Topical application of insulin accelerates and improves the quality of healing
To study the effects of insulin on wound healing, 7 mm diameter excision wounds were performed on the back of C57BL/6J mice, and locally treated with 0.03 U of insulin. This dose of insulin was chosen because it significantly stimulated healing (Fig. 1A) without affecting blood glucose levels (unpublished data). We analyzed the wound area throughout the healing process to monitor the time-dependent effects of insulin on healing, and found that in wounds treated with insulin the wound area was significantly decreased at several time points (Fig. 1B), as was the time to closure (control 10.25 ± 1.26 d, insulin 8.9 ± 0.32 d, P < 0.01). We found that insulin significantly decreased wound area by day 3 after injury. In order to elucidate the effects of insulin during this early stage of healing, we took skin samples from control and insulin-treated wounds, and compared the histological characteristics of these two wounds. At day 3, we found that in insulin-treated wounds the keratinocyte tongue was much longer than that in the control wounds, suggesting that insulin stimulates keratinocyte migration (Fig. 1C). When we measured the extent of migration of the keratinocytes by determining the length of migration of the tongue from the margin of the wound to the tip of the migrating keratinocytes, we found that there was a significant increase in migration distance of the keratinocytes in insulin-treated wounds (Fig. 1D). Furthermore, we also found that following wound closure, the epidermis of insulin-treated wounds is better defined and is characterized by an increased number of epidermal reticular ridges and dermal papilla that are not evident in the control (Fig. 1E).
Insulin stimulates keratinocyte migration in a time- and dose-dependent manner
Histological observation of wounds treated with insulin suggests that this protein stimulates keratinocyte migration. Although the effects of insulin on keratinocyte proliferation are well established [20], its effect on migration of these cells is not clear. To study the latter process, we used HaCaT keratinocytes in culture. Cells were plated in cloning rings and allowed to attach; the rings were removed after marking their positions and the cells were then treated with insulin. Migration distances from the initial edge of the cells to the new edge of the cells were measured at 24, 48, and 72 hr. At each time point, keratinocytes treated with insulin showed increased migration over the control (Fig. 2A, B). To determine whether this effect is dose-dependent, we performed the migration assays with different doses of insulin. A concentration as low as 10-8 M insulin was able to increase keratinocyte migration, which was highly significant after 48 hr of treatment. However, concentrations ranging from 10-7 M to 10-5 M significantly enhanced keratinocyte migration by 24 h (Fig. 2C). To eliminate the possibility that this migration is dependent on proliferation, we treated the cells with mitomycin C, a potent DNA crosslinker and hence inhibitor of cell proliferation, in the presence or absence of insulin. Cells were pre-treated with 5 μg/ml of mitomycin C for 3 hrs and then exposed to 10-7 M insulin for 24 and 48 hrs. Although we observed that proliferation was halted, insulin-induced migration was not (Fig. 2D), strongly suggesting that the two processes are independently regulated by this hormone.
Insulin stimulates keratinocyte migration in an insulin receptor-dependent manner but in an EGF-independent manner
It has been shown that insulin activates both its own receptor and the IGF-1 receptor, albeit with different affinities [21]. These transmembrane proteins are both tyrosine kinases, share 60% homology, and activate a number of insulin receptor substrates which then initiate signals that lead to gene expression. These genes are involved in many of the different effects of insulin on cells and ECM molecules, as well as their receptors such as integrins, which provide critical signals to guide cell movement. To determine whether insulin-induced keratinocyte migration is dependent on one or both receptors, each group of cells was pre-treated with either the neutralizing insulin receptor Ab, 29B4, or the IGF-1 receptor tyrosine kinase inhibitor, picropodophyllin and then treated with 10-7 M or 10-6 M insulin. Pre-treatment of keratinocytes with the insulin receptor Ab followed by 10-7 M insulin treatment completely abolished insulin-induced migration, suggesting that at this concentration these effects are primarily mediated by the insulin receptor itself (Fig. 3A). When cells were treated with the insulin receptor Ab and 10-6 M insulin, the Ab only partially blocked insulin-induced migration (Fig. 3A), suggesting that this concentration of insulin may induce migration through both the insulin and IGF receptors. To confirm these results, we pre-incubated the cells with picropodophyllin for 1 hr and then treated them with insulin. At 10-7 M, insulin-induced keratinocyte migration was not affected, but the inhibitor did decrease cell migration induced by 10-6 M insulin (Fig. 3B). Moreover, the keratinocyte migration resulting from 10-6 M insulin was abrogated when inhibiting both insulin and IGF receptors using the 29B4 Ab and picropodophyllin (Fig. 3B). Taken together, these data show that the effect of high concentration (10-6 M) of insulin on keratinocyte migration is mediated by both insulin and IGF-1 receptors, whereas the effect of lower concentrations of insulin (10-7 M) is primarily mediated by the insulin receptor. To study the effects of insulin that are mediated only through the insulin receptor and its associated downstream signaling pathways, 10-7 M insulin was chosen for the subsequent studies except when otherwise indicated.
Our previous studies showed that insulin stimulates EGF expression in wound marginal keratinocytes of deep partial thickness scald wounds in rats [16]. In order to exclude the potential autocrine effects of EGF secretion on insulin-induced keratinocyte migration, we treated the keratinocytes with AG1478, a selective inhibitor of EGF-R kinase, prior to treatment with insulin and measured migration distances at 24 and 48 h. This EGF-R inhibitor did not inhibit the effects of insulin on keratinocyte migration (Fig. 3C), suggesting that this process does not require EGF or its receptor.
PI-3K and Akt mediate insulin-induced keratinocyte migration
In order to determine the signal transduction pathways in insulin-induced keratinocyte migration, we examined Akt phosphorylation/activation by immunoblot analysis [22, 23], and found that the levels of phosphorylation of this signal transduction mediator increased after 5 min of insulin treatment, and remained elevated for at least 60 min (Fig. 4A). This effect was also dose-dependent (Fig. 4B). To determine whether Akt phosphorylation/activation was involved in insulin-induced keratinocyte migration, we infected keratinocytes with recombinant adenovirus expressing the constitutively active mutant of Akt (Akt-CA) or the dominant-negative mutant of Akt (Akt-DN). Higher levels of p-Akt were found in the keratinocytes expressing Akt-CA even without insulin treatment when compared to cells expressing the Akt-DN (Fig. 4C). To determine whether Akt phosphorylation is important in keratinocyte migration, we used the scratch wound migration assay. This assay was used to avoid trypsinizing and re-plating the cells because the viral infected cells have decreased survival. We compared keratinocyte migration distances with or without insulin treatment, in cells expressing the Akt mutants (Fig. 4D, E). Uninfected cells and cells infected with the null vector and then treated with insulin as well as cells infected with the Akt-CA, with or without insulin treatment, displayed the longest migration distance. In contrast, cells infected with Akt-DN exhibited a significant decrease in migration, even with insulin treatment, compared with all other groups, illustrating the requirement of Akt for insulin-induced keratinocyte migration.
PI3K is often involved in AKT phosphorylation. Therefore, to determine whether insulin stimulation of keratinocyte migration is dependent on PI3K activity, we performed the migration assays in the presence of LY294002, an inhibitor of PI3K. This treatment completely blocked keratinocyte migration stimulated by insulin (Fig. 4F), showing the importance of PI3K in this process. The dose of LY294002 we used does not inhibit S6 kinase, which is the effector of mTOR, nor does it affect MAP kinase, PKC, or PI4K [24].
Insulin stimulates translocation of Rac1, but not RhoA, to the plasma membrane; this process requires PI3K-Akt activation and is involved in insulin-induced keratinocyte migration and wound healing
Small GTPases of the RhoA family play important roles in cell motility. Therefore, we tested the possibility that the PI3K-Akt pathway stimulates RhoA activation during insulin-induced keratinocyte migration. Using immunolabeling, we show that there was no significant difference in RhoA distribution shortly after insulin treatment (Fig. 5A–C) nor did a change occur at least for 4 h. However, 3 min after insulin treatment, Rac1, another member of the RhoA family of GTPases, translocated from the cytosol to the plasma membrane, indicating its activation (Fig. 5E). This redistribution effect was also seen after 5 min of insulin treatment (Fig. 5F). In addition, plasma membrane ruffling was observed at the leading edge of migrating keratinocytes, with Rac-1 being present in the membrane of ruffles (Fig. 5F; arrow). To determine whether this translocation is consistent with the activation of Rac1, we performed Rac1 pull down assays, which can specifically pull down the active form of Rac, and found elevated levels of active Rac1 after insulin treatment (Fig. 5G). In order to confirm that Rac1 activation is important in insulin-induced keratinocyte migration, we transfected keratinocytes with plasmids containing mutant forms of Rac1, and then observed the effects of these mutants on insulin-stimulated cell migration. Keratinocytes were transfected with either the constitutively active form of Rac1 (V12, Rac1-CA), dominant-negative mutant Rac1 (N17, Rac1-DN), or wild type Rac1 (Rac1-WT). 24 h after transfection, scratch wounds were made in the cell cultures, and cell migration distances were measured in non-transfected and transfected cells. After treatment for 24 h, insulin stimulated migration in non-transfected cells, as well as in cells transfected with Rac1-CA or Rac1-WT. Cells transfected with Rac1-CA showed increased migration, even without insulin treatment. However, insulin-induced migration was eliminated in cells transfected with Rac1-DN (Fig. 5H). To determine whether the PI3K-Akt pathway is required for insulin-induced Rac1 activation/translocation, LY294002 was used to pre-treat the cells before insulin treatment. Insulin-induced Rac1 activation (Fig. 5I) and translocation (Fig. 5J) was inhibited by this PI3K inhibitor.
Insulin stimulates integrin a3 and LN332 production, which contributes to insulin-induced keratinocyte migration and wound healing
It is well known that the Rac1 GTPase is critical in cytoskeleton re-organization, that the cytoskeleton interacts with integrins on the cell surface, and that these integrins interact with ECM molecules. It is also known that the integrin α3β1 and LN332, a basement membrane (BM) component, are important in both keratinocyte migration and BM development. Therefore, we investigated the possibility that insulin modulates LN332 and integrin α3β1 expression in vitro and in vivo, and that these proteins are involved in insulin-induced wound healing. Keratinocytes were seeded in cloning rings in order to observe the integrin α3 and LN332 at the migration edge. Immunolabeling for integrin α3 showed higher levels of this protein on the cell membrane after insulin treatment (Fig. 6A, B). The increased integrin α3 levels were confirmed by immunoblot analysis (Fig. 6C).
Using similar methodology, LN332 was also found to be elevated after insulin treatment, particularly at the migrating edge, with deposition of LN332 along the cell membrane in some cells (Fig. 6D, E and inserts). Immunoblot analysis showed an increase in the LN332 protein after insulin treatment (Fig. 6F). To determine whether these observations translate into changes in the migratory behavior of keratinocytes, we observed the effect of function-inhibiting Abs to integrin α3 and LN332 on insulin-induced keratinocyte migration (Fig. 6G). Cell migration was inhibited when integrin α3 or LN332 were blocked with these Abs, while basal migration remained virtually unaffected. The effects of these blocking antibodies were more obvious after 48 h of insulin treatment; at this time point, functional blocking of α3 and LN332 affected both basal and insulin-induced keratinocyte migration. Moreover, additional inhibition was observed when insulin treatment was accompanied by treatment with both integrin α3 and LN332 function-inhibiting Abs (Fig. 6G).
These results suggest that in vivo both α3 and LN332 are involved in insulin-induced acceleration of wound closure. To test this possibility, we applied the function-blocking Abs against both proteins to mouse excision wounds, and found that inhibition of LN332 (Fig. 7A, B) or integrin α3 (Fig. 7C, D) resulted in delays in healing, primarily at early times after wounding. Histological examination shortly after wound closure showed that both inhibition of LN332 or integrin α3 resulted in a less mature epidermis (Fig. 8A). When the antibodies to these two molecules were applied, no reticular ridges were seen, the basal cells were not well defined, the interactions of the epidermis with the dermis were less well defined, and appendages were not seen. Furthermore, staining for Collagen IV (Fig. 8B), a component of basement membrane, showed that in the wounds treated with the function-inhibiting Abs for LN332, the basement membrane was not well developed, and when treated with α3 integrin, the basement membrane was irregular containing many dense areas of Collagen IV deposition (compare with the staining for wounds treated with insulin alone). When we stained for keratin 10 (Fig. 8C), a marker of keratinocyte differentiation [25], the basal keratinocytes of the wounds treated with the inhibiting antibodies did not express keratin 10. These results suggest that insulin promotes epithelial basement membrane deposition and keratinocyte differentiation.