Noninvasive Tracking of Hematopoietic Stem Cells in a Bone Marrow Transplant Model.

The hematopoietic stem cell engraftment depends on adequate cell numbers, their homing, and the subsequent short and long-term engraftment of these cells in the niche. We performed a systematic review of the methods employed to track hematopoietic reconstitution using molecular imaging. We searched articles indexed, published prior to January 2020, in PubMed, Cochrane, and Scopus with the following keyword sequences: (Hematopoietic Stem Cell OR Hematopoietic Progenitor Cell) AND (Tracking OR Homing) AND (Transplantation). Of 2191 articles identified, only 21 articles were included in this review, after screening and eligibility assessment. The cell source was in the majority of bone marrow from mice (43%), followed by the umbilical cord from humans (33%). The labeling agent had the follow distribution between the selected studies: 14% nanoparticle, 29% radioisotope, 19% fluorophore, 19% luciferase, and 19% animal transgenic. The type of graft used in the studies was 57% allogeneic, 38% xenogeneic, and 5% autologous, being the HSC receptor: 57% mice, 9% rat, 19% fish, 5% for dog, porcine and salamander. The imaging technique used in the HSC tracking had the following distribution between studies: Positron emission tomography/single-photon emission computed tomography 29%, bioluminescence 33%, fluorescence 19%, magnetic resonance imaging 14%, and near-infrared fluorescence imaging 5%. The efficiency of the graft was evaluated in 61% of the selected studies, and before one month of implantation, the cell renewal was very low (less than 20%), but after three months, the efficiency was more than 50%, mainly in the allogeneic graft. In conclusion, our review showed an increase in using noninvasive imaging techniques in HSC tracking using the bone marrow transplant model. However, successful transplantation depends on the formation of engraftment, and the functionality of cells after the graft, aspects that are poorly explored and that have high relevance for clinical analysis.


Introduction
Studies from the early 1950s established that total body irradiation in animal models causes death from hemorrhage and infection, indicating that the hematopoietic system is primarily affected [1].

Inclusion Criteria
Eligibility criteria were established a priori. This review included only original articles written in English, published between 2000 and 2020, that had used (i) in vivo models of bone marrow transplantations, (ii) tracking agents to analyze the hematopoietic stem cells labeled with these agents, and (iii) noninvasive techniques to allow the in vivo homing of hematopoietic stem cells in the bone marrow transplantation model.

Exclusion Criteria
Reasons for excluding studies were as follows: (i) reviews, (ii) clinical articles, (iii) book chapters, (iv) protocols, (v) editorials/expert opinions, (vi) letters/communications, (vii) publications in languages other than English, (viii) did not analyze stem cell homing with noninvasive techniques, (ix) indexed articles published in more than one database (duplicates), (x) only used invasive techniques to analyze hematopoietic stem cell homing. independently and randomly selected (in pairs), revised, and evaluated the titles and abstracts of the publications identified by the search strategy in the databases cited above, and all potentially relevant publications were retrieved in full. These same reviewers evaluated the full text articles to decide whether the eligibility criteria were met. Discrepancies in study selection and data extraction between the two reviewers were discussed with a third reviewer and resolved.  The analysis process and table plots were carried out by a full consensus of peers, respecting the distribution above. In cases of disagreement, a third, independent author decided to add or subtract data. The final inclusion of studies into the systematic review was by agreement of all reviewers.

Data Analysis
All results were described and presented using the percentage distribution for all variables analyzed in the tables.

Selection Process of the Articles Identified According to the PRISMA Guidelines
We searched publications between January 2000 and January 2020, indexed in PubMed, Scopus, and Cochrane Library, and a total of 2191 articles were identified. Of the 50 articles identified in Cochrane Library, none were included because, out of the articles, 42 were reviews, 5 were only protocols, and 3 were clinical articles. Of the 447 articles identified in Pubmed, 114 were excluded after screening (82 reviews, 24 publications before 2000, and 8 publications in other languages) and 325 articles were excluded after assessing eligibility (91 reported no data about HSC, 99 reported no data about animal models, and 135 reported no data about noninvasive imaging); thus, only 8 articles were included from this database. Of the 1694 articles identified in Scopus, after screening, 771 articles were excluded (416 reviews, 106 publications before 2000, 50 publications in other languages, and 199 duplicated in Pubmed search), and after assessing eligibility, 910 articles were excluded (287 reported no data about HSC and 623 reported no data about noninvasive imaging); thus, only 13 articles were included from this database. In total, only 21 nonduplicate full text articles were included in this review [9,11,[16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34], as depicted in Figure 1.
The main method used for harvesting stem cells from animals in the selected studies was flushing (29%) [18,29,30,32], followed by maceration (21%) [22,23,28]. The other studies reported the methods of differentiation from embryonic cells [34] and femoral bone marrow aspiration in live animals [31]. Of the studies that used humans as cell donors, only one study [26] reported the apheresis method for harvesting stem cells.
Concerning the medium and supplementation used during the harvesting of stem cells, two studies [29,30] used Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and heparin. Three other studies [17,22,23] used Phosphate-buffered saline (PBS), and the study by Parada-Kusz [23] supplemented the PBS with 0.5% bovine serum albumin (BSA)/FBS and 2 mM EDTA (Ethylenediaminetetraacetic acid). Furthermore, the study by Lopez [22] used 0.75× PBS supplemented with 5% BFS (APBS, axolotl PBS); the study by Lin [18] used minimum essential medium Eagle alpha modification (α-MEM) supplemented with 10% heparin; the study by Sweeny [34] used StemPro34 medium with a hematopoietic cytokine selection; and in the study by Niemeyer [33], the harvesting of stem cells was performed with citrate buffer.

Lentiviral Transduction of Hematopoietic Stem Cells
One of the noninvasive techniques used for HSC tracking after bone marrow transplantation is bioluminescence imaging. This technique depends on HSCs expressing the luciferase enzyme that, added to its substrate under ideal conditions, can emit light. This technique was used in 7 out of the 21 studies [9,11,[16][17][18][19][20] for HSC migration homing and tracking analysis ( Table 2). Three studies [16,17,20] used genetically modified animals that expressed the luciferase enzyme, and the other studies [9,11,18,19] performed luciferase transduction in hematopoietic stem cells to express the bioluminescence signal. The transduction process was performed with different vectors; for instance, the study by Lin [18] used a lower multiplicity of infection (MOI) ranging from 0.5 to 1, and the study by Ohmori [19] used an MOI of 20. Thus, already, the cell dose ranges from 3-30 × 10 2 [11] to 3-8 × 10 5 per well. For improving the transduction efficiency, three of four studies used the transfection agent during the process, such as polybrene (8 µg/mL) [18,19] and retronectin (100 µg/mL) [11], and the culture medium and supplementation varied among studies, as well as the cytokines used in the transduction process. The time of incubation was mainly 24 h, with the exception of the study by Lim [18], which used 48 h, and the study by Wang [9], which used two cycles of 24 h with an interval of 8 h. Luciferase expression evaluation was performed mainly, in three studies [11,18,19], by the flow cytometry technique (FCT); however, the study by Lin [18] also used the polymerase chain reaction (PCR) and colony forming unity (CFU) techniques, and the study by Wang [9] conducted only immunohistochemistry using monoclonal antiluciferase antibody.

Labeling of HSCs with Radioisotopes/Radiopharmaceuticals
HSC labeling with radioisotopes was reported in six studies [26][27][28][29][30][31], and all of the details of this process are described in Table 3. The common radioisotope used in three of the selected studies (50%) [29][30][31] was the 99mTc-exametazime (99mTc), which has a half-life of 6.03 h and it is associated with hexamethylpropyleneamine oxime (HMPAO) molecules (Ceretec, GE Healthcare). Regarding the labeling process, two of these studies [29,30] incubated 2 × 10 6 cells with 37 MBq 99mTc-HMPAO for 30 min and the yield was 15 ± 3% after labeling, with an activity to administration of 5.55 MBq; the study by Lange [31] has already labeled with the range of 550-583 MBq 99mTc-HMPAO, and after labeling, the radioisotope activity was between 669-1350 MBq.
Another two studies [26,28] used the zirconium-89 (89Zr) radioisotope obtained by Cyclotron, and one study [28] added an oxine molecule to the radioisotope, which has a half-life of 78.4 h, but the labeling process differed between studies. The study by Asiedu [28] incubated 10 6 cells with 0.01-5.55 MBq 89Zr-oxine for 20 min, resulting in a cell-associated radioactivity of 0.0036-1.7 MBq and a yield of 26-30%; the study by Pantin [26] already incubated 2 × 10 8 cells with 0.37 MBq 89Zr for 30 min, and after labeling, the cell activity was 0.28 MBq. The study by Faivre [27] used [18F] fluorodeoxyglucose (18F-FDG) obtained by Cyclotron, which has a half-life of 1.83 h, and labeling was performed with a radioisotope activity ranging from 301.8 to 945.9 MBq after labeling of 5-10 MBq, with a purity of 94.6 ± 6%.

Labeling of HSCs with Fluorophore
HSC labeling with fluorophore was reported in four of the selected studies [21,[23][24][25], although in the study by Lopes [22], genetically modified animals were used that express green fluorescent protein (GFP), which, thus, required no labeling process, as described in Table 4. Most of the selected studies used fluorophore that emits radiation in the range of visible light wavelength, with the exception of the study by Ushiki [25], which used fluorophore that emits in the near-infrared fluorescence wavelength. The agent concentration and time of incubation during the labeling process varied among the studies, and two of the studies [22,25] reported a cellular toxicity evaluation after labeling, using mainly the flow cytometry technique.
3.6.1. Bioluminescence Table 7 shows that the bioluminescence images were acquired after luciferin administration; this substrate was applied in most of the studies (86%) by intraperitoneal administration [11,[16][17][18][19][20], using a dose that ranged from 0.75 to 150 mg/kg with different time intervals (2-15 min) to begin the image acquisition. The Xenogen IVIS system equipment was used in all studies, and although the model varied, the same software was used to process the images; that is, the Living Image software. The selected studies that used mice [9,11,16,18,20] acquired images mainly in the dorsal and ventral animal positions; the study by Steiner [11] also used the lateral position. Among the few BLI acquisition parameters reported by the selected studies, the exposure time used in the image acquisition varied among studies in accordance with the signal finding, but three studies [9,11,16] reported a decrease in the exposure time over the weeks (3 to 1 min/1 s) due to rise of signal intensity. Binning acquisition parameters was reported only in the study by Astuti [17], with the 4 or 8 values. The HSC tracking was performed from 1 day after bone marrow transplant in three studies (43%) [16,17,20], and in another four studies [9,11,18,19], bioluminescence imaging only was detectable at the end of the first week. Bioluminescence imaging was used for follow-up evaluation of longitudinal HSC tracking in all studies [9,11,[16][17][18][19][20].

PET/SPECT
Of the six studies that tracked cells with radioisotope agents (Table 8), three of them (50%) [29][30][31] used the single-photon emission computed tomography (SPECT) technique with 99mTc-HMPAO by gamma camera equipment, two [29,30] used a window energy of 5% centered over 140 KeV and dynamic acquisition, and the final study [31] used whole-body acquisition. The time of the cell homing evaluation of these studies was 30 min [29,30] and 24 h [31], detecting uptake distribution commonly to the heart, liver, lung, spleen, and other regions.
Another three studies (50%) [26][27][28] used hybrid positron emission tomography-computed tomography (PET-CT) scanning by a micro-PET-CT scanner [27,28] or a Philips clinical PET-CT scanner [26]. The studies by Asiedu [28] and Pantin [26] used 89Zr radioisotopes, a whole-body image with a 400 to 700 KeV energy window, and the homing evaluation was reported in intervals of 2-4 h or 1-7 days in the former study and 5-15 h in the latter study. The study by Faivre [27] used 18F-FDG with dynamic image acquisition and a 511 KeV energy window for 3 h of homing evaluation. Regarding the uptake distribution of cells after transplantation, the selected studies [26][27][28][29][30]

that implanted cells by systemic administration (intravenous) reported greater concentrations of HSCs labeled in the lung.
In two studies [26,29] that implanted the cells intraosseously, the cell distribution to the lung was lower, and posterior cell migration for other organs was reported.

Fluorescence Imaging
Fluorescence imaging was performed in five of the studies [21][22][23][24][25] for HSC tracking analysis after the bone marrow transplant (Table 9). In the studies that used zebrafish or axolotl models, microscopy was used for imaging acquisition [21][22][23][24], and in the study by Ushiki [25], which used mice as receptors, the HSCs were tracked using the IVIS Spectrum system (Xenogen, Alameda, CA, USA). FLI acquisition parameters were reported only in the study by Ushiki [25]. Two studies reported the acute evaluation of HSCs after transplantation (until 24 h), but the study by Lopes reported longer homing evaluation (until 6 days) with positive detection until three days. Only the study by Ushiki [25], which used mice as cell receptors, was the uptake distribution reported in the bone marrow, lung, spleen, liver, and kidney, corroborating the ex vivo analysis [25]; in the studies that used zebrafish, the uptake distribution was reported mainly for the tail [21,23,24].

Magnetic Resonance Imaging
Of the selected studies, MRI was performed in three studies [32][33][34] to evaluate HSC tracking after the bone marrow transplant (Table 10). The study by Niemeyer [33] only evaluated the acute stage from 2-24 h, using low magnetic field equipment (1.5 Teslas), but the other two studies [32,34] reported a long evaluation, from 3 h [34] to 14 days [32], using equipment with a higher magnetic field (4.7 [34] and 17.6 [32] Teslas). These three studies [32][33][34] reported uptake distribution for the bone marrow, liver, and spleen. MRI acquisition parameters used in the selected studies showed that the T2 weighted image was widely used with variations in the field of view, matrix, slice thickness, and others. In addition, this technique was the most detailed in terms of image acquisition parameters among the selected studies.
All aspects of the analysis of the noninvasive tracking of hematopoietic stem cells in a bone marrow transplant model are shown in Figure 2.

Discussion
The results of this review show an increase in using noninvasive imaging techniques in HSC tracking in the bone marrow transplant model. The most selected studies used bioluminescence, followed by fluorescence, PET/SPECT, and MRI. The use of noninvasive imaging techniques can reveal that hematopoietic activities in both steady-state and pathological conditions are dynamic, and that their sequence is regulated spatiotemporally by interaction with the niche [55]. Successful clinical trials show that the engraftment depends on adequate HSC numbers, their homing, and the subsequent short-term and long-term engraftment of these cells in the niche (bone marrow). Enhancing the homing capability of HSCs could have a great impact on improving transplantation procedures and patient survival [56]. The main challenge of leukemia or the dysfunctional or depleted bone marrow preclinical model is the mimicking of the bone marrow niche or the environment. Experimental findings in mice often correlate with human biology and, as such, they serve as a research stand-in for human patients [57]. In this review, most of the selected studies used mice as the animal model, and regarding the type of graft, allogeneic cells were the most used among the studies. In exploring this aspect, it was reported that when comparing allogeneic and xenogeneic grafts, using the cell route by the tail vein, the allogeneic cells arrive at the niche faster (3 h after transplantation) than in the xenogeneic cells (15 h after transplantation), and in a greater amount, being detectable in the endosteal region of the femur (13% of cell quantity) and in the central marrow region (58% of cell quantity), associated with efficiency and marrow repopulating ability [58].
Irradiation is an important aspect of HSC homing analysis. The study by Xie [59] showed that only 7% of cells after 6 h of transplantation arrive at the bone marrow niche compared to the irradiated hosts, which obtained better efficiency with 35% of cells detectable in the niche. In addition, immunoassays identified the endosteal zone of irradiated recipients as a site of increased HSC proliferation after transplantation. In this review, it was observed that most rodent models of the selected studies applied bone marrow irradiation with cesium-137 or X-ray, and cell homing analysis was possible and detectable by noninvasive image and other techniques, particularly the flow cytometry technique.
Optical imaging certainly holds utility in assessing hematopoietic stem cell tracking of xenografts, allografts, autologous grafts applied in the bone marrow transplant model [35]. The same study reported that bioluminescence imaging is certainly more sensitive than fluorescence imaging. However, one strength of fluorescent fluorophores is the ability to perform high-resolution microscopy of GFP-expressing cells. As a result, if tracking of small numbers of cells at high magnification and perhaps in real-time is desired, fluorescence-based methods can be considered. Besides these techniques, this review reported the used of other techniques, such as PET/SPECT and MRI in cell homing, where PET/SPECT has a small temporal window in signal detection and MRI requires a magnetic contrast agent that may interfere with cellular growth biology and the interaction of the environment in a niche (bone marrow), decreasing the sensitivity of this technique.
Most of the selected studies of this review used bioluminescence due to its reported high sensitivity, with a wide temporal detection window (0 days to 1 year), but a limitation of this technique is that the light propagation through tissue currently restricts the application of bioluminescence imaging to small animals, where the signal can easily penetrate at all depths, and it is a technique that has not yet been well studied in vivo [9]. Another prevalent technique used by six of the selected studies was PET/SPECT imaging, which provides three-dimensional information with high sensitivity, excellent penetration depth, and a capability for whole-body imaging, but which requires radioisotopes. An advantage of this technique is that it can also be applied in small animals as well as in humans, and the disadvantages include the high cost of the cyclotron and limited temporal follow-up due to the radioisotopes' half-life-a factor that was reported in the selected studies, where cell homing was commonly evaluated at 15 h. In addition, in the selected studies, the radioisotope most used was the 99mTc-HMPAO, which has high translation capacity, but it is important to bear in mind the potential effects of radiation on therapeutic/biology cell function [60].
The fluorescence imaging technique has a high sensitivity (lower than bioluminescence), a high temporal resolution, a low cost, and it is activatable, but the disadvantages in relation to the other image techniques described before include the attenuation of sensitivity by overlying tissues, low spatial resolution, and poor penetration depth. Therefore, this technique is the most important imaging modalities for live-cell imaging at single-cell resolution using light sources to excite the fluorescent molecules [61]. In this review, the fluorescence technique observed HSC homing into the niche up until 24 h.
MRI has a high spatial and temporal resolution, no tissue penetrating limit, and no radiation, but the disadvantages include relatively low sensitivity and low contrast, requiring a high load of cells with a magnetic label and comparatively long imaging times, with the possibility of affecting cell viability or biological interaction with the contrast agent [43,[62][63][64]. In this review, three studies used HSC tracking analysis by MRI with different magnetic fields, in which the highest magnetic field (17.6 Teslas) provided longer follow-ups (14 days after transplantation) when using low nanoparticle concentrations and incubation times in the cell labeling process. Our group provided evidence, in previous studies, that a low nanoparticle concentration in the labeling process was enough to detect stem cell migration homing and tracking. Additionally, nanoparticle sizes of between 35 and 200 nm used in our previous studies showed magnetic characteristics adequate for detection by MRI [65][66][67]. This review corroborates the nanoparticle size findings among the selected studies (60 to 180 nm).
Regarding the molecular imaging techniques approaches in this review, their physical principles, applicability, advantages, and limitations ( Figure 3) [43,46,47,[68][69][70][71][72][73], show a wide potentiality not only for in vitro studies and pre-clinical applications but also in the translation of some techniques in clinical studies, such as nuclear images (PET and SPECT) and MRI [47,73]. Comparing the molecular imaging technique characteristics and the HSC tracking evaluation shown in Figure 3, the optical techniques allowed a live cell tracking for a long period of time by BLI (at 1 day to 12 months)) [9,11,[16][17][18][19][20] and short time by FLI (at 30 min to 6 days) [21,[23][24][25], but in the latter, the signal is not necessarily from living cells. In nuclear techniques, the time window varies with the radioisotopes used due to their half-life, in SPECT the radioisotopes have a shorter half-life, and therefore, the time allowed for tracking the cells was also shorter (at 30 min to 1 day) [29][30][31], compared with cell tracking made by PET (at 1 h to 7 days) [26][27][28], in which the radioisotopes used have an average half-life of 78.4 h, as is the case with 89 Zr. MRI has high spatial resolution and a wide temporal window for cell tracking; however, for all this, a high nanoparticle load is required for its detection, as its sensitivity is low (10 −3 -10 −5 mol/L) [72] compared to the other techniques, in addition to decreasing the signal temporally due to cell division, making it difficult to keep track of the cells [74].
Therefore, for preclinical studies of the cell transplantation model, all molecular imaging techniques covered in this review have good applicability in cell tracking for early assessment, but the BLI optical technique stands out in the prolonged tracking assessment [11]. In the clinical application of oncological diseases, the PET nuclear technique and MRI would be the most suitable for early assessment of cell tracking [26][27][28]33,34]. However, cell graft was not evidenced by any of the molecular techniques mentioned in this review, and for cell graft to occur, we must have good control of cell tracking. The studies selected in this review used other techniques to quantify chimerism upon transplantation. The technique most used to access chimerism in the studies selected in this review was the FCT to determine the percentage of grafted donor cells grafted using specific cell line markers such as CD45 expressed by all lymphocytes [9,11,19,23,24,27,28,33], or GFP expressed by cells after genetic modification [17,22,25,34]. The CD45, in addition to making it possible to analyze the occurrence of engraftment, also allows checking the percentage of chimerism through discrimination of its two different alleles (CD45.1 and CD45.2), which are functionally identical. CD45.1+ donor cells can be readily detected when transplanted into CD45.2 mice, as reported in two studies in this review [19,28]. In these studies, the efficiency assessment was performed at early (2 to 4 days after implantation) [23,25,33] and late time points after transplantation (after 20 days to 365 days after implantation) [9,11,16,19,20,27,28]. The evaluation of cell grafting is one of the most relevant aspects of clinical application in cancer patients, as the biodistribution and efficiency of implantation are the targets for the success of the treatment, aspects that still have many gaps in clinical research. Successful transplantation depends on the formation of engraftment, in which donor cells are integrated into the recipient's cell population, as well as a supportive hematopoietic stromal microenvironment [75]. Evaluation of this was performed in most of the selected studies, which showed that the studies that used allogeneic cell transplantation had greater graft efficiency in a shorter time period than xenogeneic transplantation in the experimental model. Another relevant aspect of the success of transplantation is the functionality of cells after the graft, because graft failure and poor graft function are important issues in the current comprehensions of the interaction between the immune and hematopoietic compartments in these conditions [76]. Therefore, besides the quantitative analysis of the graft, the graft functionality should be adequate in the evaluation of graft efficiency, but this aspect was poorly approached in most of the selected studies.
Unfortunately, there are not enough clinical HSCs available to set up phase I/II clinical trials to test the tracking of these new cells by noninvasive imaging techniques. Most investigators conducting such trials are "wed" to their personal favorite procedure. If, in the future, we can deal with this problem and find means for additional clinical efforts, it is possible that several new procedures could be used together [34], such as hybrid equipment or imaging systems. This, however, adds additional logistical problems versus the use of one procedure alone, such as requiring multifunctional probes, smart processes, and improvement of the technical limitations of imaging equipment [77]. Therefore, there is viability to translate these experimental findings into bed-side application. The cell tracking improvement of niches provides a reduction in therapeutic time with high efficiency of hematopoietic cellular renewal.