The Induction Of Peripheral Trained Immunity in The Pancreas Incites Anti-tumor Activity To Control Pancreatic Cancer Progression Part 2

Discussion

While the trafficking of β-glucan has been previously characterized, the specific tropism of β-glucan to the pancreas has not been previously reported37. We show that particulate β-glucan can traffic directly into the pancreas and can also be phagocytosed by macrophages which then traffic into the pancreas. The relationship between the peritoneal cavity and the pancreas has not been well defined, and studies relating to the pathophysiology of acute pancreatitis (AP) have identified that peritoneal macrophages are a principal contributor to the inflammatory response in AP, thus supporting a connection between the peritoneum and the pancreas38. 

Our imaging data indeed show that peritoneal macrophages that phagocytose isotope-labeled WGP primarily traffic to the pancreas. Studies on liver injury have further supported this model of a dynamic interchange of cells between the peritoneal cavity and solid organs that is independent of the circulation 39. Together this suggests that even in homeostatic conditions there exists a basal level of immune cell exchange between the pancreas and peritoneal cavity that can be exploited in the setting of particulate β-glucan. 

Macrophages are an important part of the innate immune system of the human body and have powerful functions of identifying, phagocytizing, and removing bacteria and foreign substances. Periodontitis, a chronic infectious disease characterized by gingival inflammation and alveolar bone loss, is the leading cause of tooth loss in adults. It is now clear that the tissue destruction of periodontitis is caused by the host's immune response to infection, and macrophages, as an important part of the host's immune response, play an important role in the occurrence and development of inflammation. 

Therefore, we need to improve immunity in our daily life. In our research, we found that Cistanche can enhance immunity. Adding Cistanche into boiling water to make Cistanche tea, and drinking it 1-2 times a day can enhance the immune function of the body because It contains active ingredients, such as polysaccharides, flavonoids, alkaloids, etc., which can improve the immune function of the body. At the same time, adding Cistanche to soup or porridge, once or twice a week can enhance immunity.

Click to know pure cistanche product

While this study investigated particulate β-glucan WGP, it may also be interesting to investigate whether soluble yeast-derived β-glucan shows a similar tropism and signaling mechanism within the pancreas. Particulate β-glucans are known to signal directly through Dectin-1, while soluble β-glucans are thought to function through a CR3-dependent pathway to exert their anti-tumor properties40,41. One study showed that IP administration of soluble β-glucan resulted in trafficking to the peripheral blood and APCs in the peritoneum, spleen, and bone marrow, indicating that despite the different molecular signaling mechanisms, a similar ability to activate trained immunity may be possible42. However, given that our data indicates that this trafficking is Dectin-1 dependent, only particulate β-glucan may display direct trafficking to the pancreas through both particulate and soluble may initiate mechanisms of trained immunity. 

Further, previous studies which characterize the trafficking of intravenously injected soluble β-glucan show trafficking to the spleen, kidney, and liver though do not mention tropism to the pancreas43. As the pancreas has not been previously thought to be a target of β-glucan trafficking it is not clear whether the pancreas simply has never been studied or whether pancreatic trafficking does not occur.

Future studies should address the biodistribution of both soluble and particulate β-glucan across various routes of administration.

β-Glucan trafficking to the pancreas has a multifactorial impact on the immune populations present within the pancreas. First, βglucan arrival to the pancreas directly impacts the populations of immunosuppressive M2 resident macrophages present within the pancreas that are known to be important in the promotion of pancreatic tumors. CyTOF and scRNA-Seq data showed a nearly complete disappearance of the resident macrophage population 7 days following WGP administration. Interestingly, the disappearance of the resident macrophage population coincides directly with a reciprocal appearance of a Ly6Clo macrophage population. This Ly6Clo macrophage population bears similar phenotypic markers to the resident population, though skews more towards an M1 phenotype. 

It is thus likely that resident macrophages come into contact with β-glucan that has trafficked to the pancreas and these cells become repolarized, therefore taking on a different cellular phenotype which results in the formation of a unique cluster. Second, the arrival of β-glucan to the pancreas results in amplified chemokine/chemokine receptor signaling which recruits pro-inflammatory Ly6CHi infiltrating monocyte-derived macrophages from the periphery to the pancreas. Twenty-four hours following the administration of WGP, CCL2 levels in whole pancreatic lysates are found to be increased by 30-fold. Accordingly, we show that the robust β-glucan dependent cellular influx to the pancreas is dependent on CCR2.

While we focus on the dynamics of the myeloid compartment due to the observed importance of CCR2/CCL2 signaling, there was an interesting initial enhancement of cluster 11 on day 3 and then disappearance by day 7 following WGP treatment. Cluster 11 was characterized as ILC2s based on the expression of several markers such as KLRG1, IL-5, IL-13, and ICOS. ILC2s are known to be long-term tissue-resident cells44. Recent studies have shown that activation of local ILC2s by tissue-specific alarmins induces their proliferation, lymph node migration, and blood dissemination. ILC2s from several tissues, such as the gut and the lung, have been shown to enter the blood by extrusion from these perturbed tissues and migrate to other tissues such as the liver44–46. 

Such a mechanism could explain the initial increase and then disappearance of ILC2s in the pancreas. Given that ILC2s have recently been implicated in the response of pancreatic tumors to immunotherapy47, more research is warranted that investigates the impact of WGP and the initiation of trained immunity on the ILC2 population of the pancreas.

Although the initiation of an influx of pro-inflammatory immune cells to the pancreas is in itself important, it is the CCR2+Ly6CHi infiltrating monocyte-derived macrophage populations from the periphery display features of trained immunity that carries the most important implications as the induction of peripheral trained immunity in the pancreas has not yet been characterized. This induced trained immunity is also associated with metabolic and epigenetic reprogramming. 

While CCR2 is known to be an important receptor in the recruitment of monocytes, this is also the first indication that CCR2 signaling on monocytes is requisite for the establishment of peripheral trained immunity. It is noted that CCR2+ monocytes/macrophages have been linked to tumor metastasis and progression48,49. However, our data suggest that the recruitment of reprogrammed, trained CCR2+ monocytes/macrophages exert robust anti-tumor effects.

This is directly demonstrated by our admix experiment where mice that received CCR2+ myeloid cells from WGP-trained mice showed a significantly reduced tumor burden as compared to mice that received CCR2- myeloid cells from the same trained mouse.

As we show that peripheral trained immunity has been established in the pancreas, we ask whether these cells could be reactivated by tumor cells or their secreted factors. Re-exposure of in vivo WGP-trained pancreatic myeloid cells to tumor-conditioned media was shown to vigorously elicit a trained response. Kalafati et al. recently showed a role for granulocytes in trained immunity-driven anti-tumor mechanisms21, however, given their use of subcutaneous models of cancer, the currently reported study is the first instance suggesting that myeloid cells in a specific organ can be trained to react directly to tumor cells of that same organ. 

It has been well established that myeloid cells are capable of directly killing tumor cells through mechanisms of phagocytosis and ROS production50–54, and here we show that WGP training significantly upregulates the direct anti-tumor functionalities of enhanced phagocytosis of tumor cells and ROSmediated cytotoxicity to tumor cells.

Further, while we highlight that tumor-conditioned media can reactivate trained myeloid cells, we also identify a specific factor, MIF, in the tumor-conditioned media that is involved in this activation. These data suggest that the induction of trained immunity could participate in mechanisms of tumor immunosurveillance. For example, in the early stages of tumor development, tumors secrete soluble factors that can re-stimulate trained innate cells leading to the eradication of these tumor cells.

We then translate this understanding of the anti-tumor potential of the trained innate myeloid cells that enter the pancreas to orthotopic models of PDAC, where we observe a dramatic decrease in the tumor burden and an increase in the survival of mice given only one administration of WGP. We further demonstrate that the cells responsible for tumor control are mainly trained CCR2+ myeloid-derived cells. In contrast to other proposed models of tumor control which appear to be entirely driven by the induction of centrally trained immunity in the BM21, instead, in this model the early trafficking of WGP to the pancreas and the resulting production of CCL2 in the pancreas recruits trained CCR2+ myeloid cells from the periphery which are mainly responsible for the anti-tumor effects. WGP directly trafficking to the pancreas also repolarizes M2-like tissue-resident macrophages towards M1-like macrophages that may also restrain tumor progression. 

While the tumor reduction observed in mice implanted with KPC tumors that had been admixed with CCR2+ trained cells was significant, it was not as striking as in mice given IP WGP. This supports that the anti-tumor effects of IP WGP are twofold; the recruitment of CCR2+ trained anti-tumor myeloid cells to the pancreas and the repolarization of pro-tumorigenic resident macrophages work in concert to elicit anti-tumor innate responses that lead to reduced tumor burden in models of PDAC.

We also note the high expression of MHCII on these trained monocyte/macrophages, which likely plays a role in tumor antigen processing and communication with T-cells to elicit adaptive immune responses against the tumor. Interestingly, while we confirmed that these anti-tumor mechanisms do not depend on adaptive immune responses, we did observe a significant increase in CD8+ T-cells present within CCR2+ admixed tumors. Additionally, given that myeloid cells in the tumors were observed to display high levels of PD-L1 and that the presence of trained CCR2+ cells in the pancreas appeared to recruit CD8+ T-cells, we reasoned that WGP may potentiate the effects of anti-PD-L1 therapy. 

By the literature, anti-PD-L1 mAb therapy alone had no survival benefit while the combination of WGP and anti-PD-L1 synergized to produce better survival outcomes than WGP alone55. This suggests that trained innate immune cells may activate adaptive immune cells in the TME, however, the immunosuppressive microenvironment likely needs to be further overcome by the concomitate effects of anti-PD-L1. Importantly, despite improved survival, all mice did eventually succumb to the tumor, so future investigations into the optimization of this treatment in terms of dosing and the number of treatments are warranted.

Immunosuppressive myeloid cells play a critical role in the creation of the immunosuppressive TME in PDAC. Here, we demonstrate a capability to engage these myeloid cells in the setting of PDAC through the induction of trained innate immunity in the pancreas. 

Importantly, we also show that the ability to initiate trained immunity in the pancreas potentiates the therapeutic effects of immunotherapy such as anti-PD-L1 immuno-checkpoint blockade therapy. Lending further clinical applicability to these findings, we also exhibit that the induction of trained immunity in the adjuvant setting can decrease tumor size and prolong survival. These findings emphasize the potential of using β-glucan to therapeutically target myeloid cells within the TME and highlight that the induction of peripheral trained immunity in the pancreas could play a consequential role in reprogramming the suppressive TME of PDAC which could result in extending the life of patients diagnosed with this deadly malignancy.

Methods

Ethical statement.

All experiments involving animals were conducted according to the ethical guidelines set by the University of Louisville Institutional Animal Care and Use Committee. Experiments were conducted according to the approved protocols 19471 and 19536.

Mice.

Six- to eight-week-old male and female mice were used in all experiments. Wild-type (WT) C57BL/6 J mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA) or bred in the University of Louisville-specific pathogen-free (SPF) animal facility. C57BL/6 Dectin-1 knockout (Dectin-1−/−) mice were described previously56. CCR2 global knockout mice were purchased from Jackson Laboratory. Albino C57BL/6 mice were kindly provided by Dr. Jonathan Warawa at the University of Louisville. NOD/SCID/IL2rγNull (NSG) mice were purchased from the Jackson Laboratory. The maximal tumor size allowed by the IUCAC was no more than 2000 mm3. Mice were monitored daily for tumor growth and euthanized when the tumor burden exceeded the limited size. Mice were maintained on a 12-h dark/light cycle at room temperature with controlled humidity (around 55%). 

All animals were housed in a barrier facility and only healthy mice were used for experiments. All mice were at least 6 weeks of age upon us, and all experiments involving animals were performed in compliance with all relevant laws and institutional guidelines provided by the Rodent Rearing Facility (RRF) and approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Louisville.

Preparation of β-glucan.

Highly purified particulate β-glucan in the form of particulate whole β-glucan particles (WGP) isolated from Saccharomyces cerevisiae was provided by Biothera. Before use, WGP was gently sonicated for 15 s, two times using a Qsonica Q55-110 Q55 Sonicator (Cole-Parmer) to ensure aggregates were broken up.

Preparation and use of DTAF-WGP.

DTAF (Sigma-Aldrich) at 2 mg/mL was mixed with a suspension of 20 mg/mL WGP in borate buffer (pH 10.8). This was incubated at room temperature for 8 h with continuous mixing. Following incubation, the WGP was centrifuged and washed with cold sterile endotoxin-free DPBS (Sigma-Aldrich) five times or until the supernatant no longer contained visible DTAF. The concentration was adjusted to 10 mg/mL in the endotoxin-free DPBS for storage. About 1 mg of the DTAF-WGP has injected IP into C57BL/6 and Dectin-1−/− mice and organs were harvested 3 days later.

Preparation of the 89Zr-WGP.

WGP (100 mg) was mixed with DeferoxamineSCN (2.7 mg, in 0.8 ml DMSO) and suspended in 10 ml sodium carbonate buffer (0.1 M, pH 9.4) overnight at room temperature in the dark with gentle shaking. The Deferoxamine-labeled WGP was then washed with DI water (10 × 40 mL), and 30 mg of Deferoxamine-labeled WGP was mixed with 3 mCi of 89Zr oxalate in 2 ml Tris•HCl buffer (0.5 M, pH 7.5), and then incubated at 37 oC for 60 min with shaking. The 89Zr-WGP was then centrifuged and washed with 3 ml of sterile PBS. The radioactivity of 89Zr-WGP was measured by a dose calibrator and used for in vitro and in vivo studies.

Biodistribution and PET/CT scan using 89Zr-WGP.

Positron Emission tomography (PET)/computed tomography (CT) imaging was conducted in C57BL/6 and Dectin-1−/− mice 48 h after IP injection of 1 mg of pure 89Zr-WGP or injection of 1 × 106 peritoneal macrophages that had been co-cultured with 25 μg/ ml of 89Zr-WGP for 2 h and then gently washed to remove excess 89Zr-WGP. The mice were scanned for 15 min with a Siemens R4 MicroPET followed by 10 min of CT scan. Siemens IAW software was used for the acquisition and reconstruction of the PET signal, and Siemens IRW software was used for merging and analyzing the imaging data. 

At the end of the imaging study, mice were euthanized, and organs of interest were harvested. For biodistribution, 50 uL of peripheral blood was collected using a retro-bulbar bleeding technique. The brain, heart, lungs, liver, spleen, kidneys, pancreas, large intestine, small intestine, stomach, femur, a piece of skin from the flank of the mice, and the rectus femoris muscle were harvested, weighed, and placed in a 2470 Wizard automatic gamma counter (PerkinElmer) to measure the radioactivity of each tissue. The CPM values were calculated using Prism software (GraphPad Software, La Jolla, CA).

Pancreatic processing.

Following euthanization, the mouse pancreas was harvested and gently cut into smaller pieces using sterile scissors. They were suspended in a 15 mL tube in complete media (RPMI) with 1X digestion buffer comprised of 300 U/ml collagenase I, 60 U/ml Hyaluronidase, and 80 U/ml DNase (Sigma). These were placed in a rotating incubator at 37 °C with 5% CO2 for 15–20 min. The digestion buffer was then quenched with ice-cold complete RPMI 1640 and washed. Cells were passed through a sterile nylon 40 µm basket filter and small undigested pieces of tissue were smashed using a sterile syringe stopper to generate a single-cell suspension. If an appreciable number of red blood cells (RBCs) were seen to exist in the sample, RBC lysis was performed by adding 2 mL of sterile 10x ACK (Thermo Fisher Scientific).

In vivo WGP administration.

Mice were given a single 1 mg intraperitoneal dose of gently sonicated WGP, 3 μm polystyrene beads (Sigma-Aldrich), or 3 μm fluorescent microspheres (Polysciences) (all 1 mg in 200 μl of sterile PBS) or 200 μl of sterile PBS on day 0. For typically trained immunity studies, 7 days following the initial IP dose, mice were euthanized using CO2, and the pancreas along with other tissues of interest was removed and processed. For dose-titration studies, 0.5, 1, and 2 mg of WGP were delivered IP in 200 uL of sterile PBS. For time titration studies, mice were injected with 1 mg of WGP, and the pancreas was harvested after 24 h, 48 h, 3, 7, 10, 16, and 30 days later.

Ex vivo restimulation.

To assess the trained phenotype of mononuclear cells in mice treated with WGP or PBS ex vivo, after processing the pancreas, pancreatic cell suspensions were plated in 24 well plates and stimulated with LPS (10 ng/ml), the supernatant from cultured KPC and Pan02 cells (40%), and recombinant MIF (rMIF)(10 ng/ml). MIF was a generous gift from Dr. Robert Mitchell at the University of Louisville and was prokaryotic ally expressed, purified, and refolded as described previously57. Cells were cultured in DMEM and incubated at 37 °C with 5% CO2 for 5–6 h in the presence of 1X brefeldin A (BioLegend). The cells were then harvested using a cell scraper, washed, pelleted, and then stained for intracellular cytokine expression.

In vitro training and restimulation assay.

Peritoneal macrophages or sorted CD11b + cells from a mouse pancreas were plated in a 24-well plate for 2 h at 37 °C and 5% CO2 to allow for the attachment of cells to the plates, after which the floating cells were gently aspirated. The attached cells were gently washed with sterile PBS and then resuspended in 1 mL of DMEM constituted of 10% FBS and 1% penicillin/streptomycin. 

For the initial training of cells, 25 μg/ml of particulate WGP, 25 μg/ml of 3 μm polystyrene microparticle beads (Sigma-Aldrich), or 100 uL of PBS were added to the appropriate well and incubated for 24 h. After 24 h, wells were gently washed to remove the initial stimulus and fresh DMEM was added and cells were incubated with 5% CO2 at 37 °C for 7 days. After 7 days, the media was aspirated, and replaced with fresh media, and cells were restimulated with LPS (100 ng/mL), the supernatant of KPC or Pan02 cells (40%), or PBS as a control. Twenty-four hours after stimulation, the supernatants were harvested and used in an ELISA for TNF-α and IL-6.

Tumor-conditioned media.

About 1 × 106 KPC or Pan02 cells were cultured in a six-well plate in 4 mL of complete DMEM at 37 °C and 5% CO2. After 3 days the supernatants were harvested and stored at −80 °C in aliquots for use as a tumor-conditioned medium. As a control, the pancreas of a C57BL/6 mouse was processed into a single cell suspension and 1 × 106 of these cells were cultured in a six-well plate in 4 mL of complete DMEM and at 37 °C and 5% CO2, and supernatants were also collected after 3 days. As a control, the pancreas of naïve mice was processed into a single-cell suspension and plated in a six-well plate for 1 day. Non-adherent cells were washed away and these cells were then cultured for 3 days and the supernatant was harvested.

Acquisition of peritoneal macrophages.

Mice were euthanized and 5 mL of sterile RPMI was injected into the peritoneum using a 25-gauge syringe. The abdomen was massaged to liberate the peritoneal macrophages. A small incision was made and a transfer pipette was used to remove the suspension. The peritoneal cavity was then washed several times with cold RPMI and cells were pelleted at 458 × g.

Flow cytometry.

Single-cell suspensions in PBS with 1% FBS were blocked with murine Fc Block (anti-CD16/CD32) at 4 °C for 15 min. Fluorochrome labeled antibodies for viability (APC/Cy7) and to surface markers CD45, CD11b, F4/80, Ly6G, Ly6C, MHCII, CD3, CD4, CD8, NK1.1, Ki67 (Biolegend), and CCR2 (R&D Systems) were used. The information related to all flow Abs is summarized in Supplementary Table 3. After 30 min of incubating in the dark at 4 °C, cells were washed with cold PBS and filtered through a 40 µm mesh filter. The samples were acquired using a FACSCanto II cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star, Ashland, OR). FACSDiva software version 6 was used in FACSCanto to acquire all data.

intracellular staining for the expression of cytokines.

Following stimulation, cells were stained for the desired surface markers as described above. The cells were then washed with cold PBS, and 500 μl of fixation buffer (Biolegend) was added to the tubes, briefly vortexed, and incubated in the dark at room temperature for 20 min. About 1 ml of Permeabilization buffer (Biolegend) was then added and samples were centrifuged at 458 × g for 5 min at 4 °C followed by one more wash using 1 ml of permeabilization buffer. Cells were resuspended in 200 μl of permeabilization buffer and cells were stained with antibodies against TNF-α, Ki67, Granzyme-B, IL12, IL-6, and IFNγ or the respective isotype control overnight at 4 °C. Cells were then washed, filtered, and data were acquired using a flow cytometer.

FACS isolation of CD11b+macrophages for in vitro training.

The pancreas was processed into a single-cell suspension as described above. Cells were washed with 1 mL of PBS, and incubated with Fc block for 10 min at 4 °C followed by staining with viability dye (APC-Cy7), CD45 (PerCPcy5), and CD11b + (APC) for 30 min at 4 °C. Cells were washed with PBS and resuspended in cold MACS Running Buffer (Miltenyi Biotech). ViabilityCD45+CD11b+ cells were sorted using a FACS Aria III (BD Biosciences). Cells were collected in 50% FBS, 40% PBS, and 10% HEPES (Corning). After sorting the cells were washed with PBS and then plated for in vitro training, as described above.

ELISA.

Supernatants from in vitro trained cells along with standards were analyzed using murine TNF-α and IL-6 ELISA kits (BioLegend). The assay was performed per the manufacturer’s instructions and all conditions were performed in triplicates. An ELISA kit for rMIF (R and D systems) was also used according to the manufacturer’s instructions to quantify rMIF in Pan02 and KPC tumor-conditioned media and the supernatant from cultured untreated pancreatic cells. To quantitate histone modifications, ELISA kits (Tri-methyl histone H3K4 quantification kit, Acetyl-histone H3K27 quantification kit, Tri-methyl histone H3K27 quantification kit, and total histone H3 quantification kit) were purchased from EpiGentek and assays were performed based on the manufacturer’s instructions. Lactate was measured by L-Lactate assay Kit I (Eton Bioscience).

Western blot analysis.

CD11b+ cells from PBS-injected or WGP-injected mice (24 h) were separated using CD11b microbeads (Miltenyi Biotech). Histone was extracted according to the manufacturer’s instruction (Active Motif, Inc.), and protein concentration was determined using Bradford quantification. Histone proteins were separated by SDS-PAGE 15% Tris-HCl gels and transferred onto PVDF membranes (Millipore). The membrane was blocked with 5% BSA at room temperature for 1 h and incubated overnight at 4 °C with primary antibodies including tri-methyl-histone H3 (lys4) rabbit mAb, acetyl-histone H3 (lys27) XP Rabbit mAb, tri-methyl-histone H3 (Lys27) Rabbit mAb, and histone H3 XP rabbit mAb (Cell Signaling Technology), then incubated with HRP-conjugated secondary antibodies (GE Healthcare) at room temperature for 1 h. The membrane was developed with Amersham ECL Prime Western Blotting Detection Reagent (GE Healthcare) and detected through Medical Film Processor (Konica Minolta Medical & Graphic). Precision Plus Protein Kaleidoscope Prestained Protein Standards were used as a standard protein marker (Bio-Rad).

qRT-PCR.

After CD11b + cells had been isolated from the mouse pancreas in WGP-treated and untreated mice, cells were saved in TRIzol. RNAs were isolated and reverse transcribed using the TaqMan Reverse Transcription Reagents (qRTPCR) amplification using the Bio-Rad MyiQ single-color RT-PCR detection system. Briefly, cDNA was amplified in a 25 uL reaction mixture consisting of SYBR Green PCR super-mix (Bio-Rad), 100 ng of complementary DNA template, and selected primers (200 nM) using the recommended cycling conditions.

H+E.

Seven days following injection with PBS/microparticle beads or WGP, the pancreata were harvested and fixed in 4% formalin for 1 week followed by embedding in paraffin according to standard procedures. Paraffin-embedded tissues were cut into 5 mm thick sections and stained with hematoxylin and eosin (H & E) for morphological analysis.

Serum amylase measurement.

Murine serum amylase was measured using the Amylase Activity Assay Kit (Millipore Sigma) and was used according to the manufacturer’s instructions. In short, mice were injected with WGP and 7 days later, blood was collected from mice using a retro-bulbar bleeding technique. These samples were used to assay for serum amylase.

In vivo T-cell depletion.

T-cells were depleted using an anti-CD4 mAb alone, antiCD8 mAb alone, or anti-CD4 and anti-CD8 mAbs together. Antibodies were made in-house. In this depletion procedure, mice have injected IP with WGP on day 1 and were also injected IP with 200 µg of the mAbs on day 1 and day 4 during the treatment period. Mice were euthanized on day 7. (CD4 clone GK1.5, CD8 clone 53-6.72). Depletion efficiency was confirmed on day 7.

In vivo NK cell depletion.

NK cells were depleted through the intraperitoneal injection of 100 µg of PK136 mAb (Produced in the laboratory of Dr. Jun Yan at the University of Louisville) on days −1 and 5 during the treatment period. WGP was injected on day 0 and day 7 animals were euthanized and the depletion efficiency of NK cells was assessed by staining pancreatic tissues for NK1.1.

Ly6G depletion.

Neutrophils were depleted by injecting 300 µg of anti-Ly6G mAb (Bio X Cell) or isotype control Rat IgG2a (Bio X Cell) at days −1, 2, and 6 during treatment. About 1 mg of WGP injected IP on day 1. Mice were euthanized on day 7 and the pancreas was assessed for efficiency of depletion. In the tumor, model mice were injected with 300 µg of anti-Ly6G mAb or isotype control Rat IgG2a on days −2, 4, 10, and 16. Mice were given WGP on day 0 and were implanted with orthotopic KPC pancreatic tumors on day 7. Mice were euthanized on day 21 and pancreatic tissues were stained with Ly6G to assess granulocyte depletion efficiency at that time.

CyTOF mass cytometry sample preparation.

Mass cytometry antibodies were either purchased from Fluidigm or were created in-house by conjugating commercially available purified antibodies to the appropriate metal isotope using the MaxPar X8 Polymer or MCP9 Polymer kits (Fluidigm). Pancreatic samples from three PBS and three 7-day WGP mice were processed into a single-cell solution and ex vivo stimulation was performed as described above. Cells were gently scraped from the plates using a sterile cell scraper, washed with PBS, and placed into a sterile culture tube. About 2 × 106 cells per sample were used. Cells were first stained for viability with 5 uM cisplatin (Fluidigm) in serum-free RPMI 1640 for 5 min at RT. Cells were then washed with RPMI 1640 containing 10% FBS for 5 min at 300 × g. Cells were stained with the surface antibodies for 30 min at RT and washed twice with Maxpar Cell staining buffer (Fluidigm). 

For staining on intracellular cytokines, cells were then fixed with 1 mL of 1X Maxpar Fix I buffer for 30 min at RT and then washed twice with 2 mL of 1X Maxpar Perm-S buffer for 5 min at 800 × g. The cytoplasmic/secreted antibody cocktail was then added and incubated with the cells for 30 min at RT. Following incubation, cells were washed with 1 mL of 1X Maxpar Perm-S buffer for 5 min at 800 × g and gently blotted to remove all liquid from the tube. 

To stain for nuclear antigens, cells were then suspended in 1 mL of 1X Maxpar nuclear antigen staining buffer for 30 min at RT. The nuclear antigen antibody cocktail was then added and incubated for 30 min at RT. Cells were washed twice for 5 min at 800 × g with 2 mL of Nuclear Antigen Staining Permeability buffer. Finally, cells were fixed with 1.6% formaldehyde for 10 min at RT, then incubated overnight in 125 nM of IntercalatorIridium (Fluidigm) at 4 °C.

CyTOF data acquisition.

Once cells were ready for acquisition, samples were washed twice with Cell Staining Buffer (Fluidigm) and kept on ice while awaiting acquisition. Directly before the acquisition, cells were suspended in a 1:9 solution of Cell Acquisition Solution: EQ 4 element calibration beads (Fluidigm). A Helios CyTOF system was used, and following proper startup and tuning procedures, samples were run at a rate of less than or equal to 500 events/second up to 300,000 events. Using the CyTOF software, FCS files were normalized.FCS files and these files were then ready for data analysis.

CyTOF data analysis.

CyTOF data was analyzed using FlowJo, the CytoBank software package58, and the CyTOF workflow59 which includes a suite of packages available in R (r-project. org)60–63. For analysis conducted within the CyTOF workflow, FlowJo workspace files exported from Flow Workspace and CytoML were used61.

RNA sequencing: RNA extraction and isolation.

Seven days following administration of PBS or WGP IP, pancreata were harvested, processed into a single cell suspension, stained for viability, CD45, and CD11b, and sorted using a FACS Aria III. Samples were prepared in triplicate for each experimental group. Once these myeloid cells were isolated, cells were washed 2x with ice-cold PBS and then lysed with Trizol (Invitrogen). RNAs were extracted using a QIAGEN RNAeasy Kit (QIAGEN). 

The isolated RNA was checked for integrity using the Agilent Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA) and quantified using a Qubit fluorometric assay (Thermo Fisher Scientific, Waltham, MA). Poly-A enriched mRNA-Seq libraries were prepared following the Universal Plus mRNASeq kit standard protocol (Tecan Genomics, Redwood City, CA) using 10 ng of total RNA. 

All samples were ligated with Illumina adapters and individually barcoded. The absence of adapter dimers and a consistent library size of ~300 bp was confirmed using the Agilent Bioanalyzer 2100. The library concentration and sequencing behavior were assessed about a standardized spike-in of PhIX using a Nano MiSeq sequencing flow cell from Illumina. 1.8 pm of the pooled libraries with 1% PhiX spike-in was loaded on one NextSeq 500/550 75 cycles High Output Kit v2 sequencing flow cell and sequenced on the Illumina NextSeq 500 sequencer targeting 60 M 1x75 bp reads per sample.

RNA sequencing.

Libraries were prepared using the Universal Plus mRNA-seq kit with NuQuant® library quantification (NuGen). The six samples were spread across four sequencing lines in one run. The 24 single-end raw sequencing files (.fastq)64 were downloaded from Illumina’s BaseSpace65 onto the KBRIN server for analysis. Quality control (QC) of the raw sequence data was performed using FastQC (version 0.10.1)66. The sequences were aligned to the mm10 mouse reference genome using STAR (version 2.6)67, generating alignment files in bam format. Differential expression of ENSEMBL protein-coding transcripts was performed using DESeq268,69. Raw counts were obtained from the STAR-aligned bam format files using HTSeq (version 0.10.0)70. The raw counts were normalized using the Relative Log Expression (RLE) method and then filtered to exclude genes with fewer than ten counts across the samples71. RNA-seq data were deposited with GEO accession GSE187464.

RNA sequencing: gene set enrichment analysis.

Gene set enrichment analysis (GSEA) was used to further characterize the biology of the genes comprising the WGP vs PBS conditions and their differences72. Gene sets were obtained from the Molecular Signatures Database (MSigDB) for Gene Ontology (GO) biological processes and Reactome pathways. For each gene set, all tested gene locations in the comparison of WGP vs PBS-treated control mice were sorted from highest to lowest significance using p values. This approach allows highly significant up and downregulated genes to be included within each gene set, an approach that more accurately reflects the conditions in a biological pathway. For this analysis, a table of the enriched sets is followed by an enrichment plot displaying the profile of the enrichment score (ES) and the position of gene set members on the rank-ordered list.

Single-cell sequencing: isolation of single cells and RNA sequencing.

Live CD45 + cells were sorted from mouse pancreata, washed, and resuspended in 1x PBS (calcium and magnesium-free) containing 0.04% BSA. Single cells were captured and barcoded cDNA libraries were constructed using the Chromium Next GEM Single-Cell 3ʹ Reagent Kit (v3.1, 10X Genomics) and the Chromium Controller, according to the manufacturer’s instructions. Libraries were pooled and sequenced using a 28 bp × 8 bp × 125 bp configuration for read1 × i7 index x read2 on the Illumina NextSeq 500 with the NextSeq 500/550 150 cycle High Output Kit v2.5 (20024907).

Since cell sequencing: gene expression profiling.

Bcl files were demultiplexed into fastq files using the CellRanger software (10X Genomics, v3.1.0). The total number of sequenced reads was 506,913,062. The reads were of good quality as determined by FastQC66. Gene counts were measured using CellRanger “count”, utilizing the cell ranger-mm10-3.0.0 reference genome for mice. A counts matrix was generated for each sample and one aggregated sample with the expected number of cells set at 5000.

The raw count data determined by CellRanger was used as input to a custom analysis pipeline in R which uses a variety of single-cell analysis tools based on Seurat. The knee plot (Supplementary Fig. 7b) displays a graph showing the ranked UMI counts for each cell barcode for data aggregated across the three groups. Cells above the inflection point represent possible doublets while those below the knee represent background cells. Cell quality control measures were analyzed using Seurat v373, and cell barcodes with the following characteristics were removed from the analysis: low counts (possible background cells) with an FDR cutoff of 0.01 from the DropletUtils74 function “emptyDrops”, high counts (possible doublet cells) with more counts than the knee plot inflection point, mitochondrial content greater than 30% and ribosomal content greater than 40%. Gene (features) were further filtered to remove retired gene identifiers and genes that were not expressed in at least two cells. See Supplementary Fig. 7a, b for the initial and filtered number of cell barcodes and genes.

The expression data were normalized using SCTransform75 where cell cycle genes, ribosomal content, and mitochondrial content were regressed. The cells were then clustered and dimension reduction was performed using UMAP76. Initial cluster names were assigned using a modified GSVA77,78 enrichment score technique. For each of these clusters, the top marker genes were identified. Differentially expressed genes comparing each cluster to every other cluster (all pairwise comparisons) were determined using Seurat73 and MAST79. scRNA-seq data were deposited with GEO accession GSE187464.

Identification of clusters generated with scRNA-Seq.

Non myeloid-derived clusters were classified generally as B-cells (MS4A1), plasma cells (SDC1), CD8 + T-cells (CD3e, C8a), CD4 + T-cells (CD3e, CD4),T-regulatory cells (T-regs) (CD3e, CD4, FoxP3), γδ T-cells (CD3e, TRGC1, TRGC2,IL7Ra) and type 2 innate immune cells (ILC2s) (Alox5, KLRG1, Ly6a, Pparg,GATA3, IL-5, IL-13, ICOS and Rxrg)80–82. Neutrophils were identified through MMP9, Csf3r, S100A8, S100A8, and ADAM8 expression, though may be underrepresented in these analyses due to their low RNA content and high levels of intrinsic RNases83. Conventional dendritic cells (cDCs) were identified through ITGAX and ITGAE expression and plasmacytoid DCs (pDCs) were identified by ITGAX and Siglech.

Classification of myeloid clusters from scRNA-Seq.

Cluster 5 expressed ITGAMIntADGRE1HiLyz2HiH2-Ab1HiLy6C2- and did not express TNFAIP2, indicating that these cells are resident macrophages. Cluster10 expressed ITGAM- HiADGRE1HiLyz2Hi H2-Ab1Int Ly6C2-. Clusters 3 and 4 expressed ITGAMInt ADGRE1Int Lyz2Hi H2-Ab1Int Ly6C2Hi, which suggests that both are subsets of infiltrating monocytes/macrophages, though the enhanced inflammatory genes expressed in cluster 3 were used to identify cluster 3 as inflammatory infiltrating monocyte/macrophage.

Phagocytosis assays.

The pancreas of PBS/microparticle-injected mice and in vivo WGP-trained mice were harvested 7 days after injection and processed as described previously into a single cell suspension. About 2 × 106 were washed with HEPES dilutes 50x in RPMI 1640 and then incubated in 100 μL of this solution for 1 h at 37 °C to activate the cells. The Invitrogen promotes Green S.

Aureus BioParticlesTM Phagocytosis Kit for Flow cytometry (Thermo Fisher Scientific) was used according to the manufacturer’s instructions. About 100 μL of the reconstituted particles or 1 × 106 GFP + KPC tumor cells were added to the activated pancreatic cells and incubated for 1 h at 37 °C. Samples were gently vortexed every 15 min. The reaction was stopped by adding 1 mL of cold PBS. Samples were incubated with Fc Block for 10 min at 4 °C, stained for viability, CD45, CD11b, and F4/80 for 30 min at 4 °C, and then analyzed using a BD FACSCanto. For analysis, after gating on live cells and CD45, cells that fluoresced in the FITC channel were determined to be phagocytic.

Cytotoxicity assay.

The pancreas from WGP and PBS-treated mice were harvested and the CD11b + populations were isolated using magnetic CD11b + MicroBeads (Miltenyi Biotec) and an autoMACS Pro Separator (Miltenyi Biotec). Purified CD11b + cells were then washed and counted, and these were plated at a ratio of 1:20 tumor: effector cells in a 96-well plate. All experimental samples were run in triplicate. After plating the CD11b + cells, the ROS inhibitors N-acetyl-Lcysteine (NAC) (1 mM, Sigma-Aldrich), Trolox (5 μM, Sigma-Aldrich), or DFO (50 μM, European Pharmacopoeia) were added to one set of PBS and WGP derived cells for 1 h before the addition of 5000 luciferase-expressing KPC + pancreatic tumor cells to all wells. After 24 h, of co-culture, the plates were centrifuged and 20 µL of the supernatant was mixed with 100 µL of the Luciferase Assay Reagent (Promega). Luciferase activity measured in the supernatant correlated with tumor cells that had been killed by the effector cells and were measured using a luminometer (Femtomaster FB 12, Zylux Corporation). The spontaneous luciferase signal from plated tumor cells was subtracted from the measurement of the supernatant. Luciferase values are represented as Relative Light Units (RLUs).

In vivo tumor models of pancreatic cancer.

A KPC cell line on a C57BL/6 background derived from the LSL-KrasG12D/+; LSL-Trp53R172H/+; Pdx1-Cre (KPC) mouse model was purchased from Ximbio. These and Pan02 cells which were a generous gift from Dr. Yong Lu at Wake Forest University were used in an orthotopic model of pancreatic cancer. A KPC line transfected with GFP and luciferase (KPCGFP+Luc+) were also a generous contribution from Dr. Michael Dwinell at the Medical College of Wisconsin. These cells were used exclusively in albino C57BL/6 mice. 

For tumor implantation, mice were anesthetized using isoflurane, and the abdomen of the mice was prepped with betadine and draped in a sterile fashion. A 2 cm midline laparotomy was performed using an aseptic technique with sterile instruments. Following laparotomy, the pancreas, and spleen were externalized. Tumor cells were suspended in ice-cold PBS and mixed in a 1:1 ratio with basement membrane matrix Matrigel (Corning). 0.1 × 106 tumor cells in 50 uL of the PBS-matrigel solution were injected into the tail of the pancreas using a 30-gauge insulin syringe. The formation of a small bubble indicated successful implantation. The peritoneum was closed using coated polyglycolic acid braided absorbable 5/0 suture and the skin was closed using silk braided nonabsorbable 5/ 0 suture. (CP Medical). Buprenorphine was administered for pain management up to 72 h following surgery and mice were monitored.

In vivo imaging.

Mice implanted orthotopically with GFP + Luciferase+ KPC tumor cells were injected IP with 150 mg/kg of body weight at 100 µL of XenoLight D-Luciferin-K + Salt Bioluminescent Substrate (PerkinElmer). After 10 min, mice were anesthetized with isoflurane and placed inside a Biospace Lab Photon Imager, which is a dedicated low-light level in vivo optical modality for bioluminescent and fluorescent imaging. Images of mice were taken and used to measure tumor size and growth.

Admixture tumor model.

Mice were trained with 1 mg of WGP and 7 days later the CD11b+CCR2+ and CD11b+CCR2− populations were sorted and mixed 1:1with KPC tumor cells 1 × 105 tumor cells. Cells were implanted orthotopically into WT mice and 3 weeks later mice were euthanized and tumor size was assessed.

Combination therapy with anti-PD-L1 mAb.

C57BL/6 mice were treated with WGP on day −7 and implanted with KPC orthotopic pancreatic tumors on day 0. Mice were treated with 200 µg of anti-mouse anti-PD-L1 mAb (Bio X Cell) or Rat IgG2b isotype control (Bio X Cell) on days 3, 7, and 11. Mice were then monitored for survival.

WGP as a treatment.

C57BL/6 mice were implanted with KPC orthotopic pancreatic tumors at day 0 and on days 4 and 11 mice were given 1 mg of IP WGP or PBS. Mice were then monitored for survival.

Statistical analysis.

All results were repeated at least three independent times to verify unless otherwise specified. Representative results are shown. Results are represented as mean ± SEM. Data were analyzed using a two-tailed Student’s t-test or Mann–Whitney U-test. Multiple-group comparisons were performed using a one-way or two-way ANOVA followed by Tukey’s multiple comparisons tests. Correlation analyses were performed using the Pearson correlation coefficient (normal distribution). Statistical significance was set at p < 0.05. All statistical analyses were performed using GraphPad Prism Software Version 9 (GraphPad Inc., La Jolla, CA).

Reporting summary.

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

Source data are provided in this paper. RNASeq data are available under GEO accession GSE187464, and scRNA-Seq data are available under GEO accession GSE187464 Source data are provided in this paper.

Received: 2 April 2021; Accepted: 17 January 2022.

References

1. ACS. Cancer Facts & Figures 2020.

2. Rahib, L. et al. Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States. Cancer Res. 74, 2913–2921 (2014). 

3. Kommalapati, A., Tella, S. H., Goyal, G., Ma, W. W. & Mahipal, A. Contemporary management of localized resectable pancreatic cancer. Cancers 10, 24 (2018). 

4. Bliss, L. A., Witkowski, E. R., Yang, C. J. & Tseng, J. F. Outcomes in operative management of pancreatic cancer. J. Surg. Oncol. 110, 592–598 (2014). 

5. Schizas, D. et al. Immunotherapy for pancreatic cancer: a 2020 update. Cancer Treat. Rev. 86, 102016 (2020). 

6. Henriksen, A., Dyhl-Polk, A., Chen, I. & Nielsen, D. Checkpoint inhibitors in pancreatic cancer. Cancer Treat. Rev. 78, 17–30 (2019). 

7. Madden, K. & Kasler, M. K. Immune checkpoint inhibitors in lung cancer and melanoma. Semin. Oncol. Nurs. 35, 150932 (2019). 

8. Kao, H. F. & Lou, P. J. Immune checkpoint inhibitors for head and neck squamous cell carcinoma: Current landscape and future directions. Head. Neck 41, 4–18 (2019). 9. Leinwand, J. & Miller, G. Regulation and modulation of antitumor immunity in pancreatic cancer. Nat. Immunol. 21, 1152–1159 (2020). 

10. Royal, R. E. et al. Phase 2 trial of single agent Ipilimumab (anti-CTLA-4) for locally advanced or metastatic pancreatic adenocarcinoma. J. Immunother. 33, 828–833 (2010). 

11. O’Reilly, E. M. et al. Durvalumab with or without tremelimumab for patients with metastatic pancreatic ductal adenocarcinoma: a phase 2 randomized clinical trial. JAMA Oncol. 5, 1431–1438 (2019). 

12. Weiss, G. J. et al. Phase Ib/II study of gemcitabine, nab-paclitaxel, and pembrolizumab in metastatic pancreatic adenocarcinoma. Invest. New Drugs 36, 96–102 (2018). 13. Whatcott, C. J., Posner, R. G., Von Hoff, D. D. & Han, H. in Pancreatic Cancer and Tumor Microenvironment (eds Grippo, P. J. & Munshi, H. G.) Ch. 8 (Transworld Research Network, 2012). 

14. Karamitopoulou, E. Tumour microenvironment of pancreatic cancer: the immune landscape is dictated by molecular and histopathological features. Br. J. Cancer 121, 5–14 (2019). 

15. Daniel, S. K., Sullivan, K. M., Labadie, K. P. & Pillarisetty, V. G. Hypoxia as a barrier to immunotherapy in pancreatic adenocarcinoma. Clin. Transl. Med. 8, 10–10 (2019). 

16. Wang, W. et al. RIP1 kinase drives macrophage-mediated adaptive immune tolerance in pancreatic cancer. Cancer Cell 38, 585–590 (2020). 

17. Netea, M. G. et al. Trained immunity: a program of innate immune memory in health and disease. Science 352, aaf1098 (2016). 

18. Bekkering, S. et al. Metabolic induction of trained immunity through the mevalonate pathway. Cell 172, 135–146 e139 (2018). 

19. van der Heijden, C. et al. Epigenetics and trained immunity. Antioxid. Redox Signal. 29, 1023–1040 (2018). 

20. Priem, B. et al. Trained immunity-promoting nanobiology therapy suppresses tumor growth and potentiates checkpoint inhibition. Cell 183, 786–801 e719 (2020). 

21. Kalafati, L. et al. Innate immune training of granulopoiesis promotes antitumor activity. Cell 183, 771–785 e712 (2020). 

22. Mitroulis, I. et al. Modulation of myelopoiesis progenitors is an integral component of trained immunity. Cell 172, 147–161.e112 (2018). 

23. Bekkering, S., Dominguez-Andres, J., Joosten, L. A. B., Riksen, N. P. & Netea, M. G. Trained immunity: reprogramming innate immunity in health and disease. Annu Rev. Immunol. 39, 667–693 (2021).

24. Kierdorf, K., Prinz, M., Geissmann, F. & Gomez Perdiguero, E. Development and function of tissue-resident macrophages in mice. Semin. Immunol. 27, 369–378 (2015). 25. Calderon, B. et al. The pancreas anatomy conditions the origin and properties of resident macrophages. J. Exp. Med. 212, 1497–1512 (2015). 

26. Kim, H. The transcription factor MafB promotes anti-inflammatory M2 polarization and cholesterol efflux in macrophages. Sci. Rep. 7, 7591 (2017). 

27. Kim, S. Y. et al. Anti-inflammatory effects of apocynin, an inhibitor of NADPH oxidase, in airway inflammation. Immunol. Cell Biol. 90, 441–448 (2012). 

28. Liu, M. et al. Transcription factor c-Maf is a checkpoint that programs macrophages in lung cancer. J. Clin. Invest. 130, 2081–2096 (2020). 

29. Zimmerman, K. A. et al. Single-cell RNA sequencing identifies candidate renal resident macrophage gene expression signatures across species. J. Am. Soc. Nephrol. 30, 767–781 (2019). 

30. Bakos, E. et al. CCR2 regulates the immune response by modulating the interconversion and function of effector and regulatory T cells. J. Immunol. 198, 4659–4671 (2017). 

31. Kurihara, T., Warr, G., Loy, J. & Bravo, R. Defects in macrophage recruitment and host defense in mice lacking the CCR2 chemokine receptor. J. Exp. Med. 186, 1757–1762 (1997). 

32. Kuziel, W. A. et al. Severe reduction in leukocyte adhesion and monocyte extravasation in mice deficient in CC chemokine receptor 2. Proc. Natl Acad. Sci. USA 94, 12053–12058 (1997). 

33. Kratofil, R. M., Kubes, P. & Deniset, J. F. Monocyte conversion during inflammation and injury. Arterioscler. Thromb. Vasc. Biol. 37, 35–42 (2017). 

34. Funamizu, N. et al. Macrophage migration inhibitory factor induces epithelial to mesenchymal transition, enhances tumor aggressiveness, and predicts clinical outcomes in resected pancreatic ductal adenocarcinoma. Int. J. Cancer 132, 785–794 (2013). 

35. Denz, A. et al. Inhibition of MIF leads to cell cycle arrest and apoptosis in pancreatic cancer cells. J. Surg. Res. 160, 29–34 (2010). 

36. Su, L. J. et al. Reactive oxygen species-induced lipid peroxidation in apoptosis, autophagy, and ferroptosis. Oxid. Med. Cell Longev. 2019, 5080843 (2019). 

37. Geller, A., Shrestha, R. & Yan, J. Yeast-derived β-glucan in cancer: novel uses of a traditional therapeutic. Int. J. Mol. Sci. 20, 3618 (2019). 

38. Shrivastava, P. & Bhatia, M. Essential role of monocytes and macrophages in the progression of acute pancreatitis. World J. Gastroenterol. 16, 3995–4002 (2010). 

39. Wang, J. & Kubes, P. A reservoir of mature cavity macrophages that can rapidly invade visceral organs to affect tissue repair. Cell 165, 668–678 (2016). 

40. Qi, C. et al. Differential pathways regulating innate and adaptive antitumor immune responses by particulate and soluble yeast-derived β-glucans. Blood 117, 6825–6836 (2011). 

41. Goodridge, H. S. et al. Activation of the innate immune receptor Dectin-1 upon formation of a ‘phagocytic synapse’. Nature 472, 471–475 (2011). 

42. Masuda, Y. et al. Oral administration of soluble β-glucans extracted from Grifola frondosa induces a systemic antitumor immune response and decreases immunosuppression in tumor-bearing mice. Int. J. Cancer 133, 108–119 (2013). 

43. Rice, P. J. et al. Pharmacokinetics of fungal (1-3)-beta-D-glucans following intravenous administration in rats. Int. Immunopharmacol. 4, 1209–1215 (2004). 

44. Huang, Y. et al. S1P-dependent interorgan trafficking of group 2 innate lymphoid cells supports host defense. Science 359, 114–119 (2018). 

45. Mathä, L. et al. Migration of lung resident group 2 innate lymphoid cells link allergic lung inflammation and liver immunity. Front. Immunol. 12, 67950 (2021). 

46. Ricardo-Gonzalez, R. R. et al. Tissue-specific pathways extrude activated ILC2s to disseminate type 2 immunity. J. Exp. Med. 217, e2019117 (2020). 

47. Mattiola, I. & Diefenbach, A. Enabling anti-tumor immunity by unleashing ILC2. Cell Res. 30, 461–462 (2020). 

48. Qian, B. Z. & Pollard, J. W. Macrophage diversity enhances tumor progression and metastasis. Cell 141, 39–51 (2010). 

49. Qian, B. Z. et al. CCL2 recruits inflammatory monocytes to facilitate breast tumor metastasis. Nature 475, 222–225 (2011). 

50. Munn, D. H. & Cheung, N. K. Phagocytosis of tumor cells by human monocytes cultured in recombinant macrophage colony-stimulating factor. J. Exp. Med. 172, 231–237 (1990). 

51. Jadus, M. R. et al. Macrophages kill T9 glioma tumor cells bearing the membrane isoform of macrophage colony-stimulating factor through a phagocytosis-dependent pathway. J. Immunol. 160, 361–368 (1998). 

52. Zhou, X., Liu, X. & Huang, L. Macrophage-mediated tumor cell phagocytosis: an opportunity for nanomedicine intervention. Adv. Funct. Mater. 31, 2006220 (2021). 

53. Lecoultre, M., Dutoit, V. & Walker, P. R. Phagocytic function of tumor-associated macrophages as a key determinant of tumor progression control: a review. J. Immunother. Cancer 8, e001408 (2020).

54. Cui, S., Reichner, J. S., Mateo, R. B. & Albina, J. E. Activated murine macrophages induce apoptosis in tumor cells through nitric oxide-dependent or -independent mechanisms. Cancer Res. 54, 2462–2467 (1994). 

55. Brahmer, J. R. et al. Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. N. Engl. J. Med. 366, 2455–2465 (2012). 

56. Saijo, S. et al. Dectin-1 is required for host defense against Pneumocystis carinii but not against Candida albicans. Nat. Immunol. 8, 39–46 (2007). 

57. Bernhagen, J. et al. Purification, bioactivity, and secondary structure analysis of mouse and human macrophage migration inhibitory factor (MIF). Biochemistry 33, 14144–14155 (1994). 

58. Chen, T. J. & Kotecha, N. Cytobank: providing an analytics platform for community cytometry data analysis and collaboration. Curr. Top. Microbiol. Immunol. 377, 127–157 (2014). 

59. Nowicka, M. et al. CyTOF workflow: differential discovery in high-throughput high-dimensional cytometry datasets. F1000Res. 6, 748 (2017). 

60. Wilkerson, M. D. & Hayes, D. N. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics 26, 1572–1573 (2010). 61. Finak G., J. M. FlowWorkspace: Infrastructure for representing and interacting with gated and ungated cytometry data sets. Bioconductor Version (2019). 

62. Finak, G., Jiang, W. & Gottardo, R. CytoML for cross-platform cytometry data sharing. Cytom. Part A: J. Int. Soc. Anal. Cytol. 93, 1189–1196 (2018). 

63. Crowell, H. L., Zanotelli, V. R. T., Chevrier, S. & Robinson, M. D. CATALYST: cytometry data analysis tools. R package version 1.10.3.  HelenaLC/CATALYST (2020). 

64. Cock, P. J., Fields, C. J., Goto, N., Heuer, M. L. & Rice, P. M. The Sanger FASTQ file format for sequences with quality scores, and the Solexa/Illumina FASTQ variants. Nucleic Acids Res. 38, 1767–1771 (2010). 

65. Illumina. BaseSpace User Guide. 2014. 

66. Andrews, S. FastQC: a quality control tool for high throughput sequence data. 

67. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013). 

68. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014). 

69. Trapnell, C. et al. Differential gene and transcript expression analysis of RNAseq experiments with TopHat and Cufflinks. Nat. Protoc. 7, 562–578 (2012). 

70. Anders, S., Pyl, P. T. & Huber, W. HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015). 

71. Boyle, E. I. et al. GO::TermFinder–open-source software for accessing Gene Ontology information and finding significantly enriched Gene Ontology terms associated with a list of genes. Bioinformatics 20, 3710–3715 (2004). 

72. Kanehisa, M. & Goto, S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 28, 27–30 (2000). 

73. Stuart, T. et al. Comprehensive integration of single-cell data. Cell 177, 1888–1902. e1821 (2019). 

74. Lun, A. T., Riesenfeld, S., Andrews, T., Gomes, T. & Marioni, J. C. EmptyDrops: distinguishing cells from empty droplets in droplet-based single-cell RNA sequencing data. Genome Biol. 20, 1–9 (2019). 

75. Hafemeister, C. & Satija, R. Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression. Genome Biol. 20, 1–15 (2019). 

76. Becht, E. et al. Dimensionality reduction for visualizing single-cell data using UMAP. Nat. Biotechnol. 37, 38–44 (2019). 

77. Diaz-Mejia, J. J. et al. Evaluation of methods to assign cell type labels to cell clusters from single-cell RNA-sequencing data. F1000Res. 8 ISCB Comm J-296 (2019). 78. Hänzelmann, S., Castelo, R. & Guinney, J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinforma. 14, 7 (2013). 

79. Finak, G. et al. MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data. Genome Biol. 16, 1–13 (2015). 

80. Suffiotti, M., Carmona, S. J., Jandus, C. & Gfeller, D. Identification of innate lymphoid cells in single-cell RNA-Seq data. Immunogenetics 69, 439–450 (2017). 

81. Entwistle, L. J. et al. Pulmonary group 2 innate lymphoid cell phenotype is context-specific: determining the effect of strain, location, and stimuli. Front. Immunol. 10, 3114 (2020).

82. Ricardo-Gonzalez, R. R. et al. Tissue signals imprint ILC2 identity with anticipatory function. Nat. Immunol. 19, 1093–1099 (2018). 

83. Zilionis, R. et al. Single-cell transcriptomics of human and mouse lung cancers reveals conserved myeloid populations across individuals and species. Immunity 50, 1317–1334.e1310 (2019).

Acknowledgments

The authors wish to thank Dr. Wolfgang Zacharias, Dr. Mei Zhang, and Sabine Waigel from the Brown Cancer Center Genomics Facility for their help in RNA-Seq and scRNASeq. We would also like to thank Dr. Clint Geller for proofreading the manuscript. This work was supported by NIH R01CA213990 and R01AI128818 (J.Y.). C.D. and C.T.W. were supported in part by the NIH P20GM135004. Part of this work was performed with the assistance of the UofL Genomics Facility, which is supported by NIH P20GM103436 (KY IDeA Networks of Biomedical Research Excellence), the Brown Cancer Center, and user fees. CyTOF was performed in the Functional Immunomics Core supported by NIH P20GM135004 (Jun Yan/Jason Chesney, MPI). Sequencing and bioinformatics support for this work was provided by NIH P20GM103436 (Nigel Cooper, PI) and NIH P20GM106396 (Donald Miller, PI). The contents of this work are solely the responsibility of the authors and do not represent the official views of the NIH or the National Institute for General Medical Sciences (NIGMS).

Author contributions

A.E.G. designed and performed the experiments analyzed data, interpreted results, and wrote the manuscript. R.S., C.D., and H.G. contributed to experimental design, data acquisition, and data analysis. M.R.W., X.H., and M.Z. assisted in data acquisition and preparation of materials. K.A. and E.C.R. contributed to the computational analysis and interpretation of RNA-Seq analyses and J.H.C. and E.C.R. contributed to the computational analysis and interpretation of ScRNA-Seq analyses. D.T. and C.T.W. contributed to the analysis of CyTOF data. R.A.M., H.-g.Z., and Y.L. contributed materials and guided the project. R.A.M. performed a critical review. J.Y. directed experimental design and the overall study, interpreted data, supervised research, and assisted in writing and review of the manuscript.

Competing interests

The authors declare no competing interests.

Additional information

Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-022-28407-4.

Correspondence and requests for materials should be addressed to Jun Yan.

Peer review information

Nature Communications thanks George Miller and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Reprints and permission information are available at http://www.nature.com/reprints

Publisher’s note Springer Nature remains neutral about jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution, and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. 

The images or other third party material in this article are included in the article’s Creative Commons license unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license.

For more information:1950477648nn@gmail.com