Variable Intrinsic Expression of Immunoregulatory Biomarkers in Breast Cancer Cell Lines, Mammospheres, and Co-Cultures

Advances in immunotherapy have increased interest in knowing the role of the immune system in breast cancer (BC) pathogenesis. Therefore, immune checkpoints (IC) and other pathways related to immune regulation, such as JAK2 and FoXO1, have emerged as potential targets for BC treatment. However, their intrinsic gene expression in vitro has not been extensively studied in this neoplasia. Thus, we evaluated the mRNA expression of tumor-cell-intrinsic CTLA-4, PDCD1 (PD1), CD274 (PD-L1), PDCD1LG2 (PD-L2), CD276 (B7-H3), JAK2, and FoXO1 in different BC cell lines, derived mammospheres, and co-cultures with peripheral blood mononuclear cells (PBMCs) by real-time quantitative polymerase chain reaction (qRT-PCR). Our results showed that intrinsic CTLA-4, CD274 (PD-L1), and PDCD1LG2 (PD-L2) were highly expressed in triple-negative cell lines, while CD276 was predominantly overexpressed in luminal cell lines. In contrast, JAK2 and FoXO1 were under-expressed. Moreover, high levels of CTLA-4, PDCD1 (PD1), CD274 (PD-L1), PDCD1LG2 (PD-L2), and JAK2 were found after mammosphere formation. Finally, the interaction between BC cell lines and peripheral blood mononuclear cells (PBMCs) stimulates the intrinsic expression of CTLA-4, PCDC1 (PD1), CD274 (PD-L1), and PDCD1LG2 (PD-L2). In conclusion, the intrinsic expression of immunoregulatory genes seems very dynamic, depending on BC phenotype, culture conditions, and tumor-immune cell interactions.


Introduction
Breast cancer (BC) is the most common cancer in women and the leading cause of cancer-related death in this gender worldwide [1]. Although clinical management has improved over the last few years, the efficacy of traditional therapeutic strategies such as surgery, radiation, endocrine therapy, and chemotherapy depends on the patient and the intrinsic tumor subtype. The latter supports that BC is a heterogeneous disease and explains why it behaves more aggressively, leading to recurrences, drug resistance, metastasis, and a worse prognosis [2][3][4]. Therefore, there is a need to identify new therapeutic targets to improve BC patients' treatment and outcomes. in neoplastic cells. In addition, they also participate through intrinsic functions in the proliferation, migration, and invasion of neoplastic cells [25,26], regulation of epithelialto-mesenchymal transition (EMT) [26,27], acquisition of stem cell characteristics [27][28][29], inhibition of apoptosis, and resistance to chemotherapy [30].
Of note, ICs are not the only proteins that regulate the immune response. Other molecules, such as JAK2 and FoXO1, also play an essential role. Thus, JAK2 is a tyrosine kinase protein that, together with signal transducer and activator of transcription (STATs) proteins, is responsible for initiating the transcription of different target genes involved in various cellular processes [41]. In addition, JAK2 has also been implicated in activating critical signaling pathways such as the PI3K/Akt or RAS/RAF/MAPK [42]. Its expression in neoplastic cells has been associated with poor prognosis and clinicopathological features [43,44].
Further, it has been directly involved in cell proliferation, migration and invasion, angiogenesis, resistance to treatment and apoptosis, and the maintenance of stemness and EMT in different neoplasms [45][46][47]. It is also essential for PD-L1 and PD-L2 expression in neoplastic cells and for promoting non-immunological functions after interacting with B7-H3 [37,48,49]. On the other hand, FoXO1 is a protein belonging to the FoXO subfamily of transcription factors, which also includes the FoXO3, FoXO4, and FoXO6 proteins [50]. Their primary function is to transcribe different genes involved in various physiological processes [51], such as the migration of dendritic cells [52], the production of inflammatory mediators, and the polarization of monocyte-macrophage cells [53], the maturation of B cells, and Tregs [54], as well as the formation and subsequent function of memory T cells [55]. Likewise, its expression in neoplastic cells is an independent factor of good prognosis [56] and limits the migration and invasion of neoplastic cells by inhibiting the EMT process [57]. Thus, FoXO1 is considered by many authors as a tumor suppressor protein. However, there are contradictory data since, depending on the type of neoplasm and the tumor microenvironment, it could also act as an oncogene, promoting cell proliferation and drug resistance in different neoplasms [58,59].
Therefore, there is a need to analyze the role of ICs, JAK2, and FoXO1 in regulating the immune response. Currently, the potential involvement of these molecules in the pathogenesis of BC has not been extensively investigated in different phenotypes, and data from in vitro studies and stem cell conditions are limited.

Intrinsic Gene Expression in BC Cell Lines
We detected that healthy mammary epithelial cell line 184A1 did not express CTLA-4 and PDCD1 (PD1) receptors, so the ∆Ct value was used for the analysis. In this case, the ∆Ct value acted as a relative gene expression value, where low values are indicative of higher gene expression. We observed a significant increase in CTLA-4 in MDA-MB-231 and MDA-MB-468 compared to MCF-7 (both p = 0.025) and BT-474 (p = 0.009 and p = 0.024, respectively), but not against SK-BR-3 (p = 0.193 and p = 0.385, respectively). In addition, CTLA-4 expression was not detected in T-47D ( Figure 1A). Regarding PDCD1 (PD1), the expression remained stable in all cell lines, with no significant differences found among them (all p > 0.05) ( Figure 1A).

Intrinsic Gene Expression in BC Cell Lines
We detected that healthy mammary epithelial cell line 184A1 did not express CTLA-4 and PDCD1 (PD1) receptors, so the ΔCt value was used for the analysis. In this case, the ΔCt value acted as a relative gene expression value, where low values are indicative of higher gene expression. We observed a significant increase in CTLA-4 in MDA-MB-231 and MDA-MB-468 compared to MCF-7 (both p = 0.025) and BT-474 (p = 0.009 and p = 0.024, respectively), but not against SK-BR-3 (p = 0.193 and p = 0.385, respectively). In addition, CTLA-4 expression was not detected in T-47D ( Figure 1A). Regarding PDCD1 (PD1), the expression remained stable in all cell lines, with no significant differences found among them (all p > 0.05) ( Figure 1A).  CD274 (PD-L1) and PDCD1LG2 (PD-L2) were also increased in MDA-MB-231 compared to 184A1, which was only statistically significant for CD274 (PD-L1) (p = 0.019). In the other cell lines studied, both genes were found to be significantly under-expressed (all p < 0.014) ( Figure 1B). In addition, it is important to highlight the absence of PDCD1LG2 (PD-L2) expression in BT-474. Among the tumor cell lines, MDA-MB-231 presented the highest levels of CD274 (PD-L1) and PDCD1LG2 (PD-L2) expression compared to all cell lines (all p < 0.014). Likewise, an under-expression of both genes was observed in MCF-7, T-47D, and BT-474 cell lines concerning SK-BR-3 (all p < 0.020).

Intrinsic Gene Expression in Mammospheres
Different BC cell lines were cultured under non-adherent conditions to obtain mammospheres. Enriched in cancer stem cells, these spheres are characterized by a CD44 high /CD24 low phenotype in BC [60]. Therefore, both gene expression levels were measured to verify that the generation of mammospheres had been performed correctly. In this regard, all cases showed increased CD44 expression and normal or decreased CD24 expression compared to controls ( Figure S1). However, we failed to generate mammospheres from the MDA-MB-468 cell line, so it was not considered for further analysis. Once CD44 and CD24 expression were verified, the relative gene expression of our genes of interest was quantified.
Regarding CTLA-4 and CD274 (PD-L1) expression, we observed that both genes were significantly overexpressed in the mammospheres of BT-474 and SK-BR-3 cell lines (all p < 0.029) compared to controls. This increase also occurred in the mammospheres of the MCF-7 cell line but was not statistically significant. In addition, an increase in CTLA-4 CD274 (PD-L1) and PDCD1LG2 (PD-L2) were also increased in MDA-MB-231 compared to 184A1, which was only statistically significant for CD274 (PD-L1) (p = 0.019). In the other cell lines studied, both genes were found to be significantly under-expressed (all p < 0.014) ( Figure 1B). In addition, it is important to highlight the absence of PDCD1LG2 CD276 (B7-H3) expression was significantly higher in MCF-7 (p = 0.049) and BT-474 (p = 0.021) cell lines. A significant decrease was also observed in SK-BR-3 (p = 0.003) compared to normal cell line 184A1 ( Figure 1C). When comparing the neoplastic lines, a significant increase was detected in MCF-7 and BT-474 compared to SK-BR-3 (all p < 0.028).

Intrinsic Gene Expression in Mammospheres
Different BC cell lines were cultured under non-adherent conditions to obtain mammospheres. Enriched in cancer stem cells, these spheres are characterized by a CD44 high /CD24 low phenotype in BC [60]. Therefore, both gene expression levels were measured to verify that the generation of mammospheres had been performed correctly. In this regard, all cases showed increased CD44 expression and normal or decreased CD24 expression compared to controls ( Figure S1). However, we failed to generate mammospheres from the MDA-MB-468 cell line, so it was not considered for further analysis. Once CD44 and CD24 expression were verified, the relative gene expression of our genes of interest was quantified.
Regarding CTLA-4 and CD274 (PD-L1) expression, we observed that both genes were significantly overexpressed in the mammospheres of BT-474 and SK-BR-3 cell lines (all p < 0.029) compared to controls. This increase also occurred in the mammospheres of the MCF-7 cell line but was not statistically significant. In addition, an increase in CTLA-4 expression (p < 0.029) and a decrease in CD274 (PD-L1) expression (p < 0.05) were observed in MDA-MB-231 mammospheres. Finally, in T-47D mammospheres, CD274 (PD-L1) expression was significantly under-expressed (p < 0.05), while CTLA-4 was not detected ( Figure 2A). After comparing the mammospheres with each other, we found that those derived from the MDA-MB-231 line showed the highest levels of CTLA-4. However, this difference was only significant concerning SK-BR-3 mammospheres (p = 0.035). On the other hand, CD274 (PD-L1) was significantly under-expressed in T-47D and MDA-MB-231 mammospheres concerning the rest (in all cases, p < 0.042).  Regarding PDCD1 (PD1) and PDCD1LG2 (PD-L2) expression, there were statistically significant differences between mammospheres obtained from MCF-7 and MDA-MB-231 cell lines (all cases p < 0.037) and controls. However, the levels remained close to normal in the rest of the mammospheres ( Figure 2B). Furthermore, as in the control cells, PDCD1LG2 (PD-L2) was not expressed in BT-474 mammospheres. When comparing the mammospheres, it was seen that PDCD1 (PD1) expression in MDA-MB-231 mammospheres was significantly higher than in the rest (all p = 0.021). At the same time, PDCD1LG2 (PD-L2) levels were significantly lower in SK-BR-3 mammospheres versus those derived from MCF-7 and MDA-MB-231 (both p < 0.018).
As can be seen in Figure 2C, CD276 (B7-H3) was under-expressed in mammospheres obtained from BT-474, T-47D, and SK-BR-3 cell lines, being significant in the latter two cases (both p < 0.025) compared to controls. On the other hand, the remaining mammospheres showed CD276 (B7-H3) levels close to the controls. Likewise, it was observed that the mammospheres derived from the MDA-MB-231 cell line presented a significantly higher expression (all p = 0.05) than the rest, except those derived from MCF-7 (p = 0.513).
JAK2 expression was remarkably under-expressed in T-47D and BT-474 mammospheres (both p = 0.014) and over-expressed in SK-BR-3 and MDA-MB-231 (both p = 0.014) compared to the controls ( Figure 2C). Furthermore, when comparing the results, we found that mammospheres derived from SK-BR-3 and MDA-MB-231 cell lines had significantly higher JAK2 levels than the rest (all cases p = 0.021).
Finally, FoXO1 remained statistically underexpressed in most mammospheres (all cases p < 0.026) compared to controls. The only exception was the MDA-MB-231 mammospheres that maintained a near-normal expression ( Figure 2C). After reaching the results between mammospheres, we found that those derived from the MDA-MB-231 cell line showed a significant overexpression (all p < 0.033). In contrast, those derived from T-47D showed a significant underexpression (all p < 0.020).

Intrinsic Gene Expression in Co-Cultures of BC Cell Lines and PBMCs
We developed a co-culture model with PBMCs derived from BC patients and commercial cell lines, representing the different immunophenotypes. We aimed to determine the effect of direct interaction between tumor cells and immune cells on the intrinsic expression of immunoregulatory gene expression. CTLA-4 and PDCD1 (PD1) increased expression was seen in all co-cultures compared with control cell lines ( Figure 3A,B). No significant differences in CTLA-4 expression were found for any co-culture on the different days the analysis was carried out. However, we observed that the T-47D cell line expressed this gene after interaction with PBMCs, thus representing the ∆Ct values ( Figure 3A). Regarding PDCD1 (PD1), the differences in expression observed between the second and third day in the T-47D cell line and those between the first and second day for the MDA-MB-468 cell line were statistically significant (both p = 0.043) ( Figure 3B).
Similar results were observed for CD274 (PD-L1) and PDCD1LG2 (PD-L2), showing an increase ( Figure 3C,D). Likewise, we found significant differences in CD274 (PD-L1) levels during the first day of interaction with PBMCs in the SK-BR-3 cell line compared to the second and the third day on which the measurements were carried out (p = 0.037 and 0.027, respectively) ( Figure 3C). On the other hand, PDCD1LG2 (PD-L2) was significantly elevated on the second day of interaction for the T-47D and MDA-MB-231 cell lines compared with the first day of measurements (both p = 0.043) ( Figure 3D). In addition, we observed the expression of PDCD1LG2 (PD-L2) in BT-474 cells, with significant differences between the ∆Ct values on the second and third day (p = 0.043). In contrast, CD276 (B7-H3) and JAK2 levels in co-cultures did not differ from those obtained in the cell lines used as controls ( Figure 3E,F). However, in the case of the MDA-MB-468 cell line, a significant increase of both genes was observed 72 h after interaction with PBMCs compared to the other days (all p < 0.027). Finally, FoXO1 expression remained at near-normal levels in all co-cultures. Similar results were observed for CD274 (PD-L1) and PDCD1LG2 (PD-L2), showing an increase ( Figure 3C,D). Likewise, we found significant differences in CD274 (PD-L1) levels during the first day of interaction with PBMCs in the SK-BR-3 cell line compared to the second and the third day on which the measurements were carried out (p = 0.037 and 0.027, respectively) ( Figure 3C). On the other hand, PDCD1LG2 (PD-L2) was significantly elevated on the second day of interaction for the T-47D and MDA-MB-231 cell lines compared with the first day of measurements (both p = 0.043) ( Figure 3D). In addition, we observed the expression of PDCD1LG2 (PD-L2) in BT-474 cells, with significant differences between the ΔCt values on the second and third day (p = 0.043). In contrast, CD276 (B7-H3) and JAK2 levels in co-cultures did not differ from those obtained in the cell lines used as controls ( Figure 3E,F). However, in the case of the MDA-MB-468 cell line, a significant increase of both genes was observed 72 h after interaction with PBMCs compared to the other days (all p < 0.027). Finally, FoXO1 expression remained at nearnormal levels in all co-cultures. Nevertheless, we observed significant differences between the second and the third day of interaction in the co-cultures of the MCF-7 cell line (p = 0.030). We also found differences between the first and the second day (p = 0.002), as well as the second and the third day of interaction (p = 0.036) in the co-cultures of BT-474 ( Figure 3G).

Discussion
In the current study, we analyzed the intrinsic mRNA expression of several immunoregulatory genes, such as CTLA

Discussion
In the current study, we analyzed the intrinsic mRNA expression of several immunoregulatory genes, such as CTLA  [19,[61][62][63], but not in BT-474, SK-BR-3, and MDA-MB-468 cell lines. Likewise, it has been shown that its expression in neoplastic BC cells can suppress dendritic cell maturation in vitro [61]. In addition, an independent study confirmed that CTLA-4-CD80 binding is strong enough to internalize CD80 inside the neoplastic cells, thus attenuating the ability of APCs to activate T lymphocytes [63].
Intrinsic CTLA-4 expression has also been linked to BC cell proliferation, differentiation, survival [62], and the stem-cell-like phenotype in different neoplasms [64]. Of note, its essential role in maintaining pluripotency and the ability to form mammospheres in vitro has been previously described in melanoma cells [65]. Nevertheless, cancer-cell-intrinsic CTLA-4 expression has never been previously analyzed in mammospheres. According to the current results, CTLA-4 expression on neoplastic cells may be essential for the pathogenesis of BC and the modulation of the tumor microenvironment, especially in triple-negative BC. In addition, our results suggest that its expression might be implicated in acquiring and maintaining tumor stem cell characteristics in aggressive BC phenotypes.
We also demonstrated that PDCD1 (PD1) maintained a similar mRNA level for all analyzed cell lines, with no significant differences. However, it was overexpressed in mammospheres derived from MCF-7 and MDA-MB-231 cell lines. This increase was significant in the rest of the MDA-MB-231 mammospheres. To our knowledge, this is the first study to characterize the intrinsic expression of PDCD1 (PD1) in BC cell lines and mammospheres without an immunological environment. Of note, its expression has been evaluated in other neoplasms, where in most cases, it has been described as a pro-oncogenic factor. In this regard, several independent studies have confirmed that the intrinsic expression of PD1 in melanoma, hepatocellular carcinoma, and bladder cancer cell lines promotes neoplastic growth without an adaptative immune response [66][67][68][69]. Likewise, its overexpression in pancreatic ductal adenocarcinoma cell lines was associated with increased tumor growth after cell injection in immunosuppressed mice [70]. In contrast, PD1 overexpression in the murine cell line M109 showed a significant decrease in viability and cell proliferation [71]. These data demonstrate that the intrinsic expression and function of PD1 in neoplastic cells are complex and not yet fully defined.
CD274 (PD-L1) and PDCD1LG2 (PD-L2) were intrinsically overexpressed in the triplenegative MDA-MB-231 cell line, although only significant for CD274 (PD-L1). The rest of the cell lines showed underexpressed levels compared to controls. In addition, the absence of PDCD1LG2 (PD-L2) expression in the BT-474 line was noteworthy. In agreement with our results, numerous authors have observed a significant increase of both ligands in the MDA-MB-231 cell line compared with luminal cell lines, both at the mRNA and protein levels [72][73][74][75]. On the other hand, in partial agreement, decreased mRNA PDCD1LG2 (PD-L2) has also been reported in MDA-MB-231, SK-BR-3, and MCF-7 compared to the nonmalignant MCF10A cell line by others [76]. Therefore, it can be stated that the expression of both genes varies considerably between the different BC phenotypes. Similarly, there is evidence that these genes are expressed synchronously in neoplastic cells [75], and there may be changes in response to endogenous and exogenous factors [72].
Concerning intrinsic functions, PD-L2 has been less studied in cell lines, whereas PD-L1 has been implicated in resistance to chemo-and radiotherapy in the MDA-MB-231 cell line [30] and apoptosis, as well as induction of cell proliferation, migration, and invasion processes in gastric cancer and pancreatic carcinoma cell lines [25]. These data support the idea that intrinsic CD274 (PD-L1) expression in neoplastic cells favors immune response evasion and increases neoplastic aggressiveness. Their gene expression levels have also been related to tumor stemness maintenance in different neoplasms [27,77]. We found that intrinsic CD274 (PD-L1) was significantly overexpressed in mammospheres derived from HER2-expressing cell lines (BT-474 and SK-BR-3). In contrast, compared with controls, intrinsic PDCD1LG2 (PD-L2) was upregulated in those derived from MCF-7 and MDA-MB-231. In agreement with our results, previous investigators demonstrated intrinsic PD-L2 expression by flow cytometry in MCF-7-derived mammospheres with no significant differences concerning the control cell line [78].
In contrast, a high intrinsic PD-L1 expression at the mRNA and protein level has been described in mammospheres derived from MCF-7 and MDA-MB-231 cell lines [29,78]. Furthermore, its expression has been reported to be necessary to maintain the expression of transcription factors related to stem cell phenotypes such as OCT-4A, Nanog, and BMI1 [28,29]. Likewise, an in silico study performed in the MDA-MB-231 cell line revealed an association between intrinsic PD-L1, the EGFR signaling pathway, and several factors involved in the EMT process and tumor stemness [79]. At the same time, other investigators have related the expression with the activation of Notch and/or PI3K/AKT signaling pathways [80]. Therefore, our data further support the essential role of intrinsic CD274 (PD-L1) and PDCD1LG2 (PD-L2) in maintaining BC's tumor stem cell phenotype.
We detected that intrinsic CD276 (B7-H3) was overexpressed in luminal cell lines (MCF-7 and BT-474) and underexpressed in HER2-enriched (SK-BR-3) compared to the control. In the remaining cell lines, the expression was close to normal. Conversely, BT-474, T-47D, and SK-BR-3-derived mammospheres showed decreased expression, while those derived from MCF-7 and MDA-MB-231 presented levels close to normal. The intrinsic expression of this immune checkpoint has been previously described in MCF-7 and T-47D cell lines in variable levels [31]. Furthermore, regarding its function, it has been observed that silencing its expression in the MCF-7 cell line results in an increase in vascular endothelial growth factor (VEGF), both at the mRNA and protein levels [33]. These results suggest that intrinsic B7-H3 may be involved in controlling neoplastic cell growth. How-ever, independent studies confirm that its expression in the MDA-MB-231 cell line not only favors tumor growth but also promotes the Warburg effect [81] and induces resistance to paclitaxel [82] and inhibitory molecules of the PI3K/AKT/mTOR pathway [83] through activation of JAK2/STAT3. Therefore, all these data reflect that the function of intrinsic B7-H3 is not yet fully defined in BC, as it appears to play a dual role in the pathogenesis of the molecular subtypes.
Furthermore, recent studies have shown that intrinsic CD276 (B7-H3) expression in neoplastic cells plays a crucial role in the induction of the EMT process and acquiring stem cell phenotype. For example, its intrinsic expression in colorectal cancer cell lines decreased E-cadherin expression and increased stem cell-associated markers such as N-cadherin, vimentin, CD133, CD44, and OCT4 [84]. Likewise, other authors have demonstrated its association with neural stem cell markers in glioblastomas [36] and the capacity to form tumorspheres [85]. In addition, there is evidence that intrinsic CD276 (B7-H3) is a critical factor in evading the immune response by tumor stem cells in head and neck carcinoma [86]. In BC, this gene has been associated with gene signatures related to increased cell invasion and expression of stem cell markers and signaling pathways involved in pluripotency [87]. Indeed, an independent study demonstrated that B7-H3 expression increased the number of BC stem cells through the major vault protein (MVP), thus promoting tumor development [35]. In contrast, our data do not support that B7-H3 is directly involved in maintaining the tumor stem cell properties in BC cell lines.
Intrinsic  [88]. They also found overexpression of JAK2 in the MCF-7 cell line, in contrast to our data.
Regarding its function, it is directly involved in the proliferation and aggressiveness of MDA-MB-231 and MDA-MB-468 cell lines [89]. Likewise, it would be involved in angiogenesis, cell proliferation, migration, and invasion in the tamoxifen-resistant (TAMR)MCF-7 cell line [47]. Concerning mammospheres, significantly decreased levels were observed in those derived from T-47D and BT-474, with opposite results for SK-BR-3 and MDA-MB-231. In agreement, prior studies have detected that JAK2 inhibition in the MDA-MB-231 cell line reduces the number of CD44 high /CD24 low cells and inhibits the formation of mammospheres [90,91]. Similar results were also seen in the cell lines MCF-7/ADR, resistant to doxorubicin, and MCF-7/SC, a cell line that, compared to MCF-7, shows high invasiveness and stem cell properties [91,92]. However, it should be noted that these results are not new, as previously, Marotta et al. demonstrated the involvement of the IL-6/JAK2/STAT3 axis in the growth and survival of BC stem cells [93]. Moreover, JAK2 activation is involved in Twist-1 and Twist-2 overexpression, related to the EMT process and the acquisition of stem cell phenotype [94]. In like manner, increased copy number or amplification of JAK2 in triple-negative cell lines favored mammospheres' formation and increased chemotherapy resistance [88]. This property was also observed after activating the JAK2/STAT3 pathway via PAK1 [95]. Thus, these data suggest that JAK2 expression is essential for mammosphere formation, maintenance of the stem cell phenotype in BC, and resistance to conventional treatments.
We demonstrated an intrinsic underexpression of FoXO1 in the most aggressive cell lines (BT-474, SK-BR-3, and MDA-MB-231). In addition, Liang et al. reported in the MDA-MB-231 cell line that it promotes cell proliferation but not migration or apoptosis [96]. However, luminal A cell lines (MCF-7 and T-47D) and MDA-MB-468 showed expression values close to normal. Of note, Gutilla et al. have found in the MCF-7 a large discrepancy between FoXO1 mRNA and protein levels, suggesting a robust translational control [97].
We also detected a decrease in FoXO1 expression in all analyzed mammospheres, except those from the MDA-MB-231 cell line. Similar results have been previously reported in MCF-7, SK-BR-3, and MDA-MB-231 [96,98,99]. Despite this, FoXO1 inhibits EMT and metastasis in the MCF-7 cell line [99]. Moreover, it is involved in the resistance mechanisms of tamoxifen [100] and doxorubicin [101] in TAMR-MCF-7 and MCF-7/ADR cell lines, respectively. In line with recent studies, we showed a positive correlation between FoXO1 underexpression and mammosphere formation in the MCF- [102].
In contrast, Truong et al. demonstrated that FoXO1 expression in MCF-7, T-47D, and BT-474 was directly related to progesterone receptor phosphorylation, inhibition of cell proliferation, and mammosphere formation [103]. These contradictory results may be because the cell lines were supplemented with hormones. Nevertheless, discrepancies remain regarding the function of this transcription factor. Indeed, another independent research showed that FoXO1 accumulation in BC cells promoted the expression of SOX2, a transcription factor related to the tumor stem cell phenotype [104]. The previous data suggest that FoXO1 may play a critical role in BC pathogenesis and drug resistance development. Although the published data are not entirely clear, our results show the relevance of FoXO1 in the maintenance of the tumor stem cell phenotype.
Finally, we observed in all co-cultures a significant increase in CTLA-4, PCDC1 (PD1), CD274 (PD-L1), and PDCD1LG2 (PD-L2) intrinsic expression after a direct interaction with PBMCs. Furthermore, T-47D and BT-474 cell lines recovered cell-intrinsic CTLA-4 and PDCD1LG2 (PD-L2) expression after tumor-immune cell interactions, respectively. However, CD276 (B7-H3), JAK2, and FoXO1 maintained their intrinsic expression levels close to normal, even after interaction with PBMCs. Similar results were already reported for PD-L1 by Hinterneder et al. [105], who described that direct interaction of the HCC38 cell line with activated PBMCs increased intrinsic PD-L1 protein expression. Likewise, co-culture of melanoma metastasis and their autologous TILs induced the expression of several immunomodulatory molecules, including PD-L1 or indoleamine 2,3-dioxygenase (IDO), after 24 h of interaction [106]. In this regard, these gene expressions in neoplastic cells can respond to various cytokines and growth factors in the tumor microenvironment. Indeed, IFN-γ, produced by TILs after antigenic recognition, is one of the most potent extrinsic inducers for PD-L1 and PD-L2 expression in neoplastic cells [107]. Furthermore, intrinsic PD1 expression in ovarian cancer in response to IFN-α and IFN-γ has also been described [69], although these results could not be confirmed in non-small-cell lung cancer [23]. For its part, intrinsic CTLA-4 expression in melanoma cells would also be IFN-γ-dependent [108]. Despite this, a recent study has shown that the production of soluble factors is insufficient to inhibit immune cells, which are necessary for the interaction between tumor and stromal cells [109]. Here, we demonstrated the relevance of the interaction between neoplastic and immune cells in inducing the intrinsic expression of several immunoregulatory genes as an adaptive mechanism against active antineoplastic response. Our results also showed a time-dependent variation in the BC cell line's intrinsic gene expression. In fact, for the MCF-7 cell line,  [61].
Similarly, overexpression of PD-L1 and PD-L2 in neoplastic cells would directly inhibit effector T lymphocytes [16]. Our experimental data suggest that intrinsic expression of the immunoregulatory genes is dynamic and highly dependent on the BC cell line subtypes and the inflammatory context. Nevertheless, these results are not possible to corroborate in clinical series. Therefore, co-cultures are a valuable tool for better understanding all these processes. However, further in vivo studies are necessary to confirm our results and elucidate immunoregulatory genes' role in BC pathogenesis. . Cells were incubated at 37 • C in a humidified atmosphere with 5% of CO 2 and were grown to around 80% confluence before RNA isolation.

Isolation of Human PBMCs
Peripheral blood was collected from 30 BC patients with informed consent to use their samples for research purposes and ethical approval from the ethics committee (ethic code PI2019/089 and PI2020-265) at our institution. The selected patients represented each BC immunophenotype (10 luminal A, five luminal B/HER2-positive, 5 HER2-enriched, and ten triple-negative patients). PBMCs were isolated by density gradient centrifugation. Briefly, 3 mL of Biocoll Separating Solution (Biochrom Ltd., Cambridge, United Kingdom) and 6 mL of peripheral blood samples diluted in phosphate-buffered saline (PBS) were placed in sterile tubes centrifuged at 1300× g for 15 min. Further, the PBMCs were collected from the interface between plasma and Biocoll, washed twice, and cultured in Roswell Park Memorial Institute (RPMI) 1640 medium with glutamine (Capricorn Scientific GmbH, Ebsdorfergrund, Germany) and supplemented with 10% FBS (Capricorn Scientific GmbH) and 1% antibiotics (50 U/mL penicillin and 50 mg/mL streptomycin) (Capricorn Scientific GmbH). PBMCs were incubated for 2 h at 37 • C and 5% CO 2 , allowing monocyte separation by adhering to the plastic of the culture flask [110]. After this time, PBMCs were collected and assessed for viability using 0.4% trypan blue. Finally, PBMCs were brought to the desired concentration for co-culturing.

Co-Culture of PBMCs with BC Cell Lines
For co-cultures, all tumor cell lines were cultured in RPMI 1640 medium with glutamine (Capricorn Scientific GmbH), supplemented with 10% FBS (Biowest), 50 U/mL of penicillin, and 50 mg/mL of streptomycin (Biowest). The cell lines were seeded in sterile culture flasks and incubated in a humid atmosphere at 37 • C and 5% CO 2 . Once cultures reached 70-80% confluence, tumor cells were seeded in duplicate at 100,000 cells/well in sterile 6-well plates (SPL Life Sciences). After 24 h, the isolated PBMC were co-cultured in the 6-well plates with the pre-seeded tumor cells at a concentration of 500,000 cells/well. The above was performed considering the phenotype of the cell lines and the BC patients who donated the peripheral blood sample. The co-cultures were then maintained in a humid environment at 37 • C and 5% CO 2 ( Figure 5). After co-culture for 24, 48, and 72 h, the supernatant with PBMCs was collected, and tumor cell lines were washed twice with PBS (Biowest). Before RNA isolation, tumor cells were detached using trypsinethylenediaminetetraacetic acid (EDTA) 0.05% (Capricorn Scientific GmbH), washed, and resuspended in PBS.

Isolation of Human PBMCs
Peripheral blood was collected from 30 BC patients with informed consent to use their samples for research purposes and ethical approval from the ethics committee (ethic code PI2019/089 and PI2020-265) at our institution. The selected patients represented each BC immunophenotype (10 luminal A, five luminal B/HER2-positive, 5 HER2-enriched, and ten triple-negative patients). PBMCs were isolated by density gradient centrifugation.

Mammosphere Culture
Mammosphere culture of human BC cell lines were performed using a commercial MammoCultTM medium (Stemcell Technologies, Vancouver, Canada) supplemented with 10% proliferation supplements (Stemcell Technologies), 0.48 mg/mL of hydrocortisone (Stemcell Technologies), 4 mg/mL of heparin (Stemcell technologies), 50 U/mL of penicillin, and 50 mg/mL of streptomycin (Biowest). Single T-47D, MCF-7, BT-474, SK-BR-3, and MDA-MB-231 cells were seeded in six-well plates coated with poly-2-hydroxyethyl-methacrylate (Sigma-Aldrich, St Louis, MO, USA) to prevent cell attachment at a density of 5 × 10 3 cells per well. Under these conditions, the cells grew as a non-adherent spherical cluster (Figure 4). Every two days, 2 mL of fresh medium was added to each well. After seven days in culture, mammospheres were harvested by gentle centrifugation

Mammosphere Culture
Mammosphere culture of human BC cell lines were performed using a commercial MammoCultTM medium (Stemcell Technologies, Vancouver, BC, Canada) supplemented with 10% proliferation supplements (Stemcell Technologies), 0.48 mg/mL of hydrocortisone (Stemcell Technologies), 4 mg/mL of heparin (Stemcell technologies), 50 U/mL of penicillin, and 50 mg/mL of streptomycin (Biowest). Single T-47D, MCF-7, BT-474, SK-BR-3, and MDA-MB-231 cells were seeded in six-well plates coated with poly-2-hydroxyethylmethacrylate (Sigma-Aldrich, St Louis, MO, USA) to prevent cell attachment at a density of 5 × 10 3 cells per well. Under these conditions, the cells grew as a non-adherent spherical cluster ( Figure 4). Every two days, 2 mL of fresh medium was added to each well. After seven days in culture, mammospheres were harvested by gentle centrifugation and enzymatically dissociated with trypsin-EDTA 0.05% in PBS w/o calcium w/o magnesium w/o phenol (Biowest) before RNA isolation. All experiments were performed in triplicate.

RNA Isolation and Complementary DNA Synthesis
Single-cell suspensions obtained from cell line cultures, co-cultures, and mammospheres were used for RNA isolation. Cells were lysed in 600 µL of RLT buffer (Qiagen, Hilden, Germany) supplemented with 1% β-mercaptoethanol (Merk Schuchardt OHG, Hohenbrunn, Germany) through a 29G needle. Following the RNeasy Mini kit protocol (Qiagen,), 70% ethanol was added to cell lysates, and the samples were mixed by pipetting before being transferred to the columns. Total RNA was extracted according to the manufacturer's protocol with on-column DNase digestion. RNA concentration and purity were assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and optical density ratios of 260/230 nm and 260/280 nm. Before the reverse transcription reaction, isolated RNAs were stored at −80 • C. For complementary DNA synthesis, 2 µg of total RNA was reverse-transcribed using random hexanucleotides and the High-Capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific) following the manufacturer's instructions. The reaction product was diluted to a final concentration of 20 ng/µL and stored at −20 • C until gene expression analysis.

Real-Time Quantitative Polymerase Chain Reaction (qRT-PCR)
The qRT-PCR was carried out in a 7500 Fast Real-Time PCR System (Thermo Fisher Scientific) using a total volume of 10 µL according to the manufacturer's instructions. We used TaqMan ® Fast Universal PCR Master Mix and assays based on hydrolysis probes (TaqMan ® Gene Expression Assays, Thermo Fisher Scientific), as they do not detect genomic DNA. Gene expression was performed for CTLA-4 (Hs00175480_m1), PDCD1 (PD1) (Hs00169472_m1), CD274 (PD-L1) (Hs00204257_m1), PDCD1LG2 (PD-L2) (Hs00228839_m1), CD276 (B7-H3) (Hs00987207_m1), JAK2 (Hs01078136_m1), and FoXO1 (Hs01054576_m1) for cell lines, mammospheres, and co-cultures samples. As reference genes to normalize gene expression, we used both β-ACTIN (Hs99999903_m1) and PUM1 (Hs00472881_m1). Depending on the assay type, a pool of RNA from the breast epithelial cell line 184A1 or the tumor cell lines was used as reference calibrator samples. No template controls were included in each reaction, and all experiments were duplicated. Relative changes in gene expression were calculated as fold change by the 2 −∆∆Ct method [111]. Results were analyzed with the 7500 software v2.0.6 (Thermo Fisher Scientific).

Statistical Analysis
The Kolmogorov-Smirnov test was performed to define the distribution of the variables. For the differences in gene expression, parametric variables were compared using Student's t-test, while non-parametric variables were compared using the Mann-Whitney U-test. In the co-cultures, in which repeated measurements were performed over time, the Student's t-test for paired data and the Wilcoxon test were used to compare parametric and non-parametric variables, respectively. For our in vitro study, we define increased or decreased levels of any gene expression as those above or below a relative quantification value equal to 1. In all cases, a p-value of <0.05 was considered statistically significant. Statistical analysis was performed using the SPSS version-23 statistical software package.

Conclusions
In conclusion, we have found that PDCD1 (PD1), JAK2, and FoXO1 expression are highly variable in BC cell lines independently of adaptative immunity. In addition, intrinsic CTLA-4, CD274 (PD-L1), and PDCD1LG2 (PD-L2) expression could be essential for the pathogenesis of triple-negative BC cells, whereas CD276 (B7-H3) expression would be necessary for luminal cell lines. Moreover, we have demonstrated the intrinsic expression of CTLA-4, PDCD1 (PD1), CD274 (PD-L1), PDCD1LG2 (PD-L2), and JAK2 in mammospheres derived from aggressive BC phenotypes. However, CD276 (B7-H3) and FoXO1 expression seem not to be directly involved in the acquisition and/or maintenance of tumor stem cell characteristics in vitro. Finally, we have determined that neoplastic cell-intrinsic CTLA-4, PCDC1 (PD1), CD274 (PD-L1), and PDCD1LG2 (PD-L2) expression is dynamic and varies depending on BC immunophenotype and tumor-immune cell interactions. Therefore, these data confirm our hypothesis that the expression of some immunoregulatory genes depends on BC phenotype and stem cell properties. They further support the view that this expression is intrinsic to neoplastic cells and depends on the tumor microenvironment.

Institutional Review Board Statement:
The study was conducted following the Declaration of Helsinki and approved by the Institutional Ethics Committee of Dr. Balmis University General Hospital (protocol code PI2019/089, endorsed on 29 January 2022; and protocol code PI2020-265, approved on 24 March 2021).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study by the Biobank of Dr. Balmis University General Hospital.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.