
==== Front
Proc Natl Acad Sci U S A
Proc Natl Acad Sci U S A
PNAS
Proceedings of the National Academy of Sciences of the United States of America
0027-8424
1091-6490
National Academy of Sciences

38252832
202309811
10.1073/pnas.2309811121
research-articleResearch ArticlechemChemistrymed-sciMedical Sciences410
422
Physical Sciences
Chemistry
Biological Sciences
Medical Sciences
Single-particle imaging of nanomedicine entering the brain
Wei Mian a b https://orcid.org/0000-0003-0052-304X

Qian Naixin a b https://orcid.org/0000-0001-6433-063X

Gao Xin a b https://orcid.org/0000-0002-0911-3656

Lang Xiaoqi a b
Song Donghui a b
Min Wei wm2256@columbia.edu
a b c 1 https://orcid.org/0000-0003-2570-3557

aDepartment of Chemistry, Columbia University, New York, NY 10027
bKavli Institute for Brain Science, Columbia University, New York, NY 10027
cDepartment of Biomedical Engineering, Columbia University, New York, NY 10027
1To whom correspondence may be addressed. Email: wm2256@columbia.edu.
Edited by W. Moerner, Stanford University, Stanford, CA; received June 10, 2023; accepted December 20, 2023

22 1 2024
30 1 2024
22 7 2024
121 5 e230981112110 6 2023
20 12 2023
Copyright © 2024 the Author(s). Published by PNAS.
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).

Significance

Drug delivery via nanocarriers is a groundbreaking approach. However, the in vivo behaviors of these nanocarriers remain largely unknown due to the lack of techniques. Here, we develop an optical method to image nanocarriers in tissues. Our method offers unique advantages including single-particle sensitivity, chemical specificity, and particle counting capability. With this method, we observe individual nanocarriers that cross the blood–brain barrier and count their absolute number in the brain. We find that the fate of nanocarriers varies substantially across multiple scales and the transport of nanocarriers to the brain decreases with age. This technology has the potential to greatly advance the development and translation of nanocarrier-based treatment.

Nanomedicine has emerged as a revolutionary strategy of drug delivery. However, fundamentals of the nano–neuro interaction are elusive. In particular, whether nanocarriers can cross the blood–brain barrier (BBB) and release the drug cargo inside the brain, a basic process depicted in numerous books and reviews, remains controversial. Here, we develop an optical method, based on stimulated Raman scattering, for imaging nanocarriers in tissues. Our method achieves a suite of capabilities—single-particle sensitivity, chemical specificity, and particle counting capability. With this method, we visualize individual intact nanocarriers crossing the BBB of mouse brains and quantify the absolute number by particle counting. The fate of nanocarriers after crossing the BBB shows remarkable heterogeneity across multiple scales. With a mouse model of aging, we find that blood–brain transport of nanocarriers decreases with age substantially. This technology would facilitate development of effective therapeutics for brain diseases and clinical translation of nanocarrier-based treatment in general.

nanomedicine
single particle imaging
blood–brain barrier
stimulated Raman microscopy
Chan Zuckerberg Initiative (CZI) 100014989 Dynamic Imaging 2023-321166 Wei Min HHS | NIH | National Institute of Biomedical Imaging and Bioengineering (NIBIB) 100000070 R01 EB029523 Wei Min
==== Body
pmcCentral nervous system (CNS) diseases such as clinical depression, stroke, Alzheimer’s disease, and brain tumors are among the most destructive and poorly treated diseases (1–3). Drug development for CNS diseases suffers from long development time and low success rate (3). One major challenge is to deliver drugs across the blood–brain barrier (BBB) (4–8). In fact, more than 98% of small-molecule drugs and ~100% of macromolecule drugs failed to cross the BBB (5). To this end, nanomedicine has emerged as a promising method (9–11). Encapsulation of drugs into functionalized nanocarriers can improve drug bioavailability, realize controlled release, and reduce immune response (12, 13). The first nanocarriers that achieved successful drug delivery across the BBB were poly(n-butyl cyanoacrylate) (PBCA) nanoparticles through intravenous (IV) injection to mice in 1995 (14). Various aspects of this process have been documented (15–22), making PBCA the best-studied model system. After this pioneering study, various nanocarrier systems, including polymer nanoparticles (10, 23–27), liposomes (28–32), and albumin- or chitosan-based nanoparticles (33–37), have been explored.

Despite its rapid progress, nanomedicine has yet to deliver on its promise for treating CNS diseases. Clinically, development of brain-targeted nanomedicine seems to have reached an impasse: there is still no FDA-approved CNS nanomedicine up to now (11). Scientifically, the understanding of nanomedicine for CNS diseases remains primitive. Current systems of nanomedicine–brain interaction are usually treated as a black box, where only therapeutic effects of drugs are measured as the end results. The gap of knowledge manifests itself in a basic premise: nanocarriers cross the BBB and release the drug cargo in the brain parenchyma. Numerous books and reviews have depicted such a process as the foundation of the field (10, 38). However, its validity remains debated (11, 39). Although recent experiments attempted to image nanocarriers crossing the BBB, almost all reported negative or confounding results (23–25, 29, 40–43).

We attribute the knowledge gap in nanomedicine to the lack of measurement techniques to image nano–neuro interaction. As internalized nanocarriers spread throughout the body in the form of individual nanoparticles, imaging nanocarriers inside the brain is analogous to seeking needles in a haystack. Current imaging techniques have intrinsic limitations (44). The resolution and sensitivity of MRI, PET, and ultrasound are not high enough to tell whether a nanocarrier crosses the BBB. Electron microscopy has low contrast for distinguishing organic nanocarriers from endogenous structures, and its throughput is rather low for evaluating large brain tissue. Significant progress has been made with fluorescence imaging (23, 24, 42, 45). However, as nanocarriers are biodegradable, the problem of dye leaching, independent of covalent or non-covalent labeling (39, 40, 46), makes it difficult to tell whether the observed signals come from the dyes attached to nanocarriers or the dyes that have leached out (SI Appendix, Fig. S1). Furthermore, fluorescence imaging suffers from dye perturbation and photobleaching (40), autofluorescence (25), fixation artifacts (29), and difficulty of quantification. Thus, no existing techniques have the capability of imaging nanocarriers in brain tissue in an unambiguous manner (SI Appendix, Supplementary Note 1 and Fig. S2).

Here, we develop a powerful platform to image nanocarriers, harnessing the emerging stimulated Raman scattering (SRS) microscopy. Our method achieves single-particle sensitivity, definitive chemical specificity, and particle counting capability in cells and tissues. With this technique, we attempt to ask four basic questions: 1) Do PBCA nanoparticles cross the BBB? 2) If so, how many nanoparticles cross the BBB? 3) Where do nanoparticles end up after crossing the BBB? 4) How does physiological or pathological condition affect the BBB transcytosis of nanoparticles? By imaging structures across five orders of magnitude in size, we observe nanocarriers in mouse brains at the single-particle level. We acquire direct evidence that nanocarriers cross the intact BBB to the brain parenchyma. We harness the particle counting capability of our method to quantify the number of nanocarriers that cross the BBB. In addition, the fate of nanocarriers after crossing the BBB exhibits heterogeneity at multiple scales—from specific brain regions, to major cell types, and to subcellular compartments. In a mouse model of aging, we find that blood–brain transport of nanocarriers decreases with age. Together, our technique deepens the understanding of nano–neuro interaction in diverse animal models and can aid in development of effective therapeutics in the future.

Results and Discussion

SRS Imaging of Nanocarriers at the Single-Particle Level.

Following a standard method (20, 47), we synthesized PBCA nanoparticles using emulsion polymerization. The obtained nanoparticles show a hydrodynamic diameter of ~140 nm in aqueous medium by dynamic light scattering (SI Appendix, Fig. S3A) and exhibit a spherical shape with a diameter of 101 ± 32 nm (mean ± SD, n = 214 particles) in the solid phase by transmission electron microscopy (TEM) (SI Appendix, Fig. S3B), consistent with previous literature (20, 47). The Raman spectrum of PBCA nanoparticles exhibits a prominent peak at 2,249 cm−1 with a full-width-half-maximum (FWHM) of 15 cm−1, corresponding to the C≡N stretching of the repeating unit (Fig. 1A).

Fig. 1. SRS imaging of PBCA nanoparticles at the single-particle level. (A) Raman spectrum of the nitrile peak of PBCA nanoparticles. (B) Representative single-frame image of nanoparticles immobilized in agarose gels at 2,249 cm−1 (Left) and the corresponding 50-frame-average image of the same field of view (Right). The pixel dwell time is 4 μs. PDT, pixel dwell time. (C) Theoretical curve of the SRS signal of single nanoparticles as a function of the diameter of nanoparticles. The yellow area indicates the range of the SRS signal that corresponds to the diameter range of twice the SDs from the mean. (D) Representative SRS images of nanoparticles immobilized in agarose gels at 2,249 cm−1. Signal-to-noise ratios (SNRs) are labeled near the spots of nanoparticles. The effective pixel dwell time is 200 μs. (E) Zoom-in image of a nanoparticle spot in D. (F) Intensity profile of the dashed line in E. (G) Frames 1, 15, and 30 of consecutive 30 frames of imaging in the same field view as D. Each frame is an average of five frames with 4-μs pixel dwell time, resulting in 20-μs effective pixel dwell time. (H) Intensity ratio as a function of frame number in consecutive 30 frames of imaging. (Scale bars: 2 μm in B, D, and G and 500 nm in E.)

Using this characteristic peak, we next explored the detectability for nanoparticles with conventional Raman microscopy. The number of the repeating units in a 100-nm PBCA nanoparticle is around 2 × 106, which is also the number of the C≡N bonds (SI Appendix, Supplementary Note 2). The Raman cross-section of the C≡N bonds is around 6 × 10–29 cm2 (SI Appendix, Supplementary Note 3). Therefore, the Raman cross-section of a nanoparticle is about 1.2 × 10–22 cm2, which is 106 times weaker than a single-molecule fluorescence cross-section of ~10–16 cm2 (48). The laser can be focused to a waist area of 2 × 10–9 cm2 under a microscope. Hence, the probability of Raman scattering per excitation photon is (1.2 × 10–22 cm2)/(2 × 10–9 cm2) = 6 × 10–14. Assuming a moderate laser power of 10 mW and a long acquisition time of 100 ms, only ~300 photons can be generated for each nanoparticle. Considering the quantum yield of the whole instrument (including the objective, filters, pinhole, gratings, and camera) is around 1%, approximately three photons can be eventually detected. This signal can be easily overwhelmed by noises from autofluorescence and other backgrounds.

Compared to spontaneous Raman, SRS achieves ~108-fold enhancement in vibrational transition rate via quantum amplification (49). Practically, SRS operates more than 1,000 times faster than spontaneous Raman, which is required for imaging large tissue samples. However, whether SRS with fast acquisition provides a better detection limit than spontaneous Raman and whether the detectability of SRS reaches a single PBCA particle remains unknown. Then we studied this matter by performing SRS microscopy of PBCA nanoparticles embedded in agarose gels. We found a typical single-frame image with 4-μs pixel dwell time exhibits only noise without any discernible features (Fig. 1B). To push SRS microscopy to its limit, we adopted a strategy of imaging a seemingly “empty” field of view with a maximized exposure time. To avoid over-exposure and reduce photo-damage and heating, we implemented fast scanning acquisition and increased the effective exposure by averaging multiple frames. In the seemingly empty field of view, multiple spots eventually showed up with low noise in the 50-frame-averaged image (effective pixel dwell time of 200 μs) (Fig. 1B).

Theoretical calculations help understand our signal (SI Appendix, Supplementary Note 4). Under an effective pixel dwell time of 200 μs, the imaging noise can be lowered to 2 × 10−7 (SI Appendix). The SRS signal is proportional to the cubic power of the particle diameter (Fig. 1C). In the diameter range of twice the SDs from the mean (37 to 165 nm), the SNR is calculated to vary from 0.4 to 32 for individual nanoparticles. Using a SNR~3 as a threshold, SRS should be able to detect the majority of the synthesized nanoparticles (diameter > 74.6 nm) at the single-particle level (Fig. 1C). Therefore, we used the strategy of averaging 50 frames with 4-μs pixel dwell time for all the SRS images of nanoparticles unless specified otherwise.

The spots we identified experimentally exhibit SNRs in the expected range for individual nanoparticles (Fig. 1D). Multispectral SRS imaging confirmed that these spots were PBCA nanoparticles, exhibiting the C≡N peak at 2,249 cm−1 (SI Appendix, Fig. S4). In addition, these spots exhibit a symmetric shape with a FWHM of 400 to 500 nm (Fig. 1 E and F), consistent with the diffraction limit of our microscope. Based on the SNR and the shape under SRS microscopy, these spots are most likely to be single PBCA nanoparticles. Furthermore, unlike fluorescence microscopy, SRS imaging of nanoparticles is free from photobleaching, allowing accurate quantification and repetitive imaging (Fig. 1 G and H).

Multispectral SRS Imaging of Nanocarriers in Tissue with Definitive Spectral Evidence.

We next developed highly specific SRS microscopy for nanoparticles in cells and tissues. Unlike the simple, clean environment of agarose gels, cells and tissues contain a plethora of endogenous structures that can interfere with the identification of nanoparticles. Therefore, we leveraged vibrational features for the unique specificity to distinguish nanoparticles from endogenous structures. The Raman spectrum of PBCA nanoparticles is distinct from that of macrophages or mouse brain tissue (Fig. 2A), indicating the feasibility of Raman microscopy in detecting nanoparticles in cells and tissues. Among several characteristic Raman peaks of PBCA nanoparticles, the C≡N peak at 2,249 cm−1 is the most pronounced (Fig. 2A). As this peak resides in the cell-silent region (1,800 to 2,600 cm−1) in which endogenous biomolecules have no Raman signals (Fig. 2A), targeting the C≡N mode provides a background-free imaging contrast in biological tissues, a concept known as bioorthogonal chemical imaging (50).

Fig. 2. Imaging PBCA nanoparticles with definitive chemical specificity. (A) Raman spectrum of PBCA nanoparticles, macrophages, and brain tissue. The strong peak of PBCA nanoparticles at 2,249 cm−1 corresponds to the C≡N stretching mode. (B–E) Imaging PBCA nanoparticles in macrophages. (B) Representative SRS images of nanoparticles in macrophages. SNRs are labeled near possible spots of nanoparticles. (C) Spectral images of nanoparticles at different wavenumbers. The spots are labeled in the image of 2,249 cm−1. (D) SRS spectra of the spots as labeled in C. (E) The “On–Off” image is the resultant image of on-resonance image at 2,249 cm−1 subtracting off-resonance image at 2,227 cm−1. Protein and lipid images derive from linear unmixing of two spectral images in the CH stretching region. Arrows indicate nanoparticle spots. Dotted lines delineate the cell border of the macrophage with nanoparticles. (F–J) Imaging PBCA nanoparticles in brain tissue. (F) Representative protein and lipid images of mouse brain tissue. (G) Single-frame image of nanoparticles at 2,249 cm−1 in the dotted box of F (Left) and the corresponding 50-frame-average image of the same field of view (Right). SNRs are labeled near possible spots of nanoparticles. The pixel dwell time is 4 μs. (H) Spectral images of nanoparticles at different wavenumbers. The spots are labeled in the image of 2,249 cm−1. (I) SRS spectra of the spots as labeled in H. (J) The On–Off image is the resultant image of on-resonance image at 2,249 cm−1 subtracting off-resonance image at 2,088 cm−1. Arrows indicate nanoparticle spots. (Scale bars: 2 μm in B, C, E, G, and H and 10 μm in F.)

We started this development by visualizing PBCA nanoparticles engulfed in macrophages. We treated a macrophage cell line (RAW 264.7) with nanoparticles for 24 h and then fixed cells before SRS imaging. Consistent with our calculations and results in gels, we observed spots of putative single nanoparticles (Fig. 2B). However, we found multiple spots with similar SNRs in the image at 2,249 cm−1 and could not tell whether these spots were nanoparticles or endogenous structures (e.g., lipid droplets) (Fig. 2B). We then acquired multiple images of the same field of view at six additional wavenumbers around 2,249 cm−1 (Fig. 2C). This strategy of multispectral imaging confirmed three spots have the characteristic C≡N peak (Fig. 2 C and D). In contrast, other spots only show “flat” spectra appearing in all spectral channels (Fig. 2 C and D), which is typical for cross-phase background (51, 52) from endogenous structures with high refractive index (likely lipid droplets). As the C≡N peak has a narrow linewidth (15 cm−1, Fig. 1A) compared to common fluorescence peaks (~1,500 cm−1), SRS signal of nanoparticles can be switched off if the pump laser was tuned away from the peak by ~10 cm−1 in either direction, rendering high chemical specificity (Fig. 2C). We thus subtracted the off-resonance image from the on-resonance image (SI Appendix), and the resultant On–Off image shows three nanoparticles with high specificity (Fig. 2E). The overlay of the nanoparticle image with the protein (CH3) and lipid (CH2) SRS images indicates that these nanoparticles were engulfed into a macrophage (Fig. 2E). Similar to the results in gels, these intracellular nanoparticles show almost the same diffraction-limited shape and are free from photobleaching (SI Appendix, Fig. S5).

Next, we applied multispectral SRS microscopy for nanoparticles in tissues. Although two studies from us and others have demonstrated SRS imaging of nanoparticles (53, 54), these studies were conducted with neither live animal experiments relevant to drug delivery nor single-particle imaging. In contrast, we injected polysorbate 80–coated nanoparticles into mice through the tail vein, killed the mice 1 h after injection, and harvested, fixed, and sectioned the brains for SRS imaging. Polysorbate 80 coating has been established as a gold standard to facilitate PBCA nanoparticles crossing the BBB (21). It is believed that polysorbate 80 promotes adsorption of apolipoprotein E (apoE) in blood onto nanoparticle surface, and binding of apoE to the low-density lipoprotein receptor (LDLR) and the low-density lipoprotein receptor-related protein-1 (LRP1) on brain endothelial cell membrane facilitates BBB crossing via receptor-mediated transcytosis (10). We acquired protein and lipid images to see brain tissue structures and focused on small regions to search for nanoparticles (Fig. 2F). Consistent with the results in gels, we observed mostly noise with hardly any structures in a single-frame image, but found multiple spots with expected SNRs in the 50-frame-averaged image (Fig. 2G). Similar to the situation in macrophages, multispectral imaging provided high chemical specificity to distinguish nanoparticles from endogenous structures in a definitive fashion (Fig. 2 H and I). The On–Off image shows three particles with high confidence (Fig. 2J).

Single-Particle Counting of Ultra-Purified Nanocarriers.

To fully exploit the quantification capability of SRS, we attempted to count the number of nanoparticles. This is feasible with SRS, whose signal strength is linearly dependent on the number of chemical bonds under excitation. However, the particle size has to be nearly uniform to allow for particle counting. To make nanoparticles with a narrower size distribution, we incorporated filtration and centrifugation procedures to remove larger particles and smaller particles, respectively (SI Appendix) (55). PBCA nanoparticles prepared with this refined method indeed exhibit a more uniform size distribution: a SD of only 10 nm around 105 nm (Fig. 3 A and B).

Fig. 3. Purification of PBCA nanoparticles by size enables single-particle counting. (A) Representative TEM images of nanoparticles prepared by a modified method with purification. (B) Histogram of the size distribution of nanoparticles measured by TEM. (C) Representative SRS image of nanoparticles. This is an On–Off image, i.e., the resultant image of on-resonance image at 2,249 cm−1 subtracting off-resonance image at 2,088 cm−1. The pixel size is 200 nm, which is consistent with brain tissue imaging. (D) Histogram of the SRS intensity distribution of nanoparticles. (E) Histogram of the SRS intensity distribution of nanoparticles fitted with three Gaussian functions. (F) Representative SRS images of nanoparticles and the intensity profiles along the dashed lines in the images. The steps in the intensity profiles are due to pixelation. The numbers in the SRS images indicate the quantity of nanoparticles in the specific spots. (Scale bars: 500 nm in A and F and 5 μm in C.)

We next developed single-particle counting of these ultra-purified nanoparticles embedded in agarose gels. We acquired the intensity distribution of >1,200 diffraction-limited spots and found that the distribution can be fitted with three Gaussian peaks (Fig. 3 C–E). Remarkably, these three peaks have quantized signal strengths: The intensity of the second peak doubles that of the first peak, and the intensity of the third peak triples that of the first peak (Fig. 3D), suggesting spots that consist of a single particle, double particles, and triple particles, respectively. Such quantization would not be possible without ultra-purified nanoparticles, photobleaching-free imaging (Fig. 1 G and H and SI Appendix, Fig. S5 C and D), and the linear concentration dependence of SRS.

We then calculated the expected SRS signal strength of a single PBCA nanoparticle given the size measured by TEM (Fig. 1C and SI Appendix, Supplementary Note 4). The theoretical value (ΔI/I=1.7×10-6) is close to the observed signal of the first quantized peak (ΔI/I=2×10-6), further corroborating that the first peak derives from single nanoparticles. The signal strength of single nanoparticles corresponds to a SNR of 10, which can be reliably detected. These results confirmed that the detectability of our method indeed reaches the level of single nanoparticles. Moreover, we were able to count the number (1, 2, or 3) of nanoparticles in each spot through reading its intensity profile (Fig. 3F). In summary, we have achieved single nanocarrier counting by developing ultra-purified PBCA nanoparticles, which allows us to study the process of drug delivery in a highly quantitative manner.

Comparison of the SRS-Based Method to a Traditional Fluorescence Approach.

We have developed this SRS-based method for imaging nanocarriers with single-particle sensitivity, chemical specificity, and particle counting capability. To benchmark its performance against a traditional approach, we compared our method with fluorescence microscopy on nanoparticles loaded with fluorescence dyes. We prepared PBCA nanoparticles labeled with fluorescein isothiocyanate-dextran (FITC-dextran), embedded dye-labeled nanoparticles in agarose gels, and performed correlative SRS and fluorescence imaging. As shown in SI Appendix, Fig. S6A, the SRS images overlap well with the fluorescence images for every spot in the entire images, indicating that the fluorescence labeling of nanoparticles was successful. Next, we performed time-lapse fluorescence imaging of nanoparticles for 50 frames. We found that the fluorescence intensity of nanoparticles decreased rapidly (SI Appendix, Fig. S6 B and C): There are only a few nanoparticle spots with weak fluorescence in Frame 15 and almost no visible nanoparticle spots in Frame 30 (SI Appendix, Fig. S6B). The fluorescence intensity declined to half of the original intensity within the first five frames (SI Appendix, Fig. S6C). In contrast to the constant intensity of time-lapse SRS imaging (Fig. 1 G and H), these results show the severe problem of photobleaching in fluorescence imaging of nanoparticles.

We next compared SRS with fluorescence imaging of dye-labeled nanoparticles in macrophages. We incubated macrophages with dye-labeled nanoparticles for 24 h, fixed and washed cells, and performed correlative SRS and fluorescence imaging. As shown in SI Appendix, Fig. S7 A–C, the autofluorescence of macrophages exhibits spot-like patterns and the fluorescence intensity of nanoparticles is within the intensity range of autofluorescence. As a result, it is impossible to identify nanoparticles using fluorescence images alone. Here, we used multispectral SRS imaging to pinpoint nanoparticles (SI Appendix, Fig. S7 D and E). The background-subtracted On–Off images show nanoparticles with high specificity, free from the interference of autofluorescence or endogenous structures (SI Appendix, Fig. S7 A–C). Furthermore, we found some fluorescence spots of nanoparticles exhibit a complex structure with a bright core and a less bright, asymmetric halo around the core (SI Appendix, Fig. S7F). Based on the corresponding SRS image, the bright core is the nanoparticle spot (SI Appendix, Fig. S7F). As shown in the merged image of FITC and proteins, the left contour of the asymmetric halo coincides with the cell border of a macrophage (SI Appendix, Fig. S7F). Based on its intensity, shape, and location, the asymmetric halo is likely to be from FITC-dextran that leached out of the nanoparticle matrix and diffused into the cytoplasm of a macrophage (SI Appendix, Fig. S7F), as FITC-dextran can be fixed in tissue with standard fixation methods (56). Besides, the fluorescence intensity of nanoparticles decreased rapidly in time-lapse imaging: The fluorescence was bleached to half of the original intensity within the first six frames (SI Appendix, Fig. S7 G and H). In summary, these results illustrate autofluorescence, dye leaching, and photobleaching as limitations of fluorescence microscopy and demonstrate the advantages of our method for imaging nanoparticles.

Single-Particle Imaging of Nanocarriers Crossing the BBB.

Equipped with our imaging platform, we aimed to directly image nanocarriers entering the brain. The prerequisite to visualize nanocarriers crossing the BBB is the specific staining and imaging of brain vasculature (44). Following a well-established method (57), IV injection of fluorophore-conjugated lectin achieved specific staining of blood vessels in the mouse brain (SI Appendix, Fig. S8). The pattern labeled by lectin was confirmed by immunostaining against laminin, a major component of the basement membrane (SI Appendix, Fig. S8 A and B). The BBB is composed of endothelial cells forming the vascular wall surrounded by pericytes and astrocyte end-feet embedded in the basement membrane (8). Based on this structure, laminin staining should have a slightly larger profile to enclose the lectin staining, which is exactly observed in both the volume-rendered images (SI Appendix, Fig. S8A) and the cross-section views (SI Appendix, Fig. S8 B–D).

We next combined our platform with vasculature staining to image nanoparticles in the context of the BBB. Specifically, we performed IV injections of polysorbate 80–coated PBCA nanoparticles and lectin into adult mice and killed them 1 h after nanoparticle injection. Mice brains were fixed, sectioned, and stained with NucGreen and ActinRed. Nanoparticles were imaged by SRS while lectin-labeled vasculature and NucGreen-labeled nuclei were imaged by fluorescence. For the whole procedure, PBCA nanoparticles can undergo two phases of degradation: a 1-h period in living mice with esterases leading to ~5% degradation (58) and a period in fixed tissues without active enzymes until imaging (59). We found that PBCA nanoparticles remained stable in the second period if kept at 4 °C: The SRS intensity of the same nanoparticles had no significant difference before and after even 5 d storage (SI Appendix, Fig. S9). Therefore, degradation should have a minimal effect on the signal of nanoparticles.

Consistent with previous results (Figs. 1B and 2G), we observed mostly noise with barely discernible structures in a single-frame image at 2,249 cm−1 (SI Appendix, Fig. S10A). Remarkably, with our multiple-frame-average strategy, multiple spots eventually emerged with reduced noise (SI Appendix, Fig. S10A). The specificity challenge was also encountered: There were many spots within the SNR range and we could not tell the identity of these spots from the single image at 2,249 cm−1 (SI Appendix, Fig. S10A). Multispectral imaging clearly verified that five spots were nanoparticles with the characteristic C≡N peak at 2,249 cm−1, while other spots were endogenous structures (SI Appendix, Fig. S10 B and C). The On–Off image shows five nanoparticles in a definitive manner (SI Appendix, Fig. S10D). In addition, all these five nanoparticles were far away from the vasculature (SI Appendix, Fig. S10D), suggesting that they crossed the BBB. However, such a single-plane image cannot capture the 3D extended structure of the BBB. Volumetric imaging is hence needed.

SRS microscopy exhibits intrinsic optical sectioning capability (49). Specific brain regions are determined by comparing whole-brain-section images of ActinRed staining (Fig. 4A) with the Allen Brain Atlas (mouse.brain-map.org) (Fig. 4B). We then zoomed in on specific brain regions for volumetric imaging through the entire depth of the brain section (~40 μm in depth). Volumetric imaging captured nanoparticles in the 3D context of the BBB in various brain regions, including the cerebral cortex (CCX), the dentate gyrus of the hippocampus (DG), the CA1 region of the hippocampus (CA1), and the thalamus (THA) (Fig. 4 C–F). We classified these nanoparticles into three categories: “in vasculature,” “in perivascular space,” and “in brain parenchyma” (Fig. 4 G–I; SI Appendix, Supplementary Note 5). Some nanoparticles were closely associated with vasculature (Fig. 4C, arrowhead; Fig. 4G), suggesting that they were either adsorbed on or resided inside endothelial cells. Some nanoparticles were observed in the perivascular space (Fig. 4H; nanoparticles outside the vasculature but within a distance of 1 μm from the nearest blood vessels). More importantly, we captured PBCA nanoparticles away from blood vessels in all these brain regions (Fig. 4 C–F, arrows; Fig. 4I; nanoparticles outside the vasculature with a distance larger than 1 μm from nearest blood vessels), indicating these particles have truly crossed the BBB into the brain parenchyma. Multispectral SRS imaging confirmed the observed signals indeed derived from the C≡N vibrational peak (SI Appendix, Fig. S11). In contrast, endogenous structures such as a lipid droplet show a constant intensity (SI Appendix, Fig. S11 C and D). Together, our technique visualizes structures spanning five orders of magnitude—from a centimeter-scale whole brain section to 100-nm-scale nanoparticles—with chemical specificity to identify nanocarriers that cross the BBB.

Fig. 4. Direct visualization of nanoparticles in brain tissue. (A) Large-area mosaic image of a whole brain slice using ActinRed staining. (B) Reference atlas of a coronal slice from Allen Mouse Brain Atlas (mouse.brain-map.org) at a sectioning position similar to A. Four brain regions are labeled on the atlas: the cerebral cortex (CCX); the dentate gyrus of the hippocampus (DG); the CA1 region of the hippocampus (CA1); and the thalamus (THA). (C–F) Volume-rendered images of nanoparticles by SRS and vasculature and nuclei by fluorescence microscopy in CCX (C), DG (D), CA1 (E), and THA (F). Arrows indicate nanoparticles in the brain parenchyma away from blood vessels. Arrowheads indicate nanoparticles that are closely associated with blood vessels. White frame lines indicate the border of a volume-rendered image. (G–I) Volume-rendered images of nanoparticles in three categories based on the locations relative to the BBB: in vasculature (G), in perivascular space (H), and in brain parenchyma (I). Arrows indicate nanoparticles that belong to the specific categories. The labeled spots in G and I correspond to the same spots labeled in C and E. (Scale bars: 500 μm in A, 20 μm in C–F, and 5 μm in G–I.)

Next, we performed a control experiment (n = 3 mice) with uncoated PBCA nanoparticles lacking polysorbate 80, which was reported to be unable to deliver drugs to the brain (14, 20). The experimental parameters of this control experiment are the same as those for coated nanoparticles (Fig. 4), except for the absence of polysorbate 80 coating (SI Appendix). We found no signals of uncoated PBCA nanoparticles in the brain (SI Appendix, Fig. S12A), although they were still found in the liver (SI Appendix, Fig. S12 B and C). These results support that the nanoparticles observed in the brain parenchyma (Fig. 4 C–F, arrows; Fig. 4I) did cross the BBB through receptor-mediated transcytosis, instead of artifacts from sample preparation. Thus, our imaging platform provides definitive evidence that PBCA nanoparticles cross the BBB in different brain regions. As previous studies show that PBCA nanoparticles can deliver drugs to the brain parenchyma (14, 15, 17, 20, 47), our results settle the controversies and validate the basic premise in nanomedicine that nanocarriers cross the BBB and release drugs in the brain parenchyma (10, 11, 38, 39). Besides, the results of nanoparticles with distinct surface properties (Fig. 4 and SI Appendix, Fig. S12) demonstrate that further experiments on nanocarriers of various formulations, including different surface modifications and core compositions, would promote the design of better nanomedicine for drug delivery.

Quantification of the Absolute Number of Nanoparticles that Cross the BBB.

After confirming that PBCA nanocarriers cross the BBB qualitatively, we then addressed the likelihood of nanoparticles crossing the BBB by counting individual nanocarriers. To exploit the particle counting capability of our method in the context of the BBB, we used ultra-purified nanoparticles (Fig. 3) for IV injection into adult mice. As highlighted in Fig. 5 A and B, nanoparticles appeared in the brain parenchyma far away from the vasculature: The upper two spots (a and b in Fig. 5C) were PBCA nanoparticles with the characteristic Raman peak at 2,249 cm−1, while the lower two spots (c and d in Fig. 5C) were lipid droplets with no vibrational feature (Fig. 5 C and D). Notably, we were able to count the two spots: Spot a was a single particle, and Spot b contained two particles (Fig. 5E). The signal strengths of these nanoparticles in the brain remained the same as those measured in the gel (Fig. 5 E and F). These results indicate that we were able to observe a single nanoparticle crossed the BBB with definitive evidence, demonstrating the power of our imaging platform.

Fig. 5. Single-particle counting of PBCA nanoparticles that cross the BBB. (A) Volume-rendered images of nanoparticles, vasculature, and nuclei in different brain regions. Circles indicate nanoparticle spots. (B) Single-plane images of nanoparticles, proteins, lipids, vasculature, and nuclei in the cerebral cortex. Circles indicate nanoparticle spots. (C) Multispectral SRS images of the golden box in B. Spots are labeled in the image at 2,249 cm−1. (D) SRS spectra of the spots as labeled in C. (E) SRS image of the two nanoparticle spots in the upper part of the images in C and the intensity profile along the dashed line in the image. The numbers in the SRS image indicate that there is one nanoparticle in the Lower-Left spot and there are two nanoparticles in the Upper-Right spot. (F) SRS image of another nanoparticle spot in the brain parenchyma and the intensity profile along the dashed line in the image. The number in the image indicates that there are two nanoparticles in the spot. (G) The distributions of the nanoparticle density in different brain regions (n = 33 fields of view for DG, 31 fields of view for CA1, 30 fields of view for CCX, 29 fields of view for THA, from four mice). The (mean ± SEM) values of the density distributions are indicated on the histograms. (Scale bars: 20 μm in A, 10 μm in B, 1 μm in C, and 500 nm in E and F.)

We then proceeded to count more nanoparticles in the brain parenchyma across diverse brain regions. We found the nanoparticle density of each field of view (FOV, ~200 × ~200 × ~40 μm3) had a broad distribution with a peak density below 2 × 104 mm−3 (Fig. 5G). The average nanoparticle density is (16 ± 5) × 103 mm−3 for DG, (35 ± 11) × 103 mm−3 for CA1, (39 ± 13) × 103 mm−3 for CCX, (14 ± 5) × 103 mm−3 for THA (mean ± SEM) (Fig. 5G). The average density shows a 2.8-fold difference between CCX and THA. Thus, these results represent heterogeneity in nanocarrier delivery both within a specific brain region and between different brain regions: Nanoparticles distribute in the brain parenchyma in a highly non-uniform manner. As pathological changes are not exhibited uniformly in the brain, some brain regions are especially vulnerable to CNS diseases. For example, the hippocampus is particularly susceptible to damage of Alzheimer’s disease at an early stage (60). The capability of our method to precisely quantify nanocarriers in a region-specific manner would help the development of nanomedicine targeting specific diseased tissues, which is a major goal for drug delivery (61, 62).

To evaluate the efficiency of nanocarrier delivery for individual mice, we estimated the number of nanoparticles that cross the BBB at the whole brain level. Based on the fraction of volume we have imaged, the total number of nanoparticles in the whole brain was extrapolated to be (1.0 ± 0.6) × 107 (mean ± SD, n = 4 mice, 4,824 nanoparticles, and SI Appendix, Supplementary Note 6). These results indicate that our method has the unique advantage of quantifying the absolute number of nanoparticles in the brain.

Fate of Nanocarriers After Crossing the BBB Shows Heterogeneity at Multiple Scales.

As cell-specific targeting is critical for the therapeutic efficacy of nanomedicine (38, 61, 62), we next asked where these nanoparticles might end up in the brain after crossing the BBB. We used antibodies against GFAP, NDRG2, NeuN, and Iba1 to label the processes of astrocytes, the somas of astrocytes, the somas of neurons, and microglia, respectively (Fig. 6 A–H and SI Appendix, Fig. S13). Using single-particle counting, we quantified the number of nanoparticles in each cell type across different brain regions (Fig. 6I). The distribution of nanocarriers shows heterogeneity at the levels of both brain regions and cell types (Figs. 5G and 6I). Intriguingly, the specific pattern of cell-type heterogeneity can differ for distinct brain regions (Fig. 6I). For DG, CA1, and CCX, the majority of nanoparticles localized in neurons, while only a few nanoparticles were in astrocytes or microglia (Fig. 6I). These results are consistent with the apoE’s natural pathway to transport cholesterol from astrocytes to neurons (63). For the region of THA, about half of nanoparticles localized in microglia and the other half were in neurons (Fig. 6I). As microglia transport nanoparticles from the brain parenchyma to the glymphatic system (64), the high proportion of nanoparticles in microglia suggests rapid clearance of nanocarriers from the brain parenchyma in the thalamus, which is consistent with the low nanoparticle density in the thalamus compared to other brain regions (Fig. 5G). These results demonstrate that our method is able to accurately count nanoparticles in a cell type–specific manner.

Fig. 6. Fate of nanocarriers after crossing the BBB shows heterogeneity at multiple scales. (A) Representative volume-rendered image of nanoparticles in neurons. (B) Single-plane zoom-in image of the golden box in A shows nanoparticles in neurons. Arrowheads indicate nanoparticles in the somas of neurons. Arrows indicate nanoparticles in the processes of neurons. (C) Representative volume-rendered image of nanoparticles in the somas of astrocytes. (D) Single-plane zoom-in image of the golden box in C shows nanoparticles in the somas of astrocytes as indicated by arrows. (E) Representative volume-rendered image of nanoparticles in the processes of astrocytes. (F) Single-plane zoom-in image of the golden box in E shows nanoparticles in the processes of astrocytes as indicated by arrows. (G) Representative volume-rendered image of nanoparticles in microglia. (H) Single-plane zoom-in image of the golden box in G shows nanoparticles in microglia as indicated by arrows. (I) Pie charts for the number of nanoparticles in three types of cells across different brain regions (n = 248, 663, 154, and 117 nanoparticles in DG, CA1, CCX, and THA, respectively). (J) Percentage of cells that contain nanoparticles for neurons, astrocytes, and microglia (n = 12,660, 1,785, and 240 cells for neurons, astrocytes, and microglia, respectively). Statistical significance is determined by the chi-square test. (K) Distributions of nanoparticle numbers per cell for neurons, astrocytes, and microglia that contain nanoparticles (n = 256, 22, and 19 cells for neurons, astrocytes, and microglia that contain nanoparticles, respectively). (L) Bar diagrams for subcellular localizations of nanoparticles in three types of cells across different brain regions (n = 1,052, 53, and 77 nanoparticles in neurons, astrocytes, and microglia, respectively). (Scale bars: 20 μm in A, C, E, and G, 10 μm in B, and 5 μm in D, F, and H.)

Next, we quantified the number of nanoparticles in individual cells. We found only a small proportion of cells contained nanoparticles: Approximately 2% of neurons, 1% of astrocytes, and 8% of microglia had nanoparticles (Fig. 6J). On average, neurons had 0.08 nanoparticles per cell, astrocytes had 0.03 nanoparticles per cell, and microglia had 0.3 nanoparticles per cell. For all the cells with nanoparticles, microglia had higher variation in nanoparticle uptake between individual cells than neurons or astrocytes (Fig. 6K). Further molecular profiling of these cells using spatial genomics and transcriptomics (65) may reveal the genomic determinants of this cell–cell heterogeneity in nanocarrier delivery both between different cell types and within a specific cell type.

As an ideal nanocarrier delivers drugs to not only designated cell types but also specific subcellular compartments (61, 62, 66), we went one step further to identify the subcellular locations of nanoparticles within each cell type. Besides a high intensity in the somas of neurons, the contrast of NeuN shows a low but discernible intensity in the processes of neurons in CA1 (Fig. 6 A and B). Among the nanoparticles in the neurons in CA1, about 3/4 were in the somas while about 1/4 were in the processes (Fig. 6L). Nanoparticles in astrocytes were observed mostly in the processes in DG and CA1, while nanoparticles in microglia were mostly found in the somas in THA (Fig. 6L). These results demonstrate our method has the capability to quantitatively capture the heterogeneity in the fate of nanocarriers at multiple scales, allowing detailed evaluation of targeted nanocarrier delivery in a hierarchical manner.

Blood–Brain Transport of Nanocarriers Decreases with Age.

The properties of the BBB are closely associated with the physiological or pathological state of the organism. On one hand, the integrity of the BBB is crucial for the protection, homeostasis, and functions of the CNS; on the other hand, BBB dysfunction is intimately related to the pathogenesis of various neurological disorders and diseases including stroke, epilepsy, multiple sclerosis, and Alzheimer’s disease (67). As a progressive decline in physiological condition, aging induces BBB changes (68) and remains a leading risk factor for neurodegenerative diseases (69). Although previous studies reported increasing age was associated with enhanced BBB permeability (70, 71), whether the BBB is disrupted in aging has been controversial for decades (71). How aging influences the blood–brain transport of nanocarriers remains an open question.

To tackle this question, we studied the interaction of nanocarriers with the BBB in aged mice. The unique particle counting capability of our method allows us to quantify the number of nanocarriers in the brain parenchyma of aged (18 to 22 mo, Fig. 7) and young (4 to 6 mo, Fig. 5) mice. Compared to young mice, we observed a decreased density of nanocarriers that crossed the BBB across different brain regions in aged mice (Fig. 7 A and B). The extrapolated nanoparticle number in the whole brain is approximately sevenfold lower in aged mice than that in young mice (Fig. 7 C and D). These results are consistent with a recent study that blood–brain transport of plasma proteins decreases with age via a shift from specific receptor-mediated transcytosis to non-specific caveolar transcytosis (72). As PBCA nanoparticles require specific binding of LDLR and LRP1 to trigger clathrin-dependent transcytosis and thus cross the BBB, the decreased expression of these receptors and clathrin in brain endothelial cells of aged mice (72) may underlie the molecular basis of the reduced blood–brain transport of nanocarriers. These quantitative comparisons between young and aged mice demonstrate the feasibility of using our method to study nanocarrier transport in various mouse models, facilitating the development of nanomedicine in a specific physiological or pathological context.

Fig. 7. Blood–brain transport of nanocarriers decreases with age. (A). Representative volume-rendered images of nanoparticles, vasculature, and nuclei across different brain regions in aged mice (18 to 22 mo). Circles indicate nanoparticles in the brain parenchyma away from blood vessels. (Scale bar: 20 μm.) (B) Nanoparticle density in different brain regions of young and aged mice (n = 33 and 32 fields of view for DG of young and aged mice, 31 and 31 fields of view for CA1 of young and aged mice, 30 and 31 fields of view for CCX of young and aged mice, 29 and 31 fields of view for THA of young and aged mice, from 4 young mice and 3 aged mice; mean ± SEM). (C) Summary of the number of nanoparticles observed and the extrapolated number of nanoparticles in the whole brain for 4 young mice and 3 aged mice. (D) Bar diagram for the extrapolated number of nanoparticles in the whole brain (n = 4 young mice and 3 aged mice; mean ± SD). Statistical significance is determined by the two-sided two-sample t test. *P < 0.05 and **P < 0.01. The results of young mice in B–D are from the same raw data as Fig. 5.

Conclusion

The development of nanomedicine for CNS diseases depends on the understanding of fundamental nano–neuro interaction, which has been hampered by the immense technical difficulty of imaging individual nanocarriers in a native tissue environment (SI Appendix, Supplementary Note 1). Here, we developed a SRS-based method to image nanoparticles in tissues with single-particle detectability, chemical specificity, and particle counting capability. Using this method, we observed single nanocarriers crossing the BBB, resolving the controversies on the basic premise in nanomedicine. In this work, we chose to study polysorbate 80–coated PBCA nanoparticles in healthy mice for its generality. The intact BBB serves as a standard model (10, 11), and it is also common in many CNS diseases (73, 74). Especially for highly infiltrative, “whole-brain” diseases such as glioblastoma multiforme, all patients have considerable tumor regions with an intact BBB: The BBB disruption occurs only in a localized region of the tumor core rather than the tumor margins (73). Therefore, drug delivery across the intact BBB is critical for effective therapeutics. Meanwhile, polysorbate 80–coated PBCA nanoparticles are the first successful and arguably the most studied nanomedicine for drug delivery to the brain. Its apoE-involved binding to LDLR and LRP1 has been a model for receptor-mediated transcytosis, and the more recent nanocarriers are mostly extending to other receptors through similar transcytosis mechanisms (8).

Our method exploits vibrational features of nanocarrier composition to image nanocarriers in tissues. This concept is generally applicable to other nanoparticles. For nanoparticles with moieties such as nitrile and alkyne that generate Raman peaks in the cell-silent region, our method can be readily used for label-free imaging by targeting these intrinsic moieties. This category includes the entire family of poly(alkyl cyanoacrylate) nanoparticles (75). For nanoparticles without these moieties, our method could possibly be used to image the nanoparticles by targeting a vibrational feature of the nanoparticles that is distinct from endogenous molecules. As long as the chemical composition of the nanoparticles is different from that of endogenous molecules, this vibrational feature could be distinct in terms of peak position, peak width, spectral shape, or polarization properties. Even if a distinct vibrational feature is not available or too weak, one can label the nanoparticles with small, minimally perturbative, vibrational tags such as C-D, alkyne or nitrile (53, 54), creating vibrational features. Noteworthily, our method only requires a vibrational feature that is distinct and strong enough; other aspects of nanoparticles such as surface coating or functionalization have little effect on the detection.

Our demonstration at the current stage can be further improved along several directions. First, the imaging throughput is limited by the multiple-frame-average acquisition due to the low signal of a single nanoparticle. Exploring large-cross-section Raman probes (51) to label nanoparticles may allow us to image whole brain sections. Second, we demonstrate our method only in 40-μm thin brain sections because the imaging depth is limited by light scattering. Combining our method with tissue clearing (76, 77) may extend the imaging depth to the millimeter scale. Third, the requirement of specialized instruments might hinder the widespread use of this method, although this situation has been improved by the commercialization of SRS microscopes.

Materials and Methods

PBCA nanoparticles were prepared using emulsion polymerization and characterized using dynamic light scattering and TEM. The animal experimental protocol (AC-AABN0554) was approved by the Institutional Animal Care and Use Committee at Columbia University. IV injections of nanoparticles and lectin were performed on adult mice. The mice were killed 1 h after nanoparticle injection. The brains were harvested, fixed, and sectioned for SRS and fluorescence imaging. Full details are given in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Click here for additional data file.

We thank Lingyan Shi, Lixue Shi, Bin Zhu, Zhilun Zhao, Xinwen Liu, Carli Canela, Elsy El Khoury, and Rongqin Li for discussion. We thank Beverly Shelton and Leslie Mueller for help with mouse experiments. We thank Jia Ma, Amirali Zangiabadi, and Manju Rajeswaran for help with characterization experiments. We acknowledge the use of facilities and instrumentation supported by NSF through the Columbia University, Columbia Nano Initiative, and the Materials Research Science and Engineering Center DMR-2011738. We thank support from the NIH (R01 EB029523) and the Chan Zuckerberg Initiative (Dynamic Imaging 2023-321166).

Author contributions

M.W., N.Q., X.G., and W.M. designed research; M.W. and N.Q. performed research; M.W., N.Q., X.G., X.L., and D.S. contributed new reagents/analytic tools; M.W. and W.M. analyzed data; W.M. oversaw the project; and M.W., N.Q., and W.M. wrote the paper.

Competing interests

The authors declare no competing interest.

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.

Supporting Information

This article is a PNAS Direct Submission.
==== Refs
1 V. L. Feigin , Global, regional, and national burden of neurological disorders during 1990–2015: A systematic analysis for the Global Burden of Disease Study 2015. Lancet Neurol. 16 , 877–897 (2017).28931491
2 S. Naqvi, A. Panghal, S. J. S. Flora, Nanotechnology: A promising approach for delivery of neuroprotective drugs. Front. Neurosci. 14 , 494 (2020).32581676
3 X. Dong, Current strategies for brain drug delivery. Theranostics 8 , 1481–1493 (2018).29556336
4 W. A. Banks, From blood–brain barrier to blood–brain interface: New opportunities for CNS drug delivery. Nat. Rev. Drug Discovery 15 , 275–292 (2016).26794270
5 W. M. Pardridge, The blood–brain barrier: Bottleneck in brain drug development. NeuroRX 2 , 3–14 (2005).15717053
6 Y. Chen, L. Liu, Modern methods for delivery of drugs across the blood–brain barrier. Adv. Drug Deliv. Rev. 64 , 640–665 (2012).22154620
7 R. Pandit, L. Chen, J. Götz, The blood–brain barrier: Physiology and strategies for drug delivery. Adv. Drug Deliv. Rev. 165 , 1–14 (2020).31790711
8 G. C. Terstappen, A. H. Meyer, R. D. Bell, W. Zhang, Strategies for delivering therapeutics across the blood–brain barrier. Nat. Rev. Drug Discovery 20 , 362–383 (2021).33649582
9 C. Saraiva , Nanoparticle-mediated brain drug delivery: Overcoming blood–brain barrier to treat neurodegenerative diseases. J. Controlled Release 235 , 34–47 (2016).
10 J. Kreuter, Drug delivery to the central nervous system by polymeric nanoparticles: What do we know? Adv. Drug Deliv. Rev. 71 , 2–14 (2014).23981489
11 Y. Hu, M. Hammarlund-Udenaes, Perspectives on nanodelivery to the brain: Prerequisites for successful brain treatment. Mol. Pharm. 17 , 4029–4039 (2020).33064009
12 J. Nicolas, S. Mura, D. Brambilla, N. Mackiewicz, P. Couvreur, Design, functionalization strategies and biomedical applications of targeted biodegradable/biocompatible polymer-based nanocarriers for drug delivery. Chem. Soc. Rev. 42 , 1147–1235 (2013).23238558
13 Y. Zhou, Z. Peng, E. S. Seven, R. M. Leblanc, Crossing the blood–brain barrier with nanoparticles. J. Controlled Release 270 , 290–303 (2018).
14 J. Kreuter, R. N. Alyautdin, D. A. Kharkevich, A. A. Ivanov, Passage of peptides through the blood–brain barrier with colloidal polymer particles (nanoparticles). Brain Res. 674 , 171–174 (1995).7773690
15 J. Kreuter, V. E. Petrov, D. A. Kharkevich, R. N. Alyautdin, Influence of the type of surfactant on the analgesic effects induced by the peptide dalargin after its delivery across the blood–brain barrier using surfactant-coated nanoparticles. J. Controlled Release 49 , 81–87 (1997).
16 S. E. Gelperina , Toxicological studies of doxorubicin bound to polysorbate 80-coated poly(butyl cyanoacrylate) nanoparticles in healthy rats and rats with intracranial glioblastoma. Toxicol. Lett. 126 , 131–141 (2002).11751017
17 J. Kreuter , Direct evidence that polysorbate-80-coated poly(butylcyanoacrylate) nanoparticles deliver drugs to the CNS via specific mechanisms requiring prior binding of drug to the nanoparticles. Pharm. Res. 20 , 409–416 (2003).12669961
18 A. Ambruosi, H. Yamamoto, J. Kreuter, Body distribution of polysorbate-80 and doxorubicin-loaded [14C]poly(butyl cyanoacrylate) nanoparticles after i.v. administration in rats. J. Drug Targeting 13 , 535–542 (2005).
19 B. Petri , Chemotherapy of brain tumour using doxorubicin bound to surfactant-coated poly(butyl cyanoacrylate) nanoparticles: Revisiting the role of surfactants. J. Controlled Release 117 , 51–58 (2007).
20 S. Wohlfart, A. S. Khalansky, S. Gelperina, D. Begley, J. Kreuter, Kinetics of transport of doxorubicin bound to nanoparticles across the blood–brain barrier. J. Controlled Release 154 , 103–107 (2011).
21 S. Wohlfart, S. Gelperina, J. Kreuter, Transport of drugs across the blood–brain barrier by nanoparticles. J. Controlled Release 161 , 264–273 (2012).
22 S. Gao , Polybutylcyanoacrylate nanocarriers as promising targeted drug delivery systems. J. Drug Targeting 23 , 481–496 (2015).
23 A. Zybina , Nanoparticle-based delivery of carbamazepine: A promising approach for the treatment of refractory epilepsy. Int. J. Pharm. 547 , 10–23 (2018).29751140
24 O. Maksimenko , Doxorubicin-loaded PLGA nanoparticles for the chemotherapy of glioblastoma: Towards the pharmaceutical development. Int. J. Pharm. 572 , 118733 (2019).31689481
25 I. Khalin , Dynamic tracing using ultra-bright labeling and multi-photon microscopy identifies endothelial uptake of poloxamer 188 coated poly(lactic-co-glycolic acid) nano-carriers in vivo. Nanomed. Nanotechnol. Biol. Med. 40 , 102511 (2022).
26 A. S. Joshi, V. Singh, A. Gahane, A. K. Thakur, Biodegradable nanoparticles containing mechanism based peptide inhibitors reduce polyglutamine aggregation in cell models and alleviate motor symptoms in a Drosophila model of Huntington’s disease. ACS Chem. Neurosci. 10 , 1603–1614 (2019).30452227
27 H. Wu , Sequential targeting in crosslinking nanotheranostics for tackling the multibarriers of brain tumors. Adv. Mater. 32 , 1903759 (2020).
28 C. Tapeinos, M. Battaglini, G. Ciofani, Advances in the design of solid lipid nanoparticles and nanostructured lipid carriers for targeting brain diseases. J. Controlled Release 264 , 306–332 (2017).
29 K. B. Johnsen , Targeting transferrin receptors at the blood–brain barrier improves the uptake of immunoliposomes and subsequent cargo transport into the brain parenchyma. Sci. Rep. 7 , 10396 (2017).28871203
30 B. dos Santos Rodrigues, S. Lakkadwala, T. Kanekiyo, J. Singh, Dual-modified liposome for targeted and enhanced gene delivery into mice brain. J. Pharmacol. Exp. Therapeutics 374 , 354 (2020).
31 J. D. Lewicky , Mannosylated glycoliposomes for the delivery of a peptide kappa opioid receptor antagonist to the brain. Eur. J. Pharm. Biopharm. 154 , 290–296 (2020).32717389
32 Y.-C. Kuo, I. H. Wang, R. Rajesh, Use of leptin-conjugated phosphatidic acid liposomes with resveratrol and epigallocatechin gallate to protect dopaminergic neurons against apoptosis for Parkinson’s disease therapy. Acta Biomater. 119 , 360–374 (2021).33189953
33 T. Lin , Blood–brain-barrier-penetrating albumin nanoparticles for biomimetic drug delivery via albumin-binding protein pathways for antiglioma therapy. ACS Nano 10 , 9999–10012 (2016).27934069
34 Y. Dou, D. Zhao, F. Yang, Y. Tang, J. Chang, Natural phyto-antioxidant albumin nanoagents to treat advanced Alzheimer’s disease. ACS Appl. Mater. Interfaces 13 , 30373–30382 (2021).34180234
35 J. Gu, K. Al-Bayati, E. A. Ho, Development of antibody-modified chitosan nanoparticles for the targeted delivery of siRNA across the blood–brain barrier as a strategy for inhibiting HIV replication in astrocytes. Drug Deliv. Trans. Res. 7 , 497–506 (2017).
36 S. Yu, X. Xu, J. Feng, M. Liu, K. Hu, Chitosan and chitosan coating nanoparticles for the treatment of brain disease. Int. J. Pharm. 560 , 282–293 (2019).30772458
37 S. Saleem, R. Banerjee, R. Rajesh Kannan, Chrysin-loaded chitosan nanoparticle-mediated neuroprotection in Aβ1–42-induced neurodegenerative conditions in zebrafish. ACS Chem. Neurosci. 13 , 2017–2034 (2022).35696319
38 S. Ruan, Y. Zhou, X. Jiang, H. Gao, Rethinking CRITID procedure of brain targeting drug delivery: Circulation, blood brain barrier recognition, intracellular transport, diseased cell targeting, internalization, and drug release. Adv. Sci. 8 , 2004025 (2021).
39 R. L. Cook , A critical evaluation of drug delivery from ligand modified nanoparticles: Confounding small molecule distribution and efficacy in the central nervous system. J. Controlled Release 220 , 89–97 (2015).
40 D. X. Medina , Optical barcoding of PLGA for multispectral analysis of nanoparticle fate in vivo. J. Controlled Release 253 , 172–182 (2017).
41 I. Khalin , Ultrabright fluorescent polymeric nanoparticles with a stealth pluronic shell for live tracking in the mouse brain. ACS Nano 14 , 9755–9770 (2020).32680421
42 I. Reimold , Delivery of nanoparticles to the brain detected by fluorescence microscopy. Eur. J. Pharm. Biopharm. 70 , 627–632 (2008).18577452
43 A. Vilella , Insight on the fate of CNS-targeted nanoparticles. Part I: Rab5-dependent cell-specific uptake and distribution. J. Controlled Release 174 , 195–201 (2014).
44 R. A. Yokel, Nanoparticle brain delivery: A guide to verification methods. Nanomedicine 15 , 409–432 (2020).31999236
45 K. Kucharz , Post-capillary venules are the key locus for transcytosis-mediated brain delivery of therapeutic nanoparticles. Nat. Commun. 12 , 4121 (2021).34226541
46 R. Münter , Dissociation of fluorescently labeled lipids from liposomes in biological environments challenges the interpretation of uptake studies. Nanoscale 10 , 22720–22724 (2018).30488936
47 R. M. Koffie , Nanoparticles enhance brain delivery of blood–brain barrier-impermeable probes for in vivo optical and magnetic resonance imaging. Proc. Natl. Acad. Sci. U.S.A. 108 , 18837 (2011).22065785
48 W. E. Moerner, D. P. Fromm, Methods of single-molecule fluorescence spectroscopy and microscopy. Rev. Sci. Instrum. 74 , 3597–3619 (2003).
49 C. W. Freudiger , Label-free biomedical imaging with high sensitivity by stimulated Raman scattering microscopy. Science 322 , 1857–1861 (2008).19095943
50 L. Wei , Live-cell bioorthogonal chemical imaging: stimulated Raman scattering microscopy of vibrational probes. Acc. Chem. Res. 49 , 1494–1502 (2016).27486796
51 F. Hu, L. Shi, W. Min, Biological imaging of chemical bonds by stimulated Raman scattering microscopy. Nat. Methods 16 , 830–842 (2019).31471618
52 D. Zhang, P. Wang, M. N. Slipchenko, J.-X. Cheng, Fast vibrational imaging of single cells and tissues by stimulated Raman scattering microscopy. Acc. Chem. Res. 47 , 2282–2290 (2014).24871269
53 F. Hu, S. D. Brucks, T. H. Lambert, L. M. Campos, W. Min, Stimulated Raman scattering of polymer nanoparticles for multiplexed live-cell imaging. Chem. Commun. 53 , 6187–6190 (2017).
54 S. Vanden-Hehir , Alkyne-tagged PLGA allows direct visualization of nanoparticles in vitro and ex vivo by stimulated Raman scattering microscopy. Biomacromolecules 20 , 4008–4014 (2019).31408325
55 J. D. Robertson , Purification of nanoparticles by size and shape. Sci. Rep. 6 , 27494 (2016).27271538
56 N. Thorball, FITC-Dextran tracers in microcirculatory and permeability studies using combined fluorescence stereo microscopy, fluorescence light microscopy and electron microscopy. Histochemistry 71 , 209–233 (1981).6165704
57 R. T. Robertson , Use of labeled tomato lectin for imaging vasculature structures. Histochem. Cell Biol. 143 , 225–234 (2015).25534591
58 D. Scherer, J. R. Robinson, J. Kreuter, Influence of enzymes on the stability of polybutylcyanoacrylate nanoparticles. Int. J. Pharm. 101 , 165–168 (1994).
59 C.-Y. Huang, Y.-D. Lee, Core-shell type of nanoparticles composed of poly[(n-butyl cyanoacrylate)-co-(2-octyl cyanoacrylate)] copolymers for drug delivery application: Synthesis, characterization and in vitro degradation. Int. J. Pharm. 325 , 132–139 (2006).16857330
60 Y. Mu, F. H. Gage, Adult hippocampal neurogenesis and its role in Alzheimer’s disease. Mol. Neurodegener. 6 , 85 (2011).22192775
61 W. Poon, B. R. Kingston, B. Ouyang, W. Ngo, W. C. W. Chan, A framework for designing delivery systems. Nat. Nanotechnol. 15 , 819–829 (2020).32895522
62 M. J. Mitchell , Engineering precision nanoparticles for drug delivery. Nat. Rev. Drug Discovery 20 , 101–124 (2021).33277608
63 J. M. Dietschy, S. D. Turley, Thematic review series: Brain Lipids. Cholesterol metabolism in the central nervous system during early development and in the mature animal. J. Lipid Res. 45 , 1375–1397 (2004).15254070
64 X. Gu , Clearance of two organic nanoparticles from the brain via the paravascular pathway. J. Controlled Release 322 , 31–41 (2020).
65 X. Zhuang, Spatially resolved single-cell genomics and transcriptomics by imaging. Nat. Methods 18 , 18–22 (2021).33408406
66 L. Chen , Tailored reconstituted lipoprotein for site-specific and mitochondria-targeted Cyclosporine A delivery to treat traumatic brain injury. ACS Nano 14 , 6636–6648 (2020).32464051
67 M. D. Sweeney, Z. Zhao, A. Montagne, A. R. Nelson, B. V. Zlokovic, Blood–brain barrier: From physiology to disease and back. Physiol. Rev. 99 , 21–78 (2019).30280653
68 G. Bieri, A. B. Schroer, S. A. Villeda, Blood-to-brain communication in aging and rejuvenation. Nat. Neurosci. 26 , 379–393 (2023).36646876
69 Y. Hou , Ageing as a risk factor for neurodegenerative disease. Nat. Rev. Neurol. 15 , 565–581 (2019).31501588
70 A. J. Farrall, J. M. Wardlaw, Blood–brain barrier: Ageing and microvascular disease—Systematic review and meta-analysis. Neurobiol. Aging 30 , 337–352 (2009).17869382
71 W. A. Banks, M. J. Reed, A. F. Logsdon, E. M. Rhea, M. A. Erickson, Healthy aging and the blood–brain barrier. Nat. Aging 1 , 243–254 (2021).34368785
72 A. C. Yang , Physiological blood–brain transport is impaired with age by a shift in transcytosis. Nature 583 , 425–430 (2020).32612231
73 J. N. Sarkaria , Is the blood–brain barrier really disrupted in all glioblastomas? A critical assessment of existing clinical data. Neuro-Oncology 20 , 184–191 (2018).29016900
74 K. S. Taskar , Lapatinib distribution in HER2 overexpressing experimental brain metastases of breast cancer. Pharm. Res. 29 , 770–781 (2012).22011930
75 C. Vauthier, A journey through the emergence of nanomedicines with poly(alkylcyanoacrylate) based nanoparticles. J. Drug Targeting 27 , 502–524 (2019).
76 M. Wei , Volumetric chemical imaging by clearing-enhanced stimulated Raman scattering microscopy. Proc. Natl. Acad. Sci. U.S.A. 116 , 6608 (2019).30872474
77 L. Shi , Highly-multiplexed volumetric mapping with Raman dye imaging and tissue clearing. Nat. Biotechnol. 40 , 364–373 (2022).34608326
