Tải bản đầy đủ - 0 (trang)
Methods: Two-Photon Fluorescence Imaging in Living Tissue

Methods: Two-Photon Fluorescence Imaging in Living Tissue

Tải bản đầy đủ - 0trang


3.1. Some Practical

Guidelines on How

to Improve Signal

Generation and

Its Detection

3.2. Application

of the Technique

3.2.1. Calcium Imaging

Functional Imaging Using Two-Photon Microscopy in Living Tissue


Use a proper cage: Sometimes scientists do not want to waste

time for proper shading of their microscopes or the stage of the

shading box made in the lab is just not stable enough. The

price for bad shading is high background noise that will

decrease image quality.

Water-cooled detectors are superior in some very low light


Use a local amplifier instead of sending weak signals through

long cables: high quality amplifiers with preferentially local AD

converters provide good measurement quality.

Novel GaAsP PMTs provide high QE and low noise.

Check the input aperture of the detectors: it should be large

enough for imaging scattered photons.

Order optimal filter sets matching the requirement of your

measurements. But beware: delivery of custom-made filters

takes time and prices might be high.

Detectors should be as close as possible to the objective (and

condenser, if applies).

Use the same PMT voltage for a given measurement type but

remember that changing from one measurement type to the

other might be useful. For example, high voltages are required

for line scans but lower voltages are preferred for z-stacks.

PMT values in a given measurement type could be optimized

for SNR in a measurement series by using background noise

SD and the peak amplitude of signal to calculate SNR as function of voltage. Some well repeatable signal such as backpropagating action potentials (bAPs) is ideal for these

measurements. Once the voltage is determined no further

changes are required.

Use lenses with high NA for optimal photon collecting but

large enough working distances to be practical.

Thermal drift and normal aging of materials can affect detection efficiency. Track and record detection efficiency at least on

a monthly basis, using standard samples (fluorescent beads).

Lasers are also subject to “aging.” Check the laser power at

objective level at least monthly.

Keep the working environment clean, thermally stable and—

above all—dust-free. Filtered air and thermally stabilized room

temperature are highly desirable.

Calcium imaging is probably the most common application of

functional two-photon microscopy (Fig. 4). Calcium imaging is

primarily used to study the spiking activity of a neuron or a small

network of neurons. Indeed, it consists of loading neurons with


I. Vanzetta et al.




laser light

800 nm,

80 Hz, < 100 fs

Imaging through

thinned skull

60 µm depth

Water immersion


Dye ejection




Thinned skull

90 µm depth

Stained area


300 µm


Spontaneous activity


0.2 ΔF/F

130 µm depth

Cell 1

Cell 2

Cell 3

20 µm

Cell 4

40 s


Skull removed

60 µm depth

130 µm depth

170 µm depth

240 µm depth

20 µm

Fig. 4. In vivo calcium imaging of neuronal populations. (a) Schematic drawing of the experimental arrangement. (b)

Images taken through a thinned skull of a postnatal day 13 mouse at increasing depth. (c) Spontaneous Ca2+ transients

recorded in a different experiment through a thinned skull in individual neurons (P5 mouse) located 70 mm below the cortical surface, from a region similar to that shown in (b). (d) Images obtained as in (b) in an experiment (P13 mouse), in which

the skull was removed before imaging. Reproduced with permission from (56). Copyright (2003), National Academy of

Sciences, U.S.A.


Functional Imaging Using Two-Photon Microscopy in Living Tissue


specific markers that bind to calcium, the marker acquiring its

fluorescent properties only upon binding. Since ionic calcium

channels open when the cell emits an action potential, the intracellular calcium concentration temporarily increases and spiking activity of the cell results in increased fluorescence signals. The typical

calcium response to a single action potential is a transient with a

rise time of ~20 ms, and a slow decay with a time constant as long

as ~1/4 s (Fig. 5), due to the slow evacuation of calcium ions back

toward the extracellular medium.

Using simultaneous electrophysiology recordings, it has been

shown that single action potentials can be detected in vivo (57).

Such resolution is possible only in cells with low enough spiking

rate. Otherwise, the smoothing effect of the slow calcium decays

prevents detection of single action potentials. Nevertheless, subsecond resolution can be obtained using specific deconvolution

techniques ((58), Supplemental Materials of (59)), and even in the

case of cells with high spiking activity, the monotonous (possibly

linear) relation between calcium levels and number of spikes allows

a precise quantification of the output activity of the cell. Note that

functional calcium signals have also been observed in astrocytes,

which probably play a role in the metabolic-vascular coupling to

neural activity (60, 61). With this respect, it is advisable to stain the

preparation with sulforhodamine, a specific glial marker, such as to

differentiate between those cells and neurons.

An advantage of two-photon calcium imaging as compared to

microelectrode recordings is that it allows to acquire signals simultaneously from many cells: imaging 50–100 cells can be achieved

without excessive difficulty and imaging of up to as much as 300

cells has been demonstrated using specific three-dimensional scanning techniques (4).

On the other hand, the use of these optic probes to monitor

electrical activity in the cortex has limitations. Simply getting them

into the cortex can be damaging, e.g., a mechanical stress occurs

when the dye is injected into cells or into the extracellular space

(tissue). Moreover, a number of side effects limit dye usage, including photobleaching, toxicity to the cell, and photodynamic damaging. An even more intrinsic limitation is that these probes act as an

artificial calcium buffer, because their very principle is to bind to

intracellular Ca2+ ions. Therefore, calcium probes can alter intracellular signaling pathways and ultimately induce an abnormal behavior of the cell.

For the above reasons, considerable research efforts are devoted

to improving the calcium markers. Commonly used, commercially

available, calcium markers include: Oregon Green (perhaps the

most used one), Fluo-4, Fura-2, and others. For a review on calcium imaging that enters the details of different markers, their

chemical machinery and their particularities see (62). A guide on

fluorescent proteins can be found in (63).


I. Vanzetta et al.


130 µm depth





10 µm


tau = 286 ms

0.2 ΔF/F

Cell 1

Cell 2

500 ms

Whisker deflection

0.05 ΔF/F


tau = 240 ms

Cell 2

0.5 s

0.4 ΔF/F

Trial 1

Trial 2

Trial 3

Whisker deflection

500 ms

Fig. 5. In vivo recordings of Ca2+ transients evoked by whisker deflection. (a) A high-magnification image of layer 2–3

neurons in vivo in the barrel cortex of a P13 mouse. (b) Line-scan recordings of Ca2+ transients evoked in two neurons by

a deflection of the majority of whiskers on the contralateral side of the mouse’s snout. The position of the scanned line and

the cells analyzed are indicated in (a). Note that the Ca2+ transients occurred 17–22 ms after the termination of the stimulus and therefore probably represent stimulus-offset responses. Here and in (c), the solid line represents a mono-exponential fit of the decay phase of the transient. (c) Ca2+ transients evoked in cell 2 during three consecutive trials. The top trace

is from the trial illustrated in (b). (Inset) A Ca2+ transient in a P14 layer 2–3 neuron evoked in vivo by single-shock electrical

stimulation (70 V, 0.2 ms) average of five consecutive trials. Reproduced with permission from (56). Copyright (2003),

National Academy of Sciences, U.S.A.


Functional Imaging Using Two-Photon Microscopy in Living Tissue


To minimize the buffering capacity of a dye, its concentration

in the cell has to be kept low, which goes at the expense of imaging

quality. Whatever the aimed compromise, it has to be considered

that the upper limit of imaging quality is attained at a dye concentration at which its calcium buffering capacity equals the intrinsic

calcium buffering capacity of the cell. See (4) for details, and in

particular on some relevant properties of the probe.

The problems resulting from invasive delivery of the dyes into

the cortex can be avoided using genetically encoded markers,

which, in addition, in certain cases even allow to selectively target

specific cell types. Large research efforts are therefore being devoted

to develop new ones. We will describe this promising approach in

more detail below.

Still, in many cases, neurons have to be stained with calcium

markers and there are multiple ways to do so, which we will detail

below. Particular care will be taken in detailing the popular method

of multicell bolus loading (MCBL) (56) used for studying small

networks of neurons.

(a) Filling of single cells

A common calcium imaging method is to stain a single neuron

by allowing the diffusion of the calcium marker directly into its

cytoplasm through a glass pipette. Since the dye will diffuse

into the whole arborescence of the neuron, this allows acquiring signals from different components of the same neurons. As

the method requires the user to punch a small opening on a

patch of the cell membrane, the technique is naturally applied

when using combined whole cell patch clamp and calcium

imaging, yielding correlated local calcium data and either local

dendritic or global, somatic electrophysiological recordings.

Examples are back-bAPs observed in the dendrites of a cell, as

shown on Fig. 7.

More sophisticated techniques of electroporation allow

delivering the marker in the extracellular medium in the cell’s

proximity, while an electric field is created that temporarily

breaks the cell membrane and allows the neuron to take up the

dye with minimal mechanical damage (64).

(b) Bolus loading in the extracellular medium

The so-called technique of bolus loading (56, 65) consists of

delivering the calcium marker into the extracellular medium

(tissue injection).

In such case, an acetoxymethyl (AM) ester is added to the

fluorescent molecule, which allows it to enter neighboring cells

through ordinary uptake mechanisms (62). Importantly, the

calcium-binding site of the molecule is masked; only inside the

cell this part is cleaved by enzymes and the molecule’s calcium

binding ability is activated.


I. Vanzetta et al.

A solution containing the dye is pressure-injected directly

inside the tissues through a glass micropipette. The complex

mechanisms briefly described above result in the staining of all

cells and cell processes inside a sphere of diameter ~300 mm,

including axons and dendrites of cells whose soma is even more


Since both neurons and astrocytes are stained using this

technique, it is useful to dissolve a second dye in the pipette

solution, sulforhodamine, which selectively stains astrocytes

(55). Using the appropriate set of optical filters to select the

appropriate fluorescence emission wavelength, sulforhodamine

fluorescence signals at 605 nm can be observed in a separate

channel from calcium fluorescence, which is usually around

520 nm.

The success of the staining depends strongly on a large set

of physiological parameters that are usually collectively termed

as the “quality of the preparation.” A detailed protocol can be

found in (65), and a video publication at the Journal of Visual

Experiments allows visualizing a calcium experiment in detail

(66). This preparation consists in three steps. First, the surgical

procedures: animal anesthesia, craniotomy (and in general

duraectomy), and building a chamber fixed on the skull such as

to stabilize the brain with agarose pressed by a cover glass.

Second, the injection of the fluorescent calcium marker, after

which, it takes about 1 h for the dye to stain the cortical region.

Third, two-photon calcium fluorescence signals can be

acquired. In Box 1, we highlight a number of points that critically affect the quality of the two-photon calcium signal upon

bolus-loading the sample. Box 2 shows an application of bolus

loading the hippocampal in toto preparation.

Box 1

Critical Steps for Two-Photon Imaging Using Bolus-Loaded

Ca2+ Probes

The quality of the acquisitions primarily relies on the quality of

the staining, and this requires the correct preparation of the dye

solution. Then, to allow normal dye uptake, a healthy and stable

physiological condition of the cells is of primary importance.

These and additional factors that contribute to the quality of

imaging are listed below.

Factors contributing to the quality of the staining:

– Dye preparation: as the dye is hydrophobic, it is necessary

to first dissolve it in a solution of DMSO/Pluronic

(Invitrogen, or can be prepared by dissolving



Functional Imaging Using Two-Photon Microscopy in Living Tissue


Box 1


20% Pluronic acid in DMSO); its mixing can be facilitated by a combination of vortexing, centrifuging and

sonicating; then it has to be diluted into a pipette solution, and mixed again.

– Quality of the surgery: mechanical or thermal shock to

the cortex during craniotomy has to be avoided. Use

cold ACSF to cool upon drilling the craniotomy/thinning the skull; do not touch the cortical surface; bleeding from the brain can severely affect the quality of

two-photon imaging; fast preparation is preferable, in

particular the dura mater removal (when needed), during which the brain surface is exposed.

– Careful injection: since the dye molecules and the

DMSO/Pluronic solution can be toxic to cells, it is preferable to minimize the quantity that is injected; on the

other hand, this decreases the quantity of fluorescence

signal; a good injection protocol is given in (65).

– Physiological conditions of the animal: monitor temperature, ECG, respiration, etc. to ensure that the animal is in

optimal physiological condition.

Factors contributing to the quality of imaging:

– Shot noise: the system should be optimized, in particular

for light collection efficiency; in general shot noise can

then be sufficiently reduced by increasing the laser power

(provided imaging is not too deep); do not reduce shot

noise below the level of other noise (such as pulsation) to

avoid unnecessary bleaching and phototoxicity due to

excessive (and unnecessary) dye excitation.

– Cardiac pulsation: open a small cranial window, maintain

the brain stable with low melting temperature agarose

that is applied in liquid phase and covered with a coverslip before its solidification; different groups have developed different solutions for building a chamber, in some

cases a gap is left free of coverglass, to allow insertion of

an electrode for injection through the agarose or for

simultaneous electrophysiological recording.

– Spatial resolution (and in particular, the z resolution) is

highly important to minimize the contamination of the

neuronal signals by the neuropil signals: best resolutions

are obtained with a large numerical aperture objective and



I. Vanzetta et al.

Box 1


complete filling of its back aperture by the laser beam (see

the optics part in this chapter for more details); to completely benefit from high numerical aperture, stain and

image in a region devoid of blood vessels as much as possible and do not image on the side of the craniotomy: consider that for a numerical aperture of 0.95, a laser beam

focused on a point situated at depth z below the surface, is

penetrating the surface through a disk of radius equal to

z—which thus should be as devoid of blood vessel as


– Compromise between temporal resolution, spatial resolution,

extent of the region imaged, and shot noise; theoretically, the

key factor here is the dwell time, i.e., the time spent in each

pixel. The more it can be reduced, the better. Indeed,

short dwell times allow acquiring more pixels per second,

thus allowing increasing either the rate at which images are

acquired, or the total number of pixels in the image, or

both. However, a lower limit on dwell results from the

need to collect enough light to overcome shot noise limitations. In practice, limitations on temporal resolution

arise also from scanning requirements (speed of scan, possibility to access random locations and thus gather signal

only from relevant points, e.g., neurons), and from the low

temporal resolution of calcium transients.

Box 2

Multicell Bolus Loading Applied to In Toto Hippocampal


Staining of neuronal populations is usually applied in vivo, but

there are no practical limitations to using the technique for

in vitro loading of cells (67). Here, we describe a method for

staining a large neuronal population in the whole hippocampal

preparation. The dye preparation and loading has been detailed

above and is well described in (10, 65). Figure 6 shows an example of this preparation, as through a two-photon fluorescence




Functional Imaging Using Two-Photon Microscopy in Living Tissue


Box 2


Fig. 6. In toto hippocampal preparation with MCBL and patch clamping. An example loading is shown below on

(a), where neurons are green (2 mM Oregon green BAMTA-AM) and glial cells are red (180 mM Supforhodamine-101).

(b) Shows a whole cell patch clamped neuron filled with 200 mM Fluo5 and 50 mM Alexa-594.

Extraction of the whole hippocampus:

Mice are briefly anesthetized with isoflurane and swiftly


The brain is gently but rapidly removed from the skull and

placed into a Petri-dish containing carbogen-bubbled, icecold ACSF (for composition, see (49)).

The hemispheres are separated and cut sagittally with a

surgical scalpel.

With the help of two plastic spatulae, the hippocampus of

one hemisphere is gently exposed by blunt dissection

starting from the third ventricle.

While gently keeping the brain in position, the hippocampus is disconnected from the rest of the brain at the septal

and dorsal areas (the septum can be kept by using gentle

dissection around it).

Using the long edge, one of the spatulae is placed under the

longitudinal axis of the hippocampus and used to gently roll

out the whole structure towards the cortex.

The long edge is pushed down to cut loose the preparation (along with the entorhinal cortex).

The whole hippocampus is sucked up using a wide mouth

Pasteur pipette and placed in a submerged tissue holding

chamber filled with ACSF at room temperature, bubbled

with carbogen.



I. Vanzetta et al.

Box 2


Advantages of using the in toto hippocampal preparation:

1. It can be placed in any conventional in vitro slice chamber and the dye loading can be performed. The process

does not require any modification to the in vitro microscope setups and the calcium imaging and patch clamping can commence as with brain slices.

2. As the preparation is quite translucent, both trans- and

epifluorescent photons can be collected, greatly increasing the SNR.

3. A major advantage and a major difference to brain slices is

that the in toto hippocampal preparation contains most of

the morphological connections that the hippocampus had

in vivo. Most of the axonal projections toward the septum

and the entorhinal cortex are intact or much less damaged

than in brain slices.

(c) Calcium-dependent autofluorescence obtained through

genetic modifications

Genetically encoded calcium indicators were developed recently

(68), and allow raising strains of animals with specific cell types

expressing Ca2+-sensitive fluorescent proteins. These calcium

probes are thus delivered in a minimally invasive manner, thus

allowing to image calcium dynamics in the intact brain by twophoton imaging through a thinned bone. A short but comprehensive review on genetically encoded sensors of neural activity

can be found in (69), whereas a more detailed review focusing

on calcium sensors and detailing their performances and optimization can be found in (70).

At the current stage of their development, the performance

of genetic indicators remains inferior to those of synthetic dyes

directly delivered to the cells with respect to several criteria,

including SNR, kinetics, linearity, photostability, and ion selectivity. However, it has been shown that genetic indicators can

be used in vivo (in the mouse cortex) to study Ca2+ dynamics

at single-cell and even subcellular resolution. Moreover, they

appear to allow detecting suprathreshold depolarizations consisting of as few as two to three action potentials (71, 72).

Additional advantages of genetic indicators are the possibility of selectively targeting specific types of cells, or even combining the genetic modifications expressing Ca2+-sensitive


Functional Imaging Using Two-Photon Microscopy in Living Tissue


fluorescence with modifications of other genes, e.g., involved

in the plasticity of the cell, or in neurodegenerative diseases.

Another way of using genetically engineered calcium indicators is to encode them into the genome of a virus, and then use

this virus to contaminate cells with the indicator. The use of

viruses that transfect from a cell to the next by crossing synapses is already quite well established for tracing synaptic connections and network connectivity in the cortex (73). Using

genetically encoded calcium sensors in conjunction with them

allows reporting of the activity of connected neurons (74).

3.2.2. 3D Imaging of

Dendritic Elements Using

Roller Coaster Scanning

The spatial and temporal resolution and S/N characteristic of

Roller Coaster Scanning are illustrated by the data shown in Fig. 7.

A spiny dendritic segment of a CA1 pyramidal neuron was followed by a 3D trajectory crossing 18 regions in 12 spines and the

neighboring mother dendrite at 150 Hz. Back-bAP induced Ca2+

transients in all of the total 18 regions had a SNR comparable to

that of the transients measured in one region by a single line scan

using the same conditions as a classical two-photon microscope.

Variability in the local dendritic geometry and misplacement of

the user-selected 3D trajectory results in inhomogeneities in the

basal fluorescence of the 3D measurement (Fig. 8a, b).

Therefore, during measurement of long dendritic segments a

real-time normalization for the first 500 ms (or whatever is available before the relevant event under study) of the raw 3D Ca2+

traces are required (Fig. 8c). For better visualization of the 3D

Ca2+ response, the trajectories are projected into two dimensions

(space along the dendrite and time). Ca2+ transients (Fig. 8d) were

Fig. 7. 3D two-photon dendritic imaging of a CA1 pyramidal cell at 150 Hz temporal resolution. (a) Dendritic segment of a

CA1 pyramidal cell. The z-stack was deconvolved and thresholded for surface fitting. The blue curve (light and dark blue)

shows the repeatedly scanned 3D trajectory. The size of the blue box is 15 × 15 × 37 mm3. (b) Enlarged view of the blue box

in (a). A total of 18 regions including 12 spines located on the 3D scanning trajectory were measured in one sweep, which

repeated at 150 Hz. The 12 spines were measured with one part (dark blue curve) of the trajectory while the whole parent

dendritic segment was measured with the second, approximately parallel running part of the 3D trajectory (light blue

curve). Ca2+ transients were induced by back-propagating action potentials (bAPs) elicited by somatic current injection

steps (5 APs, 35 Hz, average of 5 traces). Figure modified with permission from (49).

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

Methods: Two-Photon Fluorescence Imaging in Living Tissue

Tải bản đầy đủ ngay(0 tr)