TRA-1-60⁺, SSEA-4⁺, Oct4A⁺, Nanog⁺ Clones of Pluripotent Stem Cells in the Embryonal Carcinomas of the Ovaries

Abstract

Introduction

Embryonal carcinoma of the ovary (ECO), pure or admixed to other tumors, is the deadly gynecological cancer.

Specific aim

The specific aim of this work was identification, isolation, clonal expansion, and molecular profiling of the pluripotent cells in the embryonal carcinomas of the ovaries.

Patients. Methods

The samples were acquired from the patients, who were clinically and histopathologically diagnosed with the advanced, pure embryonal carcinomas of the ovaries. The cell surface display of the TRA-1-60 and SSEA-4 was analyzed by flow cytometry (FCM), immunoblotting (IB), multiphoton fluorescence spectroscopy (MPFS), nuclear magnetic resonance spectroscopy (NMRS), and total reflection x-ray spectroscopy (TRXFS). The transcripts of the Oct4A and Nanog were analyzed by qRTPCR and MPFS and the products by MPFS. The human pluripotent, embryonic stem cells (ESC), human pluripotent, embryonal carcinoma of the testes (ECT), healthy tissues of the ovary (HTO), healthy tissue of testes (HTT), peripheral blood mononuclear cells (PBMC), and bone marrow mononuclear cells (BMMC) served as the controls.

Results

The studied embryonal carcinomas of the ovaries (ECOs) contained the cells with the strong surface display of the TRA-1-60 and SSEA-4, which was similar to the pluripotent ESC and ECT. Their morphology was consistent with the histopathological diagnosis. Moreover, these cells showed strong expression of the Oct4A and Nanog, which was similar to the pluripotent ESC and ECT. The ECO cells formed embryoid bodies, which differentiated into ectoderm, mesoderm, and endoderm. These cells were induced to differentiate into muscles, epithelia, and neurons.

Conclusion

Herein, we revealed presence and identified molecular profiles of the clones of the pluripotent stem cells in the embryonal carcinomas of the ovaries. These results should help us with refining molecular diagnoses of these deadly neoplasms and design biomarker-targeted, patient-centered, personalized therapy.

Introduction

Cancer was the primary cause of deaths for women between the ages of 20-85 in the USA in 2010 [1]. Among them, more than 21,880 women were diagnosed with cancers of the ovaries (COs) and 13,850 of them died of that cause. The essential factor for the longest survival was the earliest diagnosis of cancer and prompt therapy. This was well reflected in the statistics, which showed the 84.1% 10-year survival rate for women diagnosed at the FIGO's early clinical stage Ia, but down to the 10.4% 10-year survival rate for those diagnosed at the advanced stage III [2,3]. Unfortunately, 63% of women were diagnosed after the cancers have already progressed to the advanced stages. These findings have branded ovarian cancer “silent killer”, as the most deadly among all gynecological neoplasms [1,2].

At the early stages of this disease, women may not feel any symptoms. Later, the symptoms may include transient abdominal discomforts, bloating, pains, urinary urgencies, as well as other symptoms non-specific for genital system [4-6]. These symptoms prompt them to visit physicians followed by referrals to clinical laboratories.

The lab test of choice is measurement of CA125 in sera of the patients; however, its poor sensitivity prompts efforts to seek other biomarkers by analysis of proteins, microRNA, and circulating tumor cells (CTCs) [7-18]. Initial screenings with ultrasonography (USG) are very promising; however, its relatively poor resolution leads to follow-up imaging with high resolution CT or MRI, which do not yet define the cancer cells’ lineages [19-21]. Cancer progression is defined through clinical staging according to FIGO, whereas after initial progression in situ, which is denoted as the stage I, the cancers grow as pelvic masses. From the moment of the cells’ break out into the peritoneal cavity, they become detected in ascites, which is specifically denoted as the stage Ic. Thereafter, the cancer cells invade the pelvic organs - stage II and subsequently metastasize to distant organs denoted as the stages III and IV [3]. The final diagnosis is based upon histopathology, which identifies the tumor cells’ lineages. Almost 90% of the ovarian neoplasms have epithelial origins. Although rare, the germ cell tumors (GCTs) are very malignant. Among them, pure or admixed embryonal carcinomas of the ovary (ECO) are most deadly malignant tumors [2, 3, 22-23]. Moreover, they are most difficult to diagnose with lab tests, since they do not secrete AFP and hCG, as the other GCTs. The ECO cells retain morphological features of pluripotent, undifferentiated, embryonic cells in the pure ECOs and in admixes to other compound tumors. However, they frequently differentiate into teratomas, which resemble various somatic cell lineages. These characteristics make patomorphology based diagnoses difficult. They make diagnoses even harder in cases of anaplastic tumors. Therefore, molecular profiling of these cells should help not only with distinction between the epithelial and germ cell tumors, but also with search for clones of therapy resistant stem cells, as essential steps towards targeted, personalized therapies [24-30].

Several biomarkers were identified as biomarkers of multipotent cells in epithelial ovarian cancers (EOC), including standard and variants of the CD133, CD44, MyD88, and EpCAM [31-42]. However, none of them identified pure populations of the pluripotent stem cells, nor defined molecular profiles of the germ cell tumors of the ovaries. Moreover, in our hands, sorting with the aid of those markers resulted in heterogeneous populations of the cells; thus wide varieties of molecular profiles and biological properties. The neoplasms, which could be identified as the closest to the embryonal carcinomas of the ovaries, were the embryonal carcinomas of the testes [22-23, 43-58]. Testicular, extragonadal, and ovarian embryonal carcinomas, all share the same morphology. Moreover, pure embryonal carcinoma cells of the testes have pluripotency equal to that of the human embryonic stem cells (hESC) [43-50]. This included their ability for self-renewal and differentiation into the three germ lineages. The ECO cells have not yet been characterized in this regard.

Our interest was focused on the stem cells biomarkers: TRA-1-60 and SSEA-4 [43-58]. They were defined as the stem cells’ hallmarks of pluripotency. They were shown to be the unique biomarkers of the pluripotent testicular embryonal carcinoma cells, which ceased to express upon their differentiation. Moreover, TRA-1-60 was identified on cells in sections from the ECT biopsies [57]. It was also detected being shed into blood of the patients with the ECT [14, 58]. However, it was not displayed on the healthy, differentiated cells. Similarly, SSEA-4 was uniquely expressed on the undifferentiated, pluripotent ESCs’ and ECTs’ cells only, but completely absent from the cells upon their differentiation [45, 48]. These biomarkers were never tested on the ECOs. Furthermore, of our great interest was also the triad of the transcription factors being indicators of the cells’ pluripotency: Oct4, Sox2, and Nanog [59-62]. In particular, from at least three alternative splicing transcripts (Oct4A, Oct4B, Oct4B1) and four alternative translation isoforms (Oct4A, Oct4B-190, Oct4B-265, and Oct4B-164), only Oct4A was shown to be the master switch of pluripotency. While strong expression of the Oct4 was measured in the ESCs and ECTs, it was not detected in the healthy tissues. Molecular profiles of the pure embryonal carcinomas of the ovary (ECO) have not yet been defined.

Accordingly, the specific aim of this work was identification, isolation, clonal expansion, and molecular profiling of the pluripotent cells in the embryonal carcinomas of the ovaries using aforementioned biomarkers.

Materials and Methods

Patients. Samples

All the samples were obtained in accordance with the Declaration of Helsinki with the Patients’ Informed Consent and with the Institutional Review Boards’ approval. The samples from the patients, who were being clinically and histopathologically diagnosed with the germ cell tumors (GCT): women with ovarian GCTs (n=43); men with testicular GCTs (n=103); women (n=3) or men (n=3) with the extragonadal GCTs were included into this study. All the samples were encoded to protect the patients’ privacy. The six cases of the pure embryonal carcinomas of the ovaries (ECO) were selected for this study. Collection of the samples from the pelvic mass, ascites, metastases, healthy tissue margins, bone marrow, and blood was performed according to the standard surgical procedures. The batches of the samples were either immediately labeled with the single or dual chain variable fragment (Fv) antibodies, or incubated for cultures / clonal expansion, or rapidly cryoimmobilized or chemically preserved as described [63-70].

Genetically engineered single or dual chain variable fragment (Fv) antibodies

The single or dual chain variable fragment (Fv) antibodies were prepared as described previously [29, 48, 69-70], thus only briefly outlined below. The pooled B cells from the patients suffering from cancers were used to isolate mRNA, reverse transcribe, and create the cDNA libraries of complementarity determining regions (CDR) and framework regions (FWR) for anti-cancer-antibodies (ACA) coding sequences. The cds, after insertion into the plasmids containing chelates’ harboring coding sequences under the CMV promoters, were propagated and expressed in human myelomas as described (Fv clones TRA-1-60 ₂₄, SSEA-4 ₃₇ were used in this project). The TRA-1-60 and SSEA-4 antigens were purified, which followed by modification with biotin or digoxigenin. The modified TRA-1-60 or SSEA-4 antigens were anchored onto anti-biotin or anti-dig saturated pans and served as baits for selection of the Fv clones from the ACA libraries. The chelates were saturated with Gd, Tb, Eu. The elemental compositions were quantified with the scanner S2 Picofox (Bruker AXS, Fitchburg, WI). The fluorescent properties were measured with the RF-5301PC spectrofluorometer (Shimadzu, Tokyo, Japan). The relaxivities were measured on the DMX 400 WB or AVANCE II NMR spectrometers (Bruker Optics, Dallas, TX). The specificity and sensitivity of the Fvs were tested with the EELS and EDXS [67]. The monoclonal antibodies targeting TRA-1-60 and SSEA-4 on the ESC and ECT served as the controls as characterized [45].

Cultures and Clonal Expansion of Embryonal Carcinoma Cells of the Ovaries

The cancer cells were grown in semi-fluid cultures as described and thus only briefly outlined here [43-44, 63]. Cell clusters were separated into single cell suspension by short treatment with the PIPES buffered DNase, RNase, hyaluronidase, trypsin, and collagenase. The cells were isolated by MACS as described [29]. The substrates for cultures consisted of two layers of 1.2% and 0.6% matrigel in DMEM medium supplemented with 20% human serum, 2 mM L-glutamine, 100 U/mL of Penicilin, 100ug/mL of Streptomycin, 1% nonessential amino acids. The cultures were maintained in the incubator at 37°C with controlled 0.1% O2, 5% CO2, 94.9% N2 and saturated humidity environment. The cultures of the human embryonic stem cells - hESC H1, H7, H9 and human metastatic embryonal carcinoma of the testis cells - hECT NT2D1, muscle cells - RD, brain neuronal cells - HCN-2 cultures were the controls as originally described (ATCC, Manassas, VA) [45]. Embryoid bodies were cultured as described originally for the testicular carcinomas [43-44]. Moreover, the healthy ovarian tissue (HTO) from prophylactic oopherectomy or margins surrounding the tumor, healthy tissue of testes (HTT) dissected during orchiectomy, healthy margins of the brain removed during the brain surgery, and healthy margins of cardiac tissue excised during cardiac surgery were the controls. For testing differentiation of the embryoid bodies into the three germ layers, they were labeled with the antibodies for myosin – as the marker of the mesoderm formation, keratin – as the marker of the endoderm formation, and 68kDa neurofilaments protein- as the marker of the ectoderm formation. For inducing differentiation, the cells were exposed to 10^-5 M retinoic acid, 1 % DMSO, 3 mM hexamethylene bisacetamide (HMBA), 1% dimethylsulphoxide (DMSO), 250 ng/ml nerve growth factor, epidermal growth factor, or vascular endothelial growth factor as described (Sigma-Aldrich, Saint Louis, MO, USA) [51-52]. The evaluation of differentiation was based upon labeling with the Fvs anti-myosin - to validate differentiation into muscle, anti-NF68 neurofilaments – to validate differentiation towards neurons, and anti-CK18 cytokeratins – to validate differentiation towards epithelia [48].

Cryo-immobilization

The cells were injected into the sterile, gold planchettes, loaded, and cooled down to –196°C within 10 ms at 2000 atm within the high pressure freezing machine HPM10 (Bal-Tec, Balzers, Liechteinstein) [67]. Alternatively, they were frozen with the melting ethane-based cryo-immobilization machine (the cryo-station built with the NSF funds – PI: Dr M. Malecki). The frozen samples were promptly transferred to and stored in liquid nitrogen dewars. For cultures, the samples were thawed directly into the culture media. For ultrastructural studies, they were prepared by freeze-substitution in the FSD010 (Bal-Tec, Balzers, Liechteinstein), low temperature embedding, and cryo-sectioning or freeze-fractured and freeze-dried on the BAF400 (Bal-Tec, Balzers, Liechteinstein). Evaluation of the freezing procedure was pursued on the cryo-sections cut on the cryo-ultratome (Leica, Buffalo Grove, IL) and transferred within the cryo-transfer system on the energy filtering system 912 Omega (Zeiss, Oberkochen, Germany) to assure absence of ice segregation determined by checking diffraction patterns.

Chemical fixation

For the ultrastructural overview, the cells were fixed by adding up to 4% formaldehyde, 0.1% glutaraldehyde, 0.2 % tannic acid (Pella, Redding, CA) [44]. Thereafter, they were either frozen, freeze-substituted, embedded in epon 812 (Ladd, Williston, VT), and sectioned on the ultratome (Leica, Buffalo Grove, IL) or rinsed in the physiological buffer attached to the silane coated gold or carbon carriers, dried in the critical point of CO₂, and cryo-sputter coated with a monomolecular layer of Cr or Pt/C on the BAF400 (Bal-Tec, Balzers, Liechteinstein).

Immunoblotting (IB)

The cells were either frozen in liquid nitrogen, crushed, and thawed or/and disintegrated with ultrasonicator (Branson Ultrasonic, Danbury, CT) within the sample buffers for native protein analysis. They were stored in liquid nitrogen or electrophoresed in the native buffer (Invitrogen, Carlsbad, CA). They were vacuum or electro-transferred onto the PVDF membranes (Amersham, Buckinghamshire, UK). The membranes carrying transferred proteins were soaked within human serum and labeled with the Fvs. The samples of muscle myosin, neuronal NF68, and keratin CK18 served as the controls. The images of the blots were acquired and quantified with Fluoroimager (Molecular Dynamics, Sunnyvale, CA) or Storm 840 (Amersham, Buckinghamshire, UK). The levels of the products were also calculated as the ratio between the protein concentration in the examined patient's cells and the controls.

Quantitative Reverse Transcription and Polymerase Chain Reaction (qRTPCR)

Nucleic acids were isolated using the Nucleic Acid Extractor Model 340A (ABI). The total isolated mRNA served as the template to generate cDNA through reverse transcription using random hexamers and reverse transcriptase (ABI, Foster City, CA, USA) as described [29]. The cDNAs’ and amplicons’ quality was tested by polymerase chain reaction of beta actin and GAPDH as the reference genes (ABI, Foster City, CA). For evaluation of the gene expression levels for Oct4A, Oct4B, Oct4B, the primers sets were designed using Primer Express (ABI, Foster City, CA) based upon the sequences imported from the Public Domain GenBank (NCBI), and synthesized on the 380A DNA Synthesizer (ABI, Foster City, CA). The PCR and qPCR reactions were carried using the mix of the cDNA, the synthesized primers, dNTPs, and Taq DNA polymerase (Hoffmann–La Roche, Basel, Switzerland) on the Robocycler (Stratagene, San Diego, CA), Mastercycler (Eppendorf, Hamburg, Germany), or 7500, 7900 HT qPCR systems (ABI, Foster City, CA). The images of the gels were acquired and quantified with Fluoroimager (Molecular Dynamics, Sunnyvale, CA) or Storm 840 (Amersham, Buckinghamshire, UK). The levels of the transcripts were first normalized against GAPDH or actin and thereafter calculated as the ratio between the transcript concentration in the examined patient's cells versus the cells from the HTO, HTT, BMMCs, PBMC, hESCs, or hECT cells.

Flow cytometry (FCM). Fluorescently activated cell sorting (FACS). Multiphoton Fluorescence Spectroscopy (MPFS)

The cell clusters were thoroughly disintegrated. The negative selection involved depletion of white blood cells with the superparamagnetic Fvs anti-CD45, CD34; the apoptotic cells were removed with the Fvs against phosphatidylserine, the dead cells were eliminated with the Fvs against dsDNA [70]. The remaining samples were further enriched by the positive selection with the superparamagnetic Fvs for TRA-1-60 or SSEA-4. The side populations were determined with the Hoechst 33342 in Verapamil tests (Sigma-Aldrich, Saint Louis, MO). The enriched populations of cells labeled with the fluorescent Fvs targeting TRA-1-60 or SSEA-4 were measured with the Calibur, Vantage SE, or Aria (Becton-Dickinson, Franklin Lakes, NJ) or the FC500 (Beckman-Coulter, Brea, CA). The fluorescently labeled cells were imaged with the Axiovert (Zeiss, Oberkochen, Germany) equipped with the Enterprise argon ion (457 nm, 488 nm, 529 nm lines) and ultraviolet (UV) (364 nm line) lasers; Odyssey XL digital video-rate confocal laser scanning imaging system operated up to 240 frames/s under control of Intervision software (Noran, Madison, WI), and the Diaphot (Nikon, Tokyo, Japan) equipped with the Microlase diode-pumped Nd:YLF solid state laser (1048 nm line), the pulse compressor with the pulses’ rate 300 fs at 120 MHz and the MRC600 scanning system under control of Comos software (the multi-photon fluorescence station built with the NIH funds – PI: Dr J. White). Images were deconvolved after their import to the Indy workstation (Silicon Graphics, Fremont, CA).

Total Reflection X-ray Fluorescence Spectroscopy (TRXFS)

In this study, the ICP standard of 1000 mg/l of mono-element Gallium (CPI International, Denver, CO) was added to 500 microL of each sample to the final concentration of 10 mg/l. The data were generated from the S2 Picofox TXRF spectrometer equipped with a molybdenum (Mo) X-ray target and the Peltier cooled Xflash Silicon Drift Detector (Bruker AXS, Fitchburg, WI). Scan times ranged up to 1000 seconds. The automatic sample changer, which can hold up to 25 samples was also used along with the SPECTRA 7 software for the instrument control, data collection, and analysis (Bruker AXS, Fitchburg, WI).

Nuclear Magnetic Resonance (NMR). Magnetically Activated Cell Sorting (MACS)

The cells were labeled for positive selection with the superparamagnetic Fvs targeting TRA-1-60 and SSEA-4, and for the negative selection targeting CD45, CD34, dsDNA, and PS, while suspended in the physiological buffer supplemented with serum and glucose [68]. The small aliquots were dispensed into the magnetism-free NMR tubes (Shigemi, Tokyo, Japan). The relaxation times T1 were measured in resonance to the applied FLAIR pulse sequences on the NMR spectrometers: DMX 400 WB or AVANCE II NMR (Bruker, Billerica, MA) or the Signa clinical scanners (GE, Milwaukee, WI). The superparamagnetic Fvs were also used to isolate the labeled cells from the solution using the 1.5 T magnetic sorter (the sorter designed and built based upon the NSF funds – PI: Dr M. Malecki) [29].

Electron Energy Loss Spectroscopy (EELS). Energy Dispersive X-Ray Spectroscopy (EDXS)

The samples, which were cryo-immobilized presented the life-like supramolecular organization. Molecular imaging was pursued as described [67]. The field emission, scanning transmission, electron microscope FESTEM HB501 (Vacuum Generators, Kirkland, WA) was equipped with the energy dispersive x-ray spectrometer (EDXS) (Noran, Middleton, WI) and post-column electron energy loss spectrometer (EELS) (Gatan, Pleasanton, CA). The cryo-energy filtering transmission electron microscope 912 Omega was equipped with the in-column, electron energy loss spectrometer (EELS) (Zeiss, Oberkochen, Germany). The cryo-energy filtering transmission electron microscope 430 Phillips was equipped with the post-column, electron energy loss spectrometer (EELS) (Noran, Middleton, WI, USA) and the energy dispersive x-ray spectrometer (EDXS) (Noran, Middleton, WI). The images were acquired using the ccd camera operating under the image acquisition and processing software (SIS, Herzogenrath, Germany, EU or Emispec Systems, Tempe, AZ, USA).

Statistical analysis

Fisher's exact test was used to examine the association of the gene expression between the human ECO cells versus the controls: cultured human embryonal carcinoma cells of the testis (NT2D1), embryonic stem cells (H1, H7, H9), healthy tissue of the ovaries (HTO), healthy tissue of the testis (HTT), peripheral blood mononuclear cells (PBMC), or bone marrow mononuclear cells (BMMC). Average gene expression measurements were run in triplicates for each patient, which were used for gene expression statistical analysis. For the comparisons, Wilcoxon signed rank test was used, and Wilcoxon rank sum test for the comparison of two independent groups of the ECO versus the ECT and ESC cells. A two-sided p-value was computed in each comparison. The graphs were displayed using GraphPad software (GraphPad Software, Inc, La Jolla, CA).

Results

The cell populations from biopsies were enriched by negative selection in MACS with the superparamagnetic Fvs targeting CD45, CD34 to deplete the samples of the white blood cells (WBC), with the Fvs against dsDNA and phosphatidylserine to deplete from the dead and apoptotic cells. They were further enriched by positive selection with the superparamagnetic Fv targeting TRA-1-60 and SSEA-4. As the first step of the study, the evaluation of the surface displayed molecules on the ECO versus the ESC and ECT cells was performed on the cells labeled with the fluorescent Fv against TRA-1-60 and imaged with multiphoton fluorescence microscopy as shown in the figure 1. The labeling of the cells is very intense and specific. The background is clean. The positive controls, the cells of cultures of the hESC H7 and hECT NT2/D1, were identified by the same labeling protocol. This morphological evaluation was followed by the statistical analysis of the labeling kinetics presented in the figure 2. The TRA-1-60 display on the ECO cells was statistically equivalent or higher relative to the ESC and ECT cells. The TRA-1-60 display was not detected on the PBMCs and BMMCs.

Figure 1. Display of TRA-1-60 on the ECO cells.

The ECO, ESC, and ECT cells were labeled with the fluorescent scFv targeting TRA-1-60. They were imaged with multiphoton fluorescence. The kinetics of labeling was determined upon the digital quantum intensity profiles presented in the figure 3. Intensity and pattern of labeling of the ECO cells (A) is identical to the ESC (B) and ECT (C) cells. Magnification: 1,000 x.

Figure 2. Quantified kinetics of the TRA-1-60 display on the ECO cells.

The images of the ECO, ESC, ECT cells acquired as described in the figure 2 were quantified by measuring pixel intensity values across the cells’ diameters to reveal the kinetics of the scFv display. They show identical kinetics of labeled the ECO (A), ESC (B), and ECT (C) cells. They also demonstrate the identical, high ratio between the nucleoplasm and cytoplasm is all these cells.

Homogenous populations of cells are really necessary to determine precisely their molecular profiles. For that reason, the populations were enriched by sorting and their purities were evaluated by flow cytometry as shown in the figure 3. It revealed that the enrichment resulted in more than 99% purity of the sorted cell populations. The statistical analysis of enrichment was presented in the figure 4. Sorted populations of the ECO cells appeared to have identical display profiles to those of the hESC and hECT cells used as the positive controls. The populations of the PBMC and BMMC cells, which served as the negative controls showed no specific labeling. Purity of the enriched populations of the cells and labeling intensity values were contingent upon the specificities of labeling for the TRA-1-60. These were measured at the full width at half-maximum (FWHM) for each run, averaged, and calculated for statistical significance. The statistical significance was approved at the p < 0.001.

Figure 3. TRA-1-60⁺ enriched populations of the ECO cells.

The ECO cells were labeled with the fluorescent scFv targeting TRA-1-60 after population enrichment selection by MACS and studied by flow cytometry. Labeling: the patients’ ovarian embryonal carcinoma cells (A-F), the human embryonic stem cells (G) and the patients’ bone marrow mononuclear cells (H). Statistical significance was accepted for the p < 0.001 with the values of the FWHM between the patients’ readings and the negative controls, while there was no significant statistical difference in the intensity between the patients’ cells and the positive controls as quantified in the figure 4.

Figure 4. Purity of the enriched TRA-1-60⁺ ECO cells.

The samples were enriched by negative selection with the superparamagnetic scFv targeting CD45, CD34, CD14, CD3, CD4, CD16, CD19, CD20, dsDNA, phosphatidylserine. That was followed by positive selection with the superparamagnetic scFv targeting TRA-1-60. This resulted in the sorts’ purities > 99% with the statistical significance p < 0.001. Labels: ECO – patients’ samples (n=6); PBMC – peripheral blood mononuclear cells (n=6); BMMC – bone marrow mononuclear cells (n=6); ESC – cultures of the human embryonic stem cells H1, H7, H9 (n=3); ECT -cultures of the human testicular embryonal carcinoma cells NT2D1 (n=1).

For the further analysis of the cell surface displayed molecular profiles, the cells were rapidly cryoimmobilized. This followed by homogenization, electrophoresis, and transfers onto the PVDF membranes. The molecules on membranes were labeled with the Fv targeting TRA-1-60 and imaged as illustrated in the figure 5. The labeling is very strong and specific. There is no other labeling along the lanes except the single, strong, specific bands. The labeling is identical to that seen for the molecules gained from the ESC and ECT cells used as the positive controls. The PBMC and BMMC cells, which served as the negative controls, did not absorb any labeling. The statistical analysis was performed by quantification of the labels on the blots as presented in the figure 6. The intensity of the surface display is equivalent or higher relative to the ECT and ESC cells included as the positive controls. The PBMC and BMMC cells’ labeling was not detected.

Figure 5. TRA-1-60⁺ display on the ECO cells.

The ECO cells were disintegrated, electrophoresed, transferred onto the PVDF membranes, and labeled with the Fvs targeting TRA-1-60. Labels: ECO – patients’ samples (encoded A-F); ESC – culture of the human embryonic stem cells H7 (G); ECT -culture of the human testicular embryonal carcinoma cells NT2D1 (H); PBMC – peripheral blood mononuclear cells; BMMC – bone marrow mononuclear cells (empty lanes not shown). The samples were run and quantified in triplicates. All the immunoblots were quantified as presented in the figure 6.

Figure 6. Quantified ECO cell surface display of TRA-1-60.

The pixel intensity values were measured on the TRA-1-60 scFv labeled immunoblots of the cell lysates as described in the figure 5. Labels: ECO – patients’ samples (n=6); PBMC – peripheral blood mononuclear cells (n=6); BMMC – bone marrow mononuclear cells (n=6); ESC – cultures of the human embryonic stem cells H1, H7, H9 (n=3); ECT -cultures of the human testicular embryonal carcinoma cells NT2D1 (n=1). Statistical significance accepted at the p < 0.001. The samples were run and quantified in triplicates.

The analysis of the ECO display molecular profiles was also performed by immunoblotting with the Fvs targeting SSEA-4 as illustrated in the figure 7. The labeling of the ECO cells is again very strong and specific. There is no label in the background. The labeling pattern is identical to that of the ESC and ECT cells. The PBMC and BMMC lanes had no labels on them. The statistical analysis is presented in the figure 8. The statistical significance was approved at the p < 0.001. It confirms the data from immunoblots that the display of the SSEA-4 on the ECO cells is statistically equivalent or higher relative to the ESC and ECT cells. Having the specific Fv validated in classical procedures, these Fvs were further used for initial and follow up screening for the presence of the TRA-1-60 and SSEA-4 positive cells performed with the NMR and TRXF. Measurements of the concentrations of the cell surface receptors were pursued using nuclear magnetic resonance (NMR). That was possible through measurements of the relaxation times T1 induced by labeling of the human, ovarian embryonal carcinomas with the superparamagnetic Fvs. Labeling of the cultured human embryonic stem cells hESC H1, H7, H9, and testicular ECT cells NT2D1 were again the positive controls. The PBMC and BMMC cells were negative controls. On average, the relaxation times of the samples labeled with the superparamagnetic Fvs for TRA-1-60 or SSEA-4 were falling down to milliseconds, i.e., two orders of magnitude lower, than the relaxation times of unlabeled cells or cells’ depleted sera. These data were further validated by measurements of the receptors’ density by EDXS or TRXFS. The concentrations of Gd, Tb, or Eu were indicative of the number of chelated Fvs, thus the number of receptors per cell. This way of measuring the cell receptors was far less cell traumatic, faster, simpler, more sensitive alternative than running the measurements by FCM or RIA. The sensitivity of these methods was similar to those of autoradiography and scintillation, but far safer in the laboratory practice.

Figure 7. SSEA-4⁺ display on the ECO cells.

The ECO cells were disintegrated, electrophoresed, transferred onto the PVDF membranes, and labeled with the Fvs targeting SSEA-4. Labels: ECO – patients’ samples (encoded A-F); ESC – culture of the human embryonic stem cells H7 (G); ECT -culture of the human testicular embryonal carcinoma cells NT2D1 (H); PBMC – peripheral blood mononuclear cells; BMMC – bone marrow mononuclear cells (empty lanes not shown). The samples were run and quantified in triplicates. All the immunoblots were quantified as presented in the figure 8.

Figure 8. Quantified ECO cell surface display of SSEA-4.

The pixel intensity values were measured on the SSEA-4 scFv labeled immunoblots of the cell lysates as described in the figure 7. Labels: ECO – patients’ samples (n=6); PBMC – peripheral blood mononuclear cells (n=6); BMMC – bone marrow mononuclear cells (n=6); ESC – cultures of the human embryonic stem cells H1, H7, H9 (n=3); ECT -cultures of the human testicular embryonal carcinoma cells NT2D1 (n=1). Statistical significance accepted at the p < 0.001. The samples were run and quantified in triplicates.

Transcripts of Oct4A – the main transcription factor controlling pluripotency, were studied by extraction and quantification of the total RNA followed by reverse transcription (RT) and quantitative polymerase chain reaction (qRTPCR). Specificity of the amplification was determined on agarose gels as illustrated in the figure 9. It shows the clean bands of the Oct4A transcripts, after being reverse transcribed and amplified. The lanes carrying amplicons for the ECO cells carry identical bands to those for the hESC and hECT cells, used as the positive controls. The lanes carrying amplicons for the BMMC and PBMC were empty. This quality control on the gels prior to qPCR helped to eliminate risks of including for sequencing the products of mispriming and mutations. The gels were scanned and digitized, which was followed by quantitative analysis. Quantification of the gene expression was also accomplished with the RT qPCR as shown in the figure 10. Quantities of the Oct4A transcripts were normalized against the beta actin and GAPDH transcripts. Thereafter, the transcripts’ levels were compared between the patients’ versus hESC and hECT, or versus BMMC and PBMC. These measurements revealed the statistically equivalent or higher relative to levels of transcripts in the ECO cells relative to the hECT and hESC cells. The amplifications in the PBMCs and BMMCs were below the thresholds. The data were approved with the meaningful statistical significance at the value p < 0.001. Products of the Oct4A translation were determined on the cells labeled with the Fvs as illustrated in the figure 11. The Oct4A transcription factors localized into the nuclei within the ECO cells on the identical manner, as in the ESC and ECT cells. Transcripts of Nanog were analyzed the same way. After reverse transcription and amplification, the amplicons were electrophoresed and imaged as shown in the figure 12. The amplicons from the ECO samples were identical to those from the ECT and ESC. The PCR of the PBMC and BMMC did not result in detectable products. The levels of transcripts were also quantified by the RT qPCR relative to the GAPDH or actin, while being compared to the ESC and ECT cells as a reference. The figure 13 demonstrates that the ECO cells expressed equivalent or higher levels of the transcripts relative to the ESC and ECT. These transcripts were not detected in the PBMC and BMMC. The data were accepted with the statistical significance the value p < 0.001.

Figure 9. OCT4A gene expression in the ECO cells.

The RNA from the ECO cells was reverse transcribed into cDNA, which was then amplified by polymerase chain reaction with the primers for OCT4A. Labels: ECO – patients’ samples (encoded A-F); ESC – culture of the human embryonic stem cells H7 (G); ECT -culture of the human testicular embryonal carcinoma cells NT2D1 (H); PBMC – peripheral blood mononuclear cells; BMMC – bone marrow mononuclear cells (no amplicons - empty lanes not shown). Intentionally the entire lanes are shown to demonstrate specificity of amplification with the entire absence of mispriming. All the gels of amplicons were normalized and calculated with the statistical significance at the P < 0.001. The samples are representative for all amplifications. The samples were run in triplicates.

Figure 10. Quantitative analysis of the OCT4A expression in the ECO cells.

The RNA from the cells was reverse transcribed into cDNA, which was then amplified by RT qPCR with the primers for the OCT4A. The amplification was calculated in relation to GAPDH and compared relative to the ESC expression. Labels: ECO – patients’ samples (n=6); PBMC – peripheral blood mononuclear cells (n=6); BMMC – bone marrow mononuclear cells (n=6); ESC – cultures of the human embryonic stem cells H1, H7, H9 (n=3); ECT -cultures of the human testicular embryonal carcinoma cells NT2D1 (n=1). Statistical significance was accepted at the p < 0.001. The samples were run in triplicates.

Figure 11. OCT4A gene expression products in the nuclei of the ECO cells.

The cells were cryoimmobilized and labeled with the antibodies targeting OCT4A. The cells were imaged by multiphoton fluorescence microscopy. The OCT4A transcription factors are restricted to the cell nuclei. Labels: ECO – patient's sample (encoded A); ESC – culture of the human embryonic stem cells H7 (B); ECT -culture of the human testicular embryonal carcinoma cells NT2D1 (C); PBMC – peripheral blood mononuclear cells; BMMC – bone marrow mononuclear cells (no labeling – empty screens not shown).

Figure 12. Nanog gene expression in the ECO cells.

The RNA from the ECO cells was reverse transcribed into cDNA, which was then amplified by polymerase chain reaction with the primers for NANOG. Labels: ECO – patients’ samples (encoded A-F); ESC – culture of the human embryonic stem cells H7 (G); ECT -culture of the human testicular embryonal carcinoma cells NT2D1 (H); PBMC – peripheral blood mononuclear cells; BMMC – bone marrow mononuclear cells (no amplicons - empty lanes not shown). Intentionally, the entire lanes are shown to demonstrate specificity of amplification with the entire absence of mispriming. All the gels of amplicons were normalized and calculated with the statistical significance at the P < 0.001. The samples were run in triplicates. The samples are representative for all amplifications.

Figure 13. Quantitative analysis of the NANOG expression in the ECO cells.

The RNA from the cells was reverse transcribed into cDNA, which was then amplified by RT PCR with the primers for Nanog. The amplification was calculated in relation to GAPDH and compared to the ESC expression. Labels: ECO – embryonal carcinoma of the ovary; BMMC – bone marrow mononuclear cells; PBMC – peripheral blood mononuclear cells; ESC – human embryonic stem cells; ECT – human embryonal carcinoma of the testis. The samples were run in triplicates. The data were accepted with the statistical significance p < 0.001.

Functional tests of the ECO cells’ pluripotency involved culturing embryoid bodies and checking their ability to differentiate into the three germ layers as summarized in the figure 14. The biomarkers of differentiation were clearly detected. Antibodies to the NF68 neurofilaments, cardiac muscle myosin, and CK18 cytokeratins reported differentiation towards ectoderm, mesoderm, and endoderm respectively. The pluripotency of the ECO cells was further confirmed after plating them into the matrigel matrix and inducing their differentiation. Thereafter, the cells were homogenized, electrophoresed, and transferred onto the PVDF membranes. The samples on the membranes were labeled with the Fvs and imaged as illustrated in the figures 16-18. The DMSO and HMBA induction of the ECO cells led to differentiation into the muscle, which was demonstrated by presence of very strong and specific bands of the cardiac muscle myosin, as shown in the figure 15. The specificity was validated by blots of the human cardiac muscle myosin and the cardiac muscle, which were used as the positive controls. The EGF induced the ECO cells to differentiate into the epithelia, as presented in the figure 16. The specificity of labeling was validated by the positive controls of the CK18 and the ovarian tissue lysate. The RA and NGF induced differentiation of the ECO cells into neurons. This resulted in the expression of the NF68 neurofilaments revealed as the strong band on the immunoblots in the figure 17. The purified NF68 and the brain tissue excised within the healthy margins during brain surgery facilitated the validation of the data.

Figure 14. Formation of embryoid bodies and differentiation into three main germ layers.

The embryoid bodies were tested by labeling with the antibodies targeting neurofilaments protein (68kDa) as the biomarker for the forming ectoderm; muscle myosin for the mesoderm; and cytokeratins for the endoderm. Labels: the embryoid bodies from the ECO cells (A-F), embryonic stem cells (H7) (G); (negative controls not included); +: positive reaction; -: negative reaction; L: lost sample.

Figure 16. Differentiation of the ECO cells into epithelium.

The ECO cells were induced to differentiate into the epithelium. The cells were homogenized, electrophoresed, and transferred onto the PVDF membranes. The transfers on the membranes were labeled with the antibodies against CK18 keratins. Labels: ECO – patients’ samples (encoded A-C); CK18 cytokeratin (D); healthy ovary tissue (E); PBMC – peripheral blood mononuclear cells; BMMC – bone marrow mononuclear cells (no labeling - empty lanes not shown).

Figure 15. Differentiation of the ECO cells into muscle.

The ECO cells were induced to differentiate into the cardiac muscles. The cells were homogenized, electrophoresed, and transferred onto the PVDF membranes. The transfers on the membranes were labeled with the antibodies against cardiac muscle myosin. Labels: ECO – patients’ samples (encoded A-C); cardiac muscle myosin (D); human cardiac muscle (E); PBMC – peripheral blood mononuclear cells; BMMC – bone marrow mononuclear cells (no labeling - empty lanes not shown).

Figure 17. Differentiation of the ECO cells into neurons.

The ECO cells were induced to differentiate into the neurons. The cells were homogenized, electrophoresed, and transferred onto the PVDF membranes. The transfers on the membranes were labeled with the antibodies against the NF68 neurofilaments. Labels: ECO – patients’ samples (encoded A-C); NF68 neurofilaments (D); human healthy brain tissue (E); PBMC – peripheral blood mononuclear cells; BMMC – bone marrow mononuclear cells (no labeling - empty lanes not shown).

Discussion

A few elements of biotechnology, which we applied here, are worth stressing. First, we used rapid cryo-immobilization to capture the living-like cells’ morphology and molecules’ native properties. It is a common misconception that freezing in liquid nitrogen offers the fast freezing rate associated with the better preservation of viability and antigenicity. As the matter of fact, due to the Leidenfrost's effect, the boiling around the sample delays cooling significantly. Therefore, we used the instruments, which provided a much faster freezing rate, thus assured instant preservation of molecules’ native properties. Second, it is a well known fact that the chemical fixation or slow freezing may dramatically change epitopes, so that the antibodies, which were developed for the native molecules may not work with the denatured proteins. Therefore, the antibodies used for Western blotting are rarely useful for native immuno-blotting. For these reasons, we preferred working with the native or rapidly frozen specimens, which offered retention of native properties; thus allowed for the direct extrapolation of the data towards in vivo environment. Third, FACS on fluorescently labeled cells inflicts a severe shearing stress due to pressurized tubes guiding the cells, which results in a very low viability in addition to the poor staining. Plans for clonal expansion, inducing differentiation, and drug testing, could be jeopardized by low viability of the isolated cells. Therefore, we engineered superparamagnetic variable fragment antibody biotags and a magnetic sorter, which allowed us to gently isolate the labeled cells by means of magnetic forces, thus assuring their high viability. Fourth, the small size of our Fvs helped in overcoming steric hindrance forces and assuring high density packing onto the receptors. Their high specificity and sensitivity resulted in heavy labeling of the receptors and practical absence of non-specific labeling of other cells, which translated in the high signal to noise ratios.

Using these technologies, we revealed the populations of the pluripotent, cancer stem cells in the embryonal carcinomas of the ovaries. Moreover, we demonstrated that the levels of gene expression for the Nanog and Oct4A transcription factors were equivalent or higher in the ECO cells, than they were in the cultured embryonic stem cells and testicular embryonal carcinoma cells. Importantly, they were statistically significantly higher, if compared to healthy control cells. These differences were detected with the highly sensitive means of detection possible, i.e., after cycles of amplification by PCR. The mononuclear cells from the same patients were showing no expression, while serving as the additional controls for these diagnostic tests. Therefore, they could be considered for applying on the standard FFPE sections or alternatively on cryo-sections.

Moreover, these data contribute to presenting hypotheses on the processes initiating neoplasms. First, neoplasms of the ovaries may develop de novo from one or a few pluripotent stem cells present in the ovaries. With the tumor progression, some of the cells partially differentiate, which may prompt determination of their lineages. If it happens, they may be used for immunohistopathological classification. Nevertheless, the stem cells are thought to be at the core of such processes in carcinoma in situ. Second, progression of the neoplasms leads to progressive accumulation of the mutations, which may turn on the transcription factors of pluripotency and cause reprogramming some cells into clones of stem cells. Third, therapies effectively regulate the selection process to eliminate sensitive clones, while leaving room for the resistant cancer cells or cancer stem cells. The cancer stem cells have to compete for spaces and nutrients with other cancer cells. These processes of selection may occur at the consecutive stages of the tumor progression and consecutive regimes of therapy.

These systemic therapy resistant cells should be the targets of specifically designed therapies. Moreover, some clones breaking away from the primary tumors may find suitable environment for advancing along the FIGO clinical stages I to II. Among them or from them, raise the clones capable to attach to the peritoneal cavity walls and break through it, what leads to the progression of the disease to the FIGO stages III and IV – metastases to lymph nodes and distant organs. In all these scenarios, capturing and detailed molecular profiling of the cancer cells is the most essential step in blocking progression of cancer.

The current approach in administering the therapies to the patients is based upon trials and errors, while watching for the responses and adjusting the therapeutic regimes ex iuvantibus. The novel approaches towards cancer molecular profiling should help us to craft the patient individualized therapy and pursue effective biomarker-targeted, patient-centered personalized therapy.

Conclusion

Herein, we revealed presence and identified molecular profiles of the clones of the pluripotent stem cells in the embryonal carcinomas of the ovaries. These results should help us with refining molecular diagnoses of these deadly neoplasms and design biomarker-targeted, patient-centered, personalized therapy.

Abbreviations

ECO

embryonal carcinoma of the ovary

GCT

germ cell tumor

ECT

embryonal carcinoma of the testis

TRA-1-60

tumor resistance antigen 1-60

SSEA-4

stage specific embryonic antigen 4

Oct4A

octamer-binding transcription factor 4A

Sox2

sex determining region Y-box 2

Nanog

homeobox transcription factor

scFv

single chain variable fragment antibody

dcFv

dual chain variable fragment antibody

FCM

flow cytometry

IB

immunoblotting

MACS

magnetically activated cell sorting

FACS

fluorescently activated cell sorting

NMRS

nuclear magnetic resonance spectroscopy

TRXFS

total reflection x-ray fluorescence spectroscopy

MPFS

multiphoton fluorescence spectroscopy

EELS

electron energy loss spectroscopy

EDXS

energy dispersive x-ray spectroscopy