Aerosolized immunotherapeutic nanoparticle inhalation potentiates PD-L1 blockade for locally advanced lung cancer

Abstract

Despite therapeutic advancements, the prognosis of locally advanced non-small cell lung cancer (LANSCLC), which has invaded multiple lobes or the other lung and intrapulmonary lymph nodes, remains poor. The emergence of immunotherapy with immune checkpoint blockade (ICB) is transforming cancer treatment. However, only a fraction of lung cancer patients benefit from ICB. Significant clinical evidence suggests that the proinflammatory tumor microenvironment (TME) and programmed death-ligand 1 (PD-L1) expression correlate positively with response to the PD-1/PD-L1 blockade. We report here a liposomal nanoparticle loaded with cyclic dinucleotide and aerosolized (AeroNP-CDN) for inhalation delivery to deep-seated lung tumors and target CDN to activate stimulators of interferon (IFN) genes in macrophages and dendritic cells (DCs). Using a mouse model that recapitulates the clinical LANSCLC, we show that AeroNP-CDN efficiently mitigates the immunosuppressive TME by reprogramming tumor-associated macrophage from the M2 to M1 phenotype, activating DCs for effective tumor antigen presentation and increasing tumor-infiltrating CD8⁺ T cells for adaptive anticancer immunity. Intriguingly, activation of interferons by AeroNP-CDN also led to increased PD-L1 expression in lung tumors, which, however, set a stage for response to anti-PD-L1 treatment. Indeed, anti-PD-L1 antibody-mediated blockade of IFNs-induced immune inhibitory PD-1/PD-L1 signaling further prolonged the survival of the LANSCLC-bearing mice. Importantly, AeroNP-CDN alone or combination immunotherapy was safe without local or systemic immunotoxicity. In conclusion, this study demonstrates a potential nano-immunotherapy strategy for LANSCLC, and mechanistic insights into the evolution of adaptive immune resistance provide a rational combination immunotherapy to overcome it.

Graphical Abstract

Inhalation of aerosolized nanoparticles loaded with cyclic dinucleotide (AeroNP-CDN) synergizes with irradiation and anti-programmed death-ligand 1 (PD-L1) immunotherapy for locally advanced non-small cell lung cancer (LANSCLC).

1. Introduction

Lung cancer is the second most common cancer in the United States, and non-small cell lung cancer (NSCLC) comprises the majority of cases [1, 2]. Locally advanced NSCLC (LANSCLC) with tumor cells invading lymph nodes and other lobes or the other lung, but without distant metastasis, represents about one third of the total NSCLC at diagnosis [3]. Often, patients at this stage are inoperable, and chemo-radiotherapy is the standard of care treatment. However, the prognosis for these patients is poor, with five-year survival rate of approximately 20% [4]. Immunotherapy, in particular, immunotherapeutic antibodies (Abs) against the immune checkpoint molecules cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), programmed death 1 (PD-1) and its ligand, programmed death-ligand 1 (PD-L1), have shown impressive clinical efficacy in several solid tumors including NSCLC [5–7]. However, only a fraction of NSCLC patients benefit from the immune checkpoint blockade (ICB) therapy.

Clinical data suggest that tumor PD-L1 expression and tumor-infiltrating CD8⁺ lymphocytes (TILs) are biomarkers predictive of ICB responsiveness, and the inflamed NSCLC tumors with more TILs and the PD-L1-positive status generally show better response to anti-PD-1/PD-L1 antibodies [8]. Thus, immunomodulation of the tumor microenvironment (TME) to convert the immunosuppressive tumors into proinflammatory tumors likely benefits ICB immunotherapy in NSCLC. Much effort has been made to develop new immunomodulators as well as non-conventional pharmacological strategies in order to maximize the immune responses in tumors while reducing the off-target immunotoxicity to normal tissues. One of the potent immunostimulants, stimulators of interferon genes (STING) agonist, has recently been identified to play an important role in induction of antitumor immunity through its activation of STING signaling and type I interferons (IFNs) production [9–11]. While intratumoral injection of STING agonist, cyclic dinucleotide (CDN), has been investigated in preclinical studies, and is being evaluated in clinical trials for solid tumors, this approach is largely limited to those accessible tumors. Attempts are being made to develop nanoparticle or microparticle-loaded immunostimulants to facilitate extended in situ application [12–18] or systemic delivery to visceral tumors and metastases [19–28]. However, efficient delivery and possibly repeated applications of these immunostimulants to deep-seated tumors while avoiding systemic immunotoxicity remain technically challenging.

Previous studies have shown that activation of STING signaling and type I IFNs production within tumor–resident antigen-presenting cells (APCs) is necessary for induction of adaptive CD8⁺ T cell anticancer immunity [10, 29–31], whereas non-targeted free CDN may induce T cell apoptosis [32–34] or increase the resistance of tumor cells to ICB [35, 36]. Recently, we have developed a liposomal nanoplatform to efficiently load CDN and target CDN to macrophage and dendritic cells [37]. We have also established a nebulizer system to aerosolize the liposomal CDN (AeroNP-CDN) for inhalation delivery to deep-seated lung tumors. To further explore its utility, we established an orthotopic NSCLC mouse model that developed multifocal lung tumors in both lungs, which recapitulates the clinical LANSCLC. We tested the combination of AeroNP-CDN and irradiation (IR) in the LANSCLC mice. Although the combination treatment led to durable survival gains, a large fraction of the mice relapsed. We then examined the molecular and immunological response to the treatment and detected increased levels of PD-L1 in tumor tissues, which resulted plausibly from IFN-γ-mediated feedback inhibition [35, 38]. While the increased PD-L1 could attribute to the evolvement of T cell exhaustion and consequently treatment resistance, it may also provide an opportunity for a rational combination with anti-PD-1/PD-L1 ICB to reinvigorate the exhausted T cells. We thus investigated our hypothesis that AeroNP-CDN-induced proinflammatory TME and upregulation of PD-L1 would set the stage for response to anti-PD-L1 immunotherapy against lung tumors in the LANSCLC mouse model.

2. Results and discussion

2.1. Uptake of NP-CDN by macrophages and dendritic cells (DCs) and upregulation of type I IFN genes

The CDN-calcium phosphate (CaP)-loaded, phosphatidylserine (PS)-coated NP-CDN has an average diameter of ~ 120 nm (Figs. 1(a) and 1(b)) and a negative surface charge of −40 mV. High-performance liquid chromatography (HPLC) analysis indicated high encapsulation efficiency of cGAMP (74.6% ± 5.1%) and high payload of cGAMP (32.5 ± 1.7 μg/mg lipid). Both the size and polydispersity index (PDI) of NP-CDN showed little change in phosphate buffered saline (PBS) with 10% fetal bovine serum (FBS) over a time course of 120 h (Fig. 1(c)). NP-CDN was stable at a physiological pH (7.4), but released CDN over time at 6.5 or 5.0 (Fig. 1(d)), indicating an acidic pH-responsive dissociation of the CDN-CaP complex. In vitro incubation of NP-DiR with APCs including bone marrow-derived macrophage (BMDM) and dendritic cell (BMDC), and alveolar macrophage (AM), or NSCLC Lewis lung carcinoma cells stably transfected with luciferase (LLC-Luc) and CMT-167 cells or vascular endothelial cells, bEnd.3 clearly showed preferential uptake of the nanoparticles by APCs over the other cell types (Fig. 1(e)). It is well known that membrane-exposed PS serves as an “eat me” signal for phagocytes via their surface PS receptors [39, 40]. After ingestion by APCs, real-time quantitative polymerase chain reaction (qPCR) measurements of mRNA levels of inflammation-related cytokines/chemokines in APCs revealed that NP-CDN (100 nM CDN) induced marked upregulation of type I IFNs and other inflammatory response genes such as Il-12b and Cxcl10 but minimal change in anti-inflammatory Il-10 and Tgfb1 (Fig. 1(f)). NP-CDN also promoted activation and maturation of APCs by overexpressing surface markers of CD86 and major histocompatibility complex (MHC) class II (Fig. 1(g)). By contrast, free CDN at 100 nM led to a smaller increase in the genes of inflammatory cytokines/chemokines and expression of CD86 and MHC-II (Figs. 1(f) and 1(g)). Structurally, CDN contains two phosphodiester bonds that restrict its penetration through the plasma membrane [11], which likely accounts for its inferior activity to NP-CDN.

Figure 1.

Characterization of NP-CDN and in vitro uptake studies. (a) Size distribution assessed by dynamic light scattering (DLS). (b) A representative transmission electron microscopy (TEM) image of NP-CDN, scale bar = 100 nm. (c) and (d) Stability of NP-CDN at pH 7.4 in 10% FBS (37 °C), but pH-dependent release of CDN. (e) In vitro uptake study by a 30 min-incubation of NP-DiR with APCs including BMDM, BMDC, and AM or NSCLC LLC and CMT167 cells or normal mouse vascular endothelial cells, bEnd.3. Immunofluorescence microscopy observed intracellular localization of NP-CDN in APCs. Scale bar = 25 μm. (f) Real-time qPCR measurements of mRNA levels in BMDM, BMDC, and AM after 4 h incubation with NP-CTR (red), free CDN (green), or NP-CDN (purple; 100 nM). NP-CTR: NP-2′ 5′ -GpAp as a control of NP-CDN. (g) NP-CDN induced increased expression of surface markers, CD86 and MHC-II on APCs. n = 3 biologically independent experiments. ***p < 0.001.

2.2. AeroNP-CDN targets CDN to macrophages and DCs in individual tumors of both lungs

An orthotopic LANSCLC model that develops multiple lesions in both lungs has been established in this study, as confirmed by magnetic resonance imaging (MRI) and histology (Figs. 2(a) and 2(b)). The LANSCLC mice were placed in a chamber connected with a nebulizer system to expose to AeroNP-DiR. The nebulizer generates aerosols with a mass mean aerodynamic diameter (MMAD) of ~ 1.4 μm and geometric standard deviation (GSD) of 1.25, which falls into the size range optimal for deep lung deposition [37, 41]. Ex vivo IVIS optical imaging revealed exclusive accumulation of AeroNP-DiRs in both lungs at 24 h, which sustained at 48 h (Figs. 2(c) and 2(d)). HPLC analysis of major organs and tissues confirmed that the AeroNPs confined in lungs and tumor draining lymph nodes (DLNs) with minimal amount detected in blood or other organs (Fig. 2(e)). Immunohistochemical co-staining of Luc⁺ tumor cells detected intratumoral AeroNP-DiRs, where many of them were co-localizing with CD11c⁺ APCs (Fig. 2(f)). To determine which subsets of APCs were capturing AeroNPs, we conducted flow cytometry analysis and found more AeroNPs in AMs than DCs in tumor-bearing lung tissues (Figs. 2(g) and 2(h)). AeroNPs were also detected in DCs and CD11b⁺ macrophages at DLNs, and found more in DCs than macrophages at 48 h (Fig. 2(i)). We have previously demonstrated that we can control the dose of AeroNPs delivered to lung tissues by changing inhalation doses or durations [37]. Under the current protocol with a 28 min-inhalation of AeroNP-CDN (37 μM CDN) in 5 mL PBS, we estimated 0.1 μg CDN deposited in lungs of a mouse, which was less than one-hundredth of the dose of free CDN used for intratumoral injection in several studies [42, 43]. Taken together, our data demonstrate that AeroNP-CDN inhalation enables effective delivery of CDN to macrophages and DCs in individual lung tumors and their draining lymph nodes in both lungs.

Figure 2.

AeroNP targets macrophages and DCs in individual lung tumors. (a) and (b) An orthotopic LANSCLC mouse model bearing multiple lung tumors (arrows in an MRI and in hematoxylin and eosin (H&E) stained sections) in both lungs mimics the clinical LANSCLC. Scale bar = 200 μm. (c) and (d) Ex vivo IVIS imaging of major organs from LLC-Luc tumors-bearing mice detected DiR signals exclusively from both lungs 24 and 48 h post AeroNP-DiR. LU: Lung, HT: heart, LV: liver, SP: spleen, KD: kidney. (e) In addition to lung tissues, HPLC detected high concentrations of NPs in tumor draining lymph nodes at 24 h (red) and 48 h (green). (f) Co-staining of CD11c⁺ APCs and Luc⁺ tumor cells in tumor-bearing lung tissue sections revealed intratumoral NP-DiRs, many of which co-localized with CD11c⁺ APCs. Scale bar = 20 μm. (g) Flow cytometry gating for the subsets of APCs, CD11c⁺ DCs and CD11c⁺F4/80⁺ AMs and (h) their differential uptake of AeroNP-DiRs in tumor-bearing lung tissues. (i) AeroNP-DiRs were also detected in CD11c⁺ DCs and CD11b⁺F4/80⁺ macrophages in DLNs. n = 3 biologically independent experiments.

2.3. AeroNP-CDN remodels the immune landscape in individual lung tumors and enhances radiation therapy

Radiotherapy is a standard of care treatment for lung cancer. Previous studies by others and us have shown that STING agonist in combination with focal IR elicits robust antitumor immune responses in multiple cancer models [37, 42]. Here, we investigated the therapeutic efficacy of AeroNP-CDN in combination with IR in the LANSCLC mouse model. After imaging confirmation of lung tumor development on day 14, a region bearing lung tumors in the right lung was irradiated with 6 Gy/every other day for 3 doses, while AeroNP-CDN was inhaled (a 28 min-inhalation of AeroNP-CDN (37 μM CDN) in 5 mL PBS) a day after each IR (Fig. 3(a)). AeroNP-CDN + IR significantly prolonged survival of the mice, 40% of which survived for at least 120 days (Fig. 3(b)). By contrast, focal IR alone showed minimal survival benefit (Fig. 3(b)). While radiation is known to cause immunogenic cell death by releasing tumor antigens (TAs), it rarely induces local or systemic immune response against non-irradiated tumors [44–46].

Figure 3.

AeroNP-CDN enhances radiation treatment. (a) Treatment of LLC-bearing mice with AeroNP-CDN, focal IR (to a tumor-bearing region in the right lung) or both (n = 6–8/group) was initiated on day 14 after confirming development of lung tumors. (b) Therapeutic efficacy was evaluated by Kaplan–Meier survival assay. (c)–(e) Flow cytometry detected significantly increased numbers of tumor-infiltrating CD4⁺ and CD8⁺ T cells in the left and right lung 48 h after the first dose of AeroNP-CDN alone or combined with IR (c). Quantitative data of CD4⁺ T cells and CD8⁺ T cells were presented in (d) and (e), respectively. (f) Regulatory T cells (Treg; Foxp3⁺CD4⁺) were quantified and (g) ratios of CD8⁺ T/Treg were plotted. (h)–(k) ELISA measurements of IFN-β, IFN-γ, IL-12p40, and CXCL10 in tumor-bearing lung tissues from the left and right lung. n = 3 biologically independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001.

To interrogate the treatment-induced immune responses, we analyzed the tumor-bearing lung tissues from both lungs by flow cytometry (Fig. S1 in the Electronic Supplementary Material (ESM)). Our data showed that AeroNP-CDN alone or in combination with IR induced increased expression of CD86 and MHC-II in both AMs and CD11c⁺CD103⁺ DCs (Fig. S2 in the ESM). Notably, CD103⁺/CD8α⁺ DCs have been indicated as the most competent APCs for cross-priming CD8⁺ T cells in mice [47, 48]. AeroNP-CDN alone or the combination treatment also led to a significant increase in tumor-infiltrating CD4⁺ and CD8⁺ T cells in both lungs (Figs. 3(c)–3(e)). Whereas IR induced some increase in regulatory T cells (Treg; Foxp3⁺CD4⁺) in the irradiated right lung, AeroNP-CDN decreased the number of Treg, and consequently a significantly increased ratio of CD8⁺ T/Tregs in both lungs (Figs. 3(f) and 3(g)). A high ratio of CD8⁺ T/Tregs has been indicated in clinic to correlate with favorable response to immunotherapy in NSCLC and other cancer types [49, 50]. In line with the flow cytometry results, enzyme-linked immunosorbent assay (ELISA) measurements of cytokines/chemokines in tumor tissues showed significant increase in type I IFN, IFN-β, and other pro-inflammatory IL-12p40, IFNγ, and CXCL10 (Figs. 3(h)–3(k)). CXCL10 is an important chemokine responsible for recruitment of T cells [48], which likely contributed to the increased number of TILs in not only the irradiated but also the non-irradiated lung. Collectively, our data indicate that AeroNP-CDN promotes the proinflammatory TME, activates both the innate APCs and adaptive cytotoxic T cells, and synergizes with focal irradiation against the LANSCLC.

2.4. AeroNP-CDN upregulates PD-L1 and blockade of PD-L1 enhances antitumor immune responses

To further explore its therapeutic potential for enhanced anticancer efficacy, we looked into the immunological responses to AeroNP-CDN. AeroNP-CDN stimulates APCs to produce type I IFNs and IL-12 (Figs. 3(h) and 3(k)), which can activate T cells to secret IFN-γ and exert their antitumor cytotoxic effects [10, 35]. Indeed, our flow cytometry data showed that AeroNP-CDN significantly increased frequency of IFN-γ-secreting CD8⁺ T cells, which was further enhanced with the combination of IR (Figs. 4(a) and 4(b)). However, it has been recognized that certain proinflammatory cytokines, in particular, IFNγ, induce PD-L1 expression as a feedback inhibitory mechanism that limits local immunity [35, 51]. We thus investigated if AeroNP-CDN could induce PD-L1 expression in tumor tissues. As expected, AeroNP-CDN alone or in combination with IR induced a drastic increase in tumor cells expressing PD-L1 in both lungs, compared to the control PBS treatment (Figs. 4(c) and 4(d)). Besides tumor cells, there was significantly increased PD-L1 expression on CD11c⁺F4/80⁺ AM (Figs. 4(e) and 4(f)). PD-L1 on stromal cells such as macrophages also contributes to T cell exhaustion via the PD-1/PD-L1 axis [52].

Figure 4.

AeroNP-CDN activates effector CD8⁺ T cells while upregulating PD-L1. (a) and (b) AeroNP-CDN alone or in combination with focal IR activated CD8⁺ T cells to secret IFN-γ. (c)–(f) Upregulation of PD-L1 was detected on the surface of tumor cells ((c) and (d)) and AMs ((e) and (f)) in both lungs. n = 3 biologically independent experiments. **p < 0.01; ***p < 0.001.

Increased expression of PD-L1 in tumor tissues could attribute to evolvement of T cell exhaustion and a sequel to treatment resistance, as observed in Fig. (3). However, we reasoned that a combination with PD-L1 blockade would augment antitumor immunity of AeroNP-CDN by counteracting the overexpressed PD-L1. Indeed, therapeutic efficacy of anti-PD-L1 Abs combined with AeroNP-CDN was found comparable to that of IR plus AeroNP-CDN, and the triple combination treatment achieved long-term survival in 80% of the mice in the wild type but not STING−/− mice bearing LLC-Luc lung cancer (Figs. 5(a)–5(e)). Similar to IR, PD-L1 blockade alone showed a little survival benefit. To interrogate if the combination treatment could induce systemic antitumor immune responses, we examined both DLNs and spleens of the treated mice and found that there was indeed a significant increase in IFN-γ⁺ CD8⁺ T cells in both tissues in response to AeroNP-CDN in combination with anti-PD-L1 Ab and/or IR (Figs. 5(f)–5(h)). We also observed marked increase in the number of memory CD8⁺ T cells including both CD44⁺CD62L⁺ central memory and CD44⁺CD62L− effector memory T cells in spleens of the mice treated with the combination immunotherapy with/without IR (Fig. 5(i)). Moreover, the long-term surviving mice from the triple combination treatment group resisted secondary tumor challenge (Fig. 5(j)), further supporting that the treatment triggers systemic antitumor memory. Together, combination immunotherapy with blockade of PD-L1, which was upregulated by AeroNP-CDN, represents a rational strategy.

Figure 5.

Anti-PD-L1 Ab enhances antitumor activity of AeroNP-CDN. (a) LLC-Luc-bearing mice were treated with IR (morning) and AeroNP-CDN (afternoon), and followed by anti-PD-L1 Ab on next day. (b)–(d) Treatment efficacy was monitored by IVIS imaging ((b) and (c)) and Kaplan–Meier survival assay (d). (e) Kaplan–Meier survival curves of indicated treatment in wild type mice or STING knockout mice. n = 8/treatment. (f)–(h) Spleens and DLNs from the mice with indicated treatment were analyzed by flow cytometry for IFN-γ⁺ activated CD8⁺ effector T cells. (i) Memory CD8⁺ T cells of both CD44⁺CD62L⁺ central memory T cells (Tcm) and CD44⁺CD62L− effector memory T cells (Tem) detected in spleens on day 26. (j) Mice treated with AeroNP-CDN ⁺ IR or AeroNP-CDN + IR + anti-PD-L1 Ab were re-challenged 80 days later with LLC-Luc cells. Naive mice challenged at the same time served as positive controls. Data showed the percent of mice rejecting LLC-Luc tumor re-challenge in each group. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

2.5. AeroNP-CDN in combination with ICB is safe

We have examined the safety of inhalation of AeroNP-CDN. Our initial studies showed that AeroNP-CDN was well-tolerated by unanesthetized mice, and the same inhalation dose as used in this study with/without focal irradiation did not cause any short or long-term adverse events in healthy or lung metastasis mice [37]. Immunotherapy with anti-PD-1/PD-L1 and anti-CTLA4 Abs is frequently associated with immunopathology. We thus investigated if the combination immunotherapy of AeroNP-CDN with anti-PD-L1 Ab may induce immunotoxicity locally to lungs or systemically to distant organs. Histopathological examinations of tissues from lung, kidney, heart, and liver revealed no morphological difference between the control PBS, AeroNP-CDN alone, and the combination treatment (Fig. 6(a)). There was no significant change in liver enzymes aspartate aminotransferase (AST) and alanine aminotransferase (ALT) or serum levels of various cytokines (Figs. 6(b) and 6(c) and Fig. S3 in the ESM). Liposome is biocompatible and widely used as a drug carrier [53–55]. Altogether, the low dose of AeroNP-CDN, the inhalation approach, and APC-targeted delivery led to negligible amounts of CDN entering other types of lung cells or the blood circulation (Fig. 2(c)–2(g)), which contributed to its excellent safety.

Figure 6.

Immunotherapy of AeroNP-CDN with anti-PD-L1 Ab is safe. (a) H&E staining of major organs from healthy mice with indicated treatment on day 26 showed no morphological change in these organs between the PBS control and AeroNP-CDN alone or in combination with anti-PD-L1 Ab. (b) and (c) Serum levels of AST and ALT were measured in the lung cancer mice with indicated treatment (n = 6).

3. Conclusion

In summary, we have demonstrated an inhalable nanoparticle system that enables delivery of immunotherapeutics to deep-seated lung tumors for targeted activation of APCs-mediated antitumor immune responses. We show that AeroNP-CDN transforms the immune landscape in lung tumors and their draining lymph nodes in both lungs in a mouse model mimicking the locally advanced NSCLC. Therapeutically, AeroNP-CDN synergizes with focal irradiation to engage both the innate and adaptive antitumor immunity. Furthermore, combination immunotherapy with anti-PD-L1 Abs, which counteract IFN-γ signaling-induced overexpression of PD-L1 in tumor tissues, elicits durable systemic antitumor immune responses, conferring long-term survival of the LANSCLC mice. Because it is non-invasive, accessible to multiple lung lesions/lobes at the same time and feasible for repeated uses without immunotoxicity, this nanoparticle system may have the translational potential for treating lung cancer.

4. Methods

4.1. Materials

1,2-Distearoyl-sn-glycero-3-phosphocholine (DSPC, 18:0 PC), cholesterol, brain PS (L-α-phosphatidylserine), and Rhod-b (18:1 Liss Rhod PE, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl)) were purchased from Avanti Polar Lipids. CDN (2′3′-cGAMP, cyclic [G(2′,5′)pA(3′,5′)p]) and its control 2′ 5′-GpAp were purchased from InvivoGen. Anti-PD-L1 (10F.9G2) was purchased from BioXcell.

4.2. Preparation and characterization of AeroNP-CDN

The liposomal CDN was prepared in two steps using the water-in-oil reverse microemulsion method, as detailed in our previous study [37]. In brief, 150 μL CaCl₂ (2.5 M in ddH₂O, Ph = 7.0) was added to 5 mL mixed oil phase 1 (cyclohexane:igepal co-520 = 80:20) and stirred to form a well-dispersed micro-emulsion phase. Oil phase 2 was prepared by adding 50 μL lipid mixture (20 mM, PS:DSPC:cholesterol = 5:4:1) to 5 mL mixed oil phase (cyclohexane:igepal co-520 = 80:20). A similar micro-emulsion containing sodium phosphate was prepared by adding 150 μL Na₂HPO₄ (25 mM in ddH₂O, pH = 9.0) to 5 mL mixed oil phase 2. The CaP core with the single lipid coating was formed by mixing oil phase 1 and oil phase 2. After washing with ethanol three times, the collected CaP cores were dispersed in 1 mL chloroform and centrifuged at 2,000g for 5 min. The supernatants were further mixed with 70 μL of the lipid mixture, followed by chloroform evaporation under reduced pressure to form a lipid film. NP-CaP nanoparticles with bilayer lipid coating were formed by adding 1 mL PBS and rehydrating under water bath sonication for 5 min and sonic probe at 20 W for 2 min at 70 °C. The resulting nanoparticles were further filtered with 0.45 μm membrane to remove the free lipid aggregates and stored at 4 °C. To load CDN, half of the desired content CDN was mixed with each of the CaCl₂ and Na₂HPO₄ solutions. For fluorescently labeling NPs, 18:1 Liss Rhod PE (Rhod-b) was added to the second lipid mixture with a molar ratio of 1%. DiR was further used for labeling the NP (NP-DiR) by adding DiR directly to the second lipid mixture at a molar ratio of 5%. The size, size distribution, and zeta potential of AeroNP-CDN in aqueous solution were measured by a Malvern Zetasizer Nano ZS90. TEM measurements were performed on an FEI Tecnai Bio Twin transmission electron microscope. To determine the CDN loading efficiency and release kinetics, HPLC analyses were conducted using an Agilent 1100 HPLC system.

4.3. Cells

NSCLC cells including luciferase-expressing Lewis lung carcinoma cells (LLC-Luc, provided by Dr. Yong Lu, Wake Forest) and CMT-167 (ATCC), and mouse vascular endothelial cells, bEnd.3 (ATCC) were maintained in Dulbecco’s modified Eagle medium (DMEM), supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. BMDMs were generated and cultured in complete α-MEM containing 10% heat-inactivated FBS, penicillin/streptomycin, and 50 ng/mL macrophage colony-stimulating factor (M-CSF). BMDCs were induced and cultured in RPMI-1640 containing 10% heat-inactivated FBS, penicillin/streptomycin, 20 ng/mL granulocyte-macrophage colony-stimulating factor (GM-CSF), 5 ng/mL interleukin 4 (IL-4), and 1× 2-mercaptoethanol. AMs from bronchoalveolar lavage (BAL) were cultured in DMEM with 10% heat-inactivated FBS, 2-mercaptoethanol (1×), and penicillin/streptomycin. All the cells were maintained in a humidified atmosphere containing 5% CO₂ at 37 °C.

4.4. Cellular uptake of PS-coated nanoparticles in vitro

Briefly, 6,000 BMDM, BMDC, AM, LLC-Luc, CMT-167, or bEnd.3 cells were seeded in 24 well plates with poly-lysine coated coverslips for 24 h before adding NP-DiR. After 0.5 h, cells were then washed twice with cold PBS and fixed with 4% formaldehyde for 15 min. After 1% Triton X-100 treatment for another 15 min, the cells were co-stained with Alexa Fluor 488 phalloidin (Thermo Fisher) for 30 min and then with 4’,6-diamidino-2-phenylindole (DAPI) for 5 min. The cellular uptake of NP-DiR was observed by fluorescence microscope.

4.5. Real-time qPCR of type I IFN and other inflammatory genes

BMDM, BMDC, or AM cells (2 × 105) were seeded into 35-mm culture dishes and cultured for 24 h. The cells were then incubated at 37 °C with free CDN (100 nM) or AeroNP-CDN (100 nM CDN) suspended in complete culture medium for 4 h. All cellular RNA was collected using TRIzol reagent (Thermo Fisher), of which 1 μg of total RNA was transcribed into cDNA using High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher) and real-time qPCR was performed using PowerUp SYBR Green Master Mix (Thermo Fisher) with primers listed (Table S1 in the ESM). mRNA levels were normalized against the housekeeping gene GAPDH.

4.6. Flow cytometry

Flow cytometry was performed on a BD Canto II flow cytometer and analyzed using FlowJo software (BD Biosciences). CD3-FITC (17A2), CD4-PE (GK1.5), CD8α-PerCP-Cyanine5.5 (53–6.7), CD45-APC-Cyanine7 (30-F11), CD45-Pacific Blue (30-F11), CD11b-FITC (M1/70), CD11b-PE-Cyanine7 (M1/70), F4/80-PE (BM8), CD11c-FITC (N418), PD-L1-PerCP-Cyanine5.5 (10F.9G2), PD-L1-PE (10F.9G2), PD-L1-APC (10F.9G2), CD86-APC-Cyanine7 (GL-1), MHC-II-APC (M5/114.15.2), CD103-Pacificl Blue (2E7), CD44-APC (IM7), CD62L-PE-Cyanine7 (MEL-14), IFN-γ-Brilliant Violet 510 (XMG1.2), and CD16/CD32 (93) were purchased from Biolegend (San Diego, CA). FOXP-3-Alexa Fluor 647(MF-14) and intracellular staining kit for Foxp-3 were from eBioscience (Thermo Fisher Scientific, Waltham, MA). For intracellular staining of IFN-γ, fresh isolated cells were treated with phorbol 12-myristate 13-acetate/ionomycin cocktail according to the manufacturer’s specification (BioLegend). Doublets and debris of dead cells were excluded before various gating strategies were applied. Gates and quadrants were set based on isotype control staining, and the mean fluorescence intensity (MFI) values were calculated by minus the MFI of isotype control antibodies.

4.7. ELISA assay

Cytokines were analyzed with ELISA Kit from BioLegend by following the manufacturer’s instructions. After adding the horseradish peroxidase (HRP) substrates, optical densities were determined at a wavelength of 450 nm in an ELISA plate reader (Bio-Rad).

4.8. Liver enzyme assay

Serum aspartate transaminase activity (Enzychrom Aspartate Transaminase Assay kit, BioAssay Systems, Haymard, CA) and alanine transaminase activity (Enzychrom Alanine Transaminase Assay kit, BioAssay Systems) were performed following manufacturer’s instructions.

4.9. Establishment of an orthotopic NSCLC mouse model

All animal experiments complied with all relevant ethical regulations for animal testing and research and were performed with approved of the Institutional Animal Care and Use Committee at the Wake Forest University School of Medicine. Anesthetized C57BL/6 mice (6–8 weeks, female:male at 1:1; Charles River Laboratories, Wilmington, MA) or STING−/− mice (The Jackson Laboratory, Bar Harbor, MA) were instilled intratracheally with LLC-Luc cells (2 × 106) in 50 μL PBS buffer with 0.1 M ethylene diamine tetraacetic acid (EDTA) [56]. The Intubation Illumination System (Braintree Scientific, Braintree, MA) with fiber-optic light was used for guiding an intratracheal catheter insertion. By day 14, multifocal primary lung carcinomas in both sides of lungs were confirmed by MRI and bioluminescence imaging (BLI). Development of LLC-Luc lung cancer was monitored by longitudinal BLI. In vivo studies of pharmacokinetics, immunological effects and treatment were subsequently conducted on the mice, as described below in details.

4.10. Aerosolized PS-coated nanoparticles inhalation

Aerosolized nanoparticles were generated by a custom-built nebulizing system. A clear plastic box with a wire-netting floor was used for inhalation treatment. Aerosol was generated via a medical-grade nebulizer (MEDNEB Compressor Nebulizer, REF MQ5600, Drive DeVilbiss Healthcare, NY) attached by medical tubing to the animal chamber. Aerosol was discharged from an opening in the opposite end of the cage. For in vivo biodistribution study, DiR- or Rhod-b-labeled PS-coated NP-CaP (DiR 700 μg; Rhod-b 100 μg) in 5 mL PBS was loaded into the nebulizer. For in vivo treatment in LLC lung cancer mouse model, AeroNP-CDN (37 μM CDN) in 5 mL PBS was nebulized and delivered to unanesthetized mice via aerosol inhalation.

4.11. In vivo biodistribution and quantification

C57BL/6 mice bearing multifocal LLC-Luc lung tumors were placed in the chamber to inhale NP-DiR for 28 min. The animals were sacrificed at 24 and 48 h (n = 3/time) and major organs were dissected and ex vivo imaging was conducted using an IVIS^® Lumina system (Caliper Life Sciences). The signal intensity was analyzed using Living Image^® 3.1 software. Immediately after ex vivo imaging, the tumors-bearing lung tissues were preserved and sectioned for immunofluorescence microscopy. Cryosections (6 μm thick) were co-stained with anti-mouse CD11c-FITC (1:200; BioLegend, N418) and anti-luciferase (1:500; Sigma-Aldrich, L0159) followed by cy3-anti-rabbit secondary antibody (1:800; Jackson Immuno) and observed using a fluorescence microscope. DiR signals were recorded and merged with the CD11c image and the luciferase-stained image of the same field. DiR⁺ cells in lung tissues and DLNs were also assessed by fluorescence-activated cell sorting (FACS). To quantify tissue concentrations of PS-coated nanoparticles, the LLC-Luc lung cancer mice were sacrificed at 24 and 48 h after inhaling NP-Rhod-b (n = 3/time). Major organs and blood were collected for HPLC analyses.

4.12. Focused irradiation

Animals were anesthetized with inhaled 3% isoflurane and positioned with the lung 50 cm below the aperture of an X-RAD 320 orthovoltage irradiator (Precision X-ray). Fractionated irradiation (6 Gy × 3 daily fractions) was delivered at a rate of 176 cGy/min (300 kVp voltage and 10 mA current) through a custom fabricated lipowitz alloy shield. The semi-circular aperture in the lipowitz alloy shield with a radius of 5 mm was placed~ 3.5 cm above the animal. This shield yields a semi-circular field of approximately 1.08 cm in diameter, allowing radiation beam to a specified area of the right lung with tumors while minimizing the radiation dose to important mediastinum and normal tissues. To ensure good reproducibility, the light-field was marked on the animal for the second and third irradiation.

4.13. MRI

MRI was conducted blinded on a 7T Bruker BioSpec small animal scanner (Bruker Biospin, Rheinstetten, Germany). Mice were anesthetized with isoflurane (3% induction, 1.5% maintenance). Respiration was monitored with a respiratory bulb under the chest and a SHARPII animal monitoring system was used for respiratory gating. Anatomical T2-weighted imaging was conducted using a RARE sequence with repetition time/echo time (TR/TE) = 1,600/23 ms, echo train length (ETL): 8, number of signal averages/acquisitions (NSA): 8, matrix size: 128 × 128, and scan time: 3 min 53 s.

4.14. In vivo experimental treatment

For the LLC-Luc-bearing mice, multifocal lung lesions in both lungs were visualized by BLI and MRI before treatment on day 14. The mice (n = 8/group) were then randomly grouped and treated as follows: (i) PBS (inhalation), (ii) inhalation of AeroNP-CDN, (iii) anti-PD-L1 (10F.9G2; 100 μg/mouse × 3, i.p.), (iv) IR (6 Gy × 3, to the tumor-bearing right lung), (v) IR (6 Gy × 3) + AeroNP-CDN, (vi) AeroNP-CDN + anti-PD-L1, and (vii) IR + AeroNP-CDN + anti-PD-L1. Orthotopic lung cancer burden was monitored longitudinally by BLI. Data were quantified with the Living Imaging software by using absolute photon counts (photons/s/cm2/Sr) in a region of interest (ROI), manually drawn to outline the BLI signal of the chest. Survival of the mice was followed for up to 120 days. For tumor re-challenge study, surviving mice from the combination treatment group (80 days) were injected subcutaneously with LLC-Luc cells (2 × 106). Additional group of mice without previous treatment were also injected and observed for tumor growth.

4.15. Histology

H&E staining was performed on cryosections (10 μm) of different tissues including normal heart, lung, liver, and spleen, as well as lung cancer bearing lungs.

4.16. Statistical analysis

Statistical analysis was performed using Microsoft Excel and Prism 7.0 (GraphPad). No priori sample size calculation was performed. No inclusion and exclusion of data were conducted. Treatment groups were not blinding to investigators. Confounders were not controlled during animal study. Data were presented as mean ± SD. Statistical significance was determined by Student’s t-test. All t-tests were one-tailed and unpaired and were considered statistically significant if p < 0.05. The survival assay was analyzed using a log-rank test and considered statistically significant if p < 0.05.