Background
Immune checkpoint inhibitors (ICI) have shown impressive clinical results against a variety of highly aggressive tumors, although durable benefits have only been observed in a limited fraction of patients [
1‐
3]. The main reason for this non-response is the lack of functional tumor-infiltrating T cells [
4]. Given the great potential of ICI, it is of paramount importance to improve their efficacy by populating less infiltrated tumors with functional immune cells.
Compelling evidence shows that in order to function, T cells required a favourable immune microenvironment in which other immune actors play a crucial role [
5]. Indeed, tumors that respond to immunotherapy are enriched not only in T cells, but also in other immune cell populations, including conventional type 1 dendritic cells (DC1) and natural killer (NK) cells, which, by interacting with each other, create an environment suitable for the priming and expansion of tumor-specific T cells [
5]. Consistently, similarly to CD8
+ T cells, intratumoral DC1 and NK cells have also been associated with a strong antitumor response and favourable clinical outcome in various tumors [
6]. The development of novel approaches that can restore the function of the entire immune cycle is therefore of crucial importance to increase the number of patients who can benefit from immunotherapy. In this regard, several chemotherapeutic treatments have been shown to recall immune cells in the tumor microenvironment (TME) through different mechanisms [
7]. For examples, some chemotherapeutics such as doxorubicin (DX), mitoxantrone (MTX), oxaliplatin (OXP), and cyclophosphamide, when administrated at low doses are able to induce immunogenic cell death (ICD) by promoting DC activation, antigen presentation and priming of tumor-specific CD8
+ T cells [
7‐
10].
However, an increase in intratumoral immune cell infiltrate may not be conclusive because of the multiple mechanisms adopted by tumors to prevent patients’ immune system from targeting and eliminating their own tumor cells [
3]. Recently, transcriptional profiling of melanoma and metastatic urothelial carcinoma patients unresponsive to anti-PD-1 therapy revealed an enrichment in pathways associated to TGFβ [
11,
12], a cytokine produced by the TME that reduces immune cell recall within the tumor and inhibits the function of both T cells and NK cells [
13]. Consistently, recent use of TGFβ-blocking antibodies has been shown to overcome resistance to anti-PD-1 therapy by reactivating the antitumor immune response in colorectal cancer and melanomas [
12,
14].
Approaches aimed at improving the effectiveness of treatments in a increasing number of patients, in combination with therapies that enhance the recruitment of innate and adaptive immune cells, could prove valuable for currently intractable tumors, such as neuroblastoma (NB), a paediatric cancer of the peripheral sympathetic system that accounts for 15% of all childhood cancer-related deaths [
15‐
19]. Despite advanced and intensive multidisciplinary therapeutic approaches based on induction chemotherapy, myeloablative chemotherapy, surgery, radiotherapy and treatment of minimal residual disease, mortality of patients with high-risk NB remains significant, and those who survive experience severe long-term side effects. Therefore, it is important to develop more effective and tolerated therapeutic approaches.
Extensive pathological studies of over than 100 cases of primary and metastatic human NBs examined by immunohistochemistry (IHC) revealed the importance of the immune content in the NB microenvironment [
20‐
22]. NB highly infiltrated by T cells are also enriched with DC and NK cells, and the abundance of these immune cell populations is positively correlated with a favourable clinical outcome [
20‐
22]. The importance of the intratumoral immune context in NB is further strengthened by the identification of intratumoral gene signatures specific to these immune cell populations that correlate with PD-1 and PD-L1 expression levels [
22].
Similarly to other tumors, NB exploits a variety of immune evasion strategies including expression of immune checkpoint molecules, induction of immunosuppressive cells, as well as secretion of immunomodulatory mediators, including TGFβ [
23‐
26]. The development of novel multiple combined immunotherapeutic protocols to recruit innate and adaptive immune cells and convert the immunosuppressive environment to an immunostimulatory one may be effective in the treatment of NB.
In this work, we employ a combinatorial in situ immunomodulation strategy based on the administration of i) low dose of chemotherapy to promote ICD and mobilize immune cells into the TME, ii) anti-TGFβ to overcome immunosuppression in the TME, iii) and anti-PD-1 to restore the function of tumor-infiltrating immune cells. Using ex vivo approaches with murine- and patients-derived NB tissues, as well mouse models of NB grown subcutaneously or orthotopically in the adrenal gland, we identify a novel combination therapy that recruits a variety of innate and adaptive immune cells into the TME, and significantly reduces the growth of this aggressive tumor.
Methods
Mice, cell lines and reagents
Six to 8-week-old female C57BL/6 black and albino mice (Charles River Laboratories) were housed under pathogen-free conditions in the Plaisant (Rome, Italy) and in Ospedale Policlinico San Martino (Genova, Italy) animal facilities, respectively. In vivo experiments were performed in accordance with the 3Rs policy and reviewed and approved by the Italian Ministry of Health (authorization n. 755/2019-PR).
Transgenic NB cell lines 9464D and 975A2 were derived from spontaneous tumors arising in TH-MYCN transgenic mice on a C57BL/6 background [
27] and kindly gift by Dr. Crystal Mackall (Stanford University, CA). Tumor cells were grown under standard conditions (RPMI with 10% FCS with Pen/Strep/Glut at 37 °C and 5% CO
2) on tissue-culture treated plastic plates, splitted every other day prior to injection into mice, passaged no more than four times since thawing and routinely tested for the absence of mycoplasma.
DX hydrochloride, MTX dihydrochloride and OXP were from Sigma-Aldrich. Cisplatin (CDDP, Accord Healthcare Limited), vincristine (VINC, Pfizer) and irinotecan (IRI, Campo, Pfizer) were kindly provided by the pharmacy of the Children’s Hospital Bambino Gesù (Rome, Italy).
Lentiviral infection
9464D cells stably expressing luciferase were obtained by infection with CMV Lentivector Plasmid expressing Luciferase-EF1a-copGFP (BLIV511PA-1, System Biosciences), as previously described [
21]. GFP-positive cells were sorted by a BD FACS Aria II and used for in vivo experiments.
Flow-cytometry
All antibodies were purchased from BD Biosciences, eBioscience, Biolegend and R&D system (listed in Supplementary Table S
1). For surface staining, cells were stained with fluorescent labelled antibodies in PBS with 2% FCS for 30 minutes on ice. Viability was assessed by staining with fixable Live/Dead Zombie (Biolegend) or DAPI. For intracellular staining, cells were seeded (1 × 10
6 cells per well) in 96-well U-bottomed plates and stained with antibodies against surface markers, fixed with 2% PFA for 10 minutes at 25 °C, permeabilized with 0.2% Saponin and then stained with anti-FOXP3, anti-IFNγ and anti-granzyme B using Fixation/Permeabilization Concentrate and Diluent kit (eBioscience). Samples were analyzed on a BD Fortessa flow cytometer and FlowJo software (Treestar, version 10.7.2).
MDOTS and PDOTS generation, drug-treatment and co-culture experiments
Murine-derived organotypic tumor spheroids (MDOTS) were obtained from tumor masses as previously described [
28]. Briefly, tumors were mechanically dissected with sterile forceps and scissors, minced against a 70-μm pore filter with a syringe plunger, and washed in 5 ml of RPMI medium (1500 rpm, 5 minutes). The cell pellet was re-suspended in 2–5 ml RPMI medium, depending on the amount of sample, and passed through a 40-μm pore filter. The cell suspension was seeded into ultra-low attachment (ULA) plates at a density of 2000 cells per well in 96-well plates with 200 μl medium, 100,000 cells per well of 24-well plate with 750 μl medium, 300,000–500,000 cells per well of 6-well plate with 2 ml medium (Corning, Costar #3471, 3473, 3474).
Patient-derived organotypic tumor spheroids (PDOTS) were obtained from tumor samples of 7 patients with NB diagnosed between 2020 and 2021 at Bambino Gesù Children’s Hospital (Rome, Italy). Written informed parental consent was obtained for each patient in accordance with the Declaration of Helsinki. The study was approved by the institute’s Ethics Committee. Clinical and genetic information is given in Supplementary Table S
2. Diagnosis and histology were performed according to the International Neuroblastoma Risk Group (INRG) staging system and the International Neuroblastoma Pathology Classification (INPC) [
29,
30], respectively. MYCN status was assessed according to current guidelines [
31]. PDOTS were obtained from tumor masses as previously described [
28]. Briefly, human fresh tumor specimens were minced in a 10-cm dish using sterile forceps and scalpel. Minced tumors were re-suspended in NB PDOTS-medium (RPMI with 20% FCS, Pen/Strep/Glut 1X, Hepes 1X, NaPYR 1X, NEAA 1X and B27 1X at 37 °C and 5% CO
2), strained on 70-μm filter and cultured in ULA tissue culture plates. MDOTS and PDOTS were treated overnight with the indicated drugs (2 μM DX, 2.5 μM OXP, 5 μM IRI for MDOTS, and 3 μM MTX, 5 μM anti-TGFβ and 5 μM anti-PD-1, for both MDOTS and PDOTS). The day after, MDOTS and PDOTS were washed to remove the drug-containing medium. Next, drug-treated and untreated MDOTS and PDOTS were co-cultured 24 hours with splenocytes derived from tumor-bearing C57BL/6 mice and autolougous human peripheral blood mononuclear cells (PBMC), respectively. The functional status of the immune cells was assessed by flow-cytometry using BD LSR Fortessa X20 with FACSDiva Software (BD Bioscences) and FlowJo software (version 10.7.2).
Microfluidic device migration assay
To evaluate the immunomodulatory ability of drugs, we carried out experiments in microfluidic devices made of polydimethylsiloxane (PDMS), a biocompatible silicone elastomer, following a well-established replica moulding procedure [
32,
33]. Prior to cell loading, the devices were sterilized under UV light for 30 minutes. MDOTS and PDOTS were re-suspended in Matrigel (2 mg/ml; BD Biosciences) and treated with drugs at the indicated concentrations. Drug-treated and untreated MDOTS and PDOTS were loaded into the side chambers of the devices (1 × 10
4 cells in 3 μl), which are separated from the central one by microchannels, and incubated at 37 °C for 30 minutes to allow gel solidification. Subsequently, 1 × 10
6 splenocytes derived from tumor-bearing C57BL/6 and autologous PBMC for MDOTS and PDOTS, respectively, were labelled with Cell Tracker Red (Invitrogen), re-suspended in complete RPMI medium and loaded into the central chamber of the device. The size of the microchannels allows splenocytes to migrate from the central chamber to the two lateral chambers, but not MDOTS/PDOTS to move into the central chamber. The reservoir chambers were filled with medium. Phase-contrast, visible and fluorescence microphotographs of the devices were taken with a LEICA DMi8 microscope by collecting photos at 24 hours after loading. The migration of splenocytes/PBMC towards treated and untreated MDOTS/PDOTS was assessed by counting the red-labeled cells in the two side chambers with the ImageJ software (
http://imagej.nih.gov/ij). The extent of splenocytes/PBMC migration to treated versus untreated MDOTS/PDOTS was then analyzed in terms of fold change ± SD [
32‐
34].
Subcutaneous tumor model and therapeutic studies
9464D and 975A2 cells (1 × 106) were inoculated subcutaneously into the flank of C57BL/6 mice. Survival of mice was monitored daily and tumor growth was measured twice weekly using a caliper. Mice were randomized into control and treatment groups (10 mice/group) at day 5, when the tumor volume reached 30–50 mm3. Drug treatment started at day 7 or 8. DX (2.9 mg/Kg) and OXP (2.5 mg/Kg) were injected intratumorally, whereas CDDP (0.25 mg/Kg), MTX (5.2 mg/kg), IRI (2.5 mg/Kg), VINC (0.5 mg/Kg), anti-PD-1 (clone RMP1–14, BE0146, BioXCell, 0.3 mg/mouse) and anti-TGFβ (clone 1D11.16.8, BP0057, BioXCell, 10 mg/Kg) were injected intraperitoneally. Control mice received an equivalent volume of PBS alone or isotype control antibody. Mice were sacrificed after 1 day or 7 or 12 days from the start of drug treatment for analysis of the tumor’s immune infiltrate. All experiments contained 5 to 10 mice per group and were performed at least 2 times, yielding similar results.
Orthotopic tumor model and therapeutic studies
After anesthetization with a mixture of xylazine-ketamine (Xilor 2% Bio98 Srl, Milan, Italy) and Imalgene 1000 (Merial SpA, Italy), six-week-old C57BL/6 albino mice were subjected to laparotomy and inoculated with 0.7 × 10
6 9464D-luc cells in 10 μL culture medium, in the left adrenal gland capsule, as previously described [
35]. Luc activity was confirmed by bio-luminescent imaging (BLI, Lumina-II, Caliper Life Sciences, Hopkinton, MA) after a 10-minute incubation with 150 μg/mL d-luciferin (Caliper Life Sciences) diluted in cell culture medium, as previously described [
36]. BLI monitoring was used as the main criterion for determining the start of treatment. Mice body weight and general physical status were daily recorded. When any sign of discomfort or poor health arose (i.e., abdominal dilatation, dehydration, paraplegia, ≥20% weight loss) mice were anaesthetized with Xilor 2% and sacrificed by CO
2 inhalation. When the tumor volume reached 1 × 10
7 ROI measurement, the mice were randomized into control and treatment groups (6 mice/group). MTX (5.2 mg/kg), anti-PD-1 (0.3 mg/mouse) and anti-TGFβ (10 mg/Kg) were injected intraperitoneally. Control mice received an equivalent volume of PBS alone. Seven days after drug treatment, mice were sacrificed for analysis of the tumor immune infiltrate.
Tumor dissection
Tumors and spleens were dissected from mice and total weight of removed tumor masses was determined. Tumors were cut into small fragments with scissors and then digested in medium containing 325 KU/ml DNAse I (Sigma) and 1 mg/ml Collagenase III (Worthington Biochemicals) per 30 minutes at room temperature in agitation followed by 0.1 M EDTA pH 7.2 for additional 5 minutes. Samples are then filtered through a 70 μm filter, spun down and re-suspended for staining.
Immunofluorescence and immunohistochemistry
Immunofluorescence (IF) and IHC stainings were performed in 2 μm of formaldehyde-fixed paraffin embedded serial tissue sections following deparaffinization and antigen retrieval as previously described [
21,
22]. For double IF staining of NK cells and granzyme B, slides were blocked for 1 hour with 1% BSA and 5% normal goat serum and then antibodies (Supplementary Table S
1) were added consecutively as follow. Sections were firstly incubated with anti-granzyme B antibody overnight at 4 °C, followed by 1-hour incubation with Alexa Fluor 594 goat anti-rabbit IgG. Next, slides were incubated with anti-NK1.1 overnight at 4 °C, followed by 1-hour incubation with Alexa Fluor 488 goat anti-mouse IgG. After staining, slides were counterstained for 5 minutes with Hoechst (H3570, Invitrogen) and cover-slipped with 60% glycerol in PBS. Confocal microscopy imaging was performed by Leica TCS-SP8Xlaser-scanning confocal microscope (Leica Microsystems) equipped with tunable white light laser source, 405 nm diode laser, 3 (PMT) e 2(HyD) internal spectral detector channels. Sequential confocal images were acquired using a HC PLAPO 40× oil immersion objective (1.30 numerical aperture, Leica Microsystems) with a 1024 × 1024 image format, scan speed 400 Hz. The density of intratumoral NK cells was recorded by two blinded examiners as the number of positive cells per unit tissue surface area (mm
2). The mean of the positive cells detected in 5 fields for each sample was used in the statistical analysis.
For calreticulin (CALR) staining, slides were blocked for endogenous peroxidase for 10 minutes with a peroxidase blocking solution (Dako), followed by 30 minutes with 5% PBS/BSA, and then incubated (overnight at 4 °C) with anti-CARL primary antibody (Supplementary Table S
1). This step was followed by incubation with secondary antibody coupled with peroxidase (Dako) for 20 minutes. Bound peroxidase was detected with diaminobenzidine solution and EnVision FLEX Substrate buffer containing peroxide (Dako). Tissue sections were counterstained with EnVision FLEX hematoxylin (Dako). Iso-type-matched mouse mAbs were used as negative controls. Stained slides were analyzed using an image analysis workstation (Nikon Eclipse E600), scanned using the NanoZoomer S60 Digital slide scanner C13210–01 (Hamamatsu Photonics) and viewed with Hamamatsu Photonics’s image viewer software (NDP.view2 Viewing software U12388–01). The density of CALR staining was obtained by evaluating integrated optical density by Color Deconvolution plugin through ImageJ, measured in independent slide images acquired with the same optical microscopic parameters such as magnification, light exposure, and acquisition time. The mean of positive cells detected in 5 fields for each sample was used in the downstream statistical analysis.
For histological characterization, MDOTS were seeded on a layer of Matrigel in 8-well chambers and grown for 5 days at 37 °C and 5% CO2. MDOTS were then fixed in 4% PFA at room temperature for 2 hours and subsequently washed with H2O. After harvested, MDOTS were transferred into Tissue-Tek®Cryomolds® already coated with histogel base (Epredia HistoGel), and then covered with a further histogel layer. The solidified blocks were transferred into formalin overnight and then in 70% ethanol for 24 hours before embedding in paraffin. Sections of 2 μm were deparaffinized, rehydrated in water and stained with hematoxylin, eosin and synaptophysin.
Chemokine analysis
Total protein extract (150 μg) quantified by the bicinchoninic acid (BCA) assay (Thermo Fisher Scientific) were used. Chemokines were detected using the Proteome Profiler Mouse Chemokine Array Kit (R&D Systems) according to the manufacturer’s instructions. The signal was detected using Western Lightning ECL Pro (PerkinElmer) and individual chemokine spots quantified using Image Studio Lite software (version 5.2).
Statistical analysis
Graphpad prism 8.0.2 software was used to calculate significance between the samples. Statistical tests are indicated in each figure legend. Unless specifically stated all data are representative of > 3 separate experiments. Error bars represent SEM and are derived from triplicate experimental conditions. P values ≤0.05 were considered significant.
Discussion
This study demonstrates that low-dose MTX curbs in vivo NB growth leading to substantial tumor regression and remodeling of the tumor’s immune landscape when combined with PD-1 and TGFβ blockade. This combined treatment is also able of inducing the recruitment of immune cells and the activation of CD8+ T cells and NK cells against PDOTS generated from human NB patients.
ICIs have been extensively shown to trigger T-cell effector function to control tumor growth in both mice and human cancers [
46,
47]. However, antitumor immune responses to individual immunotherapeutic agents remain limited to a subset of patients [
1]. This is because immunotherapy is ineffective for tumors that, as high-risk NB, are both lacking immune cells and characterized by an immunosuppressive TME [
1]. Therefore, developing combination therapies to induce recruitment of immune cells and target the immunosuppressive factors released in the TME could be a successful strategy to improve the efficacy of treatment regimens [
48].
Chemotherapy agents, originally known to directly inhibit or kill malignant cells, have recently been found to promote antitumor immunity by increasing tumor immunogenicity, enhancing T-cell infiltration, or reducing immunosuppressive cell populations [
49]. Some of these chemotherapy drugs are known to activate tumor-specific T cells by inducing ICD when administered at low doses [
7]. The effectiveness of these drugs increases in association with immune checkpoint therapies [
50]. Indeed, combined treatment of DX with PD-1 or PD-L1 antibodies resulted in a significant increase in efficacy in metastatic triple-negative human breast cancer and in various murine tumors, such as melanoma and breast cancer [
51,
52]. Similarly, OXP treatment has been found to increase the efficacy of anti-PD-L1 therapy in murine colorectal cancer [
53].
NB cells are capable of secreting a variety of soluble mediators that can suppress lymphocyte activation, including TGFβ [
23,
38,
54]. Indeed, high levels of tumor TGFβ have been associated with reduced event free survival [
43]. Activation of the TGFβ pathway has also recently emerged as a potential factor responsible for primary resistance to immune checkpoint blockade therapy [
11,
12]. Combined anti-TGFβ/anti-PD-1 treatment led to profound and durable antitumor responses in urothelial, melanoma, and breast cancer models, and promoted the establishment of immunological memory in a tumor rechallenge model [
55,
56]. The induction of long-term tumor-specific T-cell memory is attributed to immune checkpoint PD-1 and PD-Ll blockade [
55,
56]. In this context, TGFβ blockade has been found to restore an immunity-friendly environment in NB capable of unleashing the full potential of reactive immune cells and increasing their persistence [
57]. Based on these considerations, here we evaluated the immunomodulatory effect of 6 chemotherapy agents administered at low doses, including those used in the therapy of high-risk NB and/or able to induce ICD, in two transplantable 9464D and 975A2 murine models that closely recapitulate the molecular and biological features of high-risk human NBs [
27,
58‐
61]. Low-dose MTX was found to have the greatest immunostimulatory capacity able to attract immune cells into the NB microenvironment in ex-vivo and in vivo approaches. To date, MTX has been approved for the treatment of several tumors, including acute non-lymphoblastic leukemia and some advanced forms of prostate cancer (clinicaltrials.gov). We found that low-dose MTX triggers T-cell priming in NB through induction of ICD [
62], making the tumor more susceptible to immune-mediated attack by up-regulating the expression of MHC class I molecules (Supplementary Fig. S
12). Interestingly, low-dose MTX also results in a concomitant increase in PD-L1 levels (Supplementary Fig. S
12), thus strengthening the idea that chemotherapy alone may not provide a lasting therapeutic effect. This finding is consistent with the recent evidence indicating that treatment regimens combining multiple immunotherapeutic strategies, some of which even engaging innate and adaptive immunity, are more effective in NB patients than monotherapies [
63‐
65]. We found that low-dose MTX in combination with TGFβ and PD-1 blockers is able to remodeling the landscape of tumor-infiltrating immune cells by compensating for the lack of immune cells that characterize high-risk NBs [
20‐
22] (Fig.
7F). Consistently, dual PD-1/TGFβ blockade has recently been shown to i) make human tumor cells more sensitive to different chemotherapeutic agents by altering their plasticity [
66]; ii) enhance the cytolytic activity of NK and T cells towards tumor cells [
67‐
69]; and iii) up-regulate the expression of immune response genes, including those encoding multiple chemokines, such as CCL5, associated with immune cell infiltration and enhanced anti-tumor activity [
70]. Indeed, MTP treatment resulted in an enrichment of DCs and activated CD8
+ T cells and NK cells in both mouse models of NB, supporting the key role of these immune cells in controlling NB growth, as observed in human NB speciments [
20,
22]. We noted that this treatment also induced an enrichment of immunosuppressive populations, such as Treg and neutrophils. The increase of Tregs is unexpected considering the role of TGFβ in maintaining peripheral Treg cells [
71]. We interpret this increase as indicative of a strong immune response that results in the recruitment of immune cells, including Treg, to sites of inflammation. As seen for other tumors, the presence of activated effector cells is indicative that the increase in Treg is not sufficient to create a substantially immunosuppressive TME [
39]. Like macrophages, murine neutrophils are distinguished into antitumor and protumor neutrophils [
72‐
74]. In the early phase of tumorigenesis, neutrophils appear to contribute to the antitumor immune response, perhaps through stimulation of adaptive immunity and activation of CD8
+ T cells [
72]. Cross-talk between neutrophils and activated T cells resulted in substantial upregulation of the costimulatory molecules on the surface of neutrophils, capable of enhancing T-cell proliferation in a positive feedback loop [
75].
Recent evidence highlighted that, even if the adaptive immune system is compromised [
76], or T-cell function cannot be fully recovered by PD-1 inhibitors under specific circumstances [
77], PD-1/PD-L1 inhibitors can still increase antitumor efficacy. This is because other types of immune cells, such as DCs, TAMs and NK cells, are also responsive to PD-1/PD-L1 antagonists [
76,
78‐
82], thus strengthening the use of these agents to increase antitumor efficacy.
In this regard, we can assume that the addition of anti-TGFβ along with checkpoint blockade interrupts a hierarchy of immunosuppressive events, consisting of TGFβ that dampens the initial immune response by preventing immune cells from infiltrating tumors, and PD-1/PD-L1 signaling that operating at a later stage suppresses the effector functions of immune cells causing their depletion [
83].
Immune checkpoint blockade is known to reactivate pre-existing intratumoral T cells in human cancer lesions. This reactivation is accompanied by an increase in the chemoattractant production by different immune cell populations [
4,
46,
84]. Interestingly, we noted that, compared with MTX alone, treatment with MTP resulted in a simultaneously increase of different chemokines involved in the recruitment of both lymphoid and myeloid cell populations (Fig.
6G). In addition to promoting recruitment and effector function of tumor-specific CD8
+ T cells, these chemokines are also able to enrich the TME of functional DC and NK cells, two key immune components associated with improved survival of both adult and pediatric cancer patients [
5]. Particularly important is the increase of CCL5, produced by CD8
+ T cells, NK cells and innate lymphoid cells, that is crucial for recruitment of cDC1s, macrophages and Treg cells in the TME [
85]; CXCL9, CXCL10 and CXCL16, produced by DCs and macrophages, that induce the recruitment and activation of NK cells, NKT cells and CD8
+ T cells [
5,
86‐
88]; CCL21, that significantly increase the proportion of T cells, NK cells and DCs within the tumor [
89]; CD40 and FLT3L, produced by intratumoral NK cells, which supports the viability and functions of cDC1s within the TME by promoting their local differentiation from precursor cells [
19,
90]. Interestingly, CXCL9 expression by cDCs has been previously described to mediate the clustering of DC-T cells within lymph nodes [
91], and since interaction between these two immune cell populations is quite rare in tumors [
22,
92], it is possible that increased CXCL9 expression by facilitating these interactions may promote T-cell effector function.
Finally, our findings in human NB specimens strengthen the evidence found in mouse models, providing proof of principle that the proposed combinatorial strategy may serve as a potential alternative therapy in the NB clinic. So far, the treatment protocols under investigation have mostly been derived from therapeutic regimens formulated for adult tumors [
93]. This is a major limitation because childhood tumors are genetically different from their adult counterparts [
94], suggesting the need for alternative therapeutic approaches. Up to now, few studies have combined immunotherapy and chemotherapy in pediatric cancers, mainly due to the original idea that chemotherapy being immunosuppressive may act by inhibiting the beneficts achieved with immunotherapy [
95]. Instead, we believe that the use of metronomic low dose of chemotherapeutics such as MTX, capable of exposing the host immune system to large amounts of tumor antigens and damage-associate molecular patterns (DAMPs), could overcome the relative coldness of childhood cancer, as high-risk NB. Therefore, this study provides a rational approach based on intra-tumoral recall of immune effector cells that allows for greater efficacy of the proposed immunotherapy [
96,
97]. Further investigation in a prospective study with a larger number of human NB samples, will be crucial to confirm therapeutic efficacy of MTP treatment.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.