Although companies, investors and academics are all racing to develop chimeric antigen receptor-based T cells for cancer, numerous basic science questions still surround the technology. The way forward will likely involve combining chimeric antigen receptors with antibodies, ligands or other small molecules to add specificity and safety to current therapies.

Last week's deal between Pfizer Inc. and Cellectis S.A. to develop chimeric antigen receptor (CAR) T cell therapies for up to 15 targets is the latest in a series of collaborations aiming to translate the technology to the clinic.

Among the most high-profile CAR deals are Novartis AG's partnership with the University of Pennsylvania in 2012 and Celgene Corp.'s partnership with bluebird bio Inc. and the Baylor College of Medicine in 2013, both to develop and commercialize cancer immunotherapies.

In addition, Juno Therapeutics Inc. was formed in early 2014 with a $120 million financing to commercialize immunotherapeutic discoveries by scientists at three institutions: Memorial Sloan-Kettering Cancer Center (MSKCC), the Fred Hutchinson Cancer Research Center and the Seattle Children's Research Institute.

In March, five MSKCC trials involving CD19-specific CAR T cells were halted after two treatment-related deaths. Although the trials resumed after the FDA approved MSKCC's amended protocols, the incident shone a light on the many unknowns that still surround the space.

A think tank convened by SciBX has put together a road map for tackling basic science questions and hurdles for clinical implementation of T cell-based therapies.

The think tank, comprising academic, clinical, biotech, pharma and VC stakeholders, discussed priorities for the field at the SciBX Summit on Innovation in Drug Discovery and Development in Boston. In addition, key opinion leaders interviewed by SciBX before and after the summit added their thoughts and concerns to the road map.

Summit participants identified three research areas in which work is needed to enable innovation in T cell-based immunotherapy: achieving on-target, on-tissue specificity of T cells; attaining the right balance between persistence and safety of T cells; and figuring out how T cells enter and interact with the ever-shifting tumor microenvironment.

As with many cancer therapies, eliminating on-target, tumor-specific tissue while sparing normal tissue emerged as a top priority. Although a target might be mutated or more highly expressed on tumor tissues, even trace amounts on normal cells can lead to serious side effects.

The potential to hit the right target on the wrong cell also makes it difficult to find a balance between T cell persistence and safety. The greater the persistence, the more likely patients will not relapse. But with greater persistence, there is an increased chance that T cells would ultimately find and destroy normal tissues with low levels of antigens or antigen sequence motifs.

Finally, summit participants agreed that strategies need to either prevent an immunosuppressive tumor microenvironment from developing or block the environment from shutting down T cell activity.

The good news is that the three issues could have a common solution: combining traditional CARs with secondary effectors such as antibodies, ligands or other small molecules, delivered with or expressed by the CAR-based T cells. The effectors would provide an additional mechanism to fine-tune specificity, a built-in safety switch to neutralize the engineered cells if adverse events emerge or an immunostimulatory capability to overcome the generally immunosuppressive microenvironment (see "Building a better T cell").

For example, tackling tumor specificity might require using T cells engineered to contain multiple cell-identifying mechanisms that distinguish tumor cells from nontumor cells. Several approaches, including the addition of inhibitory CARs (iCARs) or chimeric co-stimulatory receptors (CCRs), already are in development.

Specific strategies to control persistence and improve the safety profile of engineered T cells include incorporating suicide genes in the cells or using transient expression systems of mRNA that is intrinsically short lived.

The panel said that incorporating immunostimulatory adjuvants into treatment regimens-or designing T cells that present or secrete these molecules-could help T cells counter the immunosuppressive effects of tumor microenvironments.

Beyond the scientific hurdles, the CAR space is grappling with both IP challenges and the commercial viability of the treatments-the products are individualized and thus are not off the shelf.

In 2012, St. Jude Children's Research Hospital filed suit against UPenn in the U.S. District Court for the Eastern District of Pennsylvania alleging breach of materials transfer agreements regarding CAR technology. UPenn counterclaimed that its cell constructs are different from St. Jude's. Novartis and Juno have filed motions on behalf of their respective partners, UPenn and St. Jude.

Although the legal issues will eventually be sorted out, commercial issues also need to be addressed. The panel said that allogeneic cells should continue to be explored as a platform for making CAR- or T cell receptor (TCR)-based T cell therapies more readily available as off-the-shelf solutions.

In addition, the panel said that companies with autologous T cell therapies should take advantage of the existing infrastructure for transfusion medicine such as bone marrow transplantation. The method is still the most widely used form of cell therapy and could help flesh out the therapeutic potential and challenges in developing other cell-based therapies.

"There will most likely be more than one path to success, with strategies needing to be tuned for each different type of cancer," said think tank member Renier Brentjens, director of cellular therapeutics and a medical oncologist at MSKCC and a scientific cofounder of Juno.

In addition to Brentjens, the SciBX Summit think tank consisted of panelists Stewart Abbot, Gwen Binder-Scholl, Bruce Booth, Aya Jakobovits and respondent Madhusudan Peshwa.

Abbot is executive director of integrative research at Celgene. Binder-Scholl is EVP at Adaptimmune Ltd. Booth is a partner in the life sciences group at Atlas Venture. Jakobovits was president and CEO of cancer immunotherapy company Kite Pharma Inc. and is now a venture partner at OrbiMed Advisors LLC. Peshwa is EVP of cellular therapies at MaxCyte Inc.

Eye on the target

Immunotherapeutic T cells are engineered to express either a CAR or TCR that interacts with a tumor-associated antigen (see "Receptor expression and engagement of antigen"). The T cells become activated when they encounter antigen-expressing tumor cells. The result is elimination of the malignant cells and expansion and proliferation of the T cells to provide persistent protection.

In 2010 and 2011, adoptive T cell therapies grabbed headlines when second-generation CARs were used to treat CD19+ leukemias and lymphomas and showed unprecedented success in small clinician-led trials at the National Cancer Institute (NCI), UPenn and MSKCC.1-4

These second-generation molecules were more potent than their predecessors because they included two T cell signaling domains rather than one.

In December 2013, UPenn and The Children's Hospital of Philadelphia published results from a Novartis-backed clinical trial of CD19-targeting T cells in 54 adults and children with intractable leukemia. In that trial, 19 of 22 pediatric patients with acute lymphoblastic leukemia (ALL) experienced complete remission (CR). Out of those, 5 relapsed during the following 20 months.5

Among 32 adult patients with chronic lymphocytic leukemia (CLL), 7 had CR.6

Although the collective clinical results for second-generation CARs are impressive, the CD19-targeting T cells eradicate normal and malignant B cells indiscriminately. As a result, patients require lifelong immunoglobulin replacement therapy.

Thus, improving target specificity was a central talking point at the SciBX Summit.

"Target specificity can be seen as a precarious balance that is attained or lost due to the designed sequence of the CAR, TCR or T cell signaling molecules and the variability of patient-specific target expression on both cancer and normal cells," said Abbot. "Right now, we want to make the T cell and target interaction more potent. But in doing so, we have to worry about the antigen being expressed on critical normal tissues."

One idea is to pursue targets that were uncovered during decades of mAb development.

"You'd want to take another look at antibodies that target antigens expressed at low levels on cancer cells that showed limited success when incorporated into antibody conjugates-where the conjugate was not enough to supply a kick to eliminate the tumor cells," said Thomas Schuetz, a consultant in the life sciences group at Atlas Venture. "That might be where a CAR-based T cell could supply not only cytotoxic T cell function but also proliferation and expansion of the therapeutic T cells."

"People want to know what the right target is. If other studies, such as those with antibodies, have validated the target already, that's a good lead to follow," added Malcolm Brenner, a professor of molecular and human genetics and director of the Center for Cell and Gene Therapy at Baylor.

According to Booth, one of the big questions in the space is whether the technology can break into solid tumors effectively. "To date, there's not been any solid evidence in support of it," he said. "Finding suitable tumor antigens that can be targeted safely with engineered T cells is the key to that."

Brenner has had success developing CAR T cells against a known antibody target-a neuroblastoma-related disialoganglioside called GD2. In an investigator-led trial at Baylor, 6 of 11 patients with refractory or relapsed neuroblastoma had responses to the GD2-specific CAR T cell therapy, including 3 CRs.7

A moving target

T cell therapeutics may themselves induce changes in a tumor by applying selective pressure that facilitates downregulation of the target cancer antigen or causes proliferation of tumor cells that lack the target cancer antigen. In some cases, these changes result from altered conditions in the tumor microenvironment.

In all cases, the result can be populations of cancer cells that do not express the original antigen target.

"Antigen escape-changes in expressed antigens-is a real problem that needs to be considered. If the tumor cell stops producing the target, there will be incomplete killing," said Dario Campana, a professor of pediatrics at the National University of Singapore.

Carl June's group highlighted this concern in a study using CD19-targeting CAR T cells to treat two patients with ALL.8 CR is ongoing in 1 patient at 11 months' post-treatment. The other patient initially had CR but relapsed after about 2 months because escape variants appeared that did not express the CD19 antigen.

June is a professor in the Department of Pathology and Laboratory Medicine at the Perelman School of Medicine at the University of Pennsylvania and director of the translational research program at the Abramson Family Cancer Research Institute at the University of Pennsylvania.

One solution for CAR-based therapies that no longer recognize cancer cells is a second set of T cell therapeutics focused on a second target expressed by the escaped T cells. Alternatively, combinations of CAR-based T cells could be used as a first regimen, before antigen escape occurs.

"We think antigen escape can be dealt with by using a CAR fleet-an infusion of CAR T cells that have several targets," said June. "We have seen this strategy work in preclinical trials, but we haven't yet tried it in clinical trials. For ALL, one possibility might be the co-infusion of CAR T cells redirected to precursor-B acute lymphoblastic leukemia specificities, such as CD22, in addition to CD19."

The National University of Singapore and St. Jude are taking a different tack by using T cells expressing a universal CAR designed to cause antibody-dependent cell cytotoxicity.9 The CAR does not target a cancer-associated antigen and instead targets the Fc portion of antibodies that bind to cancer-associated antigens.

The result is a tool that can target many different antigens by selecting an alternative tumor-targeting antibody without requiring researchers to redesign the CAR T cells.

Combinatorial CARs

Many opinion leaders who attended the summit said that combinatorial strategies provide the best option so far for controlling T cell activity and suggested that T cells should be designed to interact with two antigens instead of one.

There are two ways to go about the two-pronged attack. One is T cell cytotoxicity that is activated upon encountering two antigens. The other involves blocking cytotoxicity when the T cells encounter one tumor antigen and a normal cell self-antigen.

"A single CAR or TCR going after a single antigen is just a little too hopeful," said Brentjens.

An MSKCC team has included in its T cell-based therapeutics-in addition to the traditionally engineered CAR-an iCAR that fuses a single-chain variable fragment (scFv) to a T cell inhibitory signaling domain (see Figure 1, "Building a better T cell"). When both antigens are present, T cell cytotoxicity is blocked.10

Michel Sadelain, director of MSKCC's Center for Cell Engineering and a scientific cofounder of Juno, has designed T cells expressing two different receptors-a CAR and a co-stimulatory CAR-that ramp up T cell function when encountering on-target cells while sparing off-target cells.11

UPenn researchers have designed T cells that express one CAR with one T cell signaling domain and another CAR with a second T cell signaling domain.12 When both CARs are engaged by antigen, the T cells are activated.

"These strategies are important first steps to provide a molecularly encoded control for T cell-based therapeutics," said Abbot. "Finding tumor-associated antigens or combinations of antigens, with limited display on crucial healthy tissues, is key."

He said that despite the possibilities provided by combination strategies, "from a clinical perspective, what we need first is validation of top-ranked antigens and, more importantly, clinical outcomes that recapitulate the type of efficacy that has been seen with CD19 CAR-based therapies."

Avoiding collateral damage

According to Seth Ettenberg, head of oncology biologics at the Novartis Institutes for BioMedical Research, no single model exists to reveal everything that needs to be known about toxicity or durable responses to therapy.

"Humanized mice allow you to begin to ask questions using the ultimate cells your therapy will reside in, but questions regarding toxicity and even ultimate dissection of the underlying biology may in some cases be very different than the human setting," Ettenberg said.

"The immunodeficient mice, while least representative, clearly allow the distinction of candidate therapies and the fastest and most amenable approach to look at human tumors," he added. "So data from each of these models can help guide selection and a better understanding of the therapy."

Two different strategies are being employed to provide an added layer of safety when first moving to patients: suicide genes and transient CARs.

A suicide gene engineered in the T cell can be chemically induced, which kills the transferred T cells and shuts down the immunotherapy. The first clinical study using CAR T cells incorporating such a safety feature was started in March by researchers from the NCI to treat pediatric patients with osteosarcoma and other blastomas of non-neural origin.

The NCI researchers are using GD2-targeting CAR T cells that incorporate Bellicum Pharmaceuticals Inc.'s CaspaCIDe safety switch technology. The technology consists of an inducible caspase-9 (CASP9; MCH6) gene that can be activated with the small molecule activator AP1903. AP1903-mediated induction of CASP9 leads to rapid destruction and elimination of T cells engineered with the suicide gene. The CaspaCIDe technology has been used to selectively eliminate harmful immune cells in patients receiving stem cell transplants and to block acute graft-versus-host disease (GvHD).13

Another way to improve safety is to use T cells with transient CAR expression.

"Some groups are considering using electroporated mRNA to establish transient CAR expression in T cells," said Abbot. "This might make a good tool to determine potential toxicities without the challenges of long-term persistence of CAR T cells. If anything goes wrong, the problem is short lived." The downside, said Abbot, is that transient CARs may have muted efficacy.

Isabelle Rivière, director of the cell therapy and cell engineering facility at MSKCC and a scientific co-founder of Juno, agreed. "Some investigators are looking at transient CAR-expressing T cells as a tool for the preliminary validation of novel target antigens. Others are exploring safety switches and combinatorial targeting strategies to achieve selective tumor targeting. It would also be extremely useful to have means to modulate T cell numbers and persistence."

"More information about how and which T cells survive in the body is needed," said Booth. "Should persistence be considered the ultimate goal? Successful studies with examples of complete remission have shown that the T cells stick around, and that may be telling us something."

"Results from UPenn have demonstrated that persistence of CAR T cells in patients is a hallmark of successful therapy," noted Ettenberg. He added that it would be hard to get the benefits of transient CARs without sacrificing potency.

Peshwa was not ready to consign transient CARs to basic research. He said that transient CAR T cells, generated using mRNA electroporation, might be persistent enough to battle cancer while showing additional benefits not related to persistence.

The shifting tumor microenvironment

As more information is being uncovered about how T cells interact with the tumor microenvironment, it is becoming clear that initial T cell success can be dampened as tumor microenvironment conditions change during the course of treatment.

Solid tumors create their own microenvironment. They are heterogeneous, structurally complex and can recruit a variety of cell types, including fibroblasts, immune inflammatory cells and endothelial cells, through production and secretion of stimulatory growth factors and cytokines.14

Hematological cancers also may form tumors that can be found within the microenvironment of bone marrow or secondary lymphoid organs.

The cancers can employ a host of mechanisms that can occur concurrently within the microenvironment to tamp down the immune response and reduce the effectiveness of immunotherapy.14-16

CARs might need to be combined with checkpoint inhibitors such as molecules that target programmed cell death 1 (PDCD1; PD-1; CD279) or CTLA-4 (CD152).

An alternative strategy being tested is to convert suppressive checkpoint signals found on T cells into stimulatory signals.

For example, T cells have been engineered to express an extracellular checkpoint receptor, PD-1, linked to an intracellular CD28. The transmembrane complex enhances T cell function in the presence of the tumor-associated checkpoint ligand programmed cell death 1 ligand 1 (PD-L1; B7-H1; CD274) rather than suppressing T cell function.15

"Right now, many of the clinical trials using engineered CAR T cells in blood cancers are simultaneously analyzing the tumor microenvironment and the transferred T cells to determine the potential for cells to be functionally suppressed and to understand the mechanisms by which they are suppressed," Stanley Riddell told SciBX. "This may provide insights into appropriate combinatorial therapies."

Riddell is a member of the Clinical Research Division at the Fred Hutchinson Cancer Research Center and a professor of oncology at the University of Washington. He also is a scientific cofounder of Juno.

Novartis has already put itself at the ready by acquiring CoStim Pharmaceuticals Inc. and its portfolio of late-stage discovery programs focused on immune checkpoint proteins, including PD-1. Novartis said that the deal will provide the potential to combine checkpoint inhibitors with targeted therapies in the pharma's pipeline, including CAR immunotherapies.

Manipulating cytokine profiles also can be used to enable T cells to resist and change the tumor microenvironment. T cells expressing IL-12 have been shown to have enhanced antitumor function, resist immunosuppression by Treg cells and change myeloid cell composition within the tumor microenvironment from immunosuppressive to immunostimulatory.

Make it work

As the scientific issues are being sorted out, manufacturing and clinical protocol issues also require attention.

"In reality we're really working with a Model A Ford," said Brentjens. "We as clinician scientists don't want to generate false hope. The technology works as a proof of principle, ALL is okay, but we need to more fully develop the technology to treat other types of cancer."

The panelists and key opinion leaders at the Summit said that the first order of business should be figuring out where cell-based therapies fit into current treatment regimens.

"Once safety profiles have improved, T cell therapeutics might be used earlier rather than later in treatment to avoid the accumulation of multiple mechanisms of resistance by tumor cells after several relapses," said Gianpietro Dotti, an associate professor of medicine at Baylor.

"It will most likely be cancer specific and dependent on what options are available to the patient. For instance, for kids with leukemia there is a 90% success rate using current chemotherapy regimens," he continued. "That would not be the instance to move the treatment up unless we clearly demonstrate a therapeutic advantage, reduced toxicities and costs compared to conventional treatments."

Regardless of when T cell therapies get used, Campana said that such treatments need to be part of a physician's thought process "before cancer patients start their first day of cancer treatment. Right now autologous T cells are the adoptive cell of choice, and sometimes it is a challenge to obtain functional T cells from people with cancer that are undergoing treatment. T cells could be banked before any type of treatment begins."

Once CAR-based T cells become more prominent as a treatment option, T cell manipulation and manufacturing is going to have to be streamlined and more automated.

Novartis is arguably the furthest along in creating manufacturing infrastructure for CAR T cells. In 2012, the pharma bought Dendreon Corp.'s Morris Plains, N.J., immunotherapy manufacturing facility.

Bruce Levine, an associate professor in cancer gene therapy and director of the Clinical Cell and Vaccine Production Facility at UPenn, is training Novartis personnel at the Morris Plains site to engineer CAR T cells.

"Penn will conduct early phase clinical trials not only targeting CD19 but additional tumor targets in solid tumors. Novartis will conduct later-phase and pivotal clinical trials at its Morris Plains facility, and then, following FDA approval, commercial production will take place at this facility as well," said Levine.

Novartis also agreed to provide $20 million to establish the Center for Advanced Cellular Therapies at UPenn to co-develop CAR-based therapies. The effort is being led by Carl June.17

Borrowed T cells

In addition, the panel recommended further exploration of allogeneic cells as a platform for making CAR- or TCR-based T cell therapies more readily available as off-the-shelf solutions.

Researchers had first focused on autologous T cells because they wanted to avoid the possible complications of the body attacking the T cells (host-versus-graft disease) or of the T cells attacking host tissues (GvHD) that could accompany the nonself allogeneic T cells.

If a balance can be found that allows allogeneic T cells to proliferate in the body without attacking the host, off-the-shelf CAR T cells could be a possibility. One donor could provide T cells that could be used to engineer CAR-based T cells for multiple recipients without any manufacturing lag time. Off the shelf is considered the Holy Grail for cell-based therapies.

"With low-dose irradiation, patients are already prepped for allogeneic T cell transfer," noted Peshwa. "Indeed, allogeneic T cells have been shown to eventually be rejected but no graft-versus-host disease develops."

Brenner has first-hand experience setting up clinical trials using allogeneic T cells. "We've had some good experience in using allogeneic T cells with donor CD19 T cells and have seen no problem with graft-versus-host disease. These types of therapies aren't that far away from reality. Maybe in the next year or so," he said. "I always thought that allogeneic, CAR-based T cells wouldn't work, but now the data is showing it might be a possibility."

Fighting the IP fight

Even as scientific and clinical progress continue, it still remains murky who owns the IP to the various components or procedures needed to create engineered, CAR-based T cells.

While Novartis and Juno are jostling over who owns the rights to CAR T cells that target CD19, MSKCC is moving forward with a clinical study of CARs in collaboration with UPenn. The study is designed to see which of the two centers has a superior CAR.

"Cross-licensing and sleeper patents are some intricacies that are going to have to be worked out," said Booth. "Right now researchers are working with reagents or methodologies that they might not be able to get access to, but they are proceeding with a pragmatic wait-and-see mentality."

Binder-Scholl agreed. "IP is limiting in the field," she said. "The engineered T cell therapy can be considered a process, but there will be IP for the target, IP for the receptor placed on the cells, IP for the technology to produce the T cells and IP for the technology to expand the T cells."

Adaptimmune has chosen to carve out its niche in TCRs. In June, the company partnered with GlaxoSmithKline plc to co-develop Adaptimmune's lead immunotherapy-a TCR-based T cell targeting cancer/testis antigen 1B (CTAG1B; NY-ESO-1).

"IP is incredibly complicated and far from resolved. Basically, lawyers are going to have to put on the armor and get ready to fight it out in a really complicated IP situation," said Booth. "They are going to be making a lot of money from IP battles."

"I'd like to believe that CAR therapy competitors will be rational and cross-license relevant IP rather than fight it out, but that's not been the history of biotech," noted Booth. "Even in the antibody space, smart cross-licensing only happened after years of fighting and stacks of lawyer bills."

 "But this is going to be what is necessary to advance these types of therapies," said Booth. "We're not talking pills for the rest of your life- we're talking about a bona fide cure."

Baas, T. SciBX 7(25); doi:10.1038/scibx.2014.725Published online June 26, 2014


1.   Kochenderfer, J.N. et al. Blood 116, 4099-4102 (2010)

2.   Kalos, M. et al. Sci. Transl. Med. 3, 95ra73 (2011)

3.   Porter, D.L. et al. N. Engl. J. Med. 365, 725-733 (2011)

4.   Brentjens, R.J. et al. Blood 118, 4817-4828 (2011)

5.   Grupp, S.A. et al. T cells engineered with a chimeric antigen receptor (CAR) targeting CD19 (CTL019) produce significant in vivo proliferation, complete responses and long-term persistence without GvHD in children and adults with relapsed, refractory ALL. Presented at the American Society of Hematology annual meeting, Dec. 7-10, 2013.

6.   Kalos, M. et al. Long-term functional persistence, B cell aplasia and anti-leukemia efficacy in refractory B cell malignancies following T cell immunotherapy using CAR-redirected T cells targeting CD19. Presented at the American Society of Hematology annual meeting, Dec. 7-10, 2013.

7.   Louis, C.U. et al. Blood, 118, 6050-6056 (2011)

8.   Grupp, S.A. et al. N. Engl. J. Med. 368, 1509-1518 (2013)

9.   Kudo, K. et al. Cancer Res. 74, 93-103 (2014)

10. Fedorov, V.D. et al. Sci. Transl. Med. 5, 215ra172 (2013)

11. Kloss, C.C. et al. Nat. Biotechnol. 31, 71-75 (2013)

12. Lanitis, E. et al. Cancer Immunol. Res. 1, 43-53 (2013)

13. Di Stasi, A. et al. N. Engl. J. Med. 365, 1673-1683 (2011)

14. Pardoll, D.M. Nat. Rev. Cancer 12, 252-264 (2012)

15. Devaud, C. et al. Oncoimmunology 2, e25961 (2013)

16. Prosser, M.E. et al. Mol. Immunol. 51, 263-272 (2012)

17. McCallister, E. BioCentury 20(33), A7-A8; Aug. 13, 2012


University of Pennsylvania, Philadelphia, Pa.

Adaptimmune Ltd., Abingdon, U.K.

Atlas Venture, Cambridge, Mass.

Baylor College of Medicine, Houston, Texas

Bellicum Pharmaceuticals Inc., Houston, Texas

bluebird bio Inc. (NASDAQ:BLUE), Cambridge, Mass.

Celgene Corp. (NASDAQ:CELG), Summit, N.J.

Cellectis S.A. (Euronext:ALCLS), Paris, France

The Children's Hospital of Philadelphia, Philadelphia, Pa.

Dendreon Corp. (NASDAQ:DNDN), Seattle, Wash.

Fred Hutchinson Cancer Research Center, Seattle, Wash.

GlaxoSmithKline plc (LSE:GSK; NYSE:GSK), London, U.K.

Juno Therapeutics Inc., Seattle, Wash.

Kite Pharma Inc. (NASDAQ:KITE), Los Angeles, Calif.

MaxCyte Inc., Gaithersburg, Md.

Memorial Sloan-Kettering Cancer Center, New York, N.Y.

National Cancer Institute, Bethesda, Md.

National University of Singapore, Singapore

Novartis AG (NYSE:NVS; SIX:NOVN), Basel Switzerland

Novartis Institutes for BioMedical Research, Cambridge, Mass.

OrbiMed Advisors LLC, New York, N.Y.

Perelman School of Medicine at the University of Pennsylvania, Philadelphia, Pa.

Pfizer Inc. (NYSE:PFE), New York, N.Y.

Seattle Children's Hospital, Seattle, Wash.

St. Jude Children's Research Hospital, Memphis, Tenn.

University of Pennsylvania, Philadelphia, Pa.

University of Washington, Seattle, Wash.