|Home | About | Journals | Submit | Contact Us | Français|
Noninvasive positron emission tomography (PET) imaging of reporter gene is combined with quantitative real-time polymerase reverse transcription (RT-PCR) method to study the time course of death and proliferation of stem cells transplanted in the myocardium.
Male murine embryonic stem cells (ESCs) were stably transfected with a mutant version of herpes simplex virus type 1 thymidine kinase (HSV1-sr39tk) reporter gene; 5×106 such cells were injected into the myocardium of female athymic rats. While the transplanted cells was monitored by in vivo 9-(4-[F-18]fluoro-3-hydroxymethylbutyl)guanine ([F-18]FHBG) PET imaging of the heart, their absolute number was estimated by RT-PCR from hearts harvested at 3–5 h, 24 h, days 4, 7, and 14 after transplantation.
(1) Forty percent of injected cells were retained in the heart while majority of injected cells were lost within a few hours after injection. Cell death was peaked at 24 h when 18% of donor cells retained in the heart were dead. (2) The substantial cell loss was reversed by significant proliferation of ESCs. This led to the recovery of cell number to 3.4 million (70% of injected dose) at day 4 and first visual observation of in vivo [F-18] signal in the heart. (3) A robust correlation (R2=0.9) between percent of injected dose per gram of tissue derived from in vivo PET signal and the number of donor cells estimated by RT-PCR was revealed.
The time course of transplanted stem cells surviving in the heart reveals a process of substantial cell loss within 24 h of injection and subsequent recovery of cell number through proliferation. Such proliferation can be noninvasively monitored by reporter gene imaging.
The time course of death and proliferation of stem cells after they are grafted in the myocardium would reveal their ability to survive and to regenerate in the harsh environment of a beating heart. A variety of imaging methods have been utilized to examine this time course noninvasively. Methods available to visualize cells fall into two categories : direct labeling and reporter gene approach. The former loads into stem cells an imaging-detectable probe that would remain inside the cell during tracking. The latter inserts into stem cells a gene that expresses an enzyme or receptor detectable by an imaging modality. The direct labeling method would be less sensitive to monitor the proliferation because cell division leads to dilution or loss of the imaging probe. In contrast, reporter gene approach can potentially overcome this limitation by stable integration into each stem cell [2, 3]. Nuclear imaging modalities due to their high sensitivity have been leading the effort of cell tracking in the heart: by transfecting stem cells with herpes simplex virus type 1 thymidine kinase (HSV1-tk), Wu et al. first demonstrated the feasibility of reporter gene approach to visualize the stem cells in the heart by positron emission tomography (PET) imaging  and to monitor their proliferation in the heart upon stable transfection . Other reporter genes such as human sodium/iodide symporter have also been examined for cardiac applications [5, 6]. However, the percentage of injected cells that retained and survived in the heart over time was not estimated in previous studies [3, 7] in which the imaging was typically initiated a few days after injection. Furthermore, the correlation between in vivo imaging signal and the absolute number of grafted cells surviving in the heart was not examined. Finally, the detection limit of the reporter gene-based method has not been reported and will be an important consideration when determining the feasibility of using this approach to detect a specific number of cells.
To address these issues, we have combined with the reporter gene approach a genomic real-time polymerase reverse transcription (RT-PCR) method to estimate the number of grafted cells in a sex-mismatched transplantation in which stem cells of male origin are grafted into a female recipient heart [8, 9]. This approach allows us to examine the time course of stem cell death and proliferation and is a necessary step to validate the reporter gene approach.
Murine embryonic stem cells (ESCs; R1 line) stably transfected with HSV1-sr39tk are designated as R1-sr39tk. Details of transfection, uptake of 9-(4-[F-18]fluoro-3-hydroxymethylbutyl)guanine ([F-18]FHBG) tracer by R1-sr39tk cells are provided in the Supplemental materials.
To improve the survival of grafted cells , R1-sr39tk cells were heat shocked by incubation at 43°C for 30 min 1 day prior to injection.
All animal procedures were approved by the local Institutional Animal Care and Use Committee. Adult athymic rats (female, 150–200 g) were purchased from Frederick Cancer Center (Frederick, MD, USA). Five million R1-sr39tk cells suspended in 100 µL serum-free culture media were injected in one spot into the mid-anterior wall of the left ventricle. An insulin syringe with a 29.5-gauge needle that is bent 45° at 3 mm from the tip was used for injection; the needle tip penetrated about 2 mm into the epicardium to deliver the cells.
Three animals were used for the autoradiographic study described in the Supplemental materials.
In a pilot study, PET imaging was performed at days 0 (3–5 h), 1, 2, 3, 4, and 7 (n=2 for each time point) after injection of 5 × 106 R1-sr39tk cells. This study identified that day 4 is the earliest time to detect in vivo [F-18] PET signal in the heart.
In the study of combined PET imaging with RT-PCR, rats were randomly assigned into five groups (n=6 in each group) for follow-ups at 3–5 h, 24 h, 4 days, 7 days, and 14 days after the injection of five million R1-sr39tk cells. Four rats from each group were euthanized for RT-PCR analysis after imaging. Based on the pilot study, PET imaging was performed at days 4, 7, and 14 only. The remaining two rats in each group were used to estimate the percentage of dead donor cells by Terminal dUTP Nick-End Labeling (TUNEL) staining. Fifteen additional animals (not receiving cell injection) were used to generate the standard curve for the RT-PCR assay. A total of 60 rats were used.
PET imaging was performed on a small animal PET (A-PET) scanner based on the Philips Mosaic System (Philips Medical Systems, Cleveland, OH, USA) [11, 12]. A discrete 2 × 2 × 10 mm3 L-YSO Anger-logic detector housed in a 21-cm diameter bore provides a transverse field of view (FOV) of 12.8 cm. A-PET operates exclusively in 3D volume imaging mode and images were reconstructed using the row action maximum likelihood algorithm . A spatial resolution of 2 mm in the central region of the FOV and system sensitivity of 5.45 cps/kBq was achieved. Animals were sedated by isoflurane (1% in oxygen at flow rate of 1 L/min) during imaging. Vital signs including electrocardiography and core temperature were monitored (SA Instruments, Stony Brook, NY, USA), and core temperature was maintained at 35–37°C using a heating pad.
[N-13]ammonia and [F-18]FHBG  were synthesized at the Penn Cyclotron Facility. [N-13]ammonia at a dose of 1.2±0.2 mCi in 0.5 mL saline was injected into the tail vein and the acquisition started immediately and lasted for 15 min. Thirty minutes after the injection of ammonia, [F-18]FHBG was injected and data acquisition started 1 h later and lasted for 30 min. To examine whether a larger tracer dose leads to earlier detection of injected cells, a range of FHBG dose from 1.5 to 2.75 mCi (representing an 83% increase) in 1 mL saline was used in the pilot study, while a fixed dose of 1.6±0.2 mCi [F-18] FHBG was used for imaging before the RT-PCR study.
In order to estimate the activity concentration present in the grafted cells at the time of their first visual observation on the PET images, we performed a simple phantom study using two groups of three thin wall capillaries with ID of 0.75 and 1.75 mm, respectively, placed in a cylinder. The activity concentration ratio between the capillary and background is 4:1, 6:1, and 8:1 for each capillary size (Fig. 2f). Dynamic scans for 1, 2, 3, 4, 5, and 30 min were collected.
PET images were analyzed using the AMIDE software (http://amide.sourceforge.net/). For the pilot study, each data set was examined to identify the [F-18] hot spot within the myocardial wall that is delineated by [N-13]ammonia in three orthogonal views. When such a hot spot was detected, an ellipsoidal volume of interest was placed to encompass the greatest dimension of the hot spot in three orthogonal views of the [F-18] FHBG images (VOIh) and was also placed on septal region of the myocardial wall (VOIref). Both VOIs have the same size. The image contrast was estimated by:
When no hot spot was detected, VOIh was placed on the mid-anterior region of the left ventricle (where cells were injected) while VOIref remained at the septum.
Phantom images from the 30-min scan were analyzed by placing a VOI on the hot spot (as VOIh) and also on the background (as VOIref) to calculate the contrast as defined above. The size of the VOI used for phantoms is similar to the VOI used for in vivo images.
Average counts per minute per voxel from the VOIh was converted to tracer activity (millicuries per milliliter) using a calibration constant obtained by scanning of a 50-mm sphere filled with known [F-18] activity. The tracer activity was divided by the injected dose to obtain percent injected dose per gram of tissue (%ID/g), an index of [F-18]FHBG uptake in grafted cells.
The design of the Taqman probes for the internal standard (Xbp1) and male Sry gene are described in Supplemental materials. The internal standard is necessary to correct variations in DNA loading in samples.
Triplicate samples of the following three categories were prepared: (1) To test the sensitivity and specificity, DNA extracted from donor cells was mixed with DNA extracted from female hearts at 0:1, 1:0, 1:3, 2:3, 1:30, 1:300, and 1:3,000 (w/w). (2) To obtain a standard curve between the numbers of donor cells versus the amount of cellular DNA determined by RT-PCR, 1, 5, 10, 20, and 40 million donor cells (in a pellet) were mixed with the heart tissue and homogenized together to extract DNA (Qiagen, Valencia, CA, USA). (3) To estimate the number of donor cells retained in the heart, the left ventricle was dissected and snap-frozen in liquid nitrogen. The frozen tissue was homogenized to extract DNA. DNA concentration was determined on a Lambda-20 spectrophotometer (PerkinElmer, Waltham, MA, USA).
Each sample contains 30 ng of DNA from homogenates, 1× Sry (or Xbp1) primer, 1× Taqman universal master mix, and water in a total volume of 20 µL. All samples were run together on 364-well plates placed in an ABI 7900HT Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Wells without any DNA were run as negative controls. Denaturing was initiated at 95°C for 10 min, and the following steps were repeated for 40 cycles: denaturing at 95°C for 15 s and annealing/extension at 60°C for 45 s.
The amount of DNA from donor cells in a given homogenate sample was estimated by multiplying total DNA extracted from the sample by the ratio of donor DNA/total DNA obtained from RT-PCR. The number of donor cells was then obtained from the calibration curve.
Ten-micrometer sections from three levels (with a gap of 30 sections) were cut from a slab tissue containing the injection site. TUNEL staining was performed on nine sections (three sections per level) per heart using an in situ cell death detection kit (Roche, Indianapolis, IN, USA) to label the 3′-end of fragmented DNA with fluorescein-conjugated dUTPs. Positive control slides with DNase treatment and negative control slides without Tdt were included. Nuclei were identified by 4′,6-diamidino-2-phenylindole (DAPI) counterstaining.
Under a Nikon fluorescence microscope, three FOVs per section were captured into the image analysis software (ImageJ, NIH, Bethesda, MD, USA). The number of donor cells in the FOV was estimated by applying an appropriate threshold to generate a binary image in which nuclei of individual donor cells were outlined and counted automatically by the software. The number of TUNEL-positive cells in the same FOV was counted manually. The percent of TUNEL staining positive cells was averaged over three to nine sections for each heart (the number of sections containing donor cells is fewer at earlier time points); two hearts were used for each time point.
Data was presented as the mean (±SD). A two-tailed Student’s t test was performed when comparing two groups and a P value of less than 0.05 was considered to be statistically significant.
The retention of FHBG in R1 cells stably expressing sr39tk was 40-fold of that in nontransfected R1 cells (Fig. s 1 of the Supplemental materials). Autoradiography and histological staining confirmed the presence of R1-sr39tk ESCs in the myocardium (Fig. s 2 of the Supplemental materials).
High specificity and sensitivity in the detection of male DNA was obtained by RT-PCR (Fig. 1): as few as 10 pg of male DNA (equivalent of two donor cells) was detected by Sry primer/probe which is unresponsive to DNA from the female heart tissues (Fig. 1a). In contrast, Xbp1 detects both male and female DNA regardless of their ratio (Fig. 1b). Consequently, threshold cycles (Ct) are inversely proportional to the ratio of male/female DNA in samples for Sry detection, but independent of this ratio for Xbp1 detection (Fig. 1c). A standard curve that relates the number of donor cells to the amount of DNA reveals a linear relationship with a strong correlation (R2=0.94; Fig. 1d).
The time course of the image contrast (Fig. 2) reveals that the contrast was well below 1 at days 0, 1, 2, and 3 while it increased sharply to 1.3 at day 4 when a hot spot in the heart was first visualized on [F-18] PET images. The hot spot indeed localized in the myocardium as revealed by overlaying the [F-18]FHBG and [N-13]ammonia images. The [F-18] signals became stronger and expanded at days 7 and 14.
To estimate the sensitivity of our instrument in the detection of a hot spot of 1–2 mm in diameter, a simple phantom measurement was performed (Fig. 2f). For a 1.75-mm diameter (the larger capillary) region, the detection sensitivity of our PET scanner is at a 6:1 uptake level, while for a smaller, 0.75-mm diameter region, it can detect about an 8:1 uptake level. It is also noted that the contrast value of 1.3 was obtained from the larger capillary with 8:1 concentration activity; the same contrast value was achieved for the hot spot on day 4 [F-18] images acquired at the same scanning time, suggesting that the uptake ratio between the hot spot and surrounding myocardium could reach or exceed 8:1 at day 4. Thus, the phantom study allows a simple scaling of our measured contrast and estimation of the in vivo uptake level.
The %ID/g was quantified from in vivo images acquired from days 4 to 14 and was presented in Fig. 3a. The time course of donor cells surviving in the heart suggests that two million cells (40% of the injected dose) were retained in the heart within 3–5 h after injection. The cell number recovered to 3.5 million (71% of injected dose) at day 4 when in vivo PET signal from donor cells was first detected. A robust correlation (R2=0.9) was observed between the number of donor cells quantified by RT-PCR and the %ID/g quantified from [F-18] PET images (Fig. 3b).
By labeling the free 3′-OH termini of cleaved genomic DNA, TUNEL assay detects both apoptotic and necrotic cells . As shown in Fig. 4, at 3–5 h after injection, the fraction of dead cells is small (1.3%). Cell death reached a peak of 17.8% at 24 h and subsequently dropped to 3.5% at day 4, 2% at day 7, and less than 1% at day 14. After correction for the percentage of dead cells, the number of surviving cells at each time point is summarized in Table 1.
The major findings of this study are the following. (1) The majority (60%) of injected cells were not retained in the heart after intramyocardial injection. Cell death that peaked at 24 h after injection further decreased the number of surviving cells. (2) The substantial cell loss occurred within 24 h after injection was reversed by significant proliferation of ESCs. This led to the recovery of the cell number to 3.4 million (70% of injected dose) at day 4 and first detection of in vivo signals. (3) A robust correlation (R2=0.9) between % ID/g derived from in vivo PET images and the number of donor cells estimated by RT-PCR was obtained.
The rapid loss of the majority of injected cells within hours after injection is not likely associated with cell death since the fraction of dead cells at 3–5 h was quite small (<2%). Instead, cells might be washed out into circulation and might efflux (back up) along the needle track. The dead space of the syringe and needle would retain a fixed number of cells each time during injection. Increasing the initial cell retention in the heart would be necessary to enhance the graft formation and could be achieved by delivering cells in a scaffold. Cell death peaked at 24 h after injection leading to a nadir point of donor cells surviving, 16% of the injected dose. The substantial cell loss within 24 h of injection is consistent with previous reports [8, 15] and led to inability to detect the surviving donor cells by PET imaging. The proliferation of ESCs, however, enabled the recovery of cell number to 70% of injected dose at day 4, when 3.4 million cells were surviving in the heart and the in vivo [F-18] signal was first visually detected. The robustness of this detection is supported by the phantom study, which suggests that the activity concentration ratio between the hot spot and background was ≥8. The detection threshold of stem cells is determined by the level of reporter gene expression in each cell and its access to the tracer. Aspects of our experimental design that would affect the threshold are summarized below. First, while the stable transfection allows long-term tracking and the contribution from each cell is accounted for, it results in a lower level of reporter gene expression per cell compared to transient transfection, which leads to multiple copies of episomal DNA in a single cell. Second, as the result of intramyocardial transplantation of cells, the tracer is not uniformly accessible to individual cells; and increase of tracer dose did not lead to earlier detection in our hands. These factors might have led to a relatively high detection threshold (3.4 million cells) obtained in this study compared to that achieved when the cells are transiently transfected or are directly labeled with a radioactive probe such as [Cu-64]PTSM . Since the dose of stem cells ranging from 0.5 to 10 million has been used in rat models of myocardial infarction [4, 17], the HSV1-sr39tk combined with [F-18]FHBG PET may have limited utility in tracking a small number of cells. However, it provides a true survival marker for monitoring transplanted cells once their number reaches the detection threshold.
Our study has limitations. First, while the %ID/g obtained from static images is strongly correlated with the absolute number of donor cells, a kinetic study is required to calculate the flux of the phosphorylation reaction catalyzed by the HSV1-TK enzyme. The flux is proportional to the enzyme concentration, i.e., the level of HSV1-tk expression, whereas the concentration of a product in the enzymatic reaction is not . Recent work from Green et al.  demonstrated that dynamic imaging combined with kinetic modeling is able to estimate the phosphorylation rate of HSV1-TK, and compared to %ID, such rate seems to better correlate with TK activity measured by the standard assay. Second, cardiac gating and attenuation correction were not applied in this study. Such procedures are expected to reduce errors in estimations of %ID/g. Finally, grafted ESCs undergo differentiation, which may lead to variations in reporter gene expression level; therefore, the reporter gene approach is more suitable for tracking a single population of cells such as ESC-derived cardiomyocytes.
In conclusion, quantitative RT-PCR method permitted the estimation of the absolute number of grafted cells surviving in the heart and revealed a process of substantial cell loss within 24 h of injection followed by recovery afterwards through proliferation. Once reaching the in vivo detection threshold, a robust correlation between cell number and % ID/g validates that the index derived from imaging can serve as a surrogate for the number of surviving cells in the heart.
We thank UPenn Small Animal Imaging Facility (SAIF) and Eric Blankemeyer for the technical assistance with the A-PET scanner and UPenn Cyclotron Facility and Drs. Alexander Schmitz and Richard Freifelder for the assistance with [F-18]FHBG production. The research was supported by NIH grants, R21EB-2473 and R01-HL081185 (RZ), and was funded, in part, under a grant with the Pennsylvania Department of Health (RZ). The department specifically disclaims responsibility for any analyses, interpretations, or conclusions.
Electronic supplementary material The online version of this article (doi:10.1007/s11307-009-0222-3) contains supplementary material, which is available to authorized users.