Gene reporters for magnetic resonance imaging
Kevin M. Brindle
Abstract
MRI-based gene reporters allow imaging of gene expression at depth (tens of centimetres) and at relatively high resolution (~10–100 μm) and have the potential to be translated to the clinic. The reporters exploit either endogenous contrast mechanisms or they modulate the response to an introduced exogenous contrast agent. MRI-based gene reporters allow imaging of gene expression at depth (tens of centimetres) and at relatively high resolution (~10–100 μm) and have the potential to be translated to the clinic. The reporters exploit either endogenous contrast mechanisms or they modulate the response to an introduced exogenous contrast agent. Most biologists are familiar with the concept of gene reporters, where a specific promoter is linked to a reporter gene, typically a gene encoding GFP and its variants or various luciferases, which are detected using fluorescence or bioluminescence imaging respectively. Reporters that are expressed constitutively can be used for cell tracking purposes whereas inducible reporters can be used to interrogate various signalling pathways and transcription factor activity. Detection of optical reporters in model organisms is limited by light scattering and absorption, which in the case of fluorescence restricts imaging to within approximately 1 mm of the tissue’s surface. To image deeper than this we need to turn to near-IR fluorescence, bioluminescence, ultrasound, photoacoustic, radionuclide (SPECT, PET), or magnetic resonance imaging (MRI), which in the case of MRI can be used to image reporter activity at depths of tens of centimetres with a spatial resolution in the 100-μm range and a temporal resolution that can be as high as a few tens of ms. The use of these techniques to image gene reporters has been the subject of recent reviews [1.Farhadi A. et al.Genetically encodable materials for non-invasive biological imaging.Nat. Mater. 2021; 20: 585-592Crossref PubMed Scopus (13) Google Scholar,2.Serganova I. Blasberg R.G. Molecular imaging with reporter genes: has its promise been delivered?.J. Nucl. Med. 2019; 60: 1665-1681Crossref PubMed Scopus (23) Google Scholar]. The aim of this forum article is to briefly chronicle the development of MRI-based gene reporters and to highlight some more recent developments.Many of the most sensitive MRI-based reporters are designed to affect the intensity of the tissue water proton resonance. The product of the reporter gene, which may be present in micromolar concentration, is detected indirectly via the intense resonance from the ~80 M tissue water protons. To understand how these reporters work it is necessary to have an understanding of the basic physical principles of NMR (Figure 1 and Box 1).Box 1NMR basicsThe nuclear spins of a spin ½ nucleus populate two allowed energy levels (Figure 1A), with the magnetic dipoles aligned with (lower energy level) or against (higher energy level) the applied magnetic field (B0). Because these energy levels are nearly equally populated Mz is very small (Figure 1B). The nuclear spins precess about the B0 field at a frequency (ω) that is a function of the nucleus’s gyromagnetic ratio (γ) and the magnitude of the B0 field experienced by the nucleus (ω = γB0) (Figure 1B). The application of an oscillating magnetic field resonant with this precession frequency (ω) tips Mz into the xy-plane (Mxy), where it induces a signal in the detector coil (the same coil can be used to apply the excitation pulse). Fourier transformation of this decaying signal converts the time domain signal into a frequency domain spectrum, where the amplitude of the signal is proportional to the number of spins and the frequency depends on their chemical environment, which determines the effective B0 field experienced by the nucleus. The exponential loss of phase coherence in the xy-plane, resulting in the decay of Mxy, is described by the T2 or spin–spin relaxation time (which contributes to the resonance linewidth in the spectrum), while restoration of the magnetization along the z-axis (Mz) is described by the T1 or longitudinal relaxation time. Loss of phase coherence is also caused by magnetic field inhomogeneity; however this additional loss of phase coherence can be refocussed by the 180° pulse in a spin echo experiment (Figure 1C). Magnetization created in the xy-plane by the 90° pulse is refocussed by the 180° pulse to produce an echo at time 2τ. In a diffusion-weighted experiment, magnetization will be dephased by the first applied magnetic field gradient pulse (G) and will be rephased by the second gradient pulse provided that it does not diffuse to a different part of the gradient in the intervening period (Δ). Loss of signal intensity with increasing gradient amplitudes can be used to determine the diffusion coefficient of the nuclear spin (Figure 1D). Magnetization can also be lost by diffusion of spins through endogenous magnetic field gradients. If two nuclei with different frequencies (chemical shifts) are in slow exchange such that they have separate peaks in the spectrum (Figure 1E), selective saturation of one of the resonances (equalization of the spin populations) will result in loss of intensity of the other resonance as it loses z magnetization to the saturated resonance without receiving any magnetization in return (Figure 1F). The magnitude of this decrease in intensity can be used to estimate an exchange rate constant.All of the NMR contrast mechanisms described in Figure 1 have been exploited in the development of MRI-based gene reporters. For reporters that can be used in vivo their development can be traced back to two landmark papers published in 2000. Louie et al. [3.Louie A.Y. et al.In vivo visualization of gene expression using magnetic resonance imaging.Nat. Biotechnol. 2000; 18: 321-325Crossref PubMed Scopus (1013) Google Scholar] microinjected into Xenopus embryos a reporter expressing β-galactosidase and a gadolinium-based contrast agent in which the Gd3+ ion was shielded from solvent water by a galactopyranose blocking group. Cleavage of the sugar residue by β-galactosidase exposed the cell water to the Gd3+ ion, which by virtue of its seven unpaired electrons is paramagnetic, leading to a reduction in the water proton T1 and an increase in signal intensity. The decrease in T1 increases the signal intensity by reducing the degree of signal saturation. However, the requirement to microinject the contrast agent coupled with the relatively modest enhancement of signal intensity limited the further use of the reporter. Weissleder et al. [4.Weissleder R. et al.In vivo magnetic resonance imaging of transgene expression.Nat. Med. 2000; 6: 351-354Crossref PubMed Scopus (747) Google Scholar] used an engineered transferrin receptor, which mediated cell uptake of human holotransferrin conjugated to iron oxide nanoparticles. Implanted subcutaneous tumors in mice that expressed the reporter showed substantial decreases in signal intensity in T2*-weighted images following intravenous injection of the modified transferrin; the intracellular presence of the superparamagnetic iron oxide nanoparticles decreasing the water proton T2*. In a spin echo image (Figure 1C) loss of phase coherence and the resulting signal loss is due to both T2 relaxation, which is irreversible, and magnetic field inhomogeneity, the sum of which is given by T2* (1T2*=1T2+1Tinhom.). Loss of coherence due to inhomogeneity can be reversed by the 180° pulse in the spin echo experiment unless there is diffusion through the applied (as in Figure 1C) or endogenous magnetic field gradients, such as those created by iron oxide nanoparticles. The signal intensity changes produced by transferrin-conjugated nanoparticles persisted for up to 14 days, which means that it would not be possible to monitor rapid changes in gene expression, and the requirement to inject a macromolecular contrast agent limits use of the reporter to those tissues, such as tumors, that have a leaky vasculature that allows access of the contrast agent. Subsequent studies that used the iron storage protein ferritin as the reporter removed the requirement to inject a contrast agent [5.Genove G. et al.A new transgene reporter for in vivo magnetic resonance imaging.Nat. Med. 2005; 11: 450-454Crossref PubMed Scopus (383) Google Scholar,6.Cohen B. et al.Ferritin as an endogenous MRI reporter for noninvasive imaging of gene expression in C6 glioma tumors.Neoplasia. 2005; 7: 109-117Crossref PubMed Scopus (293) Google Scholar]. However, as with the transferrin receptor, the expression of ferritin produces negative contrast in T2*-weighted images, which can lack specificity; for example, in tumors negative contrast can also arise due to haemorrhage. Moreover, the changes in image contrast relatively gene reporter that can high levels of contrast is one on the protein which is expressed in and the uptake of a contrast agent of in in a nearly increase in signal intensity in images, which at and to of reporter the Gd3+ ion for the allowed imaging with The also uptake of bioluminescence by a factor of bioluminescence imaging of its expression in vivo et gene reporter for in vivo A. PubMed Scopus Google and the uptake of which allows near-IR fluorescence imaging of reporter expression et as a reporter gene for fluorescence and magnetic resonance PubMed Scopus Google this a and image it the injection of a contrast which may not be to of those expressing the reporter. reporters described which the use of an contrast exploit in the diffusion coefficient of tissue water in the and The reporters, A. et imaging using reporter water 7: PubMed Scopus Google Scholar] and the et of increases in water exchange of a gene Biotechnol. PubMed Scopus Google rapid exchange of water the expression was detected from of an increase in the using the pulse in Figure whereas expression was detected by the water exchange rate by the water at various of the proton magnetization of reporters that not the injection of an exogenous contrast agent the chemical exchange saturation reporter expression is detected via the exchange of with water (Figure et reporter gene MRI contrast on proton Biotechnol. PubMed Scopus Google Scholar]. which are produced by some as the magnetic field in the same that iron oxide nanoparticles and can be detected in T2*-weighted The in magnetic the and the are for their of the magnetic the negative MRI contrast created by the can be from endogenous of negative contrast by the with A. et imaging using reporter water 7: PubMed Scopus Google Scholar]. that been with a reporter expressing the for to be in vivo using A. et imaging of gene expression in 2019; PubMed Scopus Google Scholar] and also be detected by MRI et magnetic resonance imaging of protein Mater. PubMed Scopus Google of the reporters have been in that they not allow the of reporters Reporters detected using have the potential to be they to the of contrast that have different chemical In a recent example, engineered to have expression as reporters to the and of their which the on these different chemical that reporter be detected using et designed MRI reporters for noninvasive imaging of transgene expression.Nat. Biotechnol. PubMed Scopus Google Scholar]. can also be in that their and determine their NMR such that they can be from other using MRI et magnetic resonance imaging of protein Mater. PubMed Scopus Google Scholar]. their determines their they can also be by the intensity of the used to resonance is a relatively the two energy levels for a spin ½ nucleus are nearly equally populated and the magnetization (Mz) with the of spins is very small (Figure spin techniques the of which can result in a in to can as for and can be detected at using where saturation of the resonance from in the in a decrease in the intensity of the which has a different chemical from different showed different chemical for in the the of et al.Genetically reporters for magnetic resonance imaging.Nat. 6: PubMed Scopus Google Scholar]. Detection of in vivo using would of and that this is this has to be MRI-based gene reporters have been of which not the injection of an exogenous contrast agent and some of which allow imaging with different contrast of of these reporters is that in they produce images with relatively resolution with optical However, the of MRI to image at depth means that for some of these reporters there is potential for to be used in as has been for where a gene reporter was used to in I. Blasberg R.G. Molecular imaging with reporter genes: has its promise been delivered?.J. Nucl. Med. 2019; 60: 1665-1681Crossref PubMed Scopus (23) Google of are Most biologists are familiar with the concept of gene reporters, where a specific promoter is linked to a reporter gene, typically a gene encoding GFP and its variants or various luciferases, which are detected using fluorescence or bioluminescence imaging respectively. Reporters that are expressed constitutively can be used for cell tracking purposes whereas inducible reporters can be used to interrogate various signalling pathways and transcription factor activity. Detection of optical reporters in model organisms is limited by light scattering and absorption, which in the case of fluorescence restricts imaging to within approximately 1 mm of the tissue’s surface. To image deeper than this we need to turn to near-IR fluorescence, bioluminescence, ultrasound, photoacoustic, radionuclide (SPECT, PET), or magnetic resonance imaging (MRI), which in the case of MRI can be used to image reporter activity at depths of tens of centimetres with a spatial resolution in the 100-μm range and a temporal resolution that can be as high as a few tens of ms. The use of these techniques to image gene reporters has been the subject of recent reviews [1.Farhadi A. et al.Genetically encodable materials for non-invasive biological imaging.Nat. Mater. 2021; 20: 585-592Crossref PubMed Scopus (13) Google Scholar,2.Serganova I. Blasberg R.G. Molecular imaging with reporter genes: has its promise been delivered?.J. Nucl. Med. 2019; 60: 1665-1681Crossref PubMed Scopus (23) Google Scholar]. The aim of this forum article is to briefly chronicle the development of MRI-based gene reporters and to highlight some more recent of the most sensitive MRI-based reporters are designed to affect the intensity of the tissue water proton resonance. The product of the reporter gene, which may be present in micromolar concentration, is detected indirectly via the intense resonance from the ~80 M tissue water protons. To understand how these reporters work it is necessary to have an understanding of the basic physical principles of NMR (Figure 1 and Box The nuclear spins of a spin ½ nucleus populate two allowed energy levels (Figure 1A), with the magnetic dipoles aligned with (lower energy level) or against (higher energy level) the applied magnetic field (B0). Because these energy levels are nearly equally populated Mz is very small (Figure 1B). The nuclear spins precess about the B0 field at a frequency (ω) that is a function of the nucleus’s gyromagnetic ratio (γ) and the magnitude of the B0 field experienced by the nucleus (ω = γB0) (Figure 1B). The application of an oscillating magnetic field resonant with this precession frequency (ω) tips Mz into the xy-plane (Mxy), where it induces a signal in the detector coil (the same coil can be used to apply the excitation pulse). Fourier transformation of this decaying signal converts the time domain signal into a frequency domain spectrum, where the amplitude of the signal is proportional to the number of spins and the frequency depends on their chemical environment, which determines the effective B0 field experienced by the nucleus. The exponential loss of phase coherence in the xy-plane, resulting in the decay of Mxy, is described by the T2 or spin–spin relaxation time (which contributes to the resonance linewidth in the spectrum), while restoration of the magnetization along the z-axis (Mz) is described by the T1 or longitudinal relaxation time. Loss of phase coherence is also caused by magnetic field inhomogeneity; however this additional loss of phase coherence can be refocussed by the 180° pulse in a spin echo experiment (Figure 1C). Magnetization created in the xy-plane by the 90° pulse is refocussed by the 180° pulse to produce an echo at time 2τ. In a diffusion-weighted experiment, magnetization will be dephased by the first applied magnetic field gradient pulse (G) and will be rephased by the second gradient pulse provided that it does not diffuse to a different part of the gradient in the intervening period (Δ). Loss of signal intensity with increasing gradient amplitudes can be used to determine the diffusion coefficient of the nuclear spin (Figure 1D). Magnetization can also be lost by diffusion of spins through endogenous magnetic field gradients. If two nuclei with different frequencies (chemical shifts) are in slow exchange such that they have separate peaks in the spectrum (Figure 1E), selective saturation of one of the resonances (equalization of the spin populations) will result in loss of intensity of the other resonance as it loses z magnetization to the saturated resonance without receiving any magnetization in return (Figure 1F). The magnitude of this decrease in intensity can be used to estimate an exchange rate The nuclear spins of a spin ½ nucleus populate two allowed energy levels (Figure 1A), with the magnetic dipoles aligned with (lower energy level) or against (higher energy level) the applied magnetic field (B0). Because these energy levels are nearly equally populated Mz is very small (Figure 1B). The nuclear spins precess about the B0 field at a frequency (ω) that is a function of the nucleus’s gyromagnetic ratio (γ) and the magnitude of the B0 field experienced by the nucleus (ω = γB0) (Figure 1B). The application of an oscillating magnetic field resonant with this precession frequency (ω) tips Mz into the xy-plane (Mxy), where it induces a signal in the detector coil (the same coil can be used to apply the excitation pulse). Fourier transformation of this decaying signal converts the time domain signal into a frequency domain spectrum, where the amplitude of the signal is proportional to the number of spins and the frequency depends on their chemical environment, which determines the effective B0 field experienced by the nucleus. The exponential loss of phase coherence in the xy-plane, resulting in the decay of Mxy, is described by the T2 or spin–spin relaxation time (which contributes to the resonance linewidth in the spectrum), while restoration of the magnetization along the z-axis (Mz) is described by the T1 or longitudinal relaxation time. Loss of phase coherence is also caused by magnetic field inhomogeneity; however this additional loss of phase coherence can be refocussed by the 180° pulse in a spin echo experiment (Figure 1C). Magnetization created in the xy-plane by the 90° pulse is refocussed by the 180° pulse to produce an echo at time 2τ. In a diffusion-weighted experiment, magnetization will be dephased by the first applied magnetic field gradient pulse (G) and will be rephased by the second gradient pulse provided that it does not diffuse to a different part of the gradient in the intervening period (Δ). Loss of signal intensity with increasing gradient amplitudes can be used to determine the diffusion coefficient of the nuclear spin (Figure 1D). Magnetization can also be lost by diffusion of spins through endogenous magnetic field gradients. If two nuclei with different frequencies (chemical shifts) are in slow exchange such that they have separate peaks in the spectrum (Figure 1E), selective saturation of one of the resonances (equalization of the spin populations) will result in loss of intensity of the other resonance as it loses z magnetization to the saturated resonance without receiving any magnetization in return (Figure 1F). The magnitude of this decrease in intensity can be used to estimate an exchange rate of the NMR contrast mechanisms described in Figure 1 have been exploited in the development of MRI-based gene reporters. For reporters that can be used in vivo their development can be traced back to two landmark papers published in 2000. Louie et al. [3.Louie A.Y. et al.In vivo visualization of gene expression using magnetic resonance imaging.Nat. Biotechnol. 2000; 18: 321-325Crossref PubMed Scopus (1013) Google Scholar] microinjected into Xenopus embryos a reporter expressing β-galactosidase and a gadolinium-based contrast agent in which the Gd3+ ion was shielded from solvent water by a galactopyranose blocking group. Cleavage of the sugar residue by β-galactosidase exposed the cell water to the Gd3+ ion, which by virtue of its seven unpaired electrons is paramagnetic, leading to a reduction in the water proton T1 and an increase in signal intensity. The decrease in T1 increases the signal intensity by reducing the degree of signal saturation. However, the requirement to microinject the contrast agent coupled with the relatively modest enhancement of signal intensity limited the further use of the reporter. Weissleder et al. [4.Weissleder R. et al.In vivo magnetic resonance imaging of transgene expression.Nat. Med. 2000; 6: 351-354Crossref PubMed Scopus (747) Google Scholar] used an engineered transferrin receptor, which mediated cell uptake of human holotransferrin conjugated to iron oxide nanoparticles. Implanted subcutaneous tumors in mice that expressed the reporter showed substantial decreases in signal intensity in T2*-weighted images following intravenous injection of the modified transferrin; the intracellular presence of the superparamagnetic iron oxide nanoparticles decreasing the water proton T2*. In a spin echo image (Figure 1C) loss of phase coherence and the resulting signal loss is due to both T2 relaxation, which is irreversible, and magnetic field inhomogeneity, the sum of which is given by T2* (1T2*=1T2+1Tinhom.). Loss of coherence due to inhomogeneity can be reversed by the 180° pulse in the spin echo experiment unless there is diffusion through the applied (as in Figure 1C) or endogenous magnetic field gradients, such as those created by iron oxide nanoparticles. The signal intensity changes produced by transferrin-conjugated nanoparticles persisted for up to 14 days, which means that it would not be possible to monitor rapid changes in gene expression, and the requirement to inject a macromolecular contrast agent limits use of the reporter to those tissues, such as tumors, that have a leaky vasculature that allows access of the contrast agent. Subsequent studies that used the iron storage protein ferritin as the reporter removed the requirement to inject a contrast agent [5.Genove G. et al.A new transgene reporter for in vivo magnetic resonance imaging.Nat. Med. 2005; 11: 450-454Crossref PubMed Scopus (383) Google Scholar,6.Cohen B. et al.Ferritin as an endogenous MRI reporter for noninvasive imaging of gene expression in C6 glioma tumors.Neoplasia. 2005; 7: 109-117Crossref PubMed Scopus (293) Google Scholar]. However, as with the transferrin receptor, the expression of ferritin produces negative contrast in T2*-weighted images, which can lack specificity; for example, in tumors negative contrast can also arise due to haemorrhage. Moreover, the changes in image contrast relatively gene reporter that can high levels of contrast is one on the protein which is expressed in and the uptake of a contrast agent of in in a nearly increase in signal intensity in images, which at and to of reporter the Gd3+ ion for the allowed imaging with The also uptake of bioluminescence by a factor of bioluminescence imaging of its expression in vivo et gene reporter for in vivo A. PubMed Scopus Google and the uptake of which allows near-IR fluorescence imaging of reporter expression et as a reporter gene for fluorescence and magnetic resonance PubMed Scopus Google this a reporter. and image it the injection of a contrast which may not be to of those expressing the reporter. reporters described which the use of an contrast exploit in the diffusion coefficient of tissue water in the and The reporters, A. et imaging using reporter water 7: PubMed Scopus Google Scholar] and the et of increases in water exchange of a gene Biotechnol. PubMed Scopus Google rapid exchange of water the expression was detected from of an increase in the using the pulse in Figure whereas expression was detected by the water exchange rate by the water at various of the proton magnetization of reporters that not the injection of an exogenous contrast agent the chemical exchange saturation reporter expression is detected via the exchange of with water (Figure et reporter gene MRI contrast on proton Biotechnol. PubMed Scopus Google Scholar]. which are produced by some as the magnetic field in the same that iron oxide nanoparticles and can be detected in T2*-weighted The in magnetic the and the are for their of the magnetic the negative MRI contrast created by the can be from endogenous of negative contrast by the with A. et imaging using reporter water 7: PubMed Scopus Google Scholar]. that been with a reporter expressing the for to be in vivo using A. et imaging of gene expression in 2019; PubMed Scopus Google Scholar] and also be detected by MRI et magnetic resonance imaging of protein Mater. PubMed Scopus Google Scholar]. The of the reporters have been in that they not allow the of reporters Reporters detected using have the potential to be they to the of contrast that have different chemical In a recent example, engineered to have expression as reporters to the and of their which the on these different chemical that reporter be detected using et designed MRI reporters for noninvasive imaging of transgene expression.Nat. Biotechnol. PubMed Scopus Google Scholar]. can also be in that their and determine their NMR such that they can be from other using MRI et magnetic resonance imaging of protein Mater. PubMed Scopus Google Scholar]. their determines their they can also be by the intensity of the used to resonance is a relatively the two energy levels for a spin ½ nucleus are nearly equally populated and the magnetization (Mz) with the of spins is very small (Figure spin techniques the of which can result in a in to can as for and can be detected at using where saturation of the resonance from in the in a decrease in the intensity of the which has a different chemical from different showed different chemical for in the the of et al.Genetically reporters for magnetic resonance imaging.Nat. 6: PubMed Scopus Google Scholar]. Detection of in vivo using would of and that this is this has to be In MRI-based gene reporters have been of which not the injection of an exogenous contrast agent and some of which allow imaging with different contrast of of these reporters is that in they produce images with relatively resolution with optical However, the of MRI to image at depth means that for some of these reporters there is potential for to be used in as has been for where a gene reporter was used to in I. Blasberg R.G. Molecular imaging with reporter genes: has its promise been delivered?.J. Nucl. Med. 2019; 60: 1665-1681Crossref PubMed Scopus (23) Google Scholar]. of are are