Anti-apoptotic potential of several antidiabetic medicinal plants of the eastern James Bay Cree pharmacopeia in cultured kidney cells

Background Our team has identified 17 Boreal forest species from the traditional pharmacopeia of the Eastern James Bay Cree that presented promising in vitro and in vivo biological activities in the context of type 2 diabetes (T2D). We now screened the 17 plants extracts for potential anti-apoptotic activity in cultured kidney cells and investigated the underlying mechanisms. Methods MDCK (Madin-Darnby Canine Kidney) cell damage was induced by hypertonic medium (700 mOsm/L) in the presence or absence of maximal nontoxic concentrations of each of the 17 plant extracts. After 18 h’ treatment, cells were stained with Annexin V (AnnV) and Propidium iodide (PI) and subjected to flow cytometry to assess the cytoprotective (AnnV−/PI−) and anti-apoptotic (AnnV+/PI−) potential of the 17 plant extracts. We then selected a representative subset of species (most cytoprotective, moderately so or neutral) to measure the activity of caspases 3, 8 and 9. Results Gaultheria hispidula and Abies balsamea are amongst the most powerful cytoprotective and anti-apoptotic plants and appear to exert their modulatory effect primarily by inhibiting caspase 9 in the mitochondrial apoptotic signaling pathway. Conclusion We conclude that several Cree antidiabetic plants exert anti-apoptotic activity that may be relevant in the context of diabetic nephropathy (DN) that affects a significant proportion of Cree diabetics.


Background
Diabetes mellitus (DM), first officially recorded in ancient Egypt, was initially considered as a rare condition in which patient urinated excessively and lost weight [1]. Clinically, DM refers to a group of metabolic diseases in which there is a chronic hyperglycemic condition as a result of defects in insulin secretion, insulin action or both [2,3]. Due to compromised insulin secretion or sensitivity, postprandial metabolism is disturbed, resulting in less efficient uptake and utilization of glucose by the targeted tissues, further leading to elevated blood glucose level, more utilization of proteins and fats [4].
The prevalence of DM is rapidly increasing worldwide as a result of sedentary lifestyle, obesity, ageing, smoking, high blood pressure, genetic predisposition, psychiatric disorder such as depression and so on [5]. DM is predicted to affect 366 million individuals by the year 2030 [5][6][7][8]. Notably, Aboriginal populations worldwide are disproportionally affected by T2D. This has been linked to rapid environmental changes and to potential genetic predisposition to higher conservation of food calories [9][10][11][12][13]. For instance, the age-adjusted prevalence of T2D in the Cree populations of Eeyou Istchee (CEI -Eastern James Bay area of Quebec, Canada) averaged 29% in 2009 [11,14]. Cree communities also suffer from higher prevalence of diabetic complications, notably DN. This is due at least in part to the cultural inappropriateness of modern drug treatments [15,16].
In western countries, end-stage renal disease (ESRD) is mainly caused by DN [17]. Development of proteinuria, considered as the key feature of DN, is related to the decline of glomerular filtration rate (GFR), eventually leading to a progressive loss of renal function [18]. Hypertension and poor glycaemic control are usually associated with DN [19,20]. More importantly and according to numerous recent studies, it is well established that renal tubular cell apoptosis contributes to the development of DN, leading to gradual loss of renal mass [21][22][23]. In this context, different cell models can be used to study renal cell apoptosis, such as MDCK cells subjected to hypertonic stress [24].
The Canadian Institutes of Health Research Team in Aboriginal Antidiabetic Medicines (CIHR-TAAM) was instated in 2003 in an effort to find culturally relevant complementary and alternative approaches to T2D prevention and management for Canadian Aboriginal diabetics. Seventeen plant species stemming from the James Bay Cree traditional pharmacopeia were identified through ethnobotanical surveys and tested using a comprehensive platform of bioassays and animal models of obesity and diabetes to identify the plants' capacity to improve glycemic control [11,[25][26][27][28]. Cree community members also encouraged us to study the potential of the plants to protect kidney cells, since DN is quite prevalent in the James Bay area [15,16].
In current study, we therefore sought to determine the activity of the same 17 plant extracts to afford renal protection and, hence, their potential to mitigate DN. To achieve this, we developed a bioassay based on well-known MDCK cells, a kidney cell of distal tubule origin, that we stressed with hypertonic medium to induce apoptosis [24,[29][30][31][32][33]. We then used flow cytometry and staining reagents for apoptosis (AnnV) and necrosis (PI). Finally, we assessed the role of several caspases in order to begin understanding mechanisms underlying the anti-apoptotic activity of certain plant species.

Plant extract preparation
A total of 17 Cree medicinal plant species were the object of the current study (Table 1) and their respective maximal nontoxic concentrations (see below) used in MDCK cell line are listed in Table 2. Each species of the 17 plant samples were collected in CEI territory and prepared from air-dried and ground plant material according to previously published methods [27,28]. Authorization for plant sample collection was obtained and managed by a comprehensive research agreement convened between the three Canadian universities involved (Université de Montréal, McGill University, University of Ottawa), the participating Cree First Nations and the Cree Board of Health and Social Services of James Bay [34]. Dr. Alain Cuerrier, a seasoned taxonomist, ascertained the botanical identity of the plant species and voucher specimens have been deposited at the herbarium of the Montreal Botanical Garden [11]. The collected plant samples were extracted with a standard 80% aqueous ethanol protocol as previously described [27,28]. All the plant extracts have been well characterized in terms of their phytochemical content in previous studies from our laboratory and that of others (shown in Table 1) . All extracts were freezedried overnight using a Super Moudylo freeze-dryer (TheromFisher, Brockville, ON, Canada). They were then stored at 4°C in amber containers in a dessicator, both of which were flushed free of oxygen. In such conditions, extracts maintained a stable phytochemical profile and biological activity for several years, as tested using HPLC analyses with different detection systems (e.g. diode array and mass spectrometer) and using several cellbased bioassays.
Stock solutions of each of the 17 plant extracts were prepared in DMSO with concentrations ranging from 5 to 200 mg/mL. They were diluted in culture medium to the working concentrations shown in Table 2 (the final concentration of DMSO was 0.1% for all the treatments).

Determination of maximal nontoxic plant extract concentrations
Before screening the plants, maximal nontoxic concentrations were determined by cytotoxicity test that reflects the level of lactate dehydrogenase (LDH) release (LDH Colorimetric kit; Roche, Mannheim, Germany). Cells were seeded at the density of 1.5 × 10 5 /well on 6-well plates. Medium was refreshed around 20 h later when 70% confluence was reached, followed by the addition of each of 17 plant extracts with a series of concentrations, respectively. The range of concentrations selected was based on our previous experience with the 17 plant extracts in other cell lines [11,27,28]. After 18 h' incubation, Results were analyzed and used to determine the optimal nontoxic concentration for each extract. Each experiment was repeated three times. Hypertonic stress protocol To screen the 17 plant extracts, cells were seeded in 6well plates at the density of 1.5 × 10 5 /well. Around 20 h later, 70% subconfluence was reached. Cells were then switched for 18 h to 700 mOsm/L hypertonic (made by addition of 200 mM sodium chloride to EMEM culture medium) or isotonic EMEM medium without FBS [24], which contained or not each of the 17 plant extracts at their respective nontoxic maximum concentration. Z-VAD-FMK, a pan caspase inhibitor, was selected as cytoprotective positive control.

Flow cytometric analysis of AnnV/PI staining
Apoptosis assays were performed using AnnV-FITC (fluorescein isothiocyanate, BD Bioscience, Mississauga, ON) and PI (Thermo Fischer Scientific). 5 mL polystyrene tubes (BD Bioscience, Mississauga, ON)) placed in a laminar flow hood were labeled and preloaded with 500 μl FBS prior to assay. After 18 h' treatment, medium was transferred and collected in prepared polystyrene tubes; adherent cells were washed once with 1 mL PBS then harvested with 250 μl 0.25% trypsin followed by another 1 mL PBS wash; all polystyrene tubes were subjected to centrifugation at 250 g, 4°C, 5 min. Subsequently, supernatant was discarded and cell pellets resuspended in 500 μL of ice cold binding buffer (10 mM Hepes pH 7.4, 150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 ). Cell suspensions were then mixed with 2 μL of AnnV-FITC (AnnV) and 1 μL of PI. After incubation on ice for 5 min in the dark, healthy and damaged cells were determined and analyzed using a FACSCanto II (BD, Biosciences, Bedford, MA, USA) flow cytometer and FlowJo software. Positive staining was confirmed under fluorescence microscopy. Each experiment was repeated three times. As illustrated in the Results section below, PI + Quadrants Q1 and Q2 respectively represent necrosis and late stage apoptosis/secondary necrosis; Q4 Quadrant represents viability (AnnV − /PI − ); Q3 Quadrant (AnnV + /PI − ) represents early stage apoptosis.
Flow cytometry analysis of caspase 3, 8, 9 As described in the Results section, we selected a subset of plant extracts to study potential underlying mechanisms of MDCK cell protection. Consequently, caspase 3, 8, 9 assays were performed according to manufacturer's protocol by using the CaspGLOW™ Fluorescein Active Caspase Staining Kit (eBioscience, San Diego, CA, USA). Briefly, cells were detached as described above for AnnV and PI staining before CaspGLOW™ Fluorescein Active Caspase staining was applied. After centrifugation at 250 g, 4°C, 5 min, 300 μL EMEM containing 1 μL CaspGLOW™ Fluorescein Active Caspase was added to each tube, then transferred in a 37°C incubator with 5% CO 2 for 45 min. Subsequently, cells were subjected three times to a course of wash and spin-down (centrifugation at 500 g, 5 min) at room temperature. Finally, all tubes were subjected to flow cytometry for analysis of respective caspase staining by FACSCanto II (BD, Biosciences, Bedford, MA, USA) flow cytometer and FlowJo software. This assay utilizes inhibitors specific for cleaved caspase-3, − 8 and − 9, respectively. They are directly conjugated to FITC for the detection system. These reagents are cell permeable, non-toxic and irreversibly bind to the cleaved caspase 3, 8, and 9, respectively. Detection of the labeled cells can be determined by flow cytometry.

Statistical analysis
Results are presented as mean ± SEM of 3 independent experiments with triplicate for each sample. Statistical comparisons between the experimental groups were analyzed by one-way ANOVA and the Bonferroni test as appropriate in software Prism 6 (GraphPad Software Inc., San Diego, CA, USA). A P value below 0.05 was considered statistically significant.

Determination of maximal nontoxic plant extract concentrations
MDCK cells were treated with 17 plant extracts at different concentrations, respectively. After an overnight incubation (18 h), LDH release was measured. Maximal nontoxic concentrations were determined and results are listed in Table 2. At concentrations listed, all plant extracts exhibited LDH release values that did not exceed 5-10% beyond values obtained with DMSO alone.
In vitro screen of renal protective potential of 17 antidiabetic medicinal plant extracts and analysis by flow cytometry In order to identify the respective renal protective and anti-apoptotic potential of the 17 plant extracts in vitro, AnnV/PI double staining was performed. Examined by fluorescent-based flow cytometry and compared to isotonic EMEM (Fig. 1a), 700 mOsm/L hypertonic stress (Fig. 1b) induced substantial PI + (necrosis and late stage apoptosis/secondary necrosis, Q1 and Q2 Quadrants) as well as AnnV + /PI − (early stage apoptosis, Q3 Quadrant) signals in MDCK cells. We also used Z-VAD-FMK as a positive cytoprotective control. It is a commonly recognized pan caspase inhibitor, exerting its protective role by irreversibly binding to the catalytic site of caspase proteases [85]. Accordingly, Z-VAD-FMK efficiently protected MDCK cells from the deleterious effects of hypertonic stress (Fig. 1c). In the presence of such 700 mOsm/L hypertonic stress (itself alone resulting in only 53.8% AnnV − /PI − staining -viable cells), all plant extracts, except S. purpurea and K. augustifolia, variably yet significantly improved cell viability as assessed by the change in AnnV − /PI − staining (Fig. 1d). Among the 17 plant extracts, G. hispidula, S. planifolia and A. balsamea were most active, respectively increasing AnnV − /PI − staining (viable cells) by 33.5%, 27.7% and 26.6%, relative to hypertonic stress alone (P < 0.0001, Fig. 1d). In addition, the difference in AnnV − /PI − staining observed in the G.
We then sought to carry out further analysis to begin understanding potential underlying mechanisms, notably in relation to renal cell apoptosis. Based on the results presented above, we therefore selected a subset of five species. We selected G. hispidula and A. balsamea as representatives of the best plant species to improve AnnV − /PI − staining and reduce AnnV + /PI − staining. R. tomentosum and R. groenlandicum were selected as species that had moderate effects and S. purpurea as one of the least effective plant species.

Mechanism investigation Cleaved caspase 3 activity
The activity of cleaved (active) caspase 3 following various treatments was then assessed by flow cytometry analysis (Fig. 2, panels a-d). The proportion of cells exhibiting staining for active caspase 3, a hallmark of apoptosis, was significantly increased in the hypertonic stress treated group (Fig. 2b). In comparison, the isotonic EMEM group (Fig. 2a) had a proportion of caspase 3 positive cells that was 48.3% lower than hypertonic stress (P < 0.01, Fig. 2a-e). As a positive control, pan caspase inhibitor Z-VAD-FMK largely suppressed the hypertonically induced activity of cleaved caspase 3 (71.2% lower than hypertonic stress, P < 0.001, Fig. 2c and e), this effect reaching levels even lower than the isotonic EMEM group (Fig. 2e). Among the representative plant species selected from our primarily screen, a statistically significant reduction in activity of cleaved caspase 3 was observed for A. balsamea (Fig. 2d and e) and R. tomentosum (59.8% and 42.3% lower than hypertonic stress, respectively, P < 0.01, Fig. 2e). G. hispidula and R. groenlandicum afforded weak protection against cleaved caspase 3 activation induced by hypertonic stress; their caspase 3 activity lying between isotonic EMEM and hypertonic stress group (N.S. versus either group, Fig. 2e). In contrast, S. purpurea exhibited more cleaved caspase 3 activity compared to hypertonic stress treated group (44.5% higher than hypertonic stress, P < 0.001, Fig. 2e). Aside from R. tomentosum, these results are in accordance with the rank order of cytoprotection observed in our primarily screen.

Cleaved caspase 9 activity
We finally analyzed the activity of cleaved caspase 9, indicative of the mitochondrial apoptotic pathway, using fluorescent staining and flow cytometry (Fig. 4, panels a-d).

Discussion
The aim of the current study was to begin evaluating the nephroprotective capacity of 17 Cree medicinal plants that were identified as having significant antidiabetic potential in several bioassays related to glucose and lipid homeostasis as well as in animal models of obesity and diabetes [26,86]. As mentioned, our CEI partners highlighted the fact that their population was suffering disproportionately from DN [34,87] and wanted us to look into renal protection stemming from their traditional medicinal plants.
We therefore selected the MDCK cell line, which is a very well characterized renal tubular cell model that can serve to assess cytoprotection against various insults [24,[29][30][31][32][33]. We also selected hypertonic stress to cause cell death, notably involving apoptosis [24].
In preliminary experiments, we challenged MDCK cells with different medium osmolarities (including 300, 400, 500, 600, 700 and 800 mOsm/L hypertonic medium, data not illustrated) and found that 700 mOsm/L was the ideal concentration to induce substantial apoptosis, which could be used to test the anti-apoptotic potential of the 17 plant species. This also corresponded with the concentration used successfully by other researchers in the same cell line [24]. This is pertinent in the context of the present studies, since renal tubular apoptosis is recognized as a major contributor for the development of DN [21][22][23]. Our results with AnnV/PI double staining and flow cytometry clearly demonstrate that a 700 mOsm/L hypertonic stress resulted in significant increases in cell death (PI + quadrants Q1 and Q2; respectively representing necrosis and late apoptosis). Hypertonic stress also significantly enhanced the proportion of cells with AnnV + /PI − staining, indicative of early apoptotic cell damage.
Also, importantly, we chose the pan caspase inhibitor Z-VAD-FMK as a positive cytoprotective control introduced in the hypertonic medium. It was very efficient in returning the pattern of AnnV/PI staining to that seen in isotonic EMEM (no hypertonic stress). Thus, our MDCK cells provide an adequate model in which to screen for potential nephroprotective and anti-apoptotic activities of Cree antidiabetic medicinal plants.
When MDCK cells were treated with the various antidiabetic Cree plants, we obtained a wide range of effects; a number of plant extracts almost completely prevented the deleterious effects of hypertonic stress, several were moderately cytoprotective, a few were weakly so and one plant actually enhanced nephrotoxicity, namely K. angustifolia. Notably, G. hispidula and A. balsamea plant extracts were amongst the best performers in maintaining high viability and suppressing apoptosis, being as powerful as Z-VAD-FMK. As mentioned, Z-VAD-FMK is a pan-caspase inhibitor that acts against apoptosis by irreversibly binding to the catalytic site of caspase proteases. As also discussed, apoptosis is involved in cell damage caused by hypertonic stress and is implicated in DN [21][22][23][24]. We therefore sought to determine the potential role that caspases could play in the variable cytoprotective activity that we observed for the Cree plants.
Caspases (cysteine-aspartic proteases or cysteinedependent aspartate-directed proteases) belong to a family called cysteine proteases. They play an essential role in cell apoptosis or programmed cell death and have thus been termed "executioner" proteins [88,89]. Caspases are regulated at a post-translational level, which allows them to be activated rapidly. Caspases are synthesized as inactive preforms and are cleaved next to aspartate residues upon activation [90]. There are socalled initiator and effector caspases. Initiator caspases possessed specific domains not encountered in effector caspases, such as caspase activation and recruitment domains (CARDs) (e.g., caspase-2 and caspase-9) or a death effector domain (DED) (caspase-8 and caspase-10). These ensure that the caspases can interact with other molecules that regulate their activation. These regulating molecules receive signals from extracellular and intracellular stimuli and interact with initiator caspases, causing their clustering. Such clustering allows initiator caspases to auto-activate and to proceed to activate effector caspases, eventually leading to the amplification of caspase activity through a protease cascade [90], considered as a positive feedback [91]. As mentioned, the activation of caspases can be initiated from extracellular stimuli involving death receptors on the plasma membrane (receptor pathway) or through intracellular stimuli centered in mitochondria (mitochondrial pathway) [92].
Death receptor stimulation activates procaspase-8, whereas the mitochondrial pathway involves the release of cytochrome c and other factors that activate procaspase-9 [24,92]. Both signaling pathways will eventually lead to downstream activation of caspase-3, which is a major effector in the caspase cascade. Since caspase-3 serves as a convergence point for different signaling pathways, it is therefore well suited as a read-out in an apoptosis assay. In the current studies, we used fluorescent substrates of these three major caspases to probe their activation by flow cytometry. Our results clearly show an increased activity of cleaved caspase 3, 8 and 9 activities in MDCK cells subjected to a 700 mOsm/ L hypertonic stress for 18 h. Moreover, the proportion of caspase 8 and 9 "positive" cells was 22.5% and 23.4%, respectively, which indicates that the death receptor and mitochondrial pathways were similarly activated in cells challenged by hypertonic medium.
When we studied the five plant species that were selected to represent strong, medium and low cytoprotective potential in AnnV/PI assays, their rank order of inhibition of hypertonic stress-induced cell damage was mostly maintained for cleaved caspases 3, 8 and 9, except for G. hispidula that was ineffective at reducing the activity of caspases 3 and 8, and R. tomentosum that was unexpectedly effective in reducing the activity of caspases 3 and 9.
Rather limited knowledge is available regarding G. hispidula. In previous studies from our group, the plant extract exerted some cytoprotective potential in preneuronal cells subjected to hyperglycemic stress [27] and moderately stimulated AMPK in cultured hepatocytes [11]. The current studies uncovered a strong cytoprotective potential for renal tubular cells as observed through improved viability (annV − /PI − staining) and reduced early stage apoptosis (annV + /PI − staining). However, as mentioned G. hispidula did not succeed in significantly modulating the activities of caspases 3 and 8 in hypertonically stressed MDCK cells. This suggests that the nephroprotective activity of the plant extract may occur principally through the mitochondrial pathway (significant reduction of cleaved caspase 9) or through pathways other than the classic apoptotic signaling pathways (death receptor or mitochondrial) [93].
Meanwhile, the non-cytoprotective plant S. purpurea actually enhanced the activity of all cleaved caspases in hypertonically stressed MDCK cells. Hence, despite the fact that S. purpurea previously exhibited a positive impact on muscle cell glucose uptake [34] and, notably, cytoprotection of pre-neuronal cells (potential benefit in diabetic neuropathy [37]), the plant appears to be potentially more harmful to renal cells. Indeed, Elders have cautioned us that the pitcher plant needed to be used carefully for it is considered very powerful.
Another of the weaker nephroprotective plants in our Ann V/PI assay, namely R. groenlandicum, appeared to increase the activation of caspase 8 compared to hypertonic stress whereas it had no apparent effect on cleaved caspase 3 and caspase 9 activities. This result is surprising since Labrador tea (R. groenlandicum) was previously observed by our group to exhibit several beneficial antidiabetic activities when tested in both in vitro bioassays [36,94] and in vivo animal models of obesity and mild diabetes [95,96]. Such antidiabetic activities would be expected to improve renal function (for instance, through reduction of glycemia and improvement of insulin resistance). In fact, we observed improved micro-albuminuria, reduced fibrosis and diminished expression of Bcl2-modulating factor (Bmf ) in renal tissues of diet-induced obese mice treated with R. groenlandicum in vivo [95]. Interestingly, in the current studies, R. tomentosum, a close relative of R. groenlandicum also known as Northern Labrador tea, exhibited significant anti-apoptotic potential that was expressed more potently through the suppression of the mitochondrial apoptotic pathway than that of the death receptor pathway. Further studies will thus be necessary to ascertain the complete impacts of Labrador or Northern Labrador tea consumption on the kidney.
In contrast, A. balsamea clearly stands out as one of the most powerful renal protective Cree plants that exhibited important anti-apoptotic activities, especially at the level of caspases 3 and 9. Our previous work with A. balsamea extracts demonstrated that they can significantly enhance basal and insulin stimulated glucose uptake in cultured skeletal muscle cells and adipocytes [28]. The plant was also the most powerful of Cree species to mitigate liver cell glucose production mechanisms in vitro through both insulin-dependent and insulin-independent mechanisms [86]. In skeletal muscle cells, A. balsamea exerted its action through a mechanism similar to that of metformin, involving the activation of AMPK secondary to metabolic stress induced by the disruption of mitochondrial energy transduction (energy depletion), albeit with only mild effects on cell pH or ATP levels [97]. It is interesting to speculate that A. balsamea's effects on mitochondria can also result in triggering anti-apoptotic events but further studies will be necessary to address this point.
Lastly, it is acknowledged that our study provides a good assessment of the efficacy (maximal effect) of the various plant extracts, used at maximal non-toxic concentrations, but does not provide values of EC50 (improvement of cell viability) or IC50 (reduction of apoptosis and caspase activities) that can offer an indication of the relative potencies of the various plant extracts tested. This approach was selected to be consistent with our numerous previous screening studies with the same 17 plant species, carried out in various bioassays related to diabetes and/or its complications [11,27,28,34,97]. In addition, EC50 and IC50 values are usually applied to drugs and pure compounds acting on a defined target, generally a receptor, and are thus of limited interpretation when applied to complex mixtures such as crude plant extracts used herein.

Conclusion
In summary, this study demonstrates that A. balsamea and G. hispidula exhibit the greatest potential of all Cree antidiabetic plants tested to protect kidney cells against damage induced by tubular apoptosis. A. balsamea appears to do so by suppressing apoptosis, most likely through actions on the mitochondrial caspase signaling pathway, whereas G. hispidula may exert its renal protective potential through mitochondrial and other pathways. In addition, the data generated in the current study suggest that this bioassay involving MDCK cells and flow cytometry may represent a useful and novel tool to screen for nephroprotective agents.
Finally, these findings may have clinical significance for the mitigation of DN that is affecting CEI communities. Indeed, Cree medicinal plant species may contain phytochemical components that can provide novel pharmacological avenues to improve DN treatment and management. More importantly, Cree traditional plant preparations hold promise as complementary therapies that are culturally adapted and respectful. As such, future clinical studies should be encouraged.