Antibodies targeting tumors can be powerful weapons to fight cancer , if they reach their intended target. But, particularly in solid tumors, it can be difficult or impossible to see where the antibodies end up after they are introduced into the body.
Now Stanford Medicine researchers have devised a way to see exactly where in a tumor antibody drugs are located, and even which cells they are binding to — an order of resolution many times better than current methods.
They've done this by overlaying a technique developed in the laboratory of Garry Nolan , PhD, professor of pathology, that enables the detection of dozens of proteins and cell types in a slice of tissue sample with another approach that tracks a fluorescent tag attached to a therapeutic antibody. Combining the two methods allows researchers to evaluate the cellular architecture around the tumor, the precise location of the antibody and even whether the antibody has attached to its target.
When the researchers applied the technique, which they term single-cell spatial pharmacobiology, or SSP, to tissue samples from head and neck cancers or pancreatic cancer, they found wide variability in the tumor microenvironment, or cellular neighborhood, among individual tumors. Those with more stromal tissue — nontumor cells like blood vessels and immune cells and extracellular scaffolding, or matrix, into which cancer cells nestle — were more resistant to antibody infiltration.
"Conventional techniques to assess antibody delivery rely on how much of the antibody is in the blood or on low-resolution imaging of radioactively tagged antibodies in the body," said Guolan Lu , PhD, an assistant professor of urology. "Neither of these approaches reflect how much drug has made it into the tumor, whether it is binding to the targets or if it is inducing the expected pharmacological effects. With SSP we can see what cell types the antibody is interacting with and whether and how the tumor microenvironment regulates drug delivery."
Lu is the lead author of the research , which was published June 8 in Nature Biotechnology. Nolan, the Rachford and Carlota A. Harris Professor, is a senior author of the study along with Eben Rosenthal, MD, professor and chair of the department of otolaryngology – head and neck surgery at Vanderbilt University Medical Center.
Solid tumors resistant to antibody treatment
Antibody-based immunotherapy has transformed the treatment of some types of cancers, including those of the blood and some types of breast cancers. But only about 20% of solid cancers respond to the treatment. Researchers have suspected that the dense tissue within and around many solid tumors blocks antibody infiltration. But classic positron-emission tomography, or PET scans, which rely on radioactively labeled antibodies, can give researchers only a rough idea of the antibodies' location.
"PET imaging shows a blob of brightness, but no fine detail at all," Lu said. "We can see if it gets to the general location of the tumor, but we don't know if it is trapped in the vasculature or the extracellular matrix. Maybe it's not working because it hasn't reached its cellular target. In contrast, with microscopic imaging of a fluorescently tagged antibody we can clearly see which cells the drug is interacting with and understand the spatial context of that interaction."
The study capitalized on ongoing Phase 1 clinical trials at Stanford Medicine testing whether giving cancer patients intravenous infusions of a fluorescently labelled antibody that targets a protein called EGFR found on the surface of many cancer cells could help surgeons remove tumors more completely. The trials tested whether a special camera that detects that fluorescence could guide their scalpels better than conventional techniques.
After the surgery, Lu and her colleagues collected the tumor tissue and used the antibody fluorescence to reveal the drug's precise location. Then, the same tissue was analyzed using CODEX — a technique pioneered in Nolan's lab that can photograph more than 50 different proteins simultaneously using DNA-tagged antibodies and fluorescent dyes. Finally, a computer algorithm overlaid the antibody's location onto the protein map with subcellular precision — aligning them so accurately that individual cell boundaries match to within 1 micrometer, roughly one-hundredth the width of a human hair.
A clear image
The result was a single, unified picture of where the drug went, which cells it fraternized with, whether it successfully blocked its intended target and what the surrounding cellular landscape looks like in intact human tumor tissue.
Lu and her colleagues found that the antibodies were more successful in infiltrating head and neck squamous cell tumor tissue (a cancer type that responds moderately to immunotherapy) than they were in the pancreatic tumor tissue. Pancreatic cancer, which tends to be more dense, is much more resistant to antibody treatment.
A closer look of tumor samples from 18 patients with head and neck cancers revealed that only about 16% of tumor cells both had EGFR on their surface and were reached by the labeled antibody. These cells were most often found on the edge of the tumor near blood vessels. In contrast, 36% of EGFR-positive tumor cells were not bound by the antibody. These cells were located primarily in the interior of the tumor. (About 43% of tumor cells were both EGFR- and antibody-negative, and about 5% of tumor cells were EGFR-negative and antibody positive.)
Further investigation identified specific stromal and extracellular matrix compositions that correlated with the ability of the labeled antibody to reach the tumor cells. One, a protein called periostin that normally helps tendons and bones maintain their structure, forms a dense mesh around tumors. Another is a high prevalence of a subtype of a cell called a fibroblast that makes proteins, including periostin, that make up the extracellular matrix.
"We found that among these two tumor types, those with reduced antibody infiltration have a barrier of stroma that consists of cancer-associated fibroblasts and specific proteins that encapsulate the tumor and prevent the drug from penetrating," Lu said. "We believe this can guide future drug development efforts. Perhaps we can develop drugs targeting these stromal barriers to increase the effectiveness of antibody drugs to treat cancers. We are very excited about the possibilities."
Researchers from Duke University, Johannes Gutenberg University, Vanderbilt University Medical Center, Tulane University, Cleveland Clinic, Heidelberg University, the German Cancer Research Center, the Barcelona Institute of Science and Technology, and the University of Michigan contributed to the work.
The study was funded by the National Institutes of Health (grants K99CA267171, R00CA2677171, T32CA118681, R01CA273035 and R01CA190306), a Stanford Translational Research and Applied Medicines grant, the Stanford Cancer Institute, the Deutsche Forschungsgemeinschaft, the International Myeloma Society, and the American Cancer Society.
