1.0na objective

Multiphoton microscopy for patient specimens was performed at the University of Wisconsin Laboratory for Optical and Computational Instrumentation (Conklin et al, 2011). Imaging was done on a Prairie Technologies Ultima IV multiphoton microscope at an excitation wavelength of 890 with a Spectra Physics Insight laser. All images were collected with 20× 1.0 NA Objective (Olympus) and using a 445‐nm bandpass filter to discriminate for SHG collagen. SHG images at a resolution of 1,024 × 1,024 pixels were taken in triplicates, and two tumor border ROIs per image were analyzed using the CurveAlign (version 2.2 R2012a) software, covering 700–1,000 collagen fibers (curvelets) analyzed over 6 ROIs per sample. Curvelets with an angle against the border below 40° were assigned a score of 1, between 40° and 60° a score of 2, and above 60° a score of 3. The fraction of curvelets belonging to each of these three angle ranges was multiplied by the respective scores and added to obtain the final TACS‐3 score per patient. n = 18 for nulliparous, n = 28 for parous. For PAPP‐A and IGFBP‐5 immunohistochemistry, a score of 0 was assigned to an internal negative control stain in invasive regions, score 1 to regions with weak staining, and score 2 to regions with strong staining. A section was considered positive if more than 20% of cells displayed a score 2 stain.

For widefield imaging, a white light source (LED Engin LZ1-10CW00) was used for illumination. Illumination light and fluorescence signals were filtered through a GFP filter cube (460/50 excitation, 540/50 emission). A 4× 0.13 NA objective (Olympus) and a Photometrics Evolve 512 Delta camera were used to collect the emitted light. The field of view was roughly 5.5 × 5.5 mm². The frame rate was 20 Hz.
Two-photon Ca²⁺ imaging was performed with an Ultima system (Prairie Technologies) built on an Olympus BX61W1 microscope. A mode-locked laser (Coherent Chameleon XR Ti:Sapphire) tuned to 950 nm was raster scanned at 5 Hz for excitation while emitted GCaMP6s fluorescence was collected through a green filter (525/70 nm). Laser power at the sample was 20–80 mW. Dwell time was set to 2 μs. To increase imaging speed, resolution along the y-axis was reduced by a factor of 4. The final pixel size was 0.9 × 3.6 μm. A 20× 1.0 NA objective (Olympus) was used to yield a 465 × 465 μm² field of view. Imaging depths were between 150–350 μm.

Three to four weeks after virus injection, we performed two-photon glutamate imaging. The isolated retina was placed under the microscope (SliceScope, Scientifica) equipped with a galvo-galvo scanning mirror system, a mode-locked Ti:Sapphire laser tuned to 940 nm (MaiTai DeepSee, Spectra-Physics), and an Olympus 20 × (1.0 NA) objective. The retina was superfused with oxygenated Ringer’s medium. The iGluSnFr signals emitted were passed through a set of optical filters (ET525/50 m, Chroma; lp GG495, Schott) and collected with a GaAsP detector. Images were acquired at 8-15 Hz using custom software developed by Zoltan Raics (SELS Software). Temporal information about scan timings was recorded by TTL signals generated at the end of each scan, and the scan timing and visual stimulus timing were subsequently aligned during offline analysis.