- Email: [email protected]
DNA microarray analysis of the expression profiles of luteinized granulosa cells as a function of ovarian reserve
REPRODUCTIVE ENDOCRINOLOGY
FERTILITY AND STERILITY威 VOL. 77, NO. 6, JUNE 2002 Copyright ©2002 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.
DNA microarray analysis of the expression profiles of luteinized granulosa cells as a function of ovarian reserve Khew-Voon Chin, Ph.D.,a,b,c David B. Seifer, M.D.,d Bo Feng, Ph.D.,d Yong Lin, Ph.D.,a and Wei-Chung Shih, Ph.D.a The Cancer Institute of New Jersey, UMDNJ—Robert Wood Johnson Medical School, New Brunswick, New Jersey
Received August 10, 2001; revised and accepted November 1, 2001. a The Cancer Institute of New Jersey, UMDNJ—Robert Wood Johnson Medical School. b Department of Medicine, UMDNJ—Robert Wood Johnson Medical School. c Department of Pharmacology, UMDNJ— Robert Wood Johnson Medical School. d Department of Obstetrics, Gynecology and Reproductive Sciences, UMDNJ—Robert Wood Johnson Medical School. Studies were supported in part by National Institutes of Health grants AG15425 (to Dr. Seifer) and CA67722 (to Dr. Chin). Reprint requests: KhewVoon Chin, Ph.D., The Cancer Institute of New Jersey, UMDNJ—Robert Wood Johnson Medical School, 195 Little Albany Street, New Brunswick, NJ 08901 (FAX: 732-235-7493; E-mail: [email protected]). 0015-0282/02/$22.00 PII S0015-0282(02)03114-x
1214
Objective: To examine the expression profiles of luteinized granulosa cells isolated from women with normal or diminished ovarian reserve undergoing in vitro fertilization. Design: Expression profiling by complementary DNA microarray analysis. Setting: Women undergoing in vitro fertilization-embryo transfer in a university-based fertility clinic. Patient(s): Eighteen women with normal or decreased ovarian reserve. Intervention(s): All patients were given gonadotropin stimulation in preparation for IVF with granulosa cells isolated at the time of follicular aspiration. Main Outcome Measure(s): Expression profiles of luteinized granulosa cells isolated from each woman were determined by using DNA microarray analysis. Result(s): Changes in patterns of gene expression in granulosa cells were observed between women with normal and diminished ovarian reserve. These genes included several anonymous expressed sequence tags and also expressed sequence tags that correspond to known genes. Conclusion(s): Expression profiling of granulosa cells by DNA microarray may yield signature patterns that reflect the status of ovarian reserve and the competence of granulosa cells. (Fertil Steril威 2002;77:1214 – 8. ©2002 by American Society for Reproductive Medicine.) Key Words: DNA microarray, gene expression, expression profile, ovarian reserve, granulosa cell
The decline in fecundity in women with age is associated with loss of follicles from the ovary and decrease in oocyte quality. Age increases apoptosis in the follicular granulosa cells and consequently decreases ovarian fecundity (1, 2). Therefore, the viability of the follicular granulosa cells may be essential for development of the oocyte. Granulosa cell competence may be the “pacemaker” for the onset and progression of reproductive senescence in women. Reduction in the number of developing follicles and granulosa cells are characteristics of diminished ovarian reserve, the earliest sign of decreased cellular competence. These individual granulosa cells produce less steroid and less inhibin, have diminished proliferation, and undergo apoptosis at an increased rate, thus negatively affecting follicle development and oocyte quality (3). Apoptotic changes in
granulosa cells increase with age, resulting in the decline in ovarian fecundity (4). Therefore, the viability of the companion granulosa cells may substantially affect the function and growth of oocytes. Transcriptional activity of the oocytes can be modulated when cultured in the presence of granulosa cells, but not as denuded oocytes (5). Therefore, oocyte growth and differentiation requires bidirectional communication between germ cells and companion granulosa cells. Consequently, impaired function or increased apoptosis in granulosa cells with age is expected to negatively affect the competence of the oocytes. It seems likely that age-related changes that affect the competence of the granulosa cells might significantly affect oocyte development and fecundity. We raised the hypothesis that age-related changes may be associated with differential
TABLE 1 Clinical profile of 18 participants selected for DNA microarray analysis. Profile
Women with normal ovarian reserve (n ⫽ 9)
Women with diminished ovarian reserve (n ⫽ 9)
29.6 ⫾ 2.5 3.9 ⫾ 1.4 2913.2 ⫾ 869.5 21 ⫾ 6 44
38 ⫾ 2 10.8 ⫾ 2 1212.9 ⫾ 364.2 7⫾2 11
Age (y)* Day 3 FSH level (IU/L) Peak serum E2 level (pg/mL) No. of oocytes retrieved Clinical pregnancy rate (%) Note: Data with the plus/minus sign are the mean (⫾ SD). Chin. DNA microarray and ovarian reserve. Fertil Steril 2002.
granulosa cell gene expression. Recent use of DNA microarray analysis for molecular identification and classification of diseases that traditional pathologic examination failed to pinpoint strongly suggest that functional identification based on gene expression is feasible using this approach (6, 7). Thus, we prospectively examined the expression patterns of luteinized granulosa cells isolated from follicles of women with either normal or diminished ovarian reserve using DNA microarray. We found alterations in the expression of some genes that implicate their potential role in influencing the transcription program and other functions in luteinized granulosa cells. Our results suggest that the gene changes detected by microarray analysis may predict the competence of the granulosa cells in supporting development of the oocyte.
MATERIALS AND METHODS Patients Isolated luteinizing granulosa cells from women undergoing in vitro fertilization were used for microarray analysis. Samples were obtained from women with normal (n ⫽ 9) and diminished ovarian reserve (n ⫽ 9). The clinical profiles of the study participants are summarized in Table 1. All participants were given GnRH agonist followed by gonadotropin stimulation until they received hCG. The institutional review board at the University of Medicine and Dentistry of New Jersey—Robert Wood Johnson Medical School approved this study.
granulosa cell suspension was further purified with antiCD45 immunomagnetic beads (Beckman Coulter, Miami, FL) to remove residual lymphocytes from the cell suspension. The purified granulosa cells were collected by centrifugation at 7,000 rpm for 5 minutes and frozen in liquid nitrogen for RNA extraction.
Microarrays Human GeneFilters arrays, GF223, were obtained from Research Genetics, Inc. (Huntsville, AL). These arrays, printed on 5 cm ⫻ 7 cm nylon membrane, contained approximately 5,300 expressed sequence tags and complementary DNA elements, with about 2,500 genes corresponding to known genes in the GenBank database. All expressed sequence tag clones have been sequence verified. The array also contains 192 spots of total genomic DNA and 168 housekeeping genes, which serve as reference points for image analysis (Pathways; Research Genetics Inc.), normalization, and for verifying the homogeneity of the hybridization.
RNA Extraction, Probe Preparation and Labeling, Hybridization, and Scanning
Isolation of Granulosa Cells
The RNA was extracted from granulosa cells by using Triazol reagent (Life Technologies, Inc.). Qualities of all RNA samples were monitored by gel electrophoresis before further use. The labeling procedures were conducted as specified by the manufacturer; details of the protocols can be downloaded from the Web site of Research Genetics, Inc. (http://www.resgen.com).
Follicular aspirates from each patient were pooled, and the follicular fluid was centrifuged for 10 minutes at 700 ⫻ g and 20°C. The top layer of cells enriched with granulosa was carefully collected into Hank’s balanced salt solution (free of Ca⫹⫹ and Mg⫹⫹) (Life Technologies, Inc., Carlsbad, CA). The cell suspension was carefully overlaid on top of 50% Percoll medium and centrifuged for 20 min at 700 ⫻ g. After Percoll separation, the granulosa cell– enriched interface layer was collected and washed in Hank’s balanced salt solution. The resultant granulosa cell pellet was incubated in Hank’s medium containing 2 mM ethylenediamene tetreacetic acid for 5 minutes and vortexed at low speed. The
In brief, complementary DNA targets were synthesized from total RNA with 33P-dCTP by oligo dT-primed polymerization by using Superscript II reverse transcriptase (Life Technologies, Inc.). Approximately 5 g of total RNA samples were used in each labeling reaction. The pool of nucleotides in the labeling reaction was 0.5 mM dGTP, dATP, and dTTP and 0.2 mM dCTP. Probes were purified by gel chromatography (BioSpin 6; BioRad, Hercules, CA) and precipitated in ethanol. The probes were then resuspended in 100 L of TE buffer, and an aliquot was drawn for determination of incorporation efficiency. Before hybridization, the solution was boiled for 2 minutes, then allowed to cool to
FERTILITY & STERILITY威
1215
FIGURE 1 Cluster analysis of luteinized granulosa cells isolated from women with normal or diminished ovarian reserve. Each column represents the expression of a participant, and each row represents a single gene. Red squares represent higher than median levels of gene expression (up-regulation), and green squares represent lower than median levels (down-regulation). Participants 01, 13, 24, 29, 38, 40, 41, 42, and 54 have diminished ovarian reserve, and participants 22, 27, 30, 46, 47, 48, 50, 52, and 53 have normal ovarian reserve. The yellow box indicates cluster of genes specific for participants with similar patterns of gene expression that seem to segregate by diminished reserve.
room temperature. An equal number of counts per minute (cpm) was applied from each sample for hybridization at 46°C. At the end of hybridization, filters were washed for 20 minutes in 2 ⫻ sodium chloride/sodium citrate (SSC) and 0.2% sodium dodecyl sulfate at room temperature, and then for 30 minutes in 0.1 ⫻ SSC and 0.2% sodium dodecyl sulfate at 60°C. Arrays were then exposed on phosphorimage screen for the appropriate time and scanned on a Molecular Dynamics Storm Phosphorimager. The scanned images were analyzed by using Pathways software.
Statistical Methods Gene expression data from each participant were normalized by the total counts or the total intensity of gene expression (cpm) and set to a level of 1 ⫻ 107 cpm. For each set of normalized data, the two-sample unequal-variance t-tests were performed for each gene expression. The set of genes that differed between young and older women at a significance level of .01 was then log-transformed, centered by median, and subjected to cluster analyses by using the Cluster and TreeView software suite (8). The similarity metric of centered correlation and average linkage clustering method were used.
RESULTS We chose two groups of patients that would represent extremes of the reproductive age spectrum to increase the power of detection of a given difference (i.e., messenger RNA expression) and to explore whether subtle differences may exist between aging groups. Women selected for microarray analysis had reduced fecundity, diminished ovarian reserve, and increased age compared with controls (Table 1). The day 3 levels of FSH were significantly higher in women with diminished ovarian reserve. In addition, the number of retrievable oocytes and clinical pregnancy rate was lower in the group with diminished ovarian reserve and increased age (Table 1).
Chin. DNA microarray and ovarian reserve. Fertil Steril 2002.
1216 Chin et al.
DNA microarray and ovarian reserve
To assure adequate cell count for the microarray analysis, luteinizing granulosa cells from each patient were pooled and RNA was isolated, labeled, and subjected to hybridization on the array. The expressions of approximately 5,300 genes were analyzed for each patient. We previously compared the coefficient of variations for gene expression within patient by normalizing the microarray data using the total measured intensity of the array, on the assumption that fluctuations should average out over thousands of spots in the array and the total integrated intensity across all the spots should be equal for two samples (9). The normalized data were then analyzed by using the two-sample unequal-variance t-tests, and genes that exhibited a 1.5-fold change in the levels of expression were further subjected to cluster analysis, a hierarchical clustering algorithm that grouped together genes on the basis of similarities in their patterns of expression (8). The data are displayed as cluster diagram and genes Vol. 77, No. 6, June 2002
TABLE 2 Altered gene expression in diminished ovarian reserve. Gene name
Accession number
Transcription factors Forkhead-like 13, transcription factor
AA454609
RING zinc finger protein homologue Contains seven cysteine-rich zinc finger C2H2 zinc finger protein Zinc finger protein 25 Putative RING zinc finger protein Zinc finger protein 195 RNA polymerase II Protein synthesis Pumilio 2 Cell adhesion Notch3 Similar to -1 integrin isoform C Tissue factor pathway inhibitor 2
AA167382 N59422 W80702 AA430741 AA463982 N66117 AA663075
AA496663 W47159 AA418724 AA399473
Function
Differentiation of oviduct and ampulla epithelial cells, effector of cell death
Compensatory up-regulation in response to decreased transcription in aging tissues Regulates oogenesis
Associates with extracellular matrix (ECM), may promote ECM remodeling and turnover
Fold increase
4.2 3.5 2.8 2.6 2.0 2.0 2.0 1.6
1.6 2.0 1.9 1.6
Signal transduction STAT-induced STAT inhibitor 3 Growth regulation Weakly similar to host cell factor C1 p53BP1
AA001219
Up-regulation may be due to treatment with LH
1.5
AA488367 AA521389
Expressed in G(0) phase of cell cycle Functions in signal transduction pathways to promote p53 activity
1.7 1.7
Ovarian function -glucuronidase
AA028921
1.6
AA156251
-glucuronidase is increased in superovulated oocytes derived from abnormal follicles, which would not ovulate under normal physiologic stimuli Mediates progesterone-induced antimitogenic effect
AA446103
Cargo transport receptor for glycoproteins
1.6
Weakly similar to putative progesterone binding protein Lysosomal enzyme function Mannose-specific lectin (MR60) ERGIC-53 Anonymous EST EST (KIAA0266) EST EST EST EST EST
R54592 R63929 AA044565 T52847 H61552 AA464688
2.5
3.2 3.0 2.8 2.3 2.2 2.2
Note: EST ⫽ expressed sequence tags. Chin. DNA microarray and ovarian reserve. Fertil Steril 2002.
having the most similar patterns of expression are placed adjacent to each other along the vertical axis. As shown in the cluster diagram in Figure 1, cluster analysis revealed a subset of genes that were differentially expressed between women with normal ovarian reserved and those with diminished ovarian reserve. These distinct gene changes include several anonymous expressed sequence tags, whose identity and function are unknown and are not found in the GenBank database. Furthermore, the expression of Pumilio, which is involved in oogenesis (10, 11), and the prolactin-regulated STAT-induced STAT inhibitor 3 (12), suggests a potential function of these proteins in luteinized FERTILITY & STERILITY威
granulosa cells. Increased expression of the forkhead-like 13 transcription factor in the women with diminished ovarian reserve is also of note because the forkhead proteins are effectors of cell death and cell cycle arrest (13, 14). Table 2 shows some of the genes expressed in women with diminished ovarian reserve that are distinct from those in the group with normal ovarian reserve. The altered expression and functions of some of these genes seem to be consistent with a decline in the viability of granulosa cells with age. Patient 24 seemed to have an expression profile that 1217
differed from the group with diminished ovarian reserve (Fig. 1), suggesting that other underlying pathologic causes may be responsible for the diminished ovarian reserve in this patient. Similarly, patients 50 and 54 had different expression profiles compared with controls.
DISCUSSION We found that differential gene expression in isolated luteinizing granulosa cells from women with normal or diminished ovarian reserve can be distinguished by DNA microarray analysis. Although the expressions of a large number of genes do not have immediate functional implications in the underlying mechanisms of diminished ovarian reserve, the increased expression in others, including the forkhead-like 13 protein, an effector of cell death and cell cycle arrest (13, 14), is consistent with increased apoptosis in granulosa cells with advancing age. In addition, increased expression of the lysosomal enzyme -glucuronidase, which may be associated with superovulated oocytes derived from abnormal follicles (15), may be a compensatory mechanism for reduced fertility. Therefore, the distinct expression profiles in the isolated luteinizing granulosa cells shown here, which are associated with diminished ovarian reserve, may indicate a decline in the competence of granulosa cells. It is also possible that the expression patterns of luteinizing granulosa cells in women with diminished ovarian reserve may not indicate a state of decline of the granulosa. Some may argue that these changes in gene expression may occur in response to the luteinizing hormone given to these patients. However, the distinct differences in expression between these two groups of women after luteinizing hormone treatment are more likely due to an altered response in the granulosa cells with age. Expression profiles in one patient with diminished ovarian reserve (patient 24) and two patients (patients 50 and 54) in the control group differed from those of their respective clusters. We speculate that these differences in gene expression may be the result of genetic polymorphisms between individuals that give rise to idiosyncratic variation in gene expression patterns, difference in the pathogenic mechanisms of diminished ovarian reserve in patient 24 compared with the other patients, or differences in the genetic response to luteinizing hormone between individuals. Patients with diminished ovarian reserve generally receive more aggressive ovulation induction than do women with normal ovarian reserve, in an effort to reach the point of egg retrieval. In a clinical study involving a group of patients with diminished ovarian reserve, such patients may not receive the same ovulation induction protocol as a group without diminished ovarian reserve, owing to concern that the former group would not reach egg retrieval. We cannot
1218 Chin et al.
DNA microarray and ovarian reserve
rule out that the observed differential gene expression might be in part a result of differences in drug treatment between the two groups. Given that the two study groups have different clinical profiles in terms of ovarian reserve, the observed gene expression patterns may reflect the compromised response of the granulosa cells to the agents involved in ovulation induction (i.e., GnRH agonist and gonadotropins) in women with diminished ovarian reserve. These changes may have significant implications for the mechanisms of diminished ovarian reserve and reduced fecundity related to the functional status of granulosa cells. Despite these limitations, our data suggest that expression profiling of granulosa cells by DNA microarray analysis yields signature patterns that reflect the status of ovarian reserve with age. Future analysis of a larger cohort and a further increase in number of gene elements on the arrays will yield fruitful insights on the differential gene expression that we observed. References 1. Seifer DB, Naftolin F. Moving toward an earlier and better understanding of perimenopause. Fertil Steril 1998;69:387– 8. 2. Bopp BL, Seifer DB. Oocyte loss and the perimenopause. Clin Obstet Gynecol 1998;41:898 –911. 3. Seifer DB, Gardiner AC, Lambert-Messerlian G, Schneyer AL. Differential secretion of dimeric inhibin in cultured luteinized granulosa cells as a function of ovarian reserve. J Clin Endocrinol Metab 1996;81: 736 –9. 4. Seifer DB, Gardiner AC, Ferreira KA, Peluso JJ. Apoptosis as a function of ovarian reserve in women undergoing in vitro fertilization. Fertil Steril 1996;66:593– 8. 5. De la Fuente R, Eppig JJ. Transcriptional activity of the mouse oocyte genome: companion granulosa cells modulate transcription and chromatin remodeling. Dev Biol 2001;229:224 –36. 6. Golub TR, Slonim DK, Tamayo P, Huard C, Gaasenbeek M, Mesirov JP, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 1999;286:531–7. 7. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 2000;403:503–11. 8. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci U S A 1998;95:14863– 8. 9. Hegde P, Qi R, Abernathy K, Gay C, Dharap S, Gaspard R, et al. A concise guide to cDNA microarray analysis. Biotechniques 2000;29: 548 –56. 10. Asaoka-Taguchi M, Yamada M, Nakamura A, Hanyu K, Kobayashi S. Maternal Pumilio acts together with Nanos in germline development in Drosophila embryos. Nat Cell Biol 1999;1:431–7. 11. Subramaniam K, Seydoux G. nos-1 and nos-2, two genes related to Drosophila nanos, regulate primordial germ cell development and survival in Caenorhabditis elegans. Development 1999;126:4861–71. 12. Peters CA, Maizels ET, Robertson MC, Shiu RP, Soloff MS, Hunzicker-Dunn M. Induction of relaxin messenger RNA expression in response to prolactin receptor activation requires protein kinase C delta signaling. Mol Endocrinol 2000;14:576 –90. 13. Tang ED, Nunez G, Barr FG, Guan KL. Negative regulation of the forkhead transcription factor FKHR by Akt. J Biol Chem 1999;274: 16741– 6. 14. Nakamura N, Ramaswamy S, Vazquez F, Signoretti S, Loda M, Sellers WR. Forkhead transcription factors are critical effectors of cell death and cell cycle arrest downstream of PTEN. Mol Cell Biol 2000;20: 8969 – 82. 15. Banos ME, Rosales AM, Ballesteros LM, Hernandez-Perez O, Rosado A. Changes in lysosomal enzyme activities in pre-ovulatory follicles and endometrium of PMSG superovulated rats. Arch Med Res 1996;27:49 –55.
Vol. 77, No. 6, June 2002
FERTILITY AND STERILITY威 VOL. 77, NO. 6, JUNE 2002 Copyright ©2002 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.
DNA microarray analysis of the expression profiles of luteinized granulosa cells as a function of ovarian reserve Khew-Voon Chin, Ph.D.,a,b,c David B. Seifer, M.D.,d Bo Feng, Ph.D.,d Yong Lin, Ph.D.,a and Wei-Chung Shih, Ph.D.a The Cancer Institute of New Jersey, UMDNJ—Robert Wood Johnson Medical School, New Brunswick, New Jersey
Received August 10, 2001; revised and accepted November 1, 2001. a The Cancer Institute of New Jersey, UMDNJ—Robert Wood Johnson Medical School. b Department of Medicine, UMDNJ—Robert Wood Johnson Medical School. c Department of Pharmacology, UMDNJ— Robert Wood Johnson Medical School. d Department of Obstetrics, Gynecology and Reproductive Sciences, UMDNJ—Robert Wood Johnson Medical School. Studies were supported in part by National Institutes of Health grants AG15425 (to Dr. Seifer) and CA67722 (to Dr. Chin). Reprint requests: KhewVoon Chin, Ph.D., The Cancer Institute of New Jersey, UMDNJ—Robert Wood Johnson Medical School, 195 Little Albany Street, New Brunswick, NJ 08901 (FAX: 732-235-7493; E-mail: [email protected]). 0015-0282/02/$22.00 PII S0015-0282(02)03114-x
1214
Objective: To examine the expression profiles of luteinized granulosa cells isolated from women with normal or diminished ovarian reserve undergoing in vitro fertilization. Design: Expression profiling by complementary DNA microarray analysis. Setting: Women undergoing in vitro fertilization-embryo transfer in a university-based fertility clinic. Patient(s): Eighteen women with normal or decreased ovarian reserve. Intervention(s): All patients were given gonadotropin stimulation in preparation for IVF with granulosa cells isolated at the time of follicular aspiration. Main Outcome Measure(s): Expression profiles of luteinized granulosa cells isolated from each woman were determined by using DNA microarray analysis. Result(s): Changes in patterns of gene expression in granulosa cells were observed between women with normal and diminished ovarian reserve. These genes included several anonymous expressed sequence tags and also expressed sequence tags that correspond to known genes. Conclusion(s): Expression profiling of granulosa cells by DNA microarray may yield signature patterns that reflect the status of ovarian reserve and the competence of granulosa cells. (Fertil Steril威 2002;77:1214 – 8. ©2002 by American Society for Reproductive Medicine.) Key Words: DNA microarray, gene expression, expression profile, ovarian reserve, granulosa cell
The decline in fecundity in women with age is associated with loss of follicles from the ovary and decrease in oocyte quality. Age increases apoptosis in the follicular granulosa cells and consequently decreases ovarian fecundity (1, 2). Therefore, the viability of the follicular granulosa cells may be essential for development of the oocyte. Granulosa cell competence may be the “pacemaker” for the onset and progression of reproductive senescence in women. Reduction in the number of developing follicles and granulosa cells are characteristics of diminished ovarian reserve, the earliest sign of decreased cellular competence. These individual granulosa cells produce less steroid and less inhibin, have diminished proliferation, and undergo apoptosis at an increased rate, thus negatively affecting follicle development and oocyte quality (3). Apoptotic changes in
granulosa cells increase with age, resulting in the decline in ovarian fecundity (4). Therefore, the viability of the companion granulosa cells may substantially affect the function and growth of oocytes. Transcriptional activity of the oocytes can be modulated when cultured in the presence of granulosa cells, but not as denuded oocytes (5). Therefore, oocyte growth and differentiation requires bidirectional communication between germ cells and companion granulosa cells. Consequently, impaired function or increased apoptosis in granulosa cells with age is expected to negatively affect the competence of the oocytes. It seems likely that age-related changes that affect the competence of the granulosa cells might significantly affect oocyte development and fecundity. We raised the hypothesis that age-related changes may be associated with differential
TABLE 1 Clinical profile of 18 participants selected for DNA microarray analysis. Profile
Women with normal ovarian reserve (n ⫽ 9)
Women with diminished ovarian reserve (n ⫽ 9)
29.6 ⫾ 2.5 3.9 ⫾ 1.4 2913.2 ⫾ 869.5 21 ⫾ 6 44
38 ⫾ 2 10.8 ⫾ 2 1212.9 ⫾ 364.2 7⫾2 11
Age (y)* Day 3 FSH level (IU/L) Peak serum E2 level (pg/mL) No. of oocytes retrieved Clinical pregnancy rate (%) Note: Data with the plus/minus sign are the mean (⫾ SD). Chin. DNA microarray and ovarian reserve. Fertil Steril 2002.
granulosa cell gene expression. Recent use of DNA microarray analysis for molecular identification and classification of diseases that traditional pathologic examination failed to pinpoint strongly suggest that functional identification based on gene expression is feasible using this approach (6, 7). Thus, we prospectively examined the expression patterns of luteinized granulosa cells isolated from follicles of women with either normal or diminished ovarian reserve using DNA microarray. We found alterations in the expression of some genes that implicate their potential role in influencing the transcription program and other functions in luteinized granulosa cells. Our results suggest that the gene changes detected by microarray analysis may predict the competence of the granulosa cells in supporting development of the oocyte.
MATERIALS AND METHODS Patients Isolated luteinizing granulosa cells from women undergoing in vitro fertilization were used for microarray analysis. Samples were obtained from women with normal (n ⫽ 9) and diminished ovarian reserve (n ⫽ 9). The clinical profiles of the study participants are summarized in Table 1. All participants were given GnRH agonist followed by gonadotropin stimulation until they received hCG. The institutional review board at the University of Medicine and Dentistry of New Jersey—Robert Wood Johnson Medical School approved this study.
granulosa cell suspension was further purified with antiCD45 immunomagnetic beads (Beckman Coulter, Miami, FL) to remove residual lymphocytes from the cell suspension. The purified granulosa cells were collected by centrifugation at 7,000 rpm for 5 minutes and frozen in liquid nitrogen for RNA extraction.
Microarrays Human GeneFilters arrays, GF223, were obtained from Research Genetics, Inc. (Huntsville, AL). These arrays, printed on 5 cm ⫻ 7 cm nylon membrane, contained approximately 5,300 expressed sequence tags and complementary DNA elements, with about 2,500 genes corresponding to known genes in the GenBank database. All expressed sequence tag clones have been sequence verified. The array also contains 192 spots of total genomic DNA and 168 housekeeping genes, which serve as reference points for image analysis (Pathways; Research Genetics Inc.), normalization, and for verifying the homogeneity of the hybridization.
RNA Extraction, Probe Preparation and Labeling, Hybridization, and Scanning
Isolation of Granulosa Cells
The RNA was extracted from granulosa cells by using Triazol reagent (Life Technologies, Inc.). Qualities of all RNA samples were monitored by gel electrophoresis before further use. The labeling procedures were conducted as specified by the manufacturer; details of the protocols can be downloaded from the Web site of Research Genetics, Inc. (http://www.resgen.com).
Follicular aspirates from each patient were pooled, and the follicular fluid was centrifuged for 10 minutes at 700 ⫻ g and 20°C. The top layer of cells enriched with granulosa was carefully collected into Hank’s balanced salt solution (free of Ca⫹⫹ and Mg⫹⫹) (Life Technologies, Inc., Carlsbad, CA). The cell suspension was carefully overlaid on top of 50% Percoll medium and centrifuged for 20 min at 700 ⫻ g. After Percoll separation, the granulosa cell– enriched interface layer was collected and washed in Hank’s balanced salt solution. The resultant granulosa cell pellet was incubated in Hank’s medium containing 2 mM ethylenediamene tetreacetic acid for 5 minutes and vortexed at low speed. The
In brief, complementary DNA targets were synthesized from total RNA with 33P-dCTP by oligo dT-primed polymerization by using Superscript II reverse transcriptase (Life Technologies, Inc.). Approximately 5 g of total RNA samples were used in each labeling reaction. The pool of nucleotides in the labeling reaction was 0.5 mM dGTP, dATP, and dTTP and 0.2 mM dCTP. Probes were purified by gel chromatography (BioSpin 6; BioRad, Hercules, CA) and precipitated in ethanol. The probes were then resuspended in 100 L of TE buffer, and an aliquot was drawn for determination of incorporation efficiency. Before hybridization, the solution was boiled for 2 minutes, then allowed to cool to
FERTILITY & STERILITY威
1215
FIGURE 1 Cluster analysis of luteinized granulosa cells isolated from women with normal or diminished ovarian reserve. Each column represents the expression of a participant, and each row represents a single gene. Red squares represent higher than median levels of gene expression (up-regulation), and green squares represent lower than median levels (down-regulation). Participants 01, 13, 24, 29, 38, 40, 41, 42, and 54 have diminished ovarian reserve, and participants 22, 27, 30, 46, 47, 48, 50, 52, and 53 have normal ovarian reserve. The yellow box indicates cluster of genes specific for participants with similar patterns of gene expression that seem to segregate by diminished reserve.
room temperature. An equal number of counts per minute (cpm) was applied from each sample for hybridization at 46°C. At the end of hybridization, filters were washed for 20 minutes in 2 ⫻ sodium chloride/sodium citrate (SSC) and 0.2% sodium dodecyl sulfate at room temperature, and then for 30 minutes in 0.1 ⫻ SSC and 0.2% sodium dodecyl sulfate at 60°C. Arrays were then exposed on phosphorimage screen for the appropriate time and scanned on a Molecular Dynamics Storm Phosphorimager. The scanned images were analyzed by using Pathways software.
Statistical Methods Gene expression data from each participant were normalized by the total counts or the total intensity of gene expression (cpm) and set to a level of 1 ⫻ 107 cpm. For each set of normalized data, the two-sample unequal-variance t-tests were performed for each gene expression. The set of genes that differed between young and older women at a significance level of .01 was then log-transformed, centered by median, and subjected to cluster analyses by using the Cluster and TreeView software suite (8). The similarity metric of centered correlation and average linkage clustering method were used.
RESULTS We chose two groups of patients that would represent extremes of the reproductive age spectrum to increase the power of detection of a given difference (i.e., messenger RNA expression) and to explore whether subtle differences may exist between aging groups. Women selected for microarray analysis had reduced fecundity, diminished ovarian reserve, and increased age compared with controls (Table 1). The day 3 levels of FSH were significantly higher in women with diminished ovarian reserve. In addition, the number of retrievable oocytes and clinical pregnancy rate was lower in the group with diminished ovarian reserve and increased age (Table 1).
Chin. DNA microarray and ovarian reserve. Fertil Steril 2002.
1216 Chin et al.
DNA microarray and ovarian reserve
To assure adequate cell count for the microarray analysis, luteinizing granulosa cells from each patient were pooled and RNA was isolated, labeled, and subjected to hybridization on the array. The expressions of approximately 5,300 genes were analyzed for each patient. We previously compared the coefficient of variations for gene expression within patient by normalizing the microarray data using the total measured intensity of the array, on the assumption that fluctuations should average out over thousands of spots in the array and the total integrated intensity across all the spots should be equal for two samples (9). The normalized data were then analyzed by using the two-sample unequal-variance t-tests, and genes that exhibited a 1.5-fold change in the levels of expression were further subjected to cluster analysis, a hierarchical clustering algorithm that grouped together genes on the basis of similarities in their patterns of expression (8). The data are displayed as cluster diagram and genes Vol. 77, No. 6, June 2002
TABLE 2 Altered gene expression in diminished ovarian reserve. Gene name
Accession number
Transcription factors Forkhead-like 13, transcription factor
AA454609
RING zinc finger protein homologue Contains seven cysteine-rich zinc finger C2H2 zinc finger protein Zinc finger protein 25 Putative RING zinc finger protein Zinc finger protein 195 RNA polymerase II Protein synthesis Pumilio 2 Cell adhesion Notch3 Similar to -1 integrin isoform C Tissue factor pathway inhibitor 2
AA167382 N59422 W80702 AA430741 AA463982 N66117 AA663075
AA496663 W47159 AA418724 AA399473
Function
Differentiation of oviduct and ampulla epithelial cells, effector of cell death
Compensatory up-regulation in response to decreased transcription in aging tissues Regulates oogenesis
Associates with extracellular matrix (ECM), may promote ECM remodeling and turnover
Fold increase
4.2 3.5 2.8 2.6 2.0 2.0 2.0 1.6
1.6 2.0 1.9 1.6
Signal transduction STAT-induced STAT inhibitor 3 Growth regulation Weakly similar to host cell factor C1 p53BP1
AA001219
Up-regulation may be due to treatment with LH
1.5
AA488367 AA521389
Expressed in G(0) phase of cell cycle Functions in signal transduction pathways to promote p53 activity
1.7 1.7
Ovarian function -glucuronidase
AA028921
1.6
AA156251
-glucuronidase is increased in superovulated oocytes derived from abnormal follicles, which would not ovulate under normal physiologic stimuli Mediates progesterone-induced antimitogenic effect
AA446103
Cargo transport receptor for glycoproteins
1.6
Weakly similar to putative progesterone binding protein Lysosomal enzyme function Mannose-specific lectin (MR60) ERGIC-53 Anonymous EST EST (KIAA0266) EST EST EST EST EST
R54592 R63929 AA044565 T52847 H61552 AA464688
2.5
3.2 3.0 2.8 2.3 2.2 2.2
Note: EST ⫽ expressed sequence tags. Chin. DNA microarray and ovarian reserve. Fertil Steril 2002.
having the most similar patterns of expression are placed adjacent to each other along the vertical axis. As shown in the cluster diagram in Figure 1, cluster analysis revealed a subset of genes that were differentially expressed between women with normal ovarian reserved and those with diminished ovarian reserve. These distinct gene changes include several anonymous expressed sequence tags, whose identity and function are unknown and are not found in the GenBank database. Furthermore, the expression of Pumilio, which is involved in oogenesis (10, 11), and the prolactin-regulated STAT-induced STAT inhibitor 3 (12), suggests a potential function of these proteins in luteinized FERTILITY & STERILITY威
granulosa cells. Increased expression of the forkhead-like 13 transcription factor in the women with diminished ovarian reserve is also of note because the forkhead proteins are effectors of cell death and cell cycle arrest (13, 14). Table 2 shows some of the genes expressed in women with diminished ovarian reserve that are distinct from those in the group with normal ovarian reserve. The altered expression and functions of some of these genes seem to be consistent with a decline in the viability of granulosa cells with age. Patient 24 seemed to have an expression profile that 1217
differed from the group with diminished ovarian reserve (Fig. 1), suggesting that other underlying pathologic causes may be responsible for the diminished ovarian reserve in this patient. Similarly, patients 50 and 54 had different expression profiles compared with controls.
DISCUSSION We found that differential gene expression in isolated luteinizing granulosa cells from women with normal or diminished ovarian reserve can be distinguished by DNA microarray analysis. Although the expressions of a large number of genes do not have immediate functional implications in the underlying mechanisms of diminished ovarian reserve, the increased expression in others, including the forkhead-like 13 protein, an effector of cell death and cell cycle arrest (13, 14), is consistent with increased apoptosis in granulosa cells with advancing age. In addition, increased expression of the lysosomal enzyme -glucuronidase, which may be associated with superovulated oocytes derived from abnormal follicles (15), may be a compensatory mechanism for reduced fertility. Therefore, the distinct expression profiles in the isolated luteinizing granulosa cells shown here, which are associated with diminished ovarian reserve, may indicate a decline in the competence of granulosa cells. It is also possible that the expression patterns of luteinizing granulosa cells in women with diminished ovarian reserve may not indicate a state of decline of the granulosa. Some may argue that these changes in gene expression may occur in response to the luteinizing hormone given to these patients. However, the distinct differences in expression between these two groups of women after luteinizing hormone treatment are more likely due to an altered response in the granulosa cells with age. Expression profiles in one patient with diminished ovarian reserve (patient 24) and two patients (patients 50 and 54) in the control group differed from those of their respective clusters. We speculate that these differences in gene expression may be the result of genetic polymorphisms between individuals that give rise to idiosyncratic variation in gene expression patterns, difference in the pathogenic mechanisms of diminished ovarian reserve in patient 24 compared with the other patients, or differences in the genetic response to luteinizing hormone between individuals. Patients with diminished ovarian reserve generally receive more aggressive ovulation induction than do women with normal ovarian reserve, in an effort to reach the point of egg retrieval. In a clinical study involving a group of patients with diminished ovarian reserve, such patients may not receive the same ovulation induction protocol as a group without diminished ovarian reserve, owing to concern that the former group would not reach egg retrieval. We cannot
1218 Chin et al.
DNA microarray and ovarian reserve
rule out that the observed differential gene expression might be in part a result of differences in drug treatment between the two groups. Given that the two study groups have different clinical profiles in terms of ovarian reserve, the observed gene expression patterns may reflect the compromised response of the granulosa cells to the agents involved in ovulation induction (i.e., GnRH agonist and gonadotropins) in women with diminished ovarian reserve. These changes may have significant implications for the mechanisms of diminished ovarian reserve and reduced fecundity related to the functional status of granulosa cells. Despite these limitations, our data suggest that expression profiling of granulosa cells by DNA microarray analysis yields signature patterns that reflect the status of ovarian reserve with age. Future analysis of a larger cohort and a further increase in number of gene elements on the arrays will yield fruitful insights on the differential gene expression that we observed. References 1. Seifer DB, Naftolin F. Moving toward an earlier and better understanding of perimenopause. Fertil Steril 1998;69:387– 8. 2. Bopp BL, Seifer DB. Oocyte loss and the perimenopause. Clin Obstet Gynecol 1998;41:898 –911. 3. Seifer DB, Gardiner AC, Lambert-Messerlian G, Schneyer AL. Differential secretion of dimeric inhibin in cultured luteinized granulosa cells as a function of ovarian reserve. J Clin Endocrinol Metab 1996;81: 736 –9. 4. Seifer DB, Gardiner AC, Ferreira KA, Peluso JJ. Apoptosis as a function of ovarian reserve in women undergoing in vitro fertilization. Fertil Steril 1996;66:593– 8. 5. De la Fuente R, Eppig JJ. Transcriptional activity of the mouse oocyte genome: companion granulosa cells modulate transcription and chromatin remodeling. Dev Biol 2001;229:224 –36. 6. Golub TR, Slonim DK, Tamayo P, Huard C, Gaasenbeek M, Mesirov JP, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 1999;286:531–7. 7. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 2000;403:503–11. 8. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci U S A 1998;95:14863– 8. 9. Hegde P, Qi R, Abernathy K, Gay C, Dharap S, Gaspard R, et al. A concise guide to cDNA microarray analysis. Biotechniques 2000;29: 548 –56. 10. Asaoka-Taguchi M, Yamada M, Nakamura A, Hanyu K, Kobayashi S. Maternal Pumilio acts together with Nanos in germline development in Drosophila embryos. Nat Cell Biol 1999;1:431–7. 11. Subramaniam K, Seydoux G. nos-1 and nos-2, two genes related to Drosophila nanos, regulate primordial germ cell development and survival in Caenorhabditis elegans. Development 1999;126:4861–71. 12. Peters CA, Maizels ET, Robertson MC, Shiu RP, Soloff MS, Hunzicker-Dunn M. Induction of relaxin messenger RNA expression in response to prolactin receptor activation requires protein kinase C delta signaling. Mol Endocrinol 2000;14:576 –90. 13. Tang ED, Nunez G, Barr FG, Guan KL. Negative regulation of the forkhead transcription factor FKHR by Akt. J Biol Chem 1999;274: 16741– 6. 14. Nakamura N, Ramaswamy S, Vazquez F, Signoretti S, Loda M, Sellers WR. Forkhead transcription factors are critical effectors of cell death and cell cycle arrest downstream of PTEN. Mol Cell Biol 2000;20: 8969 – 82. 15. Banos ME, Rosales AM, Ballesteros LM, Hernandez-Perez O, Rosado A. Changes in lysosomal enzyme activities in pre-ovulatory follicles and endometrium of PMSG superovulated rats. Arch Med Res 1996;27:49 –55.
Vol. 77, No. 6, June 2002