Dual-Axes Confocal Microscope

Figure 1 (A) Composite dual-axes reflectance image of the squamocolumnar junction with depth up to 1 mm. Vertical cross-sectional image at l = 1345 nm of human esophagus via biopsy shows squamous epithelium (left half of image). (B) Corresponding histology. From reference 2.

Barrett's esophagus is identified by its characteristic salmon pink appearance on white light endoscopy. Horizontal ( en face ) dual-axes confocal reflectance images, taken of normal esophagus and Barrett's esophagus with low grade dysplasia, are shown in Figures 2 and 3, respectively. The nuclei, in the case of Barrett's, are much larger and the reflectance from the cytoplasm appears more intense and heterogeneous than that of normal epithelium.

Figure 2 Horizontal image of normal esophageal mucosa with (A) dual axes reflectance image showing the cell nuclei (arrow) and cell membrane (arrowhead), scale bar 20 mm, and (B) corresponding histology. From reference 1.

Figure 3 Horizontal images of Barrett's esophagus show clearly resolved nuclei with (A) reflectance at d = 10 mm, scale bar 20 mm, and (B) histology.

Figure 4. Breadboard reflectance dual axes prototype. A broadband infrared (IR) source provides illumination through a 99/1 coupler into a single mode fiber (SMF1 ) and is focused by lenses CL1 and FL1 onto the scanning mirror (SM). An image is created from backscattered light (reflectance) collected by the lenses CL2 and FL2, and focused into the fiber (SMF2 ) The coherence gate is achieved by routing 1% of the illumination to the reference of an interferometer. Balanced detection (D1, D2 ) cancels out power fluctuations in the light source. From reference 2.

We demonstrate the feasibility of the confocal theta architecture with post-objective scanning to collect in vivo fluorescence images from small animal models with our tabletop prototype using live cells and Drosophila embryos and freshly excised mouse tissue that express GFP . GFP is a well established marker of gene expression that can target the behavior of molecules in intact cells and organisms. For example, GFP has been used to monitor the development of mammalian embryos and metastatic tumors in the host tissue. These are some of the research areas that we will target for our validation studies. In addition, the distribution of GFP within the cell can be post-translationally modulated by the biochemical environment and by protein-protein interactions. Also, two or more variants of GFP may be used to provide multiple labels of sub-cellular processes. Furthermore, these types of processes may be observed with two-photon imaging.

On the fluorescence images, the SNR will be determined by taking a ratio of the mean intensity and standard deviation of a 3x3 pixel array in the region of interest. The contrast ratio will be determined by the ratio of the means of a 3x3 pixel array of two adjacent regions of interest. The fluorescence images shown are the average of 9 frames collected at 10 second per frame for a total acquisition time of 90 seconds.

Living Cells and Organisms

The tabletope prototype was used to collect fluorescence images from live cells in culture. DSH-293 cells, transfected with enhanced GFP (S65T) expressed with a CMV promotor were grown in culture media on a chamber slide consisting of a #1 cover slip (160 m) under sterile conditions. A fluorescence image of these cells is shown in Fig. 1A, and demonstrates clearly defined nuclei and cells borders. The SNR of the cytoplasm is 13 ± 5, and the contrast ratio between the cytoplasm and nuclei is 1.6.m

In addition, we collected fluorescence images from live Drosophila melanogaster embryos expressing GFP fused to the developmentally regulated engrailed gene. Engrailed, driven by the UAS/GAL4 promoter-enhancer, is expressed on the outer surface of the embryo. Embryos in stage 10 and 11 were rinsed with tap water, and then soaked in 50% bleach to remove the chorion. The embryos were then placed in a chamber slide with a #1 cover slip, and covered with a drop of fluorescence mounting medium (Vectashield) for index matching. The embryos are cylindrical in shape with dimensions of approximately 200 m in diameter by 800 m in length. A series of transverse scans were collected at increasing axial depths in 10 m m increments through the thickness of the embryo, ranging from z = 0 to 200 m. An axial section at a depth of 100 m is shown in Fig. 1B (left), scale bar 50 m. There are 14 stripes of engrailed with single lines of cells marking the posterior parts of the segments. The SNR of the stripes is 15 ± 3, and the contrast ratio between the stripes and the embryo protoplasm is 115. A 3 dimensional reconstruction of the set of axial sections using volume modeling software is shown in Fig. 1B (right).

Figure 1. HEK cells (A) Reflectance image (B) Fluorescence image. Scale bar 50 mm. From reference 3.

Freshly Excised Mouse Specimens

Furthermore, the tabletope prototype was used to collect fluorescence images from live from freshly excised specimens of transgenic mice expressing GFP. First, we resected skeletal muscle from the quadraceps of a transgenic mouse that expresses GFP under the control of a b -actin-CMV promoter-enhancer. The fluorescence image shown in Fig. 2A is an average of 9 frames collected over a FOV of 500 m, scale bar 50 m. The individual muscle fibers and cell nuclei can be seen. The SNR of the cytoplasm is 4.8 ± 0.5, and the contrast ratio between the cytoplasm and nuclei is 3.5.

Figure 2. (A) Freshly excised skeletal muscle; (B) Embryo (day 12) from transgenic mouse, scale bar 50 mm. From reference 3.
We also collected fluorescence images from mouse embryos to demonstrate the feasibility of studying embryonic brain development. A pregnant dam was sacrificed at plug day 12. Several embryos were dissected from the uterine horns. The decidua was removed with a scalpel, and the embryo was placed in a chamber slide over a #1 cover slip. PBS was placed into the well to keep the embryo moist, and for index matching. The fluorescence image, shown in Fig. 2B, was collected over the rostral end of the embryo, and reveals neuronal cell bodies, scale bar 50 mm.

MEMS biaxial scanning mirrors are currently being fabricated for a 5-mm diameter dual axes confocal microscope at l = 780 nm for reflectance imaging. Ongoing work is being done to complete the design, fabrication, and assembly of a MEMS prototype with front-viewing optics to be contained within a 5-mm diameter package. This size is compatible with a special therapeutic upper endoscope (Olympus XT30) that has a 6-mm diameter instrument channel - the 5-mm system will be integrated into the Olympus XT30 for the first set of clinical experiments. The collection of reflectance images from normal and dysplastic mucosa in the esophagus during routine endoscopy and comparison with histopathological evaluation of standard pinch biopsy specimens taken from the same sites will be a primary goal of clinical evaluations.

Figure 1 depicts the basic elements of the miniature dual-axes confocal microscope. Radiation from and to the transmission and collection fibers, respectively, are collimated with a pre-aligned collimating optic. The collimated beams are focused with an annular parabolic mirror, which provides some insensitivity to misalignment. A biaxial scanning mirror is used to scan the transmission and collection beams under the surface of the tissue for optical sectioning. Our MEMS collaborators will fabricate, package and wire-bond the MEMS scanning mirror. The alignment of various micro-optical components will be accomplished in a step-wise procedure to insure diffraction-limited imaging performance.

Optical modeling software packages such as Zemax and ASAP will be used to optimize the micro-optics and to calculate the aberrations and imaging resolution. In addition, these programs can be used to determine the effects of mechanical misalignments on the imaging quality, which will provide tolerance specifications for the assembly of the device.