Obesity and metabolic syndrome related macrophage promotes PD-L1 expression in TNBC through IL6/JAK/STAT pathway and can be reversed by telmisartan

Figures & data

Figure 1. PD-L1 expression profile of breast cancer on oncomine platform and of breast epithelial and cancer cell lines

(A-a), Comparison of PD-L1 mRNA expression among normal breast, invasive ductal breast carcinoma and invasive lobular breast carcinoma in TCGA breast dataset. (A-b), Comparison of PD-L1 DNA copy number among blood, normal breast, invasive ductal breast carcinoma and invasive lobular breast carcinoma in TCGA breast dataset. (B), Comparison of PD-L1 mRNA expression between non-TNBC and TNBC breast cancer in Curtis Breast, TCGA Breast, Kao Breast and Bittner Breast, respectively. Data in (A, B) were acquired from the Oncomine platform and shown by median, 90th, 75th, 25th and 10th percentiles. P value was calculated by Luo’s methods. (C), Western blot detection of PD-L1 in MCF-10A and seven breast cancer cell lines with different ER, PR, and HER2 status.

Table 1. Clinical characteristics distribution in BMI groups

Characteristics	Case	BMI			χ^2	P
		<23	23≤ BMI<28	≥28
Age
≤50	17	6	6	5	0.020	.887
>50	40	14	13	13
Menopausal Status
pre	13	5	5	3	1.311	.909
peri	4	1	1	2
post	40	14	13	13
Tumor Size
≤2 cm	18	8	6	4	1.360	.244
2 cm<T ≤ 5 cm	39	12	13	14
Lymph Node Invasion
no	31	13	9	9	0.886	.347
yes	26	7	10	9
Metabolic Complication
no	32	15	12	5	8.290	.004*
yes	25	5	7	13
Hypertension
no	35	15	12	8	3.641	.056
yes	22	5	7	10
Hyperglycemia
no	47	18	16	13	2.023	.341
yes	10	2	3	5
Dyslipidemia
no	45	17	17	11	4.598	0.107
yes	12	3	2	7

(<23, 23≤ BMI<28, and ≥28). Data are shown in real case numbers. All variables except the occurrence of metabolic complication distribute equally among the three groups. The χ2 and P values are determined by Linear-by-Linear Association two-tailed Chi-square test or Fisher’s exact test. The symbol * represents statistical significance (P < .05).

Table 2. Spearman correlation between clinical variables and PD-L1 expression in TNBC patients

Characteristics	Case	PD-L1			Spearman Correlation	P
		Negative	<20%	≥20%
Age
≤50	17	8	5	4	0.001	.229
>50	40	15	21	4
Menopausal Status
pre	13	6	5	2	0.046	.759
peri	4	2	1	1
post	40	15	20	5
Tumor Size
≤2 cm	18	5	12	1	−0.074	.112
2 cm<T ≤ 5 cm	39	18	14	7
Lymph Node Invasion
no	31	13	14	4	0.042	1.000
yes	26	10	12	4
BMI
<23	20	10	8	2	0.249	.055
23≤ BMI<28	19	8	11	0
≥28	18	5	7	6
Metabolic Complication
no	32	16	16	0	0.357	.001*
yes	25	7	10	8
Hypertension
no	35	17	16	2	0.287	.054
yes	22	6	10	6
Hyperglycemia
no	47	20	23	4	0.213	.058
yes	10	3	3	4
Dyslipidemia
no	45	20	21	4	0.242	0.097
yes	12	3	5	4

Data are shown in real case numbers. Metabolic complication has a significant correlation with PD-L1 expression. The P value is determined by Fisher’s exact test. The symbol * represents statistical significance (P < .05).

Figure 2. IHC staining of PD-L1 in TNBC specimen, and the relationship of PD-L1 expression with macrophage-derived cytokine corresponding receptors and with obesity/MS promoting genes in TCGA breast dataset

(A-a), PD-L1 expression at the edge of TNBC specimen. (A-b, c, d), Images which are in red, blue, and green frames are, respectively, representative part of mesenchyme tissue, tumor edge, and tumor inner tissue captured from (A-a). (B), The correlations heatmap of expression of macrophage-related receptors and obesity/MS-related genes with their PD-L1 expression counterpart in breast cancer samples. The expression data were downloaded from the UCSC Xena database and shown by log2(norm_count+1). Color red represents high expression and blue represents low. The receptors and genes were arranged by the order of their Pearson’s correlation coefficients, which are shown by bar graph below the heatmap, between them and PD-L1 expression. Symbols * and ^ represent statistically significant (P < .01 and P < .05, respectively).

Figure 3. Effects of macrophages and telmisartan exert on PD-L1 expression in breast epithelial and cancer cell lines

(a), Western blot and qPCR detection of PD-L1 expression in breast cancer cell lines and normal mammary cell line in different co-culture systems involving various macrophages (THP-1 derived M0, M1, telmisartan pretreated M1 (Tel-M1) and Wistar rat model derived NWM, ORM and telmisartan pretreated ORM (Tel-ORM)). (b), Concentration of IL6 and TNF-α was evaluated by ELISA in culture supernatant of THP-1, M0, M1, and Tel-M1 as well as in that of NWM, ORM, and Tel-ORM. (c), Western blot and qPCR detection of PD-L1 expression in breast cancer cell lines and breast epithelial cell line, respectively, treated with IL6, TNF-α and telmisartan. (d), Western blot and qPCR detection of PD-L1 expression in MDA-MBA-231 in coculture groups comprising M1, control lentivirus transfected M1 (M1-shcon), shIL6 lentivirus and transfected M1 (M1-shIL6), and humanized IL6 competitive inhibitor Tocilizumab treated M1 (M1+ Tocili). (e), PD-L1 expression of MDA-MBA-231 treated with different concentrations or different time duration of IL6. Column: mean; bar: SD. The symbol * represents a significant difference (P < .05).

Figure 4. JAK-STAT pathway is involved in MDA-MB-231 PD-L1 expression influenced by M1 and telmisartan

Involved protein nodes between IL6 and CD274 (PD-L1) illustrated in STRING protein-protein-interaction (PPI) map. The size and color of the protein nodes represent their connection degree with other proteins in the map. Lines between the nodes represent interactions between proteins. Red lines mark out the protein nodes which directly interact with CD274. (b), Western blot analysis of PD-L1, JAK1/2, and STAT1/3 as well as JAK1/2 and STAT1/3 phosphorylation in MDA-MB-231 in the control groups and coculture groups with various M1 (M1 alone, Tel-M1, Tel-M1 together with IL6 (Tel-M1+ IL6), M1-shcon and M1-shIL6). (c), Western blot detection of PD-L1 expression of MDA-MB-231 treated with IL6 or transfection by non-targeting siRNA (si-control), specific siRNAs targeting JAK1, JAK2, STAT1, or STAT3 (siJAK1, siJAK2, siSTAT1 or siSTAT3). The symbol * represents a significant difference (P < .05).

Figure 5. Telmisartan downregulates IL6 expression in M1 by activation of PPARG and inhibition of NFKB-p65

(a), Involved protein nodes between IL6 and CD274 (PD-L1) illustrated in STRING PPI map. Nodes size and color represent their connection degree with other proteins in the map. Lines between nodes represent the interactions between them. Red lines mark out the pathway that connects PPARG, NFKB, and IL6. (b), Western blot detection of PPARG, t-NFKB p65, pi-NFKB p65, and ELISA detection of IL6 in M0 and M1 treated with telmisartan, PPARG selective inhibitor T0070907 or NFKB transcriptional activity inhibitor JSH-23, respectively. The symbol * represents a significant difference (P < .05).