected from light, for further analyses.

Optical tissue clearing with ethyl cinnamate

Whole constructs were formalin-fixed and subjected to a modified ethyl cinnamate (ECi) tissue clearing protocol optimized for vascular imaging.^10,11 Constructs were dehydrated in a graded ethanol series (50%, 80%, 100%, 100%; 30 min per step) and then incubated in ECi (ethyl-3-phenylprop-2-enoate) for 2 h. Cleared samples were stored in ECi, protected from light. All processing steps were performed using a tissue processor (Leica TP1020).

3D imaging of cleared constructs

For confocal imaging, cleared constructs were affixed (UHU adhesive, Bolton Adhesives) at the center of 60 mm tissue culture dishes (Orange Scientific, Belgium) and immersed in refractive index (RI)-matched oil (RI 1.5; Thermo Fisher Scientific, USA). Images were acquired using a Leica SP8 confocal microscope equipped with a 16×/0.50 objective HC PL FLUOTAR IMM CORR (Leica, Wetzlar, Germany). MHI148-PEI was excited at 638 nm and detected via a Cy7 emission filter.

For light-sheet fluorescence microscopy (LSFM), constructs were sectioned into 2 mm-thick slices (Heart Matrix 65-2100; AgnTho’s AB, Sweden), placed on Parafilm (Carl Roth, Germany) within 60 mm dishes, and embedded in ECi (RI 1.558). Imaging was performed with a Leica TCS SP8 DLS system using a 5×/0.15 objective HC PL FLUOTAR IMM DLS (Leica Microsystems, Mannheim, Germany). Excitation wavelength was 638 nm, and fluorescence was collected via a Cy7 filter. Alternating bilateral illumination was applied at 1400 Hz acquisition speed. The final voxel dimensions were 0.719 µm (x), 0.719 µm (y), and 2 µm (z).

3D nuclear staining and post-clearing imaging

For nuclear labeling, SYTOX™ Green (Thermo Fisher Scientific, USA) staining was performed according to a modified protocol.¹² The ECi-cleared constructs were rehydrated through descending ethanol concentrations, permeabilized in 1% Triton X-100 for 24 h at 37°C, washed in 1× phosphate-buffered saline (PBS), and incubated with SYTOX™ Green (1:1000) for 12 h at 37°C (Table 1). Samples were again dehydrated and cleared in ECi for 1 h. All incubation steps were conducted in opaque Falcon tubes (Eppendorf Amber Conical Tubes) under constant rotation at room temperature (roller incubator, Thermo Fisher Scientific).

Table 1.

Nuclear Staining Protocol of Tissue Constructs

Solution	Time
Rehydration in 100% ethanol	15 min
Rehydration in 100% ethanol (repeat)	15 min
Rehydration in 80% ethanol	15 min
Rehydration in 50% ethanol	15 min
Permeabilization in 1% Triton X-100	24 h (37°C)
Wash in 1× PBS	12 h
Nuclear staining (SYTOX™ Green, 1:1000)	12 h (37°C)
Wash in 1× PBS	1 h
Dehydration in 50% ethanol	15 min
Dehydration in 80% ethanol	15 min
Dehydration in 100% ethanol	15 min
Dehydration in 100% ethanol (repeat)	15 min
Optical tissue clearing in	ECi 1 h
Total duration	∼52 h

Staining protocol with descending and increasing ethanol concentrations and incubation times followed by the ECi-clearing step.

Eci, ethyl cinnamate.

After staining, regions of interest (ROIs) were acquired using a Leica TCS SP8 confocal microscope with a 63×/1.32 oil objective HCX PL APO CS (Leica Microsystems, Mannheim, Germany). Fluorescence excitation was performed using 488 nm (SYTOX™ Green) and 638 nm (MHI148-PEI) lasers. The system operated in line-sequential mode with bidirectional x-scanning at 400 Hz. Images were acquired at 1024 × 1024 resolution, and voxel dimensions were 0.114 µm × 0.114 µm × 0.6 µm.

Data processing and management

Image acquisition, Z-stack stitching, and 3D rendering were conducted using LAS X software (Leica Microsystems). All image and video data were stored in the Scientific Data Storage (SDS) system of Heidelberg University, enabling fast and secure access for downstream analysis.

This study is designed as a proof of concept for high-resolution 3D perfusion imaging within axially vascularized artificial tissue. No quantifications or statistical comparisons between study groups were performed.

Experiment

Experimental setup, explantation, and macroscopic assessment

The AV loop operation within the rat is a well-known and established model of in vivo axial vascularization of artificial tissue constructs. Eight animals underwent the AV loop procedure in total and were included in this study. All animals tolerated the anesthesia and surgical procedures well, with no observed complications such as surgical site infections, hematomas, or wound dehiscences. After 7 (N = 5) or 28 days (N = 3), the animals underwent antegrade reperfusion with MHI148-PEI. The isolation chambers were explanted, and the artificial constructs were examined macroscopically. No loop thromboses or perfusion-related main vessel damage were observed.

Each animal underwent MHI148-PEI reperfusion followed by ECi-based OTC as described in the Materials and Methods section, allowing for high-resolution 3D imaging of vascular networks within the constructs. The experimental workflow and steps, as illustrated in Figure 1, outline the four major steps: AV loop model operation, intravascular fluorescence perfusion with MHI148-PEI, OTC with ECi, and 3D imaging via advanced fluorescence microscopy with whole-mounted (confocal laser scanning microscope, LCMC) and sectioned constructs (LSFM). Sectioned constructs were then rehydrated through decreasing ethanol concentrations, and nuclear staining was performed with SYTOX™ Green. One scan per animal or segment was obtained for every imaging modality used in this study (shown in Table 2).

FIG. 1.

Overview of the protocol. The experimental workflow is divided into four major steps. Following the AV loop operation (A), we performed an intravascular fluorescence staining with MHI148-PEI (B). After (C) a clearing method using a modified ethyl cinnamate (ECi)-based protocol, we performed high-resolution 3D fluorescence imaging modalities (D) for detailed visualization of the microcirculatory network. AV, arteriovenous; PEI, polyethylenimine (Created in BioRender. Köpple, C. https://BioRender.com/pk5wn8i).

Table 2.

Number of Acquired LCMC and LSFM Scans (One Scan per Animal and Segment)

Imaging	Day 7	Day 28
LCMC overviews of whole-mounted constructs	N = 5	N = 3
LSFM segments	N = 15	N = 9
Nuclear staining of segmented constructs	Not analyzed	N = 9

LCMC, confocal laser scanning microscope; LSFM, light-sheet fluorescence microscopy.

3D overviews of AV loop constructs

We previously demonstrated that MHI148-PEI intravascular perfusion imaging enables the detailed assessment of microvascular networks in PLA-based constructs in a 2D manner. MHI148-PEI perfusion was superior to CD31 staining of vessel walls in these constructs. To achieve 3D vascular visualization, a whole-mount scanning of AV loop constructs using confocal laser scanning microscope (16× objective) was performed after the ECi-clearing process. The images were reconstructed and depth color-coded using Leica LAS X software.

At day 7, the patent venous graft and the main vessel axis of the AV loop construct were clearly visible. Due to the high intravascular staining signal and low background noise, we achieved an unmatched signal-to-noise ratio. In addition to the clear differentiation between the arterial inflow, venous interposition graft, and outflow tract within the construct, minor regions at the arteriovenous anastomosis site are clearly lacking fluorescence signal, indicate the placement of the surgical knots. No significant peripheral vessel sprouts were detected at this stage of development at day 7. No extrinsic vascularization sprouts from the adjacent local tissue are visible at the opening of the chamber (Fig. 2A). By day 28, we observed advanced microvascular growth, characterized by a dense capillary network predominantly originating from the main vessel axis (Fig. 2B). A minor portion of additional capillaries entered the construct from the surrounding tissue and chamber opening, demonstrating progressive vascular integration in the surrounding tissue in an extrinsic manner (blue arrows). Some minor areas at the surgical anastomosis are lacking fluorescence staining, indicating placement of surgical knots. Within the center of the vascularized artificial tissue, there are quadratically arranged regions that are lacking fluorescence staining. These regions correspond to the plastic pins securing the loop in place.

FIG. 2.

Depth-encoded LCMC overview images of two AV loop constructs on day 7 (A) and day 28 (B) after intravascular MHI148-PEI staining and tissue clearing with ECi using a 16× objective in immersion oil. The arterial inflow (marked as Ar), venous graft (marked as Gr), and venous outflow (marked as Ve) are clearly visualized, but vascular sprouting is absent on day 7. The segments of the vessel axis differ in their diameter as a result of their respective size mismatch. Microsurgical knots bordering the venous graft are marked with white arrows. Asterisks (*) mark interruptions in vascular growth and tissue formation caused by the plastic pins of the AV loop chamber while blue arrows indicate superficially located capillaries extending from the chamber surroundings into the proximal chamber entry (panel B). A well-developed vascular network is visible on day 28, with capillary formation originating from the main vascular axis. LCMC, confocal laser scanning microscope.

3D detailed imaging of AV loop constructs

To obtain higher-resolution information about vascular structures within the constructs, we analyzed zoomed-in sections of LCMC overview images from day 28 constructs (Fig. 3). Following MHI148-PEI perfusion, the endothelium of the main vessel axis, higher-order branches, and microvessels were effectively stained, allowing for precise 3D characterization of microvascular features. Several branching vessels are interconnecting with each other and the capillary plexus (Fig. 3C). Vascular sprouts arising from the venous graft could be identified and continuously tracked. Similarly, branching vessels of the capillary plexus were adequately stained and tracked, allowing for precise 3D assessments of microvascular networks. Capillary vessel formation is limited to the interposed venous graft. No vessel sprouts are visible from the arterial inflow (Fig. 3B).

FIG. 3.

Zoomed-in sections of depth-encoded LCMC images of the AV loop construct on day 28 after ECi clearing. Image capture was performed using a 16× objective in immersion oil. Vessel walls are fully stained by MHI148-PEI. (A) Segment with the interposed venous graft (marked as Gr): capillary formation is visible from the graft with a thick vascular branch below and above the venous interposition graft (marked as Br). (B) Arterial inflow (Ar): capillary vessel formation is limited to the interposed venous graft (Gr). No vessel sprouts are visible from the arterial inflow. (C) Several interconnecting hierarchical branches are tracked within the graft segment of the construct.

3D overviews of AV loop segments

ECi-cleared AV loop constructs were precisely cut into thirds, corresponding to either the part involving the venovenous anastomosis, the interposition graft, or the arteriovenous anastomosis (Fig. 4A–C, respectively). These segments were then first individually scanned by light-sheet microscopy (LSMC) in order to obtain an even more detailed look with zoomed-in reconstructions. In each segment, the corresponding part of the axial vessel was clearly visible. High and evenly distributed fluorescence signal was limited to the vessel walls of the microcirculatory network with low autofluorescence background noise. In Figure 4A, the vessel wall was cut perpendicular just before the venovenous anastomosis, providing an intraluminal view into the main vessel axis and into an adjacent branching vessel running parallel to it. As previously stated, we observed vessel sprouting mainly from the loop axis and predominantly from the interposition graft, with multiple interconnecting higher-order branches (Fig. 4B, C). The dilation of the interposition graft directly after the arteriovenous anastomosis indicates the presence of a venous valve.

FIG. 4.

Segments of an ECi-cleared construct on day 28 imaged by LSMC using a 5× objective in ECi. (A) Segment of the venovenous anastomosis, which is marked with a white arrow (placement of microvascular knots is indicated by the lack of fluorescence staining within the vessel wall). Blue arrows indicate a tangential disruption of the vessel wall of the venous graft (Gr). Directly adjacent, the lumen of a branching vessel is also visible. (B) Segment centrally within the venous interposition graft (Gr) with branching vessels originating from it. (C) Segment of the arteriovenous anastomosis (white arrow) connecting the venous interposition graft (marked as Gr) with the inflow tract (a venous valve is downstream of the arteriovenous anastomosis). LSMC, light-sheet microscopy.

3D detailed imaging of AV loop segments

After successfully imaging the overviews of the AV loop segments in high resolution, we further tracked the branching patterns of higher-order microvessels with zoomed-in images of these sections. In Figure 5A, several interconnecting vessel branches originating from the interposition graft are shown. These images were obtained using the central section of the artificial tissue construct containing the interposition graft. Further scaling reveals the individual branching points of the capillary network in detailed resolution.

FIG. 5.

Zoomed-in sections of depth-encoded LSMC images (5× objective in ECi) within the venous interposition graft of the AV loop construct on day 28 (A) and detailed view of capillaries (B). Capillaries originating from the venous interposition graft are shown (the venous interposition graft itself is not visible).

Nuclear staining of AV loop segments (postprocessing)

To gain further cell-specific and molecular insights within the artificial constructs and the microvessel network, a scalable protocol for subsequent immunohistochemistry is needed. In a first step to that end, we removed the ECi (RI = 1.558) of cleared AV loop segments to adjust for RI differences needed in standard immunohistochemistry protocols. We removed the ECi using a decreasing alcohol gradient, which allows the rehydration of the constructs and to adjust the RI. In Figure 6, we present images of segments of AV loop constructs that have been counterstained with SYTOX™ Green after ECi removal. Using these steps, we were able to visualize the subcellular details within artificial constructs and to differentiate cellular nuclei from microstructures, for example, vessel wall components and cross-sections of capillaries.

FIG. 6.

3D visualization of nuclear and vascular structures in AV loop constructs on day 28 with LCMC (63× objective in immersion oil). The imaging workflow consists of an additional rehydration step of the ECi-cleared segments prior to nuclear staining with SYTOX™ Green. (A) Overview image (above) and magnified detailed section (below) of the arterial inflow section of an AV loop construct at day 28 after AV loop surgery. (B) This shows the lumen of the venous interposition graft (above) with a magnified detailed view of the region around the graft (below). (C) Magnified inflow section of the AV loop with a high cell count and a clearly visible capillary vessel (marked *). SYTOX™ Green nuclear staining appears in green, while MHI148-PEI staining is depicted in red.

Discussion

Recently, we demonstrated the suitability of PLA scaffolds for axially vascularized tissue engineering, providing an optimized environment for vessel ingrowth and microvascular network formation. Using conventional 2D fluorescence imaging, we showed that intravascular perfusion with the cationic near-infrared dye MHI148-PEI enables clear and specific staining of vascular structures in PLA-based constructs. 2D imaging alone cannot sufficiently capture the complex architecture of newly formed microvasculature within tissue-engineered constructs. By using OTC, different RI within the sample can be matched to overcome the penetration limit of light, thus enabling 3D imaging of complex and thick tissue samples. In this study, we therefore established and validated an integrated workflow: intravascular MHI148-PEI perfusion combined with ECi-based tissue clearing, applied to PLA-based AV loop constructs, followed by advanced fluorescence microscopy (confocal and light-sheet imaging). The result is a reproducible and robust protocol for 3D microvascular visualization, maintaining the anatomical integrity of vessels and offering the potential for both quantitative morphometric analyses and future functional imaging applications.

3D imaging of microvascular networks was routinely performed with high-resolution µCT scans after intravascular perfusion with a contrasting agent, like, for example, Microfil.¹³ As a high-viscosity agent, this approach suffers from significant drawbacks: shrinkage and loss of the structural integrity, endothelial damage, limited capillary penetration due to high viscosity, and possible incompatibility with downstream molecular analyses.^14,15 Additionally, 3D assessments of vascular networks, especially via µCTs, are inferior to 2D histomorphometrical quantification in dense and structurally heterogeneous tissue types.¹⁶ Here, we demonstrate for the first time that aqueous perfusion of axially vascularized tissue using MHI148-PEI, followed by ECi-based OTC, enables high-resolution 3D imaging of the entire microvascular network, including small capillaries and higher-order branches. This approach represents a major step forward, particularly for enabling downstream immunohistochemical analyses or functional fluorescence-based assessments, which are dependent on the anatomical integrity of the vessel wall. Our results indicate that MHI148-PEI perfusion overcomes several limitations inherent to established fluorescence tracing methods. While intravascular injection of Fluorescein isothiocyanate (FITC)- or Tetramethylrhodamine (TRITC)-dextran is standard for intravital microscopy, these high-molecular-weight polymers function as fluid-phase markers.^7,8,17 They do not bind to the vessel wall and are therefore unsuitable for the solvent-based dehydration steps required by ECi clearing, where they would be rapidly washed out. In contrast, MHI148-PEI leverages the electrostatic attraction between the polycationic PEI and the anionic endothelial glycocalyx, effectively “painting” the lumen in a manner that survives fixation and dehydration. Furthermore, when compared with lectin perfusion, our approach offers distinct advantages regarding tissue engineering applications. Lectins such as L. esculentum agglutinin are highly specific but prohibitively expensive for large-volume perfusion protocols. Moreover, lectin binding can be heterogeneous, with reports of reduced affinity in pathological or immature angiogenic vessels often found in engineered constructs. The nonspecific electrostatic binding mechanism of MHI148-PEI ensures uniform coverage of the entire patent vascular tree, regardless of the endothelial receptor phenotype. The aqueous nature of MHI148-PEI ensures deep capillary penetration without the risk of lateral dye diffusion or luminal blockage like in lipophilic dyes or viscous fluorescent gels, bridging between structural mapping and downstream molecular histology.¹⁷

While in ink- and Microfil cast-based visualizations of the vascular tree, the hardening and cooling of the perfusate lead to shrinkage of the vascular representation and disruption of the endothelial lining, Eci, as a solvent-based clearing method, is also known to induce tissue shrinkage due to the necessary dehydration steps.¹⁸ This shrinkage is not uniform or isotropic, rather than inhomogeneous, leading to structural distortion which might compromise morphometric measurements. However, clearing with ECi is a straightforward process, cost-effective, and a nontoxic agent for generating qualitative 3D visualizations of larger tissue samples.^19,20

MHI148-PEI binds to endothelial surfaces through a nonspecific electrostatic interaction between the positively charged cationic polymer (PEI) and the negatively charged glycosaminoglycans that form the endothelial glycocalyx, lacking the high specificity of ligand-receptor binding seen with probes like fluorescently labeled antibodies against endothelial markers such as CD31.²¹ While the mechanical disruption of the vessel wall is significantly reduced by aqueous MHI148-PEI perfusion in comparison with high-viscosity formulations, it is noteworthy that the cationic carrier PEI has potential intrinsic cytotoxic characteristics with acute disruption of plasma membrane integrity that may alter vessel permeability and induce membrane blebbing or endothelial cell death.²² While our results suggest that these potential cytotoxic effects are not a primary concern for the quality of the 3D reconstruction of the microvascular system, they might very well be an issue for further downstream molecular analysis.

3D Microfil casts of the microcirculatory architecture and cross-sections of India ink–perfused AV loop constructs offer long-term stability and can be archived virtually indefinitely. MHI148-PEI belongs to a class of fluorophores known for their limited stability, especially in aqueous solutions. A secondary absorption peak is attributed to a partially degraded form of the dye that continuously grows under illumination and indicates a breakdown of fluorescent signals.²³ Variations in processing times and light exposure during storage, preparation, and analysis could result in significant variability in signal intensity and thereby confounding comparative quantifications of vascular density. Despite these theoretical concerns, it is noteworthy that the effective fluorescence signal and dye stability of MHI148-PEI within an ECi-based clearing protocol have been studied before and showed that over the course of 3 years only a slight decrease in fluorescent signal occurred.¹²

In this study, we only performed SYTOX™ Green nuclear staining to outline potential functional downstream analysis. The challenges of performing multichannel antibody-based immunohistochemistry (IHC) on large, cleared tissue blocks are significantly more complex, and achieving a deep and homogenous penetration of antibodies into dense, cleared tissue is a major technical hurdle, creating a staining gradient that makes quantification unreliable. The process of rehydrating the tissue out of the ECi and back into aqueous buffers for IHC might cause mechanical stress, potential morphological distortion, and damage to antigenicity. Despite these concerns, we would like to point out that these challenges have already been overcome specifically within ECi-cleared tissue samples. Subsequent rehydration after an ECi protocol does not affect the efficacy of the clearing and simultaneously allows further 3D IHC staining.^11,12

Beyond structural imaging of the vascular tree, this approach might lay the groundwork in the future for functional microvascular imaging through double-fluorophore perfusion. The use of two complementary intravascular fluorophores allows for functional readouts. Combining MHI148-PEI with, for example, a hypoxia-sensitive fluorescence probe would enable the visualization of regional oxygenation differences, capillary perfusion heterogeneity, and local differences in tissue metabolism.^24,25 Besides using multiple fluorescent agents, our approach is in principle compatible with downstream immunohistochemistry after clearing and rehydration, enabling the investigation of functional and molecular markers in specific microvascular subdomains. Following ECi removal and rehydration, standard antibody labeling protocols can be applied¹² and give valuable insights into angiogenesis-relevant pathways, hypoxia-related signaling cascades within identified capillary networks, and inflammation responses and cellular infiltration patterns. By overlaying intravascular perfusion imaging data with immunofluorescence maps, subcompartment-specific analyses become possible. This paves the way for quantitative multiparameter characterization of engineered microenvironments and vessel formation at unprecedented resolution.

Expansion microscopy (ExM) physically enlarges the sample by embedding it in a swellable polymer matrix, enabling nanoscale resolution with conventional microscopes.²⁶ While our current workflow achieves high-resolution 3D imaging of microvascular networks, integrating ExM in future studies of axial vascularization can further enhance spatial resolution and allow ultra-detailed imaging of the construct in combination with IHC and thereby analyze the vascular branching, construct remodeling, and biointegration of the soft tissue into their surroundings.^12,27

In addition to experimental developments and refinements, the data generated from high-resolution 3D fluorescence imaging are ideally suited for automated image processing, vessel segmentation, and AI-assisted quantitative analysis. Manual vessel analysis is inherently limited due to time constraints and observer bias. With the availability of large volumetric datasets from cleared constructs, machine learning-based segmentation pipelines can extract complete microvascular trees, quantify vessel length and diameters, as well as individual differences in branching patterns, spatial distributions, and density across scaffold regions. AI-assisted vessel quantification might be a necessary advancement to harness the full potential of 3D datasets. Automated quantification tools will become essential for comparative studies, scaffold evaluation, and preclinical screening, and they will accelerate the translation of engineered vascularized constructs into clinical application.

Conclusion

We present a comprehensive methodological report demonstrating how aqueous MHI148-PEI fluorescence perfusion imaging, combined with ECi-based OTC, provides a novel and reproducible high-resolution approach for mapping microvascular networks within AV loop-engineered tissue constructs. Unlike previous perfusion and imaging techniques that are based on high-viscosity formulations, this protocol protects the endothelial integrity by reducing the perfusion pressure and mechanical forces in the microcirculatory system. Although not directly shown by comparing various imaging modalities with each other, we assume that the reduction of perfusion pressure and viscosity of the perfusate results in a more accurate representation of the capillary system and allows for both structural and functional downstream analyses. Future studies need to address these questions and build on these findings. Ultimately, these methods may advance our ability to study and optimize vascularization in TE constructs. Identification and spatial correlation of key regulators associated with axial vascularization may open up new avenues for the refinement of scaffold design and translational regenerative medicine applications.

Authors’ Contributions

C.K., V.J.S., and N.G.: Conceptualization. C.K. and L.P.: Data curation. C.K., L.P., and N.G.: Formal analysis. C.K., N.G., and V.J.S.: Funding acquisition. C.K., L.P., and N.S.P.: Investigation. C.K., L.P., and V.J.S.: Methodology. C.K. and N.G.: Project administration. C.K., N.G., and V.J.S.: Resources. C.K. and L.P.: Software. N.G., D.J.S., U.K., and V.J.S.: Supervision. N.G., U.K., D.J.S. and V.J.S.: Validation. C.K. and L.P.: Visualization. C.K.: Writing—original draft. N.S.P., U.K., D.J.S., N.G., and V.J.S.: Writing—review and editing.

Footnotes

Acknowledgments

ChatGPT 4o has been used to proofread the article and to improve readability. The authors assure that no alteration in regard to content or any other violation of publication ethics has been produced.

Funding Information

No funding was received for this article.

Disclosure Statement

The authors declare that there are no financial interests, commercial affiliations, or other potential conflicts of interest that could have influenced the objectivity of this research or the writing of this article.

References

Schmidt

, Hilgert

, Covi

, et al. Flow increase is decisive to initiate angiogenesis in veins exposed to altered hemodynamics. PLoS One, 2015; 10(1):e0117407; doi: 10.1371/journal.pone.0117407

Koepple

, Zhou

, Huber

, et al. Expression of connexin43 stimulates endothelial angiogenesis independently of gap junctional communication in vitro. Int J Mol Sci, 2021; 22(14):7400; doi: 10.3390/ijms22147400

Schmidt

, Hilgert

, Covi

, et al. High flow conditions increase connexin43 expression in a rat arteriovenous and angioinductive loop model. PLoS One, 2013; 8(11):e78782; doi: 10.1371/journal.pone.0078782

Nagy

, Benjamin

, Zeng

, et al. Vascular permeability, vascular hyperpermeability and angiogenesis. Angiogenesis, 2008; 11(2):109–119; doi: 10.1007/s10456-008-9099-z

Padera

, Kadambi

, di Tomaso

, et al. Lymphatic metastasis in the absence of functional intratumor lymphatics. Science, 2002; 296(5574):1883–1886; doi: 10.1126/science.1071420

Jilani

, Murphy

, Thai

SNM

, et al. Selective binding of lectins to embryonic chicken vasculature. J Histochem Cytochem, 2003; 51(5):597–604; doi: 10.1177/002215540305100505

Robertson

, Levine

, Haynes

, et al. Use of labeled tomato lectin for imaging vasculature structures. Histochem Cell Biol, 2015; 143(2):225–234; doi: 10.1007/s00418-014-1301-3

Zhu

, Deng

, Yu

, et al. Optimal combinations of fluorescent vessel labeling and tissue clearing methods for three-dimensional visualization of vasculature. Neurophotonics, 2022; 9(4):e045008; doi: 10.1117/1.NPh.9.4.045008

Koepple

, Pollmann

, et al. Microporous polylactic acid scaffolds enable fluorescence-based perfusion imaging of intrinsic in vivo vascularization. Int J Mol Sci, 2023; 24(19):14813; doi: 10.3390/ijms241914813

10.

Huang

, Brenna

, Khan

, et al. A cationic near infrared fluorescent agent and ethyl-cinnamate tissue clearing protocol for vascular staining and imaging. Sci Rep, 2019; 9(1):521; doi: 10.1038/s41598-018-36741-1

11.

Klingberg

, Hasenberg

, Ludwig-Portugall

, et al. Fully automated evaluation of total glomerular number and capillary tuft size in nephritic kidneys using lightsheet microscopy. J Am Soc Nephrol, 2017; 28(2):452–459; doi: 10.1681/ASN.2016020232

12.

Brenna

, Khan

, Picascia

, et al. New technical approaches for 3D morphological imaging and quantification of measurements. Anat Rec (Hoboken), 2020; 303(10):2702–2715; doi: 10.1002/ar.24463

13.

Arkudas

, Beier

, Pryymachuk

, et al. Automatic quantitative micro-computed tomography evaluation of angiogenesis in an axially vascularized tissue-engineered bone construct. Tissue Eng Part C Methods, 2010; 16(6):1503–1514; doi: 10.1089/ten.tec.2010.0016

14.

Margolis

, Merlo

, Chanthavisay

, et al. Comparison of micro-CT image enhancement after use of different vascular casting agents. Quant Imaging Med Surg, 2024; 14(3):2568–2579; doi: 10.21037/qims-23-901

15.

Zagorchev

, Oses

, Zhuang

, et al. Micro computed tomography for vascular exploration. J Angiogenes Res, 2010; 2(1):7; doi: 10.1186/2040-2384-2-7

16.

Ghanavati

, Yu

, Lerch

, et al. A perfusion procedure for imaging of the mouse cerebral vasculature by X-ray micro-CT. J Neurosci Methods, 2014; 221:70–77; doi: 10.1016/j.jneumeth.2013.09.002

17.

, Song

, Zhao

, et al. Direct labeling and visualization of blood vessels with lipophilic carbocyanine dye DiI. Nat Protoc, 2008; 3(11):1703–1708; doi: 10.1038/nprot.2008.172

18.

Masselink

, Reumann

, Murawala

, et al. Broad applicability of a streamlined ethyl cinnamate-based clearing procedure. Development, 2019; 146(3):dev166884; doi: 10.1242/dev.166884

19.

Brenna

, Simioni

, Varano

, et al. Optical tissue clearing associated with 3D imaging: Application in preclinical and clinical studies. Histochem Cell Biol, 2022; 157(5):497–511; doi: 10.1007/s00418-022-02081-5

20.

Sun

, Tiziana

, Khan

, et al. A simple optical tissue clearing pipeline for 3D vasculature imaging of the mediastinal organs in mice. Int J Exp Pathol, 2021; 102(4–5):218–227; doi: 10.1111/iep.12399

21.

Vajpayee

, Picascia

, Casciano

, et al. Fluorescent water-soluble polycationic chitosan polymers as markers for biological 3D imaging. Chem Biomed Imaging, 2024; 2(10):721–730; doi: 10.1021/cbmi.4c00028

22.

Moghimi

, Symonds

, Murray

, et al. A two-stage poly(ethylenimine)-mediated cytotoxicity: Implications for gene transfer/therapy. Mol Ther, 2005; 11(6):990–995; doi: 10.1016/j.ymthe.2005.02.010

23.

Yang

, Tian

, Xiao

, et al. Photodegradation of cyanine and merocyanine dyes. Dyes Pigments, 2001; 49(2):93–101; doi: 10.1016/S0143-7208(01)00012-2

24.

Mirabello

, Cortezon-Tamarit

, Pascu

. Oxygen sensing, hypoxia tracing and in vivo imaging with functional metalloprobes for the early detection of non-communicable diseases. Front Chem, 2018; 6:512; doi: 10.3389/fchem.2018.00027

25.

Sandhu

, Kydd

, Jaworski

. Luminescent probe based techniques for hypoxia imaging. J Nanomed Res, 2017; 6(3):e00160; doi: 10.15406/jnmr.2017.06.00160

26.

Sun

, Picascia

, Khan A Ul

, et al. Application of ethyl cinnamate based optical tissue clearing and expansion microscopy combined with retrograde perfusion for 3D lung imaging. Exp Lung Res, 2020; 46(10):393–408; doi: 10.1080/01902148.2020.1829183

27.

Unnersjö-Jess

, Scott

, Sevilla

, et al. Confocal super-resolution imaging of the glomerular filtration barrier enabled by tissue expansion. Kidney Int, 2018; 93(4):1008–1013; doi: 10.1016/j.kint.2017.09.019