Macrophages Under the Influence of Tumor Mesothelin Weaken Host Defenses against Pancreatic Cancer Metastasis

Pancreatic cancer is now the third biggest cause of cancer-related deaths in the United States and is expected to surge to second-place on the cancer mortality list within the next decade despite our best efforts to identify more effective interventions. Pancreatic ductal adenocarcinoma (PDAC), which accounts for most cases, has limited treatment options and a mortality rate that nearly equals its incidence. Approximately half of pancreatic cancer patients already have detectable distant metastases at diagnosis. Detecting PDAC earlier will certainly contribute to improved survival. However, less than 50% of patients diagnosed with early-stage PDAC are cured by current treatments because metastasis is an integral part of the natural history of PDAC even in patients who appear to have disease confined to the pancreas on initial diagnostic imaging. Notably, patients with PDAC with so-called early-stage disease who have complete resection without residual disease (R0) but do not receive adjuvant chemotherapy to ablate microscopic deposits of metastases have a rate of <5% disease-free survival at 5 years (1). The culprit is most often recurrence at distant metastatic sites such as the liver, peritoneum, and lung, demonstrating metastatic seeding is nearly universal even in patients with early-stage PDAC.

In this issue of Cancer Research, Luckett and colleagues set out to determine factors that contribute to the high metastatic potential of PDAC (2). Using an immunocompetent orthotopically implanted murine PDAC model to assess spontaneous metastasis, they found that a more metastatic murine PDAC cell line had comparable intrinsic growth characteristics in culture to a less metastatic cell line of similar genetic background but decreased cancer cell apoptosis when grown in vivo. As no differences in tumor fibroblast or lymphocyte populations were noted that could explain the increased survival of the highly metastatic line, the investigators next interrogated tumor-associated macrophages, the most populous resident immune cells in PDAC tumors. Interestingly, the macrophage population in tumors from the highly metastatic line had shifted from an MHC II⁺ tumor-restraining to a CD206⁺ tumor-permissive phenotype, both within the primary tumor and the “normal” lungs of mice bearing the more metastatic line. Realizing that cancer cells must be secreting specific factors that affected macrophage polarization, the investigators next examined the secretome and found that protein concentration of mesothelin (MSLN) was enriched in conditioned medium from the more metastatic cancer cells. Moreover, higher MSLN protein expression was associated with increased CD206⁺ cells in human PDAC patient surgical specimens and poorer patient survival. Taken together, these data indicate, first, that the MHC II⁺ macrophage population significantly contributes to immune surveillance that suppresses PDAC expansion and, second, that PDAC promotes the formation of a tumor permissive metastatic niche in the lungs through action of MSLN. The role of MHC II⁺ macrophages in tumor control is not further explored in the article but could be an important topic of future research to prevent PDAC metastasis.

The investigators move on to elegantly define the molecular consequences of tumor MSLN on macrophage signaling. Using tumor-conditioned medium and bone marrow–derived macrophages as well as genetic knockout or rescue of MSLN in cancer cells implanted into mice, the investigators demonstrate that MSLN promotes macrophage expression of VEGF and the neutrophil chemoattractant S100A9. Increased VEGF expression was observed in primary tumors and lung tissue of mice bearing the highly metastatic cell line compared with the less metastatic line, as well as in mice bearing the wild-type MSLN-expressing cell line compared with MSLN genetic knockout cells. Interestingly, the higher VEGF concentration did not cause higher vessel density in the mature tumors; however, exposure of cancer cells to VEGF promoted colony formation in a VEGFR2-dependent fashion. In comparison, overabundance of S100A9 was not detected in primary tumors with higher MSLN expression but only in the lungs of mice bearing these tumors, the most common site of metastasis in this model. Higher lung S100A9 expression correlated with increased formation of neutrophil extracellular traps (NET), structures known to support lung metastasis. The data clearly demonstrated that macrophages under the influence of tumor MSLN cease to guard against metastasis and instead help to promote cancer cell survival.

MSLN expression has long been appreciated as a poor prognostic marker in patients with PDAC with early-stage disease. While most PDAC tumors make at least some MSLN, those with the highest expression had an odds ratio >12 of killing their hosts within the first year after potentially curative resection as compared with tumors that lacked detectable MSLN (3). A definitive mechanistic cause for such a large MSLN dose effect has remained elusive. Some studies have shown modest increases in tumor intrinsic growth properties upon MSLN expression in PDAC cells due to activation of the NFκB pathway or increased MAPK or matrix metalloprotease signaling mediated by MSLN binding partner mucin-16 (4, 5). My laboratory has found that MSLN does not promote primary tumor growth but does contribute to metastasis by increasing angiogenesis in nascent metastatic deposits through a paracrine signaling mechanism that we failed to elucidate at the molecular level (6). Using the PDAC TCGA (The Cancer Genome Atlas) dataset, this study again showed that higher MSLN expression correlated with poorer survival in patients with PDAC, but this work transcends prior studies by identifying the complex network of molecular signals passing between immune and cancer cells that contribute to the survival disadvantage.

There are limitations to the study that elicit questions about some details of the proposed MSLN-driven mechanism. MSLN is a glycophosphoinositol (GPI)-linked protein that is initially trafficked to the cell surface. It is later released by enzymatic cleavage to become a shed product present in blood, intratumoral fluid, and ascites. Prior work by Dangaj and colleagues demonstrated that MSLN shed by phospholipase action retains a terminal mannose residue in the GPI-linkage that can directly bind to CD206, the mannose receptor on macrophages (7). Luckett and colleagues circumstantially implicate this form of shed MSLN in their mechanism by showing that the more highly metastatic PDAC cell line has increased mRNA expression of GPI-specific phospholipase D (GPLD1), yet the investigators do not demonstrate the presence of the mannose-containing MSLN product in tumor cell–conditioned medium. Importantly, a recent study of multiple MSLN-expressing human cancer cell lines, including PDAC, failed to identify this mannose-containing form as a major MSLN product (8). Proteolytic cleavage within the MSLN amino acid backbone by extracellular proteases is presently recognized as the dominant mechanism of MSLN release from the cancer cell surface, although studies of MSLN cleavage products found in pancreatic cancer patients have not been performed. It is also unclear how feasible it would be for MSLN to generate a premetastatic niche in the lung of human patients. Human PDAC sheds insufficient amounts of MSLN into patients’ circulation to raise serum MSLN concentrations (9). KPC-derived PDAC cell lines such as those used in this study have MSLN expression that is far stronger and more uniform than that seen in most PDAC patients, and elevated shed MSLN can easily be detected in the circulation of mice bearing orthotopically implanted tumors that express MSLN (10). Perhaps the low concentrations of mannose-containing MSLN anticipated in intratumoral fluid and systemic circulation are sufficient to drive the described processes; however, it would be worthwhile to consider alternate candidates for a macrophage MSLN receptor.

Overall, these new data have uncovered a novel paracrine signaling mechanism that potently promotes PDAC metastasis and is strongly associated with early recurrence and decreased survival in patients. As MSLN is a highly targetable surface antigen with favorable differential expression between tumor and normal cells, pharmacologic inhibition applied in the relevant setting could prove advantageous to patients with PDAC as understanding of this MSLN signaling pathway continues to improve.

C. Alewine reports nonfinancial support from AstraZeneca, Minneamrita, and ProDa, and other support from HCW Biologics outside the submitted work. In addition, C. Alewine has a patent for employee invention report licensed and with royalties paid to NCI by source unknown to author.

The author is grateful to Dr. Ira Pastan for critically evaluating this manuscript.