Macrophage Wnt-Calcineurin-Flt1 signaling regulates mouse wound angiogenesis and repair

Recently a great deal of research has been devoted to understanding the underlying mechanisms behind wound repair. Many studies have carefully documented the cells involved in wound repair, their function, and the mediators by which they perform such functions.^1-3 

One process critical to wound repair is angiogenesis, the development of new blood vessels.⁴  One of the most potent proangiogenic molecules is vascular endothelial growth factor A (VEGF-A). Produced by a variety of cells including macrophages,⁵  VEGF-A can bind to Flk1 (VEGF-R2) on endothelial cells to induce migration and proliferation. Mice deficient in VEGF-A^6,7  or Flk1⁸  die early in development from a lack of blood vessel formation. Importantly, another VEGF-A receptor Flt1 (VEGFR1) binds VEGF-A, but has severely deficient signaling capacity. Therefore, Flt1, which can be spliced into a soluble Flt1 or membrane-tethered Flt1, can act as a suppressor of angiogenesis.^9,10  Consistent with this, mice deficient in Flt1 die in utero of excessive angioblast proliferation.^11,12 

In uninjured skin, VEGF-A is expressed at basal levels, but in response to injury, VEGF-A levels rise.¹³  When VEGF-A is neutralized during wound repair, granulation tissue is diminished and wound repair is perturbed.¹⁴  Many of the known proangiogenic factors are products of cells that infiltrate during the inflammatory phase.⁴  Macrophages, for instance, are sources of both proangiogenic and antiangiogenic factors. Such factors can affect blood vessel development by modifying the extracellular matrix,¹⁵  affecting the growth factor milieu,^16,17  and guiding vessel anastomosis.¹⁸  During wound repair, mice deficient in macrophages have abnormally overgrown vasculature that surprisingly results in enhanced repair rate.¹⁹ 

One of the mechanisms by which macrophages regulate vessel branching in development involves a noncanonical Wnt-Flt1 signaling pathway.¹⁷  Wnt signaling is one of the core pathways regulating developmental patterning.²⁰  Wnt ligands are transported to the cell surface for secretion by Wntless/GPR177 (Wls).^21,22  In canonical Wnt signaling, Wnt ligands (eg, Wnt3a) bind to a Frizzled-Lrp5 receptor complex that results in the stabilization and nuclear translocation of β-catenin. Noncanonical Wnt signaling (eg, Wnt5a) involves the many downstream Wnt signaling responses independent of β-catenin.²⁰  One important noncanonical response is the enhancement of intracellular calcium, activation of calcineurin, and dephosporylation of nuclear factor of activated T cells (NFAT).²³ 

Because macrophages are known to regulate developmental angiogenesis via noncanonical Wnt signaling,¹⁷  and because wound angiogenesis has been linked to macrophage presence,¹⁹  it is quite likely that macrophages are recapitulating their developmental role during wound repair. Here we show that wound macrophages use the Wnt-Flt1 pathway to regulate wound responses, and that calcineurin is the critical mediator.

RAW264.7 and EOC-2 cells (ATCC) were exposed to recombinant Wnt5a, RNA was isolated (RNeasy; Qiagen, Hilden, Germany), and either quantitative reverse-transcription polymerase chain reaction or enzyme-linked immunosorbent assay (Quantikine; R&D Systems, Minneapolis, MN) was performed. Aortic ring assay (ARA) was conducted as described.²⁴ 

Wounds were made with 2-mm dermal biopsy punches. For immunofluorescence quantification, wounds were sectioned, labeled with platelet endothelial cell adhesion molecule (PECAM-1; BD Biosciences, San Jose, CA), Iba1 (Wako, Richmond, VA), or Flt1 (R&D Systems), imaged using a Zeiss ApoTome or dissecting microscope and AxioCam camera (Zeiss, Oberkochen, Germany), and quantified using Image J (NIH, Bethesda, Maryland). PECAM intensity was normalized to the avascular epidermis of the respective section.

Breeding and genotyping of Wls^fl,²⁵ cfms-icre,²⁶ Flt1^fl,²⁷  and CNB1^fl28  were performed as previously described. All experiments used littermates as controls. Mouse husbandry and experimentation was approved by the Cincinnati Children's Hospital Research Foundation Institutional Animal Care and Use Committee.

To investigate the role of macrophage Wnt signaling during wound repair, we compromised Wnt ligand secretion by inducing deletion of the conditional Wls allele²⁵  with the transgene cfms-icre.²⁶  In response to full thickness dermal wounding, animals lacking macrophage Wls (Wls^fl/-;cfms-icre) repaired faster than littermate controls (Wls^fl/-) (Figure 1A-C). Significant differences were detected at both 4 and 5 days after wounding (Figure 1A-C), but no difference was detected in the number of wound macrophages (Figure 1D). Importantly, cfms-icre expression was not responsible for the phenotype (supplemental Figure 1A). Because many articles have documented a correlation between wound repair and angiogenesis,^29-36  vascular density was assessed using PECAM antibody labeling. The Wls^fl/-;cfms-icre animals displayed enhanced angiogenesis with more intact vascular structures (Figure 1E,F).

To determine whether the role for macrophage Wnts in dermal wound repair was a more general response, we implemented the ARA, an in vitro analysis of angiogenic wound responses.²⁴  Vessel growth in the ARA requires endogenous macrophages.³⁷  In response to aortic wounding, many more vessels were seen in Wls^fl/-;cfms-icre animals relative to controls (Figure 1G,H). Taken together, these data indicated that macrophage Wnt ligands normally suppress angiogenesis.

Myeloid Wnts have been shown to suppress retinal angiogenesis by inducing the secretion of Flt1.¹⁷  To test the role of macrophage Flt1 during wound repair, mice with a loxP-flanked Flt1 allele²⁷  were crossed to the cfms-icre animals. ARA analysis revealed significantly increased angiogenesis in Flt^fl/-;cfms-icre relative to controls (Figure 2A). Because macrophage Flt1 suppresses retinal and ARA angiogenesis, and because Wnt signaling upregulates Flt1,¹⁷  we quantified Flt1 immunolabeling in Wls^fl/fl;cfms-icre wound macrophages. Importantly, macrophages from Wls^fl/fl;cfms-icre wounds had diminished Flt1 labeling (Figure 2B,C). Furthermore, when Flt1^fl/fl;cfms-icre mutant animals were exposed to full-thickness dermal wounds, they demonstrated enhanced repair (Figure 2D,E). It is important to note that these animals are deficient in both soluble Flt1 and membrane-tethered Flt1, and future work should serve to elucidate the relative role of these 2 splice variants.

To further define the Wnt-Flt1 signaling pathway, an in vitro culture system was used to assess whether certain inhibitors could prevent noncanonical Wnt5a from inducing Flt1. One of the best candidate mediators of this noncanonical Wnt response is Calcineurin-NFAT signaling. When myeloid EOC-2 cells were exposed to Wnt5a, soluble Flt1 expression was upregulated (Figure 2F). Importantly, this effect was not observed in the presence of Cyclopsorine A or NFAT activation inhibitor III (INCA-6), a potent inhibitor of Calcineurin-NFAT interactions but not an effector of other calcinuerin functions.³⁸  Similar findings were observed in the myeloid-like RAW264.7 cells (supplemental Figure 1B,C). To determine whether the in vivo wound repair process also required calcineurin, animals were generated with a conditional deletion in CNB1,²⁸  the subunit of the Calcineurin-NFAT complex that is required for signaling.²⁸  Analysis of dermal wounds in these animals revealed enhanced wound repair in the absence of calcineurin signaling (Figure 2G), a response consistent with both the Wls and Flt1 mutant wound responses. Interestingly, ARA angiogenesis was enhanced in the presence of Cyclosporine A.³⁹  Taken together, these data suggest that macrophages of the wound stroma use a Wnt-Calcineurin-NFAT-Flt1 pathway to suppress wound angiogenesis and slow wound repair.

One important caveat in the analysis presented here is the somewhat promiscuous activity of the cfms-icre transgene.²⁶  The cfms-icre efficiency is nearly 100% in macrophages, but approximately 50% in granulocytes and lymphocytes.²⁶  In wound repair, several lines of reasoning suggest macrophages are the principle effector: (1) responses are seen 3 to 4 days after injury when macrophages are abundant but lymphocytes are rare; (2) Flt1 protein levels were diminished in Wls-deficient macrophages; (3) endogenous immune cells in the ARA are predominantly macrophages³⁷ ; and (4) similar wound vascular responses are observed in PU.1^−/− mice that have relatively normal lymphocyte populations.⁴⁰ 

In the wound, it seems counterintuitive that natural mechanisms would exist to suppress angiogenesis and slow repair rates. One hypothesis is that increasing angiogenesis may increase repair rates, although it might also make the wound weaker and more susceptible to a second injury during repair. Interestingly, wounds in patients treated with cyclosporine were significantly weaker.⁴¹  Therefore, it is possible that the Wnt-Calcinuerin-Flt1 pathway identified here is used by macrophages to suppress wound angiogenesis and consequently increase the transient strength of the repairing wound. However, in a context in which wounds can be protected during repair, therapeutic targeting of this pathway may elucidate novel mechanism by which wound repair rates could be increased.

The authors thank Paul Speeg for his technical assistance and Gerald R. Crabtree for the CNB1^fl mice.

This work was supported by grants from the National Institutes of Health (R01CA131270) (J.P. and R.A.L) and (T32GM063483-08S1) (J.A.S).

Contribution: J.A.S and R.A.L designed the experiments and wrote the manuscript. J.A.S., S.R., and K.B. performed the experiments. A.C.A. and R.F.N. provided technical assistance and advice. J.W.P. and N.F. developed critical reagents.

Conflict-of-interest disclosure: N.F. is an employee of Genentech Corporation. The remaining authors declare no competing financial interests.

Correspondence: Richard A. Lang, Division of Pediatric Ophthalmology, Cincinnati Children's Hospital Medical Center, 3333 Burnet Ave, Cincinnati, OH 45229; e-mail: richard.lang@cchmc.org.