Imaging the Immune Tumor Microenvironment to Monitor and Improve Therapy

See also the article by Ng et al in this issue.

Gary Luker is a professor of radiology, biomedical engineering, and microbiology and immunology at the University of Michigan. He serves as the associate chair for clinical research in the Department of Radiology. He is the editor of Radiology: Imaging Cancer. Research in the Luker Laboratory focuses on multiscale, quantitative imaging in cancer.

Tumors in both primary and metastatic sites consist of more than just cancer cells. Multiple other cell types, including different subsets of immune cells, establish heterogeneous local ecosystems referred to collectively as the tumor microenvironment. Nonmalignant stromal cells not only contribute to the volume of a tumor detected at anatomic imaging but also regulate cancer biology and response to therapy. In particular, tumor microenvironments typically suppress antitumor immunity, allowing cancer cells to hijack functions of immune cells to promote tumor growth, metastasis, and drug resistance (1).

Overcoming immunosuppression in tumor microenvironments now represents one of the most promising new additions to cancer therapy, but current immune-targeted therapies benefit only limited numbers of patients and types of cancer. Efforts to successfully treat patients with cancer by using current immunotherapies are complicated by the heterogeneity in tumor microenvironments. Cellular composition of tumor microenvironments differs between primary tumors and metastases, and even different metastases in the same organ have heterogeneous numbers, distributions, and types of immune cells. Tumor microenvironments evolve during therapy because of cell death, recruitment of new cells, and changes in tumor vascularity. Heterogeneity and temporal changes in tumor environments may alter local drug delivery, contributing to drug resistance. A tissue biopsy provides only a snapshot view of a single tumor microenvironment that almost always is the primary tumor. Patients rarely have repeat biopsies to detect changes in the immune tumor microenvironment over the course of therapy, hindering efforts to monitor effects of therapy and implement new treatment plans as needed. Approaches to noninvasively image composition and functions of key immune cells in primary and metastatic tumors over time promise to improve uses of existing cancer drugs and reveal new vulnerabilities for targeted therapy.

Macrophages, one of the major immune cell types present in anaplastic thyroid cancer and multiple other tumors, are classified along a continuum of functional states defined as M1 or M2. M1 macrophages exhibit antitumor, cell-killing functions that correlate with favorable outcomes for patients. M2 macrophages promote tumor growth by inhibiting inflammation, suppressing activation of other antitumor immune cells, and promoting growth of blood vessels. Because the M2 state predominates in a tumor, these tumor-promoting macrophages also are known as tumor-associated macrophages (TAMs) (2). TAMs express cell surface receptors CD206 and colony stimulating factor 1 receptor (CSF1R), which can be used to detect these cells. Signaling through CSF1R drives essential functions of TAMs, making this cell surface receptor a target to block tumor-promoting effects of TAMs in cancers. TAMs also retain the ability of macrophages to phagocytose extracellular molecules, a property exploited for selective drug delivery and functional imaging.

Many tumors paradoxically contain cytotoxic killer T lymphocytes (CD8 lymphocytes) that potentially could eliminate malignant cells. However, immunosuppression in tumor environments blocks functions of these cells, allowing cancer cells to survive and grow. Immune checkpoint molecules, such as the cell surface receptor programmed cell death protein 1 (PD-1) and its activating ligand (PD-L1), function as a major mechanism that blocks functions of cytotoxic T lymphocytes (3). At normal conditions, immune checkpoint molecules prevent excess activation of T lymphocytes to prevent autoimmunity. By expressing cell surface PD-L1, cancer cells and other cell types in tumor microenvironments activate signaling through receptor PD-1 on T lymphocytes, thereby co-opting a normal brake on the immune system to inactivate antitumor immunity. Antibodies now used for cancer immunotherapy in patients block signaling through PD-1 and PD-L1, reactivating cytotoxic T lymphocytes and even eradicating cancer in some patients.

Past studies documented limitations and pitfalls of anatomic imaging for monitoring response to current immunotherapies in cancer (4). Pseudoprogression refers to the clinical phenomenon that an increase in tumor size or even appearance of new lesions may precede successful response to immunotherapy and tumor regression. The transient increase in lesion size likely reflects recruitment of immune cells to a tumor. Although it occurs in 10% or fewer patients, pseudoprogression may complicate treatment of an individual patient, particularly because the time between pseudoprogression and tumor regression remains uncertain. Cancer immunotherapy can produce a sarcoid-like reaction, generating new nodules and lymphadenopathy that can be confused for cancer progression. Pseudoprogression also can show increased uptake of fluorine 18 fluorodeoxyglucose on PET studies, although quantitative analysis of standard uptake values potentially can identify patients with progressive disease (5).

In this issue of Radiology, Ng et al (6) present an integrated set of clinically translatable functional imaging methods to investigate tumor microenvironments in anaplastic thyroid cancer (ATC) and responses of primary and metastatic lesions to targeted therapy. ATC is a rare, highly aggressive malignancy that frequently metastasizes to lung. Median survival is less than 6 months, with only 20% survival at 1 year (7). ATC commonly has mutations that constitutively activate an intracellular protein, B-Raf kinase, that promotes unrestrained cell proliferation and invasion. One such mutation, which changes amino acid 600 in B-Raf from valine to glutamate (B-Raf^V600E), is the focus of this study. Experience with ATC and other malignancies with B-Raf^V600E shows that treatment with drugs specifically targeting this kinase, alone or combined with checkpoint immunotherapy, produces transient responses and frequent recurrences.

To understand functions of TAMs in resistance of ATC to the B-Raf inhibitor vemurafenib and anti-PDL1 antibody therapy, Ng et al used an immune-competent mouse model of ATC. They used standard methods to generate a tumor in the thyroid (direct implantation of mouse ATC cells into the organ), modeling the local environment of a primary ATC tumor. They also injected ATC cells intravenously to generate lung metastases based on trapping of injected cells in lung capillaries with subsequent growth of metastatic lesions. Whereas they acknowledge limitations of the approaches, the use of immune-competent mice with mouse tumor cells rather than human ATC cells implanted into immunocompromised mice is essential to fully investigate tumor immunity. To image TAMs in thyroid tumors and lung metastases, the authors exploited the ability of TAMs to efficiently ingest polysaccharide-coated nanoparticles. Ferumoxytol, a magnetic iron oxide nanoparticle, produces a quantitative change in MRI signal that correlates with numbers of TAMs in a tumor (8). Ng et al also used a second nanoparticle, macrin, to view TAMs at single-cell resolution of optically cleared tissues removed from tumor-bearing mice. Optical clearing uses chemicals to make tissues nearly transparent, increasing imaging depth and resolution for fluorescence microscopy. Whereas this study used a fluorescent macrin nanoparticle, macrin also can be labeled with 64Cu for PET studies of TAMs (9).

Both imaging methods showed more TAMs in ATC tumors in the thyroid compared with lung metastases, clearly demonstrating how imaging can depict heterogeneity among different tumor microenvironments. Combination therapy with vemurafenib and anti-PDL1 increased TAMs in tumor sites, although treatment modestly shifted macrophages away from a tumor-promoting, M2 state. Adding an antibody to CSF1R, a major driver for infiltration of TAMs into a tumor, blocked the increase in TAMs and change in macrophage polarization without altering tumor size. The local increase in TAMs without inhibiting CSF1R likely contributes to incomplete response to therapy, drug resistance, and eventual treatment failure. However, greater density of TAMs in primary and metastatic ATC opens potential new vulnerabilities for therapy. In addition to phagocytosing nanoparticle imaging agents, TAMs also ingest nanoparticle carriers for drugs. Imaging showed that increasing TAMs in a tumor correlated with greater uptake of a nanoparticle carrier for chemotherapy drugs independent of tumor vascularization. Whereas the authors did not load the nanoparticle carrier with a drug, these results show the potential to exploit the increase in TAMs to preferentially deliver drugs to local tumor microenvironments. This approach could improve treatment efficacy and reduce systemic side effects of therapy.

Although moving from preclinical to clinical studies always presents challenges, the authors present a direct path toward imaging TAMs in human cancers and analyzing spatial and temporal changes in these immunosuppressive cells during therapy. Most excitingly, the study demonstrates how imaging can identify potential causes of drug resistance and leverage this information to direct new treatments. Because TAMs are part of almost every tumor microenvironment, the MRI and PET nanoparticle imaging methods can extend to multiple other malignancies. Because drugs and cell-based immunotherapies for cancer continue to expand, there is a clear need and opportunity for functional imaging methods to quantify heterogeneity of tumor microenvironments more comprehensively and to help detect how these environments evolve during treatment. Capitalizing on this opportunity will advance knowledge of cancer therapy and ultimately increase numbers of patients actually cured of cancer.