Study design
The overall objective of this study was to demonstrate that AXL is a suitable target for CAR T cell therapy against NSCLC, alone or in combination strategy. We designed three types of AXL-targeted chimeric antigen receptors that, when transduced into human T cells, provided tumor antigen recognition and antigen-specific effector function. In vitro, we investigated (i) AXL expression in NSCLC cell lines, non-tumor tissues, matched tumor and paracarcinoma tissues via flow cytometry, western blot (WB), or immunohistochemistry (IHC), (ii) cytotoxicity and, (iii) cytokine secretion. All in vitro assays were performed with at least triplicate samples. We identified the best CAR constructs for in vivo studies based on their in vitro killing potency. We mainly analyzed the killing efficacy of YW327.6S2-CAR T cells, alone or in a combination manner. We developed subcutaneous, lung and abdominal metastatic tumor models via NSG immunodeficient mice with human NSCLC cells (n≥5), in order to mimic different clinical metastatic situation. Tumor challenged mice were randomized divided into treatment groups. Subcutaneous and lung metastatic tumor model were treated by CAR T cells via tail vein, while abdominal metastatic were administrated by intraperitoneal or intravenous injection, to optimize delivery methods. We also evaluated the efficacy, safety and synergistic mechanism of YW327.6S2-CAR T cells combined with thermal ablation (MWA) or erlotinib, in order to validate and facilitate the translation value of AXL-CAR T to the clinical application by combination modality. Tumor burden was monitored via calipers or living image. CAR T cell infiltration was detected via flow cytometry. Veterinary staff, independent of the researchers and studies, monitored mice daily and alerted researcher when a humane endpoint had been reached. All animal experiments were carried out according to the applicable guidelines and regulations approved by the Second Affiliated Hospital of Guangzhou Medical University Experimental Animal Care Commission.
- Cells lines
A549, HCC827, and HEK-293T cell lines were obtained from the American Type Culture Collection. Erlotinib-resistant HCC827-ER3 cell line was established in Case Western Reserve University (Cleveland, Ohio, USA) as previously described17. Cells were maintained in a humidified atmosphere containing 5% CO2 at 37 ℃. DMEM (Gibco) or RPMI-1640 (Gibco) with 10% FBS (Gibco), 100 IU/mL of penicillin (Gibco), and 100 IU/mL of streptomycin (Gibco), was used to culture all cancer cell lines. For in vitro cytotoxicity and in vivo living image experiment, A549, HCC827 and HCC827-ER3 cells were lentivirally transduced with pWPXLD-Luc(+)/eGFP virus expressing the GFP and luciferase (GL) genes. All cells were routinely tested for mycoplasma contamination.
- Affinity verification of Anti-AXL scFv
The anti-AXL scFv was expressed and purified according to the instruction as previously described37. Briefly, the anti-AXL scFv was expressed by transient transfection of the plasmids with polyethyleneimine (MW25000, Polyscience) in Free-style 293-F cells (Invitrogen). The expressed scFv proteins were purified by protein A agarose beads (GE Healthcare). Affinity verification was achieved via ELISA.
- Vector design and lentivirus production
To generate CARs targeting AXL, the genes of three anti-AXL scFv18,19,23,24 and anti-CD19 (control scFv), followed by CD8 leader, CD8 hinge, CD28 transmembrane, and composite CD28-CD137-CD3ζ intracellular signaling domains, under the control of an EF-1α promoter, were codon optimized and synthesized by Genscript Co. Ltd. (Nanjing, China), then subcloned into the backbone lentiviral vector pWPXLd-2A-eGFP. The sequence of each cloned CAR was verified via sequencing.
Lentivirus particles were produced in HEK-293T cells after polyethyleneimine (Sigma-Aldrich)-mediated transfection with the constructed vector, together with two auxiliary packaging plasmids, psPAX2 and pMD.2G. Supernatants were harvested at 48 and 72 h post-transfection and filtered through a 0.45-μm filter (Millipore) to remove cell debris.
- Isolation, transduction, and expansion of primary human T lymphocytes
Peripheral blood mononuclear cells (PBMCs) were isolated from the buffy coats of healthy donors using Lymphoprep (Stemcell). T cells were positively selected from PBMCs using a MACS Pan T Cell Isolation Kit (Miltenyi Biotec) and activated using microbeads coated with anti-human CD3/CD28, (Miltenyi Biotec) at a 1:1 bead:cell ratio for 24-72 h in T cell medium (Miltenyi Biotec). Transduction were performed as previously described38. Transduction efficiency was determined after 72 h by the percentage of GFP+ cells via flow cytometry. T cell phenotype was examined before and after transduction. Informed consents from healthy PBMC donors were obtained, and all procedures were approved by the Research Ethics Board of The Second Affiliated Hospital of Guangzhou Medical University.
- Quantitative real-time polymerase chain reaction (qPCR)
The expression of different CAR mRNAs was detected by quantitative real-time polymerase chain reaction (qPCR), as previously described. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene was used as an endogenous control. The primers for different CARs were:
GAPDH - sense: 5’- GCACCGTCAAGGCTGAGAAC-3’
GAPDH - antisense: 5’- TGGTGAAGACGCCAGTGGA -3’
scFv of 3E3E8 - sense: 5’- GCACAGCAACGGCAACACCTA-3’
scFv of 3E3E8 - antisense: 5’- CCGCTAAACCTATCGGGCACT -3’
scFv of YW327.6S2 - sense: 5’- GCAGCATCTGGCTTTTCTC -3’
scFv of YW327.6S2 - antisense: 5’- TTCACGGAGTCGGCATAGT -3’
scFv of 20G7D9 - sense: 5’- TACAATGAGAAGTTCAACGACCG -3’
scFv of 20G7D9 - antisense: 5’- GTCCCCAGTATGCGAACCAG -3’
scFv of CD19 - sense: 5’- ACTACATCCTCCCTGTCTGCC - 3’
scFv of CD19 - antisense: 5’- CCACTGCCACTGAACCTTGA - 3’
- Flow cytometry
All samples were analyzed using a NovoCyteTM (ACEA Biosciences), and data were analyzed using FlowJo software (FlowJo). The antibodies used included anti-AXL-PE (clone DS7HAXL) (Invitrogen), anti-human CD3-FITC (clone UCHT1), anti-human CD4-APC (clone RPA-T4), anti-human CD8a-PE (clone HIT8a). All FACS-related staining procedures were performed on ice for 30 min, and cells were then washed with PBS containing 1% FBS before cytometry analysis. PB and tumor samples from mouse xenografts were treated with red blood cell lysis buffer (Thermo Fisher Scientific), and the cells were stained with the corresponding antibodies.
- In vitro cytotoxicity and cytokine release assay
A549 GL, HCC827 GL, and HCC827-ER3 GL target cells were incubated with Mock, CD19- or three AXL-CAR T cells at the indicated effector to target (E:T) ratios of 8:1, 4:1, 2:1, and 1:1 in triplicate wells of U-bottomed 96-well plates at 37℃. Target cell viability was monitored 24 h later by adding 100 μL/well D-Luciferin (Cayman Chemical) resolved at 150 μg/mL. Then, the viability percentage (%) was calculated as experimental signal/maximal signal × 100, and the specific cell lysis was calculated using 100%-viability percentage.
Target cells (1 × 104) were incubated with effector cells (1 × 104) in U-bottomed 96-well plates for 24 h. The culture supernatants were then collected and analyzed for the secretion of IL-2, TNF-α, IFN-γ, GM-CSF, Granzyme B, and Perforin using enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems) according to the manufacturer’s protocol.
- IHC and Western blot analysis
IHC and Western Blot procedures were performed following the standard protocols. Sample collecting was approved by the Research Ethics Board of The Second Affiliated Hospital of Guangzhou Medical University.
Different levels of AXL expression in 90 paraffin-embedded NSCLC tissue (including 40 erlotinib resistant cases) and 22 kinds of non-tumor paraffin-embedded samples (detailed in Supplementary Table 1), were evaluated by one experienced pathologist using a 4-point scale. Score 0 means no AXL expression; scores of 1+, 2++, and 3+++ mean weak to strong expression of AXL. The percentages of AXL-positive staining with different scores were also recorded.
For Western Blot (WB), membranes were probed with anti-human AXL (Cell Signaling Technology), anti-human CD3 (Abcam) primary antibodies.
- In Vivo Antitumor Studies
Six- to eight-week-old female NSG (NOD-PrkdcscidIL2rgtm1/Bcgen, Biocytogen, Beijing, China) mice were housed and treated under specific pathogen-free conditions and were provided autoclaved food and water at the Experimental Animal Center of The Second Affiliated Hospital of Guangzhou Medical University (Guangzhou, China). All animal experiments were carried out according to the applicable guidelines and regulations approved by The Second Affiliated Hospital of Guangzhou Medical University Experimental Animal Care Commission. For the following in vivo studies, tumor volume was measured every three days with a caliper and calculated with the following equation: tumor volume=(length×width2)/2, or monitored via bioluminescence imaging (BLI) for tumor progression. To improve the CAR T cell percentage of primary T cells, transduction by lentivirus was performed twice or thrice before in vivo assay.
For the cell line-based/derived NSCLC subcutaneous xenograft models, 2×106 A549 or HCC827-ER3 cells in 100 μL PBS were subcutaneously injected into the right flanks of NSG mice on day 0. When tumor nodes reached about 50 mm3, the mice were divided into three groups: Mock, CD19-CAR T and AXL-CAR T (n=6), and received PBS or 1×107 CAR T (CD19- or YW327.6S2-CAR T) cells intravenously. On day 39 (A549) and 45 (HCC827-ER3) after tumor inoculation, all mice were sacrificed. Tumor weight was measured.
In order to construct pulmonary and intraperitoneal metastasis models, 1×106 A549 GL cells in 100 μL PBS were injected into NSG mice intravenously (i.v.) or intraperitoneally (i.p.), respectively, on day 0. Two weeks after tumor cells injection, the mice were subjected to BLI. For lung metastasis, mice were randomly divided into (n=5): Mock, CD19-CAR T and AXL-CAR T. Mice were administrated with PBS (Mock), or 1×107 effector cells (CD19- or YW327.6S2-CAR T) suspended in 100 μL PBS intravenously injected on day 14. For intraperitoneal metastasis models, mice were randomly assigned into (n=5): Mock, YW327.6S2-CAR T (i.v.), and YW327.6S2-CAR T (i.p.). Mice were administrated with PBS (Mock), or 1×107 YW327.6S2-CAR T cells (i.v. or i.p.) suspended in 100 μL PBS (i.p.) on day 14. Mice were monitored with BLI frequently for disease progression. T cell percentages in peripheral blood was detected by flow cytometry.
To assess synergistic antitumor efficacy of combination therapy, subcutaneous tumor model by inoculating HCC827-ER3 or HCC827-ER3 GL cells was established (day 0). For YW327.6S2-CAR T cells combined with MWA modality (Vision-China Medical Devices R&D center), experiment was initiated when mean tumor volume reached about 200 mm3, mice were then randomly allocated to four groups (n=5): Mock (PBS), MWA (10w, 45s, on day 30), YW327.6S2-CAR T (1×107 YW327.6S2-CAR T cells, on day 30, i.t.), and combination groups (ablation, on day 30, combined with 5×106 YW327.6S2-CAR T cells, on day 33, i.t.). To reach partial ablation, a time-power gradient experiment was performed, in order to faithfully mimic the tumor residue or marginal recurrence after ablation. Tumor volume and weight was measured. Blood was collected for biochemistry analysis. T cell infiltration in peripheral blood and tumor was tested by flow cytometry. Ki-67 expression in tumor tissue was detected via IHC. At the same time, all important mouse organs, including heart, liver, spleen, lung, kidney, cerebrum, stomach, small intestine and colon, were harvested, fixed with 4% paraformaldehyde and stained with hematoxylin-eosin (H&E) or corresponding antibodies (IHC). For YW327.6S2-CAR T cells combined with erlotinib tactics, experiment was carried out when mean tumor volume reached about 100 mm3, mice were randomly divided into four groups (n=5): Mock, erlotinib, YW327.6S2-CAR T (1×107 YW327.6S2-CAR T cells, i.v.), and combination groups (erlotinib combined with 5×106 YW327.6S2-CAR T cells, i.v.). Erlotinib was administrated by gavage (100mg/kg/day) on day 21 to 42, YW327.6S2-CAR T cells were injected intravenously on day 21. AXL expression were detected in dissected tumor after euthanasia.
- Bioluminescence imaging
Isoflurane-anesthetized animals were imaged using a cooled CCD camera system (IVIS 100 Series Imaging System, Xenogen), followed by the intraperitoneal injection of 75 mg/kg D-luciferin (Cayman Chemical). Quantification of total and average emissions were quantified using Living Image software (Xenogen).
- Statistics
Data were presented as mean ± SD or SEM. Differences between groups were analyzed by one-way ANOVA with Bonferroni post-tests. Gray-scale analysis of WB was achieved by ImageJ. The Kruskal-Wallis test was utilized to compare the non-normally distributed endpoints. For survival data, Kaplan-Meier curves were plotted and compared using a log-rank test. GraphPad Prism 7.0 was used for the statistical calculations. *p < 0.05, **p < 0.01, and ***p < 0.001 were considered statistically significant.