development of an integrated bioprocess for …
TRANSCRIPT
§ NK-‐92 cells exhibit cytotoxicity against a variety of cancer cell lines, including leukemia’s, lymphoma’s and malignant melanoma’s. This acAvity has lead to several small clinical trials as an anA-‐cancer immunotherapy, and larger mulA-‐centre Phase II trials are planned.
§ Our GOAL is to enable large-‐scale manufacture of clinical grade NK-‐92 cells, and to
reduce manufacturing costs. ProducAon using perfusion culture in sArred-‐tank bioreactors (STR) would enable a straight-‐forward high producAvity process that is readily scaled for commercial manufacture.
Xvivo%10(
Culture(Bags(
Human(AB(serum(
32%
19% 25%
5 x 109 cells/m2.dose
3 x 1010 cells/cycle.paAent [NK-‐92]max≈1 x 106 cells/mL
10 Gy
Current manufacturing strategy
Ricardo P. Bap1sta, Midori E. Buechli, Lesley Y. Chan, Shahryar KhaAak, Peter W. Zandstra and Nicholas E. Timmins
Centre for Commercializa1on of Regenera1ve Medicine, Toronto, Canada
§ We successfully translated NK-‐92 culture to STR, achieving significantly higher cell densiAes. Scale-‐up to 1-‐L cultures is currently in progress, as well as expansion using serum-‐free medium formulaAon.
§ Flow-‐cytometry and real-‐Ame imaging based assays are effecAve alternaAves to 51Cr-‐release potency assay. The BiostaAon CT (Nikon)
image based approach enables acquisiAon of kineAc data and observaAon of cell-‐cell interacAons.
BACKGROUND AND MOTIVATION
DEVELOPMENT OF AN INTEGRATED BIOPROCESS FOR PRODUCTION OF NK-92 CELLS FOR IMMUNOTHERAPY
Significance and on-‐going work
MEDIUM SCREENING
QC-‐PHENOTYPE AND POTENCY
0.0E+00
2.5E+05
5.0E+05
7.5E+05
1.0E+06
1.3E+06
1.5E+06
VCD (cells/m
L)
Xvivo S1 S9 DMEM
VCDmax cell/mL
Medium and consumables ($/cycle)
~1 x106 $11,400
TRANSLATION TO STR AND PROCESS INTENSIFICATION
0 1 2 3 4 5 6 7
0
20
40
60
80
100
0.0E+00
5.0E+05
1.0E+06
1.5E+06
2.0E+06
Viab
ility (%
)
VCD (cells/m
L)
0 1 2 3 4 5 6 7
0
20
40
60
80
100
0.0E+00
5.0E+05
1.0E+06
1.5E+06
2.0E+06
Viab
ility (%
)
VCD (cells/m
L)
2.5% HS 2.5% BIT
Figure 4. Effect of feeding-‐regime and culture system in NK-‐92 cell expansion. (A) Growth curves were obtained in 100 mL of sArred suspension culture (DasGip cellferm-‐pro). A 7.0-‐fold increase in cell numbers was observed when (B) a perfusion-‐like feeding-‐regime (1.0V) was adopted in the STR. (C) Metabolite analysis indicate that accumulaAon of lactate levels above 25 mM are inhibitory to cell growth.
0
20
40
60
80
100
0.0E+00
1.5E+06
3.0E+06
4.5E+06
6.0E+06
0 1 2 3 4 5 6
Viab
ility (%
)
VCD (cell/mL)
Time (d)
Batch
1.0V
1100 750 450 300
0.0
0.5
1.0
1.5
2.0
2 3 4 5 6 7
RPM
Fold Expan
sion
Inoculum Density (x105 cells/mL)
1100 750 450 300
0
20
40
60
80
100
2 3 4 5 6 7
RPM
Viab
ility (%
)
Inoculum Density (x105 cells/mL)
0
1
2
3
4
0 2 4 6 8
D (/d)
Time (d)
Batch 1.0V
0
10
20
30
40
0 2 4 6 8
Lac (m
M)
Time (d)
BASAL FORMULATION SERUM-‐REPLACEMENT CARBON SOURCE
INOCULUM DENSITY FED-‐BATCH CULTURE
Figure 1. Effect of basal medium formula1on, serum-‐replacement cocktail, and carbon source in NK-‐92 cell expansion. Several commercial basal formulaAons were able to support cell growth with viabiliAes (>80%) comparable to those observed in Xvivo (control) (n=3). A serum-‐replacement cocktail comprising BSA, rh-‐Insulin and rh-‐Transferrin was able to sustain cell expansion in staAc culture (n=3). Carbon source (S) was screened for cell growth and low secreAon of lactate (n=2). All cultures were seeded at approx. 5 x 105 cells/mL and carried out in a 6-‐well plate format for a minimum period of 6 days. Every 2 days, cells were diluted (passaged) to a viable cell density (VCD) of 5 x 105 cells/mL by replacing sufficient culture volume with fresh media. All medium formulaAons were supplemented with 450 U/mL of IL-‐2 and L-‐Gln. Scale bars = 200 µm.
PROCESS INTENSIFICATION
LACTATE
0
20
40
60
80
100
0.0E+00
1.0E+08
2.0E+08
3.0E+08
4.0E+08
5.0E+08
6.0E+08
0 2 4 6 7
Viab
ility (%
)
Total viable cells
Time (d)
XVivo
DMEM
STAT
IC
1
10
0 2 4 6 8
Log Fold Expan
sion
Time (d)
XVivo-‐STR XVivo-‐StaAc DMEM-‐STR DMEM-‐StaAc
Xvivo + 2.5% BIT
Culture system VCDmax x106 cells/mL
µ (d-‐1)
td (d)
Fold expansion
qLac (pmol/cell.d)
Xvivo – Sta1c (n=5) 1.1±0.1 0.24±0.09 3.0±1.1 6.1±0.8 7.1±1.3
Xvivo – STR (n=4) 1.6±0.2 0.33±0.08 2.3±0.7 11.3±1.6 7.9±2.5
DMEM – Sta1c (n=3) 1.1±0.2 0.29±0.04 2.5±0.4 6.8±0.5 8.2±1.3
DMEM – STR (n=3) 1.7±0.9 0.32±0.16 2.8±1.6 12.3±6.7 9.1±1.9
NK-‐92 PHENOTYPE
2.5% HS
Figure 2. Effect of inoculum density and agita1on rate in NK-‐92 cell expansion in STR. (A) Fold expansion and (B) cell viability across different seeding densiAes and 4 agitaAon rates) using the ambr bioreactor system (TAP Biosystems). The greatest increase in viable cell number was observed when cells were seeded at 5 x 105 cells/mL and sArred at 300-‐450 rpm.
Figure 3. Effect of basal medium and culture system in NK-‐92 cell expansion. (A,B) Growth curves were obtained in staAc (T-‐75 flask) and sArred suspension (DasGip cellferm-‐pro) culture. Medium comprised Xvivo10 (Lonza) or DMEM (LifeTech) supplemented with human AB serum and L-‐glutamine, and cultures were iniAated at a cell density of approx. 5 x 105 cells/mL. Culture volume was doubled every 2 days by addiAon of fresh media, up to a final volume 200 mL. (C) Specific growth rate (µ) was higher in STR compared to staAc culture in both medium formulaAons. (D) While NK-‐92’s grow as loose clumps in staAc culture, STR culture yielded a single-‐cell suspension in Xvivo and smaller aggregates in DMEM. Scale bars = 100µm.
STR
Xvivo DMEM DAY 6
0
20
40
60
80
100
0 6
CD45
+ /CD
56+ /CD
3-‐/Perf.+ (%
)
Time (d)
Xivo-‐STR
Xvivo-‐StaAc
DMEM-‐STR
DMEM-‐StaAc
Figure 5. Effect of culture system and basal medium on NK-‐92 phenotype. (A) Monitoring of NK-‐92 phenotype shows maintenance of a CD45+/CD56+/CD3-‐/Perforin+ parent at both basal formulaAons and culture systems. (B) A method for co-‐staining of intracellular and cell surface markers was developed based on the protocol of Wang et al. 2013 (PMID: 23168618).
Figure 6. Comparison of flow-‐cytometry with real-‐1me imaging for the assessment of NK-‐92 cell cytotoxicity. (A) CTO-‐labeled K562 cells (Target) were incubated with NK-‐92 cells (Effector) at different raAos, and stained with calcein before analysis. (B) Lysis of K562 cells by acAon of NK-‐92’s was quanAfied by the increase of CTO+/Calcein-‐ populaAon. (C) Assay was developed based on the protocol of Thakur et al. 2012 (PMID: 22187077). (D) Target cells were labelled with Calcein and incubated with effector cells in the BiostaAon CT (Nikon), an integrated cell culture observaAon system. (E,F) Decay in fluorescence intensity (F.I.) as result of K562 lysis was analysed with CL-‐Quant sosware (Nikon). %K562 lysis measured using BioStaAon CT (Nikon), was comparable to that obtained by flow-‐cytometry (42%).
NK-‐92 POTENCY
0
4
8
12
Xvivo S9 S10 Glucose
qLac (p
mol/cell.d
ay)
VCDmax cell/mL
Medium and consumables ($/cycle)
Cost reduc1on
~1.5 x106 $6,185 45,7 %
0 0.2 0.4 0.6 0.8 1
1.2
0 1 2 3 4 5
Total F.I.
t (h)
Control E:T(1:10)
t=4h
t=4h
(con
trol)
K562 t=
0h
Real-‐Time Imaging by BioSta1on CT (Nikon) Flow-‐Cytometry
% K562 lysis 66±2.9
% K562 spontaneous lysis
24±11
K562 t=
0h
t=4h
t=4h
(con
trol)
0
20
40
60
80
100
0.1 1 5 10
K562 lysis (%)
E:T
t=4h
Cost break down -‐ $114,000 /treatment
% K562 lysis 44±4.3
% K562 spontaneous lysis
4±0.6
NIKON CORPORATION Instruments Company
A
B
A
B
D
C
A B C
A
95.3%
CD3
CD45/CD56
PERFORIN
CD3-‐/CD45+/CD56+/PERF.+
B
A B
C
D E
F