Abstract
Background/Aim: Dendritic cells (DCs) are important immune mediators following allogeneic hematopoietic stem cell transplantation (HSCT). We screened for DC frequency in the cornea and oral mucous membranes after HSCT by confocal laser scanning microscopy (CLSM) in a canine model. Materials and Methods: In vivo CLSM images of the epithelia were taken the day before and on days 28, 56 and 112 following HSCT. Peripheral blood counts and chimerism were determined. Results: An increase of DCs after HSCT was detected in each animal in both investigated tissue types. Highest DC numbers in the flew and the gingiva were detected on day 28 and in the corneal epithelium on day 56 after HSCT, respectively. Conclusion: Changes of DCs in ocular and non-ocular mucous membranes can be monitored by CLSM in vivo. The DC frequency in the cornea and mucosa changes following HSCT. DC recovery is rapid and their numbers correlate with the development of chimerism of peripheral blood mononuclear cells.
Dendritic cells (DCs) play a major role in the transition from the innate to the adaptive immune system. These cells are commonly located in a variety of tissues. In their role as the main antigen-presenting cells (APCs), DCs are highly capable of capturing and processing antigens. In general, mature DCs migrate to lymphatic organs in order to connect and activate antigen-specific T-cells (5). These cells infiltrate tissues in response to foreign antigens after immunological challenges such as infection or transplantation. In allogeneic hematopoietic stem cell transplantation (HSCT), DCs play a key role in graft versus leukemia and in graft versus host interactions. Following HSCT, DC repopulation appears rapidly by donor cells and occurs simultaneously with the blood myeloid engraftment (2). Although the DC recovery appears immediate [reviewed in (35)], graft versus host disease and use of immunosuppressive drugs contribute to qualitative and quantitative disturbances in DC homeostasis. Collin et al. observed the remaining epithelial DCs after HSCT relating to different conditioning regimens. While the nadir was similar in both regimens examined at 14 to 21 days after HSCT, DCs under the myeloablative regimen depleted more rapidly than that with non-myeloablative effect (14).
The eye was considered to have an entire absence of APCs until 2002, when Brissette-Storkus et al. (8) and Hamrah et al. (19) discovered the presence of stromal cluster of differentiation (CD) 11b-positive monocytes, macrophages and epithelial CD11c-positive, major histocompatibility complex (MHC) class II-negative DCs in murine corneas. Several researchers have confirmed that significant heterogenous populations of DCs appear in normal cornea, although their specific location and distribution remains a controversial issue (16, 19, 24, 26, 32). Recently, Hattori et al. showed that DCs with CD11c and the C-type lectin Langerin, the so-called Langerhans cells, are mainly found in the corneal epithelium, whereas the population in the corneal stroma featured mostly CD11c-negative langerin-positive DCs (20).
Confocal microscopy offers the possibility to analyze cell morphology without interference by dispersible irradiance from outside of the focus position (34). The first publication of in vivo confocal images of the full-thickness human cornea was represented by Lemp and colleagues in 1985 (27). In the past 25 years, this technique was re-defined for in vivo clinical and research applications, particularly in the front eye region (30, 33, 37, 45). The Rostock cornea module modification of the Heidelberg Retina Tomograph assists the refinement of confocal microscopy for analysis of corneal cellular structures and functions feasible at ocular surface in a non-invasive manner (41). Further well-testable tissue besides the cornea is the oral mucosa, which also offers an easy and non-invasive possibility for examination. These mucous membranes are detectable up to depths of 200 μm by in vivo confocal laser scanning microscopy (CLSM) (9).
Concerning DCs, Zhivov and colleagues (2005) characterized the density and distribution of such APCs in the corneal epithelia in 35 out of 112 healthy volunteers. By CLSM, DCs are distinguishable as interdigitating hyper-reflective cells between epithelial cells with or without typical irregular dendrites, indicating the state of maturation. These cells were found in both the peripheral and the central corneal epithelium by in vivo confocal microscopy (48). Two other studies evaluated and compared corneal DC density and distribution in healthy volunteers and contact lens wearers, finding higher DC counts in lens wearers (40, 47). Contact lenses as foreign bodies, in combination with the chronic mechanical irritation of the cornea, were suggested to lead to a higher DC incidence (47). Lin et al. (28) and Mastropasqua et al. (31) reported on the eye pathophysiology in terms of DC changes in the central corneal epithelium. In both studies, the mean DC densities in inflamed eyes were significantly higher, than in healthy volunteers. The presence of DCs in human healthy central corneas was only detectable with minimal numbers, but larger cell counts occurred in response to immunological challenges (28, 31, 40, 47). Regarding other epithelial cell types such as mucous membranes, changes in DC densities were also observed in pathological and inflammatory processes (9, 13).
Further characterization and monitoring of DCs in immunological changes such as HSCT could offer new insights concerning the development of graft versus host interactions.
Materials and Methods
Animal model. A group of five random-bred beagle dogs, purchased from a commercial animal research provider holding a license for animal breeding and husbandry according to paragraph 11 of the German Animal Protection Law (Harlan-laboratories, Horst, the Netherlands), was selected by means of matches for highly polymorphic MHC class I and class II microsatellite markers (10, 44). The median age of the dogs was 2.6 years (range=2.1-2.8 years) at HSCT. The animals were housed in an accredited facility in standard indoor and outdoor runs, and were provided with commercial dog chow and tap water ad libitum. All experimental procedures and examinations were performed under general anesthesia. The performed anesthetic protocols were approved by the Review Board of the State Institute for Agriculture, Food Safety and Fishery Mecklenburg-Vorpommern, Germany (LALLF M-V/TSD/722.13-1.1-030/07).
Hematopoietic stem cell transplantation. All recipients were treated with 2 Gy total body irradiation one day before HSCT. All animals received intra-bone marrow HSCT as described in Knueppel et al. (25). In brief, grafts of the respective donors were aspirated from front legs, iliac crests and hind legs. HSCs were enriched via buffy coat and ficoll density-gradient centrifugation. This graft was transplanted intraosseously by injection of equal parts (total volume of 10 ml) into the left humerus and femur. CD34+ cells were transplanted at a range of 1.1-6.4×106 per kg body weight. A combination of cyclosporin A (15 mg/kg bid per os; days −1 to 35; Sandimmun® Optoral, Novartis, Nürnberg, Germany) and mycophenolate mofetil (20 mg/kg bid per os; days 0 to 27; CellCept®, Roche, Grenzach-Wyhlen, Germany) was used for immunosuppression post-graft.
Analysis of hematopoietic chimerism. Genomic DNA was isolated from peripheral blood mononuclear cells (PBMCs) and chimerism investigated by polymerase chain reaction of polymorphic nucleotide repeats followed by capillary gel electrophoresis (21). The samples were taken once before and then bi-weekly after transplantation over a time-span of 16 weeks. The definition of engraftment depicted a detection of >5% of donor-derived DNA in the PBMCs. Graft rejection was defined as a lack of donor-derived DNA in two subsequent analyses of chimerism following engraftment.
CLSM of canine cornea and oral mucous membranes. The combination of Heidelberg Retina Tomograph (HRT II, Heidelberg Engineering GmbH; Heidelberg, Germany) with the Rostock Cornea Module was used for the in vivo imaging (17, 41). This module relocates the original focal plane of imaging from the retina to the cornea and increases the magnification by the application of a water immersion objective with long working distance and high numerical aperture (Achroplan ×63/0.95, W/AA, 1.45 mm; Carl Zeiss GmbH, Göttingen, Germany). The microscope is equipped with a diode laser (wavelength of 670 nm). The connection between the immersion objective, which was directed vertically on the specimen, and the investigative tissue was established by a polymethyl methacrylate cap (TomoCap®; Heidelberg Engineering GmbH, Heidelberg, Germany) combined with contact gel (Vidisic®, n=1.35; Dr. Mann Pharma, Berlin, Germany). The microscope head, attached to a customized stand, featured movability in the x-, y- and z-directions. While the external z-scan supplied the focusing inside the tissues, the final adaption of the objective to the respective tissue was conducted manually. The size of all images was 300 μm × 300 μm with a vertical optical resolution of 2 μm. The exact positioning of the investigative tissue for imaging was maintained by placement under the microscope perpendicular to the objective.
The applicability of CLSM in veterinary ophthalmology has already been evaluated (23, 38). Our study included in vivo CLSM images of both the central cornea (Figure 1) and oral mucous membranes (Figure 2) from each of five beagle dogs.
Image analysis. Each measurement time point, namely one day before and 28, 56 and 112 after HSCT, was recorded with at least two sequences, comprising 100 images in z direction and/or three one-image sections in x/z direction, particularly through the epithelial layers of the respective tissue. Some measurements on day −1 (dog 1, dog 2) and on day 28 (dog 3) were not available due to oral inflammation or excessive eye movement which made measurement impossible. The density and distribution analysis was performed manually by counting the DCs according to their location, size, shape and hyper-reflectivity. Further analyses were performed by calculating the cellular density via a software tool implemented in the Heidelberg Eye Explorer, already applied in several other studies (28, 31, 40, 47, 48). For the latter, adequate cells of each analyzed tissue were marked in a pre-determined image area. The data were separately expressed by mean of cell counts and by density±standard deviation (cells±SD per mm2). Both analytical methods, sequences and sections, were compared internally at the pre-determined different time-points and according to the respective images of the other examined animals. For descriptive statistics, the mean, grand mean and standard deviation were used.
In vivo confocal laser scanning microscopy images of canine cornea before allogeneic hematopoietic stem cell transplantation. Schematic view of the eye, with the epithelial section of the cornea demonstrating the area of analysis at 30 to 60 μm depth. The images (300×300 μm) show different layers inside the cornea: a: superficial cells with polygonal shape and projecting nuclei as top layer; b: intermediate and basal cells with smaller size, mostly unapparent nuclei, clearly restricted cell membranes, and one hyper-reflective dendritic cell (DC) indicated by a white arrow; c: Bowman's layer, as acellular membrane with homogenous structure and reflectivity; d: stroma as solute netting of keratocytes (scattered fibrocytes) with rod-shaped nuclei, permeated by dark collagen lamellae and acellular matrix; e: Descemet‘s membrane as turbid layer with uncellular pattern; f: the endothelium with bright polygonal-shaped cells as basal layer. The analysis of this study was focused on the epithelium (a-c). DCs were typically found in layer b. The bar in image b serves as an example scale for all displayed images.
In vivo confocal laser scanning microscopy images of the canine flew. Schematic view of the flew, one of the investigated oral mucous membranes, which is located opposite the outlined tooth with the gingiva. The images demonstrate the area of analysis at depths of 45 to 65 μm. The epithelial flew consists of three typical cell layers: a: Superficial cells with polygonal shape as top layer; b: intermediate cells with smaller size and clearly restricted cell membranes; c: basal cells with smaller shape and higher cell density, pervaded by nerve fibers. Dendritic cells are typically found in layer b. The bar in image b serves as example scale for all displayed images.
Results
Characteristics of epithelial DCs in different untreated tissues. Confocal images revealed typical features of each layer (38) concerning cell morphology, cell size and cell distribution in the epithelia of the cornea and non-ocular mucous membranes in all untreated animals. The imaged areas of the corneal epithelial sections were focused between 30 to 60 μm depth (Figure 1). DCs, clearly distinguishable from other epithelial cells by cell size, shape and hyper-reflectivity, were infrequently detected in intermediate and basal corneal cell layers. Images of the investigated mucous membranes were recorded at depths of 45 to 65 μm, with typical morphological characteristics (Figure 2).
Epithelial DC distribution pre- and post-HSCT. In untreated animals the DCs of epithelia in the central cornea and oral mucous membranes exhibited a scattered pattern, with only few or no immunological cells in analyzed the areas, same as above. After HSCT, DCs were detectable in every animal at each time point by both analytical methods in the cornea and oral mucous membranes. Changes of the epithelial DC distribution in the cornea revealed a rapid increase, a slight clustering and occasionally the appearance of hyper-reflective cells with long dendrites that gave rise to the term DCs (Figure 3, black arrows). Localization and increase of DCs in the epithelia of canine flew and gingiva after HSCT were also defined in the experimental animals, with cell counts using both analytical modes at different time points.
Epithelial DC distribution and density in the cornea. In the cornea, DCs were detected on day −1 in one (dog 4) and two (dog 4, dog 5) out of three dogs by the sequential and sectional mode, respectively (Table I). Sequential images revealed an increase of hyper-reflective cells in the epithelial cornea layers in four dogs, where peaks were found on day 28 (dog 5), 56 (dog 3) and 112 (dog 2; dog 4). In comparison to the DC appearance of nine DCs per sequence (dog 4) before HSCT, the DC counts were three-fold higher on day 28 (27 DCs per sequence) and ten-fold higher on day 112 (89 DCs per sequence) after HSCT (Figure 4a). Before HSCT, analysis of the sectional mode of two dogs out of three provided an appearance of epithelial DCs, with 43 DCs (dog 4) and 33 DCs (dog 5), respectively. The time-based measurements of DC distribution after HSCT revealed heterogeneous results, with a mean of 31 DCs/mm2 (dog 5) and 140 DCs/mm2 (dog 4) (Figure 4b). The summary of sectional images of epithelial layers in the central cornea demonstrated an increase in the DC density in two dogs (dog 3, dog 4) .
Changes of the corneal intermediate epithelium after allogeneic hematopoietic stem cell transplantation. Dendritic cells are represented as hyper-reflective spots with heterogeneous morphology in the images before (day −1) and at three different time points after HSCT (day 28, 56, 112; white arrows). Dendrites of DCs were detectable in the image on day 56 (black arrows). The images for this dog (dog 4) display the typical progression of DC distribution in all animals. The bar serves as an example scale for all displayed images.
Epithelial DC distribution and density in mucous membranes. Sequential images showed DC appearance at all measurements of non-ocular membranes with peaks on day 28 (dog 4), 56 (dog 2, dog 3) and 112 (dog 1, dog 5). Available measurements before HSCT allowed images with detectable DCs for three dogs (dog 3, dog 4, dog 5) (Table I). However, the DC counts after HSCT were four- to six-fold higher according to the DC rates on day −1 (range: 13-94 DCs per sequence) (Figure 4c). Sectional images of all animals identified DCs with peaks on day 28 (dog 5, dog 4), 56 (dog 3) and 112 (dog 1, dog 2). Images of the epithelial layers of three dogs (dog 3, dog 4, dog 5) before HSCT provided detectable DCs (range 15-56 DCs/mm2) just as demonstrated in the other mode. After HSCT, the DC counts for those three dogs depicted higher levels compared to the previous measurements, with a two-fold increase concerning the DC rates on day −1 (peak range: 36-101 DCs/mm2) (Figure 4d). In the gingival epithelia (sectional mode), images revealed an increase of DC counts in three out of five dogs during the 16 weeks of measurements with peaks on day 112 (range: 52-75 DCs/mm2; dog 1, dog 2 and dog 5) after HSCT (data not shown).
Dendritic cell densities of sequential and sectional mode pre- and post-allogeneic hematopoietic stem cell transplantation in epithelia of cornea and flew. Expression of data by cells±SD per mm2 per section and cells per sequence. --, measurement not performed (see text).
Coherence between chimerism and epithelial DC distribution after HSCT. Changes of DC distribution in vivo within the epithelia of all investigated tissues were observed simultaneously with the development of PBMC chimerism. Initial engraftments were seen in all examined animals, with highest donor chimerism rates on day 28 after HSCT in all five dogs (mean: 14.3%; range: 9.7-21.6%). The data of the averaged donor chimerisms of all dogs indicated a detectable decrease on day 56 below the defined graft rejection line. With a time lag compared to donor PBMC chimerism, the highest DC density in corneal epithelia was detected on day 56 (56±4 DCs/mm2). The cell numbers decreased at the next measurement (day 112, 42±3 DCs/mm2), but remained above the value before HSCT (Figure 5). In the epithelial layers of flew and gingiva, the grand mean DC counts showed a similar development as peaks were found in both on day 28 (62±6 DCs/mm2 and 182±9 DCs/mm2, respectively). The last CLSM measurement on day 112 provided cell numbers with a lower value than on day 28 [48±5 DCs/mm2 (flew) and 54±6 DCs/mm2 (gingiva)], but still higher than before HSCT (Figure 5).
Discussion
The main APCs in the epidermis and mucous membranes, including the ocular surface, are represented by DCs (42). These cells induce both immunity and tolerance as a result of immunological changes such as inflammation and transplantation (5). In general, pathological processes can be determined at-diagnosis during treatment and recovery by CLSM and display changes such as high-density infiltration of DCs (11, 46).
Herein, we studied the coherence of chimerism development and linked changes in DC density and distribution in different canine epithelia after HSCT. The appearance and temporary increase of DCs in all investigated tissues indicated consistency with transient engraftments in the transplanted animal group. For non-invasive in vivo visualization of cellular changes within tissues, CLSM is an excellent tool. We focused on the central epithelial layers of the cornea, flew and gingiva. Scanning through the whole mucous membranes, we did not detect any further changes of DC distribution. Nor were other cellular changes in these tissues detectable. According to the morphology of DCs, Mastropasqua et al. reported that in inflamed corneas, these immune cells presented different morphological features (31). The group imaged larger cell sizes and reflectivity, more evident development of dendrites, frequent clustering and possible association with co-existing globular cells in comparison to normal corneas (31). Lin et al. detected DCs in inflamed processes with cell bodies longer than 12 μm, whereas DCs in normal tissues exhibited a morphology with cell bodies even shorter than 12 μm without dendritic shape (28). These findings are comparable to ours, as infrequent epithelial DCs in the cornea exhibited phenotypically-different subtypes, no dendritic processes and smaller size in all dogs before irradiation and HSCT. With regard to the DC morphology found in the post-HSCT corneas, criteria of hyper-reflectivity, development of more distinct dendrites and a slight tendency to form clusters may reflect activation and therefore maturation of DCs in response to inflammatory stimuli.
Analysis of dendritic cell distribution in canine epithelial cornea (a, b) and flew (c, d) after allogeneic hematopoietic stem cell transplantation by two different methodologies. a: Analysis by sequential method providing images in the vertical direction with manual counts of hyper-reflective DCs. This method generates a whole scan through a fixed multilayered area as a volume stack. b: Analysis by sectional mode providing horizontal images in one layer with calculations of the DC densities as extensive counts via the Heidelberg Eye Explorer software. c: Analysis by sequential method providing images in the vertical direction with manual counts of hyper-reflective DCs. This method generates a whole scan through a fixed multilayered area as a volume stack. d: Analysis by sectional mode providing horizontal images in one layer with calculations of the DC densities via the Heidelberg Eye Explorer software. Data represent mean absolute cell counts±SD. Measurements on day −1 (dog 1 and dog 2) and on day 28 (dog 3) were not available for both analytic methods.
In our study, CLSM images were evaluated by two different analytical methods – the sequential mode, providing images in the vertical direction, and the sectional mode, which offered horizontal images in a defined layer. While the latter method was already established in the analysis of corneal surface layers (40, 47), a modification especially for the analysis of the intermeshed oral mucous membrane layers was required in terms of the sequential mode. Offering an insight through a whole tissue block, the sequential mode revealed potential cellular changes spreading in different layers. The results of both methods confirmed that in all dogs, DCs were detected via their hyper-reflectivity at each time point and in each epithelial cornea and oral mucous membrane after HSCT. The density of DCs in corneal epithelial layers in untreated humans (34±3 cells/mm2) (48) was similar here (25±3 cells/mm2). Equally, the appearance of DCs in oral mucosal sites in healthy human subjects gave similar cell counts (1). In our study, canine corneal DCs after HSCT were found at a range of 7.3-205.3 DCs/mm2 (grand mean=55.6±3.5 cells/mm2, with the highest peak on day 56). It is interesting to note that some research groups found an increase both of epithelial DC distribution after inflammatory processes and of DC density located in the central cornea compared to the periphery (28, 47).
Epithelial dendritic cell counts of all investigated tissues with donor peripheral blood mononuclear cells chimerism rates during the time curve before and after allogeneic hematopoietic stem cell transplantation. The DC counts in the epithelia of cornea, flew and gingiva were analyzed as combined values for each time point of measurement. The donor PBMC chimerism values are provided for each time point of measurement as average before and after HSCT (n=5). The grand mean centering (CGM) values are shown as bars and the chimerism values are illustrated as lines.
After allogeneic HSCT, successful immune recovery is essential for minimizing infections post-transplantation and relapse, without extending the potential for the graft versus host disease. It has been known for years that cells of innate immunity recover within weeks, while cells of adaptive immunity take months to years [reviewed in (7)] to recover. Aufferman-Gretzinger and colleagues demonstrated that the majority of migratory human epithelial DCs of non-ocular tissue were of donor origin and detectable as early as day 18 after HSCT (3). Regarding the cornea, mouse chimeras have shown that 75% of the corneal myeloid cell population recovered within eight weeks (12). This time lag in DC reconstitution beween the ocular and non-ocular mucous membranes after HSCT is confirmed by our data. The delay of hyper-reflective cells in the cornea after HSCT could be explained by the lack of blood and lymphatic vessels in the cornea (16). Monocytic cells can rapidly enter inflamed sites from the bloodstream and differentiate into both DCs and macrophages (4). Therefore an infiltration of APCs would last longer in the cornea than in well blood-supplied tissues such as the oral mucous membranes. Blood circulation rapidly enables the appearance of DCs in the investigated oral epithelia at the first measurement on day 28 after HSCT. Their capacity to reach and migrate to epithelial layers of the cornea contributed to this delayed turnover rate on day 56 post-transplantation.
In connection with the development of graft versus host disease as a severe complication after HSCT, it was shown that only host-derived APCs initiate the acute disease (15, 39). Collin et al. investigated the temporal implication of immune recovery of DCs post-transplantation. Immune cell recovery occurred promptly within 40 days in the absence of acute graft versus host disease, whereas there was a delay of more than 100 days in its presence (14). Therefore, monitoring and evaluation of epithelial DC changes after HSCT in different tissues could provide information of graft versus host disease outcome.
There are some limitations to our study. Hyper-reflectivity and cell morphology are not convincing markers for the definitive identification of DCs. The immature state of DCs, which is demonstrated by the lack of elongations, might also be ascribed to macrophages. However, the latter are solely restricted to the lower stromal part of the cornea (18, 19) and are present at lower densities (18). Regarding location, microstructural cellular changes post-HSCT and concordance with existing studies (22, 28, 40, 47) are consistent; we firmly believe that detections of immunological cells concerned DCs. A larger number of subjects would have allowed for detection of more differences with statistical significance. Due to the time-consuming process of adequate animal housing and the extensive medical care required pre- and post-transplantation, we conducted this study including five beagle dogs. In contrast to other transplantation models, we used the canine model according to the good transferability to humans in genetic fundamentals (29) and in characteristics of CD34 + bone marrow cells, the precursor cells of DCs (43).
Although the DC source was not identified, based on published data of several research groups, we believe that the replenishment after HSCT is of donor origin (3, 6, 36). Immune recovery is generally provided by the de novo production of immunological cells generated by the transferred progenitors and HSCs, as well as the mature cells transmitted with the graft (3). Auffermann-Gretzinger investigated the turnover rate of DCs after HSCT under two different regimens. The DC re-generation on day 14 at approximately 80% was of donor origin and continued to increase to more than 95% by day 56 under the non-myeloablative regimen. With myeloablative conditioning, even earlier time points of donor chimerism were detected. Despite deviations, the donor DC chimerism remained over 50% until day 14 (2).
In conclusion, we monitored and evaluated the changes in the distribution and density of DCs of ocular and oral mucous membranes before and after HSCT to our knowledge for the first time. CLSM is an excellent tool for non-invasive in vivo visualization of cellular changes within tissues in this context. Our data suggest a correlation between donor chimerism development and the increase of DC density in the investigated epithelia. Visualizing DC changes in this non-invasive manner could be a sensitive measure for detecting engraftment or rejection of the transplant. Furthermore, such changes might act as predictive marker for graft versus host or graft versus leukemia reactions.
Acknowledgements
The Authors are very grateful to the highly dedicated technicians of the shared animal facility. This work was kindly supported by the German Research Council – SFB/Transregio37 “Micro- and Nanosystems in Medicine” .
Footnotes
-
Conflicts of Interest
RG is a consultant to Heidelberg Engineering (Heidelberg, Germany). The other Authors of this manuscript have no conflicts of interest or commercial interest to declare.
-
Funding Sources
Christian Junghanss, Doreen Killian, Hugo Murua Ecobar, Oliver Stachs and Rudolf Guthoff received grant by the German Research Council – SFB/Transregio37 “Micro- and Nanosystems in Medicine”, subproject A2 (Laser-based Transfection of Hematopoietic Stem Cells).
- Received August 22, 2013.
- Revision received October 11, 2013.
- Accepted October 14, 2013.
- Copyright © 2013 The Author(s). Published by the International Institute of Anticancer Research.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.