Unleashing T cell anti-tumor immunity: new potential for 5-Nonloxytryptamine as an agent mediating MHC-I upregulation in tumors

Abstract

Background: New therapies are urgently needed in melanoma, particularly in late-stage patients not responsive to immunotherapies and kinase inhibitors. To uncover novel potentiators of T cell anti-tumor immunity, we carried out an ex vivo pharmacological screen and identified 5-Nonyloxytryptamine (5-NL), a serotonin agonist, as increasing the ability of T cells to target tumor cells.

Methods: The pharmacological screen utilized lymphocytic choriomeningitis virus (LCMV)-primed splenic T cells and melanoma B16.F10 cells expressing the LCMV gp33 CTL epitope. In vivo tumor growth in C57BL/6 J and NSG mice, in vivo antibody depletion, flow cytometry, immunoblot, CRISPR/Cas9 knockout, histological and RNA-Seq analyses were used to decipher 5-NL's immunomodulatory effects in vitro and in vivo.

Results: 5-NL delayed tumor growth in vivo and the phenotype was dependent on the hosts' immune system, specifically CD8⁺ T cells. 5-NL's pro-immune effects were not directly consequential to T cells. Rather, 5-NL upregulated antigen presenting machinery in melanoma and other tumor cells in vitro and in vivo without increasing PD-L1 expression. Mechanistic studies indicated that 5-NL's induced MHC-I expression was inhibited by pharmacologically preventing cAMP Response Element-Binding Protein (CREB) phosphorylation. Importantly, 5-NL combined with anti-PD1 therapy showed significant improvement when compared to single anti-PD-1 treatment.

Conclusions: This study demonstrates novel therapeutic opportunities for augmenting immune responses in poorly immunogenic tumors.

Keywords: 5-Nonyloxytryptamine (5-NL); Antigen-presenting machinery; CD8+ T cells; Cold tumors; Immunotherapy; cAMP response element-binding protein (CREB).

Fig. 1

A pharmacological screen identifies the serotonin agonist 5-Nonyloxytryptamine (5-NL) as potentiating T cell mediated anti-tumor immunity. A Screen schematic is shown. B-F Mice were infected with 2 × 10⁵ pfu of LCMV-Armstrong. 14 days post infection, splenic pan-T cells were purified. B Splenic T cells were analyzed using FACS to obtain the ratio of CD8⁺, CD4⁺ and CD4⁺CD25⁺FoxP3⁺ T cells shown as percent composition (n = 3). C CD8⁺ T cells were evaluated for various surface markers using flow cytometry (n = 3). D Co-cultured LCMV-primed splenic pan-T cells and B16.GP33 cells were treated with 770 pharmacological compounds at a concentration of 1 μM. T cells were removed from co-culture at 16 h. Tumor-cell viability was assessed using the MTT assay 48 h post-treatment. Viability for each compound was expressed as a fraction relative to control (B16.GP33 cells + T cells). Potential hit compounds below the cut-off of 0.7 are shown in red. (E, Left Panel) B16.GP33 and B16 cells were co-incubated with LCMV-primed splenic pan-T cells with and without 5-NL (1 μM) as described in D and tumor cell viability was assessed using the MTT assay (n = 3–5). (E, Right Panel) B16.GP33 cells were co-incubated with LCMV-primed splenic pan-T cells with and without T-cells for 72 h and the IC50 was determined using the MTT assay (n = 4). F B16.GP33 and B16 murine melanoma cells were co-incubated with LCMV-primed splenic pan-T cells and 5-NL for 16 h. Intracellular staining of CD8⁺ T cells for TNFα, Granzyme B (GZMB) and IL-2 was measured using flow cytometry and normalized to its own respective cell line controls (n = 6). G C57BL/6 J or (H) NSG mice were subcutaneously injected with 5 × 10⁵ B16.GP33 cells. 7 days post-tumor injection, mice were randomized and into two groups and treated daily with 6.25 mg/kg of 5-NL or vehicle for five consecutive days and tumor volume was measured (n = 5–12). Error bars indicate SEM; *P < 0.05 as determined by a Student´s t-test (unpaired, 2 tailed), one or two-way ANOVA with a Dunnett’s post-hoc test

Fig. 2.

5-Nonyloxytryptamine (5-NL) improves T cell anti-tumor immunity in vivo. A-E C57BL/6 J mice were subcutaneously injected with 5 × 10⁵ B16.GP33 cells. 7 days post-tumor injection, mice were randomized into two groups and treated daily with 6.25 mg/kg of 5-NL or with vehicle for five consecutive days. Mice were sacrificed on 13 days post tumor-inoculation. A Tumor sections were stained for CD8⁺ T cells using immunofluorescence (a representative image of n = 4 is shown, scale bar indicates 50 µm). B Numbers of tumor infiltrating CD8⁺ and CD4⁺ T cells, Treg’s (CD4⁺CD25⁺FOXP3⁺), monocytes (CD11b⁺Ly6C^highLy6G⁻), granulocytes (CD11b⁺Ly6G^highLy6C^low), tumor associated macrophages (TAMs, CD11b⁺F4/80^highLy6C^lowLy6G⁻) and dendritic cells (DCs, CD11c⁺MHC-II⁺) were assessed using flow cytometry (n = 6–10). C In addition to the tumor inoculation and 5-NL treatment described in (A), C57BL/6 J mice were also treated with a CD8⁺ T cell depleting antibody (anti-CD8) on days -2, -1 and 7 pre and post tumor cell inoculation. Tumor volume was measured (n = 4–8). D-E Tumor and tumor-draining lymph node infiltrating CD8⁺ T cell markers, intracellular GZMB as well as tetramer were assessed by flow cytometry from mice sacrificed at day 13 (upper panel) or day 20 (bottom panel) post tumor inoculation (n = 5–11). F C57BL/6 J mice were infected with 2 × 10⁵ pfu of LCMV Armstrong and treated daily with 6.25 mg/kg of 5-NL or vehicle for 5 consecutive days starting at day 1 post-infection. 10 days post-infection, cells from the blood, spleen and liver were re-stimulated with LCMV-specific gp33 epitope followed by staining for IFNγ using FACS analysis (n = 5). Tet-gp33⁺ CD8⁺ T cells in the blood, spleen and liver were measured 10 days post-infection (n = 5). Error bars indicate SEM; *P < 0.05 as determined by a Student´s t-test (unpaired, 2 tailed), or a two-way ANOVA with a Tukey’s post-hoc test

Fig. 3

5-Nonyloxytryptamine (5-NL) upregulates antigen presenting machinery in human and murine tumors in vitro and in vivo. A Basal expression levels of H2-Db and H2-Kb in mouse cells and HLA A-C in human cells were assessed using flow cytometry (n = 5). B-D C57BL/6 J mice were subcutaneously injected with 5 × 10⁵ B16.GP33 cells. 7 days post-tumor injection mice were randomized into two groups and treated daily with 6.25 mg/kg of 5-NL or with vehicle for five consecutive days. Mice were sacrificed on day 13 post tumor-inoculation. B Tumor sections were stained for MHC-I using immunofluorescence (representative images of tumors harvested from 4 mice are shown; scale bar indicates 50 µm). C H2-Db/Kb and PD-L1 protein expression on CD45.2⁻ and (D) H2-Db/Kb expression on tumor infiltrating CD8⁺ and CD4⁺ T cells, Treg’s (CD4⁺CD25⁺FOXP3⁺), monocytes (CD11b⁺Ly6C^highLy6G⁻), granulocytes (CD11b⁺Ly6G^highLy6C^low), TAMs (CD11b⁺F4/80^highLy6C^lowLy6G⁻) and DCs (CD11c⁺MHC-II⁺) was assessed using flow cytometry (n = 5–16). E H2DB and H2KB mRNA expression in B16 cells treated with 5-NL for 18 h was assessed using RT-PCR. Expression was normalized to β-Actin (n = 4). F H2-Db/Kb and PD-L1 protein expression was assessed by flow cytometry following treatment with 5-NL for 18 h in B16 cells (phenotype was recapitulated in B16.GP33 cells, data not shown) (n = 5). G H2-Db/Kb (mouse cell lines) and HLA A-C (human cell lines) protein expression was assessed using flow cytometry following treatment with 5-NL (5 μM for RPMI-7591 and 3 μM for MC-38, SW620, A549, MDA-MB-231 and MOPC cells) for 24 h (n = 5–9). Error bars indicate SEM; *P < 0.05 as determined by a Student´s t-test (unpaired, 2 tailed) or a one-way ANOVA with a Tukey’s post-hoc test

Fig. 4

5-Nonyloxytryptamine (5-NL) and other inducers of CREB activation upregulate antigen presenting machinery in vitro and in vivo. A Protein expression of HTR1D in cancer cell lines was assessed using immunoblot analysis (a representative immunoblot of n = 3 is shown, cropping is indicated by a black frame). Lysates harvested from the mouse brain were used as a positive control. B Treatment with other HTR1D agonists for 18 h (Sumatriptan and L694247, both 3 μM) did not increase H2-Db/Kb protein expression in B16 cells as assessed using FACS (n = 3). C C57BL/6 J mice were subcutaneously injected with 5 × 10⁵ B16.GP33 cells. 7 days post-tumor injection mice were randomized into two groups and treated daily with 12.5 mg/kg of Sumatriptan or with vehicle for five consecutive days. Tumor volume was measured (n = 6–7). (D, left panel) Representative immunofluorescent images of B16 cells treated with 5-NL (3 µM) or forskolin (10 µM) for 18 h and stained for phosphorylated CREB (p-CREB Ser-133) and H2-Db are shown (representative images of n = 3–4 are shown; scale bar indicates 50 µm) and fluorescent signal is quantified in D, right panel. E The adenylyl cyclase activator forskolin (20 μM) increased H2-Db/Kb protein expression in B16 cells following 18 h of treatment (n = 6) as measured by FACS. F B16 cells were treated with 5-NL and forskolin at the indicated doses for 72 h. Apoptosis was assessed using Annexin V/7AAD staining (n = 5). Percent apoptosis was ascertained by summing up the Annexin V⁺/7AAD⁻ and Annexin V⁺/7AAD⁺ populations. G MC-38 cells were pre-treated for 30 min with the p-CREB inhibitor 3i (8 μM) followed by treatment with 3 µM of 5-NL for 24 h. Cells were analyzed using FACS for expression of H2-Db /Kb (n = 3–5). H C57BL/6 J mice were subcutaneously injected with 5 × 10⁵ B16.GP33 or MC-38 cells. 7 days post-tumor injection mice were randomized into two groups and treated daily with 6.25 mg/kg of 5-NL or with vehicle for five consecutive days. Mice were sacrificed on 13 days post tumor-inoculation and tumor tissue was stained for p-CREB using immunofluorescence. Scale bars indicate 50 µm and 20 µm (region of interest outlined in red) (representative images of tumors harvested from 4–6 mice are shown). Fluorescent signal from p-CREB channel was quantified in the right panel. Error bars indicate SEM; *P < 0.05 as determined by a Student´s t-test (unpaired, 2 tailed) or a one-way ANOVA with a with a with a Dunnett’s post-hoc test

Fig. 5

5-Nonyloxytryptamine (5-NL) induces differential gene expression and activates the AMPK pathway in tumor cells. B16.GP33 cells were treated with 5-NL (3 µM) for 18 h and RNA was assessed using RNA-seq analysis. A A Volcano Plot of the fold change gene distribution is shown. B-C GSEA analysis of pathways altered in 5-NL treated B16.GP33 cells is shown with arrow pointing up indicating pathway activation and arrow pointing down indicating downregulation. Fold changes in individual genes in select pathways are shown in C (n = 4). (D) Level of AMPK phosphorylation (Thr172) was assessed using immunoblot analysis in B16 and MC38 cells treated with 3 µM of 5-NL at the indicated time points quantified in the right panel (black frames indicate cropped immunoblot; n = 7–9). Error bars indicate SEM; *P < 0.05 as determined by a one-way ANOVA with a Dunnett’s post-hoc test

Fig. 6

5-Nonyloxytryptamine (5-NL) can be successfully combined with immunotherapy in vivo. Transcriptomic data from melanoma samples of therapy naïve patients was mined from the Cancer Immunome Atlas. A Responders to checkpoint inhibitors (anti-PD1 and anti-CTL4) expressed higher mRNA levels of B2M. B Expression of CREB1 positively correlated with HLA A-C and B2M. C C57BL/6 J mice were subcutaneously injected with 5 × 10⁵ B16.GP33 cells. 7 days post-tumor injection mice were randomized and treated daily with 6.25 mg/kg of 5-NL or with vehicle for five consecutive days. Additionally, mice were intravenously injected with murine anti-PD1 antibody or isotype control on days -1, 1, 3, 5 and 7 pre and post tumor inoculation. Tumor volume was measured (n = 5–9). Data is pooled from two independent in vivo experiments. D Combination index (CI) was calculated from dose response curves of human melanoma cell lines treated with 5-NL, Vemurafenib or in a combination in ratio 1:1. CI < 1 indicates synergy, CI = 1 indicates additivity, and CI > 1 indicates antagonism. The EC50 (50% effective concentration) and EC75 (75% effective concentration) are shown (n = 3). Error bars indicate SEM; *P < 0.05 as determined by a Student´s t-test (unpaired, 2 tailed) or a two-way ANOVA with a Tukey’s post-hoc test. E A schematic diagram summarizing 5-NL’s anti-tumoral and pro-immune effects is shown. The diagram created with BioRender.com