Table 1.
Patients’ characteristics.
Fig 1.
Tumor adjacent lymph nodes had lower CD8+ T cells cytotoxicity.
Patients’ samples were received from trans-urethral resection of the bladder (TUR-B) or cystectomies (n = 27). (A) Comparison of CD8+ T cells distribution was done between peripheral blood mononuclear cells (PBMC), sentinel nodes (SN), and tumor after immunophenotyping by flow cytometry. CD8+ T cells distribution was calculated from CD3+ lymphocytes population. The data are means with the error bars indicating SEM. One-way ANOVA was used as the statistical test. (B) The cytotoxic phenotype of CD8+ T cells in different tissues from UBC patients was analyzed by flow cytometry. The frequencies of granzyme B expressing CD8+ T cells were compared among PBMC, SN, and tumor. (C) Same as in (B) but the analysis was done on perforin expressing CD8+ T cells. The black middle lines indicate median. Kruskal-Wallis was used as the statistical test. * p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.
Fig 2.
Perforin deficiency in CD8+ T cells from sentinel nodes.
(A) The expression of granzyme B and perforin in CD8+ T cells of different tissues were phenotyped by flow cytometry. The co-expression pattern in CD8+ T cells was shown in dot plots and gated for distinguishing between double and single expression of granzyme B and perforin. The gate was based on isotype control and the frequency of granzyme B and perforin expression was counted out of CD8+ T cells. Dot plots showed a representative data from a patient underwent transurethral resection of the bladder (TUR-B) and cystectomy. (B) The frequency of granzyme B+/perforin+ CD8+ T cells from PBMC, SN, and tumor tissues was shown in graphs and was calculated out of CD8+ T cells (n = 27). (C) Same as in (B) but the analysis was done on granzyme B+/perforin- CD8+ T cells. The data are means with the error bars indicating SEM. Kruskal-Wallis was used as the statistical test. (D) The expression of gene responsible in encoding granzyme B (GZMB) and perforin (PRF1) in CD8+ T cells isolated from PBMC, SN, and tumor (n = 6). RT-qPCR was done to analyze the gene expression followed by quantification using 2-ΔΔCt method. The fold change was calculated in regards of PBMC as control, with RPII gene used as the housekeeping gene. The data are the means of Log2 of fold change (2-ΔΔCt) with the error bars indicating SEM. Kruskal-Wallis was used as the statistical test on each gene. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.
Fig 3.
Production and secretion of perforin in SN CD8+ T cells are low after in vitro reactivation.
Lymphocytes isolated from peripheral blood (PBMC) and sentinel node (SN) were cultured for seven days with addition of autologous tumor homogenate. (A) Flow cytometry was done to phenotype the co-expression in CD8+ T cells from PBMC and SN before and after reactivation. The results were shown in dot plots and gated based on isotype control. The frequency of granzyme B and perforin expression was counted out of CD8+ T cells. Dot plots showed data from a representative cystectomized patient. (B) Intracellular perforin was measured by Median Fluorescence Intensity (MFI) post 7-day culture using flow cytometry from (A). The data are means with error bars indicating SEM. Mann-Whitney was used as the statistical test. (C) The concentrations (pg/ml) of secreted granzyme B and perforin after seven days of culture were analyzed by ELISA and compared between in vitro culture supernatants of PBMC and SN. The data are means with error bars indicating SEM. Mann-Whitney was used as the statistical test. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.
Fig 4.
Sentinel node CD8+ T cells with perforin deficiency are exhausted Tc2 cells.
(A) CD8+ T cells isolated from sentinel node (SN) were further phenotyped using flow cytometry to demonstrate the difference in T cells exhaustion markers expression (PD-1) and Tc1 transcription factor (T-bet) between granzyme B+/perforin−CD8+ T cells (green box) and granzyme B+/perforin+ CD8+ T cells (red box). The expression of PD-1 and T-bet were shown in dot plots from a representative patient and gated based on isotype control. (B) The frequency of PD-1 and T-bet from (A) was calculated either out of granzyme B+/perforin−or granzyme B+/perforin+ CD8+ T cells. The data are means with the error bars indicating SEM. Mann-Whitney was used as the statistical test. (C) The expression of T-bet, encoded by TBX21 gene, was compared among CD8+ T cells sorted from peripheral blood mononuclear cells (PBMC), sentinel node (SN), and tumor. mRNA was extracted from the sorted cells and the TBX21 gene expression was analyzed by RT-qPCR. The expression of TBX21 was quantified using 2-ΔΔCt method and the fold change was calculated in regards of PBMC as control. RPII gene was used as housekeeping gene. The data are means with error bars indicating SEM. Kruskal-Wallis was used as the statistical test. (D) Same as in (C), but the analysis was done on GATA3 gene expression. (E) The frequency of naïve T cells (CD45RA+ CCR7+), central memory T (TCM) cells (CD45RA- CCR7+), effector memory T (TEM) cells (CD45RA- CCR7-), and effector memory T with CD45RA expression (TEMRA) cells (CD45RA+ CCR7-) was calculated either out of granzyme B+/perforin−or granzyme B+/perforin+ CD8+ T cells. The data are means with the error bars indicating SEM. Kruskal-Wallis was used as the statistical test. * p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.
Fig 5.
Tc1 conditions can restore perforin expression in CD8+ T cells from sentinel nodes.
CD8+ T cells sorted from sentinel node (SN) were cultured in Tc1 conditions in vitro for seven days in order to rescue perforin expression. These SN-derived CD8+ T cells were stimulated with anti-CD3 and anti-CD28 stimulating antibodies with the presence of IL-12 and IL-2 cytokines, as well as an anti-IL-4 neutralizing antibody. At the end of the culture, cells were analyzed by flow cytometry and RT-qPCR. (A) Dot plots showed the flow cytometry data from a representative patient for granzyme B vs. perforin expression, before and after the stimulation. The gate was based on isotype control and the frequency of granzyme B and perforin expression was counted out of CD8+ T cells. (B) Flow cytometry result of T-bet expression percentage from CD8+ T cells pre- and post-stimulation was analyzed. The frequency of T-bet expression was calculated from CD8+ T cells. (C) TBX21 and GATA3 gene expression analysis was done by RT-qPCR from cells in different culture conditions. RPII gene was used as housekeeping gene and the fold change was calculated based on cells without IL-12 and anti-IL-4 as control using 2-ΔΔCt method.
Fig 6.
ICAM-1 and TGFβ2 signal from muscle invasive UBC causes perforin downregulation.
Culture supernatants of urothelial bladder cancer (UBC) cell lines were acquired from RT4 (non-muscle invasive) and 5637 (muscle invasive) cell lines. CD8+ T cells were then isolated from peripheral blood of healthy donor and cultured in vitro with these supernatants for five days. (A) Analysis of perforin coding gene (PRF1) expression was done by RT-qPCR. mRNA was extracted post-culture from the cells of the culture groups. Bar graphs show different expression of PRF1 in CD8+ T cells cultured in vitro between RT4 (non-muscle invasive) and 5637 (muscle invasive) supernatant. RPII gene was used as housekeeping gene and the fold change was calculated in regards of RT4 medium using 2-ΔΔCt method. The data are means with error bars indicating SEM. Paired-t-test was used as the statistical test. (B) Flow cytometry analysis of CD8+ T cells at the end of culture was done. The results comparing three groups were shown in dot plots from a representative healthy donor and gated based on isotype control. (C) The frequency of perforin- CD8+ T cells from (B) was counted out of CD8+ T cells. The data are means with the error bars indicating SEM. One-way repeated-measure ANOVA was used as the statistical test. (D) Mass spectometry (MS) analysis identified proteins expressed by RT4 and 5637 cell line. Proteins under the category “immune system process” on the GO (Gene Ontology) term were selected for network analysis based on STRING database. Size represented differential expression between RT4 and 5637 supernatants and the color represented betweenness which marked the influence of the protein to the network. Color indicators: blue = low, yellow = average and red = high. (E) The expression of ICAM-1 was validated by flow cytometry on RT4 and 5637 cell line. RT4 and 5637 cells were identified by EpCAM expression. (F) Validation of perforin downregulation by ICAM-1 and TGFβ2 was done in vitro on CD8+ T cells isolated from healthy donors in the presence of anti-CD3 stimulating antibody for 5 days. Perforin coding gene (PRF1) expression was done by RT-qPCR. mRNA was extracted post-culture from the cells. Bar graphs show different expression of PRF1 in CD8+ T cells cultured in vitro between control and TGFβ2 + ICAM-1 + αCD3. RPII gene was used as housekeeping gene and the fold change was calculated in regards of blank medium using 2-ΔΔCt method. The data are means with error bars indicating SEM. Paired-t-test was used as the statistical test. (G) Flow cytometry analysis of CD8+ T cells at the end of culture was done. The results were shown in dot plots and gated based on isotype control from a representative healthy donor. The frequency of granzyme B and perforin expression was counted out of CD8+ T cells. (H) The frequency of granzyme B+/perforin- expressing cells from (G) was counted out of CD8+ T cells. The data are means with error bars indicating SEM. Paired-t-test was used as the statistical test. * p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.