Ultra-sensitive Nanoprobe Modified with Tumor Cell Membrane for UCL/MRI/PET Multimodality Precise Imaging of Triple-Negative Breast Cancer

A biomimetic nanoprobe was built with cancer cell membrane-coated and Gd3+-doped upconversion nanoparticles. The nanoprobe could be applied to in vivo UCL/MRI/PET multimodality precise imaging and successfully differentiated MDA-MB-231 tumor models through in vivo tri-modality imaging, which may be used for breast cancer molecular classification. A biomimetic nanoprobe was built with cancer cell membrane-coated and Gd3+-doped upconversion nanoparticles. The nanoprobe could be applied to in vivo UCL/MRI/PET multimodality precise imaging and successfully differentiated MDA-MB-231 tumor models through in vivo tri-modality imaging, which may be used for breast cancer molecular classification. Triple-negative breast cancer (TNBC) is a subtype of breast cancer in which the estrogen receptor and progesterone receptor are not expressed, and human epidermal growth factor receptor 2 is not amplified or overexpressed either, which make the clinical diagnosis and treatment very challenging. Molecular imaging can provide an effective way to diagnose TNBC. Upconversion nanoparticles (UCNPs), are a promising new generation of molecular imaging probes. However, UCNPs still need to be improved for tumor-targeting ability and biocompatibility. This study describes a novel probe based on cancer cell membrane-coated upconversion nanoparticles (CCm-UCNPs), owing to the low immunogenicity and homologous-targeting ability of cancer cell membranes, and modified multifunctional UCNPs. This probe exhibits excellent performance in breast cancer molecular classification and TNBC diagnosis through UCL/MRI/PET tri-modality imaging in vivo. By using this probe, MDA-MB-231 was successfully differentiated between MCF-7 tumor models in vivo. Based on the tumor imaging and molecular classification results, the probe is also expected to be modified for drug delivery in the future, contributing to the treatment of TNBC. The combination of nanoparticles with biomimetic cell membranes has the potential for multiple clinical applications.


Introduction
Breast cancer (BC) is a malignant disease leading to approximately two million new cases (11.6%) and 620,000 deaths worldwide in 2018 [1]. It is also a highly heterogeneous disease requiring molecular classification for treatment and prognosis [2]. Triple-negative breast cancer (TNBC) is a subtype of BC in which the estrogen receptor (ER) and progesterone receptor (PR) are not expressed, and the human epidermal growth factor receptor 2 (HER2) is not amplified or overexpressed either [3]. At present, the diagnosis of TNBC mostly relies on tissue biopsy, which is affected by sampling error and invasiveness [4]. Therefore, the clinical diagnosis of TNBC remains challenging.
Molecular imaging can in principle provide powerful tools for identifying cancers with greatly improved specificity and sensitivity [5], and functional inorganic nanoparticles have shown great potential as imaging probes owing to the intrinsic physical properties of inorganic particles apart from the well-known enhanced permeation and retention effect of tumors for nano-objects. Among different kinds of functional inorganic nanoparticles, upconversion nanoparticles (UCNPs) are superior as light-emitting, and rareearth ions can facilely be combined with paramagnetic ones such as Gd 3+ , for simultaneously visualizing tumors through upconversion luminescence (UCL) [6] and magnetic resonance imaging (MRI) [7,8]. Besides, they also provide an excellent platform to further combine photoacoustic imaging (PAI) [9], single-photon emission computed tomography (SPECT) [10], and positron emission tomography (PET) imaging [11], with above imaging modalities.
Every imaging modality has its intrinsic advantages and disadvantages. To acquire multi-dimensional biological information, rationally combining different imaging modalities is essentially required, thus receiving increasing attention [7,[11][12][13]. The UCL of rare-earth nanoparticles typically requires near-infrared (NIR) lasers of 800 or 980 nm as excitation source which can reach deeper tissues and avoid autofluorescence of the biological tissues [14], particularly suitable for BC detection. The excellent photostability, lack of photoblinking, and ultra-sensitive detection of upconverting visible emissions make UCNPs promising for bioimaging [14], even for real-time imaging [15,16]. MRI can provide high spatial resolution of soft tissues [17]. PET exhibits high sensitivity and unlimited detection depth [13]. UCNPs can well combine these strengths by being properly incorporated with Gd 3+ and SPECT/PET nuclides [7,11,12]. However, the nanoparticles are prone to stimulate the mononuclear phagocyte system, shortening their blood residence time [18]. Therefore, different types of surface engineering approaches have been developed by modifying the nanoparticles with either artificial materials or natural substances. Poly(ethylene glycol) (PEG) [19] and cell membranes [15,20,21] are two representative examples of the aforementioned approaches.
In this study, we used the cancer cell membrane of MDA-MB-231, a kind of TNBC cell, to modify Gd 3+ -doped upconversion nanoparticles NaGdF 4 :Yb,Tm@NaGdF 4 (CCm 231 -UCNPs) for in vivo UCL/MRI/PET tri-modality tumor imaging of BC and further differentiate between breast cancer subtypes of MDA-MB-231 and MCF-7 to demonstrate the potential of CCm-UCNPs in BC molecular classification (Scheme 1).

Preparation of Cancer Cell Membranes
Cancer cell membrane-derived vesicles (CCm) were prepared according to the previous report [12,16,29]. MDA-MB-231 human breast cancer cells were maintained in DMEM supplemented with 10% FBS and 1% penicillin-streptomycin. To harvest membranes, cancer cells were grown in T-175 culture flasks to full confluence and detached with 2 mM EDTA in PBS and washed in PBS three times by centrifuging at 1000 rpm for 4 min. The cells were suspended in a hypotonic lysing buffer consisting of 20 mM Tris-HCl, 10 mM KCl, 2 mM MgCl 2 , and 1 EDTA-free mini protease inhibitor tablet per 10 mL of solution and disrupted at 4 °C overnight. The entire solution was ultrasonically dispersed by an ultrasonic homogenizer (Scientz-IID, Ningbo Scientz Biotechnology Co., Ltd., Ningbo, China), before spinning down at 3200 g for 5 min. The supernatants were saved while the pellet was resuspended in hypotonic lysing buffer and dispersed by ultrasound again. The supernatants were pooled and centrifuged at 20,000 g for 20 min, after which the pellet was discarded and the supernatant was centrifuged again at 100,000 g for 1 h using an ultra-speed centrifuge (Optima XPN-100 Ultracentrifuge, Beckman Coulter, Miami, FL, USA). The pellet containing the plasma membrane material was then washed once with 10 mM Tris-HCl and 1 mM EDTA. The final pellet was collected and used as purified cancer cell membranes for subsequent experiments.

Characterization of Cancer Cell Membrane Protein
SDS-PAGE gel electrophoresis displayed for protein characterization. All samples were prepared at a final protein concentration of 1 mg mL −1 in loading buffer as measured by a bicinchoninic acid (BCA) assay. CCm 231 -UCNPs were purified by centrifugation at 13,000 rpm to pellet the coated particles but not free vesicles or protein. The mixture of samples and loading buffer with the volume ratio of 4:1 was heated to 100 °C for 10 min, and the mixture of samples and loading buffer with the volume ratio of 3:1 was heated to 37 °C for 30 min. The denatured sample was loaded into each well in an Electrophoresis System (Cat #DYZC-24DN,   A CCK-8 assay was used to evaluate the cytotoxicity of nanoparticles to MDA-MB-231 cancer cells. Cells were seeded in 96-well plates at a density of 3 × 10 3 cells per well and cultured for 12 h. Then UCNPs and CCm 231 -UCNPs at various concentrations (i.e., 5, 25, 50, 100, and 500 μg mL −1 ) were added to the medium, and the cells were incubated for another 24 h. The cells grown without any nanoparticles were used as a control. At the end of the incubation, 5 mg mL −1 CCK-8 PBS solution was added, and the plate was incubated for another 4 h. Finally, the absorbance values of the cells per well were determined with a microplate reader (Bio-Rad, Hercules, CA, USA) at 450 nm for analyzing the cell viability. The background absorbance of the well plate was measured and subtracted. The cytotoxicity was calculated by dividing the optical density (OD) values of treated groups (T) by the OD values of the control (C) (T/C × 100%).

Animals and Tumor Model
Animals received care under the instruction of the Guidance Suggestions for the Care and Use of Laboratory Animals. Four-to six-week-old female BALB/c nude mice (Beijing HuaFuKang Bioscience Co. Ltd, China) were subcutaneously injected with 100 μL serum-free PBS containing 5 × 10 6 MDA-MB-231 cells or MCF-7 cells into the upper limbs or lower limbs of each mouse, according to the imaging situation. About 1 week after the injection, when the tumor volume reached 60-250 mm 3 , the tumor-bearing mice model would be used for further experiments.

In Vivo Imaging
When the volumes of MDA-MB-231 tumor or MCF-7 tumor reached 60-250 mm 3 , the BALB/c nude mice were divided into groups randomly and received an i.v. injection of 200 μL PBS or PBS containing the different nanoparticle preparations (i.e., UCNPs, RBC-UCNPs, and CCm 231 -UCNPs) at the concentration of 5 mg mL −1 via the tail vein. All mice were anesthetized by isoflurane. For in vivo UCL imaging, the fluorescence signals were obtained by an ex/in vivo imaging system (IVIS Lumina XRMS Series III, Perki-nElmer Inc.) equipped with fluorescent filter sets (excitation/ emission = 980/790 nm) at 0, 3, 6, 12, 24, 36, and 48 h after the injection. Then all mice were killed to obtain the tumor and major organs (heart, liver, spleen, lung, kidneys) to conduct the ex vivo UCL signals through the ex/in vivo imaging system (IVIS Lumina XRMS Series III, PerkinElmer Inc.).

Biodistribution Studies
MDA-MB-231 or MCF-7 tumor-bearing mice injected with N 3 -DSPE-PEG-CCm 231 -UCNPs and Al[ 18 F]F-L-NETA-DBCO to evaluate whole-body elimination and excretion were killed by cervical dislocation under anesthesia using isoflurane; samples of blood, normal tissues (including brain, heart, lung, liver, spleen, kidney, stomach, small intestine, large intestine, muscle, bone), and tumor were collected and weighed; and the radioactivity in each was measured in a γ-counter. Tissue radioactivity was calculated as percent injected dose/g (% ID/g) and then converted to % ID/organ using previously determined standard organ weights. The concentration of Gd 3+ in each organ was then measured by inductively coupled plasma-atomic emission spectroscopy (ICP-AES). For this analysis, samples were prepared as described by Rao et al. [16]. ICP-AES was performed using a Prodigy 7 (Leeman Labs Inc., Hudson, NH, USA) instrument, and the concentration (mg L −1 ) of Gd 3+ in the tissues was obtained by reference to a standard curve.

Statistical Analysis
Results are expressed as mean ± standard error of the mean. Data analyses were conducted using the software GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA). The differences among groups were analyzed using one-way ANOVA followed by Tukey's post-test. P value of < 0.05 indicates statistical significance.

Preparation and Characterization of CCm-UCNPs
To construct CCm-UCNPs, MDA-MB-231 cells, as the membrane source, were processed by using the method reported by Rao et al. [16]. Briefly, the membrane derivation was achieved through a combination of hypotonic lysis, mechanical membrane disruption, and differential centrifugation. With the collected membranes, CCm vesicles were then formed by physical extrusion through a 400-nm porous polycarbonate membrane on a mini extruder. Then, mixing UCNPs with the membranes by ultrasound and repeating the physical extrusion through a 200-nm pore, CCm was then coated by physical extrusion to form CCm 231 -UCNPs. As measured by dynamic light scattering (DLS), the hydrodynamic size of the UCNPs was approximate 75 nm (Fig. 1a, b), while that of CCm 231 was around 400 nm. Upon optimization of UCNPs:CCm 231 ratio, CCm 231 -UCNPs of approximately 200 nm were obtained, which presented the best hydrodynamic size and polydispersity in comparison with those obtained with different UCNPs:CCm 231 ratios (Fig. S1a). The zeta-potential measurements shown in Fig. 1c exhibited that the surface potential of CCm 231 -UCNPs was much closer to CCm 231 rather than that of the mother UCNPs, indicating that CCm 231 coating was successfully achieved. Transmission electron microscopy (TEM) is shown in Fig. 1d-g, the membrane coating around the UCNPs can be visualized with a thickness of around 3 nm, and the CCm 231 evenly was coated around the UNCP cores of ~ 25 nm. The as-prepared CCm 231 -UCNPs presented good colloidal stability in 1 × phosphate-buffered saline (PBS) for 3 days (Fig. S1b). Also, the upconversion luminescence CCm 231 -UCNPs remained nearly unaltered in comparison with that of UCNPs under excitation at 980 nm (Fig. S1c), indicating that CCm coating did not affect the upconversion luminescence of UCNPs.
To confirm the existence of the CCm proteins on the UCNPs, the CCm 231 -UCNPs were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blotting analysis. The SDS-PAGE results shown in Fig. 1h revealed that CCm 231 -UCNPs presented a protein profile very similar to those of CCm 231 and MDA-MB-231 cell lysate. The western blotting analysis confirmed that the CCm 231 -UCNPs well inherited the membrane-specific markers such as Na + /K + -ATPase and positive antigen EGFR from the MDA-MB-231 cells, while the control nuclear protein marker (i.e., histone H3) and the cytosol marker (i.e., glyceraldehyde 3-phosphate dehydrogenase (GAPDH)) were almost undetectable from the final CCm 231 -UCNPs, as shown in Fig. 1i.

In Vitro Homologous Targeting and Cytotoxicity of CCm 231 -UCNPs
It was expected that CCm 231 coating would endow the  Table S1), and human cervical cancer cells (HeLa) were adopted to show the binding affinity of Cy5-DSPE-PEG-CCm 231 -UCNPs after co-incubated with the corresponding cells. The confocal laser scanning microscopy (CLSM) results given in Fig. 1j revealed that only MDA-MB-231 cells presented the strongest fluorescence after incubation with Cy5-DSPE-PEG-CCm 231 -UCNPs. These confirmed the specific homologous-targeting ability of CCm 231 -UCNPs to MDA-MB-231 cells. There is some evidence proving that cancer cell membranes play an important role in immune tolerance in the tumor microenvironment [29]. To investigate the low immunogenicity of CCm 231 -UCNPs, murine macrophagelike cells (Raw264.7) were used. The Raw264.7 cells had nearly no uptake of Cy5-DSPE-PEG-CCm 231 -UCNPs ( Fig. 1j), indicating that CCm-UCNPs can obscure their identification by the mononuclear phagocyte system and indeed decreased immunogenicity. And MDA-MB-231 cells showed no uptake of DSPE-PEG-Cy5, which could exclude the effect of DSPE-PEG-Cy5 (Fig. S2).
The cholecystokinin (CCK)-8 assay was used to evaluate the cytotoxicity of UCNPs and CCm 231 -UCNPs to MDA-MB-231 cells. MDA-MB-231 cells were co-incubated with UCNPs and CCm 231 -UCNPs at concentrations up to 500 μg mL −1 for 24 h. The results revealed that the survival rates of cells in the UCNPs group and the CCm 231 -UCNPs group were both > 80% (Fig. 1k), indicating UCNPs and CCm 231 -UCNPs exerted no obvious toxicity toward MDA-MB-231 cells.

In Vivo UCL/MRI/PET Imaging
In this work, we speculated that CCm 231 -UCNPs would exhibit the homologous-targeting ability to MDA-MB-231 cancer cells in vivo. We set red blood cell membrane-coated UCNPs (RBCm-UCNPs) in PBS as a control to compare the results of homologous targeting and passive targeting. PBS or PBS containing CCm 231 -UCNPs, RBCm-UCNPs, or UCNPs at the same concentration were injected into MDA-MB-231 subcutaneous-tumor-bearing BALB/c nude mice through the tail vein. To avoid the impact of liver uptake on tumor area imaging, we injected cancer cells into the lower limb of the mice models [12,15,16,29].
For in vivo UCL imaging, the fluorescence signals were obtained after the injection. As shown in Fig. 2a, the CCm 231 -UCNPs group displayed the strongest UCL signal in the tumor site, demonstrating the homologous binding ability of CCm 231 -UCNPs to target MDA-MB-231 tumors. The RBCm-UCNPs group showed similar tumor accumulation to the UCNPs group attributed to the EPR effect, but lower fluorescence signal compared to that of the CCm 231 -UCNPs group. Liver is one of the primary organs of the phagocyteenriched reticuloendothelial system (RES) [30] and can accumulate the most nanoparticles [31]. Liver accumulation by the CCm 231 -UCNPs group was much lower than that of the UCNPs group, indicating that CCm coating can indeed decrease the RES uptake. In the CCm 231 -UCNPs group, the liver signal was of short duration, almost disappeared in 24 h, while the tumor signal lasted much longer, up to 48 h. The best imaging time was at 24 h after the injection with high uptake by the tumor and low uptake by the liver, indicating the feasibility of using the CCm 231 -UCNPs for imaging and treatment delivery. To investigate the homologous-targeting ability, CCm 231 -UCNPs were injected into MDA-MB-231 and MCF-7 tumor-bearing mice. The UCL signal in the MDA-MB-231 group was much higher than that of the MCF-7 group (Fig. 2a), further showing the enhanced specificity accomplished by CCm coating. At 48 h after injection, all the mice were killed and the tumors and other major organs were collected for ex vivo UCL imaging. As shown in Fig. 2b, c, CCm 231 -UCNPs injected into MDA-MB-231 tumor-bearing mice displayed the highest rate of tumor accumulation. The amount of CCm 231 -UCNPs accumulated in the liver and spleen was much lower than that of UCNPs. For in vivo MR imaging, the groups were set the same as those for UCL imaging. T1-weighted MR images were obtained at 24 h after the injection. The images of the Gd-DTPA group were acquired immediately after the injection. The group injected with CCm 231 -UCNPs showed slight enhancement in the tumor region compared to the groups injected with RBCm-UCNPs and UCNPs (Fig. 3a). Similarly, the tumor regions of CCm 231 -UCNPs injected into the MDA-MB-231 tumor groups were slightly enhancing than the CCm 231 -UCNPs injected into the MCF-7 groups (Fig. 3a).

MDA-MB-231
Gd-DTPA  (Figs. 3b-d and S4). Also, the liver and spleen uptake were low. Injecting N 3 -DSPE-PEG-CCm 231 -UCNPs into MDA-MB-231 and MCF-7 tumor-bearing mice, respectively, the tumor site uptake of the MDA-MB-231 group was much higher than that of the MCF-7 group (Fig. 3b,  c). The results were in accordance with that of UCL and MR imaging. Considering all these multimodality imaging results, we demonstrated that CCm 231 -UCNPs targeted to MDA-MB-231 tumors and were of greatly reduced immunogenicity.

Biodistribution of CCm 231 -UCNPs
To quantitatively analyze the biodistribution of the probe, all the mice were killed, and blood samples, tumors, and major organs were collected for biological distribution by an automatic gamma counter. We injected  (Table S2), CCm 231 -UCNPs to MDA-MB-231 model group (Table S3) showed significantly higher accumulation in the tumor (P < 0.001) (Fig. 4a, b); the results were in accord with that of imaging. Furthermore, the much higher tumorto-blood ratios (T/B) and tumor-to-muscle ratios (T/M) of MDA-MB-231 model group were seen than those of MCF-7 model group (P < 0.05) (Table S2). It proved that CCm-coated nanoparticles had homologous-targeting ability. It could be inferred that the homologous-targeting ability of CCm-UCNPs can be widely applied to other types of tumors.
The detection of the 18 F signal can only indirectly determine the biodistribution of nanoparticles, while directly detecting the distribution of Gd 3+ in various tissues in the body must be achieved by inductively coupled plasma-atomic emission spectroscopy (ICP-AES). The results showed that the biodistribution of CCm 231 -UCNPs was mainly concentrated in the liver and spleen, and was rarely distributed in blood, heart, and lung ( Fig. 4c and Table S4). The Gd 3+ uptake in the liver measured by ICP-AES for UCNPs was 1.82-fold higher than CCm 231 -UCNPs (P < 0.001), and splenic uptake of Gd 3+ for UCNPs was 1.79-fold greater than for CCm 231 -UCNPs (P < 0.05) (Fig. 4d). These results indicated that the uptake of CCm 231 -UCNPs by the liver and spleen was greatly reduced after coating with CCm, which proved their lowered immunogenicity. Tumor uptake of Gd 3+ for CCm 231 -UCNPs was 2.62-fold higher than UCNPs (P < 0.001) (Fig. 4e). Also, the Gd 3+ uptake in the tumor for CCm 231 -UCNPs to the MDA-MB-231 model group was 2.70-fold higher than that in CCm 231 -UCNPs to MCF-7 model group (P < 0.001) (Fig. 4e). There were significant differences in tumor uptake of Gd 3+ , showing that CCm 231 -UCNPs can homologously target tumors. The minimal uptake of UCNPs in the tumors may be attributed to the EPR effect. The difference between the MDA-MB-231 and MCF-7 model groups proved that CCm 231 -UCNPs could be used for the molecular classification of breast cancer in the future.

In Vivo Toxicity Evaluation
There is always a concern about potential toxicity and whether cancer cells or their membranes confer a cancer risk. For evaluating systematic toxicity, BALB/c nude mice received an injection of PBS, or PBS containing UCNPs, CCm 231 -vesicles, or CCm 231 -UCNPs. Neither death nor significant differences in body weight among the four groups were observed after 30 days (Fig. 5a). On the 30th day after the injection, all the mice were euthanized and their blood and major organs were collected for blood biochemistry, hematology tests, and histology analysis. We observed no significant differences between the treatment groups and control groups in the blood parameters and blood biochemistry indicators (Fig. 5b-l), and no significant organ damage on hematoxylin and eosin (H&E)-stained slices (Fig. 5m). Above all, the results suggested no obvious toxicity or carcinogenicity of CCm 231 -UCNPs in vivo.

Conclusion
In conclusion, we designed a novel probe using MDA-MB-231 cancer cell membrane-mimic Gd 3+ -doped upconversion nanoparticles (CCm-UCNPs). In this probe, natural cell membranes isolated from cancer cells were coated into the UCNPs. The probe exhibited homologoustargeting and immune escaping abilities. Together with the upconversion luminescence of UCNPs, the paramagnetism of Gd 3+ , and click chemistry with surface modification to label 18 F, CCm-UCNPs were used for ultra-sensitive in vivo UCL/MRI/PET multimodality precise imaging of TNBC, and differentiating MDA-MB-231 and MCF-7 tumor-bearing mice models in vivo. Based on these results, this probe can also be modified for drug delivery, contributing to the treatment of TNBC. It may also be a potential method to achieve integration of diagnosis and treatment [35], as well as to monitor and evaluate therapeutic effects. In addition, apart from cancer cell membranes, various membrane functions, including chemotaxis of platelets to atherosclerotic plaque [25][26][27]36], chemotaxis of leukocytes to inflammation [37], phagocytosis of macrophages [38], and cancer-targeting capabilities of stem cells [39], have the potential to be used for targeting different lesions. We believe this field has promising application prospects.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.