Optical fluorescence imaging with shortwave infrared light emitter nanomaterials for in vivo cell tracking in regenerative medicine

Abstract In vivo tracking and monitoring of adoptive cell transfer has a distinct importance in cell‐based therapy. There are many imaging modalities for in vivo monitoring of biodistribution, viability and effectiveness of transferred cells. Some of these procedures are not applicable in the human body because of low sensitivity and high possibility of tissue damages. Shortwave infrared region (SWIR) imaging is a relatively new technique by which deep biological tissues can be potentially visualized with high resolution at cellular level. Indeed, scanning of the electromagnetic spectrum (beyond 1000 nm) of SWIR has a great potential to increase sensitivity and resolution of in vivo imaging for various human tissues. In this review, molecular imaging modalities used for monitoring of biodistribution and fate of administered cells with focusing on the application of non‐invasive optical imaging at shortwave infrared region are discussed in detail.

by application of therapeutic cells instead of organ transplantation. 5 The success of cell-based therapies and their clinical translation to humans depends on two properties of adaptive cell transferred: safety and efficacy. 6 Despite promising cell therapy studies stating improvement and recovery of damaged organs, [7][8][9] there are still controversial findings in the literatures regarding effectiveness 10,11 and safety. 12,13 Thus, tremendous challenges have been come up in the application of this kind of treatment in regenerative medicine, which are discussed below.

| FAC TOR S AFFEC TING CELL FATE
Biodistribution pattern, viability and fate of therapeutic cells in the target tissue after infusion are main causes of contradictory results among published studies. 14

| Cell size
It is suggested that increasing number of cell passages during in vitro expansion leads to the enlargement and widening of the cell size. This issue is considered as one of the important reasons for cell entrapment in lung and obstruction of subsequent small capillaries after intravenous cell infusion. [15][16][17]

| Route of cell delivery
Cell delivery route has also a major effect on the localization and fate of transplanted cells in the living body.

| Systematic cell delivery
Cell transplantation through the systemic circulation is achieved via intravenous, intra-arterial and intraperitoneal routes. Various animal studies have demonstrated that the vascular bed of the lung is the first place where intravenously administered cells convene, which can cause small venule obstruction. 14 Consequently, subsequent interaction with lung vascular endothelial cells affects their viability, biodistribution and clinical efficiency. 16,[18][19][20] Eggenhofer et al studied the viability and biodistribution of intravenously infused mesenchymal stem cells (MSCs) after 5 minutes and 1, 24 and 72 hours. The transplanted cells could be found viable in the lung tissue only in 24 hours, but after 24 hours post-cell injection, no viable cells in the lung or other tissues such as liver, spleen or heart were found. 21 Administration of cells through the arterial route can bypass the pulmonary pathway and facilitate the translocation of cells to the intended organs. 15,22,23 This route of infusion can enhance the cell localization and engraftment at ischaemic brain 24 and damaged kidneys. 22 However, intra-arterial administration of cells may compromise arterial blood supply and cause accumulation in small arteries, [24][25][26] leading to organ infarction. 24 Li et al demonstrated that, though, intra-arterial neural progenitor stem cell delivery produces successful biodistribution and engraftment of infused cells in the brain, but yielded to a significant mortality of animals during the procedure. The reason of high mortality during cell administration may be associated with decreased blood supply to brain parenchyma, predisposing it to ischaemia, thrombosis, oedema, high intracranial pressure and consequently death of animals. 27 Vulliet et al have investigated the safety of MSC delivery to intracoronary blood flow for treatment of myocardial diseases. They infused MSCs into coronary artery of healthy animal models, and 7 days after cell infusion, healthy dogs exhibited signs of myocardial infarction. Histologic evaluation of myocardial tissue proved acute ischaemia and subacute microinfarction likely due to enlargement of MSC size during in vitro expansion or high dosage of MSCs. 28 In another study, high percentage of intra-arterially infused MSCs were entrapped at the precapillary level due to greater size of these cells compared to the diameter of microvessels. 25 Precapillary occlusion results in blood flow disturbance and ischaemia, which leads to consequent death. 25 It has been also claimed that MSC infusion through the arterial route can increase the localization of cells to the target tissue (such as ischaemic brain of animal models), but it resulted in failure of functional recovery of the damaged parenchyma. 29 Surprisingly, low-dose cell delivery for treatment of ischaemic stroke through intra-arterial pathway leads to the improvement of inflammation and decreases rate of embolus formation in vessels. 30 Another undesirable side effect of intra-arterial cell administration is the fragmentation of infused cells due to the shear forces of arterial blood flow. These damaged cells may be rapidly removed from the circulation through the liver and spleen, causing shorter blood half-life of infused cells. 15 Intraperitoneal delivery is another pathway for systematic delivery of cells to the living body with controversial results. 14 It is thought that cell administration through the intraperitoneal cavity causes circumventing of pulmonary passage and consequently can lead to an increase in the number of transferred cells to the target organs. 31 However, it has been shown that cell delivery using this route leads to the aggregation of transplanted MSCs with the host immune cells after several minutes. These small and large aggregates adhere on the peritoneal membranes including omentum and mesentery. 32 These masses cannot enter the blood circulation, and only very small subsets of MSCs that do not aggregate can be visualized in the mesenteric lymph node and spleen in the initial minutes after transplantation. Moreover, no trace of infused MSCs can be found in the other organs such as heart or liver. 32 Nonetheless, the results of another study emphasize on the localization of transplanted cells in the inflamed colon, which opens up new way for treatment of inflammatory bowel disease using stem cells. 33

| Local injection
Theoretically local infusion of therapeutic cells to the parenchyma may increase the number and retention of the transplanted cells in the target tissue 34 but with certain concerns. 22 Local injection in the parenchyma is an invasive method and may lead to further damage to the target tissue. 35 Direct intramyocardial cell delivery developed cardiac arrhythmias 36 and had deteriorating effect on the heart. 37 Conversely, other studies have reported that infusion of high dosage of therapeutic cells directly to the myocardial tissue results in increased localization of transferred cells, 37 but due to the safety issues related to cell dose, implementation of this technique is not feasible. In addition, direct intraparenchymal cell delivery for treatment of kidney diseases results in accumulation of transplanted cells at the site of infusion and did not distribute throughout the renal parenchyma. 35 Eventually, administration of large amount of cells into the hepatic parenchyma produced cell embolus formation in the lung. 38 Surprisingly, there are reports implying that this pathway of injection cannot increase the cell viability and engraftment in target tissues. 22,39

| Time-point
In addition, time-point of cell transplantation into damaged tissues can have a significant effect on cellular localization, engraftment and regeneration of damaged tissues. 40 Erpicum et al demonstrated that timing of administration of MSCs has important effect on outcome of kidney ischaemia/reperfusion (I/R) injury in small animal models. 41 Findings from their study show that administration of MSCs before I/R injury has nephroprotective effect compared to MSC administration after injury. 41 MSC infusion before liver damage has significant impact on promoting liver fibrosis. On the contrary, injection of MSCs in resolution phase speeds up liver regeneration. 42,43

| Cell dose
Also, characterization of optimal cell density that can regenerate the damaged tissue without adverse effects such as tumorigenicity 44 or embolus formation 45 is controversial and there is no comprehensive consensus on optimal infused cell density. 40 This lack of consensus is due to several factors that are involved in the determination of cell dose such as type of transplanted cells, recipient's disease and route of cell transplantation. 40 However, investigators demonstrated that embolic stroke that results from intra-arterial cell delivery is due to accumulation of cells in the blood vessels and depends on the cell numbers that are transferred. 46,47

| Cell infusion rate
In addition, cell infusion rate must be adjusted in such a timing that maximal cell viability is maintained during injection. [48][49][50] High injection rates increase shear forces, resulting in cell damages and viability reduction. 15,40

| Host bio-immunological factors
It is also believed that majority of administered cells may encounter rapid clearance from the body due to the harsh and unfavourable environmental conditions such as anoikis, ischaemia, inflammation [51][52][53] and host immune reactions. 54,55 For instance, chronic inflammation at the target tissue may inhibit regeneration process by preventing transplanted cell recruitment to the damaged tissue. 51 Also, it may lead to the cellular membrane damage through production of free radicals and cytokines. 52 Consequently, the success and efficacy of cell-based therapy may be hindered.
In summary, route of migration, biodistribution, dosages, mechanical entrapment of transplanted cells due to enlarged size during successive in vitro expansion, infusion rate and host immunological factors might have detrimental effects on cell engraftment and fate in accordance with Figure 1. Therefore, proper cell tracking and determination of homing by cell imaging is critical to optimize cell administration methods and to characterize the efficacy and safety of cell-based therapies. 40 Accurate tracking and in vivo real-time monitoring of the injected cells will solve the discrepancies between various studies regarding localization, engraftment and interaction of cells with surrounding microenvironment. 56

| MOLECUL AR IMAG ING
Information about therapeutic cell function and fate is mostly obtained from fluorescence microscopy and immunohistochemical methods after obtaining biopsy samples from the patients. However, these methods are relatively invasive techniques and may lead to F I G U R E 1 Important factors that affect cell fate and efficacy after administration to living body tissue damages and disruption of cellular structures. 57,58 In addition, these experimental techniques are limited by not being able to trace cells in a real-time manner. 57 Molecular imaging technology is a growing and powerful platform that can provide valuable information about localization site and fate of cells after transplantation. 6 During the last decades, several in vivo imaging modalities have been developed for researchers to trace delivered cells ( Figure 2). However, each of them has its disadvantages that impede their applications as a perfect non-invasive in vivo imaging technique. 58 The ideal modality for molecular in vivo imaging must be able to offer accurate information about the survival, biodistribution and engraftment of cells as well as longitudinal functional real-time response of damaged tissue to cell-based therapy. 6,59,60 Furthermore, it also must show a high degree of specificity and sensitivity to obtain information about the adaptively transferred cells without inducing any harmful effects to the body. To address these requirements, it is essential to develop a multifaceted imaging technique that can reach to rapid clinical adoption. 56

| Molecular imaging and cell labelling
To track and monitor translocation and fate of administered cells, target cells have to be labelled by contrast agent or molecular probes that can act as tracers. Two main methods could be used for cell labelling in molecular imaging: direct and indirect labelling. By direct labelling, nanoparticles or chemical agents are delivered into the cell structure prior cell administration into the body. Although the ex vivo labelling of administered cells for various imaging modalities is simple and allows accumulation or internalization of dye in cell surface or internal structure (unless nucleus), there are several challenges. One major obstacle is that intensity of signals produced by labelled cells reduces with cell division over time; thus, direct cell labelling is not appropriate for long-term tracing of transferred cells in target organs. Other challenges of direct cell labelling are toxicity, bleaching and limited sensitivity of chemical agents used for cell labelling. Indirect labelling is carried out through genetic engineering of cells by reporter genes such as green fluorescent protein (GFP) or bioluminescent luciferase. Genetic modification of cells using exogenous reporter gene that target the cell nucleus results in stable expression of detectable proteins (bioluminescent or fluorescent proteins, enzymes and receptors) in target cells and future progeny.
However, this labelling is hampered to find clinical importance due to stable integration of transgene into cellular genome and risk of mutagenesis. 61,62

| Direct cell labelling
Direct cell labelling in molecular in vivo imaging can be done by various compounds including radioactive, paramagnetic or fluorescent agents. 63 For MRI (magnetic resonance imaging), the nanoparticles consist of superparamagnetic iron oxide (SPIO) nanoparticles, perfluorocarbon nanoparticles, gadolinium-filled microcapsules and liposomes. 61,64 Direct cell labelling for nuclear imaging will be implemented with radioisotopes such as 111 Indium ( 111 In 67 For establishment of various compounds as safe materials for cell labelling, several characteristics are mandatory, including lack of cellular toxicity, optimal renal clearance and stability in biological fluid together with stability during cell division. 62

| Indirect cell labelling
The indirect cell labelling allows visualization of the administered cells by the use of various reporter genes such as iron-storage protein, ferritin, in MRI detection, the herpes simplex virus thymidine kinase type 1 (HSV1-tk) and human membrane protein sodium-iodide symporter (NIS) have also been used for positron emission tomography (PET) and hybrid SPECT/CT, respectively. 61,68,69 For bioluminescent optical imaging, firefly luciferase, Renilla luciferase, Gaussia luciferase, Metridia luciferase, Vargula luciferase or Bacterial luciferase has been employed as reporter genes. 70 Finally, indirect cell labelling technique for optical fluorescence imaging is achieved by reporter genes, which express detectable proteins such as green fluorescent protein (GFP). 6,70

| Computed tomography (CT)
Imaging in computed tomography relies on differential absorption of ionizing X-rays by various tissue components in the body. 71 However, utilization of the ionizing X-rays has mutational risks and may damage DNAs. 61 Necessary instruments for CT imaging include the Xray source and rotating detector around the imaged subject. 72 Low cost compared to other non-optical imaging modalities and excellent temporal resolution are the advantages of CT scan that make it a potential technique to visualize and track stem cells. 73,74 The image contrast (differences between attenuation of the X-ray photons by various tissue) in the CT scan is relatively low for soft tissues; thus, it is imperative to use the contrast agents to distinguish between the various soft tissues. 72,73 CT scan has potential application in the cell tracking and monitoring particularly in brain and lungs whose development is relatively slower than MRI due to lower contrast of soft tissue. 73,74 Nonetheless, different studies have shown that gold nanoparticles (AuNP) can be used safely to label, monitor and detect mesenchymal stem cells by conventional CT imaging in vivo. [73][74][75] However, high dose of ionizing X-ray radiation requirements is the major disadvantage of CT scan imaging to monitor cellular localization and engraftment. 74

| Nuclear medicine: PET and SPECT
Positron emission tomography (PET) imaging is based on radiotracers that emit positron. After production, radiotracers are unstable, immediately lose their energy and generate some particles named as positrons. These particles interact with neighbouring electrons via annihilation process, and two produced photons (each having 511 keV energy) can be detected by PET scanners. 61 Because penetration in tissue depth in PET and SPECT has no limitation, their cell tracking sensitivity is high, and PET is more sensitive than SPECT. 78,79 Although labelling procedure of therapeutic cells with PET and However, because HSV1-tk has non-human origin its structure induces the immune response in host tissue. In addition, blood-brain barrier is the main obstacle for intracerebral use of this reporter gene in humans. 57,61,68 In spite of some problems concerning to genetic manipulations of therapeutic cells, indirect labelling by reporter genes provides a better choice for cell fate tracing in comparison with direct method. 5 For example, findings from previous study have revealed that NIS reporter gene imaging either by PET or SPECT can be implemented in animal studies for assessment of biodistribution, survival and engraftment of cardiac-derived stem cells in the myocardium. 78 But, in spite of high potential of PET reporter gene imaging for cell tracking, application of this technique is restricted to preclinical studies due to low resolution of PET imaging modality at cellular level 85 and genetic manipulation of transferred cells. 5

| Magnetic resonance imaging (MRI)
Magnetic resonance imaging is a kind of non-invasive imaging technique that uses a powerful magnetic field to induce polariza- By integrating reporter genes with the genome of cells, they can stably express luciferase proteins and can be monitored longitudinally for in vivo imaging. Therefore, bioluminescence imaging does not require additional excitation light source, and light scattering would be minimal due to administration of substrate inside the body.
Also, imaging depth of tissue will be possible in live small animals. 5 Additionally, as mammalian cells do not express endogenous luciferase, in vivo bioluminescence imaging offers the greatest sensitivity compared to tomographic imaging technique. 5

| B I OIMAG ING IN SWIR REG I ON
The use of visible region of electromagnetic spectrum in the range of 400-650 nm is suitable to get image from accessible or superficial tissues such as colon and skin, but not for structures locating in the deeper parts of the body such as nucleus or stem of the brain due to the scattering and absorption by tissue components. 107 During the last two decades, many efforts have been made to increase image contrast by diminishing between the tissue scattering and absorption of light photons along with reducing disruptive autofluorescence signals due to increasing the tissue depth to avoid potential deleterious effects of tissue parameter on in vivo fluorescence imaging. 108 The image with increasing depth of tissue cannot be optimal choice due to the high level of light scattering by biological tissues. 121 Light scattering in the tissue depth eventuates higher background noise-to-signal ratio and minimizes the sensitivity of NIR I light to deep scanning of tissue. 121 Thus, for further penetration of light inside the opaque tissues, imaging in the NIR I optical window cannot satisfy the clinical needs. Acquired data from water absorption characteristic in NIR I region and longer wavelength regions show a strong peak in these regions that consequently leads to reducing in image contrast. 115,118 Image contrast depends on the absorption and scattering of light photons, and in the visible and NIR I region of electromagnetic spectrum, scattering phenomenon in tissue is a Mie-type. 115,118,121 The previous studies have demonstrated that the scattering phenomena can be decreased by using longer wavelengths beyond 1000 nm. 115,121 Hence, NIR I optical imaging extends to longer wavelengths known as short wave infrared (SWIR) region results in better penetration of light to the opaque tissues. It is good to be noted that wavelength of SWIR biological window is approximately between 1000 and 2500 nm. 122 In addition to the decrease in the scattering of light in the SWIR region, autofluorescence emanated from biological tissue has reached to minimal level or could not be seen. 104,106 There are three other biological windows in the SWIR region: NIR II window (second window, at 1100-1350 nm), NIR III window (third window also called golden window, at 1600-1870 nm, ideal for brain imaging) and NIR IV window (fourth NIR window, ranging from 2100 to 2350 nm, suitable for the optical imaging of bone). 122 The third window is called as the golden window because the transparency of brain tissue is maximum in this region of the spectrum due to the higher absorption of lipid in comparison with other windows. 113 that as the lipid is the major chromophore in the second and third NIR windows, these regions can be optimal for the imaging and studying organs containing lipids such as brain, normal prostate and normal breast. Also, third and fourth windows are appropriate for normal and abnormal bone tissue assessments because of higher collagen content of bone, which acts as the main chromophore and has large absorption peak. 122 It has been demonstrated that deep tissue imaging could be possible using SWIR optical imaging due to deep photon penetration that allows higher resolution imaging compared to other modalities. 122 Non-invasive in vivo optical imaging in the SWIR region is in its beginnings and should be explored by further efforts. Deeper penetration of photon is necessary for appropriate spatial and temporal resolution at the cellular level that is an essential prerequisite for more advances in the cell-based therapies. Further advances in optical imaging using the SWIR region of spectra rely on development of powerful laser sources, sensitive camera and suitable SWIR emitter fluorophores. 68,92,107,118,122,125 Until 2014, development of in vivo optical imaging in SWIR region had been prevented mainly due to the lack of high sensitive, low-cost, high quantum yield detectors (cameras) and SWIR emitter fluorophores together with advanced laser source. Thus, SWIR technology encountered with several issues that led to the restricted development of this field. This may be mainly due to regulations pertaining to national defence such as International Traffic in Arms Regulations (ITAR). 126 91 and Pt nanowires. 142,143 In spite of various SWIR emitting fluorophore production during recent years, quantum yields of these fluorescent probes are very low. 144 (Table 1). Therefore, optical imaging in SWIR region of the electromagnetic spectrum is currently being pursued as a potential replacement for conventional imaging technique.
However, our prospect of cell tracing using SWIR imaging is that this modality can be critical in addressing of obstacles related to acceleration of cell-based therapy to clinic. Further evolution in the SWIR emitter fluorophores could allow researchers to obtain high-quality images that lack artefacts at cellular level from tissue depth without causing harmful effects on living body.

ACK N OWLED G M ENTS
The authors would like to express their sincere gratitude to Mohammad Majidi from National Cell Bank Department, Pasteur Institute, Tehran, Iran, for his useful help in editing of manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest. Optical fluorescence imaging at SWIR region of spectrum~2 5 µm up to ~3 cm Lower light scattering or absorption, negligible autofluorescence, higher signal-to-noise ratio and consequently higher image quality compared to visible and NIR I region Low penetration depth, lack of FDA-approved fluorophores for clinical use 125,149,156