Associations between Parity, History of Breastfeeding, and T-cell Profile of Ovarian Tumors.

A total of 1,706 women were included in our analyses of parity, history of breastfeeding, and ovarian tumor immune cell profiles ( Table 1 ). Ovarian cancer cases were diagnosed between 1997 and 2015, with a median year of diagnosis of 2004. Compared to nulliparous women, parous women were older at the time of ovarian cancer diagnosis (60.0±9.6 vs. 52.0±10.1 years), more likely to have undergone tubal ligation (20.4% vs. 2.6%) or ever used oral contraception (66.3% vs. 55.2%) and more likely to be diagnosed with high grade serous tumors (71.2% vs. 47.7%). Among parous women, the majority reported having ever breastfed (64.4%) and these women were slightly younger at ovarian cancer diagnosis (59.0±9.8 years) compared to those who never breastfed (61.0±9.0 years). Women who had breastfed reported having a greater number of children (2+ children, 83.1% vs. 79.7%) and were more likely to use oral contraception (69.1% vs. 61.1%) than women who never breastfed. Tumor histology also varied slightly according to breastfeeding status, where women who breastfed had a slightly higher frequency of endometrioid (13.3% vs. 11.4%) and clear cell tumors (6.2% vs. 3.7%) than those who never breastfed. The mean time between last birth (presumably the last breastfeeding episode) and ovarian cancer diagnosis was 26.0±10.8 years. Compared to ovarian tumors of nulliparous women, there were no statistically significant differences in T cell abundance among parous women and in high-grade serous or endometrioid histotypes ( Table 2 ). Among women with clear cell tumors, parity was associated with increased odds of T cell cytotoxicity (OR: 1.57, 95% CI: 1.01–2.44) and recently activated cytotoxic T cell abundance (OR: 1.94, 95% CI: 1.15–3.29) compared to nulliparity. No differences were observed in the association between parity and T cell abundance according to the number of pregnancies ( Supplemental Table 2 ). To examine the associations between a history of breastfeeding and ovarian tumor T cell abundance, we conducted two analyses: 1) comparing nulliparous women to parous women who never or ever breastfed, and 2) comparing never vs. ever breastfeeding among parous women. Compared to nulliparous women, parous women who had breastfed had increased overall T cell abundance (OR:1.20, CI:1.00–1.43), but we did not observe significant associations when comparing parous women who never breastfed to nulliparous women ( Supplemental Table 3 ). Among women with clear cell tumors, there was a significant increase in the odds of cytotoxic and recently activated cytotoxic T cell abundance in parous women who never breastfed (OR:1.74, CI:0.99–3.06, OR:2.74, CI:1.40–5.36, respectively) and parous women who had breastfed (OR:1.58, CI:0.99–2.52, OR:1.58, CI:0.90–2.79, respectively). These results support that among clear cell tumors, parity is associated with increased odds of cytotoxic and recently activated T cell abundance, independent of breastfeeding history. When restricting to parous women, history of breastfeeding was associated with increased odds of higher abundance of all T cell types (i.e., total T cells, helper, cytotoxic, regulatory, and exhausted T cells) in all histotypes combined, with ORs ranging from 1.11 to 1.42 ( Table 3 ) compared to parous women who never breastfed. Among women with high-grade serous tumors who breastfed, there was a significant increase in the odds of recently activated helper T cells (OR: 1.23, 95% CI: 1.00–1.50), recently activated cytotoxic T cells (OR: 1.18, 95% CI: 2.02–2.38), regulatory T cells (OR: 1.29, 95% CI: 1.07–1.55), and exhausted T cells (OR: 1.55, 95% CI: 1.19–2.01), and nonsignificantly increased odds of all other T cell types. Although not statistically significant, the there was an inverse association recently activated helper T cells, recently activated cytotoxic T cells, and regulatory T cells among women with clear cell tumors who had ever breastfed compared to those who had never breastfed. In addition, we stratified by the number of children to address the impact of multiparity and observed comparable results for ever breastfed compared to never breastfed among women with two or more children ( Supplemental Table 4 ). For every 6 months of breastfeeding among parous women, we observed a potential increase in infiltration of activated helper T cells (OR 1.13, 95% CI: 0.99–1.29) and regulatory T cells (OR:1.10, 95%CI: 0.96–1.25) overall ( Table 4 ). Among women who breastfed for 12 months or more with all histotypes combined, there were increasing odds of cells being positive for all T cell types, with ORs ranging from 1.08 1.41. Among women with high grade serous tumors, similar associations were observed with a non-significant increase in activated helper T cell (OR 1.13, 95% CI: 0.97–1.31) and regulatory T cell (OR: 1.14; 95% CI: 0.98–1.31) abundance with every 6 months of breastfeeding. In contrast, among clear cell tumors, increased breastfeeding duration per 6 months was associated with a lower abundance of helper T cells (OR: 0.45; CI:0.30–0.67), activated helper T cells (OR:0.43, CI:0.21–0.87), cytotoxic T cells (OR: 0.63, CI:0.44–0.90), and regulatory T cells (OR:0.27, CI:0.13–0.57). To assess potential differences in time between the last breastfeeding episode and ovarian cancer diagnosis in our study sample, we stratified the data by time since last birth. Using the mean time since last birth among parous women who had never breastfed, more than 75% of women who had ever breastfed and had information on the timing of their last birth had given birth 20 years or more prior to their diagnosis of ovarian cancer ( Supplemental Table 5 , DOVE, NECC, NHSII only). We did not observe clear differences in the associations between breastfeeding and tumor T-cell abundance over time since the last birth. Since the age of tumor samples could influence the antigenicity of the immune markers, we conducted a sensitivity analysis stratified by tumor sample age or time from specimen collection to staining. Using the mean tumor sample age (20 years), we observed stronger associations among samples less than 20 years old compared to samples that had been archived for 20 years or more ( Supplemental Table 6 ). When stratified by study site, while most associations were no longer significant owing to the small sample size per stratum, the direction of associations for each T cell subtype was similar across study sites with no statistically significant heterogeneity ( Supplemental Table 7 )

We conducted a case-only analysis of 1,706 women with epithelial ovarian cancer from four studies: Diseases of the Ovary and their Evaluation Study (DOVE), New England Case Control Study (NECC), Nurses’ Health Study (NHS), and NHSII. All participants provided implied consent for questionnaires and signed a medical release forms for acquisition of tissue and medical records. DOVE is a population-based ovarian cancer case-control study that enrolled 1,502 ovarian cancer cases ages 35–74 years from 13 counties in Western Washington State between 2002 and 2009 ( 25 ). Ovarian cancer cases were identified using a population-based cancer registry that participated in the Surveillance, Epidemiology, and End Results Program of the National Cancer Institute ( 26 ). Participants completed in-person interviews regarding events occurring prior to their ovarian cancer diagnosis, and provided information on demographic and lifestyle factors, family history of cancer, and reproductive history. The NECC is a population-based ovarian cancer case-control study that enrolled 2,203 women with ovarian cancer aged 18–80 years residing in Eastern Massachusetts and New Hampshire over three phases (1992–1997, 1998–2003, 2003–2008) ( 27 ). Ovarian cancer cases were identified using hospital records and state cancer registries. All participants completed in-person interviews to collect information on demographic and lifestyle factors, family history of cancer, and reproductive history with respect to a reference date of one year prior to ovarian cancer diagnosis. The NHS is a prospective cohort study established in 1976 that enrolled 121,701 female registered nurses, 30–55 years old, residing in 11 US states who completed and returned a mailed questionnaire ( 28 ). The NHSII began in 1989 and enrolled 116,429 female registered nurses, aged 25 to 42 years, from 14 US states who completed a similar questionnaire ( 29 ). The updated exposure status and disease outcomes were assessed using biennial questionnaires. Incident ovarian cancer cases were identified using self-report questionnaires, reports from family, or linkage with the National Death Index ( 30 ). Ovarian cancer diagnoses were confirmed by a medical record review or linkage to state cancer registries. The study was conducted in accordance with the ethical guidelines of the Declaration of Helsinki. The DOVE protocol was approved by the Fred Hutchinson Cancer Research Center Institutional Review Board. All DOVE participants provided written informed consent for study participation. The NHS and NHSII protocol were approved by the Institutional Review Boards of the Brigham and Women’s Hospital, Harvard T.H. Chan School of Public Health. Completion of the self-administered questionnaire was considered implied consent. The NECC protocol was approved by Institutional Review Boards of the Brigham and Women’s Hospital and Dartmouth Medical School. All NECC participants provided written informed consent for study participation. In DOVE and NECC, parity and history of breastfeeding for each pregnancy lasting six months or longer were assessed, including age at each pregnancy, ever breastfed for any duration (yes/no), and breastfeeding duration (months). In the NHS and NHSII, parity was assessed as having completed a pregnancy of six or more months. Parity was categorized as 0, 1, 2, or more children. In the NHS, participants were asked about the total lifetime duration of breastfeeding for all births combined in 1986 (never breastfed, <1, 1–3, 4–6, 7–11, 12–17, 18–23, 24–35, 36–47, and ≥48 months). In the NHSII, the total lifetime duration of breastfeeding was assessed in 1993 using the same questionnaire as the NHS; in 1997, the duration of breastfeeding was assessed for each birth. For the current analysis, parous women were categorized as having ever/never breastfed, and among those who breastfed, the total lifetime duration of breastfeeding across all children was calculated and categorized as 0, 1 to ≤6 months, >6 to <12 months, and ≥12 months. Additional information on risk factors relevant to the association of parity and history of breastfeeding with ovarian cancer included age at ovarian cancer diagnosis, race (white, non-white), year of birth (1949), body mass index (BMI) 1–2 years prior to ovarian cancer diagnosis in kg/m 2 (BMI; <20, 20 to <25, 25 to <30, ≥30), smoking status (never, former, current), highest education status (less than high school, high school, some college, bachelor degree, graduate or professional degree, unknown), oral contraceptive use (never, <1 year, 1 to <5 years, ≥5 years), parity (1, 2+), tubal ligation (yes, no, unknown), endometriosis (yes, no, unknown), and family history of breast or ovarian cancer (yes, no, unknown)( 1 ). Formalin-fixed paraffin-embedded, treatment-naïve ovarian tumor tissue samples were collected from ovarian cancer patients who consented to contribute to tumor tissue samples. There were 926 cases from DOVE that provided tumor tissue, 398 from NECC, and 528 from NHS/NHSII. All ovarian tumor tissues were reviewed by a gynecologic pathologist (JLH, TRS, MK), information on tumor characteristics (e.g., morphology, grade, histology) was recorded, and ovarian tumor tissue microarrays (TMAs) were created (21 TMAs in total: 8 TMAs in DOVE, 6 TMAs in NECC, and 7 TMAs in NHS/NHSII). TMAs were constructed using four 0.6 mm (DOVE), up to three 1 mm (NECC), or three 0.6 mm (NECC/NHS/NHSII) cores per case. The process of collecting ovarian tumor tissue in NECC and NHS/NHSII has been previously described ( 31 ). Ovarian cancer cases included on the TMAs had similar ovarian cancer risk factor distributions to all ovarian cancer cases included in the study ( 32 ). Using multiplex immunofluorescence (mIF) assays, TMA slides were analyzed using a T cell panel measuring CD3+ (total T cells, Thermo Fisher Scientific Cat# MA5–14524, RRID:AB_10982026), CD3+CD4+ (helper, Cell Marque Cat# 104R, RRID:AB_1516770), CD3+CD8+ (cytotoxic, Agilent Cat# M7103, RRID:AB_2075537), CD3+CD4+FOXP3+ (regulatory), and CD3+CD4+CD69+/CD3+CD8+CD69+ activation (Abcam Cat# ab233396, RRID:AB_2922929) T cell markers in DOVE, NECC, and NHS/NHSII TMAs ( Supplemental Table 1 )( 31 ). An additional panel, including markers of T cell exhaustion, CD3+TIM3+ (Cell Signaling Technology Cat# 45208, RRID:AB_2716862), and CD3+PD1+TIM3+ (irreversible exhaustion, Cell Signaling Technology Cat# 60333, RRID:AB_2943233), was measured on NECC and NHS/NHSII TMAs. All mIF assays were performed at Moffitt Cancer Center using the Akoya Biosciences Opal ™ 7-Color Automation immunohistochemistry kit (Akoya Biosciences Cat# NEL811001KT, RRID:AB_3665660) and the Vectra ® 3 Automated Quantitative Pathology Imaging System (RRID:SCR_025828). The Opal panel allowed for staining of five immunohistochemistry antibodies plus 4′,6-diamidino-2-phenylindole (DAPI) and pan-cytokeratines (Agilent Cat# M3515, RRID:AB_2132885) on the same slide using the OPAL 7-color kit tyramide signal amplification (TSA)-conjugated to individual fluorophores. The multispectral Vectra ® microscope was used to capture slide images for specialized image analysis and data consolidation using HALO software (Indica Labs, New Mexico, RRID:SCR_018350). Epithelium and stroma compartments were defined using a random forest machine learning algorithm based on DAPI and pan-cytokeratin staining ( 31 ). T cell subpopulations were defined by marker co-expression, e.g. T-cytotoxic (CD3+CD8+), and were scored separately for tumor stroma and epithelium ( 33 ). Our analysis evaluated the positive cell counts within the tumor epithelial compartment. For each marker, a positivity threshold within the nucleus or cytoplasm was determined according to the visual intensity and compared with published staining patterns for each antibody. Cells above the positivity threshold were considered positive for this marker. Colocalized cells were considered positive if the fluorescence intensity for all relevant stains exceeded the established positivity threshold. Of the 1,852 ovarian cancer cases with tumor tissue on the TMAs with unfolded cores and complete staining, we excluded ovarian cancer cases with unknown parity or breastfeeding status prior to diagnosis (n=146), resulting in 1,706 ovarian cancer cases being included in the current analyses. All histotypes of epithelial ovarian cancer were included in the primary analysis. Analyses were performed using beta-binomial models with a random effect accounting for multiple cores per tumor to estimate odds ratios (OR) and 95% confidence intervals (CIs), interpreted as the ratio of odds that a cell was found to be positive for each T-cell marker compared to those with the exposure of interest (parity, breastfeeding) to those without (nulliparous, never breastfed)( 34 ). Models for all markers, except for total T cells (CD3+), were adjusted for tertiles of total T cells (CD3+) as a fixed effect to better estimate the association between parity or history of breastfeeding with each T cell subset, independent of the total number of T cells. All models were adjusted for age at ovarian cancer diagnosis, oral contraceptive use (<1 year, 1–5 years, ≥5 years), and study (DOVE, NEC, NHS, NHSII). The models were additionally adjusted for breastfeeding status (yes, no, or nulliparous) when evaluating the effects of parity. When evaluating the effects of ever breastfeeding and lifetime total breastfeeding duration among parous women, the models were adjusted for parity (1, 2+). Models including all ovarian cancers were adjusted for histotype (i.e., high-grade serous, low-grade serous, mucinous, endometrioid, clear cell, and other epithelial tumors), and histotype-specific estimates were reported when the stratum size was sufficient. Inclusion of other risk factors including BMI, smoking status, highest education status, tubal ligation, endometriosis, family history of breast or ovarian cancer, and stage did not result in appreciable changes to the observed estimates and therefore were not included in the model. Additional sensitivity analyses included the assessment of parity with and without breastfeeding compared to nulliparity, number of children compared to nulliparity, and history of breastfeeding by the number of children. We also stratified by the time since last birth and age of the tumor sample as potential effect measure modifiers. Site-specific associations of parity and breastfeeding were estimated, and Cochrane’s Q statistic was used for heterogeneity for random-effects meta-analysis to assess potential heterogeneity by study site. NHS/NHSII data cannot be publicly shared because of participant confidentiality and privacy concerns. According to standard controlled access procedures, applications to use NHS/NHSII resources will be reviewed by our External Collaborations Committee to verify that the proposed use maintains the protection of the privacy of participants and confidentiality of the data. Further information including the procedures to obtain and access data from the Nurses’ Health Studies is available at https://www.nurseshealthstudy.org/researchers (contact email: [email protected] ). NECC data that support the findings of this study are available upon request and reviewed by the study leadership. Applications to use DOV data can be submitted according to standard controlled access procedures by contacting the DOV PIs. Other data generated in this study are available upon reasonable request from the corresponding author.

In a very large study of ovarian cancer, we observed that parity without a history of breastfeeding was not associated with tumor infiltration of helper T cells overall, while among parous women, ever having breastfed was associated with increased tumor immune infiltration of all T cell types in high-grade serous tumors, but not in endometrioid or clear cell tumors. Previously, we observed that breastfeeding is associated with a lower risk of high-grade serous tumors, the most fatal histotype, than other histotypes( 1 ). Major immunologic changes occur during pregnancy and postpartum periods, and even long-term years after the last pregnancy episode. During pregnancy, the number of circulating T, B, and natural killer (NK) cells decreases, which helps prevent immune rejection of the fetus ( 35 – 38 ). In contrast, increased immune activity is observed during the postpartum period, which is likely to promote the transfer of maternal macrophages, neutrophils, lymphocytes, and antibodies to the infant ( 35 , 36 , 39 – 41 ). One study reported that the number of circulating lymphocytes, T cells (CD3+), and B cells (CD19+) increased in postpartum women up to 10 weeks following pregnancy compared to non-pregnant women ( 21 ). In terms of the potential long-term effect of parity on systemic immunity, two studies observed increased levels of inflammatory markers (e.g., C-reactive protein and fibrinogen) with higher parity measured years after the last pregnancy among women with a mean age of 55–60 years ( 23 , 24 ). Despite this evidence suggesting a potential long-term effect of parity on systemic immunity, we did not observe associations between parity and overall ovarian tumor T-cell abundance for all histotypes combined with high-grade serous or endometrioid subtypes. Among parous women with clear cell tumors, some results show increased odds of cytotoxic T cells and T cell exhaustion with parity, although the wide confidence intervals should be interpreted with caution. Furthermore, clear cell tumors are known to be associated with endometriosis, which may reflect the pre-existing inflammatory microenvironment in nulliparous women ( 42 ). In our study, we observed associations between breastfeeding for 12 or more months (compared to never) and increased infiltration of all T cell types, which was likely a reflection of the association among high-grade serous tumors which represented nearly 70% of all tumors in our study. We also observed a trend of greater tumor infiltration of activated helper T cells (CD3+CD4+CD69+) every 6 months of breastfeeding among all histotypes, except for clear cells, suggesting the potential impact of a history of breastfeeding on ovarian tumor immunity in most histotypes. However, it is important to note that the sample size within each histotype stratum except high grade serous was likely insufficient to detect significant histotype-specific associations. There are several ways in which breastfeeding can result in long-term changes to the systemic immune system in parous women that can be observed for years following the episode of breastfeeding which may influence the body’s immune response to tumor development and alterations in the tumor immune microenvironment. Cytokine levels decrease with longer breastfeeding ( 43 – 46 ). Changes in prolactin levels during pregnancy and breastfeeding may also play an important role in shaping the immune landscape. Prolactin levels peak towards the end of pregnancy and then decrease after birth ( 47 ). While prolactin levels are overall lower postpartum, slight surges in prolactin and oxytocin occur during episodes of breastfeeding to stimulate milk production, which may also lead to an enhanced immune response and lowered inflammation through reduced production of pro-inflammatory cytokines ( 48 , 49 ). Specifically, prolactin has been shown to enhance production of the anti-inflammatory cytokine, interleukin-10 ( 50 ). It is possible that the enhanced immune response from episodes of breastfeeding could result in immune reprogramming and observed long-term, as in our analyses. Prolactin may also be associated with the increased production of interleukin-12 and interferon-gamma. Although pro-inflammatory cytokines are considered, downstream effects include the activation and proliferation of natural killer and T cells ( 51 – 53 ). Therefore, it is possible that prolactin leading to enhanced systemic immunity may mediate the association between breastfeeding and an increased abundance of T cells in ovarian tumors among parous women, except for women who develop clear cell tumors that may differ in etiology and behavior. Given that ovarian cancer is a heterogeneous disease ( 54 ), we reported associations of parity and breastfeeding with tumor immune infiltration of T cells overall and by histotype (high-grade serous, endometrioid, and clear cell). Among parous women, a history of breastfeeding was associated with increased odds of tumor T-cell infiltration among high-grade serous and potentially endometrioid tumors, while there was no or an inverse association between clear cell tumors. While this may be due to the smaller number of endometrioid and clear cell cases, it should be noted that high-grade serous tumors often exhibit higher levels of immune infiltration than other histotypes ( 55 , 56 ). Hormonal, inflammatory, and immune changes associated with breastfeeding may be particularly relevant to the development of high-grade serous tumors. During breastfeeding episodes, the increase in prolactin leads to the reduction of gonadotropins, including follicle-stimulating hormones, which have proliferative effects and may be involved in high-grade serous carcinogenesis ( 57 , 58 ). While it is known that high grade serous tumors have distinct immune profiles compared to other histotypes and may be more influenced by immune modulation, sample size for endometrioid and clear cell histotypes were limited and the findings should be interpreted with caution. Although we observed that breastfeeding was associated with increased T cell infiltration among parous women, a history of breastfeeding has not been associated with ovarian cancer survival ( 59 – 61 ). In parous women who had breastfed compared to those who had never breastfed, we observed increases in both cytotoxic T cells, which have been associated with better prognosis, and in regulatory T cells and T cell exhaustion markers, which have been linked to worse survival outcomes in some studies ( 62 – 64 ), although the presence of increased levels of T cells, including activated and exhausted phenotypes, support an overall immune response. Another possibility is that the observed increase in T cell infiltration related to breastfeeding may not be sufficient to affect survival. Importantly, the current study only evaluated the abundance of T cells in the tumor microenvironment; therefore, there may be other immune cells involved or relevant spatial context of immune cells that may need to be considered in understanding the relationships between history of breastfeeding, tumor immune profile, and ovarian cancer prognosis. T cells were also measured exclusively in the tumor compartment, and the location of T cells may be more informative than density ( 62 , 65 ). However, these findings still help shed light on potential pathways to linking these factors to ovarian cancer development. To our knowledge, this is the first study to investigate the association between a history of parity and breastfeeding on tumor immune profiles. A major strength of this study is the large number of ovarian cancer cases with detailed exposure information. Using the case-only analysis approach, we were able to adjust for and stratify by histotype. This study leveraged data from multiple studies to increase the generalizability of the findings. However, approximately 90% of the study population was identified as white or presumed white; therefore, these findings should be evaluated in more diverse populations. Information on the initiation of breastfeeding for each child was not available; therefore, we were unable to account for time since the last breastfeeding episode in our analyses. However, we were able to consider the time since the last birth as less than 20 and 20 years or more since the last birth. The range of time from diagnosis and ovarian tumor sample collection to the assay also varied, and immune markers may be less likely to stain positive in older samples due to possible reduced antigenicity ( 32 ), although we and others have performed sensitivity analysis limited to samples that were <20 years old and observed similar results ( 31 ). Batch effect was also minimized due to the use of the same staining protocol performed at the same institution for all samples and mIF was performed on tissues from all studies across two batches. We acknowledge the limitations of not directly accounting for potential sources of variation from batch or TMA slide, however, models were adjusted for study site which is highly correlated with TMA slide. It is also important to note that causality cannot be inferred, as this was an observational study, and mechanistic studies are warranted to elucidate the biological mechanisms. The process through which parity and breastfeeding impact the ovarian tumor immune microenvironment likely involves involve complex, long-term immune reprogramming and would not be a direct, immediate effect. There is potential for residual confounding as well, and the observed association may be confounded due to differences in health behaviors (e.g., BMI, education, smoking) among those who breastfed compared to those who did not, although adjusting for these covariates did not result in meaningful changes in the observed estimates. Diagnosis stage may also be a contribute to the immune microenvironment, however, when we additionally adjusted for stage, we observed similar results and given that stage is highly correlated with histology and opted to limit the inclusion of additional factors to the most parsimonious model. Overall, our findings indicate that a history of breastfeeding may have immunomodulatory effects on the ovarian tumor immune microenvironment. These insights have the potential to serve a foundation for future research exploring mechanisms linking parity and breastfeeding with prognosis. Replication in independent studies with more diverse populations and additional immune markers is needed to better understand the association between parity and breastfeeding with the ovarian tumor immune landscape. An advanced understanding of how and why the history of breastfeeding may potentially lead to changes in immune profiles could open new opportunities for interventions to enhance immunogenicity of ovarian tumors.

Studies on breastfeeding have consistently shown inverse associations with the risk of ovarian cancer ( 1 – 16 ). However, beyond altering ovulation patterns and hormone levels, relatively little is known about the potential mechanisms by which these factors lower the risk of ovarian cancer. Recent evidence suggests that these reproductive factors may influence the host immune response, which has been recognized as increasingly important for preventing or promoting cancer development ( 17 ). Pregnancy is associated with changes in both maternal innate and adaptive immunity to prevent immune reactions against the fetus ( 18 – 20 ). In addition to changes in immune and inflammatory markers across trimesters, breastfeeding after birth continues to alter the distribution and function of T-cells to promote immune reconstitution after pregnancy ( 18 , 21 ). Breastfeeding episodes contribute to changes in circulating maternal T cell populations, including helper, cytotoxic, and regulatory T cells ( 21 ). Furthermore, a longer duration of breastfeeding has been associated with lower systemic inflammation in middle-aged women with a history of gestational diabetes ( 22 ). Although evidence is limited, several studies have reported possible long-term systemic effects of parity on immune function ( 23 , 24 ), suggesting that these reproductive events could influence the antitumor immune response in ovarian carcinogenesis even many years later. However, to our knowledge there is no study that have investigated the associations between parity, breastfeeding and ovarian tumor immunity, which may shed light on potential immunological pathways linking these reproductive factors to ovarian cancer development. Here, we conducted a study to evaluate the hypothesis that parity and history of breastfeeding are associated with alterations in the ovarian tumor immune microenvironment, specifically T-cell infiltration into the tumor.