Going Small to Attack Cancers

Q&A with Michelle Bradbury, M.D., Ph.D.
Kayt Sukel
December 18, 2018
Bradbury portrait

Michelle Bradbury, M.D., Ph.D.
Director of Intraoperative Imaging
Co-Director, MSK-Cornell Center for Translation of Cancer Nanomedicine
Memorial Sloan Kettering Cancer Center
Dana Foundation Grantee:  2012-2015

Over the past few years, several studies have heralded the promise of nanomedicine, or the use of different nanotechnologies to deliver therapeutics, to help treat progressive cancers. Malignant brain tumors, like high-grade gliomas (a type of tumor involving the brain’s glial cells) and central nervous system metastases can aggressively invade the brain; the wide-spreading tumors are notoriously difficult to treat due to their size and infiltrative nature, as well as the challenges getting therapeutic molecules across the blood-brain barrier. Michelle Bradbury, M.D., Ph.D., co-director of the Memorial Sloan Kettering-Cornell Center for Translation of Cancer Nanomedicines, has been researching novel techniques to not only deliver a promising cancer drug to these tumors, but simultaneously image them to see whether or not the drug is working.

What first sparked your interest in nanomedicine?

In my early years at Memorial Sloan Kettering, I worked with Dr. Eric Holland’s laboratory. They were always testing new drugs in mouse models. And I’d always hear, “We just tried this new drug, but we have no idea how much got to the tumor. We don’t know if it solubilized well [dissolved more easily].” It’s a big issue: If you inject a drug into the bloodstream, how do you know that it is successfully delivered to the tumor, as against “off-target” sites, such as the liver? I kept hearing this over and over again, and I finally said, “We’ve got to create drug delivery vehicles that enable imaging, in addition to improving their biological properties. This is important so that when the probe carries the therapy, it will preferentially accumulate and effectively treat the target site while clearing the body and eliminating toxicity to normal organs. You will also be able to see where it goes.”

No matter what application you pursue, whether it’s diagnostic or therapeutic, it should have an imaging label so that you can sensitively monitor its delivery, penetration, and distribution in the tumor and to other body organs over time, as well as other information that you might need to determine whether a treatment is effective.

Later, I began working with the neurosurgery and neurology teams, in particular, a neurosurgeon named Cameron Brennan. We knew that lung cancers and other solid tumors in the body seemed to be responsive to certain kinds of cancer drugs. But once those same cancers metastasized to the central nervous system (CNS), we would need a lot more of that drug to try to fight the tumors. That could lead to problems with toxicity and efficacy. And even using more of the drug, we wouldn’t see the same regression in tumor size. In fact, the tumors would often increase in size. It seemed to be a problem with both penetration and permeability.

That raised the question: Could we use an ultra-small nanoparticle to more effectively deliver drugs and better penetrate and distribute within these brain tumors? Such nanoparticles are small, they are neutrally charged, they can move across the altered blood-brain barrier and they can diffuse across the tumor. Maybe their size would be an asset to help us improve drug delivery to these brain tumors—so we could successfully treat the tumor while keeping the doses as small as possible and limiting toxicity. And, additionally, they come with an imaging label. We can follow them and see where they’re going.

What is it about this nanoparticle that you thought could overcome the penetration and permeability issues oncologists have faced in the past?

This particle is less than 8 nanometers in diameter and encapsulates fluorescent dyes for optical imaging. It’s made of silica shell that is coated with these polyethylene glycol chains that not only neutralize its charge but also serve as a scaffold to which we can attach small molecule peptides or antibody fragments, as well as cancer drugs that we want to deliver. This base particle was developed by my long-standing collaborator Professor Ulrich Wiesner, at Cornell University Ithaca. At MSK, we worked to co-develop the surface chemistry for imaging and therapeutic applications—adding contrast labels to its surface do positron emission tomography (PET) labeling. We also put special peptides on the surface that target proteins expressed by glioma tumors along with PET labels that would allow us to do multimodal imaging. So you can see where the particles are going and then can harvest the tumors and look at the fluorescent spread to see where the particle ended up later.

Most drug delivery vehicles out there are greater than 20 nanometers in size, which drives its uptake and clearance through the liver, spleen, and bone marrow. That may lead to unwanted toxicity. But since this one is less than 8, you can redirect the particle away from these organs to clear through the kidneys without problems, in addition to improving tumor targeted uptake. A smaller particle helps to drive renal clearance and limit off-target effects, thereby eliminating toxicity.

How did you test the particle?

We took this very small nanoparticle and injected it into a well-characterized genetically-engineered mouse model of glioma to see how well the particles worked as a drug delivery vehicle.

How might you determine that? One property to evaluate using imaging is whether or not—after you inject the particle without a drug—it diffuses throughout the tumor volume. You can look at the fluorescent signal from the particle to see if it gives you the right tumor coverage. That’s exactly what we did—and we found we had really good tumor coverage.

But mice models can’t tell us everything. So we also are using these particles in a clinical trial with brain tumor patients to evaluate its properties. It’s just the particle, no cancer drugs, but, again, early testing is showing that the particle diffuses over time. In fact, diffusion of the particles in a human CNS metastatic lesion appears similar to that found in preclinical studies.

Did the results surprise you?

It’s funny because, when we started, one of the neurosurgeons said, “Why would you expect this particle to overcome a significant number of in vivo challenges to improve the treatment of brain tumors?” You inject it, it has to cross a number of barriers, and then it has to diffuse throughout the tumor. I don’t think he was all that optimistic. But we saw that, over extended time intervals, the particle was fully filling out the tumor in small animal models.

The thing about nanotechnology is you just don’t know until you do the experiments. I was pleasantly surprised at how well it filled out the tumors in these mice. The real challenge is to demonstrate the same properties in larger tumor sizes, and where there’s more tissue heterogeneity. And that’s why we are doing a clinical trial, to see if it works as well in cancer patients.

What are some of the challenges of translating this technology for clinical use?

This is a diagnostic study—there are no drugs that will benefit the patient. We are just trying to find better agents that we can put cancer drugs on in the future that can deliver those drugs more effectively than the agents we have now. It can be a challenge to explain that to patients.

A second challenge is figuring out the dosing. We want to keep the dose of the imaging agents to the patient as small as possible. But even with microdoses of the particle, we saw very nice contrast when we did our PET imaging. We need to make sure that it gets out of the body, make sure that it doesn’t make any abnormal changes to a patient’s metabolic profile, blood counts, or urine. And we need to make sure that particle is stable, so it doesn’t fall apart in the body after injection. So there are some challenges where, what might be true in mice could look different in human patients.

We currently have a clinical trial where we are recruiting 20 patients. In one study, we will surgically remove the tumor from the patients and check how much the particles diffused within it. Basically, it looks at the same thing we did in our animal models but in human tumors. But then there is a second study arm that is non-surgical. There we are looking at the distribution of the particles over longer periods of time. We need both of these studies to investigate the time-dependent distribution in both the non-surgical patients and then in the surgical patients.

We can also use the imaging data to help us see if there are any features that can help us predict which patients might better benefit from particle-based therapies. Not all tumors are the same. Some may have a lot of edema, or swelling, around them. Some may have more heterogenous tissue. So we’re using deep learning algorithms to see if any features can help us predict which patients are the best candidates for further treatments with this platform. People get really excited about genomics. And gene expression profiles can be really useful. But sometimes the genomics may not stratify patients as well as the radiology features do.

Ultimately, we want to do a trial where we treat brain tumor patients with an actual cancer drug using the particle. It takes time—but we are getting there.

How can a relatively small grant, like the one you received from the Dana Foundation, help pave the way for larger projects such as this one?

That grant really gave us our start. At the time, it was hard to get nano projects funded. But from there, we received NIH funding for a nanomedicine center—and now we have a company, of which I am a co-founder, to helping to fund our clinical trials. But it was that first grant that gave us the data we needed to get to where we are now. I always tell people that it was the Dana Foundation that gave us our start—and I’m very thankful.

What is the most important thing you hope people understand about this approach for brain tumor therapeutics?

First, it is important for the technology to be safe—principally target the tumor while reducing off-target effects typically seen with larger size particles. This may lead to toxicity and poor therapeutic responses. A smaller nanoparticle may overcome these limitations by improve its distribution in the body in relation to the target site. That is, you can have an effective drug delivery vehicle that delivers drugs to the brain tumor, yet clears out of the body without toxic effects.

Many people thought that a small particle size would be a limitation – that we wouldn’t be able to get enough drug-linkers on the surface of this particle to have an effect. But after more than 12 years of working on this platform, along with the right team of investigators, we have made significant gains in creating the best surface chemical properties to substantially improve tumor targeting and other biological properties of this particle in vivo. These improvements would not be possible without the contributions and teamwork of a large group of scientific and clinical investigators, including neurosurgeons, neurologists, radiologists, materials scientists, chemists, and biologists. The field has learned a lot more about these particle technologies and their interactions with biological systems. We are making significant strides to put out the next wave of technologies that can address the limitations of prior generations. It’s really exciting.


Ultrasmall Duel-Modality Silica Nanoparticle Drug Conjugates: Design, Synthesis, and Characterization, Bioorganic & Medicinal Chemistry, Nov 15, 2015.