Scientists claim that angiogenesis inhibitors designed to kill cancers by starving them of oxygen may be hampered by the fact that a hypoxic environment promotes the recruitment of regulatory cells that both dampen the immune system’s ability to recognize and attack tumor cells and also, paradoxically, promotes angiogenesis directly.
The University of Pennsylvania-based team found that tumor hypoxia triggers the local recruitment of regulatory T cells (Tregs) through induction of expression of CCL28 (also known as mucosae-associated epithelial chemokine; MEC), which in turn promotes tumor tolerance and angiogenesis through increased levels of vascular endothelial growth factor.
George Coukos, M.D., and colleagues suggest that combining anti-angiogenesis inhibitors with new therapeutic approaches that block the production of Tregs could therefore boost the effectiveness of treatment. They report their findings in Nature in a paper titled “Tumor hypoxia promotes tolerance and angiogenesis via CCL28 and Treg cells.”
The Penn team incubated different ovarian cancer cell lines in either hypoxic or oxic condition for 16 hours and then used quantitative PCR to analyze the changes in expression of chemokines and their receptors as well as of other genes implicated in immune regulation. They found that CCL28 was the most highly upregulated chemokine gene under hypoxic conditions.
Further to these findings the researchers confirmed the production of the CCL28 protein and its regulation by hypoxia inducible factor 1α (HIF1α) in ovarian cell lines in vitro. In vivo, CCL28 expression levels varied among tumors but was localized mainly to tumor cells. In tumor xenografts CCL28 was upregulated in areas of tumor hypoxia, and CCL28 expression also correlated with HIF1α expression in ovarian cancer samples. Significantly, the authors write, CCL28 and HIFIα overexpression correlated with poor outcome in ovarian cancer patients.
Because CCL28 has been shown to recruit CC-chemokine receptor 10 (CCR10)-positive Treg cells during liver inflammation, the Penn researchers moved on to investigate whether hypoxic tumor cells actively recruit human Treg cells through CCL28 in vitro. Chemotaxis assays with human peripheral blood mononuclear cells (PBMCs) showed that CD4+CD25+ forkhead box P3 (FOXP3)+ cells migrated toward tumor cell supernatants in hypoxic medium far more readily than in oxic medium.
This preferential recruitment of cells was halted on the addition of either a CCL28-neutralizing antibody or an antibody to its receptor CCR10. Conversely, antibodies to the other known CCL28 receptor, CCR3, had no effect. “Thus, hypoxic tumor cells recruit Treg cells through CCL28, mostly through its binding to CCR10,” the authors write.
Human CCL28 is highly homologous to its murine counterpart, and in confirmation of the results in human cells, the researchers showed that the mouse ovarian cancer cell line ID8 also upregulated CCL28 under hypoxic conditions, and CCL28 expression correlated positively with expression of a hypoxia marker in orthotopic, intraperitoneal, ID8 tumors.
The team then transduced ID8 cells with mouse CCL28 (designated ID8-ccl28) and transplanted these cells into experimental animals. This resulted in significantly higher levels of CCL28 protein in the intraperitoneal tumors and peritoneal fluid of ID8-ccl28 mice compared with levels of CCL28 in the tumors and peritoneal fluid of mock-transduced ID8 mice, mimicking human ovarian cancer with a high and low level of CCL28 expression, respectively, the authors note.
Interestingly, studies in these animals demonstrated that ID8-ccl28 tumors accumulated significantly more CD4+CD25+FOXP3+ cells than ID8 tumors as a result of direct cell recruitment. Orthotopic, intraperitoneal ID8-ccl28 tumors in addition showed significantly faster growth and induced faster ascites development than ID8 tumors. Depleting CCR10+ cells suppressed tumor growth and abrogated the effects of CCL28 overexpression, whereas depletion of CCR3+ cells had no such effect.
Because leukocytes other than Treg cells could be recruited by CCL28, the researchers tested the contribution of CD4+CD25+FOXP3+ Treg cells specifically to the rapid growth of ID8-ccl28 tumors by using an anti-CD25 antibody to deplete most of CD4+CD25+FOXP3+ Treg population. CD25+ T-cell depletion hindered tumor growth and significantly reduced the effects of CCL28 overexpression in vivo.
Importantly, the researchers stress, neither ID8 nor ID8-ccl28 cells themselves express CD25, and the anti-CD25 antibody had no direct effect on their growth in vitro. “Thus, a direct link exists between tumor CCL28 upregulation and accelerated tumor growth, which is specifically attributable to Treg-cell recruitment in vivo through CCR10,” the authors state.
Consistent with a more tolerogenic environment, the team also observed markedly higher interleukin-10 (IL-10) levels in the ascites of mice with ID8-ccl28 tumors than in those with control ID8 tumors.
Further investigation showed significantly higher levels of vascular endothelial growth factor A (VEGFA) in the ascites of mice with ID8-ccl28 solid tumors that those with ID8 tumors, while ID8-ccl28 tumor nodules also exhibited increased microvascular density. Depletion of CCR10+ cells from the ID8-ccl28 tumors led to a significant reduction in tumor VEGFA levels and a significant reduction in tumor microvascular density, “confirming that excess VEGFA was contributed by CCR10+ hematopoietic cells recruited by CCL28,” the researchers note.
Further supporting the role of Treg cells in promoting tumor vasculature, mice administered with an anti-CD25 antibody demonstrated a significant reduction in tumor VEGFA levels and tumor microvascular density. “Thus, Treg-cell recruitment has a key role in establishing a VEGFA-rich tumor microenvironment and increasing tumor angiogenesis, whereas the depletion of Treg cells reduces tumor VEGFA levels and tumor vascularisation,” they write
In a final set of experiments designed to support their findings, the team discovered that Treg cells can also directly contribute to the VEGFA pool in the tumor microenvironment. CD4+CD25+ cells purified from fresh human donor PBMCs secreted markedly more VEGFA than CD4+CD25- cells under oxic or hypoxic conditions.
Similar results were obtained with mouse Treg cells purified from the spleen. In addition, medium conditioned by hypoxic human peripheral blood CD4+CD25+ cells induced a significantly larger expansion of human umbilical vein endothelial cells than medium conditioned by hypoxic human CD4+CD25- cells, while the addition of VEGF receptor neutralizing antibodies blocked this effect.
“For the first time, we are realizing the two programs, angiogenesis and immune suppression, are co-regulated and the two programs are mediated by the same cell types," states Dr. Coukos.
This scenario creates new therapeutic opportunities, he suggests, as approaches designed to suppress angiogenesis could be developed hand in hand with those designed to block regulatory T cells.
“The other implication of this study is that if antiangiogenesis therapy induces tumor hypoxia, that could create a rebound increase in regulatory T cells,” he continues. “That rebound could account for some of the resistance that is commonly seen in the clinic after antiangiogensis therapy is instituted.”