Endocrine-targeting therapies shift the breast microbiome to reduce estrogen receptor-α breast cancer risk

Summary

Studies indicate that breast tissue has a distinct modifiable microbiome population. We demonstrate that endocrine-targeting therapies, such as tamoxifen, reshape the non-cancerous breast microbiome to influence tissue metabolism and reduce tumorigenesis. Using 16S sequencing, we found that tamoxifen alters β-diversity and increases Firmicutes abundance, including Lactobacillus spp., in mammary glands (MGs) of mice and non-human primates. Immunohistochemistry showed that lipoteichoic acid (LTA)-positive bacteria were elevated in tamoxifen-treated breast tissue. In B6.MMTV-PyMT mice, intra-nipple probiotic bacteria injections reduced tumorigenesis, altered metabolic gene expression, and decreased tumor proliferation. Probiotic-conditioned media selectively reduced viability in estrogen receptor-positive (ER+) breast cancer cells and altered mitochondrial metabolism in non-cancerous epithelial cells. Human tumor samples revealed that LTA-positive bacteria negatively correlated with Ki67, suggesting that endocrine therapies influence tumor-associated microbiota to regulate proliferation. Our data indicate that endocrine-targeting therapies modify the breast microbiome, corresponding with a shift in tissue metabolism to potentially reduce ER+ breast cancer risk.

Graphical abstract

Highlights

• •Tamoxifen shifts breast tissue-resident microbiome populations
• •Elevating mammary gland Lactobacillus reduces tumorigenesis in the B6.MMTV-PyMT model
• •Mammary gland probiotic bacteria reshape tissue metabolism favoring glycolysis
• •LTA-positive bacteria inversely correlate with proliferation in ER+ tumors

Arnone et al. reveal that tamoxifen alters breast tissue microbiota, increasing commensal Lactobacillus and Streptococcus abundance. Elevating mammary gland probiotic species reduces tumor burden and improves tumor-free survival. Findings suggest that breast tissue-resident microbiota is influenced by endocrine-targeting therapies, shifts non-cancerous tissue metabolism, and may reduce breast cancer risk.

Introduction

Breast cancer is the most commonly diagnosed type of cancer in women, and despite significant progress in diagnosis and treatment, there are still more than 40,000 deaths per year.¹ Hormone receptor-positive breast cancer expresses estrogen receptor-α (ER) and/or progesterone receptor and accounts for 60%–70% of all breast cancer diagnoses.² Several risk factors are related to breast cancer development, including genetics, hormone replacement therapy, lifestyle, reproductive history, and diet.³ The microbiome, which represents the totality of microorganisms living on and within the body, is implicated as a risk factor for breast cancer.^4,5 Furthermore, the breast has a distinct microbiome, which may be a potential modulator of tumorigenesis and therapeutic response.^6,7,8

Analysis of breast tissue collected from women with and without breast cancer revealed diverse populations of bacteria, including Bacillus, Acinetobacter, Enterobacteriaceae, Pseudomonas, Staphylococcus, Propionibacterium, and Prevotella.⁶ Dysbiotic breast microbiota and microbial-associated molecular patterns influence chronic localized inflammation, inflammation-mediated carcinogenesis processes, immune invasion of the breast tissue, and genetic instability contributing to malignant progression.^9,10,11 Microbial differences in the breast tissue of patients with breast cancer as compared to non-cancerous controls have been observed.^7,9 Urbaniak et al. found higher abundances of Enterobacteriaceae and Staphylococcus in patients with breast cancer as compared to non-cancer controls, which are associated with DNA double-strand breaks.⁷ Differences in breast microbial diversity between women with benign versus malignant breast cancer have also been identified.¹² Alterations in the microbial community suggest that certain tissue-resident microbes are associated with the development of cancers.^9,10

Alterations in microbial diversity were observed in breast tumors compared to adjacent normal tissue, with breast tumors having a distinct microbiota population.¹² Furthermore, the presence of tumors, breast tumor receptor subtype, and tumor grade influence microbial composition.^7,13 Emerging evidence suggests that racial differences also affect intratumoral microbiota, with breast tumors from Black women overexpressing Staphylococcus compared to breast tumors from White women.¹⁴ These microbial populations have a role in both benign and malignant diseases, as the microbial composition of benign tumors is more similar to healthy tissues compared to malignant tumors.⁷ Breast tumor tissue has higher bacterial load and bacteria richness compared to adjacent normal breast tissue, with breast tumor tissue expressing a higher abundance of Proteobacteria compared to normal adjacent breast tissue, which is enriched in Actinobacteria.^15,16 These studies further suggest a potential link between tumor microbiota and cancer development.

Diet, one of the main determinants of gut microbial diversity, has been implicated in breast cancer development, and alterations in breast microbiota due to diet have been observed.^17,18,19 Mediterranean diet consumption increased breast Lactobacillus abundance compared to Western diet-fed non-human primates (NHPs), suggesting that diet affects microbiota populations outside of the gut, potentially mediating breast carcinogenesis.²⁰ Using the same cohort of NHPs, it was shown that NHPs with elevated leaky-gut parameters (regardless of dietary pattern) displayed a significantly different breast microbiome, suggesting that gut health influences the breast microbiome.²¹ In a carcinogen-induced murine model, fecal microbiota transplantation from high-fat diet-fed mice to control diet-fed mice increased mammary tumorigenesis, resulting in shifted gut and mammary tumor microbiota populations¹¹. Omega-3 polyunsaturated fat supplementation in patients with breast cancer during a window-of-opportunity clinical trial decreased intratumoral Ruminococcus and Bacteroidales genus proportional abundance, further supporting the potential effects of diet on breast microbiota populations and tumorigenesis.¹¹ Obesity, which is associated with worse breast cancer-specific survival overall, also shifts the breast tumor microbiome.^22,23 Understanding breast and tumor microbiota dynamics and their associations with diet, obesity, and other lifestyle factors could have significant implications for breast cancer development and prevention.

Anti-cancer chemotherapies shift breast and tumor microbiota populations by potentially acting as a selection pressure.^23,24 Specifically, neoadjuvant and adjuvant chemotherapy influence microbiota populations in the gut, breast, and tumor.^23,24,25 These studies suggest that cancer therapeutics can shift microbiota populations to modify treatment efficacy. We currently do not know the influence of endocrine-targeted therapies on the microbiome and its impact on breast tumorigenesis.

In the current study, we found that tamoxifen (TAM) administration shifted mammary gland (MG) and breast microbiota populations in mice and ovariectomized (OVX) NHPs. Interestingly, TAM increased MG and breast Firmicutes phyla proportional abundance in both mice and NHPs. Alterations in MG microbial diversity were shifted by diet and drug administration; Western diet-fed mice administered TAM displayed increased Proteobacteria phyla proportional abundance compared to healthy diet-fed mice administered TAM. This suggests that diet-drug interactions affect microbiota diversity in the MG. At the species level, TAM administration increased several commensal probiotic Lactobacillus, Streptococcus, and Staphylococcus species in mice and OVX NHPs. Intra-nipple probiotic bacteria injections increased lipoteichoic acid (LTA)-positive bacteria, including Lactobacillus species, as well as altered metabolism-related genes and protein expression. In B6.MMTV-PyMT mice, intra-nipple injections of probiotic bacteria decreased tumor multiplicity and increased tumor-free survival associated with a decrease in tumor proliferation and shifted glucose metabolism, suggesting that breast-specific bacteria illicits anti-tumor effects. We also demonstrate that probiotic conditioned media (Pro-CM) significantly reduced ER+ breast cancer cell growth, but not non-neoplastic mammary epithelial cell growth in vitro. Additionally, we demonstrate that probiotic-derived metabolites differentially modulate cellular bioenergetics in normal versus cancer cells. Lastly, we show that intratumoral LTA-positive bacteria localization negatively correlated with proliferation in hormone receptor-positive tumors from patients with breast cancer treated with endocrine-targeted therapies in the neoadjuvant clinical setting. Understanding the breast microbiota and its relationship with diet, obesity, and treatment is a promising avenue of research that may provide insights into breast cancer development and prevention.

Results

TAM shifts breast microbiota populations in an NHP model of menopause

Breast tissue has a distinct microbiome, with microbial diversity influenced by several factors, including diet.^6,11,24 However, the effect of TAM on breast microbiota populations is unknown. Therefore, 16S bacterial sequencing was performed on DNA isolated from breast tissue of OVX NHPs to determine the effects of TAM on breast bacterial abundance in an NHP model of menopause. Sequencing revealed that OVX TAM-administered NHPs had distinct breast microbiota populations compared to untreated OVX NHPs (Figure 1A). However, TAM administration did not significantly affect overall breast bacterial α-diversity (Figure 1B). The breast microbial populations at the phyla level are predominately derived from Firmicutes, Proteobacteria, and Bacteroidetes. The proportional abundance of these bacterial phyla in OVX NHP breast tissue samples is visualized by bar plots. Each colored box represents a bacterial phylum. The height of a color box represents the proportional abundance of that bacterial phylum within the sample (Figure 1C). TAM administration significantly increased Firmicutes phylum proportional abundance compared to the untreated OVX NHPs (Figure 1D). There was no significant difference in Proteobacteria phylum proportional abundance in breast tissue from TAM-treated NHPs (Figure 1E). Immunohistochemistry performed on NHP breast tissue displayed a significant increase in LTA-positive cells (Gram-positive bacteria marker) in TAM-treated OVX NHPs (Figure 1F), supporting the observed increase in Firmicutes proportional abundance. No significant difference was observed in OVX NHP breast tissue stained for lipopolysaccharide (LPS; Gram-negative bacteria marker) with TAM administration (Figure 1G).

Figure 1.

TAM administration modulates NHP breast microbiome phyla

(A) Principal coordinates analysis (PCoA) showing TAM-treated OVX NHPs have distinct microbiota populations compared to untreated OVX NHPs.

(B) TAM administration did not significantly change α-diversity compared to untreated OVX NHPs (n = 10–14 NHPs per group, unpaired t test with Welch’s correction).

(C) Relative proportional abundance of bacterial phyla in untreated OVX NHPs and TAM-treated OVX NHPs visualized by a bar graph. Each colored box represents a bacterial phylum. The height of a color box represents the relative abundance of that bacteria within the sample. “Other” represents lower abundance taxa.

(D and E) TAM-treated OVX NHP breast tissue displayed significantly increased Firmicutes proportional abundance (D) (n = 10–14 NHPs per group,∗p < 0.05, unpaired t test with Welch’s correction) and a trending decrease in Proteobacteria proportional abundance (E) (n = 10–14 NHPs per group, p = 0.1074, unpaired t test with Welch’s correction).

(F) TAM led to a significant increase in LTA+ cells per ductal region in OVX NHP breast tissue. Representative micrographs of LTA+ cells in untreated OVX NHP and TAM-treated OVX NHP breast tissue (n = 10–14 NHPs per group, ∗p < 0.05, unpaired t test with Welch’s correction).

(G) TAM treatment did not significantly affect LPS+ cells per ductal region in NHP breast tissue. Representative micrographs of LPS+ cells in untreated OVX NHP and TAM-treated OVX NHP breast tissue (n = 10–14 NHPs per group, unpaired t test with Welch’s correction). Scale, 100 μm. Data are shown as the mean ± SEM.

Altered breast genus and species composition were observed with TAM administration in OVX NHPs. Bar plots display the relative proportional abundance of bacterial genera in breast tissue samples (Figure 2A). At the genus level, TAM significantly increased Lactobacillus genus proportional abundance in OVX NHP breast tissue (Figure 2B). Lactobacillus, the largest genus within the group of lactic acid bacteria, is an essential probiotic bacterium.²⁶ Trending increases in Streptococcus (Figure 2C) and Staphylococcus (Figure 2D) genera proportional breast tissue abundance were also observed with TAM administration. Many Streptococcus species are not pathogenic and form part of the commensal microbiota of the mouth, skin, intestine, and upper respiratory tract.²⁷ Alterations at the species level within these genera were also observed with TAM treatment. Specifically, Lactobacillus of an unidentified species proportional abundance was significantly increased in TAM-treated OVX NHPs breast tissue (Figure 2E). Lactobacillus reuteri proportional abundance, a probiotic bacterium present in breast milk, displayed a trending increase with TAM administration (Figure 2F).²⁸ Furthermore, the proportional abundance of Streptococcus lutetiensis (Figure 2G) and Staphylococcus sciuri (Figure 2H) was significantly increased with TAM treatment. Altered OVX NHP breast microbial populations with TAM treatment suggest potential drug-bug interactions that may impact TAM efficacy.

Figure 2.

TAM administration shifts NHP breast microbiota genus and species populations

(A) Relative proportional abundance of bacterial genera in untreated OVX NHPs and TAM-treated OVX NHPs visualized by a bar graph. Each colored box represents a bacterial genus. The height of a color box represents the relative abundance of that bacteria within the sample. “Other” represents lower abundance genera.

(B–D) TAM significantly increased Lactobacillus (B) genus proportional abundance in OVX NHP breast tissue (n = 10–14 NHPs per group, ∗p < 0.05, unpaired t test with Welch’s correction) and led to a trending increase in Streptococcus (C) (n = 10–14 NHPs per group, p = 0.1303, unpaired t test with Welch’s correction) and Staphylococcus (D) genus proportional abundance (n = 10–14 NHPs per group, p = 0.0859, unpaired t test with Welch’s correction). Breast microbiota were shifted at the species level with TAM administration.

(E) TAM significantly increased Lactobacillus of an unidentified species proportional abundance in OVX NHP breast tissue (n = 10–14 NHPs per group, ∗p < 0.05, unpaired t test with Welch’s correction).

(F) Lactobacillus reuteri proportional abundance was increased with TAM (n = 10–14 NHPs per group, p = 0.0790, unpaired t test with Welch’s correction).

(G and H) Streptococcus lutetiensis (G) and Staphylococcus sciuri (H) proportional abundance were significantly increased with TAM administration in OVX NHP breast tissue (n = 10–14 NHPs per group, ∗p < 0.05, unpaired t test with Welch’s correction). Data are shown as the mean ± SEM.

TAM and diet interact to modify MG microbiota populations in a murine model

Diet and drug interactions can significantly impact the gut microbiome, leading to alterations in its composition and function.²⁹ Therefore, to elucidate the potential interactions between TAM and diet on breast microbiota, 16S bacterial sequencing was performed on DNA isolated from MG collected from mice fed either a healthy control diet or a Western diet treated with TAM. See Table S1 for murine diet information. Mouse MG microbiota populations were shifted by both diet and TAM administration (Figure S1A); however, the overall bacterial α-diversity was not significantly affected by diet or TAM (Figure S1B). The proportional abundance of bacterial phyla in MG tissue samples is visualized by bar plots (Figure 3A). Consistent with results observed in OVX NHP breast tissue, TAM administration significantly increased Firmicutes proportional abundance in healthy diet-fed mice, but not in Western diet-fed mice (Figure 3B). Additionally, Proteobacteria proportional abundance was significantly increased in the Western + TAM diet-fed mice compared to the healthy + TAM diet-fed mice (Figure 3C). The differential effects observed in the proportional abundance of these bacterial phyla suggest key diet-drug interactions that influence the mammary tissue microbiome.

Figure 3.

Mouse MG microbiota are shifted by diet and TAM

(A) Relative proportional abundance of bacterial phyla in MG of healthy control, healthy + TAM, Western control, and Western + TAM-fed mice is visualized by a bar graph. Each colored box represents a bacterial taxon. The height of a color box represents the relative abundance of that bacteria within the sample. “Other” represents lower abundance taxa.

(B) TAM administration significantly increased Firmicutes proportional abundance in healthy-fed mice MG, but not Western-fed mice MG (n = 8 mice per group, ∗p < 0.05, ordinary one-way ANOVA followed by Fisher’s LSD test).

(C) Proteobacteria proportional abundance was significantly increased in Western + TAM-fed mice compared to healthy + TAM-fed mice MG (n = 8 mice per group, ∗p < 0.05, ordinary one-way ANOVA, followed by Fisher’s LSD test).

(D) No significant differences in Lactobacillus genus proportional abundance were observed (n = 8 mice per group, ordinary one-way ANOVA, followed by Fisher’s LSD test).

(E) Streptococcus genus proportional abundance was significantly increased in healthy + TAM-fed mice MG compared to healthy control-fed mice (n = 8 mice per group, ∗p < 0.05, Kruskal-Wallis test).

(F) Staphylococcus genus proportional abundance was significantly increased in Western+ TAM-fed mice (n = 8 mice per group, ∗p < 0.05 ordinary one-way ANOVA, followed by Fisher’s LSD test).

(G) TAM significantly increased Lactobacillus of an unidentified species proportional abundance in healthy + TAM-fed mice compared to healthy control (n = 8 mice per group, ∗p < 0.05, ordinary one-way ANOVA, followed by Fisher’s LSD test).

(H and I) Lactobacillus casei (H) and Lactobacillus zeae (I) proportional abundance displayed trending increases in Western + TAM mice (n = 8 mice per group, ordinary one-way ANOVA, followed by Fisher’s LSD test).

(J) Streptococcus lutetiensis proportional abundance was significantly increased in healthy + TAM mice compared to healthy control (n = 8 mice per group, ∗p < 0.05, Kruskal-Wallis test).

(K) Staphylococcus sciuri proportional abundance was significantly increased in Western + TAM-fed mice compared to Western control. (n = 8 mice per group, ∗p < 0.05, ordinary one-way ANOVA, followed by Fisher’s LSD test). Data are shown as the mean ± SEM.

See also Figure S1.

As with OVX NHPs, TAM treatment significantly shifted MG bacterial abundance at both the genus and species level in mice. TAM administration had no overall effect on MG Lactobacillus genus proportional abundance (Figure 3D). Streptococcus genus proportional abundance was significantly increased with TAM in healthy diet-fed mice, which was not observed in Western diet-fed mice (Figure 3E). Conversely, Western + TAM-fed mice displayed elevated Staphylococcus genus proportional abundance (Figure 3F). Alterations in microbial composition were also observed at the species level. TAM administration significantly increased Lactobacillus of an unidentified species proportional abundance in healthy control + TAM-fed mice (Figure 3G). Additionally, probiotic bacteria Lactobacillus casei (Figure 3H) and Lactobacillus zeae (Figure 3I) proportional abundance displayed trending increases in Western + TAM-fed mice. Streptococcus lutetiensis proportional abundance was elevated in healthy + TAM-fed mice MG (Figure 3J), while Staphylococcus sciuri proportional abundance was significantly increased in Western + TAM-fed mice MG (Figure 3K).

MG probiotic bacteria impact breast cancer risk

Probiotics, such as lactic acid-producing bacteria Lactobacillus and Bifidobacterium, are live, nonpathogenic microorganisms typically administered to improve microbial composition.³⁰ The effects of probiotic bacteria colonization of the gut have been extensively studied; however, probiotic bacteria colonization of the breast tissue remains underexplored. To explore this, control diet-fed non-tumor-bearing BALB/c mice received intra-nipple injections of probiotic Lactobacillus and Bifidobacterium bacteria (study schematic Figure 4A). To specifically determine which species were incorporated into the MG of BALB/c mice that received intra-nipple probiotic injections, 16S sequencing was performed on DNA isolated from snap-frozen MG (Figure 4B). Analysis revealed an enrichment in L. casei strain 1316, L. zeae, and an unspecified species of Lactobacillus (Figures 4C–4E). 16S sequencing on aerobic live cultured MG displayed similar changes in microbiota, with enrichment in L. plantarum (7.7% proportional abundance), L. zeae (12.3% proportional abundance), and Lactobacillus unspecified (6.4% proportional abundance; Figure 4F) in samples from probiotic-injected MG. We also performed quantitative reverse-transcription PCR (RT-qPCR) to confirm presence of species and abundance of each species in the probiotic, in which L. paracasei was 36.78%, Bifidobacterium lactis was 35.36%, L. acidophilus was 17.92%, L. plantarum was 6.85%, L. rhamnoses was 2.12%, Bifidobacterium bifidum was 0.29%, and Bifidobacterium breve was 0.68% (Figure S2A). RT-qPCR of 16S and Lactobacillus species-specific primers further show the incorporation of bacteria into the MG on day 14 post injection. A significant increase in 16S rRNA gene expression was observed in the BALB/c MG 14 days post probiotic intra-nipple injection (Figure 4G). L. paracasei and L. rhamnoses gene expression were present as part of the normal MG microbiome and also expressed on day 14 post injection (Figures 4H and 4I). While no significant differences in L. acidophilus were observed (Figure 4J), L. plantarum was significantly enriched in MG tissue by day 14 post injection (Figure 4K). While B. lactis was the second most abundant bacterium present in the probiotic, there were no significant differences in B. lactis gene expression from day 0 to day 14 post probiotic injection, potentially due to the oxygen availability within the MG (Figure S2B). Immunohistochemistry on paraffin-embedded MG revealed elevated LTA-positive bacteria from probiotic bacteria-injected mice on day 14 post injection (Figure 4L). Reduced LPS-positive bacteria were observed in probiotic-injected mice MG on day 7 and 14 post injection (Figure S3A). Interestingly, no significant differences in F4/80+ macrophages were observed with probiotic injection (Figure S3B), detracting from the potential impact of immune-mediated effects.

Figure 4.

Probiotic bacteria administration shifts MG microbiota and cancer risk

(A) Non-tumor-bearing BALB/c probiotic intra-nipple injection study schematic.

(B) Proportional abundance of bacterial species in saline-injected and probiotic-injected snap-frozen MG visualized by a bar graph. Each colored box represents a bacterial taxon. The height of a color box represents the relative abundance of that bacteria within the sample. “Other” represents lower abundance taxa (n = 8–12 mice per group).

(C) Probiotic bacteria injection significantly increased L. casei strain 1316 proportional abundance in BALB/c MG (n = 8–12 mice per group, ∗p < 0.05, unpaired t test).

(D) Probiotic bacteria injection significantly increased L. zeae proportional abundance in BALB/c MG (n = 8–12 mice per group, ∗p < 0.05, unpaired t test).

(E) MG unspecified Lactobacillus species proportional abundance was significantly increased with probiotic bacteria injection in BALB/c mice (n = 8–12 mice per group, ∗p < 0.05, unpaired t test).

(F) Proportional abundance of bacterial species in saline-injected and probiotic-injected MG with aerobic live culture is visualized by a bar graph. Each colored box represents a bacterial taxon. The height of a color box represents the relative abundance of that bacteria within the sample. “Other” represents lower abundance taxa (n = 3–5 mice per group).

(G) 16S rRNA gene expression from non-tumor-bearing BALB/c MG that received saline or probiotic bacteria intra-nipple injections (n = 5–6 mice per group, ∗p < 0.05, ordinary one-way ANOVA, followed by Fisher’s LSD test).

(H–J) L. paracasei (H), L. rhamnoses (I), and L. acidophilus (J) gene expression are detectable on day 0 and not significantly modulated by probiotic bacteria injection at day 14 (n = 5–6 mice MG per group).

(K) L. plantarum gene expression was significantly increased on day 14 post probiotic injection (n = 5 mice MG per group, ∗p < 0.05, unpaired t test).

(L) LTA staining in control (day 0; n = 6 mice MG) and probiotic-injected MGs on days 7 and 14 post injection (n = 4 mice MG per day, ∗p < 0.05, ordinary one-way ANOVA).

(M) B6.MMTV-PyMT treatment schematic.

(N) Probiotic bacteria significantly improved tumor-free survival in Western diet-fed mice (n = 6–7 mice per group, ∗p = 0.01, log rank Mantel-Cox test).

(O) Tumor multiplicity was significantly decreased in probiotic-treated mice (n = 6–7 mice per group, ∗p < 0.05, unpaired t test).

(P) L. acidophilus gene expression in saline and probiotic bacteria-injected B6.MMTV-PyMT MG (n = 4–6 mice MG per group, unpaired t test).

(Q) Probiotic bacteria injection decreased tumor proliferation assessed by Ki67 (n = 7 mouse tumors per group, ∗p < 0.05, unpaired t test). Representative micrographs of Ki67 in saline-injected and probiotic-injected B6.MMTV-PyMT mouse mammary tumors.

(R) Probiotic bacteria injection led to an increase in LTA positivity in probiotic-treated mice (n = 6–13 mouse tumors per group, ∗p < 0.05, unpaired t test). Representative micrographs of LTA immunoreactivity in saline-injected and probiotic bacteria-injected B6.MMTV-PyMT mouse mammary tumors.

(S) No changes in LPS positivity with probiotic bacteria injection were observed in B6.MMTV-PyMT mammary tumors (n = 6–11 mouse tumors per group). Representative micrographs of LPS immunoreactivity in saline-injected and probiotic bacteria-injected B6.MMTV-PyMT mouse mammary tumors.

(T) Probiotic bacteria injection significantly increased F4/80+ macrophages in B6.MMTV-PyMT mouse mammary tumors (n = 7 mouse tumors per group, ∗p < 0.05, unpaired t test). Representative micrographs of F4/80+ macrophages in saline-injected and probiotic-injected B6.MMTV-PyMT mouse mammary tumors. Data are shown as the mean ± SEM. Scale, 100 μm

See also Figures S2–S4.

To further determine the effects of localized probiotic bacteria on breast cancer development, Western diet-fed B6.MMTV-PyMT mice received intra-nipple injections of probiotic bacteria at 5, 7, 9, and 11 weeks of age. Tumor development was monitored for an additional 5 weeks post last injection. We observed a reduced relative proportional abundance of Lactobacillus with Western diet in our non-tumor-bearing model; therefore, we wanted to investigate the effects of probiotic bacteria supplementation on tumorigenesis when MG Lactobacillus abundance was low. The treatment schematic is shown in Figure 4M. Probiotic bacteria intra-nipple injection significantly increased tumor-free survival (Figure 4N) and reduced tumor multiplicity in B6.MMTV-PyMT mice (Figure 4O). L. paracasei, L. reuteri, L. rhamnoses, and L. plantarum were detected in MG but not significantly regulated by treatment at the end of study (Figures S4A–S4D); however, there was a significant increase in L. acidophilus with probiotics in the non-tumor-bearing MG tissue of B6.MMTV-PyMT mice (Figure 4P). Tumor proliferation, assessed by Ki67, was significantly decreased in probiotic bacteria-treated mice (Figure 4Q). Probiotic bacteria injections lead to elevated intratumoral LTA-positive bacteria (Figure 4R). No significant difference in intratumoral LPS positivity was observed (Figure 4S). Probiotic bacteria injections lead to elevated intratumoral F4/80+ macrophages (Figure 4T), potentially contributing to the reduced tumorigenesis observed in B6.MMTV-PyMT mice.

MG probiotic bacteria shift metabolism

Studies suggest that probiotics and their metabolites could regulate metabolic processes to reduce disease progression.³¹ RNA sequencing revealed 600 differentially regulated genes in the MG of Western diet-fed mice injected with probiotic bacteria (Figure 5A), including upregulating various metabolism-related genes in MG tissue such as glucokinase (3.24 log2FoldChange, p = 0.001), hexokinase 2 (1.42 log2FoldChange, p = 0.028), lactate dehydrogenase A (1.25 log2FoldChange, p = 0.014), and protein kinase AMP-activated alpha 2 catalytic subunit (1.84 log2FoldChange, p = 0.044), among others. Kyoto Encyclopedia of Genes and Genomes (KEGG) gene enrichment highlights multiple metabolism pathways (glucagon, central carbon metabolism, glycolysis/gluconeogenesis, and insulin metabolism) modulated by probiotic bacteria in MG tissue (Figure 5B). Untargeted metabolomics performed on non-cancerous MG collected from saline and probiotic-treated B6.MMTV-PyMT mice displayed significantly different metabolite patterns as indicated by principal component analysis (Figure 5C). Enrichment analysis indicates that probiotics increased tricarboxylic acid (TCA) cycle and glycolysis metabolic biomolecules (Figure 5D). While no significant differences were observed in glucose metabolite concentrations with probiotic injection (Figure 5E), significant increases in glycolytic metabolites glucose-6-phosphate (Figure 5F) and fructose 1,6-bisphosphate (F-1,6-BP; Figure 5G) were observed. There was a trend for increased 3-phosphoglycerate (Figure 5H). Phosphoenolpyruvate (PEP; Figure 5I) was significantly increased while pyruvate (Figure 5J) was unchanged with probiotic treatment. Lactate was significantly reduced in MG tissue from probiotic bacteria-injected mice (Figure 5K). Probiotic injection also significantly increased pyruvate kinase M-1/2 protein levels in B6.MMTV-PyMT MG tissue (Figure 5L). Changes in taurine-conjugated and unconjugated bile acid metabolites were also observed in probiotic bacteria-injected MG tissue (Figures S5A–S5D), with higher taurine-conjugated/unconjugated bile acid metabolite ratios observed in saline-injected MG compared with probiotic bacteria-injected MG tissue consistent with the bile salt hydrolase activity of Lactobacillus species.³² Tissue-resident probiotic bacteria regulate MG glycolytic signaling pathways at the gene, protein, and metabolite levels overviewed in Figure 5M.

Figure 5.

Probiotic MG microbiome shifts molecular signaling and cancer risk

(A) RNA sequencing revealed 600 differentially regulated genes in non-tumor-bearing MG injected with probiotic bacteria (n = 3 per group).

(B) KEGG pathway analysis revealed significant increases in several metabolism-related signaling pathways.

(C) Principal component analysis (PCA) of metabolite profiles from MG tissue injected with saline or probiotic bacteria from B6. MMTV-PyMT mice (n = 4).

(D) KEGG analysis revealed enrichment of several metabolic pathways with probiotic injection.

(E–I) No significant differences were observed in glucose (E) levels in probiotic-injected B6.MMTV-PyMT MG (n = 4 mice per group). Probiotic bacteria injection led to a significant increase in glucose-6-phosphate (F) metabolite levels, fructose-1,6-bisphosphate (G) metabolite levels, a trending increase in 3-phosphoglycerate (H) metabolite levels, and a significant increase in phosphoenolpyruvate (I) metabolite levels (n = 4 mice per group, ∗p < 0.05, unpaired t test).

(J) Pyruvate was unchanged with probiotic bacteria injection in B6.MMTV-PyMT MG (n = 4 mice per group, unpaired t test).

(K) Probiotic bacteria injection led to a significant decrease in lactate metabolite levels (n = 4 mice per group, ∗p < 0.05, unpaired t test).

(L) Pyruvate kinase M1/2 protein levels were elevated with probiotic injection in B6.MMTV-PyMT mice MG (n = 5 mice per group, ∗p < 0.05, unpaired t test). Representative western blot images (see Figure S7 for full blot images).

(M) Glycolysis signaling pathway regulation at the gene, protein, and metabolite levels in MG shifted by probiotic bacteria. Data are shown as the mean ± SEM.

See also Figure S5.

Probiotic bacteria differentially modulate cellular bioenergetics and viability

To further support the role of probiotic bacteria modulating metabolism, we show that Pro-CM increased both oxygen consumption rate (OCR, Figures 6A and 6B) and extracellular acidification rate (ECAR, Figure 6C) in S1 cells (a non-neoplastic human breast epithelial cell line), demonstrating that bacteria-produced metabolites promote breast metabolism. Conversely, Pro-CM had minimal effects on OCR and ECAR in MCF-7 ER+ human breast cancer cells (Figures 6D–6F), and Pro-CM significantly reduced OCR (Figures 6G and 6H), with no significant differences observed in ECAR (Figure 6I) in murine 4T1.2ER+ breast cancer cells, suggesting that probiotic bacteria-secreted metabolites only promote metabolism in non-cancerous breast epithelial cells.

Figure 6.

Probiotic bacteria-secreted metabolites differentially modulate cellular bioenergetics

(A–C) Mitochondrial respiration as measured by oxygen consumption rate (OCR) (A) in non-neoplastic S1 mammary epithelial cells treated with probiotic conditioned media (Pro-CM; n = 4 biological replicates). Pro-CM significantly increased basal OCR (B) and basal ECAR (C) in non-neoplastic S1 mammary epithelial cells (n = 4 biological replicates, ∗p < 0.05, unpaired t test).

(D and E) No significant differences in OCR were observed with Pro-CM treatment in human ER+ MCF-7 breast cancer cells (n = 5–12 biological replicates, unpaired t test).

(F) No significant differences were observed in basal ECAR with Pro-CM treatment in human ER+ MCF-7 breast cancer cells (n = 5–12 biological replicates, unpaired t test).

(G) Bioenergetic profile of murine 4T1.2ER+ breast cancer cells treated with Pro-CM.

(H) Pro-CM treatment significantly reduced basal OCR in murine ER+ 4T1.2ER+ breast cancer cells (n = 5–12 biological replicates, ∗p < 0.05, unpaired t test).

(I) No significant differences with Pro-CM were observed in basal ECAR in murine ER+ 4T1.2ER+ breast cancer cells (n = 5–12 biological replicates, unpaired t test). Data are shown as the mean ± SEM.

Consistent with published studies,^33,34 we demonstrate that Pro-CM significantly reduced the ER+ breast cancer cell line MCF-7 and ZR-75-1 in vitro. A dose-dependent reduction in cell viability of human ER+ breast cancer cell lines MCF-7 and ZR-75-1 (reported as relative cell index) was observed with increasing doses of Pro-CM, with 5% Pro-CM significantly reducing cell viability within 36 h of treatment (Figures 7A and 7B). Pro-CM did not significantly reduce S1 cell viability at any dose (Figure 7C). Next, we combined Pro-CM with 4-hydroxytamoxifen (4-OHT), the active metabolite of TAM, to determine if probiotic bacteria metabolites could enhance TAM efficacy on ER+ breast cancer cells. We demonstrated that 1% Pro-CM in combination with 1 μM 4-OHT significantly reduced both MCF-7 and ZR-75-1 cell viability (Figures 7D and 7E). Untargeted metabolomics was then performed on cell culture media only (control), unfiltered Pro-CM, and filtered Pro-CM to identify potential bacterial-processed metabolites contributing to the reduction in cell viability. Metabolomic analysis revealed differently regulated metabolites between media only, unfiltered, and filtered Pro-CM (Figure 7F). KEGG analysis of control and Pro-CM revealed enrichment of several metabolic pathways, including pentose phosphate pathway, galactose metabolism, and starch and sucrose metabolism (Figure 7G). Bacterial-derived metabolite trehalose was significantly increased in both unfiltered and filtered Pro-CM (Figure 7H). 1 μM trehalose significantly reduced both MCF-7 (Figure 7I) and ZR-75-1 (Figure 7J) cell index. Exogenous administration of trehalose at 100 mg/kg reduced tumor volume in the syngeneic 4T1.2ER+ model when compared with untreated control and maltose-administered animals, suggesting that probiotic bacteria-derived metabolites have anti-cancer effects (Figure 7K).

Figure 7.

Probiotic bacterial-derived metabolites reduce ER+ breast cancer viability

(A) 5% Pro-CM significantly decreased human ER+ MCF-7 breast cancer cell growth within 36 h (n = 4 biological replicates, ∗p < 0.05, ordinary one-way ANOVA, followed by Tukey test).

(B) 5% Pro-CM significantly decreased human ER+ ZR-75-1 breast cancer cell growth within 36 h (n = 4 biological replicates,∗p < 0.05, ordinary one-way ANOVA, followed by Tukey test).

(C–E) No significant differences were observed with Pro-CM treatment in non-neoplastic S1 mammary epithelial cells (n = 4 biological replicates, ordinary one-way ANOVA, followed by Tukey test). Combination of 1% Pro-CM and 1 μm 4-OHT significantly decreased human ER+ MCF-7 (D) and ZR-75-1 (E) breast cancer cell growth within 72 h (n = 4 biological replicates, ∗p < 0.05, ordinary one-way ANOVA, followed by Tukey test).

(F) Untargeted metabolomics revealed significantly different metabolites between media only (control), unfiltered, and filtered Pro-CM (n = 5 per group).

(G) KEGG pathway analysis revealed significant increases in several signaling pathways between media-only (control), unfiltered, and filtered Pro-CM (n = 5 per group).

(H–J) Bacterial-derived metabolite trehalose (H) was significantly increased in unfiltered and filtered Pro-CM (n = 5 per group, ∗p < 0.05, ordinary one-way ANOVA). Treatment with 1 μm trehalose significantly reduced human ER+ MCF-7 (I) and ZR-75-1 (J) breast cancer cell viability within 72 h (n = 4–5 biological replicates, ∗p < 0.05, unpaired t test).

(K) 100 mg/mL trehalose significantly reduced murine ER+ 4T1.2ER+ tumor growth (n = 5–15 per group, ∗p < 0.05, two-way ANOVA, followed by Holm-Šídák’s multiple comparisons test).

(L) Representative regions of interest (ROIs) images of ER+ breast tumors from women treated with neoadjuvant aromatase inhibitors and/or Faslodex were stained against Ki67 and LTA.

(M) LTA positivity negatively correlates with Ki67 in ER+ primary tumors that were treated with neoadjuvant endocrine-targeting therapies (n = 10 patients, 5 ROIs per patient; Pearson’s correlation coefficient r). Data are shown as the mean ± SEM. Scale, 100 μm.

See also Figure S6.

Primary breast tumor microbiota populations are associated with reduced tumor proliferation in patients with breast cancer treated with endocrine-targeting therapies

To determine potential associations between tumor proliferation and intratumoral bacterial populations, immunohistochemistry (IHC) was performed on primary tumors from patients with hormone receptor-positive breast cancer treated with aromatase inhibitors, Faslodex, or a combination of aromatase inhibitors and Faslodex in the neoadjuvant setting. Tumors were assessed for immunoreactivity against the proliferation marker Ki67, along with bacterial marker LTA (representative images of regions of interest are shown in Figure 7L). LTA positivity was negatively correlated with Ki67, suggesting that LTA-positive bacteria present in the breast tumor are associated with reduced proliferation (Figure 7M).

Primary breast tumor microbiota populations are associated with distant metastasis development in endocrine-targeting therapy-treated patients with breast cancer

16S sequencing was performed on DNA isolated from primary breast tumor tissue from patients with hormone receptor-positive breast cancer who did or did not develop distant metastases within 10 years while administered adjuvant TAM and/or aromatase inhibitors. See Table S2 for patient demographic information. Sequencing results do not indicate alterations at the phyla level (Figures S6A and S6D). However, genus (Figures S6B and S6F) and species (Figures S5C, S5G, and S5H) proportional abundance in hormone receptor-positive primary tumors from patients who did or did not develop distant metastases later on in life did display significant shifts. At the genus level, no differences in Lactobacillus proportional abundance were observed in primary tumors that did metastasize compared to primary tumors without distant metastasis after treatment (Figure S6E). However, a significant decrease in Streptococcus genus proportional abundance was observed in primary tumors from patients who developed distant metastasis (Figure S6F). Regarding species differences, Streptococcus gordonii proportional abundance (Figure S6G) and Streptococcus of an unspecified species proportional abundance (Figure S6H) were significantly decreased in primary tumors from patients who later on developed distant metastasis while undergoing or after endocrine-targeting therapy.

Discussion

Breast tissue harbors a diverse microbiome, with tissue microbial populations contributing to metabolic, inflammatory, and immune responses.³⁵ Diet-induced dysbiosis of breast microbial populations contributes to breast cancer initiation, progression, and therapeutic resistance.^8,11 However, the effects of oral endocrine-targeting therapies used to treat ER+ breast cancer on breast microbiota populations and cancer risk are unknown. In both murine and OVX NHPs, TAM administration did not significantly affect the overall breast bacterial α-diversity; however, shifts in β-diversity with TAM were observed in both models. Firmicutes phylum, which is primarily LTA-positive microbes,³⁶ was elevated with TAM administration in both healthy diet-fed mice and OVX NHPs. Consistently, shifts in LTA-positive genera Lactobacillus, Streptococcus, and Staphylococcus were observed in both models. However, differential shifts in these genera were observed with diets in our mouse model. Healthy diet + TAM-fed mice displayed significantly elevated Streptococcus genus abundance, while Western diet + TAM-fed mice had significantly elevated Staphylococcus genus abundance, suggesting that diet-drug interactions may affect MG microbial populations.

Similar species alterations were also observed in both models. Interestingly, TAM treatment elevated an unspecified Lactobacillus species in both OVX NHPs and healthy diet-fed mice. L. reuteri, which was elevated in OVX NHPs treated with TAM, is a probiotic species of bacteria with anti-inflammatory effects mediated by reducing the production of pro-inflammatory cytokines.³⁷ Studies have shown that L. reuteri also inhibits early-stage carcinogenesis and induces apoptosis in murine breast cancer.³⁸ L-asparaginase, isolated from L. reuteri, induced apoptosis via Cas8 and Cas9 upregulation in breast cancer cell lines.³⁹ L. casei, which was increased in Western + TAM mice, reduced transplanted tumor growth and increased survival through Th1 cytokine production.⁴⁰ Milk fermented with L. casei has been demonstrated to stimulate the immune response against breast tumors, reducing tumor growth in a murine model.⁴¹ Furthermore, metabolites produced by Lactobacillus and Streptococcus species display tumor-suppressing effects.^42,43,44 L. plantarum produces several postbiotic metabolites that promote apoptosis in MCF-7 breast cancer cells.⁴³ S. thermophilus inhibits colorectal tumorigenesis by secreting β-galactosidase, which has also been shown to inhibit MCF-7 growth in vitro.^44,45 Lactobacillus species also produce short-chain fatty acids (SCFAs), which have anti-tumor effects by inducing apoptosis in breast cancer cells.^46,47 In addition to butyrate, propionate inhibits tumor growth and the epithelial-to-mesenchymal transition and induces apoptosis in breast cancer cells by binding to GPR43 and GPR41 receptors.⁴⁶ Additionally, SCFA exhibits HDAC inhibitory activity to downregulate ERα activity. These SCFAs exhibit selective ER downregulator activity in MCF-7 and T-47D cell lines, further suggesting that microbial-derived metabolites have anti-cancer effects.⁴⁸ Therefore, increased probiotic species with TAM administration may further reduce breast carcinogenesis through the production of microbial-derived metabolites, which in turn may increase TAM efficacy.

Breast microbial alterations observed in both models align with findings reported by several studies. Lactobacillus is an abundant genus present in the breast tissue microbiome of healthy women.⁴⁹ Furthermore, specific Lactobacillus species like L. vini, L. paracasei, and L. gasseri are enriched in normal breast tissue compared to breast tumor tissue and adjacent non-tumor tissue.⁴⁹ Consistently, enrichment of Lactobacillus species with TAM was observed in both models. Lactobacillus appears to be a commensal member of the normal breast microbiome with potential protective effects. Its depletion may contribute to dysbiosis associated with breast pathologies like cancer and infections.^35,49 Increased Streptococcus relative abundance has been documented in healthy women compared to those with cancer, consistent with outcomes observed in our OVX NHPs and healthy diet-fed mice treated with TAM. Streptococcus species present in the breast microbiome are associated with the upregulation of genes involved in immune response and metabolism, suggesting that they may play a beneficial role.⁵⁰ However, lower abundances of Streptococcus and other lactic acid bacteria have been observed in the breast microbiome of women with breast cancer compared to healthy controls.^7,35 We also show increased intratumoral Streptococcus in primary tumors from women who did not display recurrence on endocrine-targeting therapies, further supporting the anti-cancer role of this genera. Elevated Staphylococcus abundance was reported in patients with breast cancer compared to healthy controls,⁷ as observed in OVX NHPs and Western diet-fed mice treated with TAM. S. sciuri, elevated with TAM administration in the OVX NHPs and Western diet-fed mice, was reclassified as Mammaliicoccus sciuri. The lactic acid-producing, non-pathogenic M. sciuri GMN01 strain has demonstrated probiotic activity, surviving at a wide range of pH levels.⁵¹ However, due to the potential pathogenic activity of M. sciuri and the lack of isolates with attributed Generally Regarded as Safe status in the United States, the question of the health-promoting effects should be considered with great caution.⁵²

Probiotic bacteria injection incorporated LTA-positive bacteria into the MG, as evidenced by increased LTA positivity in MG on day 14 post injection, increased proportional abundance of Lactobacillus spp. by 16S, elevated Lactobacillus spp. proportional abundance in DNA isolated from aerobic live cultured MG, and Lactobacillus gene expression by RT-PCR. L. paracasei and L. rhamnoses expression was found to be naturally present in the MG on day 0 before probiotic injection and unchanged by probiotic injection. Lactobacillus species including L. acidophilus (B6.MMTV-PyMT model) and L. plantarum (BALB/c model) were found elevated in MG after probiotic injection. 16S sequencing further validates IHC and RT-qPCR results. 16S sequencing on DNA isolated from aseptically collected frozen MG revealed an enrichment in L. plantarum (0.5%), L. zeae (6.5%), and an unspecified species of Lactobacillus (6.7%), suggesting that probiotic bacteria can colonize the breast tissue. Live cultured probiotic bacteria-injected MG displayed increased Lactobacillus species compared to live cultured saline-injected MG, including an enrichment in L. plantarum (7.7%), L. zeae (12.3%), and an unspecified species of Lactobacillus (6.4%) species, consistent with IHC, RT-qPCR, and sequencing data on frozen MG. Interestingly, L. zeae was not present in our probiotic injection, suggesting that probiotic injection may influence endogenous bacterial populations in the breast tissue.⁵³ These results suggest that exogenous probiotic bacteria can be incorporated into the breast microbiome; however, further study is needed to determine if orally administered probiotics also colonize the breast tissue, as intra-nipple injections are not feasible in a clinical setting.

Numerous murine models have demonstrated that intestinal probiotic bacteria strains can inhibit tumor development, reduce tumor volume, and suppress angiogenesis and metastasis.^40,42 L. acidophilus has been shown to increase survival time in tumor-bearing BALB/c mice compared to controls.⁵⁴ However, probiotics were orally administered in these studies; therefore, we wanted to determine the effects of localized probiotic bacteria on mammary tumorigenesis. Different Lactobacillus species and strains can have varying functional effects, such as producing antimicrobial substances, modulating the immune response, and preventing pathogen adherence.⁵⁵ Therefore, we wanted to investigate the effects of probiotic supplementation on tumorigenesis in Western diet-fed animals when MG Lactobacillus abundance was low. Intra-nipple MG probiotic bacteria injection elevated tumoral LTA-positive bacteria. Elevated intratumoral probiotic LTA-positive bacteria were associated with reduced tumor multiplicity, reduced tumor proliferation, and increased tumor-free survival. Although L. paracasei, L. reuteri, L. rhamnoses, and L. plantarum were not significantly regulated by treatment at the end of study, we did observe significant MG enrichment of L. acidophilus in probiotic bacteria-injected B6.MMTV-PyMT mice. Differences between the specific species enriched in MG tissue between models (time course and tumor model) may be due to timing (2 weeks in time-course study, while MGs were measured 5 weeks after last injection in MMTV study), dietary background (time course was performed in control diet consuming mice, while B6.MMTV-PyMT mice were consuming a Western diet), or murine species background strain (time course was performed in BALB/c mice and the B6.MMTV-PyMT tumorigenesis study was in the C57BL/6 background; BALB/c and C57BL/6 strains differ in immune bias, which could influence MG tissue bacterial surveillance).⁵⁶ Collectively, these results suggest that locally administered probiotic bacteria colonize the MG and reduce tumorigenesis in the spontaneously cancer-prone B6.MMTV-PyMT mouse model.

Altered metabolism is one of the hallmarks of cancer cells, in which tumors rewire their metabolism to promote growth, survival, proliferation, and long-term maintenance. Dysregulated metabolism can also lead to aberrant synthesis and accumulation of biosynthetic molecules, such as nucleotides and lipids, promoting tumor progression.⁵⁷ Maintaining normal cellular metabolism is therefore essential for preventing cancer by supporting proper energy production, biosynthesis, redox homeostasis, immune surveillance, and cellular signaling. Probiotic bacteria play a key role in nutrient metabolism, metabolite production, and energy balance regulation in the gut, which may contribute to their anti-cancer properties.⁵⁸ We now report that localized tissue-resident probiotic bacteria influence breast metabolic pathways in both non-tumor-bearing BALB/c mice and tumor-bearing B6.MMTV-PyMT.

Probiotics differentially regulated 600 genes in the MG of non-tumor-bearing BALB/c mice, with significant increases observed in several pathways, including glucagon signaling, central carbon metabolism, glycolysis, gluconeogenesis, and insulin signaling. In normal tissue, energy production relies primarily on mitochondrial oxidative phosphorylation (OXPHOS), which produces more adenosine triphosphate (ATP) than glycolysis.⁵⁹ However, normal tissues utilize both glycolysis and OXPHOS to produce ATP, cooperating to maintain cellular energetic balance. Studies suggest that glycolysis is necessary to supply rapid energy demands, while OXPHOS supplies chronic energy demands.⁶⁰ Additionally, we previously reported reduced glucose metabolism in the breast tissue of ovariectomized NHPs when compared with ovary-intact NHP breast tissue, suggesting that menopause may reduce breast glycolysis dysregulating breast tissue energy metabolism to impact cell stress responses and tumorigenesis.⁶¹

Untargeted metabolomics on non-cancerous MG tissue from B6.MMTV-PyMT mice further revealed significantly different metabolite abundances with probiotic bacteria injection. Of note, several glycolytic intermediates were regulated by probiotics in the MG. Probiotic bacteria increased MG F-1,6-BP, an endogenous intermediate of the glycolytic pathway produced by phosphorylation of fructose-6-phosphate by phosphofructokinase-1, as well as PEP, a high-energy metabolite in the final step of glycolysis in tumor-bearing B6.MMTV-PyMT mice.^62,63 Exogenous administration of F-1,6-BP reduces inflammation by inhibiting the production of proinflammatory molecules, including TNF-α, reactive oxygen species, and prostaglandin E₂.^64,65 While increased glycolysis is associated with carcinogenesis, under normal conditions, glycolysis plays a critical physiological role by rapidly responding to energetic demands.⁶⁰ Our respirometry data in human breast cancer (MCF-7), murine ER+ breast cancer (4T1.2ER+), and non-cancerous breast epithelial (S1) cell lines indicate that probiotic bacteria-secreted metabolites only promote mitochondrial respiration and glycolysis in non-cancerous breast cells. Therefore, these results suggest that probiotics may maintain normal MG metabolism, as the induction of glycolysis is only observed in normal MG tissue and cells.

Lactobacillus species, including L. plantarum, possess bile salt hydrolase enzymes that enable them to deconjugate bile acids like glycocholate and taurocholate in the gut. In breast cancer, the hepatic synthesis of primary bile acids and the bacterial conversion to secondary bile acids in the large intestine are suppressed.⁶⁶ Certain secondary bile acids like lithocholic acid have been shown to negatively correlate with breast cancer cell proliferation markers like Ki67, suggesting that bile acids may have anti-proliferative effects in breast cancer. This is further supported by the literature indicating that bile acid receptor FXR activation reduced triple-negative breast cancer (TNBC) growth.⁶⁷ We previously observed elevated bile acid metabolites and Lactobacillus abundance in breast tissue of NHPs consuming a Mediterranean diet when compared with Western diet consuming NHPs,²⁰ suggesting that breast microbiota may regulate tissue-specific levels of bile acid metabolites. We observed decreased taurine-conjugated/unconjugated bile acid concentrations in MG tissue after probiotic bacteria injection, suggesting that MG L. acidophilus may influence bile acid profiles and activity through deconjugation to impact breast cancer risk. Overall, the results from these two studies suggest that probiotics influence the composition of the breast microbiome and introduce beneficial functions to breast microbial communities, resulting in amelioration or prevention of disease phenotypes.

Considering the varied anti-cancer properties of probiotic bacteria, we aimed to assess whether probiotic bacteria-secreted metabolites could enhance the effectiveness of TAM treatment against ER+ breast cancer growth. Data indicated that treatment with Pro-CM reduced ER+ breast cancer cell viability in a dose-dependent manner. Interestingly, the treatment of nonneoplastic S1 mammary epithelial cells with Pro-CM did not lead to any alterations in cell viability. Furthermore, combining Pro-CM with 4-OHT, TAM’s active metabolite, resulted in a significant reduction in breast cancer cell viability compared to either Pro-CM or 4-OHT alone, suggesting that probiotic bacteria-produced metabolites can increase TAM efficacy against ER+ breast cancer cells.

Many anti-cancer properties of probiotics are attributed to metabiotics. “Metabiotics” refers to the structural components of microorganisms, their metabolites, and signaling molecules having a defined chemical structure that can optimize host-specific physiological functions as well as regulatory, metabolic, and behavior reactions associated with the activity of the host.^68,69 These metabolites can exert direct or indirect effects on cancer cells and the tumor microenvironment. Some examples of anti-cancer metabolites produced by probiotic bacteria include SCFAs,⁷⁰ bacteriocins,⁷¹ antioxidants,⁷² exopolysaccharides,⁷³ and conjugated linoleic acid.⁷⁴ Untargeted metabolomics performed on Pro-CM revealed significantly different metabolite profiles between media only (control), unfiltered Pro-CM, and filtered Pro-CM. Interestingly, several metabolites were upregulated with both unfiltered and filtered Pro-CM, including trehalose, a non-reducing disaccharide.

Trehalose is ubiquitous, found in a variety of different organisms, such as yeast, fungi, insects, and bacteria.⁷⁵ In these organisms, trehalose has several functions. It acts as a storage compound for carbohydrates, as well as protects cells from environmental stresses, including heat, cold, desiccation, dehydration, and oxidation.⁷⁶ Lactobacillus synthesizes trehalose to survive during the freeze-drying process during probiotic capsule formation.^77,78 Trehalose is not present in mammalian cells, due to the lack of trehalose biosynthesis genes. Trehalose has been shown to dramatically reduce the development and progression of several diseases, including neurodegenerative disorders, cardiovascular diseases, and metabolic diseases.⁷⁹ Recent studies have shown that trehalose induces autophagy in TNBC, reducing cell growth in vivo. A study by Li et al. demonstrated trehalose, but not sucrose, significantly delayed and reduced TNBC tumor growth.⁸⁰ Our data indicate that trehalose reduces ER+ breast cancer growth both in vitro and in vivo, indicating that probiotic bacteria-derived metabolites may serve as potential treatment options to complement conventional cancer therapies, which we have begun to explore. However, further study is needed to identify the exact mechanisms by which trehalose exerts its anti-cancer effects.

Conclusions

Overall, the results of this study demonstrate that oral endocrine-targeting therapy TAM shifts breast microbiota populations, increasing beneficial probiotic Lactobacillus and Streptococcus bacteria. Elevating MG probiotic bacteria reduced tumor multiplicity and increased tumor-free survival in cancer-prone B6.MMTV-PyMT mice, which was associated with decreased tumor proliferation and modulation of normal MG metabolism. We also observed increased intratumoral LTA-positive bacteria that negatively correlated with proliferation in tumors from endocrine-targeted therapy-treated patients. Further studies are needed to determine successful interventions to increase breast LTA-positive bacterial abundance and determine the molecular signaling mechanisms by which Lactobacillus and Streptococcus bacteria may reduce breast tumorigenesis in conjunction with oral endocrine therapies.

Limitations of the study

To study the effects of TAM on breast microbiota in a model of menopause, we utilized ovariectomized NHPs. Studies demonstrate that the ovariectomy-induced estrogen deficiency in these animal models closely mimics the hormonal changes and physiological effects observed during natural menopause in women.⁸¹ Nonetheless, utilizing ovariectomized NHPs for menopausal research presents certain limitations, including disparities in the timing and hormonal fluctuations compared to the human menopausal transition. Additionally, the age of the animals plays a role, as older animals (>12 years) may exhibit distinct responses and better represent human menopause, compared to younger ovariectomized NHPs.⁸²

Next-generation sequencing comparing the microbiota of malignant, non-malignant, and healthy breast tissue identified potential causative links between bacteria, host tumors, and normal tissue that can be explored. However, there are still some major challenges associated with accurately characterizing these microbiota populations. Tumor and breast samples are regions of known low microbial biomass, a feature that complicates any metagenomic analysis. The high ratio of host to bacterial DNA can lead to bias and can make whole-genome sequencing impossible without a microbial enrichment strategy.⁸³ Furthermore, environmental contamination is likely to be inherent given the sampling process, which, due to the low biomass of these tissues, has the potential to obscure tumor-originating bacteria.⁸⁴ For our studies, we utilized aseptic techniques to reduce environmental contamination and sent them as batched samples to be sequenced.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed and will be fulfilled by the lead contact, Katherine L. Cook (klcook@wakehealth.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •All source data reported in this paper will be shared by the lead contact upon request.
• •16S sequencing data have been deposited at NCBI:PRJNA1138780 and are publicly available as of the date 9/1/2024. Accession numbers are listed in the key resources table.
• •Metabolomics data have been deposited at Figshare and are publicly available as of the date of publication, https://doi.org/10.6084/m9.figshare.27292473.v1.
• •This study did not report new original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This work was funded in part by Breakthrough Awards from the Department of Defense Breast Cancer Research Program (W81XWH-20-1-0014 and W81XWH-22-1-0055 to K.L.C.). A.A.A. is supported by an NIH-NIAID T32 Training Program in Immunology and Pathogenesis (5T32AI007401). We would like to acknowledge the support provided by the Proteomics and Metabolomics Shared Resource and the Tumor Tissue and Pathology Shared Resource of the Wake Forest School of Medicine and Atrium Health Wake Forest Baptist Comprehensive Cancer Center (NIH/NCI P30 CA12197).

Author contributions

A.A.A., A.S.W., and Y.-T.T. performed experiments and obtained the data included in this manuscript. A.A.A., B.W., D.R.S.-P., and K.L.C. analyzed the data. A.A.A. and K.L.C. wrote the manuscript. K.L.C. and J.M.C. contributed to the conceptual design of experiments. K.L.C. supervised the study. Clinical sample selection, study design, and supervision were provided by M.E.S., A.C., M.H.-M., E.A.L., and A.T. All authors discussed the results and commented on the manuscript.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
F4/80 (D2S9R) XP® Rabbit mAb	Cell Signaling	Cat #: 70076 RRID:AB_2799771
GLUT1 Recombinant Rabbit Monoclonal Antibody	Invitrogen	Cat #: SA0377; RRID:AB_2941981
Lipoteichoic Acid Monoclonal Antibody	Invitrogen	Cat #: MA1-40134 RRID:AB_1076514
Lipopolysaccharide Polyclonal Antibody	LSBio	Cat #: LS-C375096
Ki-67 (D3B5) Rabbit mAb	Cell Signaling	Cat #: 12202 RRID:AB_2620142
Hexokinase I (C35C4) Rabbit mAb	Cell Signaling	Cat #: 2024 RRID:AB_2116996
PKM1/2 (C103A3) Rabbit mAb	Cell Signaling	Cat #: 3190 RRID:AB_2163695
alpha-Tubulin Antibody	Cell Signaling	Cat #: 2144 RRID:AB_2210548
Bacterial and virus strains
Lactobacillus paracasei NI327	Over-the-counter probiotic	N/A
Bifidobacterium breve NI312	Over-the-counter probiotic	N/A
Bifidobacterium lactis NI349	Over-the-counter probiotic	N/A
Lactobacillus acidophilus NI317	Over-the-counter probiotic	N/A
Lactobacillus plantarum NI329	Over-the-counter probiotic	N/A
Lactobacillus rhamnoses NI332	Over-the-counter probiotic	N/A
Bifidobacterium bifidum	Over-the-counter probiotic	N/A
Biological samples
Postmenopausal ER + tumors	Wake Forest University Comprehensive Cancer Center	Table S2 IRB00026366
Chemicals, peptides, and recombinant proteins
(Z)-4-Hydroxytamoxifen	Tocris	Cat #: 0999
Critical commercial assays
DNeasy PowerLyzer PowerSoil Kit	Qiagen	Cat #: 12855-100
16S Sequencing	CosmosID	N/A
Untargeted Metabolomics	Metabolon	N/A
Untargeted Metabolomics	Proteomics and Metabolomics Shared Resource of the Wake Forest School of Medicine and Wake Forest Baptist Comprehensive Cancer Center (NIH/NCI P30 CA12197)	N/A
Deposited data
Raw data of 16S rRNA sequencing	This paper	NCBI Sequence Read Archive:SUB14616768
Raw data of untargeted metabolomics	This paper	Figshare: https://doi.org/10.6084/m9.figshare.27292473.v1
Experimental models: Cell lines
MCF-7	ATCC	Cat #: HTB-22 RRID:CVCL_0031
ZR-75-1	ATCC	Cat #:CRL-1500 RRID:CVCL_0588
HMT-3522 S1	Dr. Pierre Vidi (Institut de Cancerologie de l’Ouest, Angers, France)	Holmes et al.85
4T1.2ER+	Bethany Kerr and Katherine L. Cook Lab	Langsten et al., 202386
Experimental models: Organisms/strains
Mouse: C57BL/6J	Jackson Laboratories	RRID:IMSR_JAX:000664
Mouse: BALB/cJ	Jackson Laboratories	RRID:IMSR_JAX:000651
Mouse: B6.FVB-Tg(MMTV-PyVT)634Mul/LellJ	Jackson Laboratories	RRID:IMSR_JAX:022974
Non-human primate: Cynomolgus Macaques (Macaca fascicularis)	Institut Pertanian Bogor or Charles River Primates	Arnone et al.61 Cline et al.87
Software and algorithms
inForm Tissue Analysis Software	Akoya BioSciences	RRID:SCR_019155
VIS Image Analysis Software	Visiopharm	RRID:SCR_021711
Image Lab Software	Bio Rad	RRID:SCR_014210
GraphPad Prism	GraphPad	RRID:SCR_002798
Seahorse Wave	Agilent Technologies	RRID:SCR_014526
RTCA Software	Agilent Technologies	RRID:SCR_014821
MetaboAnalyst	McGill University; Montreal; Canada	RRID:SCR_015539
Biorender	Biorender	RRID:SCR_018361
Other
Mantra Quantitative Pathology Image System	Akoya Biosciences	N/A
Agilent Seahorse XFe96 Analyzer	Agilent Technologies	RRID:SCR_019545
Seahorse XF Cell Mito Stress Test Kit	Agilent Technologies	Cat #:103015-100
Seahorse XF DMEM medium, pH 7.4, 500 mL	Agilent Technologies	Cat #:103575-100
Seahorse XF 1.0 M glucose solution, 50 mL	Agilent Technologies	Cat #: 103577-100
Seahorse XF 100 mM pyruvate solution, 50 mL	Agilent Technologies	Cat #: 103578-100
Seahorse XF 200 mM glutamine solution, 50 mL	Agilent Technologies	Cat #: 103579-100
Seahorse XF Calibrant Solution 500 mL	Agilent Technologies	Cat #: 100840-000
Seahorse FluxPaks	Agilent Technologies	Cat #: 103792-100
Agilent xCELLigence RTCA DP - Cell Invasion & Migration	Agilent Technologies	RRID:SCR_019037
Agilent xCELLigence RTCA DP E-Plate 16; 36 Plates	Agilent Technologies	Cat #: 5469813001
Life Technologies QuantStudio 3 Real-Time PCR System	ThermoFisher Scientific	RRID:SCR_020238
BioRad Trans Blot Turbo system	Biorad	RRID:SCR_023156
Bio Rad ChemiDoc MP Imaging System	Biorad	RRID:SCR_019037
DMEM, high glucose, pyruvate	Gibco	Cat #: 11995065
RPMI 1640 Medium	Gibco	Cat #: 11875093
DMEM/F-12, HEPES	Gibco	Cat #: 11330057
Ovine Prolactin	NIDDK	Cat #: NIDDK-oPRL-21
Insulin solution human	Sigma	Cat #: I9278
Hydrocortisone	Sigma	Cat #: H0888
β-Estradiol	Sigma	Cat #: E2758
Sodium selenite	Corning	Cat #: 354201
apo-Transferrin human	Sigma	Cat #: T2252
SuperSignal™ West Femto Maximum Sensitivity Substrate	ThermoFisher Scientific	Cat #: 34095
Pierce™ Bovine Serum Albumin Standard Pre-Diluted Set	ThermoFisher Scientific	Cat #: 23208
Blotting-Grade Blocker	Bio-Rad	Cat #: 1706404
PCR Plate, 96-well, semi-skirted, flat deck	ThermoFisher Scientific	Cat #: AB1400
ABsolute qPCR Plate Seals	ThermoFisher Scientific	Cat #:AB1170
PowerUp™ SYBR™ Green Master Mix for qPCR	Applied Biosystems	Cat #: A25742
Nuclease-Free Water (not DEPC-Treated)	Invitrogen	Cat #: AM9937
SuperScript™ III Reverse Transcriptase	Invitrogen	Cat #: 18080044
RNaseOUT™ Recombinant Ribonuclease Inhibitor	Invitrogen	Cat #: 10777019
TRIzol™ Reagent	Invitrogen	Cat #: 15596018
D-(+)-Maltose monohydrate	Sigma	Cat #: M5885
D-(+)-Trehalose dihydrate	Sigma	Cat #: T9531
Healthy low fat control diet (6% OO, 2% FO, 0.6% NaCl, G)	inotiv	Cat #: TD.190100
Western 45/Fat diet (7% Fruct, 3.4% NaCl, R)	inotiv	Cat #: TD. 190099
Healthy+ Tamoxifen Citrate Diet (37 ppm, 6% OO, 2% FO, 0.6% NaCl, O)	inotiv	Cat #: TD. 200645
Western+ Tamoxifen Citrate (37 ppm, 45/Fat Diet, 7% Fruc, 3.4% NaCl, P)	inotiv	Cat #: TD. 200647
LB Broth (1X)	Gibco	Cat #: 10855001
Peroxidase and Alkaline Phosphatase Blocking Reagent (Dual Endogenous Enzyme-Blocking Reagent)	Agilent	Cat #: S200389-2
Protein Block, Serum-Free	Agilent	Cat #: X090930-2
Antibody Diluent	Agilent	Cat #: S080983-2
EnVision+/HRP, Mouse, HRP. Mouse, Immunohistochemistry Visualization	Agilent	Cat #: K400111-2
EnVision+/HRP, Rabbit, HRP. Rabbit, Immunohistochemistry Visualization	Agilent	Cat #: K400311-2
Antigen Retrieval Citra (pH-6.0)	BioGenex	Cat #: HK0865K-GP
2-Mercaptoethanol	Biorad	Cat #: 1610710
4 x Laemmli Sample Buffer	Biorad	Cat #: 1610747
Precision Blue Protein All Blue Standards	Biorad	Cat #: 1610373
MagicMark™ XP Western Protein Standard	Invitrogen	Cat #: LC5602
Restore™ Western Blot Stripping Buffer	Thermo Scientific	Cat #: 21063
Pierce™ 660nm Protein Assay Reagent	Thermo Scientific	Cat #: 22660
10x Tris/Glycine/SDS Buffer	Biorad	Cat #: 1610732
Trans-Blot Turbo 5X Transfer Buffer	Biorad	Cat #: 10026938
4–15% Mini-PROTEAN® TGX Stain-Free™ Protein Gels, 10 well, 30 μL	Biorad	Cat #: 4568083
L. rhamnosus (GCC GAT CGT TGA CGT TAG TTG G; CAG CGG TTA TGC GAT GCG AAT)	IDT	N/A
L. plantarum (GCT GGC AAT GCC ATC GTG CT; TCT CAA CGG TTG CTG TAT CG)	IDT	N/A
L. paracasei (CAA TGC CGT GGT TGT TGG AA; GCC AAT CAC CGC ATT AAT CG)	IDT	N/A
L. acidophilus (CCT TTC TAA GGA AGC GAA GGA T; ACG CTT GGT ATT CCA AAT CGC	IDT	N/A
16S rRNA (TCCTACGGGAGGCAGCAGT; GGACTACCAGGGATCTAATCCTGTT)	IDT	N/A
B. breve (TCA TCA CGGCAA GGT CAA GA:GGC CAG AAC AGC TGG AAC AA)	IDT	N/A
B. bifidum (CTG GCA GCC GTG ACA CTA CT, TGA ACT GGC CGT TAC GGT CT)	IDT	N/A
B. lactis (ACC TCA CCA ATC CGC TGT TC, GAT CCG CAT GGT GGA ACT CT)	IDT	N/A
HPRT (CATAACCTGGTTCATCATCGC; TCCTCCTCAGACCGCTTTT).	IDT	N/A

Experimental model and study participant details

Cell culture

Cells were grown at 37°C in a humidified, 5% CO2:95% air atmosphere. MCF-7 human mammary epithelial cells were cultured in phenol red-containing DMEM medium supplemented with 10% FBS and defined as basal growth conditions (Gibco, #1995065). ZR-75-1 human mammary epithelial cells were cultured in phenol red-containing RPMI supplemented with 10% FBS and defined as basal growth conditions (Gibco, #11875093). These cell lines were authenticated by IDEXX BioAnalytics using short tandem repeat analysis. HMT-3522 S1 (S1) non-neoplastic mammary epithelial cells were obtained in collaboration with Dr. Pierre Vidi (Institut de Cancerologie de l’Ouest, Angers, France) and propagated between passages 54 and 60 in DMEM/F12 (Gibco, #11330057) modified with ovine prolactin (NIDDK, #NIDDK-oPRL-21), insulin (Sigma, #I9278-5m), hydrocortisone (Sigma, #H0888-1G), β-estradiol (Sigma, #E2758-250MG), sodium selenite (Corning, #354201), and transferrin (Sigma, #T2252-100MG).⁸⁵

Oral endocrine therapy and diet model

Female six-week-old C57BL/6 mice (n = 32) were purchased from Jackson Laboratories (Bar Harbor ME, USA). This animal study was approved by the Wake Forest University School of Medicine Institutional Animal Care and Use Committee and all procedures were carried out in accordance with relevant guidelines and regulations. Mice were placed on either a healthy control diet (HC; 21% kcal from fat derived from olive oil and fish oil; TD. 190100) or a Western diet (45% kcal from fat derived from corn oil, lard, milk-fat: TD. 190099) for 6 weeks. See Table S1 for diet composition. All Teklad custom diets were purchased from Envigo. Mice were then randomized based on body fat composition determined by Echo-MRI (Houston TX, USA) into 4 different treatment groups: healthy control; healthy control +TAM, Western diet, and Western +TAM. Body weight and food consumption was monitored weekly. Tamoxifen citrate (TAM) was formulated in both the healthy and Western diets at 37 ppm (Healthy TD. 200645; Western TD.200647) representing the human equivalent dose in mice based upon the FDA body surface area calculations.

BALB/c probiotic intra-nipple injection time course model

Female 8–12-week-old BALB/c mice (n = 37) were purchased from Jackson Laboratories (Bar Harbor ME, USA). This animal study was approved by the Wake Forest University School of Medicine Institutional Animal Care and Use Committee and all procedures were carried out in accordance with relevant guidelines and regulations. Mice were placed on a control diet for the duration of the study. Mice were injected with 50 μL saline (control; n = 9) or common over-the-counter probiotic bacteria species (10⁷ CFU mixture of Bifidobacterium lactis; Bifidobacterium bifidum; Bifidobacterium breve; Lactobacillus acidophilus; Lactobacillus rhamnosus; Lactobacillus paracasei; Lactobacillus plantarum; n = 28) into the right 2/3, left 4/5 and right 4/5 mammary fat pad on day 0 of study. Left 2/3 mammary glands were uninjected to use as a control. Mammary gland tissue was fixed, frozen, or live cultured on 0 (baseline) 1-, 3-, 7-, and 14 days post-injection.

MMTV-PyMT probiotic intra-nipple injection model

Female 5-week-old MMTV-PyMT (B6.FVB-Tg(MMTV-PyVT)634Mul/LellJ; Strain #022974; n = 13) mice were purchased from Jackson Laboratories (Bar Harbor ME, USA). This animal study was approved by the Wake Forest University School of Medicine Institutional Animal Care and Use Committee and all procedures were carried out in accordance with relevant guidelines and regulations. Mice were placed on a Western diet (TD.190099) for the duration of the study. Mice were injected with either saline (control; n = 6) or commonly used over-the-counter probiotic bacteria (strains: L. paracasei, L. casei, Bifidobacterium breve, L. rhamnosus, B. lactis, B. longum, L. acidophilus, L. plantarum, L. bulgaricus, L. helveticus; 10⁷ CFU; n = 7) into all (R1-5 and L1-5) mammary glands at 5, 7, 9, and 11 weeks of age. Tumor-free survival, tumor multiplicity, tumor latency, and tumor weight were recorded. Non-tumor-bearing MG tissue was used for RNA sequencing (Novogene; Sacramento, CA) and untargeted metabolomics (Metabolon; Raleigh, NC).

Ovariectomized non-human-primate model

Methods were adapted from Arnone et al. and Cline et al..^61,87 In brief, adult female cynomolgus macaques (Macaca fascicularis) were imported from Indonesia (Institut Pertanian Bogor or Charles River Primates, Port Washington, NY). Bilateral ovariectomies were performed on animals 3 months before treatment. Ovariectomized animals were untreated (control group) or continuously treated with tamoxifen for 2.5 years. Treatment was administered in the diet at doses equivalent on a caloric basis to 20 mg/day for tamoxifen. The animal’s age was determined at randomization by dentition, with a mean age of 7.5 years at the end of the study. Animals were housed in social groups of 4–6 monkeys in an AAALAC-accredited facility. All experimental protocols were approved by the Institutional Animal Care and Use Committee.

BALB/C 4T1.2ER + trehalose dose response model

Female 8-week-old BALB/c mice (n = 35) were purchased from Jackson Laboratories (Bar Harbor ME, USA). This animal study was approved by the Wake Forest University School of Medicine Institutional Animal Care and Use Committee and all procedures were carried out in accordance with relevant guidelines and regulations. Mice were placed on a Western diet (TD.190099) for the duration of the study. After one week on diet, mice were orthotopically injected with 1 X10⁶ syngeneic murine 4T1.2ER + breast cancer cells into the left 4/5 mammary gland fat pad to induce tumors. Once tumors reached 75–100 mm³, mice were randomized into the following groups: untreated control group, 100 mg/mL maltose (control for trehalose), 11 mg/mL trehalose, 33 mg/mL trehalose, and 100 mg/mL trehalose. Maltose (Sigma-Aldrich; M5885) and trehalose (Sigma-Aldrich; T9531) were administered in the drinking water, which was monitored and changed weekly. Mice were on treatment for 21 days.

Adjuvant and neoadjuvant-treated ER + breast cancer samples

This study was approved by the Institutional Review Board of the Wake Forest School of Medicine (Winston-Salem, NC) in accordance with the U.S. Department of Health and Human Services regulations for the protection of human research subjects. The use of the tissue for analysis was covered by IRB00045734. Patient demographic information is provided in Table S2. Breast tumor specimens are snap-frozen. A designation of tumor or normal is made by visual gross inspection by the pathologist before tumor samples are submitted to the Tumor Bank.

Method details

Cell Viability Assays

MCF-7 breast cancer cells, ZR-75-1 breast cancer cells, or S1 non-neoplastic mammary epithelial cells were seeded into 16-well ACEA plates. Cells were treated with 0.5%, 1%, or 5% filtered probiotic conditioned media (Pro-CM) 4 h after plating. Pro-CM was made by adding one over the counter 20 × 10⁹ CFU (L. paracasei NI327, L. paracasei NI320, Bifidobacterium breve NI312, L. rhamnoses GG, Bifidobacterium lactis NI349, Bifidobacterium longum NI316, L. acidophilus NI317, L. plantarum NI329, L. bulgaricus NI321, L. helveticus NI325) probiotic capsule in 5 mL DMEM (-P/S, +L-Glut, +10% FBS). Pro-CM was placed in a 37°C water bath for 3 h, and then sterilely filtered. Cells were treated with either 1% Pro-CM, 1μM 4-hydroxytamoxifen (4-OHT; Tocris # 68047-06-3) or combination 1% Pro-CM and 1 μM 4-OHT. Cells were also treated with 1 μM trehalose. Cell growth was dynamically monitored using xCelligence system (ACEA Biosciences) for 72 h. Measurements were automatically collected by the analyzer. Cell cytotoxicity was calculated using xCELLigence software set to collect impendence data (reported as cell index).

Seahorse bioenergetic flux assay

ER + breast cancer MCF-7 and 4T1.2ER + cells, or S1 nonneoplastic mammary epithelial cells were plated at 50,000 cells/well in seahorse microplates and treated with 1% probiotic conditioned media (Pro-CM) for 24 h. See Pro-CM decroption in Cell Viability Assay description. Mitochondrial metabolism was measured as the oxygen consumption rate (OCR). OCR was measured in XF media (non-buffered DMEM containing 2mM Glutamine pH 7.4) under basal condition and in response to Oligomycin 1μM, FCCP 1μM + Rotenone/Antimycin A 1μM with the XF-96 Extracellular Flux Analyzer (Agilent Technologies). Glycolytic flux expressed as extracellular acidification rate (ECAR) and analyzed with the Wave software.

Immunohistochemistry

Paraffin-embedded sections from BALB/c mouse mammary glands that received intra-nipple probiotic injections were stained for LTA (lipoteichoic acid; Gram-positive bacteria; Invitrogen #MA1-40134), LPS (lipopolysaccharide; Gram-negative bacteria; LSBio #LS-C375096) and pan-macrophage marker F4/80 (Cell Signaling #70076). Paraffin-embedded sections from B6.MMTV-PyMT mouse mammary tumors were stained for antibodies against Ki67 (proliferation marker; Cell Signaling #12202), LTA, and LPS, and F4/80. Neoadjuvant endocrine therapy treated patient breast tumor tissue was stained for LTA, LPS, or Ki67. The Dako Dab chromogen kit was used for IHC. Staining was visualized by the Mantra Quantitative Pathology Image System with a 20x objective. Percent positive cells per field were quantified using the inForm tissue analysis software (Akoya Biosciences, Marlborough, MA). A minimum of 4 images were analyzed for each antibody. Non-human primate breast tissue was stained using antibodies against LTA, and LPS. Positive staining was analyzed using Visiopharm pathology image analysis software (Visiopharm, Broomfield, CO). The percent DAB positive cells/total number of nuclei was used for quantification. Intralobular ducts were used for ROIs.

Protein isolation and immunoblotting

Proteins isolated from B6.MMTV-PyMT MG were used to determine whether probiotic injection induced glycolytic signaling. Briefly, protein was isolated from snap-frozen MG using RIPA buffer. Protein was size fractionated by gel electrophoresis and transferred to a nitrocellulose membrane. Membranes were incubated overnight at 4°C with primary antibodies Pyruvate Kinase (PKM1/2; Cell Signaling, #3190) and α-tubulin (Cell Signaling, # 2144) at a 1:1000 dilution. The next day, membranes were washed and incubated with polyclonal horseradish peroxidase-conjugated secondary antibodies. Immunoreactive products were visualized by chemiluminescence (SuperSignal West Femto; Thermo Scientific # 34095) and quantified by densitometry using the Bio-Rad digital densitometry Image Lab software. Western blots are shown in Figure 5 as cropped images. Whole blot images are shown in Figure S7.

Bulk RNA-sequencing

MG from B6.MMTV-PyMT mice injected with saline or probiotics were collected 3 days after injection. RNA was isolated from snap-frozen B6.MMTV-PyMT MG tissue using a TRizol reagent protocol (Thermo Fisher, Catalog # 15596026). Preparation of library and transcriptome sequencing was performed by Novogene (Sacramento, CA). RNA-seq data provided by Novogene (Ensembl output) were filtered using Entrez conversion in DAVID.^88,89 All genes with a DE p-value <0.1 and an average count of 64 for all subjects were analyzed for Discrete Correlation Summation (DCS) Clustering.⁹⁰ Enrichment analysis for the KEGG pathway database was performed using ShineyGO 0.77.⁹¹ Pathways with FDR ≤0.05 cutoff were explored.

Untargeted metabolomics

Non-tumor bearing MG from B6.MMTV-PyMT mice injected with saline or probiotics were snap-frozen at the end of the study and untargeted metabolomics was performed by Metabolon (Raleigh, NC) as previously described.^61,92

Ultrahigh performance liquid chromatography-tandem mass spectroscopy (UPLC-MS/MS)

A Waters ACQUITY ultra-performance liquid chromatography (UPLC) system and a Thermo Scientific Q-Exactive mass spectrometer interfaced with a heated electrospray ionization (HESI-II) source and Orbitrap mass analyzer was used. Compounds were identified by comparison to library entries of purified standards or recurrent unknown entities. Peaks were quantified using area-under-the-curve.

Metabolomics, bioinformatics, and statistics

The informatics system consisted of the Laboratory Information Management System (LIMS), the data extraction and peak-identification software, data processing tools for QC and compound identification, and a collection of information interpretation and visualization tools.⁹³ Log transformation and imputation of missing values were performed with the minimum observed value for each compound.

Media only (control), unfiltered Pro-CM, and filtered Pro-CM were prepared for LC-MS/MS analysis and untargeted metabolomics performed by Wake Forest Proteomics and Metabolomics Shared Resources. In brief, the LC-MS analysis was performed on a Q Exactive HF hybrid quadrupole-Orbitrap mass spectrometer (Thermo Scientific, Waltham, MA, USA) coupled with a Vanquish UHPLC system (Thermo Scientific, Waltham, MA, USA). Samples were analyzed on three different columns, a Hypersil GOLD pentafluorophenyl (PFP) column (2.1 × 100mm, 1.9 μm, Thermo Scientific, Waltham, MA, USA), an Accucore Vanquish C18+ (2.1 × 100 mm, 1.5 μm, Thermo Scientific, Waltham, MA, USA) column, and an Acquity BEH Amide HILIC column (2.1 × 150 mm, 1.7 μm, Waters, Milford, MA, USA). Data were acquired by collecting full mass spectra (MS1) using polarity switching (positive/negative) at a resolution of 120K. The metabolite peak features were extracted and integrated within the MSMLS Discovery software (IROA Technologies LLC, Sea Girt, NJ, USA) in combination with an in-house compound library prepared using the Sigma Mass Spectrometry Metabolite Library. To eliminate redundancy in compound identification, the most abundant ion in each metabolite was selected⁹⁴ (https://www.sigmaaldrich.com/US/en/technical-documents/protocol/research-and-disease-areas/metabolism-research/msmls). Metabolite pathway enrichment was performed by MetaboAnalyst (https://www.metaboanalyst.ca).

Live cultured mammary glands

Un-injected left 2/3 mammary glands and injected right 2/3 mammary glands from non-tumor bearing BALB/c mice that received intranipple probiotic injections were sterilely removed and placed in 15 mL of autoclaved LB broth (Gibco # 10855001) in a tissue culture hood on days 0, 1,3,7, or 14 post-injection. Flasks for environmental control (LB broth only), the left un-injected mammary gland, and the right injected mammary gland were placed in a 37°C shaker for 24 h. Samples were spun down and DNA was then isolated from the pellet for 16S sequencing.

16S sequencing

DNA was isolated from live-cultured non-tumor bearing BALB/c mouse MG, non-tumor bearing BALB/c mouse MG, B6.MMTV-PyMT tumors, OVX NHP breast, and ER + human breast tumors using the Qiagen DNeasy PowerLyzer Powersoil Kit (Valencia, CA) and 16S sequencing was performed by CosmosID Inc. (Rockville, MD). Briefly, DNA libraries were prepared using Illumina 16s Metagenomic Sequencing kit (Illumina, Inc., San Diego, CA, USA) according to the manufacturer`s protocol. The V3-V4 region of the bacterial 16S rRNA gene sequences were amplified using the primer pair containing the gene-specific sequences and Illumina adapter overhang nucleotide sequences.

The full-length primer sequences are: 16S Amplicon PCR Forward Primer (5′- TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG) and 16S Amplicon PCR Reverse Primer (5′-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC).

Amplicon PCR was performed to amplify template out of input DNA samples. Briefly, each 25 μL of polymerase chain reaction (PCR) reaction contains 12.5 ng of sample DNA as input, 12.5 μL 2x KAPA HiFi HotStart ReadyMix (Kapa Biosystems, Wilmington, MA) and 5 μL of 1 μM of each primer. PCR reactions were carried out using the following protocol: an initial denaturation step performed at 95°C for 3 min followed by 25 cycles of denaturation (95°C, 30 s), annealing (55°C, 30 s) and extension (72°C, 30 s), and a final elongation of 5 min at 72°C. PCR product was cleaned up from the reaction mix with Mag-Bind RxnPure Plus magnetic beads (Omega Bio-tek, Norcross, GA). A second index PCR amplification, used to incorporate barcodes and sequencing adapters into the final PCR product, was performed in 25 μL reactions, using the same master mix conditions as described above. Cycling conditions were as follows: 95°C for 3 min, followed by 8 cycles of 95°C for 30 min, 55°C for 30 min and 72°C for 30 min. A final, 5-min elongation step was performed at 72°C. The library ∼600 bases in size was checked using an Agilent 2200 TapeStation and quantified using QuantiFluor dsDNA System (Promega). Libraries were normalized, pooled, and sequenced (2 x 300 bp paired end read setting) on the MiSeq (Illumina, San Diego, CA).16S sequencing data were analyzed by the CosmosID 16S pipeline and database. Results were presented as an operational taxonomic unit (OTU) table, visualized as heatmaps, stacked bar charts, alpha diversity plots, and beta diversity network graphs. Zymo community standard (D6305) is used as a positive control and lab-grade DEPC (diethylpyrocarbonate) treated water is used as a negative control.

RT-qPCR

RNA was extracted from B6.MMTV-PyMT mammary glands and from non-tumor bearing BALB/c mice mammary glands and intestines that received intranipple injections with probiotics using TRIzol following the manufacturer’s protocol. cDNA was synthesized from 2-5μg of total RNA using Superscript first-strand RT-PCR reagents described by the manufacturer. qRT-PCR was then performed using the SYBR green kit. Primers for Lactobacillus-species specific primers were used for the following gene sequences: L. rhamnosus (GCC GAT CGT TGA CGT TAG TTG G; CAG CGG TTA TGC GAT GCG AAT), L. plantarum (GCT GGC AAT GCC ATC GTG CT; TCT CAA CGG TTG CTG TAT CG), L. paracasei (CAA TGC CGT GGT TGT TGG AA; GCC AAT CAC CGC ATT AAT CG), L. acidophilus (CCT TTC TAA GGA AGC GAA GGA T; ACG CTT GGT ATT CCA AAT CGC), B. breve (TCA TCA CGGCAA GGT CAA GA:GGC CAG AAC AGC TGG AAC AA), B. bifidum (CTG GCA GCC GTG ACA CTA CT, TGA ACT GGC CGT TAC GGT CT), B. lactis (ACC TCA CCA ATC CGC TGT TC, GAT CCG CAT GGT GGA ACT CT), 16S rRNA (TCCTACGGGAGGCAGCAGT; GGACTACCAGGGATCTAATCCTGTT), HPRT (CATAACCTGGTTCATCATCGC; TCCTCCTCAGACCGCTTTT).

Quantification and statistical analysis

Data are presented as the mean ± SD or SEM. Statistical differences were evaluated by one-way ANOVA followed by Tukey or Fisher LSD posthoc analysis, or by a two-tailed unpaired t-test with or without Welch’s correction. Statistical differences were evaluated on non-normal data using a nonparametric Kruskal-Wallis test. Principal coordinates analysis (PCoA) of bacterial beta-diversity based on the Bray-Curtis dissimilarity matrix using relative abundance was used to distinguish groups. Tumor-free survival data were analyzed by log rank Mantel–Cox test. Tumor volume data were analyzed by two-way ANOVA followed by Holm-Šídák’s posthoc analysis. Statistical significance was set at p ≤ 0.05.

Published: December 31, 2024