Completing the Cell Therapy Revolution

In 2017, the pharmaceutical industry erupted in celebration as the FDA approved the first two CAR T-cell therapies, Yescarta and Kymriah. Until then, the prospect of extracting a patient's cells, modifying them to express chimeric antigen receptors on the surface, and reinfusing them into the patient to latch onto specific antigens to kill tumors had seemed like science fiction to many. But FDA approvals answered the doubters: CAR T works.

However, other questions remained unanswered. Ideal drug manufacturing and logistic processes are closed and automated to eliminate the risks associated with human intervention and manual operations – but this is not the case with autologous CAR T. So how would companies handle living, breathing cells in transit? Would healthcare systems be able to cope? Pricing too was a concern. Would all stakeholders embrace evidence-based pricing?

Though these questions are yet to be fully resolved, we are seeing a conversational shift back to where it all began: scientific efficacy. We know cell therapy works in liquid tumors (leukemia and lymphoma), but what about solid tumors, which represent approximately 90 percent of adult human cancers and, therefore, a huge area of unmet need. In short, what’s the hold up?

“We all thought solid tumors might be a little bit harder – but how hard could it really be?” asks Bruce Levine, Barbara and Edward Netter Professor in Cancer Gene Therapy at the University of Pennsylvania, and President of the International Society for Cell & Gene Therapy (ISCT). “Quite a lot harder, it turns out.”

The central challenge is antigen specificity. The first CAR T-cell therapies were approved for beta cell-malignancies, which have easily identifiable surface markers, such as CD-19 or BCMA. An anti-CD-19 CAR T-cell therapy may wipe out most of a patient’s B-cells in addition to their cancer, though this isn’t a major problem. “But if you found a target that was unique to lung tissue, for example, you couldn’t easily treat it with a T cell therapy because you’d run the risk of also seriously damaging the patient’s lungs,” says Elliot Norry, Chief Medical Officer at Adaptimmune. 

Some targets, such as EGFRviii, are tumor specific – so attacking these does not risk wiping out the patient’s organs. However, they’re only present in about a third of glioblastomas. Finding a target that is both tumor specific and homogeneously expressed has vexed developers looking to target solid tumors. The first blood cancer cell therapies were far less challenging. 

“You need to be looking at multiple targets,” says Levine. “But you also need to titrate those targets.” He raises the example of mesothelin, which is expressed in pancreatic adenocarcinomas, mesotheliomas, ovarian cancers, and about half of lung cancers – plus others. The catch is that mesothelin also exists at lower levels in the pleural cavity, which means any potential cell therapy targeting it could be destructive to a certain degree if not titered or controlled. 

Another hurdle is the highly immunosuppressive solid tumor microenvironment, which includes the expression of checkpoint ligands, the secretion of immunosuppressive mediators like TGF-beta, the presence of regulatory T cells, and myeloid-derived suppressor cells – all of which conspire to prevent the immune system from detecting and killing the tumor. To make matters worse, solid tumors aren’t well vascularized; the stroma is tightly packed and resistant to penetration by immune cells because of a matrix of cancer-associated fibroblasts.

But Levine thinks we have strategies to combat each of these problems. “It’s going to take a combination of strategies and targets, including synthetic biology. The route of administration may be important too,” he argues. “I’m optimistic because we do see evidence of clinical activity in both preclinical models and some early clinical trials.” 

“Solid tumors are the field’s holy grail right now,” says Tony Ting, Chief Scientific Officer at Bone Therapeutics. “This is something people have been focused on for quite some time, but we’re all hopeful of strong clinical results in the near future.” 

Signs of scientific efficacy 

So how might developers go after multiple targets? Levine cites a paper by Anna Wing, Carl June, and colleagues from 2018 (1); their approach targets two antigens at once using both CAR T-cells and an oncolytic virus-driven bispecific antibody. “It’s one of my favourite papers,” says Levine, who worked with Carl June on developing the first CAR T-cell therapies. “Essentially, you get three for one; you have the two antigens targeted as well as the antigens released by the oncolytic vector.”

With regard to synthetic biology, Levine highlights the integration of switch receptors. “This involves turning a negative signal into a positive,” he explains. “You can make a switch receptor with PD-1, extracellularly, and then a signal-transducing co-stimulatory signal like CD-28. So when the tumor delivers a negative signal, the engineered T-cell sees it as a positive signal. That’s really clever.” 

In April, the University of California San Francisco published two papers on their “SynNotch” system. In the first paper, they found that SynNotch-CAR-T cells could completely clear human patient-derived tumors from the brains of mice – safely and without recurrence (2). In a second paper, another set of researchers showed how components of the system can be switched out to target other cancers, such as ovarian and lung (3). 

The new approach has two steps. The first step uses SynNotch to grant CAR-Ts the ability to “judge” whether they are in a tumor. The second step uses a different set of SynNotch sensors to ensure a strong tumor-killing response. “Our approach allows us to prime the expression of the CAR against broad tumor antigens only in conditions where the T cells see tumor-specific or brain-specific signals,” says Hideho Okada, co-author of the first paper. “As such, the SynNotch-T cells are safer and more effective.” 

Okada and his team are actively working on moving into the clinic. “We’re also developing brain-specific priming,” he says. “In the paper, we described priming by MOG, but there may be other brain-specific antigens that may work as well.” 

Levine is also excited about local administration of CAR T-cell therapy. His team at the University of Pennsylvania are locally administering mesothelin-targeted CAR Ts to tumors. “MD Anderson and Sloan Kettering are also looking into this approach,” he says, moving on to describe how City of Hope researchers have also incorporated local administration into the CNS. “That’s technically challenging, but they did see some evidence of clinical efficacy.” They’ve also used lentiviral transfer of CAR, targeting mesothelin. “We saw clinical activity in one-out-of-five pancreatic cancer patients using that approach,” says Levine. UPenn and City of Hope have also targeted EGFRviii in glioblastoma. In the University of Pennsylvania clinical trial, investigators saw tumor necrosis and downregulation of the target in patient tumor tissue.

Another promising area is macrophage-based cell therapy. In 2020, University of Pennsylvania researchers used genetically engineered macrophages to kill solid tumors in both mouse models and human samples (4). Then, in March 2021, Carisma Therapeutics – a company founded by researchers at the University of Pennsylvania – announced that it had dosed its first human participant in a phase I clinical study assessing the safety of CAR macrophages (5).

“Engineered macrophages may be particularly suited to the very challenging microenvironment of solid tumors,” says Levine. In a review of recent developments in CAR-macrophage-based treatments for solid tumors from Anhui Medical University, China, researchers cited “great potential” when it came to migration to tumor and recruitment of immune effector cells (6). 

However, a central challenge with engineered cell therapy is the potential for toxicity and cytokine release syndrome. Tmunity recently suffered a serious setback after the company was forced to shut down and modify their lead program for prostate cancer after two patients died following CAR T-cell therapy. The researchers had taken PSMA-specific and TGFβ-resistant CAR-modified autologous T cells into an 18-subject phase I prostate cancer trial in 2017. Tmunity then began a second, larger study late in 2019. President and CEO Oz Azam and co-founder Carl June explained that they were initially shocked at how well the therapy was performing in a recent interview with Endpoint News (7). But the two deaths in the small study forced a rethink. 

“What we are discovering is that the cytokine profiles we see in solid tumors are completely different from hematologic cancers,” said Azam, during the interview. “We observed immune effector cell-associated neurotoxicity – ICANS. And we had two patient deaths as a result of that.” 

“We didn’t see this coming until it happened,” said June. “But I think we’ll engineer around just like we did with tocilizumab back in 2012.”

“We’ve been lulled into a false sense of security by the rapid progress with blood cancers,” says Levine. “But with solid tumors, while we’re making progress – we have more centers working on the problem, as well as new tools and technologies – we’re going to need long attention spans.” 

The CAR alternatives 

Another set of promising non-CAR-based approaches to the development of solid tumor therapy involves T-cell receptors (TCRs). CAR technology uses an artificial receptor introduced into the immune effector cells to recognize tumor cell surface proteins (such as CD-19 or EGFRviii). In contrast, TCR-engineered effector cells use naturally occurring (or minimally modified) TCR’s that have been selected for their ability to recognize tumor-specific epitopes presented by the major histocompatibility complex (MHC) molecules on the tumor cell surface. 

“Here, you’re targeting peptide fragments from intracellular targets expressed on the cell surface in the context of HLA, which only TCRs can address,” says Norry, who has been actively researching this area alongside his colleagues at Adaptimmune. “This increases the number of potential targets and allows for greater specificity – you can more readily differentiate between cancer and healthy tissue.”

Our T-cell receptors may be recognizing malignant cells all the time and destroying them without us ever realizing. Some malignant cells avoid this protective mechanism and become tumors. By enhancing the affinity of these receptors, researchers can give TCRs the ability to recognize a tumor as foreign – and then attack it.   

In addition to enhancing the affinity of the T-cell receptor, researchers are also focused on improving the potency of T cells as a whole. “We and other groups are focusing on increasing the ability of T cells to overcome the inhibitory features of the tumor microenvironment,” says Norry. “We’ve also shown, in a laboratory setting, that we can enhance their ability to recruit the rest of the immune system once activated.”   

The ability to recognize intracellular antigen fragments presented by MHC molecules increases the number of targets available to TCR therapies; however, it also makes the therapy “MHC restricted,” which means their activity depends on presentation by MHC molecules to recognize targets and activate T cell functions. “This is a potential limitation because we all have our own MHC (or HLA) types – some are more or less common,” says Norry. “This means that a given TCR may only work in a certain sub-population.” 

Norry and his team are developing TCRs that work across various HLA types. “We’re also developing something called an HLA-independent TCR, which would expand the applicability of the therapy to a broader population.” 

Researchers from the MD Anderson Cancer Center recently reviewed the current technology and early clinical development of TCR-based therapy in patients with solid tumors, concluding that, while still early stage, TCR therapies may prove to be a “more effective option for solid tumors where intracellular antigens presented in MHC.” The researchers also thought it “plausible” that TCR therapies could be cheaper, given the “substantially lower costs” associated with the manufacturing processes. However, Levine is skeptical of the costs being substantially lower. “I’m not aware of how this would be true for TCRs and not for CARs,” he says.

“We’re very optimistic about TCRs,” says Norry. “We have a first-generation TCR in the clinic for patients with sarcoma, which we believe will become the first registered TCR-based therapy for solid tumors. We also have a next-generation TCR therapy in the clinic that incorporates a CD8-alpha cofactor, which enhances the killing capability of the product (giving it enhanced killer T-cell properties).”

In addition to TCR therapy, researchers are also interested in tumor-infiltrating lymphocyte (TIL) therapy, which involves harvesting infiltrated lymphocytes from tumors, then culturing and amplifying them in vitro, and finally infusing them back to treat patients.

“I remember listening to a talk by Steve Rosenberg about TILs in 1986,” says Ting, who also recounted how Rosenberg isolated TIL’s from multiple mouse tumor models in 1982 – the first time in history. In fact, the earliest attempt at TIL therapy in the clinic goes back to 1988, in which a 60 percent objective response rate in metastatic melanoma was achieved. “Now they’re being used to treat solid tumors in clinical trials.” 

Because TILs are composed of T cells with multiple TCR clones capable of recognizing an array of tumor antigens, a TIL-based approach may allow researchers to tackle tumor heterogeneity more easily than in CAR T and TCR T cell therapy.

A recent review of TIL therapy for solid tumors found that there have been 79 trials of TIL therapy, including 22 kinds of TIL products between 2011 and 2020 – and factoring in two successful phase II trials by Iovance in 2018 (8). The researchers highlighted “impressive clinical benefits” in metastatic melanoma and advanced cervical cancer, even in patients treated with checkpoint inhibitors, while emphasizing that “the laborious, expensive, and time-consuming tissue collection and production process” means TILs are only currently being developed at a few leading research institutions and companies in a handful of countries.

“It has become increasingly apparent that TIL therapies will have a role to play in selected indications,” says Norry. “This is why we are working with the CCIT in Denmark to develop a next-generation TIL product. We believe the ability to modify TILs with our next-gen scientific capabilities to potentially enhance efficacy has great promise.”

Allo versus auto

So far, the CAR T-cell therapies that have made it to the market have been autologous (the patient’s own cells are taken out of their body, modified so they target cancer cells, and then reinjected). But treating hundreds of thousands – or even millions – of solid tumor cancer patients using this relatively complicated, somewhat manual manufacturing process seems unlikely. One alternative is allogeneic cell therapy – an off-the-shelf alternative in which donor cells (rather than the patient’s own cells) are modified, which can reduce production time, cost, manufacturing delays, and dependence on the functional fitness of patient T cells. 

The major downside of allogeneic cell therapy is the potential for graft-versus-host disease, and host allorejection. There are, however, several approaches to overcome or at least ameliorate this difficulty, such as the generation of TCR-deficient T cells using genome editing tools such as CRISPR/Cas9. Researchers are also evaluating repeated rounds of administration, using chemotherapy-resistant CAR T-cells or genetically eliminating key molecules governing CAR T-cell immunogenicity (9).

Besides T cells, other cells are also being explored to generate allogeneic cell therapies. Most commonly this applies to NK cells because of their potent cytotoxic anti-tumor activity and favorable safety profile. NK cells tend to possess a smaller risk of inducing GVHD because (as opposed to T cells) NK cells kill independently of MHC expression – though one of the ways by which NK cells kill is by sensing the absence of self MHC. In 2020, Fate Therapeutics announced encouraging preliminary phase I data for their iPSC-derived allogeneic NK-cell therapy in advanced solid tumors – the first study in the US to evaluate an iPSC-derived cell product. Among 15 heavily pre-treated patients (nine of whom were refractory to prior therapy), 11 had a best overall response of Stable Disease (10).

So, is allogeneic the answer? 

“There are some great qualities to allogeneic therapies,” says Levine. “They can be made in advance, stored in the freezer, and ready to go within days. And there are certainly patients from whom we cannot collect or generate enough quality CAR T or even CAR-NK cells for autologous cell therapy.

“But I think it’s going to be both – I just can’t see allo-therapies ever reaching the potency of autologous therapies. For me, it’s more a question of how these therapies will evolve together – because they aren’t being developed independently of one another.” 

But Norry believes that allogeneic approaches are particularly exciting: “The product can be more consistent from patient to patient, and you have the ability to gene edit rather than using a viral vector to introduce a piece, or multiple pieces of genetic material into the cell.

“Really, all of the various iterations of TCR therapy can be made using an allogeneic platform, and we – alongside several other companies – are making good progress in the allogeneic space. Ultimately, it’s about making a real difference to the patient and I think both allogeneic and autologous approaches can do that for solid tumors.” 

In the end, the successful approach may be something totally out of the box. “There’s got to be a revolution,” says Levine. “When we’re thinking about autologous therapy: integrating automation for sure, but maybe even going beyond that and generating CAR T-cells in vivo. There are several companies – probably a dozen now – using viral vectors or nanoparticles to create CAR T-cells in the patients without having to extract, modify, and readminister.”

Recently, researchers from Nanjing University generated CAR T-cells in vivo using AAV vectors carrying the CAR gene. This “AAV delivering CAR gene therapy” (ACG) resulted in tumor regression in a mouse model of human T-cell leukemia (11).

“Just look at the disruption we’ve seen in the vaccine field with the development of mRNA lipid nanoparticles,” says Levine. “I think the in vivo approach has the potential for massive disruption, and we’ll soon see clinical data from some of these therapies.

“When one looks at solid tumors, treating hundreds of thousands of patients with the current autologous manufacturing methods wouldn’t be sustainable. I don’t know how it’s going to shake out, but I think we’ll find out by the latter end of this decade.”