Anti-PD1 Therapy for Pediatric Sarcomas


In this study we investigated the use of a novel chimeric antigen receptor (CAR) targeting the disialoganglioside GD2 for sarcoma immunotherapy. In addition to the well known expression of GD2 on neuroblastoma, we generated data that demonstrated GD2 to be highly expressed on human osteosarcoma and rhabdomyosarcoma cell lines and patient samples. Assays conducted ex vivo verified that the same cell lines were targets of GD2-CAR-mediated killing. We then explored whether this immunotherapy treatment could eradicate osteosarcoma in mice.

Remarkably, despite high cytolytic activity ex vivo, the GD2-CAR transduced T-cells failed to impact tumor growth in mice. Upon further examination of the transduced T-cells, we discovered that they expressed an important molecule that was potentially responsible for negative regulation of the T-cell response, programmed death-1 (PD-1).

Normally, the PD-1 checkpoint serves an important role in dampening aberrant T-cell responses and is thus critical for preventing autoimmunity. Many solid tumors have exploited this pathway by expressing PD-L1, PD-1’s major ligand, in effort to turn off T-cell activity and escape immune recognition. We hypothesized that blocking antibodies specific for PD-1 would disrupt this negative regulatory pathway and would result in enhanced CAR T-cell effector function. Indeed, we found that blocking PD1 enhanced cytolytic activity of CAR-transduced T-cells when they were co-incubated with GD2-expressing tumor cell lines ex vivo. We also found that when we treated mice bearing human osteosarcoma with a combination of CAR T-cells and anti-PD-1 antibody we observed a trend toward enhanced function of the CD8 T-cells when compared to CAR T-cells left untreated as measured by production of a cytokine, interferon-gamma. These results led us to predict that treatment with anti-PD1 would increase the effectiveness of CAR T-cells against GD2+ tumors in mice.

Unfortunately, we did not witness a significant difference in tumor growth or survival between groups that were treated with anti-PD1 and those that were not. During the course of this project we discovered a tumor-induced expansion of suppressor cells termed myeloid-derived suppressor cells (MDSC). MDSCs are well known to promote T-cell hyporesponsiveness and contribute to tumor evasion in both mouse and man, and, in fact, we discovered that host (mouse) MDSCs readily suppressed human T-cell proliferation.

Hence, our current studies are incorporating two approaches. We have devised a strategy by which we are able to neutralize MDSC suppressive capability on CAR transduced T-cells and are currently exploring the effects of PD-1 blockade therapy in this setting in the mouse model.

In conclusion, our results did not demonstrate that anti-PD1 was sufficient to fully restore the effectiveness of GD2-CAR T cells against osteosarcoma in our xenograft mouse model. Whether this is due to the fact that PD-1 is not an important mediator limiting the success in this setting or whether this is due to the fact that PD-1 is one of many important mediators limiting effectiveness in this setting is not yet known. Ongoing studies will determine whether combining anti-PD1 with approaches that alter the tumor microenvironment by neutralizing suppressive MDSCs will synergistically augment anti-GD2 CAR function and improve the efficacy of these immune-mediated killers.

Previous studies in our lab using mouse sarcoma cell lines orthotopically injected into immune competent mice have shown that disruption of sarcoma-induced MDSC migration to the tumor resulted in significantly enhanced T-cell function after anti-PD1 administration. The extent to which suppressive murine MDSC were expanded in our CAR model by human xenogeneic tumors was surprising. To our knowledge, this is a novel finding that needs to be taken into consideration when using xenograft models to determine the efficacy of adoptive T-cell therapy, especially for solid tumors.


Optimizing Genetically Engineered Lymphocytes for Immunotherapy of Pediatric Sarcoma via Blockade of Negative Regulatory Pathways


Tumor cells express proteins, lipids, and carbohydrates on their cell surface that are sometimes truly unique and often increased in number compared to that found on non-malignant cells. Immunologist can target tumor-specific or tumor-associated cell surface molecules for destruction by the immune system. One carbohydrate highly expressed on the surface of tumor cells is the disialoganglioside GD2. Although first described on the surface of neuroblastoma, studies show that GD2 expression in human osteosarcoma and rhabdomyosarcoma cell lines is equivalent to, or higher than, the prototypical neuroblastoma cell line LAN5.unpublished observation Thus, GD2 is also a potential target for pediatric sarcomas.

Disialogangliosides in Cancer

The membrane of all cells is composed of a lipid bilayer. A ganglioside is a lipid that also expresses a unique sugar (carbohydrate) residue. An array of gangliosides are found on the surface of normal brain and nerve cells. The ganglioside GD2 is normally expressed at very low levels in the central nervous system and peripheral nerves but some childhood cancers, such as neuroblastoma, express high levels of GD2. Anti-GD2 antibodies bind GD2 with high affinity. Once bound, other immune cells, such as natural killer (NK) cells, engage the antibody and kill the antibody-decorated cancer cells. Clinical trials using anti-GD2 antibodies have shown promise for neuroblastoma, specifically following bone marrow transplantation for high-risk patients, but they have not yet been able to induce tumor shrinkage in the presence of measurable disease. Nonetheless, the positive clinical results seen with anti-GD2 antibody absolutely demonstrates that activation of the immune system can be used to treat solid tumors.

Our lab is focused on using cytotoxic T-cells to eliminate sarcoma. Patient-derived T-cells can be engineered to express chimeric antigen receptors (CAR) ex vivo and then re-infused for therapeutic anti-tumor effect. CARs are created by fusing an extracellular binding domain derived from an antibody (in our case anti-GD2) with intracellular T-cell signaling domains (i.e. those derived from the TCR zeta chain, 4-1BB, or CD28). GD2 CAR-engineered T-cells recognize tumor in a non-MHC restricted manner, then become activated, expand and mediate cytolysis against tumor cell targets.1

Engineering T-cells to Target Cancer: Chimeric Antigen Receptor (CAR) Modified T-cells

Figure 1.

Figure 1: Basic Chimeric Antigen Receptor (CAR).

One limitation for using adoptive T-cell transfer for the treatment of cancer is that it is extremely difficult to obtain enough T-cells from patients that are specific for molecules expressed by the cancer. Scientists have devised a method by which we can use retroviral gene vectors to force-express receptors on normal T-cells that allow them to target specific molecules on the surface of various cancers. In general, CARs consist of 1) a region that imparts tumor-cell specificity which is a single chain variable fragment (scFV) derived from the variable sequences of an antibody, 2) an internal signaling region (CD3z), 3) a region that acts to amplify the internal activation signal of the molecule (such as CD28, 4-1BB, or OX40). This method allows for the efficient production of a large number of T-cell clones all capable of targeting and destroying cancer cells.

We have generated a series of GD2-specific CARs and have optimized both extracellular domain structure and the intracellular signaling required to generate highly active anti-tumor reactive T-cells. First generation GD2-CAR transduced T-cells (expressing only one signaling motif) have already been tested in clinical trials for the treatment of neuroblastoma and results demonstrate that this therapy is safe, and can induce meaningful antitumor effects even in patients with advanced disease.2,3 Second and third generation GD2-CAR-transduced T-cells generated in our lab exhibit potent anti-tumor responses in vitro. Cytotoxic T-cells assays show that GD2-CAR anti-tumor activity is comparable between sarcoma cell lines and the neuroblastoma cell line (LAN5).unpublished observation

We will next evaluate the in vivo efficacy of GD2-CAR transduced human T-cells using xenogeneic murine models. In order to allow for human tumor growth, mice lacking a functional immune system (NSG mice: NOD.Cg-PrkdcscidIl2rgtm1Wjl) will be injected intramuscularly with human osteosarcoma (143B) and rhabdomyosarcoma (RH18) cells and GD2-CAR transduced T-cells will be administered once the tumor becomes established. We will monitor tumor area twice weekly to determine the direct effects of the CAR-engineered T-cells. Preliminary data gathered by our laboratory shows that CAR-T-cells express a surface molecule termed programmed death-1 (PD-1). PD-1 is an immune checkpoint receptor that mediates immune suppression and is expressed by activated T-cells. T-cells expressing PD-1 are rendered nonfunctional upon encounter with PD-1 ligands (primarily PDL-1) which are highly expressed by many tumor cells,4 including pediatric sarcomas.

Programmed Death-1 (PD-1): A ‘Turn-Off’ Switch for the Immune System

The first report which clearly demonstrated the importance of PD-1 in negatively regulating immune responses was performed in mice lacking the ability to express PD-1 (Pdcd1 KO). It was discovered that these mice exhibited a hyperactive immune system and developed a severe lupus-like autoimmune disease (Nishimura H, 1999). Not long after, the ligands for PD-1 were identified and it was found that PD-1 prevents autoimmunity by inhibiting the activation of auto-reactive T-cells. Researchers soon realized that many human tumors express high levels of PD-1 ligands (PDL-1 and PDL-2) and wondered if this was a mechanism cancers had evolved that would allow them to evade the immune system. Experiments in mice with numerous types of cancers had shown improved overall responses with single agent PD-1 blockade (either blocking PD-1 or PDL-1) and led to the clinical trials in humans. Recently, the first reports were published on the effectiveness of this new therapy which showed impressive results in the ability to shrink lung, kidney and skin cancers (Topalian SL, NEJM 2012; Brahmer JR, NEJM 2012). This research will seek to determine if PD-1 blockade can enhance the effect of adoptive cell therapy for the treatment of cancers.

Simultaneous to conducting these studies, we had developed a murine model to test our hypothesis that the PD1-PDL1 pathway is an active pathway in sarcoma that mediates T-cell dysfunction. Blockade of this axis has recently demonstrated clinical activity in trials targeting patients with advanced cancer, presumably by blocking inhibitory signals on tumor specific T cells.5,6 We utilized the murine embryonal rhabdomyosarcoma (eRMS) cell line, M3-9-M, which was derived from a C57BL/6 mouse transgenic for hepatocyte growth factor and heterozygous for mutated p53.7 This murine cell line expresses high levels of PDL-1, and also induces the expression of PD-1 on peripheral CD4+ and CD8+ T-cells when implanted into the gastrocnemius muscle in WT C57BL/6 mice.unpublished observation Tumor immunotherapy using PD-1 blocking antibodies had antitumor effects but these were limited in the setting of bulk disease. Taken together, these murine data identify PD1-PDL1 as an important negative regulatory axis in sarcoma and suggest that disruption of this axis via PD-1 blockade results in an improvement in T-cell function and an enhancement in their ability to target tumor cells.

Aim of the Study

Figure 2: Overview of all 7 MTAP splice forms. RT-PCR shown on the right.

Figure 2: Proposed model of CAR T-cell impact on sarcoma progression.

In the current proposal, our aim is to determine the efficacy of GD2-CAR T-cells to target and kill human sarcoma cell lines in vitro and in vivo and to determine if blockade of the negative costimulatory molecule Programmed Death 1 (PD-1) can improve the potency of this cell-mediated therapy. We hypothesize that by blocking this negative regulatory axis on a highly enriched population of tumor-specific T-cells we will dramatically improve the efficacy of GD2-CAR therapy.

Experimental Outline

1.1 CAR construction - CARs are generated by de novo synthesis using primary amino acid sequences encoding an antigen binding domain and intracellular co-stimulatory signaling domains with or without the addition of a human immunoglobulin heavy chain constant region IgG1 CH2CH3 region. Antigen binding domains consisting of the single chain variable fragment (scFv) sequence were obtained from previously published sequences.3,8 Intracellular signaling motifs comprised of second generation (CD28 and CD3-zeta chain) or third generation (CD28, CD137, and CD3-zeta chain) signaling domains are included in retroviral backbone vectors.9 Tumor-binding domains, with or without CH2CH3 sequences are synthesized using codon-optimization algorithms for mammalian gene expression.

1.2 Transduction and activation of human lymphocytes - Retroviral supernatants will be generated by transfecting 293GP packaging cells with an MSGV retroviral vector plasmid encoding the GD2-CAR, retroviral packaging elements, and a second plasmid encoding the RD114 envelope using Lipofectamine 2000 reagent.9,10 For T-cell transduction, peripheral blood mononuclear cells (PBMC) will be thawed and activated with OKT3 and recombinant interleukin-2 (rIL-2) for 2 days in complete AIM-V media. Non-tissue culture-treated 6-well plates will be coated with Retronectin for two hours at room temperature (RT), then used to capture retrovirus from supernatants. Activated human lymphocytes will be added to vector-coated plates in cAIM-V plus rIL-2 and cultured overnight. Transduction will be repeated on the following day and transduced cells will be expanded in cAIM-V plus rIL-2 for 5 additional days. Flow cytometry using anti-idiotype GD2-CAR antibody will be used to determine the expression and transduction efficiency. Anti-human PD-1 antibodies will be used to measure the expression levels of PD-1.

1.3 GD2-CAR T-cell functional assay and blockade of PD-1 in vitro – Here, we will determine the capacity of GD2-CAR T-cells to lyse sarcoma cell lines in vitro and how blockade of PD-1 affects this function. GD2-CAR T-cells generated above will be tested for their ability to lyse a panel of sarcoma cell lines using a standard chromium-51 (51Cr) release assay. This panel will consist of cell lines known to express GD2 including rhabdomyosarcoma (Rh18, Rh30, Rh41), osteosarcoma (143b, G292, MG63.2), Ewings sarcoma (Tc71), and neuroblastoma (LAN5, Sy5y, SKNSH). All tumor lines will also be tested for their expression of the PD-1 ligands (PDL-1 and PDL-2). For the 51Cr assay, target tumor cells from the panel will be labeled with 51Cr for 1 hour. These target cells will then be co-incubated with either GD2-CAR transduced or mock-transduced T-cells at varying effector:target (E:T) cell ratios (30:1, 15:1, 7.5:1, 3.75:1, 1.88:1, 0.94:1). After 4 hours, supernatants from the co-incubation will be collected and assayed for 51Cr using a Top Count Reader. Tumor cell lysis will be calculated as follows: % lysis=(experimental lysis)-(spontaneous lysis)/(maximum lysis-spontaneous lysis)x100. Functional lysis assays will be performed with and without the addition of PD-1 blocking antibody. Anti-human PD-1 (clone J116) will be added to GD2-CAR T-cell products just prior to the addition of tumor cell targets. Tumor cells found to be negative for the expression of PDL-1 and PDL-2 will serve as negative controls for PD-1 blockade. We predict that the addition of PD-1 blocking antibodies will lead to a significant boost the lytic activity of these CAR-modified T-cells.

1.4 Therapeutic activity of GD2-CAR T-cells against human sarcoma in vivo and blockade of PD-1– Immune-compromised NSG Mice (NOD scid gamma, NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ JAX®, Jackson Laboratories, Bar Harbor, ME) will be injected into the gastrocnemius muscle with human tumor cells that have varying expression levels of both GD2 and PDL-1 (from 1.3). On day 3, GD2-CAR transduced or mock transduced T-cells will be injected i.v. Anti-human PD-1 or anti-human IgG control antibody will be given 2x/weekly starting at day 3 i.p. at 200ug/mouse until day 30. Tumor measurements will be taken twice weekly using a caliper and survival will be monitored daily. A separate cohort of mice will be euthanized on days 7, 10, 14 and 21 and peripheral blood, spleen, and tumor will be analyzed via flow cytometry for the presence of GD2-CAR transduced T-cells. We will also use flow cytometry at these same time points to determine the activation status and the ability of these T-cells to secrete proinflammatory cytokines (IFNγ, TNFα, IL2). We predict that PD-1 blockade will enhance the effectiveness of GD2-CAR therapy for multiple pediatric sarcomas.


Our preliminary data demonstrate that disialoganglioside GD2 is highly expressed on pediatric sarcomas and can therefore be a target for immune therapy. We will develop and test GD2-specific CAR T-cells for their ability to lyse sarcoma in vitro. Unpublished observations conducted in our laboratory indicate that GD2-CAR T-cell function may be disrupted due to the binding of PD-1 on the CAR T-cells to PDL-1 on the sarcoma. We propose to enhance the efficacy of these T-cells by blocking the negative regulatory axis PD1-PDL1. These studies may not only have a positive impact on sarcoma, but also may have a more broad influence as the results may effect the way in which CAR T-cell therapy is applied to many different types of cancers.

By Steven L Highfill, PhD
and Rimas J Orentas, PhD
and Crystal L Mackall, MD
Pediatric Oncology Branch
National Cancer Institute


1. Jena B, Dotti G, Cooper LJ. Redirecting T-cell specificity by introducing a tumor-specific chimeric antigen receptor. Blood. Aug 19 2010;116(7):1035-1044.

2. Pule MA, Savoldo B, Myers GD, et al. Virus-specific T cells engineered to coexpress tumor-specific receptors: persistence and antitumor activity in individuals with neuroblastoma. Nat Med. Nov 2008;14(11):1264-1270.

3. Louis CU, Savoldo B, Dotti G, et al. Antitumor activity and long-term fate of chimeric antigen receptor-positive T cells in patients with neuroblastoma. Blood. Dec 1 2011;118(23):6050-6056.

4. Okazaki T, Honjo T. The PD-1-PD-L pathway in immunological tolerance. Trends Immunol. Apr 2006;27(4):195-201.

5. Brahmer JR, Tykodi SS, Chow LQ, et al. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. N Engl J Med. Jun 28 2012;366(26):2455-2465.

6. Topalian SL, Hodi FS, Brahmer JR, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med. Jun 28 2012;366(26):2443-2454.

7. Meadors JL, Cui Y, Chen QR, et al. Murine rhabdomyosarcoma is immunogenic and responsive to T-cell-based immunotherapy. Pediatr Blood Cancer. Dec 1 2011;57(6):921-929.

8. Rossig C, Bollard CM, Nuchtern JG, Merchant DA, Brenner MK. Targeting of G(D2)-positive tumor cells by human T lymphocytes engineered to express chimeric T-cell receptor genes. Int J Cancer. Oct 15 2001;94(2):228-236.

9. Zhao Y, Wang QJ, Yang S, et al. A herceptin-based chimeric antigen receptor with modified signaling domains leads to enhanced survival of transduced T lymphocytes and antitumor activity. J Immunol. Nov 1 2009;183(9):5563-5574.

10. Miller AD, Rosman GJ. Improved retroviral vectors for gene transfer and expression. Biotechniques. Oct 1989;7(9):980-982, 984-986, 989-990.

Grant Funds Immunotherapy Research on Pediatric Sarcomas

The Liddy Shriver Sarcoma Initiative has awarded a $50,000 grant to fund promising research on pediatric sarcomas by investigators at the National Cancer Institute. In the study, Steven Highfill, PhD; Rimas Orentas, PhD; and Crystal Mackall, MD will use state-of-the-art immunotherapeutics to target osteosarcoma and rhabdomyosarcoma cell lines and mouse models.

Immunotherapy holds much promise in treating and curing cancers. According to Dr. Mackall, "Immunotherapy has the potential to slow tumor growth, shrink tumors and potentially cure patients of sarcoma." This project works on improving immunotherapies that are specifically targeted to treat pediatric sarcomas.

Part I of the Study

In the first part of this study, investigators plan to alter immune cells (T cells) so that they recognize and kill sarcoma cells as if they were virus-infected cells.

According to Dr. Orentas, there is a unique signature on the surface of osteosarcoma and rhabdomyosarcoma cells called GD2 (disialoganglioside). The GD2 tumor signature, or antigen, is not present in normal cells. In order to help a patient's body attack the cancer, the scientists will add an artificial receptor to the patient's T cells that will recognize the GD2 on the surface of the cancer cells. What will happen next? Dr. Highfill predicts that "once the receptor on the T cell engages the cancer cell, the T cell will destroy it."

The Chimeric Antigen Receptor (CAR)

By Steven Highfill, PhD

The receptor we are engineering to be expressed on the surface of the T cell is called a chimeric antigen receptor (CAR). The ability to create a CAR is based on the last two decades of basic biology studies where we have learned what structures the immune system uses to recognize foreign entities, and how that recognition activates immune cells to respond.

Part II of the Study

The second part of the investigation will address one of the obstacles to the success of immunotherapy. Dr. Highfill explains, "The human immune system has a built in 'off switch' that is normally used to prevent auto-immunity. Unfortunately, cancer cells are capable of hijacking this switch to prevent their own demise. The second goal of our research study is to thwart the cancer cells' ability to turn off these engineered T cells."

Enhancing the Effectiveness of the Immune Response

By Steven Highfill, PhD

Solid tumors are extremely good at evading the immune system through a number of mechanisms. Recent clinical results with adult solid tumors have shown tumor shrinkage after blockade of one of the most utilized pathways (programmed cell death protein 1 or PD-1). We hope to be able to combine the effects of CAR-immunotherapy with blocking negative signals to see an enhanced effect in tumor clearance.

The researchers plan to use blocking antibodies, also known as "checkpoint inhibitors," to prevent sarcoma from dampening the T cells' immune reaction. Dr. Orentas adds, "We are exploring the ability of blocking the PD1 interaction as a means of keeping T cells active and maintaining the anti-sarcoma immune response."

The Promise of this Research

Engineered T cells are already being used as a treatment option for other types of cancers, such as leukemia, with much success. This study is expected to accelerate advances in immunotherapy for sarcoma patients in particular. Dr. Mackall believes that studies like this one are crucial for sarcoma patients. She says, "Cancer is not one disease but many diseases, and within the larger framework of cancer, sarcoma is not one disease but many diseases. If we don't study the disease, we have no hope of curing it."

Better Treatments Will Come From Sarcoma-Specific Research

By Rimas Orentas, PhD

The era of large scale, highly toxic chemotherapy has come full circle. Now with the completion of the human genome project, we have the ability to understand individual tumors at the genetic level, which may vary from one pediatric malignancy to another, and in some cases from one patient to another. Therefore, we have the responsibility to keep doing the basic biological studies that will discover new gene-specific or mutation-specific drugs. This has been termed "targeted therapy" or "personalized medicine."

In this study, we will target a known tumor antigen on the surface of sarcoma, GD2. Our hypothesis is that more tumor specific antigens remain to be found. So, just as pediatric tumors are very different from adult carcinomas, sarcomas are likely to display a unique cell surface antigenic signature that we can exploit for immunotherapy.

The power of targeted therapy is that we have a huge increase in the number of specific and less-toxic approaches to therapy. The downside is that the more specific treatments become, the less broadly applicable they will be to all cancers. Thus, in this age of targeted therapy, it is sarcoma-specific research that will discover the targets most applicable to sarcoma patients.

From the Investigators

Steven Highfill, PhDSteven Highfill, PhD: I am the grandson of a colon cancer victim and the son of a breast cancer survivor. Although I did not have the ability to make a difference in these cases at the times they occurred in my life, I have since been able to learn a great deal about cancer and the immune system in general. Cancer is nondiscriminatory and attacks even the things we hold most precious, our kids. As a father, I understand how difficult it is to learn of a child's illness, and my heart goes out to the fathers who have to hear a diagnosis of cancer. I am devoted to finding a cure and to alleviating some of the pain cancer brings to patients and families.

Rimas Orentas, PhDDr. Orentas: I have been focused on understanding the basic biology of the anti-tumor immune response since graduate school.As our understanding of the immune system has matured, I have focused increasingly more on translating basic advances in tumor immunology to the clinical research arena. Being based at the National Institute of Health allows for the formation of a unique investigative team where both lab-focused people (such as myself), can interact with world-class physicians treating the most severe pediatric malignancies. I am driven both by raw curiosity about how the immune response works (we are still learning!) and by a deep animosity to cancer and the suffering it brings to patients and their families.

Crystal Mackall, MDCrystal Mackall, MD: As a treating oncologist who has lost too many young people to sarcoma, I am personally committed to improving outcomes for patients with sarcoma. I also believe that we have not yet tapped into what the immune system is capable of when it comes to fighting cancer. I am excited about the new tools that we have available to harness the immune system to fight cancer, and I truly believe it can make a difference in the lives of people battling this disease.

  • Figure 1: Basic Chimeric Antigen Receptor (CAR)
  • Figure 2: Proposed model of CAR T-cell impact on sarcoma progression.