Critical role of CD11b⁺ macrophages and VEGF in inflammatory lymphangiogenesis, antigen clearance, and inflammation resolution

Peripheral tissues, such as skin and mucosa, are exposed to external environments and are constantly challenged by antigens. Some antigens, such as pathogen molecules derived from bacteria and viruses, evoke inflammatory responses in these tissues through a consecutive set of processes: activation of residential macrophages and mast cells leading to secretion bioactive amines, cytokines, chemokines, and lipid mediators; recruitment of leukocytes including neutrophils, monocytes, and T lymphocytes from circulation; activation of macrophages and dendritic cells (DCs) to engulf dead leukocytes, pathogens, and other inflammatory debris; and drainage of these cells into draining lymph nodes (DLNs) via local lymphatic vessels.^1,–3  Throughout these processes, residential and recruited monocyte-derived macrophages are critical for evoking and resolving the acute inflammation caused by pathogen entry.^1,,,–5  The pathophysiologic, molecular, and cellular mechanisms governing initiation and promotion of inflammatory processes have been extensively characterized; however, it is yet to be clarified how acute inflammation resolves, rather than leads to, chronic inflammation. Particularly, the roles of local lymphatic vessels and DLNs in antigen clearance and inflammation resolution are poorly understood.

During inflammation, macrophages actively participate in inducing the formation of new lymphatic vessels.^6,,,,,,–13  In adult animals, new lymphatic vessels are formed mainly through lymphangiogenesis, which is a sequence of processes that include sprouting, migration, proliferation, and tubule formation by preexisting lymphatic endothelial cells (LECs).^14,–16  Additional LECs may also be derived through transdifferentiation or incorporation of circulating bone marrow–derived cells and some macrophages.^9,11  The discovery of specific lymphangiogenic growth factors (VEGF-C and VEGF-D), receptors (VEGF receptor-3 [VEGFR3]), a transcription factor (Prox-1), and markers (lymphatic vessel endothelial hyaluronan receptor-1 [LYVE-1], podoplanin) for LECs has led to a renewed interest in the roles of lymphatic vessels and lymphangiogenesis in tissue fluid homeostasis, tumor metastasis, wound healing, antigen presentation, and digested lipid absorption.^14,,,,–19  Among the roles of lymphatic vessels and lymphangiogenesis, the functional importance of lymphatic vessels in inflammation was initially observed a long time ago.²⁰  Lymph nodes, in concert with the lymphatic vessels, play a pivotal role in initiating adaptive immune responses.²¹  Notably, local tissue inflammation causes DLNs to undergo a transient but profound remodeling, with volume expansion, lymphoid hyperplasia, and markedly increased lymphatic vessel densities.^22,23  Lymphatic vessels in DLNs act as essential conduits for antigen-presenting cells such as epidermal Langerhans cells and dermal DCs to T lymphocytes.²¹  Up-regulated induction of inflammatory adhesion molecules in the lymphatic vessels within inflamed DLNs enhances DC mobilization to the T lymphocytes.^24,–26  The main mediators and factors governing inflammation-induced lymphangiogenesis (IIL) are yet to be defined. Intriguingly, our pilot experiment revealed that in bacterial pathogen–induced inflamed skin and DLNs, CD11b⁺ macrophages were profoundly infiltrated around expanded lymphatic vessels. Based on this observation, we have speculated that macrophages, including CD11b⁺ macrophages, play a pivotal role in the expansion of lymphatic vessels in the inflamed skin and DLNs by secreting lymphangiogenic growth factors. Moreover, we have sought to determine whether macrophage-induced lymphangiogenesis may be actively involved in antigen clearance, in addition to initiating the adaptive immune response.

In this study, we investigated the role of IIL in antigen clearance and inflammation resolution. To this end, we generated an inflammation model by the introduction of representative etiologic agents capable of inducing septic shock. These agents are lipopolysaccharide (LPS) from Gram-negative bacteria and a mixture of lipoteichoic acid (LTA) from Gram-positive bacteria plus muramyl dipeptide (MDP),^27,28  which were injected into the dorsal side of ear dermis of mice. At the peak day of acute inflammation, we analyzed changes in lymphatic vessels, macrophage infiltration, lymph flow, inflammatory cell migration, antigen clearance, and inflammation resolution from the inflamed skin and DLNs in normal adult mice and in skin-specific VEGF-C–overexpressing transgenic mice.²⁹  To determine the role of macrophages and lymphangiogenic growth factors in these processes, specific agents that block macrophage and lymphangiogenic growth factors were applied. Our results indicate that macrophages, including CD11b⁺/Gr-1⁺ macrophages, play a critical role in antigen clearance and inflammation resolution, and VEGF-C, -D, and -A derived from the macrophages, local inflamed tissue, and DLNs contribute to accelerate antigen clearance and inflammation resolution.

C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME). GFP⁺ mice (C57BL/6J genetic background) were a gift from Dr Masaru Okabe (Osaka University, Osaka, Japan).³⁰  K14-VEGF-C transgenic mice (FVB/N genetic background) were generated and maintained as previously described,²⁹  and transferred to KAIST. Mice were bred in our pathogen-free animal facility, and male mice aged 7 to 8 weeks were used. Animal care and experimental procedures were performed under approval from the Animal Care Committee of KAIST.

LPS (from Escherichia coli 0111:B4) and MDP were purchased from Sigma-Aldrich (St Louis, MO). Highly purified and structurally intact LTA was prepared from Staphylococcus aureus (ATCC, Manassas, VA) as previous described.²⁸  Mice were anesthetized by intramuscular injection of a combination of anesthetics (80 mg/kg ketamine and 12 mg/kg xylazine), then a single intradermal injection of LPS (10 μg in 10 μL PBS) or LTA + MDP (10:1 ratio, 10 μg LTA + 1 μg MDP in 10 μL PBS)²⁷  was given into the dorsal side of the ear using a 31-gauge syringe.

For systemic depletion of macrophages, mice were given an intravenous injection of clodronate liposome (CDL, 25 mg/kg)³¹  through the tail vein at 1 day before and after the intradermal injection of LPS or LTA. To block VEGF-C/D, mice were treated with a single intravenous injection of 10⁹ pfu Ad-sVEGFR3³²  at 12 hours before the intradermal injection of LPS or LTA. To block VEGF-A, mice were given a subcutaneous injection of VEGF-Trap (25 mg/kg)^33,34  or anti–VEGF-A blocking antibody (4 mg/kg; R&D Systems, Minneapolis, MN) under the back skin at 1 day before and after the intradermal injection of LPS or LTA.

Mice were anesthetized by intramuscular injection of anesthetics, as described in “Ear skin inflammation model,” at the indicated times after the intradermal injection of LPS or LTA. The histologic and morphologic analyses in the inflamed ear and DLNs were performed as described in Document S1 (available on the Blood website; see the Supplemental Materials link at the top of the online article).

After anesthesia, the ears and DLNs were harvested and dissected into small pieces by a microscissor, and the pieces were incubated with 2 mL Hank balanced salt solution (Sigma-Aldrich) containing 0.2% collagenase type-II (Worthington, Lakewood, NJ) for 1 hour at 37°C. Flow cytometric analysis of CD11b⁺ macrophages from ear skin and DLNs was performed as described in Document S1.

CD11b⁺ macrophages in the DLN were enriched using anti–mouse CD11b antibody-coupled MicroBeads (Miltenyi Biotec, Auburn, CA) and a magnetic cell sorter (magnetic-activated cell sorting [MACS]; Miltenyi Biotec) according to the manufacturer's instructions as described in Document S1.

Ear thickness was measured by a Flat Anvil type caliper (cat no. 7301; Mitutoyo, Kawasaki, Japan). Swelling and erythema were graded by severities as described in Tables S1,S2.

The semiquantitative and quantitative real-time reverse-transcription–polymerase chain reaction (RT-PCR) of total RNA isolated from ear skin, whole DLNs, and MACS-enriched CD11b⁺ cells from the DLNs was performed with appropriate primers (Table S3) as described in Document S1.

Monitoring lymph flow was performed using 2 intravital imaging methods as described in Document S1.

GFP⁺ mice were intraperitoneally injected with 1 mL 3.0% thioglycolate (Sigma-Aldrich) in saline, and inflammatory cells were collected 3 days later by peritoneal washing with ice-cold DMEM culture medium. Approximately, 10⁶ GFP⁺ inflammatory cells in 10 μL PBS were adaptively transferred by injection at the inflamed site of the ear. Twelve hours later, DLNs were collected and sectioned for immunohistology, and a cell suspension was made by collagenase digestion for the flow cytometric analysis.

The monitoring antigen clearance was performed as described in Document S1.

Values presented are means plus or minus standard deviation (SD). Significant differences between means were determined by Student t test or analysis of variance (ANOVA) followed by the Student-Newman-Keuls test. Statistical significance was set at a P value less than .05.

Acute skin inflammation induced by intradermal injection of LPS (10 μg in 10 μL PBS; hereafter referred to as LPS) or LTA (10:1 ratio, 10 μg LTA plus 1 μg MDP in 10 μL PBS; hereafter referred to as LTA) evokes robust lymphangiogenesis and marked recruitment of CD11b⁺/Gr-1⁺ macrophages in the inflamed skin and DLNs (Figures S1Figure S2. Intradermal LPS or LTA increase lymphatic and blood vessel densities, lymphatic sprouting, thickness, and recruitment of CD11b+/Gr-1+ macrophages (JPG, 112 KB)Figure S3. Intradermal LPS or LTA increase lymphatic vessel densities and sprouts, and recruitment of CD11b+/Gr-1+/F4/80+ macrophages in the DLN (JPG, 133 KB)–S4). These results led us to explore the involvement of macrophages, including CD11b⁺/Gr-1⁺ macrophages, in lymphangiogenesis in the inflamed skin and DLNs by depletion of whole macrophages by intraperitoneal treatment with clodronate liposome (CDL).³¹  The CDL treatment, but not control liposome (CL) treatment, efficiently depletes macrophages, including CD11b⁺/Gr-1⁺ macrophages, in all organs, including inflamed skin and DLNs (Figures 1A and 2A, and data not shown) by inducing apoptosis of macrophages.³¹  In this situation, the LPS- or LTA-induced increased lymphatic vessel densities and sprouting and infiltration of CD11b⁺/Gr-1⁺ macrophages were largely attenuated, not only in the inflamed skin but also in the DLNs (Figures 1,2). However, unexpectedly, the skin thickness was increased to a greater extent in LPS + CDL and LTA + CDL than in LPS + CL and LTA + CL (Figure 1A,B). In comparison, the inflamed skin and DLNs treated with LPS + CL or LTA + CL showed no significant differences in any parameters examined compared with the skin and DLNs treated with only LPS or LTA. In addition, the inflamed skin and DLNs treated with the CL or CDL showed no significant differences compared with the skin and DLNs treated with PBS (Figures 1,2). To determine the role of macrophages in inflammation resolution, we further examined LPS- or LTA-induced inflammatory reactions in the ear skin after treatment with CL or CDL. Compared with the inflamed skin treated with LPS + CL and LTA + CL, the inflamed skin treated with LPS + CDL and LTA + CDL exhibited exaggerated inflammatory reactions including severe swelling, erythema formation, and corresponding histologic changes (Figure S5). Local depletion of macrophages by a localized injection of CDL, but not CL, into the inflamed skin produced similar findings (data not shown). These data suggest that LPS- or LTA-induced recruitment of macrophages, including CD11b⁺/Gr-1⁺ macrophages, plays a critical role not only in lymphangiogenesis in the inflamed skin and DLNs, but also in inflammation resolution in the inflamed skin.

To elucidate the underlying mechanism of how macrophages play a role in inflammation resolution, we first monitored lymph flow from the inflamed ear skin to DLNs. As a monitoring system, we obtained simultaneous vital images for 30 minutes using an in vivo imaging system (IVIS) imaging system and a fluorescence stereomicroscope to track the movement of a lymphatic tracking fluorescent dye, fluorescein isothiocyanate (FITC)–dextran, at the injection site on day 3 after the LPS or LTA injection. Fluorescence intensities in the DLNs were similar in PBS, PBS + CL, and PBS + CDL, whereas they were slightly but significantly higher in LPS + CL and LTA + CL (Figure 3A-C), indicating greater lymphatic flow during LPS- and LTA-induced inflammation. However, fluorescence intensities in the DLNs were lower in LPS + CDL and LTA + CDL than in LPS + CL, LTA + CL, and PBS (Figure 3A-C), indicating that lymphatic flows were decreased by depletion of macrophages, even during LPS- and LTA-induced inflammation. Second, we determined the mobilization rate of inflammatory cells from the inflamed ear skin to DLNs by the intradermal injection of GFP⁺ inflammatory cells (∼ 10⁶ cells). Twelve hours after the injection of green fluorescent protein–positive (GFP⁺) inflammatory cells, the mobilized inflammatory cells were localized mainly in the pericapsular region of DLNs and rarely detectable in the paracortical region of DLNs in PBS (Figure 3D). Compared with PBS, the number of the mobilized inflammatory cells substantially increased both in the pericapsular region (∼ 3.3-fold and ∼ 2.8-fold by LPS + CL and LTA + CL, respectively) and paracortical region (∼ 6.1-fold and ∼ 5.5-fold by LPS + CL and LTA + CL, respectively; Figure 3D,E). However, compared with LPS + CL and LTA + CL, the numbers of the mobilized inflammatory cells were markedly lower in both the pericapsular region (∼ 63% and ∼ 62% by LPS + CDL and LTA + CDL, respectively) and the paracortical regions (∼ 90% and ∼ 85% by LPS + CDL and LTA + CDL, respectively; Figure 3D,E). Flow cytometric analyses on the total DLNs exhibited similar findings (Figure 3F). These data suggest that macrophages play a substantial role in increases of lymph flow and inflammatory cell mobilization from the inflammatory site to the DLNs for antigen clearance and inflammatory resolution.

To investigate the source and expression of lymphangiogenic factors that drive LPS- or LTA-induced lymphangiogenesis, we performed a series of reverse-transcription–polymerase chain reaction (RT-PCR) assays in the skins and their DLNs with and without macrophage depletion by CDL treatment, and in the enriched CD11b⁺ macrophages from DLNs. Compared with PBS, LPS + CL and LTA + CL increased VEGF-A₁₆₄ (2.3- and 2.2-fold, respectively), VEGF-A₁₂₀ (2.6- and 1.9-fold, respectively), VEGF-C (2.1- and 1.9-fold, respectively), and VEGF-D (1.5- and 1.4-fold, respectively) in the ear skins (Figure S6A,C). Compared with PBS, LPS + CL and LTA + CL increased VEGF-A₁₆₄ (2.4- and 2.3-fold, respectively), VEGF-A₁₂₀ (3.1- and 3.2-fold, respectively), VEGF-C (2.6- and 2.5-fold), respectively, and VEGF-D (1.3- and 1.3-fold, respectively) in the DLNs (Figure S6A,C). CDL treatment markedly reduced LPS + CL- and LTA + CL-induced increases in expression of VEGF ligands in the ear skin and DLNs, and the relative levels were similar to those of PBS. In CD11b⁺ macrophages enriched by magnetic-activated cell sorting (MACS), LPS and LTA increased VEGF-A₁₆₄ (4.7- and 4.2-fold, respectively), VEGF-A₁₂₀ (3.2- and 3.0-fold, respectively), VEGF-C (4.8- and 3.4-fold, respectively), and VEGF-D (3.6- and 3.1-fold, respectively), compared with PBS (Figure S6B,D). Consistently, quantitative real-time RT-PCR analyses revealed that LPS and LTA increased VEGF-A, VEGF-C, and VEGF-D mRNA levels, whereas they did not change the placental growth factor (PlGF) mRNA level in the skin, DLNs, and enriched CD11⁺ macrophages (Figure S7). Thus, macrophages, including CD11b⁺ macrophages, could be the main mediators or sources of lymphangiogenic VEGF ligands in the skin and DLNs during acute IIL.

The pacarine role of macrophages as a source of VEGFs led us to explore the role of VEGF ligands on LPS- or LTA-driven lymphangiogenesis, lymph flow, inflammatory cell migration, and inflammation resolution. To block VEGF-C and VEGF-D (VEGF-C/D), an adenovirus encoding the soluble extracellular domain of VEGFR-3 (Ad-sVEGFR3) was used.³²  Compared with the control, which was treated with 10⁹ pfu adenovirus encoding β-galactosidase (Ad-βgal), treatment with 10⁹ pfu Ad-sVEGFR3 markedly attenuated the LPA- and LTA-increased lymphatic density (∼ 47% and ∼ 48%, respectively, in the skin and ∼ 50% and ∼ 52%, respectively, in the DLNs), lymph flow (∼ 44% and ∼ 30%, respectively), and inflammatory cell migration to DLNs (∼ 59% and ∼ 58%, respectively, in the pericapsular region and ∼ 78% and ∼ 75%, respectively, in the paracortical region), and delayed inflammation resolution (Figures 4,S8). The mice treated with the Ad-βgal showed no significant differences from the mice treated with PBS. Compared with the controls, which were treated with dimeric-Fc protein, mice receiving a blockade of VEGF-A with either VEGF-Trap^33,34  or anti–VEGF-A antibody also had significant attenuation of the LPA- and LTA-increased lymphatic density in the skin and DLNs, increased lymph flow and inflammatory cell migration from the inflamed skin to DLNs (pericapsular and paracortical regions), and delayed inflammation resolution (Figures 5,S9,S10). The mice treated with the dimeric-Fc protein showed no significant differences from the mice treated with PBS. In addition, flow cytometric analysis of the DLNs revealed that infiltration of CD11b⁺/Gr-1⁺ macrophages to the DLNs was markedly decreased by blockade of VEGF-C/D or VEGF-A (Figure S11). These results indicate that VEGF-C, -D, and -A derived from inflamed skin and DLNs make a major contribution to the induction of dermal and DLN lymphangiogenesis by LPS and LTA, as well as to the enhancement of lymph flow and inflammatory cell migration, and to the resolution of inflammation.

To ascertain whether VEGF-C, -D, and -A are involved in antigen clearance through local lymphatic vessels and DLNs, a single intradermal injection of FITC-LPS (10 μg/ear in 10 μL PBS) into the ear skin was performed in the mice treated with the AdsVEGFR3, Ad-βgal, VEGF-Trap, and dimeric-Fc, and the clearance of FITC-LPS was monitored. Compared with treatment with Ad-βgal, treatment with Ad-sVEGFR3 attenuated the disappearance of FITC-LPS from the ear (∼ 19.8% at day 1 and ∼ 11.1% at day 3; Figure 6A,C). Compared with treatment with dimeric-Fc protein, treatment with VEGF-Trap attenuated FITC-LPS disappearance from the ear (∼ 18.9% at day 1 and ∼ 7.5% at day 3), and treatment with anti–VEGF-A antibody also attenuated FITC-LPS disappearance from the ear (∼ 25.2% at day 1 and ∼ 4.4% at day 3; Figures 6B,D, S12). Histologic analyses revealed that more FITC-LPS remained in the injected ear skin of mice treated with Ad-sVEGFR3, VEGF-Trap, or anti–VEGF-A antibody compared with the mice treated with Ad-βgal or dimeric-Fc (Figures 6E, S12), confirming that blocking VEGF-C/D or VEGF-A delays antigen clearance from the inflamed skin.

Compared with wild-type littermates, K14-VEGF-C transgenic mice displayed profoundly denser and larger lymphatic vessel networks in the ear skin and DLNs (Figure 7A). Moreover, vital fluorescence lymphangiography by injection of FITC-dextran revealed an immediate and diffuse distribution of the FITC-dextran from the injection site in the transgenic mice, compared with limited and selective distribution of the FITC-dextran from the injection site in normal littermates (Figure 7B). In the LPS-induced inflamed skin, no differences were seen in the rate of lymphatic drainage to the DLNs between the normal littermates and the transgenic mice (Figure 7C,D). However, inflammatory cell migration into the DLNs was higher (∼ 7.8-fold) in the K14-VEGF-C transgenic mice than in the normal littermates (Figure 7E,F). Furthermore, the inflammatory responses to LPS, including swelling and erythema, were significantly less in the transgenic mice than in the normal littermates (Figure 7G,H), suggesting that K14-VEGF-C transgenic mice have faster antigen clearance in the skin. In fact, the disappearance of FITC-LPS from the ear was significantly faster over time in the transgenic mice than in the normal littermates (Figure 7I,J), further indicating faster antigen clearance in the skin. To determine the role of macrophages in lymphatic function in the K14-VEGF-C mice, lymph flow and inflammatory cell migration were measured after treatment with CL or CDL. Depletion of macrophages by CDL did not affect lymph flow or inflammatory cell migration from the skin to DLNs in the K14-VEGF-C mice (Figure S13A-D). Furthermore, depletion of macrophages by CDL did not significantly exaggerate or alleviate the LPS-induced inflammatory reactions, including severe swelling and erythema formation (Figure S13E,F), indicating that the profoundly denser and larger lymphatic vessel networks in the ear skin and DLNs are sufficient to resolve inflammation and can compensate for the role of macrophage-induced lymphangiogenesis in the inflamed skin.

As part of the innate immune response during skin inflammation, locally generated interstitial fluid containing activated immune cells is drained unidirectionally into the DLNs via afferent lymphatic vessels. The cellular and molecular mechanisms of DC migration from inflamed skin to DLNs with respect to antigen presentation have been well characterized.^21,,,,–26  However, little is known about the underlying cellular and molecular mechanisms for clearance of pathogen molecules and the resolution of inflammation. In this study, we found that VEGF ligands secreted from infiltrated macrophages, including CD11b⁺/Gr-1⁺ macrophages, could be major mediators for pathogen clearance and inflammation resolution through profound expansion of lymphatic vessels, enhanced lymph flow, and mobilization of inflammatory cells from the inflamed skin to the DLNs (Figure S14).

Macrophages play diverse roles in IIL.^6,,,,,,–13  For example, CD11b⁺ macrophages are actively involved in IIL in the cornea,¹⁰  and VEGF-A stimulates IIL mainly via macrophage recruitment.⁶  Accordingly, depletion of macrophages abrogates this IIL.¹⁰  Moreover, normal wound healing requires macrophage-induced lymphangiogenesis.^12,13  In a diabetic mouse model, administration of macrophages promotes wound healing with enhanced lymphangiogenesis,¹³  demonstrating the critical role of macrophages in IIL in inflamed tissues. Our recent study in an ovarian cancer model indicated that VEGF ligands derived from CD11b⁺ macrophages induce profound but dysfunctional lymphangiogenesis in inflammatory carcinogenesis.³⁴  The present results show that acute inflammation induced by intradermal administration of virulence factors, LPS and LTA, results in the abundant recruitment of CD11b⁺/Gr-1⁺ macrophages, and these macrophages appear to induce a temporary but profound lymphangiogenesis in the inflamed skin as well as in the DLNs. Furthermore, the subsequent experiments demonstrated that the CD11b⁺/Gr-1⁺ macrophages were mainly F4/80⁺ cells. These cells could have substantial phagocytic activity because they were depleted by CDL, and F4/80⁺ macrophages are known to have a relatively high phagocytic activity.^31,35  Because of the phagocytic activity of these cells, they could actively participate in clearing dead leukocytes, pathogens, and cellular debris by phagocytosis. Then these cells are transmigrated into the lumen of local lymphatic vessels and finally drained into the DLNs.³⁶  These phagocytic macrophages are also known to be the main cell type for secreting major cytokines, chemokines,^1,,,–5  and lymphangiogenic growth factor.^6,,,,,,–13  Indeed, upon phagocytosis of the apoptotic bodies, macrophages secrete abundant VEGF-A, which promotes endothelial cell growth.³⁷  Our RT-PCR analyses revealed that activated macrophages, including CD11b⁺ macrophages, are the main producers of lymphangiogenic VEGF ligands in the inflamed skin and DLNs during acute IIL. Furthermore, depletion of macrophages largely attenuated IIL, indicating that macrophages could be the major source of lymphangiogenic growth factors. Thus macrophages, including CD11b⁺/Gr-1⁺/F4/80⁺ macrophages, help themselves by laying the groundwork for their own transportation to the DLN by secreting lymphangiogenic growth factors (Figure S14).

Our results also show that profound increases of lymphatic vessel density in the inflamed skin and DLNs result from active proliferation and sprouting of LECs. The proliferation and sprouting of these cells are essential processes of lymphangiogenesis.^14,–16  The increased lymphatic vessel densities obviously increase the surface area of the initial, collecting, and conduit lymphatic vessels, and these vessels would facilitate entry and transport of interstitial fluid and immune cells from the inflamed skin to the conducting lymphatic vessels and the DLNs. Johnson et al showed that the expression of the key leukocyte adhesion molecules, intercellular adhesion molecule, and vascular cell adhesion molecule, are up-regulated in LECs in inflamed skin, and these adhesion molecules facilitate DC transmigration.³⁸  Furthermore, Baluk et al elegantly described the preferential entry of MHC-II⁺ DCs and macrophages and CD45⁺ leukocytes into the proximal half of initial lymphatic vessels having “buttonlike junctions” in tracheal mucosa during LPS-induced acute inflammation.³⁹  Likewise, activated macrophages, including CD11b⁺/Gr-1⁺/F4/80⁺ macrophages, could actively transmigrate to inflamed lymphatic vessels by virtue of these up-regulated adhesion molecules, preferentially through the proximal half of the initial lymphatic vessels. The robust expansion of CD11b⁺/Gr-1⁺ macrophages observed in the DLNs during acute skin inflammation implies recruitment of these cells from the outside to the DLNs. Previous studies^21,40,–42  indicate that recruitment of circulating monocyte-driven macrophages to the DLNs during skin inflammation is mediated through high endothelial venules (HEVs) by “remote control” of chemokines including “macrophage chemoattractant protein-1” and “monokine induced by gamma interferon” secreted from the inflamed skin tissues. However, our compelling evidence favors a model in which draining lymphatic vessels, rather than HEVs, are responsible for the monocyte-driven macrophage recruitment to the DLNs during acute inflammation. The evidence in favor of this model is (1) our immunohistologic analyses revealing that robust increases of CD11b⁺ macrophages are localized mainly around subcapsular afferent lymphatic vessels, which are directly connected to peripheral tissues, and (2) the adaptively transferred inflammatory cells are also localized mainly around subcapsular afferent lymphatic vessels, with only a few localized in the HEV-rich regions of the DLNs during the inflammation. Thus, the draining lymphatic vessels that lead from the inflamed tissue to the DLNs are the main routes for monocyte-driven macrophage recruitment to the DLNs during acute inflammation.

Recently, lymphangiogenesis in the DLNs has emerged as a significant factor in several pathophysiologic conditions. During inflammation, B cell–driven DLN lymphangiogenesis enhances DC mobilization,²⁵  whereas tumor-induced DLN lymphangiogenesis enhances lymph flow and metastasis.^43,44  These intriguing reports led us to examine whether inflammation-induced DLN lymphangiogenesis promotes inflammatory cell mobilization and lymph flow, and subsequently enhances antigen clearance and inflammation resolution. Indeed, our findings exhibited that DLN lymphangiogenesis is closely linked to accelerated mobilization of inflammatory cells and to a lesser extent, enhanced lymph flow, from inflamed tissue to DLNs. IIL at the inflamed tissue could be an additional factor in accelerating inflammatory cell mobilization and lymph flow. Interestingly, Halin et al⁴⁵  found that DLN lymphangiogenesis during chronic inflammation is due to VEGF-A secreted from the inflamed tissue, but not from the DLNs. However, our results show a clear correlation between abundant macrophage infiltration, up-regulated expression of VEGF-C, -D, and -A, and increased lymphatic densities within the DLNs. Moreover, the enriched CD11b⁺ macrophages from the DLNs displayed markedly up-regulated expression of VEGF-C, -D, and -A. Furthermore, depletion of macrophages including CD11b⁺/Gr-1⁺, and blockade of VEGF-C/D or VEGF-A, largely attenuated DLN lymphangiogenesis. These data indicate that the lymphangiogenic growth factors secreted from robustly infiltrated macrophages in the DLNs could be the major driving force for DLN lymphangiogenesis. Therefore, we propose that ILL in DLNs could be caused not only by VEGF ligands from inflamed tissue, but also by VEGF ligands secreted from CD11b⁺/Gr-1 macrophages within DLNs in acute inflammation.

The most intriguing and novel finding in this study is that inflammation-induced up-regulation of VEGF-C, -D, and -A from the macrophages and local tissues plays an active role in antigen clearance and inflammation resolution by promoting IIL and lymphatic function. Accordingly, depletion of macrophages or blocking of VEGF-C/D or -A strikingly reduced inflammatory cell migration, lymph flow, antigen clearance, and inflammation resolution with attenuated DLN lymphangiogenesis. Thus, these components largely contribute to regulate inflammatory cell migration and lymph flow directly and indirectly with the expansion of lymphatic surface area through local and DLN lymphangiogenesis. (Figure S14). Accordingly, the K14-VEGF-C transgenic mice with abundant lymphatic vessels in skin and DLNs displayed higher inflammatory cell migration, and faster antigen clearance and inflammation resolution than the wild-type mice. In fact, VEGF-C is known not only to augment lymphatic growth, but also to maintain or enhance lymphatic functions including lymph drainage and flow from the interstitial compartment to the lymphatic lumen and DLNs,⁴⁶  and VEGF-A is one of the most potent factors that promotes permeability from the blood vascular lumen to the interstitial compartment.⁴⁷  Thus, because VEGF-C, -D, and -A may control and contribute to homeostasis between lymph drainage and flow and blood vascular permeability, disturbances in this homeostasis could be involved in delayed antigen clearance and inflammation resolution in the inflamed skin tissue. However, it should be noted that VEGF-Trap treatment attenuates the ability of transgenic VEGF-A to induce chronic inflammation in skin, corneal inflammation, and the release of inflammatory cytokines induced by polymicrobial sepsis.^6,48,49  Therefore, the blockade of VEGF-A with VEGF-Trap can reduce or promote inflammation, depending on the type and location of the inflammation.

In conclusion, our findings shed light on an additional role of macrophages during acute inflammation. The lymphangiogenic growth factors secreted from infiltrated macrophages in inflamed tissue and DLNs appear to be critical in lymphatic vessel expansion, antigen clearance, and inflammation resolution through enhancement of lymphangiogenesis, lymph flow, and mobilization of inflammatory cells from inflamed tissue to the DLNs. A subset of CD11b⁺/Gr-1⁺/F4/80⁺ macrophages and the lymphangiogenic growth factors VEGF-C, -D, and -A could be the key components in promoting antigen clearance and inflammation resolution during bacterial pathogen–induced acute inflammation. Disruptions in these orchestrated cellular and molecular actions may disrupt early inflammation resolution, leading to uncontrolled chronic inflammation.

This work was supported by the Korea Science and Engineering Foundation (KOSEF) through the National Research Laboratory Program (2004-02376, G.Y.K.) funded by the Ministry of Science and Technology and BioGreen 21 Program, Rural Development Administration (20080401034061, S.H.H.), Republic of Korea.

Contribution: R.P.K., S.H.H., K.A., and G.Y.K. designed and organized the experiments; R.P.K., C.J., and G.Y.K. performed the animal studies, analyzed the data, generated the figures, and wrote the paper; K.J. and H.Y. performed histologic analyses; and R.A.S. and J.E.B. prepared materials for the experiments.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Gou Young Koh, Department of Biological Sciences, KAIST, 373-1, Guseong-dong, Daejeon, 305-701, Republic of Korea; e-mail:gykoh@kaist.ac.kr.