The leaves, roots and seeds of two Impatiens species were collected in August 2015. I. glandulifera Royle (no. IG-0815) were collected in Józefów, near Biłgoraj (Poland) at an altitude of 240 m a.m.s.l. (coordinates N 50°29’06; E 23°02’12’’) and I. noli-tangere L. (no. INT-0815) were gathered in Zalesie Górne near Warsaw (Poland) at an altitude of 115 m a.m.s.l. (coordinates N 52°2’16’’; E 21°1’55’’). Voucher specimens were deposited in the Department of Pharmaceutical Botany, Faculty of Pharmacy, Medical University of Lublin. The plants were identified by Prof. Tadeusz Krzaczek.

All chemical reagents used in the experiment were purchased from various commercial suppliers and were of the highest purity available. 2,2-diphenyl-1-picrylhydrazyl (DPPH), ferrozine (3-(2-pyridyl)-5,6-bis (4-phenyl-sulfonic acid)-1,2,4-triazine), α-tocopherol, 2,6-di-tert-butyl-4-methylphenol (BHT), were purchased from Sigma-Aldrich (St. Louis, MO, USA). Phosphate-buffered saline (PBS) was purchased from Gibco (Carlsbad, CA, USA). Other chemicals used for the preparation of the extracts were of analytical grade, and were obtained from Polish Reagents (POCH, Gliwice, Poland).

The air-dried, ground leaves (100.0 g), roots (100.0 g) and seeds (20.0 g) of two plants were extracted with petroleum ether (bp 45–60 °C) in a Soxhlet apparatus for over 50 h. The obtained lipid extracts were evaporated under vacuum. Subsequently, yellowish oil residues were saponified with a 10% ethanol solution of potassium hydroxide for 9 h, and the solutions were reduced to about half of their volume under vacuum, diluted with water, and extracted with diethyl ether. The ether phases were washed with distilled water until neutral, dried, and concentrated to yield crude unsaponifiable fractions. The sterols were first separated from these fractions in the typical manner by precipitation with digitonine (Jerzmanowska, 1967). The obtained digitonides were filtered and washed with ethanol, and acetylation was performed (acetic acid anhydrous, with anhydrous sodium acetate, 2 h, 100 °C). Sterol acetates were crystallized from 50% ethanol (Jerzmanowska, 1967; Harrabi et al., 2016). Saponification of these mixtures yielded the free sterols which were identified by GC-MS. The total amount of the phytosterols was determined by means of the weight method (Jerzmanowska, 1967).

Triterpenoids were then isolated from the filtrates remaining after the crystallization of sterol digitonides. The eluates were reduced to about half of their volume under vacuum, and methanol was added for the crystallization of triterpenoids. The obtained precipitates were again crystallized from ethanol, acetylated (acetic acid anhydride, anhydrous pyridine, 20 min), and again crystallized from ethanol. Hydrolysis of the triterpenoid acetate mixtures (5% KOH in anhydrous ethanol, benzene, 7 h) yielded free triterpenoids (Jerzmanowska, 1967). Both free and triterpenoid acetates were analyzed using GC-MS. A schematic diagram of the isolation sterol and triterpenoid fractions is presented in Scheme 1.

In the next stage, powdered plant materials were sonicated with n-hexane (3 × 30 min) at a controlled temperature (40 ± 2 °C). The supernatants were concentrated to dryness under vacuum at a controlled temperature. Then, 5 mL of methyl-tert-butyl ether were added to 0.1 g oil. The fatty acid methyl esters (FAMEs) were obtained by adding a trimethylsulfonium hydroxide solution (TMSH). The mixtures were then incubated (60 min, 60 °C), and analyzed using GC-MS.

For antioxidant and cytotoxic assays, 0.1 g of triterpenoid and fatty acid fractions were diluted in 80 and 95% ethanol (10 mL), respectively.

The analyses were performed on a Trace GC Ultra coupled with a DSQII mass spectrometer (Thermo Electron Corporation). GC-FID and MS analyses was performed using a MS-FID splitter (SGE Analytical Science). Mass range: 33-550 amu, ion source-heating: 200 °C, EI: 70eV, He (p: 300 kPa for phytosterols; 91 kPa for FAMEs). Operating conditions for derivatives of phytosterols: BPX5 (30 m×0.25 mm i.d., film thickness 0.25 μm), temperature program 100 °C (1 min, 10 °C/min) – 250 °C (15 min, 4 °C/min) – 300 °C (30 min). Injector temperature: 310 °C, and detector: 300 °C. Operating conditions for FAMEs: Stabilwax-DA, Restek (30m x 0.25 mm i.d., film thickness 0.25 μm), 50 °C (3 min) – 250 °C (40 min), 4 °C/min. Injector: 250 °C, detector 260 °C.

Phytosterols were analyzed as TMS (trimethylsilyl ethers; Thanh et al., 2006) and fatty acids as FAME derivatives along with their mass spectra, compared with the MS of the NIST/EPA/NIH Mass Spectral Library 2009 and Wiley Registry of Mass Spectral Data, 8th edition. The retention times (Rt) were compared with a standard mixture Rt.

Area percent was obtained electronically from the GC-FID response without the use of an internal standard or correction factors. Quantitative determinations corresponded to the means of three replicates ± SD.

Human leukemia cell lines HL-60 and HL-60/MX2 were cell lines of AML (acute myeloid leukemia) origin and were obtained from the ECACC (the European Collection of Cell Cultures).

The HL-60 and HL-60/MX2 cells were maintained in RPMI 1640 medium (Biomed Lublin) supplemented with 20% and 10% of fetal bovine serum (FBS; PAA Laboratories, Linz, Austria), respectively and the antibiotics were 100 U/mL penicillin, 100 μg/mL streptomycin, and 2.5 μg/mL amphotericin B (Gibco, Carlsbad, USA), incubated at 37 ºC in a humidified atmosphere of 5% CO_2.

The effect of the obtained triterpenoid and fatty acid fractions on HL-60 and HL-60/MX2 cell lines was measured using the trypan blue assay. These cells were seeded on 12-well plates (Sarstedt, Wiener, Austria) at a density of 2×10⁵ and 3×10⁵ cells per well, respectively. After 24 hours, the cell suspensions were treated with plant samples at concentrations ranging from 10 to 5000 μg/mL for triterpenoid and from 10 to 2500 μg/mL for fatty acid fractions, and incubated for 24 hours. Then, cell suspensions were centrifugated (800 rpm, 5 min), washed with PBS and centrifugated again. The cells were stained with a 0.4% solution of trypan blue (Bio-Rad) and counted with a TC10^TM Automated Cell Counter (Bio-Rad). Each experiment was repeated three times. Dose response curves were made and IC₅₀ values were found.

To determine the antioxidant activity of the obtained triterpenoid and fatty acid fractions, two methods were used. DPPH (2.2-diphenyl-1-picryl-hydrazyl) free radical scavenging activity was measured according to the Brand–Williams et al., (1995) method. The changes in color from deep-violet to light-yellow were measured at 515 nm in a UV/visible light spectrophotometer. The metal chelating power was determined using the Guo et al., (2001) method. Absorbance was measured at 562 nm and the percentage of inhibition of ferrozine- Fe²⁺ complex formation was calculated according to the formula:

% inhibition = [1-(A_s/A _b )] x100

where A _b - absorbance of the blank, A _s – absorbance in the presence of the test sample

Both antioxidant activities were expressed as an efficient concentration EC₅₀, the extract solution concentration provided 50% of the activity in a dose-dependent manner. α-Tocopherol and BHT were used as a positive control.

All extractions and determinations were done in triplicate. The results are presented as experimental means and ± SD. The obtained data were subjected to statistical analysis using Statistica 10.0. (StatSoft, Cracow).

The mixtures of sterol acetates and free sterols were obtained in typical way, and then analyzed using the GC-MS method. The total amount of sterols was determined using the weight method and the results are given in Table 1. On the basis of GC-MS analysis, in the sample from the roots of I. Glandulifera (IGR) fifteen phytosterols were identified. In the leaves of both species (IGL; INL) in the seeds of I. noli-tangere (INS), in the seeds of I. Glandulifera (IGS), and the roots of I. noli-tangere (INR),eleven and twelve compounds were observed, respectively.

The major components of the total sterol fractions obtained from the leaves of I. glandulifera were spinasterol + chondrillasterol and ∆⁷-sitosterol. The co-eluted compounds were separated in the GCxGC analysis (Figure 1). Similarly, spinasterol was observed in the root cultures of I. balsamina (Panichayupakaranant et al., 1995). This compound is known as an antimutagen agent that was tested using the mouse skin tumor assay (Villaseñor and Domingo, 2000). α-Spinasterol, together with glanduliferins A and B (belonging to a cholestane subgroup), were isolated from I. glandulifera (Cimmino et al., 2016). In the rhizomes of I. pritzellii, some sterols, such as α-spinasterol, spinasteryl-3-one, α-spinasteryl-3-O-β-ᴅ-glucopyranoside, and 3-O-[6’-O-palmitylo-β-ᴅ-glucosyl]-spinasterol were noticed previously (Zhou et al., 2007). Moreover, α-spinasterol was isolated from the seeds, roots, leaves and fruits of I. balsamina (Wang et al., 2011).

In other samples in our research, the most important sterol was β-sitosterol (ranging from 23.25 to 59.00% of total sterols). High levels of campesterol and sitostanol were also observed in these fractions. Other compounds such as ergost-8(14)-en-3-ol, campestenol, ∆⁷-stigmasterol, ∆⁵- avenasterol, and cholesterol occurred in low concentrations.

The appropriate GC-MS procedure allowed for the identification of the triterpenoid acetate fraction from the leaves of I. glandulifera (IGLt).Thirteen compounds such as 5α-ergost-7-en-3β-ol, taraxasteryl, α-spinasterol, 13,27-cycloursan-3-ol, (3β,13β,14β)-, β-amyrin, cycloeucalenyl, β-simiarenol, stigmast-7-en-3-ol, (3β,5a)-, 5α-lup-20(29)-en3β-ol, 9,19-cyclolanostan-3-ol, 24-methylene-, (3β)-, Ψ-taraxasteryl, lupan-3-ol, and olean-12-ene-3,28-diol, (3β)- were detected. The occurrence of eleven compounds was observed in the fraction from the roots of I. glandulifera (IGRt). In the roots of I. noli-tangere (INRt) and the seeds of I. glandulifera (IGSt), eight compounds were observed. In the leaves of I. noli-tangere (INLt), ten compounds were detected; and in the seeds of I. noli-tangere (INSt), seven compounds were found. As can be seen in Table 2, phytosterols were also present in these fractions, with the exception of triterpenoids. An exemplary GC-MS chromatogram of the triterpenoid acetates from the leaves of I. Glandulifera is shown in Figure 1.

In the GC-MS analysis of the triterpenoid acetate fraction from the leaves of I. glandulifera, peak 1 had an identical retention time to 5α-ergost-7-en-3β-ol acetate. The mass spectrum was similar to that obtained for authentic 5α-ergost-7-en-3β-ol and showed a molecular ion at m/e 426 and other important ions at m/e 411, 393, 355, 341, 302, 287, 269, 257, 218, and 204. This compound was only present in this sample. 13,27-cycloursan-3-ol, (3β,13β,14β)- was observed in the leaves of both species, and cycloeucalenyl was detected only in the leaves and roots of I. glandulifera. Ψ-Taraxasteryl acetate was absent in the analyzed seeds. The taraxasteryl, α-spinasterol, β-amyrin, β-simiarenol, stigmast-7-en-3-ol, (3β,5a)-, and 5α-lup-20(29)-en3β-ol acetates were identified in all the examined fractions from both Impatiens species.

The GC-MS method permitted an estimation of the contents of triterpenoid acetates determined in the leaves, roots and seeds of I. glandulifera and I. noli-tangere, based on the total fraction (Table 2). The quantitative analysis was done in triplicate. Amongst triterpenoid acetate fractions, the most abundant compounds in all samples were α-spinasterol (from 2.42 to 33.91%), 5α-lup-20(29)-en3β-ol (5.73 - 22.91%) and β-amyrin (6.34 – 17.08%).

In the hexane extracts from the leaves, roots and seeds of I. glandulifera and I. noli-tangere ten fatty acids were identified (Table 3). The saturated fatty acids comprised from 12.2 (roots of I. noli-tangere; INRf) to 27.2% (leaves of I. glandulifera; IGLf), monounsaturated – 16.9 (IGLf) to 34.1% (leaves of I. noli-tangere; INLf), and polyunsaturated fatty acids comprised 40.3 (INRf) to 55.8% (IGLf). The unsaturated fatty acids α-linolenic and oleic acids were dominant compounds in all the examined samples. In I. noli-tangere γ-linolenic acid was also found at a high level, from 5.8% for the roots to 7.9% for the leaves. Trace fatty acids such as capric acid (C10:0) was detected only in the roots and seeds of I. noli-tangere (INSf). Small amounts of arachidonic acid (C20:4) were noticed in the leaves and seeds of I. glandulifera (IGLf, IGSf). The presence of azelaic acid in the studied plants is interesting, as this acid has antibacterial, anti-inflammatory, keratolytic and sebostatic and tyrosinase-inhibiting properties, which is quite rare among plants (Reszke and Szepietowski, 2016).

It was observed that the amounts of the saturated fatty acids from the leaves and seeds of both examined species were higher than in the roots (13.8% in I. glandulifera and 12.2% in I. noli – tangere). Differences in the values of the unsaturated / saturated ratios were observed from 2.7 for the leaves and seeds of I. glandulifera to 5.1 for the roots of I. noli-tangere. The polyunsaturated fatty acids were the most abundant of the unsaturated fatty acids in all organs of both species. The polyunsaturated (PUFAs) / monounsaturated (MUFAs) fatty acids ratios were from 1.2 for the leaves of I. glandulifera to 3.3 for the leaves of I. noli-tangere.

All the examined samples contained both ω-3 and ω-6 fatty acids. It has been reported that the ideal intake ratio of ω-6 to ω-3 fatty acids is between 1:1 and 4:1 (Simopoulos, 2006). According to Simopoulos (2006), increased levels of ω-3 fatty acids exert suppressive effects on many diseases such as cancer, cardiovascular disease, and inflammatory and autoimmune diseases. In our study, the ratio of ω-6 to ω-3 fatty acids is from 1/2.5 (roots of I. glandulifera; IGRf) to 1/12.8 (roots of I. noli-tangere; INRf). The obtained lower ratio of omega-6 to omega-3 fatty acids suggests that I. glandulifera and I. noli–tangere may be considered for the prevention and management of chronic diseases. The richer source of omega-3 fatty acids is I. glandulifera. A comparison of the present results with those of other authors has revealed that some of the identified compounds in the leaves, roots and seeds of I. glandulifera and I. noli-tangere have been reported in I. glandulifera. Ortin and Evans, (2013) noticed that the hydrophobic extracts from flower stalks with seed pods of I. glandulifera contained linolenic, palmitic, stearic, arachidic and trans-tetradec-2-enoic acids. According to Kaufmann and Keller, (1948), the Impatiens species contained oils with acetic and parinaric acid glycerides. They found that the seeds of I. glandulifera contain 50% and I. noli-tangere 55% of these types of oils. Saponification provided 13% glycerol, 10% acetic acid, 40% parinaric acid, about 3% palmitic acid, about 3% stearic acid and about 20% mixture of oleic, linoleic and linolenic acid. Nisar et al., (2012) analyzed a n-hexane extract of I. bicolor. Their study showed that the fatty acid esters, such as trans-methyl 13-octadecenoate, methyl heptadecanoate, methyl octadecanoate, methyl docosanoate, methyl tetracosanoate, and methyl eicosanoate are the major compounds of this extract.

Because plant triterpenoids and fatty acids have been reported to exhibit a variety of antioxidant, anti-inflammatory, antimicrobial, and antitumor promoting biological activities (Topçu, 2006; Dzubak et al., 2006), the antioxidant and cytotoxic properties of fractions obtained from the leaves, roots and seeds of I. glandulifera and I. noli-tangere were also evaluated.

In this study, the effect of the triterpenoid and fatty acid fractions in increasing concentrations on two types of cancer cell lines, HL-60 and HL-60/MX2, were investigated.

The cytotoxicity was estimated using trypan blue vital staining. The experiment was performed in triplicate and the mean values were calculated from the given values. The IC₅₀ (Half Maximal Inhibitory Concentration, the inhibitor concentation when cell viability is 50%) values of the examined samples were determined using MS Excel. The cells of both cancer lines exposed to these fractions presented diverse cytotoxicity depending on the dose of IC₅₀.

Based on the obtained results, it was found that the analyzed fractions from I. glandulifera and I. noli-tangere induce apoptosis of the cells of the both tested cell lines. The results, which are given in Table 4, showed that the triterpenoid fraction from I. noli-tangere seeds significantly inhibited HL-60 human leukemia cells, and was the most potent fraction with IC₅₀ values of 11.69 μg/mL, followed by fatty acid fractions from the roots and leaves of I. glandulifera with IC₅₀ of 41.54 and 61.81 μg/mL, respectively. Moreover, the fatty acids from the roots of I. noli-tangere and triterpenoids from the roots of I. glandulifera showed a moderate cytotoxicity against HL-60 with IC₅₀ values of 65.37 and 65.56 μg/mL.

On the other hand, the triterpenoid fractions from the seeds of I. glandulifera and I. noli-tangere showed high inhibition activity against the HL-60/MX2 cell line with IC₅₀ of 33.92 and 43.35 μg/mL, respectively. Relatively high cytotoxicity against HL-60/MX2, fatty acid fractions from the leaves of I. noli-tangere (IC₅₀ = 50.82 μg/mL) and roots of I. glandulifera (IC₅₀ = 52.81 μg/mL) was shown.

The weakest cytotoxic activity was found in the triterpenoid fraction from the leaves of I. glandulifera. The lowest IC₅₀ doses against both cell lines used in the research was determined.

Based on the IC₅₀ values, it can be concluded that the HL-60 cell line is more sensitive to the Impatien fractions studied. The highest IC₅₀ value was observed after the exposure of HL-60 cells to leaf extract (IC₅₀ = 963.69 μg/mL).

The presence of 5α-lup-20 (29)-en3β-ol and α-spinasterol may be responsible for the cytotoxic activity of the studied fractions. It has been reported that 5α-lup-20(29)-en3β-ol inhibits skin cancer in CD-1, and induces apoptosis in HL-60 human leukemia cells (Saleem et al., 2004; Zhang et al., 2009). Spinasterol has shown antitumor effect against ovarian, skin and breast cancer cells (Jeon et al., 2006).

Previous studies report the cytotoxic and antitumor activities of extracts and compounds isolated from I. balsamina. For example, the ethanol extract of I. balsamina was investigated for in vitro cytotoxic and in vivo antitumor activities against transplantable tumors and human cell lines. The obtained results showed significant antitumor and cytotoxic effects against Dalton’s ascites lymphoma and human cancer cell lines (Baskar et al., 2012). Moreover, balsaminone C (dinaphthofuran-7,12-dione derivative) isolated from the seeds of I. balsamina exhibited cytotoxicity against A549, Bel-7402 and Hela cancer cell lines (Pei et al., 2012).

The antioxidant activities of triterpenoid and fatty acid samples were evaluated using the DPPH radical-scavenging test and metal chelating power. As seen in Table 5, EC₅₀ values for the analyzed fractions and standards on the DPPH radical were found in the range of 12.81 to 27.11 μg/mL and from 39.36 to 48.74 μg/mL for triterpenoid fractions from I. glandulifera and I. noli-tangere, respectively, from 9.43 to 18.18 μg/mL and from 11.61 to 22.04 μg/mL for fatty acid fractions from I. glandulifera and I. noli-tangere, respectively, 29.16 μg/mL for BHT and 1.50 μg/mL for α-tocopherol. Most of the samples tested showed better antioxidant activity than BHT (the lower EC₅₀ values) but weaker than α-tocopherol. The fatty acid fraction from the leaves of I. glandulifera showed the strongest DPPH radical scavenging activity with EC₅₀ at 9.43 μg/mL, followed by the fatty acid fraction from the seeds of I. glandulifera and the leaves of I. noli-tangere with EC₅₀ of 10.56 and 11.61μg/mL, respectively. The strong antioxidant activity may be attributed to the presence of high amounts of omega-3 fatty acids in these fractions.

EC₅₀ values for the Fe²⁺ chelating capacities of the analyzed fractions were found in the range from 9.62 to 15.13 μg/mL and from 16.00 to 19.83 μg/mL for triterpenoid fractions from I. glandulifera and I. noli-tangere, respectively. For fatty acid fractions the values of EC₅₀ ranged from 5.49 to 11.86 μg/mL and from 6.01 to 13.62 μg/mL for I. glandulifera and I. noli-tangere, respectively. The positive controls BHT and α-tocopherol showed EC₅₀ values of 5.30 and 13.04 μg/mL, respectively. The results obtained in this test showed that metal chelating power was similar to positive controls.

Similar findings were obtained for the n-hexane extract of I. bicolor in the DPPH assay. Nisar et al., (2012) found that IC₅₀ values ranging from 23.22 to 59.00 μg/mL were comparable to ascorbic acid. Moreover, the antioxidant activity was comparable with the content of the fatty acid esters in the tested samples.

In another study, it was found that ethanolic (Baskar et al., 2012) and water (Sha et al., 2013) extracts from I. balsamina showed antioxidant activity which was measured using different assays. On the other hand, our previously research showed that the essential oils obtained from I. glandulifera, I. parviflora, I. balsamina, and I. noli-tangere had moderate antioxidant activity (Szewczyk et al., 2016a). Furthermore, hydromethanolic extracts from the aerial parts of. I. balfourii, I. glandulifera, I. parviflora, I. balsamina, I. noli-tangere, and I. walleriana also possessed moderate antioxidant potential in the DPPH radical scavenging and reducing power assays (Szewczyk et al., 2016b).