Metastatic sites pose sizable challenges for arriving disseminated cancer cells. No more so than the cerebrospinal fluid (CSF)-filled leptomeninges, which is the secondary site of leptomeningeal metastasis (LM). This harsh microenvironment is typically devoid of cells, hypoxic and lacking metabolic intermediates and micronutrients. The infiltration of this space by cancer cells brings about the arrival of immune cells. Therefore, cancer cells not only face survival on limited resources but must escape activation of an immune response. A new study suggests that cancer cells can accomplish this by outcompeting macrophages for an essential micronutrient, iron.
Clinical samples of human LM originating from breast cancer show that immune cells, predominantly macrophages and lymphocytes within the CSF, are far more numerous than cancer cells. To determine the transcriptional programmes of cancer cells that would enable them to cope under these conditions, the authors utilized single-cell RNA sequencing to profile the cellular components in the CSF of five patients with LM arising from either primary breast cancer or non-small-cell lung cancer relative to CSF from patients without LM. This revealed that all cells contained within the CSF of patients with LM had increased expression of iron transport genes indicative of iron being a limiting factor in this microenvironment. However, whilst immune cells expressed canonical iron transporter genes, the cancer cells expressed a range of genes related to iron binding and transport including an iron-capturing system comprising the iron-binding protein lipocalin 2 (LCN2) and its receptor SLC22A17.
As LCN2 and SLC22A17 transcript as well as protein expression was restricted to cancer cells and not macrophages or lymphocytes, the authors sought to investigate the consequence of this phenotype in an amenable mouse model. To mimic human LM, three independent mouse models were created through repeated in vivo selection from parental mouse lung adenocarcinoma, human lung adenocarcinoma and human breast adenocarcinoma cells. These mouse models confirmed that the LM subpopulation of cancer cells (LeptoM cells) but not macrophages within the leptomeningeal space exclusively expressed LCN2. Consistent with LCN2 being functionally important, depletion of LCN2 with short hairpin RNA (shRNA) was sufficient to reduce the growth of LeptoM cells in the leptomeninges in all three mouse models as well as confer a survival benefit in the mice. Yet, in other iron-rich anatomical locations in vivo, loss of LCN2 had no effect on the growth of LeptoM cells.
How might LCN2 expression be upregulated in cancer cells within the leptomeninges? To begin to address this question, the authors noted that concentrations of inflammatory cytokines IL-6, IL-8 and IL-1β were increased in the CSF of patients with LM compared with those without LM. This observation indicated that macrophage-derived cytokines might induce LCN2 expression in cancer cells in the leptomeninges. Indeed, conditioned media from macrophages isolated from the CSF of mice bearing mouse lung adenocarcinoma LeptoM cells stimulated LCN2 expression in LeptoM cells cultured in vitro, and, furthermore, LCN2 was specifically induced in LeptoM cells and not CSF-derived macrophages. Additional evidence for the role of inflammatory cytokines in LM came from treating the CSF derived from patients with LM with neutralizing antibodies to IL-6 or IL-6 and IL-8. This had the effect of inhibiting LCN2 induction in human lung adenocarcinoma LeptoM cells.
Iron levels as well as the amount of iron bound to LCN2 was found to be higher in the CSF of patients with LM than in those without LM. To show that the effects of LCN2 expression in mediating cancer cell growth in the CSF were attributed to its function in iron regulation, shRNA knockdown of LCN2 or SLC22A17 in LeptoM cells was performed. The outcome was a reduction in intracellular iron accumulation and cell growth, which could be reversed upon addition of exogenous transferrin.
Iron is important for maintaining immune function. Thus, it might be expected that macrophages in the CSF would have lower levels of intracellular iron and hence impaired function as a result of cancer cells siphoning off the scarce iron present in the CSF. Comparing CSF-derived macrophages with those collected from the spleen of mice bearing LM revealed that macrophages from the CSF had lower iron stores, which could be rescued upon shRNA-mediated depletion of LCN2 from cancer cells. Moreover, the ability of CSF-derived macrophages to generate reactive oxygen species and undergo phagocytosis was compromised in an LCN2-dependent manner.
Reasoning that iron abundance in the CSF creates a metabolic vulnerability that could be exploited therapeutically, the LM mouse models were treated with the iron chelator deferoxamine (DFO) or a copper chelator d-penicillamine. Only DFO reduced the iron content of LeptoM cells and their growth as well as extended the survival of the mice.
“iron abundance in the CSF creates a metabolic vulnerability”
This study exemplifies a mechanism of cell competition to circumvent iron deprivation, serving as an efficient survival strategy for metastatic cells.
Chi, Y. et al. Cancer cells deploy lipocalin-2 to collect limiting iron in leptomeningeal metastasis. Science 369, 276–282 (2020)