Engineered immune cells using gene editing to target solid tumors

As AAVS1 is a safe genomic location for integrating a gene, we designed guide RNAs (gRNAs) targeting AAVS1 and generated a CRISPR/Cas9 vector. Following transfection into 293T cells, the specific gene knockout was evidenced by PCR and subsequent T7 endonuclease 1 assay (Supplementary Fig.). 2a,b

PBMCs were isolated and after culture for a short period, T cells were enriched and activated for 48 h. The activated T cells were electroporated with pX330-sgRNA targeting the AAVS1 locus and donor-AAVS1 vector. PCR was performed to detect the inserted homologous recombination genes (Mock, CD19 CAR, or CD105 CAR) Identification of the Mock, CD19 CAR, and CD105 CAR fragments are ~740, 1460, and 1100 bp, respectively (Supplementary Fig. 2c). The results were consistent with the expected target band size, which proved that CAR gene fragment had been successfully integrated into the the AAVS1 locus. Fluorescent microscopy indicated GFP signals in different groups of activated T cells at 24 h post electroporation (Fig. 2c). Flow cytometry was used to detect the expression of the CAR receptor with 41.3% (Fig. 2d) and the ratios of CD4⁺ to CD8⁺ T cells among the different groups of CAR-T cells (Supplementary Fig. 2d).

The levels of CD105 expression in HCC cells (Bel7404, HepG2, SMMC7721, and MHCC97H), human umbilical vein endothelial cell (HUVEC), and 293T cells were characterized by flow cytometry. CD105 was highly expressed on the surface of Bel7404, HepG2, SMMC7721, and HUVEC cells, but little on MHCC97H cells and was not expressed on 293T cells (Supplementary Fig.). 3

The different groups of CAR-T cells were stimulated with the same number of irradiated Bel7404, HepG2, and MHCC97H for 16 h. The levels of interferon-γ (IFN-γ), tumor necrosis factor-α (TNF-α), interleukin (IL)-2, macrophage inflammatory protein-1α (MIP-1α), and IL-10 in the supernatants of individual groups of cells were measured by enzyme-linked immunosorbent assay (ELISA) (Fig. 4b, c). The levels of IFN-γ, TNF-α, MIP-1α, and IL-2 in the supernatants of cultured anti-CD105 CAR-T cells were significantly higher than that of the Utd, Mock, and CD19 CAR groups of cells following challenged with Bel7404 and HepG2 cells expressed high levels of CD105. There was no significant difference of levels of above cytokines, except for IFN-γ produced by anti-CD105 CAR-T cells cultured with CD105-low expressing MHCC97H cells. However, there was no significant difference in levels of IL-10 among the different groups of cells. Enzyme-linked immune absorbent spot (ELISPOT) revealed that the numbers of IFN-γ-secreting spot-forming cells in the anti-CD105 CAR-T cells were significantly greater than that of the Utd, Mock, and CD19 CAR groups after challenged with CD105⁺ cells (Fig. 4d, e). These independent lines of data demonstrated that anti-CD105 CAR-T cells have activated T cells and secreted higher levels of pro-inflammatory cytokines after challenged with CD105⁺ cells.

To test whether anti-CD105 CAR-T cells could specifically kill CD105⁺ target cells, the Utd, Mock, anti-CD19 CAR-T cells, and anti-CD105 CAR-T cells were challenged with PKH26-labeled HCC cells (Bel7404, SMMC7721, HepG2, or MHCC97H), HUVEC, and 293T cells at E/T ratios of 1 : 3, 1 : 1, and 3 : 1 for 16 h. The PKH26-labeled cells were stained with propidium iodide (PI) and the percentages of PKH26⁺PI⁺ dead cells were analyzed by flow cytometry. We found that the anti-CD105 CAR-T cells had potent cytotoxicity against HCC cells (Bel7404, HepG2, and SMMC7721) that expressed high levels of CD105 and CD105-positive HUVEC cells, but little cytotoxicity against CD105-low expressing MHCC97H cells, whereas the Utd, Mock, and anti-CD19 CAR-T cells had low or little cytotoxicity against such HCC cells. All groups showed no cytotoxicity against CD105-negative 293T cells (Fig. 4f). The specific cytotoxicity of anti-CD105 CAR-T cells against CD105⁺ HCC cells was mitigated by pretreatment with with CD105 protein; the results show that the killing effect of CD105 CAR-T on target cells is reduced after incubated with CD105 protein (Fig. 4g). Hence, the anti-CD105 CAR-T cells effectively and specifically killed CD105⁺ cells in vitro.

We compared lentivirus encoded CD105 CAR-T cells (LV-CD105 CAR-T) with T cells inserted with CD105 CAR at AAVS1 locus (AAVS1-CD105 CAR-T). We first detected the expression of exhaustion markers in CAR-T cells culturing with the presence of antigen. The expression of exhaustion markers in LV-CD105 CAR-T cells was higher than that in AAVS1-CD105 CAR-T cells, indicating that LV-CD105 CAR-T cells had shown signs of depletion (Supplementary Fig. 5a). LV-CD105 CAR and AAVS1-CD105 CAR-T cells differed markedly in their antitumor activity; AAVS1-CD105 CAR-T cells inhibited tumor growth and induced greater responses (Supplementary Fig. 5b). Immunohistochemistry exhibited that the numbers of human anti-CD3-stained T cells in the tumors from the mice receiving AAVS1-CD105 CAR-T cells were significantly higher than those in LV-CD105 CAR-T cells and Utd groups. The result revealed that human CD3⁺ T cells had accumulated in residual tumors after intravenous AAVS1-CD105 CAR-T administration, but less accumulation in LV-CD105 CAR (Supplementary Fig. 5c)

Human HCC cells HepG2, HUVEC, and non-tumor embryonic kidney (293T) cells were obtained from the American Type Culture Collection. HCC cell lines Bel7404, SMMC7721, and MHCC97H cell were obtained from the Cell Bank of Chinese Academy of Sciences (Shanghai, China). The cells were cultured in Dulbecco’s modified Eagle’s medium medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 μg/ml streptomycin (complete medium, Invitrogen) at 37 °C in 5% CO₂.

A healthy young camel was injected subcutaneously with extracellular domain of CD105 protein (1 mg in PBS) mixed with an equal volume of Freund’ s complete adjuvant on 7-day interval for 6 weeks. Six weeks later, PBMCs were collected, the total RNA was extracted, and cDNA was synthetized with a RevertAid First Strand cDNA Synthesis Kit. The VHHs were amplified by nested PCR.⁴⁹ The pMECS phagemid vector and the nested PCR product were digested with NotI and PstI restriction enzymes, and then ligated with T4 DNA ligase enzyme at 16 °C overnight. The ligated vector was electroporated into electrocompetent Escherichia coli TG1 cells. The electroporated cells were cultured. The VHH libraries were displayed on phages after infection with VCSM13 helper phages. The Nbs against CD105 were screened by phage display; after three rounds of screening, the positive clones were identified by the periplasmic extract ELISA, the plasmids from positive colonies were extracted, and sent for sequencing.⁵⁰ The positive clones were classified into different families, based on the diversity of their amino acid sequences in the CDR3 region. The recombinant pMECS plasmids of the different Nb families were extracted from TG1 cells and transformed into E. coli WK6 cells by electroporation. The expression of Nbs was induced by isopropyl β-d-1-thiogalactopyranoside. The specific Nbs containing His-tags were purified from the cell lysate by immobilized metal affinity chromatography using a His-select column and analyzed by 15% SDS-PAGE.

The affinities of selected anti-CD105 Nbs were measured by the kinetic analysis by surface plasmon resonance experiment with PlexArray® HT system.⁵¹ First, CD105 Nbs (10 μg mL⁻¹) were immobilized on PlexArray® Nanocapture® Sensor Chips. Then, to take note of the signals, different concentrations were injected with a flow rate of 3.0 μL s⁻¹. Finally, a regeneration step was performed after the affinity binding reaction. Briefly, the sensor chips were regenerated with 0.05 mol L⁻¹ NaOH for 2 min and then PBS buffer was injected over the sensor surface for about 5 min. All signals were recorded as sensorgrams and steps were performed in triplicate at 25 °C.

Three 20 bp gRNA (Supplementary Table 1) were designed for the PPP1R12C gene at the AAVS1 locus on chromosome 19. These gRNA oligonucleotides were synthesized by Shanghai Biotech (Shanghai, China). After annealing the gRNA, it was ligated to the plasmid pX330. After being sequenced, the generated pX330-sgRNAs were transfected into 293T cells and their genomic DNA was extracted for PCR amplification of the DNA fragment covering the knockout site. The specific gene-editing efficiency was analyzed by T7 endonuclease 1 assay.

DNA sequence of the desired CAR was cloned into KpnI and HindIII sites of the plasmid Puc-Mdonor-AAVS1 (Genomeditech, Shangai, China) containing the homologous recombinant arms to construct homology-directed repair (HDR) vectors. As shown in Fig. 2a, the CD105 CAR was composed of the CD105 Nb, human CD8a hinge, and transmembrane domain (nucleotides 412–609), the intracellular signaling domain of CD137 (4-1BB, nucleotides 640–765), and CD3ζ molecules (nucleotides 154–492). Individual CAR sequences were linked to GFP through IRES.

Peripheral blood samples were collected from healthy human donors and PBMCs were prepared by Ficolle-Paque density gradient centrifugation. PBMCs were cultured in RPMI-1640 medium (Gibco) containing 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin (complete medium). After 2 h culture, the suspended lymphocytes were collected and cultured in complete RPMI-1640 medium containing 100 U/ml of recombinant human IL-2 (Peprotech, USA). The cells were stimulated for 24 h using the Easy-T Stimulation Kit (Jikai, Shanghai), according to the manufacturer’s instructions. Two days after stimulation, lymphocytes (1 × 10⁶ cells/tube) were transfected with 2 µg pX330-sgRNA and 4 µg individual types of Puc-Mdonor-AAVS1 vector containing Mock, CD19 CAR, or CD105 CAR by electroporation in the Lonza electroporation mixing buffer using a Lonza 4D Necleofector^TM electroporation system with preset program. Subsequently, the cells were cultured in pre-warm complete RPMI-1640 medium for 4 h and continually cultured in the medium containing 500 U/ml of IL-2 for 48 h. The fluorescent signals were observed under a fluorescent microscope (Nikon, Japan).

For identification of CAR genes in T cell, total DNA was extracted from anti-CD105 CAR-T (CD105 CAR), anti-CD19 CAR-T (CD19 CAR), or Mock-T cells, and the CAR genes inserted into the AAVS1 locus were identified by PCR using specific primers. CAR fragment primers with the sequences were CMV-F: 5′-CGCAAATGGGCGGTAGGCGTG-3′ (located upstream of the CAR gene insertion site) and AAVS1-R: 5′-ATGGGGCTTTTCTGTCACCA-3′ (located downstream of the CAR gene insertion site). The PCR reactions were performed in duplicate at 94 °C for 2 min and subjected to 35 cycles of 94 °C for 30 s, 58 °C for 30 s, and 72 °C for 30 s, followed by a final extension at 72°C for 2 min. The PCR products were visualized by 1% agarose gel electrophoresis.

Lentivirus encoded CD105 CAR-T cells (LV-CD105 CAR-T) were prepared as follows: recombinant lentiviral particles were produced by a calcium phosphate transfection system. Lentiviral particles were concentrated 30-fold by followed by ultracentrifugation at 28,000 r.p.m. 16 °C for 120 min. Activated T cells were then transduced with the lentiviral vector at a multiplicity of infection of 10 U/cell. The transduced T cells were cultured at a concentration of 5 × 10⁵ cells/mL in the presence of recombinant human IL-2 (100 U/mL) every other day.

Four days after electroporation, T cells incubated with Phycoerythrin (PE)-conjugated anti-His-tag antibody, flow cytometry was used to detect the expression of the CAR receptor. Furthermore, individual groups of CAR-T cells and control Utd T cells were stained with PE-anti-CD25 (BD, Cat555431), PE-anti-CD69 (BD, Cat555531), or PE-anti-CD62L antibody (BD, Cat555543). Some cells from each group were stained with PE-anti-CD4 (BD, Cat555347) and percp-cy5.5-anti-CD8 antibody (BD, Cat565310). The percentages of CD4⁺ and CD8⁺ T cells in individual groups of CAR-T cells were determined by flow cytometry. In addition, human HCC cells (Bel7404, HepG2, SMMC7721, and MHCC97H), HUVEC, and 293T cells were stained with PE-anti-CD105 antibody (eBioscience, Cat12-1057-42). The percentages of CD105⁺ target cells was examined by flow cytometry. The data were analyzed using FlowJo software (7.6).

Proliferation of T cells was determined by PKH26 staining in vitro. Individual groups of CAR-T and control Utd T cells (1 × 10⁶ cells/tube) were labeled with PKH26 (Sigma-Aldrich) at 37 °C for 5 min. After being washed, individual groups of cells were stimulated with the same number of CD105⁺ target cells that had been irradiated (100 Gy) for 120 h. The suspended T cells were collected and the percentages of proliferative T cells were determined by flow cytometry.

For cytotoxicity assays, individual groups of CAR-T and control Utd T cells were cultured in triplicate with each type of PKH26-labeled target cells at a ratio of 3 : 1, 1 : 1, or 1 : 3 for 16 h. After being washed, the adhered target cells were collected and stained with PI stain (Sigma), and the percentages of PKH26⁺PI⁺ dead cells in individual groups were determined by flow cytometry as % of specific lysis. In addition, anti-CD105 CAR-T cells were co-cultured with Bel7404 cells at a ratio of 3 : 1 for 16 h in the presence or absence of CD105 protein (eBioscience, Cat14-1051-85). The percentages of lysed target cells were determined by flow cytometry. Moreover, individual groups of CAR-T and control Utd T cells were cultured in triplicate with Bel7404 cells at a ratio of 3 : 1 for 12 h and stained with PE-anti-CD107a antibody (Biolegend,Cat304124). The percentages of CD107a⁺ T cells were determined by flow cytometry.

Individual groups of CAR-T and control Utd T cells (1 × 10⁵ cells/well) were stimulated with 1 × 10⁵ cells/well of Bel7404, HepG2, or MHCC97H cells for 16 h. The levels of IFN-γ, TNF-α, IL-2, MIP-1α, and IL-10 in the supernatants of cultured cells were determined by ELISA kits (BD), according to the manufacturer’s protocols.

To test the density of IFN-γ-secreting cells, ELISPOT assay was performed. Briefly, individual groups of CAR-T and control Utd cells (3 × 10⁵ cells/well) were stimulated in triplicate with 10⁵ irradiated Bel7404 cells at 37 °C overnight in 96-well plates that had been coated with anti-IFN-γ. After being washed, the captured IFN-γ in individual wells was reacted with biotinylated anti-IFN-γ at 4 °C overnight. Subsequently, the IFN-γ-specific immunocomplex was detected with streptavidin-AP and visualized using a substrate solution 5-Bromo-4-Chloro-3-Indolyl Phosphate (BCIP)/Tetranitroblue tetrazolium chloride (NBT). The numbers of spot-forming cells in individual wells were analyzed on a CTL ImmunoSpot S6 Ultimate-V analyzer using Immunospot software, version 5.1.

Female NOD/SCID mice were purchased from Vitalriver (Beijing, China) and housed in a specific pathogen-free facility with free access to autoclaved food and water. The mice were randomized and injected subcutaneously with 5 × 10⁵ Bel7404 cells in their right flank to induce a solid tumor. When individual mice developed a tumor with a volume of ~100 mm³, the mice were randomized and treated intravenously with PBS, 2 × 10⁶ Utd T, Mock CAR-T, anti-CD19 CAR-T, and anti-CD105 CAR-T twice. The dynamic growth of implanted tumors was monitored every 3 days using a caliper through measurement of the length (L) and width (W), and the tumor volumes were calculated by the formula of (L × W²)/2. Post treatment, the mice were killed and their tumors were dissected, photo-imaged, and weighed. Another set of mice was continually monitored until death and the survival of individual groups of mice was calculated. Similarly, tumor-bearing mice were treated with 2 × 10⁶ LV-CD105 CAR-T cells injected into the tail vein. The experimental protocols were approved by the Animal Research and Care committee of Guangxi Medical University, Nanning, China.

To generate PDX mouse models of HCC, surgically removed specimens were obtained from HCC patients with informed consent and institutional approval. The human cells and tissue relevant studies were approved by the Ethics Committee of Guangxi Medical University and the First Afliated Hospital of Guangxi Medical University. The collected primary HCC tumor tissues were diced into thin pieces of 3–4 mm and then subcutaneously inoculated into the right flank of 6-week-old male NOD/SCID mice. For continuous transplantation, tumor-bearing animals were anesthetized with ether and killed by cervical dislocation. Tumors were shredded under sterile conditions and serially transplanted into NOD/SCID mice as described above. When tumor volume of individual mice expandedto ~80 ~ 100 mm³, the mice were randomized and treated intravenously with 2 × 10⁶ of CD105 CAR-T cells or control CD19 CAR-T cells twice. CD105 CAR-T was incubated with CD105 protein for 30 min and then were injected into the tumor-bearing mice through the tail vein. The size of the tumor was measured every 3 days, and the temperature and body weight of mice were also monitored.

The tumor tissues from individual mice were dissected, fixed in 10% neutral formalin, and paraffin-embedded. Furthermore, the tumor tissue sections (4 µm) were stained with anti-Ki67 antibody (kit-0005), anti-CD105 antibody (MAB-0344), anti-Bax antibody (MAB-0254), or anti-CD3 antibody ((kit-0005) from Maixin Biotech. The bound antibodies were reacted with biotinylated second antibodies and detected using the Streptavidin-Peroxidase kit and 3,3′-Diaminobenzidine (Maixin Biotech). In addition, the frequency of apoptotic cells in tumor tissue sections was determined by TUNEL assay using the In Situ Cell Death Detection Kit (fluorescein isothiocyanate (FITC)), Roche, Switzerland), according to the manufacturer’s instruction. Images were obtained under a microscope (Nikon, Japan).

The flow cytometry data was analyzed with FlowJo 7.6 software. Statistical analysis was performed by using GraphPad Prism software 6.0. One-way analysis of variance (ANOVA) with Tukey’s multiple comparison was performed to assess the differences among different groups in the in vitro assays. Tumor growth curve was analyzed using two-way ANOVA with correction for Tukey’s multiple comparison. The survival curve of animals in the difference between groups was calculated by the Kaplan–Meier analysis (log-rank test) log-rank test. A P-value < 0.05 was considered statistically significant.