Preliminary Study of a Ninj1‐Loaded Bimodal Ultrasound/NIR Fluorescence Targeted Molecular Probe for Diagnosing Early‐Stage Inflammation in Coronary Microvascular Dysfunction

ABSTRACT

Coronary microvascular dysfunction (CMD) exhibits a high prevalence and is associated with adverse clinical outcomes, underscoring the critical importance of early detection. Early identification of CMD can significantly improve patient prognosis. This study aims to provide a novel strategy for the precise diagnosis of CMD during its early inflammatory phase. Focusing on this key inflammatory stage in the pathological progression, we sought to identify stage‐specific molecular biomarkers. Through proteomic screening, we identified Nerve Injury‐Induced Protein 1 (Ninj1). During the inflammatory response, Ninj1 promotes leukocyte migration and macrophage transendothelial migration, thereby influencing the trafficking and distribution of inflammatory cells. IR780 is a novel near‐infrared (NIR) fluorescent agent characterized by excellent photostability and low toxicity. Loading IR780 onto nanoparticles enhances its in vivo biocompatibility and photostability while prolonging its circulation time. This project proposes a dual‐modal molecular imaging probe targeting Ninj1 and loaded with IR780, which integrates NIR fluorescence and ultrasound imaging capabilities. This probe is designed to enable the early screening of CMD.

Coronary microvascular dysfunction (CMD) targeting remains a challenge for precise diagnosis. This work presents a dual‐modal nanoprobe (T‐IR780‐NBs) that combines ultrasound contrast with near‐infrared fluorescence. This technology utilizes proteomics‐derived antibodies that specifically localize to inflamed and injured cardiac tissue, enabling precise visualization of inflammatory myocardial damage. It provides a highly promising tool for the dual‐modal diagnosis of CMD.

1. Introduction

Coronary microvascular dysfunction (CMD) arises from functional or structural abnormalities within the coronary microcirculation, leading to reduced coronary blood flow, diminished myocardial perfusion, and subsequent myocardial ischemia, heart failure, and other adverse cardiovascular events. Its presence is strongly associated with poor cardiovascular prognosis [1, 2, 3]. Core pathological features of CMD include microvascular endothelial injury, inflammatory response, and perfusion abnormalities. CMD represents a dynamically evolving process. Although a globally unified staging standard is currently lacking, it is typically categorized based on pathophysiological progression and clinical research into the endothelial impairment phase, inflammatory response phase, and microthrombotic phase [4, 5, 6, 7, 8]. The inflammatory response phase constitutes one of the primary early alterations. Inflammation in CMD exacerbates tissue damage and contributes to adverse cardiac remodeling. Timely diagnosis and early intervention during the inflammatory phase can effectively halt disease progression. Exploring inflammatory biomarkers holds significant clinical value for the early detection of the inflammatory phase and addresses the current limitations in clinical diagnostic capabilities for CMD. In recent years, research into the pathogenesis of acute myocardial infarction has increasingly focused on associated genes, proteins, and metabolic pathways [9, 10, 11, 12, 13]. This study employed 4D label‐free quantitative proteomics technology to investigate differential protein expression in cardiac tissues at distinct time points (1, 3, and 7 days) post‐myocardial infarction, using sham‐operated mouse cardiac tissue as a control. We conducted a detailed screening for molecular markers exhibiting specificity or high expression during different phases of CMD. Notably, nerve injury‐induced protein 1 (Ninj1) was found to be significantly upregulated in the CMD 3‐day group, suggesting its potential as a novel target associated with the inflammatory response. Literature reports indicate that Ninj1 is a cell surface adhesion molecule. Current research implicates Ninj1 in the inflammatory processes of various diseases, such as atherosclerosis and diabetic vasculopathy, where it participates in critical stages of inflammatory regulation [14, 15, 16, 17, 18]. However, its role and application in CMD remain unreported. Through preliminary screening and validation via Western blotting (WB) and quantitative real‐time PCR (qPCR), we confirmed that Ninj1 is highly expressed in the CMD 3‐day group (inflammatory response phase) and in inflammation‐injured human umbilical vein endothelial cells (HUVEC), consistent with the proteomics results. Consequently, Ninj1 was selected as the target for ultrasound molecular imaging to enable detection during the early inflammatory phase of CMD. Furthermore, achieving precise localization and in vivo visualization of early CMD lesions remains challenging due to the limitations of single imaging modalities and the lack of specific molecular probes. Advanced imaging techniques currently applied for the clinical screening and diagnosis of CMD include ultrasound imaging, magnetic resonance imaging (MRI), single‐photon emission computed tomography (SPECT), and positron emission tomography (PET) [19, 20, 21, 22, 23, 24, 25]. Near‐infrared (NIR) fluorescence imaging, recognized for its high sensitivity, is also widely utilized in fundamental CMD research [26, 27, 28, 29]. Therefore, we incorporated the NIR fluorescent dye IR780 into microbubbles, thereby enabling dual‐modal diagnostic imaging combining ultrasound and NIR fluorescence.

Based on the above, we constructed Ninj1‐targeted bimodal microbubbles (designated as Ninj1‐IR780‐NBs, abbreviated as T‐IR780‐NBs) as a novel probe to achieve dual‐modal molecular imaging combining NIR fluorescence and ultrasound. This platform is designed for diagnosing CMD and subsequent therapeutic efficacy assessment in mouse models, with a schematic diagram shown in Figure 1. The Ninj1‐IR780‐NBs imaging probe primarily consists of: Ninj1 (targeting moiety), IR780 (NIR fluorophore), and Nanobubbles (ultrasound contrast agent). This probe exhibits the following significant advantages: (a) High Biocompatibility: Demonstrates excellent biosafety in both in vitro and in vivo evaluations. (b) Dual‐Modal Imaging: Combines high‐sensitivity NIR fluorescence imaging with ultrasound imaging, offering complementary capabilities. This synergy enhances the capture of both the location and structure of inflammation‐damaged sites, enabling the visualization of molecular information. (c) Multifunctional Application: Ninj1‐IR780‐NBs can be utilized for both the diagnosis of CMD and the assessment of therapeutic efficacy for drug‐based treatments. The developed Ninj1‐IR780‐NBs offer a novel technical platform for the early diagnosis of CMD.

FIGURE 1.

Schematic representation of dual‐mode molecular imaging and efficacy evaluation of CMD using Ninj1‐IR780‐NBs.

2. Results and Discussion

2.1. Target Screening by Proteomics

We first performed principal component analysis (PCA) to assess the statistical consistency of protein quantification across biological or technical replicates. The PCA results for all samples are presented in Figure 2A, in which the degree of clustering among the samples represents the degree of difference among the samples. A heatmap was generated by ranking differentially expressed proteins based on their fold changes, visualizing both upregulated and downregulated proteins (Figure 2B). The relative quantitative ratio between the model and sham groups was defined as the fold change (FC). Proteins with p‐value < 0.05 and an FC > 2.5 were considered significantly upregulated, while those with an FC < 1/2.5 were considered significantly downregulated. All differentially expressed proteins between the 3‐day model and sham groups are summarized in the volcano plot (Figure 2C). Notably, Ninj1 appeared as an outlier in the top right quadrant, displaying a high fold change and thereby confirming its significant upregulation in our dataset.

FIGURE 2.

Proteomics analysis. (A) Principal Component Analysis (PCA) plots of protein quantification for all samples. The degree of clustering among the samples represents the degree of difference among the samples. (B) Heatmap of protein expression comparing the CMD group versus the Sham group. Columns represent individual samples, rows represent individual proteins, and color intensity indicates the expression level of each protein across samples. (C) Volcano plot of differentially expressed proteins between the CMD 3‐day and Sham groups. Red dots denote significantly upregulated proteins, blue dots denote significantly downregulated proteins, and grey dots represent proteins with no significant change in expression. (D) Gene Ontology (GO) enrichment analysis of the filtered differentially expressed proteins. A bar graph displays the number of proteins involved in significantly enriched biological pathways. (E) Protein–protein interaction (PPI) network analysis of the filtered proteins (interaction confidence score: 0.4). n = 3.

Given that our previous studies and existing literature indicate that 3 days post CMD is primarily characterized by inflammatory changes, we filtered differentially expressed proteins using thresholds of FC > 2.5 and p < 0.01. The resulting proteins were subjected to protein–protein interaction (PPI) network analysis (confidence score: 0.4; Figure 2E). As shown, Ninj1 functions as a hub protein within the network, indicating high connectivity. The PPI network revealed that Ninj1 closely interacts with proteins involved in proteasomal subunits (e.g., Psmb4, Psme4), endoplasmic reticulum stress (e.g., Pdia4, Hspa5), and classic inflammatory and apoptotic signaling molecules (e.g., NF‐κB, Gsk3b). Importantly, several of these interacting partners have been reported to play critical roles in myocardial ischemia. For instance, Psmb4 can inhibit cardiomyocyte apoptosis by activating the NF‐κB pathway [30], while endoplasmic reticulum stress is a key trigger of post‐ischemic cell injury [31].

To further investigate the functional relevance of Ninj1, we performed Gene Ontology (GO) enrichment analysis on genes co‐expressed with Ninj1. The resulting bar graph shows that the enriched genes are predominantly clustered in cell adhesion(239 genes), positive regulation (158 genes), and biological adhesion (242 genes), as shown in Figure 2D. Key steps in cardiovascular inflammation—monocyte rolling, firm adhesion, and trans‐endothelial migration—all rely on cell adhesion and vascular structural remodeling. This enrichment pattern suggests that Ninj1 may participate in cardiovascular inflammatory processes such as myocardial ischemia and atherosclerosis by modulating adhesion signaling and vascular development pathways.

Based on the above analyses, we preliminarily identified differentially expressed proteins closely associated with inflammatory alterations in CMD. Among them, Ninj1 showed prominent differential expression and was linked to key processes such as endothelial inflammation, which aligns well with the pathophysiological mechanisms of CMD. Further PPI network analysis indicated that Ninj1 interacts extensively with proteins related to inflammation and apoptosis, implying that Ninj1 may play an important role in the inflammatory response of CMD. To validate this hypothesis, subsequent cellular and animal experiments were conducted to explore the mechanisms and potential value of Ninj1 in CMD.

2.2. Validation of Screening Targets

2.2.1. Validation of Cell Model

1) Establishment of damaged HUVEC model: under the light field observation of an inverted fluorescence microscope, with the increase of TNF‐α concentration, the degree of HUVEC injury was gradually aggravated, and there was no significant difference in the degree of HUVEC injury between the 10 ng/mL group and the 20 ng/mL group. The results of flow cytometry showed that the expression of CD44 on the surface of damaged HUVEC was significantly increased when the concentration of TNF‐α reached 10 ng/mL. The expression level of CD44 in the 10 ng/mL group was (78.30 ± 1.10)%, and the expression level of CD44 in the 20 ng/mL group was (80.80 ± 2.70)%, p > 0.05, and the difference was not statistically significant (Figure 3A). Ten (nanograms per milliliter) was selected as the treatment concentration of HUVEC in the co‐culture model.

FIGURE 3.

Expression of the target protein in cell models. (A) Flow cytometric analysis of CD44 expression on the surface of injured HUVEC and the quantification of CD44 expression levels. (B) Cellular immunofluorescence staining of Ninj1 protein expression on the surface of injured HUVEC versus normal HUVEC. Scale bar: 75 µm. (C) RT‐qPCR analysis of Ninj1 mRNA expression in injured HUVEC compared to normal HUVEC. (D,E) Western blot (WB) analysis of Ninj1 protein expression in inflamed/injured HUVEC versus normal HUVEC (n = 3). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, and ns is no significant difference.

2) Ninj1 Expression in the Cell Model: Immunofluorescence analysis revealed significantly higher expression of Ninj1 protein in TNF‐α treated HUVEC (Treated group) compared to untreated controls (Control group), as visualized in Figure 3B. This finding was further corroborated by both Western blotting and quantitative real‐time PCR (qPCR) analyses. The qPCR results are presented in Figure 3C. Representative Western blots and quantitative data are shown in Figure 3D,E, respectively. These results collectively demonstrate the upregulation of Ninj1 in the inflammation‐injured HUVEC model. These findings provide preliminary evidence that Ninj1 is closely associated with endothelial responses under inflammatory conditions, thereby offering robust experimental support for further investigating the feasibility of targeting Ninj1 in CMD.

2.2.2. Verification of Animal Models

The CMD mouse model was established by ligation of the left anterior descending coronary artery. Preoperative and postoperative cardiac and masson's trichrome staining analysis confirmed the success of the model construction. As shown in Figure 4, a small amount of myocardial fiber necrosis near the epicardium of the left ventricle of mice in the 1 day model group was observed in the HE staining section, with nuclear pyknosis, deeply stained or fragmented dissolution, cytoplasm fragmentation, dissolution, and light color, and perivascular inflammatory cell infiltration in the epicardium. In the 3‐day model group, there were more patchy necrosis of cardiomyocytes, enhanced cytoplasmic eosinophicity, mild hemorrhage, and inflammatory cell infiltration. A large number of myocardial cells in the left ventricular wall of the 7‐day model group were necrotic in sheets, with nuclear pyknosis and deep staining or lysis. Some of the necrotic cells were dissolved and replaced by proliferative connective tissue, accompanied by more inflammatory cell infiltration. In the masson staining, the heart tissue of the sham group was stained uniform red, the 1 day group showed disorder of myocardial arrangement and uneven myocardial color, the 3 days group showed blue fiber accumulation and most myocardial cells flaky necrosis, and the 7 days group showed a large number of blue fiber accumulation and serious myocardial fibrosis. The cardiac pathological tissue of the model mice in the 3‐day group was mainly inflammatory [23, 24, 25]. Western blot analysis revealed a markedly thicker Ninj1 band in the CMD 3 days group compared with all other groups, and densitometric quantification confirmed a significant increase in protein abundance, mirroring the trend observed for Ninj1 mRNA levels by real‐time PCR. Immunofluorescence staining further showed a robust enhancement of Ninj1 fluorescence within the myocardium of the CMD 3‐day group, whereas only faint signals were detected in sham‐operated and CMD 1‐day animals. Although the signal intensity declined in the CMD 7‐day group relative to the 3‐day time point, it remained higher than that in the sham group (Figure 4D–G). Western Blot, real‐time PCR and immunofluorescence were used to analyze the expression of Ninj1 in the sham‐operated mice, CMD 1 day group, CMD 3 days group and CMD 7 days group, and the results showed that: In these four groups, the relative expression of Ninj1 and Ninj1 mRNA was significantly increased in the CMD 3 days group (p < 0.01), indicating that Ninj1 was highly expressed in the inflammatory stage of CMD.

FIGURE 4.

Pathological staining and related detection results of mice at different pathological progression periods. (A,B) Representative images of H&E staining. (A) and Masson's trichrome staining (B) in the Sham group and the CMD 1‐, 3‐, and 7‐day groups. (C) Immunohistochemical staining for CD31 in mouse models at different time points. (D) Immunofluorescence images showing Ninj1 expression in mouse models across different stages. Note the strongest Ninj1 expression in the CMD 3‐day group. Blue indicates nuclei (DAPI staining). (E) Western blot (WB) bands showing Ninj1 protein expression levels in mouse models at different time points. (F) Quantitative analysis of band intensity (gray value) from WB. (G) RT‐qPCR analysis of Ninj1 mRNA expression levels in mouse models at different time points. Data are displayed as mean ± SD and n = 3. Statistical significances: *p < 0.05, **p < 0.01, ***p < 0.001,****p < 0.0001, and ns is no significant difference.

These findings demonstrate that Ninj1 is markedly up‐regulated during the inflammatory phase of coronary microvascular dysfunction (CMD), and its expression level correlates closely with the severity of myocardial injury and the magnitude of the inflammatory response. To further substantiate the potential of targeting Ninj1 in CMD, we will next systematically evaluate the targeting efficiency and in vitro/in vivo safety profile of the Ninj1‐IR780‐NBs nanoprobes.

2.3. Preparation and Characterization of Ninj1‐IR780‐NBs

IR780‐NBs and Ninj1‐IR780‐NBs were synthesized using the thin‐film hydration method. Ninj1 was conjugated to the IR780‐NBs surface via a biotin‐avidin‐biotinylated antibody bridging strategy. The Ninj1‐IR780‐NBs imaging probe primarily comprises: Ninj1 (targeting moiety), IR780 (NIR dye), and Nanobubbles (ultrasound contrast agent) (Figure 5). Fluorescence imaging offers the advantages of rapid acquisition and high sensitivity but suffers from limited tissue penetration depth. Conversely, ultrasound imaging provides deep tissue penetration and high spatial resolution but has lower sensitivity. Therefore, combining these two modalities enables rapid, highly sensitive, and high‐resolution imaging in the mouse CMD model [21, 22]. The designed material features a core–shell microbubble structure: The core contains C₃F₈ gas to enhance ultrasound imaging. The shell incorporates the IR780 dye for near‐infrared (NIR) fluorescence imaging (Figure 5A). As shown in Figure 5B, the prepared T‐IR780‐NBs solution appeared as a pale green liquid. Following mechanical agitation using an amalgamator for 45 s, it formed a pale green suspension (Figure 5C). Visualization under confocal microscopy and scanning electron microscopy (SEM) confirmed the microbubbles exhibited a regular spherical morphology (Figure 5D,E). The mean diameter of the IR780‐NBs was (495.30 ± 1.72) nm. After conjugation with Ninj1, the diameter of the Ninj1‐IR780‐NBs increased to (579.90 ± 1.55) nm (Figure 5G,H; detailed parameters are provided in Figure 5K). Owing to a size approximately one‐tenth that of clinically approved microbubbles, T‐IR780‐NBs can remain in the bloodstream for prolonged periods and penetrate compromised physiological barriers [32].

FIGURE 5.

Comprehensive characterization of T‐IR780‐NBs. (A) Schematic illustration of the composition and architecture of Ninj1‐IR780‐NBs. (B,C) Macroscopic appearance: (B) pale‐green Ninj1‐IR780‐NBs at rest; (C) homogeneous dispersion after gentle agitation. (D) Confocal laser‐scanning micrograph of Ninj1‐IR780‐NBs. (E) Representative scanning electron micrograph revealing spherical morphology. (F) The connection of NB nanoparticles and Ninj1 antibody was observed by a fluorescence microscope. For Dil‐labeled nanoparticles and AF488‐labeled Ninj1 antibody, yellow fluorescence indicated that the antibody was successfully bound to the surface of NBs. (G,H) Size‐distribution histograms of IR780‐NBs (G) and Ninj1‐IR780‐NBs (H) obtained by dynamic light scattering. (I) Variation in the size of Ninj1‐IR780‐NBs measured over 4 h at room temperature (25°C). (J) Long‐term stability at 4°C for 1–24 h. (K) The key parameters of the primary microbubbles. (L,M) Normalized fluorescence spectra: excitation at 780 nm (L) and emission at 810 nm (M) for free IR780, IR780‐NBs, and Ninj1‐IR780‐NBs. (N) Normalized UV—Vis–NIR absorption spectra of free IR780, IR780‐NBS, and Ninj1‐IR780‐NBs. (O,P) Antibody binding efficiency increased with the increase of antibody amount. When the amount of antibody reached 30ul, the binding rate reached saturation. (Q,R) Representative ultrasound contrast enhancement images of microbubbles in vitro. (S,T) Representative near‐infrared (NIR) fluorescence images of microbubbles in vitro. Data are displayed as mean ± SD and n = 3. Statistical significances: *p < 0.05, **p < 0.01, ***p < 0.001,****p < 0.0001, and ns is no significant difference.

To verify the successful conjugation of the Ninj1 antibody, Ninj1‐IR780‐NBs exhibited distinct yellow fluorescence under the fluorescence microscope, whereas IR780‐NBs alone did not, indicating that AF488‐labeled anti‐Ninj1 antibody had been efficiently immobilized onto the surface of IR780‐NBs. Additionally, fluorescence spectrophotometry revealed characteristic excitation and emission peaks at 780 and 810 nm for both IR780‐NBs and Ninj1‐IR780‐NBs, which were identical to those of free IR780 (Figure 5L,M). UV–Vis–NIR spectroscopy further demonstrated that all three formulations—free IR780, IR780‐NBs, and Ninj1‐IR780‐NBs—displayed superimposable absorption spectra (Figure 5N). Collectively, these data confirm that IR780 dye was successfully incorporated into the microbubble shell and retained its intrinsic NIR optical properties.

The stability of the prepared Ninj1‐IR780‐NBs system was systematically evaluated. We evaluated the stability of the prepared Ninj1‐IR780‐NBs. The particle size of Ninj1‐IR780‐NBs increased slightly from 0 min (584.30 ± 13.61 nm) to 4 h (610.40 ± 8.87 nm) at room temperature. The particle size of Ninj1‐IR780‐NBs decreased slightly from 0 min (584.30 ± 13.61 nm) to 10 h (573.7 ± 46.18 nm) at 4°C, but the difference was not statistically significant, indicating that Ninj1‐IR780‐NBs had good stability at room temperature for 4 h and 4°C for 10 h, providing a crucial foundation for their potential biomedical applications (Figure 5I,J).

In order to optimize the antibody carrying rate of IR780‐NBs, the binding rate of IR780‐NBS with different doses of Ninj1 antibody was further detected by flow cytometry. As shown in Figure 5O,P, when the amount of antibody was 5, 10, 20, 30, 40 µL (1: 100), the binding rates to 100 µL (1 × 10⁸) nanobubbles were 19.70% ± 1.57%, 36.63% ± 0.67%, 69.70% ± 1.06%, 81.27% ± 0.81% and 82.67% ± 0.64%, respectively. This indicates that the binding rate of the antibody to the microbubbles increases continuously with the increase of the antibody dose. However, when the amount of antibody was 30 µL, the binding rate of antibody to microbubbles was not significantly different from that of 40 µL. The results of this study showed that when the amount of antibody reached 30 µL, the binding rate reached saturation. The optimal antibody binding rate of NBs was 81.27% ± 0.81%, and the optimal ratio of antibody and nanobubbles was 1:3.3. By optimizing the optimal antibody carrying rate of targeted nano‐microbubbles, it will be more helpful to play its targeting performance and save the amount of antibody.

The ultrasound signal intensity of optimized Ninj1‐IR780‐NBs was enhanced with increasing microbubble concentration, as shown in Figure 5Q,R. The NIR fluorescence signal of Ninj1‐IR780‐NBs was enhanced with the increase of microbubble concentration. When the microbubble concentration was 50 times diluted, there was no significant difference in fluorescence intensity between 0 and 30 min, 0 and 60 min. When the microbubble concentration was 100 times diluted, there was no significant difference in fluorescence intensity between 0, 30, and 60 min. This indicates that the fluorescence intensity is stable when microbubbles are diluted 50 times and 100 times, as shown in Figure 5S,T.

2.4. Specific Targeting of Ninj1‐IR780‐NBs Probe In Vitro

2.4.1. Static Specific Adhesion Experiments of Targeted Microbubbles

To determine the binding specificity of Ninj1‐IR780‐NBs, we further investigated the binding specificity of Ninj1‐IR780‐NBs, IgG‐IR780‐NBs, and IR780‐NBs to impaired inflammation endothelial cells under bright‐field microscopy. The results (Figure 6A) showed that the binding specificity of Ninj1‐IR780‐NBs to impaired inflammation endothelial cells was significantly higher than that of IgG‐IR780‐NBs and IR780‐NBs, and Ninj1‐IR780‐NBs and IR780‐NBs did not adhere to the Impaired inflammation endothelial cells. The number of nanobubbles around each endothelial cell in the Ninj1‐IR780‐NBs group (15 ± 0.82) was significantly higher than that in the IR780‐NBs group (3.75 ± 0.96) and the IgG‐IR780‐NBs group (5 ± 0.82) (****: p < 0.0001).

FIGURE 6.

In vitro static adhesion of T‐IR780‐NBs to endothelial cells. (A) Representative bright‐field micrographs showing endothelial cell‐bound nanobubbles and (B) quantitative analysis of the number of adherent nanobubbles per cell. (C) Representative confocal laser scanning microscopy (CLSM) images of endothelial cell‐bound nanobubbles (blue: DAPI‐stained nuclei; red: IR780‐labeled nanobubbles; scale bar: 75 µm). (D) Flow cytometric analysis further demonstrates that endothelial cell attachment is significantly greater for T‐IR780‐NBs compared to IgG‐IR780‐NBs or IR780‐NBs. (E) Quantitative data derived from the analysis of (D). Data are displayed as mean ± SD and n = 3. Statistical significances: *p < 0.05, **p < 0.01, ***p < 0.001,****p < 0.0001, and ns is no significant difference.

Subsequently, the targeting capability of different nanobubble ultrasound contrast agents toward both inflamed inflammation and normal endothelial cells was evaluated using confocal laser scanning microscopy (CLSM). As shown in Figure 6C, CLSM results showed that the IgG‐IR780‐NBs group and IR780‐NBs group could show a little red fluorescence around the inflamed inflammation endothelial cells, while the Ninj1‐IR780‐NBs could show obvious red fluorescence around the inflamed inflammation endothelial cells. The range of red fluorescence was the largest. The Ninj1‐IR780‐NBs group and IR780‐NBs group had no obvious fluorescence around the undamaged endothelial cells, while the PBS group had no obvious fluorescence around the inflamed inflammation endothelial cells. Then, we quantitatively detected the targeted binding rate of T‐IR780‐NBs, IgG‐IR780‐NBs, and IR780‐NBs to damaged HUVEC by flow cytometry. As presented in Figure 6D, the binding rate of IR780‐NBs to inflamed inflammation HUVEC was 23.77% ± 1.33%. The binding rate of IgG‐IR780‐NBs to damaged HUVEC was 25.535 ± 0.91%, the binding rate of T‐IR780‐NBs to damaged HUVEC was 49.90% ± 0.96%, and the binding rate of T‐IR780‐NBs to normal HUVEC was 21.13% ± 0.25%. The binding rate of IR780‐NBs to normal HUVEC was 19.30% ± 0.40%. The binding rate of Ninj1‐IR780‐NBs to damaged HUVEC was higher than that of IR780‐NBs and IgG‐IR780‐NBs (P < 0.001).

2.4.2. Dynamic Specific Adhesion Experiments of Targeted Microbubbles

To further test the dynamic adhesion ability of the bubbles, a parallel plate flow chamber was used to create different flow environments, and HUVEC cells intervened with TNF‐α were seeded at the bottom of the flow chamber. Under light field (Figure S1A) and fluorescence microscope (Figure S1B), the accumulation of Ninj1‐IR780‐NBs was always higher than that of IR780‐NBs and IgG‐IR780‐NBs under all shear stresses (p < 0.05), as shown in Figure S1C. The accumulation of Ninj1‐IR780‐NBs tended to decrease with time, as shown in Figure S1D. According to statistics, there was no significant difference between the groups at 1–4 min of flow, suggesting that the bubble binding was relatively stable. There was a difference in the number of microbubbles binding at 5, 1, and 2 min of flow, but the overall number was not low. The following figure shows the dynamic adhesion diagram of Ninj1‐IR780‐NBs in the injured endothelium. When τ = 1, Ninj1‐IR780‐NBs adhered most, and when τ = 12, a few microbubbles remained. When PBS was used to infuse the injured HUVEC with τ = 12 for 10 s, a few microbubbles could still be observed, indicating that T‐IR780‐NBs had a better adhesion ability to injured HUVEC.

In summary, the present study validated the specific targeting capability of Ninj1‐IR780‐NBs toward inflamed inflammation endothelial cells using multiple complementary experimental approaches.

2.5. T‐IR780‐NBs were used for Targeted NIR Fluorescence Imaging at Inflamed Damaged Myocardium In Vivo

Demonstrating excellent in vitro targeting capability of T‐IR780‐NBs, we further evaluated their targeting ability toward inflammation‐damaged sites in vivo. The targeting of T‐IR780‐NBs to inflammatory sites was assessed using NIR fluorescence imaging in the CMD mouse model. A cohort of 30 mice was randomly assigned to six experimental groups (n = 5 per group) and treated as follows: (a) CMD model mice + T‐IR780‐NBs; (b) CMD model mice + IgG‐IR780‐NBs; (c) CMD model mice + IR780‐NBs; (d) Sham‐operated mice + T‐IR780‐NBs; (e)Sham‐operated mice + IgG‐IR780‐NBs; (f) Sham‐operated mice + IR780‐NBs (Injection dose: 0.3 mg/kg IR780, concentration: ∼6 µm). Bright NIR fluorescence signals were detected specifically within the cardiac region of CMD model mice as early as 0.5 h post‐intravenous injection of T‐IR780‐NBs (Group a). In contrast, only minimal fluorescence was observed in the cardiac areas of mice from all other control groups (b–f). Quantitative analysis revealed that the average signal‐to‐background ratio (SBR) in Group (a) at 4 h post‐injection was 82.2. This value was 2.30‐fold, 1.94‐fold, 3.89‐fold, 5.12‐fold, 4.49‐fold higher than the SBR values measured in Groups (b), (c), (d), (e), and (f) respectively (Figure 7B). Following injection from 1 to 12 h, NIR fluorescence in the cardiac region of group (a) progressively intensified. After 12 h, the fluorescence signal intensity gradually diminished over time but remained detectable at the 72‐hour mark, indicating signal retention for up to 72 h. This provides a sufficient time window for localizing myocardial injury tissue (Figure 7). Ex vivo cardiac imaging following heart extraction corroborated these findings.

FIGURE 7.

In vivo near‐infrared (NIR) fluorescence imaging of T‐IR780‐NBs targeting inflamed myocardium. (A,B) NIR fluorescence imaging of CMD and sham‐operated mice following vein injection of T‐IR780‐NBs, IgG‐IR780‐NBs, or IR780‐NBs, along with quantitative comparison of fluorescence intensities. (C,D) Time‐course NIR fluorescence imaging of CMD model mice at various time intervals (1, 4, 8, 12, 24, 48, and 72 h) post‐intravenous injection of T‐IR780‐NBs. (E) Ex vivo NIR fluorescence images of excised hearts from each experimental group, and (G) ex vivo images of hearts from the CMD model group at 2, 12, and 24 h after T‐IR780‐NBs injection. (F,H) Quantitative analysis of the NIR fluorescence signal intensities corresponding to the excised heart tissues shown in panels E and G, respectively. Scale bars: 1 cm. Data are displayed as mean ± SD and n = 5. Statistical significances: *p < 0.05, **p < 0.01, ***p < 0.001,****p < 0.0001, and ns is no significant difference.

2.6. T‐IR780‐NBs Was Used for Targeted Ultrasound Imaging at Inflamed Damaged Myocardium In Vivo

Complementing the highly sensitive NIR fluorescence imaging, ultrasound imaging offers the advantages of excellent soft tissue penetration depth and real‐time imaging capabilities. Molecular ultrasound imaging utilizes targeted contrast agents that are selectively enriched at the disease site via specific ligands, thereby enhancing imaging of the pathological region. The combination of ultrasound and NIR fluorescence imaging holds promise for achieving complementary molecular imaging of coronary microvascular dysfunction (CMD). To this end, we investigated the in vivo targeting efficacy of different nanobubble contrast agents in both CMD and sham‐operated mouse models using a high‐frequency small‐animal ultrasound imaging system (Figure 8A). The animal grouping scheme was identical to that described in Section 2.5. Following intravenous administration of the probes (T‐IR780‐NBs, IgG‐IR780‐NBs, or IR780‐NBs, Injection dose: 0.3 mg/kg IR780, concentration: ∼6 µm), contrast‐enhanced ultrasound (CEUS) imaging was performed to visualize the cardiac ventricles and myocardium. The efficacy of contrast enhancement and the temporal dynamics of the ultrasound signal intensity were comparatively analyzed across groups. The CEUS results indicated that nanobubbles from all groups appeared in the cardiac region within 5 s, rapidly filling the cardiac chambers and generating intense echo signals, confirming their general suitability for cardiac imaging (Figure 8D). The time‐intensity curves exhibited a characteristic parabolic profile (rise followed by decay). In CMD mice injected with T‐IR780‐NBs, the signal intensity in the anterior ventricular wall was higher than in all other groups, peaking at approximately 2 s. Although the intensity subsequently declined, it remained at a relatively high level, with noticeable contrast enhancement still observable at around 120 s post‐injection. Quantitative analysis revealed that the peak intensity in the CMD + T‐IR780‐NBs group was approximately 1.8‐fold and 3.6‐fold greater than that in the CMD + IgG‐IR780‐NBs and CMD + IR780‐NBs groups, respectively. In contrast, the anterior wall intensities in the Sham + T‐IR780‐NBs and Sham + IgG‐IR780‐NBs groups showed no significant difference and were roughly half that of the CMD + IgG‐IR780‐NBs group. Notably, during the 2–120 s observation window, the CMD + T‐IR780‐NBs group demonstrated the most efficient imaging performance Figure 8B), maintaining significant contrast enhancement in the cardiac region at 120 s. Furthermore, within this group, the echo intensity in the inflamed zone increased more rapidly and decayed more slowly compared to the remote normal zone. This imaging profile—characterized by rapid uptake and slow clearance—is likely attributable to the Ninj1 antibodies conjugated on the surface of T‐IR780‐NBs. Finally, the excised heart tissues collected post‐injection were subjected to immunofluorescence staining. As shown in Figure 8F, Ninj1 (green) was distributed within the cardiac injury areas. The fluorescence signal from T‐IR780‐NBs (red) co‐localized with the anti‐Ninj1 antibody signal (green), whereas no distinct fluorescence was observed in the injury zones of hearts from non‐targeted treatment groups. The concordance between the in vivo and ex vivo imaging results in CMD mouse hearts demonstrates that T‐IR780‐NBs enable precise spatial targeting of inflammatory‐injured sites within the cardiac tissue via the microbubble platform.

FIGURE 8.

In vivo ultrasound molecular imaging. (A) Schematic diagram illustrating the positioning for cardiac ultrasound imaging in mice, showing a long‐axis view of the left ventricle. LV, left ventricle; LVAW, left ventricular anterior wall; LVPW, left ventricular posterior wall. (B) Quantitative analysis of cardiac contrast enhancement across different experimental groups at various time points post‐injection under contrast‐enhanced ultrasound (CEUS) mode. (C) Schematic illustration of T‐IR780‐NBs targeting the injured cardiac tissue for in vivo ultrasound molecular imaging. (D) Ultrasound molecular images of the cardiac injury region in mice following intravenous injection of the respective nanobubble formulations (scale bar: 2 mm). The red and yellow dotted lines in the left ventricular long‐axis view outline the cardiac contour, and the white dotted line delineates the cardiac injury area. (E) Time‐intensity curves (TICs) depicting echo intensity over time in the cardiac region for each experimental group. (F) Immunofluorescence imaging of mouse heart tissue after T‐IR780‐NBs injection. Red: IR780 signal; Green: Ninj1; Blue: DAPI‐stained nuclei. Scale bar: 50 µm. Data are displayed as mean ± SD and n = 5. Statistical significances: *p < 0.05, **p < 0.01, ***p < 0.001,****p < 0.0001, and ns is no significant difference.

2.7. Dual‐Modality Molecular Imaging to Evaluate the Efficacy of Drug Therapy for Myocardial Injury

Atorvastatin is a statin that is widely used to lower cholesterol levels and has anti‐inflammatory and plaque‐stabilizing effects. In the treatment of CMD, atorvastatin exerts anti‐inflammatory effects through a variety of mechanisms, including inhibiting the production of inflammatory cytokines and improving endothelial function [33, 34]. Atorvastatin was used to evaluate the efficacy of the CMD model mice. Atorvastatin (dose: 10 mg/kg, frequency: once daily) was given orally for 28 days according to the routine clinical treatment cycle. The treated mice were then subjected to tail vein injection of T‐IR780‐NBs (dose: 0.3 mg/kg, IR780, concentration: ∼6 µm) for NIR fluorescence imaging and ultrasound imaging to evaluate the treatment effect. After 4  weeks of drug treatment, the CEUS signal intensity at the myocardial injury decreased from 130 to 60 db (Figure 9B). The NIR fluorescence signal in the mouse heart region was also reduced by 1.8‐fold relative to that before treatment (Figure 9C). These results suggest that atorvastatin may have reduced the degree of inflammation at the site of myocardial injury. Cardiac tissues were collected from CMD mice after 4 weeks of treatment for HE staining analysis and correlation analysis of Ninj1 expression. HE results showed that the inflammatory cell infiltration was significantly relieved after atorvastatin treatment, WB and immunofluorescence showed that the expression of Ninj1 decreased after treatment (Figures 9D,H,J), which was consistent with the results of dual‐modality imaging, and proved that T‐IR780‐NBs can be used for fluorescence/ultrasound dual‐modality molecular imaging to evaluate the efficacy of atorvastatin in the treatment of CMD.

FIGURE 9.

Therapeutic efficacy evaluation of the NIR fluorescence/ultrasound dual‐modal targeted molecular imaging probe T‐IR780‐NBs in the mouse model of CMD. (US, ultrasound; FL, fluorescence). (A) Timeline schematic illustrating CMD model establishment, therapeutic administration, and imaging evaluation. (B) Ultrasound imaging of myocardial injury in mice. Scale bar: 2 mm. (C) In vivo near‐infrared (NIR) fluorescence molecular imaging of mice. Scale bar: 1 cm. D) H&E staining analysis before and after atorvastatin treatment. (E–G) Quantitative analysis of the ultrasound (US), fluorescence (FL), and H&E staining results, respectively. (H) Immunofluorescence staining of Ninj1 expression in injured cardiac tissue before and after treatment. Scale bar: 25 µm. (I) Statistical quantification of Ninj1 immunofluorescence intensity. (J,K) Western blot (WB) analysis of Ninj1 protein expression in cardiac tissues before and after treatment (J) and the corresponding grayscale quantification of protein bands (K). Data are displayed as mean ± SD and n = 5. Statistical significances: *p < 0.05, **p < 0.01, ***p < 0.001,****p < 0.0001, and ns is no significant difference.

2.8. Biocompatibility of T‐IR780‐NBs

The high biocompatibility of T‐IR780‐NBs is a prerequisite for their clinical translation. Optical microscopy revealed no morphological alterations in HUVEC incubated with T‐IR780‐NBs for either 24 or 48 h. Furthermore, cell viability assays demonstrated that after incubation with varying concentrations of T‐IR780‐NBs for 24 or 48 h, cell survival rates consistently remained above 90% at microbubble concentrations ranging from 10³ to 10⁷/mL, and above 65% even at the high concentration of 10⁸/mL. This confirms the low cytotoxicity of the microbubbles (Figure 10B). Subsequently, we further evaluated the in vivo biosafety profile of T‐IR780‐NBs using C57BL/6J mice. Blood samples were collected at different time points (days 1, 3, 7, and 14) for both biochemical analysis and routine hematological examination. Key indicators, including: Liver function: Alanine Aminotransferase (ALT), Aspartate Aminotransferase (AST), Total Bilirubin (TBIL); Kidney function: Urea (UREA), Creatinine (CREA), Blood Urea Nitrogen (BUN); Cardiac function: Creatine Kinase (CK), Lactate Dehydrogenase (LDH); Hematological parameters: white blood cell count (WBC), Mean Corpuscular Volume (MCV), Monocyte count (Mon), Neutrophil count (NE). all fell within normal reference ranges and were consistent with the control group (Figure 10C–N). These results indicate that T‐IR780‐NBs exhibit low toxicity within the tested concentration range. Finally, major organs (heart, liver, spleen, lung, kidney) were harvested for histopathological evaluation via H&E staining following euthanasia after T‐IR780‐NBs injection. The results showed no significant organ damage or inflammatory lesions in the T‐IR780‐NBs‐treated group compared to the control group (Figure 10A). Collectively, these findings demonstrate that T‐IR780‐NBs possess excellent biocompatibility both in vitro and in vivo, highlighting their significant potential for clinical translation and offering a novel perspective for the diagnosis of CMD.

FIGURE 10.

Biocompatibility and biosafety assessment of T‐IR780‐NBs. (A) Representative H&E staining analysis of major organs (heart, liver, spleen, lung, and kidney) harvested from mice 24 h post‐injection of either PBS or T‐IR780‐NBs (dose: 0.5 mg/kg IR780), scale bar: 50 µm. (B) Cell viability of HUVEC incubated with varying concentrations of T‐IR780‐NBs for 24, 48, and 72 h. (C–N) Hematological and biochemical analysis of the control group (Day 0, intravenous PBS injection) and the experimental groups (Days 1, 3, 7, and 14, intravenous T‐IR780‐NBs injection). Parameters analyzed include: White Blood Cell count (WBC), Mean Corpuscular Volume (MCV), Monocyte count (Mon#), Neutrophil count (NE), Alanine Aminotransferase (ALT), Aspartate Aminotransferase (AST), Total Bilirubin (TBIL), Urea (UREA), Creatinine (CREA), Blood Urea Nitrogen (BUN), Creatine Kinase (CK), and Lactate Dehydrogenase (LDH). All error bars represent mean ± SD (n = 5 per group).

Our study is not without its limitations. First, this work constitutes a preliminary proof‐of‐concept investigation. Subsequent studies should systematically optimize the imaging parameters for T‐IR780‐NBs to achieve maximal diagnostic efficacy. Second, while the biotin‐avidin system enables robust conjugation of targeting ligands, its intrinsic immunogenicity may pose a challenge for clinical translation. Future research should focus on developing preparation protocols that are more translatable to clinical settings, along with longitudinal monitoring at more frequent time points. This will generate a more comprehensive dataset to inform the potential clinical application of this strategy for CMD.

3. Conclusion

Coronary microvascular dysfunction (CMD) is well established as a significant contributor to poor cardiovascular outcomes. Current quantitative techniques for functional assessment, comprising both invasive and non‐invasive modalities, are not widely adopted in clinical settings. This limited adoption stems from their substantial cost, procedural complexity, and insufficient diagnostic accuracy. Concurrently, the field of medical imaging has witnessed rapid progress in dual‐modal nanobubbles (NBs) as contrast agents, offering a promising avenue to address these diagnostic shortcomings. To capitalize on this opportunity, we engineered T‐IR780‐NBs, a targeted microbubble contrast agent designed to specifically bind to inflammatory markers during the CMD progression, with the aim of facilitating its early‐stage detection. Incorporating IR780 into the microbubble lipid shell endowed the microbubbles with NIR fluorescence imaging capability. Furthermore, T‐IR780‐NBs demonstrated active targeting toward damaged sites during the CMD inflammatory phase, as observed in both an in vitro model of inflammation‐injured endothelial cells and an in vivo mouse model of CMD. We successfully utilized the prepared T‐IR780‐NBs to achieve high‐sensitivity NIR fluorescence imaging alongside real‐time, high‐resolution ultrasound molecular imaging in CMD mice. Moreover, we successfully employed this dual‐modal imaging technology to conduct a dual‐modal therapeutic assessment of the efficacy of Atorvastatin in treating ischemic heart disease. The prepared T‐IR780‐NBs also demonstrated favorable biocompatibility in both in vitro and in vivo settings. This actively targeted probe, combined with a dual‐modality imaging strategy, not only offers a novel approach for the early diagnosis of CMD but may also open new research directions and provide fresh perspectives for the clinical diagnosis and therapeutic evaluation of other vascular injury‐related diseases.

4. Experimental Section

4.1. Materials

1,2‐Distearoyl‐sn‐glycero‐3‐phosphatidylcholine (DSPC), 1,2‐distearoyl‐sn‐glycero‐3‐phosphoethanolamine‐N‐[methoxy(polyethyleneglycol)‐2000] (DSPE‐PEG2000), and 1,2‐distearoyl‐sn‐glycero‐3‐phosphoethanolamine‐N‐[biotinyl(polyethyleneglycol)‐2000] (DSPE‐PEG2000‐biotin) was purchased from Avanti Polar Lipids Inc. (Alabaster, AL, USA). Biotinylated AFF488‐labeled Ninj1 antibody and IgG isotype control antibody were synthesized by Bioss (Beijing, China). IR780 was obtained from Sigma Aldrich. Perfluoropropane (C₃F₈) gas was obtained from NEWRADAR Company (Wuhan, Hubei, China) and avidin from Sigma‐Aldrich (St Louis, MO, USA). Cell culture components were obtained from Qida Biological Technology (Shanghai, China). Cell Counting kit‐8 (CCK‐8) was purchased from Dojindo Laboratories (Shanghai, China). AlexaFluor‐488 conjugated goat anti‐mouse secondary antibodies were purchased from Proteintech (Wuhan, Hubei, China).

4.2. Screening of Molecular Markers—Proteomics

Proteomics enabled the investigation of altered protein expression in diseased hearts and guided the development of novel diagnostic and therapeutic strategies. This study employed liquid chromatography‐tandem mass spectrometry (LC‐MS/MS) technology to investigate the expression patterns of proteins in cardiac tissues during different phases of coronary microvascular dysfunction (CMD) (1‐, 3‐, and 7‐days groups), using sham‐operated mouse cardiac tissue as a control. The study aimed to identify differentially expressed proteins in these cardiac tissues, explore proteins associated with CMD and their potential theoretical biomarker significance, and analyze their biological information. This provided potential targets for the subsequent development of earlier‐stage bimodal molecular probes for CMD. The specific methodology was as follows: ①Sample Processing: Cardiac tissue samples exhibiting microcirculatory dysfunction at different time points and normal microcirculation samples were homogenized in chilled protein extraction buffer. High‐abundance proteins were depleted using a dedicated kit. Proteins were then enzymatically digested, and the resulting peptides were labeled using I TRAQ reagents. The labeled peptides were subsequently analyzed by liquid chromatography‐high‐resolution mass spectrometry (LC‐HRMS). ②Screening of Differentially Expressed Proteins: Raw data were preprocessed using Progenesis QI software. Protein identification and quantitative analysis were performed. Statistically significant differentially expressed proteins were screened using statistical analysis. ③Bioinformatics Analysis: Differentially expressed proteins underwent enrichment analysis, Gene ontology (GO) enrichment analysis, pathway analysis (utilizing pathway databases), and PPI network analysis. These analyses were conducted using the Gene Ontology database, pathway databases, and the Clusters of Orthologous Groups (COG) database to elucidate molecular functions and signaling pathways. ④Correlation Analysis: Proteomic data were correlated with transcriptomic data. A protein search database was constructed using transcriptomic data to enhance the number of peptide and protein identifications. Hierarchical clustering analysis of expression patterns across different samples was performed using Cluster 3.0 and Java Treeview software to analyze relationships between proteins.

4.3. Validation of Screening Targets

4.3.1. Immunofluorescence Staining

The expression level of Ninj1 in cells or tissues was detected. The tissues were fixed in 4% paraformaldehyde and blocked with 10% goat serum. Then, they were incubated with anti‐Ninj1 antibody (Invitrogen, PA5‐95755) overnight. APC or AF488‐labeled secondary antibodies were added, and the slides were incubated for 60 min at 37°C. Finally, cells or sections were imaged using a CLSM (Leica).

4.3.2. Western Blot

Total protein was isolated from HUVEC and mouse tissue samples. The samples were homogenized in ice‐cold RIPA lysis buffer (Solarbio, Beijing, China) supplemented with protease inhibitors (Solarbio, Beijing, China). Protein concentration in the tissue homogenates or cell lysates was quantified. Proteins were separated by electrophoresis on SDS‐PAGE gels and subsequently transferred onto PVDF membranes. The membranes were blocked with 5% skim milk at room temperature for 1 h. After washing, the membranes were incubated overnight at 4°C with the primary antibody against Ninj1 (Santa Cruz, sc136295). Following three washes with TBST (Tris‐buffered saline with Tween 20), the membranes were incubated with secondary antibody at room temperature for 2 h. Protein bands were visualized using an enhanced chemiluminescence (ECL) substrate (TransGen Biotech). Band intensity was quantified using E blot software to determine expression levels.

4.3.3. RNA Extraction and Real‐Time Quantitative PCR Methods

Total RNA was isolated from HUVEC and tissue samples using Trizol reagent (Life Technologies). Gene expression levels were determined using the 2× Universal SYBR Green Fast qPCR Mix (ABclonal) on a real‐time PCR system.

The following primers were used for qPCR analysis:Ninj1: Forward 5’‐GGAACCGGCCCATCAATGT‐3’, Reverse 5’‐GGCACGAAGAAGGCGAAATC‐3’

GAPDH: Forward 5’‐TGGAAAGCTGTGGCGTGATG‐3’, Reverse 5’‐GTCAGATCCACGACGGACAC‐3’. All experiments were performed in triplicate.

4.4. Preparation of Ninj1‐IR780‐NBs

IR780‐NBs were prepared using the thin‐film hydration method [35, 36, 37]. First, DSPC (10 mg), DSPE‐PEG2000 (4.5 mg), DSPE‐PEG2000‐Biotin (2.5 mg), and IR780 (0.5 mg) were dissolved in 1 mL of chloroform (CHCl₃). The mixture was vortexed for 1 min in a 65°C water bath until complete dissolution. Then, the CHCl₃ was evaporated under a stream of nitrogen gas, followed by vacuum treatment, to form a thin lipid film on the inner wall of the vessel. Then, the film was hydrated with 5 mL of Tris‐HCl buffer (pH 7) to dissolve the lipids. The vessel was then sonicated in an ultrasonic cleaner until the solution became visibly translucent and free of precipitates. Next, the resulting liposome suspension was aliquoted into glass vials. Finally, C₃F₈ gas (99.999% purity) was introduced into each sealed vial for 10 s, followed by evacuation using a vacuum pump for 45 s. This gas exchange cycle was repeated 4–5 times to form gas‐filled microbubbles (IR780‐NBs). Ninj1‐targeted microbubbles were constructed using a biotin‐avidin bridging strategy. Avidin was added in sufficient quantity to saturate the biotin groups on the IR780‐NBs surface. After incubation, the avidinylated microbubbles were washed. Subsequently, either AF488‐fluorescently labeled, biotinylated anti‐Ninj1 antibody or an isotype control IgG antibody was added and co‐incubated to allow binding to the avidin. Unbound antibodies were removed by washing three times with PBS. This yielded the targeted microbubbles: Ninj1‐IR780‐NBs (targeted) or IgG‐IR780‐NBs (isotype control).

4.5. Characterization of Ninj1‐IR780‐NBs

The morphology of Ninj1‐IR780‐NBs was characterized by fluorescence microscopy (Leica). The particle size and potential of Ninj1‐IR780‐NBs were measured using MalvernPanalytical. NB concentrations were determined using a hemocytometer. The measurements were repeated three times for each sample. The binding efficiency of NBs to AF488‐labeled AMH antibody or control IgG was examined by fluorescence microscopy and flow cytometry. The stability of Ninj1‐IR780‐NBs was analyzed by measuring the particle size of microbubbles in PBS at different time points at room temperature (25°C) (0 min, 20 min, 40 min, 1 h, 2 h, 3 h, and 4 h) and 4°C (1 h, 2 h, 4 h, 8 h, 12 and 24 h). To evaluate the in vitro ultrasound imaging ability of Ninj1‐IR780‐NBs, ultrasound images of microbubbles with different concentrations (10⁴, 10⁵, 10⁶, 10⁷, 10⁸ bubbles/mL) in AGAR gels were acquired using a high‐resolution ultrasound imaging system (VisualSonicsVevo3100, Canada). For the resulting ultrasound images, the intensity of the region of interest (ROI) was quantified by ImageJ software. To determine the NIR fluorescence imaging performance of the prepared microbubbles, fluorescence images of Ninj1‐IR780‐NBs at different concentrations (10⁴, 10⁵, 10⁶, 10⁷, 10⁸ bubbles/mL) were recorded on a fluorescence imaging system (IVIS Spectrum, Perkin Elmer, USA). The excitation wavelength (Ex) was 780 nm, and the emission wavelength (Em) was 810 nm. The fluorescence intensity of different concentrations of microbubbles was quantified using the Living Image IVIS software. The fluorescence spectra of the three probes (Ninj1‐IR780‐NBs, IR780‐NBS, and free IR780) with the same IR780 concentration were recorded using a fluorescence spectrophotometer (Shimatsu RF‐5301 PC type fluorescence spectrophotometer). The absorption spectra of Ninj1‐IR780‐NBs, IR780‐NBS, and free IR780 were recorded using UV–vis–NIR spectroscopy (UV‐1900, Shimadzu, Japan).

4.6. Ninj1‐IR780‐NBs Probe Specific Targeting In Vitro

4.6.1. Establishment of the Inflammation‐Injured Cell Model

Human umbilical vein endothelial cells (HUVEC) were purchased from Qida Biological Technology (Shanghai, China), catalogue number: CD0290. The cells were seeded into 6‐well plates at a density of 5 × 10⁵ cells/mL. After culturing for 3 days to form a confluent monolayer, the culture medium was replaced with basal medium containing human tumor necrosis factor‐α (TNF‐α) at concentrations of 5, 10, 15, and 20 ng/mL. HUVEC cultured without TNF‐α served as the control group. The optimal intervention concentration for establishing the inflammation‐injured cell model was determined based on the level of CD44 expression. After 12 h of incubation, the cells were harvested, washed twice with PBS, and incubated with an APC‐conjugated anti‐human CD44 antibody. The cell suspension was mixed thoroughly and incubated on ice for 30 min in the dark. Unbound antibody was removed by washing with PBS. The cell suspension was then filtered into flow cytometry tubes. Prior to analysis, the tubes were wrapped in aluminum foil to protect from light and stored at 4°C. The surface expression level of CD44 on HUVEC was analyzed using flow cytometry. Based on the results, the optimal TNF‐α concentration for inducing the inflammation‐injured cell model was selected.

4.6.2. Static Specific Adhesion Assay of Targeted Microbubbles

HUVEC, either untreated or treated with TNF‐α, were incubated with the designated nanobubbles (NBs) for 30 min according to the following groups: (a) Untreated HUVEC + IR780‐NBs; (b) Untreated HUVEC + Ninj1‐IR780‐NBs; (c) TNF‐α‐treated HUVEC + PBS; (d) TNF‐α‐treated HUVEC + IR780‐NBs; (e) TNF‐α‐treated HUVEC + IgG‐IR780‐NBs; (f) TNF‐α‐treated HUVEC + Ninj1‐IR780‐NBs. Lipid bubble adhesion was observed under bright‐field microscopy. The targeting capability of Ninj1‐IR780‐NBs was further evaluated using the CLSM, and flow cytometry (FCM). Prior to the assay, IR780‐NBs, IgG‐IR780‐NBs, and Ninj1‐IR780‐NBs were diluted 10‐fold in sterile 1× PBS. After 24 h of cell culture, the old medium was aspirated, and cells were gently washed 2–3 times with sterile 1× PBS. Sterile suspensions of IR780‐NBs, IgG‐IR780‐NBs, or Ninj1‐IR780‐NBs (1 × 10⁷ bubbles/mL) were then added to the culture dishes. The dishes were gently swirled to ensure even distribution of the contrast agent solution across the bottom and incubated at room temperature in the dark for 40 min. After incubation, the contrast agent solution was aspirated, and cells were gently washed three times with sterile 1× PBS. Subsequently, 1 mL of DAPI solution was added to counterstain the nuclei, followed by incubation at 37°C in the dark for 30 min. The DAPI solution was then aspirated, cells were washed three times with sterile 1× PBS. At least three images were captured. All experiments were performed in triplicate.

4.6.3. Dynamic Specific Adhesion Assay of Targeted Microbubbles

This assay utilized a TNF‐α‐treated HUVEC model under fluid shear stress. Experiments were conducted by applying varying levels of wall shear stress (1, 2, 4, 8, and 12 dyn/cm²) for 5 min. Following each experiment, five random microscopic images were captured at 200× magnification. The number of Ninj1‐IR780‐NBs adherent to the cell surface per field of view per minute was quantified to assess the number of microbubbles retained on the substrate. Each group was tested in triplicate. Additionally, the adhesion kinetics were assessed by varying the incubation time (1, 2, 3, 4, and 5 min) under a constant shear stress of 4 dyn/cm². After each time point, the number of retained microbubbles was evaluated using the same quantification method described for the static co‐incubation assay.

4.7. In Vivo NIR Fluorescence Imaging to Evaluate the Targeting Effect of Ninj1/IR780‐NBs

This study was approved by the Animal Ethics Committee of First Affiliated Hospital of Xinjiang Medical University (Approval Number: IACUC‐20200318‐103). Male C57BL/6J mice aged 8—10 weeks, weighing 22–25 g, were used to establish the CMD model. Sham‐operated mice of the same age, sex, and body weight were used as controls. The animal rooms were sterilized by ultraviolet (UV) light, maintained at a temperature of 22°C–24°C and a relative humidity of 50%–60%, with a ventilation rate of 14 air changes per hour. The animals were given a standard diet and distilled water under standard conditions. CMD model mice were divided into 6 groups (n = 5) and treated as follows: (a) CMD model mice + T‐IR780‐NBs; (b) CMD model mice + IgG‐IR780‐NBs; (c) CMD model mice + IR780‐NBs; (d) Sham‐operated mice + T‐IR780‐NBs; (e) Sham‐operated mice + IgG‐IR780‐NBs; (f) Sham‐operated mice+IR780‐NBs (dose: 0.3 mg/kg, IR780, concentration: ∼6 µm), in vivo NIR fluorescence imaging (Ex: 780 nm, Em: 810 nm) was performed using the small animal imaging system at 1, 4, 8, 12, 24, 48, and 72 h after injection. The mean fluorescence intensity values of the cardiac regions were measured and analyzed using Image J software.

4.8. In Vivo Ultrasound Imaging to Evaluate the Targeting Effect of T‐IR780‐NBs

CMD model mice were divided into 6 groups (n = 5) and treated as follows: (a) CMD model mice + T‐IR780‐NBs; (b) CMD model mice + IgG‐IR780‐NBs; (c) CMD model mice + IR780‐NBs; (d) Sham‐operated mice + T‐IR780‐NBs; (e) Sham‐operated mice + IgG‐IR780‐NBs; (f) Sham‐operated mice + IR780‐NBs (dose: 0.3 mg/kg, IR780, concentration: ∼6 µm). Contrast‐enhanced ultrasound (CEUS) was performed in the mice. The ultrasound imaging parameters were configured as follows: a transmit frequency of 18 MHz, an output power of 10%, a dynamic range of 40 dB, a mechanical index (MI) of 0.1, and a scanning depth of 4 mm. For all scanning regions, the image acquisition was performed at the maximum available frame rate, while the time gain compensation and focus settings were optimized to achieve the best image quality. After intravenous (i.v.) administration of nanobubbles, long‐axis images of the mouse left ventricle were obtained with a high‐frequency ultrasound probe, showing the ventricles and myocardium of the mouse heart. At the same time, the location of cardiac damage was visualized using enhanced imaging. After image acquisition, the acquired data were corrected for respiratory motion artifacts and measured for myocardial contrast intensity using Vevo CQ software. Each set of ultrasound images was analyzed by two experienced sonographers using Vevo CQ software in a double‐blind manner, recording time‐intensity curves. Mice in groups (a), and (c) were euthanized after injection, and the corresponding heart tissues were perfused with 4% paraformaldehyde and then cut into 5 µm sections using a freezing microtome. Heart sections were incubated with Ninj1 antibody and then stained with Alexa Fluor488‐labeled secondary antibody, and nuclei were labeled with DAPI. The distribution of microbubbles and Ninj1 in heart sections was observed by confocal microscopy.

4.9. Dual‐Modality Molecular Imaging for Efficacy Evaluation

The CMD model mice were treated according to the conventional clinical treatment cycle. Atorvastatin (dose: 10 mg/kg, frequency: once daily) was administered orally for 28 days. The treated mice were then subjected to tail vein injection of T‐IR780‐NBs (dose: 0.3 mg/kg, IR780) for NIR fluorescence imaging and ultrasound imaging to evaluate the treatment effect. The imaging procedure was based on the methods described above.

4.10. Cell Culture

Human umbilical vein endothelial cells (HUVEC) were purchased from Qida Biological Technology (Shanghai, China). The complete medium for HUVEC contains endothelial cell growth medium (ECGM), fetal bovine serum (FBS), and penicillin/streptomycin solution (P/S). The cells were cultured in an incubator at 37°C in a 5% CO₂ environment. When the cells had grown to cover 90% of the bottom of the flask, the cells were digested with 0.125% trypsin, subcultured once every 2–3 days, and the cells in the logarithmic growth phase were taken for subsequent experiments.

4.11. In Vitro and In Vivo Toxicity Tests

Human umbilical vein endothelial cells were cultured and seeded in 96‐well plates at a concentration of 5 × 10³ per well. Endothelial cell medium (100 µL) containing 10% fetal bovine serum was added to each culture medium and incubated for 24 h at 37°C with 5%CO₂. Then, the same amount of fresh culture medium was changed, and T‐IR780‐NBs were added to the culture medium according to the concentration gradient (5 × 10³ – 5 × 10⁸/mL). Six secondary Wells were set up in each gradient, and a blank control group, a negative control group, and a positive control group were set up at the same time. The culture medium was discarded, and 100 µL of medium containing 10 µL of CCK8 reagent was added to each well after washing three times with PBS. After incubation for 2 h at 37°C with 5%CO₂, the absorbance (OD) value at 450 nm was determined by a microplate reader to determine the cell viability.

In order to study the biosafety of T‐IR780‐NBs in vivo, blood samples were collected from experimental animals in each group for blood biochemical tests. The control group (intravenous injection of PBS on day 0) and the experimental group (intravenous injection of targeted microbubbles, 0.5 mg/kg, IR780 on day 1, 3, 7, and 14) were analyzed for hematological and biochemical parameters (blood routine, liver, kidney, and heart function) 24 h after injection. The main organs (heart, liver, spleen, lung, and kidney) were collected for HE staining analysis.

4.12. Statistical Analysis

Origin Lab (2018, 64‐bit) and Graphpad Prism 9 software were used for statistical analysis. Data were processed and expressed as mean ± standard deviation (SD). Statistical significance between the two groups was evaluated using Student's t‐test statistics, and multiple variables were compared using one‐way ANOVA analysis.

Author Contributions

X. Xu: conceptualization, data curation, formal analysis, writing – original draft. L. Guan: formal analysis, investigation, visualization. Y. Mu: funding acquisition, supervision, writing – review, and editing. B. Tayier: formal analysis, investigation, methodology. C. Yuan: software, validation, visualization. S. Chen: visualization. Q. Shi: investigation.

Funding

The National Natural Science Foundation of China (Grant No. 32071459、82572221、82560344), the 2024 Major Scientific Research Cultivation Project of Xinjiang Medical University (Grant No. XYD2024ZX07).

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File: adhm70847‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.