A novel ternary heterostructure with dramatic SERS activity for evaluation of PD-L1 expression at the single-cell level

Abstract

Surface-enhanced Raman scattering (SERS) probes based on a charge transfer (CT) process with high stability and reproducibility are powerful tools under open-air conditions. However, the key problem ahead of practical usage of CT-based SERS technology is how to effectively improve sensitivity. Here, a novel ternary heterostructure SERS substrate, Fe₃O₄@GO@TiO₂, with a significant enhancement factor of 8.08 × 10⁶ was first synthesized. We found the remarkable enhanced effect of SERS signal to be attributed to the resonance effect of CuPc, CT between GO and TiO₂, and enrichment from a porous TiO₂ shell. In addition, we developed a robust SERS probe with good recyclability under visible light illumination on Fe₃O₄@GO@TiO₂ nanocomposites toward ultrasensitive detection of cancer cells down to three cells. We have now successfully applied this probe for in situ quantification and imaging of programmed cell death receptor ligand 1 (PD-L1) on triple-negative breast cancer cell surface at the single-cell level and for monitoring the expression variation of PD-L1 during drug treatment.

Fig. 1. The design and illustration of MGT.

Schematic illustration of the synthesis process and enhancement mechanism of the MGT substrate.

Fig. 2. Characterization of MGT nanocomposites.

(A) SEM image of the MG substrate. (B) TEM image of the MG substrate. Inset: Size distribution of MG by dynamic light scattering measurement. a.u., arbitrary units. (C) SEM image of MGT nanocomposites. (D) TEM image of MGT nanocomposites. Inset: Size distribution of MGT by dynamic light scattering measurement. (E) N₂ sorption isotherms and (F) corresponding pore size distribution curves of the MGT. (G) (I) Bright-field EDS images, (II) high-magnification image of (I) (marked with a green square), (III) high-resolution image of a select region in (II) (marked with a red square), and (IV to VIII) low-/high-magnification EDS images of MGT. (H) XRD patterns of (I) Fe₃O₄ NPs, (II) MG, (III) MGT, and (IV to VI) standard patterns of (IV) Fe₃O₄, (V) GO, and (VI) porous TiO₂ NPs. (I) Raman spectra and (J) FTIR spectra of (I) Fe₃O₄ NPs, (II) GO, (III) MG, (IV) Fe₃O₄@TiO₂ (MT), and (V) MGT. (K) The magnetic hysteresis loops at 300 K of the MGT. Inset: Photograph of the magnetic separation process of the MGT from the solution. emu, electromagnetic units.

Fig. 3. Enhancement effect of MGT.

Raman spectra of (A) CC, (B) PTCDA, (C) CV, and (D) CuPc on (I) Fe₃O₄ NPs (M), (II) MT, (III) MG, and (IV) MGT substrates under excitation of a 633-nm laser.

Fig. 4. Raman enhancement mechanism of MGT.

(A) XPS survey spectra of MG and MGT nanocomposites. High-resolution XPS spectra of (B) C 1s, (C) O 1s, and (D) Ti 3p of MG and MGT nanocomposites. TDR spectroscopic measurement for (E) MG, (F) MGT, and (G) CuPc after irradiation with a 633-nm laser flash. OD, optical density. (H) Time profiles of normalized transient absorption at 750 nm. (I) Two-dimensional integrated density difference plotted along the slab’s perpendicular axis. The bottom dashed line at 16 indicates the position of graphene, and the upper dashed line at 28 denotes that of the molecule. The red line indicates the charge difference on the GO and the green line denotes that on the TiO₂-supported GO. (J) Conclusion of calculated results and the mechanism in MGT platforms. (K) Raman spectra of CuPc on Fe₃O₄, TiO₂ (T), MT, MG, MGT with compact TiO₂ shell (cMGT), Fe₃O₄@TiO₂@GO (MTG), and MGT with the porous TiO₂ shell (MGT) substrates under a 633-nm laser (5 mW) with an exposure time of 0.5 s.

Fig. 5. Evaluation of PD-L1 expression on cell face.

(A) Site-specific recognition and detection of PD-L1 on TNBC cells. (B) Raman spectra of MGT-Abs-CuPc obtained with HCC38 cells at various concentrations. (C) Plot of Raman scattering intensity versus logarithm of cell numbers. Each data point represents the average value from six replicate SERS spectra (SD, n = 6). Error bars represent SDs. (D) Plot of I versus the effect of PD-L1 concentration (m) on the decrease of Raman intensity at 1530 cm⁻¹ (ΔI) of HCC38 cells on the cytosensor. (E) Raman spectra of MGT-Abs-CuPc obtained with MDA-MB-231 cells at various concentrations. (F) Plot of Raman scattering intensity versus logarithm of cell numbers. Each data point represents the average value from six replicate SERS spectra (SD, n = 6). Error bars represent SDs. (G) Effect of PD-L1 concentration (m) on the decrease of Raman intensity at 1530 cm⁻¹ (ΔI) of MDA-MB-231 cells on the cytosensor. (H) Raman spectra of MGT-Abs-CuPc obtained with MCF-7 cells at various concentrations. (I) Plot of Raman scattering intensity versus logarithm of cell numbers. Each data point represents the average value from six replicate SERS spectra (SD, n = 6). Error bars represent SDs. (J) Effect of PD-L1 concentration (m) on the decrease of Raman intensity at 1530 cm⁻¹ (ΔI) of MCF-7 cells on the cytosensor. (K) The expression of PD-L1 on the cell surface. (L) Raman scanning mapping images obtained from HCC38, MDA-MB-231, and MCF-7 incubated with 0, 4, 10, 16, and 20 ng of IFN-γ for 48 hours. (M) Raman intensity of the peak at 1530 cm⁻¹ obtained from HCC38, MDA-MB-231, and MCF-7 without stimulation. (N to P) Relationship between the Raman intensity and IFN-γ concentration for (N) HCC38, (O) MDA-MB-231, and (P) MCF-7 cells.