Enhanced detection of metastatic prostate cancer cells in human plasma with lipid bodies staining

LNCaP and RWPE1 cell lines were obtained from ATCC. LNCaP cells were maintained in RPMI media and RWPE1 cells were grown in Keratinocyte media supplemented with bovine pituitary extract and epidermal growth factor (Invitrogen, Carlsbad, CA) in a humidified incubator with 5% CO₂. All cell culture media were supplemented with 10% FBS. Human sodium citrate pooled plasma was purchased from Bioreclamation Inc. (Westbury, NY). The peripheral blood mononuclear cells (PBMC) were isolated from whole heparinized blood by Ficoll plaque gradient centrifugation; buffy coat was collected and washed with PBS to remove platelets.

Total protein extracts were separated on 10% SDS-PAGE gels, transferred to nitrocellulose membranes, incubated first with primary antibodies against proteins of interest and then with secondary antibodies conjugated with HRP or labeled with Infrared Dye. Membranes were stripped, and re-incubated with antibodies against GAPDH for evaluation of loading controls. Primary antibodies were anti-acetylated lysine (1:1000, Cat. No. 9441S), GAPDH (1:10,000, Cat. No. 10R-G109A), and anti-O-linked N-acetylglucosamine (1:1000, Cat. No. 9875S) from Cell Signaling (Danvers, MA), Sigma Aldrich (Saint Louis, MO), Fitzgerald (Acton, MA) and Thermo Fisher Scientific (Lafayette, CO) respectively. Infrared fluorescently labeled secondary antibodies conjugated with IR dye 680 (Cat. No. 926–68070) and IR dye 800 (Cat. No. 926–32211) from LICOR Biosciences (Lincoln, NE) were used for detection using Odyssey CLx Imager. The experiments were repeated three times and one representative experiment is shown in Figure 1. In case of anti-O-linked N-acetylglucosamine HRP-linked IgM antibody was provided with the kit.

2D Western blots were performed by Kendrick Laboratories (Madison, WI). The blots were wet in 100% methanol, rinsed briefly in Tween-20 Tris buffered saline (TTBS), and blocked for two hours in 5% bovine serum albumin (BSA) in TTBS. The blots were then incubated in primary antibodies including anti-acetylated lysine (Cat. No. 9441S) and anti-O-GlcNAc monoclonal antibody (Cat. No. 9875S) from Cell Signaling diluted 1:1000 in 2% BSA in TTBS overnight and rinsed 3 × 10 minutes in TTBS. The blots were then placed in secondary antibody (anti-mouse IgG HRP, GE, Cat. No. NA931V) diluted 1:2,000 in 2% BSA in TTBS for two hours, rinsed in TTBS as above, treated with ECL, and exposed to X-ray film. The spots to be sequenced were excised and sent to Applied Biomics (Hayward, CA) for protein identification with MALDI-TOF-MS.

RWPE1 and LNCaP (30,000 to 40,000) cells were plated on the poly-D-Lysine (Sigma Aldrich-Saint Louis, MO) coated XL24 Seahorse bioanalyzer tissue culture 24 well plates. At 5 hours after plating, cells were incubated for 12 hours with 50% plasma or used as controls. Bioenergetics of LNCaP/RWPE1 was determined using the XF Cell Mito Stress Test Kit and a XF24-3 Analyzer (Seahorse Bioscience, North Billerica, MA) following published protocols [15]. Bioenergetics experiments were performed at the UCLA’s Cellular Bioenergetics Core Facilities. At least 48 repeated measurements were performed per experimental condition. The concentration of Oligomycin, Carbonyl cyanide 4-trifluoromethoxy phenylhydrazone (FCCP), Rotenone and Myxothiazol used was 0.5 μM, 0.75 μM, 0.75 μM and 0.75 μM for LNCaP and 0.75 μM, 0.25 μM, 0.75 μM and 0.75 μM for RWPE1 respectively.

To measure glucose and lipid uptake, fluorescently labeled glucose analogue and lipophilic dyes were used, respectively. All dyes were purchased from Molecular Probe (Life Technologies, Grand Island, NY). For glucose uptake measurement, 2-NBD glucose (Cat. No. N13195) at 12.5 μg/ml was incubated for 0 to 30 minutes at room temperature, washed and analyzed on a C6 Accuri flow cytometer (BD Bioscience, San Jose, CA). For lipid uptake measurement, either DiD dye (Cat. No. V-22887) at 1.25 μM or Bodipy FL C16 dye (Cat. No. D3821) at 1 μg/ml was used. Cells were incubated at room temperature for 0 to 30 minutes, then washed and analyzed with flow cytometry.

For flow cytometry, cells were first stained with FITC conjugated CD45 antibody (Cat. No. 555482, BD Bioscience), then fixed and permeabilized with Intrasure kit (Cat. No. 641776, BD Bioscience), and finally labeled with AlexaFluor647 conjugated pan-cytokeratin antibody (Cat. No. 4528, Cell Signaling, Danvers, MA). For fluorescent imaging, cytokeratin was visualized by staining first with unconjugated primary antibodies (Cat. No. 8018, Santa Cruz Biotechnology, Santa Cruz, CA), then with TRITC-conjugated secondary antibodies.

Vibrational frequency used for lipid bodies imaging was fixed at 2851 cm^-1 using a custom-built multimodal CARS microscope described previously [16]. CARS microscopy is a sensitive label-free method for visualization of lipid bodies [17, 18]. CARS lasers were also used for simultaneous two-photon fluorescence imaging. Epi-reflected signals were collected using a three-channel detector. Bandpass filters for FITC, TRITC, and CARS were 510/42 nm, 579/34 nm, and 736/128 nm, respectively.

Protein acetylation and glycosylation are nutrient-sensitive post-translational modifications important for the regulation of cellular energy metabolism [5]. Using 1D Western blots, protein lysine acetylation profiles of LNCaP cells were examined following incubation with 50% human plasma for 24 hours (Figure 1A). Surprisingly, there was no significant change to protein acetylation profiles between untreated and plasma treated LNCaP cells. Compared to RWPE1 cells, LNCaP cells exhibited hyper-acetylation of proteins with low molecular weight. In contrast, non-transformed prostate epithelial RWPE1 cells exhibited significant changes to the protein lysine acetylation profiles between untreated and plasma treated cells. Interestingly, plasma treated RWPE1 cells exhibited de-acetylation of low molecular weight proteins compared to untreated RWPE1 cells.

1D Western blots were employed to examine protein O-linked glycosylation of LNCaP cells following incubation with 50% human plasma for 24 hours (Figure 1B). With the exception of a slight increase in O-linked glycosylation of a protein band at 52 kD, there was no other significant change to protein O-linked glycosylation profiles between untreated and plasma treated LNCaP cells. In contrast, RWPE1 cells exhibited several significant changes to the protein O-linked glycosylation profiles between untreated and plasma treated cells. Most notable are the de-glycosylation of a protein band at approximately 60 kD and a protein band at 52 kD following incubation of RWPE1 cells in 50% human plasma.

2D Western blots were employed for high resolution analysis of protein lysine acetylation profiles (Figure 2A-D). Compared to RWPE1 cells, LNCaP cells exhibited protein hyper-acetylation with significantly more immuno-positive spots for proteins with low molecular weight of 30 kD or less (Figure 2A, C). Plasma treatment did not significantly change the protein lysine acetylation profile of LNCaP cells (Figure 2A, B). In contrast, plasma treatment reduced lysine acetylation of four low molecular weight protein spots of RWPE1 cells (Figure 2C, D). Selective 2D gel spots matching the positions of lysine acetylated proteins were excised and used for protein identification with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). Identified acetylated proteins and their biological functions are listed in Table 1 and Additional file 1: Table S1. Expectedly, a number of acetylated proteins participate in energy metabolism pathway including mitochondrial fatty acid β-oxidation and glycolysis. Other acetylated proteins participate in antioxidation, stress response, cytoskeletal structures, and other biological functions. Interestingly, four low molecular weight protein spots that got de-acetylated following plasma incubation in RWPE1 cells were identified as enoyl-CoA hydratase (mitochondrial fatty acid β-oxidation), mitochondrial carrier homolog 2 (transport/apoptosis), actin (cytoskeletal structure), and peroxiredoxin-4 (antioxidation). Many proteins identified did not migrate with predicted molecular weights or isoelectric points due to possible protein post-translational modifications or degradation.

2D Western blots were employed for high resolution analysis of protein O-linked glycosylation profiles (Figure 3A-D). Plasma incubation induced significant changes to the protein O-linked glycosylation profiles of LNCaP cells. Most notably are the de-glycosylation of protein spots 2 and 3 of ~14 kD and glycosylation of protein spot 7 of ~29 kD (Figure 3A, B). Plasma incubation also induced significant changes to the protein O-linked glycosylation profiles of RWPE1 cells. Most notably are the de-glycosylation of protein spot 9 of ~80 kD and the reduced glycosylation of protein spots 5 and 6 of ~60 kD (Figure 3C, D). In general, RWPE1 cells exhibited more O-linked glycosylated protein spots than LNCaP cells. Selective 2D gel spots matching the positions of O-linked glycosylated proteins were excised and used for protein identification with MALDI-TOF-MS. Identified O-linked glycosylated proteins and their biological functions are listed in Table 2 and Additional file 1: Table S2. For LNCaP cells, protein spots 2 and 3 were identified as histone 2B type 1-M, a component of the nucleosome. Protein spot 7 was identified as phosphoglycerate mutase 1, an enzyme of the glycolytic pathway. On the other hand, protein spot 9 of RWPE1 cells was identified as mitochondrial trifunctional enzyme subunit alpha, which is a critical enzyme for fatty acid β-oxidation pathway. Protein spots 5 and 6 were identified as keratin, type II cytoskeletal 6A and 6B, respectively. Proteomics data using 2D Western blots couldn’t identify any specific protein spot at 52 kD, whose O-linked glycosylation changed as a function of plasma incubation in 1D Western blot. It is likely that the glycosylation profile of the 1D gel band at 52 kD was the collective profile of multiple proteins with the same molecular weight.

Bioenergetics of LNCaP and RWPE1 cells were evaluated as a function of plasma incubation. Using an extracellular flux analyzer, mitochondrial function was evaluated by measuring in real time the oxygen consumption rates (OCR) and glycolytic function was evaluated by measuring in real time the extracellular acidification rates (ECAR) (Additional file 1: Figure S1A, B). Incubation with plasma slightly reduced ATP production of LNCaP cells while having no statistically significant effect on their maximal mitochondrial respiration capacity (Figure 4A, B). Incubation with plasma also slightly increased glycolysis of LNCaP cells while having no statistically significant effect on their maximal glycolytic capacity (Figure 4C, D). In contrast, incubation with plasma slightly reduced ATP production of RWPE1 cells while severely reduced their maximal mitochondrial respiration capacity by over 50% (p-value <0.05) (Figure 5A, B). Incubation with plasma also reduced both the average glycolysis rate and maximal glycolytic capacity of RWPE1 cells by over 30% (p-value <0.05) (Figure 5C, D).

Next, glucose and lipid uptake of LNCaP cells were evaluated using a fluorescent glucose analog and a lipid stain dye (DiD) and conventional flow cytometry. The difference between glucose and lipid uptake of LNCaP cells, RWPE-1 cells, and peripheral blood mononuclear cells (PBMC) was also evaluated to determine whether such difference could be exploited for their detection (Figure 6A, B). We found that plasma incubation repressed uptake of fluorescent glucose analog of more than 90% in LNCaP cells and less than 30% of RWPE1 cells (Figure 6C, Additional file 1: Figure S2, Additional file 1: Figure S3). Nearly 70% of PBMC exhibited uptake of the fluorescent glucose analog. In contrast, plasma incubation enhanced DiD staining of LNCaP cells by 14% (86% staining without plasma incubation versus 100% staining with plasma incubation) (Figure 6D, Additional file 1: Figure S4, Additional file 1: Figure S5). On the other hand, plasma incubation repressed DiD staining of RWPE1 cells by approximately 17%. Nearly 80% of PBMC stained positively with DiD (Figure 6B, Additional file 1: Figure S6). Our data revealed that plasma incubation repressed glucose uptake of LNCaP cells while enhanced their lipid uptake.

To improve lipid staining, DiD was replaced with Bodipy, a fluorescent fatty acid analog that undergoes native like transport and metabolism. Using flow cytometry, we found that both LNCaP cells (incubated over night with plasma) and PBMC achieved nearly 100% staining with Bodipy after 5 minutes of staining (Figure 7A). However, after fixation and permeabilization, Bodipy staining was retained for more than 90% of LNCaP cells compared to less than 20% of PBMC. A second stain with cytokeratin conjugated to Alexa Fluor 647 confirmed that up to 90% of LNCaP cells and only 5% of PBMC were stained positively for both Bodipy and cytokeratin following fixation and permeabilization protocol (Figure 7B, C). Clearly, while both LNCaP cells and PBMC exhibited strong affinity for Bodipy, only cytokeratin-positive LNCaP cells were able to significantly retain Bodipy after fixation and permeabilization.

Widefield fluorescent microscopy confirmed the ability to retain Bodipy staining after fixation and permeabilization of LNCaP cells and lack-there-of in PBMC (Figure 8A). Interestingly, Bodipy staining revealed granulated structures in the cytoplasm of LNCaP cells. Using coherent anti-Stokes Raman scattering (CARS) microscopy, a label-free imaging technique for lipid visualization [17], and two-photon fluorescent (TPF) microscopy, we observed the present of lipid bodies in LNCaP cells but not in PBMC (Figure 8B). Taken together, our data revealed that Bodipy was taken up by both LNCaP cells and PBMC along with plasma lipids. However, only in LNCaP cells did Bodipy and plasma lipids formed into lipid bodies. These lipid bodies remained within LNCaP cells even after cell fixation and permeabilization. In contrast, Bodipy and plasma lipids did not form into lipid bodies in PBMC, thus, escaped PBMC following cells fixation and permeabilization (Figure 8B & Figure 7D). The difference in the ability to form lipid bodies between LNCaP cells and PBMC permits clear differentiation of these cell types through the use of a fluorescent fatty acid analog Bodipy.

Both RWPE1 and LNCaP cells accumulated intracellular lipid bodies after incubation with plasma (Additional file 1: Figure S7) [19]. The presence of lipid bodies caused lipophilic stains, such as Bodipy (data not shown) and Oil Red O, to remain even after cell fixation in both RWPE1 and LNCaP cells. Therefore, the use of lipid staining together with cell fixation and permeabilization did not permit the differentiation of RWPE1 from LNCaP cells.