A 3D in vitro cancer model as a platform for nanoparticle uptake and imaging investigations

Abstract

In order to maximize the potential of nanoparticles (NPs) in cancer imaging and therapy, their mechanisms of interaction with host tissue need to be fully understood. NP uptake is known to be dramatically influenced by the tumor microenvironment, and an imaging platform that could replicate in vivo cellular conditions would make big strides in NP uptake studies. Here, a novel NP uptake platform consisting of a tissue-engineered 3D in vitro cancer model (tumoroid), which mimics the microarchitecture of a solid cancer mass and stroma, is presented. As the tumoroid exhibits fundamental characteristics of solid cancer tissue and its cellular and biochemical parameters are controllable, it provides a real alternative to animal models. Furthermore, an X-ray fluorescence imaging system is developed to demonstrate 3D imaging of GNPs and to determine uptake efficiency within the tumoroid. This platform has implications for optimizing the targeted delivery of NPs to cells to benefit cancer diagnostics and therapy.

Keywords: 3D in vitro model; X-ray fluorescence; cancer model; gold nanoparticles; imaging.

Figure 1

TEM image of a 3T3 fibroblast (left) and an HT29 cancer cell (right) incubated with 1.9 nm GNPs (incubation time 24 h). The cells appear viable; the dark appearance of the cytoplasm indicates presence of ground substances and a healthy state; organelles required for healthy cell functioning are present (such as the nucleus, endoplasmic reticulum and mitochondria). GNPs (black round spheres) appear either dispersed or as aggregates, mostly encapsulated in cytoplasmic vesicles.

Figure 2

TEM image of the cell membrane surface of an HT29 cancer cell incubated with 1.9 nm GNPs (incubation time 24 h). The image has captured each step of endocytosis: 1) GNP-membrane interaction, 2) membrane invagination to take up GNPs in to the cell, and 3) resulting endosome transports GNPs towards the lysosomes (4).

Figure 3

Quantitative XRF measurement of GNP concentration within 3D constructs. a) Measured spectra of 4 mgAu/mL GNP incubation dose uncompressed HT29 cellular construct and 0 mgAu/mL (control) acquired over 40 mins; b) Calibration curve relating XRF signal to GNP concentration. The XRF signal has been normalized to the Compton peak and acquisition time. A weighted linear fit (solid line) and boundary levels (dashed lines) are shown. The boundary levels were fit to fully include 95% of the data points and their error bars; the latter calculated using Poisson statistics. Measurements were made over a range of known GNP concentration solutions; c) GNP concentrations measured in 3T3 and HT29 3D uncompressed cell-populated collagen gel constructs over a range of initial incubation doses. The boundary levels of (b) were used to determine the error on each sample GNP concentration measurement, which was then added in quadrature to the error in the XRF measurement resulting from statistical fluctuations.

Figure 4

Tumoroid consisting of an artificial cancer mass (ACM):stromal surrounding GNP concentration ratio 5:1, the ACM of approximate concentration 0.02 mgAu/mL. a) Microscopic appearance (TEM) of tumoroid photographed in (b). The boundary between ACM and stroma is evident as a difference in collagen density, the greatest difference in collagen greyscale mapped as a dashed line to estimate the cancer/stromal boundary; collagen is more dense within the ACM (above dashed line) and stromal collagen less dense (below dashed line). GNP containing vesicles are visible within the cells.

Figure 5

XRF image of three different cross sections of a tumoroid composed of an ACM and surrounding 3T3-embedded collagen gel with a challenging GNP concentration ratio 5:1 between the ACM and surrounding. The blue-to-red colour scale presents gold concentration (red representing the highest gold concentration).