Prednisolone modulates CD8⁺ and regulatory T-cell activity to dampen response to immune checkpoint inhibitor therapy in melanoma

ABSTRACT

Immune checkpoint inhibitors (ICIs) have transformed the treatment of advanced melanoma, yet their efficacy is limited by high-grade immune-related adverse events that often require treatment with systemic corticosteroids. Although corticosteroids are widely used, their impact on anti-tumor immunity remains poorly defined. Using an ICI-responsive murine melanoma model, we show that tapered systemic prednisolone administered after three cycles of combined anti-CTLA4 and anti-PD1 therapy compromises ICI-mediated tumor control, leading to delayed progression in one-third of initially responding animals. Mechanistically, prednisolone selectively suppressed CD8⁺ effector T-cell activation in tumor-draining lymph nodes and in the circulation, while expanding activated regulatory T-cells. These changes increased the Treg:CD8⁺ effector ratio, reduced cytotoxic T-cell function and blocked the early ICI-mediated induction of cytokines, including IL-2, IFNγ, VEGF, CCL3/4, IL-13, IL-3, and GM-CSF. Importantly, despite these early immunosuppressive effects, long-term tumor-specific memory responses were preserved. Autologous melanoma:T-cell cocultures validated these findings. Overall, systemic prednisolone disrupts early CD8⁺ T-cell-mediated anti-tumor activity but spares durable immunity, highlighting the critical importance of timing and context in the introduction of corticosteroids during ICI therapy.

Introduction

Immune checkpoint inhibitors (ICIs) have markedly improved the outcomes for patients with advanced melanoma. The combination of anti-CTLA4 (ipilimumab) and anti-PD1 (nivolumab) produces durable clinical benefit, with melanoma-specific survival rate exceeding 50% at 10 years.¹ Despite these advances, combination ICI is limited by immune-related adverse events (irAEs), autoimmune-like events that can affect multiple organ systems and vary widely in their presentation, severity and onset.² Low-grade events are observed in up to 90% of patients treated with ICI, while severe (grade 3-4) irAEs are reported in 20-60% of patients and are most commonly observed with combination anti-CTLA4 plus anti-PD1.² Severe irAEs are typically managed with ICI termination and high-dose corticosteroids.³ In the CheckMate-067 trial, 59% of patients receiving combination ICI therapy experienced grade 3 or 4 irAEs, often necessitating corticosteroid treatment, and in some cases, permanent treatment discontinuation.¹

High-dose systemic corticosteroids, such as prednisolone (PRED), remain the standard-of-care for managing irAEs and are typically administered over a 6–8 week tapering schedule.⁴ In the cytoplasm, PRED binds the glucocorticoid receptor (GR), forming a complex that translocates to the nucleus to regulate gene expression via direct DNA binding and modulation of other transcription factors.⁵ PRED can also exert rapid non-genomic effects via membrane intercalation and cytoplasmic or mitochondrial GR signalling, independent of transcription.⁵ GR activation induces anti-inflammatory mediators such as Annexin A1, suppresses pro-inflammatory transcription factors including NF-κB and AP-1,⁶ reduces the expression of interleukin-2 (IL-2), a cytokine critical for T-cell proliferation and memory.^6,7 While corticosteroids are indispensable for the management of irAEs, their broad immunosuppressive effects may also compromise the activation, expansion and cytotoxic activity of tumor-reactive T-cells. In animal models, expansion and function of low-affinity memory CD8⁺ T cells were reported to be selectively suppressed by corticosteroids, resulting in compromised long-term protection.⁸ Clinical studies assessing the impact of corticosteroid use on ICI efficacy have reported conflicting findings. The early use of corticosteroids, particularly high doses near the start of ICI therapy, has been associated with lower response rates and reduced progression-free survival and overall survival in many cancer types.^9-13 In contrast, other studies have found no significant effect on ICI efficacy, particularly after adjusting for steroid indication (cancer-related versus irAEs) and cumulative dose.^14-16These discrepancies highlight the need for mechanistic insights into the effects of corticosteroids on anti-tumor immunity.

In this study, we examined the immunological effects of systemic PRED administration following combination ICI in a mouse melanoma model. We evaluated tumor control and comprehensively profiled systemic and tissue-specific immune alterations induced by PRED. Following ICI, PRED impaired CD8⁺ effector T-cell activation, decreased circulating immune-modulatory cytokines and chemokines, and promoted early expansion of activated regulatory T-cells (Tregs), leading to relapse in 30% of mice that initially responded to treatment. Despite these early changes, long-term tumor-specific immunological memory was preserved in complete responders administered combination ICI and PRED. These findings provide experimental evidence that systemic corticosteroids can alter the kinetics of ICI response, underscoring the importance of steroid timing and immune context in the management of irAEs during combination immunotherapy.

Materials and methods

In vivo experiments

The study was carried out between June 2023 and May 2025. Female C57BL/6J mice (8-12 weeks old) were purchased from the Animal Resources Centre (ARC) or Ozgene (Perth, Australia) and acclimated for 2-4 weeks prior to experiments. Mice were housed under specific pathogen-free conditions in temperature- and humidity-controlled central animal facility, with a 12 h light/dark cycle. For tumor engraftment, the right flank was shaved and Y3.3^UVRc34 melanoma cells¹⁷ were injected subcutaneously under isoflurane anaesthesia (5 × 10⁵ cells in 50 µL saline). Animals were examined daily for general condition, and tumors were measured every 2–3 days with digital callipers. Tumor volumes were calculated as width² × length/2. Animals were euthanized when tumor volume reached 1000 mm³, when tumors became necrotic or as determined by experimental design. Euthanasia was carried out using metered carbon dioxide exposure.

Combination immune checkpoint inhibitor therapy (Combi-ICI) was initiated once the tumor volume reached 145 ± 20 mm³. Mice received three cycles of antibodies against PD1 (Cat# BE0146, clone RMP1-14, 300 µg/mouse; BioXCell, Lebanon, NH) and CTLA4 (Cat# BE0131, clone 9H10, 200 µg/mouse, BioXCell) or matched isotype control antibodies (Cat# BE0089, clone 2A3 and Cat# BE0087, Syrian hamster IgG; both from BioXCell). Treatments were delivered via intraperitoneal injection into alternating flanks on days 1, 4, and 7. This schedule was selected based on recent clinical trials in which two cycles of neoadjuvant anti-PD1 plus anti-CTLA4, followed by surgery and response-driven adjuvant therapy, resulted in longer event-free survival than surgery followed by 12 cycles of adjuvant anti-PD1.¹⁸ On the day following the final dose, mice were started on 1 mg/kg prednisolone (PRED; Cat# S1737, SelleckChem, Houston, TX) administered via oral gavage in 0.1 mL of water using a flexible feeding tube (Cat# FTP-20-30, Walker Scientific, Australia). This PRED dose was consistent with other studies where long-term PRED demonstrated measurable immunomodulatory effects while avoiding steroid-related toxicity.^19-21 Our steroid treatment regimen followed clinical tapering practices,²² starting at an initial dose of 1 mg/kg tapered by 0.1 mg/kg every three days until completion on day 30. Control mice received vehicle (DMSO) at matched dilutions and intervals.

Responses were classified into three categories based on tumor growth patterns: Complete response (CR), defined as complete tumor clearance with no recurrence; Delayed progression (DP), defined as initial tumor shrinkage (>30% reduction in peak tumor volume) followed by tumor regrowth; and Non-response (NR), defined as continuous linear tumor outgrowth without intermittent tumor shrinkage.

Tissue histology

Excised tissues (liver and colon) were fixed in 10% neutral buffered formalin, then routinely processed for paraffin embedding (graded ethanol dehydration, xylene clearing, and paraffin infiltration) and embedded as FFPE blocks. Sections (4 μm) were cut using a microtome and placed onto charged glass slides, baked at 60°C for 1 hour, de-paraffinised in xylene and rehydrated through graded ethanol to water. Slides were stained for 5 min with Harris haematoxylin (Leica Cat# 3801560), rinsed, differentiated in acid alcohol, and blued in Scott’s tap water prior to counter staining in eosin Y (Sigma Cat #230251) for 45 sec. Sections were dehydrated through graded ethanol, cleared in xylene, and cover-slipped in DPX mountant for histology (Leica Cat# 3801600). Grading of immune infiltration across tissue sections was performed by a pathologist blinded to treatment groups.

Melanoma cell culture

The murine melanoma cell line Y3.3^UVRc34¹⁷ was thawed and briefly expanded in complete RPMI-1640 medium containing 10% heat-inactivated fetal bovine serum (FBS), 10 mM HEPES, 1% non-essential amino acids (all from Sigma-Aldrich, St. Louis, MA), 2 mM L-glutamine, and 55 µM 2-mercaptoethanol (Gibco, Waltham, MA). The cells were cultured at 37 °C in a humidified 5% CO₂ incubator and passaged three times before engraftment. Cells tested negative for mycoplasma (MycoAlert Mycoplasma Detection Kit, Lonza, Basel, Switzerland).

Processing of blood and tissues for flow cytometric analysis

Blood was collected longitudinally via submandibular venipuncture at five experimental time points: baseline, pre-ICI, on-ICI, early during treatment (EDT; 0.9 mg/kg PRED), and late during treatment (LDT; 0.5 mg/kg PRED) (Figure 1(a)). Blood was collected into EDTA-coated tubes, with a minimum of 14 days between collections. Peripheral blood mononuclear cells (PBMCs) were derived by incubating blood with red cell lysis buffer (Cat# 11814389001, Sigma-Aldrich) for 10 min at room temperature, followed by centrifugation at 200 × g. PBMC pellets were washed with saline before cryopreservation in RPMI medium containing 20% FBS and 10% DMSO. Tumors were excised, mechanically dissociated, enzymatically digested using the Mouse Tumor Dissociation Kit (Cat# 130-096-730, Miltenyi Biotec, Bergisch Gladbach, Germany) and passaged through a 70 µm stainless-steel mesh to obtain single-cell suspensions. Tumor-draining lymph nodes (TDLNs: ipsilateral brachial, axillary, and inguinal) were processed as described above, without the digestion step. All single-cell suspensions were cryopreserved for subsequent flow cytometric analysis.

Figure 1.

PRED administration following ICI cessation promotes delayed progression without affecting immunological memory. (a) Experimental overview. Y3.3^UVRc34 melanoma cells were subcutaneously grafted, and Combi-ICI was initiated when tumor volume reached 145 mm³. After 3 cycles of Combi-ICI, daily oral PRED was started at 1 mg/kg, tapered every 3 days by 0.1 mg/kg until completion. Complete responders were rested for 100 days and re-challenged with parental tumor to test long-term protection. No additional treatment was given prior to re-challenge. Blood was collected longitudinally at baseline, Pre-ICI, On-ICI, early during PRED treatment (EDT; 0.9 mg/kg PRED or DMSO), and late during PRED treatment (LDT; 0.5 mg/kg PRED or DMSO). (b-c) Tumor growth curves in mice treated with isotype control antibodies or Combi-ICI followed by DMSO (b) or PRED (c). Isotype control (ISO) or ICI treatment intervals are indicated above the graphs. All graphs were normalized to treatment start. Red, non-responders (NR); blue, complete responders (CR); orange, delayed progressors (DP). Data shown are pooled from 2 independent experiments. (d) Survival outcomes of initial responders (animals with a >30% reduction in peak tumor volume). (e) Animal survival following tumor challenge after 3 months resting (memory response). Kaplan-Meier survival curves were generated in Prism (d, and e); Log rank test was used to compare the outcomes (*p < 0.05; **p < 0.001; ****p < 0.0001).

Flow cytometry

Cells were stained in flow cytometry buffer (PBS supplemented with 5% FBS, 10 mM EDTA and 0.05% sodium azide). Cell surface staining was performed for 30 min on ice in the dark. Fc block (Cat# 553142, BD Biosciences) was used to prevent non-specific antibody binding to the Fc receptors. Dead cells were stained with a near-infrared (NIR) fixable viability dye (Cat# L34975, Invitrogen). Following surface staining (reagents detailed in Supplementary Table S1), cells were fixed and permeabilized using the eBioscience Transcription Factor Buffer Set (Cat# 00-5523-00, Thermo Fisher Scientific, Waltham, MA) and then stained with antibodies against intracellular markers in permeabilization buffer for 45 min at room temperature. After extensive washing, samples were acquired on a 5-laser LSRII Fortessa X-20 flow cytometer (BD Biosciences, Franklin Lakes, NJ) and analyzed using FlowJo v10.10 software (RRID:SCR_008520). Detailed gating strategies are provided in the Supplementary data. To ensure the robustness of the statistical analyses, cell number threshold was applied (populations containing <200 events within the respective gates, were excluded from the downstream analyses).

Mouse plasma cytokine/chemokine analysis

EDTA blood (processed within 4 hours of collection) was centrifuged at 400 × g for 10 min to separate plasma. Plasma was cleared by centrifugation at 10000 × g for 15 min, transferred to new vials and immediately stored at -80 °C. Plasma samples were diluted 2-fold in PBS pH 7.5 and kept on dry ice before analysis with mouse cytokine/chemokine 32-plex discovery assay (MD32, Eve Technologies, Canada). Analysis was conducted by Eve Technologies using the Luminex^TM 200 system (Luminex, Austin, TX, USA) and each analyte was analyzed in duplicate. Fluorescence intensity (FI) values were derived from the discovery assay and were in direct proportion to the amount of protein in the samples. FI values were log2-transformed, to achieve homogeneity of variance and to produce a less skewed distribution for analysis. Heatmap was visualized with Morpheus online analysis tool (https://software.broadinstitute.org/morpheus/; RRID:SCR_014975). Statistical analysis was performed in RStudio version 2025.05.1+513 (RRID:SCR_000432) using limma package version 3.50.3 (Bioconductor; RRID:SCR_006442). P values were adjusted for multiple testing using the Benjamini-Hochberg (BH) method, and adjusted p-values are reported as q-values.

In vitro human melanoma and TIL co-culture assay

The human melanoma cell line SMU17-0263²³ was cultured in DMEM medium supplemented with 10% heat-inactivated FBS, 4 mM L-glutamine, and 20 mM HEPES. Autologous CD8⁺ tumor-infiltrating lymphocytes (TILs) were maintained in complete TIL medium (RPMI-1640 supplemented with 10% heat-inactivated AB human serum, 25 mM HEPES, 100 U/mL penicillin, 100 µg/mL streptomycin, 10 µg/mL gentamycin, 2 mM L-glutamine, and 1000 U/mL IL-2; Peprotech) at 37 °C in 5% CO₂. Melanoma cells were transferred to TIL medium prior to co-culture. Co-cultures were established at a 1:1 ratio (1 × 10⁴ cells) in 24-well plates for 72 h in the presence of DMSO or PRED (0.1, 0.5, or 1 µM). The plates were imaged every 4 h using an IncuCyte ZOOM live-cell imaging system (Sartorius, Goettingen, Germany). At endpoint, supernatants were collected, cleared of debris by centrifugation (13000 rpm, 5 mins) and stored at −30 °C for downstream IFNγ ELISA. Non-adherent cells and trypsinized adherent cells from duplicate wells were combined and stained for immunophenotyping by flow cytometry, as described above. Viable melanoma cells and T cells were quantified, and T cell/melanoma ratio was calculated for each sample and normalized to the respective DMSO control.

IFNγ ELISA

IFNγ levels in co-culture supernatants were quantified using a human IFNγ DuoSet ELISA kit (R&D Systems, Cat# DY285B) as per manufacturer’s instructions. Samples were tested at 1:6 and 1:10 dilution. The minimum level of detection was 9.38 pg/mL.

Statistical analysis

Cytokine analysis was performed using limma package as described above, values considered significant if q value < 0.05. GraphPad Prism v10 (RRID:SCR_002798) was used for other statistical comparisons. Values were considered significant if p < 0.05. Statistical analyses used are specified in the respective figure legends.

Results

Systemic immunosuppression following ICI impacts melanoma outcomes

A syngeneic mouse melanoma model was established to evaluate the impact of systemic steroids, administered after ICI cessation, on ICI treatment outcomes (Figure 1). Y3.3^UVRc34 melanoma tumors were established in immunocompetent C57BL/6 mice by subcutaneous grafting of tumor cells. When tumors reached a median volume of 145 mm³ (range: 125–165 mm³), mice were randomised to receive either three doses of combination ICI (anti-CTLA4 + anti-PD1; Combi-ICI) or isotype control antibodies. On completion of Combi-ICI or isotype treatment, mice were started on oral PRED (starting dose 1 mg/kg tapered by 0.1 mg/kg every three days) mimicking clinical steroid tapering protocols; for vehicle control, DMSO in water was administered (Figure 1(a)).

All animals that received isotype control antibodies followed by oral vehicle control (isotype + DMSO) or oral PRED (isotype + PRED) displayed linear tumor growth with rapid progression (Figure 1(b-c), left panels). Among mice treated with Combi-ICI followed by oral vehicle control (Combi-ICI + DMSO), 14/20 (70%) achieved complete response (CR) and 6/20 (30%) were non-responders (NR; Figure 1(b), right panel). In contrast, three distinct patterns of tumor growth were observed in the Combi-ICI + PRED group; 12/21 (57%) of mice achieved CR, 3/21 (14%) were NR, and 6/21 (29%) exhibited delayed progression (DP) defined as an initial response (>30% reduction from peak tumor volume) followed by late recurrence (Figure 1(c), right panel) this pattern was unique to mice receiving oral PRED (Supplementary Figure S1a-b). Critically, all Combi-ICI + DMSO-treated mice that had initially responded to immunotherapy (>30% reduction in peak tumor volume) had cleared tumors and remained tumor-free, compared to only 66% of initial responders in the PRED treated group (Figure 1(d)). Despite the distinct response patterns in the two Combi-ICI-treated groups (Supplementary Figure S1a-b), the difference in median overall survival was not statistically significant (Combi-ICI + PRED group, 29 days; Combi-ICI + DMSO control group, 12 days; Supplementary Figure S1c).

Taken together, our findings highlight that administration of systemic steroids following Combi-ICI disrupts anti-tumor control in a subset of initially responding mice, leading to tumor progression.

Immunological memory remains intact after delayed PRED administration

Given the altered short-term tumor control dynamics observed with PRED, we evaluated whether these effects extended to memory responses assessed in a long-term protection setting. Complete responders from two independent experiments (Combi-ICI ± PRED, n=43) were rested for at least three months and challenged with parental Y3.3^UVRc34 melanoma cells; no ICI or PRED treatment was administered prior to re-challenge (Figure 1(a)). Upon re-challenge, 19/22 (86%) mice in the Combi-ICI + DMSO group were protected while 3/22 (14%) progressed (Supplementary Figure S1d, left). Similarly, in the Combi-ICI + PRED group, 18/21 (86%) mice were protected and 3/21 (14%) progressed (Supplementary Figure S1d, right). Thus, 86% of mice in each control and PRED treatment groups exhibited tumor-specific immunological memory (Figure 1(e)). These data demonstrate that although PRED promoted delayed progression in the primary setting, it did not compromise the establishment of long-term tumor-specific immunity in mice that achieved complete tumor clearance on immunotherapy, regardless of PRED administration.

Prednisolone diminishes systemic T-cell activation early during steroid treatment

To gain insight into systemic immunological changes on steroid therapy, we first compared levels of 32 circulating cytokines and chemokines early during steroid treatment (EDT; 0.9 mg/kg PRED or vehicle equivalent as per Figure 1(a)). As shown in Figure 2(a), treatment with PRED blocked Combi-ICI-mediated induction of soluble mediators across multiple immune activation pathways, including cytotoxic (IL-2, IFNγ) and effector programs (IL-13), myeloid cell differentiation and recruitment (IL-3, GM-CSF), immune cell migration (CCL3/4) and angiogenesis (VEGF). These changes highlight the systemic dampening of effector T-cell activation and immune cell trafficking, along with broader systemic effects on angiogenic and myeloid recruitment pathways in response to PRED (cumulative dose 3.9–4.9 mg/kg; Figure 2(a) and Supplementary Table S2). The levels of effector cytokines were restored to control (ICI+DMSO) levels late during PRED treatment (LDT; 0.5 mg/kg PRED or vehicle equivalent, Supplementary Figure S2a), suggesting a transitory effect of PRED. To assess ICI-mediated toxicity, tissue sections from mice treated with Combi-ICI (+/− PRED) or isotype control (+/− PRED) (n = 2 per group) were examined. Liver and colonic sections showed no definitive evidence of toxicity; scattered small foci of inflammation observed in the hepatic lobules²⁴ were regarded as a background change (results not shown).

Figure 2.

Systemic immune activation is impaired early during treatment with PRED. (a) Left: Plasma levels of indicated cytokines early during PRED treatment (EDT; 0.9 mg/kg PRED or vehicle control). Each column represents a single mouse. Fluorescence intensity (FI) values were log₂ transformed and a heatmap was generated in Morpheus. The color scale indicates relative marker expression. Right: Volcano plot displaying differential cytokine expression between Combi-ICI + PRED vs Combi-ICI + DMSO-treated mice. Cytokine expression values were compared using limma’s moderated t-tests with Benjamini-Hochberg false discovery rate (BH-FDR) adjustment; p-values and q-values (BH-adjusted p-values) are shown in Supplementary Table S2. Red dots indicate downregulation and black dots, upregulation. Dotted lines indicate cut off values for y-axis log₁₀(q values; <0.05) and x-axis; log₂(Fold change; >1.5). Significant cytokine changes highlighted in a circle. (b−h) PBMCs collected early during treatment (EDT; 0.9 mg/kg PRED or vehicle control) or late during treatment (LDT; 0.5 mg/kg PRED or vehicle control) were profiled by flow cytometry for T-cell activation. (b) Representative flow profiles (left) and a summary (right) of proliferation (Ki67⁺) or activation (Ki67⁺PD1⁺) of CD8⁺ T-cells EDT and LDT. (c) CD8⁺ Tem gating (left panel), and a summary of proliferation (Ki67⁺) or activation (Ki67⁺PD1⁺) of CD8⁺ Tem cells EDT and LDT (right panels). (d−h) Proliferation (Ki67⁺) or activation (Ki67⁺PD1⁺) of CD8⁺ Tcm (d), naïve CD8⁺ T-cells (e), CD4⁺ Foxp3^– Tconv cells (f), naïve CD4⁺ Foxp3^– Tconv (g) and NK cells (h). Orange dots indicate delayed progressors (DP). Data pooled from two independent experiments, mean ± SD is shown. Statistical comparisons were determined using ordinary ANOVA with post-hoc Bonferroni multiple comparison test; significant p-values are shown.

To determine cellular alterations induced by PRED, T-cell phenotypes were assessed in longitudinal blood samples collected at baseline, pre-immunotherapy (Pre-ICI), during-immunotherapy (On-ICI), early during PRED treatment (EDT, 0.9 mg/kg PRED or vehicle equivalent) and late during PRED treatment (LDT; 0.5 mg/kg PRED or vehicle equivalent) (Figure 1(a)). Flow cytometry was used to profile naïve (CD44⁻CD62L^±), effector/memory-like (hereafter Tem, CD44⁺CD62L⁻) and central memory-like (hereafter Tcm, CD44⁺CD62L⁺) CD8⁺ and conventional CD4⁺Foxp3⁻ T cells (Tconv), and assess activation of T cells in circulation (Supplementary Figure S2b). Proliferation (Ki67⁺) and activation (Ki67⁺PD1⁺) of CD8⁺ T cells (Figure 2(b)) was significantly reduced at EDT but not LDT, and this was observed in both CD8⁺ Tem (Figure 2(c)) and CD8⁺ Tcm subsets (Figure 2(d)), while proliferation of naïve CD8⁺ T cells was also reduced (Figure 2(e)). Proliferative suppression extended to CD4⁺ Tconv cells (Figure 2(f)) with effect confined to naïve rather than effector subsets (Figure 2(g) and Supplementary Figure S2c-e), however the overall frequencies of effector CD4⁺ and CD8⁺ T-cells remained unchanged (Supplementary Figure S2c-e). A reduction in NK cell activation (Ki67⁺PD1⁺) was also observed (Figure 2(h)).

Taken together with survival data, treatment with PRED following Combi-ICI transiently blocked the cytotoxic and chemotactic programs, leading to compromised CD8⁺ effector response and trafficking. This is supported by the reduction in activated effector CD8⁺ T cells in systemic circulation at EDT, while largely sparing CD4⁺ effector T-cells.

PRED remodels lymph node responses via CD8⁺ T-cell suppression and Treg activation

To characterize tissue-specific immune alterations, tumor-draining lymph nodes (TDLNs) and the tumor microenvironment (TME) were analyzed by flow cytometry (Supplementary Figure S3) early during PRED treatment (EDT, 0.9 mg/kg PRED or vehicle equivalent). CD8⁺ Tem and Tcm subsets within the TDLNs displayed decreased proliferation (Figure 3(a) and Supplementary Figure S4a), while early activation (CD69⁺) in Tcm subsets was also decreased in PRED-treated mice compared to controls (Figure 3(b)). Additionally, PRED reduced the frequency of CD8⁺ Tem cells expressing tissue retention markers CD69⁺ and CD103⁺ (Figure 3(b)). Importantly, reduced CD8⁺ T-cell activation and retention were accompanied by a concomitant increase in Treg activation and retention early during PRED treatment (EDT; 0.9 mg/kg PRED or vehicle equivalent, Figure 3(c)), manifesting in the increased Treg to CD8⁺ Tem ratio within TDLN, suggesting a shift toward a more immunosuppressive milieu (Figure 3(d)). In contrast, the effect on CD4⁺ Tconv cells was minimal (Supplementary Figure S4b-c). A mild reduction in Treg and CD8⁺ Tcm proliferation was observed in the TME, with other immune subsets unaffected (Supplementary Figure S4). These results indicate that a systemic decrease in CD8⁺ T-cell activation early during PRED treatment is due to Treg-mediated suppression of effector CD8⁺ T-cell responses in the lymphoid tissue, while not ruling out a direct effect on T-cell activation and/or effector function.

Figure 3.

PRED impairs activation of effector CD8⁺ T-cells and promotes lymph node retention of regulatory T-cells. Tumor-draining lymph nodes (TDLNs) were collected early during treatment (EDT; 0.9 mg/kg PRED or vehicle control) from mice grafted with bilateral tumors and profiled with flow cytometry. (a) Proliferation of CD8⁺ Tem in control and PRED-treated mice. Left, representative flow plots; right, summary. (b-c) Early activation (CD69⁺) and tissue residency (CD69⁺CD103⁺) of CD8⁺ Tem (b) and Tregs (c) in control and PRED-treated mice. Left, representative flow plots; right, summary. (d) Ratio of Tregs to CD8⁺ Tem in TDLNs. Data are shown as mean ± SD, significant p-values are shown (Mann-Whitney test).

Prednisolone diminishes CD8⁺ T-cell activation and effector function in melanoma-TIL co-culture

To assess whether PRED directly impairs CD8⁺ T-cell effector function in vitro, we used a reactive human melanoma-T-cell pair.²³ Melanoma cells (SMU17-0263) were co-cultured with autologous CD8⁺ tumor-infiltrating lymphocytes (TILs) expanded from the same tumor specimen, in the presence of PRED or DMSO. T cell clustered around melanoma cells and after 72 hours, DMSO-treated cultures demonstrated complete or near-complete elimination of tumor cells (Supplementary movie S1, Supplementary Figure S5 and Figure 4(a)). In contrast, PRED-treated cultures demonstrated reduced T cell clustering and retained viable melanoma cells after 72 hours (Supplementary movie S2, Supplementary Figure S5 and Figure 4(a)). Flow cytometry confirmed a dose-dependent increase in viable melanoma cells and a reduction in CD8⁺ T-cell:melanoma ratio in PRED-treated samples relative to DMSO controls (Figure 4(b) and Supplementary Figure S6(a)). PRED treatment reduced expression of CD69, CD137 (4-1BB) and MHC class II, T-cell activation markers known to correlate with CD8⁺ cytotoxicity, in a dose-dependent manner (Figure 4(c) and Supplementary Figure S6(a)). Frequency of activated cytotoxic (CD69⁺4-1BB⁺) CD8⁺ T-cells was significantly reduced across all PRED doses (Figure 4(c) and Supplementary Figure S6(a)). Consistent with reduced T-cell activation and increased melanoma cell content, IFNγ concentrations in co-culture supernatants were significantly decreased in response to PRED, in a dose-dependent manner (Figure 4(d)). Together, these data indicate that exogenous PRED directly suppresses T-cell activation, effector cytokine producton and cytotoxicity of fully differentiated effector CD8⁺ T-cells, compromising anti-tumor immunity.

Figure 4.

PRED suppresses human CD8⁺ effector T-cell activation and limits autologous melanoma killing in vitro. (a) Melanoma cells (SMU17-0263) were co-cultured with autologous CD8⁺ tumor-infiltrating lymphocytes (TILs) for 72 hours in the presence of PRED at indicated concentrations. Images were captured using the IncuCyte system. Scale bar, 100 μm. (b-c) Flow cytometry analysis of melanoma and T cell content (b) and T-cell activation (c) in the presence of PRED. Relative content of viable melanoma and CD8⁺ T-cells and melanoma/TIL ratio (b) normalized to control (DMSO). (c) Downregulation of T-cell activation markers CD69 and/or CD137 (4-1BB) following PRED exposure (c). Data represent mean ± SD of five independent co-culture experiments performed in duplicate or triplicate. Statistical analysis was with one-way ANOVA with Dunnett's multiple comparisons; significant p-values are shown. (d) IFNγ levels in supernatants collected 72 hours after co-culture of SMU17-0263 melanoma cells with autologous TILs were measured by ELISA. ELISA absorbance (OD) values were blank-subtracted and converted to pg/mL by interpolation from standard curve (4-parameter logistic fit). Bars show mean ± SD, dots indicate biological replicates (n=5 per group). Statistical analysis was conducted using ordinary one-way ANOVA with Dunnett’s multiple comparisons vs DMSO; significant p-values are shown.

Discussion

Up to 59% of patients with advanced melanoma treated with ICIs develop severe irAEs,¹ and systemic corticosteroids, including PRED, remain the standard treatment for managing these toxicities. Recent clinical studies assessing patients with advanced melanoma have linked high-dose steroid use, especially within two months of ICI initiation, with reduced progression-free and overall survival.^12,25,26 However, the functional effects of systemic immunosuppression following ICI therapy remain poorly understood. Existing knowledge is derived mainly from retrospective analyses, highlighting the critical need for controlled experimental models to dissect these immunological consequences.^13,15,27,28

Using a syngeneic mouse melanoma model and a tapered PRED regimen designed to reflect the clinical management of irAEs, we demonstrate that systemic PRED initiated after three cycles of Combi-ICI, impedes initial tumor control and promotes tumor progression after an initial ICI response without compromising long-term tumor-specific immunological memory. Relapse was consistently observed across biological replicates, and variability not attributable to PRED treatment was comparable between groups, indicating that relapse differences reflected treatment effects rather than uncontrolled experimental or environmental factors.²⁹

Mechanistically, the attenuated early response was associated with selective suppression of CD8⁺ T-cell activity, characterised by reduced activation/retention of effector CD8⁺ T-cells in the TDLNs and in the circulation. These findings align with cytokine data and prior reports that corticosteroids transiently suppress lymphocyte proportions and inhibit IL-2 signalling.^30,31 The sensitivity of differentiated CD8⁺ T-cell subsets, under chronic TCR-engagement conditions, to GR-mediated suppression likely contributes to this phenotype.³² Our results are consistent with the observation that GR expression increases along the CD8⁺ T-cell differentiation axis, from low levels in naïve cells, to high levels in effector, terminally differentiated, and dysfunctional CD8⁺ T-cells.³² Supporting GR-linked attenuation of T-cell activation, systemic levels of cytotoxic and chemotactic proteins (IL-2, IFNγ, CCL3/4, GM-CSF, and IL-13) were profoundly suppressed in response to PRED early during treatment (0.9 mg/kg PRED). IFNγ is strongly associated with clinical response to ICI,³³ while IL-2 contributes to the restoration of functional T-cell pool in responding patients.³⁴ CCL3/4 facilitate effector cell trafficking into the TME while GM-CSF promotes DC activation and antigen cross-priming, and collectively, reduced presence of these cytokines is associated with poor CD8⁺ effector cell activation and infiltration into TME.^35-37 The concomitant suppression of VEGF is in agreement with a broader anti-inflammatory and anti-angiogenic effect of prednisolone.^38,39

 The effects of PRED on CD8⁺ T-cells within the TDLNs are particularly important because these nodes serve as critical hubs for orchestrating anti-tumor immunity. The TDLN microenvironment is sensitive to shifts in the Treg:CD8⁺ T-cell balance, which has been shown to correlate with immunotherapy outcomes.⁴⁰ Our data suggest that PRED reinforces this immunosuppressive milieu, by skewing the Treg:CD8 ratio in TDLNs toward an immunosuppressive and regulatory environment through enhancement of activated Tregs. Although the tumor microenvironment appeared less affected, modulation in the sentinel lymph node may impact the quality and quantity of the anti-tumor T-cell response. Systemic corticosteroids are known to impair T-cell priming in TDLNs, a process essential for robust anti-tumor responses.^41,42 This disruption likely contributes to the attenuated efficacy of ICIs observed in some patients, by shifting immune regulation toward a more immunosuppressive and pro-tumorigenic state. It is important to emphasize that in our longitudinal tapered model, ICI-mediated T-cell priming was allowed to proceed normally before PRED initiation,¹⁷ which likely accounts for the relatively mild effects - in contrast to the significant impact on survival in early steroid treatment models.^21,43

Divergent outcomes in ICI-responding animals after receiving PRED, highlight the importance of host immune factors in determining complete tumor control. The reasons why steroid-induced suppression of T-cell activity does not universally impair ICI efficacy, remains uncertain. This heterogeneity likely reflects the threshold-dependent nature of the anti-tumor immunity: in some cases, immune activation is sufficiently established to resist subsequent immunosuppression while in others, particularly those with suboptimal priming or marginal T-cell responses, corticosteroids may interrupt key phases of effector priming, tipping the balance toward tumor escape and progression.^8,44 Supporting this, patients with non-small-cell lung cancer receiving early glucocorticoids often exhibit poor ICI outcomes, frequently in the context of additional adverse prognostic features such as brain metastases and poor ECOG performance status.¹⁴ While lack of ICI-driven immunotoxicity is a limitation of our model, the rationale of this study was to determine the functional effect of systemic immunosuppression following ICI.⁴⁵ Current mouse models of ICI-toxicity require additional interventions such as chemical disruption of the colonic mucosa (Dextran Sulfate Sodium colitis), colonic microbiota transfers⁴⁶ and/or use autoimmune-prone animal strains.²¹ These interventions induce intestinal inflammation that is driven primarily by widespread activation of IFNγ-producing CD4 T cells,⁴⁵ manifesting in bystander T-cell activation with the potential to mask any other antigen-specific responses and confound the interpretation of the results.

These findings support the growing interest in alternative immunosuppressive strategies that mitigate irAEs without compromising ICI efficacy and patient outcomes. Janus kinase inhibitors can reduce inflammatory signalling while preserving T-cell activation.⁴⁷ Similarly, cytokine-targeting therapies, such as anti-IL-6 (Tocilizumab) or anti-TNFα (Infliximab), have demonstrated utility in managing irAEs with less disruption to anti-tumor immunity.^3,48,49 In parallel, evolving ICI regimens may reduce the need for high-toxicity combinations currently in use. For instance, the RELATIVITY-047 trial showed that anti-LAG3 + anti-PD1 produced response rates similar to anti-CTLA4 + anti-PD1 (48% vs 50%), with markedly fewer high-grade irAEs (23% vs 61%) and treatment discontinuations (17% vs 41%).⁵⁰

Overall, our findings demonstrate that early steroid administration selectively impairs CD8⁺ T-cell proliferation and antigen-driven activation across both circulatory and lymphoid compartments, while concurrently promoting early activation and expansion of Tregs. This immunosuppressive reprogramming likely disrupts effective CD8⁺ T-cell priming, thereby undermining the therapeutic efficacy of ICI in a subset of responding patients. Together, these findings support a tipping-point model in which post-ICI corticosteroid administration destabilizes CD8⁺ T-cell immunity in a subset of responders, shifting the balance toward immune suppression and tumor progression.

Supplementary Material

Brown et al Supplemental material

Brown et al Supplemental material

Brown et al ARRIVE checklist.pdf

Brown et al ARRIVE checklist.pdf

Brown et al Supplementary movies 1 and 2.pptx

Brown et al Supplementary movies 1 and 2.pptx

Funding Statement

This work was supported by the Macquarie University, Melanoma Institute Australia and National Health and Medical Research Council of Australia (NHMRC; grant 2012860 and 2028055), the Australian Cancer Research Foundation (ACRF) Centre for Advanced Cancer Modelling.

Disclosure of potential conflicts of interest

G.V.L. is a consultant advisor for Agenus, Amgen, Array Biopharma, AstraZeneca, Aulos Bioscience Inc, Bayer HealthCare Pharmaceuticals Inc, BioNTech SE, Boehringer Ingelheim International GmbH, Bristol Myers Squibb, Evaxion Biotech A/S, Fortiva Biologics (USA) Inc, GI Innovation Inc, Hexal AG (Sandoz Company), Highlight Therapeutics S.L., IOBiotech, Immunocore Ireland Limited, Innovent Biologics USA Inc, Iovance Biotherapeutics Inc, Merck Sharpe & Dohme, Novartis Pharma AG, OncoSec Medical Australia, PHMR Limited, Pierre Fabre, Regeneron Pharmaceuticals, Scancell Limited, SkylineDX B.V. No disclosures reported by the other authors.

Data availability statement

Data supporting the findings of this study are available from the authors upon reasonable request.

Ethics statement

All animal procedures were performed in accordance with the Australian Code for the Care and Use of Animals for Scientific Purposes (National Health and Medical Research Council, NHMRC)⁵¹ and were approved by the Macquarie University Animal Ethics Committee (approval number ARA #2019-020). The study was conducted according to the ARRIVE 2.0 guidelines.⁵²

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/2162402X.2026.2643494.