Estrogen promotes the growth of mammary tumour in mice undergoing mammary involution but not in nulliparous mice
We have reported previously that estrogen promotes ERα-negative mammary tumour growth in BALB/cAnNTac mice undergoing mammary involution [15]. In this study, we compared the effect of estrogen on tumour growth between Inv mice and age-matched Null. In contrast to the Inv mice, where estradiol valerate (E2V) stimulated tumour growth significantly, E2V treatment in Null mice had no significant effect on the weight of 4T1-Luc2 tumour (Figure 1). This suggests that tissue microenvironment during mammary involution is necessary for estrogen to stimulate the growth of ERα-negative mammary tumour. One of the factors in the tissue microenvironment is estrogen-stimulated neutrophil activity in the mammary gland [15]. In the following studies, we determined how estrogen influences the activities of circulating neutrophils.
Estrogen significantly reduces circulating neutrophils in tumour-bearing mice
OVX mice inoculated with tumour were treated with control vehicle (Ctrl) or E2V on alternative days. Blood samples were analysed on day 14 post-tumour cell inoculation. The levels of circulating neutrophils were evaluated by flow cytometry. The percentage of circulating neutrophils (CD45+CD11b+Ly6G+) doubled in mice with tumour compared to tumour-free mice under both Null and Inv conditions (Figure 2A). More interestingly, treatment with E2V resulted in marked reductions of the circulating neutrophils in both tumour-bearing Null and Inv mice. To confirm that the reduction of neutrophils percentage in response to E2V is due to decrease of absolute number of neutrophils, the number of neutrophils per ml blood was determined by cell counting. E2V reduced the number of neutrophils per ml by ~10 times (Figure 2B; Null+Ctrl, 35.2x107/ml; Null+E2V, 3.8x107/ml; Inv+Ctrl, 36.9x107/ml; Inv+E2V, 3.5x107/ml). There are also significant increases in the percentage of circulating monocytes (CD45+CD11b+Ly6Chi), CD4 T-cells (CD45+CD3e+CD4+), and CD8 T-cells (CD45+CD3e+CD8+) out of CD45+ cells in the tumour-bearing mice with E2V. We believe that these are likely due to decreases of the proportion of neutrophils (Figure 2C).
Estrogen-induced reduction of circulating neutrophils in tumour-bearing mice is due to reduction in LDN
Using Percoll gradient centrifugation, blood cells were separated into the mononuclear cells, granulocytes, and RBCs (Figure 3A). The LDN reported in various disease conditions are present in the mononuclear cell fraction. Flow cytometry analysis of the Percoll isolated mononuclear cell layer revealed an expected increase of circulating LDN (CD45+CD11b+Ly6G+) in both tumour-bearing Null and Inv mice as compared to their non-tumour counterpart. Interestingly, this increase in LDN in tumour-bearing mice was significantly reduced by E2V in both percentage and in number per ml (Figure 3B, 3C). Based on the number of LDN per ml of blood, this reduction was about 10-folds (Inv+Ctrl, 1.8x107/ml; Inv+E2V, 1.7x106/ml) in Inv and 25-folds (Null+Ctrl, 1.5x107/ml; Null+E2V, 5.9x105/ml) in Null mice (Figure 3Ci). Although there is a decreasing trend, E2V did not significantly decrease the number of HDN in both tumour-bearing Null and Inv mice (Figure 3Cii; Null, p=0.1858; Inv, p=0.1466). Morphological analysis showed that less than 25% of the isolated LDN are immature neutrophils with banded or ringed nucleus (Figure 3D, 3Ei), in contrast to the 65% immature neutrophils reported by Sagiv et al using the same syngeneic BALB/c-4T1 tumour model [7]. HDN fraction also had ~10% immature neutrophils (Figure 3Eii). In addition, E2V treatment did not significantly affect the distribution of mature and immature neutrophils in the circulating LDN despite causing marked down-regulation (Figure 3Ei).
We next evaluated whether the marked down-regulation of LDN is the accumulated effect of the prolonged (14 days) treatment. As similar estrogenic effect was observed in both tumour-bearing Inv and Null mice, we opted to evaluate the effect in the Inv mice only. OVX Inv mice were inoculated with tumour and treated with one dose of E2V at 48h prior to harvest at 14 days post-tumour inoculation. In contrast to that observed with a 14 days treatment, E2V induced a significant increase in the percentage of circulating neutrophils (Figure 3F). However, despite the overall increase, the number of circulating LDN was reduced significantly by 48h E2V treatment (Figure 3G, 3Hi). In contrast, the amount of HDN remain unchanged with E2V treatment (Figure 3Hii).
Taken together, estrogen treatment for 14 days in tumour-bearing mice resulted in substantial reduction in circulating neutrophils. This reduction is due to the reduction in LDN. Furthermore, this reduction in LDN appears to be accumulative as treatment with E2V for 48h resulted in a lesser reduction.
Estrogen-induced LDN reduction was associated with increased LDN cell death and reduced neutrophil production from bone marrow
We next determined the mechanism leading to the reduction of circulating LDN by E2V. OVX Inv mice bearing 4T1-Luc2 tumour were treated with either Ctrl or E2V for 48h and labelled with BrdU for 24h prior to sample collection. The percentages of BrdU+ cells in the circulating LDN, HDN, and bone marrow neutrophils were evaluated using BrdU incorporation assay [17]. After 48h E2V treatment, there were no significant differences in BrdU+ cells in the circulating LDN and HDN between Ctrl and E2V-treated groups (Figure 4A). However, a significant decrease in BrdU+ neutrophils (CD45+CD11b+Ly6G+BrdU+PI+) in the bone marrow was observed (Figure 4B). Newly synthesized neutrophils can remain in bone marrow for 4 – 6 days before moving to the circulation [18]. A reduction in mitotic neutrophils in the bone marrow suggests that there is a reduced LDN output.
LDN cell death was also determined by flow cytometry analysis of annexin V staining, which detect membrane phosphatidylserine of apoptotic cells on the outer surface [19]. Treatment with estrogen for 48h resulted in a significant increase in Annexin V+ Dead stain+ cells (Figure 4Ci). However, we did not detect a significant increase in the percentage of early apoptotic cells (Annexin V+ Dead stain-) (Figure 4Cii). It is possible that LDN apoptosis triggered after 48h E2V treatment has already reached the late stages (Annexin V+ Dead stain+).
It has been reported that estrogen treatment increased both blood and splenic Ly6C+/Ly6G+ neutrophils in mice with lung or mammary tumour [12]. We looked into the possibility that estrogen-induced decrease of circulating LDN is resulted from the diversion of neutrophils into the spleen or tumour. We found ~50-folds increase in the percentage of splenic Ly6G+ neutrophils in tumour-bearing mice compared to tumour-free Null mice. This increase is greater than the increase in circulating neutrophils. With 14 days treatment of E2V, we observed a ~50% reduction of splenic neutrophils (CD45+CD11b+Ly6G+) in both tumour-bearing Null and Inv mice (Figure 4D). This 50% reduction in neutrophils is higher than the percentage of reduction of LDN in the blood (~30% and 15% in Null and Inv mice, respectively).
There is a slight increase of tumour-associated neutrophils (CD45+CD11b+Ly6G+) in Null mice and decrease in Inv mice with E2V treatment, but the changes are not statistically significant (Figure 4E). The data indicate that the observed reduction in circulating LDN by E2V was not due to increased mobilization into the tumour.
Estrogen increases the expression of pro-tumoral markers in both circulating LDN and HDN of Inv mice
Accumulation of LDN has been described in both the murine model and in breast cancer patients [7, 20, 21]. LDN are generally believed to be pro-tumoral, while its counterpart, the HDN was generally anti-tumoral. LDN’s pro-tumoral action mainly comes from its immune suppressive nature [8]. Since estrogen is known to elicit varying immunomodulating effect [22], we evaluated the effect of estrogen on the gene expression of known pro-tumoral genes, arginase-1 (Arg1), interleukin 1b (Il1b), and C-C motif chemokine ligand 5 (Ccl5) in neutrophils by qPCR. E2V treatment for 48h had no significant effect on the expression of Arg1, Il1b, and Ccl5 in both LDN and HDN (Figure 5A). However, E2V treatment for 14 days resulted in increased expression of Arg1, Il1b in LDN and HDN of Inv mice, while Ccl5, Arg1 were significantly upregulated in the LDN and HDN of Null mice (Figure 5B). In non-tumour bearing Inv mice, only HDN were analysed because there were very few circulating LDN for isolation in non-tumour-bearing mice. E2V treatment for 14 days did not significantly affect the expression of these genes in neutrophils (Figure 5B). Intriguingly, the relative expressions of all these genes are substantially higher in neutrophils of tumour-free mice than tumour-bearing mice. This suggests that tumour may alter the epigenetics of circulating neutrophils.
Transforming growth factor beta (Tgfb) has been reported to have a pro-tumoral role in cancer development by promoting tumour metastasis and immune evasion [23]. Furthermore, it was also discovered that Tgfb can polarize tumour-associated neutrophil towards the pro-tumour phenotype [2, 24]. In this study, we observed that estrogen treatment for 14 days resulted in a significant increase in Tgfb1 gene expression in both LDN and HDN of tumour-bearing Inv mice while having no effect in tumour-bearing Null mice. Interestingly, in non-tumour Inv mice, similar treatment instead led to a significant reduction of Tgfb1 expression (Figure 5B). Like the other genes tested, treatment for 48h had no effect on Tgfb1 expression in both LDN and HDN (Figure 5A).
Taken together, one dose of E2V did not significantly influence the gene expression of circulating neutrophils, but estrogen exposure for 14 days enhanced the expression of putative pro-tumoral genes Arg1, Il1b and Tgfb1 in Inv mice, potentially contributing to the increased tumour growth observed only in Inv mice (Figure 1). In contrast, only Ccl5 and Arg1 is upregulated by E2V in Null mice.