Sex-based considerations in the choice for a TLR9 or TLR7/8 agonist to arm the sentinel lymph node in early-stage melanoma

ABSTRACT

Intradermal delivery of the Toll-like receptor (TLR)-9 agonist agatolimod/CPG7909, prior to sentinel lymph node (SLN) biopsy was previously shown to induce locoregional and systemic immunity, reduce tumor-involved SLN rates, and improve recurrence-free survival in patients with early-stage melanoma. Remarkably, men exhibited superior dendritic cell (DC) maturation. Here, we report on further sex-based differences in the immune response after intradermal administration of CPG7909, which included higher CD80/CD83 expression levels in conventional (c) DC subsets in men’s as compared to women’s SLN, as well as higher ex-vivo release levels of IL-1β, TNF, and IL-6 (all contributors to cDC activation) and Th1/Th2 cytokines. In an effort to identify a more effective DC-activating therapy for women, we compared the in-vitro effects of CPG7909 with those of the TLR7/8 agonist resiquimod/R848 on SLN single cells from female patients. R848 induced superior cDC subset activation and TNF, IL-6, IL-10, IL-12, IFNγ, and CXCL10 release. Correlation analyses suggested that IFNα, TNF, and IL-6 were key for CPG7909-induced LNR-cDC activation, whereas R848’s effect appeared more cytokine-independent. We conclude that combining locally delivered CPG7909 and R848 in early-stage melanoma will ensure full-range DC subset activation and robust pro-inflammatory T-cell responses in melanoma SLN, independent of sex.

Introduction

As the first-line tumor-draining lymph nodes, melanoma sentinel lymph nodes (SLN) are the preferred site of early loco-regional metastasis. As such, they carry prognostic significance; moreover, they bear the brunt of the primary tumor-induced immune suppression that supports metastatic niche formation, and their immune status may predict both tumor progression and response to immunotherapy.¹ The SLN microenvironment thus presents an excellent target for enhancing immune responses and overcoming tumor-induced immune suppression, in particular for immune modulatory strategies aimed at promoting dendritic cell (DC) activation and maturation.^2-4 Toll-like receptors (TLRs), a family of pattern recognition receptors (PRRs), are well-established for their ability to elicit innate immune responses by inducing the maturation of DCs, a critical step in initiating antigen-specific adaptive immune responses.^5,6

Our previous work involving 52 patients with clinical stage I/II melanoma demonstrated immune-activating effects in the melanoma SLN, including the mobilization and activation of lymph node-resident conventional dendritic cell (LNR-cDC) and plasmacytoid dendritic cell (pDC) subsets, after intradermal delivery of the TLR9 agonist CPG7909/agatolimod, a CpG type-B oligodeoxynucleotide (CpG-B ODN), one week prior to the sentinel node biopsy (SNB). This was well-tolerated, with minimal side effects, including mild flu-like symptoms and injection site induration, all of which were transient.^7,8 In addition, intradermal CPG7909 induced systemic type I interferon (IFN) responses, elevated frequencies of local and systemic melanoma-specific CD8⁺ T cells, lower rates of tumor-positive SLN, and improved recurrence-free survival (RFS) compared to patients who received a saline placebo.^7,9,10 Interestingly, data following treatment with CpG suggested a sex-based difference in the maturation state of specific DC subsets. In males, CpG monotherapy led to the maturation of both CD14⁻ LNR-cDC and pDC in the SLN, as evidenced by increased CD83 expression levels. However, among female patients, the combined administration of CpG and GM-CSF was required to reach equivalent CD83 levels.¹¹

Sex-based differences in the innate detection of nucleic acids by PRRs have been previously described.¹² The genes TLR7 and TLR8, both located on the X chromosome and involved in recognizing viral single-stranded ribonucleic acid (RNA), can escape X chromosome inactivation (XCI). This results in bi-allelic expression, leading to greater protein expression of these TLRs and an enhanced immune response in females compared to males,^13,14 thereby making R848, a TLR7/8 agonist, a noteworthy candidate for further clinical exploration in women.

Here, we report more extensively on sex-based clinical and immunological differences following the local administration of CPG7909 in patients with early-stage melanoma. Additionally, we compared the immune-potentiating effects of 2-day in-vitro stimulation with CPG7909 and R848, using single-cell suspensions of healthy donor lymph nodes (HLN) and melanoma SLN obtained from females. Our findings support further clinical investigation into immune potentiation of melanoma SLN through local delivery of TLR7/8 agonists, possibly in combination with CpG, in female patients who respond less favorably to CpG-B monotherapy.

Materials and methods

Clinical study design and clinical procedures

Viable SLN cells were analyzed from 43 patients with early-stage melanoma who had previously received treatment with intradermal CPG7909 monotherapy or saline between June 2004 and June 2007 in two consecutive single-blinded and placebo-controlled randomized phase II trials (registered as ISRCTN63321797.^7,8 All patients had received CPG7909 injections (8 mg in trial 1 and 1 mg in trial 2) at the tumor excision site one week before the SNB, and a second of 1 mg in trial 2 two days before SNB, followed by wide local excision (WLE). In addition, treatment-naive SLN single-cell suspensions, obtained from 5 female patients with early-stage melanoma who underwent SNB and WLE at Amsterdam UMC between February 2018 and March 2021, and axillary HLN were obtained from 5 BRCA-1/−2 positive female patients undergoing a prophylactic mastectomy in Antoni van Leeuwenhoek Hospital between January 2012 and September 2014. Single-cell suspensions from both HLN and melanoma SLN of all female participants were in vitro and were stimulated with the TLR7/8 agonist R848/resiquimod and the TLR9 agonist CPG7909/agatolimod. These studies were approved by the Institutional Review Boards of the VU University Medical Center and the Antoni van Leeuwenhoek Hospital.

Lymph node handling and sampling

All SLNs were identified and retrieved by the standard triple technique consisting of preoperative lymphoscintigraphy, blue-dye injection, and intraoperative lymphoscintigraphy with a hand-held gamma-detecting probe.¹⁵ Immediately after removal, SLN were collected in a dry sterile container and taken to the pathology department of Amsterdam UMC, location VUmc, for retrieval of viable cells under sterile conditions. Before routine histological examination and after confirmation by the pathologist that the SLN was suitable for harvesting (i.e. >0.5 cm diameter), viable cells were scraped from the SLN using a previously described method, without interfering with standard diagnostic procedures.¹⁶ In short, after measuring the size of the SLN, it was bisected crosswise with a surgical scalpel, and the cutting surface of the SLN was scraped 10 times with a surgical blade (size no. 22, Swann Morton Ltd., Sheffield, England). SLN cells were rinsed from the blade with medium containing 0.1% DNAse, 0.14% collagenase A (Boehringer), 5% fetal calf serum (FSC), and penicillin-streptomycin L-glutamine (PSG), then transferred to a sterile flask, and subsequently incubated for 30–45 min at 37 °C. Finally, the collected SLN cells were washed twice in culture medium (CM), comprising Iscove’s modified Dulbecco’s medium (IMDM; BioWhittaker, Verviers, Belgium), supplemented with 10% heat-inactivated FCS, PSG, and gentamycin. The obtained viable SLN single-cell samples were either immediately used for ex vivo immune monitoring or cryostored until use for in vitro culture with TLR agonists. After SLN cell sampling, the bisected SLN was further processed by the pathologist according to the SLN protocol of the pathology department of Amsterdam UMC, location VU. Axillary HLN were retrieved from prophylactic mastectomy specimens in the Antoni van Leeuwenhoek Hospital. No additional skin incision or radio isotope or patent blue injection was used. The HLN were collected in a sterile test-tube containing complete medium (CM) and transported to Amsterdam UMC, location VUmc. HLNs were cut into 2-mm³ pieces and further processed into single-cell suspensions using the same dissociation method as described above. Harvested cells were cryopreserved until use for in vitro cultures with TLR agonists.

Lymph node cultures

Collected SLN or HLN cells were plated in a 48-well plate at 5 × 10⁵ cells/well and cultured for 2 d in CM or CM supplemented with either 5 µg/ml CPG7909 (Coley Pharmaceutical Group, Wellesley, Massachusetts, USA) or 10 µg/ml R848 (InvivoGen, San Diego, CA). After 2 d of culture, 100 µl of supernatant from each well was removed and stored at –20 °C for cytokine release assessment. SLN cells were harvested from the wells and subsequently washed and resuspended in 100  µl of fluorescence-activated cell sorting (FACS) buffer (phosphate-buffered saline [PBS] with 0.1% bovine serum albumin (BSA) and 0.02% sodium azide [NaN3]) for subsequent flowcytometric phenotyping.

Flowcytometric phenotyping

Four-color flow cytometry of freshly isolated SLN cell suspensions from the clinical trials was performed as previously described.¹⁷ For methods, gating strategies, and statistics we refer to previously published work.^7,8,18 Available FACS data were (re)analyzed in a separate comprehensive analysis exploring DC maturation and activation state, compared to placebo, and subsequently assessed for sex disparities.

For in vitro stimulation of single-cell suspensions derived from HLN and SLN with R848 and CpG, the following monoclonal antibodies (mAbs) were used in 8-color panels: CD1a-FITC, CD14-PerCP-Cy5.5, CD1c-PE-Cy7, CD11c-APC, CD45-AF700, CD19-PE-CF594, CD83-PE, CD80-BV421, PD-L1-BV786, CD40-BV711, and CD123-BV650. For an overview of the mAb clones and manufacturers, we refer to Table S1. The frequency and activation state of DC subsets were assessed by membrane staining: LN cell suspensions were stained in FACS buffer and incubated with mAbs for 30  min at 4 °C. After incubation, the cells were washed in FACS buffer to remove excess antibodies and used for flowcytometric analysis. Fluorescence minus one (FMO) negative controls were included. For each measurement, a minimum of 2 × 10⁵ cells were required. Multicolor flow cytometry was performed using the BD LSR Fortessa. Data were analyzed using FlowJo (v.10.7.1) analysis software. DC subset gating strategies post-culture were carried out as previously reported.¹⁹

Type-I IFN response transcript analysis

Total RNA was isolated from pre- and post-treatment peripheral blood mononuclear cells (PBMCs) from patients participating in the CPG7909 clinical trials and reverse transcribed as previously described.¹¹ Transcript levels of IFN response genes (IRGs) in PBMC were analyzed before treatment (pre-treatment; t = 0), one week after treatment (t = 7), and three weeks after treatment (t = 21). Post-treatment IRGs were defined as the maximum IRG expression levels at any of the time points t = 7 or t = 21 d. For a list of the 33 IRGs included in the analysis, we refer to Koster et al.¹¹

Cytokine profiling

SLN inflammatory and T cell cytokine detection in SLN single-cell suspension culture supernatants from in vivo CPG7909- or placebo-treated patients were previously described.^7,8,20 Released cytokines from SLN and HLN cultures at day 2 of culture with either CM, CPG7909 or R848 were measured by cytometric bead array (CBA) analysis using multiple human Flex kits (TNF, IL-6, IL-10, IL-12p70, CXCL10, and IFNγ). CBA analyses were performed following the manufacturer’s instructions, and the results were analyzed by flow cytometry (all from BD Biosciences). Concentration in the supernatants (pg/mL) of the respective cytokines and chemokines was calculated using FCAP array software (Soft Flow Hungary Ltd.). A human IFNα platinum enzyme-linked immunosorbent assay (ELISA) kit (catalog number BMS216CE; Thermo Fisher Scientific) was used to determine IFNα levels after in vitro TLR stimulation according to the manufacturer’s instructions. The Synergy™ HT spectrophotometer was used to read the plate.

Statistical analysis

Normal distribution was assessed by means of the Shapiro‒Wilk normality test. Differences between placebo and CpG7909 treatments, for both female and male patients, were evaluated for statistical significance using an unpaired (two-tailed) t-test when normally distributed or the Mann‒Whitney test when non-normally distributed. To assess differences in immune subsets and cytokine/chemokine levels between the medium condition and treatment (CpG7909 and R848), the results were analyzed by the paired t-test (two-tailed) when the parameters showed a normal distribution or alternatively analyzed by the non-parametric Wilcoxon test. Similarly, for comparisons between CpG7909 and R848 treatment groups the paired t-tests (two-tailed) or Wilcoxon test was used. A correlation analysis was performed using R V.4.0.3. The correlation coefficients were calculated using the Spearman rank-order correlation analysis.

Differences and correlations were considered statistically significant when p ≤ 0.05. Significance is presented as p < 0.05*, < 0.01**, < 0.001***, 0.0001****. Data were plotted and analyzed using GraphPad Prism software (version 8), and (correlation matrix) heatmaps were created using R (version 4.3.2).

Results

Local CPG7909 delivery in two randomized phase II studies: patient-related clinical data

In total, 43 patients with early-stage melanoma, enrolled in two randomized phase II trials received intradermal injections around the primary tumor excision site of either CPG7909 (n = 21; 9 female and 12 male patients) or a saline placebo (n = 22; 9 female and 13 male patients) and were selected to retrospectively assess sex disparities in terms of DC subset activation and pro-inflammatory modulation of the SLN. At baseline, there were no significant differences in clinical tumor characteristics between the treatment and placebo groups for either sex (Table 1). In terms of clinical efficacy, men benefited more from i.d. CPG7909 than women, as evidenced by both SLN status (Figure 1A) and RFS (Figure 1B). Nevertheless, the relatively low patient numbers call for caution in drawing too firm conclusions.

Table 1.

Clinical and pathological characteristics of patients (n = 43) treated with TLR9 agonist CPG7909 (n = 21) vs. Placebo (n = 22) according to sex. Data are n (%) of median [IQR]. Percentages may not total 100 because of rounding. The T category is according to TNM staging.

		Female			Male
		CPG7909* (n = 9)	placebo* (n = 9)	p value¥	CPG7909* (n = 12)	placebo* (n = 13)	p value¥
Age (years)
	Mean ± SD	53·3 ± 12·7	47·8 ± 10·5	0·340	55·7 ± 12·7	57·4 ± 12·9	0·437
Melanoma site	Head and Neck	1 (11%)	0	0·415	1 (8%)	0	0·242
	Trunk	2 (22%)	4 (44%)		10 (83%)	8 (62%)
	Upper extremities	0	0		1 (8%)	3 (23%)
	Lower extremities	6 (67%)	5 (56%)		0	2 (15%)
Histological subtype	Superficial spreading melanoma (SSM)	16 (76%)	8 (89%)	0·576	10 (83%)	9 (69%)	0·645
	Nodular	0	0		2 (17%)	4 (31%)
	Other	3 (33%)	1 (11%)		0	0
Breslow (mm)
	Median [IQR]	1·4 [0·9−2·1]	1·2 [0·5−1·7]	0·258	1·6 [0·9−2·2]	1·6 [1·0−3·0]	0·168
Ulceration	Yes	1 (11%)	1 (11%)	1·000	3 (25%)	3 (23%)	1·000
	No	8 (89%)	8 (89%)		9 (75%)	10 (77%)

^¥ Mann–Whitney U test was used to compare continuous variables between different patient groups; the X² test or Fisher's exact test was used to assess associations between two categorical variables.

^* NB: There were also no significant differences between female and male patients for both treatment arms.

Figure 1.

SLN tumor positivity rates (A) and recurrence-free survival (B) in the saline versus CPG7909 (CpG) arms divided by sex. A: *p<0.05 in a Fisher’s exact test​​​​​​.

Sex-based differences in DC subset activation after local treatment with CPG7909 vs. placebo in early-stage melanoma

Maturation and activation state of conventional DC (cDC) subsets in SLN after intradermal delivery of the TLR9 agonist CPG7909 were assessed by means of FACS analysis, as previously described by van de Ven et al.¹⁷ and van den Hout et al.²¹ cDC subsets in the SLN comprise two migratory CD1a⁺ cDC subsets (i.e., CD1a⁺CD11c^intCD1a^hi Langerhans cells [LCs] and CD1a⁺CD11c^hiCD1a^int dermal-like DCs [dDCs]) and two LNR CD1a⁻ cDC subsets (i.e., CD1a⁻CD11c⁺CD14⁻ and CD1a⁻ CD11c⁺CD14⁺ LNR-cDC, the first consisting mostly of cDC2 and for up to 25% of cDC1.^17,19 We previously showed these migratory and LNR-cDC subsets to be related to loco-regional and distant RFS, respectively,²¹ and we showed their activation state to be increased after local CPG7909 treatment.^8,22 We now show by unsupervised clustering analysis (Figure 2A) that concertedly high expression levels of the maturation and activation markers CD83 and CD80 were consistently observed post-treatment in all four cDC subsets in men. In contrast, consistently lower levels were observed in women. As cDC activation was assumed to be induced by type-I IFN released by pDC upon direct activation by CPG7909, we also compared their post-treatment activation between men and women and found significantly higher levels of the CD83 and CD86 activation markers on pDC in men (Figure 2B). However, comparing pre- and post-treatment type-I IFN response signatures, based on expression levels of 33 IFN response genes in PBMC, IFN responses were found to be induced to a similar extent in men and women (Figure 2C), negating differences in type-I IFN release as a root cause for the observed difference in cDC subset activation. Assessment of CPG7909-induced expression of other pro-inflammatory cytokines in overnight cultures of SLN-derived single-cell suspensions showed a more significant and consistent release of TNF, IL-6, IL-10, and IL-1β in men over women (Figure 2C). This finding suggests the possible involvement of CPG7909-triggered cytokines other than IFNα or IFNβ in the observed superior cDC activation in men.

Figure 2.

Enhanced dendritic cell (DC) activation and cytokine release in men as compared to women upon intradermal (i.d.) injection of CPG7909. A: Clustered heat map of cDC activation marker expression in SLN post 2x1 mg or 1x8 mg i.d. CPG7909; blue = male, pink = female. B: CD83 and CD86 expression rates on pDC post i.d. CPG7909 delivery in women (CpG-F, n = 9) versus men (CpG-M, n = 13). The gray line denotes the mean expression levels in saline-treated patients (NB: no significant differences in the saline groups between women and men). C: type-1 IFN: maximum IFN response gene score (based on 33 genes by qRT-PCR) in PBMC from CPG7909-treated patients (2 × 1 mg i.d.) 14 d post SNB, F (pink) n = 4, M (blue) n = 5. TNF, IL-6, IL-12p70, IL-10, and IL-1β cytokine release upon in-vitro stimulation with saline or CpG (O/N) in i.d. CPG7909-treated patients (2x1 mg, 7 and 2 d prior to SLN excision), F (pink) n = 4, M (blue) n = 6 *P < 0.05,**P < 0.01 by 2-sided unpaired (B) or paired (C) t-test.

Sex-based disparities in T-cell cytokine profiles after i.d. treatment with CPG7909

To determine differences in T-cell differentiation in the SLN between men and women after local delivery of CPG7909, SLN-derived single-cell suspensions were subjected to CD3- and CD28-mediated polyclonal stimulation. After 24 h, the supernatants were harvested, and T-cell effector cytokines were measured. Figure 3 shows superior effector T-cell activation in men, as evidenced by significantly higher induction of IFNγ, IL-2, IL-5, and IL-10 release levels.

Figure 3.

Enhanced effector cytokine release levels in T-cells following intradermal (i.d.) injection of CPG7909. SLN T-cell cytokine release in i.d. saline- or CPG7909-treated patients (1 mg, 7 and 2 d prior to SLN excision) after O/N anti-CD3/CD28 stimulation of SLN-derived single-cell suspensions. Saline ♀ n = 5; ♂ n = 4; CpG-F (females) n = 4; CpG-M (males) n = 6. *P < 0.05,**P < 0.01 by 1-sided ANOVA with Tukey post-hoc test.

Comparative assessment of the in vitro immune modulatory effects of R848 versus CPG7909 in HLN and SLN of women

As CPG7909 clearly favored SLN cDC and T-cell activation in men, we explored the relative efficacy of the TLR7/8 agonist in this respect as a therapeutic alternative for women. TLR7 triggering was previously reported to induce superior type-I IFN responses in women as compared to men through the activation of pDC.²³ We confirmed relatively high expression levels of TLR7 and TLR9 transcripts in pDC; in contrast, LNR-cDC expressed relatively high levels of TLR8 (Figure S1). We cultured single-cell suspensions derived either from HLN (n = 5) or melanoma SLN (n = 5) from women in complete medium alone or supplemented with either CPG7909 (5 µg/ml) or R848 (10 µg/ml) over a period of 2 d, after which the supernatants were collected, and the pDC and cDC activation state was assessed by flow cytometry. Table 2 shows the relevant clinical characteristics of the healthy donors and melanoma patients from whom HLN and SLN single-cell suspensions were derived for in vitro TLR stimulation cultures. The median age was 35 y (IQR 25–46). As similar trends were observed for the HLN and SLN, their data were pooled to increase statistical power (Figures 4 and 5; HLN and tumor-positive and -negative SLN are indicated by different symbols in the scatter graphs, and statistically significant differences for the SLN only are separately listed). Of note, we omitted the CD14⁺ LNR-cDC subset from these analyses because its frequency after culture was too low for reliable assessment; the CD14^- LNR-cDC subset will hereafter be referred to as LNR-cDC. Whereas CPG7909 and R848 induced similar phenotypic activation levels in pDC, R848 induced superior activation in all cDC subsets, and dDC and SLN-cDC in particular, based on CD80, PD-L1, or CD40 expression levels (see Figure 4, which lists statistical differences versus medium or CPG7909 for each of these markers). Of note, CPG7909 did selectively induce increased expression of the maturation marker CD83 on all DC subsets. Cytokine release profiles showed significantly higher levels of IFNα upon CPG7909 exposure, whereas significantly higher levels of TNF, IL-6, and IFNγ were observed upon R848 conditioning (Figure 5). Additionally, higher levels of CXCL10, IL-10, and IL-12p70 were observed in the R848 cultures, although these did not reach statistical significance in a direct comparison to CPG7909. In a small panel of available untreated melanoma SLN from men (n = 3), we also performed a head-to-head comparison of CPG7909 to R848 in 2-d single-cell suspension cultures. Although these very small numbers precluded any definitive conclusions, similar effects were observed for both TLR agonists in terms of pDC/cDC activation and cytokine/chemokine release, without clear evidence of superior dDC and LNR-cDC activation or cytokine release upon R848 stimulation, as was observed in women (see Figures S2 and S3). To assess whether our observations were restricted to tumor-draining lymph nodes, we also tested the effects of CPG7909 and R848 in 5-d cultures of peripheral blood mononuclear cells from healthy men and women (n = 5 for either sex). Under these conditions, we did not find significant differences in favor of either sex in terms of cDC or pDC activation by either CPG7909 or R848 (see Figure S4). In contrast to SLN cultures, on the whole weaker activation was found upon exposure to R848 than to CPG7909. From these divergent data, we conclude that our observations for DC subset activation in (tumor-draining) lymph nodes may either be specifically related to melanoma conditioning or to the lymph node microenvironment – or both.

Table 2.

Clinical and pathological characteristics of the donors and patients from whom single-cell suspensions from HLN (n = 5) and melanoma SLN (n = 5) were derived for in vitro TLR stimulation.

Patient	Lymph node	Sex	Site of primary tumor	Breslow (mm)	Ulceration	SLN metastasis
1	HLN	Female	−	−	−	−
2	HLN	Female	−	−	−	−
3	HLN	Female	−	−	−	−
4	HLN	Female	−	−	−	−
5	HLN	Female	−	−	−	−
6	Melanoma SLN	Female	Trunk	1.5	No	No
7	Melanoma SLN	Female	Extremities	0.85	Unknown	No
8	Melanoma SLN	Female	Trunk	2.38	No	No
9	Melanoma SLN	Female	Extremities	0.85	Unknown	Yes
10	Melanoma SLN	Female	Extremities	3.2	No	Yes

HLN: healthy lymph node; SLN: sentinel lymph node.

Figure 4.

More robust in vitro activation of cDC subsets by R848 than by CPG7909 in lymph nodes from women. Phenotypic data were obtained by flow cytometry 2 d after in vitro stimulation with either R848 or CPG7909 in cultures of single-cell suspensions derived from healthy (n = 5) or sentinel lymph nodes (n = 5) (HLN and SLN) from women either undergoing prophylactic mastectomies or with early-stage melanoma, respectively. SLN+: tumor-positive SLN, SLN−: tumor- negative SLN. *P < 0.05,**P < 0.01 by 2-sided paired t-test. P-values for separate comparisons between conditions for the SLN are indicated in boxes.

Figure 5.

Higher in vitro pro-inflammatory cytokine release levels were induced by R848 than by CPG7909 in HLN and SLN from women. Data were obtained by ELISA (IFNα) or cytometric bead array 2 d after in vitro stimulation with either R848 or CPG7909 in cultures of single-cell suspensions derived from healthy (n = 5) or sentinel lymph nodes (n = 5) (HLN and SLN) from women either undergoing prophylactic mastectomies or with early-stage melanoma, respectively. SLN+: tumor-positive SLN, SLN−: tumor-negative SLN. *P < 0.05,**P < 0.01 by 2-sided paired t-test.

Correlation matrix analyses

Correlation matrix analyses were performed to compare co-regulation of cytokines, chemokines, and DC markers between CPG7909 and R848, and try and identify possible differential mechanisms underlying cDC subset activation (Figure 6). In CPG7909-conditioned lymph nodes, there appeared to be a dichotomy between IFNα on the one hand, and IL-6 and TNF on the other, each clustering with different cDC activation markers: whereas the former mostly clustered and was more strongly correlated with the B7 family members CD80 and PD-L1, the latter mostly clustered and showed stronger positive correlations with CD40 and CD83. In the R848-modulated lymph nodes, IFNα did co-cluster with the other pro-inflammatory cytokines, but the majority of cDC activation markers, including CXCL10, clustered separately from these cytokines, revealing a more cytokine-independent cDC activation mechanism for R848. Based on relative TLR expression patterns (Figure S2), most likely this would have entailed direct TLR8-mediated activation. Whereas IL-10 expression levels were consistently negatively correlated to expression of the B7 family members in the CPG7909-conditioned lymph nodes, this correlation was far less the case for R848, with IL-10 even being positively correlated with PD-L1 levels on dDC and LNR-cDC. This raised the question of whether this apparent difference in IL-10 responsiveness could have explained the lower activation levels of cDC observed upon CPG7909 exposure. We previously observed enhanced in vitro CPG7909-mediated cDC and pDC activation in SLN from patients with breast cancer upon STAT3 inhibition by AG490, an upstream JAK2-inhibitor. As JAK2/STAT3 inhibition is known to interfere with IL-10R signaling, we also tested the effects of this combinatorial treatment on pDC, dDC, and LNR-cDC in melanoma SLN and HLN cultures. As shown in Figure S5A, rather than up-regulating activation markers on the cDC subsets, the addition of the JAK2 inhibitor AG490 did not significantly alter their expression levels. In pDC there was even a trend for down-regulating co-stimulatory/inhibitory molecules, reaching significance only for PD-L1 (p = 0.0358). Notably, while IL-6 and IL-10 release levels both went down, IFNα release also dropped significantly, possibly accounting for the failure of AG490 to increase CPG7909-induced cDC activation (Figure S5B).

Figure 6.

Correlation matrix analysis of DC and cytokine release data after in vitro stimulation of lymph nodes from women with either CPG7909 or R848. Data were obtained by flow cytometry, ELISA (IFNα) or cytometric bead array 2 d after in vitro stimulation with either R848 or CPG7909 in cultures of single-cell suspensions derived from healthy (n = 5) or sentinel lymph nodes (n = 5) (HLN and SLN) from women either undergoing prophylactic mastectomies or with early-stage melanoma, respectively. Correlation matrices showing correlation coefficients between 1 and −1, correlation coefficients were calculated using Spearman rank-order correlation analysis. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by 2-sided paired t-test.

Discussion

DCs play a crucial role in initiating antitumor immunity and, as a result, serve as important targets of early melanoma-induced immune suppression. This also makes them attractive therapeutic targets in early-stage melanoma. As the melanoma SLN is both the site where priming of adaptive immunity against the tumor will occur, as well as the stage for early metastatic events, it is of particular importance to bolster DC functionality there. Indeed, in a previous analysis, we showed that early inhibition of migratory cDC was related to primary tumor burden and decreased local RFS, whereas LNR-cDC suppression was related to SLN invasion, increased Treg rates in the SLN, and decreased distant RFS.²¹ These observations provide a clear rationale for early therapeutic inventions aimed at cDC activation in the SLN. pDC are legitimate targets in this respect, as they are suppressed by melanoma-derived soluble factors, leading to the defective release of type-I IFNs, which are vital for LNR-cDC activation and recruitment, cross-priming, and effector T-cell and NK-cell activation.^8,24 TLR9 agonists can be used to overcome this tumor-induced suppression and drive an antitumor response.²⁵ We observed increased frequencies and activation state of LNR-cDC subsets in the SLN, accompanied by the induction of local and systemic melanoma antigen (Ag)-specific CD8⁺ T-cell responses upon intradermal delivery of the TLR9 agonist CPG7909 at the primary melanoma excision site, which was associated with a reduction in the SLN metastatic rate and improved RFS.⁹ Here, we show that this clinical benefit was mostly observed in men, and we also observed higher post-treatment cDC activation levels in men. Most importantly, this also resulted in higher ex vivo release of pro-inflammatory cytokines as well as T-cell effector cytokines in men as compared to women. This raised the question if women would have benefited more from an alternative cDC-targeted treatment, which we addressed here by performing a head-to-head comparison of the DC-activation properties of CPG7909 and the TLR7/8 agonist R848 in in vitro cultures of melanoma SLN and skin-draining HLN, both of which were obtained from women.

Although sex-based immunological differences in the innate detection of nucleic acids by PRRs and type-I IFN responses have previously been described,²⁶ sex, as a biological factor, is often overlooked in (pre-)clinical cancer research.^12,27 Clear links have been found between women and improved responsiveness to TLR7 stimulation. This is in part due to the escape of XCI, leading to higher TLR7 (and TLR8) expression levels in women, as well as to sex hormone influences.^13,14,28-30 Additionally, other genetic mechanisms affecting the expression and function of TLR7 and TLR8, such as single-nucleotide polymorphisms (SNPs) and genomic copy number variation (CNV),¹² have also been described. In general, estrogen has immune-stimulatory properties, enhancing T-cell receptor signaling, whereas androgens have more immune-suppressive properties, driving T-cell exhaustion.³¹ However, it must be noted that in terms of antitumor immunity, the role of estrogens in cancer immunotherapy is more complex, since they have also been implicated in the enhanced induction of myeloid-derived suppressor cells and the skewing of macrophages to a more immune suppressive M2-like state.^26,31 pDCs express higher levels of ER and estrogens can enhance IFNα release by pDCs upon TLR7 triggering.³⁰ Additionally, signaling elements driving IFNα expression, such as IRF5, are elevated under the influence of estrogens, revealing that an added regulatory level contributing to increased type-I IFN responses in women. These intrinsic sex-based differences directing TLR-mediated type-I IFN responses explain enhanced immunity against viruses in women, which may be exploited for the induction of anti-cancer immunity.^26,32 It may also at least in part explain an overall better prognosis observed for female patients with melanoma, as evidenced by higher overall survival rates,²⁴ lower risk of metastasis,²⁴ and better disease-free survival.²⁶ These differences are influenced by various clinico-pathological factors such as tumor thickness, ulceration, and anatomical melanoma site, but also by potential biological differences, such as immune competence.^23,27 Of note, we also observed a better outcome in terms of RFS in the women as compared to men in the placebo arms of our Phase-II clinical studies on local CPG7909 delivery.^7,8

In keeping with the notion of superior TLR7/8-driven immunity in women, we found stronger migratory- and LNR-cDC activation in the female HLN and SLN cultures and concurrent higher release levels of pro-inflammatory cytokines like TNF, IL-6, IL-12, as well as IFNγ and CXCL10, indicating enhancement of effector cell-mediated immunity upon R848 exposure as compared to CPG7909. Of note, various myeloid and lymphoid subsets in the SLN single-cell suspension cultures may have contributed to these increased cytokine and chemokine release levels, in concert generating a more pro-inflammatory antitumor microenvironment. Interestingly, maturation of all DC subsets, as determined by CD83 expression levels, was greater upon CPG7909 conditioning, which was related to higher IFNα release levels in the cultures. pDC activation was comparable between CPG7909 and R848. As type-I IFN expression is related more to TLR7 signaling, whereas the expression of other pro-inflammatory cytokines is related to TLR8 activation,³³ our data suggest a dominant TLR8-mediated effect of R848 in terms of cDC activation and subsequent pro-inflammatory conditioning of the SLN microenvironment. Predominant direct activation through TLR8 ligation in cDC rather than an indirect IFNα-mediated effect upon TLR7 triggering in pDC, is also associated with the separate clustering of cDC activation markers from IFNα in the R848 correlation matrix analysis. This is in contrast to TLR9 triggering, where post-conditioning CD80 and PD-L1 levels were correlated with IFNα release. Of note, R848 also induced higher release levels of the immune suppressive cytokine IL-10. However, little evidence of significant inverse correlations with cDC activation markers was observed, indicating relative resistance to the potential suppressive effects of IL-10. This contrasts somewhat with the IL-10 released upon TLR9 ligation in these female HLN and SLN cultures, for which more inverse correlations with cDC activation were observed. This may be related to the interference of IL-10 in IFNα responses. CpG-induced IL-10 has previously been shown to interfere with its anticancer efficacy, prompting its combined application with STAT3 siRNA to silence IL-10 signaling.^34,35 This raises the question if CPG7909-induced IL-10 might have been the reason for the inferior cDC activation observed in the SLN of female melanoma patients. However, we found no evidence for this by combining CPG7909 with the JAK2 inhibitor AG490 in melanoma SLN, despite the observed reductions in IL-10 release. Of note, this combination previously resulted in superior cDC activation in breast SLN.³⁵ This may have been due to the fact that SLNB in breast cancer occurs while the suppressive effect of the primary tumor still prevails, whereas the melanoma SLN are usually removed one or even several months after the primary tumor has been excised, arguably reducing the effects of tumor-derived immune suppressive soluble factors such as IL-10 in the SLN. Importantly, the JAK2 inhibitor AG490 is also known to interfere with STAT1 phosphorylation, which is vital for mediating type-I IFN responses as well as driving IFNα transcription.^36,37 This finding could explain the significantly reduced IFNα release observed after combined CPG7909/AG490 conditioning of the SLN cultures, which might also explain the failure of this combination therapy to optimize cDC activation. In future studies, more targeted STAT3 silencing methodologies should be used to positively identify the possible interference of the IL-10/STAT3 axis in the cDC-activating effects of CPG7909 in melanoma SLN, and the relative effects in men vs. women.

Whereas the evidence for superior TLR7/8 stimulation and downstream immune effects in women is very strong, for TLR9 this is not so clear. Our finding of apparent superior antitumor immunity triggered by CPG7909 in men (both in terms of cDC activation and pro-inflammatory conversion of the SLN microenvironment) is in keeping with observations in mice, showing that increased TLR9 expression levels in males led to improved clearance of a viral infection.³⁸ It is notable that we did not find differences in the magnitude of the systemic type-I IFN response in men versus women upon local intradermal delivery of CPG7909, suggesting to other root causes than differential activation of the P1-pDC subset (defined by the combination of PDL1 and CD80 expression) and its ability to produce IFNα.³⁹ Of note, whereas CpG-B oligodeoxynucleotides (ODN) like CPG7909 in contrast to the CpG-A ODN are generally considered poor inducers of IFNα,⁴⁰ the induction of a type-I IFN response signature in peripheral blood was strong, durable (up to 3 weeks post-injection) and significant in patients of both sexes receiving CPG7909. This finding is actually consistent with a previous study demonstrating a superior type-I IFN response to in vivo CpG-B treatment over CpG-A in immunologically humanized mice.⁴¹ In contrast to IFNα, we found more robust expression of other pro-inflammatory cytokines in men as compared to women, such as TNF, IL-6, and, in particular, IL-1β, which are all capable of contributing to cDC activation and the transcription of which can be triggered by alternative signaling pathways downstream of the TLR9 receptor involving MAPK/NF-κB rather than IRF7-mediated IFN transcription.⁴² In particular, IL-1β was more significantly elevated in the SLN of men after local injections of CPG7909 and ex vivo re-stimulation with CPG7909; this may also be causally related to the elevated TNF and IL-6 levels, as they have been identified as downstream effectors of IL-1β. TLR9 induces NF-κB-dependent transcription of the IL-1β pro-cytokine, which subsequently needs to be cleaved by the NLRP3 inflammasome to ensure its maturation and subsequent secretion.⁴³ In the context of sex-specific expression, it is compelling that testosterone can activate the NLR3P inflammasome and that its over-activation under pathological conditions (e.g. in COVID-19) can lead to a hyper-inflammatory response and cytokine storm more often in men than in women.⁴⁴ The possible role of IL-1β in the superior cDC-activating effects of TLR9 triggering by CPG7909 in the SLN of men certainly warrants further investigation.

To our knowledge, we are the first to investigate sex-based differences related to a TLR9 agonist in the context of immunotherapy for early-stage melanoma, and to in vitro compare the effects of a TLR9 and a TLR7/8 agonist in HLN and SLN of women. However, in light of the small number of patients in the CPG7909 clinical trials and the experimental size of the in-vitro study, we should interpret the obtained data with caution, considering them as hypothesis-generating for further studies. First, owing to the limited cell numbers we were unable to compare multiple TLR9 and 7/8 agonists. Moreover, owing to the in vitro study design of our comparative CPG7909 and R848 SLN analysis, we were unable to ascertain the effects of the mobilization of precursors from the blood for both TLR agonists.

Taken together, our findings provide a compelling rationale for further exploration of the combined use of R848 and CPG7909 as a potential treatment strategy to effectively arm SLN in both male and female patients with early-stage melanoma. Moreover, preclinical and clinical studies investigating sex-based immunological differences in melanoma immunotherapy are critically important for enhancing our understanding of these factors, which can help tailor treatment strategies to improve outcomes for both males and females.

Supplementary Material

Supplementary material

Supplementary Material

Figure S3_02122025.jpg

Figure S3_02122025.jpg

Figure S2_02122025.jpg

Figure S2_02122025.jpg

Figure S4_17032026.jpg

Figure S4_17032026.jpg

Figure S1_02122025.jpg

Figure S1_02122025.jpg

Figure S5_17032026.jpg

Figure S5_17032026.jpg

Table S1_02122025.docx

Table S1_02122025.docx

Funding Statement

This work was funded in part by Oncode Accelerator, a Dutch National Growth Fund project under grant number NGFOP2201.

Disclosure of potential conflicts of interest

TDdG has received research funding from Idera Pharmaceuticals and served as an advisor to Mendus and Vivaldi Therapeutics. No other potential conflict of interest was reported by the authors.

Data availability statement

Data are available from the corresponding authors upon reasonable request.

Ethics approval statement

The studies were approved by the Institutional Review Boards of the VU University Medical Center (IRB Ethic approval number 2003/199 and 2012-421) and the Antoni van Leeuwenhoek Hospital (IRB Ethic approval number PTC11.0300/N11ISN). SLN samples were collected and handled according to the ethical and medical guidelines described in the Code of Conduct for Proper Use of Human Tissue of the Dutch Federation of Biomedical Scientific Societies. Written informed consent was obtained from all patients prior to SLN sampling, in accordance with the Declaration of Helsinki.

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/2162402X.2026.2662708.