Erythroid-transdifferentiated myeloid cells promote portal vein tumor thrombus in hepatocellular carcinoma

Rationale: Hepatocellular carcinoma (HCC) is primarily characterized by a high incidence of vascular invasion. However, the specific mechanism underlying portal vein tumor thrombus (PVTT) in HCC remains unclear. As a consequence of myeloid cell developmental arrest, CD71+ erythroid progenitor cells (EPCs) and myeloid-derived suppressor cells play important roles in HCC; however, their roles in PVTT remain unclear. Methods: The role of CD71+ EPCs in the HCC tumor microenvironment (TME) was evaluated via morphological, RNA-sequencing, enzyme-linked immunosorbent assay, and flow cytometric analyses. Co-culture techniques were employed to assess the CD45+ EPCs and their vascular compromising effect. Additionally, the PVTT-promoting function of CD45+ EPCs was explored in vivo in a murine model. Results: The CD45+EPCs in HCC tissues exhibited increased myeloid cell features, including morphology, surface markers, transforming growth factor (TGF)-β generation, and gene expression, compared with those in circulation. Hence, a large proportion of CD45+EPCs, particularly those in TMEs, comprise erythroid-transdifferentiated myeloid cells (EDMCs). Additionally, the expression of C-C chemokine receptor type 2 (CCR2) mRNA was upregulated in CD45+EPCs within the TME. Tumor macrophages from HCC tissues induced substantial migration of CD45+EPCs in a dose-dependent manner. Meanwhile, results from immunofluorescence analyses revealed that these two cell types are positively associated in the TME and circulation. That is, EDMCs are chemoattracted by HCC macrophages mainly via CCR2 from CD45+ EPCs in the circulation. Additionally, the expressions of FX, FVII, FGB, C4b, CFB, and CFH were elevated in CD45+EPCs within the TME compared with those in the spleen. The CD45+EPCs from the HCC TME promoted vessel endothelial cell migration and compromised tube formation through TGF-β and FGB, respectively. Additionally, CD45+EPCs from the TME induced HCC cell migration. HCC macrophage-induced CD45+EPCs to exhibit higher levels of FX, FVII, FGB, and TGF-β. Meanwhile, upregulation of CCAAT/enhancer binding protein beta expression induced FGB and TGF-β generation in CD45+EPCs in the TME. WTAP, a major RNA m6A writer, stabilized FX and FVII mRNA and enhanced their nuclear export in CD45+EPCs from the TME. CD45+EPCs from the TME were positively associated with PVTT and poor prognosis. Splenectomy reduced the level of CD45+EPCs in the circulation and TME, as well as the incidence of microvascular invasion. The incidence of microvascular invasion increased following the transfer of HCC tissue CD45+EPCs to splenectomized HCC-bearing mice. Conclusions: The CD45+EPCs enriched in the HCC microenvironment are EDMCs, which are induced by HCC macrophages to migrate from the circulation to the TME. Subsequently, EDMCs promote PVTT by compromising the blood vessel endothelium, aggravating coagulation, and promoting HCC cell migration.


Orthotopic HCC mouse model
These experiments were approved by the Institutional Animal Care and Use Committee of the Third Affiliated Hospital of SYSU.C57BL/6J mice (male, aged 6-8 weeks) were purchased from the Guangdong Medical Laboratory Animal Center (Guangzhou, China).All mice were housed and maintained under pathogen-free conditions.Preparation and transportation of Hepa1-6 cell Matrigel matrix glue (Corning, Corning, NY, USA) suspension: Hepa1-6 cells were routinely cultured and transferred 3-5 generations after recovery.Hepa1-6 cells in the logarithmic growth stage were digested and centrifuged, and then prepared into a mixture containing 1×10 6 cells, 10 μL PBS, and 10 μL Matrigel matrix gel.All devices in contact with the Matrigel matrix glue were pre-cooled in a 4 ℃ refrigerator, and the prepared Hepa1-6 cell Matrigel matrix glue suspension was placed on ice for transport.
Operation: 10-15 min before operation, C57BL/6J male mice were anesthetized by intraperitoneal injection with 160 μL 1% pentobarbital per 25 g weight.Then, the response of the animals to anesthesia was observed.The mice should breathe slowly and shallowly and not respond to squeezing of the back feet.The abdominal skin of the mice was shaved with a shaver, and the fine fur attached to the skin was wiped with a wet gauze with 70% (v/v) ethanol to prevent the fur from entering the peritoneum.The animals were placed in the supine position and fixed on to the lid of a 100 mm Petri dish.Cotton swabs were placed on the backs of mice to better expose the liver.The skin was sterilized with povidone iodine solution, and an incision of approximately 1 cm was made in the middle of the abdomen under the xiphoid process with ophthalmic scissors.Another 1-cm incision was made along the alba line of the lower abdomen of the xiphoid process, and the middle liver was gently pulled onto the peritoneum with a cotton swab moistened with normal saline and placed on the outer surface of the peritoneum.A cotton swab moistened with normal saline was placed on the dorsal side of the middle lobe of the liver of mice, and 20 µL Hepa1-6 cells mixed with Matrigel matrix glue were absorbed with 0.5 mL insulin injection needle.The cells were inserted near the lower edge of the liver in the middle lobe of the liver and slowly injected into the left lobe of the liver under the capsule.After injections, needles were withdrawn along their direction, and a cotton swab was used to place pressure for 1 min to stop the bleeding.Before suture, no bleeding or fluid leakage was observed at the injection site.The cotton swab was removed and the liver returned to the enterocoelia; then, the peritoneum and skin incision were closed by intermittent 5-0 suture.Then, the skin was partially wiped with a sterile gauze to remove any blood.The animal was placed on its side on a thermal blanket until the anesthetic wore off (30-60 min).When the mice were resuscitated and could roll over on their own, they were placed in a new cage.

Splenectomy mouse model
Mice were placed in the right lateral position.A 1.0-2.0cm skin incision was made with scissors along the left side, midway between the last rib and the hip joint.The connective tissue was loosened using the blunt end of forceps.Subsequently, a 1.0-2.0cm incision was made through the peritoneal wall.The spleen was gently placed and placed on the exterior surface of the peritoneum.The artery attached to the spleen was tied off with a 4-0 suture by looping the suture through the mesentery.A single knot at the tip of the spleen was made, and the mesentery and connective tissue were cut to remove the spleen.The peritoneal wall was closed with one or two separate sutures, and the skin was closed with an additional two or three sutures.

CD45 + EPC transfer experiment
CD45 + EPCs were isolated by FACS from the tumor or spleen of tumor-bearing mice.
Then CD45 + EPCs were stained with CFSE (2.5 µM) for 10 min at 37 ℃.Next, 1.5 × 10 6 CD45 + EPCs from the tumor or spleen were transferred into splenectomized tumor-bearing mice via caudal vein injection.After 12 h, mice were sacrificed, and CFSE + CD45 + EPCs were assessed in the blood and tumor tissue.Other mice were sacrificed after 20 days for analysis of MVI or further survival assessment.

CD45 + EPC migration analysis
The bottom of a 24-well plate was coated with 160 μL complete medium + 20 μL Matrigel matrix glue, 160 μL complete medium containing 1×10 6 Hepa1-6 cells + 20 μL matrix glue, 160 μL complete medium containing 1×10 6 F4/80 positive cells + 20 μL matrix glue, or 160 μL containing 1×10 6 /well of F4/80 negative cells + 20 μL matrix glue.The plate was incubated at 37 ℃ for 30 min to solidify the matrix glue.Then, 420 μL complete medium was mixed with the substrate glue, and 200 μL suspension complete medium containing 5.0×10 5 CD45 + EPCs was added to each well of a Transwell chamber.The chamber was removed 48 h later, and the cell density of the indoor medium was counted.
After counting, the medium was frozen at -80 ℃.

Cell migration and invasion analysis
Migration or invasion assays were performed using 24-well plates with 8.0-μm pore size Transwell filter inserts (Corning) with or without pre-coated diluted Matrigel (Corning).
Then, 2×10 4 C166 or Hepa1-6 cells with serum-free medium were placed into the upper chamber, and medium containing 20% (v/v) FBS was added to the bottom chamber.After incubation at 37 °C for 24 h (migration) or 72 h (invasion), cells on the underside of the membrane were immobilized and stained with crystal violet (Solarbio Life Sciences, Beijing, China).Then, penetrated cells were counted in five random fields under the microscope.

Vascular endothelial cell tube formation analysis assay
The 96-well plates were coated with 20 μL per well of Matrigel (Corning).C166 cells were serum-depleted overnight in DMEM and then seeded at 1.5×10 4 cells per well and allowed to attach for 45 min.The medium was then replaced with conditioned medium of CD45 + EPCs from the spleen and tumor tissues of tumor-bearing mice.C166 cells on this matrix migrated and formed tubules within 12 h of plating.Tubule formation was monitored at 40× magnification using a light microscope (Nikon Eclipse TS100, Surbiton, UK) For the experiments using αvβ3, the conditioned medium of CD45 + EPC from the spleen and tumor tissues of tumor-bearing mice was incubated with recombinant αvβ3 for 1 h before addition to the Matrigel assay.

Construction of stable knockdown and overexpressed cells
Lentiviruses expressing shWtap, shCebpb, shNC, AAV-Wtap, AAV-Cebpb, or empty AAV vector were purchased from GeneChem, Co. Ltd. (Shanghai, China).Mouse tumor CD45 + EPCs were selected to establish stable Wtap-knockdown models (Table S4), while the spleen CD45 + EPCs of cancer mice were used for stable AAV-Wtap overexpression experiments.According to the supplier's instructions, 10 5 cells in 100 μL StemSpan SFEM II medium per well were planted into a 96-well plate and transfected with the indicated virus (final MOI = 50) for 24 h.The transfection efficiency was determined by flow cytometric analysis.The cells were then washed twice with 1× PBS and replated in StemSpan SFEM II medium at 250,000/mL.
We cultured 1×10 5 CD45 + EPCs from mouse BM, liver, or spleen in 100 μL StemSpan TM SFEMII medium supplemented with 10 IU/mL human recombinant erythropoietin.The cultures were maintained at 37 °C in a 5% CO2-humidified atmosphere in 96-well plates for 6 days.Wright-Giemsa staining was performed to examine cell morphology.

Flow cytometry analysis and sorting
Cell phenotypes were evaluated on CytoFLEXLX (Beckman Coulter Life Sciences, Brea, CA, USA), and data were analyzed with the CytExpert and FlowJo V10.0.7 (FlowJo, Ashland, OR, USA).Data were acquired as the fraction of labeled cells within a live-cell gate set to a minimum of 50,000 events.For flow cytometric sorting, a BD FACSAria III cell sorter (BD Bioscience) was used.The antibodies used are listed in Table S2.The gating strategy is shown in Fig. S6-8.
Isolation of PBMCs, cord blood mononuclear cells (CBMCs), and mononuclear cells from liver, BM, and spleen.
Blood and tissue samples were analyzed within 6 h after sampling.PBMCs and CBMCs were isolated from whole blood or cord blood through Ficoll centrifugation.Liver samples were cut into approximately 1-mm 3 pieces and were triturated with a 1-mL syringe plunger on a 70-μm Cell-Strainer (BD) in RPMI-1640 medium (Invitrogen) containing 10% (v/v) FBS until uniform cell suspensions were obtained.After the supernatant was removed, the pellets were washed twice with PBS.Later, cells were resuspended with 7 mL 30% Percoll (GE Healthcare, 17089109-1) and then paved upon 3 mL 70% Percoll at the bottom of a 15-mL tube.The cells were sequentially centrifuged in an Eppendorf 5810R centrifuge at 600 g for 20 min.Liver mononuclear cells were collected from the cell layer at the interface between 30% and 70% Percoll.Mouse BM cells were harvested by femoral bone flushing and filtered through a 70-μm pore size cell strainer (Corning).Spleens were mechanically dissociated and filtered.

Intracellular staining
Freshly isolated mononuclear cells were incubated with Cell Stimulation Cocktail (eBioscience, Cat #00-4970-30) for 4 h at 37 °C.Then, the mononuclear cells were harvested, stained with antibodies for surface markers, and permeabilized using the Intracellular Fixation & Permeabilization BUFFER Set Kit (eBioscience, Cat#88-8824-00) overnight at 4 °C.After washing with PBS, the cells were incubated and additionally stained for intracellular cytokine antibodies.The labeled cells were analyzed via flow cytometry.

Immunofluorescence analysis
Liver tissues were successively cut in one direction from the liver right lobe for mouse experiments and used for paraffin embedding.Liver tissues of patients with HCC were obtained from those who underwent surgery.Normal liver tissues from surgery of patients with liver hemangioma were used as the control.Paraffin-embedded sections from liver tissues were deparaffinized with xylene and rehydrated with descending grades of ethanol.
Immunofluorescence was performed after antigen retrieval.Sections were microwaved in 10 mM citric acid for 2 min at high power followed by 5 min at 50% power.The sections were then cooled for 60 min at room temperature, washed in PBS, and blocked with 0.5% BSA/PBS for 30 min 37 °C.Later, the sections were incubated with primary antibodies (listed in Table S2) overnight at 4 °C and then incubated for 30 min at 37 °C.After washing in PBS, the sections were incubated with an appropriate secondary antibody, 1:400 diluted goat/donkey anti-rat Alexa Fluor 488, goat/donkey anti-rabbit Alexa Fluor 594, or goat/donkey anti-mouse Alexa Fluor 647 (listed in Table S2) for 45 min at 37 °C.Finally, the sections were washed and mounted with DAPI (Mounting Medium with DAPI -Aqueous, Fluor shield Abcam, Cambridge, MA, USA; ab104139) and stored in dark at -30 °C until imaging.Identical confocal settings were used to limit intra-experimental variability.
Fluorescent signals were detected using a laser scanning confocal microscope (ZEISS LSM 780) and automatic slide slice scan (Panoramic MIDI, 3DHISTECH).For the quantification of immunofluorescence results, the images of the indicated tissues were analyzed with Imaris 8.4 (BITPLANE).Alternatively, the absolute numbers of the CD71 and CD235a/Ter119 double positive cells/mm 2 in nine fields of interested areas were statistically analyzed.
Investigators were blinded to the experimental group when performing these studies.

Measurement of intracellular ROS
ROS levels were determined through 2′,7′-dichlorofluorescein diacetate (DCFDA) staining.Briefly, mononuclear cells were incubated with 2.5 μM DCFDA at 37 °C for 30 min.Cells were then washed three times, resuspended in PBS, and stained with surface markers at 4 °C for 30 min.ROS levels were analyzed using a flow cytometer at excitation

Table S2 :
Reagents and resources.

Table S4 .
Baseline demographic and clinical characteristics of hepatocellular carcinoma patients with high or low CD45 + erythroid progenitor cells in circulation.