Background
Osteosarcoma (OS) is the primary malignant bone tumor affecting children and adolescents. It is inclined to occur in the metaphysis of long bones, including distal femur and proximal tibia [
1]. The treatment strategy for OS has advanced from amputation to the current neoadjuvant chemotherapy, surgery and the subsequent adjuvant chemotherapy. Despite the introduction of various chemotherapy regimens, the overall survival rate for advanced or recurrent OS patients still remains low [
2,
3]. Therefore, novel treatment regimens with improved therapeutic benefits need to be investigated.
Cellular immunotherapy is the emerging strategy that is of great interest in oncology. Adoptive cell transfer (ACT), in particular, is a promising regimen attributed to the recent success of CD19-chimeric antigen receptor (CAR)-T cells against acute lymphoblastic leukemia [
4‐
6]. Specifically, the genetic addition of CARs enables T cells to target tumor cells in a major histocompatibility complex (MHC)-unrestricted manner [
7]. Despite its consistently remarkable antitumor activity against hematological malignancies, CAR-T therapy in the treatment of solid tumors remains challenging mainly due to the lack of proper tumor associated antigen (TAA) [
8].
Activated leukocyte cell adhesion molecule (ALCAM, CD166) is a 105 kDa trans-membrane glycoprotein that belongs to the immunoglobulin superfamily. The binding of ALCAM specifically to CD6 mediates the interaction between adjacent cells [
9]. ALCAM is considered to function in a various biological activities including neuronal outgrowth, hematopoiesis and inflammatory responses [
10]. Previous studies have demonstrated its association with tumorigenesis of many malignancies, including breast cancer, prostate cancer, melanoma and OS [
11‐
14]. ALCAM can also function as the therapeutic target using anti-ALCAM monoclonal antibody, conjugated to the nanoparticles to eliminate prostate cancer and OS cells [
11,
15]. All these evidence qualifies ALCAM as the promising candidate to target OS in the adoptive cell immunotherapy.
In the present work, we found that ALCAM was expressed in four different human OS cell lines at levels ranging from 36.9 to 96.7%. The generated CD166 CAR-T cells incorporated with 4-1BB demonstrated their cytotoxic activity against OS in vitro and in vivo, the cytotoxicity of which correlated well with the expression levels of CD166. Our results support future investigations of CD166 CAR-T cells and rational combinations with other immunotherapy for the treatment of OS and other CD166 positive malignancies.
Methods
Cell lines and cell culture
The human osteosarcoma lines MNNG/HOS, U2OS, MG-63 and Saos-2, normal human osteoblasts hFOB 1.19, normal human fetal lung fibroblasts HFL1 were obtained from Cell Bank of Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). Normal human hepatocytes HL-7702 were obtained from Procell (Wuhan, China). Their identity was verified by short tandem repeat analysis. MG-63, MNNG/HOS, Saos2 cells were cultured in Dulbecco’s modified Eagle’s medium (Gibco, Rockville, MD, USA), U2OS cells and HL-7702 cells were cultured in RPMI 1640 medium (Gibco), HFL1 cells were cultured in F-12 K medium (ATCC), hFOB 1.19 cells were cultured in DMEM/F-12 medium (Gibco). All the medium was supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA) and 100 μg/mL streptomycin–penicillin. All the cell lines were maintained at 37 °C in a humidified incubator containing 5% CO2. The cells were always passaged when they approximately reached 80% confluence.
Construction of the anti-CD166 CAR
The chimeric CD166/CAR is composed of CD166 scFv and a 4-1BB-CD3ζ expression cassette that was designed and synthesized by the GeneChem Biotechnology Company (Shanghai, China), as is shown in Fig.
2a. The CD166 scFv was derived from a high-affinity monoclonal antibody. The 4-1BB-CD3ζ expression cassette contains the hinge and transmembrane (TM) region of CD8α. CD166 scFv and 4-1BB-CD3ζ were connected in-frame by overlap PCR. The generated CD166/CAR was verified by DNA sequencing and cloned into the BamHI sites of a lentiviral vector (Genechem Biotechnology, China); the resultant product was named CD166.BBζ CAR. The intracellular domain of the CARs has the self-cleaving 2A peptide connected to a GFP green fluorescent label. The sequences of all PCR primers are available upon request.
Transduction of lentiviral CD166/CAR
Peripheral blood mononuclear cells (PBMCs) were isolated using the Ficoll density gradient centrifugation methods from whole blood of healthy volunteer donors. T cells were transfected with an Easy-T kit (GeneChem Biotechnology, China). Briefly, isolated T cells/PBMCs were activated on a plate pre-coated with S buffer (EASY-T cell infection activation kit, catalog no. LCR6018, GeneChem) at a concentration of 0.7 × 106 cells/mL in complete RPMI 1640 medium (Gibco) supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA), 50 IU/mL IL-15 and 200 IU/mL IL-2 (PeproTech). Two days later, the stimulated T cells were resuspended at 0.5 × 106 cells/mL with the Trans B buffer (EASY-T cell infection activation kit, catalog no. LCR6018, GeneChem). CAR-encoding lentivirus (CD166.BBζ CAR) was thawed and added into the cells resuspension solution (virus titer: 5 × 108 TU/mL, MOI = 5). The cells were seeded onto plates that had been coated for 16 h with Trans A buffer (EASY-T cell infection activation kit, catalog no. LCR6018, GeneChem). After 24 h transduction, equal volume of fresh medium supplemented with IL-2 (300 IU/mL) and IL-15 (100 IU/mL) was added into the plates. T cells were then fed with fresh media every 2 days and were used within 28 days of expansion in all experiments.
Cytotoxicity assays
The cytotoxic activity of the CD166.BBζ CAR and non-transduced T cells (NTD) was evaluated using the CytoTox 96® Non-Radioactive Cytotoxicity Assay (Promega). Lactate dehydrogenase (LDH) release was evaluated after 4 h in the supernatant with effector-to-target (E:T) ratios of 20:1, 10:1 and 1:1.
Cytokine release detection
CD166.BBζ CAR and non-transduced T cells were plated at 1 × 106 cells per well on a 96-well plate at a 1:1 ratio with Saos-2 and U2OS cells. Interleukin-2 (IL-2), interleukin-4 (IL-4), interleukin-6 (IL-6), interleukin-10 (IL-10), tumor necrosis factor (TNF-α), and interferon-γ (IFN-γ) cytokine release after 24 h of culture was measured using the cytometric bead array (CBA) human Th1/Th2 cytokine kit (BD Bioscience).
Flow cytometry
FITC-, PE-, perCP/cy5.5-, APC-, AF700- and PE/Cy7- conjugated anti-CD3, CD4, CD8, CD56, CD44, CD62L, CD86, CD197, CD25, PD-1, CD45RO monoclonal antibodies were used to stain lymphocytes (all from Biolegend), whereas the anti-CD166 mAb was used to label osteosarcoma cells.
Immunochemistry
For the detection of T cells infiltration into the tumors, animals were sacrificed and tumors were harvested embedded in OCT medium, snap frozen or formalin fixed and paraffin-embedded. The specimens were cut into serial sections of 6 μm. To show the intratumoral T cells, staining of anti-CD3 (Abcam) was performed on consecutive tissue sections. Images were obtained using confocal microscope (Nikon A1, Japan) or optical microscopes, respectively.
The in vivo antitumor activity of CD166/CAR-T cells in an orthotopic osteosarcoma model
7-week-old NOD/SCID (non-obese diabetic and severe combined immunodeficiency) mice (6–8 weeks old; 18–22 g) were obtained from Experimental Animal Center of the Zhejiang Chinese Medical University and maintained under SPF grade conditions and supplied with sterilized food and water. The use of all mice in this study were approved by the Animal Care and Use Committee of Zhejiang University, China. Saos-2 cells were transfected with luciferase (Saos2-fLuc) for in vivo imaging. The orthotopic osteosarcoma model was established according to the previous studies [
16]. Briefly, the mouse was anesthetized by isoflurane, and a 30G needle was inserted to the proximal tibia through the cortex of the anterior tuberosity. Then Saos2-fLuc cells (5 × 10
6 in 25 μL PBS) were injected slowly into the medullary cavity using Hamilton Syringe fitted with a 26G needle. Mice were randomly separated into three groups (five mice in each group). After 7 days, tumors in different groups reached the same volume, which was confirmed by In Vivo Imaging System (IVIS) (Lumina Series III, Caliper life sciences), treatments of each group were then set as follows: (1) untreated mice, receiving 100 μL PBS, (2) non-transduced-T cells (1 × 10
7) in 100 μL PBS, (3) CD166.BBζ CAR-T cells (1 × 10
7) in 100 μL PBS. Tumor progression was confirmed by the measurement of bioluminescence intensity using the IVIS for up to 28 days.
Statistical analysis
All the data were analyzed using the SPSS software (version 16.0, SPSS, Chicago, IL, USA) and presented as mean ± SD. The statistical differences were detected by Student’s t-test, one-way analysis of variance (ANOVA) with Dunnett’s test or two-way ANOVA analysis. p < 0.05 was considered to be statistically significant.
Discussion
OS is an aggressive malignancy of bone characterized by surrounding calcified osteoid extracellular matrix and frequent lung metastases [
17]. The prognosis of OS patients has achieved little improvement since the advent of chemotherapy. The 5-year overall survival remains dismal and stagnant for the last five decades [
18]. Hence, there is an urgent need for the development of new therapeutic regimens. Several immunotherapies have been carried out in clinical trials against OS, including interferon α2b and muramyl tripeptide [
19,
20]. However, these trials were plagued with different obstacles. ACT is another alternative strategy for the treatment of OS. Earlier efforts have been put on ACT for cytotoxic T lymphocytes and γδ T lymphocytes [
21,
22], while recent studies focused mainly on genetic engineering of T lymphocytes with new antitumor specificities, including TCR-T Cells and CAR-T cells [
23,
24]. Despite its favorable outcomes in treating melanoma and metastatic synovial cell sarcoma [
24], the TCR-engineered T cell therapy still confronts many challenges, including low MHC complex binding affinity and decreased TCRs expression. In contrast, the single-chain variable fragment from the CAR-T cells enables them to bind and recognize targeting antigens in an MHC-independent way, thus overcoming barriers such as HLA downmodulation-related tumor escape and low epitope density-related T cell inactivation [
25]. Due to its great advantages over traditional immunotherapies, CAR-T therapy has now been widely explored and adopted [
26,
27].
Appropriate TAA selection is quite essential for the successful CAR-T therapy. Our results indicate that genetically modified T cells transduced to recognize CD166 may have therapeutic potential against orthotopic OS. Firstly, we demonstrated that CD166 was expressed by the OS cell lines with varying levels. CD166 has previously been identified in primary OS biopsy specimens with high frequency of expression [
11]. Due to its vague role in the correlation between expression level and overall survival [
13,
28], CD166 might have its limitation to serve as the prognostic marker in OS. Instead, it has the great potential to be adopted as the targeting molecule against OS. What’s more, the therapeutic potential of targeting CD166 has been shown by exploiting polymerized liposomal nanoparticles conjugated with corresponding antibody [
11]. All these evidence taken together with our findings favor the idea of CD166 to serve as the promising targeting molecule for CAR-T therapy against OS.
It has been noted that intracellular signaling of T cells is essential for the activation of effector function and persistence of T cells. Meanwhile, just as most malignancies, the paucity of costimulatory molecules expression on the OS cells would make it insufficient for the complete activation of T cells while binding TAA via chimeric receptor. Previous studies have confirmed the advantages of 4-1BB (CD137) costimulation over CD28 for the improvement of T cell proliferation and reduction of exhaustion markers expression [
29,
30]. Therefore, dual endodomains (4-1BB and CD3ζ) were introduced to form the second generation CAR. In the current work, we characterized the effects of CD166.BBζ CAR-T cells in the immunodeficient mice models of OS. Our findings showed that CD166.BBζ CAR-T cells were fully activated upon engagement with CD166-positive OS cells as demonstrated by their cytotoxicity and release of T helper type 1 (Th1) cytokines, which is crucial for the recruitment and maturation of antigen-presenting cells and improvement of cytotoxic T cell responses [
31,
32]. Interestingly, despite 4-1BB’s contribution to the great levels of Th1 cytokines [
33], Th2 distortive responses were also observed. Due to the possible suppressive immune responses from Th2-biased cytokine production, it is necessary to evaluate and improve the CAR construct before further application.
A major concern in immunotherapy research is the potential “on-target, off-tumor” toxicity [
34]. CD166 expression is closely correlated with a wide variety of human cancers, including melanoma, head and neck squamous cell carcinoma, rectal cancer [
35]. Meanwhile, it is also expressed on the surface of epithelial cells, fibroblasts and neurons [
36‐
38]. It is worth noting that CD166-mediated interaction of antigen presenting cells with CD6 on T cells is considered to assign a role in T cell activation [
39]. In the present study, the CD166-specific CAR-T therapy was safe and had no obvious off-tumor effect on mice as proved by histological examination. Nevertheless, we observed that CD166.BBζ CAR-T cells exhibited a low level of cytotoxicity against HFL1 (normal fetal lung) cells and hFOB 1.19 (normal osteoblasts) cells in vitro. For HFL1 cells, since placenta is believed to have immunosuppressive role, the cytotoxicity against normal fetal lung that we have observed might not be extrapolated to their behaviors in vivo [
40]. As for hFOB 1.19 cells, since they were tranfected with vector pUCSVtsA58 and pSV2-neo, they may not be considered as an entirely “normal” cells, which could explain the cytotoxicity observed. All these data suggest that CD166.BBζ CAR-T cells therapy might be safe in humans. However, safety issues should be further evaluated in the immunocompetent mice models as human CD166.BBζ CAR-T cells were administered in NOD/SCID mice bearing human CD166
+ OS in our studies. In addition, more strategies should be developed in the future work to reduce possible adverse events brought by CD166.BBζ CAR-T cells, including the engineering of multi-specific CAR-T cells and the optimization of therapeutic dose. Since there are other potential targets (such as Her-2) proposed for therapy against OS [
23,
41], construction of bi-specific T cells with CD166 and Her-2 CARs may prove beneficial and limit their activation to tumor sites.
In the current work, OS progression in our mice models were partially inhibited following CD166.BBζ CAR-T cells administration. Herein, the efficacy to kill tumor cell lines in xenografts were modest compared with other CAR-T regimens [
42,
43]. Therefore, it is necessary to exploit more regimens to improve anticancer efficacy in solid tumors, among which the combination of CAR-T therapy and other immunotherapies is quite promising. Emerging efforts seek to utilize success with immune checkpoint inhibitors in other malignancies by extending these strategies to OS. Despite the partial success in murine models, the results turned out to be disappointing in human trials [
44,
45]. It has been suggested that the low mutational burden in OS makes it inadequate to generate adoptive immune responses [
46]. In this context, it will be wise to combine these two immunotherapeutic regimens for the maximum of T cell activation in tumor lesions and their efficacy against solid tumors.
There are some limitations in our study that requires further acknowledgement. For instance, the NOD/SCID mice model are unable to fully simulate system reaction to human antigens like CD166, which might lead to the eruption of cytokine release syndrome and relevant toxicity. Thus, the safety issues related to off-tumor toxicity needs further studies to address.