Immune cell populations recruited to the tumour microenvironment (TME) can profoundly influence cancer progression. Equally, immune cells can be reprogrammed by the TME itself to promote cancer progression. Now, Han et al. have discovered a new immune cell population induced outside of the TME by primary tumours; a circulating factor secreted by these cells can in turn support tumour growth.

Credit: Lara Crow/Macmillan Publishers Limited

To search for novel cellular subsets originating in the main lymph organs that could have a systemic role in tumour progression, the researchers used the common leukocyte marker CD45 to screen for known immune cells and previously unidentified non-leukocyte populations in the spleen, draining lymph node and bone marrow of tumour-bearing mice. Using the Hepa cell line to create a syngeneic orthotopic mouse model of hepatocellular carcinoma (HCC), it was shown that the proportion of all known immune cell subsets in the enlarged spleen significantly decreased during HCC progression. However, 4 weeks after tumour challenge, an expanded population of CD45 non-leukocyte cells was detected in the spleen of mice. The phenotype of these cells was defined by the expression of erythrocyte and megakaryocyte markers, including Ter-119, CD41, CD71 and CD44. The authors named this cell population Ter erythroblast-like, as their transcriptome shared similarities with megakaryocyte–erythroid progenitor cells from the bone marrow. Importantly, Ter-cells were specifically induced by tumours as they were not found in the spleens of normal mice and, moreover, originated exclusively within the spleen as they were not induced in HCC-bearing mice that had undergone splenectomy before injection of Hepa cells. Ter-cells were also found in the spleens of mice bearing EG7 T cell lymphoma and B16 melanoma tumours, high-lighting that the generation of this cell subset was not exclusive to HCC.

Investigating the emergence of Ter-cells revealed that SMAD3 activation and transforming growth factor-β (TGFβ) signalling were required, as both Smad3-knockout mice bearing Hepa cell tumours and mice injected with Tgfb-knockout Hepa cells had a reduced ability to form splenic Ter-cells. The authors went on to show that whilst Ter-cells did not affect host antitumour immunity, they did exhibit increased expression of the neurotrophic factor artemin (encoded by Artn). Accordingly, elevated levels of serum artemin were detected during HCC progression in the mouse model, which could be effectively abolished with splenectomy before tumour challenge.

Using in vitro assays with Ter-cells demonstrated that Ter-cells could increase Hepa and EG7 cell migration and invasion and that this effect could be reversed with a neutralizing antibody to mouse artemin. Similarly, in vivo injection of Ter-cells increased Hepa and EG7 tumour growth and reduced survival of mice, a phenotype that could be opposed by caudal vein injection of a neutralizing antibody to mouse artemin. Additionally, Hepa tumour growth was inhibited in Artn-knockout mice implying that Ter-cell-derived artemin and not cancer cell-derived artemin was responsible for the observed tumour progression. Mechanistically, the authors found that artemin promoted HCC survival through phosphorylation of caspase 9 and promoted invasion through expression of TRIO and F-actin-binding protein (TRIOBP) and integrin β5.

These findings in cancer mouse models prompted the authors to examine samples from patients with HCC. This analysis confirmed the presence of CD235A (the human orthologue of Ter-119) and artemin+ cells in the spleen. Furthermore, increased artemin and TGFβ levels in the serum and higher expression of the artemin receptor GFRα3 in HCC tissues all correlated with poor prognosis.

This study proposes that splenectomy might be a viable clinical option for patients with advanced HCC to treat not only the associated symptom of splenomegaly but also to slow tumour progression.