Th cells sensitized against autoantigens acquire pathogenicity following two sequential events, namely activation by their target Ag and a process named “licensing.” In this study, we analyzed these processes in a transgenic mouse system in which TCR-transgenic Th cells specific to hen egg lysozyme (HEL) are adoptively transferred to recipients and induce inflammation in eyes expressing HEL. Our data show that the notion that the lung is the organ where “licensing” for pathogenicity takes place is based on biased data collected with cells injected i.v., a route in which most transferred cells enter via the lung. Thus, we found that when donor cells were activated in vitro and injected intraperitoneally, or were activated in vivo, they migrated simultaneously to the lung, spleen, and other tested organs. In all, tested organs donor cells undergo “licensing” for pathogenicity, consisting of vigorous increase in number and changes in expression levels of inflammation-related genes, monitored by both flow cytometry and microarray analysis. After reaching peak numbers, around day 3, the “licensed” donor cells migrate to the circulation and initiate inflammation in the HEL-expressing recipient eyes. Importantly, the kinetics of increase in number and of changes in gene expression by the donor cells were similar in lung, spleen, and other tested organs of the recipient mice. Furthermore, the total numbers of donor cells in the spleen at their peaks were 10- to 100-fold larger in the spleen than in the lung, contradicting the notion that the lung is the organ where “licensing” takes place.

Studies with animal models of autoimmune diseases have established the critical role played by Th cells in the pathogenic processes of many of these diseases (13). The crucial involvement of Th cells in autoimmune processes has been directly demonstrated by the finding that autoimmune animal disease models could be adoptively transferred to naive recipients by lymphocytes sensitized against Ags specifically expressed by the target organ (46). Further analysis of the process revealed that to be effective, the Th cells need to be activated prior to their being adoptively transferred (79). The onset of the transferred disease was not immediate, however, and early publications suggested that the delay is required to allow a small fraction of the transferred cells to enter the target organ and promote the later invasion of most transferred cells by affecting the tissue–blood barriers and inducing efficient expression of MHC molecules on local APCs (10, 11). A later study, however, by Flügel et al. (12), provided evidence to show that the delayed onset of disease could be attributed to a process in which prior to invading the target organ the transferred T cells enter lymphoid organs, where they acquire a new phenotype, which apparently is critical for their invasiveness into target organs. In a more recent publication, however, Odoardi et al. (13) identified the lung as the organ where the “licensing” for pathogenicity by the transferred cells is taking place.

The present study was aimed at analyzing the process of “licensing” for pathogenicity in an experimental system in which Th lymphocytes specific against hen egg lysozyme (HEL) are adoptively transferred into recipients in which HEL is a neo–self-antigen, selectively expressed in the recipient eyes. When activated in vitro (1416) or in vivo (17, 18), the transferred Th cells initiate inflammation in the recipient eyes. Our observations with these two modes of Th activation, recorded in the present study, show that the major “licensing” organ in the recipient animal is the spleen, not the lung. Moreover, our data show that the involvement of the lung in the “licensing” process is marginal in systems in which the donor cells are administered via routes other than i.v. The study also yielded new information concerning inflammation-related molecules involved in the lymphocyte activation and “licensing” processes.

All mice used in this study were (FVB/N × B10.BR) F1 hybrids, transgenically expressing either HEL in their eyes (HEL transgenic [HEL-Tg]), or HEL-specific TCR by their T cells (3A9) (for details, see Refs. 16, 19). All mice were kept under specific pathogen-free conditions in the National Eye Institute’s animal facility, and all experimental procedures were performed in compliance with the National Institutes of Health Resolution on the Use of Animals in Research (under protocol NEI-555).

HEL was purchased from Sigma-Aldrich. The following reagents were from BD Biosciences: 7-aminoactinomycin D, PE–anti-CD4, IgG isotype control, PE–anti-CD25, PE–anti-CD44, PE–anti-CD69, PE–anti-CD62L, and allophycocyanin–anti-CD4. PE–anti-CCR7 Abs were purchased from e-Bioscience. A clonotypic mAb specific for the TCR of 3A9 mice, designated 1G12, a gift from E. Unanue (Washington University, St. Louis, MO), was conjugated with FITC. LPS was from Sigma-Aldrich, and Percoll and Ficoll were purchased from GE Healthcare. Endotoxin-free phosphorothioate oligodeoxynucleotide (ODN) was synthesized at the Core Facility of the Center for Biologics Evaluation and Research, Food and Drug Administration (Bethesda, MD). The sequence of the CpG ODN used here (1555) is 5′-GCTAGACGTTAGCGT-3′.

Naive CD4+ T cells were purified from spleen and lymph nodes of 3A9 mice, using T cell columns (R&D Systems), followed by MACS microbeads (Miltenyi Biotec). The naive CD4 cells were cultured in 12-well plates (Corning) at 25 × 104 cells/ml in a volume of 2 ml of RPMI 1640, supplemented with 10% FCS, antibiotics, and 50 μM 2-ME. Activation was induced by HEL (2 μg/ml) and irradiated (30 Gy) syngeneic wild-type naive splenocytes, serving as APCs (250 × 104/2 ml).

CD4 cells (5 × 106) activated for 4 d were harvested and injected either i.v. or i.p. into HEL-Tg mice. Recipient mice were sacrificed at different time points (5 min, 2 h, 1, 2, 3, 4, or 7 d) after cell injection and their lungs, spleen, liver, and parathymic lymph nodes were collected. Additionally, blood samples of the mice were collected and the lymphocyte population was isolated by Ficoll following the manufacturer’s instructions. Spleen and parathymic lymph node lymphocytes were isolated by conventional procedures. Lymphocytes located in the lung were harvested by cutting the organ into small pieces with scissors, followed by incubation with 0.3% collagenase (Roche) for 1 h at 37°C. Subsequently, single-cell preparations were obtained by filtration through a 40-μm cell strainer (Falcon), washed in RPMI 1640 medium, and the lymphocytes were isolated by Ficoll gradients. Liver lymphocytes were collected from recipient mouse livers following perfusion with PBS. The liver tissue was cut to small pieces that were pressed through a 70-μm strainer and lymphocytes and erythrocytes were isolated on a 37% Percoll gradient, for 20 min, at 2000 rpm at room temperature. The erythrocytes were eliminated by lysing buffer.

Naive CD4 cells were isolated from spleens and lymph nodes of 3A9 mice as described above and were injected i.v. into naive HEL-Tg mice, at 5 × 106 per mouse. On the following day, recipient mice were injected i.v. with HEL and TLR ligands. The doses we used per mouse were 100 μg of HEL, 50 μg of LPS, or 240 μg of CpG ODN. At different time points (1, 2, 3, 4, or 7 d after TLR ligand injection), recipients’ lungs and spleens were collected and lymphocytes were isolated as described above.

Cell preparations, as indicated, were analyzed by flow cytometry for expression of surface molecules using the manufacturers’ instructions. The individual Abs are listed in the “Materials” paragraph. Donor cells were gated according to their expression of the 1G12 molecule. Cells were acquired on a MACSQuant analyzer (Miltenyi Biotec) and the data were analyzed by FlowJo software (Tree Star).

NanoString technology is based on standard hybridization between the target gene and target-specific capture and reporter probes (20). Lymphocytes isolated from the spleen and lungs of recipient mice were sorted on a BD FACSAria II, to isolate donor cells, by gating on CD4+1G12+ cells. Total RNA was isolated from sorted cells of three or four recipient mice per time point using an Arcturus RNA isolation kit (Applied Biosystems) according to the manufacturer’s protocol. In brief, 100  ng of total RNA was hybridized to nCounter reporter probe sets for at least 12 h at 65 °C. The hybridized samples were transferred to the nCounter sample preparation station, and color-coded barcodes on the reporter probes were read and quantified by using an nCounter digital analyzer (NanoString Technologies). Expressions of 560 genes were analyzed using an nCounter GX mouse immunology kit.

NanoString data normalization was conducted with nSolver analysis software 2.0. Data were normalized using positive and negative controls and housekeeping gene probes. Hierarchical cluster analysis was performed using Pearson correlation for distance measure algorithm to identify samples with similar patterns of gene expression (MultiExperiment Viewer v4.8).

Combined data of repeated experiments are shown as the mean ± SEM. The software GraphPad Prism was used to perform the statistical analyses of the data with a two-tailed Student t test. Differences were considered significant at p < 0.05.

Th lymphocytes sensitized against HEL induce ocular inflammation when adoptively transferred to recipients expressing HEL in their eyes (14, 16). Supplemental Fig. 1 demonstrates the typical histopathological changes in eyes of recipient mice. The changes mainly include infiltration of inflammatory cells in various eye components, including the retina, vitreous, and anterior chamber.

To identify the homing organs for the adoptively transferred activated lymphocytes, we used a system in which the transferred lymphocytes (3A9) are traced in the recipient mouse by a mAb (1G12) specific against their HEL-specific TCR (21). The combined data of three experiments are recorded in Fig. 1A and show that the i.v. transferred Th donor cells immediately migrated to the recipient lung: at the 5 min posttransfer time point the transferred cells constituted ∼20% of the total CD4 cells in the lung. The great majority of donor cells exited the lung, however, within 24 h (Fig. 1A). The small number of donor cells retained in the lung (5% at day 1 posttransfer) increased vigorously in number, reaching their peak on day 3 posttransfer, and their numbers gradually declined thereafter. Donor cells that migrated to the recipient spleen (0.5% of total CD4 cells at 1 d) rapidly increased in number in this organ as well. Importantly, as seen in Fig. 1A, the kinetics of increase and gradual decrease of donor cell (1G12+) proportions among the CD4 populations in the spleen closely resemble those in the lung when determined at the time points beyond day 1.

FIGURE 1.

Similar kinetics of changes in donor cell populations in recipients’ lungs and spleens during the “licensing” period. In vitro–activated 3A9 mouse donor cells were adoptively transferred by the i.v. route to HEL-Tg recipients (expressing HEL in their eyes) (5 × 106 per recipient) and monitored in the lungs and spleens of the recipients at different time points after transfer. (A) Analysis of percentage of donor cells (1G12+) among the total CD4 populations in the tested organs. Note that the percentage of donor cells in the spleen (green) and the lung (purple) are recorded on separate y-axes. The data demonstrate the immediate accumulation of donor cells in the lung, shown at the 5 min time point, followed by a sharp decrease during day 1. Importantly, similar kinetics of increase in number of donor cells in the lung and spleen of the recipient mice are seen when determined from day 1 after cell injection, reaching their peaks on day 3 and gradually decreasing thereafter. (B) Total numbers of donor cells in the recipients’ spleens and lungs, presented as mean cell number per organ. Note that the number of donor cells in the spleen was larger by far than that in the lung. Combined data of three separate experiments, with 12 mice being used for each experiment and 2 mice of each group sacrificed for each time point, are shown. *p < 0.05, ***p < 0.001.

FIGURE 1.

Similar kinetics of changes in donor cell populations in recipients’ lungs and spleens during the “licensing” period. In vitro–activated 3A9 mouse donor cells were adoptively transferred by the i.v. route to HEL-Tg recipients (expressing HEL in their eyes) (5 × 106 per recipient) and monitored in the lungs and spleens of the recipients at different time points after transfer. (A) Analysis of percentage of donor cells (1G12+) among the total CD4 populations in the tested organs. Note that the percentage of donor cells in the spleen (green) and the lung (purple) are recorded on separate y-axes. The data demonstrate the immediate accumulation of donor cells in the lung, shown at the 5 min time point, followed by a sharp decrease during day 1. Importantly, similar kinetics of increase in number of donor cells in the lung and spleen of the recipient mice are seen when determined from day 1 after cell injection, reaching their peaks on day 3 and gradually decreasing thereafter. (B) Total numbers of donor cells in the recipients’ spleens and lungs, presented as mean cell number per organ. Note that the number of donor cells in the spleen was larger by far than that in the lung. Combined data of three separate experiments, with 12 mice being used for each experiment and 2 mice of each group sacrificed for each time point, are shown. *p < 0.05, ***p < 0.001.

Close modal

Importantly, note that because the total number of CD4 cells in the spleen is larger by far than that in the lung, the total number of donor cells (1G12+) in the spleen was profoundly greater than that in the lung, despite the higher percentage of donor cells among the lung’s CD4 population (Fig. 1A). Thus, the total number of donor cells in the recipient spleen was ∼10-fold higher than that in the lung, as early as day 1 after cell transfer (Fig. 1B).

To further learn about the destination and fate of the injected activated donor cells in the recipient mice, we analyzed the percentage of donor cells (1G12+) in the liver and blood, as well as in the spleen (for comparison), at certain time points after cell injection by the i.v. route. The data of three combined experiments are summarized in Fig. 2A. The pattern of changes in percentage of donor cells in the blood resembled that seen in the spleen, reaching the peak on day 3 after cell injection and sharply declining on day 4. A partially different pattern was found in the liver: the peak percentage of donor cells was reached on day 2, with a slight reduction on day 3, followed by a sharp decline on day 4.

FIGURE 2.

Patterns of changes in the populations of donor cells in the liver, spleen, and blood of donor cells, activated in vitro and injected i.v. Procedures were carried out as detailed in the legend for Fig. 1. (A) Proportions of donor cells (1G12+) among the total CD4 populations in the three organs. Note that the percentages of donor cells in the liver (purple), spleen (green), and blood (blue) are recorded on separate y-axes. The data show that the proportions of donor cells increased in the liver faster than in the spleen, whereas the changes in the blood resembled those in the spleen. (B) Total numbers of donor cells in the spleen, liver, and blood circulation at the different time points. Combined data of three separate experiments, with 12 mice being used for each experiment and 2 mice of each group been sacrificed for each time point, are shown.

FIGURE 2.

Patterns of changes in the populations of donor cells in the liver, spleen, and blood of donor cells, activated in vitro and injected i.v. Procedures were carried out as detailed in the legend for Fig. 1. (A) Proportions of donor cells (1G12+) among the total CD4 populations in the three organs. Note that the percentages of donor cells in the liver (purple), spleen (green), and blood (blue) are recorded on separate y-axes. The data show that the proportions of donor cells increased in the liver faster than in the spleen, whereas the changes in the blood resembled those in the spleen. (B) Total numbers of donor cells in the spleen, liver, and blood circulation at the different time points. Combined data of three separate experiments, with 12 mice being used for each experiment and 2 mice of each group been sacrificed for each time point, are shown.

Close modal

Fig. 2B summarizes the total numbers of donor cells in the liver, spleen, and blood circulation at the different time points. The largest numbers of donor cells were found in the spleen, but considerable numbers of these cells were also present in the liver and blood, at the tested time points. It is noteworthy that the number of donor cells in the blood was negligible on day 1, suggesting that all injected cells already homed to the “licensing” organs at that time point. The number of donor cells in the blood increased sharply on day 2, when “licensed” cells began their migration toward the target tissue harboring HEL.

In view of the lung being the site of entrance for cells injected i.v. (Fig. 1A) (13), we examined the role of the lung in homing patterns of Th cells injected to recipients via another route, that is, i.p. In vitro–activated Th cells transferred by the i.p. route efficiently transfer experimental autoimmune encephalomyelitis (12, 22) and were found to transfer ocular inflammation in our experimental system (data not shown). In parallel with the lung, we determined the homing patterns to the spleen and the parathymic lymph nodes; these nodes were found by Flügel et al. (12) to be, in the rat, the first homing lymphoid organ for i.p. injected lymphocytes. Fig. 3 summarizes combined data of three experiments in which 3A9 cells activated in vitro were administered to HEL-Tg recipients by the i.p. route. Unlike the immediate homing to the lung of considerable numbers of cells injected i.v. (∼20% of total CD4 cells at the 5 min time point; Fig. 1A), only ∼1% donor cells injected i.p. were found in the lung at that time point. Furthermore, in contrast to the sharp decline in donor cell proportions in the lung of the i.v. injected recipients (Fig. 1A), a moderate increase in the proportion of donor cells was measured in lung on day 1 after cell injection (Fig. 3A), followed by a sharp increase (∼20%) on day 2 and by a gradual decrease on day 3. The kinetics of homing and increase in number of donor cells in the spleen of recipients injected i.p. resembled those seen in mice injected i.v. (Fig. 1A). Of interest are the observations with the parathymic lymph nodes: considerable numbers of donor cells were detected in these nodes at the 1 d time point (Fig. 3B), increasing rapidly and reaching their peak on day 2 after cell transfer. A gradual decrease in donor cell proportions was observed at the later time points. It is noteworthy that the peak percentage of donor cells among the total CD4 cells in the parathymic lymph nodes and the lung was reached ∼1 d prior to that in the spleen. It is also notable that the total number of donor cells in the spleen at the peak time was >100-fold larger than that in the lung, and even the total number in the parathymic lymph nodes was ∼10-fold larger than that in the lung.

FIGURE 3.

Donor cells injected by the i.p. route differ from those injected by the i.v. route in their patterns of homing to recipient organs. (A) Donor 3A9 cells, activated in vitro, were injected into HEL-Tg recipients by the i.p. route (5 × 106 per recipient) and were tracked in the lung, spleen, and parathymic lymph nodes by the 1G12 Ab at the indicated time points. Note that 1) just small numbers of donor cells migrate to the lung at the early time points and their numbers increase only after day 1 postinjection, and 2) donor cell proportions in the lung and parathymic lymph nodes reached their peak on day 2, whereas in the spleen the peak was reached on day 3. (B) Total numbers of donor cells in the three organs at the different time points. Note the drastically small number of donor cells in the lung as compared with those in the spleen and parathymic lymph nodes. Combined data of three separate experiments, with 12 mice being used for each experiment and 2 mice of each group being sacrificed for each time point, are shown.

FIGURE 3.

Donor cells injected by the i.p. route differ from those injected by the i.v. route in their patterns of homing to recipient organs. (A) Donor 3A9 cells, activated in vitro, were injected into HEL-Tg recipients by the i.p. route (5 × 106 per recipient) and were tracked in the lung, spleen, and parathymic lymph nodes by the 1G12 Ab at the indicated time points. Note that 1) just small numbers of donor cells migrate to the lung at the early time points and their numbers increase only after day 1 postinjection, and 2) donor cell proportions in the lung and parathymic lymph nodes reached their peak on day 2, whereas in the spleen the peak was reached on day 3. (B) Total numbers of donor cells in the three organs at the different time points. Note the drastically small number of donor cells in the lung as compared with those in the spleen and parathymic lymph nodes. Combined data of three separate experiments, with 12 mice being used for each experiment and 2 mice of each group being sacrificed for each time point, are shown.

Close modal

Early studies on adoptive transfer of autoimmune disease revealed that to be pathogenic, sensitized lymphocytes require in vitro activation prior to transfer to the recipient animal (79). Little is known, however, about the migration and homing patterns of adoptively transferred naive cells, and we used our experimental system to collect information on these issues.

Fig. 4A demonstrates the homing and recovery patterns of naive 3A9 CD4 cells injected i.v. to recipient mice and traced in the spleen and lung at different time points. The injected cells migrated immediately to the lung, where they constituted a high proportion of total CD4 lung cells at the 5 min time point (∼30%), but rapidly exited the lung, declining to only ∼3% of the CD4 population at the 2 h time point and to ∼1% at the 24 h time point. The proportion of naive cells in the spleen increased to ∼5% of total CD4 cells at the 2 h time point, but declined to ∼2% at the 24 h time point. Importantly, unlike the pattern seen with the injected activated donor cells, the naive donor cells did not increase in number in either the lung or the spleen at any tested later time point (Fig. 4A). Fig. 4B shows the total numbers of donor cells in the recipient lung and spleen. In line with the values recorded in Fig. 4A, the highest number of donor cells in the lung was found at the 5 min time point and rapidly declined at the later time points. The number of donor cells in the spleen reached its peak at the 2 h time point and declined at the later time points.

FIGURE 4.

Naive CD4 cells remarkably differ from in vitro–activated donor cells in their homing and proliferation patterns. (A) Naive CD4 cells of 3A9 mice were injected i.v. to HEL-Tg recipients (5 × 106 per recipient) and their percentages among the CD4 populations of the spleen or lung were determined at the indicated time points. Note the profound differences between the profiles of naive cells recorded here and those of the in vitro–activated CD4 donor cells recorded in Fig. 1A. (B) Total numbers of donor cells in the lungs and spleens of recipients of the naive CD4 cells. Note the absence of increase in the number of donor cells in the spleen and lung at the later time points, in contrast to observations with activated donor cells, recorded in previous figures. Combined results of two separate experiments, with closely similar data, are shown. *p < 0.05, ***p < 0.001.

FIGURE 4.

Naive CD4 cells remarkably differ from in vitro–activated donor cells in their homing and proliferation patterns. (A) Naive CD4 cells of 3A9 mice were injected i.v. to HEL-Tg recipients (5 × 106 per recipient) and their percentages among the CD4 populations of the spleen or lung were determined at the indicated time points. Note the profound differences between the profiles of naive cells recorded here and those of the in vitro–activated CD4 donor cells recorded in Fig. 1A. (B) Total numbers of donor cells in the lungs and spleens of recipients of the naive CD4 cells. Note the absence of increase in the number of donor cells in the spleen and lung at the later time points, in contrast to observations with activated donor cells, recorded in previous figures. Combined results of two separate experiments, with closely similar data, are shown. *p < 0.05, ***p < 0.001.

Close modal

We have previously reported that naive 3A9 donor cells that have no pathogenic capacity do acquire this capacity in recipient mice treated with TLR ligands, such as LPS or the ODN CpG (17, 18). In the in vivo activation experimental system, the TLR ligands are injected, along with HEL, to the recipient mice 1 d following the adoptive transfer of the naive 3A9 cells, when only trace numbers of the transferred naive cells are retained in the lung (Fig. 4A). Tracing the donor cells in the recipient mice in this system revealed that, unlike the cells activated in vitro and injected i.v. (Fig. 1A), the donor cells activated in vivo (Fig. 5A) did not exhibit early homing to the lung. Instead, small numbers of donor cells were found to accumulate, simultaneously, in the lung and spleen, when tested on day 1 after injection of the TLR ligand and HEL. These donor cells similarly increased rapidly in number in both organs, reaching their peaks on day 3 after cell injection and declining gradually thereafter. It is of interest that donor cells activated in vivo constituted a higher proportions among the CD4 population in the recipient spleen, as compared with the proportions reached with the in vitro–activated donor cells (∼15 versus ∼5% at the peak). Importantly, the total numbers of donor cells in the recipient spleen were ∼100-fold higher than those in the lung throughout the tested time periods (Fig. 5B).

FIGURE 5.

CD4 cells activated in vivo differ from those activated in vitro by their pattern of activities. Naive 3A9 CD4 cells (5 × 106 per recipient) were injected to HEL-Tg mice on day −1, followed on day 0 by injections of LPS, or CpG, along with HEL. (A) Percentage of donor cells among CD4 populations of spleen and lung of recipient mice at different time points. Parallel patterns of donor cell changes in number in recipients’ lungs and spleens, with increased values during days 2 and 3, followed by gradual decrease at the later time points are shown. (B) Total number of donor cells in the spleen and lung of recipient mice, determined at the indicated time points. Note the actual absence of donor cells in the lung on day 1. Combined results of two experiments with closely similar data are shown. In each experiment, 24 or 30 mice were used and 3 or 4 mice were sacrificed for each time point for each group. *p < 0.05, **p < 0.005, ***p < 0.001.

FIGURE 5.

CD4 cells activated in vivo differ from those activated in vitro by their pattern of activities. Naive 3A9 CD4 cells (5 × 106 per recipient) were injected to HEL-Tg mice on day −1, followed on day 0 by injections of LPS, or CpG, along with HEL. (A) Percentage of donor cells among CD4 populations of spleen and lung of recipient mice at different time points. Parallel patterns of donor cell changes in number in recipients’ lungs and spleens, with increased values during days 2 and 3, followed by gradual decrease at the later time points are shown. (B) Total number of donor cells in the spleen and lung of recipient mice, determined at the indicated time points. Note the actual absence of donor cells in the lung on day 1. Combined results of two experiments with closely similar data are shown. In each experiment, 24 or 30 mice were used and 3 or 4 mice were sacrificed for each time point for each group. *p < 0.05, **p < 0.005, ***p < 0.001.

Close modal

To determine the activation status of donor and recipient cells at different stages of the pathogenic process, we analyzed by flow cytometry the expression of cell markers specific for activation or resting statuses. Fig. 6 records the mean data collected in two independent experiments with high similarity. CD25, CD44high, and CD69 are markers for activation, whereas CD62L and CCR7 are typically expressed by T cells at the resting state. There are three observations of interest. First, remarkable differences are seen between the activation profiles (i.e., expression levels of the five markers) of the cells following their activation in vitro and when collected from the recipient spleen on day 4 after cell injection. We consider these differences to be components of the “licensing” process in the spleen. Second, recipient cells are inferior to donor cells in their expression of the activation marker CD44high, but they express higher levels of CD62L and CCR7, the markers for the resting status. Third, the marker profiles of both donor and recipient cells collected from the recipient eyes exhibit higher profiles of activation than do those of the donor cells in the recipient spleens, that is, higher levels of CD25 and CD69 and lower levels of CD62L and CCR7.

FIGURE 6.

Flow cytometric analysis of expression levels of activation markers by donor and recipient CD4 cells. The tested cell preparations include: CD4 lymphocytes collected from spleens of naive 3A9 mice and include both 1G12+ and 1G12 cells; spleen cells of 3A9 mice following activation (act) in vitro with the Ag; recipient spleen cells on day 7 after cell injection; and recipient mouse eyes on day 7 after cell injection. The flow cytometric procedure is detailed in 2Materials and Methods.

FIGURE 6.

Flow cytometric analysis of expression levels of activation markers by donor and recipient CD4 cells. The tested cell preparations include: CD4 lymphocytes collected from spleens of naive 3A9 mice and include both 1G12+ and 1G12 cells; spleen cells of 3A9 mice following activation (act) in vitro with the Ag; recipient spleen cells on day 7 after cell injection; and recipient mouse eyes on day 7 after cell injection. The flow cytometric procedure is detailed in 2Materials and Methods.

Close modal

Odoardi et al. (13) interpreted their data to show that pathogenic donor cells found in the recipient spleen were initially “licensed” in the lung, before migrating to the spleen. To further scrutinize this assumed sequence of events, we adoptively transferred i.v. in vitro–activated 3A9 cells and sorted out the donor cells from the recipients’ lungs and spleens to monitor expression changes in the transcript levels of inflammation-related genes. The tested organs were collected from the recipient mice at different time points, and the transcript expression analysis was carried out by NanoString technology. Fig. 7 shows data collected with 9 selected genes known to participate in cell migration and tissue invasion (Fig. 7A), 12 genes involved in mediation of immunological responses (Fig. 7B), as well as 5 genes of molecules commonly used as activation markers (Fig. 7C). Importantly, a parallel is seen in general between the profiles of expression levels of the tested transcripts by donor cells collected from the recipients’ spleens and lungs at the indicated time points. The changes seen in gene expression by donor cells are assumed to be components of the process of “licensing” for pathogenicity (12), and our findings thus provide evidence to show that this process takes place simultaneously in the recipient’s lung and spleen, ruling out the notion that the donor cells enter the spleen following “licensing” in the lung, as put forward by Odoardi et al. (13).

FIGURE 7.

Similarities in gene expression profiles of inflammation-related molecules by donor cells in lung and spleen of recipient mice. CD4 cells of 3A9 mice activated in vitro with HEL were adoptively transferred i.v. into HEL-Tg mice (5 × 106 cells per mouse), and spleens and lungs of the recipient mice were collected at the indicated time points. Donor cells (1G12+) in these two organs were sorted out and their total RNA was isolated and analyzed by NanoString technology for the expression of the indicated 26 gene transcripts. The data are expressed as NanoString reading units. The figure summarizes data of two separate experiments, combined by NanoString technology after showing close similarities. In each experiment, 18 or 24 recipient mice were used, with tissues of 3 or 4 mice collected for each time point. Group (A) includes data of eight chemokine receptors and one chemokine, group (B) shows data of 12 additional inflammation-related genes, and group (C) records data of five activation markers. Note the correlation between the patterns of changes in expression levels of the tested genes in donor cells collected from the lungs or spleens of the recipient mice.

FIGURE 7.

Similarities in gene expression profiles of inflammation-related molecules by donor cells in lung and spleen of recipient mice. CD4 cells of 3A9 mice activated in vitro with HEL were adoptively transferred i.v. into HEL-Tg mice (5 × 106 cells per mouse), and spleens and lungs of the recipient mice were collected at the indicated time points. Donor cells (1G12+) in these two organs were sorted out and their total RNA was isolated and analyzed by NanoString technology for the expression of the indicated 26 gene transcripts. The data are expressed as NanoString reading units. The figure summarizes data of two separate experiments, combined by NanoString technology after showing close similarities. In each experiment, 18 or 24 recipient mice were used, with tissues of 3 or 4 mice collected for each time point. Group (A) includes data of eight chemokine receptors and one chemokine, group (B) shows data of 12 additional inflammation-related genes, and group (C) records data of five activation markers. Note the correlation between the patterns of changes in expression levels of the tested genes in donor cells collected from the lungs or spleens of the recipient mice.

Close modal

The data collected by NanoString technology, recorded in Fig. 7, also made it possible for us to identify specific patterns of changes in expression of individual gene transcripts of the 26 selected CD4 lymphocyte marker molecules during the process of “licensing” in vivo.

Fig. 7A demonstrates the transcript expression profiles of eight chemokine receptors and one chemokine, ccl5 (RANTES). Note that the transcript for ccl5 was the only chemokine transcript included in the NanoString probe set that exhibited expression changes due to the activation processes investigated in this study; transcripts of the other 24 tested chemokines monitored by the NanoString system were expressed at trace levels by the CD4 cells and did not change during the tested “licensing” process (data not shown).

Fig. 7B shows the expression profiles of 12 additional transcripts that include nine cytokine receptors, the adhesion molecule cd2, as well as Itga4 and Itgb1, the two components of the integrin α4β1 (VLA-4).

Fig. 7C records the expression profiles of genes of five molecules commonly used to indicate T cell activation, by exhibiting sharp increase (Il2ra/Cd24, Cd44, and Cd69) or decrease (Sell/Cd62l and Ccr7) in their expression. These changes in the transcript expression during the 4 d of activation are in line with the changes in expression of the corresponding proteins during the activation process (Fig. 6). Furthermore, a correlation is seen between the changes in profiles of the five activation markers in naive CD4 cells following the 4 d of activation in vitro as detected by flow cytometry (Fig. 6) and by the NanoString technique (Supplemental Fig. 2).

In line with their incapacity to proliferate and undergo “licensing” in recipient mice (Fig. 4), naive CD4 cells of 3A9 mice are nonpathogenic and do not induce inflammation in recipient eyes (17, 18). The NanoString system made it possible for us to identify changes in transcript expression of the 26 inflammation-related genes in naive CD4 cells following the 4 d of activation in vitro. As shown in Fig. 8, expression of the great majority of these genes underwent considerable changes of either increase or decrease.

FIGURE 8.

Activation of naive cells in vitro remarkably modifies the expression levels of inflammation-related genes. (AC) Naive CD4 cells were isolated from pooled spleens and lymph nodes of naive 3A9 mice and activated in vitro with HEL for 4 d (Ac CD4), as detailed in 2Materials and Methods. Total RNA from the naive and activated populations was isolated and analyzed by NanoString technology for expression of the 26 inflammation-related genes. The recorded data, presented as NanoString reading units, are values of three separate experiments, combined after showing close similarities. Note the remarkable changes in expression of most genes due to the activation process. Grouping of the genes is the same as that in Fig. 7.

FIGURE 8.

Activation of naive cells in vitro remarkably modifies the expression levels of inflammation-related genes. (AC) Naive CD4 cells were isolated from pooled spleens and lymph nodes of naive 3A9 mice and activated in vitro with HEL for 4 d (Ac CD4), as detailed in 2Materials and Methods. Total RNA from the naive and activated populations was isolated and analyzed by NanoString technology for expression of the 26 inflammation-related genes. The recorded data, presented as NanoString reading units, are values of three separate experiments, combined after showing close similarities. Note the remarkable changes in expression of most genes due to the activation process. Grouping of the genes is the same as that in Fig. 7.

Close modal

Next, we analyzed the patterns of changes in gene transcript expression by naive 3A9 cells during their acquisition of pathogenicity following activation in vivo. We collected the donor cells (1G12+) in the lung and spleen, at different time points following injection with LPS, and analyzed the expression levels of the 26 genes related to inflammation. Fig. 9 records the combined NanoString data collected in two separate experiments. A parallel is generally seen between the pattern of expression changes of the tested genes by donor cells collected from the lung and spleen of the recipient mice. Note, however, that only a partial correlation is seen between the profiles of gene expression by the donor cells activated in vivo (Fig. 9) and those activated in vitro (Fig. 7). These differences could be attributed to the disparities between the two biological systems.

FIGURE 9.

Patterns of changes in transcript expression of inflammation-related genes by Th cells activated in vivo. (AC) Donor cells were collected from lung and spleen of recipient mice at the indicated time points following the injection of LPS, and expression levels of the indicated genes were determined by the NanoString method. The procedures are detailed in 2Materials and Methods section. The recorded data are combined values of two separate experiments with similar results. The grouping of the genes is detailed in the legend for Fig. 7.

FIGURE 9.

Patterns of changes in transcript expression of inflammation-related genes by Th cells activated in vivo. (AC) Donor cells were collected from lung and spleen of recipient mice at the indicated time points following the injection of LPS, and expression levels of the indicated genes were determined by the NanoString method. The procedures are detailed in 2Materials and Methods section. The recorded data are combined values of two separate experiments with similar results. The grouping of the genes is detailed in the legend for Fig. 7.

Close modal

Data presented in the present study provide new information concerning the processes whereby autoantigen-specific in vitro–activated or naive CD4 cells acquire pathogenicity. The seminal study of Flügel et al. (12) established the requirement for a 2- to 3-d process during which pathogenicity is acquired by T cells, and a more recent publication by that group (13) named this process “licensing.” This latter cited paper focused on the transferred cells that accumulate and increase in number in the recipient lung and adjacent lymph nodes and concluded that the lung is the organ where “licensing” for pathogenicity takes place. Our data show, however, that the spleen, not the lung, is the major “licensing” organ. Furthermore, our data suggest that the conclusion of the cited study (13), that is, that the lung is the “licensing” organ, was biased by the authors transferring the donor cells only by i.v. injection; such cells enter the recipient animal mostly through the lung (Fig. 1) (13). However, only a small fraction of the transferred cells is retained in the lung, whereas the great majority of these cells quickly exit the lung and migrate to other organs of the recipient animal, in particular the spleen, where the “licensing” process takes place concurrently with that in the lung, as shown in the present study.

Our conclusion, that the spleen, not the lung, is the major organ where donor cells acquire pathogenicity, is based on the following observations. First, donor cells activated in vitro and injected i.v. exhibited paralleled kinetics of increase in numbers in the lung and spleen when monitored from day 1 after cell injection, similarly reaching peaks on days 3–4 postinjection, followed by gradual decline (Fig. 1) Second, further analysis of the changes in transcript expression of molecules involved in cell migration and immune response during the “licensing” period demonstrated similar and parallel patterns of changes in donor cells in the recipient’s lung and spleen (Fig. 7), ruling out the statement (13) that donor cells in the spleen were previously “licensed” in the lung. Third, donor cells activated in vitro and injected by the i.p. route, or those activated in vivo, accumulated concurrently in the lung and the spleen and then increased in number in the two organs by similar kinetics (Fig. 1 versus Fig. 5, respectively). Finally, and most importantly, the total number of donor cells in all tested experimental systems was larger by far in the spleen than that in the lung (Figs. 1, 3, 5). Moreover, note that the number of donor cells in the lung was ∼100-fold lower than that in the spleen in mice injected with donor cells by the i.p. route, or those in which the donor cells were activated in vivo, whereas in mice injected by the i.v. route the ratio was only ∼10-fold lower. The relatively large number of donor cells in the lungs of the i.v. injected mice could be attributed to the large number of cells retained in their lung.

In addition to the spleen and lung, homing patterns of transferred donor cells were examined in the liver and parathymic lymph nodes following injection by the i.v. or the i.p. routes, respectively. The liver is the organ where dying and dead cells are eliminated, but a small number of live donor cells homed to this organ and apparently underwent “licensing” there, as indicated by their typical increase in number, followed by a decline. Donor cells migrating to the parathymic lymph nodes, reported to be the earliest homing organ for i.p. injected cells (12), also underwent “licensing” in these lymph nodes. It is also of interest that the total numbers of donor cells in the parathymic lymph nodes were ∼10-fold larger than those in the lung of these i.p. injected mice, indicating, again, that the lung plays a minor role in the “licensing” process.

It is of note that the system in which naive cells are activated in vivo to become pathogenic partly imitates the physiological process of immunization, whereas the system in which the cells are activated in vitro is nonphysiological. Our data with the in vivo activation system showing simultaneous migration of the activated cells to the lung and spleen, with a much smaller number of donor cells migrating to the lung than to the spleen (100-fold) in these mice, thus further underscore the notion that the lung plays a minor role in the physiological “licensing” process.

We assume that the increase in number of donor cells in the recipient organs tested in this study could be attributed to their local proliferation. The elevation in the number of donor cells in the blood, however, could be attributed to an increase in the number of donor cells exiting these mentioned organs and entering the circulating blood after their “licensing” process is completed. The circulating “licensed” cells are assumed to be seeking their target Ag; cells that fail to reach their target are assumed to be eliminated, mostly in the liver.

We analyzed the activation status of donor and recipient Th cells by flow cytometry, staining the cells for markers specific for activated cells (CD25, CD44high, and CD69) and for resting cells (CD62L and CCR7) (Fig. 6). Of particular interest is the difference in profiles of surface molecule expression between cells activated in vitro, prior to adoptive transfer, and the cells collected from the recipient spleens on day 4 after cell injection. It is assumed that these differences represent major components of the “licensing” process, which endows the donor cells with the capacity to invade the target organ; activation in vitro is essential for the transferred lymphocytes to initiate the pathogenic process (Fig. 4), but not to enter the target organ (12, 22). The changes in expression level profiles between the donor cells prior to injection and on day 4 in the spleen (following “licensing”) are particularly apparent for CD25, CD69, and CCR7. Also of note is the marker profile of donor cells from the recipient eyes, showing elevated levels of activation, as compared with their profile while in the recipient spleen (Fig. 6). This observation is attributed to the process of reactivation of the donor cells when re-exposed in the recipient eyes to their target Ag, HEL.

It is also of interest that the recipient mouse CD4 cells, too, underwent changes in their expression of the tested markers (Fig. 6), despite their noninvolvement in the initiation of the inflammatory process. These changes in the recipient cells can be attributed to the cytokine environments in the recipient spleen, mediated by the ongoing “licensing” process. This notion is supported by the finding that the recipient cells collected from the inflamed mouse eyes exhibited higher levels of activation than did those in the spleen (Fig. 6), a finding attributed to the intensive cytokine stimulatory environment in the recipient mouse eyes following the reactivation of the donor cells by HEL, mentioned above.

Changes in expression of the five activation markers by donor cells at different time points were also analyzed by NanoString microarray technology (Supplemental Fig. 2). A correlation is seen in general between the two methods in showing changes in the gene expression of these five marker molecules following activation in vitro, and when the donor cells were analyzed in the spleen after the “licensing” process (namely spleen on day 4 after cell transfer).

It is noteworthy that Odoardi et al. (13) also focused on the changes in gene expression during the “licensing” process, collecting the information by Affymetrix technology. Significantly, these authors emphasized the increase in expression of certain genes that were also found to increase in expression in our study, which used NanoString technology. These genes include chemokine receptors Ccr2, Ccr5, Cxcr3, Cxcr4, chemokine Ccl5, and integrin Itgb1. This similarity in patterns of changes in gene expression in the two studies thus supports the assumption that the “licensing” processes in the two studies are basically the same, despite the difference in the tested animal species, rats and mice, respectively. It is also of note that information concerning the products of these genes would be useful for identification of new targets for treatment of pathogenic autoimmune processes. Indeed, Ab against VLA-4 (α4β1) is a powerful inhibitor of experimental autoimmune encephalomyelitis development (23, 24), and the transcript expressions of the α4 and β1 (Itga4 and Itgb1) components of VLA-4 were found in our study to increase during the “licensing” process (Fig. 7B).

Comparing naive and in vitro–activated CD4 cells revealed remarkable differences in their expression levels of the 26 genes tested in this study (Fig. 8). We assume that these differences in gene expression are responsible at least in part for the remarkable differences between these two CD4 populations in their homing and proliferation when adoptively transferred to naive recipients (Fig. 1 versus Fig. 4), as well as the observation that only the latter cell population acquires pathogenicity. To our knowledge, the present study (Fig. 8) provides for the first time data concerning the differences between naive and activated Th cell populations in their expression levels of genes assumed to be responsible for the capacity of the in vitro–activated population to initiate inflammation in the recipient animal. It is also notable that our study thus further defines the two separate processes of changes in gene expression responsible for the activation and the “licensing” for pathogenicity. Our observations also suggest that naive cells activated in vivo (Fig. 9) undergo these two separate and successive changes in gene expression to acquire pathogenicity.

In summary, the current study sheds new light on processes whereby lymphocytes specific against an autoantigen acquire pathogenic capacity (“licensing”) prior to initiating pathogenic responses in the tissues where the target Ags are expressed. Importantly, our observations indicate that the spleen, not the lung, is the major organ where the “licensing” for pathogenicity is taking place.

We thank Philip Murphy (National Institute of Allergy and Infectious Diseases) and Hong-wei Sun (National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health) for invaluable advice, and the National Eye Institute Cores of Flow Cytometry and of Histology for superb technical support.

This work was supported by the Intramural Program of the National Eye Institute, National Institutes of Health.

The online version of this article contains supplemental material.

Abbreviations used in this article:

HEL

hen egg lysozyme

HEL-Tg

HEL-transgenic

ODN

oligodeoxynucleotide.

1
Mills
K. H.
2011
.
TLR-dependent T cell activation in autoimmunity.
Nat. Rev. Immunol.
11
:
807
822
.
2
Joller
N.
,
Peters
A.
,
Anderson
A. C.
,
Kuchroo
V. K.
.
2012
.
Immune checkpoints in central nervous system autoimmunity.
Immunol. Rev.
248
:
122
139
.
3
O’Garra
A.
,
Steinman
L.
,
Gijbels
K.
.
1997
.
CD4+ T-cell subsets in autoimmunity.
Curr. Opin. Immunol.
9
:
872
883
.
4
Paterson
P. Y.
1960
.
Transfer of allergic encephalomyelitis in rats by means of lymph node cells.
J. Exp. Med.
111
:
119
136
.
5
Mochizuki
M.
,
Kuwabara
T.
,
McAllister
C.
,
Nussenblatt
R. B.
,
Gery
I.
.
1985
.
Adoptive transfer of experimental autoimmune uveoretinitis in rats. Immunopathogenic mechanisms and histologic features.
Invest. Ophthalmol. Vis. Sci.
26
:
1
9
.
6
Rangachari
M.
,
Kuchroo
V. K.
.
2013
.
Using EAE to better understand principles of immune function and autoimmune pathology.
J. Autoimmun.
45
:
31
39
.
7
Panitch
H. S.
1980
.
Adoptive transfer of experimental allergic encephalomyelitis with activated spleen cells: comparison of in vitro activation by concanavalin a and myelin basic protein.
Cell. Immunol.
56
:
163
171
.
8
Holda
J. H.
,
Welch
A. M.
,
Swanborg
R. H.
.
1980
.
Autoimmune effector cells. I. Transfer of experimental encephalomyelitis with lymphoid cells cultured with antigen.
Eur. J. Immunol.
10
:
657
659
.
9
Ben-Nun
A.
,
Wekerle
H.
,
Cohen
I. R.
.
1981
.
The rapid isolation of clonable antigen-specific T lymphocyte lines capable of mediating autoimmune encephalomyelitis.
Eur. J. Immunol.
11
:
195
199
.
10
Hickey
W. F.
,
Hsu
B. L.
,
Kimura
H.
.
1991
.
T-lymphocyte entry into the central nervous system.
J. Neurosci. Res.
28
:
254
260
.
11
Prendergast
R. A.
,
Iliff
C. E.
,
Coskuncan
N. M.
,
Caspi
R. R.
,
Sartani
G.
,
Tarrant
T. K.
,
Lutty
G. A.
,
McLeod
D. S.
.
1998
.
T cell traffic and the inflammatory response in experimental autoimmune uveoretinitis.
Invest. Ophthalmol. Vis. Sci.
39
:
754
762
.
12
Flügel
A.
,
Berkowicz
T.
,
Ritter
T.
,
Labeur
M.
,
Jenne
D. E.
,
Li
Z.
,
Ellwart
J. W.
,
Willem
M.
,
Lassmann
H.
,
Wekerle
H.
.
2001
.
Migratory activity and functional changes of green fluorescent effector cells before and during experimental autoimmune encephalomyelitis.
Immunity
14
:
547
560
.
13
Odoardi
F.
,
Sie
C.
,
Streyl
K.
,
Ulaganathan
V. K.
,
Schläger
C.
,
Lodygin
D.
,
Heckelsmiller
K.
,
Nietfeld
W.
,
Ellwart
J.
,
Klinkert
W. E.
, et al
.
2012
.
T cells become licensed in the lung to enter the central nervous system.
Nature
488
:
675
679
.
14
Lai
J. C.
,
Lobanoff
M. C.
,
Fukushima
A.
,
Wawrousek
E. F.
,
Chan
C. C.
,
Whitcup
S. M.
,
Gery
I.
.
1999
.
Uveitis induced by lymphocytes sensitized against a transgenically expressed lens protein.
Invest. Ophthalmol. Vis. Sci.
40
:
2735
2739
.
15
Kim
S. J.
,
Zhang
M.
,
Vistica
B. P.
,
Chan
C. C.
,
Shen
D. F.
,
Wawrousek
E. F.
,
Gery
I.
.
2002
.
Induction of ocular inflammation by T-helper lymphocytes type 2.
Invest. Ophthalmol. Vis. Sci.
43
:
758
765
.
16
Cox
C. A.
,
Shi
G.
,
Yin
H.
,
Vistica
B. P.
,
Wawrousek
E. F.
,
Chan
C. C.
,
Gery
I.
.
2008
.
Both Th1 and Th17 are immunopathogenic but differ in other key biological activities.
J. Immunol.
180
:
7414
7422
.
17
Fujimoto
C.
,
Yu
C. R.
,
Shi
G.
,
Vistica
B. P.
,
Wawrousek
E. F.
,
Klinman
D. M.
,
Chan
C. C.
,
Egwuagu
C. E.
,
Gery
I.
.
2006
.
Pertussis toxin is superior to TLR ligands in enhancing pathogenic autoimmunity, targeted at a neo-self antigen, by triggering robust expansion of Th1 cells and their cytokine production.
J. Immunol.
177
:
6896
6903
.
18
Shi
G.
,
Vistica
B. P.
,
Nugent
L. F.
,
Tan
C.
,
Wawrousek
E. F.
,
Klinman
D. M.
,
Gery
I.
.
2013
.
Differential involvement of Th1 and Th17 in pathogenic autoimmune processes triggered by different TLR ligands.
J. Immunol.
191
:
415
423
.
19
Tan
C.
,
Wei
L.
,
Vistica
B. P.
,
Shi
G.
,
Wawrousek
E. F.
,
Gery
I.
.
2014
.
Phenotypes of Th lineages generated by the commonly used activation with anti-CD3/CD28 antibodies differ from those generated by the physiological activation with the specific antigen.
Cell. Mol. Immunol.
11
:
305
313
.
20
Geiss
G. K.
,
Bumgarner
R. E.
,
Birditt
B.
,
Dahl
T.
,
Dowidar
N.
,
Dunaway
D. L.
,
Fell
H. P.
,
Ferree
S.
,
George
R. D.
,
Grogan
T.
, et al
.
2008
.
Direct multiplexed measurement of gene expression with color-coded probe pairs.
Nat. Biotechnol.
26
:
317
325
.
21
Zhang
M.
,
Vacchio
M. S.
,
Vistica
B. P.
,
Lesage
S.
,
Egwuagu
C. E.
,
Yu
C. R.
,
Gelderman
M. P.
,
Kennedy
M. C.
,
Wawrousek
E. F.
,
Gery
I.
.
2003
.
T cell tolerance to a neo-self antigen expressed by thymic epithelial cells: the soluble form is more effective than the membrane-bound form.
J. Immunol.
170
:
3954
3962
.
22
Shi
G.
,
Ramaswamy
M.
,
Vistica
B. P.
,
Cox
C. A.
,
Tan
C.
,
Wawrousek
E. F.
,
Siegel
R. M.
,
Gery
I.
.
2009
.
Unlike Th1, Th17 cells mediate sustained autoimmune inflammation and are highly resistant to restimulation-induced cell death.
J. Immunol.
183
:
7547
7556
.
23
Yednock
T. A.
,
Cannon
C.
,
Fritz
L. C.
,
Sanchez-Madrid
F.
,
Steinman
L.
,
Karin
N.
.
1992
.
Prevention of experimental autoimmune encephalomyelitis by antibodies against alpha 4 beta 1 integrin.
Nature
356
:
63
66
.
24
Steinman
L.
2009
.
A molecular trio in relapse and remission in multiple sclerosis.
Nat. Rev. Immunol.
9
:
440
447
.

The authors have no financial conflicts of interest.

Supplementary data