|Home | About | Journals | Submit | Contact Us | Français|
An important goal in modern biology is to understand how molecular processes commonly studied at the cellular level give rise to physiological functions in complex tissues and organisms. Noninvasive imaging of gene expression patterns in whole animals could provide information critical to this end but current methods lack sensitivity and spatiotemporal precision. Enzymatic reporter systems detectable by magnetic resonance imaging (MRI) address these limitations by combining the relatively high spatial and temporal resolution of MRI with the ability of each genetically-expressed enzyme to generate many MRI-detectable product molecules.[1, 2] A challenge with imaging-based detection of some of the most popular reporter enzymes is the need to deliver MRI probes to their sites of action within cells. Here we describe a new reporter gene system for MRI that relieves this problem by harnessing an extracellular enzyme, the mammalian secreted alkaline phosphatase (SEAP).
SEAP is a truncated, secreted variant of placental alkaline phosphatase (PLAP), and is widely used as a stable and heterologously expressable reporter enzyme in conjunction with optically absorbant, fluorescent, or luminescent substrates. To detect SEAP activity optimally by MRI, we modified an existing sensor for adenosine (Ado), the product of SEAP’s hydrolysis of phosphorylated adenosine derivatives. In this system, the reporter enzyme is therefore detected through its generation of product molecules, as opposed to its activity on an MRI contrast agent directly. The approach is reversible by removal or degradation of Ado, nondestructive to the Ado sensor, and relatively fast, because SEAP substrates can be used at concentrations well above their Km values without affecting background MRI signal (Fig. 1a).
The Ado sensor we used is actuated by an Ado-binding DNA aptamer, which, in the absence of saturating Ado concentrations, crosslinks superparamagnetic iron oxide (SPIO) nanoparticles modified with reverse-complementary DNA segments.[5, 6] Ado-dependent disaggregation of the functionalized SPIOs modulates their ability to create contrast in T2-weighted MRI scans. If nanoparticles with diameter greater than ~50 nm are used, experimental and theoretical studies show that disaggregation accompanies an increase in T2 relaxation rate (R2 = 1/T2); smaller SPIOs show the opposite change in R2.[8-10] A prototypical Ado sensor formed from SPIOs with mean diameter of 106 ± 1 nm, as measured by dynamic light scattering (DLS), showed a 50% change in relaxation rate (R2 = 1/T2) at an Ado concentration (EC50) of 1.0 ± 0.2 mM (Fig. 1b). To improve on this apparent affinity, we used hybridization rules to predict thermodynamically favorable modifications to the probe, and found that weakening the crosslinking segment’s interaction with its 5′ SPIO-conjugated binding partner resulted in a new Ado sensor with a tenfold-improved EC50 of 91 ± 14 μM at the same sensor concentration (Fig. 1b and Supporting Information). Similar Ado dependence was reported by changes in ΔR2 or by changes in the ratio ΔR2/ΔR1, which was observed to be approximately independent of SPIO concentration (Suppl. Fig. 1).
The Ado sensor was insensitive to a number of SEAP substrates (Suppl. Fig. 2), including 2′-adenosine monophosphate (2′-AMP), a compound hydrolyzed to Ado by the enzyme. In the presence of 2′-AMP, the SEAP parent enzyme PLAP induced Ado sensor responses that were detected after about ten minutes by MRI (Fig 2a). R2 changes were not observed in the absence of substrate. A second enzyme, adenosine deaminase (ADA), was used to reverse the MRI contrast changes induced by hydrolysis of 2′-AMP (Fig. 2b). After addition of ADA, R2 returned to baseline values observed before addition of substrate (Fig. 2a, shaded area). This reversal was blocked by pentostatin, an ADA inhibitor, showing that the ADA reversal was dependent on deaminase activity of the enzyme. Both PLAP- and ADA-catalyzed changes observed by MRI were closely correlated (r = −0.91) to differences in the mean particle cluster sizes recorded by DLS from the same samples (Fig. 2c), consistent with the established relationship between T2 relaxation and cluster size for these SPIOs.[7, 8] DLS also enabled time-resolved measurements of the Ado sensor’s response to alkaline phosphatase activity. Fig. 2d shows that the complete sensor response took place within 7 minutes after addition of PLAP in the presence of 2′-AMP, while no discernable changes followed addition of BSA or enzyme in the absence of 2′-AMP. These results demonstrated that SEAP/PLAP activity could be detected using the improved Ado sensor, and that reversible, nondestructive use of the system is possible in the presence of ADA.
We next sought to test the system in a cellular context where SEAP could be applied as genetically-encoded reporter enzyme. SEAP was transiently expressed in nonadherent HEK-293 cells and its activity was measured in conditioned supernatants by performing MRI in the presence of 2′-AMP and the Ado sensor. Fig. 3a shows that the four day time course of R2 changes following SEAP transfection corresponded to independent measurements of SEAP activity using fluorescence-based assays performed (r = 0.88). Changes were reversible and consistent with the dynamic range of the sensor (Suppl. Fig. 3). As a further demonstration of the SEAP reporter system in cultured cells, the SEAP gene was placed under control of a tetracycline-inducible promoter and coexpressed with the tetracycline repressor. When 2′-AMP and the Ado sensor were applied to detect SEAP activity, R2 values from tetracycline-induced cells were significantly increased with respect to uninduced cells, and similar to cells constitutively expressing SEAP in the absence of the repressor (Fig. 3b). Again, an independent measurement of SEAP activity using an optical readout demonstrated that levels of the enzyme corresponded to the changes measured by MRI (Suppl. Fig. 4). These experiments demonstrated that genetic control of SEAP expression in cell culture could be effectively detected by MRI.
We have shown that reversible detection of an established secreted reporter enzyme, SEAP, is possible using an MRI contrast agent that selectively monitors products of SEAP-mediated hydrolysis of phosphorylated purines. The MRI sensor mechanism allowed tracking of SEAP expression induced by transient transfection and tetracycline-inducible gene regulation in cultured cells. The system generates strong T2-based contrast changes, does not involve cell delivery or catalytic destruction of contrast agents, and is both reversible and moderately fast, because of its product-dependent sensing mechanism. This form of detection, and the resulting reversibility, differ from earlier efforts to measure enzymatic activity using contrast agents and conjugates that are themselves modified by reporter or endogenous marker[11-14] enzymes. MRI detection of SEAP reporter activity could be useful in opaque cell or tissue culture environments where optical assays are unreliable. MRI-based assays may be particularly beneficial for screening applications where data from three dimensional sample arrays may be acquired in parallel. MRI measurements using the new system might also be effective for monitoring alkaline phosphatase activity in widely-used SEAP or PLAP-expressing tissue and animal models.[16, 17] In these complex contexts, implants containing the Ado sensor could be used to avoid interference by endogenous factors. Systemic mapping experiments may also be feasible, supported by the possibility of performing ratiometric ΔR2/ΔR1 measurements with the SPIO-based Ado sensor (Suppl. Fig. 1).
Ado sensors were assembled by conjugating biotinylated DNA strands to steptavidin-coated magnetic nanoparticles. MRI was performed on an Avance 4.7 T scanner. Detailed protocols are available as Supporting Information.
**We thank the Raymond and Beverley Sackler Foundation for their generous support. Additional funding was provided by NIH grant DP2-OD002114 to AJ.
Supporting information for this article is available on the WWW under http://www.angewandte.org
Dr. Gil G. Westmeyer, Departments of Biological Engineering, Brain & Cognitive Sciences, and Nuclear Science & Engineering, Massachusetts Institute of Technology, 150 Albany St., NW14–2213, Cambridge, MA 02139, Fax: (+) 1 617–253–0760.
Dr. Yves Durocher, NRC Biotechnology Research Institute, 6100 Royalmount Avenue, Montréal, Quebec, H4P 2R2.
Prof. Alan Jasanoff, Departments of Biological Engineering, Brain & Cognitive Sciences, and Nuclear Science & Engineering, Massachusetts Institute of Technology, 150 Albany St., NW14–2213, Cambridge, MA 02139, Fax: (+) 1 617–253–0760.