Tumor Vascular Remodeling Affects Molecular Dissemination to Lymph Node and Systemic Leukocytes

Abstract

Angiogenic and lymphangiogenic remodeling has long been accepted as a hallmark of cancer development and progression; however, the impacts of this remodeling on immunological responses, which are paramount to the responses to immunotherapeutic treatments, are underexplored. As immunotherapies represent one of the most promising new classes of cancer therapy, in this study, we explore the effects of angiogenic and lymphangiogenic normalization on dissemination of molecules injected into the tumor microenvironment to immune cells in lymph nodes draining the tumor as well as in systemically distributed tissues. A system of fluorescent tracers, size-matched to biomolecules of interest, was implemented to track different mechanisms of tumor transport and access to immune cells. This revealed that the presence of a tumor, and either angiogenic or lymphangiogenic remodeling, altered local retention of model biomolecules, trended toward normalizing dissemination to systemic organs, and modified access to lymph node-resident immune cells in manners dependent on mechanism of transport. More specifically, active cell migration by skin-derived antigen presenting cells was enhanced by both the presence of a tumor and lymphangiogenic normalization, while both angiogenic and lymphangiogenic normalization restored patterns of immune cell access to passively draining species. As a whole, this work uncovers the potential ramifications of tumor-induced angiogenesis and lymphangiogenesis, along with impacts of interrogation into these pathways, on access of tumor-derived species to immune cells.

Impact Statement

Angiogenic and lymphangiogenic normalization strategies have been utilized clinically to interrogate tumor vasculature with some success. In the age of immunotherapy, the impacts of these therapeutic interventions on immune remodeling are unclear. This work utilizes mouse models of angiogenic and lymphangiogenic normalization, along with a system of fluorescently tagged tracers, to uncover the impacts of angiogenesis and lymphangiogenesis on access of tumor-derived species to immune cell subsets within various organs.

Introduction

Melanoma represents a significant clinical problem, leading to more than 320,000 cases in 2020 alone.¹ While treatable when localized, survival is disappointingly low for advanced disease.² Understanding the pathways that contribute to disease progression represents a crucial need to address the high mortality associated with advanced melanoma.

It is well established that tumors, including melanomas, induce angiogenic and lymphangiogenic remodeling locally^3–7: prolonged angiogenic signaling causes blood vessels to expand and become unruly and dysfunctional, while lymphatic vessels tend to collapse and lose functionality.^7–9 This remodeling has been implicated in blunting the potency of anticancer therapeutic agents,^10,11 and approaches to counter these pathways have demonstrated promise in restoring oxygen tension and the accumulation of nanomedicines.^12–19 Remodeling of the vasculature in tumors is thus implicated in influencing not only the biology of disease,^6,20–23 but also its progression and response to therapy.^24–30

While these concepts have primarily been considered in the context of chemotherapy, immunotherapy is increasingly used to treat advanced melanoma, being a cancer type that is generally considered highly immunogenic.³¹ Understanding how remodeling within the tumors may influence the local immune microenvironment as well as the potential of existing immunotherapeutic regimens to combat disease is thus of high relevance; in fact, localized angiogenic normalization has been shown to enhance the efficacy of immunotherapies in preclinical models,^24,26,32 however, the mechanisms underlying this remain underexplored. Specifically, while there are numerous attributes of the immune system and tumor microenvironment that contribute to melanoma immune regulation,^31,33–35 the role of the lymphatic system is increasingly appreciated.^36–39

In addition to the multiple immunomodulatory roles of lymphatic endothelial cells,^6,40–43 the lymphatic system plays crucial roles in regulating how antigens and immune cells transit and emigrate from the tumor microenvironment to then distribute to sentinel (tumor-draining) lymph nodes.^44–49 Of note, changes in fluid fluxes resulting from angiogenic modulation within the tumor are likely to impact the resulting lymphatic fluid flows, further altering antigen dissemination into lymph nodes (LN) and LN-resident leukocytes. How alterations in lymphatic transport that manifest due to remodeling of the tumor microenvironment, including changes in blood vasculature and fluid fluxes locally, may regulate tumor immunity, however, remains unexplored.

Numerous interesting immunological studies have been conducted and are ongoing that investigate how lymphatic function is influenced by tumors. However, fundamental measurements of how transport through the lymphatic and vascular systems are affected by remodeling of the tumor microenvironment, akin to those that established the influence of these remodeling hallmarks on tissue hypoxia and nanocarrier delivery, are lacking.

In this work, we evaluate the effects of alterations in skin angiogenesis and lymphangiogenesis that result from melanoma formation on the dissemination of molecules from the skin microenvironment. This was achieved using a panel of fluorescently labeled tracers with minimal spectral overlap that were coinjected into the lateral dorsal skin versus melanomas of C57/Bl6 mice. Levels of tracers that were selected based on their size that restricted the mechanism in which they were cleared from the tissue injection site within the tumor as well as in tissues implicated in tumor metastasis and immune regulation, including the tumor-draining lymph node (TdLN), spleen, lung, and liver, were measured in bulk tissue homogenates and by flow cytometry, separately quantifying the extent of tissue accumulation versus cell association, respectively. The influence on profiles of tracer dissemination of vascular remodeling was assessed using interventions that countered the lymphangiogenic and angiogenic remodeling observed in this melanoma model.

Results reveal that remodeling of the blood and lymphatic vasculature in this widely used tumor model diminishes the overall transit by tracers from tumors that is mediated by both lymphatic-trafficking cells and lymph drainage to LNs, both in a manner that transpires independently from tumor effects on immune cell uptake of those tracers. These results suggest how tumor vascular remodeling, previously proposed to improve cancer therapies leveraging nanotechnology and drug delivery to tumors, also has the potential to regulate both local and systemic immune signaling implicated in disease progression and immunotherapeutic response.

Methods

Cell culture

B16F10 and B16F10 cells exhibiting vasoendothelial growth factor (VEGF)-C overexpression (VC) mouse melanoma cells were cultured in Dulbecco's modified Eagle's medium with 10% heat-inactivated fetal bovine serum and 1% penicillin/streptomycin/amphotericin B from Life Technologies (Carlsbad, CA).

Animal tumor models

C57/Bl6 mice were purchased at 6 weeks of age from The Jackson Laboratory (Bar Harbor, ME). All protocols were approved by Georgia Tech's Institutional Animal Care and Use Committee. For wild-type (WT) and R2 versus VC animal cohorts, either 0.5 × 10⁶ B16F10 or B16F10-VC murine melanoma cells, respectively, were intradermally implanted into the left dorsal skin of 6- to 8-week-old mice on day 0. For cohorts termed R2, from day 2 post-cell implantation until endpoint, vandetanib at 50 mg/kg in 1% Tween 80 (Sigma-Aldrich) or vehicle alone was administered daily by oral gavage using a 22-gauge reusable small animal feeding needle (Cadence Science, Staunton, VA). Tumor dimensions were measured with calipers and reported as ellipsoidal volume.

Tissue concentrations of VEGF-A and VEGF-C

Excised tumor and skin tissues were dissected directly into Dulbecco's phosphate-buffered saline (D-PBS; VWR International, Inc., West Chester, PA) containing tubes and homogenized with 1.4 mm acid-washed zirconium grinding beads (MP Biomedicals, Santa Ana, CA). Homogenized samples were stored at −80°C until analysis. Dilutions of 1:10 of mixed tissue homogenates in reagent buffer were measured for VEGF-A and VEGF-C by enzyme-linked immunosorbent assay (R&D Systems, Inc., Minneapolis, MN) as per the manufacturer's protocol.

Immunohistochemistry and imaging

Tumor, skin, and LN tissues were frozen in an optimum cutting temperature compound (Sakura Finetek USA, Inc., Torrance, CA) in 2-methylbutane (Sigma-Aldrich) chilled by liquid nitrogen and frozen tissue blocks immediately stored at −80°C. Tissue sections were fixed with prechilled acetone for 10 min at 4°C and subjected to standard immunofluorescence protocols using the following antibodies, which were obtained from Thermo Fisher Scientific, Inc., unless otherwise specified: fluorescein isothiocyanate (FITC)-conjugated rat anti-mouse CD31 (1:50), rabbit anti-mouse Lyve-1 (1:250), Alexa Fluor 633 goat anti-rabbit (1:300), Armenian hamster anti-mouse CD3Ɛ (1:50), Alexa Fluor 647 goat anti-hamster (1:300; Abcam plc., Cambridge, MA), Alexa Fluor 488-conjugated rat anti-mouse CD169 (1:100; BioLegend, Inc., San Diego, CA), biotinylated rat anti-mouse F4/80 (1:200; Life Technologies), streptavidin-Alexa Fluor 555 (1:400; Life Technologies), biotinylated rat anti-mouse B220 (1:250), and Alexa Fluor 488-conjugated Armenian hamster anti-mouse CD11c (1:50; BioLegend, Inc.).

Blocking and antibody dilutions were performed with 10% donkey serum (Sigma-Aldrich) in D-PBS. Slides were washed with 0.1% Tween 20 (Sigma-Aldrich) in D-PBS for washing steps, counterstained with DAPI (VWR International, Inc.), and imaged using a 710 NLO confocal microscope (Carl Zeiss Microscopy Ltd, Jena, Germany) with a 20 × magnification objective.

Microcomputed tomographic imaging for vascular measurements

Animals were perfused with saline followed by neutral buffered formalin (Thermo Fisher Scientific, Inc.) for 10 min, then with saline to rinse, and lastly Microfil (Flow Tech, Inc., Carver, MA) catalyzed at a viscosity appropriate for small vessels (5 mL lead-based contrast agent: 2.5 mL diluent: 0.25 mL curing agent). Afterward, perfused mice were carefully stored at 4°C overnight to cure the contrast agent. The following day, skin or tumor samples were harvested and stored in D-PBS. Microcomputed tomographic imaging (μCT) was accomplished using SCANCO Medical μCT50 (SCANCO USA, Inc., Wayne, PA). μCT image slices were constrained using manual selection of the sample outline and processed with a Gaussian filter at a consistent global threshold via the SCANCO Medical μCT Evaluation Program before three-dimensional reconstruction.⁵⁰

Fluorescent tracers

Five hundred nanometer yellow-green and red fluorescent (505/515 and 580/605 excitation/emission, respectively) carboxylate-modified microspheres were purchased from Thermo Fisher Scientific, Inc. Forty thousand Daltons tetramethylrhodamine isothiocyanate (TRITC) dextran was purchased from Sigma-Aldrich. Five hundred thousand Daltons amino-dextran (Thermo Fisher Scientific, Inc.) was covalently labeled by incubation with Alexa Fluor 647- or 700-NHS-Ester dyes (Thermo Fisher Scientific, Inc.) in 0.1 M NaHCO₃ at pH 8.4 for 4 h on a tube rocker. AF647 and AF700 dextran-dye conjugates were purified from unreacted free dye by Sepharose CL-6B gravity column chromatography after conjugation. Purified dextran-fluorophore conjugates were further confirmed free of unconjugated dye by a second Sepharose CL-6B column analysis.⁴⁴ All reagents were used and maintained under sterile conditions. Hydrodynamic sizes were confirmed preinjection by dynamic light scattering using a Zetasizer Nano ZS (Malvern Instruments, Ltd., Malvern, United Kingdom).

Tracer injections

Fluorescent tracers suspended in saline were coinfused by syringe pump at a rate of ∼300 nL/s directly into the tumor center for tumor-bearing groups or into the skin for naive groups using a 27–31-gauge needle while mice were under isoflurane anesthesia. For biodistribution experiments, 500 nm red or yellow-green fluorescent microspheres (19 pM), 30 nm AF700 or AF647 dextran (4.8 μM), and 10 nm TRITC dextran (4.8 μM) were coinfused in 10 μL of total saline. Mice were euthanized via CO₂ asphyxiation at the prescribed times post-tracer injection for each experiment.

Tracer biodistribution analyses

At 4, 24, and 72 h post-tracer injection, mice were sacrificed and the tumor or skin injection site, tumor-draining axillary and brachial LNs, spleen, lungs, liver, and kidneys were harvested and homogenized in D-PBS using 1.4 mm acid-washed zirconium grinding beads with a FastPrep-24 Automated Homogenizer. Whole tissue homogenate fluorescence was measured with a Synergy H4 BioTek plate reader (BioTek Instruments, Inc., Winooski, VT), compensation was applied, and fluorescent tracer amounts and concentrations were calculated from standard curves made by spiking individual naive tissue homogenates with tracer solution.

Tracer flow cytometry analyses

Axillary and brachial draining LNs were pooled and incubated at 1 mg/mL collagenase D (Sigma-Aldrich) in D-PBS with calcium and magnesium for 1 h at 37°C, passed through a 70 μm cell strainer (Greiner Bio-One, Monroe, NC), washed, and resuspended in a 96-well plate (VWR International, Inc.) for staining. Lung tissues were treated with the same procedure as lymph nodes.

Spleen capsules were disrupted using 18-gauge needles, and the cell suspension was passed through a 70 μm cell strainer, pelleted, then incubated with red blood cell lysis buffer (Sigma-Aldrich) for 7 min at room temperature, diluted with D-PBS, washed, and resuspended. Liver tissues were disrupted using 18-gauge needles and then passed through a 70 μm cell strainer, centrifuged at 60 g for 1 min to remove debris, and the supernatant was collected into a different tube and centrifuged at 300 g for 5 min. The pellet was resuspended, layered on to a lymphocyte separation medium (Thermo Fisher Scientific, Inc.), and centrifuged for 20 min at 400 g. The mononuclear cell layer was recovered, incubated with red blood cell lysis buffer as before, washed, and resuspended for staining.

All antibodies for flow cytometry were from BioLegend, Inc., unless otherwise stated. Cells were blocked with 2.4G2 (Tonbo Biosciences, San Diego, CA) for 5 min on ice, washed, then stained with a fixable viability dye eFluor 455UV (1:1000; eBioscience, San Diego, CA) for 15 min on ice before quenching with 0.1% bovine serum albumin in D-PBS (flow cytometry buffer). Antibodies were prepared in flow cytometry buffer at the following dilutions based upon preliminary titrations: PerCP anti-mouse CD45 (0.5:100), PE-Cy7 anti-mouse CD11b (0.625:100), BV421 anti-mouse CD11c (5:100), BV605 anti-mouse CD169 (5:100), BV650 anti-mouse B220 (2:100), BV711 anti-mouse CD3 (1.25:100), and BV785 anti-mouse F4/80 (2.5:100). Cells were fixed with 4% paraformaldehyde (VWR International, Inc.) and kept at 4°C until analyzed with a customized BD LSRFortessa flow cytometer (BD Biosciences, Franklin Lakes, NJ). Compensation was performed with AbC compensation beads (Thermo Fisher Scientific, Inc.) and data were analyzed using FlowJo software v10 (FlowJo, LLC, Ashland, OR).

Statistical analysis

Data are represented as the mean accompanied by standard error of the mean, and statistics were calculated using Prism 6, 7, and 8 software (GraphPad Software, Inc., La Jolla, CA). Statistical significance was defined as p < 0.05, 0.01, 0.005, and 0.001, respectively, unless otherwise specified. Area under the curve (AUC) was calculated using the built-in Prism analysis tool.

Experiment

Fluorescent tracer panel for in situ tracing of molecular dissemination from tumors

Melanomas induce alterations in skin vasculature^51–54 (Fig. 1a) that play multiple barrier and transport functions relevant to malignant disease progression. These alterations thus have the potential to alter the distribution of biomolecules and cells present in the tumor microenvironment to other tissues in a manner relevant to both metastatic dissemination and antitumor immune responses.^{44,45,55–57} To interrogate how profiles of dissemination from the skin microenvironment are impacted by disease, a method to assess the way macromolecules and particles within the tissue are cleared and distribute into the lymphatic drainage basin versus systemic tissues was implemented.

FIG. 1.

Tumor-mediated alterations in vasculature alter dissemination from the local microenvironment. Tumor vasculature remodels (a). (b) Tracer system used to assess changes in tumor-disseminated transport. Concentration of 500, 30, and 10 nm tracers within dLNs assessed on a bulk basis (c) and by flow cytometry (d). (e) Frequency of live dLN-resident cells accessing each size tracer across time. (f) Counts for tracer-positive cells expressing CD11c, F4/80, B220, or CD3 in local dLN and systemic tissues, including the spleen, lungs, and liver. n = 6 mice per group. dLNs, draining lymph nodes.

In this approach, a panel of tracers comprised of bioinert polymers resistant to proteolytic and hydrolytic degradation selected for their capacity to quantitate different mechanisms of transport from the tissue interstitium relevant to melanoma⁵⁸ were coinjected as a single suspension (10 μL total volume) into the naive skin or melanomas in C57/Bl6 mice at a slow pump-controlled rate (∼300 nL/s). These tracers included 500 nm polystyrene spheres that, due to their restriction by the small pore size of the extracellular matrix after administration, require the functions of emigrating antigen presenting cells for clearance (Fig. 1b). Dextrans that are retained in the tissue site of injection due to their exclusion from the blood capillaries and capacity to traverse the small pore size of the tissue extracellular matrix, namely, those 30 and 10 nm in hydrodynamic size, were additionally used.

When administered into the naive skin, the tracer that relied on cell-mediated mechanisms to be cleared from the tissue injection site (500 nm sphere) accumulated within LNs draining the skin injection site at dramatically lower extents and slower rates compared with the nanoscale dextran tracers 30 and 10 nm in hydrodynamic diameter that can directly drain into lymphatic vessels through the direction of interstitial flow (Fig. 1c). Correspondingly, the total LN leukocyte (as defined by CD45⁺) association with 500 nm tracer over 72 h postadministration analyzed flow cytometrically using AUC, which quantifies the measured amount over a given time, was substantially lower compared with levels for that of the dextran tracers (Fig. 1d).

Among the latter two, association with LN leukocytes over this time span was greater for the 10 versus 30 nm tracer (Fig. 1d). It should be noted that although they were administered as a single suspension, the 500, 30, and 10 nm tracers dispersed among cells that reside within the LN draining the injection tissue site and those cells generally associated with one tracer type, and only very rarely two (Fig. 1e).

When evaluating the relative extent of association of administered tracers with various target cell subtypes in not only draining LNs (dLNs) but also in systemic tissues that included the spleen, lungs, and liver, differences in association specifically with dLN-resident leukocytes became starkly apparent. In interpreting these results, it is important to note that the chemistry of the 500 nm versus 10 and 30 nm tracers differs, which may influence phagocytosis. In examining access to systemic organs, 500 nm tracers appeared to accumulate in splenic CD11c⁺ cells to high extents that even exceeded that seen for dLNs (Fig. 1f).

However, the total extent of cell association was relatively low for 500 nm tracers compared with lymph-draining tracers that associated with far greater extents with all the analyzed leukocyte populations. Of the total measured fraction of CD11c⁺ cells that 30 nm tracers were found to associate with after administration, the numbers of cells localized within dLNs were greatest 24 h postadministration, and then decreased thereafter, with a large proportion of tracer association with CD11c⁺ cells being in the spleen. Thirty nanometer tracer association with F4/80⁺, B220⁺, and CD3⁺ cells was also high in the spleen, but also proportionally exhibited the highest levels of association in dLNs 24 h postadministration.

In stark contrast, 10 nm tracers, that associated with dLN cells overall to highest extents of all tested tracers (Fig. 1d), were most abundantly associated with CD11c⁺ and F4/80⁺ cells isolated from spleens and lungs at all time points of analysis (Fig. 1f). This demonstrates that despite their increased access to dLN leukocytes, their relative, or selective, association with dLN leukocytes is lower compared with the other tested tracers. Moreover, whereas the total association of 500 and 30 nm tracers with cells generally increased over time, the 10 nm tracers associated with high leukocyte numbers at the earliest analysis times.

These results show that despite distributing from the tissue microenvironment via the lymphatic vasculature, tracers with differing capacities to transit the tissue space associate with vastly differing leukocyte numbers and qualities at differing timescales, both locally in the dLN versus in systemic tissues.

Modulation of melanoma angiogenesis and lymphangiogenesis on tumor formation and tissue structure

Melanomas formed from the implantation of B16F10 cells in the lateral dorsal skin corresponding to the site of tracer injection in previous studies. In these formed tumors (Fig. 2a), concentrations of VEGF-A, an angiogenic growth factor, and VEGF-C, a lymphangiogenic growth factor, were assessed throughout melanoma development (Fig. 2b). Concentrations of VEGF-C were found to be maintained in early-stage (day 5) tumors before trending toward declining later during tumor growth (Fig. 2b). Concentrations of VEGF-A, on the contrary, showed a progressive increase with tumor progression (Fig. 2b).

FIG. 2.

Model for interrogation of angiogenesis and lymphangiogenesis within melanomas. (a) Tumor growth after implantation of WT B16F10 tumors in immunocompetent mice. (b) Intratumoral VEGF-A and VEGF-C concentrations of B16F10 (WT) tumors over the course of tumor progression. (c) VEGF-C measurements for B16F10-VC compared with B16F10 cells in culture per 1000 cells plated. (d, e) Growth of B16F10-VC (d) and B16F10 tumors with R2 inhibition (e) after implantation in immunocompetent mice. (f) Concentration of VEGF-A and VEGF-C in naive skin, WT, VC, and R2 tumors at day 7 of tumor growth. Blood and lymphatic densities quantified from immunohistochemistry staining and images; white arrows indicate vessel structures; scale bars = 200 μm (g). (h) Total blood (CD31⁺Lyve-1⁻) and lymphatic (Lyve-1⁺) structures in naive skin, and WT < VC, and R2 tumors, assessed by immunohistochemistry at day 7 of tumor growth. (i) 3D reconstructions of microcomputed tomography of the tumor blood vasculature at day 7 postimplantation for naive skin, WT, VC, and R2 tumors at day 7 of tumor growth; scale bars = 1 mm. (j) Total number of cells within LNs draining naive skin, and WT, VC, and R2 B16F10 tumors, assessed by flow cytometry. ^†Indicates significance for VEGF-A WT days 5–9 grouped versus naive by t-test; *indicates significance by multiple t-tests; *indicates p < 0.05; **indicates p < 0.01; ***indicates p < 0.005; ****indicates p < 0.001; n = 4–12 animals per group. 3D, three dimensional; LNs, lymph nodes; VC, VEGF-C overexpressing; VEGF, vasoendothelial growth factor; WT, wild type.

To assess the effects of changes in these two growth factors on the tumor microenvironment, models of altered angiogenic and lymphangiogenic signaling were also implemented. The first were mice whose melanomas were formed from a VC B16F10 cell line that expresses elevated levels of VEGF-C in vitro (Fig. 2c), and whose VEGF-C overexpression (VC) we hypothesized would counteract the diminishing levels of VEGF-C seen in melanomas formed from parental (WT) B16F10 cells. Upon implantation, VC tumors grew at similar rates (Fig. 2d, e) relative to WT B16F10 tumors (Fig. 2a). In these tumors on day 7, not only were reductions in VEGF-C levels reversed to even exceed those of naive skin, but also VEGF-A levels were restored (Fig. 2f).

These effects on total growth factor levels were furthermore confirmed immunohistochemically (Fig. 2g), with higher levels of Lyve-1⁺ structures per tissue area in VC tumors (Fig. 2h), but no measured differences in densities of CD31⁺Lyve-1⁻ structures, a result confirmed by μCT imaging demonstrating sustained hypervascularization in VC tumors (Fig. 2i).

With the intent to restore altered VEGF-A levels during tumor growth, VEGFR2, the receptor for VEGF-A responsible for angiogenesis, was inhibited through daily administration of vemurafenib. This regimen, referred to as R2, resulted in slowed B16F10 melanoma growth relative to WT tumors (Fig. 2e), an expected result given the known role of angiogenesis in promoting tumor growth.^21,59,60 Concentrations of VEGF-A were also diminished in R2 tumors by vemurafenib treatment, while VEGF-C concentrations remained unchanged (Fig. 2f). R2 tumors also exhibited decreased numbers of both CD31⁺Lyve-1⁻ and Lyve-1⁺ structures in immunohistochemically stained and imaged tumor sections compared with WT tumors (Fig. 2h). R2 was found to furthermore substantially diminish tumor vascularization to extents that more closely mirrored that of naive skin (Fig. 2i).

The number of cells in LNs draining the various skin microenvironment models was also measured. Overall, the presence of a melanoma increased the number of LN leukocytes, an effect sustained in VC tumors (Fig. 2j). LNs draining tumors in R2 mice, however, exhibited a reduction in the total number of cells compared with both LNs draining WT tumors and the naive skin (Fig. 2j). Leukocyte numbers within recovered spleens, lungs, and liver, however, remained unchanged (data not shown). Altered VEGF-A signaling and VEGF-C expression modulated both blood and lymphatic vascularization in melanomas as well as their draining LNs.

Normalization of lymphangiogenic and angiogenic tumor remodeling alters tracer retention and dissemination to systemic tissues

Using these models of altered angiogenic and lymphangiogenic melanoma microenvironments, the role of these vascular remodeling pathways on regulating molecular dissemination from the skin was explored using our tracer system (Fig. 1). The 500, 30, and 10 nm tracers were coinjected into the naive skin or lesions of WT, VC, or R2 tumor-bearing animals and bulk levels of tracer in various tissues measured in homogenates of tissues harvested from mice at various times postinjection (4, 24, 72 h). The 500 nm tracer was found to exhibit the greatest exposure in the local tissue injection site irrespective of the tested model (Fig. 3). Of the lymph-draining tracers, the 30 nm dextran was retained to greater extents within the injected tissue site compared with the 10 nm dextran (Fig. 3).

FIG. 3.

(Lymph)angiogenic normalization effects on retention of tracers at tissue site of injection. Retention at site of injection as area under curve of measurements taken 4, 24, and 72 h after administration in naive skin, or WT, VC, or R2 B16F10 tumors. *Indicates significance by two-way ANOVA with Tukey's post hoc test; *indicates p < 0.05; ***indicates p < 0.005; ****indicates p < 0.001. ANOVA, analysis of variance.

However, the extent of retention differed based on the local microenvironment: WT tumors resulted in higher retention of the lymph-draining 30 nm tracer compared with naive skin (Fig. 3). VC tumors exhibited similarly high levels of 30 nm tracer retention as WT tumors, but R2 reversed this trend such that retention was more similar to naive skin (Fig. 3). Retention of the 10 nm tracer in the tissue injection site remained largely unchanged in WT tumor bearing animals for the smallest tracer size assessed (10 nm) and was decreased by R2 inhibition (Fig. 3).

Extending the focus outside of the tissue injection site, total tracer levels within systemic tissues, as well as their relative association with resident leukocytes, were measured. When considering the summed fraction of the total amount of injection that was measured in the spleen, liver, kidneys, and lungs over 72 h postadministration, the presence of a melanoma was found to increase the total levels of 500 nm tracer in the skin (Fig. 4a, left). This increase relative to the skin was reversed in VC animals, however (Fig. 4a, left). The presence of a melanoma had no influence on levels of 30 nm tracer exposure in systemic tissues relative to naive skin but was decreased in VC tumors (Fig. 4a, middle). Exposure of 10 nm tracer in systemic tissues was decreased by a melanoma relative to naive skin, with greater decreases seen in VC tumors, but unchanged in R2 tumors (Fig. 4a).

FIG. 4.

Local (lymph)angiogenic normalization normalizes systemic dissemination of tumor-derived molecules to an extent. (a) Accumulation of 500, 30, and 10 nm tracers in systemic tissues (liver, kidneys, spleen, and lung), after injection into naive skin or day 7 melanoma. Comparison of cellular subtype exposure versus whole tissue exposure per tracer in the spleen (b), lungs (c), and liver (d). n = 4–6 mice per group. *Indicates significance by multiple t-tests (a) or linear fit of slope through all points is significantly nonzero; *indicates p < 0.05; ***indicates p < 0.005. (b–d); n = 4–9 animals.

Thus, on a bulk basis, systemic exposure to each tracer size was altered as a result of a tumor in a manner that differed based on the mechanism of transport, and tended to decrease in VC tumors while remaining unchanged in R2-inhibited tumors.

As bulk access to a tissue does not necessarily imply access to cells within that tissue, flow cytometry was used to assess the access of each tracer to leukocytes expressing CD11c, F4/80, B220, and CD3 in various tissues after administration in the healthy or malignant skin. Whether the extent of accumulation resulted in increased access (reflected by a statistically significant correlation in bulk tracer AUC versus association with each leukocyte subtype in each tissue) was assessed. Overall, access of 500 nm tracer to leukocytes was found to be largely independent of the total extent of tracer accumulation in the spleen, lungs, and liver (Fig. 4b–d), with R2-inhibited tumors often displaying a higher efficiency of access compared with the other tumor subtypes (Fig. 4b–d, left).

However, in the lungs, 500 nm access to F4/80⁺ cells correlated to the total tracer accumulation level in that tissue (Fig. 4c). Among tracers that could be transited from the tumor microenvironment directly via lymph drainage, this relationship was variable, with correlations of B220⁺ cell access in the lungs to 10 nm tracer (Fig. 4c, right). There was a general trend of inverse correlation between the total levels of accumulation by 10 nm tracer with CD11c⁺ and F480⁺ cells in the lung, among all 30 nm cell types and F4/80⁺ cells accessing 10 nm tracer (Fig. 4d, right). Overall, access to tracers administered in the skin microenvironment by leukocytes in various systemic tissues did not necessarily correlate with the total local availability of tracer, and only lowly influenced by modulation of the presence of a melanoma, despite alterations in total accumulation associated with disease.

Normalization of lymphangiogenic and angiogenic remodeling in tumors increases transport of tumor-derived molecules to draining lymph nodes

As LNs house the adaptive immune system and are increasingly acknowledged for their role in the development of antitumor immune responses and mediation of immunotherapy responsiveness,^45,61–63 we next focused on the assessment of how tumors and their vascular remodeling influence the access by LN-resident leukocytes to tracers administered into the skin. This was first assessed on a bulk basis within the skin-draining brachial and axillary LNs that both drain the lateral dorsal skin.⁶⁴ This revealed that the transport of 500 and 30 nm tracers diminished the presence of a melanoma, an effect not completely reversed in either altered model of tumor microenvironment (Fig. 5a). Accumulation of the 10 nm tracer in LNs was likewise diminished after administration in a melanoma compared with skin.

FIG. 5.

(Lymph)angiogenic normalization normalizes dissemination of tumor-derived molecules to dLNs. (a) Total accumulation of each tracer within dLN as AUC from 4 to 72 h. (b) Proportional tracer exposure represented as AUC of tracer-positive cell counts from 0 to 72 h in dLN. *Indicates significance by multiple t-tests; *indicates p < 0.05; **indicates p < 0.01; ^{^}indicates slope significantly different from 0 with p < 0.05; n = 4–6 mice per group. AUC, area under the curve.

However, in contrast to that seen for both 500 and 30 nm tracers, accumulation of 10 nm tracers in dLN was increased in VC and R2 tumors (Fig. 5a). Of note, it is possible that these observed differences are due not only to transport changes within the tumor environment. Phagocytosis within the tumor microenvironment has the potential to be altered relative to the skin microenvironment, which may also play a role in these findings.

Whether access by leukocytes resident within dLNs was altered by the extent of accumulation in the tissue itself was next assessed by a correlation analysis with flow cytometrically assessed leukocyte-tracer association levels. No statistically significant relationships were found for 500 nm tracers. However, the extent of association with various leukocyte subsets with 30 and 10 nm tracers tended to increase with the increasing extent of total accumulation within the LNs, reaching statistically significant levels for CD11c⁺ and B220⁺ cells accessing 30 nm tracers (Fig. 5b). When considering the relative extents of tracer accumulation in LNs relative to all analyzed systemic tissues in total, reversal of tumor-induced remodeling did not restore the relative selectivity of accumulation of 500 nm tracers in dLNs compared with systemic tissues 72 h after administration, but did reverse the substantial reduction in 10 nm tracer enrichment in LNs relative to systemic tissues 4 h postinjection (Fig. 6).

FIG. 6.

(Lymph)angiogenic normalization normalizes dissemination of tumor-derived molecules to systemic tissues. Ratio of dLN 500, 30, and 10 nm concentration relative to systemic concentration across 72 h after tracer injection into naive skin or day 7 melanoma. *Indicates significance by two-way ANOVA with Tukey's post hoc comparison; *indicates p < 0.05; ***indicates p < 0.005; n = 4–6 animals.

Notably, 30 nm tracers were highly enriched (∼1000-fold) within dLNs compared with systemic tissues irrespective of the tested skin microenvironment (Fig. 6). Reversing remodeling of the blood and lymphatic vasculature in tumors thus appears to improve selective drainage of tumor-derived lymph-draining species to an extent, but not cell-trafficked species.

Normalization of tumor angiogenesis and lymphangiogenesis ameliorates tumor-induced alterations in tracer transport to draining lymph nodes but not uptake by skin-resident antigen presenting cells

A modified flow cytometry panel (Supplementary Fig. S1) to assess the subtypes of various leukocyte subpopulations associating with tracers after administration was next used to evaluate the mechanisms underlying patterns of dissemination and how they are influenced by tumor microenvironmental remodeling. Specifically, dermal dendritic cells (dDCs) and Langerhans cells (LCs) were assessed using CD11c, DEC205, and granularity (by side scatter, Supplementary Fig. S1), given their roles in regulating the skin immune microenvironment and migratory phenotype. When applied to the tracer method of analysis, 500 nm tracer association with these cells within LNs was found to be enhanced after injection in a melanoma compared with naive skin (Fig. 7a).

FIG. 7.

(Lymph)angiogenic remodeling impacts cellular distribution and exposure to actively transported model antigen in draining lymph nodes. (a) Number of 500 nm+migratory cells in dLN at 72 h relative to naive dLN cells. (b) Five hundred nanometer mean fluorescence index in 500 nm+migratory cells in dLN at 72 h, fold change relative to dLN cells. (c) Five hundred nanometer signal in CD45⁺ cells in LNs draining naive skin, and WT and VC melanomas. *Indicates significance by one-way ANOVA compared with naive value with p < 0.05; n = 4–6 animals.

In VC tumors, the number of 500 nm+dDCs was further enhanced, while LCs also exhibited higher association with 500 nm tracers compared with that seen after injection in naive skin, indicating increased migration by dDCs and LCs from the malignant versus healthy skin to the LN (Fig. 7a). Examination of the mean fluorescence index (MFI) of 500 nm signal as a proxy for the number of tracers within migratory cells, however, revealed that while the overall numbers of 500 nm tracer-positive cells were higher for LNs draining the tumor, they exhibited a lower per cell MFI (Fig. 7b) that corresponded with fewer cells carrying more than one tracer (Fig. 7c). This difference was also seen after administration in VC tumors (Fig. 7b, c).

Due to limitations of the technique that rely on sufficient sample sizes to make conclusions and the decreased size of LN draining tumors seen with VEGFR2 inhibition (Fig. 2j), impacts of VEGFR2 inhibition could not be assessed. Nevertheless, tumors appear to increase migration by antigen presenting cells overall, but tracer uptake per cell is diminished. Thus, despite what appears by bulk quantification methods to correspond to normalization of cell migration as a result of VC overexpression in the tumor, the dynamics of cell processing of particles and their migration is separately regulated by the propensity for antigen presenting cells to take up 500 nm tracers versus capacity to migrate.

The association of tracers with the capacity to directly drain into lymph to distinct leukocyte subsets in LNs was next evaluated. These included T cells (CD3⁺B220⁻), B cells (B220⁺CD3⁻), cDCs (conventional dendritic cells, CD11c⁺B220⁻), plasmacytoid dendritic cells (pDC, CD11c⁺B220⁺), subcapsular sinus macrophages (SSM, CD11b^loF4/80⁻CD169⁺), medullary cord macrophages (MCM, CD11b^hiF4/80⁺CD169⁻), medullary sinus macrophages (CD11b^hiF4/80⁺CD169⁻SSC^hi), dDC (CD11c⁺DEC205^loSSC^med), and LC (CD11c⁺DEC205^hiSSC^hi) (Supplementary Fig. S1). Given the high number of dimensions and readouts, in this analysis, the “efficiency” of accumulation that is the ratio of the proportion of measured cell subset associated with the tracer to the total amount of tracer accumulation within the LN by bulk measurement was calculated.⁵⁸

We have previously proposed this metric to account for differences in the extent of distribution to any tissue and the relative access by a resident cell to that species being traced.⁵⁸ In doing so, malignancy was found to enhance the access by LN-resident pDCs, dDCs, and LCs to lymph draining 10 nm tracers compared with that of LNs draining the naive skin (Fig. 8a). These effects were reversed in VC or R2 tumors, however (Fig. 8a). Not only were the number of cells that could associate with lymph-draining tracers increased by the presence of a melanoma and normalized by modulation of vascular modeling, the extents of association on a per cell level appeared to follow similar trends.

FIG. 8.

(Lymph)angiogenic remodeling impacts cellular distribution and exposure to passively draining model antigen in draining lymph nodes. (a) Frequency of cells containing tracer relative to total 10 nm tracer accumulation in the dLN 4, 24, and 72 h after injection in naive skin or WT, VC, and R2 tumors. (b) Efficiency of tracer accumulation via comparison of % of cells containing 10 nm tracer relative to total accumulation within tumor draining lymph nodes. (c) Ten or thirty nanometer MFI of tracer+pDCs from 4 to 72 h after injection. (d) Ten and thirty nanometer MFI over time within CD11c⁺ and F4/80⁺ cells. *Indicates significance by two-way ANOVA with Tukey's post hoc test; *indicates p < 0.05; **indicates p < 0.01; ***indicates p < 0.005; ****indicates p < 0.001; ^#indicates significance by one-way ANOVA with Tukey's post hoc; ^#indicates p < 0.05; ^##indicates p < 0.01; n = 4–6 animals (per time point). MFI, mean fluorescence index; pDC, plasmacytoid dendritic cells.

More specifically, the relationship between the frequency of cells containing 10 nm tracer relative to the efficiency of tracer accumulation was normalized in VC and R2 tumors compared with WT tumors within B cells, cDC, and pDC, which make up the parenchyma of the dLN (Fig. 8b). This metric did not show significance for 30 nm tracer, which may be saturated in this system (Supplementary Fig. S2). Using MFI as an indicator of the number of tracers per cell, LN-resident pDCs were found to exhibit higher levels of association with both 30 and 10 nm tracers after they were administered in a melanoma compared with skin, an effect lost in either VC or R2 tumor model (Fig. 8c).

Likewise, MFIs of both 10 and 30 nm tracers were normalized within total F4/80⁺ and CD11c⁺ cells within LNs draining VC and R2 tumors relative to WT tumors (Fig. 8d). Altering vascular remodeling in tumors restores patterns of tracer access to LN-resident leukocytes by lymph-draining species.

Discussion

The cross talk of biomolecules and extracellular vesicles that are tumor derived with the immune system has significant ramifications on the development of anticancer immune responses.^45,58,63 While the implications of this on biological signaling has garnered considerable interest, the impacts of remodeling of the blood and lymphatic vasculature that occurs with disease on biotransport have to date remained underexplored. In this study, we used the most commonly used preclinical model of skin cancer, in which multiple immune therapeutic regimens have been developed, to explore the impacts of altered lymphangiogenic and angiogenic signaling in tumors alter the biodistribution of macromolecular tracers from the tumor to systemic tissues, as well as lymphoid tissues and leukocytes. Specifically, patterns of access to systemic tissues, and leukocytes within those tissues, differ by both the mechanism of transport and the normalization technique implemented, with trends pointing toward slight but incomplete normalization of leukocyte access by tumor vascular normalization.

In addition, we demonstrate that using a tumor model of increased lymphangiogenesis that increased transit of migratory antigen presenting cells from skin cannot counter tumor-induced suppression of their phagocytic functions, resulting in similar total levels of accumulation but altered mechanisms of transport. However, efficiency of immune cell access within LNs, frequency of access to immune cells, and MFI of tracer within LNs, are normalized for LN parenchyma-resident immune cells after implementation of angiogenic and lymphangiogenic normalization techniques. This suggests that cell-mediated transport to the dLN is affected by (lymph)angiogenic normalization in manners independent of tracer accumulation, while cellular access to passively draining species is primarily dependent on structures within the LN. As a whole, despite modifications in local retention at the tumor and systemic accumulation, (lymph)angiogenic normalization induced minor degrees of normalization of accumulation within dLNs, dependent on mechanism of lymphatic transport.

It is, however, possible that these minor impacts could have significant immunological ramifications given the role of the lymphatic system in tumor immunity and tolerance.⁴⁵

Conclusion

Through this work, we were able to utilize models of (lymph)angiogenic normalization in melanoma to improve upon current understandings of the links between blood and lymphatic vasculature and immune remodeling to show how alterations in tumor-mediated transport impact access of tumor-derived species to systemic tissues, as well as locoregional tissues with immunological ramifications. This manipulation of the VEGF-A and VEGF-C pathways revealed subtle changes to access of tumor-derived molecules to LNs and systemic tissues as well as retention in the tumor microenvironment resulting from active, cell-mediated, and passive lymphatic drainage-based mechanisms of lymphatic transport, the immunological ramifications of which remain to be determined. This has the potential to inform therapeutic strategies, in particular immunotherapies, which are one of the most promising cancer therapeutics that have not fully reached their potential clinically to date.

Authors' Contributions

M.J.O. and N.A.R. performed the experiments. M.J.O. and N.A.R. analyzed the data. M.J.O., N.A.R., and S.N.T. designed the experiments. M.J.O. and S.N.T. wrote the article.

Disclosure Statement

The authors declare that they have no competing interests.

Funding Information

This work was supported by the U.S. National Institutes of Health grants R01CA207619 (S.N.T.), U01CA214354 (S.N.T.), S10OD016264, and T32GM00843 (M.J.O. and N.A.R.).

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2