We describe a human peripheral blood mononuclear cell (PBMC) — based humanized xenograft mouse model for translational immuno-oncology research. This protocol could serve as a general guideline for establishing and characterizing similar models for I-O therapy assessment.
The discovery and development of immuno-oncology (I-O) therapy in recent years represents a milestone in the treatment of cancer. However, treatment challenges persist. Robust and disease-relevant animal models are vital resources for continued preclinical research and development in order to address a range of additional immune checkpoints. Here, we describe a human peripheral blood mononuclear cell (PBMC) — based humanized xenograft model. BGB-A317 (Tislelizumab), an investigational humanized anti-PD-1 antibody in late-stage clinical development, is used as an example to discuss platform set-up, model characterization and drug efficacy evaluations. These humanized mice support the growth of most human tumors tested, thus allowing the assessment of I-O therapies in the context of both human immunity and human cancers. Once established, our model is comparatively time- and cost-effective, and usually yield highly reproducible results. We suggest that the protocol outlined in this article could serve as a general guideline for establishing mouse models reconstituted with human PBMC and tumors for I-O research.
Immuno-oncology (I-O) is a rapidly expanding field of cancer treatment. Researchers have recently started to appreciate the therapeutic potential of modulating functions of the immune system to attack tumors. Immune checkpoint blockades have demonstrated encouraging activities in a variety of cancer types, including melanoma, renal cell carcinoma, head and neck, lung, bladder and prostate cancers1,2. Contrary to targeted therapies that directly kill cancer cells, I-O therapies potentiate the body’s immune system to attack tumors3.
To date, numerous relevant I-O animal models have been established. These include: 1) mouse tumor cell lines or tumor homograft in syngeneic mice; 2) spontaneous tumors derived from genetically engineered mouse (GEM) or carcinogen-induction; 3) chimeric GEMs with the knock-in of human drug target(s) in a functional murine immune system; and 4) mice with reconstituted human immunity transplanted with human cancer cells or patient-derived xenografts (PDXs). Each of these models have obvious advantages as well as limitations, which have been described and reviewed extensively elsewhere4.
Reconstitution of human immunity in immunodeficient mice have been growingly appreciated as a clinically relevant approach for translational I-O research. This is usually achieved through either 1) engraftment of adult immune cells (e.g., peripheral blood mononuclear cells (PMBC))5,6, or 2) engraftment of hematopoietic stem cells (HSC) from, for example, umbilical cord blood or fetal liver7,8. These humanized mice could support the growth of human tumors, thus allowing the assessment of I-O therapies in the context of both human immunity and human cancers. Despite the advantages, applications of humanized mice in I-O research were usually hindered by several concerns, such as long model development time and considerably high cost.
Here, we describe a human PBMC-based model that could be widely applied for translational I-O studies. This model is comparatively time- and cost-effective with high reproducibility in efficacy studies. It has been used in-house for the evaluations of several I-O therapeutics currently under preclinical and clinical development. BGB-A317 (Tislelizumab), an investigational humanized anti-PD-1 antibody9 , is used as the example to discuss model development, characterization, and possible applications for anti-tumor efficacy analyses.
All procedures performed in studies involving human participants were in accordance with the ethical standards of BeiGene and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study. All procedures performed in studies involving animals were approved by the Internal Review Board at BeiGene. This protocol has been specifically adjusted for the evaluation of BGB-A317 (Tislelizumab) in humanized NOD/SCID mice.
1. Establishment of human PBMC-based model
2. PBMC Donor Screen
3. Human Cancer Cell Line and PDX Screen
4. Immunohistochemistry (IHC)
5. In Vivo efficacy and Pharmacodynamics Studies in Humanized PBMC-NOD/SCID Xenograft Models
Following the procedures presented here, a PBMC-based humanized xenograft model was successfully established. In brief, CP myeloablation effects in NOD/SCID mice was determined by flow cytometry analysis of neutrophil and monocyte populations post CP and DS treatment (Figure 1). 100 mg/kg CP plus 125 mg/kg DS was determined as the optimal dose and used in later studies as the regimen results in maximum depletion of neutrophils and monocytes without causing severe toxicity to mice. Next, human PBMC and tumor transplantation was performed. Presence of human immune cell infiltrates in the tumor microenvironment was verified by IHC (Figure 2).
A panel of PBMC donors were screened in vivo to ensure relatively high immune cell infiltrations in the tumor microenvironment and acceptable tumor growth rate (section 2, Figure 3A). Meanwhile, over 30 human cancer cell lines as well as over 20 PDXs of different cancer types were screened to evaluate tumor growth rate, tumor PD-L1 expression and immune cell infiltrations (section 3). Representative results were shown in Figure 3B.
These PBMC-engrafted humanized mice were then used to examine the anti-tumor activity of BGB-A317. Human PBMCs from selected healthy donors were co-injected with human tumor cells (A431, SKOV3 and SK-MES-1) or primary tumor tissue fragments derived from cancer patients (BCCO-028 and BCLU-054) subcutaneously. The mice were treated, as indicated in Figure 4, with BGB-A317 or PBS intraperitoneally once a week from the day of tumor implantation. In all abovementioned models, BGB-A317 demonstrated significant anti-tumor activities (Figure 4).
Figure 1: Myeloablation of NOD/SCID mice using cyclophosphamide (CP) and disulfiram (DS). (A) Gating strategy used to identify myeloid cell subsets including neutrophils and monocytes. (B) Representative results of myeloid cell (CD11b+), neutrophil (CD11b+Ly6G+) and monocyte (CD11b+Ly6C+) numbers upon different dosages of CP treatment. The boxes represent the 75th, 50th and 25th percentile of the values. The top and bottom lines represent maximum and minimal data points within the 1.5x IQ (inter quarter) range, respectively. n = 3 for the vehicle group and n = 6 for CP and/or DS treated groups. Please click here to view a larger version of this figure.
Figure 2: Human PBMC transplantation and tumor engraftment. (A) Schematic diagram showing the general workflow of PBMC-based humanized xenograft model. (B) Tumor growth of A431 cells upon subcutaneous co-injection with donor PBMC with the indicated conditions (data represents mean tumor volume ± SEM, n = 6). (C) IHC analysis of tumors developed in mice treated with or without CP+DS. Please click here to view a larger version of this figure.
Figure 3: PBMC donor and human cancer cell line screen. (A) Representative summary data from PBMC donor screen. PBMC were mixed with A431 cells and inoculated subcutaneously in humanized NOD/SCID mice (see step 2). Each dot represents the mean data value of 3 mice engrafted with PBMCs from 1 donor. (B) Representative results from human cancer cell line screen. PBMC from selected donors were co-injected with A431, SK-MES-1 or SKOV3 cells. Data represents mean tumor volume ± SEM collected from 3 mice, 14 day post inoculation of the indicated cell lines. Mean IHC score ± SEM represents the average expression of human CD8, PD-1, and PD-L1 of all 3 mice. Please click here to view a larger version of this figure.
Figure 4: Anti-tumor activities and pharmacodynamics analysis of BGB-A317 in PBMC-based humanized xenograft model. The anti-tumor activity of BGB-A317 at indicated doses (i.p., QW) was assessed using human cancer cell lines (A) A431 (with 5 x 106 PBMC), (C) SKOV3 (with 5 x 106 PBMC), (D) SK-MES-1 (with 5 x 106 PBMC), and patient-derived xenografts (PDXs) (E) BCLU-054 (with 5 x 106 PBMC) and (F) BCCO-028 (with 5 x 106 PBMC). PBMC from selected healthy donors and corresponding tumor cells were co-injected subcutaneously into humanized NOD/SCID mice (n = 8 to 10). (B) Quantification of tumor-infiltrating hCD8+ and hCD8+hPD-1+ cells in BGB-A317 treated A431 tumors. n = 8-10 animals per group in A, and C to F, n = 4-6 animals per group in B; data represents mean ± SEM. The significance was evaluated using a two-tailed unpaired Student's t-test under the assumption of unequal variance. Please click here to view a larger version of this figure.
Our knowledge of cancer development and progression has advanced significantly in recent years, with focus on a comprehensive understanding of both the tumor cells and its associated stroma. Harnessing the host immune mechanisms could induce a greater impact against cancer cells, representing a promising treatment strategy. Murine models with intact mouse immune systems, such as syngeneic and GEM models, have been widely used to study checkpoint-mediated immunity. Efficacy assessments using these models depend largely on surrogate anti-mouse target antibodies13,14. However, inherent differences between human and murine immune systems and the lack of some human targets in murine models limit preclinical studies of I-O anti-tumor effects15,16. Therefore, robust mouse models that include both human immune cells and human tumors are urgently desired, which will significantly improve the translation and development of novel I-O therapeutics.
Here, we describe a human PBMC-based xenograft mouse model that could potentially be widely used for translational I-O studies. PBMC donor as well as cancer cell line/PDX screens are critical to ensure robustness and reproducibility of in vivo efficacy studies. PBMC donors were screened in vivo to ensure successful human tumor and immune cell engraftment. Meanwhile, over 50 cell lines and PDXs of various human cancer origins were screened to evaluate tumor growth rate, human PD-L1 expression, and immune cell infiltrations. Our analyses suggest that about 20% of cancer cell lines and PDXs examined demonstrate acceptable tumor growth rate while at the same time having relatively high TILs and PD-L1 staining, which are considered good models for I-O efficacy evaluations.
HLA matching is routinely used in the clinic to match patients and donors for organ or marrow transplants17. The authors, however, have only performed limited characterization on HLA typing, and this remains an interesting topic to be investigated in future studies. The authors would like to note that a PBMC donor might be suitable for one cancer cell line/ PDX but not ideal for others. Therefore, PBMC donors might need to be screened for each cancer model to ensure optimal results.
Engraftment of human PBMC into NOD/SCID or NSG mice invariably leads to a xenogeneic graft-vs-host disease (xGvHD), a post-transplant disorder that results from immune-mediated attack of recipient tissue by donor T cells18,19. Clinical observations commonly associated with xGVHD have been observed in our humanized model, such as erythema, hunched posture, weight loss and mortality (data not shown). These phenotypes were usually observed towards the end of our studies, usually at 1-2 months post engraftment, indicating the propagation and infiltration of human T cells in xGVHD target organs. This allows a 1-2 months window for therapeutic I-O therapy evaluations. Several approaches have been utilized to decrease mouse innate immunity and enhance human immune cell engraftment20,21. For example, NSG mice defective in murine MHC Class I and Class II expression support engraftment of functional human T cells in the absence of acute xGvHD following injection of PBMC22.
NOD/SCID mice were used in this protocol. Mice homozygous for the SCID mutation have impaired T and B cell lymphocyte development and the NOD background additionally results in deficient natural killer (NK) cell function20. Other more highly immunodeficient mice, such as the NSG (The Jackson Laboratory), NCG (Charles River) and NOG (CIEA) strains, have been established. When engrafted with PBMCs, these mice have been shown develop human immune cells and form an environment that resembles the human immune system23,24. Alternatively, these mice could be engrafted with CD34+ human hematopoietic progenitor cells (HPCs) and display more sustained T cell differentiation and maturation25. In addition, next generation immunodeficient mouse models with further genetic modifications have been established to support better human myeloid lineage development and increased engraftment efficiency (refer to the variants portfolio webpage of The Jackson Laboratory and Taconic Biosciences).
More details of using these new strains for human immunity reconstitution are pending further investigations. Nevertheless, the protocol outlined in this article could serve as a general guideline of establishing and characterizing immunodeficient mouse models reconstituted with human PBMC. Three human cancer cell lines and two human patient-derived xenografts, covering a range of cancer types, are demonstrated in this article, suggesting the potential broad applications of our protocol in translational I-O studies. Most humanized PBMC models, to our knowledge, have chosen IV or IP as the route of injection26,27. Our models instead provide partially reconstituted human immunity in human tumor bearing mice through subcutaneously admixing of human PBMC with cancer xenografts. This approach provides a rapid and cost-effective, yet highly reproducible alternative to full stem cell reconstitution (e.g., CD34+ hematopoietic stem cell-engrafted humanized mice). Our model has been proved to be useful for evaluating T cell-engaging cancer immunotherapies, particularly when working on short timelines or to select agents before moving to a more complex multi-lineage immunity model.
The authors have nothing to disclose.
We thank members of our laboratories for helpful discussions. This work was partially supported by the Biomedical and Life Science Innovation and Cultivation Research Program of the Beijing Municipal Science and Technology Commission under Grant Agreement No. Z151100003915070 (project "Preclinical study on a novel immune oncology anti-tumor drug BGB-A317"), and it was also partially supported by internal company funding for preclinical research.
PBMC separation /cell culture | |||
Histopaque-1077 | Sigma | 10771 | Cell isolation |
DMEM | Corning | 10-013-CVR | Cell culture |
DPBS | Corning | 21-031-CVR | Cell culture |
FBS | Corning | 35-076-CV | Cell culture |
Penicillin-Streptomycin, Liquid | Gibco | 15140-163 | Cell culture |
Trypsin-EDTA (0.25%), phenol red | Gibco | 25200-114 | Cell culture |
Matrigel | Corning | 356237 | CDX inoculation |
FACS analysis | |||
Deoxyribonuclease I from bovine pancreas | Sigma | DN25 | Sample preparation |
Collagenase Type I | Sigma | C0130 | Sample preparation |
Anti-mouse/human CD11b (M1/70) antibody | BioLegend | 101206 | FACS |
Anti-mouse Ly-6C (HK1.4) antibody | BioLegend | 128008 | FACS |
Anti-mouse Ly-6G (1A8) antibody | BioLegend | 127614 | FACS |
Anti-human CD8 (OKT8) antibody | Sungene Biotech | H10082-11H | FACS |
Anti-human CD279 (MIH4) antibody | eBioscience | 12-9969-42 | FACS |
Anti-human CD3 (HIT3a) antibody | 4A Biotech | — | FACS |
Guava easyCyte 8HT Benchtop Flow Cytometer | Millipore | 0500-4008 | FACS |
Tumor/PDX implantation /dosing / measurement | |||
Cyclophosphamide | J&K | Cat#419656, CAS#6055-19-2 | In vivo efficacy |
Disulfiram | J&K | Cat#591123, CAS#97-77-8 | In vivo efficacy |
Syringe | BD | 300841 | CDX inoculation |
Hypodermic needles (14G) | Shanghai SA Mediciall & Plastic Instruments Co., Ltd. | 0.7*32 TW SB | PDX inoculation |
Vernier Caliper (MarCal) | Mahr | 16ER | Tumor measurement |
IVC individual ventilated cages | Lingyunboji Ltd. | IVC-128 | Animal facility |
IHC | |||
Leica ASP200 Vacuum tissue processor | Leica | ASP200 | IHC |
Leica RM2235 Manual Rotary Microtome for Routine Sectioning | Leica | RM2235 | IHC |
Leica EG1150 H Heated Paraffin Embedding Module | Leica | EG1150 H | IHC |
Ariol-Clinical IHC and FISH Scanner | Leica | Ariol | IHC |
Anti-human CD8 (EP334) antibody | ZSGB-Bio | ZA-0508 | IHC |
Anti-human PD1 [NAT105] antibody | Abcam | ab52587 | IHC |
Anti-human PD-L1 (E1L3N) antibody | Cell Signaling Technology | 13684S | IHC |
Polink-2 plus Polymer HRP Detection System | ZSGB-Bio | PV-9001/9002 | IHC |