A brief review of reporter gene imaging in oncolytic virotherapy and gene therapy

Abstract

Reporter gene imaging (RGI) can accelerate development timelines for gene and viral therapies by facilitating rapid and noninvasive in vivo studies to determine the biodistribution, magnitude, and durability of viral gene expression and/or virus infection. Functional molecular imaging systems used for this purpose can be divided broadly into deep-tissue and optical modalities. Deep-tissue modalities, which can be used in animals of any size as well as in human subjects, encompass single photon emission computed tomography (SPECT), positron emission tomography (PET), and functional/molecular magnetic resonance imaging (f/mMRI). Optical modalities encompass fluorescence, bioluminescence, Cerenkov luminescence, and photoacoustic imaging and are suitable only for small animal imaging. Here we discuss the mechanisms of action and relative merits of currently available reporter gene systems, highlighting the strengths and weaknesses of deep tissue versus optical imaging systems and the hardware/reagents that are used for data capture and processing. In light of recent technological advances, falling costs of imaging instruments, better availability of novel radioactive and optical tracers, and a growing realization that RGI can give invaluable insights across the entire in vivo translational spectrum, the approach is becoming increasingly essential to facilitate the competitive development of new virus- and gene-based drugs.

Graphical abstract

Reporter gene imaging can accelerate the clinical translation and development of oncolytic virotherapy and gene therapy vectors. Here we review common reporter genes for radionuclide, magnetic resonance, and optical imaging modalities. We highlight the elegance of the sodium iodide symporter (NIS) that is amenable to SPECT or PET imaging in mice, large animals, and humans.

Introduction

Engineered viruses have emerged as the workhorses of oncolytic virotherapy (OV) and gene therapy.1, 2, 3, 4 To date, one herpes simplex-derived oncolytic virus (Imlygic), three adeno-associated virus (AAV) gene therapies (Glybera, Luxturna, and Zolgensma) and five retro- or lentiviral cellular gene therapies (Strimvelis, Kymriah, Tecartus, Yescarta, and Zynteglo) have gained marketing approval in the United States and/or the European Union. Ongoing clinical development of viral vectors for other disease indications is expected to benefit greatly from incorporating reporter gene imaging (RGI) to shorten development timelines by providing critical information on the biodistribution, magnitude, and duration of viral gene expression. RGI employs a variety of tracers and imaging modalities to detect the expression of reporter-gene-encoded proteins at their site of production in living animals. This review provides a succinct explanation of the following common RGI modalities and transgenes, emphasizing their relative strengths and limitations: three deep-tissue modalities (single photon emission computed tomography [SPECT], positron emission tomography [PET], and functional/molecular magnetic resonance imaging [f/mMRI]) and four optical imaging modalities (fluorescence, bioluminescence, Cerenkov luminescence, and photoacoustic imaging [PAI]).

We have focused exclusively on reporter genes that have been incorporated into oncolytic viruses or viral gene therapy vectors being developed for direct in vivo gene transfer (Figure 1). Not covered are the topics of imaging for ex vivo cellular therapies or imaging with quantum dots, nanoparticles, fluorescent dyes, contrast agents, or imaging of endogenously expressed proteins such as the dopamine D2 receptor (D2R), somatostatin receptor 2 (SSTR2), prostate-specific membrane antigen (PSMA), or transferrin receptor-1 (Tfr), but resources are provided for the reader’s interest.5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26 Additionally, detailed reviews of imaging for OV and gene therapy are recommended.27, 28, 29

Figure 1.

Reporter gene imaging in oncolytic virotherapy and gene therapy

Reporter genes are organized by imaging modality. Commonly used imaging substrates are indicated, as well as the use of reporter genes in human clinical trials of oncolytic virotherapy or gene therapy. ^§SPECT radiotracer; ^ƤPET radiotracer. Abbreviations: SPECT, single photon emission computed tomography; PET, positron emission tomography; NIS, sodium iodide symporter; HSV1-tk1, herpes simplex virus 1 thymidine kinase; FIAU, 1-(2-deoxy-2-fluoro-1-D-arabinofuranosyl)-5-iodouracil; FHBG, 9-(4-fluoro-3-[hydroxymethyl]butyl)guanine; FHPG, 9-[(3-fluoro-1-hydroxy-2-propoxy)methyl]guanine; GCV, ganciclovir; FPCV, 8-fluoropenciclovir; NET, norepinephrine transporter; MIBG, meta-iodobenzylguanidine; MFBG, meta-fluorobenzylguanidine; SSTR2, somatostatin receptor 2; DOTATOC, (DOTA⁰-Phe1-Tyr3)octreotide; DOTATATE, DOTA⁰-Tyr3-octreotate; DOTANOC, DOTA-1-Nal(3)-octreotide; f/mMRI, functional or molecular magnetic resonance imaging; Tyr, tyrosinase; N/A, not applicable; LRP, lysine-rich protein; GFP, green fluorescent protein.

The necessity of in vivo imaging

Oncolytic virotherapy and gene therapy depend on adequate delivery of the viral vector and effective targeting to the tissues of interest. Insufficient transduction of target tissues, leading to insufficient transgene expression, is often the primary limitation affecting the potency of these therapies. RGI can be helpful in identifying these problems in living animals, especially because it can be repeated at multiple time points in a single experimental animal, avoiding the need for euthanasia and analysis of necropsy material at multiple time points. For example, employing an AAV-9 vector encoding the sodium iodide symporter (NIS) reporter gene, Moulay et al.³⁰ used SPECT/computed tomography (CT) imaging to determine that intracoronary perfusion is a suboptimal method for transduction of heart muscles in research canines but that this can be remedied by using a novel spiral catheter and a new cardiac muscle injection technique. Such an example illustrates that it is essential to monitor and quantify infection and therapeutic gene expression throughout the course of treatment, especially in preclinical studies using large animal models. Most gene products cannot be assessed via blood assays and require other methods, such as invasive biopsies or non-invasive imaging with reporter genes. Besides the obvious drawbacks of patient/animal discomfort and surgical risks, biopsies only provide a snapshot at one time point in a single tissue.³¹ To truly understand the pharmacokinetics and pharmacodynamics of a viral vector, repeated non-invasive imaging across the entire body is necessary.³² If only the tissues of interest are sampled, infection of unsampled tissues may be missed, leading to possible side effects that could not be explained or addressed without extensive biopsies. Tracking gene expression over long periods is especially critical for gene therapy, which usually aims for lifelong expression of a therapeutic gene. Longitudinal tracking is imperative to prove the safety and efficacy of such therapies. Early gene therapy studies suffered from a lack of effective tracking mechanisms: did the therapy fail because the vector did not infect the right tissue or the transgene did not express? Or, what caused toxicity: the vector, too much transgene expression, or infection of off-target tissues?33, 34, 35 The use of reporter genes in these systems would answer these key questions.

Challenges to widespread use of reporter genes lie mainly in the technology required to image them: imaging agents and machines can be costly, and specialized training may be needed to handle these agents. However, facilities capable of producing short half-life radiotracers in-house are proliferating, and imaging instruments are becoming cheaper and more widespread, expanding access for both preclinical and clinical applications. Many manufacturers have also designed the imaging machines to be capable of multi-modality imaging, incorporating highly sensitive CT imaging for anatomical location of tissues to complement SPECT or PET modalities. With more centers adopting noninvasive imaging technologies, and increasing attention being given to the “three Rs” in animal research (replace, reduce, refine), RGI must become a fundamental step in virus drug development, as it is an invaluable tool to assess virus spread, infection, persistence, and efficacy.

Deep-tissue imaging

Several deep-tissue molecular imaging modalities exist, though the most commonly used are CT, SPECT, PET, MRI, and f/mMRI. CT uses a series of X-rays to generate high-resolution 3D images.³⁶ SPECT is a type of gamma scintigraphy where individual planar images generated by gamma photons emitted from radiotracers are compiled into a 3D volumetric image via CT.³⁷ Upon radioactive decay, SPECT radiotracers emit a single photon, with emission in any direction occurring with equal probability.³⁸ This characteristic necessitates strict collimation to ensure captured photons are directly from the decay event to maintain resolution, which increases the dose of radioisotope given to achieve sufficiently high signals and imaging.³⁸ PET imaging is more sensitive and has higher resolution than SPECT due to the unique physics of positron decay and subsequent signal analysis by the detector. Positrons emitted from the nucleus of a radionuclide such as ¹⁸F travel a short distance and combine with an electron, resulting in an annihilation reaction. Two 511-keV photons are emitted roughly 180° apart and are detected as a coincident event when they strike opposing detectors simultaneously.³⁹ It is the requisite simultaneous detection of both photons that provides PET with superior signal-to-background resolution over SPECT.⁴⁰ SPECT and PET are often combined with CT for co-registered anatomical and functional imaging.³⁶ MRI uses strong magnetic fields and radio waves to excite the nuclear spin energy transition of targeted molecules, most frequently hydrogen.⁴¹ The rate of relaxation from the excited state varies per tissue density, and this difference allows for high-resolution imaging.⁴¹ f/mMRI is used to enhance the anatomical data obtained via MRI with physiological data relating to a biological process. However, the sensitivity of f/mMRI for available imaging agents is poor compared to optical imaging and radionuclide imaging (RI).²⁸ Other drawbacks to MRI are cost of the machines and the time required for imaging (15–90 min).

The following sections are arranged by reporter gene instead of imaging modality, since most deep-tissue imaging transgenes can be used for several modalities depending on the substrate.

RI

NIS

NIS imaging is the most mature RGI method for monitoring gene and viral therapies in human clinical trials. NIS mediates the concentration of iodide into the thyroid gland for thyroid hormonogenesis. NIS can transport several other monovalent anions that can be used for both SPECT and PET (Figure 2A).^46,47 NIS is the most common human reporter gene used, and a plethora of viruses have been engineered to express NIS.48, 49, 50, 51, 52, 53, 54 Many pre-clinical and clinical trials have shown that virally delivered NIS can be used to accurately track viral infection and spread in the tumor, as well as track tumor response to treatment.⁵⁵ Several gene therapy studies have displayed the utility of NIS for longitudinal in vivo imaging of adenovirus and AAV gene therapies to identify biodistribution and durability of gene expression.^30,56, 57, 58, 59, 60, 61, 62 NIS imaging was also shown to be more sensitive and longer lasting than imaging with herpes simplex virus 1 thymidine kinase (HSV1-tk) (discussed below).⁶⁰ Development of novel NIS-compatible PET radiotracers, like ¹⁸F-tetrafluoroborate, is likely to replace ¹²³I or ^99 mTcO₄ SPECT imaging with NIS in the near future due to the superior resolution and sensitivity of PET imaging.63, 64, 65 The half-life of ¹⁸F is extremely short (109.7 min), greatly reducing long post-procedure quarantines while the radiotracer is excreted. Studies of ¹⁸F-tetrafluoroborate in healthy volunteers show safety, feasibility, and high sensitivity, significantly improving resolution of NIS RGI.⁶⁶

Figure 2.

Mechanisms of radionuclide imaging (SPECT/PET) reporter genes and examples from oncolytic virotherapy and gene therapy studies

(A) NIS expressed on the cell surface concentrates ∗I (∗I indicates it can be ¹²³I, ¹²⁴I, ¹²⁵I, or ¹³¹I) radiotracer into the cell. Decay of ∗I radiotracers directly releases (¹²³I/¹²⁵I/¹³¹I) or leads to the release (¹²⁴I) of gamma (γ) photons for SPECT or PET imaging. Imaging panel: 3D reconstruction of PET/CT imaging of mice infected with AAV-9-CAG-human NIS or minke whale NIS in the absence or presence of the NIS inhibitor perchlorate (ClO₄). (B) HSV1-tk: Radiolabeled substrates like [∗I]-FIAU can pass in and out of the cell until they are phosphorylated by HSV1-tk to yield [∗I]-FIAU-PO₄, which becomes trapped intracellularly. Upon radioactive decay, ∗I releases γ photons for capture via SPECT/PET. Imaging panel: 9-[4-[^18F] fluoro-3-(hydroxymethyl) butyl] guanine (¹⁸F-FHBG) uptake in the liver and injection site (arrow) of a self-complementary recombinant AAV-3 vector expressing HSV1-tk and kallistatin. (A) depicts coronal slices of an animal’s PET imaging (left) and image of maximum intensity projection (right). (B) depicts transverse, coronal, and sagittal images of the PET/CT. (C) NET: NET transports radiolabeled norepinephrine analogs like [∗I]-meta-iodobenzylguanidine (∗I-MIBG) into the cell, where γ photons are released upon radioactive decay. Imaging panel: HepG2 tumors in mice infected with an adenovirus expressing NET. ¹³¹I γ-camera images 30 min (a), 1 h (b), 4 h (c), and 24 h (d) after radiotracer administration. Radiotracer collects in the bladder non-specifically for excretion. (D) SSTR2: Radiolabeled peptides (such as ⁶⁸Ga-DOTATE) bind to the SSTR2 receptor and γ photons are released upon radioactive decay. SSTR2 is the only nuclear transgene depicted which cannot amplify its signal due to the 1:1 receptor:peptide binding stoichiometry. Imaging panel: Infection of HT29 tumors with oncolytic vaccinia virus expressing SSTR2 (VV-SSR2) and SPECT imaging with ¹¹¹In-octreotide 4 days after virus treatment. Figure made with BioRender (https://biorender.com/) and Photoshop CC 2019. Protein shapes are artistic interpretations. Permissions for imaging panels: (A) From Concilio et al.⁴² and reproduced under CC Attribution-NonCommercial-NoDerivs License (https://creativecommons.org/licenses/by-nc-nd/4.0/). (B) From Liu et al.⁴³ and reproduced under CC Attribution License (https://creativecommons.org/licenses/by/4.0/). (C) From Jia et al.⁴⁴ and reproduced with licensed permission. (D) From Wang et al.⁴⁵ and reproduced with permission from the journal; this research was originally published in JNM.

NIS is also unique in the RGI field, as many species of the NIS gene have been sequenced or cloned, enabling RGI in many animal models with the native NIS protein.^67,68 NIS is present in all vertebrates, enabling species-specific NIS vector development for the vast majority of model systems.⁶⁹ This has allowed longitudinal imaging in immune-competent small and large animal models due to the lack of immune response against a self-protein reporter gene. Of particular interest, NIS has enabled tracking of OV, cell, and gene therapies in large animal models, the importance of which is discussed later.^30,61,70, 71, 72, 73, 74, 75, 76 The moderate sensitivity and excellent specificity of NIS in vivo imaging requires fewer animals for small and large animal studies, reducing animal use and costs and expediting pre-clinical and translational research.

Several challenges to NIS imaging include endogenous NIS expression, efflux of radiotracer from NIS-expressing cells, and low transduction/expression of the NIS transgene in transduced target tissues. The concentration of radioisotopes in non-target tissues that naturally express NIS, such as the thyroid, salivary glands, and stomach, makes interpretation and quantitation of NIS signals technically difficult if the transduced tissue is in the vicinity of endogenous NIS-expressing tissues. Several studies have explored ways to improve NIS expression or block endogenous NIS expression to address the above concerns.77, 78, 79, 80 For example, to facilitate imaging of transduced hepatocytes, animals are fed barium sulfate to delineate the stomach/gastrointestinal tract and mask some of the endogenous NIS signal in the gastric mucosa in mice and pig models after AAV gene transfer.⁷⁹ A recent study by our group investigated the feasibility of reducing uptake by endogenous NIS via use of the NIS inhibitor perchlorate and a perchlorate-resistant NIS variant from minke whale (B. acutorostrata scammoni).⁴² Combination of minke whale NIS and perchlorate pretreatment resulted in superior imaging using low levels of AAV-9 transduction in mice with SPECT/CT and PET/CT while significantly reducing uptake in endogenous NIS-expressing tissues.⁴² NIS continues to grow in popularity in both preclinical and clinical settings due to its human origin, sensitivity, SPECT and PET utility, label-free radiotracers, and persistence of expression.

HSV1-tk

The next most advanced RGI method for OV and gene therapy in human clinical trials utilizes HSV1-tk. HSV1-tk can phosphorylate non-toxic prodrugs like purine nucleotide derivatives and acyclovir derivatives to trap them in cells, where the toxic forms can be used for radioisotope imaging or radiotherapy (Figure 2B).²⁷ Since HSV1-tk was one of the first reporter genes investigated, it enjoys a wide breadth of PET and SPECT tracers, as well as mutant forms, such as HSV1-sr39tk, which has enhanced in vivo activity.^27,29,81,82 Though HSV1-tk imaging is successful in most small animal models, the results in clinical trials have been mixed.^29,83, 84, 85, 86 Investigations into the use of HSV1-tk for human imaging are meager: out of the ten active or recruiting clinical trials currently registered with ClinicalTrials.gov using a virus expressing HSV1-tk, only one plans to perform HSV1-tk imaging (ClinicalTrials.gov: NCT04313868). The paucity of translational application of HSV1-tk for imaging hinders accurate assessment of its potential use in humans. Clinical trials should make every effort to include at least one round of imaging to assess virus spread and infection. Longitudinal imaging would reveal the optimal windows for suicide gene therapy and improve the efficacy of HSV1-tk cancer gene therapy. AAV gene therapy has been quantified in rodents and non-human primates using HSV1-tk PET imaging.^43,87 As a non-human protein, HSV1-tk could provoke an immune response, making it an unlikely candidate for long-term imaging, as would be needed for gene therapy.⁸⁸

Norepinephrine transporter (NET)

NET is a Na⁺/Cl⁻-dependent membrane protein that transports norepinephrine, epinephrine, and dopamine into the cell. The human protein (hNET) has been used as a reporter gene for SPECT and PET with several radiotracers, primarily meta-iodobenzylguanidine (MIBG), which is already approved by the US Food and Drug Administration (FDA) (Figure 2C).⁸⁹ NET has been used for nuclear imaging or radiotherapy with vaccinia virus, adenovirus, and herpes virus.^44,90,91 Use of endogenous NET for tumor imaging is still under clinical development and use, but further use of NET for imaging of OV has not been reported beyond the above-mentioned studies, likely due to the existence of other reporter genes with better expression profiles and radiotracer availability.^92,93

SSTR2

SSTR2 is one of the receptors for the peptide hormone somatostatin and is highly expressed on neuroendocrine cells and others, where it is involved in neurotransmission, hormone secretion, and cell proliferation.^27,94 The human protein (hSSTR2) has been used for SPECT imaging with indium-111-labeled synthetic peptide substrates such as octreotide, pentetreotide, and lanreotide, or PET imaging with gallium-68-labeled peptides (Figure 2D).⁹⁵ SSTR2 has been inserted into adenovirus and oncolytic vaccinia, where long-term tracking of viral infection and persistence was shown to be feasible in several tumor types.^32,45,96, 97, 98, 99 SSTR2 has also been used to track AAV gene therapy, where it permitted PET imaging 6 months post-treatment.¹⁰⁰ The major drawbacks of SSTR2 are endogenous expression, which can reduce diagnostic efficacy, and that each receptor can only bind one radiolabeled ligand, preventing signal amplification and thus limiting imaging sensitivity.^45,101

MRI

Ferritin

Ferritin is an endogenous intracellular protein that binds and coverts labile Fe²⁺ iron to the stable and non-toxic Fe³⁺ form for storage (Figure 3A).¹⁰⁵ Iron is an ideal MRI contrast agent, so ferritin overexpression in target cells permits RGI with MRI.^106,107 Ferritin RGI has been tested in lentiviral and AAV gene therapy in rodent brain.¹⁰⁸ HSV1-mediated ferritin expression in mouse brain was sufficient for imaging and permitted quantitative assessment of infectious particles.¹⁰² Ferritin and enhanced green fluorescent protein (GFP) were recently co-expressed in vesicular stomatitis virus (VSV) to assess neural connectivity with MRI.¹⁰⁹ Ferritin is attractive because it does not require any exogenous probes, since it binds to naturally bioavailable iron. However, this approach leads to negative contrast in the MR image, which exhibits low sensitivity (in the mM range) compared to positive contrast approaches like PET (low nM range).^29,110 Further methodologies are required to increase the sensitivity of ferritin RGI.

Figure 3.

Mechanisms of functional/molecular magnetic resonance imaging (f/mMRI) reporter genes and examples from oncolytic virotherapy and gene therapy studies

(A) Ferritin: Iron cations captured by ferritin induce differences in weighted MRI images. Imaging panel: MRI imaging of (a and b) 25 million particles or (c and d) 100 million particles of HSV-1 expressing ferritin in mouse brain. (B) Tyrosinase: Production of melanin induces differences in weighted MRI images. Imaging panel: MRI imaging of MCF-7 tumors with or without Tyr expression in mice. (C) Lysine-rich protein (LRP): LRP induces differences in weighted MRI images. Imaging panel: MRI chemical exchange saturation transfer (CEST) imaging of murine heart transduced with AAV-9 encoding LRP. Figure made with BioRender (https://biorender.com/) and Photoshop CC 2019. With the exception of ferritin, protein shapes are artistic interpretations. Permissions for imaging panels: (A) From Iordanova et al.¹⁰² and reproduced with licensed permission. (B) From Qin et al.¹⁰³ and reproduced under CC Attribution 3.0 Unported License (https://creativecommons.org/licenses/by/3.0/). (C) From Meier et al.¹⁰⁴ and reproduced under CC Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/).

Tyrosinase (Tyr)

Tyr can be used for fMRI in addition to PET and PAI.¹⁰³ Melanin production induced by Tyr expression also increases metal ion chelation, which causes substantial improvement in MRI contrast (Figure 3B).¹¹¹ To date, the only studies with Tyr in OV or gene therapy use the vaccinia virus strain GLV-1h68 expressing Tyr and tyrosinase-related protein 1 (Tyrp1), which further enhances melanin production.^112,113 The main problems with Tyr arise from low transgene expression and off-target effects of excess melanin production, such as production of reactive oxygen species and inhibition of virus replication.^105,112, 113, 114 Issues arising from excess melanin production can be overcome by using drug-inducible Tyr expression systems, but this adds another layer of genetic and clinical regulatory complication.¹¹³

Lysine-rich protein (LRP)

LRP is an artificial reporter gene that was generated for use with chemical exchange saturation transfer (CEST) MRI (Figure 3C).¹¹⁵ CEST MRI relies on radio-frequency-irradiated protons from a source agent (i.e., LRP), which become saturated, exchanging with protons from water in the tissue to decrease bulk water signal on the MRI.¹⁰⁵ CEST imaging has made it to the clinic in the past decade and offers unique windows into physiological changes, such as intracellular and extracellular pH, glucose uptake, and lactate production.¹¹⁶ LRP was inserted into oncolytic HSV-1 G47Δ and successfully used for CEST MRI without impacting viral replication or oncolytic efficacy.¹¹⁷ Another study generated an AAV-9 vector encoding LRP to longitudinally track cardiac transduction in mice, with robust imaging achieved 90 days post infection.¹⁰⁴ The sustained expression and lack of observed toxicity bode well for LRP CEST MRI. However, LRP imaging and CEST MRI suffer from low sensitivity and high specific absorption rate (SAR), limiting its clinical use.¹⁰⁵ Novel reporter genes for CEST MRI are needed to spur more widespread use of this interesting imaging technology.

Optical imaging

Optical imaging (OI) relies on light in the infrared, visible, or ultraviolet spectrum and thus is best suited for surface imaging and not deep-tissue imaging. OI modalities include fluorescence, bioluminescence, Cerenkov luminescence, and PAI. OI modalities are attractive due their short acquisition times, lower cost, high-throughput abilities, multispectral imaging capabilities, and convenience of use as compared to radioisotopes.^28,89,118 These attributes make OI extremely popular for in vitro studies or preclinical studies in small animals. However, light intensity and spatial resolution remain challenges for OI, as emitted photons can be absorbed or scattered by tissue.^28,119 These problems, as well as the potential for immune responses against these non-human proteins, have limited the use of OI reporter genes in the clinic.²⁸ A thorough review of OI modalities was recently published.¹²⁰

Fluorescence imaging (FI)

FI with reporter genes primarily uses the GFP from the jellyfish Aequorea victoria and its variants, such as yellow fluorescent protein (YFP), mCherry, and many others. FI requires incident light at the correct excitation wavelength to reach the fluorophore, which will then release a photon at a distinct emission wavelength to be captured by a specialized charge-coupled device camera (Figure 4A). This negates the need for exogenously applied substrates or imaging agents but is limited by light penetrance into the tissue of interest.²⁹ Autofluorescence and immunogenicity also limit the sensitivity and persistence of FI. FI can be used to image oncolytic virus infection and spread in murine tumor models and is especially useful for visualizing tumors in explanted orthotopic or subcutaneous models after necropsy. Many oncolytic viruses have been engineered to express GFP or other fluorescent proteins to track viral infection and spread, but IF is best for in vitro analysis or requires necropsy of the animal and explanting tissues for analysis under a dissecting microscope.²⁷ Novel OI genes are constantly being developed, such as infrared fluorescence proteins or the very small endogenously fluorescent protein UnaG, which can be used to monitor gene transfer in AAV vectors, which have limited genetic capacity.123, 124, 125

Figure 4.

Mechanisms of optical imaging modalities and examples from oncolytic virotherapy and gene therapy studies

(A) Fluorescence: Incident excitation wavelength (Ex. λ) excites the fluorophore within the intracellular GFP and causes an emission wavelength (Em. λ) to be released for detection. Imaging panel: GFP distribution in neonatal mice after intravenous (i.v.) administration of AAV-8-CMV-GFP. Left to right: whole body, eyeball, liver, kidney, muscle, and brain. (B) Bioluminescence: Intracellularly expressed firefly luciferase (Fluc), Renilla or Gaussia luciferase (Rluc/Gluc), or NanoLuc bind with substrates D-luciferin, coelenterazine (CTZ), or furimazine, respectively, which enters the cell from exogenous sources. Emission wavelengths (Em. λ) are released upon processing of the substrate by luciferase. Imaging panel: Firefly luciferase imaging in mice after i.v. administration of an AAV-8-NFκB-Luc-2A-GFP biosensor or the constitutively expressed control AAV-8-SFFV-Luc-2A-GFP. (C) Cerenkov luminescence: NIS expressed on the cell surface concentrates ¹³¹I radiotracer and concentrates it into the cell. Upon radioactive decay, ¹³¹I releases light (Em. λ), which can be captured by a charge-coupled camera device. Imaging panel: Medulloblastoma xenografts in mice infected with measles virus expressing NIS (MV-NIS). Top panels are tumor bioluminescence and lower panels are ¹³¹I imaging from radiotracer concentrated into the tumor by NIS. (D) Photoacoustic imaging (PAI): Laser light absorbed by tissue is translated to heat and converted into ultrasound waves (photoacoustic effect), which are used for imaging. Tyr produces melanin, which readily absorbs the laser. Imaging panel: PAI and ultrasound imaging (US) of MCF-7 tumors with or without Tyr expression in mice. Figure made with BioRender (https://biorender.com/) and Photoshop CC 2019. With the exception of GFP, protein shapes are artistic interpretations. Permissions for imaging panels: (A and B) From Karda et al.,¹²¹ modified via cropping of the original figure and reproduced under CCA 4.0 International License (https://creativecommons.org/licenses/by/4.0/). (C) From Hutzen et al.¹²² and reproduced under CC Attribution License (https://creativecommons.org/licenses/by/2.0). (D) From Qin et al.¹⁰³ and reproduced under CC Attribution 3.0 Unported License (https://creativecommons.org/licenses/by/3.0/).

Bioluminescence imaging (BLI)

BLI utilizes charge-coupled device cameras to capture light produced by the oxidation of substrates by luciferase enzymes. The most commonly used luciferase enzymes are those from the North American firefly (P. pyralis) (Fluc), the sea pansy (Renilla reniformis) (Rluc), and the marine copepod (Gaussia princeps) (Gluc). Fluc uses D-luciferin as its substrate, along with the cofactors ATP, Mg²⁺, and O₂, and releases yellow-green light. Both Rluc and Gluc use coelenterazine (CTZ) and release primarily blue light, which is less tissue penetrant than yellow-green light (Figure 4B).¹²⁶ Newly developed enzyme variants, such as NanoLuc, and novel substrates continue to improve BLI.¹²⁷ Further modifications to NanoLuc or the substrate have expanded imaging options, such as Nano-lanterns, Antares, Antares2, and NanoBiT.^126,128, 129, 130, 131, 132, 133, 134, 135, 136 BLI offers an improvement over FI due to a lack of background signal, as only the delivered transgene can catalyze the light-producing reaction with the exogenously supplied substrate. Similar to FI, BLI suffers from poor spatial resolution and depth-dependent signal attenuation, which prevents translation into large animal models and humans.¹³⁷

Though BLI is not used clinically, it is an invaluable tool for preclinical studies. Its ease of use, sensitivity, relative low cost, and relative lack of toxicity in animal models permits widespread utility in many oncolytic virotherapy and gene therapy platforms.^27,28 BLI can also be quantitative, with studies indicating that viral titers can be correlated with luminosity.^138,139 BLI provides an easy-to-use high-throughput method for estimating viral infection, replication, and persistence in mice, which are fundamental factors to monitor for accurate assessment of any given viral therapy. The limitation is lack of resolution and 3D tomography, making it difficult to discriminate between transduced tissues.

Cerenkov luminescence imaging (CLI)

Cerenkov luminescence is a unique optical phenomenon whereby light is emitted when charged particles, such as positrons (β⁺) or electrons (β⁻) travel through a dielectric medium faster than the speed of light through that particular medium.¹⁴⁰ The released light is in the ultraviolet-to-visible spectrum and can be captured with sensitive charge-coupled device cameras.^141,142 CLI is attractive due to its optical nature, the detection instrumentation for which is much cheaper than RI modalities, and many of the most suitable radioisotopes are already clinically approved.^120,140,143 Several isotopes permit simultaneous radiotherapy, RI, and CLI.¹⁴⁴ The primary limitations of CLI are low light yield and the high absorption rate of ultraviolet-to-blue light in tissue.¹⁴⁵ These issues hinder the applicability of CLI in clinical settings, but CLI RGI in animals has been reported.^140,146 CLI has been used successfully to image ¹³¹I concentration into medulloblastoma xenografts infected with an oncolytic measles virus expressing the NIS in mice (Figure 4C).¹²² CLI can allow for imaging of β⁻ radiotracers, which cannot be detected via other imaging technologies, expanding the repertoire of useful radiotracers.¹⁴⁰

PAI

Also known as optoacoustic imaging, PAI is a recently developed imaging modality. PAI makes use of the photoacoustic effect by using a short-pulse laser to irradiate the target tissue. Depending on the physical attributes of the tissue, varying amounts of light are absorbed, which causes molecular vibration and the generation of thermoelastic expansion.^120,147 The acoustic waves produced by this process are subjected to less scattering than photons as they travel though tissue, increasing the imaging depth significantly.¹²⁰ Contrast agents are sometimes used to enhance molecular specificity.^148,149 To date, only one oncolytic virus has been used for PAI, the previously described vaccinia virus expressing Tyr and Tyrp1, which facilitates melanin production (Figure 4D).^112,113 Melanin is an ideal contrast agent for PAI, as well as MRI, and overexpression in tumors permitted PAI in an animal model.¹¹² Recently, a calcium and manganese carbonate-coated oncolytic adenovirus was successfully imaged with PAI and MRI in tumor-bearing mice using the bound manganese as the PAI agent.¹⁵⁰ PAI remains underutilized in the OV and gene therapy field but will hopefully expand as more relevant PAI imaging genes are developed.

Large animal imaging

The transition from in vitro and preclinical in vivo studies to clinical trials requires the use of relevant animal models. Particularly for oncolytic virotherapy and gene therapy, adequate assessment of safety and efficacy are essential to the translational process. RGI can play a crucial role in this process, giving insight into otherwise difficult-to-assess parameters like viral spread, replication, and persistence. RGI is already commonly used in small animal models like mice and rats, but it also simplifies the use of large animal models like dogs, pigs, and non-human primates by limiting the need for extensive biopsies and tissue staining of multiple slices. Studies using larger animals with physiologies closer to that of humans are sometimes unavoidable, as they are often better predictors of toxicity or clinical efficacy than small animal models.¹⁵¹ Indeed, use of large animal models is recommended and sometimes required by the FDA, the European Medicines Agency (EMA), and the International Society for Stem Cell Research (ISSCR) to expedite testing and final commercial approval, especially for biological drugs.¹⁵¹ RGI is of particular value for large animal imaging: (1) fewer animals are needed for experiments because the transgene, and therefore the vector bearing the transgene, can be tracked noninvasively, negating the need for multiple time points for tissue collection; and (2) each animal can serve as its own control by imaging prior to vector administration, thus reducing the number of animals needed and eliminating inter-animal variability. Cost is a significant limitation for large animal studies, so reducing the number of animals required will make large animal imaging more accessible. OI techniques used for small animals cannot be applied to large animals because visible light cannot penetrate the tissue mass of large animals.27, 28, 29 Therefore, deep-tissue imaging modalities like SPECT, PET, and MRI must be used. To date, the only reporter gene to be repeatedly used for large animal deep-tissue imaging of OV or gene therapy is NIS, which has been studied in dogs, pigs, and non-human primates.^30,61,70, 71, 72, 73, 74, 75, 76 Use of the other genes discussed in this review for large animal imaging of OV and gene therapy is skewed heavily toward gene therapy and is sporadic.^87,152, 153, 154, 155, 156, 157, 158, 159, 160, 161 The one exception is GFP, which has been used widely to test ocular gene therapies in dogs, cats, and non-human primates, because OI is possible due to the unique location and physiology of the eye.162, 163, 164, 165, 166, 167, 168 Most recent research using HSV1-tk is directed toward cell therapies.169, 170, 171, 172, 173 The use of deep-tissue imaging modalities for large animals is also much more translatable to human trials than OI, increasing the value of large animal imaging for pre-clinical validation. Further development of RGI methods for large animals will facilitate efficient translation of novel therapeutics.

The future of RGI

Oncolytic virotherapy and gene therapy are poised to revolutionize modern medicine in the coming decades. RGI offers a powerful modality for tracking oncolytic viruses and gene therapies but is vastly underutilized. This has been partially due to cost and access issues but may also stem from a lack of appreciation as to what can be achieved with RGI, particularly with regard to the rapid assessment of vector or virus biodistribution and durability of transgene expression in living animals. Detection sensitivity for many RGI systems has increased to the point that transgene expression can now be readily detected below the threshold that is necessary for clinical relevance, so while increasing sensitivity may improve imaging, this may not translate to increased therapeutic benefit. Further sensitivity gains will doubtless be developed for most available RGI systems, but these enhancements may not be necessary to answer critical questions pertaining to the levels of gene expression needed for therapeutic benefit. Regardless, RGI itself has vastly improved since its inception at the turn of the century and continues to evolve and progress. As long as disease threatens human life, there will be a need to light up the therapies to eradicate it.