Illuminating the lateral organization of cell-surface CD24 and CD44 through plasmon coupling between Au nanoparticle immunolabels

Abstract

CD44 and CD24 are important cell surface glycoproteins whose relative expression levels are used to identify so-called cancer stem cells (CSCs). While current diagnostic applications of CD44 and CD24 focus primarily on their expression levels, we demonstrate here that noble metal nanoparticle (NP) immunolabeling in combination with plasmon coupling microscopy (PCM) can reveal more subtle differences, such as the spatial organization of these surface species on subdiffraction limit length scales. We quantified both expression and spatial clustering of CD44 and CD24 on MCF7 and SKBR3 breast cancer cells through analysis of the labeling intensity and the electromagnetic coupling of the NP labels, respectively. The labeling intensity was well correlated with the receptor expression, but the inspection of the labeled cell surface in the optical microscope revealed that the NP immunolabels were not homogeneously distributed. Consistent with a heterogeneous spatial distribution of the targeted CD24 and CD44 in the plasma membrane, a significant fraction of the NPs were organized into clusters, which were easily detectable in the optical microscope as discrete spots with colors ranging from green to orange. To further quantify the spatial organization of the targeted proteins, we characterized individual NP clusters through spatially resolved elastic scattering spectroscopy. The statistical analysis of the single cluster spectra revealed a higher clustering affinity for CD24 than for CD44 in the investigated cancer models. This preferential clustering was removed upon lipid raft disruption through cholesterol sequestration. Overall, these observations confirm a preferential enrichment of CD24 in lipid rafts and a more random distribution of CD44 in the plasma membrane as cause for the observed differences in protein clustering.

Figure 1

40 nm Au NP immunolabeling facilitates surface expression profiling of CD44 and CD24 on SKBR3 and MCF7 cells. Darkfield scattering images of SKBR3 cells labeled for a) CD44 and b) CD24, and MCF7 cells labeled for c) CD 44 and d) CD24. Scale bars are 5μm

Figure 2

Au NP immunolabels form discrete spots on the cell membrane for different cell/protein targets: a) SKBR3/CD24, b) MCF7/CD24, c) MCF7/CD44. The magnified view in d) further emphasizes the “spotting” after labeling with NPs. e) SEM imaging confirms the organization of the cell-bound NPs into clusters of varying sizes (here for MCF7/CD24). f) Exemplary spectra of individual spots on the cell surface. Size bars are 5μm for a–c), 1μm for d–e).

Figure 3

Characterization of NP clustering through spectral analysis. a) Cumulative distribution plots of the fitted peak resonance wavelength (λ_res) of 40 nm Au NPs in an ambient refractive index of n_r = 1.40 and targeted at CD24 and CD44 on MCF7 and SKBR3 cells in HBSS. b) Cumulative distribution plots of λ_res for NPs targeted at MCF7/CD24 before (red) and after (blue) treatment with m-β-CD. For each individual condition approximately 200 individual measurements were evaluated in a) and 400 in b). Histograms of the fitted peak resonance wavelengths are included in Figure S7.

Figure 4

SEM characterization of NP cluster size distributions. Size distribution of a-i) randomly immobilized Au NP immunolabels collected after incubation with cells, a-ii) Au NPs targeted at CD44 on MCF7 cells, and a-iii) Au NPs targeted at CD24 on MCF7 cells. b) Neutravidin funtionalized Au NP cluster size distributions before (b-i) and after (b-ii) m-β-CD treatment of MCF7/CD24. The numerical percentages of each section are summarized in Tables S1&2.