您当前的位置: 首页 > 网页快照
Human tissue cultures of lung cancer predict patient susceptibility to immune-checkpoint inhibition | Cell Death Discovery
Results . Heterogeneity and morphologic preservation of lung cancer slice . A tissue culture protocol for cultivation of patient-derived lung cancer specimens for up to 12 days ex vivo was successfully established (Fig. 1 ). To adjust for differences in auto-fluorescence characteristics and staining intensities, we combined stain-specific sequences of processing and segmentation algorithms in Image J [ 19 ]. Slice cultures showed a good preservation of morphologic features of the original tumor (Fig. 1A ). Parameters including cell formation, mucin production, and staining characteristics were assessed for comparison from 24 tumor specimens. In all samples, the pathological diagnosis of the original tumor matched the features found in baseline culture, representatively shown in Table 1 . 4/24 cases (16,6%) were excluded from the analysis due to high necrotic areas and inflammatory regions observed in the baseline tissue. Specimens not shown in the table were used for adjustments of the culture conditions. As shown in Fig. 1B TFs were maintained up to day 12 ex vivo. The overall TF of 31%?±?12 ( n ?=?3, mean?±?SD) was nonsignificantly enhanced from day 2 to day 10 up to 53%?±?12 and to 68%?±?14 at day 12. Fig. 1: Tissue maintenance and dose-dependent response ex vivo. A Slices from tumor specimen were stained with HE at baseline or after culture in control conditions for up to 12 days. Shown are representative HE staining of the two different NSCLC subtypes at baseline and after 10 days ex vivo. B The tumor fraction present in the cultured specimens was evaluated and normalized to the baseline. Shown are the results from three different patients. To control reproducibility multiple slides from the same patients were measured at the same timepoint. C Slices were cultured for 3 or 6 days with different doses of cisplatin (1, 10??M) or paclitaxel (1, 10, 100?nM). To calculate the proliferating tumor fractions tissues were stained for Ki67 (green), pancytokeratine (red), and counterstained with Hoechst (blue). Shown are pictures of the different treatment at day 3 and 6 (right) as well as the overall variations in proliferating tumor fraction within the one tissue specimen for which enough material was available to perform all conditions with at least two replicates (left). Background, yellow; Bar?=?50??m. Full size image Table 1 Patient specimens. Full size table Effects of cytotoxic drugs on tumor cell proliferation . Slice cultures were treated with different doses of the cytotoxic drugs cisplatin (1, 10??M) or paclitaxel (1, 10, and 100?nM). The effect on tumor cell proliferation was determined by antibody staining of Ki67. Whereas nonsignificant changes due to high intratumor variances were detected after 3 days of treatment, a significant trend was found after 6 days in culture. Low doses of both drugs resulted in an increased tumor proliferation rate compared to the control condition. In contrast, higher concentrations of both drugs showed a dose-dependent decrease in tumor proliferation (Fig. 1C ). Response to nivolumab is independent of lymphocyte supplementation in tumor tissue cultures . To investigate nivolumab susceptibility, the presence of T-lymphocytes in the cultivated tissue is required. Therefore, we evaluated whether immune cells still present in the NSCLC tissue would be sufficient to detect an effect of the immunotherapy or if an “exogenous” provision using lymphatic tissue would be required. The co-culture of NSCLC with autologous lymphatic tissue in the absence of additional treatment did not result in a significant reduction of the tumor cell proliferation at day 5 and day 10 (Fig. 2A ). In contrast, nivolumab supplementation caused a decreased proliferation of TF in the presence as well as in the absence of lymphatic tissue. This effect was already detected at day 5, but was more pronounced at day 10 ex vivo (Fig. 2A ). The HE staining of the co-culture in Fig. 2B demonstrated a high similarity with tissue of the tumor area. A reduced proliferation in the presence of nivolumab with respect to the control condition (Fig. 2A ) was accompanied by a diminished number of tumor cells displaying a reduced metabolism, as shown by the calcein live staining (Fig. 2C ). Fig. 2: Nivolumab tissue response is independent of lymphocyte supplementation. NSCLC tissue cultures were cultivated alone (T) or with autologous lymphatic tissue (LT) in the presence or absence of Nivolumab for 5 or 10 days. A Percentages of Ki67 proliferating tumor fractions in the different culture conditions from one patient specimen are shown upon normalization to the control culture in the absence of lymphocytes (left) Kruskal–Wallis test showed significant differences. Representative pictures of the different conditions at day 5 and 10 are shown (right). pancytokeratine, red; Ki67, green; Hoechst, blue; background, white. Bar?=?50??m. B Shown is the HE staining of a NSCLC tissue co-cultured with lymphatic tissue slice at day 10. Overview, bar?=?500??m; HE magnification, bar?=?50??m. C Calcein live staining (green) of a control tissue (top) and a nivolumab treated tissue (bottom) after 5 days in co-culture with lymphatic tissue. Bar?=?200??m. Full size image Heterogeneous response of tumor tissues to therapy . The individual susceptibility of different sNSCLC samples upon application of cisplatin, nivolumab, or paclitaxel was investigated by determination of proliferation and apoptosis (Fig. 3 ). Adjacent tissue served as a control in order to determine potential harming effects of the different treatments on the healthy epithelial tissue. HE sections are shown in Fig. 3A and B. In the tumor fraction, treatment with cisplatin or paclitaxel reduced the percentage of Ki67+ cells (Fig. 3C and D, left portion) and induced apoptosis. (Fig. 3E and F, left portion). Interestingly, the culture control conditions induced the proliferation as well as an enhanced apoptosis in healthy epithelial cells of the adjacent tissue (Fig. 3C–F ). Both chemotherapy treatment reduced the “culture-induced” proliferation whereas the apoptotic rate was further augmented by cisplatin application in contrast to paclitaxel that induced much lower apoptosis than control conditions (Fig. 3C–F , right portion). Evaluation of the effect of immunotherapy with nivolumab highlighted responder and nonresponder tissue culture: In patient #12 tumor cells still proliferated despite an increased apoptosis induction (Fig. 3C and E), whereas in patient #11 apoptosis induction was also paralleled by a significantly reduced frequency of proliferating tumor cells (Fig. 3D and F). The effects of nivolumab on the adjacent tissue did not significantly differ to that from control conditions (Fig. 3C –F, right portion) Fig. 3: Individual susceptibility in squamous cell carcinoma of the lung and in adjacent tissue. Tissue specimens were treated for 5 days with nivolumab (3??g/ml), cisplatin (10??M), or paclitaxel (100?nM, only #12), depending on tissue availability. Shown are representative outcome from patient #12 that did not respond to nivolumab ex vivo (left) and patient #11 that responded (right). Representative HE sections of the tumor as well as of the adjacent tissue of #12 ( A ) and #11 ( B ) at 6 days in culture are shown. Bar?=?100??m. Tumor tissue and adjacent healthy epithelial tissue were evaluated for proliferation ( C, D ) as well as for apoptosis ( E, F ). The adjacent tissue of #12 was not free of tumor cells but only tumor-free areas were evaluated. p ?one-way Anova tested against the control condition. Base baseline. Full size image Unspecific effects of nivolumab by applying an IgG4 antibody . To have deeper insights into the specific effect of nivolumab, exemplary samples were treated with nivolumab or with the IgG4 isotype control Ab. As shown in Fig. 4A , the tissue of patient #14 was neither significantly affected by nivolumab nor by the isotype control treatment regarding proliferation, apoptosis, the percentage of proliferating cells, or the total tumor fraction. In contrast, nivolumab, but not the IgG4 Ab reduced total tumor fraction as well as the percentage of proliferating tumor cells in specimen #15 (Fig. 4B ). Fig. 4: Altered T-cell repertoire after nivolumab treatment in NSCLC. A, B Tissue specimens #14 ( A ) and #15 ( B ) were cultured ex vivo under control condition or upon addition of nivolumab or its isotype control, an IgG4 Ab. After 5 days of treatment the tissue were stained for Ki67 or cleaved PARP to evaluate the percentage of proliferating or apoptotic cells as well as for the total fraction of tumor within the tissue. Kruskal–Wallis test was performed. C Representative staining of tissue from #14 and #15 treated with nivolumab and stained with PD-L1 (red) and CD8 (green), Hoechst, blue; background, white. Bar?=?20??m. D Representative staining of #11 with CD3 (left) or FoxP3 (right) in control (top) and nivolumab (bottom) treated conditions. CD3, bar?=?50??m, FoxP3, bar?=?20??m. E Single-cell suspension obtained from the tissue specimens, cultured for 5 days in the presence or absence of nivolumab, were evaluated by flow cytometry. Shown are the percentages of the different T-cell populations as obtained with the gating strategy shown in supplementary Fig. 1 . NR nonresponder, R responder. Full size image Altered composition of the tumor microenvironment in response to nivolumab . The responses of the T-cell populations to nivolumab treatment ex vivo was investigated using conventional immunohistochemistry (IHC), multicolor flow cytometry and multispectral imaging. Immunohistological stainings for CD3, CD4, CD8, and FoxP3 was done for consecutive tissue cuts in #11, #12, #14, and #15. In all stainings, lymphocytes differed in size and their distribution within the stromal or tumoral region. Double staining for CD8 and PD-L1 indicated that the nonresponding tissue was PD-L1 negative and the responding tissue PD-L1 positive (Fig. 4C ). Staining for CD3 together with Foxp3 indicate that CD3+ cells maintained the distribution pattern and morphological characteristics of the native tissue and showed an invasion into the tumor site after nivolumab treatment. Foxp3+ T cells were observed in tumor stroma as well as in tumor regions. However, different sizes of lymphocytes were observed (Fig. 4C and D). In order to obtain quantitative information, the tissues were also disrupted and the obtained single-cell suspension was evaluated by flow cytometry. Out of seven experiments, we obtained two tissues responding towards nivolumab treatment, one squamous cell carcinoma and one adenocarcinoma. Five nonresponding specimens (three adenocarcinomas and two squamous cell carcinomas) were also evaluated by flow cytometry, with three of them that did not show any alterations upon treatment and are consequently not shown. In parallel to the histological differences detected in the size of the immune infiltrate, one could differentiate between a small lymphocyte fraction and a fraction that displayed enhanced granularity and size within the forward versus sideward scatter plot. The gating strategy is provided in supplementary Fig. 1 . The cellular fractions of the flow cytometric analysis are shown in Fig. 4E . The NSCLaC tissue (#15) responding to nivolumab treatment showed a decreased CD3, CD4, and Th1 population, while the Treg and Th2 populations were increased within the small lymphocytes. On the contrary, there was an increase in the CD3 and Th1 population within the big lymphocyte population as well as in the Treg cells. The nonresponding NSCLaC tissue (#14) revealed an increase in the CD3 population and a decrease in the CD4 population within the small lymphocytes. In the big lymphocyte gate, nivolumab treatment increased the population of CD3+ cells up to 10% with a prevalent expansion of CD8, Treg, and Th2 cells. The nonresponding sNSCLC tissue (#12) displayed significant lower amounts of CD3 positive cells within the big lymphocyte population (CTR: 30%; nivolumab: 26%) than in the small lymphocyte fraction (CTR: 71%; nivolumab 64%). In both lymphocyte fractions the CD4 population increased, while the CD8 population decreased. In addition, the small lymphocyte fraction showed an increase of the Th17 population, while the Th1 population decreased. An increase of the CD4, Th1, and Th2 population as well as lower levels of CD8, Treg, and Th17 cells were observed in the responding #11. The big lymphocyte population showed only decreased numbers of Treg and Th17 cells. The IgG4 conditions in #14 and #15 demonstrated only minor adaptations compared to the control condition (data not shown). To further investigate the immunological response in adenocarcinoma tissues, multispectral imaging was conducted (Fig. 5 ). Evaluation of three baseline slices from the same tumor underlined a good reproducibility of the technique as shown by almost identical frequencies for the majority of the evaluated populations (Fig. 5B ). The NSCLaC specimens #14 (nonresponder) and #15 (responder) were quantitatively analyzed for the effects of nivolumab treatment (Fig. 5C ). The frequency of panCK+ tumor cells was unaltered in the NR specimen, whereas it was reduced in response to nivolumab in the responder #15. Regarding the immune cells an expansion of B cells and CD4+ T cells in response to nivolumab was found in both tissues, even if it was stronger in the responder. CD8+ T cells as well as CD163+ macrophages were slightly reduced in both cases, whereas the Treg population showed an augmentation despite low absolute numbers of FoxP3+ cells. Fig. 5: Multispectral imaging (MSI) analysis discriminates responding tissues. Tissue was stained with Ab against CD3, CD8, CD163, CD20, Foxp3, and panCK and was evaluated by MSI. A Representative staining of the single antibody upon conversion of the multispectral image into a pathological DAB-like view. B Three tissue pieces from the same patient were stained and evaluated for the percentages of the different cell populations. C The frequency of the different immune populations as well as tumor cells within the tissue are shown for the responder and nonresponder samples treated or not with nivolumab for 5 days. D, E The spatial organization of tumor cells and immune cells was evaluated by calculating the frequency of cells having a cell from another phenotype within a 25?μm radius. Data are shown for relationship involving panCK+ cells ( D ) or FoxP3+ cells ( E ). Full size image Evaluation of the reciprocal distribution of tumor and immune cells upon nivolumab treatment highlighted major changes. Whereas the nonresponding tissue (#14) did not display changes in the relative distribution of T cells with respect to tumor cells, the tumor cells have more Treg, CD4, or CD8+ T cells in their proximity (Fig. 5D ) in the responding tissue #15. In contrast despite constant numbers in the tissue, Treg were less frequently in the proximity of a tumor cell (Fig. 5D ). In addition, in the responding patient less Treg have other immune cells in their proximity, whereas Treg were more close to them in the nonresponding tissue (Fig. 5E ). Thus, the spatial distribution of immune cells was influenced by nivolumab treatment demonstrating a reorganization of the TME. Clinical correlation . For two of the specimen the clinical outcome of patients after 30 months receiving nivolumab treatment was available and thus could be compared to the ex vivo results (Fig. 6 ). The tissue from one patient (Fig. 6A ) did not respond to nivolumab treatment as evaluated by unchanged proliferation of the tumor fraction, this patient died 15 months after surgery. In contrast, the tissue from another patient, shown in Fig. 6B , demonstrated a histological reaction to nivolumab ex vivo with a loss of proliferating tumor cells. This patient is still alive 30 months after surgery and nivolumab treatment. Fig. 6: NSCLaC tissue culture susceptibility to nivolumab correlates with patients’ clinical response. Tumor specimens from two patients, A a clinical nonresponder and B a clinical responder, were treated ex vivo with nivolumab (Nivo). Cell proliferation was evaluated after 5 days. p ?Full size image .
From:
系统抽取主题     
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)  
(1)