Knockout of Angiotensin AT2 receptors accelerates healing but impairs quality.

Wounds are among the most common, painful, debilitating and costly conditions in older adults. Disruption of the angiotensin type 1 receptors (AT1R), has been associated with impaired wound healing, suggesting a critical role for AT1R in this repair process. Biological functions of angiotensin type 2 receptors (AT2R) are less studied. We investigated effects of genetically disrupting AT2R on rate and quality of wound healing. Our results suggest that AT2R effects on rate of wound closure depends on the phase of wound healing. We observed delayed healing during early phase of wound healing (inflammation). An accelerated healing rate was seen during later stages (proliferation and remodeling). By day 12, fifty percent of AT2R−/− mice had complete wound closure as compared to none in either C57/BL6 or AT1R−/− mice. There was a significant increase in AT1R, TGFβ1 and TGFβ2 expression during the proliferative and remodeling phases in AT2R−/− mice. Despite the accelerated closure rate, AT2R−/− mice had more fragile healed skin. Our results suggest that in the absence of AT2R, wound healing rate is accelerated, but yielded worse skin quality. Elucidating the contribution of both of the angiotensin receptors may help fine tune future intervention aimed at wound repair in older individuals.

AT 1 R blockers impair fibroblast migration and delay wound healing [11]. The angiotensin subtype 2 receptor is less studied, but its anti-inflammatory, anti-apoptotic and anti-proliferative effects are thought to oppose the effects of AT 1 R [12]. Virtually nothing is known about the contribution of the AT 2 R to stages of wound healing. The overarching hypothesis of this study is that a functional balance between skin expression of AT 1 R and AT 2 R is required for optimal healing. We further hypothesized that targeted deletion of AT 2 R would accelerate wound healing rate via un-opposed AT 1 R activity upregulating skin TGFβ signaling. To dissect the role of AT 2 R on wound healing we have selected the genetic knockout to avoid the effects of variations in drug delivery to wound bed. Furthermore, given that Angiotensin II binds with equal affinity to AT 1 R or AT 2 R, the knockout of the angiotensin receptors allows for better discrimination of the effects of the receptors by eliminating the possibility of remaining unblocked receptors. In this study we compared C57BL/6J wild-type (WT) mice to age-and gender-matched; AT 1 R knockout (AT 1 R -/-) mice and AT 2 R knockout (AT 2 R -/-) mice.

RESULTS
To ascertain the influence of angiotensin receptors on wound healing, downstream effectors and healed skin quality, we compared C57BL/6J wild-type (WT) mice to age-and gender-matched; AT 1 R knockout (AT 1 R -/-) mice and AT 2 R knockout (AT 2 R -/-) mice.
Delayed wound healing in AT 1 R -/and accelerated wound healing in AT 2 R -/mice RAS is a key hormonal system whose dysregulation has been linked to aging, inflammation, and impaired wound healing. We show that the AT 1 R -/mice were the most delayed in wound healing as compared to WT and the AT 2 R -/- (Figure 1, panel A, B and C) which is in agreement with previous reports on wound healing in AT 1 R -/mice [11]. In contrast, the healing rate in the AT 2 R -/mice was accelerated which was unexpected given that AT 2 R levels in general correlate with positive outcomes [12]. By day 12, 50% of the animals in the AT 2 R -/cohort achieved complete wound healing as compared to none in either the AT 1 R -/or WT mice(P<0.05). By day 12, only 5% of the size of wounds remained unhealed in the AT 2 R -/mice vs. 17% in the AT 1 R -/and 11% in WT mouse cohorts (P<0.05; Figure 1, panel C and E). All the AT 2 R -/mice were healed completely by day 16 as compared to 10% of WT and 30% of AT 1 R -/that remained with open wounds (P<0.05; Figure 1: panel C and E). Further analysis of the fastest healing group, the AT 2 R -/-, revealed a delayed healing during the inflammatory phase of wound healing (day1-7) and an accelerated healing during proliferative and remodeling phases (day8-16). ( Figure 1D) This biphasic pattern of healing observed in the AT 2 R -/group, was not seen in either the AT 1 R -/or WT mice. www.impactaging.com

Increase in wound blood flow in the AT 1 R -/mice
Given the prominent role for angiotensin receptors in tissue perfusion and to determine if changes in blood flow to the wounds contributed to the observed healing pattern in AT 1 R -/and AT 2 R -/mice, wound area blood flow was measured by non-invasive LDPI on days 7 and 11. There were no differences in wound area blood flow among the three groups by day 7, but by day 11 we observed a significantly higher value in the AT 1 R -/as compared to the other two groups (P<0.05; Figure 2). Interestingly, the increase in blood flow in the AT 1 R -/mice did not correlate with better wound healing. Further, we did not observe differences in wound blood flow the AT 2 R -/as compared with WT controls, suggesting that changes in blood flow did not contribute to the accelerated healing observed in the AT 2 R -/mice.
Down regulation of TGFβ isoforms and the downstream target proteins (Smads) in AT 1 R -/and upregulation of AT 1 R, TGFβ 1 and TGFβ 2 during later stages of wound healing in the AT 2 R -/-Although not completely characterized, all phases of wound healing appear to be greatly influenced by subtle modulation of TGF-β, which is strongly influenced by RAS [6;7;30-32]. The three isoforms of TGF-β (β1, β2, and β3) signal through the same cell surface receptor, but appear to play distinct functions during wound healing. While TGF β1 and β2 have predominantly proscarring roles, TGF β3 have mainly anti-scarring effects [32]. RAS has been tightly linked to TGFβ activity but the specific effects of AT 1 R and AT 2 R on the different TGF isoforms are not known. Using qPCR we determined differences in expression of the three different isoforms of TGFβ in the healing skin (day 20) of our mouse cohorts. Our results demonstrate a significant decrease in all the three different isoforms of TGFβ in the wound of the AT 1 R -/mice (P<0.05; Figure  3: panel A, B and C). TGF-β1 mRNA levels in AT 1 R -/mice were decreased 7.69 fold as compared to WT mice. In contrast the TGFβ1 mRNA levels increased 1.83 fold in AT 2 R -/mice as compared to WT mice (P < 0.005; Figure 3: panel A). The expression of TGFβ2 mRNA was also decreased 20 fold in AT 1 R -/mice compared with WT mice (P < 0.005; Figure 2: panel B). TGFβ3 mRNA expression was decreased 7.1 fold in AT 1 R -/mice in comparison with WT mice (P < 0.005; figure 2: panel C). In contrast, we observed an increase only in TGFβ1 in AT 2 R -/-, compared to both WT control www.impactaging.com and AT 1 R -/mice (p <0.05; Figure 3: panel A). We quantified changes in AT 1 R and the three isoforms of TGFβ in days 0, 3, 7 and 9 of wound healing to investigate if the biphasic pattern observed in wound healing of the AT 2 R -/corresponded to a stage dependent changes in AT 1 R and to quantify changes in TGF-β isoforms. Our results suggest that the expression of AT 1 R was upregulated by day 7 of wound healing (2.18 and 2.56 fold change respectively, p<0.05; Figure 3: panel D). This increase corresponded to an increase in both TGFβ1 and TGFβ2 (p<0.05; Figure 3: panel E and F). No change in TGFβ3 was observed. Changes in AT 1 R in different stages of wound healing strongly correlated with the changes in TGFβ1 (Pearson r=0.99, p=0.04).
Next we sought to determine the impact of the disruption of angiotensin receptors on the downstream target proteins of the TGFβ signaling pathway. TGFβ signals through Smad2 and Smad3 that are phosphorylated by TGFβ receptors and translocate to the nucleus with the common-mediator (co-Smad) Smad4 [33;34]. Our results showed no significant difference in Smad2 in wounds of AT 1 R -/or AT 2 R -/mice as compared to WT mice ( Figure  4). Consistent with the reduction in the three isoforms of TGFβ, we have observed a significant reduction in both Smad3 and the common-mediator Smad4 in wounds of AT 1 R -/mice as compared with WT mice (P<0.001, Figure 4). Our results also demonstrate reduction in phosphorylation of Smad2 and Smad3 only in wounds of AT 1 R -/mice as compared with WT mice (P<0.01, Figure  5). Interestingly, we have observed a similar decrease in Smad3 in AT 2 R -/mice as compared to WT mice (P<0.0001, Figure 4). There were no differences in Smad4, phospho-Smad2 or phospho-Smad3 in wounds of AT 2 R -/mice.

Reduced repair and proliferation activity in wounds of AT 1 R -/mice
The angiotensin receptors have been linked to changes in cell differentiation and proliferation. While AT 1 R have been shown to increase cell proliferation, AT 2 R have been shown critical for cell differentiation [12].   www.impactaging.com Given the observed delayed wound healing noted in AT 1 R -/mice, we wanted to determine if this delay was driven by changes in cellular proliferative activity. We quantified changes in Proliferating cell nuclear antigen (PCNA), a nuclear protein essential for DNA replication and repair and is a marker for cellular growth and proliferation. To further investigate the mitotic activity at the wound site, we studied expression levels of the phosphorylated form of the histone protein H3. Histone H3 phosphorylation is linked to cells that are actively dividing. Consistent with the delayed healing rate in AT 1 R -/mice, we have observed a significant reduction in PCNA (P<0.005, Figure 6) and in mitotic histone H3 phosphorylation at several residues, including serines 28 (P<0.005) as well as threonines 3 (0.005) and 11(0.05) (Figure 7). Surprisingly, we have observed a similar decrease in PCNA and phosphorylation of Histone 3 Threonine 3 residue in AT 2 R -/mice. The impact of the differential phosphorylation of certain histone H3 residues in AT 2 R -/mice on wound healing activity and scar quality is currently unclear.

Tensiometry shows wound fragility in AT 2 R -/mice
Given the pro-scarring and fibrotic effects of TGFβ1, we next sought to determine if there was a difference in the healed skin's physical characteristics (Peak force, total work and compliance). Our results show that despite the accelerated healing rate observed in the AT 2 R -/mice, the healed skin in the AT 2 R -/mice was more fragile, fracturing more easily (Panel 8B), being more compliant (Panel 8C), and breaking with less work (Panel 8D) than wounds from AT 1 R -/or WT mice. (P<0.05). Masson's trichrome staining of healing skin shows increase of subcutaneous fat in both AT 1 R -/and AT 2 R -/-. A reduction in dermal collagen zone in AT 1 R -/was also observed ( Figure 9).

DISCUSSION
Several lines of evidence suggest that increased RAS activity through the AT 1 R plays a crucial role in wound healing [5][6][7][8][9]. Our results further dissects the impact of angiotensin receptors on wound healing. AT 2 R antagonizes inflammatory signaling, a necessary activating function that leads to the proliferation phase. The lack of AT 2 R was associated with a slower closure rate during the early stages. This may have resulted from an unopposed pro-inflammatory AT 1 R, causing delayed resolution of the inflammatory phase and impairing the transition to the proliferative and remodeling phases [1-4;35].   We have observed a delayed healing pattern in AT 1 R -/throughout all phases of wound healing, which is consistent with the pro-inflammatory and proproliferative characteristics of AT 1 R [12], and is in agreement with previous reports [11]. This is also is supported with the significant reduction in both PCNA and phospho-Histone H3 in healing skin of the AT 1 R -/mice. In contrast, by day 8 in AT 2 R -/mice, as the healing wounds were transitioning to the proliferative phase we observed a significant upregulation of wound AT 1 R along with an accelerated rate of healing. Our combined results of accelerated healing with the upregulation of AT 1 R (in AT 2 R -/mice), contrasted with delayed healing in AT 1 R -/may suggest phasedependent role for increased AT 1 R signaling during the proliferative phase through alterations in TGF-β signaling and alterations in the extracellular matrix [30;36;37].
The relationship between the TGFβ family and angiotensin receptors is not entirely mapped out and remains mechanistically vague. Previous studies have reported that RAS activation through AT 1 R increases TGFβ signaling [38;39]. Our results demonstrate a significant down regulation of all the three isoforms of TGFβ and the downstream targeting proteins Smad3, Smad4, as well as the phospho-Smad2 and phospho-Smad3 in the AT 1 R -/mice. The impact of TGFβ on cutaneous wound healing has been well established. The release of TGFβ 1 during early stages of healing prompts the expression of key components such as fibronectin, collagen types I and III, and VEGF [32]. Additionally, TGFβ 1 improves angiogenesis to facilitate blood supply to the injured site [40] which then stimulates fibroblasts to allow for wound closure [41]. Whether the decrease in TGFβ 1 is causal of the impaired wound healing in the AT 1 R -/mice is not known. However, in AT 2 R -/mice there was a strong, positive correlation between dermal TGFβ 1 and AT 1 R expression in later stages of wound healing that corresponded to an accelerated wound closure. This is in agreement with previous reports linking AT 1 R stimulation to increased TGFβ 1 expression and collagen maturation [42] and may potentially explain the decreased compliance seen in healed skin in AT 1 R -/mice.
Similarly, the second isoform TGFβ 2 , is involved in granulation tissue formation, angiogenesis and collagen synthesis [43;44]. Impaired wound healing has been demonstrated in TGFβ 2 transgenic mice [45]. In AT 2 R -/mice, we observed a decrease in the expression of TGFβ 2 by day 3 of wound healing. This initial drop corresponded to the impaired healing seen in early phases in AT 2 R -/mice. Furthermore, we have observed www.impactaging.com a significant increase in TGFβ 2 by day 7 that was matched with a faster healing rate.
The lack of change in TGFβ 3 and blood flow in AT 2 R -/mice may suggest that these two factors do not play a significant role in modulating wound healing in response to the knocking out of the AT 2 R.
Changes in the TGFβ downstream signaling proteins (Smads) have been linked to the rate of wound healing. The down regulation of Smad3 have been linked to acceleration of wound closure [46]. In contrast, the knockdown of Smad4 was associated with aberrant wound healing [47]. Consistent with previous reports on the role of Smad3, we have noted the lowest level of Smad3 in AT 2 R -/along with the fastest wound closure rate. The down regulation of Smad4 in AT 1 R -/may have played a role in the delayed healing rate.
In summary, the silencing of AT 1 R delayed wound healing, while the interruption of AT 2 R accelerated wound healing. Furthermore, this effect in AT 2 R -/mice, was at least partially mediated by AT 1 R, TGFβ 1 and TGFβ 2. This data supports the notion of the antagonistic interaction between AT 1 R and AT 2 R.
Mitochondria provide energy and produce reactive oxygen species to drive the increased mitotic and synthetic activity necessary for wound healing. Several groups demonstrated a link between age-related mitochondrial dysfunction and impaired wound healing [48]. The identification of a functional intramitochondrial angiotensin system (MAS) [49] may provide additional insight into the RAS interface with wound healing. Activation of the intra-mitochondrial AT 2 R is coupled to increased nitric oxide generation and inhibition of mitochondrial energy production [49]. Further work is needed to determine the impact of MAS on wound healing.
There are several limitations to our current study. Structurally, mice skin differs from human skin in that mice have much thinner epidermal and dermal layers than humans. Furthermore, mice also have a large subcutaneous muscle layer, which augments wound repair by contraction making further studies in a second animal model (pigs) or humans necessary before extrapolating results to humans. Also, given that we are studying mice that are homozygote knockouts for either the AT 1 R or the AT 2 R, partial effects of the genes and the compensatory effects of one angiotensin receptor on the absence of the other receptor are still not clear.
Given the effects of aging on angiotensin receptors [12;49;50], and that many aged, frail individuals are already on angiotensin receptor blockers, this research highlights the crosstalk between AT 1 R and AT 2 R and that pinpointing the exact molecular changes in angiotensin receptors and the impact of angiotensin receptor blockers on wound healing in aged individuals is important for the progression of the field of wound healing.

METHODS
Mouse models. This study was approved by the Johns Hopkins Animal Care and Use Committee (ACUC). To ascertain the influence of angiotensin receptors on wound healing, downstream effectors and healed skin quality, we compared 28-week old male C57BL/6J wild type (WT) mice (Jackson Laboratories, Bar Harbor, Maine) to age and gender matched AT 1 R knockout (AT 1 R -/-) (Jackson Laboratories, Bar Harbor, Maine) [13] and AT 2 R knockout (AT 2 R -/mice (supplied by our collaborator Dr. Tedashi Inagami, Vanderbilt University, TN) [ 14;15]. Male mice were employed to avoid the effects of hormonal changes on wound blood flow and healing.
Wounding procedure and area calculation. Mice (N=10 in each group) were anesthetized by a mobile RC 2 nonrebreathing anesthesia machine (Vet Equip, Inc. Pleasanton, CA). Buprenorphine (1 mg/kg) was administered by a subcutaneous injection during the first 24 hours. A full thickness 8 mm wound was created by punch biopsy. On days 0, 3, 5, 7, 9, 11 and 13, the wound borders were traced in situ onto clear acetate paper. Images were digitized at 600 dpi (Hewlett Packard Company, Laser Jet 3390, Paolo Alto). Wound areas (in pixels) were calculated using Adobe Photoshop CS3 Image software (Adobe System Inc. San Jose, CA). Wound area on day 0 was taken as a 100% and a wound size ratio obtained with that measurement in each time point. The wounds were checked daily after day 10 until complete closure.
Laser Doppler Perfusion Imaging (LDPI). Blood flow in the wound areas was measured at days 7 and 11 using a 633 nm, He-Ne scanning laser Doppler imaging device (Moor Instruments, Devon, UK], which utilizes a nearinfrared laser diode to measure subcutaneous blood flow as a function of light scattering by moving red blood cells (Doppler shift), as described previously [16].
Physical measurements of tissue strength. Peak force, work to rupture and flexibility of healed skin were calculated at day 21 using a FGV-10XY tensiometer (Checkline by Electromatic, Cedarhurst, NY) to record the force generated as the skin was elongated until rupture as described previously [17].

Funding
This study was supported by the Johns Hopkins Older Americans Independence Center National Institute on Aging Grant P30 AG021334, National Institute on Aging Grants 1R01AG046441 and K23 AG035005, Nathan Shock in Aging Scholarship Award (PMA), R21AG043284 (JW and PMA), the Wound healing society foundation 3M scholarship(PMA) , NIH research grant HL58205 (TI) and NIH grant S10 OD016374 (microscope Facility).

Author contributions
Mahya Faghih-PCR measurement of angiotensin receptors and TGFβ; drafting of the manuscript; critical revision of the manuscript for important intellectual content. Sayed Mohammad Hosseini-performance of animal experiments, acquisition of data, critical revision of the manuscript for important intellectual content. Amir Mehdi Ansari-acquisition of data, critical revision of the manuscript for important intellectual content. Barbara Smith-IHC staining and quantification of www.impactaging.com SMADs and wound healing markers. Drafting of the manuscript; critical revision of the manuscript for important intellectual content. Frank Lay-acquisition of data, critical revision of the manuscript for important intellectual content. Tedashi Inagami: provided the AT 2 R -/mice. Study concept and design; critical revision of the manuscript for important intellectual content. Guy Marti-provided expertise regarding the animal model and the methods for assessing wound healing including planimetry, laser Doppler perfusion imaging of blood flow and assessment of biodynamic characteristics of the skin with tensiometry. Critical revision of the manuscript for important intellectual content. John W. Harmon-provided expertise regarding the animal model and the methods for assessing wound healing including planimetry, laser Doppler perfusion imaging of blood flow and assessment of biodynamic characteristics of the skin with tensiometry. Critical revision of the manuscript for important intellectual content. Jeremy D. Walston -study concept and design; critical revision of the manuscript for important intellectual content. Peter Abadir-study concept and design; acquisition of data; analysis and interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; administrative, technical, and material support; study supervision.