Biomedical Engineering Reference
In-Depth Information
of our quantum dot-labeling protocol are particularly emphasized in Fig.
6
because
the micrographs were taken in wide-fi eld nonconfocal mode which collects light
from the entire thickness of the tissue slice, unlike in confocal mode, where stray
light is physically excluded from the plane of focus. This is critical because, as dis-
cussed above, one of the biggest diffi culties associated with immunospecifi c quan-
tum dot labeling of neural cells is nonspecifi c interactions and clumping between
quantum dot particles that can produce false-positive results (Pathak et al.
2006,
2007
) . Methodologically, nonspecifi c binding proved to be more of an issue with the
methanol-fi xed samples; all samples shown in the results were fi xed using paraform-
aldehyde followed by Triton X-100 to remove excessive cross-linking of proteins
induced by fi xation. It is essential for the bulkier quantum dot conjugates, compared
to the smaller size of traditional fl uorophores, to experience a low amount of cross-
linking in the tissue in order to avoid clumping. It is a common mistake to assume
that quantum dot nanoparticles are smaller than fl uorescent dyes, when in fact they
are 10-20 times larger (depending on the color) than fl uorescein isothiocyanate
(FITC). Serially sectioned 10-mm slices throughout the thickness of the retina, of
which two consecutive slices are shown in Fig.
6a, b
, showed no observable nonspe-
cifi c labeling. These results are similar to control retinal sections labeled using tradi-
tional ICC with a primary antibody specifi c to the target antigen and an FITC-tagged
secondary antibody that binds to the primary antibody and acts as a fl uorescent
reporter (Fig.
7c, d
), although the FITC labeling was somewhat more qualitatively
diffuse and did show some degree of nonspecifi c labeling despite our best attempts.
Upregulation of GFAP in Müller cells and astrocytes occurs only under pathologi-
cal conditions and is considered the hallmark of the reactive glial response. GFAP
upregulation in Müller cells is particularly apparent because it spans the length of
their cell bodies throughout most of the thickness of the retina up to the inner limiting
membrane. In the rat laser-induced choroidal neovascularization (CNV) model we
used, gliosis and glial scarring occur as secondary processes and result in a strong
upregulation of GFAP. Figure
8
shows a confocal z-stack of a 10-mm tissue section
with an imaged slice thickness of 1 mm centered at a laser-induced lesion site. To the
best of our knowledge, these results represent the fi rst successful specifi c labeling in
situ of an intact neural tissue preparation. The intense upregulation of GFAP in both
Müller cells and astrocytes indicated a strong reactive response to the induced trauma.
Our quantum dot-labeling protocol was optimized to ensure even tissue penetration,
minimal nonspecifi c antigen labeling, and maximal specifi c antigen retrieval. Given
this, the fact that the upregulation in GFAP for all lesions we looked at extended over
a cross-sectional thickness of the retina of about 10 mm, as demonstrated in Fig.
8
by
the drop in fl uorescence signal in the confocal stack by slice 10, suggests that the
reactive volume of the neural retina in response to the laser-induced injury averaged
between 9 and 10 mm in cross-sectional width. The high signal-to-noise ratio of the
quantum dot-labeling procedure also putatively provides greater observable and
therefore measurable cellular detail throughout the volume of the glial response. In
the representative stack in Fig.
8
, the upregulation progresses from proximal Müller
cell processes near the boundary of the lesion site (progressively from slice 1 to 3) to
the entire length of the Müller cells and astrocyte layer near the center of the lesion
