Personalized Versus Precision Nanomedicine for Treatment of Ovarian Cancer

Abstract

The response to treatment is substantially varied between individual patients with ovarian cancer. This response largely depends on the resistance of individual tumors to the anticancer drug used for the treatment. However, chemotherapy treatment plans rarely pay sufficient attention to the mentioned factors. Instead, standardized treatment protocols are usually employed for most ovarian cancer patients. Variations in an individual’s sensitivity to drugs significantly limit the effectiveness of treatment in some patients and lead to severe toxicities in others. Human genetic makeup profoundly impacts the individual responses and side effects of drugs for each patient. A stratification of patients based on a genetic profiling of their drug resistance mechanisms (sometimes called "precision" medicine) substantially increased treatment efficacy. However, in most cases, such stratification is based on the measurement of specific average genetic characteristics in groups of cancer patients. Consequently, a patient with more than one mechanism of drug resistance or possessing unique genetic characteristics may receive sub-optimal treatment. In the present investigation, we developed and validated a nanotechnology-based approach for personalized treatment of ovarian carcinoma (the most lethal type of gynecological cancer) constructed on the individual genetic profile of the patient's tumor. Samples of the tumor were obtained from the patients during tumor debulking surgery. The expression of predefined genes and proteins was analyzed for each patient sample. Based on this analysis, several genes responsible mainly for the resistance to chemotherapy were selected. Finally, a mixture of the complex nanocarrier-based targeted delivery system (TDS) containing drug(s)/siRNA(s)/targeted peptide was selected from the pre-synthesized bank and tested in vivo on murine cancer model using cancer cells isolated from tumors of each patient. The comparison of precision (for all groups of patients) versus personalized (for each individual patient) nanomedicine treatment showed substantial advantages of the personalized approach. The selection of suppressors of drug resistance precisely for each patient enhanced the efficacy of the treatment of advanced ovarian cancer. Based on the results of the present study, we suggested and evaluated an innovative approach and protocol for personalized treatment of ovarian cancer. The results of the present study clearly show the advantages and perspectives of the proposed individual treatment approach.

Graphical Abstract

A personalized treatment of ovarian cancer tailoring therapies based on individual genetic tumor profiles is proposed. The expression of predefined genes/proteins is analyzed in samples of the patient’s tumor tissue, individual molecular targets are selected, and a mixture of nanoparticles containing drug(s)/siRNA(s)/targeted peptide is prepared and used for the treatment. The approach is validated using an orthotopic mice tumor model.

Introduction

Ovarian cancer remains a highly lethal disease, representing the second leading cause of death from gynecologic malignancies worldwide. The incidence rate of ovarian cancer in 2023 is estimated to be 115,000, with a death rate of around 34,000 (1). Most ovarian cancers are diagnosed at the advanced stage (stages III and IV), and the 5-year survival is less than 30% (2, 3). Multimodality treatment with surgery and chemotherapy remains the main treatment strategy for ovarian cancer; however, approximately 70% of patients relapse after the treatment and develop chemoresistance (4, 5). The efficacy of chemotherapy is limited by the intrinsic and acquired resistance of cancer cells to anticancer drugs (6). Development of such resistance may require an increase in the dose of a single chemotherapeutic agent or the use of several drugs with different mechanisms of action, which in turn substantially raises the possibility of developing severe side effects upon healthy organs, tissues, and cells. Recently, several approaches have been developed and tested to suppress such resistance (Figure 1).

Figure 1.

The overview of current treatment choices and the outline for the work.

Simultaneous induction of cell death and suppression of multidrug resistance (MDR) in chemotherapy:

Traditional chemotherapy uses the maximum tolerated dose (MTD) of one or several anticancer drugs and initially kills the most sensitive cancer cells in the tumor. However, the surviving cells usually are more resistant to the drug and proliferate into more resistant tumors. Finally, repeated courses of chemotherapy lead to the development of highly resistant cancer cells and tumors. Moreover, in most cases, such resistance evolves into so-called multidrug resistance (MDR), causing resistance to several different drugs after the treatment with even a single drug. MDR is involved in over 90% of deaths in cancer patients receiving traditional chemotherapeutics or novel targeted drugs (7, 8). Overall, the mechanisms of MDR are complex and include, among others, elevated metabolism of xenobiotics, enhanced efflux of drugs, growth factors, increased DNA repair capacity, and genetic factors (gene mutations, amplifications, and epigenetic alterations) (8). Consequently, using a single approach to suppress or overcome MDR (like the most frequently used inhibition of drug efflux pumps) cannot effectively prevent the development of MDR and extinguish existing resistance. Therefore, a complex advanced approach is required. It is also essential to suppress MDR by simultaneous induction of cell death with a high dose of the chemotherapeutic agent(s). During the last two decades, we proposed to treat ovarian cancer using various complex advanced drug delivery systems that combine an inducer of cell death and suppressors of different mechanisms of MDR (drug efflux pumps and antiapoptotic defense). Such systems demonstrated several orders of magnitude higher anticancer activities when compared with each component of the delivery system applied separately. Moreover, the effectiveness of treatment in both sensitive and resistant cancer cells was comparable (9-32).

Cancer-targeted chemotherapy:

Augmented toxicity afforded by advanced delivery systems and the use of high-dose or multiple-drug chemotherapy raised concerns about the severe adverse side effects of such chemotherapy upon healthy, non-target organs, tissues, and cells. In order to limit adverse side effects, further elevate the efficacy of cancer treatment, and decrease the overall doses of anticancer drugs, targeting drugs or the entire delivery system specifically to cancer cells was suggested. Many different targeting approaches have been proposed and continue to be developed to treat and image ovarian cancer (33-36). Major targeting approaches include passive targeting (enhanced permeability and retention effect), targeting specific tumor conditions, topical delivery, and active targeting, including targeting organs, cells, intracellular organelles and molecules, sandwich targeting, promoter targeting, indirect targeting, and targeting by external stimuli (37). Combining a cancer-targeting approach with a complex anticancer proapoptotic delivery system demonstrated exceptionally high efficacy in treating multidrug-resistant ovarian cancer with minimized adverse side effects (11, 31). Therefore, we proposed to use a synthetic analog of luteinizing hormone release hormone (LHRH) decapeptide as a cancer targeting moiety specific to LHRH receptors overexpressed in the plasma membrane of many types of cancer cells, while the expression of these receptors in healthy visceral organs was not detected (31, 38, 39). Some LHRH-targeted therapeutics have entered clinical trials (40-43).

Genetic profiling, stratification of patients, “precision” medicine:

Despite successes in developing modern anticancer drugs, their dosage forms, and targeted pro-apoptotic anticancer delivery systems, the efficiency of cancer treatment fluctuates significantly in different patients due to wide variations in drug response and resistance between individuals (44, 45). Some patients on identical chemotherapeutic protocols may receive relatively inefficient amounts of anticancer drugs, while other individuals obtain excessive doses of drugs that induce severe adverse side effects. Hence, the individualized selection of drugs, their doses, and specific targets to suppress mechanisms of resistance and growth based on the molecular characteristics of tumors can improve the treatment outcome and bring us closer to an era of personalized medicine. A stratification of patients based on a genetic profiling of their drug resistance mechanisms (sometimes called “precision” medicine) substantially increased the efficacy of treatment (30, 46-56). However, in most cases, such stratification is based on the measurement of certain average genetic characteristics in groups of cancer patients. Consequently, a patient with more than one mechanism of drug resistance or possessing unique genetic characteristics can still receive inadequate treatment.

Based on the aforementioned, we hypothesized that personalized selection of a cancer-targeted nanotechnology-based delivery system for chemotherapy of each individual tumor will enhance the effectiveness of treatment of primary tumors, delay or prevent the development of metastases, and limit severe adverse side effects of chemotherapy on healthy tissues and organs. Consequently, the central goal of the present study is to focus on the individual genotype and phenotype profiles of tumor tissue samples obtained during surgery from patients with ovarian cancer to select a cancer-targeted delivery system (TDS) with maximum antitumor effectiveness for each individual patient. Determining individualized treatment based on the patient’s unique set of biomarkers in the tumor represents a key innovation of the proposed approach. This set differs substantially from a conventional FDA-approved broad panel of biomarkers for ovarian cancer for diagnosis (57, 58). Therefore, a personalized treatment option for a single patient represents a major distinction of the proposed approach from the currently available treatment choices, where the selection of drug doses and treatment regimens is based on the average characteristics of a group of patients. The present investigation addresses the following two problems: (1) personalized treatment based on the distinct molecular signature of individual tumor and (2) targeted therapy that enhances the efficacy of treatment, limits adverse side effects, and increases the potential to improve clinical outcomes.

Proposed approach:

In order to address weaknesses in the rigor of the prior research, we developed and validated an innovative approach to personalized treatment of ovarian cancer (the most lethal type of gynecological malignancies) based on the individual genetic profile of the patient’s tumor (Figure 1). The main difference between our approach and currently developed “precision” treatment and patient stratification consists of individually selecting anticancer drugs and suppressors of drug resistance for each patient (not for a group of patients). While the main mechanisms of drug resistance and the type of siRNA (and TDS) for the suppression of these mechanisms will be selected from the results of a tumor analysis from an entire group of patients with ovarian cancer, the treatment of an individual patient (or animal with a model of tumor derived from that patient) will be selected based on the patient's unique profile for the suppression of only resistance mechanisms that uniquely exists in the tumor of the specific patient.

The present work aims to validate the proposed approach on a mouse model of advanced human ovarian carcinoma developed by the intraperitoneal injection of cancer cells isolated from malignant ascites obtained from patients with ovarian cancer.

Materials and Methods

Patient Samples

The primary tumor isolates from patients with ovarian cancer were provided by the Rutgers Cancer Institute of New Jersey. Discarded anonymous pathological materials that do not allow identifying patient information were used. The patient tumor samples were identified as Patient #1, Patient #2, etc. The cells were isolated from the samples, propagated, and used to create an orthotopic mouse model of human malignant cancer.

Isolation and Propagation of Primary Tumor Isolates

Tumor tissue samples were placed in a tube, washed twice with excessive phosphate-buffered saline (PBS) then shaking, and placed into a Petri dish containing 1-2 mL of RPMI 1640 medium (Sigma Chemical Co., St. Louis, MO). Tissue samples were then dissected into small pieces using a scalpel or scissors into small pieces of roughly one mm³. Homogenized tumor tissues were transferred to a 50 mL tube, and 10x the volume (w/v) of PBS was added and mixed vigorously. Tissue pieces were allowed to settle for 2 minutes, and the supernatant was aspirated. The tissue pellets were resuspended in Accumax solution (A7089, MilliporeSigma, Burlington, MA) at a concentration of 20 mL/g of tumor tissue and incubated for one hour at room temperature with constant mixing for enzymatic digestion. After settling the remaining tissue pieces, the turbid supernatants were progressively filtered using cell strainers of descending pore size down to 40 μM. The resulting single cell suspensions were diluted 1:1 with the Primary Cancer Cell Medium D-ACF (C-28081, MilliporeSigma, Burlington, MA), pelleted for 10 min at 240 x g and room temperature, washed by the media and pelleted again and propagated in flasks using the same cell medium.

Targeted Drug Delivery System (TDS)

Two types of liposomal TDS were synthesized (Figure 2). “Neutral” liposomes (with a relatively low negative charge of around −10 mV) were used for the delivery of paclitaxel (PTX). PTX (as a lipophilic drug) was incorporated into the lipid membrane of liposomes (Figure 2A). Cationic (positively charged) liposomes were used to deliver anionic (negatively charged) siRNA (Figure 2B). The outer surface of liposomes was decorated with poly(ethylene glycol), PEG, carrying LHRH peptide on distant ends as a targeting moiety. Each liposome formulation contained a drug (PTX) or single siRNA. A complex of PEGylated cationic liposomes with a single siRNA targeted to a specific mRNA was used to suppress the synthesis of a specified protein. Cationic liposomes also contain LHRH tumor-targeting peptides. The sequence of native LHRH peptide, which is similar in human, mouse, and rat, was modified to provide a reactive amino group only on the side chain of a lysine residue, which replaced Gly at position 6 to yield the super active, degradation-resistant-Lys-6-des-Gly-10-Pro-9-ethylamide LHRH analog: Gln-His-Trp-Ser-Tyr-D-Lys(D-Cys)-Leu-Arg-Pro-NHEt (38). A cocktail of TDSs (the mixture of liposomal PTX with several types of liposomes conjugated with specific siRNAs) was used to suppress several selected proteins. The list and sources of used siRNAs are shown in Table 1. The liposomes were fabricated using a previously developed procedure (59-63). Briefly, for the neutral liposome-paclitaxel formulations, DSPE-PEG (1,2-distearoyl-sn-glycero-3-phosphoetanol amine-N-aminopolyethylene glycol, Mw ~2000 ammonium salt-polyethylene glycol) was reacted with LHRH–NH2 to form an amide conjugate. Egg phosphatidylcholine, cholesterol, and DSPE-PEG–LHRH conjugates were dissolved in chloroform, evaporated to a thin film in a rotary evaporator, rehydrated with 0.9% NaCl and gradually extruded through 200 nm and 100 nm pore size polycarbonate filters with an extruder apparatus. The final phospholipid concentration was 20 mM. Paclitaxel was diluted in methanol (50mg/ml) and added to the liposome suspension in a 10%/ 90% v/v ratio. The non-encapsulated drug was separated from liposomes by extensive dialysis against saline. The amount of TAX incorporated into liposomes was determined by high-performance liquid chromatography (HPLC). To determine the stability of liposomal formulations, their size, zeta potential, and concentration of TAX and siRNA were measured after storage in PBS (pH=7.4) under 4°C within 60 days, aqueous solution with low pH=4.5 and after freezing/drying cycle.

Figure 2.

Targeted drug delivery systems. (A) Cancer targeted neutral PEGylated liposomes for the delivery of paclitaxel (PTX). (B) Cancer-targeted cationic PEGylated liposomes for the delivery of siRNA.

Table 1.

The list of gene abbreviations, color coding, and siRNA SKU/Assay numbers. Each assay (except MMP9) represents a heterogeneous pool of siRNAs targeting the same mRNA sequence.

#	Gene Number	Name/ Color Coding	siRNA SKU/ Assay #	Company
1	23	IGF1	EHU059801	Millipore Sigma
2	37	CDKN1C (p57KIP2)	AM16708	Thermo Fisher Scientific
3	49	ABCB1 (MDR1)	EHU131661	Millipore Sigma
4	51	ABCG2 (BCRP)	EHU012921	Millipore Sigma
5	52	CD44	AM16708	Thermo Fisher Scientific
6	55	MMP9	EHU010531	Millipore Sigma
7	58	ESR1 (ER-alpha)	EHU141651	Millipore Sigma
8	59	ESR2 (ER-beta)	EHU-900081	Millipore Sigma
9	60	GLI1	EHU026341	Millipore Sigma
10	62	KRT5	AM16708	Thermo Fisher Scientific
11	63	KRT8	EHU112091	Millipore Sigma

Gene Expression: Quantitative Reverse Transcription Polymerase Chain Reaction (QRT-PCR)

mRNA was isolated from tumors obtained from patients during the surgery. The gene expression profile for each tumor sample was calculated using the ΔΔCT method with normalization of the raw data to the housekeeping gene using an Excel workbook provided by the manufacturer. According to the manufacturer's protocol, mRNA was isolated using a RNeasy kit (Qiagen, Valencia, CA). First-strand cDNA was synthesized with Ready-To-Go You-Prime First-Strand Beads (Amersham Biosciences, Piscataway, NJ) with 1 μg of total cellular RNA and 100 ng of random hexadeoxynucleotide primer (Amersham Biosciences, Piscataway, NJ). The reaction mixture was immediately subjected to - QRT-PCR (Applied Biosystems, Inc., Foster City, CA). We already preselected a set of 83 genes/proteins predominately overexpressed in tumor tissues of patients with ovarian cancer and constructed a custom-made 96-well plate for a quantitative reverse transcription polymerase chain reaction that was manufactured by Qiagen, Inc. (Valencia, CA) according to our design. The plate also includes 12 standard genes commonly used by Qiagen in qRT-PCR plates (reference genes, positive, negative, and contamination controls). Initially, the expression of 191 genes in cells isolated from patient samples was analyzed using three different commercial cancer gene profiling qRT-PCR kits (Human Apoptosis and Breast Cancer RT2 ProfilerTM PCR Arrays and qBiomarker Copy Number PCR Array Human Ovarian Cancer, Qiagen, Valencia, CA). Based on the results of these measurements, 83 genes were selected to create a custom QRT-PCR array (54). The selected genes were picked up from the groups responsible for angiogenesis (13 genes), apoptosis induction and antiapoptotic defense (21 genes), cell cycle (8 genes), DNA damage (6 genes), drug resistance (5 genes), metalloproteases (2 genes), signal transduction (18 genes) and 10 genes of transcription factors. In addition, the expression of the LHRH receptor (LHRHR, gene #84) was measured. The following primers were used for LHRHR – GAC CTT GTC TGG AAA GAT CC (sense), CAG GCT GAT CAC CAC CAT CA (antisense). The manufacturers incorporated the primers for other studied genes into corresponding gene profiling kits. The detailed list of all 84 investigated genes is shown in Figure 3.

Figure 3.

Average gene expression in the mRNAs isolated from tumors of all patients. The expression of the corresponding genes is represented as relative ΔΔCT values normalized of the raw data to the housekeeping gene. Gene names, numbers, and colors are presented in the table at the bottom of the figure.

Protein Expression: Western Blotting

The Western blot was used to confirm the expression of the selected proteins. Briefly, 100 μg of samples were denatured at 95°C for 10 min and transferred to a 4-20% Mini-PROTEAN^® TGX^™ Precast Protein Gels (Bio-Rad, Hercules, CA), and electrophoresis was performed at 120 V for 60 min. Then, the band was transferred to a 0.2 μm cellulose nitrate membrane (c). Afterward, the immunoreactive bands were detected by ChemiDoc MP Imaging System (Bio-Rad, Hercules, CA). The expression of three proteins that are overexpressed in all studied patient samples - CDKN1C (p57KIP2), CD44, and ESR2 (ER-beta) – was measured. Mouse monoclonal antibodies for these proteins and HRP-conjugated goat anti-mouse antibodies were used as primary and secondary antibodies, respectively (Thermo Fisher Scientific Inc, Waltham, MA).

Cytotoxicity of Liposomal Formulations

Cytotoxicity of free PTX and various liposomal drug and siRNA was analyzed by the modified MTT assay as previously described (64, 65) after 48 h incubation of cancer cells isolated from Patient #3 with different formulations and appropriate controls. Based on the test results, the IC50 doses (the concentrations that kill 50% of cells) were calculated for free paclitaxel and PTX-containing delivery systems.

Animal Model and Treatment

Animal experiments were carried out according to protocols and procedures of animal use approved by the Institutional (Rutgers, the State University of New Jersey) Animal Care and Use Committee (IACUC). The cells isolated from human tumors (2×10⁶) were administered into the peritoneal area of 20 female athymic nu/nu mice. The progression of the intraperitoneal tumor growth was assessed by three different imaging systems (optical, MRI, ultrasound) as described previously (31, 66, 67). The part of isolated cancer cells was transfected with luciferase as described previously (31), allowing an optical bioluminescence imaging system to visualize the intraperitoneal tumor. When the tumors reached a size of about 0.4 cm³ (15-20 days after transplantation), mice were treated intraperitoneally with a cremophor-based free unbound paclitaxel or liposomal paclitaxel (PTX or LIP-PTX, respectively) with the maximum tolerated dose (MTD) of paclitaxel (2.5 mg/kg) twice per week for four weeks. The dose selection mimicked a similar approach used for the clinical trials of anticancer drugs. The average MTD was determined in a special series of experiments by exposing nu/nu mice without tumors to five increasing doses of free PTX. The body weight of mice was monitored. The dose that led to the average decrease of mouse body weight by approximately 15% was considered the MTD. The dose of siRNA was around 170 μg/kg for the single administration. Previously, we found that this dose was the most suitable for animal experiments (59, 68). The tumor volumes were measured three times per week and at the end of the experiments. LIP-PTX was mixed with liposomes conjugated with a mixture of selected siRNAs in some series. The list of the series of experiments is shown in Table 2.

Table 2.

Series of experiments.

#	Series abbreviation	Description	Mice number	Patient samples
1	Control	Untreated mice (saline was injected instead of therapeutics).	20 (5 mice per patient sample)	Patients #1-4
2	PTX	Mice treated with free non-bound cremophor-based PTX.	20 (5 mice per patient sample)	Patients #1-4
3	LHRH-LIP-PTX	Targeted Liposomes with paclitaxel	20 (5 mice per patient sample)	Patients #1-4
4	LHRH-LIP-PTX+3 siRNAs	A mixture of targeted liposomes with paclitaxel and liposomes with three types of siRNAs selected for the group of patients (precision therapy).	20 (5 mice per patient sample)	Patients #1-4
5	LHRH-LIP-PTX+11-siRNAs	A mixture of targeted liposomes with paclitaxel and liposomes with 11 types of siRNAs selected for a single patient (personalized therapy).	10 mice	Patient #3

Statistical Analysis

Data were analyzed using descriptive statistics, single-factor analysis of variance (ANOVA), and presented as mean values ± standard deviation (SD). The comparison among groups was performed by the independent sample student’s t-test. The difference between variants was considered significant if P < 0.05.

Results

Liposome Characterization and Stability

Synthesized liposomal formulations were characterized by measuring their size, zeta potential, and loading capacity. The average size of all liposomal formulations was around 120 nm; the LIP-PTX formulations have a slightly negative charge (Table S1 in Supplemental Materials). Conjugation of cationic liposomes with anionic siRNA resulted in the formation of practically neutral particles. The encapsulation efficacy of TAX ranged from 55 to 60% in different experiments. The average siRNA loading was about 2.5 nmol/mg of liposomes (Table S1). All liposomal formulations were stable during the storage at different conditions (4°C, normal pH, 60 days; room temperature, pH=4.5, three hours, with minimal loss of the drug and siRNA) and after the freezing/drying cycle (Tables S1, S2, S3 in Supplemental Materials). During these conditions, we did not register a statistically significant change in liposomal size, charge, and drug/siRNA content. Cytotoxicity of prepared liposomal formulations was accessed using cancer cells isolated from patients. It was found that unloaded liposomes and liposomal siRNA formulations were not toxic in the concentrations used (data not shown). The encapsulation of PTX into liposomes substantially increased its cytotoxicity (Table S3 in Supplemental Materials). The suppression of targeted genes encoding proteins responsible for multidrug resistance in cancer cells by the siRNA pool statistically significantly enhanced the cytotoxicity of PTX. Further enhancement of cytotoxicity was achieved by employing the LHRH peptide (Table S3) as a cancer-specific targeting moiety. Previously, we showed that LHRH receptors are overexpressed in several types of cancer cells with low expression on tissues from visceral organs and healthy reproductive organs (26, 38). We also showed a high targeting efficacy of formulations decorated with the LHRH peptide (60). We used this targeting approach in our laboratory to enhance the efficiency of the delivery of drugs, siRNA, antisense oligonucleotides to cancer cells/tumors and decrease an adverse side effect of the treatment to healthy tissues (9, 26, 32, 39, 63, 66-69).

Average and Individual Responses to Treatment

The first part of our investigation was devoted to illustrate the concept that “One size does not fit all” and showing the disadvantages of ovarian cancer treatment without personalized stratification of each patient. We treated mice as described in Table 2. Figure 4 illustrates the effect on tumor size of control mice, a conglomerate of mice treated with either drug alone or liposomal PTX, and mice with patient-specific tumors treated with liposomal PTX. We showed that all untreated mice developed intraperitoneal tumors/malignant ascites, which reached an average of 2 cm³ (the maximum tumor size allowed by the animal protocol approved by IACUC of Rutgers University) within 25-30 days after injecting cancer cells. Treatment with the free PTX decreased tumor size on average by 32% (P < 0.05). Furthermore, it increased the survival of animals on average by one week (until the total size of intraperitoneal tumors reached ~2 cm³). Treatment of mice with tumors by Lip-PTX (with the same average MTD concentration of the drug) decreased tumor size on average by 42% (P < 0.05) and increased survival by two weeks. Despite the statistically significant limitation of tumor growth in both cases, a massive deviation of statistical data was registered. The coefficient of variation of individual data increased from ~11% to 50-60% in treated animals. Such dramatic data variation usually indicates a substantial difference in individual responses to the treatment (Figure 4).

Figure 4.

Average data (tumor volume calculated together for all series of mice inoculated with cells obtained from all four patients with ovarian cancer) and individual responses (tumor volume calculated separately for each series of mice inoculated with cells obtained from Patient #1, #2, #3 and #4). (A) Representative optical bioluminescence image of intraperitoneal ovarian tumors. (B) Tumor size of untreated mice and mice treated with free unbound paclitaxel (PTX) or LHRH-targeted liposomal PTX (LHRH-Lip-PTX). Tumor volume was measured at the end of the experiments. Means ± SD are shown. *P < 0.05 when compared with untreated animals.

In fact, analysis of tumor size in mice bearing cancer cells from each individual patient supported this conclusion. The animals bearing tumors from Patient #1 were the most responsive to the treatment with Lip-PTX. The tumor size in this series decreased on average to 13% from the tumor volume in untreated mice (P < 0.05). On the other hand, cancer cells from Patient #2 were more resistant to the treatment and demonstrated a decrease in total tumor volume only to 40% when compared with untreated mice (P < 0.05). At the same time, treatment with Lip-PTX of mice bearing cancer cells from Patients #3 and #4 did not demonstrate a statistically significant decrease in the tumor volume. Consequently, cancer cells from Patient #1 responded relatively well to the chemotherapy, and cells from Patient #2 were more resistant to the treatment. In contrast, tumor cells from Patients #3 and #4 were highly resistant to chemotherapy and essentially failed to respond to the treatment with the liposomal form of PTX. Similar results were obtained with free PTX (data are not shown).

Gene Expression in Cancer Cells Isolated from Patients

One of the reasons for the different patient responses to chemotherapy is the overexpression in the tumor cells of specific genes and proteins responsible for the resistance to chemotherapy. For the initial analysis, we selected three commercially available cancer profiling qRT-PCR kits: Human Apoptosis (1) and Breast Cancer (2) RT2 ProfilerTM PCR Arrays and qBiomarker Copy Number PCR Array Human Ovarian Cancer (3) available from Qiagen (Valencia, CA). These kits included 191 genes encoded signal transduction molecules, transcription factors, and other proteins that regulate the cell cycle, cytoskeletal organization, epigenetic modifications, and ubiquitin-proteasome-dependent protein turnover in cancer cells. Based on the results of our previous investigations (15-17, 25, 30, 54, 59, 60, 64-66, 69), we selected 80 genes included in these kits that potentially can be responsible for multidrug resistance in ovarian cancer cells and were overexpressed in tumor samples obtained from patients with advanced ovarian carcinoma. These selected genes can be subdivided into the following groups based on the primary functions of corresponding proteins: angiogenesis, apoptosis induction and antiapoptotic defense, cell cycle regulation, DNA damage, multidrug resistance, signal transduction, and transcriptional factors. They also include metalloproteases responsible for deactivating xenobiotics (including anticancer drugs. In total, a set of 84 genes was selected to manufacture a custom-made 96-well plate for a quantitative reverse transcription polymerase chain reaction by Qiagen, Inc. (Valencia, CA) according to our design. The plate also includes 12 standard genes commonly used by Qiagen in qRT-PCR plates (reference genes, positive, negative, and contamination controls). The list, position in the plate (gene number), and color coding are shown in Figure 3.

we measured the expression of the preselected genes in patient samples by quantitative RT-PCR. The average results of such measurements for all four patient samples are presented in Figure 4. As illustrated, the different tumors had very similar directions in up- and downregulation of pathway genes, but there were specific quantitative differences in expression in the different tumors. Hence each patient harbored tumors with a unique quantitative expression of genes, mainly drug resistance-related genes.

While the overall tendency in the gene expression for all four patients is preserved, specific quantitative differences are observed. First, the overexpression of genes in the tumors of Patients #1 and 2 is substantially less pronounced when compared with the cells obtained from the tumors of Patients #3 and 4. At the same time, the cells obtained from Patients #1 and 2 demonstrated a lower level of resistance to chemotherapy. It is reasonable to conclude that the overexpression of some genes may be responsible for the resistance of cancer cells to chemotherapy, and these genes may determine the response of ovarian cancer cells to treatment. Therefore, we hypothesized that inhibiting the expression of such genes can improve the response to chemotherapy in drug-resistant ovarian cancer cells.

Suppression of Cellular Resistance of Cancer Cells to Chemotherapy

The genes overexpressed in studied cancer cells (with expression around or exceeding five rel. units) are presented in Figure 5. It can be seen that six genes were overexpressed in cells obtained from the tumor of Patient #1, four – in Patient #2, twelve genes – in Patient #3 and Patient #4 samples. At the same time, only three genes - CDKN1C (p57KIP2), CD44, and ESR2 (ER-beta) – were overexpressed in cancer cells obtained from tumors of all four patients. The expression of corresponding proteins was confirmed by Western blotting analysis (Figure 5, Table S1 in the supplemental material). According to the principles of “precision” medicine, these three genes/proteins should be suppressed to inhibit resistance to chemotherapy. We prepared a mixture of liposomal PTX with three liposomal formulations containing siRNAs targeted to these three mRNAs corresponding to the specified genes that are overexpressed in tumors of all four patients. We treated all mice with tumors using these formulations (Lip-siRNAs). The results of the treatment are shown in Figure 6.

Figure 5.

Genes overexpressed in cancer cells from all four patients. Three genes - CDKN1C (p57KIP2), CD44, and ESR2 (ER-beta) were overexpressed in all four patients (these genes are shown in red boxes). The overexpression of these genes was confirmed by Western blotting (insert shows a typical image of the blot). The color coding of genes corresponds to the color genes presented in Figure 3.

Figure 6.

Tumor volume of mice with malignant human ovarian carcinoma treated with the mixture of LHRH-Lip-PTX+LHRH-Lip containing three siRNAs (CDKN1C, CD44, and ESR2) that are overexpressed in the tumors of all four patients. Average data (tumor volume calculated for all mice inoculated with cells obtained from all four patients) and individual responses (tumor volume calculated separately for each series of mice inoculated with cells obtained from Patients #1, #2, #3 and #4). Means ± S.D. are shown, *P < 0.05 when compared with untreated animals.

Data obtained show that, on average, treatment of mice bearing orthotopic tumors from cancer patients with LHRH-Lip-PTX combined with the suppressors of the resistance of cancer cells to chemotherapy by exposure to three types of siRNAs substantially decreased tumor volume after the treatment almost by five folds (average data for all mice inoculated with tumors from all four patients). However, significant quantitative differences were observed between different groups of mice. Mice bearing tumors from Patients #1 and #2 responded to the treatment well. At the same time, the response of mice with tumor samples obtained from Patients #3 and #4 was suboptimal. One can conclude that suppressing only three genes (responsible for the resistance in the tumors of all four patients) is insufficient to overcome the resistance in tumors of Patients #3 and #4 because their tumors overexpress substantially more genes that confer drug resistance.

Gene Expression in Tumor Tissues Isolated from Experimental Animals

The proposed cancer treatment approach heavily depends on the tumor gene expression. The selection of siRNAs used for treating mice bearing tumors was primarily based on the expression of genes in the cells isolated from source patients’ tumor tissues. However, after propagation in vitro and growth in mouse tumors in vivo, the cells could change their properties (especially mRNA expression) compared to the original freshly isolated cells. In order to confirm that the gene expression profile in tumors isolated from experimental animals is similar to the freshly isolated cells, we measured the expression of genes selected as molecular targets for individual therapy (Table 1) in tumor tissues isolated from experimental animals bearing cells from tumors of Patient #3 by QRT-PCR. The results of such measurements are presented in Figure 7. The results are expressed as relative units as ΔΔCT values normalized housekeeping genes using an Excel workbook provided by the manufacturer. The data obtained clearly show that isolated and propagated tumor cells generally preserve the gene expression pattern.

Figure 7.

Gene expression in tumors from mice bearing cancer cells isolated from the tumor from Patient #3. (A) Data from the mouse tumor (bars, means ± SD are shown) and source freshly isolated cells (red lines). (B) Gene numbers, color coding, and gene names. Gene expression is presented as relative ΔΔCT values normalized of the raw data to the housekeeping gene.

Personalized Chemotherapy

We selected the cancer cell samples from Patient #3 as a model to show the advantages of personalized vs. precision treatment of ovarian carcinoma. The cancer cells of this patient overexpress 12 genes (three genes typical for all four patients and nine additional genes shown in Figure 5). We selected 11 genes mainly responsible for pump and non-pump resistance to chemotherapy, development of metastases and cancer cell growth in this patient's tumor. We prepared liposomal formulation of siRNA targeted to these genes for their suppression. The list and source of these siRNAs are presented in Table 1. The CDH1 gene encodes a pro-apoptotic protein that promotes cell death during chemotherapy and, therefore, was excluded from the genes to be suppressed. The influence of treatment on the gene expression before and after treatment is presented on Figure 8. It can be seen, that treatment with nanoparticles containing siRNA targeted to the specific gene significantly decreased the gene expression.

Figure 8.

The expression of the selected genes in mouse tumors before and after treatment with nanoparticles containing paclitaxel and siRNAs. The cells were obtained from Patient #3. Mice were treated with LHRH-Lip-PTX and a mixture of LHRH-Lip-PTX+LHRH-Lip containing 11 siRNAs. Means ± SD are shown (n = 5). (A) – Gene expression. (B) Gene numbers, names, and color coding. *P < 0.05 when compared with untreated animals.

The comparison of the tumor volume after treatment with different liposomal formulations is presented in Figure 9. Treatment with LHRH-Lip-PTX alone did not change the tumor volume significantly (P > 0.05). Treatment with targeted liposomal PTX and three siRNAs decreased tumor volume by approximately 50% (P < 0.05). Finally, suppressing all 11 overexpressed genes by tumor-targeted liposomal siRNAs combined with liposomal PTX led to the almost complete disappearance of intraperitoneal tumors in mice inoculated with cancer cells from Patient #3.

Figure 9.

The average response of mice with malignant intraperitoneal ovarian cancer. The cells were obtained from patient #3. Mice were treated with LHRH-Lip-PTX, mixture of LHRH-Lip-PTX+LHRH-Lip containing three siRNAs (CDKN1C, CD44, and ESR2) that are overexpressed in the tumors of all four patients and mixture of LHRH-Lip-PTX+LHRH-Lip containing 11 siRNAs (siRNAs were selected based on the expression of genes in tumor of Patient #3 and are presented in Figure 5). Means ± S.D. are shown, *P < 0.05 when compared with untreated animals.

Discussion

Our overall goal of the present study is to select gene/protein expression that can predict tumor response to anticancer drugs and to reveal correlations between the expression of selected proteins and resistance to chemotherapy of cells isolated from tumors and surrounding healthy tissues from patients with ovarian cancer. We used the cells isolated from patient tumors instead of established lines of cancer cells in order to maintain the existing individual genetic variations in tumors of patients with ovarian cancers. Paclitaxel (PTX) was selected as a model drug to evaluate individual cancer cell resistance and the correlation between the efficacy of cancer treatment and the individual genetic profile of cancer cells. This drug is currently used to treat many types of cancer, including gynecologic malignancies.

We developed and tested a multicomponent quadruple theranostic system that consists of (1) a stable, non-toxic nanocarrier; (2) cancer-targeting moiety; (3) anticancer drug; and (4) specifically selected for each patient multiple siRNAs for the suppression of mechanisms responsible for the development of drug resistance. The lipid-based carrier was used for the delivery of siRNAs, an anticancer drug, and a targeted peptide in order to improve the stability and solubility of active components, enhance drug and siRNAs content in the tumor, achieve better penetration into cancer cells, provide for excellent biocompatibility, low cyto- and genotoxicity of the carrier, avoid organic solvents, decrease the cost of the dosage form, and simplify scale-up and sterilization procedures.

We tested an individual personalized treatment of advanced ovarian cancer that combined several previously developed methods and includes the following distinct features:

Cancer targeted therapy.

Delivery of anticancer drugs specifically to tumor cells is achieved by two types of tumor targeting: passive and active. Passive targeting is accomplished by delivering anticancer drugs by a nanoscale-based delivery system. It is well known that, based on the Enhanced Permeability and Retention Effect, after systemic administration, high molecular weight substances (including nanocarriers) preferentially accumulate in solid tumors (70-74). However, such passive targeting is generally ineffective for advanced tumors and metastases. Previously, we have shown that a synthetic analog of LHRH decapeptide that targets LHRH receptors overexpressed in many types of cancer cells (including ovarian cancer), and does not demonstrate a detectable level of the expression in healthy cells of visceral organs (11, 31, 38, 39, 69). We showed that the LHRH peptide demonstrated the high targeting efficacy (68). The use of LHRH peptide improved the systemic distribution of the targeted delivery system towards preferential accumulation in the cancer cells both in solid tumors and spread metastatic cells and inside the affected organs leaving healthy cells intact (28, 38, 75). Such a distribution enhances the accumulation of anticancer therapeutics in the tumor cells, improving the efficacy of treatment and simultaneously effectively limiting adverse side effects of the treatment upon healthy cells and organs (31, 38, 63, 66).

Anticancer drug.

An anticancer drug (or a mixture of drugs) is required to induce cell death. Previously, we demonstrated that the treatment with siRNA alone used to suppress resistance mechanisms in cancer cells would induce cancer cell death to a relatively minor extent. However, only a combination of such siRNA(s) with an anticancer drug (or a mixture of drugs) has the potential to kill cancer cells effectively (17, 31, 68, 69, 74).

Suppressors of drug resistance mechanisms specific for an individual patient.

Present experimental data show that the suppression of individually selected genes substantially augmented the therapeutic effect of anticancer drugs. However, it should be reiterated again that these candidates for suppression should be selected based on the tumor's individual genetic profile and not on the average gene expression profile for the group of different patient tumors.

Major innovations of the present study and differences from the existing approaches and our previous investigations include the following points. (1) An approach to the personalized selection of efficient treatment for each patient with gynecologic cancer based on the individual genotype and phenotype profiles of tumor tissue samples obtained from the same patient (in contrast to the selection of treatment options based on the genetic profile of a group of patients with similar cancer type ad stage). (2) Combinatorial cancer-specific delivery of inhibitors of resistance pathways (siRNAs) with cell death inducer (an anticancer drug) in order to achieve a synergetic effect, overcome the resistance of cancer cells to therapy, and enhance treatment efficacy.

The three most common approaches to chemotherapy were modeled in the present study using the cancer cells isolated from tumors of patients with ovarian carcinoma. First, the most frequently used treatment approach based on the average maximum tolerated dose of the anticancer drug (paclitaxel) was tested. Second, induction of cancer cell death by PTX was combined with the suppression of drug resistance mechanisms by a mixture of siRNAs selected based on average data and used for all patients. Third, siRNAs individually selected for a single patient were used with the anticancer drug to treat the same patient. The second case imitated so-called “precision” medicine, while the last represents the proposed “personalized” treatment approach. The results of the investigation demonstrated the vast advantages of the personalized approach over both traditional treatment and “precision” medicine. Only personalized selection of inhibitors of drug resistance individually for each patient allowed for the effective shrinkage of tumor growth.

Conclusions and Further Directions

Based on the results of the present study, we propose to create a pre-synthesized "bank" of diverse delivery systems with different siRNAs and drugs available for the treatment of patients. The patient will be treated with the chosen cocktail of delivery systems (from pre-synthesized bank compounds) designed specifically for the individual's tumor. The current investigation focuses on ovarian cancer as the most common and deadly type of gynecologic malignancy. However, it is expected to extend the results of the current investigations to other forms of gynecologic and other types of cancers.

The procedure of personalized treatment appears as follows (Figure 10). First, samples of a patient's tumors and normal surrounding tissues are obtained during tumor resection surgery (debulking) and expanded using a conventional tissue culture technique. Tumor profile data (the expression of predefined genes/proteins) are obtained and analyzed. Based upon the results of this analysis, several molecular targets for the suppression of central individual mechanisms of drug resistance/metastases and the most effective anticancer drug(s) are selected individually for each patient. Finally, a mixture of complex nanotechnology-based targeted delivery systems containing drug(s)/siRNA(s)/targeted peptide is prepared. The selected systems include a tumor-targeted moiety, nanotechnology-based carrier, the most effective drug(s), and siRNAs selected for each patient based on the genetic profile of the patient's tumor. Such personalized therapy is expected to effectively suppress drug resistance, tumor growth, and the development of metastases and limit the adverse side effects of therapy. The proposed personalized approach is feasible, and the treatment selection for an individual patient can be completed within a week.

Figure 10.

The proposed approach for the personalized treatment of ovarian cancer.