Isolation and purification of glycoglycerolipids to induce apoptosis in breast cancer cells

Monogalactosyldiacylglycerol (MGDG) is the most abundant type of glycoglycerolipid found in the plant cell membrane and mostly in the chloroplast thylakoid membrane. The amphiphilic nature of MGDG is attractive in pharmaceutical fields for interaction with other biological molecules and hence exerting therapeutic anti-cancer, anti-viral, and anti-inflammatory activities. In this study, we investigated the therapeutic efficacy of cyanobacteria derived MGDG to inhibit breast cancer cell growth. MGDG was extracted from a cyanobacteria Synechocystis sp. PCC 6803 followed by a subsequent fractionation by column chromatographic technique. The purity and molecular structure of MGDG were analyzed by nuclear magnetic resonance (NMR) spectroscopy analysis. The presence of MGDG in the extracted fraction was further confirmed and quantified by high-performance liquid chromatography (HPLC). The anti-proliferation activity of the extracted MGDG molecule was tested against BT-474 and MDA-MB-231 breast cancer cell lines. The in vitro study showed that MGDG extracted from Synechocystis sp. PCC 6803 induced apoptosis in (70 ± 8) % of BT-474 (p < 0.001) and (58 ± 5) % of MDA-MB-231 cells (p < 0.001) using ~ 60 and 200 ng/ml of concentrations, respectively. The half-maximal inhibitory concentration, IC50 of MGDG extracted from Synechocystis sp. PCC 6803 were (27.2 ± 7.6) and (150 ± 70) ng/ml in BT-474 and MDA-MB-231 cell lines, respectively. Quantification of caspase-3/7 activity using flow cytometry showed (3.0 ± 0.4) and (2.1 ± 0.04)-fold (p < 0.001) higher protein expressions in the MGDG treated BT-474 and MDA-MB-231 cells, respectively than untreated controls conferring to the caspase-dependent apoptosis. The MGDG did not show any significant cytotoxic side effects in human dermal fibroblasts cells. A commercially available MGDG control did not induce any apoptotic cell death in cancer cells substantiating the potential of the MGDG extracted from Synechocystis sp. PCC 6803 for the treatment of breast cancer cells through the apoptosis-mediated pathway.

Glycoglycerolipids (GGLs) are natural products abundantly found in the cell membrane of marine algae 1,2 , cyanobacteria 3,4 , and higher plants [3][4][5][6][7] , with one or two carbohydrate units, glycerol, and diversified acyl lipid groups 7,8 . GGLs are conserved in both gram-negative and gram-positive bacteria, i.e., Bacillus pumilus 9 , Lactobacillus plantarum 10 , Microbacterium sp. 11 , Micrococcus luteus 12 , and Phormidium tenue 13 with highly conserved structures. The amphiphilic nature of GGLs is attractive in pharmaceutical fields for interaction with other biological molecules and hence exerting therapeutic anti-cancer [14][15][16] , anti-viral 17,18 , and anti-inflammatory activities 19,20 . GGLs potently inhibit angiogenesis 21 , cancer cells 14 , and solid tumor 16,22,23 growth by selectively inhibiting the replicative DNA polymerase activity both in vitro and in vivo 15,16,23 . Monogalactosyl diacylglycerol (MGDG), a GGL isolated from spinach, combined with gemcitabine anti-cancer drug revealed synergistic effects of inhibitive proliferation on human pancreatic cancer cell lines BxPC-3, MiaPaCa2, and PANC-1 through the inhibition of DNA replicative pols alpha and gamma activities, compared with MGDG or gemcitabine alone 24 . The fractions of GGLs e.g. MGDG, digalactosyl diacylglycerol (DGDG), and sulfoquinovosyl diacylglycerol (SQDG) in spinach potently affect in vitro colon cancer cells, angiogenesis, and solid tumor growth via their inhibitory activities of DNA polymerase 16,[21][22][23] . Naturally occurring sulfoquinovosylglycerolipids show promising anti-proliferative activity toward human cancer cells not only by targeting DNA polymerases 25 , but also inhibiting the mitotic centromere-associated kinesin (MCAK) 15 . Another interesting feature of these compounds is their involvement in cell recognition and signaling which make them promising agents in drug delivery systems 26 . Several GGLs have been related to the activation of natural killer T cells, which is a central event in a variety of immune responses including the development of autoimmunity, tolerance, and maintenance in defense responses  39 . Fluorescence microscopic images of PCC 6803 and the corresponding spectra analysis were captured using a scanning laser inverted confocal microscope (Ti-Eclipse; Nikon). For the growth curve analysis, the species was inoculated in a 5 ml liquid BG11 media. 150 μl aliquots of liquid samples were transferred to a 96 well plate (Corning) at different times to measure the growth kinetics of PCC 6803 after adjusting the initial absorbance at 680 nm, A 680,0 = 0.05 at 24 h. BG11 media alone without any inoculum was used as a negative control. The maximum absorbance at the wavelength of 680 nm was measured using a microplate reader (BioTek). The specific growth rate, α was computed using Eq. (3).

Materials and methods
where A 680 is the absorbance of cells at time, t . By integrating Eq. (1) from t = 0 to t = t, The doubling time, t d is calculated using Eq. (4): Total lipid extraction from Synechocystis Sp. PCC 6803. The total lipid was extracted from the Synechocystis sp. using the Bligh and Dyer Method 40 . Cell pellets from 250 ml of culture were collected at A 680,t = 3.0 after 120 h, centrifuged at 15,000 g for 15 min and suspended in deionized water for two times to remove residual medium. One ml of cell suspension was mixed with 3.75 ml of chloroform: methanol (1:2, v/v) solution for 15 min using a vortex mixer. After thorough mixing, 1.25 ml of chloroform was added and vortexed for 1 min. Two distinctive layers were obtained after adding 1.25 ml of water to the mixture. The chloroform layer (lower phase) contained all the lipids, and the aqueous methanol layer (upper phase) contained the non-lipid fractions. The upper phase was discarded carefully, and the lower phase was collected. Chloroform was then evaporated completely under vacuum in a fume hood to obtain the total lipid. Total lipid was reconstituted in 1 ml of chloroform and stored in −20 °C.
Isolation of MGDG using column chromatography. Total lipid was fractionated using the normal phase column chromatography technique. A manually packed column was used for the isolation process (SI Fig. 1a). Silica gel (particle size 10-40 microns) was used as a stationary phase. Chloroform and acetone were used as mobile phases. A gradient elution method was applied for the fractionation (SI Table 1). Samples (A-K) www.nature.com/scientificreports/ were collected for each combination of the eluent and labeled for further detection, structural analysis, and quantification (SI Fig. 1b). . The absorbance value was adjusted by subtracting the mean absorbance level of wells containing medium only. Cell viability was calculated as a means of six wells containing BT-474, MDA-MB-231, and dermal fibroblast cells. The cells treated with PBS were used as live control that was placed in every alternative column in the well plate adjacent to corresponding dosage or treatments to remove the variability in the number of seeded cells in adjacent wells. The live controls were replicated at least six times for each dosage. For dead control, we treated the cells with saponin in at least six replicates (SI Fig. 2). The % cell viability was calculated as follows Eq.

Statistical analysis.
Each experiment was carried out in independent repetitions to have at least triplicates valid measurements. The results were reported as mean ± standard deviation and analyzed using the Student's t-tests with two-tailed hypotheses and using the JMP statistical software (version 15, SAS Institute). p < 0.001 was considered as statistically significant and was denoted by ***.

Results
Synechocystis sp. growth curve kinetics and molecular structure of MGDG. Figure 1 demonstrates the confocal fluorescence images and fluorescence emission spectra analysis of Synechocystis sp. PCC 6803, indicating that the microorganism grew properly with the highest absorbance spectrum at 680 nm 44,45 . The growth kinetics of Synechocystis sp. PCC 6803 was studied to obtain the maximum total lipid yield. The growth curve shows that the lag phase extended for almost 30 h (Fig. 2a). It took the species for six days to reach the stationary phase. The specific growth rate, α was calculated as 0.04 h −1 . The doubling time ( t d ) was calculated as  NMR analysis confirms the isolation of MGDG from total lipid extract. MGDG molecule was isolated from the total lipid extract of Synechocystis sp. PCC 6803 using column chromatography. The fractions with each eluent were collected separately and analyzed individually by NMR analysis. The MGDG molecule was isolated with an eluent combination of 40% chloroform and 60% acetone. The fractions (H, I, J) were collected with the same eluent combination. The NMR analysis confirms the presence of MGDG molecule in fraction I (Fig. 3). The set of NMR peaks of fraction I were completely identical to the standard MGDG molecule suggesting the isolation of MGDG molecule with high purity. The structure of MGDG in fraction I was analyzed from the peaks obtained in the NMR analysis (Fig. 4) Detection and quantification of MGDG using HPLC-UV system. The Synechocystis sp. extracted MGDG molecule was detected and quantified using the HPLC coupled with a UV detector. The correlation of area under the peak with known concentrations of standard MGDG was observed in HPLC analysis (Fig. 5). The R 2 value of the fitted curve was 0.99. After the isolation of MGDG from total lipid extract in chromatographic fraction I, the fraction was run through the HPLC column for quantification. The peak obtained at 6.24 min confirmed the presence of the MGDG molecule (Fig. 6). The amount of detected MGDG in fraction I was quantified using the concentration correlation plot obtained in     Fig. 7a,b). Although MGDG from Synechocystis sp. induced the % cell death of breast cancer cells in a concentration-dependent manner, the cytotoxicity of fibroblast cells (normal cell control) was less (< 15%) than breast cancer cells after exposure to 100 ng/ml of MGDG for 72 h (Fig. 7c), indicating that breast cancer cells were more sensitive to the MGDG derived from Synechocystis sp. than normal cells. Taken    www.nature.com/scientificreports/ green fluorescently stained DNA indicating the activation of caspase-3/7 in BT-474 (Fig. 8a) and MDA-MB-231 cells (Fig. 8b). The combined phase and fluorescent signal images show the overlap of caspase-3/7 positive cells and the unhealthy shrinking cells suggesting the initiation of apoptosis. The induction of apoptosis was assessed by Western blot technique (SI Figs. 3 and 4). The Western blot analysis showed that the full-length caspase-3 (MW: 32 KDa) expression was reduced in BT-474 and MDA-MB-231 cells treated with MGDG from Synechocystis sp., while the protein expression did not change in cells treated with MGDG standard relative to untreated cell controls. The reduced expression of full-length caspase-3 suggests the cleavage of the protein into an active form of low molecular weight (17 kDa) active cleaved caspase-3. The cleaved caspase-3/7 expression was quantified using flow cytometry. The fold enhancement of caspase-3/7 was quantified using flow cytometry analysis (Fig. 8c,d). We observed a statistically significant increase of 1.7 ± 0.04 and 3.0 ± 0.4 (p < 0.001)-fold higher caspase-3/7 concentrations after 48 and 72 h, respectively in BT-474 cells treated with MGDG extracted from Synechocystis sp. than untreated cell controls (filled column, Fig. 8c, SI Fig. 5a,5b). The number of caspase-3/7 concentrations in MDA-MB-231 cells treated with MGDG from Synechocystis sp. increased significantly up to 2.1 ± 0.03-fold (p < 0.001) after 72 h compared to untreated cell controls (filled column, Fig. 8d, SI Fig. 5c,d). MGDG standard did not show any significant changes in cleaved caspase-3/7 expressions compared to untreated control cells. These results are in good agreement with fluorescent microscopic images indicating the cells being shrunk and round shaped with a complete halt in growth of cancer cells. In contrast, after 48 and 72 h of treatment, there was no sign of apoptosis in the standard MGDG treated compared to untreated cell controls (open column, Fig. 8c,d). The untreated cells and the cells treated with standard MGDG were compact multilayered colonized and elongated in shape in the case of BT-474 and MDA-MB-231 cell lines, respectively, and were observed healthy suggesting continuous growth capability.

Discussion
MGDG contains fatty acyl groups derived from two fatty acid molecules at sn-1 and sn-2 position and a polar head at sn-3 position in a 3-carbon glycerol scaffold (Fig. 2b). Fatty acid molecules such as n-6 GLA may exert anti-proliferative effect by regulating genes and proteins involved in cell cycle and apoptosis, altering the cellular composition of fatty acids, and by producing downstream anti-proliferative metabolites such as 1-series prostaglandins and free radical molecules from cyclooxygenase (COX) catalyzed lipid peroxidation 55,56 . The potent anti-proliferative activity of the MGDG is mostly attributed to the fatty acyl components of the MGDG molecular structure 55,57 .  48 showed that the most abundant n-6 GLA moiety in Synechocystis sp. derived MGDG is subject to desaturation when the growth temperature was shifted from 38 °C to 22 °C. In contrast to the variable sn-1 fatty acid content, the sn-2 position was shown to be conserved mostly with 16: 0 (palmitic acid) fatty acid chain in Synechocystis sp. derived MGDG and was not affected with the shifts in temperature 6,48,49 . The MGDG lipid molecule was isolated in a normal phase column using chloroform and acetone gradient eluent combinations. The column length to diameter ratio was ~ 10. Multiple samples were collected with each eluent combination for better purification. We confirmed the presence of MGDG molecule in fraction I using 1 H NMR analysis (Fig. 3). Though fraction H, I, and J were eluted with the same mobile phase combination of 40% chloroform and 60% acetone, only fraction I contained the MGDG molecule. The peaks obtained for a fraction I identical with > 99% pure standard MGDG molecule suggests the high purity of MGDG molecule in that fraction. The diallylic methylene moiety obtained at δ 2.8 in the NMR analysis shown in Fig. 4 confirmed the polyunsaturation in the long fatty acid chain at the sn-1 position. The extracted MGDG molecule in fraction I was quantified using the HPLC-UV system. We prepared a calibration curve (Fig. 5) with known > 99% pure standard MGDG molecule to quantify the MGDG in each extracted fraction (~ 94.4 μg/ml). The peak obtained at retention time 6.24 min shown in Fig. 6, upper panel corresponds to the MGDG molecule in the fraction 43 . The large peak obtained at retention time 2.36 min is the lipid pigments eluted initially by the chloroform. The peak at 7.23 min might be due to the presence of steryl glucosides present as a contaminant in the fraction 43 . Glucosides do not have any known effects on the HER2-positive and triple-negative breast cancer cells. The peaks for pigments and glucosides were also observed in the standard > 99% pure standard MGDG chromatogram (Fig. 6, lower panel). So, the contaminant should not have any effect in this study during the drug formulation used for cytotoxicity analysis. However, the GGL molecules are detected with high quality and precision using the evaporative light scattering detectors (ELSD) providing with sharp and clear peaks in the chromatogram 43,58,59 .
We investigated the in vitro anti-proliferation efficacy of cyanobacteria derived MGDG molecule in comparison with MGDG molecule from plant source on the HER2-positive and triple-negative breast cancer cell lines. We observed significant differences in the % of cancer cell death using MGDG extracted from Synechocystis sp. versus a commercially available MGDG standard. MGDG extracted from Synechocystis sp. induced apoptosis in (70 ± 8) % of BT-474 and (58 ± 5) % of MDA-MB-231 cells at 60 and 200 ng/ml, respectively, while < 15% cell death was observed in fibroblasts at 100 ng/ml (Fig. 7). While the precise mechanism by which each molecule works about their cell-killing effect is unknown, it is possible that the presence of polyunsaturated GLA in the sn-1 position and palmitic acid in the sn-2 position of MGDG from Synechocystis sp. resulted in the mitochondrial depolarization, cytochrome c release, DNA fragmentation and generation of free radicals causing specific cell death  [60][61][62][63][64] . GLA has been shown to exhibit anti-proliferative activities specifically in a variety of cancer cell lines both in vitro and in vivo. GLA inhibited the cell growth of four human neuroblastoma cell lines (GOTO, SK-N-DZ, NKP, and NCG) in vitro 65 , three human glioma cell lines (MOG, U87, U373) and a rodent glioma cell line (C6) in vitro and a rat C6 glioma model in vivo 66 . Glioma regression and apoptosis had been reported using both C18 and C20 fatty acids of the n-6 and n-3 series GLA along with the preservation of normal neural tissue and vasculature in adjacent brain 66 . GLA at a concentration of 150 μM inhibited Walker 256 cancer cell growth both in vitro and in vivo causing decrease in mitochondrial membrane potential, and increase in cytochrome c release, caspase activation, and DNA fragmentation 61 . The mitochondrial apoptosis pathway was likely induced by an increase in reactive oxygen species (ROS), lipid peroxide production, ATP generation and the deposition of large amounts of triacylglycerol in the form of lipid droplets 61,63 . A diet containing 5.5% GLA caused 45% decrease in Walker 256 tumor growth in vivo by reducing mitochondrial metabolic activity 60 . More interestingly, in vitro, in vivo and clinical study data showed that GLA has selective anti-proliferative actions in cancer cells with little or no side effects on normal cell growth. Polyunsaturated fatty acids including GLA incubated with human breast, lung, and prostate cancer cells suppressed the cancer cell growth exhibiting no adverse effects on normal human fibroblasts or normal animal cell lines 67 . Intraarterial injection of a lithium salt derivative of GLA demonstrated its ability to selectively suppress angiogenesis 62 . These reports, together with our data increases lead to the conclusion that MGDG from Synechocystis sp. is a promising cancer therapeutic agent with high selectivity of cancer cell growth inhibition leading to apoptosis and a decrease in cancer development. Also, saturated fatty acid, palmitic acid at the sn-2 position of MGDG from Synechocystis sp. plays a significant role in the elevation of calcium flux, endoplasmic reticulum stress, caspase-3, and caspase-9 activity, and thus inducing apoptosis which is in good agreement with previous reports [68][69][70] . Treatment of mouse 3T3-L1 and rat primary preadipocytes with palmitic acid induced multiple cell signaling pathways, endoplasmic reticulum stress responses, and cell cycle arrest leading to apoptosis 68 . Palmitic acid induced oxidative stress and DNA damage in rodent-derived insulin-secreting cell line RINm5F and primary human fibroblasts 70 . Spinach MGDG molecule in combination with gemcitabine in vitro 57 and radiation in vivo 71 showed enhanced suppression of MIAPaCa-2, PANC-1, and BxPC-3 pancreatic cancer cell lines compared to MGDG treatment alone. With the in vitro spinach MGDG treatment alone the IC 50 values for the mentioned pancreatic cell lines ranged from 18 to 25 μM 71,72 . MGDG molecule extracted from the spinach has been shown to inhibit human replicative and repair DNA polymerase enzymes with the IC 50 values ranging from 10 to 200 μM 57 . These results illustrate the specificity of the MGDG extracted from Synechocystis sp. PCC 6803 to breast cancer cells by caspase-dependent apoptotic pathway (Fig. 8).

Conclusion
Fatty acids are of great interest as natural anti-proliferative factors because of their selective inhibition of cancer cells. The structure and efficacy of fatty acid molecules are variable within and across the sources. We investigated the cancer cell inhibition efficacy of plant and cyanobacterial derived MGDG molecule. The marine cyanobacteria Synechocystis sp. derived MGDG molecule predominantly rich in n-6 GLA was isolated from the total lipid extract in this study. The molecule was isolated in a hand-packed low-cost normal phase silica column with the gradient elution method using chloroform and acetone as the mobile phase. We confirmed the presence of MGDG molecule in the fraction eluted with 40% chloroform and 60% acetone eluent combination. The NMR analysis confirmed the high purity of the isolated MGDG molecule from the cyanobacterial total lipid extract. The isolated MGDG molecule was quantified by the HPLC Our results support these findings showing that the cyanobacteria derived MGDG induces caspase-dependent apoptosis pathway to inhibit the HER2-positive BT-474 and triple-negative MDA-MB-231 breast cancer cell growth. Further studies involving additional mechansisms of actions, intracellular uptake, and robust screening of additional cancer cell lines will confirm the potential novel therapeutic efficacy of cyanobacteria derived MGDG molecule for a low dosage selective treatment of breast cancer cells.