Research ArticleCancer

Autologous tumor cell–derived microparticle-based targeted chemotherapy in lung cancer patients with malignant pleural effusion

See allHide authors and affiliations

Science Translational Medicine  09 Jan 2019:
Vol. 11, Issue 474, eaat5690
DOI: 10.1126/scitranslmed.aat5690

Insider attack on tumors

A major factor that limits the effectiveness of cancer therapy is the ability to get the treatment to tumor cells. A variety of microparticles and nanoparticles have been developed in the past, and each of them has its own advantages and limitations. Guo et al. have now developed a method of creating microparticles from the patients’ own cancer cells, resulting in improved intratumoral delivery. The authors embedded the chemotherapy drug methotrexate inside the microparticles and then successfully tested this approach in both mouse models and patients with pleural effusions caused by lung cancer.


Cell membrane–derived microparticles (MPs), the critical mediators of intercellular communication, have gained much interest for use as natural drug delivery systems. Here, we examined the therapeutic potential of tumor cell–derived MPs (TMPs) in the context of malignant pleural effusion (MPE). TMPs packaging the chemotherapeutic drug methotrexate (TMPs-MTX) markedly restricted MPE growth and provided a survival benefit in MPE models induced by murine Lewis lung carcinoma and colon adenocarcinoma cells. On the basis of the potential benefit and minimal toxicity of TMPs-MTX, we conducted a human study of intrapleural delivery of a single dose of autologous TMPs packaging methotrexate (ATMPs-MTX) to assess their safety, immunogenicity, and clinical activity. We report our findings on 11 advanced lung cancer patients with MPE. We found that manufacturing and infusing ATMPs-MTX were feasible and safe, without evidence of toxic effects of grade 3 or higher. Evaluation of the tumor microenvironment in MPE demonstrated notable reductions in tumor cells and CD163+ macrophages in MPE after ATMP-MTX infusion, which then translated into objective clinical responses. Moreover, ATMP-MTX treatment stimulated CD4+ T cells to release IL-2 and CD8+ cells to release IFN-γ. Our initial experience with ATMPs-MTX in advanced lung cancer with MPE suggests that ATMPs targeting malignant cells and the immunosuppressive microenvironment may be a promising therapeutic platform for treating malignancies.


Efficient drug delivery to target cells and tissues is critical for improving the efficacy of cancer therapy (1). Conventional artificial synthetic drug delivery systems (DDSs), such as polyethylene glycol (PEG) and liposomes, are useful in drug delivery. Nevertheless, dangerous clinical profiles, potential immunological responses, and off-target effects during systemic circulation limited their application (2, 3).

Cellular microparticles (MPs) are a type of extracellular vesicles with a diameter of 100 to 1000 nm, which are formed directly through the shedding of the cellular membrane in response to various physiological and artificial stimuli. The released MPs naturally function as intercellular messengers, carrying bioactive molecules such as nucleic acids (DNA, mRNA, microRNA, and other noncoding RNAs), proteins, and lipids from donor cells to recipient target cells to regulate many disease processes including cancer (46). Several studies indicate that MPs may have intrinsic tissue tropism according to their characteristics and origins, which can be used to target disease organs (79). In addition, their natural origin makes them immune-privileged, which enables decreased drug clearance and immune response compared to synthetic PEGylated nanomaterials (10, 11). Furthermore, MPs can be patient-derived, if required, and have an innate ability to cross major biological barriers including the blood-brain barrier (12). Thus, their specific tissue tropism, theoretical biocompatibility, reduced clearance, increased penetration of biological barriers, and concurrently increased drug transport to target tissues make MPs an attractive candidate DDS (13, 14).

Recently, MPs derived from various cell origins, including human monocytic leukemia cell line THP-1 cells, human umbilical vein endothelial cells, and antigen-presenting cells such as bone marrow–derived dendritic cells (BMDCs), have been exploited for small interfering RNA and chemotherapeutic drug delivery in cancer treatment and shown to be efficacious in terms of drug uptake and targeting (1517). Compared with MPs derived from noncancerous cells, MPs budding from tumor cells allow for a cancer cell–specific targeting by taking advantage of the inherent homotypic adhesion properties mediated by surface antigens of the source cell membrane, which can be used for drug delivery applications (1820). Furthermore, we have demonstrated that tumor cell–derived MPs (TMPs), carrying tumor antigen repertoires, costimulatory molecules, and DNA fragments similar to their parental cells, can elicit a potent T cell–dependent antitumor immune response and achieve therapeutic effects in mouse models of melanoma, hepatocellular carcinoma, and colon carcinoma (21). Thus, the targeting and immunostimulatory potential of TMPs resulted in the development and application of dual chemo-immunotherapeutic approaches by incorporating methotrexate (MTX) into TMPs to yield dual-functional MPs (TMPs-MTX).

Lung cancer is the leading cause of cancer-related deaths worldwide, with 5-year survival rates varying from 4 to 17% depending on stage and regional differences (22). Non–small cell lung cancer (NSCLC) accounts for >85% of all lung cancer cases (23), and about 40% of end-stage NSCLC is associated with pleural metastasis, resulting in a paramalignant effusion in the pleural cavity, referred to as malignant pleural effusion (MPE) (24). Patients with MPE from lung tumors have shorter life spans than patients with MPE caused by non-lung primary tumors (25, 26). Current palliative treatments including pleurodesis and indwelling pleural catheters for managing MPE are suboptimal in terms of efficacy and safety, and etiologic management approaches are urgently being sought (27, 28). Thus, refractory MPE secondary to advanced lung cancer, for which there is no effective management targeting tumor-promoting mechanisms, provides an attractive clinical setting to investigate the chemo-immunotherapeutic approach. Autologous malignant cells in MPE have been chosen as donor cells for the preparation of autologous TMPs loaded with the chemotherapeutic drug MTX (ATMPs-MTX) to address issues related to immune compatibility and tumor heterogeneity.

Here, we used TMPs for the delivery of the chemotherapeutic drug MTX and investigated the targeted cytotoxic effect of TMPs-MTX in immunocompetent mouse models of lung [Lewis lung carcinoma (LLC)] and colon (MC38) adenocarcinoma–induced MPE and in patients with MPE caused by primary lung cancer.


Preparation and characterization of TMPs and TMPs-MTX

In this study, we selected murine LLC (lung), murine MC38 (colon), murine B16-F10 (skin), human A549 (lung), human MCF-7 (breast), and other cancer cells as the donor cells for TMP production because primary tumors developed by these cell lines are frequently accompanied by MPE in their advanced stages. MTX was incorporated into TMPs by incubation with tumor cells at room temperature under a period of ultraviolet (UVB, 300 J m−2) irradiation. TMPs were characterized on the basis of morphology, size, and protein content (Fig. 1, A to D). TMPs had an irregular spherical or cup-shaped morphology, as shown by transmission electron microscopy (TEM) (Fig. 1A). Nanoparticle tracking analysis revealed that the isolated naive empty TMPs used as a control had diameters ranging from 10 to 730 nm, with a mean diameter of 166 nm and a peak (mode) diameter of ~93 nm, whereas TMPs-MTX ranged from 30 to 930 nm, with a mean diameter of 264 nm and a peak (mode) diameter of 196 nm (Fig. 1B and fig. S1A). A similar size change could also be observed for TMPs loaded with other medications (fig. S1B). However, TMP production remained similar in all groups except TMPs induced by doxorubicin at a concentration of 100 μg/ml (fig. S1C). Analysis by flow cytometry revealed the presence of annexin V, a commonly used MP marker, as well as two cell surface proteins expressed on A549 cell membranes, epithelial cellular adhesion molecule (EpCAM) and CD147 (Fig. 1C). Other extracellular vesicle–associated proteins, such as CD9, CD63, and tumor susceptibility gene 101 protein (TSG101), were also examined (Fig. 1D). The gating strategy for locating and counting TMPs for the flow cytometry–based method used to analyze the TMPs is shown in fig. S1D. On the basis of this counting strategy, 2 × 107 LLC cells incubated with MTX (100 μg/ml) under UVB irradiation produced ~2 × 106 TMPs-MTX. In addition, the amount of MTX loaded in the above-obtained LLC cell-derived TMPs-MTX analyzed by high-performance liquid chromatography (HPLC) was ~1 μg, which gradually increased with the initial drug concentration in the culture medium used for donor cell incubation (Fig. 1E). This very low dose of MTX was also loaded in A549 cell–derived TMPs (fig. S1E). We determined the stability of MTX under UVB irradiation in our setting. HPLC results indicated that both free MTX and MTX loaded in TMPs after 72 hours of cold storage did not show noticeable degradation when exposed to UV irradiation [UVB (300 J m−2), 60 min], suggesting that the performance of TMPs-MTX would not be affected by MTX degradation induced by UV irradiation process during this period (fig. S1, F and G).

Fig. 1 Characterization of TMPs and TMPs-MTX.

(A) TEM of (left) multiple TMPs, (middle) a single TMP of rounded shape, and (right) a single TMP of cup shape. Scale bar, 200 nm. (B) Representative size and particle distribution plots of A549-derived TMPs-MTX. (C) Flow cytometry analysis of A549-derived TMPs labeled with annexin V–fluorescein isothiocyanate (FITC), EpCAM–Brilliant Violet 421, and CD147–Alexa Fluor 647 (negative control shown as pale gray histograms). (D) Western blot of EpCAM, CD63, CD9, and TSG101 expression in A549 whole-cell lysates (positive control) and TMP pellets. (E) Standard curve of MTX (left) and concentration of MTX loaded in LLC-derived MPs. LLC cells (4 × 107) were exposed to a concentration gradient of MTX (0, 1, 10, 100, and 500 μg/ml) under UV irradiation for the preparation of TMPs (L0, L1, L10, L100, and L500, respectively). The concentration of MTX loaded in LLC-derived MPs suspended in 500 μl of filtered PBS was measured by HPLC. (F to I) C57BL/6 mice received intrapleural injection of LLC cells (day 0), followed by intrapleural treatment 5 days later with phosphate-buffered saline (PBS), TMPs induced by UV irradiation (TMP0), and TMPs secreted by tumor cells naturally without UV exposure (TMPauto) every other day for 5 days. Mice were euthanized at day 14 (n = 8 per group). (F) Representative photographs of pleural tumors on the parietal and visceral pleura. (G) Pleural tumor number. (H) MPE volume. (I) Life-span analysis in the three groups (n = 7 per group). *P < 0.05 and ***P < 0.0001.

The contribution of TMPs to cancer development has been widely documented, indicative of the cancer-promoting potential of TMPs (29, 30). Thus, we also explored the pro-tumor potential of TMPs in vitro and in MPE models. The results revealed that UV-induced TMPs had no tumor-promoting effects in vitro or in vivo, whereas the naturally secreted TMPs enhanced tumor cells’ viability and invasion in vitro and fostered MPE progression in vivo (Fig. 1, F to I, and fig. S2), indicating that inducing strategies may affect TMP potency and heterogeneity. Therefore, UVB-induced MPs were used in subsequent MPE models in the therapeutic context.

TMP-MTX uptake efficiency and cytotoxic effect in vitro

Next, we studied the ability of TMPs to deliver the drug payload into target cells with quantum dot (QD) 605–labeled TMPs in A549 cells. Confocal images revealed accumulation of TMPs in A549 cells (Fig. 2, A and B). Furthermore, flow cytometry analysis was performed on the same cells incubated with membrane-bound fluorescent dye PKH26-labeled TMPs-MTX for 1, 6, 12, and 24 hours. The results showed a time-dependent uptake of TMPs-MTX by A549 cells (Fig. 2C). The signal intensity plot revealed that TMP-MTX uptake occurred early, and the fluorescence signal was already detected at the 1-hour time point. Efficient internalization began at 6 hours after incubation, and the fluorescence signal increased in intensity with increasing time. At the 24-hour time point, fluorescence intensity indicated the presence of TMPs-MTX in about 87% of A549 cells. Similar uptake behavior was observed in LLC cells (fig. S3A). The specific time curve of TMP-MTX uptake suggested that at least a 24-hour incubation of target cells would be preferred for subsequent cytotoxicity measurement.

Fig. 2 Accumulation and cytotoxicity effect of TMPs-MTX in tumor cells and immune cells in vitro.

(A and B) Representative confocal microscopy images of A549 cells incubated with A549-derived TMPs for 24 hours. QD-labeled TMPs (red) were added into the culture medium of parental cells overnight at 37°C. A549 cells were stained by dye carboxyfluorescein diacetate succinimidyl ester (CFSE) (green), and the nucleus was counterstained with dye 4′,6-diamidino-2-phenylindole (DAPI) (blue). (A) Most A549 cells showed an accumulation of TMPs. Scale bars, 100 μm. (B) Zoomed image of a single A549 cell showed the TMPs’ internalization in the cytoplasm. Scale bars, 10 μm. (C) Flow cytometric analysis of TMPs’ internalization into A549 cells at multiple time points (n = 3). (D) Detection of apoptosis in LLC cells treated with TMPs-MTX for 48 hours. TMPs-MTX (1 × 106) packaging MTX at a total dose of ~1 μg were used for the cytotoxicity assay, compared to the naive empty TMPs group (TMP0) and to the free MTX group at concentrations of 10 μg/ml (MTX10) and 100 μg/ml (MTX100), with PBS as control. Treated cancer cells were stained with an annexin V–FITC apoptosis detection kit, and apoptotic cells were defined as annexin V–positive cells as measured by flow cytometry (n = 3). (E and F) Detection of apoptosis in MPE-derived malignant cells (E) and other target cancer cell lines such as A549/MCF-7/B16-F10 cells (F) treated with TMPs-MTX for 48 hours (n = 3). (G) Representative confocal microscopy images of CFSE-labeled RAW264.7 macrophages (green) incubated with red fluorescent TMPs overnight at 37°C (scale bars, 10 μm) and the incidence of RAW264.7 apoptosis were analyzed by flow cytometry after 48 hours of incubation (n = 3). *P < 0.05, **P < 0.01, and ***P < 0.001.

To verify whether these TMPs-MTX are cytotoxic to tumor cells, we added TMPs-MTX with a concentration gradient of MTX into tumor cell culture medium with free MTX as a control. Apoptosis in LLC cells was analyzed at different time points by flow cytometry. TMPs-MTX exerted their cytotoxicity in a concentration- and time-dependent manner, as shown in fig. S3 (B and C). At the 48-hour time point, ~2 × 106 TMPs-MTX prepared by the incubation of LLC cells with MTX (100 μg/ml) (packaging MTX at a total dose of ~1 μg determined by HPLC) resulted in higher apoptosis rates than a dose of free MTX (10 μg/ml) (Fig. 2D). However, no notable difference was observed between the TMP-MTX group and the group receiving free MTX (100 μg/ml), which may be attributed to the high MTX sensitivity of LLC cells. Next, the cytotoxic effect of TMPs-MTX was tested on malignant cells isolated from three patients with MPE caused by primary lung adenocarcinoma, and the results showed that TMPs-MTX (packaging MTX at a total dose of ~2 μg) induced higher rates of death than 100 μg of free MTX (Fig. 2E). Similar cytotoxic effects were observed in other tumor cells including A549 (lung), MCF-7 (breast), and B16-F10 (skin) (Fig. 2F). Further, the interactions of TMPs-MTX with immune cells including macrophages and T cells were explored in vitro. The results revealed that TMPs-MTX can be efficiently taken up by RAW264.7 cells, thereby inducing apoptosis (Fig. 2G), but only a handful of TMPs-MTX can be taken up by T cells isolated from the peripheral blood of healthy donors (fig. S3D). Together, these data indicated that MTX-packaging MPs can exert direct cytotoxic effects on tumor cells and macrophages but not on T cells in vitro.

We next investigated the effects of TMP-MTX–induced apoptotic tumor cells on murine BMDCs. Exposure to TMP-MTX–induced apoptotic LLC cells increased the percentage of DCs expressing costimulatory/activation markers CD80 (B7-1), CD86, and major histocompatibility complex II (fig. S3E). Our findings suggest that in addition to the reported TMP vector itself (21), apoptotic tumor cells treated with TMPs-MTX can activate DCs.

Targeting and internalization of fluorescent TMPs in vivo

The biodistribution of TMPs was first studied in healthy mice by labeling TMPs with near-infrared fluorescence dye 1, 1ʹ-dioctadecyl-3, 3, 3ʹ, 3ʹ-tetramethylindotricarbocyanine iodide (DiR). In the group that received TMPs intravenously, DiR fluorescence was mainly distributed in the abdomen (liver and spleen), whereas in the intrapleurally injected group, fluorescence mainly accumulated in the chest and abdomen (lung, liver, and spleen), which was further confirmed using ex vivo fluorescence imaging on the major organs of interest (fig. S4, A to C).

Next, MPE models were established using luciferase stably transfected LLC cells (LLC-Luc) in C57BL/6 mice to explore the biodistribution of TMPs. The dynamic distribution of TMPs in the intrapleurally treated group was detected by in vivo imaging at 1, 3, 24, and 48 hours. As shown in Fig. 3A, intrapleural delivery resulted in a rapid increase in the accumulation of pleural TMPs within 24 hours of administration. Thereafter, the fluorescence signal from TMPs began to decrease but was still maintained at a high level at the 48-hour time point, indicating that the TMPs can effectively accumulate and stay in the pleural cavity for that amount of time. At the 24-hour time point, the mice were euthanized, and organs were excised for fluorescence imaging (Fig. 3, B and C). In the intravenously injected group, TMPs primarily accumulated in livers and spleens and secondarily in lungs. However, considerable TMP accumulation was observed in tumor tissues, indicating that intravenous MPs can traffic to the tumor. In contrast to systemic administration, intrapleural injection resulted in the greatest deposition of TMPs in lungs, followed by that in the tumor. These results indicated that local intrapleural administration would be the preferred route because it enables a greater accumulation of TMPs at tumor tissues in the pleural cavity for a longer time than systemic administration in MPE models.

Fig. 3 In vivo tissue distribution and cell targeting of TMPs and TMPs-MTX.

(A) In vivo imaging of C57 MPE models induced by LLC-Luc cells at 1, 3, 24, and 48 hours after intrapleural (IP) injection with DiR-labeled LLC-derived TMPs-MTX. A scale of the radiance efficiency is presented to the right of the images. (B and C) MPE mice were treated by intravenous (IV) or IP injections with DiR-labeled LLC-derived TMPs. Major organs were excised and weighed at 24 hours after injection. Representative ex vivo fluorescence images (B) and fluorescence intensity (C) of major organs and tumors are shown (n = 4). (D and E) Mice with MPE induced by LLC cells were intrapleurally injected with PKH67-labeled LLC-derived TMPs. Twenty-four hours later, mice were euthanized. Representative confocal images of immunofluorescent staining of frozen tumor sections with CD11b (scale bars, 20 μm) (D) and F4/80 (scale bars, 20 μm) (E) primary antibodies are shown. Insets are magnified. (F) Representative flow cytometry dot plots of PKH67-labeled TMP accumulation in lymphocytes and macrophages in tumor tissues. SSA, side scatter. (G) Quantification of PKH67-labeled TMP accumulation in various cell types in tumors and MPE (means ± SEM, n = 4 mice). *P < 0.05, **P < 0.01, and ***P < 0.001.

We also explored the specific cellular internalization of PKH67-labeled TMPs in mice with MPE. In situ visualization assays on frozen sections of MPE cell pellets and pleural tumors revealed that positive PKH67 fluorescence signals were mostly located in the cells positive for CD11b (myeloid marker) and F4/80 (macrophage marker) but only in a handful of cells positive for CD3 expression (T lymphocyte marker) (Fig. 3, D and E, and fig. S4D). Consistently, flow cytometry analysis revealed that PKH67-labeled TMPs were mainly taken up by CD11b+ cells [including tumor-associated macrophages (TAMs) and polymorphonuclear neutrophils (PMNs)] and tumor cells in both the pleural effusion and the primary tumor, whereas only a few TMPs were taken up by CD3+ T cells (Fig. 3F).

Antitumor activity of intrapleurally administered TMPs-MTX in murine MPE models

To investigate the potential effect of TMPs-MTX on MPE in mice, we intrapleurally inoculated C57BL/6 mice with LLC, LLC-Luc, or MC38 cells, followed by intrapleural control or TMP-MTX treatment at 4 × 106 particles/50 μl. Mice were evaluated after 14 days (or 12 days after MC38 cell injection). Primary endpoints were pleural tumor burden and pleural fluid volume. In addition to counting pleural foci, we performed bioluminescence imaging to detect luciferase activity in LLC cells for the evaluation of pleural tumor burden. Compared with the three controls (PBS, empty TMP0, and free MTX), TMP-MTX treatment reduced pleural tumor burden, number of pleural foci, and effusion volume (Fig. 4, A to E). To determine whether the impact of TMPs-MTX is applicable to additional tumors, we used another mouse model of colon (MC38) adenocarcinoma–induced MPE. Similar to those in LLC models, MC38 cell–induced pleural tumor foci and MPE volume were significantly limited by TMP-MTX administration (P < 0.0001 for tumor foci and P = 0.0396 for MPE volume) (Fig. 4, F and G).

Fig. 4 Impact of TMP-MTX treatment on murine MPE models.

(A) Graphical outline of in vivo TMP-MTX treatment experiments. C57BL/6 mice received intrapleural injection of LLC cells or MC38 cells (white arrow) followed by intrapleural treatment with TMPs-MTX packaging 3 μg of MTX, TMP0, MTX (2.5 mg/kg), or PBS three times (on days 5, 7, and 9 after tumor inoculation) (black arrows). Mice were euthanized after 14 days for LLC models and 12 days for MC38 models (gray arrow). (B to E) In vivo bioluminescence images (B) and quantitation of luciferase activity (C) (n = 6 per group), pleural tumor number (D), and pleural fluid volume (E) of mice with LLC-induced MPE treated with TMPs-MTX (n = 14 per group). (F and G) Pleural tumor number (F) and MPE volume (G) of mice with MC38-induced MPE after TMP-MTX treatment (n = 12 per group). The graphs depict data pooled from two independent experiments for MPE volume and pleural tumor number for both LLC and MC38 models. (H and I) Comparisons of CD31-positive cells (H) and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) staining among cells (I) in the four groups (each n = 6). Microvessel density (MVD) defined as the mean number of microvessels determined by CD31 immunostaining. Scale bar, 100 μm. (J and K) Life-span analysis in four groups of LLC and MC38 models. Kaplan-Meier survival analysis shows a comparison of TMP-MTX treatment with the three control groups in LLC models (J, n = 12) and in MC38 models (K, n = 9). *P < 0.05, **P < 0.01, and ***P < 0.001.

To better understand the mechanism by which treatment with TMPs-MTX inhibited MPE formation, we measured microvessel density and apoptosis in pleural tumor cells in the four groups. Pleural tumors distributed between visceral and parietal pleural surfaces were harvested and fixed in formalin, and paraffin-embedded sections were stained with anti-CD31 antibody for microvessel density detection and subjected to TUNEL assays to determine the apoptotic index. The results demonstrated that the expression of CD31 was decreased in mice treated with TMPs-MTX (Fig. 4H), whereas the density of apoptotic cancer cells was higher in the TMP-MTX group than in the three control groups (Fig. 4I).

Life-span analysis showed that the median survival durations of PBS-, empty TMP0–, TMP-MTX–, and free MTX–treated mice bearing MPE induced by LLC cells were 17, 21, 24, and 19 days, respectively (Fig. 4J). The overall log-rank test showed a significant difference in survival among the four groups (P = 0.0081), and pairwise log-rank tests revealed a significant difference between the TMP-MTX group and the free MTX group (P = 0.015). A similar improvement in the median survival was also observed in MC38 models (Fig. 4K). Collectively, these results indicated that TMPs-MTX were effective against MPE induced by two different mouse adenocarcinomas.

TMPs-MTX–mediated shifts toward a tumor-inhibiting microenvironment in murine MPE models

Our previous report demonstrated that TMPs can act as a potent tumor vaccine to induce a specific antitumor immune response by efficiently activating type I interferon (IFN) pathway in a cyclic GMP-AMP synthase (cGAS)/stimulator of interferon genes (STING)–dependent manner (21), which prompted us to explore whether immune infiltrates were affected by TMP-MTX treatment. To this end, we first analyzed the cell constitution of MPE in MC38 cell–induced MPE-bearing mice. The TMP-MTX–treated MPE had a reduced proportion of CD45 cells, suggesting a reduction in the infiltration of tumor cells (Fig. 5A). We observed a concurrent increase in CD3+, CD3+CD4+, and CD3+CD8+ T cells in the TMP-MTX group. The changes in the CD4/CD8 ratio and the frequency of Foxp3+ T regulatory cells were not statistically significant (Fig. 5B). Further dissection of changes within the myeloid cell population (CD45+CD3CD11b+) revealed an increase in PMNs in the TMP-MTX group (Fig. 5C), which could be explained by the inflammatory responses caused by the cytotoxicity of MTX, and a significant reduction in TAMs (P = 0.0033) (Fig. 5D), confirming the selective cytotoxicity of TMPs-MTX toward TAMs. Collectively, the increased T cell populations (including CD3+, CD3+CD4+, and CD3+CD8+ T cells) and decreased TAMs in MPE indicated a resistance to microenvironmental immunosuppression upon TMP-MTX treatment.

Fig. 5 Infiltrating immune cell proportion alternation in pleural tumor and MPE after TMP-MTX treatment.

(A to D) Flow cytometry analysis of MPE samples in MC38 models after different treatments. (A) CD45 cells within the live gate. (B) Lymphocyte infiltrates including CD45+CD3+ T cells, CD45+CD3+CD4+ T cells, CD45+CD3+CD8+ T cells, CD4/CD8 T cell ratio, and CD45+CD3+CD4+Foxp3+ Tregs. The graphs depict data pooled from two independent experiments. (C and D) Myeloid cell infiltrates including PMNs (C) and TAMs (D) within the CD45+ gate. (E to J) Flow cytometry analysis of tumor samples in MC38 models receiving different treatments. (E) Total infiltrates (CD45+) within the live gate. (F) Lymphocyte infiltrates including CD45+CD3+ T cells, CD45+CD3+CD4+ T cells, and CD45+CD3+CD8+ T cells. (G) CD4/CD8 T cell ratio. (H) CD45+CD3+CD4+Foxp3+ Treg infiltrates. (I and J) Myeloid cell infiltrates including PMNs (I) and TAMs (J). Data (shown as means ± SEM) were analyzed by one-way analysis of variance (ANOVA) followed by Bonferroni post hoc analysis. (K to N) Mice bearing LLC tumors and MPE were treated intrapleurally with TMPs-MTX, clodronate liposomes, or their combination. Fourteen days later, MPEs were analyzed by flow cytometry. Representative dot plots of CD11b+F4/80+ cells of MPE samples for each treatment (K) and quantification of TAMs (L), MPE volume (M), and survival rates (N) of mice receiving different treatments are shown (P < 0.001 for combination therapy compared to the PBS group and P = 0.045 compared to the TMP-MTX group by log-rank test). *P < 0.05 and **P < 0.01; ns, no significance.

Considering that pleural solid tumor tissues could have a different distribution of TMPs-MTX compared to pleural effusion due to the physiological barriers and the relatively limited penetration through tumor tissue, we next sought to determine the immune infiltrates in the pleural tumor. The TMP-MTX–treated tumors had a greater proportion of CD45+ infiltrates than the other groups, indicating that TMP-MTX treatment resulted in a more inflammatory phenotype (Fig. 5E). An increased proportion of CD4+ and CD8+ T cells among tumor-infiltrating lymphocytes (TILs) was observed in the pleural tumors (Fig. 5F). A pronounced increase in CD8+ T cells relative to CD4+ T cells resulted in a decrease in the CD4/CD8 ratio (Fig. 5G). The percentage of regulatory T cells (Tregs) (CD4+Foxp3+) among CD4+ TILs was not significantly different (Fig. 5H). Meanwhile, the increased frequency of PMNs and the decreased proportion of TAMs were similar to those in MPE upon TMP-MTX treatment (Fig. 5, I and J). The overall alterations in immune cells in pleural tumors were consistent with an antitumor immune microenvironment.

Similar immune infiltration, including the frequencies of total immune cells, PMNs, and TAMs in MPE and pleural tumors, was seen in the MPE models induced by LLC cells (fig. S5). As for T cells, although TMP-MTX treatment failed to increase T cell frequencies in MPE relative to the PBS control, a significant difference in CD3+ and CD3+CD8+ T cells existed relative to free MTX treatment (P = 0.0389 for CD3+ cells and P = 0.0473 for CD3+CD8+ cells). In addition, a significant decrease in the CD4/CD8 ratio was observed in the pleural tumors (P = 0.003). We also investigated the corresponding immune cells in the blood. PMNs and macrophages were both increased after TMP-MTX treatment in LLC and MC38 models (fig. S6, A and B). Therefore, aside from the reduction in CD45 cells and TAMs in the tumor infiltrate, TMP-MTX treatment increased the percentage of immune effector cells and promoted a tumor-suppressive immune microenvironment.

To further explore the role of TAMs in the antitumor activity of TMPs-MTX, we used mice bearing LLC cell–induced MPE and injected them with clodronate liposomes, resulting in greatly decreased macrophage infiltrates in pleural cavity (fig. S6, C and D). Under these experimental conditions, the depletion of TAMs in MPE was comparable between the mice treated with TMPs-MTX and those treated with clodronate liposomes (Fig. 5, K and L). Combined treatment with clodronate and TMPs-MTX failed to further reduce TAMs but resulted in a significantly improved antitumor activity (P = 0.0123 for MPE volume comparison and P = 0.001 for survival analysis) (Fig. 5, M and N).

Antitumor activity of intrapleurally administered ATMPs-MTX in patients with advanced lung cancer and MPE

Given the promising preclinical results obtained with TMPs-MTX, we initiated a pilot human clinical trial to assess the safety and feasibility of intrapleural infusion of ATMPs-MTX in patients with advanced lung cancer and MPE. From January 2016 to December 2017, 12 lung cancer patients (4 females and 8 males) with histocytologically proven MPE were enrolled in this trial. All patients completed the whole treatment except for one, whose condition deteriorated rapidly from progressive malignancy without receiving ATMPs-MTX, resulting in withdrawal from the study. This patient’s data were excluded from the analysis. The characteristics of the remaining patients are shown in table S1. Strictly following the standard procedure in our registered clinical trial, the eligible patients were treated with ATMPs-MTX via intrapleural infusion six times on alternate days (fig. S7A). The initial dose of ATMPs-MTX, based on animal studies, was 5 U of ATMPs-MTX containing a total dose of 20 μg of MTX, which is several orders of magnitude below the doses used for free MTX in first-line chemotherapy (at least 10 mg/m2). The in vivo effect of ATMPs-MTX on MPE was monitored by routine examination of pleural fluid at baseline and after each infusion, and by ultrasound and computed tomography (CT) scans at baseline, at the end of the treatment (day 12 of treatment), and at 28-day follow-up. Concurrently, whole blood was taken to investigate ATMP-MTX infusion–related adverse effects and immune-related changes (Fig. 6A). Data on 11 participants were evaluated for adverse events according to the Common Toxicity Criteria version 3 from the National Cancer Institute (CTCAEv3). A total of six adverse events were recorded throughout the treatments in all patients (table S2). ATMP-MTX infusion did not produce any acute toxic effects. The most frequently reported adverse events causally related to the use of ATMPs-MTX were mild (grades 1 to 2), including dizziness, fever, nausea, and vomiting. There were no detectable hepatic, renal, pulmonary, cardiac, hematologic, or neurologic toxicities. No clinical manifestations of autoimmune reactions were observed.

Fig. 6 Therapeutic effect of ATMP-MTX in lung cancer patients with MPE.

(A) ATMP-MTX treatment and biomarker assessment schedule. Patients were intrapleurally infused with ATMPs-MTX on days 1, 3, 5, 7, 9, and 11 (blue arrows), accompanied by the routine examination of pleural fluid at baseline and on treatment after each infusion (white triangles), as well as ultrasound and CT scans at baseline, at the end of the treatment (day 12) and at 1-month follow-up (black squares). In addition, blood sampling was performed for assessment of possible adverse effects and immune responses (red triangles). (B) Top: CT images of the thorax of a 56-year-old patient with MPE. The red dashed line indicates the pleural effusion (left, before treatment; right, 28 days after treatment). Bottom: Pleural effusion volume assessment of the same patient by CT analysis (left, before treatment; right, 28 days after treatment). (C) Kaplan-Meier curve of pleurodesis success in all 11 patients. Most participants achieved the predefined pleurodesis criteria at a median of 7 days. Day 1 was defined as the first day of intrapleural treatment with ATMPs-MTX. (D) Top: Light microscope examination of MPE cells from a patient treated with ATMPs-MTX (left, before treatment; right, after treatment). Bottom: Wright-Giemsa stain of MPE cell smear from a patient treated with ATMPs-MTX under a light microscope (left, before treatment; right, after treatment). Scale bars, 100 μm. (E) The proportions of CD45 cells within the live gate in MPE samples from clinical subjects before and after ATMP-MTX treatment. Data are represented as means ± SEM. ***P < 0.001.

Clinical effects were demonstrated by decreased pleural effusion volume and objective clinical responses according to the British Thoracic Society Pleural Disease Guideline, accompanied by symptomatic improvement in 10 of 11 patients with MPE (27). For example, the CT image and pleural effusion volume measurement of a 56-year-old female lung cancer patient with refractory MPE clearly confirmed the control of the MPE by ATMP-MTX treatment (Fig. 6B). Of all 11 patients, 4 patients showed complete response (CR), 6 showed partial response, and 1 showed no response. The objective clinical response rate was 90.91% (10 of 11). The median time to pleurodesis was 7 days (Fig. 6C). The median time to pleurodesis in the four patients who achieved CR by the 28-day assessment was merely 5 days. Long-term pleural fluid control assessment revealed that 9 of 11 (81.81%) patients did not need further therapeutic pleural drainage until death. Of the remaining two patients, one patient received one therapeutic fluid drainage, and the other failed the pleurodesis treatment. Aside from the MPE amount, the characteristics of MPE including fluid color were also affected by ATMPs-MTX, as shown in fig. S7B. Although there is no universal affirmative conclusion about the correlation of the number of red blood cells with therapy response thus far, decreased red blood cells may indicate improved vascular permeability in MPE (31).

To evaluate the actual killing of tumor cells by ATMPs-MTX in situ, we assessed pleural fluid cytology in eight patients before and after ATMP-MTX administration. Clinical activity was correlated with a reduction in tumor cells in pleural effusion examined by light microscopy (Fig. 6D and fig. S7C). Meanwhile, flow cytometry analysis showed that the CD45 cells in posttreatment pleural effusion samples decreased significantly (P = 0.0463), revealing elimination of tumor cells after ATMP-MTX infusion (Fig. 6E). The concentration of carcinoembryonic antigen in MPE and blood, a reliable surrogate marker of chemotherapy efficacy in patients with advanced NSCLC (32), was also reduced after ATMP-MTX treatment, indicating the cytotoxicity of ATMPs-MTX against tumor cells (fig. S7D).

The median overall survival in this study was 240 days (95% confidence interval, from 149 days to infinity) (table S1). Because of the small number of patients involved, no firm conclusions could be drawn about the efficacy of ATMP-MTX treatment from our study.

Immunological responses in the pleural microenvironment of patients with MPE treated with ATMPs-MTX

The modulation of pleural immune infiltrates by ATMPs-MTX was further confirmed in selected patients. The proportions of T cells (including CD3+CD4+ T cells, CD3+CD8+ T cells, and Tregs) in MPE were analyzed by flow cytometry. As shown in Fig. 7A, despite a significant increase after the second infusion of ATMPs-MTX (P = 0.008), total lymphocyte infiltration was unchanged compared to the baseline. There was a slight increase in CD4+ T cells but not in CD8+ T cells in the MPE after ATMP-MTX treatment. Further profiling of TILs from MPE in the study subjects revealed that the proportion of IFN-γ–secreting CD3+ T cells increased after the treatment (Fig. 7B). ATMPs-MTX resulted in a ~2.5-fold increase in interleukin-2 (IL-2) production by CD4+ cells and a ~2-fold increase in IFN-γ produced by CD8+ TILs (Fig. 7, C and D). These results indicated that ATMPs-MTX could activate cytotoxic T lymphocyte (CTL)/T helper 1 (TH1) responses to elicit antitumor immunity. Meanwhile, the percentage of Tregs showed a significant but transient decrease among CD4+ TILs during the treatment cycles (P = 0.016) (Fig. 7E).

Fig. 7 Alteration of immune infiltrates in MPE after ATMP-MTX treatment.

MPE samples from 10 patients (MPE of 1 patient resolved after one injection) were collected to investigate the immune infiltrates by flow cytometry at the indicated time as shown [baseline, before treatment (D0); 1st, 48 hours after the first intrapleural infusion (D3); 2nd, 48 hours after the second infusion (D5)]. (A) The proportions of lymphocytes within the CD45+ gate, CD4+ cells within the CD45+CD3+ T cell gate, and CD8+ T cells within the CD45+CD3+ T cell gate. (B to D) Quantification of cytokines produced by MPE-derived tumor-infiltrating T lymphocytes. TILs were isolated from MPE and analyzed for the expression of IL-2 and IFN-γ after phorbol 12-myristate 13-acetate/ionomycin restimulation by intracellular cytokine staining. (B) IFN-γ+ T cells within the CD45+CD3+ T cell gate. (C) IFN-γ+ cells and IL-2+ cells (n = 9) within CD45+CD3+CD4+ cells. (D) IFN-γ+ cells and IL-2+ cells (n = 9) within CD45+CD3+CD8+ cells. (E) The proportions of regulatory T lymphocytes. (F) The proportions of myeloid cells (CD45+CD3CD11b+). (G) The change in the numbers of tumor-associated neutrophils (TANs) (CD45+CD11b+CD15+CD66b+) of three patients during ATMP-MTX infusions. MPE of the subjects resolved quickly after ATMP-MTX treatment, and MPE samples at 28 days were only available in three patients. Black arrows indicate the days of ATMP-MTX intrapleural infusion therapy. (H and I) (H) The proportions of monocytes (CD45+CD3CD14+) in patients’ MPE samples over time. (I) Representative flow cytometry plots of MPE samples stained for CD14 and CD163, gated on CD45+ cells from a patient with lung adenocarcinoma before and after therapy (left). Quantification of TAMs (CD45+CD14+CD163+) (middle). Percentages of TAMs in three patients at different time points during ATMP-MTX infusions (right). Black arrows indicate the days of ATMP-MTX intrapleural infusion therapy. Data (shown as median ± interquartile range) were analyzed using a repeated-measures model. *P < 0.05, **P < 0.01, and ***P < 0.001. (J) The concentrations of pleural fluid IFN-γ, IL-17, TNF-α (tumor necrosis factor–α), and MCP-1 (monocyte chemoattractant protein 1) (n = 8; the concentration of MCP-1 was below the detection range in two patients’ MPE) were measured. Red squares represent data from the only nonresponder (patient 13).

We further dissected the changes within the myeloid cell population. Myeloid cell numbers peaked on day 3 after the first infusion and then decreased, indicative of a classic acute inflammatory response due to the cytotoxic effect of ATMPs-MTX on tumor cells (Fig. 7F) (33, 34). Human “suppressive” granulocytes (CD45+CD11b+CD15+CD66b+), referred to as TANs, were detected in MPE samples from three patients (Fig. 7G) and decreased during ATMP-MTX infusions (35, 36). Analysis of monocytes and their TAM subtype (CD45+CD14+CD163+) in the MPE revealed sustained elimination, confirming that ATMP-MTX treatment–induced monocyte/macrophage changes in mice were mirrored in patients with MPE caused by advanced lung cancer (Fig. 7, H and I).

The immune profile was also studied in blood samples taken at different time points during and after ATMP-MTX treatment (fig. S8A). ATMP-MTX treatment was associated with an increase in the absolute numbers of circulating myeloid cells, including circulating neutrophils and monocytes, which reached peak values on day 12 of treatment but remained within normal limits, accompanied by a transient decrease in the absolute numbers of lymphoid cells. This change in circulating myeloid cells can be regarded as a side effect of local pleural cavity inflammation and subsequent depletion of TANs and TAMs mentioned above (Fig. 7, F to I). The relative frequencies of CD4+, CD8+, and IFN-γ+ T cells among the CD3+ cells remained unaffected by the cytotoxicity of ATMPs-MTX (fig. S8A).

In addition to the changes in the immune cell infiltrates, a pattern of cytokine/chemokine alterations was also observed in the pleural cavity of patients with MPE after ATMP-MTX treatment (Fig. 7J). Production of TNF-α and IFN-γ, the key cytokines produced by TH1, as well as IL-17, the key cytokine produced by TH17 cells, was increased after treatment. MCP (MCP-1/C-C motif chemokine 2) concentrations were reported to be correlated with fluid volume and may be a critical target in MPE therapy (37, 38). Reductions in MCP-1 concentration in MPE occurred in samples from six of eight analyzed patients treated with ATMPs-MTX. However, the concentrations of IL-4, a key cytokine of TH2 cells (39), and other important cytokines such as vascular endothelial growth factor, IL-6, and IL-10 were unchanged (fig. S8B). Together, these data indicate that in addition to the effects on macrophage depletion, ATMP-MTX treatment can increase the activation of antitumor effector cells including CTLs and TH1 cells within the tumor microenvironment, associated with chemokine secretion.


In this study, we have assessed the feasibility, safety, and efficiency of ATMPs, a drug delivery vector, in lung adenocarcinoma patients with MPE. We have demonstrated that ATMPs-MTX show clinical activity in killing tumor cells and TAMs and induce antitumor immune responses. The treatment protocol tested in this study only caused grade 1 or grade 2 toxicity in patients with MPE, suggesting that ATMPs-MTX can be safely applied in clinical trials.

In our studies, TMPs secreted by tumor cells exposed to UVB light were chosen as the drug carrier. The absence of pro-tumor potential of UVB-induced TMPs examined in our C57BL/6 MPE models is a key premise for the development of TMP-based DDS. Moreover, the preserved integrity and short-term stability of TMPs in vitro, as well as the high biocompatibility, cancer cell–specific targeting, and long circulation time observed in vivo confirmed these previously presumed advantages of UVB-induced TMPs for drug delivery. Naive TMP0 treatment was associated with an increase in CD3+ and CD3+CD8+ T cells in MPE and tumor, as well as CD8/CD4 ratio in MC38 tumor models, indicating the immunoregulatory effect of naive TMP0, which was consistent with our previous study (21). However, TMP0 treatment alone failed to control MPE despite the enhanced antitumor immunity with TMP0, which was in accordance with the idea that treatment with tumor antigens alone may not be able to overcome the immunosuppressive tumor microenvironment and inhibit tumor growth even in highly immunogenic contexts (40). Thus, concerns of efficacy have provided impetus for development of combinatorial approaches that leverage multiple modes of action including TMP0-based immunotherapy strategies as adjuvants to chemotherapy in our study.

Despite the encouraging efficacy and the satisfactory safety profile of TMPs packaging chemotherapeutic drugs in preclinical solid malignancies, great care has been taken when extrapolating these results to the clinical setting. For clinical application, autologous tumor cells rather than immortalized cell lines were used as the donor cells to prepare ATMPs packaging MTX because cell lines would not have the necessary “self” characteristics to avoid recognition by the immune system. Moreover, heterogeneity of immunogenic tumor antigens of individual patients makes ATMPs that share similar tumor antigen repertoires with malignant donor cells a favorable choice for MTX delivery (41). For the clinical setting, ATMPs-MTX were delivered through intrapleural catheters, considering the role of regional administration in the standard of care for MPE (27), the successful clinical translation of regional administration of biological agents such as cytokines (42) and an oncolytic virus (43), as well as the specific biodistribution of ATMPs-MTX delivered via intrapleural infusion, which enables the ATMPs-MTX to have easy access to tumor cells in the pleural cavity and lungs.

Enhanced targeted cytotoxic effects of cancer cells by TMP-MTX treatment have been reported (31) and were further confirmed by our in vitro, in vivo, and clinical studies. This enhanced targeted killing of cancer cells with TMPs-MTX could be explained by several mechanisms. First, because of the high biocompatibility, TMPs exhibit a relatively long circulation time, resulting in a lower combined uptake by the liver and spleen. Meanwhile, the mean particle size of 200 to 300 nm for the TMPs in our study, together with the physical deformability of these lipid-enriched vesicles, eases the extravasation of TMPs through the leaky tumor vasculature and subsequent penetration into the target organ via the enhanced permeability and retention effect. The tumor cells had greater uptake efficiency for TMPs than normal tissue cells or antitumor immune cells, enabling the active cellular targeting of TMPs-MTX. Moreover, after entering cells, TMPs not only can deliver drugs into the nucleus but also can interfere with the ATP-binding cassette transporter system to impede drug efflux, resulting in the death of tumor-repopulating cells (44). TMPs-MTX can have domino-like effects, where drug-packaging TMPs can trigger the formation of new drug-packaging TMPs after entering tumor cells, hence maintaining cytotoxicity (45).

On the basis of the cytotoxic effects mentioned above, intrapleural administration of ATMPs-MTX efficiently decreased tumor microvessel density, constrained MPE accumulation, and prolonged survival in C57BL/6 mouse models. Mirroring the findings from the animal study, ATMP-MTX treatment effectively decreased MPE volume and relieved MPE-related symptoms (cough, gasping, and respiratory distress) in patients. Although describing pleurodesis efficacy was not the primary aim of this study, the 28-day response rate (90.91%) and long-term pleural fluid control rate (81.81%) after ATMP-MTX administration were encouraging in this regard. If this efficacy is replicated in larger studies, then ATMPs-MTX will be comparable to talc pleurodesis, which has a 42-day response rate of 89% and a long-term pleural fluid control rate of 78% (46). Although survival data from our trial are not conclusive because of the limited number of patients, the median of 240 days is very promising compared with previously reported data (a median of 74 days) (26). The two most severe and most common toxicities of MTX are hepatotoxicity and bone marrow suppression (47), which were not observed in this study. The overall safety profile for ATMPs-MTX was acceptable, with the most frequent adverse event being mild nausea.

In addition to the targeted cytotoxic effect, substantial clinical and experimental evidence suggests that TMPs-MTX may modulate the tumor microenvironment. TAMs are key promoters of cancer-related inflammation (48) and are regarded as the dominant immunosuppressive cell type in MPE (28, 37). In our study, ATMPs-MTX effectively decreased the frequency of TAMs in animal models and in the clinical study. In addition, the expression of MCP-1, a marker of macrophage accumulation in tumor sites, was reduced after treatment in most cancer patients analyzed. Therefore, ATMPs-MTX can down-modulate the frequency of mononuclear phagocytes in MPE by direct cytotoxic activity and by decreased recruitment of circulating monocytes into tumors, thereby eliminating an important immunosuppressive component of the tumor microenvironment.

An increase in IFN-γ+ T cells including CD4+ T cells and CD8+ T cells in MPE was detected 48 hours after the first injection of ATMPs-MTX and lasted for at least 5 days. The assessment of long-lasting effects on IFN-γ+ T cells was not possible because of the control of MPE. The reinvigoration of T cells can be attributed to multiple factors, including DC maturation stimulated by taking up antigenic ATMPs and tumor antigens exposed by apoptotic tumor cells (21, 49), as well as the indirect influence of monocyte depletion as previously reported (50). Consistently, TH1 cytokine concentrations (TNF-α and IFN-γ) were also increased after administration of ATMP-MTX. On the basis of this evidence, we conclude that ATMP-MTX treatment can remodel the immunosuppressive microenvironment in MPE.

We revealed an enhanced anti-MPE effect by combining TMPs-MTX with liposomal clodronate. Although the specific mechanism of action for this enhanced effect has not been clarified yet, we speculate that clodronate could decrease infiltrating mononuclear macrophages in tumors, improving tumor cells’ internalization of TMPs-MTX, which otherwise would be taken up by TAMs. In addition, macrophage depletion by liposomal clodronate can be associated with reduced tumor angiogenesis and metastasis, as well as chemotherapy and immunotherapy resistance (5153).

Despite the above advantages, our study of ATMP-based DDSs has some limitations. First, although our immunocompetent mouse models of MPE have been used for a number of MPE-related mechanistic or therapeutic studies (54, 55), these models mimic the local microenvironment of the pleural cavity rather than that of the primary tumor. Interpretation of the consistency between animal models and clinical studies should consider not only the uncertainty in cross-species translation but also the pitfalls of the experimental MPE models themselves. Thus, models that can better mimic MPE formation by primary lung cancer remain to be established. Second, the key factors that contributed to our ability to translate this targeted chemo-immunotherapeutic platform from the bench to human clinical trials included the availability of autologous tumor cells in MPE and the development of a robust, scalable process for ATMP-MTX formation. However, access to autologous tumor cells is not always feasible when trying to extend this approach to other solid tumor types. To solve this problem, alternative cell sources such as the immortalized cell line A549 have been explored as donor cells for the development of TMP-based DDSs, and initial efficacy has been identified in six patients with MPE (44). However, more evidence is needed to address the risks of potentially dangerous immune responses caused by therapy with allogeneic cell-derived MPs. In addition, data on changes in pleural fluid immune infiltrates after ATMP-MTX treatment were unavailable when pleural fluid production in clinical subjects ceased entirely, which means that only a short-term evaluation of immunoregulatory effects in the pleural cavity was performed. In our study, all patients’ MPE samples were obtained after two injections of ATMPs-MTX (on day 5), except in one subject whose MPE disappeared entirely after just one injection. However, the paucity of available pleural fluid samples from the patient with the most effective pleural fluid control could have affected our results on immune alteration in MPE, resulting in underestimation of the immunoregulatory effect of ATMPs-MTX. Last, despite an acceptable safety profile and no evidence of pro-tumor potential of TMP-based vectors, there is still concern for potential long-term hazards of introducing a TMP-based vector into a human, and the long-time effects on patients’ health need to be monitored in the clinical trial follow-up.

To maximize the potential of ATMPs-MTX, a number of areas should be further explored, including optimization of the dosing schedule, which we based on the dosage used for animal experiments; identification of patient populations most likely to benefit from treatment with ATMPs-MTX based on, for example, MTX sensitivity or tumor burden in the pleural cavity; and investigation of the systemic administration of ATMPs-MTX. Last, particle size distribution is one of the main characteristics in the drug delivery context, affecting circulation kinetics, tissue biodistribution, and safety of system administration (56). Although TMPs with a wide size distribution used in our study had relatively good safety profiles and stable efficacy with local application, a narrower size distribution will be critical for the intravenous delivery of a TMP-based drug system. Extensive investigation will be required to identify the optimal size range and to manufacture desired TMPs with a specific size range.

In summary, we constructed a platform for targeted chemotherapy based on ATMPs, and our studies highlighted the potential of ATMP-based dual-functional MP technologies as chemo-immunotherapeutic agents. Our study demonstrated that ATMPs-MTX can be highly active in both C57BL/6 MPE models and lung adenocarcinoma patients with MPE. Therapeutic efficacy with prolonged survival was achieved by ATMPs-MTX in experimental MPE models, and positive clinical responses were observed in lung adenocarcinoma patients with MPE. Although clinical benefit could not be definitively determined from this small study, our results have shown that ATMPs-MTX are safe, nontoxic, and tolerable when delivered by intrapleural administration. In addition, we observed immune alterations in our animal and clinical experiments. Our study provides a promising strategy to treat advanced lung cancer with MPE by converting autologous tumor cell–released vesicles into intrinsically biocompatible drug carriers, providing early evidence for the initiation of larger and robust clinical trials to assess the efficacy of these drug vectors.


Study design

The aim of the study was to evaluate the application of TMP-based targeted chemotherapy in the setting of MPE. Our hypothesis was that TMPs-MTX would show enhanced antitumor efficacy and potentiate antitumor responses against MPE. We first used immunocompetent mouse models of MPE induced by murine LLC and colon adenocarcinoma cells to define the impact of TMPs-MTX on MPE control. Therapeutic effects of TMPs-MTX were evaluated by MPE volume, pleural tumor burden, and survival. For mechanistic studies, systemic and intratumoral immunological parameters were analyzed by flow cytometry. To explore the role of macrophages in TMP-MTX treatment, macrophages were inhibited by clodronate liposomes. Mice used for these studies were on the C57BL/6 background and were randomly assigned to treatment groups 4 days after pleural tumor inoculation. The sample sizes were selected on the basis of the results of pilot experiments to achieve a significant difference between experimental groups. Specific sample sizes for experimental groups and controls, as well as experimental replicates, are indicated in the figure legends. The details regarding injection, grouping, intervention, harvest, sample handling, and survival analysis are described in the Supplementary Materials. After observing that TMPs-MTX could effectively restrict MPE growth and provide a survival benefit in MPE models, we started a prospective, randomized, double-blind study (clinical trial identifier NCT02657460), with the primary objectives of safety and feasibility. No dose escalation was planned.

Statistical analysis

All statistical analyses were conducted using GraphPad Prism 6 (GraphPad). Data obtained from in vitro assays were subjected to unpaired two-tailed t test or one-way ANOVA with Tukey’s test for multiple comparisons. Statistical analysis involving immune parameters in mice was performed with one-way ANOVA followed by Bonferroni post hoc analysis. Survival of differentially treated tumor-bearing mice was compared by the Kaplan-Meier method and the log-rank (Mantel-Cox) test. The immune alterations in the clinical study were analyzed using the repeated-measures model of ANOVA with post hoc Bonferroni correction. All grouped data are presented as means ± SEM unless otherwise stated, and P < 0.05 was considered to indicate statistical significance. For all other Materials and Methods, see the Supplementary Materials.


Materials and Methods

Fig. S1. Characterization of TMPs and TMPs-MTX.

Fig. S2. Effect of TMPs on tumor cell viability and invasion in vitro.

Fig. S3. Cellular uptake and cytotoxicity of TMPs-MTX and the interaction of TMPs-MTX with immune cells.

Fig. S4. Biodistribution and internalization of TMPs-MTX in vivo.

Fig. S5. Analysis of immune infiltrates in pleural tumor and MPE of LLC cell–induced murine MPE models after ATMP-MTX treatment.

Fig. S6. Immune cell components in peripheral blood after ATMP-MTX treatment and macrophage depletion by clodronate liposome treatment.

Fig. S7. ATMP-MTX treatment in advanced lung cancer patients with MPE.

Fig. S8. Immune cell components in peripheral blood and cytokine concentrations in MPE of lung cancer patients treated with ATMPs-MTX.

Table S1. Patient demographics.

Table S2. Patient adverse events related to ATMP-MTX treatment.

References (5760)


Acknowledgments: We would like to thank members of the Research Center for Translational Medicine, and managers of Flow Cytometry, In Vivo Imaging core facilities at the central laboratory of the Wuhan Union Hospital for valuable help; W. Zou (University of Michigan School of Medicine, Ann Arbor, MI) for the gift of MC38 colon adenocarcinoma cells and B16-F10 melanoma cells; and S. Zhou for help with animal experiments. Funding: This study was supported by the National Natural Science Foundation of China (nos. 81572942 and 81770096), the Natural Science Foundation of Hubei Province (no. 2014CFA057), the Special Fund for Technological Innovation of Hubei (no. 2017ACA094), the Health and Planning Commission Fund of Hubei Province (WJ2017M098), the Science and Technology Support Program of Hubei Province (YSF2015001294), the Wuhan Planning Project of Science and Technology (no. 2014060101010036), and the Special Fund for Industrial Transformation and Upgrading. The study was also funded and sponsored by Soundny (Sheng-Qi-An) Biotech. Author contributions: Conceptualization: M.G., F.W., and Y.J.; methodology: M.G., X.L., K.T., Y.J., and F.W.; investigation: M.G., P.X., G.H., L.C., X.W., Y.L., S.Z., S.L., T.L., J.F., Z.L., G.Y., B.L., L.D., M.Z., and X.T.; formal analysis: M.G., F.W., and G.H.; writing—original draft: M.G. and Y.J.; writing—review and editing, M.G., F.W., G.H., Q.H., and Y.J.; funding acquisition: Y.J.; resources: F.W. and J.X.; supervision: Y.J. Competing interests: U.S. Patent No. 9,351,931 and China Patent No. ZL201110241369.8 held by Hubei Soundny Bio-Tech Co. Ltd. cover the pharmaceutical preparation for “Tumor cell-derived microparticles packaging of chemotherapeutic drugs.” All authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials. Patient-derived autologous tumor cell–derived MPs loaded with chemotherapeutic drug MTX (ATMPs-MTX) are available from Y.J. under a material transfer agreement with Hubei Soundny Bio-Tech Co. Ltd for clinical research use.

Stay Connected to Science Translational Medicine

Navigate This Article