 |
|
|
|
|
|
 |
 |
 |
 |
 |
 |
 |
||
Volume 32 |
April 2010 |
Number 2 |
|
||
|
||
Identification of Annexin A13 as a Regulator of Chemotherapy Resistance Using Random Homozygous Gene Perturbation
Heinz Reiske, Ph.D., Baoquan Sui, M.S., Huong Ung-Medoff, B.S., Robert Donahue, M.S., Wu-Bo Li, Ph.D., Michael Goldblatt, Ph.D., Limin Li, Ph.D., and Michael S. Kinch, Ph.D.
OBJECTIVE: To utilize a powerful new technology for target discovery, Random Homozygous Gene Perturbation (RHGP), and to identify novel targets that cause tumor cells to become chemoresistant.
STUDY DESIGN: RHGP was used to identify and validate genetic changes that cause chemoresistance of tumor cells to Rapamycin.
RESULTS: A series of targets was identified that allowed tumor cells to survive treatment with Rapamycin. We validated these targets and focused on Annexin A13, a target where decreased expression caused tumor cell insensitivity to Rapamycin. Ectopic overexpression of Annexin A13 was likewise sufficient to sensitize malignant breast cancer cells to treatment with Rapamycin.
CONCLUSION: These findings expand our knowledge of mechanisms that allow tumor cell drug resistance and demonstrate the power of RHGP to identify novel targets and mechanisms. (Anal Quant Cytol Histol 2010;32: 61–69)
Keywords: cancer, drug resistance.
Despite improvements in cancer detection and treatment, drug-resistant cancer generally remains fatal. While great strides have been made in the early detection and treatment of many different cancers, a great need remains to identify and intervene against those mechanisms that allow tumor cells to evade cytotoxic therapy. To this end, much investigation has sought to relate changes in particular biomarkers with differential drug sensitivity using broad-based screening approaches (genomics and proteomics).1-3 However, these types of findings often are correlative, and the causal role of relevant targets may be more subtle or hidden within the wealth of data provided by conventional sampling approaches.
An alternative to sampling for differences that distinguish among drug sensitive versus resistant therapies is to adopt an approach of actively seeking causative targets. Such approaches are feasible when a particular target or family of targets is known or suspected. In such cases, one can overexpress or block the expression of a particular gene and determine the potential impact on drug sensitivity. However, a more efficient solution would arise were investigators to be able to sample the entire genome to evaluate any genetic change (e.g., overexpression, loss of expression, production of dominant-negatives, etc.). Recent technologies have sought to address this critical need, including genome-wide antisense and RNAi approaches.4,5
Our laboratories recently pioneered a new approach, Random Homozygous Gene Perturbation (RHGP).6 RHGP provides an opportunity to sample the entire genome for genetic changes that give rise to a desired phenotype.7 The technology centers on a unique gene search vector (GSV), which can randomly integrate at any site within the genome and either induce or inhibit the expression of entire genes or gene domains (e.g., as dominant negative inhibitors). The strength of the technology is that all genes, both known and unknown, can be sampled in a single experiment. The key to the technology is the use of a reliable cellular outcome. In our present report, we utilize the acquisition of tumor cell resistance to cytotoxic chemotherapy as the desired outcome. RHGP allowed us to identify a set of targets that regulate tumor cell sensitivity to Rapamycin, a new chemotherapeutic option for many cancers.
Materials and Methods Cell Culture
Human MDA-MB231 breast cancer cells were purchased from the American Type Culture Collection. The Phoenix A cell line was a gift from Dr. Nolan, Stanford University, and mouse N2a cell line was a gift from Dr. Xu, Rockefeller University. MDA-MB231 cells were cultured in DMEM containing 10% FBS, and N2a cells were cultured in 50% DMEM and 50% Opti-MEM media containing 5% FBS.
Construction of pTet-Off Transactivator Vectors and RHGP Gene Search Vector The RHGP gene search vectors were constructed from the pBabe-puro vector. To create a gene search cassette containing an antibiotic resistant reporter gene driven by tetracycline regulated mini CMV promoter, the puromycin resistant gene was cloned downstream from, and in the same orientation as, a mini CMV promoter controlled by 14 repeats of TRE fragment in a shuttle vector, pAY48 (Figure 1).
 Construction of Tet-Off/Hygromycin Resistant Stable Cell Lines N2a cells were plated in 10 cm plates for 24 hours and transfected with 2–5 µg of pTet-Off regulator plasmid DNA using Fugene 6 (Roche, Nutley, New Jersey, U.S.A.). After 1 day, the medium was replaced with selection medium containing 100 µg/ mL hygromycin. The medium was changed every 3–4 days until all untransfected cells were killed. Individual colonies were subcultured and screened for the doxycycline (Dox) induction activity using pTRE-luc and pEGFP vectors. The stable clones with induction activity above 100-fold were used for RHGP libraries.
Selection of Breast Cancer Drug–Resistant Clones RHGP libraries were subjected to Rapamycin selection in 10 cm plates (1
 104–106 cells/plate) in medium containing FBS, 100 mg/mL hygromycin and 10 nM Rapamycin and 0.5 µg/mL of puromycin. After 4 weeks, Rapamycin-resistant clones were isolated, picked and subcultured. Identification of the Candidate Genes from Drug–Resistant Clones Genomic DNA was obtained using the BIO-RAD AquaPure Genomic DNA Isolation Kit. Self-ligated genomic DNA was concentrated and electroporated into Escherichia coli cells and selected using chloramphenicol.
Validation and Functional Analysis of the Target Genes To verify the reversibility of the Rapamycin-resistant clones, the phenotypes and the parental N2a cells were cultured in medium with Rapamycin and/or Dox (Invitrogen, Carlsbad, California, U.S.A.). After 3 days of treatment with the indicated amounts of Rapamycin, MTT assays provided readouts of cell viability.
Results The goal of the present studies was to utilize a new technology, RHGP, to identify novel targets that cause tumor cells to become chemoresistant. To accomplish this goal, it was necessary to identify an appropriate drug to model resistance. Rapamycin was selected based on recent identification of mechanisms that control tumor cell sensitivity to Rapamycin that are independent of cell lineage or specific genetic abnormalities. Furthermore, Rapamycin is not subject to elimination via conventional multidrug resistance pathways.8-10 The sensitivity of N2a neuroblastoma cells was evaluated over a range of concentrations of Rapamycin, providing a reproducible cell growth curve with which to model Rapamycin sensitivity (Figure 2). MTT-based assays established that Rapamycin inhibited N2a cells with an IC50 of 25 nM. Consequently, we defined Rapamycin sensitivity as the ability of N2a cells to propagate when cultured in the presence of 1 uM Rapamycin.
 Having established their drug sensitivity, N2a cells were stably transduced with an MLV-based GSV, and RHGP libraries were generated as indicated in the schema shown in Figure 3A. To obtain the phenotype of Rapamycin resistance, the cultures were cultured in the continuous presence of
1 uM Rapamycin, replenished every 48 hours. To begin focusing on interesting candidates, approximately 200 drug-resistant clones were isolated by limiting dilution. We emphasized clones that were resensitized to Rapamycin when the RHGP technology was inactivated by the addition of Dox to the culture medium (see Figure 3B for an example).
 Throughout these studies, we considered that artifacts unrelated to RHGP that might confer Rapamycin resistance. For example, it is generally understood that variations in cell culture conditions can alter tumor cell sensitivity to Rapamycin. To minimize this possibility, the subclones were subjected to multiple and independent rounds of challenge with Rapamycin. These studies allowed us
to prioritize 83 clones (43% of the starting population) that retained the highest levels of Rapamycin resistance.
Additional assessments were used to confirm that the RHGP technology was responsible for the drug-resistant phenotype. An important feature of the RHGP strategy is that expression of the target gene is under the control of an inducible promoter, which provided an objective means of confirming the reversibility of Rapamycin resistance using standard MTT-based assays of metabolic activity. Most candidate clones remained viable when challenged with Rapamycin under conditions of RHGP promoter activation (Figure 4A). Upon addition of Dox, these clones regained sensitivity to Rapamycin killing. Clones that remained resistant to Rapamycin in the presence of Dox were deprioritized since these might have arisen as a result of means other than the RHGP perturbation. (For an example, see Candidate 8 in Figure 4A.) We also confirmed the resistant phenotype over a wide range of Rapamycin concentrations (data not shown).
 Additional assays of tumor cell growth precluded artifacts that might have arisen based on the use of a single assay, MTT, to assess Rapamycin resistance. For example, 2 standard markers of cell cycle progression, cdc2 phosphorylation and cyclin D1 expression, confirmed that candidate clones continued through cell cycle progression regardless of Rapamycin treatment when the GSV promoter was activated via incubation without Dox (see Figure 4B for a representative example). In contrast, the cell cycle was arrested in these same clones following treatment with Dox.
After verifying that the GSV insertion was responsible for the observed Rapamycin resistance, we sought to characterize the integration event. Specifically, it was important to identify the gene into which the GSV had integrated and the orientation of the GSV promoter (to determine if the RHGP event represented a positive or negative regulation of the target gene). To identify the integration site, a unique cassette had been engineered into the GSV. The CAT gene in GSV allows us to clone the genomic DNA flanking the GSV site. Altogether, the GSV was isolated from 16 different Rapamycin-resistant cell clones (Table I). A review of the published literature indicated that 9 of the 16 genes (56%) had not yet been annotated, or no function had been linked with that particular cDNA. Of the remaining targets, 7 had been described previously (ANX13, CCRL2, PIASy, CD38, H2AFd, HMGCS2 and HSPG2), 4 had been previously linked with cancer (CD38, HMGCS2, HSPG2 and PIASy),11-15 and 2 had been linked previously with chemotherapy resistance (CD38 and PIASy11,14,15). In a separate analysis, we evaluated the orientation of the genomic insertion of the GSV. As expected by change, approximately half of the insertions (7 of 16) represented sense orientations, with the remaining representing insertions in an antisense orientation (9 of 16, Table II).
  To focus our efforts on 1 particular target, we selected RapR5 for further analysis and to identify the target of the RHGP perturbation. DNA sequencing and mouse genome mapping of the site of integration determined that the promoter of GSV for RapR5 had inserted in an antisense orientation at an intron upstream from the last domain of the Annexin A13 gene (Figure 5A and B). To confirm the outcome of the GSV integration event, western blot analyses were conducted with RapR5 cells that had been incubated in the presence of absence of Dox. Consistent with the genomic mapping of the GSV, the levels of Annexin A13, as measured by western blot analyses with specific antibodies, were virtually undetectable in the RHGP-transduced RapR5 clone. Importantly, the reversible nature of the RHGP transduction was confirmed by incubation of the RapR5 cells in the presence of Dox, which is consistent with the evidence that the Rapamycin-resistant phenotype of RapR5 cells was similarly reversible in the presence of Dox. (Note that the levels do not reach that of control cells since the integration event itself disrupted expression of one allele.)
 Although intriguing, the antisense insertion of the GSV into the Annexin A13 gene did not conclusively link this particular target to drug resistance. We also sought to exclude that antisense technology itself might have introduced an artifact causing Rapamycin resistance. Finally, we sought to model a realistic situation by using a cell system in which Rapamycin resistance had occurred naturally rather than in the laboratory. To address all these potential questions, we asked if ectopic overexpression of Annexin A13 could resensitize patient-
derived tumor cells to Rapamycin treatment. To conduct these studies, it was necessary to identify a Rapamycin-resistant tumor cell system. Published findings demonstrated that MDA-MB-231 cells display inherent resistance to Rapamycin. Therefore, an expression vector encoding for wild-type Annexin A13 was constructed and stably expressed in MDA-MB-231 via ectopic transfection. Multiple clones stably overexpressing Annexin A13 were isolated and subjected to MTT-based assays in the presence of increasing concentrations of Rapamycin. As indicated in Figure 6, ectopic overexpression of Annexin A13 was sufficient to sensitize MDA-MB-231 cells to Rapamycin. Indeed, Annexin A13-transfected clones remained sensitive to Rapamycin over a wide range of drug concentrations relative to matched controls. Comparable findings were obtained using multiple and different clones of MDA-MB-231 that overexpress Annexin A13, thus precluding that the outcome was unique to a particular clone (data not shown).  Discussion The major finding of our present study is that RHGP can be used to identify targets that control drug resistance in tumor cells. Using RHGP, we identified a set of targets that control tumor cell sensitivity to Rapamycin, a small molecule antitumor cytotoxic agent. By focusing on one particular target, Annexin A13, we demonstrated that decreased Annexin A13 decreases Rapamycin sensitivity and, in the converse experiment, that ectopic overexpression of Annexin A13 sensitizes patient-derived tumor cells to Rapamycin.
One unique feature of our present study is the use of RHGP to identify targets that regulate tumor cell sensitivity to Rapamycin. Previous studies have demonstrated that RHGP provides an opportunity to sample the entire genome for any genetic change able to confer the desired phenotype.7,16-18 In the application studied herein, we utilized RHGP to identify targets that allow tumor cells to override otherwise lethal concentrations of Rapamycin chemotherapy. By focusing on an outcome associated with survival (rather than lethality), we were able to preclude targets associated with potential toxicity (since lethal genetic changes would similarly have resulted in cell death and prevented their identification).
Another strength of the RHGP technology arises in part from the fact that it can directly ascribe biologic outcomes (such as changes in drug sensitivity) that arise either from the overexpression or loss of expression of any gene in the genome. By utilizing an objective measure that can be reliably evaluated in the laboratory, RHGP can provide novel information about both known and unknown targets. For example, 4 of the 16 targets identified herein had been previously linked with cancer, and 2 had been linked with chemotherapy resistance. Nonetheless, most of the targets had not been linked with cancer, and 9 had not yet been annotated. These findings may provide much-needed opportunities to understand or prevent drug-resistant cancer, which remains a significant cause of cancer mortality. In light of published evidence, the identification of CD38 and PIASy as regulators of cancer chemoresistance provides further validation for the idea that RHGP can identify relevant targets and pathways. Consistent with this idea, our colleagues recently utilized RHGP to identify Txr1,18 a target the overexpression of which is associated with paclitaxel resistance of breast cancer cells.
Annexin A13 is 1 member of a family of 13 homologous phospholipid binding proteins.19 Most family members share a structure consisting of 4 domains. All annexins bind calcium, which triggers a translocation from the cytosol to the cell membrane.20 The biologic actions of many annexins have not yet been described, though recent evidence suggests a role for Annexin A13 in vesicular sorting and epithelial cell polarization.19,21 In particular, different forms of Annexin A13 have been implicated in the differentiation of apical versus basolateral polarity.21 A recent study reported that several annexins were significantly down-regulated at the cDNA level in localized prostate cancer samples when compared to benign prostatic tissue.22
To our knowledge, the link between Annexin A13 and drug resistance is unique. Interestingly, a series of recent studies has linked Annexin II with breast and prostate cancers.22,23 Based on our present findings, additional investigations should be conducted to determine Annexin A13 in cancer and, in particular, if and how changes in Annexin A13 expression relate to tumor drug sensitivity. Specifically, our understanding of the role of Annexin A13 would be improved with studies to relate the expression levels in clinical specimens with disease prevalence and outcome. These studies suggest a role for Annexin II in the transmission of signals regulating cell growth, angiogenesis and metastasis. One line of investigation has linked Annexin II with the regulation of protease activity. For example, Annexin II can regulate the cleavage of fibrinogen with tissue plasminogen activator.24 This study further demonstrated that both molecules form a complex with Annexin II at the cell surface. In light of evidence that Annexin A13 is restricted to the cytoplasm and inner leaf of the cell membrane, it is unclear if this elegant model will explain the results observed herein. Annexins have also been shown to regulate the transmission of intracellular signals through multiple and complex pathways. For example, multiple annexin molecules can regulate signaling by ERK kinases,25,26 Ras,27 cyclinD1,26 nitric oxide synthase28 and matrix metalloproteinases.29 Given the widespread understanding of the importance of these different signaling pathways in the regulation of cell growth regulation, the aforementioned links with annexins may provide a foundation to increase our understanding of how Annexin A13 and related molecules can regulate cancer in general and chemoresistance in particular.
References
From Functional Genetics, Inc., Gaithersburg, Maryland, U.S.A.
Dr. Reiske is Scientist. Messrs. Sui and Donahue and Ms. Ung-Medoff are Associate Scientists.
Dr. W.-B. Li is Director. Dr. Goldblatt is Chief Executive Officer.
Dr. L. Li was Chief Scientific Officer and currently is a Member of the Board of Directors. Dr. Kinch is Vice President.
Address correspondence to: Michael S. Kinch, Ph.D., Functional Genetics, Inc., 708 Quince Orchard Road, Gaithersburg, Maryland 20878, U.S.A. (mkinch@functional-genetics.com). Financial Disclosure: All authors are or were employees of Functional Genetics, Inc.
 0884-6812/10/3202-0061/$18.00/0 © Science Printers and Publishers, Inc.
Analytical and Quantitative Cytology and Histology® Send website comments to Webmaster@aqch.com |
||