car.m: Breakthrough in Macrophage Cancer Therapy

Chimeric antigen receptor macrophage therapy, known as CAR-M, is an emerging form of cancer immunotherapy that engineers the body’s own immune scavenger cells to recognize, engulf, and destroy tumor cells. While CAR-T cell therapy has transformed treatment for blood cancers over the past decade, it has largely failed against solid tumors, the cancers that kill the most people. CAR-M therapy is designed specifically to address that gap, and the first human clinical trial, reported in 2025, showed the approach is safe and can reach tumors after intravenous infusion. The field is still early, but the science behind it represents a genuine shift in how researchers think about engineering immune cells for cancer.

Why Solid Tumors Have Resisted Engineered Cell Therapy

CAR-T cells have produced dramatic remissions in leukemia and lymphoma, but translating that success to solid tumors like breast, lung, or pancreatic cancer has been frustrating. Solid tumors throw up a dense, hostile environment that stops T cells from doing their job. The barriers are physical and biochemical: abnormal blood vessels make it hard for infused cells to reach the tumor, a thick web of structural proteins surrounds the cancer cells like scaffolding, and the tumor itself secretes signals that exhaust or shut down T cells that manage to arrive.¹ There is also the problem of antigen heterogeneity: solid tumors are a patchwork of cells carrying different surface markers, making it difficult to find a single target that covers the entire mass.²

Adding insult to injury, the tumor microenvironment actively recruits immune cells and then corrupts them. Macrophages are among the most abundant immune cells inside solid tumors, but the tumor reprograms them from their normal inflammation-promoting state into a pro-tumor state that helps cancer grow, spread, and hide from immune surveillance.³ These corrupted cells, called tumor-associated macrophages, promote blood vessel formation to feed the tumor, suppress other immune cells, and even help cancer metastasize.⁴ The insight behind CAR-M therapy is that macrophages’ natural ability to infiltrate tumors and shape their environment could be turned from a liability into a weapon.

What Makes Macrophages Attractive as an Engineered Weapon

Macrophages are the immune system’s professional cleanup crew. They patrol tissues, swallow dead cells and debris, and sound the alarm when they detect threats. Several of their built-in abilities make them well suited for anti-cancer engineering. They naturally migrate to tumors, something T cells struggle to do in solid tissue. Once there, they can physically engulf and digest cancer cells, a process called phagocytosis. And they serve as antigen presenters, chewing up tumor proteins and displaying fragments on their surface so T cells can learn what to attack.⁵

This combination of direct killing, immune activation, and natural tumor homing gives macrophages a multi-pronged toolkit that T cells lack. A comparison published in Biomolecules and Biomedicine noted that even at its earliest stage, CAR-macrophage technology shows potential advantages over CAR-T for solid tumors, including more diverse anti-tumor mechanisms and better infiltration into tumor tissue.⁶

Engineering the Macrophage

The basic idea behind a CAR-macrophage is the same as a CAR-T cell: you bolt a synthetic receptor onto the cell’s surface that recognizes a specific protein on cancer cells. When the receptor locks onto its target, it triggers the cell to attack. But because macrophages kill differently from T cells, the internal machinery of the receptor has to be redesigned from scratch.

Early proof-of-concept work showed this was feasible. Researchers screened a panel of intracellular signaling domains from natural engulfment receptors and found that domains from Megf10 and FcRγ could reliably trigger macrophages to swallow target cells, including cancer cells, when connected to an extracellular antibody fragment.⁷ These first-generation CAR-macrophages demonstrated the principle but had limited potency.

Second-generation designs have been more ambitious. One team incorporated the signaling domain from Toll-like receptor 4, a component of innate immune sensing, into the CAR architecture. The result was a macrophage that not only engulfed target cells more aggressively but also locked itself into a pro-inflammatory anti-tumor state and resisted being reprogrammed by the tumor into a pro-tumor state. A tandem design combining this domain with a T-cell signaling element gave macrophages the ability to remodel the hostile tumor microenvironment around them.⁸ Other groups have focused on enhancing the receptor’s activation strength by engineering integrin-mediated signaling components linked to Fc-gamma receptor I, one of the most potent phagocytic triggers in the human immune system.⁹

The First Human Trial

The milestone that moved CAR-M from theoretical to clinical came with CT-0508, an anti-HER2 CAR-macrophage developed by Carisma Therapeutics and tested in a phase 1 trial in patients with advanced solid tumors that overexpress HER2. The trial enrolled patients with a range of cancers, and the results, published in Nature Medicine in early 2025, were encouraging on the safety front: no dose-limiting toxicities occurred, no cases of severe cytokine release syndrome (grade 3 or higher), and no immune effector cell-associated neurotoxicity, two of the most feared side effects of CAR-T therapy.¹⁰

In terms of anti-tumor activity, the results were modest but meaningful for a first-in-human test. Among the nine patients with the highest level of HER2 expression (HER2 3+), about 44% achieved stable disease as their best response eight weeks after treatment. No meaningful activity was seen in the five patients whose tumors had lower HER2 expression (HER2 2+). Perhaps more telling than the tumor measurements were the biopsy findings: serial tissue samples confirmed that the infused CAR-macrophages reached the tumor, remodeled the local immune environment, and triggered expansion of cancer-killing T cells within the tumor.¹¹ That last point matters a lot. It suggests CAR-M therapy might work not just by eating cancer cells directly but by waking up the patient’s own adaptive immune system to join the fight.

More Than Just Eating Cancer Cells

One of the most exciting aspects of CAR-M is that the macrophage’s contribution goes well beyond swallowing individual tumor cells. Macrophages are master remodelers, and CAR engineering amplifies that trait.

When a macrophage engulfs a tumor cell, it processes the cell’s proteins and presents fragments on its surface. This functions like a wanted poster for T cells: it teaches them what the cancer looks like so they can mount a targeted attack. The phase 1 clinical data confirmed this mechanism in humans. After CT-0508 infusion, biopsies showed increases in antigen presentation gene expression, a heightened tumor inflammation signature, and greater T cell activation and clonality at the tumor site.¹²

CAR-macrophages can also be engineered to physically dismantle the structural barriers around tumors. One design uses the intracellular region of CD147 to activate enzymes called matrix metalloproteinases, which break down the dense extracellular matrix that acts as armor around many solid tumors. In preclinical testing, this approach roughly quadrupled the infiltration of T cells into the tumor compared to conditions where the matrix remained intact.¹³ If that translates to humans, it would mean CAR-M therapy could break open the door for other immune cells or even other therapies to get in.

Disabling the “Don’t Eat Me” Shield

Cancer cells have a trick that specifically defends them against macrophages. Many tumors plaster a protein called CD47 on their surface, which binds to a receptor called SIRPα on macrophages and sends a signal that says, essentially, “don’t eat me.” This checkpoint is one of the main reasons macrophages tolerate tumors even when they are sitting right next to them.

Researchers are working this into CAR-M design. One proposed approach combines a CAR targeting a gastric cancer antigen with built-in silencing of the SIRPα receptor, so the macrophage literally cannot hear the “don’t eat me” signal. The same design also overexpresses the interferon-gamma receptor to boost pro-inflammatory signaling and lock the macrophage into its anti-tumor state.¹⁴ This kind of multi-layered engineering is where the field is heading: macrophages that carry not just a targeting receptor but an entire suite of modifications to overcome the tumor’s defensive strategies.

Pairing CAR-M with Checkpoint Inhibitors

Since CAR-macrophages activate T cells as part of their mechanism, combining them with drugs that remove the brakes from T cells is a logical next step. Checkpoint inhibitors like anti-PD-L1 antibodies are already widely used in oncology, and early preclinical data suggest the pairing works well.

In a prostate cancer model, researchers tested CAR-macrophages engineered to express the costimulatory molecule CD86, which helps activate T cells, alongside anti-PD-L1 treatment. The combination produced the strongest tumor suppression of any group tested, outperforming either therapy alone. Importantly, the combination did not cause additional toxicity, with no effect on body weight or organ weight in treated animals.¹⁵ A separate study in hepatocellular carcinoma similarly found that combining CAR-macrophages with PD-L1 blockade accelerated tumor clearance and improved survival compared to either treatment alone, with the benefit driven by enhanced T cell killing.¹⁶ These results reinforce the idea that CAR-M’s real power may lie in its ability to convert immunologically “cold” tumors, those that T cells ignore, into “hot” ones where immunotherapy can gain traction.

The Safety Question

CAR-T therapy’s most dangerous side effect is cytokine release syndrome, a potentially life-threatening inflammatory storm. Ironically, research has shown that myeloid cells, the family macrophages belong to, are the major mediators of cytokine release syndrome during CAR-T therapy, primarily through release of IL-1 and IL-6.¹⁷ This raised legitimate concerns that engineering macrophages to be more activated could make inflammation worse, not better.

The phase 1 trial of CT-0508 offered reassuring early data on this front. No severe cytokine release syndrome occurred in any patient at any dose level tested.¹⁰ The reasons for this favorable profile are not entirely clear, but one likely factor is scale: CAR-M therapy currently involves far fewer cells than a typical CAR-T infusion, and macrophages do not expand in the body the way T cells do. Whether the safety profile holds at higher doses or in combination therapies remains an open question.

Where CAR-M Cells Go After Infusion

Getting engineered cells to the right place is half the battle. When CAR-macrophages are delivered intravenously, a large fraction gets trapped in the lungs, liver, and kidneys before ever reaching the tumor.¹⁸ Clinical pharmacokinetic data from the CT-0508 trial showed that the macrophages accumulated primarily in the lungs within the first 24 hours, then redistributed to the liver and spleen, with detectable levels in peripheral blood lasting a median of roughly one to two weeks.¹⁸ That relatively short persistence window is a double-edged sword: it limits the duration of potential toxicity, but it also means the therapeutic window may be narrow.

Alternative delivery routes are being explored. Intraperitoneal injection, for example, greatly reduces off-target sequestration in distant organs and increases tumor-specific delivery for abdominal cancers. For brain tumors or other localized cancers, direct intratumoral injection is another possibility. The best route will likely depend on the tumor type and its location.

Manufacturing at Scale

One of the biggest practical hurdles for CAR-M therapy is manufacturing. Macrophages do not multiply the way T cells do once infused, so you need to produce large numbers of functional cells before treatment. Drawing enough macrophages from a patient’s own blood to reach therapeutic doses is difficult. The clinical translation of CAR-M remains challenging in part because of restricted cell expansion, engineering complexity, and batch-to-batch variability.¹⁹

A promising workaround is deriving macrophages from induced pluripotent stem cells. These stem cells can be engineered with the CAR gene first and then differentiated into macrophages, providing a theoretically unlimited supply of consistent, off-the-shelf product. One group demonstrated scalable production using an automated 3D bioreactor system, reaching a stable weekly output of roughly 57 million CAR-macrophages per 40-milliliter batch, with consistent quality and functionality from harvest to harvest.²⁰ Another group showed that stem cell-derived CAR-macrophages could be coaxed into the anti-tumor state and demonstrated anti-cancer activity in animal models, providing a foundation for an “off the shelf” product that would not require using each patient’s own cells.²¹

Skipping the Factory Entirely

Perhaps the most radical idea in the field is generating CAR-macrophages inside the patient’s body without removing any cells at all. Two delivery strategies are converging to make this plausible.

The first uses lipid nanoparticles, the same basic technology behind mRNA COVID vaccines, to deliver CAR-encoding mRNA directly into cells. Researchers have shown that lipid nanoparticle-delivered mRNA can successfully produce functional CAR-macrophages in the lab, with the resulting cells showing significant killing of lymphoma cells.²² Because mRNA is temporary, the CAR expression fades over days, which could be a safety advantage but limits the duration of the effect.

The second, more ambitious approach aims to deliver these nanoparticles directly into the tumor, converting the tumor’s own macrophages into CAR-armed killers on site. A recent study used lipid nanoparticles designed to selectively transfect tumor-associated macrophages, co-delivering CAR mRNA along with a STING agonist, an immune-activating compound, to amplify the response. In a mouse melanoma model, this in situ approach produced enhanced anti-tumor effects without any ex vivo cell manufacturing at all.²³ If this kind of approach matures, it could eventually turn CAR-M from an expensive, logistically complex cell therapy into something closer to an injectable drug.

What Remains Unclear

For all the progress, CAR-M therapy is still in the very early stages of clinical development. The first human trial enrolled only 14 patients and was designed primarily to test safety, not effectiveness. Whether CAR-macrophages can produce tumor shrinkage or durable remissions in larger trials is unknown. The short persistence of infused macrophages in the body raises questions about whether repeated dosing will be needed and whether the immune system will reject repeat infusions, particularly with off-the-shelf products derived from donor stem cells.

There are also open biological questions. Macrophages are remarkably plastic cells, and even engineered versions could be reprogrammed by the tumor over time. The second-generation designs with built-in M2 resistance are promising, but no one has tested whether that resistance holds up over weeks or months in a human tumor. The field also lacks standardized potency assays: how do you measure whether a batch of CAR-macrophages is “good enough” before infusing it into a patient? That kind of quality control infrastructure still needs to be built. For now, the evidence supports cautious optimism: the biology is sound, the first safety data are reassuring, and the combination strategies look additive. How far CAR-M can go will depend on whether the clinical results in larger trials match the preclinical promise.