Bacterial imaging is an emerging technology that has many health and environmental applications.1
For example, there is an obvious need to develop highly sensitive assays that can detect very small numbers of pathogenic bacterial cells in food, drinking water or biomedical samples. In other situations, the goal is to study in vivo
the temporal and spatial distribution of bacteria in live animals.
Optical imaging of bacteria in vivo
is much less developed than methods such as radioimaging and MRI. One approach is to use bacteria that are genetically encoded to produce luciferase or green fluorescent protein.2
A second strategy, which is the focus of this study, employs a molecular probe with a fluorescent reporter group. An obvious limitation with a live animal is restricted tissue penetration of the light. However, near-infrared (NIR) dyes with emission wavelengths in the region of 650−900 nm can propagate through two or more centimeters of tissue, and may enable deeper tissue imaging if sensitive detection techniques are employed.3
Compared to MRI and radioimaging, optical imaging provides a convenient method to monitor multiple biological processes simultaneously and in real time.4
Furthermore, the technology is operationally simple, amenable to miniaturization, and potentially mobile.
Molecular imaging probes can often be deconstructed into two structural components, an affinity ligand and a reporter group. In the case of bacterial targeting, previously reported affinity ligands include antibodies,5
bacteria binding peptides,7
and antibiotic drugs.10
Recently, we discovered that fluorescent molecular probes containing synthetic zinc(II) dipicolylamine (Zn-DPA) coordination complexes as affinity groups are able to selectively stain the surfaces of bacterial cells11
and apoptotic animal cells.12
Zn-DPA affinity ligands bind strongly to the anionic surfaces that are a common characteristic of these two cell-types, whereas affinity for the zwitterionic surfaces of healthy animal cells is weak. These in vitro
results have motivated us to pursue in vivo
studies, and we report that molecular probe 1
, which has a NIR fluorophore attached to an affinity group with two Zn-DPA units, can be used for targeted, fluorescence imaging of bacterial infection in a living whole animal.
The bacterial imaging probe 1
abs: 794 nm, em: 810 nm) was prepared in straightforward fashion using a carbocyanine dye as the NIR fluorophore.13
Researchers have incorporated this fluorophore into probes for other optical imaging applications.14 In vitro
fluorescence microscopy studies proved that probe 1
can effectively stain the periphery of bacterial cells (). In contrast, the cells are not stained when treated under identical conditions with control fluorophore 2
, confirming that the Zn-DPA affinity group is required for bacteria cell targeting.
Figure 1 Left: Phase contrast image of S. aureus cells treated with probe 1 (10 μM) and viewed at 1500X. Middle: Fluoresence image of the same cells acquired with a Cy7 filter set. Right: Fluorescence image of a live mouse after thigh muscle injections (more ...)
With these results in hand, we progressed to in vivo
imaging of nude mice that were infected with injections of Gram-positive Staphylococcus aureus
NRS11. Imaging was achieved by placing an anesthetized mouse inside a Kodak 4000MM Imaging Station configured for epi-illumination. The entire animal was irradiated with filtered light of wavelength 720±35 nm and an image of emission intensity at 790±35 nm was collected by a CCD camera during a 60 s acquisition period.15
First, we examined a series of mice with thigh muscle injections of bacteria that were pre-incubated with NIR probe 1
or control fluorophore 2
. As shown in , the fluorescence from the bacteria treated with 1
can easily penetrate through the skin and muscle tissue of the living animal. Fluorescence was not detected at the site that was injected with cells that were pre-incubated with 2
Next, we conducted in vivo
targeting experiments, where the bacteria were injected into the mouse thigh, and fluorescent probe was introduced into the bloodstream. The images in show the change in fluorescence intensity emanating from one mouse, out of a cohort of four, with a S. aureus
NRS11 infection (~5×107
Colony Forming Units in 50 μ
L Luria Bertani (LB) broth) in the left rear thigh. The opposite side of the mouse was injected with only the LB vehicle as a negative control. The infections were allowed to incubate for 6 hours, followed by introduction of probe 1
L of 1 mM aqueous stock solution) into the blood stream via a tail vein injection. Panels A and B show the mouse immediately before and following injection of 1
. The fluorescent probe clears slowly from the blood stream, except for significant accumulation at the site of bacterial infection. The maximum signal contrast is reached after about 18 hours (panels E and F). With this cohort of four mice, the in vivo
fluorescence intensity from the infected muscle after 21 h was 3.7±0.6 times higher than the contralateral control muscle.17
Similar imaging results were obtained with mice that were infected with Escherichia coli
(see supporting information
). As expected, control studies using mice injected with the fluorophore 2
failed to show localized emission from the site of bacterial infection (data not shown). These bacteria targeting experiments have been repeated successfully on several separate occasions using different cohorts of nude mice, independent batches of the probe 1
, and different workers performing the infection and imaging procedures. A residual fluorescence signal from the tail of the mouse is often observed, since it is the point of injection of 1
. Another notable observation is the mice (infected or uninfected) appear to easily tolerate the presence of probe 1
over the 21 h time-course of these experiments, as judged by their continued grooming and nesting activities.
Figure 2 Optical images of a mouse with a S. aureus infection in the left rear thigh muscle. Images were acquired before (A), and immediately following (B), intravenous injection of probe 1, and at 6 h (C), 12 h (D), 18 h (E), and 21 h (F). Scale represents the (more ...)
These in vivo
imaging results are in accord with our previous in vitro
observations that appropriately designed Zn-DPA coordination complexes have a strong and selective affinity for anionic bacterial cells over healthy animal cells.11
It is also possible that probe 1
is targeting the adjacent apoptotic/necrotic tissue that is produced by the bacteria's virulent action.18
Ongoing histology experiments will address this question.
In summary, the NIR fluorescent probe 1 selectively accumulates at bacterial infection sites in the thigh muscles of nude mice. Future studies will determine if it is possible to image more medically relevant infection sites that are within the tissue penetration limit for NIR light (e.g., skin, throat, urinary tract, etc). Another goal is to conjugate affinity groups containing Zn-DPA units to other types of imaging reporter groups or therapeutic agents and produce a range of useful probes for various in vivo imaging and pharmaceutical applications.