Many cancers, including ovarian, overexpress epithelial mucin (MUC1) and promote anti-MUC1 antibodies that may correlate with more favorable prognosis. By extension, risk for ovarian cancer might be reduced by preexisting MUC1-specific immunity. We measured anti-MUC1 antibodies in 705 control women, identified events predicting antibodies, and estimated ovarian cancer risk by comparing profiles of events generating antibodies in controls with those in 668 ovarian cancer cases. Factors predicting antibodies included oral contraceptive use, breast mastitis, bone fracture or osteoporosis, pelvic surgeries, nonuse of talc in genital hygiene, and to a lesser extent intrauterine device use and current smoking. There was a significant increase in the likelihood of having anti-MUC1 antibodies from 24.2% in women with 0 or 1 condition, to 51.4% in those with five or more conditions. By the same index of events, the risk for ovarian cancer was inversely associated with number of conditions predisposing to anti-MUC1 antibodies. Compared with having experienced 0 or 1 event, the adjusted risk for ovarian cancer decreased progressively with relative risks (and 95% confidence limits) of 0.69 (0.52-0.92), 0.64 (0.47-0.88), 0.49 (0.34-0.72), and 0.31 (0.16-0.61), respectively for women with two, three, four, and five or more events related to the presence of antibodies (Ptrend < 0.0001.) We conclude that several traditional and new risk factors for ovarian cancer may be explained by their ability to induce MUC1 immunity through exposure of MUC1 to immune recognition in the context of inflammatory or hormonal processes in various MUC1-positive tissues.

Human mucin (MUC) family member, MUC1, is a high molecular weight protein expressed in a highly glycosylated form and low levels by many types of normal epithelial cells and in a hypoglycosylated form and high levels by most epithelial adenocarcinomas, including breast and ovarian cancer (1). Anti-MUC1 antibodies have been described and correlated with a more favorable prognosis (2-5) showing that patients generate immunity against MUC1 produced by their tumors and defining MUC1 as a tumor-associated antigen and candidate for cancer vaccines (6). Anti-MUC1 antibodies are also found in healthy individuals, especially in women during pregnancy and lactation. It has been hypothesized that a natural immunity against tumor MUC1 might develop and account for the long-term protective effect of pregnancy or breast-feeding on breast cancer risk (7), an elaboration on the so called β€œfetal antigen theory” (8). Indeed it has been shown that sera from multiparous women, but not from nulliparous women or from men, are able to mediate killing of breast cancer cells (9). Supporting a key role for MUC1 in these reactions, core peptide sequences from MUC1 can induce proliferation of T cells and cytotoxic T-cell responses in multiparous women (10). Recently, the β€œfetal antigen” hypothesis was extended to ovarian cancer after it was shown that sera from multiparous women also reacted with multiple antigens from ovarian cancer cells more strongly than sera from nulliparous women or men (11), although MUC1 was not specifically examined in these experiments.

In this study, we used an ELISA to determine the presence and relative amounts of MUC1-specific antibody in women from the general population who served as controls in a study of ovarian cancer. In analyses confined to these controls, we identified the predictors of anti-MUC1 antibodies and used case-control comparisons to evaluate these predictors in relation to ovarian cancer risk. We hypothesized that events predicting anti-MUC1 antibodies would be inversely associated with ovarian cancer risk and that there would be a cumulative effect of such events.

Subject Recruitment and Study

This report is based on the second phase of a population-based case-control study of ovarian cancer conducted between 7/98 and 7/03 and involving eastern Massachusetts and all of New Hampshire, approved by the Brigham and Women's Hospital and Dartmouth Medical Center's Institutional Review Boards. We identified 1,267 cases from tumor boards and Statewide Registries and excluded 119 cases who died, 110 who moved from the study area, one who had no telephone, 23 who did not speak English, and 46 found to have a nonovarian primary upon review. Of the remaining 968, physicians denied permission to contact 106 and 171 declined to participate, leaving 691 cases interviewed. Of these, 668 had an epithelial ovarian cancer (including borderline malignancies) and are included in this report. A small number of cases (n = 48) were enrolled before surgery.

Controls were identified through town books in Massachusetts and Drivers' License lists in New Hampshire and sampled to match the age and residence of previously accumulated cases. Invitations to participate were sent to 1843 potential controls. Of these 318 had moved and could not be located or had died, 197 (in Massachusetts) could not be recontacted because subjects returned an β€œopt out” postcard required by the hospital's Institutional Review Boards, and 47 no longer had a working telephone. Of the remaining 1,281 who were contacted, 152 were ineligible because they had no ovaries or were not the correct age, 59 were incapacitated or did not speak English, and 349 declined, leaving 721 who were interviewed and included in this report.

After written informed consent, an in-person interview dealing with demographic, medical, and family history was conducted. Subjects also completed a self-administered dietary questionnaire. Heparinized blood specimens were collected from subjects agreeing to provide one; separated into red cell, buffy coat, and plasma components; and stored at βˆ’80Β°C.

ELISA Assay for Anti-MUC1 Antibodies

Plasma specimens were available for measuring anti-MUC1 antibodies in 48 cases with preoperative bloods and 705 controls. Antibodies were measured against a synthetic 100-mer MUC1 peptide corresponding to five tandem repeats of the MUC1 polypeptide core tandem repeat region, according to our previously published protocol (2). Briefly, 0.5 ΞΌg of MUC1 peptide in 100 ΞΌL of PBS was added to each well of Immulon 4 plates (Dynax, Chantilly, VA) and incubated overnight at 4Β°C. Control plates without the MUC1 peptide were also prepared. The plates were washed thrice with PBS and 100 ΞΌL of 2.5% bovine serum albumin in PBS added for 1 hour at room temperature to coat remaining sites in the well (blocking step). Fifty microliters of serially diluted plasma (1:20 to 1:80 in PBS) were added to the MUC1 peptide-coated and control plates and incubated for 1 hour at room temperature. The plates were washed 5Γ— with 100 ΞΌL PBS and 0.1% Tween 20 detergent. Fifty microliters of secondary antibody, alkaline phosphatase–labeled goat anti-human polyvalent IgM, IgG, IgA (Sigma-Aldrich, St. Louis, MO), diluted 1:1,000, was added for 1 hour at room temperature, and plates again washed 5Γ— with PBS-Tween. One hundred microliters of alkaline phosphatase substrate pNPP (Sigma-Aldrich) were added at 3 mg/mL in 0.05 mol/L NaCO3 and 0.5 mmol/L MgCl2 and the plates incubated in the dark for exactly 1 hour before adding 50 ΞΌL of the stop solution (0.5 mol/L NaOH). The absorbance at 405 to 410 nm was measured using the plate reader MRX Revelation (Thermo Labsystems, Chantilly, VA). Absorbance values for each sample in the MUC1-coated plate were compared with values in the antigen-negative plates to subtract nonspecific binding. Based upon the previous responses in over 500 cancer cases and controls, absorbance reactions at the 1:20 dilution at <0.6 are scored as negative, reactions in the 0.6 to 0.79 range as low, reactions in the 0.8 to 0.99 range as intermediate, and reactions β‰₯1.0 as high. In the current study, 20 blinded replicate specimens were included and the Spearman correlation coefficient between the paired absorbances was 0.93 (P < 0.0001).

Statistical Methods

Logistic regression analysis was used to compare those with an antibody reaction at any level against those considered negative for MUC1 antibody (A < 0.6), while adjusting for potential confounding variables. Spearman rank correlations or generalized linear modeling was used to assess differences in absorbance levels (using log-transformed values of absorbance) for a more quantitative assessment of factors affecting anti-MUC1 antibody production. Combinations of factors were examined to identify the best cumulative index of experiences associated with likelihood of having antibodies. Ovarian cancer cases and controls were then categorized by the presence or absence of events found to affect the likelihood of antibodies and risk for ovarian cancer calculated using unconditional multivariate logistic regression to adjust for potential confounders. In our models, we adjusted for the matching variables, age (continuous), and study site (Massachusetts, New Hampshire), as well as ethnicity (White, non-White), religious background (Jewish, non-Jewish), and parity as a continuous variable except where noted in the text or footnotes.

The distributions of absorbance readings (corresponding to the amount of anti-MUC1 antibodies measured in the ELISA assay) seemed bimodal in cases with preoperative bloods and skewed right in controls prompting log transformation for statistical testing (Fig. 1). By a cutoff of A β‰₯ 0.6, 33.8% of controls and 45.8% of cases were positive for antibodies. By a cutoff of A β‰₯1.0, 12.3% of controls and 25% of cases had a high level of antibodies, a significant difference that likely reflects an ongoing immune response to tumor in the cases.

Figure 1.

Distribution of absorbances from anti-MUC1 antibody assay in cases with preoperative bloods and all controls.

Figure 1.

Distribution of absorbances from anti-MUC1 antibody assay in cases with preoperative bloods and all controls.

Close modal

Events Predicting Occurrence of Anti-MUC1 Antibodies

A number of demographic, reproductive, and medical conditions were examined as they affected the likelihood of controls having a low, intermediate, high level, or any anti-MUC1 antibody (Table 1). The last two columns show the (geometric) mean absorbance value, its SE, and the P from the linear regression model. Age was a strong predictor with 50% having antibodies at ages <35, declining to 29.3% at ages 55 to 64 years, and increasing back up to 32.6% in those ages β‰₯65 years, prompting age adjustment when testing for the significance of further variables. The proportion of women who were positive for anti-MUC1 antibody was similar for women who had never been pregnant (33.3%), had at least one live birth (34.1%), or had breast-fed without experiencing a mastitis (33.0%) but elevated for women who had experienced mastitis while breast-feeding (46.1%). Notably, 25.0% of women reporting mastitis had high antibody levels compared with 10% to 14% of parous women who either never breast-fed or breast-fed and reported no mastitis (P = 0.05). Women who had used oral contraceptives (OC), compared with those who had not, were more likely to have antibodies; and this was most apparent among premenopausal women in whom 40.7% of OC users had antibodies compared with 26.7% of nonusers (P = 0.05). The proportion of women with antibodies was also higher for those who reported a bone fracture or osteoporosis after age 40 or within 20 years of their age at interview (36.0%) than in those who had not (33.0%) and 17.1% of women with fracture/osteoporosis had high antibody levels compared with 10.8% of women who had not (P = 0.03). Several types of pelvic/gynecologic surgery, including tubal sterilization, cervical conization, and cesarean section increased the likelihood of a positive antibody reaction and 47.2% of women who had more than one surgery had antibodies compared with 30.9% of women who never had pelvic surgery (P = 0.01). A surprising finding was that 38.1% of women who reported no use of cosmetic talc in hygiene had antibody compared with 28.6% of women who regularly used talc in genital hygiene (P = 0.04). The final entry shows the trend for elevated anti-MUC1 antibody levels by increasing number of antibody-promoting conditions. These included all variables significant in univariate analyses, such as OC use, bone fracture, mastitis, pelvic surgery, and genital talc use (where no use was considered the β€œcondition”) as well as variables of marginal significance in the univariate analysis, which nevertheless improved the overall model including current smoking and use of an intrauterine device (IUD). A significant trend (P = 0.0005) in the likelihood of having antibodies was observed such that 24.2% of women who had zero or one of condition had antibodies compared with 51.4% of women who had experienced five or more of these conditions.

Table 1.

Occurrence of anti-MUC1 antibodies in control women by epidemiologic variables

NegativePositive*
Mean absorbance (SE)Age-adjusted P
LowIntermediateHighAny
AllΒ 467 (66.2)Β 94 (13.3)Β 57 (8.1)Β 87 (12.3)Β 238 (33.8)Β 0.44 (0.08)Β Β 
Age (y)Β Β Β Β Β Β Β Β 
    <35Β 33 (50.0)Β 12 (18.2)Β 8 (12.1)Β 13 (19.7)Β 33 (50.0)Β 0.57 (0.24)Β 0.0009Β 
    35-44Β 90 (65.7)Β 17 (12.4)Β 18 (13.1)Β 12 (8.8)Β 47 (34.3)Β 0.45 (0.18)Β Β 
    45-54Β 133 (67.5)Β 25 (12.7)Β 13 (6.6)Β 26 (13.2)Β 64 (32.5)Β 0.48 (0.13)Β Β 
    55-64Β 118 (70.7)Β 24 (14.4)Β 11 (6.6)Β 14 (8.4)Β 49 (29.3)Β 0.37 (0.19)Β Β 
    β‰₯65Β 93 (67.4)Β 16 (11.6)Β 7 (5.1)† 22 (15.9)Β 45 (32.6)† 0.41 (0.23)Β Β 
RaceΒ Β Β Β Β Β Β Β 
    WhiteΒ 454 (66.5)Β 92 (13.5)Β 55 (8.0)Β 82 (12.0)Β 229 (33.5)Β 0.44 (0.08)Β 0.07Β 
    Non-WhiteΒ 13 (59.1)Β 2 (9.1)Β 2 (9.1)Β 5 (22.7)Β 9 (40.9)Β 0.62 (0.37)Β Β 
ReligionΒ Β Β Β Β Β Β Β 
    Non-JewishΒ 448 (66.3)Β 89 (13.2)Β 55 (8.1)Β 84 (12.4)Β 228 (33.7)Β 0.44 (0.09)Β 0.68Β 
    JewishΒ 19 (65.5)Β 5 (17.2)Β 2 (6.9)Β 3 (10.3)Β 10 (34.5)Β 0.47 (0.35)Β Β 
Marital statusΒ Β Β Β Β Β Β Β 
    Never marriedΒ 40 (66.7)Β 6 (10.0)Β 6 (10.0)Β 8 (13.3)Β 20 (33.3)Β 0.51 (0.22)Β 0.49Β 
    Ever marriedΒ 427 (66.2)Β 88 (13.6)Β 51 (7.9)Β 79 (12.2)Β 218 (33.8)Β 0.44 (0.09)Β Β 
Pregnancy historyΒ Β Β Β Β Β Β Β 
    Never pregnantΒ 62 (66.7)Β 11 (11.8)Β 8 (8.6)Β 12 (12.9)Β 31 (33.3)Β 0.51 (0.19)Β 0.24Β 
    Pregnant but no live birthsΒ 20 (71.4)Β 4 (14.3)Β 3 (10.7)Β 1 (3.6)Β 8 (28.6)Β 0.45 (0.30)Β Β 
    At least one live birthΒ 385 (65.9)Β 79 (13.5)Β 46 (7.9)Β 74 (12.7)Β 199 (34.1)Β 0.43 (0.10)Β Β 
Breast-feeding (among parous women)Β Β Β Β Β Β Β Β 
    Never breast-fedΒ 159 (66.0)Β 33 (13.7)Β 16 (6.6)Β 33 (13.7)Β 82 (34.0)Β 0.43 (0.15)Β 0.95Β 
    Breast-fed and no mastitisΒ 211 (67.0)Β 41 (13.0)Β 29 (9.2)Β 34 (10.8)Β 104 (33.0)Β 0.42 (0.13)Β Β 
    Breast-fed and mastitisΒ 15 (53.6)Β 5 (17.9)Β 1 (3.6)Β 7 (25.0)‑ 13 (46.4)Β 0.55 (0.38)Β Β 
OC useΒ Β Β Β Β Β Β Β 
    NoΒ 170 (70.8)Β 25 (10.4)Β 19 (7.9)Β 26 (10.8)Β 70 (29.2)Β 0.41 (0.15)Β 0.42Β 
    YesΒ 297 (63.9)Β 69 (14.8)Β 38 (8.2)Β 61 (13.1)Β 168 (36.1)Β 0.46 (0.10)Β Β 
OC use in premenopausal subjectsΒ Β Β Β Β Β Β Β 
    NoΒ 39 (73.6)Β 4 (7.6)Β 8 (15.1)Β 2 (3.8)Β 14 (26.4)Β 0.45 (0.24)Β 0.30Β 
    YesΒ 147 (59.3)Β 43 (17.3)‑ 23 (9.3)Β 35 (14.1)† 101 (40.7)† 0.49 (0.13)Β Β 
IUD useΒ Β Β Β Β Β Β Β 
    NoΒ 379 (66.4)Β 70 (12.3)Β 51 (8.9)Β 71 (12.4)Β 192 (33.6)Β 0.44 (0.09)Β 0.60Β 
    YesΒ 88 (65.7)Β 24 (17.9)‑ 6 (4.5)Β 16 (11.9)Β 46 (34.3)Β 0.45 (0.18)Β Β 
Bone fracture/osteoporosisΒ Β Β Β Β Β Β Β 
    NoΒ 355 (67.0)Β 68 (12.8)Β 50 (9.4)Β 57 (10.8)Β 175 (33.0)Β 0.43 (0.10)Β 0.16Β 
    YesΒ 112 (64.0)Β 26 (14.9)Β 7 (4.0)‑ 30 (17.1)† 63 (36.0)Β 0.46 (0.16)Β Β 
ColitisΒ Β Β Β Β Β Β Β 
    NoΒ 440 (65.9)Β 93 (13.9)Β 54 (8.1)Β 81 (12.1)Β 228 (34.1)Β 0.44 (0.08)Β 0.97Β 
    YesΒ 27 (73.0)Β 1 (2.7)‑ 3 (8.1)Β 6 (16.2)Β 10 (27.0)Β 0.44 (0.39)Β Β 
EndometriosisΒ Β Β Β Β Β Β Β 
    NoΒ 430 (66.2)Β 86 (13.2)Β 53 (8.2)Β 81 (12.5)Β 220 (33.8)Β 0.45 (0.09)Β 0.26Β 
    YesΒ 37 (67.3)Β 8 (14.6)Β 4 (7.3)Β 6 (10.9)Β 18 (32.7)Β 0.39 (0.32)Β Β 
Pelvic surgeryΒ Β Β Β Β Β Β Β 
    No pelvic surgeryΒ 304 (69.1)Β 52 (11.8)Β 33 (7.5)Β 51 (11.6)Β 136 (30.9)Β 0.43 (0.10)Β 0.12Β 
    Hysterectomy onlyΒ 26 (70.3)Β 4 (10.8)Β 1 (2.7)Β 6 (16.2)Β 11 (29.7)Β 0.39 (0.55)Β Β 
    Tubal sterilization onlyΒ 54 (64.3)Β 15 (17.9)Β 6 (7.1)Β 9 (10.7)Β 30 (35.7)Β 0.44 (0.23)Β Β 
    Conization onlyΒ 14 (58.3)Β 3 (12.5)Β 1 (4.2)Β 6 (25.0)Β 10 (41.7)Β 0.62 (0.34)Β Β 
    Cesarean section onlyΒ 31 (64.6)Β 5 (10.4)Β 6 (12.5)Β 6 (12.5)Β 17 (35.4)Β 0.47 (0.29)Β Β 
    >1Β 38 (52.8)Β 15 (20.8)† 10 (13.9)† 9 (12.5)Β 34 (47.2)† 0.48 (0.27)Β Β 
SmokingΒ Β Β Β Β Β Β Β 
    NeverΒ 228 (66.5)Β 43 (12.5)Β 30 (8.8)Β 42 (12.2)Β 115 (33.5)Β 0.45 (0.12)Β 0.86Β 
    FormerΒ 182 (67.7)Β 36 (13.4)Β 20 (7.4)Β 31 (11.5)Β 87 (32.3)Β 0.43 (0.14)Β Β 
    CurrentΒ 57 (61.3)Β 15 (16.1)Β 7 (7.5)Β 14 (15.0)Β 36 (38.7)Β 0.45 (0.25)Β Β 
Talc useΒ Β Β Β Β Β Β Β 
    NoneΒ 208 (61.9)Β 43 (12.8)Β 41 (12.2)Β 44 (13.1)Β 128 (38.1)Β 0.46 (0.12)Β 0.08Β 
    Body use onlyΒ 117 (68.8)Β 23 (13.5)Β 9 (5.3)Β 21 (12.4)Β 53 (31.2)Β 0.46 (0.15)Β Β 
    Genital useΒ 142 (71.4)Β 28 (14.1)Β 7 (3.5)† 22 (11.1)Β 57 (28.6)† 0.39 (0.17)Β Β 
No. conditions§        
    0 or 1Β 119 (75.8)Β 15 (9.6)Β 10 (6.4)Β 13 (8.3)Β 38 (24.2)Β 0.37 (0.20)Β 0.003Β 
    2Β 155 (67.1)Β 26 (11.3)Β 20 (8.7)Β 30 (13.0)Β 76 (32.9)Β 0.44 (0.14)Β Β 
    3Β 114 (65.9)Β 25 (14.4)Β 17 (9.8)Β 17 (9.8)Β 59 (34.1)Β 0.45 (0.15)Β Β 
    4Β 61 (57.0)Β 21 (19.6)Β 6 (5.6)Β 19 (17.8)Β 46 (43.0)Β 0.51 (0.20)Β Β 
    5 or moreΒ 18 (48.6)Β 7 (18.9)† 4 (10.8)Β 8 (21.6)† 19 (51.4)βˆ₯Β 0.53 (0.41)Β Β 
NegativePositive*
Mean absorbance (SE)Age-adjusted P
LowIntermediateHighAny
AllΒ 467 (66.2)Β 94 (13.3)Β 57 (8.1)Β 87 (12.3)Β 238 (33.8)Β 0.44 (0.08)Β Β 
Age (y)Β Β Β Β Β Β Β Β 
    <35Β 33 (50.0)Β 12 (18.2)Β 8 (12.1)Β 13 (19.7)Β 33 (50.0)Β 0.57 (0.24)Β 0.0009Β 
    35-44Β 90 (65.7)Β 17 (12.4)Β 18 (13.1)Β 12 (8.8)Β 47 (34.3)Β 0.45 (0.18)Β Β 
    45-54Β 133 (67.5)Β 25 (12.7)Β 13 (6.6)Β 26 (13.2)Β 64 (32.5)Β 0.48 (0.13)Β Β 
    55-64Β 118 (70.7)Β 24 (14.4)Β 11 (6.6)Β 14 (8.4)Β 49 (29.3)Β 0.37 (0.19)Β Β 
    β‰₯65Β 93 (67.4)Β 16 (11.6)Β 7 (5.1)† 22 (15.9)Β 45 (32.6)† 0.41 (0.23)Β Β 
RaceΒ Β Β Β Β Β Β Β 
    WhiteΒ 454 (66.5)Β 92 (13.5)Β 55 (8.0)Β 82 (12.0)Β 229 (33.5)Β 0.44 (0.08)Β 0.07Β 
    Non-WhiteΒ 13 (59.1)Β 2 (9.1)Β 2 (9.1)Β 5 (22.7)Β 9 (40.9)Β 0.62 (0.37)Β Β 
ReligionΒ Β Β Β Β Β Β Β 
    Non-JewishΒ 448 (66.3)Β 89 (13.2)Β 55 (8.1)Β 84 (12.4)Β 228 (33.7)Β 0.44 (0.09)Β 0.68Β 
    JewishΒ 19 (65.5)Β 5 (17.2)Β 2 (6.9)Β 3 (10.3)Β 10 (34.5)Β 0.47 (0.35)Β Β 
Marital statusΒ Β Β Β Β Β Β Β 
    Never marriedΒ 40 (66.7)Β 6 (10.0)Β 6 (10.0)Β 8 (13.3)Β 20 (33.3)Β 0.51 (0.22)Β 0.49Β 
    Ever marriedΒ 427 (66.2)Β 88 (13.6)Β 51 (7.9)Β 79 (12.2)Β 218 (33.8)Β 0.44 (0.09)Β Β 
Pregnancy historyΒ Β Β Β Β Β Β Β 
    Never pregnantΒ 62 (66.7)Β 11 (11.8)Β 8 (8.6)Β 12 (12.9)Β 31 (33.3)Β 0.51 (0.19)Β 0.24Β 
    Pregnant but no live birthsΒ 20 (71.4)Β 4 (14.3)Β 3 (10.7)Β 1 (3.6)Β 8 (28.6)Β 0.45 (0.30)Β Β 
    At least one live birthΒ 385 (65.9)Β 79 (13.5)Β 46 (7.9)Β 74 (12.7)Β 199 (34.1)Β 0.43 (0.10)Β Β 
Breast-feeding (among parous women)Β Β Β Β Β Β Β Β 
    Never breast-fedΒ 159 (66.0)Β 33 (13.7)Β 16 (6.6)Β 33 (13.7)Β 82 (34.0)Β 0.43 (0.15)Β 0.95Β 
    Breast-fed and no mastitisΒ 211 (67.0)Β 41 (13.0)Β 29 (9.2)Β 34 (10.8)Β 104 (33.0)Β 0.42 (0.13)Β Β 
    Breast-fed and mastitisΒ 15 (53.6)Β 5 (17.9)Β 1 (3.6)Β 7 (25.0)‑ 13 (46.4)Β 0.55 (0.38)Β Β 
OC useΒ Β Β Β Β Β Β Β 
    NoΒ 170 (70.8)Β 25 (10.4)Β 19 (7.9)Β 26 (10.8)Β 70 (29.2)Β 0.41 (0.15)Β 0.42Β 
    YesΒ 297 (63.9)Β 69 (14.8)Β 38 (8.2)Β 61 (13.1)Β 168 (36.1)Β 0.46 (0.10)Β Β 
OC use in premenopausal subjectsΒ Β Β Β Β Β Β Β 
    NoΒ 39 (73.6)Β 4 (7.6)Β 8 (15.1)Β 2 (3.8)Β 14 (26.4)Β 0.45 (0.24)Β 0.30Β 
    YesΒ 147 (59.3)Β 43 (17.3)‑ 23 (9.3)Β 35 (14.1)† 101 (40.7)† 0.49 (0.13)Β Β 
IUD useΒ Β Β Β Β Β Β Β 
    NoΒ 379 (66.4)Β 70 (12.3)Β 51 (8.9)Β 71 (12.4)Β 192 (33.6)Β 0.44 (0.09)Β 0.60Β 
    YesΒ 88 (65.7)Β 24 (17.9)‑ 6 (4.5)Β 16 (11.9)Β 46 (34.3)Β 0.45 (0.18)Β Β 
Bone fracture/osteoporosisΒ Β Β Β Β Β Β Β 
    NoΒ 355 (67.0)Β 68 (12.8)Β 50 (9.4)Β 57 (10.8)Β 175 (33.0)Β 0.43 (0.10)Β 0.16Β 
    YesΒ 112 (64.0)Β 26 (14.9)Β 7 (4.0)‑ 30 (17.1)† 63 (36.0)Β 0.46 (0.16)Β Β 
ColitisΒ Β Β Β Β Β Β Β 
    NoΒ 440 (65.9)Β 93 (13.9)Β 54 (8.1)Β 81 (12.1)Β 228 (34.1)Β 0.44 (0.08)Β 0.97Β 
    YesΒ 27 (73.0)Β 1 (2.7)‑ 3 (8.1)Β 6 (16.2)Β 10 (27.0)Β 0.44 (0.39)Β Β 
EndometriosisΒ Β Β Β Β Β Β Β 
    NoΒ 430 (66.2)Β 86 (13.2)Β 53 (8.2)Β 81 (12.5)Β 220 (33.8)Β 0.45 (0.09)Β 0.26Β 
    YesΒ 37 (67.3)Β 8 (14.6)Β 4 (7.3)Β 6 (10.9)Β 18 (32.7)Β 0.39 (0.32)Β Β 
Pelvic surgeryΒ Β Β Β Β Β Β Β 
    No pelvic surgeryΒ 304 (69.1)Β 52 (11.8)Β 33 (7.5)Β 51 (11.6)Β 136 (30.9)Β 0.43 (0.10)Β 0.12Β 
    Hysterectomy onlyΒ 26 (70.3)Β 4 (10.8)Β 1 (2.7)Β 6 (16.2)Β 11 (29.7)Β 0.39 (0.55)Β Β 
    Tubal sterilization onlyΒ 54 (64.3)Β 15 (17.9)Β 6 (7.1)Β 9 (10.7)Β 30 (35.7)Β 0.44 (0.23)Β Β 
    Conization onlyΒ 14 (58.3)Β 3 (12.5)Β 1 (4.2)Β 6 (25.0)Β 10 (41.7)Β 0.62 (0.34)Β Β 
    Cesarean section onlyΒ 31 (64.6)Β 5 (10.4)Β 6 (12.5)Β 6 (12.5)Β 17 (35.4)Β 0.47 (0.29)Β Β 
    >1Β 38 (52.8)Β 15 (20.8)† 10 (13.9)† 9 (12.5)Β 34 (47.2)† 0.48 (0.27)Β Β 
SmokingΒ Β Β Β Β Β Β Β 
    NeverΒ 228 (66.5)Β 43 (12.5)Β 30 (8.8)Β 42 (12.2)Β 115 (33.5)Β 0.45 (0.12)Β 0.86Β 
    FormerΒ 182 (67.7)Β 36 (13.4)Β 20 (7.4)Β 31 (11.5)Β 87 (32.3)Β 0.43 (0.14)Β Β 
    CurrentΒ 57 (61.3)Β 15 (16.1)Β 7 (7.5)Β 14 (15.0)Β 36 (38.7)Β 0.45 (0.25)Β Β 
Talc useΒ Β Β Β Β Β Β Β 
    NoneΒ 208 (61.9)Β 43 (12.8)Β 41 (12.2)Β 44 (13.1)Β 128 (38.1)Β 0.46 (0.12)Β 0.08Β 
    Body use onlyΒ 117 (68.8)Β 23 (13.5)Β 9 (5.3)Β 21 (12.4)Β 53 (31.2)Β 0.46 (0.15)Β Β 
    Genital useΒ 142 (71.4)Β 28 (14.1)Β 7 (3.5)† 22 (11.1)Β 57 (28.6)† 0.39 (0.17)Β Β 
No. conditions§        
    0 or 1Β 119 (75.8)Β 15 (9.6)Β 10 (6.4)Β 13 (8.3)Β 38 (24.2)Β 0.37 (0.20)Β 0.003Β 
    2Β 155 (67.1)Β 26 (11.3)Β 20 (8.7)Β 30 (13.0)Β 76 (32.9)Β 0.44 (0.14)Β Β 
    3Β 114 (65.9)Β 25 (14.4)Β 17 (9.8)Β 17 (9.8)Β 59 (34.1)Β 0.45 (0.15)Β Β 
    4Β 61 (57.0)Β 21 (19.6)Β 6 (5.6)Β 19 (17.8)Β 46 (43.0)Β 0.51 (0.20)Β Β 
    5 or moreΒ 18 (48.6)Β 7 (18.9)† 4 (10.8)Β 8 (21.6)† 19 (51.4)βˆ₯Β 0.53 (0.41)Β Β 
*

Positive antibodies: low, 0.6 β‰₯ A < 0.8; intermediate, 0.8 β‰₯ A < 1.0; high A β‰₯ 1.0; any A β‰₯ 0.6.

†

P < 0.05.

‑

P between 0.05 and 0.15.

Β§

Conditions include bone fracture/osteoporosis, mastitis, pelvic surgeries, IUD use, no genital talc use, OC use, and current smoking; Also adjusted for study center (Massachusetts, New Hampshire), parity (continuous), non-White race, and Jewish religion.

βˆ₯

Ptrend = 0.0005.

Spearman (rank) correlations were calculated between the absorbance reading and several variables quantifiable on a numerical scale. No significant correlations were noted with number of live births, months of breast-feeding, or pack-years of smoking (data not shown). Weak but significant positive correlations were noted between absorbance values and months of OC use (r = 0.09, P = 0.02) and number of cesarean sections (r = 0.10, P = 0.02). A nonsignificant inverse correlation was noted between absorbance and estimated total applications of talc. When genital talc users were characterized by <weekly, weekly, and daily use, there was a trend of borderline significance (P = 0.11) for women using talc more frequently to have the lower antibody levels after adjustment for age, smoking, bone fractures, and OC or IUD use.

Risk for Ovarian Cancer Associated with Antibody-Promoting Events

The variables examined in relation to anti-MUC1 antibodies were then examined in relation to ovarian cancer risk, based upon case-control comparisons (Table 2). Odds ratios for ovarian cancer with each of these variables (except for age which was a matching variable) were calculated and adjusted for age, study site, exact parity, non-White race, and Jewish religion. Our study confirmed the influence of known ovarian cancer risk factors including parity, breast-feeding, and OC use. In addition, we identified previously unreported risk factors, including mastitis, relative risk (and 95% confidence limits) of 0.35 (0.16-0.77); IUD use, relative risk of 0.68 (0.50-0.91); and bone fracture, relative risk of 0.70 (0.53-0.91). The final entry shows the association between number of antibody-promoting conditions and ovarian cancer risk. Compared with women with zero or one condition, the risk for ovarian cancer decreased progressively with relative risks (and 95% confidence limits) of 0.69 (0.52-0.92), 0.64 (0.47-0.88), 0.49 (0.34-0.72), and 0.31 (0.16-0.61), respectively, for women with two, three, four, and five or more conditions (Ptrend < 0.0001). This pattern mirrored the effect of these same conditions on the likelihood that control women had anti-MUC1 antibody (Fig. 2). Finally, risk by number of antibody-promoting conditions was examined separately for major histologic subtypes of ovarian cancer (Table 3). The inverse association was most evident for endometrioid cancers followed by undifferentiated and then invasive serous cancers. Numbers were too limited to make any definitive comments about predictors of antibodies among the 48 cases with preoperative bloods in whom anti-MUC1 antibodies were measured.

Table 2.

Adjusted risk for ovarian cancer by epidemiologic variables in ovarian cancer cases and controls

CasesControlsAdjusted odds ratio (95% confidence interval)Adjusted P*
Age (y)Β Β Β Β Β 
    <35Β 68 (50.8)Β 66 (49.2)Β Β Β 
    35-44Β 123 (46.2)Β 143 (53.8)Β Β Β 
    45-54Β 198 (49.2)Β 204 (50.8)Β Β Β 
    55-64Β 159 (48.6)Β 168 (51.4)Β Β Β 
    β‰₯65Β 120 (46.2)Β 140 (53.8)Β Β Β 
RaceΒ Β Β Β Β 
    WhiteΒ 629 (47.4)Β 699 (52.6)Β 1.00Β 0.007Β 
    Non-WhiteΒ 39 (63.9)Β 22 (36.1)Β 2.14 (1.24-3.69)Β Β 
ReligionΒ Β Β Β Β 
    Non-JewishΒ 614 (47.1)Β 690 (52.9)Β 1.00Β 0.007Β 
    JewishΒ 54 (63.5)Β 31 (36.5)Β 1.88 (1.19-3.00)Β Β 
Marital statusΒ Β Β Β Β 
    Never marriedΒ 107 (62.9)Β 63 (37.1)Β 1.00Β 0.03Β 
    Ever marriedΒ 561 (46.0)Β 658 (54.0)Β 0.66 (0.46-0.96)Β Β 
Pregnancy historyΒ Β Β Β Β 
    Never pregnantΒ 169 (64.0)Β 95 (36.0)Β 1.00Β Β 
    Pregnant but no live birthsΒ 37 (54.4)Β 31 (45.6)Β 0.69 (0.40-1.18)Β 0.18Β 
    At least one live birthΒ 462 (43.7)Β 595 (56.3)Β 0.41 (0.30-0.55)Β <0.0001Β 
Breast-feeding (among parous women)Β Β Β Β Β 
    Never breast-fedΒ 231 (48.6)Β 244 (51.4)Β 1.00Β Β 
    Breast-fed and no mastitisΒ 219 (40.5)Β 322 (59.5)Β 0.72 (0.56-0.93)Β 0.01Β 
    Breast-fed and mastitisΒ 9 (23.7)Β 29 (76.3)Β 0.35 (0.16-0.77)Β 0.009Β 
OC useΒ Β Β Β Β 
    NoΒ 312 (56.1)Β 244 (43.9)Β 1.00Β <0.0001Β 
    YesΒ 356 (42.7)Β 477 (57.3)Β 0.55 (0.43-0.69)Β Β 
OC use in premenopausal subjectsΒ Β Β Β Β 
    NoΒ 102 (65.4)Β 54 (34.6)Β 1.00Β <0.0001Β 
    YesΒ 184 (41.6)Β 258 (58.4)Β 0.38 (0.26-0.57)Β Β 
IUD useΒ Β Β Β Β 
    NoΒ 577 (49.6)Β 586 (50.4)Β 1.00Β 0.01Β 
    YesΒ 91 (40.3)Β 135 (59.7)Β 0.68 (0.50-0.91)Β Β 
Bone fracture/osteoporosisΒ Β Β Β Β 
    NoΒ 544 (50.0)Β 545 (50.0)Β 1.00Β 0.007Β 
    YesΒ 124 (41.3)Β 176 (58.7)Β 0.70Β Β 
ColitisΒ Β Β Β Β 
    NoΒ 645 (48.6)Β 683 (51.4)Β 1.00Β 0.05Β 
    YesΒ 23 (37.7)Β 38 (62.3)Β 0.58 (0.34-1.00)Β Β 
EndometriosisΒ Β Β Β Β 
    NoΒ 611 (47.8)Β 666 (52.2)Β 1.00Β 0.92Β 
    YesΒ 57 (50.9)Β 55 (49.1)Β 1.02 (0.69-1.52)Β Β 
Pelvic surgeryΒ Β Β Β Β 
    No pelvic surgeryΒ 441 (49.4)Β 452 (50.6)Β 1.00Β Β 
    Hysterectomy onlyΒ 39 (51.3)Β 37 (48.7)Β 1.23 (0.75-2.03)Β 0.40Β 
    Tubal sterilization onlyΒ 69 (44.5)Β 86 (55.5)Β 1.00 (0.70-1.42)Β 0.98Β 
    Conization onlyΒ 15 (38.5)Β 24 (61.5)Β 0.64 (0.33-1.26)Β 0.20Β 
    Cesarean section onlyΒ 60 (55.0)Β 49 (45.0)Β 1.38 (0.91-2.09)Β 0.13Β 
    >1 surgeryΒ 44 (37.6)Β 73 (62.4)Β 0.73 (0.48-1.10)Β 0.13Β 
SmokingΒ Β Β Β Β 
    NeverΒ 338 (48.9)Β 353 (51.1)Β 1.00Β Β 
    FormerΒ 235 (46.2)Β 274 (53.8)Β 1.15 (0.83-1.60)Β 0.40Β 
    CurrentΒ 95 (50.3)Β 94 (49.7)Β 0.94 (0.74-1.19)Β 0.59Β 
Talc useΒ Β Β Β Β 
    NoneΒ 319 (48.2)Β 343 (51.8)Β 1.00Β Β 
    Body use onlyΒ 135 (43.7)Β 174 (56.3)Β 0.87 (0.66-1.15)Β 0.33Β 
    Genital useΒ 214 (51.2)Β 204 (48.8)Β 1.16 (0.90-1.49)Β 0.25Β 
No. conditions†     
    0 or 1Β 218 (57.4)Β 162 (42.6)Β 1.00Β Β 
    2Β 220 (48.1)Β 237 (51.9)Β 0.69 (0.52-0.92)Β 0.01Β 
    3Β 150 (45.9)Β 177 (54.1)Β 0.64 (0.47-0.88)Β 0.005Β 
    4Β 67 (38.3)Β 108 (61.7)Β 0.49 (0.34-0.72)Β 0.0002Β 
    5 or moreΒ 13 (26.0)Β 37 (74.0)Β 0.31 (0.16-0.61)Β 0.0007Β 
CasesControlsAdjusted odds ratio (95% confidence interval)Adjusted P*
Age (y)Β Β Β Β Β 
    <35Β 68 (50.8)Β 66 (49.2)Β Β Β 
    35-44Β 123 (46.2)Β 143 (53.8)Β Β Β 
    45-54Β 198 (49.2)Β 204 (50.8)Β Β Β 
    55-64Β 159 (48.6)Β 168 (51.4)Β Β Β 
    β‰₯65Β 120 (46.2)Β 140 (53.8)Β Β Β 
RaceΒ Β Β Β Β 
    WhiteΒ 629 (47.4)Β 699 (52.6)Β 1.00Β 0.007Β 
    Non-WhiteΒ 39 (63.9)Β 22 (36.1)Β 2.14 (1.24-3.69)Β Β 
ReligionΒ Β Β Β Β 
    Non-JewishΒ 614 (47.1)Β 690 (52.9)Β 1.00Β 0.007Β 
    JewishΒ 54 (63.5)Β 31 (36.5)Β 1.88 (1.19-3.00)Β Β 
Marital statusΒ Β Β Β Β 
    Never marriedΒ 107 (62.9)Β 63 (37.1)Β 1.00Β 0.03Β 
    Ever marriedΒ 561 (46.0)Β 658 (54.0)Β 0.66 (0.46-0.96)Β Β 
Pregnancy historyΒ Β Β Β Β 
    Never pregnantΒ 169 (64.0)Β 95 (36.0)Β 1.00Β Β 
    Pregnant but no live birthsΒ 37 (54.4)Β 31 (45.6)Β 0.69 (0.40-1.18)Β 0.18Β 
    At least one live birthΒ 462 (43.7)Β 595 (56.3)Β 0.41 (0.30-0.55)Β <0.0001Β 
Breast-feeding (among parous women)Β Β Β Β Β 
    Never breast-fedΒ 231 (48.6)Β 244 (51.4)Β 1.00Β Β 
    Breast-fed and no mastitisΒ 219 (40.5)Β 322 (59.5)Β 0.72 (0.56-0.93)Β 0.01Β 
    Breast-fed and mastitisΒ 9 (23.7)Β 29 (76.3)Β 0.35 (0.16-0.77)Β 0.009Β 
OC useΒ Β Β Β Β 
    NoΒ 312 (56.1)Β 244 (43.9)Β 1.00Β <0.0001Β 
    YesΒ 356 (42.7)Β 477 (57.3)Β 0.55 (0.43-0.69)Β Β 
OC use in premenopausal subjectsΒ Β Β Β Β 
    NoΒ 102 (65.4)Β 54 (34.6)Β 1.00Β <0.0001Β 
    YesΒ 184 (41.6)Β 258 (58.4)Β 0.38 (0.26-0.57)Β Β 
IUD useΒ Β Β Β Β 
    NoΒ 577 (49.6)Β 586 (50.4)Β 1.00Β 0.01Β 
    YesΒ 91 (40.3)Β 135 (59.7)Β 0.68 (0.50-0.91)Β Β 
Bone fracture/osteoporosisΒ Β Β Β Β 
    NoΒ 544 (50.0)Β 545 (50.0)Β 1.00Β 0.007Β 
    YesΒ 124 (41.3)Β 176 (58.7)Β 0.70Β Β 
ColitisΒ Β Β Β Β 
    NoΒ 645 (48.6)Β 683 (51.4)Β 1.00Β 0.05Β 
    YesΒ 23 (37.7)Β 38 (62.3)Β 0.58 (0.34-1.00)Β Β 
EndometriosisΒ Β Β Β Β 
    NoΒ 611 (47.8)Β 666 (52.2)Β 1.00Β 0.92Β 
    YesΒ 57 (50.9)Β 55 (49.1)Β 1.02 (0.69-1.52)Β Β 
Pelvic surgeryΒ Β Β Β Β 
    No pelvic surgeryΒ 441 (49.4)Β 452 (50.6)Β 1.00Β Β 
    Hysterectomy onlyΒ 39 (51.3)Β 37 (48.7)Β 1.23 (0.75-2.03)Β 0.40Β 
    Tubal sterilization onlyΒ 69 (44.5)Β 86 (55.5)Β 1.00 (0.70-1.42)Β 0.98Β 
    Conization onlyΒ 15 (38.5)Β 24 (61.5)Β 0.64 (0.33-1.26)Β 0.20Β 
    Cesarean section onlyΒ 60 (55.0)Β 49 (45.0)Β 1.38 (0.91-2.09)Β 0.13Β 
    >1 surgeryΒ 44 (37.6)Β 73 (62.4)Β 0.73 (0.48-1.10)Β 0.13Β 
SmokingΒ Β Β Β Β 
    NeverΒ 338 (48.9)Β 353 (51.1)Β 1.00Β Β 
    FormerΒ 235 (46.2)Β 274 (53.8)Β 1.15 (0.83-1.60)Β 0.40Β 
    CurrentΒ 95 (50.3)Β 94 (49.7)Β 0.94 (0.74-1.19)Β 0.59Β 
Talc useΒ Β Β Β Β 
    NoneΒ 319 (48.2)Β 343 (51.8)Β 1.00Β Β 
    Body use onlyΒ 135 (43.7)Β 174 (56.3)Β 0.87 (0.66-1.15)Β 0.33Β 
    Genital useΒ 214 (51.2)Β 204 (48.8)Β 1.16 (0.90-1.49)Β 0.25Β 
No. conditions†     
    0 or 1Β 218 (57.4)Β 162 (42.6)Β 1.00Β Β 
    2Β 220 (48.1)Β 237 (51.9)Β 0.69 (0.52-0.92)Β 0.01Β 
    3Β 150 (45.9)Β 177 (54.1)Β 0.64 (0.47-0.88)Β 0.005Β 
    4Β 67 (38.3)Β 108 (61.7)Β 0.49 (0.34-0.72)Β 0.0002Β 
    5 or moreΒ 13 (26.0)Β 37 (74.0)Β 0.31 (0.16-0.61)Β 0.0007Β 
*

Adjusted for age (continuous), study center (Massachusetts, New Hampshire), parity (continuous), non-White race, and Jewish religion.

†

Conditions include bone fracture/osteoporosis, mastitis, pelvic surgeries, IUD use, no genital talc use, OC use, and current smoking.

Figure 2.

Likelihood of anti-MUC1 antibodies by index of number of conditions and risk for ovarian cancer by same index.

Figure 2.

Likelihood of anti-MUC1 antibodies by index of number of conditions and risk for ovarian cancer by same index.

Close modal
Table 3.

Adjusted risks, 95% confidence intervals, and trends for ovarian cancer of different histologic types associated with number of conditions predisposing to MUC1 antibodies

No. conditionsHistologic subtype
Serous borderline (n = 91)Serous invasive (n = 261)Mucinous (n = 73)Endometrioid/CC (n = 201)Other/Undiff. (n = 42)
0 or 1Β 1.00Β 1.00Β 1.00Β 1.00Β 1.00Β 
2Β 0.74 (0.40-1.35)Β 0.80 (0.55-1.17)Β 0.62 (0.32-1.18)Β 0.68 (0.45-1.02)Β 0.57 (0.27-1.22)Β 
3Β 0.72 (0.38-1.38)Β 0.91 (0.61-1.37)Β 0.41 (0.19-0.89)Β 0.62 (0.39-0.98)Β 0.32 (0.12-0.84)Β 
4Β 0.85 (0.40-1.78)Β 0.53 (0.31-0.90)Β 0.80 (0.36-1.75)Β 0.38 (0.20-0.71)Β 0.27 (0.08-0.97)Β 
5 or moreΒ 0.38 (0.08-1.74)Β 0.38 (0.14-1.02)Β 0.65 (0.18-2.37)Β 0.17 (0.04-0.73)Β 0.31 (0.04-2.42)Β 
PtrendΒ 0.32Β 0.02Β 0.30Β 0.0002Β 0.008Β 
No. conditionsHistologic subtype
Serous borderline (n = 91)Serous invasive (n = 261)Mucinous (n = 73)Endometrioid/CC (n = 201)Other/Undiff. (n = 42)
0 or 1Β 1.00Β 1.00Β 1.00Β 1.00Β 1.00Β 
2Β 0.74 (0.40-1.35)Β 0.80 (0.55-1.17)Β 0.62 (0.32-1.18)Β 0.68 (0.45-1.02)Β 0.57 (0.27-1.22)Β 
3Β 0.72 (0.38-1.38)Β 0.91 (0.61-1.37)Β 0.41 (0.19-0.89)Β 0.62 (0.39-0.98)Β 0.32 (0.12-0.84)Β 
4Β 0.85 (0.40-1.78)Β 0.53 (0.31-0.90)Β 0.80 (0.36-1.75)Β 0.38 (0.20-0.71)Β 0.27 (0.08-0.97)Β 
5 or moreΒ 0.38 (0.08-1.74)Β 0.38 (0.14-1.02)Β 0.65 (0.18-2.37)Β 0.17 (0.04-0.73)Β 0.31 (0.04-2.42)Β 
PtrendΒ 0.32Β 0.02Β 0.30Β 0.0002Β 0.008Β 

NOTE: Adjusted for age (continuous), study site (Massachusetts, New Hampshire), ethnicity (white, non-White), religious background (Jewish, non-Jewish), and parity (continuous).

To date, this is the largest study to examine determinants of anti-MUC1 antibodies and the first to show that conditions that generally increase the likelihood of having antibodies decrease the risk for ovarian cancer. MUC1 is normally present in a glycosylated, membrane-bound form on the apical surface of most polarized epithelial cells of the respiratory, genitourinary, and digestive tracts as well as breast ducts (12). With malignant transformation, epithelial cells lose polarity and overexpress MUC1 on their entire cell surface. A soluble, underglycosylated form circulates in cancer patients, thus becoming available for recognition by the immune system (6, 13). Some healthy women and men also have detectable MUC1 (albeit much lower levels) as well as anti-MUC1 antibodies. In women mostly ages 50 to 70 years, McGuckin et al. assessed the presence of circulating MUC1 using the cancer-associated serum antigen assay. Cancer-associated serum antigen concentrations were elevated in smokers and increased progressively with age (14). In a sample of women from the same study, Richards et al. then measured anti-MUC1 antibodies and found that virtually all women less than age 40 had antibodies and this percentage declined with age (4), somewhat similar to the pattern we observed. It is well established that women have MUC1 and anti-MUC1 antibodies during pregnancy and breast-feeding, presumably due to changes within the breast or uterus that alter MUC1 expression, glycosylation, or shedding (4, 15, 16). In addition, Hinoda et al. observed antibodies specific for the peptide backbone of MUC1 in patients with ulcerative colitis and speculated that inflammation may change MUC1 glycosylation and enhance its immunogenicity (17). One difficulty in evaluating these studies is that assays both for MUC1 and anti-MUC1 antibodies may differ. In measuring antibodies, assays will vary by the specific epitope of MUC1 and the secondary immunoglobulin antibody used. The assay in our study is based on the peptide backbone of MUC1 that we believe is closer to tumor MUC1 and we assessed total immunoglobulin levels including all isotypes, IgG, IgM, and IgA.

In our data, anti-MUC1 antibodies were associated with events affecting the reproductive tract, whose epithelia heavily express MUC1 (18). Injury and/or inflammation of these tissues, surgery, and other events might allow enhancement of MUC1 expression, spillage into circulation, and presentation to the immune system. Thus, the mechanism by which tubal sterilization reduces ovarian cancer risk, previously attributed to preventing substances like talc or endometrial cells from reaching the ovaries (19, 20), may include production of protective antibodies. In our data, cervical conization involving injury and repair of endocervical tissue was also associated with a nonsignificant increase in antibody formation and decrease in risk for ovarian cancer. Antibody formation was also directly correlated with number of cesarean sections, which involve incision and repair of the uterine wall and endometrium. Endometrial expression of MUC1 might also be affected by IUD use, as suggested by biopsies showing a low-grade, chronic inflammation with enhanced mucin staining (21). We found that IUD use increased the likelihood of antibodies in the β€œlow” range and significantly decreased the risk for ovarian cancer. This is the first study to identify an inverse association between ovarian cancer and IUD use, whereas there is considerably more evidence that IUD use reduces risk for endometrial cancer (22), another tumor with high MUC1 expression (23).

An increased likelihood of MUC1-specific antibodies in the β€œhigh” range was found in women reporting bone fracture or a diagnosis of osteoporosis. Both conditions are known to be associated with high interleukin 6 levels (24, 25), an important regulatory cytokine for MUC1 expression (26). Furthermore, a bone fracture might be associated with release of hematopoetic precursors into the circulation, some of which may express MUC1 and be immunogenic (27). We also found an inverse association between bone fracture/osteoporosis and ovarian cancer risk, which to our knowledge has not been shown previously. Interestingly, bone fracture is associated with reduced endometrial and breast cancer risk (28). Whereas this may simply reflect low estrogen, an influence of anti-MUC1 antibodies should also be considered. Besides bone fracture and IUD use, a third factor, which may link the etiology of ovarian and endometrial cancer, is smoking. A decreased risk for endometrial cancer is found in smokers, especially current smokers (29, 30). The data are less clear for ovarian cancer with two recent studies suggesting that smoking may increase the risk only for mucinous histologic subtypes (31, 32). Although current smoking was not clearly related to either anti-MUC1 antibody development or ovarian cancer risk in our univariate analyses, it did improve the cumulative index models in Tables 1 and 2. Furthermore, McGuckin's observation that smokers have higher serum MUC1 levels (presumably from damaged lung epithelium) provides a basis for linking current smoking to anti-MUC1 antibody production (14).

OC use is a strong protective factor for ovarian (and endometrial) cancer and also seemed to generate anti-MUC1 antibodies, particularly among premenopausal women. CA15-3 (MUC1) levels in saliva were found to be 75% higher in OC users compared with nonusers, a nonsignificant difference in that small study (33). Other studies suggest that MUC1 expression in the endometrium is progesterone dependent (34) and up-regulated by exogenous progesterone (35). Considered together, these observations support the speculation that OC users may have higher MUC1 levels that could translate into higher antibody production.

History of mastitis was associated with both increased anti-MUC1 antibodies and decreased ovarian cancer risk in our study. We believe this is an important finding in light of our previous report of a long-term breast cancer survivor in whom MUC1-specific antibody production and mucin-specific T lymphocytes became elevated following mastitis in pregnancy (36). The lactating breast secretes a form of MUC1 that is similar to the underglycosylated form of MUC1 produced by tumors. Thus, mastitis may lead to a potent anti-MUC1 and antitumor immune response, which could explain the substantial decreased risk for ovarian cancer associated with mastitis found in our current study.

Curiously, we found that use of talc in the genital area was associated with significantly decreased levels of anti-MUC1 antibodies. Use of talc in the genital area would expose at least lower genital tract epithelia to talc and conceivably affect MUC1 expression in these tissues. In serial assays of pleural fluid in patients who received talc pleurodesis, inflammatory mediators eventually became depressed (37). Use of talc in the genital area has been consistently found to increase the risk for ovarian cancer in several meta-analyses (38-40). However, some investigators have challenged the association because of uncertainty about its biological basis and the absence of a dose-response relationship (38, 40). Although our present finding may also meet with skepticism, a testable hypothesis is now suggested by the possible link between genital talc exposure and systemic diminution of anti-MUC1 antibodies.

Existing theories of ovarian cancer pathogenesis have invoked incessant ovulation, gonadotropin excess, androgen excess, progesterone deficiency, or deleterious effects of inflammation to explain risk factors for ovarian cancer (41-44). Our findings offer an additional perspective on how OC use, tubal sterilization, and even talc use might exert their effects on ovarian cancer risk and suggests an entirely new set of protective factors such as mastitis, IUD use, and bone fracture that might be explained by the same immune-mediated mechanism. Interestingly, this mechanism may also explain the decreased risk for ovarian cancer associated with mumps parotitis noted in older studies conducted before the widespread use of vaccination (45, 46). Analogous to mastitis, infection of MUC1-rich salivary glands might also lead to an anti-MUC1 immune response and antibody production. Clearly, we have not explained all features of ovarian cancer including the β€œdose-related” decrease in risk associated with multiple pregnancies and length of breast-feeding. Based on the studies reporting anti-MUC1 antibodies in women currently pregnant or breast-feeding, we had expected, but did not observe, that antibodies would increase with the more pregnancies a woman had or the longer she breast-fed. However, it should also be appreciated that anti-MUC1 antibodies are just one of several immune-effecter mechanisms that may also include helper and cytotoxic MUC1-specific T cells that are generated by MUC1 presentation to the immune system. Indeed, the reactions described in sera and T cells from multiparous women suggest that a complete picture of the link between ovarian cancer risk and MUC1 immunity will require assessment of cell-mediated reactions. In addition, immunity to other human mucins, including MUC16 (CA 125), may also need to be examined.

The principal limitation of our study comes from its case-control design. Exposure information was collected by self-report after the diagnosis in cases, introducing the possibility of misclassification. More importantly, we were unable to directly compare anti-MUC1 antibody levels in cases and controls and directly calculate odds ratios based on antibody levels because the cancer itself leads to production of antibodies. Consequently, assessing antibodies in cases after the diagnosis is not useful for identifying earlier events that influenced antibody generation or the predictive value of such antibodies. Prospective studies, in which blood samples are obtained decades or years before the development of ovarian cancer, will be necessary to assess directly the predictive value of anti-MUC1 antibodies on ovarian cancer risk. In addition, prospective studies before and after events like tubal sterilization, IUD use, mastitis, etc. that document the precise changes in the status of anti-MUC1 antibodies will refine our β€œcumulative index model” with its crude assumption that all events might be of equal potency in ability to generate antibodies. Thus, we make no claim this model is final but rather represents a simple foundation for a paradigm shift that will incorporate MUC1 immunity as a key mechanism through which many risk factors for ovarian cancer may exert their influence.

In summary, evidence from this case-control study of ovarian cancer suggests that events predicting the presence of anti-MUC1 antibodies are inversely associated with ovarian cancer risk and that the more conditions a woman experienced to raise antibodies the lower is her risk for ovarian cancer. We believe these data support the immune response as one mechanism of action of β€œtraditional” ovarian cancer risk factors such as OC use and tubal sterilization, as well as novel ones observed in this study including mastitis, bone fracture, and IUD use. If, as we would like to propose, the immune response is a major mechanism, the implications are profound because it may eventually offer new avenues for ovarian cancer prevention through vaccines to stimulate immunity against MUC1 and perhaps other antigens expressed in ovarian cancer. Much work would need to be done, including prospective documentation of the precise changes in cell-mediated and humoral responses to MUC1 associated with pregnancy, breast-feeding, mastitis, and other events. Such studies may have implications beyond ovarian cancer and apply to other cancers with high MUC1 expression including endometrial and breast cancer.

Grant support: National Cancer Institute grants CA 054419 (D.W. Cramer), CA 86381 (D.W. Cramer), and CA 56103 (O.J. Finn).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Special thanks to Mr. Abbott Dean in memory of his wife, Sandra, and to all the participants of this study.

1
Ho SB, Niehans GA, Lyftogt C, et al. Heterogeneity of mucin gene expression in normal and neoplastic tissues.
Cancer Res
β€ˆ
1993
;
53
:
641
–51.
2
Kotera Y, Fontenot JD, Pecher G, Metzgar RS, Finn OJ. Humoral immunity against a tandem repeat epitope of human mucin MUC-1 in sera from breast, pancreatic, and colon cancer patients.
Cancer Res
β€ˆ
1994
;
54
:
2856
–60.
3
von Mensdorff-Pouilly S, Gourevitch MM, Kenemans P, et al. Humoral immune response to polymorphic epithelial mucin (MUC-1) in patients with benign and malignant breast tumours.
Eur J Cancer
β€ˆ
1996
;
32
:
1325
–31.
4
Richards ER, Devine PL, Quin RJ, Fontenot JD, Ward BG, McGuckin MA. Antibodies reactive with the protein core of MUC1 mucin are present in ovarian cancer patients and healthy women.
Cancer Immunol Immunother
β€ˆ
1998
;
46
:
245
–52.
5
Hamanaka Y, Suehiro Y, Fukui M, Shikichi K, Imai K, Hinoda Y. Circulating anti-MUC1 IgG antibodies as a favorable prognostic factor for pancreatic cancer.
Int J Cancer
β€ˆ
2003
;
103
:
97
–100.
6
Vlad AD, Kettel JC, Alajez NM, Carlos CA, Finn OJ. MUC1 immunobiology: from discovery to clinical applications.
Adv Immunol
β€ˆ
2004
;
82
:
249
–93.
7
Agrawal B, Reddish MA, Krantz MJ, Longenecker BM. Does pregnancy immunize against breast cancer?
Cancer Res
β€ˆ
1995
;
55
:
2257
–61.
8
Janerich DT. The influence of pregnancy on breast cancer risk: is it endocrinological or immunological?
Med Hypotheses
β€ˆ
1980
;
6
:
1149
–55.
9
Forsman LM, Jouppila PI, Andersson LC. Sera from multiparous women contain antibodies mediating cytotoxicity against breast carcinoma cells.
Scand J Immunol
β€ˆ
1984
;
19
:
135
–9.
10
Shields LB, Gercel-Taylor C, Yashar CM, et al. Induction of immune responses to ovarian tumor antigens by multiparity.
J Soc Gynecol Investig
β€ˆ
1997
;
4
:
298
–304.
11
Agrawal B, Reddish MA, Longenecker BM. In vitro induction of MUC-1 peptide-specific type 1 T lymphocyte and cytotoxic T lymphocyte responses from healthy multiparous donors.
J Immunol
β€ˆ
1996
;
157
:
2089
–95.
12
Gendler SJ, Spicer AP. Epithelial mucin genes.
Annu Rev Physiol
β€ˆ
1995
;
57
:
607
–34.
13
Fontenot JD, Mariappan SV, Catasti P, Domenech N, Finn OJ, Gupta G. Structure of a tumor associated antigen containing a tandemly repeated immunodominance epitope.
J Biomol Struct Dyn
β€ˆ
1995
;
3
:
245
–60.
14
McGuckin MA, Ramm LE, Joy GJ, Devine PL, Ward BG. Circulating tumor associated mucin concentrations determined by the CASA assay in healthy women.
Clin Chim Acta
β€ˆ
1993
;
214
:
139
–51.
15
Bon GG, Kenemans P, Verstraeten AA, et al. Maternal serum Ca125 and Ca 15-3 antigen levels in normal and pathological pregnancy.
Fetal Diagn Ther
β€ˆ
2001
;
16
:
166
–72.
16
Croce MV, Isla-Larrain MT, Price MR, Segal-Eiras A. Detection of circulating mammary mucin (Muc1) and MUC1 immune complexes (Muc1-CIC) in healthy women.
Int J Biol Markers
β€ˆ
2001
;
16
:
112
–20.
17
Hinoda Y, Nakagawa N, Nakamura H, et al. Detection of a circulating antibody against a peptide epitope on a mucin core protein, MUC1, in ulcerative colitis.
Immunol Lett
β€ˆ
1993
;
35
:
163
–8.
18
Gipson IK, Ho SB, Spurr-Michaud SJ, et al. Mucin genes expressed by human female reproductive tract epithelia.
Biol Reprod
β€ˆ
1997
;
56
:
999
–1011.
19
Hankinson SE, Hunter DJ, Colditz GA. Tubal ligation, hysterectomy, and risk of ovarian cancer.
JAMA
β€ˆ
1993
;
270
:
2813
–8.
20
Green A, Purdie PD, Bain C, et al. Tubal sterilization, hysterectomy and decreased risk of ovarian cancer. Survey of Women's Health Study Group.
Int J Cancer
β€ˆ
1997
;
71
:
948
–51.
21
Hester LL Jr, Kellett WW III, Spicer SS, Williamson HO, Pratt-Thomas HR. Effects of the intrauterine contraceptive device on endometrial enzyme and carbohydrate histochemistry.
Am J Obstet Gynecol
β€ˆ
1970
;
106
:
1144
–54.
22
Hubacher D, Grimes DA. Noncontraceptive health benefits of intrauterine devices: a systematic review.
Obstet Gynecol Surv
β€ˆ
2002
;
57
:
120
–8.
23
Sivridis E, Giatromanolaki A, Koukourakis MI, Georgiou L, Anastasiadis P. Patterns of episialin/MUC1 expression in endometrial carcinomas and prognostic relevance.
Histopathology
β€ˆ
2002
;
40
:
92
–100.
24
Papadopoulos NG, Georganas K, Skoutellas V, Konstantellos E, Lyritis GP. Correlation of interueukin-6 serum levels with bone density in postmenopausal women.
Clin Rheumatol
β€ˆ
1997
;
16
:
162
–5.
25
Strecker W, Gebhard F, Rager J, Bruckner UB, Steinbach G, Kinzl L. Early biochemical characterization of soft-tissue trauma and fracture trauma.
J Trauma Injury Infect Crit Care
β€ˆ
1999
;
47
:
358
–64.
26
Gaemers IC, Vox HL, Volders HH, van der Valk SW, Hilkens J. A stat-responsive element in the promoter of the episialin/MUC1 gene is involved in its overexpression in carcinoma cells.
J Biol Chem
β€ˆ
2001
;
276
:
6191
–9.
27
Brugger W, Buhring HJ, Grunebach F, et al. Expression of MUC-1 epitopes on normal bone marrow: implications for the detection of micrometastatic tumor cells.
J Clin Oncol
β€ˆ
1999
;
17
:
1535
–44.
28
Newcomb PA, Trentham-Dietz A, Egan KM, et al. Fracture history and risk of breast and endometrial cancer.
Am J Epidemiol
β€ˆ
2001
;
153
:
1071
–8.
29
Terry PD, Miller AB, Rohan TE. A prospective cohort study of cigarette smoking and the risk of endometrial cancer.
Br J Cancer
β€ˆ
2002
;
86
:
1430
–5.
30
Brinton LA, Barrett RJ, Berman ML, Mortel R, Twiggs LB, Wilbanks GD. Cigarette smoking and the risk of endometrial cancer.
Am J Epidemiol
β€ˆ
1993
;
137
:
281
–91.
31
Pan SY, Ugnat AM, Mao Y, Wen SW, Johnson KC. Association of cigarette smoking and the risk of ovarian cancer.
Int J Cancer
β€ˆ
2004
;
111
:
124
–30.
32
Zhang Y, Coogan PF, Palmer JR, Strom BL, Rosenberg L. Cigarette smoking and increased risk of mucinous epithelial ovarian cancer.
Am J Epidemiol
β€ˆ
2004
;
159
:
33
–9.
33
McIntyre R, Bigler L, Dellinger T, Pfeifer M, Mannery T, Stredkfus C. Oral contraceptive usage and the expression of CA 15-3 and c-erB-2 in the saliva of healthy women.
Oral Surg Med Pathol Radiol Endod
β€ˆ
1999
;
88
:
687
–90.
34
Hild-Petito S, Fazleabas AT, Julian J, Carson DD. Mucin (Muc-1) expression is differentially regulated in uterine luminal and glandular epithelia of the baboon (Papio anubis).
Biol Reprod
β€ˆ
1993
;
54
:
939
–47.
35
Hewetson A, Chilton BS. Molecular cloning and hormone-dependent expression of rabbit Muc1 and the cervix and uterus.
Biol Reprod
β€ˆ
1997
;
57
:
468
–77.
36
Jerome KR, Kir AD, Pecher G, Ferguson WW, Finn OJ. A survivor of breast cancer with immunity to MUC-1 mucin, and lactational mastitis.
Cancer Immunol Immunother
β€ˆ
1997
;
43
:
355
–60.
37
D'Agostino P, Camemi AR, Arcoleo F, et al. Matrix metalloproteinases production in malignant pleural effusions after talc pleuodesis.
Clin Exp Immunol
β€ˆ
2003
;
34
:
138
–42.
38
Gross AJ, Berg PH. A meta-analytical approach examining the potential relationship between talc exposure and ovarian cancer.
J Expo Anal Environ Epidemiol
β€ˆ
1995
;
5
:
181
–95.
39
Cramer DW, Liberman RF, Titus-Ernstoff L, et al. Genital talc exposure and risk of ovarian cancer.
Int J Cancer
β€ˆ
1999
;
81
:
351
–6.
40
Huncharek M, Gerschwind JF, Kupelnick B. Perineal application of cosmetic talc and risk of invasive epithelial ovarian cancer: a meta-analysis of 11,933 subjects from sixteen observational studies.
Anticancer Res
β€ˆ
2003
;
23
:
1955
–60.
41
Fathalla MF. Incessant ovulation: a factor in ovarian neoplasia?
Lancet
β€ˆ
1971
;
2
:
163
.
42
Cramer DE, Welch WR. Determinants of ovarian cancer risk. II. Inference regarding pathogenesis.
Natl Cancer Inst
β€ˆ
1983
;
1
:
717
–21.
43
Ness RB, Cottreau C. Possible role of ovarian epithelial inflammation in ovarian cancer.
J Natl Cancer Inst
β€ˆ
1999
;
91
:
1459
–67.
44
Risch HA. Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of androgens and progesterone.
J Natl Cancer Inst
β€ˆ
1998
;
90
:
1774
–86.
45
West RD. Epidemiologic study of malignancies of the ovaries.
Cancer
β€ˆ
1966
;
19
:
1001
.
46
Newhouse ML, Pearson RM, Fullerton JM, Boesen EA, Shannon HS. A case control study of carcinoma of the ovary.
Br J Prev Soc Med
β€ˆ
1977
;
31
:
148
–53.