Guide to Plant-PET Imaging Using 11CO2

Abstract

Due to its high sensitivity and specificity for tumor detection, positron emission tomography (PET) has become a standard and widely used molecular imaging technique. Given the popularity of PET, both clinically and preclinically, its use has been extended to study plants. However, only a limited number of research groups worldwide report PET-based studies, while we believe that this technique has much more potential and could contribute extensively to plant science. The limited application of PET may be related to the complexity of putting together methodological developments from multiple disciplines, such as radio-pharmacology, physics, mathematics and engineering, which may form an obstacle for some research groups. By means of this manuscript, we want to encourage researchers to study plants using PET. The main goal is to provide a clear description on how to design and execute PET scans, process the resulting data and fully explore its potential by quantification via compartmental modeling. The different steps that need to be taken will be discussed as well as the related challenges. Hereby, the main focus will be on, although not limited to, tracing ¹¹CO₂ to study plant carbon dynamics.

Keywords: 11CO2; carbon-11 (11C); guide; image analysis; image quantification; plant-PET; positron autoradiography; positron emission tomography (PET).

FIGURE 1

Schematic of branch inside a PET detector ring, i.e., field of view (FOV). Positron decay of the (orange) ¹¹C-nucleus in the branch is shown in the enlarged circle. The positron is traveling a certain distance (typically 1.2 mm for ¹¹C-positrons - black arrow) known as the positron range to eventually collide with an electron and annihilate to produce two γ-photons (red arrows) traveling in opposite (180°) direction. Subsequently, these γ-photons can be detected by two different PET detectors (red ovals) in the detector ring.

FIGURE 2

Schematic showing the multidisciplinary steps in performing PET experiments on plants.

FIGURE 3

Schematic showing a potential set-up of an air circulation system of a PET experiment using gaseous ¹¹CO₂ (top) with specific examples to hermetically seal a plant part (bottom). Air flow is typically provided by a pump or another air controlling device to the plant tissue that will be labeled with ¹¹CO₂. Photosynthesis and transpiration can be obtained by a gas analyzer measuring CO₂ and H₂O content, respectively, of the incoming and outgoing air of the labeling system. Flow meters are used for detection of undesired leaks in the labeling system. At the end of the circulation system, the air can be scrubbed from ¹¹CO₂ before being released to the atmosphere. The bottom pictures show effective methods for enclosing part of a plant organ in a labeling bag or an acrylate feeding chamber, while hermetically sealing it from the atmosphere and other plant parts without damaging the tissue. The plant organ can for instance be enveloped by (multiple) small cylindrical flexible pieces of tubing, which are lubricated with petroleum jelly on the inside (bottom left). Straps can then be tightened upon the tubing (bottom middle) to close the labeling bag or fix the plant position in the acrylate chamber. Consecutive application of a stiff semi flexible and a soft flexible piece of tubing ensures airtightness without tissue damage when cable ties are tightened. Alternatively, malleable polysiloxan material can be used (bottom right).

FIGURE 4

Schematic representation of a true coincidence event and several image degrading effects in positron emission tomography, i.e., photon noncollinearity, scattered and random coincidence. In each case, the resulting line of response (LOR) that is registered by the detectors is shown.

FIGURE 5

Photograph of a plant-PET setup for a study on visualizing phloem transport in Arabidopsis (A). The rosette of the plant was hereby labeled with one pulse of ¹¹CO₂, while the inflorescence was positioned inside the field of view. The OSEM reconstructed time frame using 4 subsets and 30 iterations per subset is inserted in the right top corner. The region of interest, indicated by the arrow, was drawn on the reconstructed time frame and used to calculate the signal-to-noise ratio (SNR - Table 5). The effect of reconstruction algorithms MLEM and OSEM as well as the effect of a varying number of OSEM subsets X (indicated by OSEM X) on plant-PET data was investigated based on the convergence of the sum of all voxel values in the reconstructed time frame as a function of the number of iterations per subset (B). Stable convergence of the total voxel value implies that more iterations will not result in a qualitatively better image, on the contrary, the noise present can be amplified with further iterations.

FIGURE 6

Example of a static (upper left rectangle) and three dynamic PET images (timestamp in minutes shown in the lower left corner) of a Populus tremula branch that was exposed to gaseous ¹¹CO₂ during a 60-min PET acquisition. Transport of the label via the petioles to the branch is visualized by dynamic PET images. The static PET image (i.e., sum of dynamic images) has a better signal-to-noise ratio and can be used for drawing regions of interest (ROIs) around the branch or petiole. These ROIs can then be copied on the dynamic PET images to obtain tracer concentrations per ROI over time, i.e., time-activity curves (TACs).

FIGURE 7

Example of a static volume rendered PET image showing xylem-transported ¹¹CO₂ in a branch segment of Populus tremula (A). By extracting the tracer concentrations within, e.g., four consecutive ROIs of the corresponding dynamic PET images (e.g., 2.5 min time frames) time-activity curves (TACs) are obtained [circles in panel (B)]. Time is expressed in minutes after pulse-labeling aqueous ¹¹CO₂ to the cut end of the branch. By means of mathematical frameworks a model, representing the molecular system under study, can be fitted continuous lines) to the measured TACs. The importance of pursuing good region of interest (ROI) drawing practices is demonstrated by knowing which corresponding plant part is inside the field of view (C). It would be straightforward to draw ROI 1 on the branch segment having the highest tracer concentration [dotted ROIs in panel (A)]. On the branch segment enclosed in these ROIs however, a petiole originates, which cannot be resolved from the branch itself due to the limited spatial resolution of the PET system. Therefore, the ¹¹C-tracer detected in the branch and the petiole is added in these ROIs, resulting in an incorrect higher signal and eventually incorrect TACs [gray ROI measurements in panel (B)]. These ROI data sets were therefore excluded from parameter calibration and thus do not have a continuous model fit. Note that PET image (A) shows the side view whereas the branch (C) is shown from above.

FIGURE 8

Schematic of a simplified compartmental model used to describe xylem-dissolved ¹¹CO₂-tracer movement in a cylindrical region of interest (ROI) within a branch segment shown in Figure 7A. The model is described by two parameters, i.e., xylem CO₂ transport speed v_CO2 (mm min^–1) and exchange fraction a (min^–1) as defined by Eqs. (S3–4) (Supplementary file 2). Through sap flow, ¹¹CO₂ enters and moves through the xylem conduits (i.e., compartment 1) of each ROI with transport speed v_CO2. Within each ROI, ¹¹CO₂ can move from the xylem to surrounding chloroplast containing cells (i.e., compartment 2) through a, where it is assimilated by woody tissue photosynthesis and stored.