The field of chimeric antigen receptor (CAR) T-cell therapy, a pillar of modern immunotherapy, is undergoing significant evolution. While ex vivo CAR T therapies—where a patient’s T cells are extracted, genetically engineered in a lab, and reinfused—have achieved remarkable success against certain blood cancers, they face considerable challenges. These include complex, costly, and time-consuming manufacturing processes, significant treatment-related toxicities like cytokine release syndrome (CRS), and limited accessibility for patients. In response, the next wave of innovation is surging toward in vivo CAR T therapy, where the genetic reprogramming of T cells occurs directly inside the patient’s body.
By potentially simplifying administration to an intravenous infusion akin to a monoclonal antibody, in vivo platforms could reduce manufacturing complexity, lower costs, shorten the time from decision to treatment, and expand access beyond major academic centers. The race is on to develop safe and effective delivery vectors, including lentiviruses, adeno-associated viruses (AAVs), and lipid nanoparticles (LNPs), each with distinct advantages for targeting, durability, and scalability.
To explore this rapidly advancing landscape, we spoke with Dr. Luke Russell, Executive Vice President at Vyriad, a clinical-stage oncology company developing in vivo CAR T development. Vyriad’s platform leverages its deep virology expertise to create precisely targeted viral vectors. Dr. Russell delves into the technical rationale for in vivo approaches, explains Vyriad’s unique platform safeguards, outlines the path to the clinic, and discusses the broader future of the field.
This interview has been edited for clarity, consistency, and length.
Dr Phalguni Deswal [PD]: The CAR T field is accelerating rapidly. Can you discuss the benefits of in vivo CAR T, especially it being less time-consuming and logistically simpler, and why is the field moving more toward in vivo approaches?
Dr. Luke Russell: The main reasons are indeed the ones you listed, but a maybe less-discussed aspect is that the generation of those CAR T cells in situ is going to be fundamentally different from what’s possible ex vivo.
The ex vivo T-cell manufacturing process includes all these activation steps that don’t have to happen with an in vivo delivered product. In the ex vivo paradigm, you pull cells out, sort the T cells, and then activate them with CD3/CD28 beads. These activated cells then upregulate the receptor for the VSV-G glycoprotein (Vesicular Stomatitis Virus glycoprotein G) used by lentiviral vectors. You transduce the T-cells, expand them, and what you are delivering back to the immuno-depleted patient is a population of hyperactivated T cells—an artificial process that would never normally happen in the body. It works very well and has proven the point that a CAR on a T cell can be curative for some patients.
But generating that same CAR T cell in vivo is wildly different. You are not immunodepleting the patient, so you are not creating an empty niche, but instead activating the already present immune cells upon entry. Our vector is decorated with an anti-CD3 molecule, providing “signal 1” upon binding. The T cells then express the CAR protein, and when the CAR binds its target, it provides “signal 2” from the intracellular domain of the CAR. The T cells then secrete cytokines like interferon gamma, providing “signal 3.” It’s a much more natural activation cascade. I think we’ll see a fundamentally different type of CAR T cell in circulation compared to what’s created ex vivo. It is a profound distinction that still needs to be assessed in clinical trials to understand the full meaning.
PD: Vyriad’s platform retargets viruses exclusively to T cells without impacting transduction efficiency. How is this critical for safety and efficacy, specifically regarding off-target effects?
Dr. Luke Russell: That’s a really important aspect of the design of these novel vectors. When you are delivering a virus into the bloodstream and asking it to seek out only T cells, specificity is paramount. If you accidentally transduce a tumor cell, you now have a tumor cell expressing both your target (e.g., BCMA) and an anti-target single-chain antibody (the CAR) on its surface. That single chain could bind to BCMA (B-cell maturation antigen) and hide it. The risk is creating a tumor subpopulation that no longer responds to the very CAR T therapy you just administered. So, solving targeting was essential for in vivo CAR T to work.
We have addressed this in a couple of ways with our lead clinical asset, VV169. We start with a lentivirus particle with an engineered VSV-G glycoprotein on its surface. We have mutated a residue that binds to the LDL receptor (the natural, ubiquitous receptor for VSV-G), effectively “detargeting” the virus from most cells. Then, on that same mutated protein, we express an anti-CD3 single-chain fragment, which redirects entry exclusively to CD3-positive T cells.
But there’s a second layer. We must manufacture this virus in a cell substrate. If the CAR payload itself were expressed on the surface of our manufacturing cells, it could be accidentally incorporated into the virus particle, potentially redirecting it to tumor cells. To avoid this, Vyriad engineered a reverse orientation, T-cell-selective promoter to drive CAR expression. This promoter is inactive in our manufacturing cell line, so no CAR protein is expressed on the surface of our viral particles. This dual strategy—detargeting from broad receptors and actively retargeting only to T-cells, while ensuring the CAR isn’t on the particle itself—gives us very specific entry.
PD: Regarding delivery mechanisms, why do you think a lentiviral approach works better compared to lipid nanoparticles (LNPs) / messenger RNA (mRNA) or other viral approaches like adeno-associated viruses (AAV)?
Dr. Luke Russell: I can’t say definitively one approach is better than all others yet, as they haven’t all been clinically validated. But there are strong theoretical reasons for lentivirus. We understand it well, and it provides durable expression through genomic integration.
In oncology, where the goal is to eliminate every single tumor cell, durable expression makes sense. The CAR is integrated into the T cell, and as long as that T cell and its progeny live, they express the CAR. This differs greatly from LNP/mRNA approaches, which lead to transient expression. mRNA is short-lived, and as a successfully transduced T cell expands in response to the tumor, the mRNA doesn’t replicate. The CAR expression dilutes with each cell division, potentially reducing efficacy over time and necessitating complex redosing schedules where you’re trying to hit an ever-expanding pool of T cells.
That dilution problem doesn’t exist with an integrating vector. Daughter cells maintain the same level of CAR expression. It just seems to theoretically align well with indications requiring long-term, durable elimination of every target cell.
PD: Vyriad has compelling preclinical data in disseminated multiple myeloma. Looking ahead to clinical trials, what would you be most keen to monitor in terms of biomarkers, efficacy, or safety readouts?
Dr. Luke Russell: Our goal is to treat patients this year with our in vivo CAR T therapy, and we have had positive pre-IND feedback from the US Food and Drug Administration (FDA). We are aiming for a first-in-human trial around mid-2026 at Mayo Clinic in Rochester, with Dr. Yi Lin, who heads Mayo’s CAR T program, as the lead principal investigator (PI).
At its core, it’s a Phase I safety study, so primary endpoints are safety-related. However, because BCMA is such a powerful target, we will definitely explore preliminary clinical responses. We will look at standard multiple myeloma endpoints: minimal residual disease (MRD) negativity, imaging of tumor deposits, and other key biomarkers.
From a drug development perspective, it’s also crucial to understand the pharmacokinetics and pharmacodynamics of this “living therapy.” We need to know: How long does the virus survive in circulation? How many T cells are transduced? How do they expand in response to each patient’s unique tumor burden? Furthermore, we are well aware of toxicities like cytokine release syndrome (CRS) seen with traditional CAR T. We’ll closely monitor the cytokine profile in patients, both in response to the virus infusion and as the CAR T cells expand and engage the tumor.
PD: A major promise of in vivo CAR T is reduced cost and complexity. How do you see Vyriad’s therapy fitting into the treatment landscape in terms of accessibility compared to monoclonal antibodies or high-cost ex vivo cell therapies?
Dr. Luke Russell: Cost is one of many potential benefits, alongside accessibility. Long-term, our goal is to pass along as many of these benefits as possible to the patient. If it costs less to manufacture, the patient should experience that. If it becomes more accessible, potentially moving from a complex inpatient procedure to a simpler outpatient infusion—that’s a major gain.
There are still things to understand, like the exact dosing and whether it can be fully outpatient. But ultimately, it should be less costly to manufacture than a traditional ex vivo CAR T therapy, which requires a bespoke, multi-week cell manufacturing process. We hope that economic and logistical benefits translate directly to the healthcare system and patients.
PD: Looking broadly at the field, what do you think the landscape for in vivo CAR T therapy will look like in a few years?
Dr. Luke Russell: I think for the next few years, we’ll see many parallel clinical trials exploring different delivery mechanisms and targets. It will take time to understand which is most efficacious. So far, clinically, BCMA has emerged as a strong beachhead for the field. Early data from a few companies have shown that you can not only deliver the construct in vivo but also generate potent CAR T cells against multiple myeloma. That early validation helps carry the whole field forward.
The platforms that succeed will be those that combine the best targeting, the least toxicity, and the best scalability. A great drug that cannot be manufactured at a massive scale won’t reach all the patients who need it. Scalability, specificity, and safety management will be the deciding factors in who leads this field. It’s an exciting time, and we look forward to seeing what becomes possible for patients.
