The Mechanistic Target of Rapamycin Complex 1 Pathway Contributes to the Anti-Tumor Effect of Granulocyte-Macrophage-Colony-Stimulating Factor-Producing T Helper Cells in Mouse Colorectal Cancer

ABSTRACT Introduction The role of granulocyte-macrophage-colony-stimulating factor-producing T helper (ThGM) cells in colorectal cancer (CRC) development remains unclear. This study characterizes the function of ThGM cells in mouse CRC. Methods Mouse CRC was induced by administrating azoxymethane and dextran sulfate sodium. The presence of ThGM cells in CRC tissues and the mechanistic target of rapamycin complex 1 (mTORC1) signaling in ThGM cells was detected by flow cytometry. The impact of mTORC1 signaling on ThGM cell function was determined by in vitro culture. The effect of ThGM cells on CRC development was evaluated by adoptive transfer assays. Results ThGM cells, which expressed granulocyte-macrophage-colony-stimulating factor (GM-CSF), accumulated in CRC tissues. mTORC1 signaling is activated in CRC ThGM cells. mTORC1 inhibition by rapamycin suppressed ThGM cell differentiation and proliferation and resulted in the death of differentiating ThGM cells. mTORC1 inhibition in already differentiated ThGM cells did not induce significant cell death but decreased the expression of GM-CSF, interleukin-2, and tumor necrosis factor-alpha while impeding cell proliferation. Furthermore, mTORC1 inhibition diminished the effect of ThGM cells on driving macrophage polarization toward the M1 type, as evidenced by lower expression of pro-inflammatory cytokines, major histocompatibility complex class II molecule, and CD80 in macrophages after co-culture with rapamycin-treated ThGM cells. Lentivirus-mediated knockdown/overexpression of regulatory-associated protein of mTOR (Raptor) confirmed the essential role of mTORC1 in ThGM cell differentiation and function. Adoptively transferred ThGM cells suppressed CRC growth whereas mTORC1 inhibition abolished this effect. Conclusion mTORC1 is essential for the anti-CRC activity of ThGM cells.


Introduction
Colorectal cancer (CRC) is one of the world's most fatal malignancies (Hossain et al., 2022).The immune system plays a fundamental role in recognizing and killing malignant cells (Kather & Halama, 2019;Mezheyeuski et al., 2021).Immune cells essential for anti-CRC immunity are CD4 + T cells, CD8 + cytotoxic T cells, natural killer cells, macrophages, and dendritic cells (Guo et al., 2020).Notably, T cells also contribute to the initiation and progression of inflammation which benefits CRC development.Instructed by antigenpresenting cells and focal milieu, CD4 + T cells become distinct effector subsets, including T helper 1 (Th1), T helper 2 (Th2), T helper 9 (Th9), T helper 22 (Th22), and T helper 17 (Th17) cells (Zielinski, 2023).Th1 and Th17 cells secrete leukocyte-recruiting chemokines and tissue-damaging factors to exacerbate bowel inflammation (Imam et al., 2018).However, Th1 cells can enhance anti-tumor immunity by stimulating tumoricidal CD8 + T cells or secreting cytokines to restrain tumor cell expansion (Basu et al., 2021;Lee et al., 2021).Th2 cells are considered to support tumor development because they negatively modulate cellular immunity (Basu et al., 2021).Th9 cells have both anti-tumor and protumor effects depending on tumor type (Chen et al., 2020).Th22 cells promote CRC development through diverse pathways (Cui, 2019).Therefore, understanding the functions and modulatory mechanisms of T helper subsets is crucial for developing CRC therapies.
In this study, we made the following original findings in a murine CRC model: 1) ThGM cells exist in CRC tissues; 2) mTORC1 signaling is activated in CRC-associated ThGM cells; 3) mTORC1 signaling promotes ThGM cell differentiation and function; 4) mTORC1 signaling is important for ThGM cells to exert the anti-CRC effect.Accordingly, we unveiled a novel regulatory mechanism underlying the function of ThGM cells in CRC.

CRC model
The study was approved by the Wuhan University Animal Care and Use Committee (Reference #: WHU2021-3356), following the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines and the National Research Council's Guide for the Care and Use of Laboratory Animals.Eight-week-old wild-type male C57BL/6J mice and green fluorescent protein (GFP) transgenic mice (C57BL/6J background) were purchased from Biocytogen Co., Ltd.CRC was induced based on previous research (Arnesen et al., 2021).Briefly, 2.5 mg/ml azoxymethane (AOM, Sigma-Aldrich) was prepared in saline, and 2% dextran sulfate sodium (DSS, MW 36,000-50,000 kDa, Sigma-Aldrich) solution was prepared in drinking water.On day 0, AOM was administered to each mouse via intraperitoneal injection (4 µl/g body weight), followed by feeding the mice regular water for 6 days.On day 7, the mice were fed 2% DSS for 7 days and then given regular water for 2 weeks.The "DSS-regular water" treatment was repeated twice (Supplementary Figure S1A).Approximately 80% of animals survived 12 weeks after the initial AOM administration.The mice were then sacrificed, and those with visible tumors in the colon and rectum were subjected to further experiments.CRC formation was confirmed by histological examination (Supplementary Figure S1B).The control group was given regular water without AOM injection.

Isolation of cells from spleens, CRC tissues, and colons
Mice were euthanized using CO 2 at week 6 and week 12 after AOM injection.Each mouse spleen was gently ground with a 5-ml syringe plunger in a 70-µm cell strainer containing 2 ml of ice-cold phosphate-buffered saline (PBS) to prepare a single-cell suspension.The CRC tissues (or the distal colons in the control mice without CRC formation) were collected, cut into 1-mm (Kather & Halama, 2019) fragments, and digested in 1 ml of digestion buffer (RPMI 1640 medium containing 3 mg/ml collagenase VIII, 5 mM HEPES, 300 U/ml deoxyribonuclease I, 2 mM CaCl 2 , 1 mM dithiothreitol, and 20% fetal bovine serum) for 30 minutes in a 37°C water bath.The digested tissues were pressed through 70-µm cell strainers to prepare single-cell suspensions, followed by centrifugation at 250 g for 5 minutes.Red blood cells were eradicated by incubating the cell pellet in the ACK lysis buffer (Gibco) for 5 minutes at room temperature.Cells were then suspended in PBS before testing.If intracellular cytokine staining ensued, the cell density was adjusted to 1 × 10 (Imam et al., 2018)/ml, followed by stimulation with 20 ng/ml phorbol 12-myristate 13-acetate (PMA), 1 μg/ml ionomycin, 2.5 mg/ml brefeldin A, and 2 μM monensin for 4 hours.Cells were then washed with PBS twice and subjected to intracellular cytokine staining.All chemicals and reagents were purchased from Sigma-Aldrich.

Flow cytometry
Fluorochrome-conjugated antibodies are listed in Supplementary Table S1.For surface marker detection, 1 × 10 (Imam et al., 2018)/ml cells were incubated with 2 µg/ml each antibody at 4°C for 15 minutes.For detecting intracellular cytokines and signaling molecules, 1 × 10 (Imam et al., 2018)/ml cells were fixed in 500 µl of Cytofix/Cytoperm buffer (BD Biosciences) for 15 minutes at room temperature and permeabilized in 1 ml of 90% methanol-PBS for 30 minutes on ice.After two washes with PBS, the cells were stained with 5 µg/ml each antibody for 1 hour at room temperature.Cell apoptosis was determined using the APC Annexin V apoptosis detection kit with propidium iodide (BioLegend).Briefly, cells were resuspended in Annexin V binding buffer at a concentration of 2.5 × 10 (Imam et al., 2018) /ml, followed by adding 5 µl of APC Annexin V and 10 µl of propidium iodide solution.Cells were then gently vortexed and incubated for 15 minutes at room temperature.Cells were analyzed on a BD CellTrace™ cytometer (BD Biosciences).Cell sorting was performed on a BD Influx cell sorter (BD Biosciences).

RNA purification, reverse transcription, and real-time polymerase chain reaction (real-time RT-PCR)
RNA was purified using the PureLink RNA Mini Kit (Invitrogen).cDNAs were synthesized using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems).PowerUp™ SYBR™ Green Master Mix (Applied Biosystems) was used to quantify gene expression on a 7300 real-time PCR platform (Applied Biosystems).The PCR procedure was as follows: Step 1: 50°C for 2 minutes; Step 2: 95°C for 10 minutes; Step 3: 95°C for 15 seconds and 60°C for 1 minute, 40 cycles.The expression of target genes was normalized to β-actin and calibrated using the 2 −ΔΔCt method.Primers are shown in Supplementary Table S2.

In vitro ThGM induction and rapamycin treatment
Naïve CD4 + CD25 − T cells were enriched from normal mouse spleens using the Dynabeads™ Untouched™ Mouse CD4 Cells Kit (Invitrogen).Briefly, 1 × 10 (Basu et al., 2021)/ml splenocytes were incubated with 100 μl of antibody mix for 20 minutes at 4°C.Cells were then washed with 10 volumes of isolation buffer and centrifuged at 350 g for 8 minutes.The cell pellet was resuspended in 4 ml of isolation buffer and 1 ml of depletion Dynabeads for 15 minutes at room temperature.After adding 5 ml of isolation buffer, the tube was placed in the magnet for 2 minutes and the supernatant containing CD4 + T cells was transferred into a new tube.

Lentiviral transduction
The mouse Rptor shRNA lentivector (42724094, also encoding GFP) and scrambled control siRNA GFP lentivector (LV015-G) were obtained from Applied Biological Materials.The GFP-encoding mouse Rptor ORF clone lentivector (MR225740L4) was purchased from Origene.Lentivirus packaging and titration were conducted by Viraltherapy Technologies Inc.
To check the effect of mTORC1 on ThGM differentiation, ThGM induction was performed as described previously.Three days after the start of ThGM induction, 6 µg/ml polybrene (Sigma-Aldrich) and corresponding lentiviral particles (multiplicity of infection (MOI) = 10 for each lentivirus) were added to incubate differentiating ThGM cells for 16 hours.The supernatant was then replaced with fresh medium containing the same ThGMinducing reagents to culture the cells for 2 days.On day 5 after the initial induction, GFP expression was measured by flow cytometry to determine transduction efficiency.GFP + cells were then sorted by flow cytometry before analyzing intracellular cytokines and apoptosis.To detect ThGM cell proliferation, CD4 + CD25 − T cells were labeled with 5 µM CellTrace™ Violet (C34571, Invitrogen) for 20 minutes at 37°C before ThGM induction.The dilution of CellTrace™ Violet was determined on day 5 after ThGM induction.
To check the effect of mTORC1 on differentiated ThGM cells, ThGM induction was performed for 5 days as described previously.On day 5, 6 µg/ml polybrene (Sigma-Aldrich) and corresponding lentiviral particles (MOI = 10) were added to incubate the differentiated ThGM cells for 16 hours.The supernatant was then replaced with fresh medium containing the same ThGM-inducing reagents to culture the cells for 2 days.After that, GFP expression, intracellular cytokines, and apoptosis were evaluated as depicted above.

Co-culture of ThGM cells with splenic macrophages
Mouse splenic macrophages were purified from the spleens of normal mice using the EasySep™ Mouse Monocyte Isolation Kit (Stemcell Technology).Briefly, 1 × 10 (Basu et al., 2021)/ml splenocytes were incubated with 50 μl of rat serum and 100 μl of selection antibody cocktail for 5 minutes at 4°C.Cells were then incubated with 100 μl of Rapidspheres for 3 minutes at 4°C, followed by adding PBS to 2.5 ml.Cells were then placed in the magnet and incubated for 5 minutes and the supernatant containing macrophages was transferred into a new tube.
In vitro differentiated ThGM cells (1×10 (Imam et al., 2018)/ml) and splenic macrophages (1×10 (Imam et al., 2018)/ml) were mixed at a ratio of 1:1 in RPMI 1640 containing 10% FCS.Two hundred microliters of the mixed cells were seeded into each well of a 48well microplate (Corning) and incubated for 24 hours in the presence of plate-bound CD3ε antibody and soluble CD28 antibody.brefeldin A (10 µg/ml) was added 6 hours before the end of co-culture.After co-culture, the plate was centrifuged at 250 g for 5 minutes, and the supernatants were carefully discarded.The cell mixture at the bottom was then incubated in 500 ml of 0.25% trypsin-EDTA (Gibco) for 5 minutes at room temperature to dissociate ThGM cells and macrophages.The dissociated cell mixture was collected and subjected to three different sets of processing: 1) Cells were stained with FITC anti-F4/80 antibody, pacific blue anti-TCRβ antibody, APC anti-I-A b antibody, and PE/Cy7 anti-CD80 antibody for 15 minutes on ice.The expression of I-A b and CD80 on TCRβ − F4/80 + cells (i.e., macrophages) was then analyzed by flow cytometry.2) Cells were first stained with FITC anti-F4/80 antibody and pacific blue anti-TCRβ antibody for 15 minutes on ice.The cells were fixed in 500 µl of Cytofix/Cytoperm buffer (BD Biosciences) for 15 minutes at room temperature and permeabilized in 1 ml of 90% methanol-PBS for 30 minutes on ice.After two washes with PBS, the cells were stained with APC anti-IL-23 p40 antibody and PE/Cy7 anti-TNF-α for 1 hour at room temperature.The expression of IL-23 p40 and TNF-α in TCRβ − F4/80 + cells was analyzed by flow cytometry.3) Cells were stained with FITC anti-F4 /80 antibody and pacific blue anti-TCRβ antibody as described above.TCRβ − F4/80 + cells were then sorted by flow cytometry and subjected to real-time RT-PCR to analyze the expression of IL-1β, CXCL1, IL-6, and IL-10.

Adoptive transfer
ThGM cells were generated in vitro as described above from GFP transgenic mice and treated with 10 ng/ml rapamycin for 24 hours as described above.C57BL/6J mice were subjected to CRC induction as previously described.From week 8 after the initial AOM injection, 1 × 10 (Imam et al., 2018) GFP + ThGM cells were intraperitoneally injected into each C57BL/6J mouse once a week for 4 consecutive weeks.At week 12 after the initial AOM injection (i.e., 1 week after the last cell transfer), the recipients were sacrificed to analyze CRC development.

Discriminating CRC cells among total CRC tissue cells
CRC tissues were digested and processed to prepare single-cell suspensions as described previously.The cell density was adjusted to 1 × 10 (Imam et al., 2018)/ml, followed by incubation with APC anti-CD45 antibody, APC anti-CD31 antibody, and unconjugated anti-MCT4 antibody (2 μg/ml each) for 20 minutes on ice.After two PBS washes, the cells were incubated with 2 μg/ml APC goat anti-rabbit IgG for 30 minutes on ice.The cells were then subjected to two sets of processing: 1) Cells were fixed with Cytofix/Cytoperm buffer and permeabilized with 90% methanol-PBS as described previously, followed by incubation with 5 μg/ml PE anti-Ki67 antibody for 30 minutes on ice.2) Cells were stained with the FITC Annexin V apoptosis detection kit with propidium iodide (BioLegend) following the vendor's manual.During flow cytometry analysis of Ki67 and apoptosis, leukocytes (CD45 + ), vascular endothelial cells (CD31 + ), and fibroblasts (MCT4 + ) were excluded, while the remaining CD45 − CD31 − MCT4 − cells were regarded as malignant CRC cells.

Histology
CRC tissues were fixed with formalin, embedded in paraffin, and cut into 5-μm sections, followed by staining with hematoxylin and eosin according to the standard procedures.The histological score was evaluated based on the criteria in Supplementary Table S3.

Statistics
The data are presented as the mean ± standard deviation.One-way ANOVA was used for comparisons.A P value < 0.05 was considered significant.

ThGM cells are found in the spleen and CRC tissues
We established a murine CRC model (Supplementary Figure S1).At the indicated time points after AOM administration, the spleen, distal colon (day 0 and week 6), and CRC tissue (week 12) were harvested from each mouse to prepare single cells.As outlined in Supplementary Figure S2, single cells and live cells were sequentially gated, followed by recognition of TCRβ + CD4 + T cells among live cells (Figure 1a).To detect ThGM cells, intracellular staining of GM-CSF and T-bet (Th1 master regulator) in TCRβ + CD4 + T cells was carried out.Total CD4 + T cells and Th1 cells (CD4 + T-bet + T cells) were significantly increased in CRC tissues (week 12) compared with the colon (day 0 and week 6) (Figure 1ad).Furthermore, CD4 + T-bet + GM-CSF + T cells were present in the spleen, colon, and CRC tissues (Figure 1b), suggesting that some Th1 cells produced GM-CSF.Notably, CD4 + T-bet − GM-CSF + T cells were rare in the spleen and colon from day 0 to week 6 but were significantly increased in the spleen and CRC tissues at week 12, accounting for 3% of week-12 splenic CD4 + T cells and 8% of week-12 CRC-associated CD4 + T cells, respectively (Figure 1b,e).RORγt + (Th17) or Foxp3 + (Treg) T cells did not express abundant GM-CSF (Figure 1f).We did not find GATA3-expressing T cells in CRC tissues.Therefore, CD4 + T-bet − GM-CSF + T cells were primarily ThGM cells.Since ThGM cells accumulated in CRC tissues at week 12, we focused on these CRC-associated ThGM cells in the following experiments.

The mTORC1 pathway is activated in ThGM cells in CRC tissues
mTORC1 signaling is crucial for the differentiation of CD4 + T cell subsets.mTORC1 inhibits Treg cell differentiation by suppressing Foxp3 expression and promotes Th1 differentiation by augmenting the STAT4 signaling, whereas mTORC1 restrains STAT6 activation and Th2 cell differentiation (Chi, 2012).Furthermore, TCR-mediated signals are transduced to mTORC1 to promote the translation of various mRNA targets (Myers et al., 2019).Therefore, mTORC1 is also critical to the activity of CD4 + T cells.Since CD4 + T cells are anti-tumor effector cells (Speiser et al., 2023) mTORC1 thus contributes to CD4 + T cellmediated anti-tumor activity.Nonetheless, the role mTORC1 of in the anti-tumor function of ThGM cells is unclear.To assess the activation status of mTORC1 signaling, we measured the phosphorylation of mTOR and its downstream target 4EBP1.Th1 and ThGM cells expressed higher phosphorylated mTOR than T-bet − GM-CSF − CD4 + T cells (i.e., other CD4 + T cells) (Figure 2a & Supplementary Figure S3).Total mTOR expression was equivalent in Th1, ThGM, and other CD4 + T cells (Figure 2b & Supplementary Figure S3).Consistently, Th1 and ThGM cells expressed more phosphorylated 4EBP1 than other CD4 + T cells (Figure 2c & Supplementary Figure S3).Interestingly, Th1 and ThGM cells also expressed higher total 4EBP1 than other CD4 + T cells (Figure 2d & Supplementary Figure S3).Therefore, Th1 and ThGM cells exhibited higher mTORC1 activation.

mTORC1 inhibition suppressed ThGM cell differentiation in vitro
To evaluate the role of mTORC1 in ThGM differentiation, we induced ThGM cell differentiation in vitro by stimulating naïve splenic CD4 + T cells with agonistic antibodies and ThGM-inducing cytokines.Rapamycin, a selective mTORC1 inhibitor, was added to the culture two days after the start of ThGM induction.Five days after the induction, over 50% of CD4 + T cells expressed GM-CSF in the absence of rapamycin, implying successful ThGM generation.Rapamycin not only decreased GM-CSF expression but also elevated IFN-γ expression in a dose-dependent manner (Figure 3a-d), suggesting that mTORC1 inhibition weakened ThGM differentiation and enhanced Th1 polarization.RORγt and GATA3 were not expressed in these cells (Figure 3e).Rapamycin also significantly inhibited ThGM cell proliferation, as evidenced by the prevention of CFSE dilution in a dose-dependent manner (Figure 3f,g).Surprisingly, rapamycin treatment resulted in substantial ThGM cell apoptosis and necrosis, suggesting that mTORC1 is essential for the survival of differentiating ThGM cells (Figure 3h,i).

mTORC1 inhibition suppresses the function of differentiated ThGM cells
To evaluate the role of mTORC1 in differentiated ThGM cells, we first induced ThGM differentiation for 5 days and then treated differentiated ThGM cells with rapamycin for 24 hours or 72 hours.As indicated in Figure 4(a-d), rapamycin markedly reduced GM-CSF expression but only mildly up-regulated IFN-γ in differentiated ThGM cells.Besides, rapamycin strongly down-regulated IL-2 and TNF-α, another two ThGM-associated cytokines in a dose-dependent manner (Figure 4e-g).Interestingly, rapamycin did not significantly cause ThGM cell death (Figure 4h,i & Supplementary Figure S4).Rapamycin inhibited the proliferation of differentiated ThGM cells in a dose-dependent manner, as evidenced by higher CFSE intensities (Figure 4j,k).Therefore, mTORC1 is essential for the function but not survival of differentiated ThGM cells.

mTORC1 inhibition impairs the activating effect of ThGM cells on macrophages
The classical activated macrophages, i.e., M1 macrophages, express high MHC-II and CD86.They induce tumor cell apoptosis through phagocytosis, antibody-dependent cellmediated cytotoxicity, and release of TNF-α and nitric oxide (Zhang et al., 2021).Besides, they promote tumor immunity by recruiting cytotoxic T cells (Zhang et al., 2021).The antitumor effect of M1 macrophages on colon cancer cells has been reported (Cheng et al., 2018;Gao et al., 2017;Xiong et al., 2018).Because GM-CSF induces the generation of M1 macrophages (Lotfi et al., 2019;Ushach & Zlotnik, 2016) we speculated that ThGM cells would induce M1 macrophage differentiation to suppress CRC growth.To this end, we cocultured splenic macrophages with vehicle-treated or rapamycin-treated ThGM cells for 24 hours and then analyzed cytokine expression in F4/80 + macrophages (Supplementary

Figure S5
).As demonstrated in Figure 5(a-c), compared with macrophages cultured alone, macrophages co-cultured with either vehicle-treated ThGM cells or rapamycin-treated ThGM cells expressed higher pro-inflammatory IL-23 p40 and TNF-α.However, rapamycin-treated ThGM cells were less potent than vehicle-treated ThGM cells in inducing these cytokines in macrophages.We also assessed the mRNA levels of other pro-inflammatory cytokines including IL-1β, CXCL1, and IL-6 in macrophages and observed similar results, whereas the expression of anti-inflammatory IL-10 was not significantly altered by ThGM cells (Supplementary Figure S6).Furthermore, I-A b and CD80 (two markers of M1 macrophages) were up-regulated on the surface of macrophages by both vehicle-treated ThGM cells and rapamycin-treated ThGM cells.However, rapamycin-treated ThGM cells induced lower expression of I-A b and CD80 than vehicle-treated ThGM cells (Figure 5d-f).Therefore, mTORC1 is essential for ThGM cells to support M1 macrophage generation.

Raptor knockdown and overexpression validate the effect of mTORC1 on ThGM cells
To validate the above findings, we performed Raptor knockdown or overexpression in ThGM cells using lentiviral transduction.Raptor is a constitutively binding protein of mTORC1 and is crucial for mTORC1 activity.Therefore, Raptor knockdown selectively inhibits mTORC1 activity while Raptor overexpression promotes mTORC1 activity (Wang et al., 2020).First, we performed ThGM induction as described previously.Three days after the start of ThGM induction, we transduced the differentiating CD4 + T cells with lentiviruses encoding Raptor shRNA or Raptor cDNA (The lentiviruses also encode GFP).On day 5 after the induction, 65% to 75% of CD4 + T cells became GFP + (Supplementary Figure S7A), suggesting they were successfully transduced.GFP + T cells were then subjected to further analysis.Compared with ThGM cells transduced with the control lentivirus, Raptor knockdown diminished GM-CSF expression while elevating IFN-γ expression.In contrast, Raptor overexpression increased GM-CSF expression (Supplementary Figure S7B &amp; S7C).Additionally, Raptor knockdown profoundly suppressed ThGM cell proliferation, as evidenced by the high CellTrace TM Violet intensity.However, Raptor overexpression accelerated ThGM cell proliferation, as evidenced by the lowest CellTrace TM Violet intensity (Supplementary Figure S7D &amp; 7E).Raptor knockdown increased ThGM cell death while Raptor overexpression mildly reduced ThGM cell death (Supplementary Figure S7F &amp; S7 G).Real-time RT-PCR confirmed Raptor knockdown or overexpression in ThGM cells (Supplementary Figure S8).
We then manipulated Raptor expression in already differentiated ThGM cells.To this end, we performed ThGM induction for 5 days first and then transduced ThGM cells with lentiviruses.Two days after the transduction, GFP + cells reached 60% to 70% (Supplementary Figure S9A), implying efficient transduction.GFP + T cells were then subjected to further analysis.Compared with differentiated ThGM cells transduced with the control lentivirus, Raptor knockdown down-regulated GM-CSF expression while boosting IFN-γ expression.In contrast, Raptor overexpression up-regulated GM-CSF expression (Supplementary Figure S9B &amp; 9C).Raptor knockdown also down-regulated the expression of TNF-α and IL-2 in differentiated ThGM cells, whereas Raptor overexpression elevated TNF-α and IL-2 in differentiated ThGM cells (Supplementary Figure S9D &amp; S9E).Neither Raptor knockdown nor Raptor overexpression altered the death of differentiated ThGM cells (Supplementary Figure S9F &amp; S9 G).
After Raptor knockdown or Raptor overexpression, differentiated ThGM cells were cocultured with macrophages as previously described.The expression of cytokines and surface markers were then analyzed in F4/80 + macrophages (Supplementary Figure S10A).Control ThGM cells, which were transduced with the control lentivirus, remarkably induced macrophages to produce TNF-α and IL-23 p40.However, Raptor knockdown in ThGM cells weakened this effect, resulting in less expression of TNF-α and IL-23 p40 in macrophages.In contrast, Raptor-overexpressing ThGM cells further increased the expression of TNF-α and IL-23 p40 in macrophages (Supplementary Figure S10B &amp; S10C).Notably, the control ThGM cells robustly induced I-A b and CD80 expression on macrophages, whereas Raptor knockdown diminished this effect.Meanwhile, Raptor overexpression further boosted the expression of I-A b but not CD80 (Supplementary Figure S10D &amp; S10E).Therefore, mTORC1 is essential for ThGM cells to induce M1-type macrophages.

mTORC1 inhibition weakens the anti-CRC effect of ThGM cells
To clarify the role of ThGM cells in vivo, ThGM cells were generated in vitro from splenic CD4 + T cells of GFP transgenic mice as described above and then treated with or without 10 ng/ml rapamycin for 24 hours.After that, these GFP + ThGM cells were infused into wild-type C57BL/6J mice that were undergoing CRC induction (Supplementary Figure S11).The recipients were euthanized at week 12 after AOM administration to examine S13).As shown in Figure 6(d-f), the transfer of vehicle-treated ThGM cells increased CRC cell apoptosis, whereas the transfer of rapamycin-treated ThGM cells failed to do so.Consistently, vehicle-treated ThGM cells significantly reduced Ki67 expression in CRC cells, suggesting that they impeded CRC growth.However, rapamycin-treated ThGM cells had little effect on Ki67 expression in CRC cells (Figure 6g,h).
We also checked the expression of GM-CSF and IL-2 in GFP + cells, i.e., exogenous ThGM cells in recipients' CRC tissues.As displayed in Figure 6(i,j), vehicle-treated ThGM cells expressed profoundly higher levels of GM-CSF and IL-2 than rapamycin-treated ThGM cells.Therefore, rapamycin treatment also impaired ThGM cell activity in vivo.Vehicle-treated ThGM cells prolonged mouse survival (median survival = 101 days) compared with saline (median survival = 90 days), while rapamycin-treated ThGM cells had little effect on mouse survival (median survival = 92 days) (Figure 6k).

Discussion
To our knowledge, this is the first study reporting ThGM cells in CRC.ThGM cells gradually increased in both the spleen and CRC tissue during CRC progression, suggesting that the CRC-related immunity caused ThGM cell generation.Notably, the ThGM cell frequency in the CRC tissue seemed higher than that in the spleen, implying ThGM cell accumulation in the tumor.However, whether this accumulation resulted from preferential chemotaxis, accelerated proliferation, or prolonged survival needs future investigations.ThGM cells are susceptible to activation-induced cell death (AICD) (Zhang et al., 2013).Perhaps the CRC milieu hinders AICD to prolong ThGM cell lifespan.
mTOR regulates the activities of distinct immune cell populations in the tumor microenvironment (TME).For example, the mTORC1 pathway polarizes tumor-associated macrophages towards the immune-suppressive M2 type to down-regulate tumor immunity, while rapamycin induces macrophage apoptosis and reduces M2 polarization (Covarrubias et al., 2015).mTORC1 inhibition impairs the anti-tumor function of effector CD8 + T cells (Zeng, 2017).mTORC1 activation promotes Treg cell proliferation but impairs Treg cell suppressive capacity in the TME (Gerriets et al., 2016).mTORC1 drives the differentiation of ant-tumor Th1 cells and tumor-promoting (although controversial) Th17 cells (Conciatori et al., 2018).Besides, γδ T cells possess anti-tumor activity mediated by tumoricidal cytokines and direct cytotoxicity.mTORC1 is essential for γδ T cell differentiation.However, mTORC1 inhibition increases the yield and cytotoxicity of γδ T cells on tumor cell lines (Cao et al., 2016;Li & Pauza, 2011;Meng et al., 2021).These findings suggest that mTORC1 exerts differential effects on different immune cell populations which benefit or suppress tumor growth.Therefore, it seems impossible to use a single mTORC1 inhibitor or agonist to uniformly modulate all immune cells in the TME.Specifically and delicately tuning mTORC1 signaling in each immune cell type is necessary for tumor therapy.
In addition, we found that mTORC1 supported ThGM differentiation from naïve CD4 + T cells by maintaining GM-CSF expression and suppressing Th1 differentiation.Because the transcription factors key to ThGM differentiation have not been identified, whether mTORC1 enhances ThGM-related transcription factor expression is unknown.Surprisingly, mTORC1 inhibition (including rapamycin treatment and Raptor knockdown) also caused profound ThGM cell death during the differentiation process, suggesting that mTORC1 is important for the survival of differentiating ThGM cells.However, the underlying mechanism is a mystery.We plan to conduct both bulk and single-cell transcriptome sequencing to analyze the change in gene expression in mTORC1-inhibited ThGM cells in our next project, trying to pinpoint genes that are critical to mTORC1-related ThGM cell differentiation.
Interestingly, the changes in IFN-γ expression and cell death caused by mTORC1 inhibition were differential between differentiated ThGM cells and differentiating ThGM cells, suggesting that mTORC1 functions differentially at various ThGM cell developmental stages.Differentiated ThGM cells were more resistant to Th1 shift and cell death than differentiating ThGM cells, probably because mature ThGM cells have a more stable transcriptome profile.However, mTORC1 inhibition lowered GM-CSF expression whereas Raptor overexpression elevated GM-CSF expression in both differentiated and differentiating ThGM cells, suggesting that mTORC1 impacts GM-CSF production to the same extent in immature and mature ThGM cells.Moreover, mTORC1 inhibition also reduced IL-2 and TNF-α expression whereas Raptor overexpression increased their expression in differentiated ThGM cells.Because GM-CSF and TNF-α activate macrophages (Kaufman et al., 2014;Parameswaran & Patial, 2010;Ushach & Zlotnik, 2016) we observed a weaker activating effect of mTORC1-inhibited ThGM cells and a potent activating effect of Raptoroverexpressing ThGM cells on macrophages.Therefore, in CRC tissues, mTORC1 is likely essential for ThGM cells to drive macrophage polarization toward the M1 type.M1 macrophages, featuring abundant pro-inflammatory cytokine production, high MHC-II expression, and increased costimulatory molecules, might be a double-edged sword: they could promote inflammation to exacerbate CRC or enhance antigen presentation to boost the anti-CRC immunity.The ultimate effect of ThGM cells on macrophage-mediated proor anti-CRC responses might depend on the focal and systemic immune networks.
The adoptive transfer assay demonstrated that ThGM cells suppressed CRC development.We thought this effect could be attributed to the following mechanisms: 1) ThGM cells might directly inhibit CRC cell growth by producing GM-CSF.GM-CSF could slow CRC cell proliferation if CRC cells express the GM-CSF receptor (Urdinguio et al., 2013).2) ThGM cells promote the antigen presentation capability of macrophages to augment tumor immunity.3) ThGM cells might promote the function of Th1 cells and cytotoxic T cells, as reported previously (Zhang et al., 2013) to amplify tumor immunity.In summary, this study

Figure 3 .
Figure 3.The effect of rapamycin on ThGM differentiation.(a) dot plots showing the expression of GM-CSF and IFN-γ in T cells on day 5 after ThGM induction.Rapamycin was added to the culture two days after the start of ThGM induction.Vehicle: DMSO.Rapa (5): 5 ng/ml rapamycin.Rapa (10): 10 ng/ml rapamycin.Cells were restimulated with PMA and ionomycin for 4 hours before the end of culture.(b -d) frequencies of the indicated cell populations on day 5 after ThGM differentiation.(e) dot plots showing the expression of RORγt and GATA3 in T cells on day 5 after ThGM induction.The images represent two independent assays.(f) histograms showing CFSE dilution in T cells on day 5 after ThGM induction.(g) statistics of CFSE intensities shown in (F).(h) dot plots showing apoptosis (Annexin V + ) and necrosis (PI + ) of T cells on day 5 after ThGM induction.(I) frequencies of apoptotic plus necrotic cells.N = 6 samples in three independent assays.One-way ANOVA.*: P < .01. **: P < .01. ***: P < .001.

Figure 4 .
Figure 4.The effect of rapamycin on the function of differentiated ThGM cells.(a) dot plots showing the expression of GM-CSF and IFN-γ in differentiated ThGM cells after 24 hours of rapamycin treatment.Vehicle: DMSO.Rapa (5): 5 ng/ml rapamycin.Rapa (10): 10 ng/ml rapamycin.Cells were restimulated with PMA and ionomycin for 4 hours before the end of culture.(b -d) frequencies of the indicated cell populations in differentiated ThGM cells after 24 hours of rapamycin treatment.(e) dot plots showing the expression of TNF-α and IL-2 in differentiated ThGM cells after 24 hours of rapamycin treatment.(f , g) frequencies of TNF-α-expressing (f) and IL-2-expressing (g) cells shown in (e).(h) dot plots showing apoptosis (Annexin V + ) and necrosis (PI + ) of differentiated ThGM cells after 24 hours of rapamycin treatment.(i) frequencies of apoptotic plus necrotic cells.(j) histograms showing CFSE dilution in T cells after 24-hour rapamycin treatment.(k) statistics of CFSE intensities shown in (J).N = 3 or 5 samples in 3 independent assays.One-way ANOVA.*: P < .01. **: P < .01. ***: P < .001.