T‐cell derived extracellular vesicles prime macrophages for improved STING based cancer immunotherapy

Abstract

A key phenomenon in cancer is the establishment of a highly immunosuppressive tumour microenvironment (TME). Despite advances in immunotherapy, where the purpose is to induce tumour recognition and hence hereof tumour eradication, the majority of patients applicable for such treatment still fail to respond. It has been suggested that high immunological activity in the tumour is essential for achieving effective response to immunotherapy, which therefore have led to exploration of strategies that triggers inflammatory pathways. Here activation of the stimulator of interferon genes (STING) signalling pathway has been considered an attractive target, as it is a potent trigger of pro‐inflammatory cytokines and types I and III interferons. However, immunotherapy combined with targeted STING agonists has not yielded sustained clinical remission in humans. This suggests a need for exploring novel adjuvants to improve the innate immunological efficacy. Here, we demonstrate that extracellular vesicles (EVs), derived from activated CD4⁺ T cells (T‐EVs), sensitizes macrophages to elevate STING activation, mediated by IFNγ carried on the T‐EVs. Our work support that T‐EVs can disrupt the immune suppressive environment in the tumour by reprogramming macrophages to a pro‐inflammatory phenotype, and priming them for a robust immune response towards STING activation.

1. INTRODUCTION

The immune system, in particular activated T cells, are important for controlling tumour growth. However, tumours are often characterised with poor infiltration of activated T cells, or presence of dysfunctional T cells incapable of killing tumour cells (Binnewies et al., 2018). Tumour associated macrophages (TAMs) are one of the most abundant immune cell populations in solid tumours (Broz et al., 2014; Franklin et al., 2014) and are essential for regulating anti‐tumoural T cell responses (Etzerodt et al., 2019). The phenotype and functions of TAMs are highly plastic. However, they can broadly be classified as either anti‐tumourigenic or pro‐tumourigenic. The majority of TAMs in solid tumours are pro‐tumourigenic and play a role in inducing the immune suppressive tumour microenvironment (TME), favouring tumour progression, cancer cell invasion and metastasis (DeNardo & Ruffell, 2019). Therapeutic reprogramming TAMs to an anti‐tumourigenic phenotype is, therefore, a highly attractive strategy to improve current anti‐cancer immunotherapy (Cassetta & Pollard, 2018).

Activation of the stimulator of interferon genes (STING) signalling pathway is an effective trigger of type I Interferon and pro‐inflammatory cytokines, supporting activation of the host immune system and polarisation of macrophages to an anti‐tumourigenic phenotype (Ohkuri et al., 2017; Wang et al., 2022). Activation of the STING pathway involves multiple factors, but most importantly, the enzyme cGAS detect accumulated cytosolic DNA, which result in the production of cyclic‐GMP‐AMP (cGAMP). Binding of cGAMP to STING initiates a signalling cascade through TBK1, leading to downstream activation of the transcription factors IRF3 and NF‐kB (Almine et al., 2017; Dunphy et al., 2018; Jønsson et al., 2017; Li et al., 2019). This support upregulation of type I interferon genes, and various cytokines and chemokines (Jønsson et al., 2017; Khoo & Chen, 2018; Li & Chen, 2018). In addition to trigger macrophages, STING activation has also shown to support infiltration and activation of cytotoxic T cells in the TME (Diamond et al., 2011; Ding et al., 2018; Woo et al., 2014) leading to increased tumour control in various tumour models (Corrales et al., 2015; Demaria et al., 2015; Li et al., 2016; Sivick et al., 2018). However, immunotherapy targeting STING has not yielded sustained clinical remission in humans (Amouzegar et al., 2021; McWhirter & Jefferies, 2020), which may suggest a need for exploring novel adjuvants to improve the innate immunological efficacy.

Extracellular vesicles (EVs) have emerged as a novel mechanism of cellular communication and play a key role in regulating antitumour immune responses (Becker et al., 2016; Marar et al., 2021). EVs are small enveloped vesicles ranging in size from 30 to 1000 nm. They contain several biologically active molecules comprising proteins, lipids and nucleic acids, to modulate the function of recipient cells (Kalluri & LeBleu, 2020; Wen et al., 2017; Yáñez‐Mó et al., 2015). They are produced by all cells and formed either by outward budding from the cell membrane as microvesicles or formed inside the cells in endosomes and released as exosomes (Colombo et al., 2014; Mathieu et al., 2019; van Niel et al., 2018). Importantly, during the interaction of T cells with antigen‐presenting cells (APCs), the T cells release high numbers of EVs (Blanchard et al., 2002) in a polarised manner at the interface with the APC (Choudhuri et al., 2014). These EVs are transferred unidirectionally to the APCs (Mittelbrunn et al., 2011) to enhance their function (Céspedes et al., 2022; Torralba et al., 2018). In addition, EVs from CD4⁺ T cells have also been demonstrated to promote antigen‐specific B cell responses (Lu et al., 2019). T‐cell derived EVs contain, among other molecules, cytokines (Fitzgerald et al., 2018) and RNA fragments (Chiou et al., 2018; Greisen et al., 2017; Okoye et al., 2014), and may therefore theoretically have the capacity to regulate surrounding immune cells by both direct receptor activation and by regulating gene expression. Some suggest T‐cell derived EVs carry DNA capable of activating STING signalling in the APCs (Torralba et al., 2018). However, there is still very limited knowledge on how such T‐cell derived EVs in fact modulates the immune response and in particular macrophages in the context of cancer (Veerman et al., 2019).

Since CD4⁺ T cells play a key role in activating machrophages (Mosser & Edwards, 2008) we investigated how EVs derived specifically from these cells modulate macrophage function. In the present study, we demonstrate that EVs derived from activated CD4⁺ T cells in particular prime the signalling capacity of the cGAS‐STING pathway in macrophages, supporting enhanced production of type I IFN and T‐cell chemotaxic cytokines. We demonstrate that EVs from activated CD4⁺ T cells can be utilised as an adjuvant to improve the efficacy of in vivo STING agonist therapy for cancer.

2. RESULTS

2.1. CD4⁺ T‐cell derived EVs sensitise macrophages to STING activation

The ability of activated CD4⁺ T‐cell derived EVs (T‐EVs) to modulate activation of the innate immune system was examined by exploring different pattern recognition pathways in macrophages after priming with T‐EVs. As some reports that T‐EVs may carry surface‐associated DNA (Torralba et al., 2018), we consequently, and throughout the entire study, pre‐treated T‐EVs with DNase I prior to stimulation, to prohibit direct STING activation. However, in our experimental settings we did not see any direct STING activation from the T‐EVs alone (Figure S1A–C). THP‐1 cells primed with T‐EVs was exposed to stimulation with the STING ligand 2′,3′‐cGAMP (cGAMP), the TLR3/RIG‐I ligand poly(I:C) or the TLR4 ligand LPS, and assessed for cytokine production after 20 h. Priming with T‐EVs resulted in a significantly enhanced production of type I IFN and IL‐6 in response to activation by STING and TLR4. By contrast, priming did not affect the response to TLR3/RIG‐I stimulation (Figure 1a,b). Next, we showed that the priming effect of T‐EVs in THP‐1 cells was controlled in a cGAMP‐concentration dependent manner. Interestingly, primed THP‐1 cells were able to produce robust type I IFN and CXCL‐10 at even very low cGAMP concentrations as compared to non‐primed cells (Figure 1c,d). This priming effect was not restricted to THP‐1 cells only, as primary human monocyte‐derived macrophages (MDMs) primed with T‐EVs also gave significantly higher cytokine production to low‐level dosage of cGAMP (Figure 1e,f). Importantly, treating THP‐1 or primary macrophages with T‐EVs alone did not result in production of type I IFN, suggesting that T‐EVs do not directly activate STING signalling (Figure 1e,f). Finally, we evaluated the dose‐dependent effect of T‐EV particles needed to support priming of the STING pathway. Here, we found that at least 5 × 10⁷ particles were needed for priming of the cells (Figure 1g). To summarise, our data indicate that EVs derived from activated CD4⁺ T cells are able to modulate the function of macrophages and in particular renders the STING pathway more sensitive to cGAMP stimulation without directly activating STING signalling.

FIGURE 1.

Activated CD4⁺ T‐cell derived EVs sensitises macrophages to STING activation. (a and b) THP‐1 cells were treated with T‐EVs (1.5–2 × 10⁹) from activated CD4⁺ T cells or EV‐free media for 1 h prior to stimulation with either cGAMP (0.5 μg), Poly(I:C) (0.1 μg) or LPS (0.5 μg/mL). After 20 h, the supernatant was analysed for secretion of (a) functional type I IFN and (b) IL‐6. Data show mean + SEM and individual replica of T‐EVs from two distinct T cell donors, each in duplicates. Unpaired t‐test (two‐tailed). (c and d) THP‐1 cells were treated with T‐EVs (3 × 10⁹) from activated CD4⁺ T cells or EV‐free media for 1 h prior to stimulation with increasing amounts of cGAMP as indicated. After 20 h, the supernatant was analysed for secretion of (c) functional type I IFN and (d) CXCL‐10. Data show mean + SEM and individual replica, and are representative of two independent experiments. (e) THP‐1 cells and (f) human monocyte derived macrophages (MDMs) were treated with T‐EVs (3 × 10⁹) from activated CD4⁺ T cells or EV‐free media for 1 h prior to stimulation with cGAMP (0.5 μg). The production of type I IFN was determined after 20 h of stimulation. Data in (e) shows mean +SD of six independent experiments and data in (f) shows mean +SD of six different experiments all with T‐EVs from distinct T cell donors. Each datapoint indicate mean value of duplicates. Wilcoxon test (two‐tailed). (g) THP‐1 cells were treated with increasing amounts of T‐EVs or EV‐free media for 1 h prior to stimulation with cGAMP (0.5 μg) as indicated. The production of type I IFN was determined after 20 h. Data show mean +SEM and individual replica and are representative of two distinct T‐EV donors. UT, untreated, and received only EV‐free media and lipofectamine. T‐EV, EVs derived from activated CD4⁺ T cells.

2.2. T‐EVs modulate STING signalling independent of cGAS and intravesicular cGAMP

Importantly, DNase I treatment of T‐EVs do not remove potential contaminating DNA inside the particles. Thus, to exclude that such mechanism could be a result of the priming effect, we used cGAS‐deficient THP‐1 cells as well as a STING‐deficient control (Figure 2a,b) generated by CRISPR‐Cas9 editing (Holm et al., 2016). The KO cell‐lines responded similarly to PolyI:C stimulation (Figure 2c) but upon priming with T‐EVs and following cGAMP stimulation, we observed that cGAS‐deficient THP‐1 cells responded equally well as WT THP‐1 cells (Figure 2d). This support that the priming of macrophages by T‐EVs is not mediated by intra‐vesicular DNA. In parallel, priming of STING‐deficient THP‐1 cells with T‐EVs and stimulating with cGAMP did not result in any IFN production (Figure 2d). Potentially, T‐EVs may also carry fractions of cGAMP, if such are produced by activated T cells. To rule out this scenario, we conducted a mass spectrometry‐based detection assay of cGAMP in the T‐EVs, but was unable to measure any detectable levels of cGAMP (Figure 2e). Together, our data suggest that the function of T‐EVs on sensitizing STING signalling is not mediated by neither DNA nor cGAMP inside the T‐EVs.

FIGURE 2.

T‐EVs modulate STING signalling independent of cGAS and intravesicular cGAMP. (a) Schematics of the cGAS‐STING pathway. (b) Western blot analysis of cell lysates from WT, cGAS^−/−, and STING^−/−, THP‐1 cells. Vinculin (VCL) was used as a loading control. Data represents 1 experiment. (c) WT, cGAS^−/−, and STING^−/− THP‐1 cells were stimulated with Poly(I:C) (0.1 μg). The production of type I IFN response was determined after 20 h of stimulation. Data show mean + SEM and individual replica of two independent experiments in duplicates or triplicates. ns, non‐significant; unpaired t‐test (two‐tailed). (d) WT, cGAS^−/−, and STING^−/− THP‐1 cells were stimulated with T‐EVs (3 × 10⁹) or EV‐free media for 1 h prior to stimulation with cGAMP (0.5 μg). The type I IFN production was determined after 20 h of stimulation. Data show mean + SEM and individual replica of two independent experiments in duplicates or triplicates. Unpaired t‐test (two‐tailed). (e) T‐EVs were analysed by mass spectrometry for presence of cGAMP. Data show cGAMP concentration in each of three distinct T‐EV samples as well as in samples spiked‐in with indicated amount of cGAMP. WT, Wild type. UT samples in (c) and (d) are identical.

2.3. T‐EVs prime THP‐1 cells for enhanced STING signalling

To further investigate the mechanism by which T‐EVs prime the STING pathway, we examined potential changes in STING expression and function. After T‐EV priming we did not observe any changes in STING mRNA expression compared to untreated cells (Figure 3a). Next, we conducted a time‐kinetic stimulation experiment with cGAMP. Cells primed with T‐EVs resulted in a robust Type I IFN production as early as 4 h after cGAMP was added to the culture (Figure 3b). We further examined if removing T‐EVs from the THP‐1 cell cultures prior to cGAMP stimulation would still support STING activation. Indeed, we observed that T‐EV priming, independent of time prior to cGAMP stimulation, resulted in significant enhanced type I IFN production (Figure 3c). This suggest that priming with T‐EVs not only make cells more sensitive to low‐level dosage of cGAMP, but also induce a more rapid initiation of the signalling pathway.

FIGURE 3.

T‐EVs prime THP‐1 cells for enhanced STING activation. (a) THP‐1 cells were stimulated with T‐EVs (3 × 10⁹) or EV‐free media. After 6 h of stimulation, the mRNA expression of STING was determined. Data shows mean + SEM and individual duplicates and is representative of two independent experiments. (b) THP‐1 cells were stimulated with T‐EVs (3 × 10⁹) for 1 h prior to stimulation with cGAMP (0.5 μg). At indicated time‐points after cGAMP stimulation the supernatant was harvested and the production of type I IFN response was determined. Data shows mean +SEM and individual replica of three independent experiments each in duplicates. Unpaired t‐test (two‐tailed). (c) THP‐1 cells were stimulated with T‐EVs (1 × 10⁹) or EV‐free media at the indicated time‐points prior to cGAMP stimulation (0.5 μg). Right before cGAMP stimulation, the T‐EVs were washed out. After 20 h of cGAMP stimulation, the supernatant was collected and the production of type I IFN was determined. Data show mean +SEM and individual replica of two independent experiments. Unpaired t‐test (two‐tailed). (d and e) THP‐1 cells were stimulated with T‐EVs (3 × 10⁹) or EV‐free media for 1 h prior to stimulation with cGAMP (0.5 μg). At indicated time‐points after cGAMP stimulation, the cells were harvested. Expression and phosphorylation of indicated proteins were analysed by Western blotting. Data are representative of three independent experiments. VCL was used as a loading control. VCL in (d) and (e) are identical.

To better understand how T‐EV priming modulates response to STING activation, we investigated the early signalling pathway events upstream of IFN and cytokine gene expression, using western blotting analysis. T‐EV priming alone did not lead to any clear signals in phosphorylation of STING and IRF3, though a minor increased intensity was observed for TBK1 (Figure 3d). However, comparing cGAMP stimulation of unprimed and primed cells, clearly showed more intense phosphorylation bands for T‐EV primed cells (Figure 3d). Furthermore, both pTBK1 and pSTING was increased already at 1 h after cGAMP stimulation, supporting our earlier results suggesting more rapid pathway activation. Activation of the STING pathway has also been reported to engage the IKK‐NF‐kB signalling pathway. Therefore, we next examined activation of p65, a component of the NF‐kB complex, following T‐EV and cGAMP stimulation. We observed both increased and more rapid phosphorylation of p65 upon priming with T‐EVs prior to cGAMP stimulation compared to cGAMP stimulation alone (Figure 3e). Also, the data indicated that T‐EV priming alone, resulted in early phosphorylation of p65 which was sustained for up to 2 h after priming (Figure 3e). In all, these results suggest that the priming effect of T‐EVs is a rapid event, and can be sustained in stimulated cells for long periods after the priming event happened.

2.4. EVs from primarily activated CD4⁺ T cells enhances STING signalling

EVs are produced continuously from T cells, whether they are activated or not. Thus, to investigate whether STING sensitisation was induced by T‐EVs from CD4⁺ T cells in general, we isolated T‐EVs from CD4⁺ T cells either (i) stimulated with anti‐CD3/anti‐CD28 together with low‐dosage of IL‐2; or (ii) left in the presence of only low‐dosage IL‐2. Activation status of the CD4⁺ T cells at the time of collection of T‐EVs, was assessed by surface expression of CD69, CD134 and CD40L (Figure S2A–F). We found that activation of CD4⁺ T cells was followed by an increased number of T‐EVs in the cell culture supernatant, which may reflect proliferation of the cells (Figure 4a). We further characterised the size distribution of the T‐EVs and presence of particular proteins associated with EVs. We found no difference in the size distribution between T‐EVs from activated and non‐activated CD4⁺ T cells (Figure 4b,c). The majority of T‐EVs from both activated and non‐activated CD4⁺ T cells were between 50 and 200 nm, indicating that they consist of a mixture of both exosomes and microvesicles. Western blot analysis of lysate from T‐EVs and their corresponding cells of origin showed that the surface protein CD9 was exclusively expressed in the EV‐fraction, whereas the cytosolic protein HSP70 and the marker for apoptotic bodies, calreticulin, was primarily expressed by the cellular fraction (Figure 4d). Transmission electron microscopy showed presence of vesicular structurers in the size range of 100–200 nm (Figure S3A). We further showed that treatment of CFSE labelled T‐EVs with Triton X‐100 completely disrupted the presence of T‐EVs on flow cytometry analysis, further confirming the vesicular nature of the T‐EVs (Figure 3b). This confirmed the identity of EVs and indicated low presence of apoptotic bodies in the T‐EV preparations. Importantly, when priming THP‐1 cells with these two batches of T‐EVs from multiple T‐cell donors, we observed that only T‐EVs from activated CD4⁺ T cells showed a significant priming of STING (Figure 4e). Consistent with this, we further showed that phosphorylation of p65 was primarily induced by T‐EVs from activated CD4⁺ T cells (Figure 4f). These data support that activation of CD4⁺ T cells leads to secretion of a distinct type of T‐EVs with specific content capable of sensitizing the STING pathway in macrophages.

FIGURE 4.

T‐EVs from primarily activated CD4+ T cells enhances STING signalling. T‐EVs were isolated from CD4⁺ T cells that were either activated (a) with anti‐CD3 (1 μg/mL) and anti‐CD28 (1 μg/mL), or left non‐activated (NA) for 48 h in presence of only IL‐2 (10 ng/mL). The T‐EVs were analysed by tunable resistive pulse sensing using a qNano for (a) concentration and (b and c) size distribution. (a) data shows T‐EVs from each of four distinct donors. Paired t‐test (two‐tailed). Data in (b) show mean size of T‐EVs from each of the four distinct donors. ns, not significant, paired t‐test (two tailed). Data in (c) show mean +SD of T‐EVs from each of the 4 distinct donors. (d) Western blot of paired cell lysate and T‐EVs from two individual donors. E) THP‐1 cells were treated with T‐EVs (3 × 10⁹) from either A or NA CD4⁺ T cells for 1 h prior to stimulation with cGAMP (0.5 μg). The production of type I IFN was determined upon 20 h of stimulation. Data shows mean +SD and individual mean values of duplicates, from six different experiments. Paired t‐test (two‐tailed). (f) THP‐1 cells were stimulated with T‐EVs (3 × 10⁹) from either A or NA CD4⁺ T cells for 1 h prior to stimulation with cGAMP (0.5 μg). After 1 h of stimulation, the cells were harvested for Western blot analysis. Data are from one experiment using T‐EVs from two distinct donors. Vinculin (VCL) was used as a loading control. NA, non‐activated. A, activated. ns, not significant.

2.5. T‐EVs transfer pro‐inflammatory cytokines to enhance macrophage function

Various pro‐inflammatory cytokines are known to trigger induction of NF‐kB activation, and there is also evidence that cytokines can be transported between cells by EVs (Fitzgerald et al., 2018). Prompted by our data, that T‐EVs priming let to rapid pP65 activation, we therefore analysed T‐EVs for the presence of 10 inflammatory‐associated cytokines using a highly sensitive multiplex immunoassay. Among these, we found that T‐EVs from activated T cells contained both IFNγ, TNFα and IL‐2 (Figure 5a). Pre‐treatment of THP‐1 cells with these recombinant cytokines prior to cGAMP stimulation showed that mainly IFNγ and to a lesser extent TNFα enhanced the cGAMP‐induced type I IFN production in a dose‐dependent manner (Figure S4A,B). Importantly, neither IFNγ, TNFα and IL‐2 alone induced production of type I IFN (Figure 4a–c). Next, we compared the priming of THP‐1 cells with a high‐dose of recombinant IFNγ or TNFα compared to T‐EV. These data showed that IFNγ alone was able to prime cells to a similar level as seen for T‐EVs, whereas there was no further additive effect of combining IFNγ and TNFα (Figure 5b). To test whether priming of the STING pathway was induced by cytokine content within T‐EVs, we next blocked the receptors on THP‐1 cells using antibodies recognizing IFNGR1 or TNFR1. To obtain efficient inhibition of IFNγ signalling we did also treat the T‐EVs with anti‐IFNγ before stimulating the cells with the T‐EVs and subsequently cGAMP. Blocking TNFα signalling alone minorily affected type I IFN production (Figure 5c) whereas blocking IFNγ signalling alone substantially abrogated the STING priming effect of T‐EVs (Figure 5d). Blocking IFNγ and TNFα signalling simultaneously, did however, not further abrogate the STING priming effect of T‐EVs (Figure 5e). This suggest that T‐EV‐associated IFNγ may play an important role for priming STING signalling. These data further support that IFNγ may be carried on the surface of the T‐EVs.

FIGURE 5.

T‐EVs transfer proinflammatory cytokines to enhance macrophage function. (a) EVs from either activated (A) or non‐activated (NA) CD4⁺ T cells were lysed in 2.5% Triton‐X‐100 and the presence of cytokines was measured using mesoscale or ELISA for IFNγ. Data show mean +SD and individual concentrations of n = 5 (A) and n = 2 (NA). (b) THP‐1 cells were stimulated with either T‐EVs (1 × 10⁹), recombinant TNFα (10 ng/mL) or recombinant IFNγ (10 ng/mL) for 1 h prior to stimulation with cGAMP (0.5 μg). The production of type I IFN was determined after 6 h of stimulation. Data show mean +SEM and individual replica from two independent experiments. (c) THP‐1 cells were treated with anti‐TNFR1 (10 μg/mL) or IgG1 (10 μg/mL) for 30 min at 37°C. The cells were stimulated with T‐EVs (1 × 10⁹) for 1 h followed by stimulation with cGAMP (0.5 μg). The production of type I IFN was determined after 6 h of stimulation. Data show mean +SEM and individual replica of three independent experiments. (d) THP‐1 cells were treated with anti‐IFNGR1 (20 μg/mL) or IgG1 (20 μg/mL) for 30 min at 37°C. T‐EVs were treated with anti‐IFNγ (10 μg/mL) or IgG1 (10 μg/mL) and incubated for 15 min prior to use for stimulations. The cells were stimulated with the T‐EVs (1 × 10⁹) for 1 h followed by stimulation with cGAMP (0.5 μg). The production of type I IFN was determined after 6 h of stimulation. Data show mean +SEM and individual replica of two independent experiments with T‐EVs from three distinct donors in total. (e) THP‐1 cells were treated with a combination of anti‐IFNGR1 and anti‐TNFR1 as described for (c) and (d). Data show mean +SEM and individual replica of four independent experiments. Unpaired t‐test (two‐tailed). Due to experimental setup, some of the datapoints for panel b) cGAMP and T‐EV+cGAMP are identical to datapoints in panel (d) and (e) ‘No block’.

2.6. T‐EVs enhances the anti‐tumoural function of cGAMP in vivo

So far, our results indicate that T‐EVs can potentially function as an adjuvant for therapies targeting the STING pathway. To explore this further we moved to an in vivo syngeneic cancer mouse model. First, we confirmed cross‐species activities, by exploring how T‐EVs derived from activated murine CD4⁺ T cells functioned compared to our observations seen for human T‐EVs. We primed murine bone‐marrow derived macrophages (BMMs) with murine T‐EVs prior to stimulation with a suboptimal dose of cGAMP. Similar to our observations in human macrophages, murine T‐EVs enhanced the STING‐induced production of IFN‐beta (Figure 6a). We further found that priming of BMMs with murine T‐EVs alone induced phosphorylation of p65 (Figure 6b). To confirm that our treatment was not toxic to cancer cells, we stimulated MC38 cells in vitro with murine T‐EVs and cGAMP, but observed no effect on proliferation or cell death (Figure S5).

FIGURE 6.

T‐EVs enhance antitumour efficacy of cGAMP. (a) Murine bone marrow derived macrophages (BMMs) were stimulated with murine T‐EVs (1 × 10⁹) or EV‐free media 1 h prior to suboptimal cGAMP stimulation (0.05 μg). The secretion of IFN‐beta was determined after 20 h of stimulation. Data show mean +SEM and individual replica of two independent experiments each in duplicates. (b) Murine BMMs were stimulated with murine T‐EVs (1 × 10⁹) or EV‐free media. At indicated time‐points after T‐EV stimulation, the cells were harvested and analysed by western blotting for phosphorylation of P65. Data is from one experiment. VCL was used as a loading control. (c) MC38 tumour bearing mice were treated with different amounts of murine T‐EVs, administered intratumourally (IT) two times with 3 days interval as indicated with black arrows on the figurers. Treatment started on day 9 after tumour cell inoculation. Data show mean ± SEM of tumour volume up to day 19 after tumour cell inoculation. n = 4 in all groups. (d) MC38 tumour bearing mice were treated with different amounts of cGAMP as indicated, administered IT three times with 3 days interval, as indicated with black arrows on the figurers. Treatment started on day 9 after tumour cell inoculation. Data show mean ± SEM of tumour volume up to day 15 after tumour cell inoculation. n = 6 in all groups. (e) and (f) MC38 tumour bearing mice were treated with either T‐EVs alone (1.5 × 10⁸), cGAMP alone (1 μg) or a combination. Mice were treated IT three times with 3 days interval as indicated with black arrows on the figurers, starting on day 9 after tumour cell inoculation. Data in (e) show mean ± SEM of tumour volume in mice treated with either Vehicle (n = 9), T‐EVs (n = 9), cGAMP (n = 9), or a combination of T‐EVs and cGAMP (n = 8) up to day 15 after tumour cell inoculation. Data in (f) show difference in tumour growth from initiation of treatment on day 9 and until day 15, calculated as ∆TG. Bars indicate mean ± SEM and individual value of ∆TG. TG, tumour growth. (g and h) MC38 tumour bearing mice were treated with either T‐EVs alone (1.5 × 10⁸), cGAMP alone (10 μg) or a combination. Mice were treated IT two times with 2 days interval as indicated with black arrows on the figurers, starting on day 10 after tumour cell inoculation. Data in (g) show mean ± SEM of tumour volume in each group up to day 19 after tumour cell inoculation. Data in (h) shows probability of survival up to day 64 after tumour cell inoculation. n = 9 in each group. (i) Mice that obtained complete tumour regression in (g) was re‐challenged with inoculation of MC38 cells subcutaneously in left flank. As controls was used C57B/6J mice (n = 4). Data show mean ± SEM of tumour volume in each group of either complete responders or control mice.

To investigate the anti‐tumoural function of T‐EVs, we next treated C57BL/6 mice bearing a subcutaneous MC38 colon adenocarcinoma with increasing amounts of murine T‐EVs administered intratumourally in two repeated doses with 3‐days interval. We found no anti‐tumoural effect of the T‐EVs alone (Figure 6c, S6). Next, we titrated intratumoural administration of cGAMP to determine a suitable suboptimal dose, which we defined to be 1μg cGAMP leading to only a minor delay in tumour growth (Figure 6d, Figure S7). We then treated MC38 tumour‐bearing mice with a combination of T‐EVs and low‐dosage of cGAMP in three repeated dosage with 3‐days interval. After the first dosage, we observed a clear tendency to delayed tumour growth in the group receiving combination therapy compared to cGAMP monotherapy (Figure 6e, S8). When we calculated the tumour growth development during the treatment intervention, we saw a clear trend toward stagnated tumour growth in the combination group (Figure 6f). To push for a better anti‐tumoural effect, we next treated MC38‐bearing mice with a combination of T‐EVs and high‐dose of cGAMP. Here we observed similar delay in tumour growth between mice receiving combined T‐EV and high‐dose cGAMP, and high‐dose of cGAMP alone (Figure 6g, S9). Intriguingly, in the group of mice recieving combinational treatment, we had 2/9 mice with complete tumour regression (Figure 6h). Importantly, these mice were resistant to re‐challenge with MC38 (Figure 6i) indicating development of anti‐tumour immunity. Altogether, our data indicate that treatment with T‐EVs can prime the tumour microenvironment for a more potent immune response to cGAMP stimulation supporting improved control of tumour growth.

3. DISCUSSION

Activation of the STING pathway in the tumour microenvironment has shown profound anti‐tumoural responses in preclinical models. However, it is unclear why these effects, to date, have not been achieved in clinical settings (Amouzegar et al., 2021). The usage of STING agonist as monotherapy and in combination with checkpoint inhibitors, lack to generate results that imply partly or complete responders. Potentially, this reflect that the dosage of STING agonists used, are either too low to induce signals or too high, which can ablate the essential immune cells in the TME needed for killing tumour cells (Sivick et al., 2018). Thus, there is a need to discover methods that expand the therapeutic index for proper STING activation in human settings.

One plausible explanation as to why STING agonists do not mount a powerful anti‐tumoural response is a physiological high threshold for activating the pathway in immune cells. A major question related to this is what endogenous level of cGAMP we can expect to measure in the TME. We currently lack compelling data demonstrating that endogenous levels of cGAMP accumulates in tumour tissue; and whether such concentrations are applicable for driving a potent STING‐dependent immune response is difficult to answer. Some reports using murine tumour models suggest that blockage of cGAMP hydrolyzation in the extracellular space by the enzyme ENPP1 can support activation of STING (Li et al., 2014). Thus, ENPP1 inhibitors are considered promising therapeutics to enhance the level of extracellular cGAMP (Carozza et al., 2020; Ritchie et al., 2022). Nonetheless, it is unresolved whether these levels of extracellular cGAMP overcomes the threshold of intracellular cGAMP needed for activating the STING pathway. In this study, we describe a promising mode‐of‐action for lowering the threshold for STING activation in immune cells. Cellular transmitters in form of EVs from activated CD4⁺ T cells were proved to prime macrophages and lower their responsiveness to even very small amounts of cGAMP. This mechanism of ‘alertness’ resulted in both a rapid and strong cytokine induction profile, which sensitised tumours in vivo to mount effective tumour growth control at suboptimal dosage of STING agonists.

Conceptionally, the usage of adjuvants to support a strong and broad immunological response has been applied for decades in vaccine settings. Thus, applying such rationale for immunotherapy combination with STING agonists is an intriguing concept. There seem to be a delicate balance between robust activation of STING and induction of apoptosis, which needs to be considered in therapeutic settings. Work from Benoit‐Lizon et al. and Larkin et al. suggest that the dosage of STING agonists to trigger type I IFN responses in various immune cells, will have detrimental effects on the viability of T cells (Benoit‐Lizon et al., 2022; Larkin et al., 2017). Thus, do we favour a powerful inflammatory cytokine response in the TME over the existence of potentially anti‐tumoural lymphocytes? Our work propose that pre‐activation of CD4⁺ T cells, which allow secretion of EVs that primes myeloid cells, can be beneficial prior to a direct STING activation. This could allow administration of STING agonists at dosages that both secure the presence of cytotoxic T cells but also elevates inflammatory cytokines in the TME.

EVs are continuously secreted from cells and it is well appreciated that they can modulate multiple aspects of the TME, where their function closely reflect the characteristics of their cell of origin (Céspedes et al., 2022; Marar et al., 2021; Seo et al., 2018). Thus, it is relevant to point out that in our study, we found that only EVs released from activated CD4⁺ T cells were able to sensitise STING signalling in macrophages. This raise the question whether EVs released from tumour infiltrating CD4⁺ T cells, which are often dysfunctional, may carry similar features or not? The usage of immune checkpoint inhibitors (ICI), that function by releasing the physiological break on T cell activation, could potentially be associated with elevated secretion of STING priming EVs. However, further work will be needed to address this.

Using a selective cytokine analysis, we found that EVs from activated T cells carried a series of known inflammatory signalling cytokines. Similar observations have been reported earlier, where a broad range of cell types seems to have the capacity to encapsulate cytokines within EVs (Fitzgerald et al., 2018; Jung et al., 2020). Interestingly, cytokines within EVs can both be surface bound and internalised, thereby mediating different signalling events to target cells. Our finding indicated that both IFNγ and TNFα was primarily surface bound, as blocking their designated receptors on macrophages prohibited the sensitisation of the STING pathway. Multiple studies have described that tumours responding to ICI can be associated with active IFNγ signalling (Ayers et al., 2017; Gao et al., 2016), and the effect of IFNγ is associated with modulation of cancer cells and their activation of the STING pathway by increasing DNA damage and cGAMP production (Hoekstra et al., 2020;Thibaut et al., 2020; Xiong et al., 2022). In parallel, IFNγ is also essential in activating and inducing a pro‐inflammatory phenotype in macrophages (DiDonato et al., 1997; Mills et al., 2000;Nathan et al., 1983). However, the therapeutic application of IFNγ in cancer have been challenging because of the broad expression of the IFNGR and pleiotropic function of IFNγ in the TME (Gocher et al., 2022). EVs carrying cytokines on their surface may provide a targeted delivery of IFNγ to APCs in the TME and could influence the functional outcome and favour some of the key anti‐tumourigenic functions of IFNγ. Thus, our work support a model where a combination of elevated cGAMP production within IFNγ‐activated cancer cells with IFNγ‐sensitised macrophages, are essential factors supporting modulation of an anti‐tumoural STING‐mediated immune response.

In conclusion, we show that IFNγ containing EVs derived from activated CD4⁺ T cells are able to sensitise macrophages, and potentially other immune cells, for enhanced STING signalling. We speculate that such EVs can disrupt the immune suppressive TME by reprogramming macrophages to a pro‐inflammatory phenotype. In addition, they may render immune cells more sensitive to low‐level STING agonist concentrations in the TME, supporting signalling pathway activation. Although clinical trials of STING agonists as cancer therapy showed mixed results, clinical development remains active (McWhirter & Jefferies, 2020). We see that EVs from activated CD4⁺ T cells may be an exciting new avenue of therapy in combination with either ICI or other therapies supporting increased intratumoural cGAMP production, including radiotherapy or chemotherapy.

4. MATERIALS AND METHODS

4.1. Isolation of human CD4⁺ T cells

Buffy coats from healthy donors were anonymously obtained from the blood bank at Aarhus University Hospital, Skejby, Denmark. Human peripheral blood mononuclear cells (PBMCs) were isolated from the buffy coats using Ficoll‐Paque PLUS (GE Healthcare Bioscience). The PBMCs were cryopreserved in 40% heat‐inactivated fetal bovine serum (FCS, Gibco), 50% RPMI‐1640 (Sigma‐Aldrich) and 10% DMSO (Sigma‐Aldrich) and stored at −150°C. CD4⁺ T cells were isolated from PBMCs by negative selection using EasySep Human CD4⁺ T Cell Isolation Kits (Stemcell) according to manufacturer's instructions.

4.2. Isolation of murine CD4⁺ T cells

The spleen was dissected from healthy C57BL/6 mice and a single cell suspension of splenocytes was prepared by pressing the spleens gently through a 70 μm nylon cell strainer. CD4⁺ T cells were isolated from the splenocytes by negative selection using EasySep Mouse CD4⁺ T Cell Isolation Kits (Stemcell) according to manufacturer's instructions.

4.3. Generation of EVs

The generation and collection of EVs was done in the best possible manner to fulfil the MISEV2018 guidelines (Théry et al., 2018). For generation of EVs the cells were cultured in medium containing EV‐free FCS prepared by ultracentrifugation at 100,000 × g for 20 h to remove EVs from the serum. CD4⁺ T cells were activated in vitro resulting in release of high numbers of EVs into the cell culture supernatant. For activation of human CD4⁺ T cells, approximately 4–6 × 10⁶ isolated human CD4⁺ T cells/well were stimulated in a 6‐well plate pre‐coated with 1μg/mL αCD3 (clone UCHT‐1, R & D System) and cultured in RPMI‐1640 supplemented with 10% EV‐free FCS, 10 mM HEPES (Gibco), 2 mM glutaMAX (Gibco), 200 IU/mL Penicillin, 100 μg/mL Streptomycin, 1 μg/mL soluble αCD28 (clone CD28.2, BD Pharmingen) and 10 ng/mL recombinant human IL‐2 (Roche). For activation of murine CD4⁺ T cells, approximately 4–6 × 10⁶ isolated murine CD4⁺ T cells/well were stimulated in a 6‐well plate pre‐coated with 1μg/ml αCD3 (clone 145‐2C11, BD Pharmingen) and 1μg/ml and αCD28 (clone 37.51, BD Pharmingen) and cultured in RPMI‐1640 supplemented with 10% EV‐free FCS, 10 mM HEPES (Gibco), 2 mM glutaMAX (Gibco), 1X non‐essential amino acids (Gibco), 200 IU/mL Penicillin, 100 μg/mL Streptomycin, 50 μM beta‐mercaptoethanol and 10 ng/mL recombinant human IL‐2 (Roche). Both human and murine CD4⁺ T cells were cultured for 46 h under humidified conditions at 37°C, 5% CO₂.

4.4. Flow cytometry analysis of cells

CD4⁺ T cells were collected for flow cytometry analysis before and after activation as described above. Two million cells were stained using the following antibodies: anti‐CD3‐BV421 (BD Horizon), anti‐CD4‐FITC (BD Pharmingen), anti‐CD45RA‐V500 (BD Horizon), anti‐CD69‐BV786 (BD Horizon), anti‐CD134‐PE‐Cy7 (BD Bioscience) and anti‐CD40L‐BB700 (BD Optibuild). LIVE/DEAD Fixable near‐IR (Thermo Fisher Scientific) was used for live/dead discrimination. Samples were analysed using the NovoCyte Quanteon and data processed in FlowJo.

4.5. Isolation of EVs

EVs were isolated from the cell culture supernatant by differential centrifugation. In brief, the cell culture supernatant was centrifuged for 10 min at 1000 × g followed by 20 min at 2000 × g and 35 min at 10,000 × g. The EVs were then pelleted from the supernatant by centrifugation for 2 h at 100,000 × g. All centrifugation steps were performed at 4°C and the ultracentrifugation was done using 14 mL thinwall polypropylene tubes (Beckmann Coulter) in a TST 41.14 rotor (Sorval). The EV pellet was resuspended in either PBS (Sigma‐Aldrich), T cell media or physiological water (Sigma‐Aldrich). To remove surface associated DNA the EVs were incubated with 10 units DNase I (Thermo Scientific) for 1 h at room temperature in complete RPMI‐1640 media. The EVs used for functional studies were immediately frozen in small aliquots at −80° and kept for up to maximum 6 months. The freeze‐thawing of the EVs were repeating for a maximum of two times.

4.6. Nanoparticle tracking analysis

Throughout the entire study, the concentration of EVs was analysed by nanoparticle tracking analysis (NTA) to adjust the amount of EVs used for functional studies. NTA was performed in scatter detection mode, using a NanoSight NS300 (Malvern Instruments) with a 405 nm laser. The EVs were diluted 100 times in PBS and measurements were performed in five times of 60 s video captures of each sample with camera level 13 and detection threshold 4–5 for all analysis. Data were analysed using the NTA software version 3.4.

4.7. EV characterisation

A qNano Gold (Izon Bioscience) equipped with a NP100 Nanopore (analysis range 50–330 nm, Izon Bioscience) for tunable resistive pulse sensing (TRPS) was used to determine size and concentration of the EVs. The EV samples were diluted in PBS and the EV samples were analysed under identical settings – the same diluent, stretch (47 mm), pressure (10Pa) and voltage (0.6 mV) as used for CPC100 calibration particles (Izon Bioscience). The EV concentration and size distribution were determined using the ‘Izon control suite’ software (Izon Bioscience).

4.8. Electron microscopy of EVs

T‐EVs isolated from app. 25 × 10⁶ activated human CD4⁺ T cells, were washed in PBS and ultracentrifuged a second time at 100,000 × g. The T‐EVs were diluted in 200 μL PBS and stored at −80°C for 1 month prior to analysis by electron microscopy. Electron microscopy was performed at the EMBION facility (embion.au.dk), iNANO, Aarhus University. Three microliters of a diluted  sample was added to a 400 mesh collodion (Sigma Aldrich) and carbon coated copper grid (Pelco) that had been glow discharged 45 s at 25 mA and 39 mbar using an EasiGlow (Pelco). After a 30 s incubation the grid was blotted using a 85 mm filter paper grade 1 (lot. no. 10302, Whatman) and washed/stained 3× with 3 μL 2% uranyl formate (Polysciences Europe GmbH) with blotting steps between each washing/staining step and a final blotting and drying step. Micrographs were collected using a Tecnai Spirit TWIN transmission electron microscope (ThermoFisherScientific) operated at 120 kV using a TemCam F416 CMOS camera and EM‐Menu software (Tvips).

4.9. Flow cytometry analysis of T‐EVs

Human T‐EVs (2 × 10⁸) and murine T‐EVs (1 × 10⁹) were diluted in 100 μL PBS and stained with 20 μM CFSE (eBioscience) for 1 h at RT and put on ice until end of flow cytometry analysis within app 2 h. The EV samples were analysed on the Novocyte Quanteon with clean water as sheath fluid and an extra 0.05 μM filter attached to the sheath fluid line. Threshold was set to SSC‐H 500 and the PMT voltage on both FSC and SSC was adjusted to 400. The PMT voltage for detecting CFSE was adjusted to 1000 based on voltage‐gain tration. The samples were analysed at a flow rate of 5 μL/min. The following control samples were analysed under similar conditions: (a) PBS alone, (b) 20 μM CFSE in PBS and (c) T‐EVs alone. After recording 5 μL of the CFSE‐labelled T‐EVs, the sample was diluted in PBS four times in a 10‐fold serial dilution, and each of the diluted samples were recorded for 1 min. The undiluted CFSE‐labelled T‐EV sample was finally treated with 1% Triton X‐100 (Sigma‐Aldrich) for 30 min at RT and recorded for 1 min. The data was visualised in FlowJo.

4.10. Mass spectrometry for cGAMP analysis

Mass spectrometry analysis for presence of cGAMP within the EVs was done according to previously described (Jønsson et al., 2017). In brief, 6–9.5 × 10⁹ EVs from activated human CD4⁺ T cells were resuspended in 1 mL v/v 80%, 2% acetic acid. The samples were incubated for 15 min at RT and added to NH₂ SPE‐columns pre‐conditioned with first methanol and water. Columns containing the EV samples were washed with methanol and subsequently water and cGAMP was eluted with alkaline methanol (v/v 80% methanol, 5% NH₄OH). The eluted samples were dried using a vacuum centrifuge and dissolved in 50 μL 0.1% formic acid. Ten microliters of the samples were injected on a HSS T3 (2.1 × 100 mm) LC‐column. cGAMP was semi‐quantified using LC‐MS/MS with the transition m/z 338 > 152 in a positive ionisation mode (ESI(+)). As external calibrators, corresponding blank samples were spiked with 0, 50, 100 and 200 nM cGAMP respectively and worked up as the remaining samples prior to analysis.

4.11. Cell culture

All cells were cultured under humidified conditions at 37°C and 5% CO_2.

Human acute monocytic leukaemia cell line (THP‐1) was cultured in RPMI‐1640 (Sigma‐Aldrich) supplemented with 10% heat‐inactivated fetal calf serum (FCS, Gibco), 200 IU/mL Penicillin, 100 μg/mL Streptomycin and 600 μg/mL glutamine (hereafter termed RPMI complete). Infection with mycoplasma was tested and ruled out on a regular basis by Eurofins. For all experiments, THP‐1 cells were differentiated into phenotypically adherent macrophages by stimulation with 100 nM Phorbol 12‐myristate 13‐acetate (PMA, Sigma‐Aldrich) in RPMI complete, for 24 h. The medium was then refreshed with normal RPMI complete, allowing the cells to further differentiate an additional day. They were hereafter defined as macrophages. THP‐1 clones harbouring knock‐out mutations in genes encoding STING and cGAS was previously described (Holm et al., 2016). These cells were cultured and activated similar to WT THP‐1 cells.

Human MDMs were differentiated from PBMCs by culturing the cells in RPMI complete medium supplemented with 10% AB‐positive human serum and 15 ng/mL M‐CSF (PeproTech, 300‐25‐100UG). After 2 days of culturing, the medium was changed to Dulbecco's Modified Eagle Medium (DMEM, Sigma‐Aldrich) supplemented with 200 IU/mL Penicillin, 100 μg/mL Streptomycin, 600 μg/mL glutamine, 10% AB‐positive human serum and 15 ng/mL M‐CSF and the cells were allowed to further differentiate for additional 6–8 days.

Murine bone marrow was extracted from femur and tibia from 8‐ to 12 week‐old healthy C57B6/J mice by centrifugation, and cryopreserved at −150°C in 40% heat‐inactivated fetal bovine serum (FCS, Gibco), 50% RPMI‐1640 (Sigma‐Aldrich) and 10% DMSO (Sigma‐Aldrich). Bone marrow derived macrophages (BMMs) were differentiated by culturing bone marrow cells in presence of L929 SN containing the necessary M‐CSF for differentiation. L929 SN was prepared by collecting the supernatant (SN) from confluent L929 cells. For differentiation, bone marrow cells were cultured in RPMI‐1640 complete medium supplemented with 20% L929 SN. After 2 days the cells were refreshed with RPMI complete medium supplemented with 40% L929 SN and on day 4 the medium was changed to RPMI complete supplemented with 20% L929 SN. After 6 days the BMMs were fully differentiated and harvested using Accutase and seeded for further stimulation.

4.12. Stimulation of cells in vitro

For stimulation of the cells in vitro were used 5 × 10⁴ cells/well in a 96‐well plate of either PMA‐differentiated THP‐1 cells, MDMs or BMMs. The cells were seeded 1 day prior to stimulation allowing the cells to adhere to the bottom of the culture well. Upon stimulation, the culture medium was removed from the cells and 50 μL EV‐free medium containing 1–6 × 10⁹ EVs, or 50 μL EV‐free medium alone was added to the cells and they were incubated at 37°C for 1 h. 2′3′‐cGAMP (InvivoGen) (0.5 μg/well) or poly:IC (0.1 μg/well) were mixed with Lipofectamine‐2000 (Invitrogen) in a ratio of 1:1 according to manufacturer's instructions. This mix was diluted with EV‐free media and added to the cells in a volume of 100 μL. The cells were incubated at 37°C for 20–24 h if not otherwise indicated in the figures. Lipopolysaccharides (LPS) (Sigma‐Aldrich) were diluted in EV‐free media to a concentration of 0.5 μg/mL and added to the cells in a volume of 100 μL. Recombinant human (rh) Interleukin‐2 (IL‐2, Sigma‐Aldrich), rhTNFα (Peprotech) and rhIFNγ (Peprotech) were diluted in EV‐free media as indicated on the figurers, and added to the cells in a total volume of 50 μL. The cells were incubated at 37°C for 1 h and stimulated with 2′3′‐cGAMP (0.5 μg/well) as described previously. For inhibiting NF‐kB signalling, cells were treated with 10 μM ML120B (Tocris) for 30 min at 37°C. For blocking cytokine signalling, cells were incubated with 10 μg/mL human anti‐TNFR1 (clone 16805) (R&D System), 20 μg/mL human anti‐IFNGR1 (clone GIR208) (R&D System) or human IgG1 (clone MOPC‐21) (Biolegend) for 30 min at 37°C. For blocking IFNγ signalling, the T‐EVs was incubated with human anti‐IFNγ (clone MD‐1) (Biolegend) for 15 min at 37°C.

4.13. Real‐time PCR

RNA was extracted from the cells using the RNeasy mini kit (Qiagen) according to manufacturer's instructions. The concentration of the isolated RNA was determined using Nanodrop and 250ng RNA was converted into cDNA using the iScript cDNA synthesis kit (Bio‐Rad) according to the manufacturer's instruction. The cDNA was diluted by a final factor 25 for the real‐time PCR analysis. Real‐time PCR analysis of STING and ACTB was performed using the TaqMan Fast advanced Mastermix (Thermo Fisher Scientific) and the following TaqMan gene expression assays: Hs00736955_g1 (STING/TMEM173) and Hs01060665_g1 (ACTB). The analysis was done using the AriaMX realtime PCR System (Agilent Technology).

4.14. Functional type I IFN

Bioactive functional type I IFN was quantified in supernatants using the reporter cell line HEK‐Blue IFN‐α/β (Invivogen) according to manufacturer's instructions. This cell line expresses secreted alkaline phosphatase (SEAP) under control of the IFN‐α/β inducible ISG54 promoter. The cell line was maintained in DMEM supplemented with GlutaMax‐I (Gibco, Life Technologies), 10% heat‐inactivated FCS, 100 μg/mL streptomycin and 200 U/mL penicillin, 100 μg/mL normocin (InvivoGen), 30 μg/mL blasticidin (InvivoGen) and 100 μg/mL zeocin (InvivoGen). For measurement of functional type IFN, cells were seeded at 3 × 10⁴ cells/well in 96‐well plates in 150 μL medium devoid of Blasticidin and Zeocin. The following day 50 μLof the supernatant for analysis were added to the HEK‐Blue cells. SEAP activity was assessed by measuring optical density (OD) at 620 nm on a microplate reader (Elx808, BioTEK). The concentration was determined from a standard curve made with IFN‐α (IFNa2 PBL Assay Science) ranging from 2 to 500 U/mL.

4.15. Enzyme‐linked Immunosorbent assay (ELISA)

Protein levels of the cytokines CXCL‐10 and IL‐6 in the cell culture supernatants were determined using the following ELISA kits: CXCL‐10 and IL‐6 (R&D System) according to manufacturer's instructions. IFN‐beta in the cell culture supernatants was determined using the mouse IFN‐beta DuoSet ELISA kit (R&D System) according to manufacturer's instructions. IFNγ in the T‐EVs were determined using the Human IFNγ DuoSet ELISA kit (R&D System) according to manufacturer's instruction. Prior to analysis, the T‐EVs were lyzed in v/v 2.5% Triton‐X‐100, 0.05% Tween‐20.

4.16. Mesoscale V‐Plex

1 × 10⁹ T‐EVs from either activated (A) or non‐activated (NA) human CD4⁺ T cells were lyzed using v/v 2.5% Triton‐X‐100, 0.05% Tween‐20. This lysate was analysed for presence of IFNγ, IL‐1β, IL‐2, IL‐4, IL‐6, IL‐8, IL‐10, IL‐12p70, IL‐13 and TNFα using V‐plex (Meso Scale Discovery®, V‐PLEX® Proinflammatory Panel 1 (human)) according to manufacturer's instructions. Samples were diluted 1:2.

4.17. Western blotting

Protein expression was analysed by Western blotting. Cell lysis buffer was made by mixing Pierce RIPA buffer (Thermo Scientific) supplemented with 10 mM NaF, 2X complete protease cocktail inhibitor (Roche), 2X protease and phosphatase inhibitor (Pierce) in a 1:1 dilution with Laemmli Lysis Buffer (Sigma‐Aldrich). Benzonase (Sigma‐Aldrich) were added to the lysis buffer in a volume of 1 μL/mL lysis buffer. For lysis of cells, 60 μL of this lysis buffer was added to a single well in a 96‐well plate. For lysis of T‐EVs, 75 μL of lysis buffer was added to the 100K T‐EV pellet. Protein concentration in the lysates was determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) according to manufacturer's instructions. The lysate was denatured at 95°C for 5 min and separated on a Criterion Precast Gel, 4%–12% Tris‐HCl (Bio‐Rad) with XT MOPS running buffer (Bio‐Rad). The separated proteins were transferred to a Trans‐Blot Turbo 0.2 μm PVDF membrane using the Trans‐Blot Turbo Transfer System (Bio‐Rad). Membranes were washed in TBS supplemented with 0.05% Tween‐20 (TBS‐T) and blocked in 5% fresh made skim‐milk (Sigma‐Aldrich) in TBS‐T. Following antibodies were diluted in 1%–5% BSA (Roche): anti‐cGAS (Cell Signaling, D1D3G), anti‐STING (Cell Signaling, D2P2F), anti‐vinculin (Sigma Aldrich, v9131), anti‐phospho‐STING (Cell Signaling, D7C3S), anti‐phospho‐TBK1 (Cell Signaling, D52C2), anti‐TBK1 (Cell Signaling), anti‐phospho‐IRF3 (Cell Signaling, D601M), anti‐IRF3 (Cell Signaling, D83B9), anti‐phospho‐P65 (Cell Signaling, 93H1), anti‐P65 (Cell Signaling, D14E12), anti‐IkBa (Cell Signaling, L35A5), anti‐CD9 (Santa Cruz, C4), anti‐HSP70 (Cell Signaling) and anti‐Calreticulin (Cell Signaling, D3E6). The membranes were incubated with primary antibodies overnight at 4°C. The following secondary antibodies were diluted in 1% skim‐milk: peroxidase‐conjugated F(ab´)2 donkey anti‐mouse IgG (H+L), peroxidase conjugated affinipure F(ab´)2 donkey anti‐rabbit IgG (H+L), and peroxidase conjugated F(ab´)2 donkey anti‐goat IgG (H+L) (all purchased from Jackson Immuno Research). The membranes were incubated with secondary antibodies for 1 h at room temperature. The membranes were exposed using Clarity Western ECL Blotting Substrate.

4.18. Cell proliferation assay

Proliferation of MC38 cells was measured using the CellTiter‐glo 2.0 Assay (Promega) according to manufacturer's instructions. In brief, 5000 MC38 cells/well were seeded in a 96‐flat well plate the day before stimulation. The cells were treated in triplicates with T‐EVs and cGAMP as previously described. As a control, the cells were treated with 0.1μM Doxorubicin hydrochloride (Sigma‐Aldrich). Luminescent signal was determined after 48 h of stimulation.

4.19. Mice

C57BL/6 mice from Janvier were housed in the animal facility at Department of Biomedicine, Aarhus University, under conditions according to the recommendations of The Animal Experiments Inspectorate under the Ministry of Environment and Food of Denmark. The Study was conducted in accordance with The Animal Ethics Council, license number: 2017‐15‐0201‐01253. Female mice with an age of 7–9 weeks at initiation of the experiments, were used for all experiments.

4.20. Generation of tumour xenografts

MC38 (Kerafast) colon adenocarcinoma cells were cultured in DMEM (Sigma Aldrich) supplemented with 10% FCS, 2 mM Glutamine, 1000 U/mL penicillin and 1000 μg/mL streptomycin, 1X non‐essential amino acids (Gibco). The cells were cultured under humidified conditions at 37°C and 5% CO₂. The cells were split before they reached 100% confluency and were detached from the cell culture flask using 0.25 % trypsin (Gibco) in PBS with 0.02% EDTA (Invitrogen). To establish a tumour in the mice, 1 × 10⁶ MC38 cells in 50 μL were injected subcutaneously into the right flank of the mice under isoflurane anaesthesia.

4.21. In vivo treatment

The in vivo experiment was conducted at day 9–10 after tumour cell inoculation, when the tumours reached a size of 20–40 mm³. The mice were grouped randomly and treated intratumourally (IT), with 30 μL containing either different amounts of 2′3′‐cGAMP Vaccigrade (Invivogen), murine T‐EVs diluted in physiological water (Sigma‐Aldrich) or a combination. The tumour size was measured regularly using a caliper and the tumour volume was calculated using the formula: Tumour Volume (mm³) = 0.5233*L*W*H where, L = Length (mm); W = Width (mm); H = Height (mm). At the end of the experiments or when the tumour reached a size of 1000 mm³, the mice were euthanised.

4.22. Statistics

Data are in general shown as bars indicating mean +SD or +SEM, and with points indicating each replica. Statistical analysis was performed with Graphpad Prism software. Comparison between different groups were analysed by either paired Students t‐test, unpaired Students t‐test or Wilcoxon test, as indicated in the figure legends. All p‐values are two‐tailed. Significance was defined as p < 0.05.

AUTHOR CONTRIBUTIONS

Aida S. Hansen: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Writing‐review & editing. Lea S. Jensen: Formal analysis; Investigation; Methodology; Writing‐review & editing. Kristine R. Gammelgaard: Formal analysis; Investigation. Kristoffer G. Ryttersgaard: Formal analysis; Methodology. Christian Krapp: Formal analysis; Investigation; Methodology. Jesper Just: Formal analysis; Investigation. Kasper L. Jønsson: Formal analysis; Investigation. Pia B. Jensen: Formal analysis; Investigation. Thomas Boesen: Formal analysis; Methodology. Mogens Johannsen: Methodology; Resources. Anders Etzerodt: Investigation; Methodology; Project administration; Supervision. Bent W. Deleuran: Conceptualization; Funding acquisition; Project administration; Supervision; Writing‐review & editing. Martin R. Jakobsen: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Project administration; Supervision; Visualization; Writing‐review & editing.

CONFLICT OF INTEREST STATEMENT

M.R.J. is shareholder and consultant for the biotech companies Stipe Therapeutics and Unikum Therapeutics who develop novel cancer immunotherapies to treat cancer. The rest of the authors have no conflicts to declare.

Supporting information

Supplementary figure 1. T‐EVs alone do not activate STING signalling

Supplementary figure 2. CD4⁺ T cells express activation‐associated markers at time of T‐EV harvest.

Supplementary figure 3. T‐EVs have a vesicular structure.

Supplementary figure 4. Recombinant IFNγ and TNFα prime STING signalling in THP‐1 cells dose‐dependent.

Supplementary figure 5. T‐EVs are not toxic to cancer cells in vitro.

Supplementary figure 6. T‐EVs alone have no anti‐tumoural function in MC38 adenocarcinoma.

Supplementary figure 7. The anti‐tumoural function of cGAMP is dose‐dependent.

Supplementary figure 8. T‐EVs enhance antitumour efficacy of low‐dosage cGAMP.

Supplementary figure 9. T‐EVs combined with high‐dosage cGAMP induce complete tumour regression.

Hansen, A. S. , Jensen, L. S. , Gammelgaard, K. R. , Ryttersgaard, K. G. , Krapp, C. , Just, J. , Jønsson, K. L. , Jensen, P. B. , Boesen, T. , Johannsen, M. , Etzerodt, A. , Deleuran, B. W. , & Jakobsen, M. R. (2023). T‐cell derived extracellular vesicles prime macrophages for improved STING based cancer immunotherapy. Journal of Extracellular Vesicles, 12, e12350. 10.1002/jev2.12350