Postpartum Breast Cancer and Survival in Women With Germline BRCA Pathogenic Variants

Importance

In young-onset breast cancer (YOBC), a diagnosis within 5 to 10 years of childbirth is associated with increased mortality. Women with germline BRCA1/2 pathogenic variants (PVs) are more likely to be diagnosed with BC at younger ages, but the impact of childbirth on mortality is unknown.

Objective

To determine whether time between most recent childbirth and BC diagnosis is associated with mortality among patients with YOBC and germline BRCA1/2 PVs.

Design, Setting, and Participants

This prospective cohort study included women with germline BRCA1/2 PVs diagnosed with stage I to III BC at age 45 years or younger between 1950 and 2021 in the United Kingdom, who were followed up until November 2021. Data were analyzed from December 3, 2021, to November 29, 2023.

Exposure

Time between most recent childbirth and subsequent BC diagnosis, with recent childbirth defined as 0 to less than 10 years, further delineated to 0 to less than 5 years and 5 to less than 10 years.

Main Outcomes and Measures

The primary outcome was all-cause mortality, censored at 20 years after YOBC diagnosis. Mortality of nulliparous women was compared with the recent post partum groups and the 10 or more years post partum group. Cox proportional hazards regression analyses were adjusted for age, tumor stage, and further stratified by tumor estrogen receptor (ER) and BRCA gene status.

Results

Among 903 women with BRCA PVs (mean [SD] age at diagnosis, 34.7 [6.1] years; mean [SD] follow-up, 10.8 [9.8] years), 419 received a BC diagnosis 0 to less than 10 years after childbirth, including 228 women diagnosed less than 5 years after childbirth and 191 women diagnosed 5 to less than 10 years after childbirth. Increased all-cause mortality was observed in women diagnosed within 5 to less than 10 years post partum (hazard ratio [HR], 1.56 [95% CI, 1.05-2.30]) compared with nulliparous women and women diagnosed 10 or more years after childbirth, suggesting a transient duration of postpartum risk. Risk of mortality was greater for women with ER-positive BC in the less than 5 years post partum group (HR, 2.35 [95% CI, 1.02-5.42]) and ER-negative BC in the 5 to less than 10 years post partum group (HR, 3.12 [95% CI, 1.22-7.97]) compared with the nulliparous group. Delineated by BRCA1 or BRCA2, mortality in the 5 to less than 10 years post partum group was significantly increased, but only for BRCA1 carriers (HR, 2.03 [95% CI, 1.15-3.58]).

Conclusions and Relevance

These findings suggest that YOBC with germline BRCA PVs was associated with increased risk for all-cause mortality if diagnosed within 10 years after last childbirth, with risk highest for ER-positive BC diagnosed less than 5 years post partum, and for ER-negative BC diagnosed 5 to less than 10 years post partum. BRCA1 carriers were at highest risk for poor prognosis when diagnosed at 5 to less than 10 years post partum. No such associations were observed for BRCA2 carriers. These results should inform genetic counseling, prevention, and treatment strategies for BRCA PV carriers.

CCL2-Mediated Stromal Interactions Drive Macrophage Polarization to Increase Breast Tumorigenesis

CCL2 is an inflammatory cytokine that regulates macrophage activity and is implicated in increased mammographic density and early breast tumorigenesis. The role of CCL2 in mediating stromal interactions that contribute to breast tumorigenesis has yet to be fully elucidated. THP-1-derived macrophages and mammary fibroblasts were co-cultured for 72 h. Fibroblasts and macrophages were analysed for phenotype, expression of inflammatory and ECM-regulatory genes and collagen production. Mice overexpressing CCL2 in the mammary glands were analysed for global gene expression by RNAseq at 12 weeks of age. These mice were cross-bred with PyMT mammary tumour mice to examine the role of CCL2 in tumorigenesis. The co-culture of macrophages with fibroblasts resulted in macrophage polarization towards an M2 phenotype, and upregulated expression of CCL2 and other genes associated with inflammation and ECM remodelling. CCL2 increased the production of insoluble collagen by fibroblasts. A global gene expression analysis of CCL2 overexpressing mice revealed that CCL2 upregulates cancer-associated gene pathways and downregulates fatty acid metabolism gene pathways. In the PyMT mammary tumour model, CCL2 overexpressing mice exhibited increased macrophage infiltration and early tumorigenesis. Interactions between macrophages and fibroblasts regulated by CCL2 can promote an environment that may increase breast cancer risk, leading to enhanced early tumorigenesis.

Isogenic Mammary Models of Intraductal Carcinoma Reveal Progression to Invasiveness in the Absence of a Non-Obligatory In Situ Stage

In breast cancer, progression to invasive ductal carcinoma (IDC) involves interactions between immune, myoepithelial, and tumor cells. Development of IDC can proceed through ductal carcinoma in situ (DCIS), a non-obligate, non-invasive stage, or IDC can develop without evidence of DCIS and these cases associate with poorer prognosis. Tractable, immune-competent mouse models are needed to help delineate distinct mechanisms of local tumor cell invasion and prognostic implications. To address these gaps, we delivered murine mammary carcinoma cell lines directly into the main mammary lactiferous duct of immune-competent mice. Using two strains of immune-competent mice (BALB/c, C57BL/6), one immune-compromised (severe combined immunodeficiency; SCID) C57BL/6 strain, and six different murine mammary cancer cell lines (D2.OR, D2A1, 4T1, EMT6, EO771, Py230), we found early loss of ductal myoepithelial cell differentiation markers p63, α-smooth muscle actin, and calponin, and rapid formation of IDC in the absence of DCIS. Rapid IDC formation also occurred in the absence of adaptive immunity. Combined, these studies demonstrate that loss of myoepithelial barrier function does not require an intact immune system, and suggest that these isogenic murine models may prove a useful tool to study IDC in the absence of a non-obligatory DCIS stage-an under-investigated subset of poor prognostic human breast cancer.

Abstract 4231: The anti-cancer effects of vitamin D are blocked postpartum due to suppression of vitamin D metabolism in the liver

Postpartum breast involution is a physiologic inflammatory process that associates with increased risk of postpartum breast cancer (PPBC). In rodents, involution promotes breast cancer progression, and treatment with anti-inflammatory agents is protective. These data provide rationale for targeting the pro-tumor window of involution with anti-inflammatory agents for PPBC prevention. Vitamin D is another anti-inflammatory agent with a good safety profile. In postpartum women, vitamin D deficiency is prevalent, suggesting vitamin D supplementation during involution may prevent development of PPBC.

To model vitamin D deficiency and supplementation, BALB/c mice were fed diets deficient or supplemented with vitamin D for 4 weeks. Blood was then collected from nulliparous (never-pregnant) and 2 days post-wean (involution) mice, and serum vitamin D (i.e. 25(OH)D) measured. As expected, nulliparous mice fed a vitamin D supplemented diet showed a >2-fold increase in serum 25(OH)D (67.4±8.1nmol/L) compared to mice fed a vitamin D deficient diet (28.7±11.7nmol/L, p<0.01). In contrast, involution mice fed a vitamin D supplemented diet did not show increased serum 25(OH)D (supplemented 46.6±8.7nmol/L vs deficient 33.7±8.5nmol/L, p>0.05).

Activation of dietary vitamin D requires hydroxylation in the liver, producing both circulating (25(OH)D) and active (1,25(OH)2D) forms. Thus, reduced serum 25(OH)D in vitamin D supplemented involution mice could be due to impaired vitamin D hydroxylation in the post-wean liver. Analysis of livers collected from involution and nulliparous mice found reduced expression of genes involved in vitamin D hydroxylation (Cyp2r1, Cyp27a1) during involution. Further, liver concentrations of 1,25(OH)2D were reduced 3-fold in involution mice (0.31±0.29ng/g) compared to nulliparous mice (0.90±0.37ng/g, p=0.04). Together, these findings suggest that impaired metabolism of vitamin D in the postpartum liver may reduce bioavailability of active vitamin D, and provide corroborative evidence for weaning-induced liver involution.

To understand how suppressed vitamin D metabolism in the postpartum liver affects the anti-cancer activity of dietary vitamin D, mammary cancer cells (D2A1, 2 × 104 cells, 20µL) were injected into mammary fatpad of involution or nulliparous BALB/c mice, and tumor growth tracked for 4 weeks. In nulliparous mice, vitamin D supplementation associated with a 3-fold reduction in tumor growth (p=0.03); whereas no anti-cancer effects were observed in involution mice. Together, these findings suggest the anti-cancer effects of vitamin D are blocked postpartum, likely due to suppression of vitamin D metabolism in the liver. Understanding how postpartum liver metabolism influences efficacy of preventative agents that require liver metabolism (such as vitamin D) is required to optimize cancer prevention strategies that target this window.

Citation Format: Sarah M. Bernhardt, Pepper Schedin. The anti-cancer effects of vitamin D are blocked postpartum due to suppression of vitamin D metabolism in the liver. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4231.

Abstract B011: The anti-cancer effects of vitamin D are blocked postpartum, due to suppression of vitamin D metabolism in the involuting liver

Postpartum mammary gland involution is a physiologic window of increased breast cancer risk. It has been proposed that the poor prognosis associated with postpartum breast cancer is due to the involuting mammary microenvironment promoting progression of indolent lesions to invasive disease. As such, the involuting gland has been implicated as a target for preventive strategies. Vitamin D has anti-cancer properties, and there are data demonstrating that vitamin D supplementation protects against breast cancer progression in mouse models. Moreover, vitamin D deficiency is prevalent in postpartum women, suggesting that vitamin D supplementation during the vitamin D-deficient, at-risk window of involution may be an approach for preventing the progression of indolent lesions. Here, we characterized how vitamin D deficiency and supplementation affect tumor growth in a mouse model of postpartum breast cancer. Vitamin D deficiency and sufficiency were established in BALB/c mice through feeding diets containing low or high levels of vitamin D. Quantification of serum 25(OH)D verified that diets resulted in vitamin D deficiency (25.8±4.3nmol/L;mean±stdev) or sufficiency (66.4±11.7nmol/L) at levels comparable to humans. Murine mammary cancer cells (D2A1, 2×104 cells, 20µL) were injected into mammary fatpads of involuting (2 days post wean), or age-matched nulliparous mice (n=5-12 mice/group), and tumor growth tracked for 4 weeks. Independent of vitamin D status, tumors exposed to the involuting mammary gland grew 1.8-fold larger than tumors exposed to the nulliparous gland (p<0.001); consistent with prior data showing that the microenvironment of the involuting gland is tumor promotional. Interestingly, while vitamin D supplementation of nulliparous mice associated with a 3.4-fold reduction in tumor growth (p=0.03), vitamin D supplementation of involution mice did not reduce tumor growth. Dietary vitamin D is metabolized to its active form in the liver. Our group has previously shown that the liver also undergoes weaning-induced involution. Thus, we tested if involution of the liver impairs metabolism of vitamin D to its active form, potentially contributing to the null effects of vitamin D supplementation in involution mice. We collected serum from mice on vitamin D deficient and supplemented diets across reproductive time points (nulliparous, lactation, involution days 2, 4, 6), and found that vitamin D supplementation during involution was insufficient to restore serum vitamin D to levels of sufficiency. Further, gene expression analysis of livers show that expression of Cyp2r1 and Cyp27a1—genes involved in vitamin D activation—were reduced during involution. Together, these findings suggest that impaired metabolism of vitamin D during involution may reduce the availability of active vitamin D, and contribute to the null effect observed in involution mice. Understanding the mechanisms by which mammary gland involution influences the anti-cancer effects of vitamin D is required to optimize cancer prevention strategies that target this window.

Citation Format: Sarah M Bernhardt, Pepper Schedin. The anti-cancer effects of vitamin D are blocked postpartum, due to suppression of vitamin D metabolism in the involuting liver [abstract]. In: Proceedings of the AACR Special Conference on Rethinking DCIS: An Opportunity for Prevention?; 2022 Sep 8-11; Philadelphia, PA. Philadelphia (PA): AACR; Can Prev Res 2022;15(12 Suppl_1): Abstract nr B011.

Abstract IA034: Compromised myoepithelial cell differentiation correlates with DCIS to IDC transition

DCIS, breast cancer confined to mammary ducts, shares genomic and transcriptomic signatures with IDC. However for DCIS patients, standard of care for localized breast cancer remains controversial, as treatment reduces local recurrence but does not reduce risk of distant recurrence or mortality. Further, as the risk of dying from DCIS is ~5-20 fold less than from early stage breast cancer, questions remain concerning treatment harms. The fact that treatment does not reduce DCIS death rates provides indirect evidence that DCIS-stage tumor cells can disseminate and serve as reservoirs for metastatic disease. Since presence of myoepithelium is the clinical determinate of non-invasive disease, death from DCIS also implies that myoepithelial cell loss is not required for tumor cell dissemination. Here, we investigate whether changes in myoepithelial cell differentiation, rather than overt cell loss, associate with DCIS tumor cell dissemination. Using murine DCIS models and human DCIS tissue, our group reported sequential loss of myoepithelial cell differentiation markers p63, calponin, and SMA prior to progression to IDC. In human DCIS lesions delineated into low and high risk (pure DCIS vs mixed DCIS with IDC), loss of myoepithelial p63, calponin and SMA occurred more frequently in high risk DCIS lesions. Mice transgenic for calponin knockdown in the myoepithelium were at increased risk for mammary cancer. Combined, these data implicate loss in myoepithelial cell differentiation as a mediator of DCIS progression and dissemination. In human DCIS, compromised myoepithelium correlates with microinvasion and an immune infiltrate enriched in PD1+CD8 T cells. To address the role of immune cells in myoepithelial cell barrier function, we investigated 6 different immune competent mouse models of intraductal cancer (DCIS). While loss of myoepithelial differentiation markers were evident, IDC lesions developed in the absence of DCIS. Similarly, in immune compromised hosts, only IDC lesions were detected after intraductal injection of murine tumor cells, whereas human tumor cells formed stable DCIS lesions. In vitro, murine mammary tumor cells were more motile than human lines, and when tumor cells were co-cultured with myoepithelial cells, fewer myoepithelial to myoepithelial cell contacts were observed in the presence of the mouse tumor cells. Finally, in women, we find the differentiation state of mammary myoepithelium to be modulated by reproductive state, with myoepithelium during weaning–induced gland involution (an established risk window for breast cancer progression) sharing similarities with myoepithelium of high-risk DCIS lesions. In rodents, recently weaned hosts support increased DCIS to IDC transition. In sum, these studies implicate intrinsic and extrinsic regulators of myoepithelial cells, highlight the need for immune competent mouse models of DCIS, and raise the question of whether species mismatch in xenograft models impacts cellular crosstalk between myoepithelial and tumor cells critical to our understanding of the DCIS to IDC transition.

Citation Format: Sarah M. Bernhardt, Elizabeth Mitchell, Tanya Russell, Sonali Jindal, Jayasri Narasimhan, Reuben J. Hoffmann, AeSoon Benson, Pepper Schedin. Compromised myoepithelial cell differentiation correlates with DCIS to IDC transition [abstract]. In: Proceedings of the AACR Special Conference on Rethinking DCIS: An Opportunity for Prevention?; 2022 Sep 8-11; Philadelphia, PA. Philadelphia (PA): AACR; Can Prev Res 2022;15(12 Suppl_1): Abstract nr IA034.

Vitamin D as a Potential Preventive Agent For Young Women's Breast Cancer

Clinical studies backed by research in animal models suggest that vitamin D may protect against the development of breast cancer, implicating vitamin D as a promising candidate for breast cancer prevention. However, despite clear preclinical evidence showing protective roles for vitamin D, broadly targeted clinical trials of vitamin D supplementation have yielded conflicting findings, highlighting the complexity of translating preclinical data to efficacy in humans. While vitamin D supplementation targeted to high-risk populations is a strategy anticipated to increase prevention efficacy, a complimentary approach is to target transient, developmental windows of elevated breast cancer risk. Postpartum mammary gland involution represents a developmental window of increased breast cancer promotion that may be poised for vitamin D supplementation. Targeting the window of involution with short-term vitamin D intervention may offer a simple, cost-effective approach for the prevention of breast cancers that develop postpartum. In this review, we highlight epidemiologic and preclinical studies linking vitamin D deficiency with breast cancer development. We discuss the underlying mechanisms through which vitamin D deficiency contributes to cancer development, with an emphasis on the anti-inflammatory activity of vitamin D. We also discuss current evidence for vitamin D as an immunotherapeutic agent and the potential for vitamin D as a preventative strategy for young woman's breast cancer.

Ovarian cycle stage critically affects 21-gene recurrence scores in Mmtv-Pymt mouse mammary tumours

BACKGROUND

The Oncotype DX 21-gene Recurrence Score is predictive of adjuvant chemotherapy benefit for women with early-stage, estrogen receptor (ER)-positive, HER2-negative breast cancer. In premenopausal women, fluctuations in estrogen and progesterone during the menstrual cycle impact gene expression in hormone-responsive cancers. However, the extent to which menstrual cycling affects the Oncotype DX 21-gene signature remains unclear. Here, we investigate the impact of ovarian cycle stage on the 21-gene signature using a naturally cycling mouse model of breast cancer.

METHODS

ER-positive mammary tumours were dissected from naturally cycling Mmtv-Pymt mice at either the estrus or diestrus phase of the ovarian cycle. The Oncotype DX 21-gene signature was assessed through quantitative real time-PCR, and a 21-gene experimental recurrence score analogous to the Oncotype DX Recurrence Score was calculated.

RESULTS

Tumours collected at diestrus exhibited significant differences in expression of 6 Oncotype DX signature genes (Ki67, Ccnb1, Esr1, Erbb2, Grb7, Bag1; p ≤ 0.05) and a significant increase in 21-gene recurrence score (21.8 ± 2.4; mean ± SEM) compared to tumours dissected at estrus (15.5 ± 1.9; p = 0.03). Clustering analysis revealed a subgroup of tumours collected at diestrus characterised by increased expression of proliferation- (p < 0.001) and invasion-group (p = 0.01) genes, and increased 21-gene recurrence score (p = 0.01). No correlation between ER, PR, HER2, and KI67 protein abundance measured by Western blot and abundance of mRNA for the corresponding gene was observed, suggesting that gene expression is more susceptible to hormone-induced fluctuation compared to protein expression.

CONCLUSIONS

Ovarian cycle stage at the time of tissue collection critically affects the 21-gene signature in Mmtv-Pymt murine mammary tumours. Further studies are required to determine whether Oncotype DX Recurrence Scores in women are similarly affected by menstrual cycle stage.

Comparison of hormone-induced mRNA and protein biomarker expression changes in breast cancer cells

PURPOSE

Protein biomarkers estrogen receptor (ER), progesterone receptor (PR), and marker of proliferation (Ki67) are routinely assessed by immunohistochemistry to guide treatment decisions for breast cancer. Now, quantification of mRNA encoding these proteins is being adopted in the clinic. However, mRNA and protein biomarkers may be differentially regulated by fluctuations in estrogen and progesterone that occur across the menstrual cycle in premenopausal breast cancer patients. This study aimed to compare how estrogen and progesterone affect mRNA and protein biomarker expression in hormone-responsive breast cancer cells.

METHODS

Hormone-responsive ZR-75-1 and T-47D human breast cancer cell lines were xenografted into the mammary fat pad of BALB/c nude mice supplemented with estrogen. Progesterone or vehicle was administered prior to dissection of tumors. Protein expression of ER, PR and Ki67 was quantified by immunohistochemistry, and mRNA encoding these proteins, ESR1, PGR and KI67, respectively, was quantified by real-time PCR. mRNA expression was also quantified in breast cancer cell lines treated with estrogen and progesterone in vitro.

RESULTS

In T-47D-xenografted tumors, estrogen and progesterone treatment reduced PGR and KI67 mRNA expression, and reduced PR and Ki67 protein positivity, compared to estrogen treatment alone. In ZR-75-1 xenografted tumors, no significant differences in protein or mRNA biomarker expression were observed. In vitro, estrogen and progesterone co-treatment significantly reduced ESR1 and PGR mRNA expression in both T-47D and ZR-75-1 cell lines.

CONCLUSIONS

Estrogen and progesterone similarly affect mRNA and protein biomarker expression in hormone-responsive breast cancer xenografts. Further research is needed to investigate concordance between protein and mRNA biomarkers in premenopausal breast cancer.

Modern, exogenous exposures associated with altered mammary gland development: A systematic review

BACKGROUND

Many women report low milk supply as the reason for premature breastfeeding cessation. Altered mammary gland development may impact a woman's lactation ability.

OBJECTIVE

This review identifies modern exogenous exposures which alter mammary gland development during embryonic life, puberty and pregnancy.

METHODS

A systematic review was undertaken whereby Medline, CINAHL and Embase articles published from January 1, 2005 to November 20, 2020 were searched using the keywords puberty or embry* or fetal or foetal or foetus or fetus or pregnan* or gestation* AND "mammary gland development" or "breast development" or "mammary development" or "mammary gland function" or "mammary function" or "insufficient glandular tissue" or "mammary hypoplasia" or "breast hypoplasia" or "mammary gland hypoplasia" or "tubular breast*" or "tuberous breast*" or "glandular tissue" or "breast composition" or "mammary composition" or "mammary gland composition". After initial screening of 1207 records, 60 full texts were assessed for eligibility; 6 were excluded due to lack of information about exposure or outcome, leaving 54 studies.

RESULTS

The review included results from 52 animal (rats and mice, monkeys, rabbits, sheep, goats pigs and cows) and 2 human studies. Various endocrine disrupting chemicals and an obesogenic diet were found to be associated with altered mammary gland morphology during key development stages.

CONCLUSIONS

To improve lactation outcomes, future studies need to focus on lactation as the endpoint and be conducted in a standardised manner to allow for a more significant contribution to the literature that allows for better comparison across studies.

Attenuated TGFB signalling in macrophages decreases susceptibility to DMBA-induced mammary cancer in mice

BACKGROUND

Transforming growth factor beta1 (TGFB1) is a multi-functional cytokine that regulates mammary gland development and cancer progression through endocrine, paracrine and autocrine mechanisms. TGFB1 also plays roles in tumour development and progression, and its increased expression is associated with an increased breast cancer risk. Macrophages are key target cells for TGFB1 action, also playing crucial roles in tumourigenesis. However, the precise role of TGFB-regulated macrophages in the mammary gland is unclear. This study investigated the effect of attenuated TGFB signalling in macrophages on mammary gland development and mammary cancer susceptibility in mice.

METHODS

A transgenic mouse model was generated, wherein a dominant negative TGFB receptor is activated in macrophages, in turn attenuating the TGFB signalling pathway specifically in the macrophage population. The mammary glands were assessed for morphological changes through wholemount and H&E analysis, and the abundance and phenotype of macrophages were analysed through immunohistochemistry. Another cohort of mice received carcinogen 7,12-dimethylbenz(a)anthracene (DMBA), and tumour development was monitored weekly. Human non-neoplastic breast tissue was also immunohistochemically assessed for latent TGFB1 and macrophage marker CD68.

RESULTS

Attenuation of TGFB signalling resulted in an increase in the percentage of alveolar epithelium in the mammary gland at dioestrus and an increase in macrophage abundance. The phenotype of macrophages was also altered, with inflammatory macrophage markers iNOS and CCR7 increased by 110% and 40%, respectively. A significant decrease in DMBA-induced mammary tumour incidence and prolonged tumour-free survival in mice with attenuated TGFB signalling were observed. In human non-neoplastic breast tissue, there was a significant inverse relationship between latent TGFB1 protein and CD68-positive macrophages.

CONCLUSIONS

TGFB acts on macrophage populations in the mammary gland to reduce their abundance and dampen the inflammatory phenotype. TGFB signalling in macrophages increases mammary cancer susceptibility potentially through suppression of immune surveillance activities of macrophages.

Timing of breast cancer surgery during the menstrual cycle-is there an optimal time of the month?

An intriguing relationship between menstrual cycle phase at the time of breast cancer surgery and clinical outcomes was first proposed in the late 1980s. Despite a number of clinical studies conducted to address this, as well as meta-analyses and systematic reviews, there remains significant controversy surrounding the effect of menstrual cycle phase at time of surgery on the prognosis of premenopausal breast cancer. While some studies have suggested that surgery performed during the luteal phase results in the most favourable outcome, other studies report the follicular phase is more favourable, and others show no association. Given the conflicting results, there remains insufficient evidence to determine whether there is an optimal time of the month to perform surgery. This issue has dogged breast cancer surgery for decades; knowledge of an optimal time of the month to conduct surgery would be a simple approach to improving patient outcomes. This review explores the potential biological mechanisms through which the hormonal milieu might contribute to differences in prognosis, and why clinical findings are so variable. It is concluded that a significant problem with current clinical research is the lack of insight from mechanistic studies. While there are a number of plausible biological mechanisms that could lead to altered survival, supporting evidence is limited. There are also variable approaches to defining the menstrual cycle phase and hormone receptor status of the tumour and few studies controlled for prognostic factors such as tumour size and stage, or addressed the impact of adjuvant treatments. Elucidation of the specific confounding factors, as well as biological mechanistic pathways that could explain the potential relationship between timing of surgery and survival, will greatly assist in designing robust well-controlled prospective clinical studies to evaluate this paradigm.

Abstract P1-10-12: Menstrual cycling critically affects the Oncotype DX 21-gene signature: Implications for predictive biomarker assays in premenopausal women

Introduction: The 21-gene Oncotype DX Recurrence Score is predictive of adjuvant chemotherapy benefit for women with early stage, hormone receptor (HR)-positive, HER2-negative breast cancer. In premenopausal women, fluctuations in estrogen and progesterone during the menstrual cycle can impact gene expression in these hormone-responsive cancers. However, the extent to which hormonal fluctuations affect Oncotype DX Recurrence Scores remains unclear.

Aim: To determine the extent to which ovarian cycling affects the Oncotype DX 21-gene signature using paired premenopausal breast cancer samples and mouse models.

Methods: To investigate menstrual variation in the Oncotype DX 21-gene signature within the same tumour, paired formalin-fixed paraffin-embedded, HR-positive invasive breast cancer samples were collected on different days of the menstrual cycle from women &lt;50 years old (n=18), and compared to women &gt;50 years old (n=11). Samples were collected an average of 18 days apart and were neoadjuvant therapy naive. To determine the effect of progesterone on gene expression, HR-positive T47D breast cancer cells were xenografted into the mammary fat pad of BALB/c nude mice, treated with exogenous estrogen±progesterone (n=12, 11 respectively). Additionally, HR-positive mammary tumours were collected from naturally cycling Mmtv-Pymt mice at the estrus or diestrus phase of the ovarian cycle (n=25, 28 respectively). The 21-gene Oncotype DX signature was assessed through quantitative RT-PCR and an experimental recurrence score (RS) was calculated using the Oncotype DX Recurrence Score algorithm.

Results: Increased discordance in RS was observed between paired samples collected from younger women (3.2±2.5; mean±stdev), compared to older women (2.0±1.7; p=0.04). In young women, discordance was primarily driven by variable expression of proliferative genes, compared to older women, where discordances were a result of variable expression of invasive genes. Variable concentrations of progesterone at the time of tissue collection may influence proliferative gene expression and contribute to discordant RS. In support of this, in HR-positive xenograft tumours, expression of proliferative genes Ki67 (p=0.03) and STK15 (p=0.04), and Ki67 protein expression (p=0.02), were reduced following progesterone treatment. Furthermore, in naturally cycling Mmtv-Pymt mice, mammary tumours collected at diestrus, analogous to the luteal phase of the menstrual cycle in humans, show significant differences in expression of 6 Oncotype DX signature genes (Ki67, Ccnb1, Esr1, Her2, Grb7, Bag1; p≤0.05) and a significant increase in RS (21.1±2.4; mean±SEM) compared to tumours dissected at estrus (15.5±1.9; p=0.03). Clustering analysis revealed a subgroup of Mmtv-Pymt mammary tumours collected at diestrus characterised by increased expression of proliferative (p&lt;0.001) and invasive (p=0.01) genes and a significant increase in RS (p=0.01). These tumours also exhibited higher expression of estrogen regulated genes (p=0.005) suggesting increased sensitivity to hormonal fluctuations during the ovarian cycle, and possibly greater variability in RS.

Conclusion: Our results suggest that menstrual cycling affects the expression of genes included in the Oncotype DX 21-gene signature and influences experimental recurrence scores. Oncotype DX may be less effective for guiding chemotherapy treatment decisions for cycling premenopausal women compared to older postmenopausal women.

Citation Format: Sarah M Bernhardt, Pallave Dasari, Danielle J Glynn, Lucy Woolford, Wendy Raymond, Lachlan M Moldenhauer, David Walsh, Amanda R Townsend, Timothy J Price, Wendy V Ingman. Menstrual cycling critically affects the Oncotype DX 21-gene signature: Implications for predictive biomarker assays in premenopausal women [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P1-10-12.

Hormonal Modulation of Breast Cancer Gene Expression: Implications for Intrinsic Subtyping in Premenopausal Women

Clinics are increasingly adopting gene-expression profiling to diagnose breast cancer subtype, providing an intrinsic, molecular portrait of the tumor. For example, the PAM50-based Prosigna test quantifies expression of 50 key genes to classify breast cancer subtype, and this method of classification has been demonstrated to be superior over traditional immunohistochemical methods that detect proteins, to predict risk of disease recurrence. However, these tests were largely developed and validated using breast cancer samples from postmenopausal women. Thus, the accuracy of such tests has not been explored in the context of the hormonal fluctuations in estrogen and progesterone that occur during the menstrual cycle in premenopausal women. Concordance between traditional methods of subtyping and the new tests in premenopausal women is likely to depend on the stage of the menstrual cycle at which the tissue sample is taken and the relative effect of hormones on expression of genes versus proteins. The lack of knowledge around the effect of fluctuating estrogen and progesterone on gene expression in breast cancer patients raises serious concerns for intrinsic subtyping in premenopausal women, which comprise about 25% of breast cancer diagnoses. Further research on the impact of the menstrual cycle on intrinsic breast cancer profiling is required if premenopausal women are to benefit from the new technology of intrinsic subtyping.