Health

Fluorescence Lifetime Imaging Microscopy (FLIM) reveals spatial-metabolic changes in 3D breast cancer spheroids

Cell lines and culture media

Metastatic MDA-MB-231 and non-tumorigenic MCF-10A human breast cell lines were purchased from American Type Cell Culture Collection (ATCC) and cultured using standardized media and conditions as previously described³¹. Briefly, MDA-MB-231 cells were cultured in DMEM (Corning, No. 10013CV), supplemented with 10% fetal bovine serum (ATCC, No. 302020). MCF-10A cells were cultured in DMEM/F-12 (ThermoFisher, No. 11330032) supplemented with 5% horse serum (Invitrogen, No. 16050122), 20 ng/ml EGF (Peprotech, AF10015; ThermoFisher, No. 10605HNAE), 0.5 mg/ml hydrocortisone (Sigma-Aldrich, No. H0888), 100 ng/ml choleratoxin (Sigma-Aldrich, C8052), 10 mg/ml insulin (Sigma-Aldrich, No. I1882). Both media recipes contained 1% penicillin/streptomycin (ATCC, No. 302300; ThermoFisher, No. 15140122). Cells were maintained at 37 °C and 5% CO₂ in a cell culture incubator.

Preparation of polydimethylsiloxane (PDMS) wells

Cylindrical PDMS wells (9 mm in diameter) were created inside 35 mm glass-bottom petri dishes (MatTek, No. P35G-0-10-C) using 3D printed self-centering cylinders as previously described³². Briefly, 1 ml of a liquid PDMS mixture (Dow Corning, Midland, MI) with a 10:1 weight ratio of silicon elastomer to curing agent was deposited inside the dishes and around the cylinders and cured at 50 °C for 2 h. After removal of the cylinder, the resulting cylindrical well was cleaned from PDMS residues and coated using poly-d-lysine (Sigma-Aldrich, St. Louis, MO) and glutaraldehyde, which provided an anchoring layer for collagen to avoid gel floating. After rinsing with 1 × PBS, the PDMS wells were sterilized under ultraviolet (UV) light for 30 min.

Spheroid formation

MDA-MB-231 and MCF-10A spheroids were generated as previously described³¹. Briefly, spheroids were generated by seeding approximately 1 × 10³ cells in each of the 96 wells of an ultra-low attachment plate (Corning, No. 7007) and allowed to form for 48 h in the presence of 2.5% v/v Matrigel. Once formed, individual spheroids surrounded by 5 µl of media were transferred onto coverslips inside PDMS wells (9 mm in diameter) created in 35 mm glass-bottom petri dishes (one spheroid per dish). Each spheroid was covered by 195 µl of ice-cold, rat-tail collagen I solution to achieve a total volume of 200 µl and a specific collagen concentration in each well. Collagen solutions were prepared by mixing acid-solubilized collagen I (Corning, No. 354249) with equal volumes of a neutralizing solution (100 mM HEPES buffer in 2 × PBS). The desired collagen concentration was reached by adding adequate volumes of 1 × PBS. Collagen solutions at different concentrations (1 and 4 mg/ml) polymerized for 1 h at 37 °C. The cell culture plates were rotated every minute for the first 10 min of polymerization to guarantee full embedding of the spheroid within the 3D collagen matrix. Finally, 2 ml of culture media (phenol-free, 50:50 v/v of MDA-MB-231 media to MCF-10A media) was added and the 3D organotypic culture was placed inside the incubator until taken out for FLIM measurements.

Multiphoton fluorescence lifetime imaging microscopy (FLIM)

Label-free multiphoton FLIM was used to monitor intracellular NAD(P)H and FAD fluorescence intensity and lifetimes of MDA-MB-231 and MCF-10A spheroids completely embedded in 3D collagen gels at different densities (1 and 4 mg/ml) over time (days 0 and 3). These spheroids were imaged at a depth of ~ 1–2 mm from the surface of the collagen gel in a single optical section. All spheroid samples were allowed to incubate in fresh phenol-free 50:50 culture media for a minimum of 3 h prior to imaging. FLIM was performed using an upright multiphoton microscope (Bruker, Ultima Investigator) equipped with a stage top incubator (Tokai Hit, STXF-UKX-SET) to maintain samples at 37 °C and 5% CO₂ with humidity during imaging.

A femtosecond titanium:sapphire tunable laser (Spectra-Physics, InSight X3) was used as the excitation source with a quarter-waveplate in the excitation path to circularly polarize the incident beam at the sample plane. The laser was tuned to either 760 or 880 nm for two-photon imaging of NAD(P)H or FAD, respectively. For collagen imaging, through second harmonic generation (SHG), the laser was tuned to 1050 nm. The emission was detected in a non-descanned geometry through a 16× long working distance (3 mm) water immersion objective (0.8 NA) (Nikon, CFl75 LWD 16X W), and separated by a dichroic mirror (Chroma, 700 nm long-pass). A 720 nm short-pass filter blocked residual excitation wavelengths in the detection path. NAD(P)H and FAD fluorescence emission was collected using a 440 ± 40 or 550 ± 50 nm bandpass filter (Chroma), respectively. A GaAs photon counting photomultiplier tube (Hamamatsu, H10770PB-50) with a time-correlated single photon counter (Becker & Hickl, SPC-150) was used to collect the temporal decays of NAD(P)H and FAD autofluorescence (120 s collection time per frame, 256 temporal bins per pixel). Images were collected at a scanning resolution of 0.8 µm/pixel and 10 µs pixel dwell time using 1024 × 1024 pixels (Fig. 1a).