A Feedback Loop Formed by ATG7/Autophagy, FOXO3a/miR-145 and PD-L1 Regulates Stem-Like Properties and Invasion in Human Bladder Cancer

Programmed cell death protein 1 (PD-1) and its ligand PD-L1 blockade have been identified to target immune checkpoints to treat human cancers with durable clinical benefit. Several studies reveal that the response to PD-1-PD-L1 blockade might correlate with PD-L1 expression levels in tumor cells. However, the mechanistic pathways that regulate PD-L1 protein expression are not understood. Here, we reported that PD-L1 protein is regulated by ATG7-autophagy with an ATG7-initiated positive feedback loop in bladder cancer (BC). Mechanistic studies revealed that ATG7 overexpression elevates PD-L1 protein level mainly through promoting autophagy-mediated degradation of FOXO3a, thereby inhibiting its initiated miR-145 transcription. The lower expression of miR-145 increases pd-l1 mRNA stability due to the reduction of its direct binding to 3′-UTR of pd-l1 mRNA, in turn leading to increasing in pd-l1 mRNA stability and expression, and finally enhancing stem-like property and invasion of BC cells. Notably, overexpression of PD-L1 in ATG7 knockdown cells can reverse the defect of autophagy activation, FOXO3A degradation, and miR-145 transcription attenuation. Collectively, our results revealed a positive feedback loop to promoting PD-L1 expression in human BC cells. Our study uncovers a novel molecular mechanism for regulating pd-l1 mRNA stability and expression via ATG7/autophagy/FOXO3A/miR-145 axis and reveals the potential for using combination treatment with autophagy inhibitors and PD-1/PD-L1 immune checkpoint blockade to enhance therapeutic efficacy for human BCs.


Introduction
Bladder cancer (BC) is the fourth most common cancer in men and the fifth most common malignancy in the US [1]. BC is categorized as either non-muscle invasive or muscle invasive, and this differentiation is key to treatment planning and prognosis [2]. Survival of patients with muscle-invasive bladder cancer (MIBC) is poor, with only 35% survival at five years in lymph node-positive tumors [3]. There have been great improvements in BC diagnosis and treatment nowadays, due to the evolution of imaging, the introduction of new diagnostic modalities, the urinary tumor markers and the improvement of the surgical skills and techniques, as well as the upgraded role of chemo-and immunotherapy. However, BC remains a major clinical concern because of its great recurrence rate and, in many cases, progression propensity [4]. Cancer stem cells (CSCs) is a subpopulation of tumor cells that are enriched for tumorigenic property and can regenerate the heterogeneity of the original tumor in immunocompromised mice [5,6]. Growing evidence suggests It has been reported that PD-L1 expression levels are correlated with the response to PD-1-PD-L1 blockade in tumor cells [29,30]. Our previous studies have discovered that ATG7 overexpressed plays a critical role in cell invasion, growth and sphere formation of human BC cells [19,20].
To test whether PD-L1 expression is responsible for ATG7 s promotion of tumorigenesis and stem cell-like properties of human BCs, we firstly transfected into highly invasive human BC cell lines with shATG7#1 and shATG7#2 ( Figure 1A-C). Knockdown of ATG7 dramatically decreased PD-L1 protein expression ( Figure 1A-C). To determine PD-L1 s biological contribution in ATG7 regulating stem-like property, invasion, and tumorigenesis, the GFP-PD-L1 constructs were stably transfected into T24T(shATG7#1) cells ( Figure 1D). ATG7 knockdown almost completely abolished sphere formation in T24T cells ( Figure 1E). In comparison to T24T(shATG7#1) cells, GFP-PD-L1 ectopic expression restored the sphere formation ( Figure 1E,F), anchorage-independent growth ( Figure 1G,H), migration and invasion of human BC cells ( Figure 1I,J). These results demonstrated that PD-L1 is a novel ATG7 downstream regulated gene and plays a critical role in ATG7-mediated positive regulation of human high invasive BC cell stem-like property, invasion, and anchorage-independent growth.
Cancers 2019, 11, doi: 3 To test whether PD-L1 expression is responsible for ATG7′s promotion of tumorigenesis and stem cell-like properties of human BCs, we firstly transfected into highly invasive human BC cell lines with shATG7#1 and shATG7#2 ( Figures 1A-C). Knockdown of ATG7 dramatically decreased PD-L1 protein expression ( Figures 1A-C). To determine PD-L1′s biological contribution in ATG7 regulating stem-like property, invasion, and tumorigenesis, the GFP-PD-L1 constructs were stably transfected into T24T(shATG7#1) cells ( Figure 1D). ATG7 knockdown almost completely abolished sphere formation in T24T cells ( Figure 1E). In comparison to T24T(shATG7#1) cells, GFP-PD-L1 ectopic expression restored the sphere formation ( Figures 1E,F), anchorage-independent growth ( Figure 1G,H), migration and invasion of human BC cells ( Figure 1I,J). These results demonstrated that PD-L1 is a novel ATG7 downstream regulated gene and plays a critical role in ATG7-mediated positive regulation of human high invasive BC cell stem-like property, invasion, and anchorageindependent growth. Figure 1. PD-L1 acted as an ATG7 downstream mediator being responsible for ATG7-promoted stemlike property, invasion, and anchorage-independent growth in human high invasive BC cells. (A-C) T24T, UMUC3, T24 cells were stably transfected with ATG7 knockdown constructs (#1 & #2), respectively. Western Blot was used to assess the ATG7 protein knockdown efficiency and its effects on other protein expression. (D) The GFP-tagged PD-L1 overexpression plasmid was stably transfected into T24T(shATG7#1) cells. (E & F) The indicated cells were subjected to determination of sphere formation abilities according to the manufacturer's instruction, the number of spheroid formed cells were counted as described in the section of "Materials and Methods". The asterisk (*) indicates a significant decrease in comparison to scramble nonsense transfectant (* p < 0.05), while the symbol (**) indicates a significant increase in comparison to T24T(shATG7#1/pEGFPc1) cells (** p < Figure 1. PD-L1 acted as an ATG7 downstream mediator being responsible for ATG7-promoted stem-like property, invasion, and anchorage-independent growth in human high invasive BC cells. (A-C) T24T, UMUC3, T24 cells were stably transfected with ATG7 knockdown constructs (#1 & #2), respectively. Western Blot was used to assess the ATG7 protein knockdown efficiency and its effects on other protein expression. (D) The GFP-tagged PD-L1 overexpression plasmid was stably transfected into T24T(shATG7#1) cells. (E,F) The indicated cells were subjected to determination of sphere formation abilities according to the manufacturer's instruction, the number of spheroid formed cells were counted as described in the section of "Materials and Methods". The asterisk (*) indicates a significant decrease in comparison to scramble nonsense transfectant (* p < 0.05), while the symbol (**) indicates a significant increase in comparison to T24T(shATG7#1/pEGFPc1) cells (** p < 0.05).
(G,H) The indicated cells were subjected to anchorage-independent soft agar assay using the protocol described in the section of "Materials and Methods". Representative images of colonies of indicated cells were photographed under an Olympus DP71 (G). The number of colonies was counted with more than 32 cells of each colony and the results were presented as colonies per 10 4 cells, and the bars show mean ± SD from three independent experiments (H). The asterisk (*) indicates a significant decrease in comparison to scramble nonsense transfectant (* p < 0.05), while the symbol (**) indicates a significant increase in comparison to T24T(shATG7#1/pEGFPc1) cells (** p < 0.05). (I) Invasion abilities of the indicated cells were determined using BD BiocoatTM matrigelTM invasion chamber. The migration ability was determined using the empty insert membrane without the matrigel, while the invasion ability was evaluated using the same system except that the matrigel was applied. (J) The invasion ability was normalized to the insert control according to the manufacturer's instruction. The asterisk (*) indicates a significant inhibition in comparison to T24T(Nonsense) cells (* p < 0.05), while the symbol (**) indicates a significant increase in comparison to T24T(shATG7#1/pEGFPc1) (** p < 0.05). Scale bars in (E,I) = 200 µm, Scale bars in (G) = 500 µm.

ATG7 Promoted pd-l1 mRNA Stability by Regulating Its 3 -UTR Activity
To elucidate the mechanisms of ATG7 promotion of PD-L1 protein expression, we first detected the effect of ATG7 s regulatory on pd-l1 mRNA level. ATG7 knockdown remarkably inhibited pd-l1 mRNA in T24T(shATG7) cells ( Figure 2A). Therefore, we exploited the possibility of ATG7 stabilization of pd-l1 mRNA. Upon treated with actinomycin D (Act D), pd-l1 mRNA degradation rates in T24T(shATG7) cells was much faster than in T24T(Nonsense) cells ( Figure 2B), revealing that ATG7 overexpression results in the stabilization of pd-l1 mRNA in human BC cells. The 3-terminal untranslated region (3 -UTR) of mRNA has been reported to participate in its mRNA stability regulation [31]. Further results showed that knockdown of ATG7 significantly inhibited pd-l1 mRNA 3 -UTR activity ( Figure 2C), suggesting that ATG7 positively modulate 3 -UTR activity of pd-l1 mRNA, which might be associated with ATG7 promotion of pd-l1 mRNA stability in human BC cells. Moreover, the results from analyses of pd-l1 mRNA 3 -UTR region showed that there were multiple ARE-elements in the pd-l1 3 -UTR sequence. Since our previous studies have reported that AUF1 protein is critical for ATG7 overexpression stabilizing rhogdiβ mRNA in human BC cells [32]. Thus, we knocked down AUF1 in UMUC3(shATG7#1) cells. The results indicated that AUF1 knockdown did not show a restoration of PD-L1 protein expression ( Figure 2E), excluding the participation of AUF1 in ATG7 stabilization of pd-l1 mRNA. MicroRNAs (miRNAs) are reported to regulate targeted gene expression by binding to mRNA 3 -UTR [33]. Putative miRNAs that could potentially target 3 -UTR of pd-l1 mRNA were bioinformatics searched by using TargetScan [34], Pictar [35] and miRANDA [36] and the results were listed in Figure 2F. We anticipated that some of these miRNAs might be involved in the ATG7 regulation of pd-l1 mRNA 3 -UTR activity. did not show a restoration of PD-L1 protein expression ( Figure 2E), excluding the participation of AUF1 in ATG7 stabilization of pd-l1 mRNA. MicroRNAs (miRNAs) are reported to regulate targeted gene expression by binding to mRNA 3′-UTR [33]. Putative miRNAs that could potentially target 3′-UTR of pd-l1 mRNA were bioinformatics searched by using TargetScan [34], Pictar [35] and miRANDA [36] and the results were listed in Figure 2F. We anticipated that some of these miRNAs might be involved in the ATG7 regulation of pd-l1 mRNA 3′-UTR activity.

ATG7 Overexpression Downregulated miR-145 and Subsequently Stabilized pd-l1 mRNA Through Directly Binding to Its 3′-UTR
Based on the results shown in the above studies, Real-time PCR was performed to evaluate the miRNAs expressions as indicated in Figure 2F. As shown in Figure 3A, miR-145 was found to be the only one that was upregulated in T24T(shATG7) cells. To determine the potential effect of miR-145 on ATG7 regulation of PD-L1 expression, a construct expressing miR-145 or miR-145 inhibitor was transfected into T24T or T24T(shATG7#1) cells as shown in Figures 3B,C. MiR-145 ectopic expression resulted in a dramatical decrease of PD-L1 protein expression ( Figure 3D), while miR-145 inhibition remarkably promoted PD-L1 protein expression ( Figure 3E). These results revealed that miR-145 played an essential role in ATG7 promotion of PD-L1 expression. Further, miR-145 ectopic expression inhibited the sphere formation of T24T cells, while the inhibition of miR-145 in T24T(shATG7#1) cells markedly promoted the sphere formation and increased pd-l1 mRNA stability ( Figure 3F) in T24T(shATG7#1/miR-145 inhibitor) cells as compared with T24T(shATG7#1/Nonsense) (Figures 3G-I). To detect whether miR-145 could directly bind to pd-l1 mRNA 3′-UTR, as indicated in Figure 3J, After synchronization, the cells were used to determine pd-l1 mRNA stability in the presence of Actinomycin D (Act D) by using Real-time PCR. gapdh was used as internal control. (C) The pMIR-pd-l1 3 -UTR reporter was transiently transfected into the indicated cells and luciferase activity of each transfectant was evaluated. The luciferase activity was presented as a relative to nonsense transfectant with normalized by using pRL-TK as an internal control. The bars show mean ± SD from three independent experiments. The symbol (*) indicates a significant decrease in T24T(shATG7) cells as compared to nonsense transfectant (* p < 0.05). (D) The predicted Class I ARE element on the 3 -UTR of pd-l1 mRNA. (E) Cell lysates extracted from the indicated cells were subjected to Western Blot to determine AUF1, and PD-L1 protein expression, GAPDH was used as a protein loading control. (F) The potential microRNAs binding sites in 3 -UTR of pd-l1 mRNA were analyzed by the TargetScan, Pictar and miRANDA database.

ATG7 Overexpression Downregulated miR-145 and Subsequently Stabilized pd-l1 mRNA Through Directly Binding to Its 3 -UTR
Based on the results shown in the above studies, Real-time PCR was performed to evaluate the miRNAs expressions as indicated in Figure 2F. As shown in Figure 3A, miR-145 was found to be the only one that was upregulated in T24T(shATG7) cells. To determine the potential effect of miR-145 on ATG7 regulation of PD-L1 expression, a construct expressing miR-145 or miR-145 inhibitor was transfected into T24T or T24T(shATG7#1) cells as shown in Figure 3B,C. MiR-145 ectopic expression resulted in a dramatical decrease of PD-L1 protein expression ( Figure 3D), while miR-145 inhibition remarkably promoted PD-L1 protein expression ( Figure 3E). These results revealed that miR-145 played an essential role in ATG7 promotion of PD-L1 expression. Further, miR-145 ectopic expression inhibited the sphere formation of T24T cells, while the inhibition of miR-145 in T24T(shATG7#1) cells markedly promoted the sphere formation and increased pd-l1 mRNA stability ( Figure 3F) in T24T(shATG7#1/miR-145 inhibitor) cells as compared with T24T(shATG7#1/Nonsense) ( Figure 3G-I).
To detect whether miR-145 could directly bind to pd-l1 mRNA 3 -UTR, as indicated in Figure 3J, we constructed the miR-145 binding site mutation in pd-l1 3 -UTR-luciferase reporter. The results exhibited that pd-l1 3 -UTR luciferase activity was decreased in T24T(shATG7) cells, whereas the mutation luciferase reporter completely reversed this inhibition in ATG7 knockdown transfectant T24T(shATG7) ( Figure 3K). Collectively, our results reveal that the direct binding of miR-145 to pd-l1 3 -UTR is important for miR-145 destabilization of pd-l1 mRNA. Thus, ATG7 overexpression inhibits miR-145 expression, which subsequently binds to pd-l1 3 -UTR and results in pd-l1 mRNA stabilization, and in turn enhancing human BC sphere formation.

ATG7 Inhibited miR-145 Transcription through Attenuating FOXO3a Protein Expression
To explore the mechanisms of the ATG7 impairing miR-145, we tested the pre-miR-145 expression level and miR-145 promoter activity. The results showed that both pre-miR-145 expression and miR-145 promoter activity were markedly elevated in ATG7 knockdown transfectants ( Figure 4A,B), indicating ATG7 regulation of miR-145 expression at the transcription level. Therefore, we bioinformatically analyzed and evaluated the expression of the potential transcriptional factors, which could bind to the -1548 to -30 region of the miR-145 promoter ( Figure 4C). Results indicated that knockdown of ATG7 consistently promoted FOXO3a protein expression and c-Jun phosphorylation at Ser63/73, while it did not show consistent effect on other transcription factors, such as Stat5, ETS1, ELK1, JunB and C-Fos ( Figure 4D,E), suggesting that FOXO3a or c-Jun might be involved in ATG7 inhibition of miR-145 expression. To explore the potential role of c-Jun, dominant negative mutant Jun (TAM67) was stably transfected into T24T(shATG7#1) cells. As shown in Figure 4F, the introduction of TAM67 had no observable effect on PD-L1 protein expression, excluding the possible involvement of c-Jun in ATG7 promotion of PD-L1 expression. We further detected the effects of ATG7 knockdown on FOXO3a nuclear translocation by using a nuclear/cytosol fractionation assay. The results showed that knockdown of ATG7 resulted in the marked translocation of FOXO3a from cytoplasmic to nuclear in comparison to T24T(Nonsense) cells ( Figure 4G), suggesting FOXO3a activation in ATG7 knockdown cells.
To detect whether FOXO3a was the transcription factor responsible for ATG7 inhibition of miR-145 transcription, we stably transfected HA-FOXO3a into T24T cells ( Figure 5A). FOXO3a overexpression markedly promoted miR-145 promoter-driven transcription activity, and miR-145 expression, and inhibited PD-L1 protein expression ( Figure 5A-C). Moreover, the chromatin immunoprecipitation (ChIP) assay was performed, and found that FOXO3a could directly bind to the miR-145 promoter ( Figure 5D). Consistently, the sphere formation ability was dramatically decreased in FOXO3a overexpressed T24T cells ( Figure 5E,F). Collectively, these results strongly demonstrate that FOXO3a is the important transcription factor responsible for ATG7 inhibition of miR-145 transcription and further promoted PD-L1 expression, as well as human BC stem-like property.
4F, the introduction of TAM67 had no observable effect on PD-L1 protein expression, excluding the possible involvement of c-Jun in ATG7 promotion of PD-L1 expression. We further detected the effects of ATG7 knockdown on FOXO3a nuclear translocation by using a nuclear/cytosol fractionation assay. The results showed that knockdown of ATG7 resulted in the marked translocation of FOXO3a from cytoplasmic to nuclear in comparison to T24T(Nonsense) cells ( Figure  4G), suggesting FOXO3a activation in ATG7 knockdown cells. To detect whether FOXO3a was the transcription factor responsible for ATG7 inhibition of miR-145 transcription, we stably transfected HA-FOXO3a into T24T cells ( Figure 5A). FOXO3a overexpression markedly promoted miR-145 promoter-driven transcription activity, and miR-145 immunoprecipitation (ChIP) assay was performed, and found that FOXO3a could directly bind to the miR-145 promoter ( Figure 5D). Consistently, the sphere formation ability was dramatically decreased in FOXO3a overexpressed T24T cells (Figures 5E,F). Collectively, these results strongly demonstrate that FOXO3a is the important transcription factor responsible for ATG7 inhibition of miR-145 transcription and further promoted PD-L1 expression, as well as human BC stem-like property.

ATG7 Overexpression Promoted Autophagic Removal of FOXO3a in Human BC Cells.
To elucidate how ATG7 inhibited FOXO3a protein expression, we assessed the effect of ATG7 on foxo3a mRNA abundance in T24T cells. ATG7 knockdown only showed a slight reduction of foxo3a mRNA as compared to its scramble nonsense transfectants ( Figure 6A), suggesting that inhibition of FOXO3a protein by ATG7 might not occur at mRNA level. We further detected whether ATG7 regulated FOXO3a protein stability. The result showed that ATG7 knockdown remarkably increased FOXO3a protein stability ( Figure 6B). Intriguingly, ATG7 plays an essential role in the process of autophagy, which delivers some proteins to lysosomes for degradation [37]. Thus, we next determined whether ATG7 overexpression mediated autophagy, and in turn promoting FOXO3a

ATG7 Overexpression Promoted Autophagic Removal of FOXO3a in Human BC Cells
To elucidate how ATG7 inhibited FOXO3a protein expression, we assessed the effect of ATG7 on foxo3a mRNA abundance in T24T cells. ATG7 knockdown only showed a slight reduction of foxo3a mRNA as compared to its scramble nonsense transfectants ( Figure 6A), suggesting that inhibition of FOXO3a protein by ATG7 might not occur at mRNA level. We further detected whether ATG7 regulated FOXO3a protein stability. The result showed that ATG7 knockdown remarkably increased FOXO3a protein stability ( Figure 6B). Intriguingly, ATG7 plays an essential role in the process of autophagy, which delivers some proteins to lysosomes for degradation [37]. Thus, we next determined whether ATG7 overexpression mediated autophagy, and in turn promoting FOXO3a protein degradation. The results showed that knockdown of ATG7 did attenuate autophagy in T24T cells, while FOXO3a was markedly upregulated after inhibition of autophagy by Baf A1 treatment in comparison to the vehicle control ( Figure 6C). Consistent with the alteration of FOXO3a protein, miR-145 expression was elevated in either ATG7 knockdown cells or autophagy inhibitor Baf A1-treated BC cells ( Figure 6C,D). Consistently, PD-L1 protein expression was dramatically decreased in both ATG7 knockdown cells or autophagy inhibitor Baf A1-treated BC cells ( Figure 6D). These results consistently reveal that ATG7 overexpression promotes autophagy, which further promotes FOXO3a protein degradation in human BC cells.
protein degradation. The results showed that knockdown of ATG7 did attenuate autophagy in T24T cells, while FOXO3a was markedly upregulated after inhibition of autophagy by Baf A1 treatment in comparison to the vehicle control ( Figure 6C). Consistent with the alteration of FOXO3a protein, miR-145 expression was elevated in either ATG7 knockdown cells or autophagy inhibitor Baf A1-treated BC cells (Figures 6C,D). Consistently, PD-L1 protein expression was dramatically decreased in both ATG7 knockdown cells or autophagy inhibitor Baf A1-treated BC cells ( Figure 6D). These results consistently reveal that ATG7 overexpression promotes autophagy, which further promotes FOXO3a protein degradation in human BC cells.

A Positive Feedback Loop was Formed by ATG7/autophagy, FOXO3a/miR-145 and PD-L1
As we have discovered that PD-L1 was an ATG7 downstream mediator for promotion of human high invasive BC cell invasion, anchorage-independent growth and sphere formation. We next tried to ask whether PD-L1 participated in the regulation of other ATG7 downstream effectors for human BC invasion/progression. The constructs that express GFP-PD-L1 were stably transfected into T24T(shATG7#1) cells. Ectopic expression GFP-PD-L1 in T24T(shATG7#1) cells promoted the conversion of LC3 from LC3-I to LC3-II, inhibited AUF1 and increased RhoGDIβ protein expression ( Figure 7A), revealing PD-L1 was the critical upstream factor for ATG7-induced human BC invasion. Moreover, ectopic expression GFP-PD-L1 promoted CD44s ( Figure 7A). Given our most recently have identified that the TET1/USP28/CD44/RhoGDIβ pathway that is crucial for ATG7′s oncogenic

A Positive Feedback Loop was Formed by ATG7/autophagy, FOXO3a/miR-145 and PD-L1
As we have discovered that PD-L1 was an ATG7 downstream mediator for promotion of human high invasive BC cell invasion, anchorage-independent growth and sphere formation. We next tried to ask whether PD-L1 participated in the regulation of other ATG7 downstream effectors for human BC invasion/progression. The constructs that express GFP-PD-L1 were stably transfected into T24T(shATG7#1) cells. Ectopic expression GFP-PD-L1 in T24T(shATG7#1) cells promoted the conversion of LC3 from LC3-I to LC3-II, inhibited AUF1 and increased RhoGDIβ protein expression ( Figure 7A), revealing PD-L1 was the critical upstream factor for ATG7-induced human BC invasion. Moreover, ectopic expression GFP-PD-L1 promoted CD44s ( Figure 7A). Given our most recently have identified that the TET1/USP28/CD44/RhoGDIβ pathway that is crucial for ATG7 s oncogenic role in human BC cell invasion, metastasis and stem-like properties [20], our findings reveal that PD-L1 was the important upstream regulator for ATG7-induced human BC sphere formation and tumorigenesis. In addition, ectopic expression GFP-PD-L1 inhibited FOXO1 and p27 protein expression ( Figure 7A), while our recent studies have indicated that ATG7 is able to promote BC tumorigenesis via ETS2/miR-196b/FOXO1/p27 axis [19]. Most importantly, we found that overexpression of PD-L1 in T24T(shATG7#1) cells attenuated FOXO3a protein expression and miR-145 expression as compared to T24T(shATG7#1/pEGFPc1) cells ( Figure 7A,B). Collectively, our current studies together with our previous findings strongly reveal that a positive feedback loop was formed by ATG7/Autophagy and FOXO3a/miR-145 to regulate PD-L1 protein expression, and all of which together to form a network for the promotion of BC stem-like properties, tumorigenesis and invasion as illustrated in Figure 7C. Collectively, this study demonstrates that ATG7 autophagic removal of FOXO3a leads to attenuation of miR-145 promoter transcription, and in turn resulting in the stabilization of pd-l1 mRNA, further promoting stem-like property, invasion, and tumorigenesis of human high invasive BC cells. Furthermore, overexpression of PD-L1 in ATG7 knockdown cells reverses the defect of autophagy activation, FOXO3A degradation, and miR-145 transcription attenuation, suggesting a positive feedback loop to promote autophagy and consequently biological outcome due to ATG7/autophagy activation in human BC cells.
Cancers 2019, 11, doi: 10 role in human BC cell invasion, metastasis and stem-like properties [20], our findings reveal that PD-L1 was the important upstream regulator for ATG7-induced human BC sphere formation and tumorigenesis. In addition, ectopic expression GFP-PD-L1 inhibited FOXO1 and p27 protein expression ( Figure 7A), while our recent studies have indicated that ATG7 is able to promote BC tumorigenesis via ETS2/miR-196b/FOXO1/p27 axis [19]. Most importantly, we found that overexpression of PD-L1 in T24T(shATG7#1) cells attenuated FOXO3a protein expression and miR-145 expression as compared to T24T(shATG7#1/pEGFPc1) cells ( Figures 7A,B). Collectively, our current studies together with our previous findings strongly reveal that a positive feedback loop was formed by ATG7/Autophagy and FOXO3a/miR-145 to regulate PD-L1 protein expression, and all of which together to form a network for the promotion of BC stem-like properties, tumorigenesis and invasion as illustrated in Figure 7C. Collectively, this study demonstrates that ATG7 autophagic removal of FOXO3a leads to attenuation of miR-145 promoter transcription, and in turn resulting in the stabilization of pd-l1 mRNA, further promoting stem-like property, invasion, and tumorigenesis of human high invasive BC cells. Furthermore, overexpression of PD-L1 in ATG7 knockdown cells reverses the defect of autophagy activation, FOXO3A degradation, and miR-145 transcription attenuation, suggesting a positive feedback loop to promote autophagy and consequently biological outcome due to ATG7/autophagy activation in human BC cells.

Discussion
In order to maintain cellular homeostasis, autophagy facilitates the turnover or clearance of misfolded, protein aggregates or long-lived proteins, and damaged organelles [38]. ATG7, a crucial protein for autophagic responses, has been found to be upregulated in BBN-induced basal MIBCs

Discussion
In order to maintain cellular homeostasis, autophagy facilitates the turnover or clearance of misfolded, protein aggregates or long-lived proteins, and damaged organelles [38]. ATG7, a crucial protein for autophagic responses, has been found to be upregulated in BBN-induced basal MIBCs and human BC tissues, which plays an oncogenic role in BC stem-like property, invasion, lung metastasis and tumorigenic growth [19,20]. ATG7 has also been reported to facilitate the transcription of OCT4, promoting prostate tumor initiation, self-renewal, and drug resistance [39]. It has been reported that ATG5 or ATG7 knockdown reduces human hematopoietic stem/progenitor cells frequencies both in vitro and in vivo [40]. The current study identified that PD-L1 was a critical downstream mediator for ATG7 overexpression inducing stem-like property, invasion, and anchorage-independent growth of human BC. Our results discover a new linker among ATG7 upregulation, PD-L1 protein, human BC stem-like property and invasion, which offers novel insight into the therapeutic targeting of ATG7/PD-L1 for advanced human BCs.
Aberrant expression of PD-L1 suppresses cancer immunity and further promotes cancer progression in tumor cells [41]. Five immune checkpoint inhibitors blocking PD-1 or PD-L1 in the management of advanced bladder cancer have been approved by US FDA. PD-L1 overexpression has been shown to enhance cancer proliferation and invasion in vitro and promote cervical cancer growth in vivo [42]. However, the effect of high PD-L1 protein expression on tumor cell invasion has never been explored yet. Herein, we found that ectopic expression of GFP-PD-L1 promoted autophagy/AUF1/RhoGDIβ and autophagy/FOXO3a/miR-145 cascade signaling, elevated CD44a protein expression, impaired FOXO1 and p27 protein expression, all of which form a positive feedback loop to strengthen PD-L1 protein expression, further exhibiting the oncogenic role of ATG7 in human BCs stem-like property, invasion, and tumorigenesis. It has been reported that MAPK, PI3K/AKT, and Wnt/β-catenin pathways induce PD-L1 transcription by triggering the binding of the transcription factors to the PD-L1 promoter [43]. EGFR activation induces EMT and myc-dependent PD-L1 expression in human salivary adenoid cystic carcinoma cells [44]. MiRNAs have been shown to target PD-L1 [45,46]. OCT4-activated miR-18a targets WNK2, PTEN, and SOX6 to repress p53 signaling or activate MEK/ERK, PI3K/AKT, and Wnt/β-catenin pathways to upregulate PD-L1 expression [42]. Our findings showed that ATG7 overexpression decreased miR-145 expression, in turn promoted pd-l1 mRNA stabilization by directly binding to its 3 -UTR. These results provide strong evidence demonstrating that miR-145 is a PD-L1 upstream regulator for mediating ATG7 elevation of BC stem-like property and invasion.
MiR-145 expression was decreased in multiple cancers including prostate, colon and lung, which is regarded as a tumor suppressor [47,48]. Multiple studies have focused on the role of miR-145 underlying carcinogenesis. MiR-145 reduces syndecan-1 and increases stem cell factors and induces cell differentiation and senescence on urothelial carcinoma cells [48]. MiR-145 could directly bind to TAGLN2 to suppress BC cell proliferation and migration [49] and silence c-Myc to inhibit the PTBP1/PKMs Axis, thereby reducing the cancer-specific energy metabolism [50]. Our previous study has discovered that isorhapontigenin (ISO) increases miR-145 expression, which inhibits SOX2 protein expression by directly binding to its 3 -UTR region. SOX2 inhibition results in the downregulation of cyclin D1 and suppression of anchorage-independent growth and sphere formation of patient-derived glioblastoma [51]. In this study, we discovered that ATG7 overexpression inhibits miR-145 expression, which further enhances the stability of pd-l1 mRNA by directly binding to its 3 -UTR, and in turn promotes ATG7-induced stem-like property of human BCs. Mir-145 could be regulated at several levels to exert its tumor suppressor role. A sponge of miR-145, Circ_0058063 has been shown to regulate CDK6 expression and promote BC progression [52]. Reduced lncRNA-LET inhibited miR-145 biogenesis and subsequently led to the accumulation of CSCs [53]. Herein, FOXO3a was found to be responsible for the promotion of miR-145 promoter transcription. Our finding further provides compelling evidence for miR-145 s tumor suppressor role in human BCs.
FOXO3a recruits transcriptional coactivators to trigger the transcription of downstream genes. FOXO3a plays crucial roles in apoptosis, cell cycle arrest, metabolism, and DNA repair, and participates in multiple pathological and physiological processes [54]. FOXO3 inhibits AGS cell growth by promoting Beclin1-medicated autophagy and inhibiting p62 protein expression in an acidic microenvironment [55]. XB130 promotes human colorectal cancer cell migration and invasion by enhancing phosphorylation of FOXO3a and epithelial-mesenchymal transition [56]. DFOG, a novel synthetic genistein analogue, decreases the level of phosphorylated FOXO3a to inhibit the ovarian cancer sphere-forming cell self-renewal activity [57]. In the current study, ATG7 overexpression inhibited FOXO3a protein expression and nuclear translocation, and further promoted stem-like properties of human high invasive BCs. A variety of cellular stimuli, including metabolic stress, growth factors, and oxidative stress tranduce signals to FOXO3a to regulate the downstream gene expression and protein-protein interactions [58,59]. Ubiquitination, phosphorylation, and acetylation have been reported to be involved in the regulation of these processes [59]. Crosstalk of several frequently deregulated signaling cascades: AKT, ERK/MAPK or IKK pathways, culminates in the inactivation of the tumor suppressive function of FOXO3a [60]. Here, we unexpectedly discovered that ATG7 exhibits autophagic removal role on FOXO3a protein, and promotes its protein degradation, and consequently inhibits FOXO3a-dependent transcriptional activity of its targeted gene miR-145.
In summary, this studies discovered a new feedback loop of ATG7/Autophagy/FOXO3a/ miR-145/PD-L1 pathway that acts as an ATG7 downstream factor being responsible for ATG7 s oncogenic role in the promotion of human BC cell stem-like property, invasion, and tumorigenesis. Specifically, ATG7 overexpression impairs FOXO3a protein expression, thereby decreases miR-145 promoter transcription and its expression by directly binding to its promoter region. In turn, downregulated miR-145 reduces its direct binding to pd-l1 mRNA3 -UTR and thereby stabilizes its mRNA and elevates PD-L1 protein expression, further resulting in promotion of human BC cell stem-like property and invasion. Besides, PD-L1 is defined to form a positive feedback loop to further enhance ATG7 s oncogenic role in human BC stem-like property, invasion, and tumorigenesis. Given the recent findings of the ATG7 s oncogenic effect on promoting BC progression, our new results provide the potential for developing a combination treatment with autophagy inhibitors and PD-1/PD-L1 immune checkpoint blockade therapeutic strategy for human BC patient treatment.

Plasmids, Reagents and Antibodies
Human miR-145 promoter (from -1548 to -30) was cloned into the pGL3 Basic luciferase reporter to make a human miR-145 promoter-driven luciferase reporter. The constructs of short hairpin RNA specific targeting ATG7 (shATG7), and AUF1(shAUF1) were purchased from Open Biosystems (Pittsburg, PA, USA). The GFP-tagged PD-L1 plasmid was constructed with Xho I and EcoR I using cDNA obtained from T24T cells. Human pd-l1 3 -UTR luciferase reporter was cloned into the pMIR luciferase assay vector. The HA-tagged FOXOA3a plasmid was obtained from Addgene (Cambridge, MA, USA). The miR-145 expression construct was kindly provided by Dr. Renato Baserga (Department of Cancer Biology, Thomas Jefferson University and Kimmel Cancer Center, Philadelphia, PA, USA) [61]. The miR-145 inhibitor was purchased from GeneCopoeia (San Diego, CA, USA). The dominant negative Jun plasmid (TAM67) was described and used in our previous studies [62]. Plasmids were prepared by the Plasmid Preparation/Extraction Maxi Kit from QIAGEN (Valencia, CA, USA). The chemical Actinomycin D was purchased from Calbiochem (Billerica, MA, USA). Proteasome inhibitor MG132 and protein synthesis inhibitor Cycloheximide (CHX) were purchased from Calbiochem. Bafilomycin A1 (Baf A1) was purchased from Santa Cruz (St. Louis, MO, USA). TRIzol reagent and the SuperScriptTM First-Strand Synthesis system were acquired from Invitrogen (Grand Island, NY, USA). PolyJetTM DNA in Vitro Transfection Reagent was purchased from SignaGen Laboratories (Rockville, MD, USA). The nuclear/cytosol fractionation kit was obtained from Biovision Incorporated (Milpitas, CA, USA).

Western Blot Analysis
Western blot was performed in our previous publication [64]. T24T, T24, UMUC3 cells and their transfectants were seeded in 6-well plates and cultured in the normal medium until 70-80% confluence. Following culture cells in 0.1% FBS medium for 12 h, the medium were replaced with 5% FBS DMEM-F12 (1:1) or 10% FBS DMEM for another 12 h. Whole cell extracts were prepared with the cell lysis buffer (10 mM Tris-HCl, pH 7.4, 1% SDS, and 1 mM Na 3 VO 4 ) and then were subjected to Western Blot analysis. The images were captured by scanning with the phosphorimager (Typhoon FLA 7000, GE, Pittsburgh, PA, USA).

Luciferase Report Assay
Luciferase reporter assays were carried out as described as previously [65].

Semiquantitative or Quantitative RT-PCR
Semiquantitative or quantitative RT-PCR was carried out to examine the mRNA or miRNA expression, as previously described [19,62], and the primers were described in Table S1.

ChIP Assay
ChIP assay was carried out as described in our previous publication [65]. To specifically amplify the region containing the putative responsive elements on the human miR-145 promoter, PCR was performed with the following pair of primers: 5 -AAG CAC AAT GAG CAG AGG AGA G-3 and 5 -TTA GGG TAC AAT CTC AGC TCA GCA C-3 . The PCR products were separated on 2% agarose gels and stained with ethidium bromide; the images were then scanned with a UV light.

Sphere Formation Assay
The indicated cells were cultured in a 6-well ultra-low attachment culture plate for 5 or 7 days, and Olympus DP71 microscopy (Olympus America Inc., Center Valley, PA, USA) was used to capture the images, and the number of spheroids formed cells in each image was counted by the"Image J" software [66] for at least 6 random images. The results presented as the percentage of the spheroid colonies.

Cell Migration and Invasion
As described in our earlier publication, cell migration and invasion assay were performed [64]. Control inserts without matrigel and permeable support for 24 well plate with 8.0 µm transparent PET membrane were purchased from Corning Incorporated (Corning, NY, USA) (353097), the invasion kit (354480) was purchased from BD Biosciences (Bedford, MA, USA). The invasion assay was performed in normal cell culture serum according to the manufacturer's instruction. The images were taken under microscopy, Olympus DP71 (Olympus America Inc. Center Valley, PA, USA), and the number of the cells in each image was counted by the Image J software. According to the manufacturer's instruction, the invasion rate was normalized with the insert control.

Anchorage-Independent Growth Assay
As described in our previous studies, anchorage-independent growth (soft agar assay) was carried out [67]. Briefly, the 1 × 10 4 cells in 10% FBS Basal Medium Eagle (BME) containing 0.33% soft agar and were seeded over the basal layer containing 0.5% agar containing 10% FBS BME in each well of 6-well plates. The plates were incubated in 5% CO 2 incubator at 37 • C for 3 weeks. Colonies were captured under a microscope and only colonies with over 32 cells were counted. The results were presented as mean ± SD obtained from three independent experiments.

Statistical Analysis
Prism 5.0 Software (GraphPad Software, San Diego, CA, USA) was used to perform the statistical analysis. All data are shown as the means of triplicate assays ± SD. Student's t-test was employed to detect the significance of differences between various groups. The differences were considered significant at p < 0.05.

Conclusions
The identification of the feedback loop of ATG7/Autophagy/FOXO3a/miR-145/PD-L1 pathway provides an important insight into understanding the oncogenic role of ATG7 in the promotion of human BC cell stem-like property, invasion, and tumorigenesis, suggesting that autophagy inhibitors and PD-1/PD-L1 immune checkpoint blockade maybe combined to be a potential therapeutic strategy for the treatment of human BC patients.