Workflow automation and parallelization improves the isolation … · 2018. 4. 15. · cells after...
1
Workflow automation and parallelization improves the isolation and analysis of tumor-infiltrating immune cell subpopulations Janina Brauner, David Agorku, Anne Langhammer, Lukas Pohlig, Jan Drewes, Thomas Dino Rockel, Oliver Schilling, Stephan Mehlhose, Wael Al Rawashdeh, Andreas Bosio, and Olaf Hardt Miltenyi Biotec GmbH, Bergisch Gladbach, Germany Permanent Abstract Number: 1672 Introduction Conclusion Immunotherapy against cancer has proven clinical efficacy and tremendous potential in multiple tumor entities. Syngeneic mouse tumor models represent the gold standard to analyze effects of immunotherapy, as they possess a fully competent immune repertoire. However, the amount and composition of tumor-infiltrating leukocytes (TILs) is highly variable, compli- cating the analysis of individual subpopulations. In particular, small subpopu- lations might escape analysis as they could get lost in the background noise. When working with large cohort sizes, even immunophenotyping of TILs by flow cytometry is time consuming and data processing highly work inten- sive. Therefore, pre-enrichment of TILs is highly desirable to increase the sensitivity of analysis and save time and effort during flow cytometry. To this end, we have established an automated workflow combining tissue dissocia- tion with specific TIL isolation. Tumor dissociation was automated using the gentleMACS™ Octo Dissociator and optimized for epitope preservation to overcome bias in immunophenotyping caused by dissociation with aggres- sive or impure enzymes. Moreover, isolation of TILs was improved by devel- oping a new CD45-specific enrichment reagent for magnetic cell sorting, based on MACS® Technology, directly from dissociated tumor tissue. The whole workflow takes only about 90 min to complete. To fulfill the need for parallelization and automation of sample processing in large cohort sizes used in in vivo studies, a fully automated version of the MultiMACS™ Cell24 Separator was developed by integrating the instrument into a robotic liquid handling platform. This new system, the MultiMACS X, can perform 24 cell separations in parallel with minimal hands-on time. Automation and parallelization of tumor dissociation has been achieved previously by developing the gentleMACS Octo Dissociator. As TIL subpopulation analysis is a common readout for pre-clinical in vivo studies in the immuno-oncology field, the method for TIL isolation had to be adapted to large cohort sizes as well. To address this need, a fully automated version of the MultiMACS Cell24 Separator was developed by integrating the instrument into a robotic liquid handling platform. This new system, the MultiMACS X (fig. 3A), can process 24 cell separations in parallel with minimal hands-on time. When compared to the manual system, equal purities were achieved, while the overall yield of target cells was increased from 70% to 90% as shown for TILs from CT26.WT tumors (fig. 3B). We have developed a new CD45-specific enrichment reagent for magnetic cell sorting (MACS Technology) directly from dissociated tumor tissue (fig. 1A). To validate this method for starting material showing variable frequencies of TIL infiltration, we used syngeneic mouse tumors induced by injection of three different murine tumor cell lines. Tumors derived from B16-F10 melanoma cells showed TIL frequencies of 2–4% among total viable cells after dissociation. CT26.WT cell–derived colon carcinoma contained 15–21%, and 4T1 cell–derived breast carcinoma 32–37% TILs. Using a manual separation system, TILs were enriched to purities above 90% at yields above 70% for CT26.WT and 4T1 tumors, and purities above 80% at high yields above 95% for B16-F10 tumors (fig. 1B and C). We used flow cytometry to detect and quantify TIL subpopulations in syngeneic mouse tumors induced by subcutaneous transplantation of B16-F10 melanoma cells. This tumor model showed TIL frequencies of only 2–4%, making it difficult to analyze smaller immune cell subpopulations in a reliable way. To directly compare the results from samples without (fig. 2A) or with (fig. 2B) prior CD45-specific TIL enrichment, we kept the number of acquired events, and therefore acquisition time, constant. In both cases, 100,000 events were acquired for subsequent analysis. In the sample without TIL enrichment, it was hard to detect and quantify immune cell subpopulations. This was particularly obvious for small subpopulations, such as plasmacytoid dendritic cells (pDCs), which could be detected reliably only in the pre-enriched sample. Importantly, the composition of infil- trating immune cell populations was not affected, excluding the risk of introducing a bias by this method (fig. 2C). An alternative approach to increase sensitivity would be to acquire more events during flow cytometric analysis. However, for this tumor, the overall TIL enrichment factor was 48-fold. This means that an acquisition of 4,800,000 events from the unseparated fraction would have been necessary to reach an equal level of detection as the acquisition of 100,000 events from the enriched fraction. Reliable and fast isolation of TILs from syngeneic mouse tumors 1 Immunophenotyping of TIL subpopulations 2 Automation and parallelization of TIL isolation 3 Results Figure 1 • We have developed an automated workflow for the isolation of TILs from mouse tumors, streamlining downstream analysis while standardizing and enhancing the detection and quantification of immune cell subpopulations. • Importantly, while the TIL enrichment significantly reduced time and reagent costs for immune cell subset analysis, the composition of infiltrating immune cell populations was not affected, excluding the risk of introducing a bias by this method. Unless otherwise specifically indicated, Miltenyi Biotec products and services are for research use only and not for therapeutic or diagnostic use. MACS, the MACS logo, gentleMACS, MultiMACS, Vio, VioBlue, and VioGreen are registered trademarks or trademarks of Miltenyi Biotec GmbH and/or its affiliates in various countries worldwide. Copyright © 2017 Miltenyi Biotec GmbH and/or its affiliates. All rights reserved. A B C A B C Figure 2 Figure 2 (continued) CT26.WT–induced tumor Bulk tumor Negative fraction Isolated TILs B16-F10–induced tumor 4T1–induced tumor Bulk tumor 90.65% 2.27 % CD4 + T cells 31.96% CD11c + 21.50% CD11b + cDCs 38.10% TAM-A 8.30% TAM-B 26.64% CD11c + 2.23% DN 93.31% 70.01% B cells 1.82% NK cells 2.74% CD11b + 27.67% CD3ε + 10.00% CD11b low 61.29% DN CD3ε + cells 49.48% CD11b + Ly6G – 91.28% CD 11b – cDCs 13.91% CD8 + T cells 16.49% CD11c – 78.16% Inflammatory DCs 42.86% DN 14.41% TAM-C 50.66% Neutrophils 7.48% pDCs 3.11% Bulk tumor Isolated TILs CD45-PE Labeling Check Reagent-VioBlue® A B Enriched TIL fraction Initial frequency Purity Yield B16-F10–induced tumor (n = 5) 100% 80% 60% 40% 20% 0% Initial frequency Purity Yield CT26.WT–induced tumor (n = 12) 100% 80% 60% 40% 20% 0% 100% 80% 60% 40% 20% 0% Initial frequency Purity Yield 4T1–induced tumor (n = 8) 0 50 100 150 Total cell count 200 3,000 6,000 9,000 12,000 15,000 CD4 + T cells CD8 + T cells NK cells B cells CD11b + cDCs Inflammatory DCs CD 11b – cDCs DN CD11b + cells TAM-A TAM-B TAM-C Neutrophils pDCs POS ORI 100% 80% 60% 40% 20% 0% Initial frequency Purity Yield 4T1–induced tumor (n = 8) Manual cell separation MultiMACS Cell24 Separator Plus MultiMACS X Manual cell separation MultiMACS Cell 24 Separator Plus MultiMACS X Side scatter Forward scatter CD45-VioGreen™ CD3ε-PE CD3ε-PE Anti-NKp46-FITC CD19-APC Anti-Ly-6C-PE-Vio 770 Anti-Ly-6C-PE-Vio 770 CD8a-APC Side scatter Anti-Ly-6G-FITC Anti-Ly-6C-PE-Vio® 770 Propidium iodide Side scatter CD11b-VioBlue® Side scatter CD11c-APC-Vio 770 Anti-MHC Class II-APC CD4-APC-Vio 770 CD11c-APC-Vio 770 Anti-MHC Class II-APC Anti-MHC Class II-APC Side scatter 88.37% 96.97 % CD4 + T cells 44.95% CD11c + 30.24% TAM-A 10.11% TAM-B 31.79% CD11c + 19.76% DN 97.49% 51.73% B cells 16.72% NKcells 1.53% DN CD3ε + cells 24.72% CD 11b – cDCs 26.78% CD8 + T cells 28.46% CD11c – 69.63% DN 12.34% TAM-C 45.07% pDCs 7.43% Side scatter Forward scatter CD45-VioGreen CD3ε-PE CD3ε-PE Anti-NKp46-FITC CD19-APC Anti-Ly-6C-PE-Vio 770 Anti-Ly-6C-PE-Vio 770 CD8a-APC Side scatter Anti-Ly-6G-FITC Anti-Ly-6C-PE-Vio® 770 Propidium iodide Side scatter CD11b-VioBlue Side scatter CD11c-APC-Vio 770 Anti-MHC Class II-APC CD4-APC-Vio 770 CD11c-APC-Vio 770 Anti-MHC Class II-APC Anti-MHC Class II-APC Side scatter CD11b + 56.97% CD3ε + 11.94% CD11b low 28.47% CD11b + Ly6G – 91.33% Neutrophils 7.98% CD11b + cDCs 51.26% Inflammatory DCs 36.71%
Transcript of Workflow automation and parallelization improves the isolation … · 2018. 4. 15. · cells after...
-
Workflow automation and parallelization improves the isolationand analysis of tumor-infiltrating immune cell subpopulations
Janina Brauner, David Agorku, Anne Langhammer, Lukas Pohlig, Jan Drewes, Thomas Dino Rockel, Oliver Schilling, Stephan Mehlhose, Wael Al Rawashdeh, Andreas Bosio, and Olaf HardtMiltenyi Biotec GmbH, Bergisch Gladbach, Germany
Permanent Abstract Number: 1672
Introduction
Conclusion
Immunotherapy against cancer has proven clinical efficacy and tremendous potential in multiple tumor entities. Syngeneic mouse tumor models represent the gold standard to analyze effects of immunotherapy, as they possess a fully competent immune repertoire. However, the amount and composition of tumor-infiltrating leukocytes (TILs) is highly variable, compli-cating the analysis of individual subpopulations. In particular, small subpopu-lations might escape analysis as they could get lost in the background noise. When working with large cohort sizes, even immunophenotyping of TILs by flow cytometry is time consuming and data processing highly work inten-sive. Therefore, pre-enrichment of TILs is highly desirable to increase the sensitivity of analysis and save time and effort during flow cytometry. To this end, we have established an automated workflow combining tissue dissocia-
tion with specific TIL isolation. Tumor dissociation was automated using the gentleMACS™ Octo Dissociator and optimized for epitope preservation to overcome bias in immunophenotyping caused by dissociation with aggres-sive or impure enzymes. Moreover, isolation of TILs was improved by devel-oping a new CD45-specific enrichment reagent for magnetic cell sorting, based on MACS® Technology, directly from dissociated tumor tissue. The whole workflow takes only about 90 min to complete. To fulfill the need for parallelization and automation of sample processing in large cohort sizes used in in vivo studies, a fully automated version of the MultiMACS™ Cell24 Separator was developed by integrating the instrument into a robotic liquid handling platform. This new system, the MultiMACS X, can perform 24 cell separations in parallel with minimal hands-on time.
Automation and parallelization of tumor dissociation has been achieved previously by developing the gentleMACS Octo Dissociator. As TIL subpopulation analysis is a common readout for pre-clinical in vivo studies in the immuno-oncology field, the method for TIL isolation had to be adapted to large cohort sizes as well. To address this need, a fully automated version of the MultiMACS Cell24 Separator was developed by integrating the
instrument into a robotic liquid handling platform. This new system, the MultiMACS X (fig. 3A), can process 24 cell separations in parallel with minimal hands-on time. When compared to the manual system, equal purities were achieved, while the overall yield of target cells was increased from 70% to 90% as shown for TILs from CT26.WT tumors (fig. 3B).
We have developed a new CD45-specific enrichment reagent for magnetic cell sorting (MACS Technology) directly from dissociated tumor tissue (fig. 1A). To validate this method for starting material showing variable frequencies of TIL infiltration, we used syngeneic mouse tumors induced by injection of three different murine tumor cell lines. Tumors derived from B16-F10 melanoma cells showed TIL frequencies of 2–4% among total viable
cells after dissociation. CT26.WT cell–derived colon carcinoma contained 15–21%, and 4T1 cell–derived breast carcinoma 32–37% TILs. Using a manual separation system, TILs were enriched to purities above 90% at yields above 70% for CT26.WT and 4T1 tumors, and purities above 80% at high yields above 95% for B16-F10 tumors (fig. 1B and C).
We used flow cytometry to detect and quantify TIL subpopulations in syngeneic mouse tumors induced by subcutaneous transplantation of B16-F10 melanoma cells. This tumor model showed TIL frequencies of only 2–4%, making it difficult to analyze smaller immune cell subpopulations in a reliable way. To directly compare the results from samples without (fig. 2A) or with (fig. 2B) prior CD45-specific TIL enrichment, we kept the number of acquired events, and therefore acquisition time, constant. In both cases, 100,000 events were acquired for subsequent analysis. In the sample without TIL enrichment, it was hard to detect and quantify immune cell subpopulations. This was particularly obvious for small subpopulations,
such as plasmacytoid dendritic cells (pDCs), which could be detected reliably only in the pre-enriched sample. Importantly, the composition of infil-trating immune cell populations was not affected, excluding the risk of introducing a bias by this method (fig. 2C). An alternative approach to increase sensitivity would be to acquire more events during flow cytometric analysis. However, for this tumor, the overall TIL enrichment factor was 48-fold. This means that an acquisition of 4,800,000 events from the unseparated fraction would have been necessary to reach an equal level of detection as the acquisition of 100,000 events from the enriched fraction.
Reliable and fast isolation of TILs from syngeneic mouse tumors1
Immunophenotyping of TIL subpopulations2
Automation and parallelization of TIL isolation3Results
Figure 1
• We have developed an automated workflow for the isolation of TILs from mouse tumors, streamlining downstream analysis while standardizing and enhancing the detection and quantification of immune cell subpopulations.
• Importantly, while the TIL enrichment significantly reduced time and reagent costs for immune cell subset analysis, the composition of infiltrating immune cell populations was not affected, excluding the risk of introducing a bias by this method.
Unless otherwise specifically indicated, Miltenyi Biotec products and services are for research use only and not for therapeutic or diagnostic use. MACS, the MACS logo, gentleMACS, MultiMACS, Vio, VioBlue, and VioGreen are registered trademarks or trademarks of Miltenyi Biotec GmbH and/or its affiliates in various countries worldwide. Copyright © 2017 Miltenyi Biotec GmbH and/or its affiliates. All rights reserved.
A
B
C
A
B
C
Figure 2
Figure 2 (continued)
CT26.WT–induced tumor
Bulk tumor Negative fraction Isolated TILs
B16-F10–induced tumor 4T1–induced tumor
Bulk tumor
90.65% 2.27 %
CD4+ T cells31.96%
CD11c+21.50%
CD11b+ cDCs38.10%
TAM-A8.30%
TAM-B26.64%
CD11c+2.23%
DN93.31%
70.01%
B cells1.82%
NK cells2.74%
CD11b+27.67%
CD3ε+10.00%
CD11blow61.29%
DN CD3ε+ cells49.48%
CD11b+ Ly6G–91.28%
CD 11b– cDCs13.91%
CD8+ T cells16.49%
CD11c–78.16%
Inflammatory DCs42.86%
DN14.41%
TAM-C50.66%
Neutrophils7.48%
pDCs3.11%
Bulk
tum
orIs
olat
ed T
ILs
CD45
-PE
Labeling Check Reagent-VioBlue®
A B
Enriched TIL fraction
Initial frequency
Purity Yield
B16-F10–induced tumor (n = 5)
100%
80%
60%
40%
20%
0%
Initial frequency
Purity Yield
CT26.WT–induced tumor (n = 12)
100%
80%
60%
40%
20%
0%
100%
80%
60%
40%
20%
0%
Initial frequency
Purity Yield
4T1–induced tumor (n = 8)
0
50
100
150
Tota
l cel
l cou
nt
200
3,000
6,000
9,000
12,000
15,000
CD4+
T ce
lls
CD8+
T cell
s
NK ce
llsB c
ells
CD11b
+ cDCs
Inflam
mator
y DCs
CD 11
b– cD
Cs
DN CD
11b+ ce
llsTA
M-A
TAM-
B
TAM-
C
Neutr
ophil
spD
Cs
POSORI
100%
80%
60%
40%
20%
0%
Initial frequency
Purity Yield
4T1–induced tumor (n = 8)
Manual cell separation MultiMACS Cell24 Separator Plus MultiMACS X
Manual cell separation
MultiMACS Cell 24 Separator Plus
MultiMACS X
Side
sca
tter
Forward scatter CD45-VioGreen™ CD3ε-PE
CD3ε-PE Anti-NKp46-FITC
CD19-APC Anti-Ly-6C-PE-Vio 770
Anti-Ly-6C-PE-Vio 770
CD8a-APC Side scatter
Anti-Ly-6G-FITC
Anti-Ly-6C-PE-Vio® 770
Prop
idiu
m io
dide
Side
sca
tter
CD11
b-Vi
oBlu
e®
Side
sca
tter
CD11
c-AP
C-Vi
o 77
0
Anti-
MH
C Cl
ass
II-AP
CCD
4-AP
C-Vi
o 77
0
CD11
c-AP
C-Vi
o 77
0
Anti-
MH
C Cl
ass
II-AP
CAn
ti-M
HC
Clas
s II-
APC
Side
sca
tter
88.37% 96.97 %
CD4+ T cells44.95%
CD11c+30.24%
TAM-A10.11%
TAM-B31.79%
CD11c+19.76%
DN97.49%
51.73%
B cells16.72%
NKcells1.53%
DN CD3ε+ cells24.72%
CD 11b– cDCs26.78%
CD8+ T cells28.46%
CD11c–69.63%
DN12.34%
TAM-C45.07%
pDCs7.43%
Side
sca
tter
Forward scatter CD45-VioGreen CD3ε-PE
CD3ε-PE Anti-NKp46-FITC
CD19-APC Anti-Ly-6C-PE-Vio 770
Anti-Ly-6C-PE-Vio 770
CD8a-APC Side scatter
Anti-Ly-6G-FITC
Anti-Ly-6C-PE-Vio® 770
Prop
idiu
m io
dide
Side
sca
tter
CD11
b-Vi
oBlu
e
Side
sca
tter
CD11
c-AP
C-Vi
o 77
0
Anti-
MH
C Cl
ass
II-AP
CCD
4-AP
C-Vi
o 77
0
CD11
c-AP
C-Vi
o 77
0
Anti-
MH
C Cl
ass
II-AP
CAn
ti-M
HC
Clas
s II-
APC
Side
sca
tter
CD11b+56.97%
CD3ε+11.94%
CD11blow28.47%
CD11b+ Ly6G–91.33%
Neutrophils7.98%
CD11b+ cDCs51.26%
Inflammatory DCs 36.71%
