QnAs with Hidde L. Ploegh

Over the past decade, immunotherapy has been slowly transforming cancer treatment. Among the treatment approaches already in the clinic are checkpoint inhibitors, which are antibodies that unshackle patients’ own tumor-targeting immune cells, and chimeric antigen receptor T cells (CAR-T cells), which are genetically engineered T cells infused into patients to attack tumors. Despite the hope they have inspired, these approaches have had no more than modest success in the clinic. By one recent estimate, only one in eight cancer patients in the United States who received Food and Drug Administration-approved checkpoint inhibitors in 2018 responded to the drugs (1). Response to CAR-T cells has been similarly sobering and so far restricted to a handful of cancer types. Part of the problem tied to the low response rates is the forbidding microenvironment around solid tumors that many therapies fail to breach. Hidde L. Ploegh, an immunologist and biochemist at Boston Children’s Hospital and a member of the National Academy of Sciences, turned to an unlikely source for a solution: camels and their close cousins. In 1993, a group of Belgian scientists chanced upon a naturally occurring form of small antibody in the blood serum of dromedary camels. These antibodies, composed exclusively of heavy chains, could be miniaturized, resulting in small (∼12 kDa) proteins with vastly improved tissue-penetrating power. Dubbed nanobodies, these miniature antibodies have found an astonishing array of research applications, such as labeling cancer cells and crystallizing challenging proteins (2, 3). Together with colleagues in Richard Hynes’ laboratory at the Massachusetts Institute of Technology’s Koch Institute for Integrative Cancer Research, Ploegh’s team engineered the nanobodies as building blocks for CAR-T cells and tested their mettle as therapeutic agents in animal models of cancer. Ploegh’s approach yielded two types of CAR-T cells: one designed to carve chinks into the tumors’ protective armor and another designed to target the cancerous cells at their core. The former, Ploegh reasoned, would cannonball into the tumors’ ramparts, and the latter would cripple the sentinels that suppress immune defense. In his Inaugural Article (4), Ploegh reports that CAR-T cells engineered in this manner beat back melanoma and colon cancer in mice. Ploegh expands on his findings.

Hidde L. Ploegh. Image courtesy of Simona Stella (Boston Children's Hospital, Boston, MA).

PNAS: First, some context. CAR-T therapy has been remarkably successful in treating blood-borne cancers in some patients, but solid tumors have proven largely refractory. Why?

Ploegh: There are many reasons why solid tumors are refractory to various forms of immunotherapy. Solid tumors are often surrounded by a dense fibrotic matrix that is difficult for lymphocytes to penetrate; by contrast, in blood-borne cancers the engineered CAR-T cells have immediate access to tumor cells in the circulation. It’s also possible that chemo-attractants that draw T cells to tumors may not be present in adequate amounts in solid tumors.

More importantly, many surface antigens found on solid tumors are also expressed on normal cells, and targeting these antigens using immunotherapy would cause collateral damage. For some B cell malignancies, you could conceivably remove all of the B cells using immunotherapy and compensate for the missing function of the B cells by infusing patients with immunoglobulins; because stem cells in the bone marrow produce new B cells, you would eventually restore the missing B cell compartment. That, of course, is not an option with solid tumors, which share many antigens with healthy tissues, which don’t necessarily self-renew.

PNAS: The approach used in your Inaugural Article (4) relies on a different type of antigen-recognition module from the standard one used in CAR-T. How did you come upon these nanobodies?

Ploegh: I have taught immunology for the better part of three decades, yet I was introduced to these nanobodies relatively late. They are a truly remarkable discovery made in 1993, but it’s only fairly recently that I became aware that camelids [camels, llamas, and alpacas] are the source of these unusual antibody fragments. Their properties continue to amaze me. We have used them in an array of applications because their small size—they are about a tenth of the size of a full antibody molecule—affords them superior tissue penetration. Their target affinities can be similar to those of conventional antibodies, and nanobodies that are not bound to the targets are rapidly cleared from the circulation. That is what gives this excellent signal-to-noise ratio and astonishingly clear images of immune cells imaged in vivo using nanobodies.

PNAS: Your approach focuses on the tumor microenvironment as a therapeutic strategy. What was the rationale for this approach?

Ploegh: For most cancers, the extent to which a tumor differs in its antigenic make-up from the cell or tissue that gave rise to it is fairly limited. (Neoantigens are an exception; melanomas carry lots of mutations, some of which generate neoantigen epitopes that can be targeted by the immune system. A similar phenomenon is seen in lung cancer among cigarette smokers; the mutagenic effects of cigarette smoke inflict DNA damage that likewise generates neoantigens).

We worked with Richard Hynes’ group at the Koch Institute. They identified nanobodies that recognize components of the tumor extracellular matrix. One component was particularly intriguing. It is a splice variant [a version of a protein produced as a result of alternative processing] of the matrix protein fibronectin that is highly expressed on tumor vasculature and stroma [matrix]. Because solid tumors require a blood supply to sustain themselves and grow, if we target newly formed blood vessels we would create an environment conducive to the delivery of a range of therapeutics, including small-molecule drugs, antibodies, and CAR-T cells. The related PNAS article from the Hynes group (5) shows that this fibronectin splice variant is highly expressed not only in the tumor neovasculature and stroma, but also in various lesions thought to be precursors of pancreatic cancer and in mouse models of metastatic melanoma. By focusing on something that tumors rely on for growth but is not abundantly expressed elsewhere in the organism, it might be possible to temporarily penetrate the tumor microenvironment without serious off-target effects.

PNAS: Your CAR-T cells also target the immune checkpoint protein PD-L1 as a way to rev up the immune system against tumors.

Ploegh: The fibronectin splice variant-specific CAR-T cells we reported in the Inaugural Article (4) serve as a sort of battering ram to open up the tumor microenvironment and allow other T cells access to the tumor. And the second type of CAR-T cells we used, PD-L1–specific T cells, were chosen because many tumors up-regulate the checkpoint molecule PD-L1, which then engages PD-1 on antigen-experienced T cells and conveys an inhibitory signal to tamp down the immune response. So the idea was to target the checkpoint.

Now, PD-L1 is also expressed on a subset of normal T cells, so this approach of targeting PD-L1 using CAR-T cells is predicated on the notion that as long as you get an antitumor effect, you could accept some collateral damage to healthy cells that express the same marker.

PNAS: So in a sense you have combined checkpoint blockade with CAR-T. How common is this approach?

Ploegh: I’m not aware of anyone else who has successfully tried PD-L1 as a target for CAR-T cells. The idea of combining PD-L1 and CAR-T was, so to speak, to get “two-for-the-price-of-one.” PD-L1–specific CAR occludes PD-L1 on tumors, preventing PD-L1 from interacting with PD-1 on otherwise tumor-specific T cells. We chose the B16 melanoma model because it’s quite aggressive.

PNAS: Do you think the nanobody approach is widely applicable across a range of tumor types?

Ploegh: With the Hynes laboratory, we have examined these fibronectin splice variants in MC38, a mouse model of colorectal carcinoma, and found that this particular model expresses lower levels of the splice variant than the melanoma model. So this approach obviously cannot be applied across all solid tumors, and we have to look for other potential targets in this and other tumor models. Our next steps are to extend the range of nanobodies, including other targets, such as integrins, which are cell-adhesion molecules that influence cancer cell invasiveness and metastasis.

PNAS: So what exactly is the clinical viability of these nanobodies for treatment?

Ploegh: As someone who works strictly in the preclinical setting, we have access first and foremost to mouse models. I would love to see this type of approach tried in humans, but that would require overcoming a number of regulatory hurdles. The standard CAR-T cells used in the clinic are based on the single-chain Fv fragments of antibodies, and these fragments are mostly derived from fully human antibodies already approved for clinical use. By contrast, the nanobodies are of camelid origin, and their immunogenicity in people remains to be determined. Moreover, if nanobody-based CAR-T cells are repeatedly administered to patients, you might generate an immune response against the CAR-T cells themselves that limits their efficacy, but that’s an issue that can be fixed by humanization [a process by which the cells are rendered nonimmunogenic].

That said, the nanobodies can be particularly useful as imaging agents to monitor disease progression and treatment response. That’s an application that would be easier to get into the clinic. In fact, other groups have used radiolabeled nanobodies that recognize the human EGF receptor-2, implicated in breast cancer; the chemistry of these nanobodies has been modified to serve as imaging agents in a “first-in-human” trial (6).

PNAS: Do you have plans to commercialize the nanobodies as clinical imaging agents?

Ploegh: We have filed for patent protection for these nanobodies as imaging agents, but have so far not partnered with any biotechnology firms for commercial production and testing. Standalone diagnostics do not appear to be lucrative propositions for biotech and pharma in the immuno-oncology space. They might welcome a companion diagnostic for an immunotherapy that is either approved or in development, but the commercial viability of standalone diagnostics is low, so I’ve been told. Still, the utility of these nanobodies as imaging agents is unquestionable.