| Optimizing In Situ Hybridization Protocols
Chris Carlson
Research Associate
Zhongting Hu, PhD
Senior Scientist
In situ hybridization (ISH), first introduced in 1969, is a method
to detect specific nucleic acid sequences within cells by hybridizing
a labeled RNA or cDNA probe to target transcripts in a tissue
sample (John et al., Gall and Pardue). Ever since its inception,
it has become clear that an optimized ISH protocol is particularly
important for detecting mRNAs. While some cells may express
a specific mRNA species at high levels (Hahn et al., 1982),
most functional mRNAs exist in small quantities. Therefore
the ISH technique must be sensitive and specific enough to
detect a few molecules in a single cell, as well as slight
variations in mRNA levels between cells.
Optimized ISH protocols share several common goals. These include
retention of tissue morphology, rendering tissue permeable to
probe, retaining target mRNA within the tissue, effective penetration
and binding of probes, and reduction of nonspecific background.
The critical parameters that result in successful ISH are type
of fixative and length of tissue fixation, method for embedding
fixed tissue, agents used for sample permeabilization, choice
of hybridization conditions, and post-hybridization treatment.
In this article, we provide some basic guidelines to simplify
the choices and manipulations of these parameters for high quality
ISH results.
Tissue Fixation
The purpose of tissue fixation is to preserve
tissue morphology and to prevent loss of mRNA from cells. Tissues
can be fixed
by either perfusion or submersion in fixative -- in most
cases perfusion is the method of choice. Perfusion is much better
at preserving tissue quality and RNA integrity because of the
rapid spread of fixative through the cells. In addition, perfusion
results in ISH data with low background due to clearance of blood
cells from the tissue. Fixation by submersion, on the other hand,
should be used when perfusion is not possible -- for example
with clinical samples or embryonic tissues.
The type of fixative used will also affect ISH sensitivity.
Two types of fixatives commonly used in ISH are precipitating
fixatives
and cross-linking fixatives. Precipitating fixatives (such
as ethanol/acetic acid or Carnoy's Solution) function by precipitating
proteins to trap the RNA inside cells. These types of fixatives
provide the best probe penetration. However, tissues fixed
by
precipitating fixatives are subject to loss of target mRNA
and the cell’s morphological structure (Lawrence and
Singer, 1985), resulting in poor ISH data quality.
Cross-linking fixatives rely on molecules that form covalent
bonds with reactive groups on proteins that are in close proximity,
thus providing better RNA retention and tissue morphology. Cross-linking
fixatives are therefore routinely used to fix tissue samples
for ISH studies. Gluteraldehyde, formalin, and paraformaldehyde
are the most commonly used cross-linking fixatives.
Buffered formalin solutions can lead to unpredictable RNA retention
due to RNA degradation. Gluteraldehyde fixation gives the best
RNA retention and tissue morphology, but probe penetration is
poor due to the strong cross-linking effect. The 4% paraformaldehyde
fixatives, conversely, provide a compromise between probe permeability
and RNA retention. We have found that paraformaldehyde fixation
provides adequate morphology retention and sufficient probe permeability
to make permeablization treatments (such as proteinase K; see
below) unnecessary. For most tissue samples, six hours to overnight
post-fixation in 4% paraformaldehyde after perfusion is sufficent
for obtaining good ISH results. Over fixation with 4% paraformaldehyde,
however, will significantly reduce ISH quality.
Tissue Embedding
Cryostat sections of frozen tissue and paraffin embedded tissue
sections have both been effectively used for ISH. In general,
paraffin-embedded tissues show better morphology than frozen
tissue. However, paraffin embedding requires more tissue processing
and can result in RNA loss and low ISH signal (Pintar and Lugo,
1985). Figure 1 compares hybridization signals obtained from
paraffin-embedded and cryostat sections from mouse brain. Significantly
stronger hybridization signals of vasoactive intestinal peptide
(VIP) mRNA were observed with the cryostat sections.
| |
|
|
| Figure 1. Cryosection
vs. Paraffin Section ISH.Comparison
of signal intensities of hybridized VIP mRNA in mouse brain
cortex on cryostat sections (A and C) and paraffin sections
(B and D). Results from both 33P labeled (A
and B) and digoxigenin-labeled probes (C and D) are shown.
Note that intensities of hybridized signals on cryostat
sections (arrows in A and C) are much stronger than those
on the paraffin sections (arrows in B and D). |
Figure 2 demonstrates that mRNAs in
cryostat sections are very stable when kept at -80°C. We saw no loss of VIP gene expression
in mouse brain sections that were stored at -80°C for
six months. Therefore, cryostat sections should be considered
as the first
choice for ISH studies. Paraffin sections should be used
with caution for ISH experiments on mammalian tissues where
sensitivity
is critical. However, paraffin sections still have particular
value in preparation of clinic and pathological samples for
long-term protection of tissue morphology.
| |
|
|
| Figure 2. Stability
of mRNAs in Cryostat Sections Kept at -80°C for Six Months. Probes
to VIP mRNA labeled with 33P (A and B) and digoxigenin
(C and D) were hybridized to fresh brain sections (A
and C) and sections preserved at -80°C for 6 months.
Note that signal intensities on fresh brain sections
(arrows in A and C) show approximately the same levels
as those on the sections stored for 6 months (B and D). |
Permeabilization of Tissue Samples
The permeabilization of tissue sections is a critical step in
ISH. The purpose of permeabilization is to facilitate probe diffusion
into and out of cells, i.e. to maximally expose target mRNAs
to probe (without loss of target mRNAs), and to reduce the nonspecific
binding of probes to tissue macromolecules. A permeabilization
step is required for samples that are over-fixed by paraformaldehyde
(but not for samples fixed overnight in 4% paraformaldehyde),
or fixed by glutaraldehyde. The most commonly used agents for
permeabilization are diluted HCl and proteinase K (PK).
The extent of permeablization required for a give tissue specimen
must be determined empirically. PK treatment for a short period
of time enhances ISH results in certain samples. Excessive PK
treatment, however, results in both loss of hybridization signal
and deterioration of morphology. Deproteination is a critical
step and special care needs to be taken to prevent excessive
loss of target mRNAs.
We have assessed permeabilization of brain tissues fixed in 4%
paraformaldehyde overnight and found that deproteination by PK
is either unnecessary or detrimental to RNA retention (Figure
3). This may be due to the fact that paraformaldehyde cross-links
the cell cytosol just enough to hold mRNAs in cells, but still
leaves RNA available for hybridization. Therefore, PK digestion
of the cell may result in loss of mRNAs. However, we noted that
addition of HCl diluted in triethanolamine increases detection
sensitivity in paraformaldehyde fixed samples, possibly due to
its ability to denature ribosomes, thus exposing additional target
mRNAs to probe.
| |
|
|
| Figure 3. Effect
of Proteinase K Treatment on ISH. The
photomicrographs show hybridized signals of VIP mRNA
on mouse brain sections treated by proteinase K for different
time. Note that as compared to that on the section without
proteinase K treatment (A), intensities of hybridized
signals (arrows) decreased with increased length of proteinase
K treatment (10 min (B), 20 min (C), and 30 min (D). |
Reduction of Nonspecific Binding
Minimizing background signal
is also a critical factor for ISH sensitivity. Background signal
arises primarily from nonspecific
retention of probe in tissue sections (due to electrostatic
interactions between probe and tissue macromolecules), and
entrapment of probes in the three dimensional lattice of the
tissue section. Several chemical functional groups in proteins
(such as amine and carboxylate groups) are believed to induce
this nonspecific binding. Scientists have tried to minimize
this source of background by treating tissue slides with acetic
anhydride and triethanolamine (Hayashi et al., 1978). Acetylation
of amine groups by acetic anhydride, routinely used in ISH
protocols, maybe important in reducing backgrounds for probes
larger than 2.0 kb (Lawrence and Singer, 1985).
Another way to decrease nonspecific probe binding is to saturate
the binding sites on proteins by incubating tissue with prehybridization
solution, which typically includes ficoll, bovine serum albumin,
polyvinyl pyrrolidone, and nucleic acids. These reagents are
also present in hybridization buffer to compete with the nonspecific
binding of probes to tissue. However, addition of the above reagents
to the hybridization buffer does not completely prevent background
signal. Nuclease treatment after hybridization is still necessary
for reducing this nonspecific signal (nuclease treatment degrades
unhybridized, single stranded probe). We have found that without
RNase treatment, the background with [33P]-labeled RNA probes
is so high that specific hybridization signal is not discernable.
Even high stringency washing did not remove this background (Figure
4). RNA probes tend to exhibit high levels of nonspecific binding,
so RNase treatment must be applied in ISH studies when RNA probes
are used (Lynn et al., 1983).
| |
|
|
| Figure 4. Effect
of Post Hybridization RNase Treatment on ISH. Photomicrographs
depict hybridization signals on brain sections without
(A and C) and with (B and D) post hybridization RNase treatment.
Panel C and D are dark-field views of panels A and D, respectively.
Without RNase treatment, background is very high and specific
hybridization signals cannot be identified (open arrows
in A and C). However, specific mRNA signals (arrows in
B and D) are present in sections treated with RNase. |
Probe and Hybridization
Antisense RNA probes are widely used for ISH because they have
been demonstrated to be more specific and sensitive than cDNA
probes (Cox et al., 1984). Both isotopically and nonisotopically
labeled probes have been used successfully for ISH. 33P and
35S are the most frequently used isotopes to label probes.
35S labeled RNA probes usually give higher backgrounds, so
when 35S labeled probes are used, dithiothreitol should be
added to all solutions used in prehybridization, hybridization,
and posthybridization washes. Based on our experience, we recommend
researchers to use 33P labeled probes for ISH studies since
they result in lower background and higher resolution as compared
to 35S labeled RNA probes.
Nonisotopic labeling systems (such as digoxigenin and biotin)
are also frequently used for ISH studies. We will address the
use of nonisotopic labeling systems for ISH in a future issue
of TechNotes.
While there are different recipes for making hybridization
buffers, the inclusion of dextran sulfate in the hybridization
solution
increases probe binding to target mRNA -- including 10%
dextran sulfate enhances ISH signal several fold. However, too
much dextran sulfate in the hybridization buffer will induce
high background, which is difficult to remove in post hybridization
washes.
back
to top
|
