Ischemia Reperfusion Injury Produces, and Ischemic Preconditioning Prevents, Rat Cardiac Fibroblast Differentiation: Role of KATP Channels

Ischemic preconditioning (IPC) and activation of ATP-sensitive potassium channels (KATP) protect cardiac myocytes from ischemia reperfusion (IR) injury. We investigated the influence of IR injury, IPC and KATP in isolated rat cardiac fibroblasts. Hearts were removed under isoflurane anesthesia. IR was simulated in vitro by application and removal of paraffin oil over pelleted cells. Ischemia (30, 60 and 120 min) followed by 60 min reperfusion resulted in significant differentiation of fibroblasts into myofibroblasts in culture (mean % fibroblasts ± SEM in IR vs. time control: 12 ± 1% vs. 63 ± 2%, 30 min ischemia; 15 ± 3% vs. 71 ± 4%, 60 min ischemia; 8 ± 1% vs. 55 ± 2%, 120 min ischemia). IPC (15 min ischemia, 30 min reperfusion) significantly attenuated IR-induced fibroblast differentiation (52 ± 3%) compared to 60 min IR. IPC was mimicked by opening KATP with pinacidil (50 μM; 43 ± 6%) and by selectively opening mitochondrial KATP (mKATP) with diazoxide (100 μM; 53 ± 3%). Furthermore, IPC was attenuated by inhibiting KATP with glibenclamide (10 μM; 23 ± 5%) and by selectively blocking mKATP with 5-hydroxydecanoate (100 μM; 22 ± 9%). These results suggest that (a) IR injury evoked cardiac fibroblast to myofibroblast differentiation, (b) IPC attenuated IR-induced fibroblast differentiation, (c) KATP were involved in IPC and (d) this protection involved selective activation of mKATP.


Introduction
Cardiac ischemia reperfusion (IR) injury describes the damage caused by reduced coronary blood flow, causing depletion of ATP, reduced partial pressure of oxygen (PO 2 ), acidosis, and build-up of toxins [1]. Reperfusion leads to further damage through generation of oxygen free radicals and a proton gradient across both the sarcolemma and the inner mitochondrial membrane [2,3].
Ischemic preconditioning (IPC) was first described by Murry et al. [4] and is classically defined as one or more cycles of brief IR injury, which protect the heart against a subsequent prolonged ischemic insult. The mechanisms responsible for generating this protection are complex and debated; however, the activation of ATP-sensitive potassium (K ATP ) channels has been strongly implicated [5]. K ATP channels were first described on the sarcolemmal membrane [6] of cardiac myocytes and were later discovered on the inner mitochondrial membrane [7]. Mitochondrial (mK ATP ) and sarcolemmal (sK ATP ) K ATP channels are postulated to evoke cardioprotection by hyperpolarizing the sarcolemma Following dissociation, IR was simulated at 37 °C under sterile conditions. Dissociated cells were centrifuged and resuspended in DMEM. Reserving a small layer of supernatant, ischemia was simulated by the addition of paraffin oil and reperfusion was simulated by the removal of paraffin oil and the replacement with DMEM, as described above (Figure 1). Pelleted cells were exposed to 30, 60-or 120-min ischemia followed by 60 min reperfusion. Time controls (TC) were run, in which equal volumes of DMEM were applied in place of paraffin oil. Cells were then placed under culture conditions.

Effect of Ischemic Preconditioning on Ischemia Reperfusion Injury Induced Fibroblast to Myofibroblast Differentiation
Dissociated cells were centrifuged and resuspended in DMEM. Reserving a small layer of supernatant, IPC was studied by exposing pelleted cells to one episode of 15 min ischemia followed by 30 min reperfusion, prior to 60 min ischemia and 60 min reperfusion (Figure 1). Time controls were also run. Following IPC and IR, cells were placed under culture conditions. While the identities of all cell types within the cultures were not characterized in the current investigation, previous studies indicated that the unpassaged (P0) cultures contained ≥95% fibroblast purity [23,24]. Staining for factor VIII indicated that less than 1% of the culture were endothelial cells, while staining for desmin indicated that less than 1% of the culture were vascular smooth muscle cells [23,24]. . Protocols utilized to determine the effect of ischemia-reperfusion (IR) injury and of ischemic preconditioning (IPC) on rat cardiac fibroblast to myofibroblast differentiation. IR injury was produced by subjecting freshly isolated cardiac cells to ischemia at 37 °C. Reperfusion lasted 60 min. IPC was induced by preceding the IR with 15 min ischemia and 30 min reperfusion. IPC was mimicked by the application of KATP channel openers pinacidil (Pin, 3rd row) or diazoxide (Diaz, 4th row). The effect of blocking KATP current on IPC was tested by including glibenclamide (Glib, 5th row) or 5-hydroxydecanoate (5HD, 6th row). Key: black boxes represent periods where ischemia was mimicked by layering sterile paraffin oil onto the cells, whereas, white boxes represent periods where cells were covered with DMEM, with or without drugs, as specified. Figure 1. Protocols utilized to determine the effect of ischemia-reperfusion (IR) injury and of ischemic preconditioning (IPC) on rat cardiac fibroblast to myofibroblast differentiation. IR injury was produced by subjecting freshly isolated cardiac cells to ischemia at 37 • C. Reperfusion lasted 60 min. IPC was induced by preceding the IR with 15 min ischemia and 30 min reperfusion. IPC was mimicked by the application of K ATP channel openers pinacidil (Pin, 3rd row) or diazoxide (Diaz, 4th row). The effect of blocking K ATP current on IPC was tested by including glibenclamide (Glib, 5th row) or 5-hydroxydecanoate (5HD, 6th row). Key: black boxes represent periods where ischemia was mimicked by layering sterile paraffin oil onto the cells, whereas, white boxes represent periods where cells were covered with DMEM, with or without drugs, as specified.

Effect of Ischemic Preconditioning on Ischemia Reperfusion Injury Induced Fibroblast to Myofibroblast Differentiation
Dissociated cells were centrifuged and resuspended in DMEM. Reserving a small layer of supernatant, IPC was studied by exposing pelleted cells to one episode of 15 min ischemia followed by 30 min reperfusion, prior to 60 min ischemia and 60 min reperfusion ( Figure 1). Time controls were also run. Following IPC and IR, cells were placed under culture conditions. While the identities of all cell types within the cultures were not characterized in the current investigation, previous studies indicated that the unpassaged (P0) cultures contained ≥95% fibroblast purity [23,24]. Staining for factor VIII indicated that less than 1% of the culture were endothelial cells, while staining for desmin indicated that less than 1% of the culture were vascular smooth muscle cells [23,24].

Role of Adenosine Triphosphate-Sensitive Potassium Channels in Ischemic Preconditioning
Dissociated cells were centrifuged and resuspended in DMEM. Cells exposed to K ATP blockers were subjected to 15 min IPC and 30 min reperfusion followed by 60 min ischemia and 60 min reperfusion ( Figure 1). Glibenclamide (Glib, 10 µM, Sigma) or 5-hydroxydecanoate (5HD, 100 µM, Sigma) were administered 15 min prior to the onset of IPC, and maintained throughout the protocol. In other cells, IPC was mimicked by adding pinacidil (Pin, 50 µM, Sigma) or diazoxide (Diaz, 100 µM, Sigma) in place of the 15 min IPC, prior to 60 min ischemia and 60 min reperfusion. Cells were then placed under culture conditions.

Characterization of the Ischemic Conditions
Data are presented as mean percent change ± SEM (n). Statistical significance was determined using one-way ANOVAs with Scheffe posthoc tests where p values of <0.05 were considered significant.

Ischemia Reperfusion Injury and Ischemic Preconditioning
Slides were coded to guard against observer bias. Five representative images were taken from each slide and cells were assessed for the presence of α-SMA stress fibers. Cells were classified 0-4 depending on the extent of α-SMA expression ( Figure 2). Undifferentiated fibroblasts did not express any α-SMA and were labelled 0. If the cell cytoplasm was occupied by greater than 75% α-SMA, the cell was classified as a fully mature myofibroblast and labelled 4. Cells with intermediate expression of α-SMA were classified as immature myofibroblasts and labelled '1-3 if α-SMA expression was as follows: 1, less than 25%; 2, between 25% and 50%; and 3, between 50% and 75%.
For most analyses, the relative percentages of fibroblasts across treatments were compared, and no differentiation between immature and fully mature myofibroblast frequencies was made. Data are presented as mean percent (of the total number of cells analyzed within the representative images) ± SEM (n) unless specified otherwise. Statistical significance was determined using paired or unpaired Student's t tests for single comparisons, and one-way or two-way ANOVAs with Scheffe posthoc tests for multiple comparisons. Significance was set at p < 0.05.
The pH, PO2 and PCO2 were also measured at 5, 15-and 30-min reperfusion, following replacement of paraffin oil with fresh DMEM. Though there was variation in the rate of recovery, especially of PCO2 at 5 min, average pH, PO2 and PCO2 were not significantly different from the preischemic conditions at any of the times sampled in reperfusion ( Figure 3B).
The pH, PO 2 and PCO 2 were also measured at 5, 15-and 30-min reperfusion, following replacement of paraffin oil with fresh DMEM. Though there was variation in the rate of recovery, especially of PCO 2 at 5 min, average pH, PO 2 and PCO 2 were not significantly different from the preischemic conditions at any of the times sampled in reperfusion ( Figure 3B).

The Effects of Ischemia Reperfusion on Cardiac Fibroblast Differentiation
When exposed to an ischemic period of 30 min followed by 60 min reperfusion, significant fibroblast to myofibroblast differentiation was observed, when compared to time control (12 ± 1% (5) vs. 63 ± 2% fibroblasts (5), respectively, Figure 4). With 60 min and 120 min ischemia followed by 60 min reperfusion, significantly fewer fibroblasts were present than in their respective time controls. Following 60 min ischemia, 15 ± 3% (12) of the culture were fibroblasts, compared to 71 ± 4% (13) of cells in 60 min time controls. Similarly, after 120 min ischemia, only 8 ± 1% (5) of the population within the cultures was fibroblasts, compared to 55 ± 2% (4) in the 120 min time control.

The Effects of Ischemia Reperfusion on Cardiac Fibroblast Differentiation
When exposed to an ischemic period of 30 min followed by 60 min reperfusion, significant fibroblast to myofibroblast differentiation was observed, when compared to time control (12 ± 1% (5) vs. 63 ± 2% fibroblasts (5), respectively, Figure 4). With 60 min and 120 min ischemia followed by 60 min reperfusion, significantly fewer fibroblasts were present than in their respective time controls. Following 60 min ischemia, 15 ± 3% (12) of the culture were fibroblasts, compared to 71 ± 4% (13) of cells in 60 min time controls. Similarly, after 120 min ischemia, only 8 ± 1% (5) of the population within the cultures was fibroblasts, compared to 55 ± 2% (4) in the 120 min time control.

The Effects of Ischemic Preconditioning on Cardiac Fibroblast Differentiation, and the Role of Adenosine Triphosphate-Sensitive Potassium Channels
Ischemic preconditioning significantly attenuated IR-induced fibroblast differentiation ( Figure  5). The IPC group had significantly more fibroblasts (52 ± 3%, (7)) compared to 60 min IR alone. These results suggest that while IR injury induced significant fibroblast-to-myofibroblast differentiation, IPC was able to protect against this effect. In time controls where cells were exposed to fresh DMEM in place of oil at each ischemic stage of the IPC/IR protocol, 72 ± 5% (4) of the cells were fibroblasts. . Ischemia reperfusion (IR) induced significant fibroblast to myofibroblast differentiation following 30, 60, and 120 min ischemia, relative to time controls (TCs). TCs are shown for 30, 60, and 120 min. The average percent of fibroblasts and myofibroblasts relative to the total number of cells analyzed are shown, ± SEM. * p < 0.05 between IR and TC for each time.

The Effects of Ischemic Preconditioning on Cardiac Fibroblast Differentiation, and the Role of Adenosine Triphosphate-Sensitive Potassium Channels
Ischemic preconditioning significantly attenuated IR-induced fibroblast differentiation ( Figure 5). The IPC group had significantly more fibroblasts (52 ± 3%, (7)) compared to 60 min IR alone. These results suggest that while IR injury induced significant fibroblast-to-myofibroblast differentiation, IPC was able to protect against this effect. In time controls where cells were exposed to fresh DMEM in place of oil at each ischemic stage of the IPC/IR protocol, 72 ± 5% (4) of the cells were fibroblasts. Pinacidil has been shown to non-selectively activate KATP channels in myocytes [8,25] and reduce infarct size [26], indicating a role for KATP channels in IPC protection in cardiac myocytes. Mitochondrial KATP channels were also implicated in this process in cardiac myocytes, as application of the mKATP channel opener Diaz also mimicked the effects of IPC [27]. While KATP channels have been demonstrated in cardiac fibroblasts [28][29][30], it is not known if they play a similar role in IPC in cardiac fibroblasts. We investigated the effects of Pin and Diaz on IR injury in cardiac fibroblasts in vitro, to determine if these potassium channel openers could mimic the effects of IPC and significantly attenuate IR-induced fibroblast differentiation.
Application of Pin (50 μM) in place of IPC prior to 60 min IR attenuated the differentiation of fibroblasts to myofibroblasts, compared to those cells exposed to 60 min IR alone (43 ± 6% (4) vs. 15 ± 3% (12) fibroblasts respectively, Figure 6) but did not cause a significant change. No significant differences were observed between the IPC and Pin-treated groups.
Similarly, Diaz significantly attenuated the IR-induced fibroblast to myofibroblast differentiation ( Figure 6). Pretreatment with Diaz (100 μM) in place of IPC prior to 60 min IR, was associated with cultures containing 53 ± 3% (4) fibroblasts, significantly more than in the IR group. The percentage of fibroblasts within the Diaz-treated cells was not significantly different than that of the IPC groups. Neither drug affected fibroblast differentiation when tested in the time controls: when Pin was added to cells and subjected to the same protocol as the IPC cells, but without ischemia, 63 ± 9% (4) of the culture were fibroblasts, compared to 72 ± 5% (5) of cells in the IPC time controls. With Diaz present but no ischemia, 56 ± 13% (3) of the population were fibroblasts. Pinacidil has been shown to non-selectively activate K ATP channels in myocytes [8,25] and reduce infarct size [26], indicating a role for K ATP channels in IPC protection in cardiac myocytes. Mitochondrial K ATP channels were also implicated in this process in cardiac myocytes, as application of the mK ATP channel opener Diaz also mimicked the effects of IPC [27]. While K ATP channels have been demonstrated in cardiac fibroblasts [28][29][30], it is not known if they play a similar role in IPC in cardiac fibroblasts. We investigated the effects of Pin and Diaz on IR injury in cardiac fibroblasts in vitro, to determine if these potassium channel openers could mimic the effects of IPC and significantly attenuate IR-induced fibroblast differentiation.
Application of Pin (50 µM) in place of IPC prior to 60 min IR attenuated the differentiation of fibroblasts to myofibroblasts, compared to those cells exposed to 60 min IR alone (43 ± 6% (4) vs. 15 ± 3% (12) fibroblasts respectively, Figure 6) but did not cause a significant change. No significant differences were observed between the IPC and Pin-treated groups.
Similarly, Diaz significantly attenuated the IR-induced fibroblast to myofibroblast differentiation ( Figure 6). Pretreatment with Diaz (100 µM) in place of IPC prior to 60 min IR, was associated with cultures containing 53 ± 3% (4) fibroblasts, significantly more than in the IR group. The percentage of fibroblasts within the Diaz-treated cells was not significantly different than that of the IPC groups. Neither drug affected fibroblast differentiation when tested in the time controls: when Pin was added to cells and subjected to the same protocol as the IPC cells, but without ischemia, 63 ± 9% (4) of the culture were fibroblasts, compared to 72 ± 5% (5) of cells in the IPC time controls. With Diaz present but no ischemia, 56 ± 13% (3) of the population were fibroblasts.

The Effects of Glibenclamide and 5-Hdroxydecanoate on Cardiac Fibroblast Differentiation
The nonselective KATP channel blocker Glib [12] and the selective mKATP channel blocker 5HD [9,31] have been shown to reduce the effectiveness of IPC in cardiac myocytes. We used Glib and 5HD to determine if blockade of KATP channels affected IPC in cardiac fibroblasts, as is seen in cardiac myocytes.
Both KATP blockers significantly reduced the effectiveness of IPC in preventing IR-induced fibroblast to myofibroblast differentiation (Figure 7). With 10 μM Glib present during IPC, the percentage of fibroblasts was reduced from 52 ± 3% (7, IPC) to 23 ± 5% (4, Glib). When cells were treated with 100 μM 5HD, only 22 ± 9% (3) of the culture were fibroblasts. There were no significant differences between the Glib-or 5HD-and IR-treated groups. Neither drug had any significant effect in the absence of ischemia. When used in time controls, Glib treatment was associated with 73 ± 2% (4) fibroblasts, while 5HD treatment was associated with 43 ± 15% (4) fibroblasts. These results were not significantly different from the IPC time control.

The Effects of Glibenclamide and 5-Hdroxydecanoate on Cardiac Fibroblast Differentiation
The nonselective K ATP channel blocker Glib [12] and the selective mK ATP channel blocker 5HD [9,31] have been shown to reduce the effectiveness of IPC in cardiac myocytes. We used Glib and 5HD to determine if blockade of K ATP channels affected IPC in cardiac fibroblasts, as is seen in cardiac myocytes.
Both K ATP blockers significantly reduced the effectiveness of IPC in preventing IR-induced fibroblast to myofibroblast differentiation (Figure 7). With 10 µM Glib present during IPC, the percentage of fibroblasts was reduced from 52 ± 3% (7, IPC) to 23 ± 5% (4, Glib). When cells were treated with 100 µM 5HD, only 22 ± 9% (3) of the culture were fibroblasts. There were no significant differences between the Glib-or 5HD-and IR-treated groups. Neither drug had any significant effect in the absence of ischemia. When used in time controls, Glib treatment was associated with 73 ± 2% (4) fibroblasts, while 5HD treatment was associated with 43 ± 15% (4) fibroblasts. These results were not significantly different from the IPC time control.

The Effects of Glibenclamide and 5-Hdroxydecanoate on Cardiac Fibroblast Differentiation
The nonselective KATP channel blocker Glib [12] and the selective mKATP channel blocker 5HD [9,31] have been shown to reduce the effectiveness of IPC in cardiac myocytes. We used Glib and 5HD to determine if blockade of KATP channels affected IPC in cardiac fibroblasts, as is seen in cardiac myocytes.
Both KATP blockers significantly reduced the effectiveness of IPC in preventing IR-induced fibroblast to myofibroblast differentiation (Figure 7). With 10 μM Glib present during IPC, the percentage of fibroblasts was reduced from 52 ± 3% (7, IPC) to 23 ± 5% (4, Glib). When cells were treated with 100 μM 5HD, only 22 ± 9% (3) of the culture were fibroblasts. There were no significant differences between the Glib-or 5HD-and IR-treated groups. Neither drug had any significant effect in the absence of ischemia. When used in time controls, Glib treatment was associated with 73 ± 2% (4) fibroblasts, while 5HD treatment was associated with 43 ± 15% (4) fibroblasts. These results were not significantly different from the IPC time control.  . Ischemic preconditioning (IPC) was reduced by blocking K ATP channels with the non-specific K ATP blocker glibenclamide (Glib) and the mK ATP -selective K ATP blocker 5-hydroxydecanoate (5HD). Blocking K ATP channels with Glib (10 µM) or 5HD (100 µM) prevented IPC from reducing ischemia reperfusion (IR)-induced fibroblast to myofibroblast differentiation. * p < 0.05, comparisons indicated by the lines.

Effects of Ischemia Reperfusion Injury and Ischemic Preconditioning on Fibroblast Differentiation into Immature Vs. Fully Mature Myofibroblasts
Upon differentiation, myofibroblasts progress through a continuum of α-SMA expression as they mature into fully differentiated myofibroblasts with well-developed stress fibers [32][33][34]. Fibroblasts and fully differentiated myofibroblasts have distinct physiology, in the spectrum of signaling molecules they secrete and in the amount of collagen they are capable of producing. In addition, the amount of force a myofibroblast may produce during scar contracture is directly proportional to the extent of α-SMA expression [32][33][34]. Accordingly, cultures were scored as to the expression of α-SMA, to determine if IR injury and IPC influenced myofibroblast maturation as well as fibroblast differentiation.
Ischemia of all durations followed by 60 min reperfusion was associated with differentiation of fibroblasts into a range of myofibroblasts with variable amounts of α-SMA-containing stress fibers ( Figure 8). For all durations of ischemia, the trend was for the myofibroblasts to be highly differentiated. Only in the 30 min IR condition was the percent of fully differentiated myofibroblasts (staining category "4") not larger than each of staining categories "1" through "3" (Figure 8A). In contrast, in all time controls, few myofibroblasts were present, and those tended to be less mature ( Figure 8).

Immature Vs. Fully Mature Myofibroblasts
Upon differentiation, myofibroblasts progress through a continuum of α-SMA expression as they mature into fully differentiated myofibroblasts with well-developed stress fibers [32][33][34]. Fibroblasts and fully differentiated myofibroblasts have distinct physiology, in the spectrum of signaling molecules they secrete and in the amount of collagen they are capable of producing. In addition, the amount of force a myofibroblast may produce during scar contracture is directly proportional to the extent of α-SMA expression [32][33][34]. Accordingly, cultures were scored as to the expression of α-SMA, to determine if IR injury and IPC influenced myofibroblast maturation as well as fibroblast differentiation.
Ischemia of all durations followed by 60 min reperfusion was associated with differentiation of fibroblasts into a range of myofibroblasts with variable amounts of α-SMA-containing stress fibers (Figure 8). For all durations of ischemia, the trend was for the myofibroblasts to be highly differentiated. Only in the 30 min IR condition was the percent of fully differentiated myofibroblasts (staining category "4") not larger than each of staining categories "1" through "3" (Figure 8A). In contrast, in all time controls, few myofibroblasts were present, and those tended to be less mature (Figure 8).
Ischemic preconditioning shifted the relative percentages of myofibroblasts toward less mature myofibroblasts ( Figure 9A), suggesting that not only were fewer fibroblasts differentiating, but maturation was delayed or prevented in the myofibroblasts that were produced. Application of Pin and Diaz had a similar effect ( Figure 9B), while preventing IPC with Glib was associated with a greater proportion of more mature myofibroblasts ( Figure 9C). A similar finding was observed when IPC was blocked by 5HD, though the trend toward mature myofibroblasts was not as pronounced as with Glib ( Figure 9C).  Ischemic preconditioning shifted the relative percentages of myofibroblasts toward less mature myofibroblasts ( Figure 9A), suggesting that not only were fewer fibroblasts differentiating, but maturation was delayed or prevented in the myofibroblasts that were produced. Application of Pin and Diaz had a similar effect ( Figure 9B), while preventing IPC with Glib was associated with a greater proportion of more mature myofibroblasts ( Figure 9C). A similar finding was observed when IPC was blocked by 5HD, though the trend toward mature myofibroblasts was not as pronounced as with Glib ( Figure 9C). significantly more fibroblasts (α-SMA staining intensity 0), and of those fibroblasts that did differentiate, the resultant myofibroblasts tended to express less α-SMA stress fibers.  1-3). These trends were also observed when IPC was mimicked by opening K ATP channels with pinacidil (Pin) (B) or diazoxide (Diaz) (B). In contrast, blocking K ATP channels with either glibenclamide (Glib) (C) or 5-hydroxydecanoate (5HD) (C) was associated with fewer fibroblasts (α-SMA staining intensity 0) than following IPC (B,C).

Discussion
Cardiac fibroblasts are both functionally and phenotypically different from the classically studied cardiac myocytes [17,18,32]. The role of cardiac fibroblasts and myofibroblasts in the development of fibrosis following myocardial ischemia makes them potential candidates to unravel the pathophysiology of ischemic heart disease. Until now, the effects and mechanisms of IR injury and IPC on cardiac fibroblast to myofibroblast differentiation have not been studied. We modified the paraffin oil method of simulating IR injury in order to investigate the role of IR injury, IPC and K ATP channels in adult rat ventricular cardiac fibroblasts.

Characterization of the Ischemic Conditions
To validate the use of paraffin oil as an effective method of simulating ischemia in vitro, we measured the pH, PO 2 and PCO 2 at 30, 60 and 120 min of ischemia, and the recovery of these parameters to pre-ischemic levels at 5, 15-and 30-min reperfusion. Our results demonstrate that the application of paraffin oil resulted in a significant increase in the PCO 2 , decrease in the PO 2 and acidosis, all of which recovered rapidly following removal of the paraffin oil ( Figure 3). In vivo studies have demonstrated that myocardial ischemia results in increased PCO 2 , decreased PO 2 [35] and acidosis [36]. Our results are consistent with ischemic conditions in vivo and suggest that paraffin oil is a valid method of simulating ischemic conditions in vitro. However, not all elements of ischemia and reperfusion in vivo can be mimicked by this in vitro assay. The in vivo tissue architecture and gap junctional communication are lost in the dissociation process. Reperfusion in vivo is associated with activation of neutrophils and free oxygen radical production by neutrophils and endothelial cells [37]; in vitro assays may lack these elements.

Ischemia Reperfusion Injury and Ischemic Preconditioning in Cardiac Fibroblasts
To the best of our knowledge, this is the first study to demonstrate that cardiac fibroblasts differentiate into myofibroblasts in response to IR injury. This is also the first study to demonstrate that IPC ameliorates the IR injury-induced differentiation of cardiac fibroblasts into myofibroblasts. Few other studies have examined cardiac fibroblasts following ischaemia and reperfusion, or hypoxia and reoxygenation. Vivar et al. [38] studied IR injury-induced death and the protective effect of insulin-like growth factor 1 in cultured neonatal rat cardiac fibroblasts. Zhou et al. [39,40] established a model of cultured rat neonatal and adult cardiac fibroblasts, in which they have considered the deleterious effects of hypoxia and reperfusion, comparing biochemical and morphological changes in cultured fibroblasts to those of cultured ventricular myocytes [39]. They also found that conditioned media from cultured cardiac fibroblasts subjected to hypoxia and reoxygenation did not protect cardiac myocytes from IR damage, while factors from other non-myocytes of mesenchymal origin did [40]; however, these authors did not consider IR injury beyond changes associated with fibroblast death, or IPC, in their model. Lefort et al. [41] reported that cultured human ventricular fibroblasts produced secretomes in response to 5 h of hypoxia and 24 h reoxygenation, which reduced cardiac myocyte death during the hypoxia/ reoxygenation challenge. These authors also reported that stimulation of the metabotropic purine P2Y11 receptor in cultured human ventricular fibroblasts at the onset of reoxygenation reduced fibroblast to myofibroblast differentiation [41], suggesting that Gq and Gs protein-coupled pathways modulate fibroblast differentiation in response to hypoxia and reoxygenation.
Ischemic preconditioning may be an effective therapeutic strategy not only to protect the heart against myocyte injury and death, but also to prevent the development of inflammation and fibrosis following myocardial infarction. Hypoxia/reoxygenation-induced secretome release from cultured human ventricular fibroblasts was associated with a pro-inflammatory response; this effect was diminished if P2Y11 receptors were activated during reoxygenation [41]. P2Y11 receptor activation also reduced fibroblast to myofibroblast differentiation in this model [42], similar to our findings ( Figures 5-9). This may be significant when considering the degree of fibrosis associated with IR injury.
Fibrosis is also strongly correlated with fibroblast to myofibroblast differentiation [31,32] and contributes to the pathogenesis of arrhythmias and heart failure [17]. Furthermore, fibroblasts have been shown to protect the myocardium via the production of 'currently undefined' substances [43]. Therefore, preserving fibroblasts in the infarct zone following an ischemic insult will not only help to reduce the number of myofibroblasts in the heart, limiting reactive fibrosis [17], it may also help to protect the myocardium against further IR injury by other mechanisms. In addition, specific cardiac fibroblast G-protein-coupled receptor kinase 2 (GRK2) knockout in mice was shown to reduce infarct size, degree of fibrosis and inflammation following IR injury [44]. These results suggests that upregulation of GRK2 contributes to fibrosis, inflammation, and myocyte death through fibroblast-specific actions. It would therefore be interesting to determine the effect of IPC on GRK2 activation in cardiac fibroblasts.

The Role of Adenosine Triphosphate-Sensitive Potassium Current in Preventing Fibroblast to Myofibroblast Differentiation
The activation of K ATP channels occurs during IPC in myocytes and has been shown to protect against IR injury [28][29][30]. Adenosine triphosphate-sensitive K channels have recently been described in cardiac fibroblasts [28][29][30].
We investigated whether K ATP channels may mediate the protective effect of IPC in cardiac fibroblasts. Our results demonstrated for the first time that the opening of pK ATP and mK ATP channels with Pin and selectively opening mK ATP channels with Diaz mimicked the effects of IPC and attenuated IR-induced fibroblast differentiation ( Figure 6). Furthermore, our results indicate that inhibition of pK ATP and mK ATP channels with Glib and selective inhibition of mK ATP channels with 5HD abolished IPC-induced protection against fibroblast differentiation. These results suggest that mK ATP channels are activated during IPC to prevent fibroblasts from differentiating into myofibroblasts.

Involvement of Mitochondrial Vs. Sarcolemmal Adenosine Triphosphate-Sensitive Potassium Channels in Ischemic Preconditioning in Fibroblasts
Activation of sK ATP channels in cardiac myocytes causes sarcolemmal hyperpolarization at rest [8]. This current also enhances membrane repolarization, shortening the action potential duration and limiting voltage-gated Ca 2+ current [12]. This action leads to reduced Ca 2+ entry into cardiac myocytes during ischemia [13].
Cardiac fibroblasts do not appear to be excitable cells [19,20,[45][46][47]. While immunohistochemical [48] and pharmacological studies [48][49][50] have suggested that L-type Ca 2+ channels may play a role in cardiac fibroblasts and myofibroblasts, none of the electrophysiological patch clamp studies on cardiac fibroblasts or myofibroblasts have provided evidence of voltage-gated Ca 2+ currents [19,20,[28][29][30][45][46][47][51][52][53][54][55][56][57][58][59]. In the absence of voltage-gated Ca 2+ currents, hyperpolarization due to pK ATP current would not be predicted to reduce Ca 2+ entry through the plasmalemma, as it would for myocytes. Conversely, as non-selective cation conductances appear to be present [48,54], hyperpolarization would be predicted to increase Ca 2+ entry through these channels. Thus, we hypothesized that cardioprotection from the activation of K ATP channels in cardiac fibroblasts, arises primarily from the activation of mK ATP channels. In fibroblast as well as myocyte mitochondria, activation of mK ATP current is predicted to reduce Ca 2+ current through the Ca 2+ uniporter and prevent Ca 2+ overload [10,11]. Hence, as K ATP channels appear to be an end-effector of signaling in IPC, it is reasonable to hypothesize that the mitochondrial channel is preferentially activated in cardiac fibroblasts. This possibility is further supported by evidence that Wistar rat left-ventricular fibroblasts do not strongly express pK ATP currents [30]. Intriguingly, Benamer et al. [30] found that K ATP currents were larger in fibroblasts isolated from infarction scars and from the border zone than in fibroblasts from the non-infarcted regions of the infarcted hearts, or from non-infarcted control hearts. These authors did not determine if the fibroblasts may have differentiated into myofibroblasts in the infarction and border zones. If so, this suggests that calcium entry into myofibroblasts may be enhanced due to pK ATP current, in the infarction scar and border zone. The significance of such calcium entry is not currently known.
Our results demonstrated that activating mK ATP channels mimicked the effects of IPC by reducing fibroblast differentiation, while blocking mK ATP channels attenuated the effects of IPC and resulted in increased fibroblast differentiation (Figure 7). The magnitude of selective activation of mK ATP channels with Diaz was not significantly different from non-selective activation with Pin, suggesting that the primary effect was mitochondrial. Similarly, non-selective blockade with Glib had the same effect as selectively blocking mK ATP channels with 5HD, suggesting that the primary effect was inhibition of the mitochondrial channel. These results support our hypothesis that mK ATP channels are important in the development of IPC-mediated protection against IR-induced fibroblast differentiation.

Immature vs. Fully Mature Myofibroblasts
Myofibroblasts vary in the amount of α-SMA they express, and the corresponding degree of mature stress fibers formed in these contractile cells [52]. In addition, when compared to fibroblasts, myofibroblasts secrete a distinct range of signaling molecules, and are much more rapidly capable of remodeling the extracellular matrix (ECM) [17,32,34]. As myofibroblasts may be associated with maladaptive fibrosis in the heart [17], the degree to which the myofibroblasts mature following IR injury is of clinical interest.
We found that IR injury was associated with more mature myofibroblasts, and that IPC tended to prevent or delay myofibroblast maturation (Figures 8 and 9). In addition, modulation of K ATP channel recruitment influenced myofibroblast maturation: treatment with Pin and Diaz tended to prevent or delay maturation, while Glib and 5HD tended to promote it following IR injury. These results suggest that IPC may be important not only in reducing the total number of myofibroblasts being generated, but also in delaying or preventing maturation. Both effects would be predicted to reduce fibrosis in the wounded heart, and so, to reduce mechanical dysfunction and the risk of arrhythmias.
While to our knowledge, this is the only study of IR injury and IPC in cardiac fibroblasts, the effects of persistent hypoxia and ischemia in skin wound healing have been investigated. In a rat model of sustained hind limb ischemia, a wound on the ischemic foot failed to heal as well as a matching wound on the non-ischemic control foot [60]. This was attributed to delayed production and maturation of myofibroblasts within the granulation tissue, preventing wound closure and contracture [60]. Hypoxia was found to have a similar effect [61,62]. These studies lead to the hypothesis that during ischemia, cardiac fibroblasts may not as readily differentiate into myofibroblasts, and that it is during reperfusion that the differentiation is triggered in the absence of K ATP channel activation by IPC. This is in agreement with the hypothesis that the mitochondrial isoform of this channel is the primary one at work in the fibroblast, as it is during reperfusion that protection of the inner mitochondrial membrane potential is important to prevent Ca 2+ overload [10,11]. Like myocytes, cardiac myofibroblasts are believed to express the Na + /Ca 2+ exchanger [48]. The Na + /H + exchanger NHE1 has been demonstrated in cardiac fibroblasts [63]. Only in reperfusion is there a pH gradient which affects the Na + /H + and Na + /Ca 2+ exchangers, increasing cytosolic and subsequently, mitochondrial Ca 2+ levels [10]. Hence, it is during reperfusion that activation of mK ATP channels is likely to have a protective effect in preventing fibroblast differentiation. The role of membrane potential in fibroblast to myofibroblast differentiation has not been studied; however, modulation of membrane potential is known to affect migration [24,48], contraction and proliferation [19,48] of myofibroblasts.

Important Considerations
One caveat of research concerning mK ATP channels is that the existence of these channels has not been conclusively demonstrated (reviewed in [8]). While currents from these channels have been recorded in rat liver mitochondrial inner membrane [7], these authors could not be absolutely certain that their mitochondrial preparations were free of contamination by the plasma membrane. Other indirect studies of mitochondrial function have provided supportive evidence of this channel's existence and that it plays a role in protecting the mitochondria from Ca 2+ overload during IR injury, but the molecular identity of this channel is still unknown (reviewed in [8]). In our studies, we have been careful to apply the selective mK ATP blockers and activators at appropriate concentrations [25,64] to avoid non-selective effects. At present, the technological constraints of recording K ATP currents from the inner membranes of individual mitochondria isolated from cardiac fibroblasts prevent us from providing direct evidence that IPC and the selective mK ATP drugs modulated mK ATP channels in these cells, as we suggest. Hence, the role and nature of K ATP channels within the mitochondria during IPC in cardiac fibroblasts are to be treated with due caution, as they are in cardiac myocytes [8].
Isoflurane is known for its ability to open K ATP channels and exert cardioprotective effects on the heart [65,66]. Our animals were only briefly (<10 min) exposed to isoflurane during the isolation procedure. Isoflurane-induced IPC was not observed after 15 min exposure in mice [66], suggesting that the duration of exposure in our studies was unlikely to evoke IPC. Our results also suggest that our cardiac fibroblasts were not under the protection of isoflurane IPC, as IR injury caused significantly more fibroblast differentiation than IPC or matched time controls. It would be interesting, however, to investigate the effects of isoflurane and other volatile gases on IR injury and cardiac fibroblast differentiation.

Conclusions
Using an in vitro model of IR injury, we demonstrated that cardiac fibroblasts differentiate into myofibroblasts in response to IR-injury. Furthermore, we showed that IPC reduced the amount of IR-induced differentiation. We are the first to demonstrate that the activation of K ATP currents mimic the protection of IPC and protect against IR-induced fibroblast to myofibroblast differentiation, and that blockade of these channels reduces the effectiveness of IPC. While we are aware that the molecular identity of mK ATP is not yet known [8], we believe that the electrophysiological evidence [7] supports the existence of this channel. Our data therefore suggest that IPC is mediated by specific activation of the mK ATP in cardiac fibroblasts. This research will hopefully help in future development of new therapeutic treatments for post-myocardial infarction fibrosis. Funding: This research received no external funding. The authors are grateful to James Cook University for the internal research funding which supported this research.