Search tips
Search criteria 


Logo of amjepidLink to Publisher's site
Am J Epidemiol. 2013 January 1; 177(1): 75–83.
Published online 2012 December 6. doi:  10.1093/aje/kws221
PMCID: PMC3590036

Assessment of Autoantibodies to Meningioma in a Population-based Study


Meningioma is an intracranial tumor with few confirmed risk factors. Recent research points to an impact on meningioma risk from factors related to immune function and development, such as allergy, immunoglobulin E, and Varicella infection status. To further explore an association with immune function, the authors assessed individual seroreactivity to meningioma tumor-associated antigens among participants enrolled in a multicenter, population-based US case-control study of meningioma (2006–2009). Serum samples from cases (n = 349) and controls (n = 348) were screened for autoantibody reactivity to 3 proteins identified in previous studies: enolase 1 (ENO1), NK-tumor recognition protein (NKTR), and nuclear mitotic apparatus protein 1 (NUMA1). Case-control differences were not strong overall (adjusted odds ratio (OR)ENO1 (continuous) = 1.1, 95% confidence interval (CI): 0.6, 1.9 (Ptrend = 0.3); adjusted ORNKTR (continuous) = 1.3, 95% CI: 0.7, 2.4 (Ptrend = 0.02); and adjusted ORNUMA1 (continuous) = 1.1, 95% CI: 0.7, 1.8 (Ptrend = 0.06)); however, antibodies to NKTR and NUMA1 were detected at higher levels in cases than in controls, particularly among men (for men, adjusted ORENO1 (continuous) = 1.6, 95% CI: 0.5, 4.7 (Ptrend = 0.24); adjusted ORNKTR (continuous) = 4.3, 95% CI: 1.2, 15 (Ptrend = 0.009); and adjusted ORNUMA1 (continuous) = 3.6, 95% CI: 1.1, 11 (Ptrend = 0.006)). These results indicate that men with meningioma commonly react with a serologic antimeningioma response; if supported by further research, this finding suggests a distinctive etiology for meningioma in men.

Keywords: autoantibodies, brain neoplasms, genetics, immunologic factors, meningioma, neurosurgery, risk factors

Meningioma accounts for 34% of brain tumors in the United States (2002–2006) and is the most frequently diagnosed primary tumor of the brain and central nervous system (1). However, risk factors for meningioma remain relatively unexplored. Besides gender (incidence is twice as high in women) and increasing age (2, 3), ionizing radiation exposure to the head is the only established risk factor. We and other investigators recently found that immune factors may influence the etiology of meningioma, including a less frequent report of allergies or certain allergic conditions and Varicella virus-associated disease (chickenpox and shingles) in meningioma cases compared with controls (410). Such observations raise the possibility that immune surveillance mechanisms may affect meningioma incidence via immune suppression of nascent tumors. If such a mechanism exists, we may be able to capitalize on it for early detection and treatment strategies.

Neoplasms express proteins, termed “tumor-associated antigens” (TAA), which are new to the immune system and may elicit a reaction. These may be embryonic proteins which are reexpressed, mutant oncogenic proteins, overexpressed proteins, or those otherwise processed in an unusual manner. When presented to the immune system with proper costimulation, these TAA may stimulate the production of autoantibodies. Such antibodies may help suppress tumor growth and can also serve as diagnostic and/or prognostic biomarkers (11). Meese and colleagues have extensively used a technique called SEREX (serologic identification of recombinantly expressed clones) to identify autoantibodies present in meningioma patients that are not expressed among controls (1215). The technique involves the expression of a “library” of proteins derived from meningioma or other brain cancers in bacteriophage colonies on a Petri plate, the transfer of such antigens to a membrane, and the subsequent screening of patient and control sera for reactive antibodies to the antigens on the membranes. The development of autoantibodies to specific proteins in individual patients depends on epitope availability (restricted by human leukocyte antigen genotype) and the state of immune costimulation in the patient and tumor. This approach yielded several dozen autoantigens, none of which individually identify meningioma patients with accuracy but as a panel had a great deal of classification accuracy (15). This has evolved into an approach that targets multiple antigens from a single serum sample with an analytical algorithm to help distinguish patients from controls (14, 15), with the goal of developing a serologic test for a tumor in the central nervous system.

Autoantibodies are typically assessed in convenience clinical samples and not in a population-based or comprehensive manner. We sought to understand the nature of autoantibody reactivity in our meningioma case-control study and to examine whether autoantibody reactions against the tumor might vary by allergy status, gender, age, and other epidemiologic characteristics related to meningioma risk factors. This use of autoantibody profiles in an epidemiologic study is unique and may shed light on the etiology of meningioma. With the use of a semiautomated flow cytometry-based detection system, 3 TAA shown previously to be particularly useful in autoantibody screens were chosen for this analysis: enolase 1 (ENO1), NK-tumor recognition protein (NKTR), and nuclear mitotic apparatus protein 1 (NUMA1). Two of these autoantibody targets had among the highest abilities to discriminate between meningioma cases and controls in a recent report (15), and the other (NKTR) was a top hit in a previous high-dimension serologic screen (12).


Study participants

Our case and control sera were derived from a larger group of study participants in the Meningioma Consortium Study (4). Eligible cases included all persons aged 20–79 years who were newly diagnosed with pathologically confirmed intracranial meningioma between May 1, 2006, and December 12, 2009, among residents of the states of Connecticut, Massachusetts, and North Carolina, as well as several counties in California (Alameda, San Francisco, Contra Costa, Marin, San Mateo, and Santa Clara) and Texas (Brazoria, Fort Bend, Harris, Montgomery, Chambers, Galveston, Liberty, and Waller). Cases were identified through the Rapid Case Ascertainment systems at most sites (Connecticut, North Carolina, and California) and through review of hospital pathology departments and statistics-based tumor registries at sites without a formal Rapid Case Ascertainment mechanism (Texas and Massachusetts). Controls were selected through random-digit dialing methods by an outside consulting firm (Kreider Research and Consulting, Waterville, Maine) and were frequency-matched to cases by 5-year age interval, gender, and state of residence. Cases or controls with a previous history of meningioma and/or a brain lesion of unknown diagnosis were excluded.

The study, consent forms, and data collection instruments were approved by the institutional review boards at Yale University School of Medicine, the Connecticut Department of Public Health, Brigham and Women's Hospital, the University of California, San Francisco, MD Anderson Cancer Center, and Duke University School of Medicine. Consent, interviews, and sample collection were performed as previously described in detail (4). Briefly, study participants were invited to participate with a letter of introduction. Following this, trained telephone interviewers called each participant and performed the interview or scheduled it for a future date. Interviews, available in English or Spanish, had a mean duration of 43 minutes, and consent for biologic sampling was discussed.

Collection of biologic samples was subsequently performed by trained phlebotomists who would visit the home, and specimens were immediately shipped to the University of California, San Francisco, for processing. One tube of blood with no additives (red top) was used to collect serum which was used for the current study. For the current analysis, we selected approximately 350 cases at random and 350 controls (from the available control group) from the sample bank, frequency-matched by gender and age (5-year categories). Because of the female predominance of the disease, all available male cases were chosen (from 858 available serum samples). Case sera were chosen at random from the sample bank and then matched to appropriate controls (selected at random from age- (in 5-year categories) and gender-matched groups). Sera were obtained from cases at a mean of 14 months (standard deviation, 8.5 months; median, 11 months; range, 5–43 months) after surgery, and no case patients were under treatment with cytotoxic chemotherapy or radiation therapy.

Measurement of serum autoantibody levels

Autoantibody targets were produced as recombinant glutathione S-transferase-tagged proteins in cell-free wheat-germ extracts (Abnova Corporation, Taipei, Taiwan). Proteins were purified on glutathione columns, and glutathione S-transferase tags were removed by proteolytic digestion and further purified using size exclusion chromatography. Five μg of protein was attached to 150 μL of carboxylated magnetic Luminex microspheres using a labeling kit (Bio-Rad Laboratories, Hercules, California). Human serum albumin (Sigma catalog A3782)-bound beads (Sigma-Aldrich Corporation, St. Louis, Missouri) were used as a control for nonspecific binding (serum “matrix effect”) (16). Therefore, 4 separate bead sets were used in multiplex. All samples were randomized in blocks of 44 samples with equal numbers of cases and controls in each block. One block of samples was tested on each plate. Incubation and washes were performed as follows: Sera were diluted and incubated in 150 μL of assay buffer with 106 labeled beads for 2 hours at room temperature with shaking. Three washes were performed with wash buffer (Bio-Plex automated wash station; Bio-Rad Laboratories). Secondary biotin-labeled mouse anti-human immunoglobulin G (BD Biosciences catalog 555869; BD Biosciences, San Jose, California) diluted 1:1,000 in 100 μL of detection antibody diluent (Bio-Rad Laboratories) was incubated for 30 minutes with shaking, followed by 3 washes in wash buffer. Assays were built by performing limiting dilutions of test sera in assay buffer (Bio-Rad Laboratories) to determine the assay titer for each target protein. A 1:100 dilution of sera:sample diluent (phosphate-buffered saline + 10% fetal bovine serum + 2.5% heterophile antibody blocking agent (catalog item CBS-K; EMD Millipore, Billerica, Massachusetts)) was determined to be optimal, and all assays were conducted singly, paired, and all together on a test series of sera to determine the potential for cross-reactivity. Lack of cross-reactivity interaction was confirmed. All study serum samples were randomized and assayed in duplicate. The resulting data were normalized across all plates to median values of total combined autoantibody levels on each plate. Standard reference samples were assayed on each plate to confirm assay consistency.

Statistical analysis

Data were analyzed using SAS 9.2 (SAS Institute Inc., Cary, North Carolina). Serum autoantibody levels were examined as the log-transformed values of the ratio of the antibody measure to the bovine serum albumin level and categorized on the basis of quartiles of the distribution in controls. The associations between discrete conditions and exposures of interest and meningioma were assessed using chi-square tests. Spearman rank correlations were used to assess pairwise correlations among autoantibody levels, and Kruskal-Wallis and Wilcoxon tests were used to assess differences between groups. Odds ratios as estimates of the relative risk of meningioma related to autoantibody level (as both continuous and discrete variables) were computed using multivariable logistic regression models. Potential confounding was assessed for education, history of allergies, asthma, and smoking history. Factors that altered point estimates by at least 10% were considered statistical confounders and included in the final model (among these factors, only allergy and education were confounders and were included in some models, as indicated in table footnotes). Statistical interaction was assessed using −2 log likelihood tests comparing the models with and without the cross-product term for the interaction of interest. For these analyses, interaction effects with P values less than or equal to 0.10 were considered statistically significant. The area under the receiver operating characteristic (ROC) curve (plotted as sensitivity vs. 1 – specificity) was computed for each autoantibody. The area under the curve (AUC) for each autoantibody was also tested to determine the difference from chance. All models included adjustment for age in addition to confounders or effect modifiers, and analyses were stratified by gender.


Among 1,914 participants with available data at baseline (of whom 858 had sera available for analysis), 697 were randomly selected for this analysis. There were 349 cases tested and 779 not tested, and 348 controls tested and 538 not tested. Cases selected for autoantibody testing were slightly older than those not selected (59.4 years vs. 57.0 years; P = 0.001), and selected controls were also slightly older than those not selected (58.4 years vs. 56.3 years; P = 0.014). A higher percentage of tested cases were men (36% of tested cases vs. 23% of nontested cases; P = 0.0001), whereas among controls the percentages of men among those tested and not tested did not differ (both 36%). Tested and nontested participants differed by race, with those tested being more likely to be white among both cases (86% vs. 82%; P = 0.2) and controls (90% vs. 82%; P = 0.004). Years of education also were not related to testing status for cases or controls. Tested cases and controls did not differ from their nontested counterparts with regard to reports of allergy, chickenpox, asthma, or smoking (all P's > 0.2). Further, none of these variables were found to be statistical confounders of the association between antibody levels and disease status in our analyses (no odds ratios changed by more than 5%).

Among the autoantibody-tested subjects, controls had more years of education than cases (P < 0.0001), and for women, controls were more likely than cases to be white (versus nonwhite) (P = 0.05; Table 1). The race/ethnicity distribution of study participants reflected the overall distribution of the study recruitment sites. Women outnumbered men by a ratio of 1.8:1. Controls were more likely to report allergy and asthma, but not eczema (Table 1). Among controls, women and men had equivalent autoantibody levels, apart from NUMA1, which was higher among women (log NUMA1control women = 0.67 and log NUMA1control men = 0.62; Kruskal-Wallis test: P = 0.005 (see Web Table 1, available at Autoantibody levels were also associated with allergies among controls, and ENO1 levels were borderline-associated with current smoking among control women only (P = 0.06; Web Table 1). There were no statistically significant associations for asthma, eczema, ever smoking, race, or ethnicity within cases or controls, or for allergies among cases only (Web Table 1).

Table 1.
Demographic Characteristics and Prevalence of Immune System-related Conditions in the Study Sample, Overall and by Gender, Meningioma Consortium, 2006–2009

Autoantibody levels to TAA were higher in cases than in controls; however, results were not statistically significant (Tables 224). Statistical analyses indicated differences by gender, and we report odds ratios stratified by gender. Specific autoantibody levels were positively associated with meningioma status only in men (Table 4). Risk of meningioma increased with increasing anti-NKTR levels among men, whereas no association was observed among women (P for interaction = 0.08). Among men, the odds of meningioma were increased 2.7-fold for persons in the highest NKTR quartile compared with the lowest, and for every unit change in NKTR, the odds of meningioma increased 4.3-fold (Table 4). Assessment for potential confounders and statistical interaction showed no interaction with allergies, asthma, or smoking. Odds ratios did change slightly (>10%) with the addition of allergy to the model, and therefore the presented results were adjusted for allergy, age, and education among men.

Table 2.
Adjusted Odds Ratios for Development of Meningioma According to Serum Autoantibody Level (Continuous and in Quartiles of the Distribution in Controls), Meningioma Consortium Study, 2006–2009
Table 3.
Adjusted Odds Ratios for Development of Meningioma According to Serum Autoantibody Level (Continuous and in Quartiles of the Distribution in Controls) Among Women, Meningioma Consortium Study, 2006–2009
Table 4.
Adjusted Odds Ratios for Development of Meningioma According to Serum Autoantibody Level (Continuous and in Quartiles of the Distribution in Controls) Among Men, Meningioma Consortium Study, 2006–2009

Risk of meningioma increased with increasing anti-NUMA1 levels among men, whereas no association was observed among women (P for interaction = 0.03). Among men, the odds of meningioma were increased 3-fold for persons in the highest NUMA1 quartile compared with the lowest, and for every unit change in NUMA1, the odds of meningioma increased 3.6-fold. Assessment for potential confounders and statistical interaction showed an interaction between allergy history and NUMA1 level among men (P for interaction = 0.10), where men without allergies had 9-fold increased odds of meningioma for every unit increase in NUMA1 (odds ratio (OR) = 9.2, 95% confidence interval (CI): 1.5, 55) and those with allergies had no increase (OR = 1.1, 95% CI: 0.16, 7.0). We constructed ROC curves, which indicated no improvement in classification of meningioma patients as compared with controls for women (Figure 1A) and some improvement among men when using NKTR and NUMA1 (AUC values of 0.58 and 0.60, respectively).

Figure 1.
Receiver operating characteristic curves for autoantibodies to 3 meningioma antigens (enolase 1 (ENO1), NK-tumor recognition protein (NKTR), and nuclear mitotic apparatus protein 1 (NUMA1)) among A) women and B) men in the Meningioma Consortium, 2006–2009. ...

We did not discover any gender differences in allergy in our prior analysis of allergy and meningioma and did not report gender-specific odds ratios (4). Because of the current strong differences in autoantibody levels, we reanalyzed our data on allergy and meningioma by gender. The relation of meningioma with allergies was virtually the same for men and women (for men, OR = 0.60, 95% CI: 0.39, 0.91; for women, OR = 0.61, 95% CI: 0.47, 0.80), and there was no evidence of statistical interaction between allergies and gender.


Given previous data indicating that immune factors are involved in meningioma and the accessibility of autoantibody targets derived from the work of other investigators (4, 12, 15), we developed and performed a small panel of autoantibody Luminex assays in a population-based case-control study of meningioma. Our results indicate that recognition of immunogenic epitopes on meningioma cells is stronger in men than in women, to the point of being a phenomenon exclusive to men for two of the epitopes chosen here. Whether this phenomenon is limited to some TAA or is true for most TAA is unknown and will await further study. Note that this result could be a type 1 error due to smaller numbers of participants within subgroups. This risk of error is increased further with regard to the allergy interaction analysis for NUMA1 (Table 4, bottom), which should be interpreted with caution. Further research should be performed to clarify this result. If true, this phenomenon may affect gender-specific incidence rates. Men, having one-half the incidence rate of meningioma as women, may be able to suppress nascent meningiomas with an effective antitumor immune reaction. Whether their lower incidence of meningioma is caused partly or mostly by immune suppression is an interesting question that requires further study and suggests therapeutic or prophylactic interventions based on enhancing antimeningioma immune reactions, particularly for women who show no evidence of an antimeningioma reaction.

The strengths of our current study include the population-based and geographically diverse recruitment of study participants and the large sample size in comparison with previous studies that assessed autoantibodies in meningioma, which were limited to fewer than 50 patients (12, 15). Additional strengths include the equivalence of sample collection between cases and controls (for both, samples were collected in the field with laboratory personnel blinded to status to ensure identical processing), the lack of chemotherapy in the patient population (which might have affected antibody production), and the large number of immune covariates available for analysis. We developed a novel assay which included clean, expressed antigens and precise, relative quantification. Inherent weaknesses of this pilot study include the limited number of targets, although we intentionally selected the most biologically promising targets from the published literature for this project, and the small numbers for analysis of relevant subgroups, particularly men. The targets chosen for this study were derived from eukaryotic cultures to ensure that they matched the same expression system used in large autoantigen discovery platforms (12, 15). Future studies should include more autoantibody targets and if possible mammalian expression systems to allow for relevant posttranslational modifications that are not possible in bacterial expression systems. Of course, such studies will need to include new discovery screening if a different antigen expression system is employed.

Prior investigations of autoantibodies to meningioma antigens did not report results of gender-stratified analyses; therefore, our gender-specific results are without precedence. Higher rates of meningioma in women are thought to result from hormonal factors, as gender differences peak at 3:1 during the late reproductive years (ages 35–44 years (1)), and there are some reported risk associations for exogenous hormone use and meningioma incidence. Anthropometric variables that correlate with endogenous hormone levels have also been linked to meningioma risk (reviewed by Wiemels et al. (3) and Cowppli-Bony et al. (17)). Our gender-specific results may be counterintuitive when one considers that hormonal factors are known to contribute to autoantibody-related diseases such as lupus erythematosus, with estrogens enhancing T helper 2-cell development and antibody affinity maturation and androgens suppressing immune reactions (18, 19). Women have higher baseline immunoglobulin levels and produce more immunoglobulin in response to infection and immunization (20, 21). Among controls, women had statistically significantly higher autoantibody levels in the current study only for NUMA1 (Web Table 1). Women did not exhibit any additional autoantibody reaction when harboring a meningioma, and the stimulation of such a reaction may represent a potential therapeutic modality in the future. Men had higher immunoglobulin E levels than did women in our study population (4); however, immunoglobulin E is a small proportion of total immunoglobulin. Our current study measured specific immunoglobulin G (immunoglobulin G being the predominant antibody subtype), and the immunoglobulin E result is likely irrelevant here.

Prior reports of meningioma autoantibodies that were based on analyses of men and women combined or unreported gender found stronger associations than we found here (12, 15). Apart from ENO1, our results were consistent with prior reports, though the differences between cases and controls were smaller in magnitude. In addition, ROC curve analysis indicated that the autoantigens assessed here performed slightly less well than in the previous study in which ROCs were computed (15). Still, the ROC curves calculated by us discriminate cases from controls at about the same level that the Gail model is predictive of breast cancer (22). The reasons for our lower AUC values compared with those of Ludwig et al. (15) and the lack of discrimination by ENO1 level may be related to assay differences. Ludwig et al. used Escherichia coli-expressed antigens (15); the antigens used in our study were produced in a eukaryotic expression system based on wheat. Protein folding differences and posttranslational modifications may account for differences in epitope availability. The exact epitopes that are immunogenic within these proteins are not known.

Many factors determine whether an individual will produce autoantibodies to specific tumor antigens and whether levels of such autoantibodies will have any relation to the levels of expression of particular antigens. Antibody specificity is governed in part by the interplay between protein expression in the tumor and an individual's capacity to present particular antigens to the immune system based on the highly polymorphic peptide binding pockets of the major histocompatibility proteins. As a meningioma grows and evolves, it may alter expression of immunogenic antigens that elicit an antitumor response, probably altering the autoantibody repertoire over time. Because of these variables, an ideal autoantibody assay would encompass enough antigens to yield consistent and correct answers both across a population of persons with different tumors and human leukocyte antigen genotypes and within an individual whose autoantibodies may be altered over time. Typically, a single serum sample will be available for any given individual, necessitating a robust multipoint assay. While the current study represents an important start, future studies will need to address a much larger antigen repertoire including possibly posttranslational modifications such as carbohydrate moieties, as well as a robust analytical rubric. Such studies would also aid in understanding the gender differences in autoantibody response and whether such differences may help account for gender differences in meningioma incidence.

Supplementary Material

Web Table:


Author affiliations: Department of Epidemiology and Biostatistics, School of Medicine, University of California, San Francisco, San Francisco, California (Joseph Wiemels, Paige M. Bracci, Jon C. Pfefferle, Mi Zhou); Department of Neurosurgery, School of Medicine, University of California, San Francisco, San Francisco, California (Margaret Wrensch, Jennette D. Sison); Department of Community and Family Medicine, School of Medicine, Duke University, Durham, North Carolina (Joellen Schildkraut); Department of Pediatrics, Baylor College of Medicine, Houston, Texas (Melissa Bondy); Department of Epidemiology and Public Health, School of Medicine, Yale University, New Haven, Connecticut (Lisa Calvocoressi); and Department of Neurosurgery, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts (Elizabeth B. Claus).

This work was supported by National Institutes of Health R01 grants CA109468, CA109461, CA109745, CA108473, CA109475, and CA151933, the Brain Science Foundation, and the Meningioma Mommas. The authors thank Sydnee Crankshaw, Katherine Wagenman, Joe Patoka, and Lisa Padilla for sample management and Koren Jones, Katherine Saunders, and Estella Kanevsky for project management expertise. They also thank the study physicians.

Conflict of interest: none declared.


1. Central Brain Tumor Registry of the United States. 2009–2010 CDTRUS Statistical Report: Primary Brain and Central Nervous System Tumors Diagnosed in Eighteen States in 2002–2006. Hisdale, IL: Central Brain Tumor Registry of the United States; 2009.
2. Claus EB, Bondy ML, Schildkraut JM, et al. Epidemiology of intracranial meningioma. Neurosurgery. 2005;57(6):1088–1095. [PubMed]
3. Wiemels J, Wrensch M, Claus EB. Epidemiology and etiology of meningioma. J Neurooncol. 2010;99(3):307–314. [PMC free article] [PubMed]
4. Wiemels JL, Wrensch M, Sison JD, et al. Reduced allergy and immunoglobulin E among adults with intracranial meningioma compared to controls. Int J Cancer. 2011;129(8):1932–1939. [PMC free article] [PubMed]
5. Schoemaker MJ, Swerdlow AJ, Hepworth SJ, et al. History of allergic disease and risk of meningioma. Am J Epidemiol. 2007;165(5):477–485. [PubMed]
6. Brenner AV, Linet MS, Fine HA, et al. History of allergies and autoimmune diseases and risk of brain tumors in adults. Int J Cancer. 2002;99(2):252–259. [PubMed]
7. Ryan P, Lee MW, North B, et al. Risk factors for tumors of the brain and meninges: results from the Adelaide Adult Brain Tumor Study. Int J Cancer. 1992;51(1):20–27. [PubMed]
8. Schlehofer B, Blettner M, Preston-Martin S, et al. Role of medical history in brain tumour development. Results from the International Adult Brain Tumour Study. Int J Cancer. 1999;82(2):155–160. [PubMed]
9. Schwartzbaum J, Jonsson F, Ahlbom A, et al. Cohort studies of association between self-reported allergic conditions, immune-related diagnoses and glioma and meningioma risk. Int J Cancer. 2003;106(3):423–428. [PubMed]
10. Wigertz A, Lonn S, Schwartzbaum J, et al. Allergic conditions and brain tumor risk. Am J Epidemiol. 2007;166(8):941–950. [PubMed]
11. Kobold S, Lutkens T, Cao Y, et al. Autoantibodies against tumor-related antigens: incidence and biologic significance. Hum Immunol. 2010;71(7):643–651. [PubMed]
12. Comtesse N, Zippel A, Walle S, et al. Complex humoral immune response against a benign tumor: frequent antibody response against specific antigens as diagnostic targets. Proc Natl Acad Sci USA. 2005;102(27):9601–9606. [PubMed]
13. Keller A, Ludwig N, Comtesse N, et al. Combining gene expression signatures and autoantibody profiles in human meningioma. Gene Ther. 2009;16(2):184–189. [PubMed]
14. Keller A, Ludwig N, Comtesse N, et al. A minimally invasive multiple marker approach allows highly efficient detection of meningioma tumors. BMC Bioinformatics. 2006;7:539. ( doi:10.1186/1471-2105-7-539) [PMC free article] [PubMed]
15. Ludwig N, Keller A, Heisel S, et al. Novel immunogenic antigens increase classification accuracy in meningioma to 93.84% Int J Cancer. 2011;128(6):1493–1501. [PubMed]
16. Khan IH, Kendall LV, Ziman M, et al. Simultaneous serodetection of 10 highly prevalent mouse infectious pathogens in a single reaction by multiplex analysis. Clin Diagn Lab Immunol. 2005;12(4):513–519. [PMC free article] [PubMed]
17. Cowppli-Bony A, Bouvier G, Rue M, et al. Brain tumors and hormonal factors: review of the epidemiological literature. Cancer Causes Control. 2011;22(5):697–714. [PubMed]
18. McClelland EE, Smith JM. Gender specific differences in the immune response to infection. Arch Immunol Ther Exp (Warsz) 2011;59(3):203–213. [PubMed]
19. Tanriverdi F, Silveira LF, MacColl GS, et al. The hypothalamic-pituitary-gonadal axis: immune function and autoimmunity. J Endocrinol. 2003;176(3):293–304. [PubMed]
20. Shames RS. Gender differences in the development and function of the immune system. J Adolesc Health. 2002;30(4 suppl):59–70. [PubMed]
21. Cutolo M, Sulli A, Seriolo B, et al. Estrogens, the immune response and autoimmunity. Clin Exp Rheumatol. 1995;13(2):217–226. [PubMed]
22. Gail MH, Brinton LA, Byar DP, et al. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst. 1989;81(24):1879–1886. [PubMed]

Articles from American Journal of Epidemiology are provided here courtesy of Oxford University Press