Localization of macrophages and dendritic cells in human thoracic lymph nodes: An immunohistochemical study using surgically obtained specimens

Abstract

Both dendritic cells (DCs) and macrophages are bone marrow‐derived cells that perform antigen presentation. The distribution of DCs and CD68‐positive macrophages were immunohistochemically examined in 103 thoracic nodes obtained from 23 lung cancer patients (50–84 years old) without metastasis. Among three antibodies tested initially—CD209/DCsign, fascin, and CD83—DCsign was chosen as the DC marker. For comparison, 137 nodes from 12 patients with cancer metastasis were also examined histologically. In patients without metastasis, DCs were found as (1) clusters along the subcapsular sinus and in a border area between the medullary sinus and cortex (mean sectional area of multiple nodes at one site, 8.4%) and, (2) rosette‐like structures in the cortex (mean number in multiple nodes at one site, 20.5). Notably, DC clusters and rosettes contained no or few macrophages and were surrounded by smooth muscle actin (SMA)‐positive, endothelium‐like cells. The subcapsular linear cluster corresponded to 5%–85% (mean, 34.0%) of the nodal circumferential length and was shorter in older patients (p = 0.009). DC rosettes, solitary, or communicating with a cluster, were usually connected to a paracortical lymph sinus. Few differences were found between nodes with or without metastasis, but DC cluster sometimes contained abundant macrophages in cancer metastasis patients. The subcapsular DC cluster is not known in the rodent model, in which the subcapsular sinus is filled with macrophages. This quite different, even complementary, distribution suggests no, or less, cooperation between DCs and macrophages in humans.

Dendritic cell (DC) clusters and rosettes contained no or few macrophages and were surrounded by smooth muscle actin‐positive, endothelium‐like cells. Few differences between nodes with or without metastasis, but DC cluster sometimes contained abundant macrophages in cancer metastasis patients.

1. INTRODUCTION

It has long been established that interdigitating dendritic cells (DCs), or simply DCs, are the key professional antigen‐presenting cells orchestrating innate and adaptative immunity in infection, chronic inflammatory diseases, and cancer (reviewed by Kvedaraite & Ginhoux, 2022). DCs so defined do not include follicular DCs. Macrophages have also long been considered distinct antigen‐presenting cells, but their primary functions are maintenance of tissue homeostasis and phagocytic clearance of various native and foreign bodies. Both DCs and macrophages originate from common progenitor cells in the bone marrow under the influence of key growth factors although macrophages are likely to originate from the yolk sac and liver in embryos. Moreover, a DC marker “DCsign” (CD209) was reported to be expressed in common among nodal DCs and a subpopulation of macrophages (Angel et al., 2009; Granelli‐Piperno et al., 2005; Park et al., 2014). These DCsign‐positive macrophages are called “monocyte‐derived DCs” (moDCs; Cheong et al., 2010; Lutz et al., 2022; Marzaioli et al., 2022).

In the last decade, there has been an apparent difference in the major focus of research on nodal DCs and macrophages, with research on the former emphasizing cancer therapy and that on the latter emphasizing diagnosis. DC‐mediated anticancer immunity in lymph nodes is suppressed in cancer patients (Heeren et al., 2021; Morandi et al., 2014), whereas sialoadhesin (CD169)‐positive macrophages in lymph nodes likely activate cytotoxic T lymphocytes and natural killer cells (Kawasaki et al., 2013; Kumamoto et al., 2021; Rakaee et al., 2019). However, in lung cancer patients, the immunological status of regional thoracic node DCs is complicated by the fact that the patients are often smokers. Possibly because of chronic inflammation, smokers have larger and/or greater numbers of nodes, with these larger nodes containing greater numbers of secondary follicles (Jin et al., 2022). Even in non‐cancer patients, inflammation is a major trigger of DC recruitment and activation in lymph nodes (Angeli & Randolph, 2006; Johnson & Jackson, 2010; Nitschké et al., 2012).

To our knowledge, only a couple of studies have assessed DC localization in the entirety of human nodal histological sections (Angel et al., 2009; Engering et al., 2004). In one, Engering et al. (2004) considered DC localization in the outer zone of the T lymphocyte‐rich paracortex. In a second study, in addition to the paracortex, Angel et al. (2009) described a linear cluster of DCs along the nodal capsule and its trabeculae insertion into the cortex. Human lymph nodes are characterized by trabeculae or intermediate sinuses, which divide the cortex into multiple islands (Grigorova et al., 2010; Murakami & Taniguchi, 2004). Here, we sought to revisit the morphology and localization of both DCs and macrophages in human thoracic nodes, initially employing three antibodies against DCs in our evaluation.

2. MATERIALS AND METHODS

A total of 240 surgically removed thoracic nodes were examined histologically. The use of surgical specimens was approved by the Ethics Committee of Kagoshima University Graduate Scholl of Medicine and Dentistry (No. 210198). Almost one‐third of specimens overlapped with those used for our recent study of nodal fibrosis (Jin et al., 2022).

First, we chose 103 nodes surgically removed from 23 lung cancer patients without metastasis. Of these patients, 16 were male and 7 were female, and ranged in age from 50 to 84 years old at the time of surgery (mean, 68 years old); 12 of the 20 patients were smokers. The sites of these nodes were lower paratracheal (14 nodes), subcarinal (28 nodes), hilar (10 nodes), and intrapulmonary (interlobar and lobar; 51 nodes; Table 1). Second, among four commercially obtained antibodies, we chose the one that was most suitable for the identification of DCs using the present materials (see subsection 2.1). Third, for comparison, we histologically examined the remaining 137 nodes from 12 patients with lung cancer metastasis. Among them, 10 were male and 2 were female, and ranged in age from 50 to 74 years old at the time of surgery (mean, 70 years old). Ten of the 12 patients were smokers, and the sites of nodes were the same as those in the non‐metastasis group (Table 1). The number of nodes detected at each site was greater in the metastasis group than in the non‐metastasis group.

TABLE 1.

Site and numbers of lymph nodes examined.

	Non‐meta patients (23) a	Meta+ patients (12) b
Lower paratracheal	14 nodes (5 patients)	30 nodes (9 patients)
Subcarinal	28 (5)	38 (5)
Hilar	10 (5)	20 (5)
Intrapulmonary	51 (15)	49 (12)
Total	103 nodes	137 nodes

^a 23 patients without lung cancer metastasis (16 men and 7 women; 1–8 nodes were examined per patient) contained 12 smokers (11 men and 1 woman).

^b 12 patients with nodal metastasis of the lung cancer (10 men and 2 women; 2–18 nodes were examined per patient) contained 10 smokers (all, men).

The removed nodes were fixed in 10% (w/w) neutral formalin solution for 7 days, followed by a routine histological procedure for paraffin‐embedded samples. Five to six serial sections were prepared from each node, including one corresponding to the maximum cross‐sectional area of the node. Two sections were stained with hematoxylin and eosin and sliver staining, whereas the others were used for immunohistochemistry (see subsection 2.1). The silver staining procedure used, referred to as “Gitter (reticular network) staining,” is based on Lillie et al. (1980). To provide greater insight into the three‐dimensional morphology, we prepared several sets of five to six serial sections at 50‐μm intervals along the 200‐ to 500‐μm thickness of the paraffin block in some nodes.

2.1. Immunohistochemistry

The primary antibodies used, together with their dilution and antigen retrieval procedure, are shown in Table 2. We tested three antibodies targeting DC markers, including mouse monoclonal antibodies against DCsign (also known as CD209) and CD83, and rabbit polyclonal antibodies against fascin; in addition, we used an antibody against the pan‐macrophage marker CD68. To identify the lymph node cortex and vascular structures, we also used antibodies against CD79a (B lymphocytes) and CD3 (T lymphocytes); for identification of blood vessels, we used antibodies against factor VIII (endothelial cells), alpha‐smooth muscle actin (SMA; smooth muscle cells) and alpha elastin. After incubating with primary antibodies, sections were incubated for 30 min with horseradish peroxidase‐conjugated secondary antibodies (Histofine Simple Stain Max‐PO; Nichirei), diluted 1:1000, after which immunoreactive proteins were detected by incubating with diaminobenzidine for 3–5 min (Histofine Simple Stain DAB; Nichire). Each sample was counterstained with hematoxylin, and a negative control without the first antibody was included for all specimens.

TABLE 2.

Primary antibodies and their dilution and specific treatment.

Legend	Ig types	Sources	Final dilution	Antigen retrieval
DC‐SIGN	Mouse monoclonal	Santa Cruz sc‐65740	1:200	Dako PT Link pH high
Fascin	Rabbit polyclonal	Proteintech 14384‐1‐AP	1:200	Dako PT Link pH high
CD83	Mouse monoclonal	Santa Cruz sc‐55536	1:50	pH 9 ER2 range cooker 12 min
CD3	Mouse monoclonal	Dako IR621	1:400	ready‐to‐use, autoclave 121°C, 5 min
CD79a	Mouse monoclonal	Dako N1592	1:50	Trypsin
CD68	Mouse monoclonal	Dako N0814	1:200	Trypsin
Factor VIII	Rabbit polyclonal	Dako GA527	1:16000	Trypsin
α Elastin	Mouse monoclonal	Abcam ab9519	1:20	Trypsin
α Smooth muscle actin	Mouse monoclonal	Dako M0851	1:800	Trypsin

Because immunostaining is sometimes difficult with the deeply fixed materials used here (Jin et al., 2022), in the first part of this study, in which we examined 16 nodes from seven patients without metastasis, we evaluated three antibodies against DC markers (Figure 1). Fascin and CD83 are generally considered classical markers of DCs (Iking‐konert et al., 2002; Kawachi et al., 2002; Kurata et al., 2005), but there is limited information on double staining with these markers (Asano et al., 2012). DCsign (DC‐specific ICAM‐3‐grabbing nonintegrin), a C‐type lectin identified as an antigen‐uptake receptor on DCs (Horrevorts et al., 2018; Khoo et al., 2008; Manzo et al., 2012; Rahimi, 2021), has been used to identify DCs in paraffin‐embedded samples using immunohistochemical procedures (Chen et al., 2019; Hollenbach et al., 2014). Park et al. (2014) reported that endothelial cells of the medullary sinus, as well as those in a region termed the “paracortical sinus”, express DCsign/CD209.

FIGURE 1.

A comparison of reactivity among three antibodies for dendritic cells (DCs) in a lower paratracheal node from a patient without cancer metastasis. Seventy‐year‐old woman, non‐smoker. Distribution of DCs candidates (panels a–c) are compared with CD68‐positive macrophages (panel d). Panels a, e and i, DCsign; panels b, f, and j, fascin; panels c, g, and k, CD83; panels d, h, and l, CD68. Panels e–h displays the negative control, while panels i–k (tonsil) and panel l (liver) exhibits the positive control, The subcapsular sinus shows reactivity of DCsign and fascin strongly (panels a, b). However, no reactivity was seen for antibody of CD83 and CD68 (panels c, d). Squares in panels a and d are shown in the upper and lower inserts at the higher magnification (four times higher than panels), respectively: arrows indicate CD68‐positive, DCsign‐negative macrophages. All panels were prepared at the same magnification (scale bar in panel a, 0.1 mm).

In addition, to identify apoptotic cells, especially in anthoracotic macrophages, we performed TUNEL (TdT‐mediated dUTP‐biotin nick end‐labeling) method using “in situ apoptosis detection kit” (Takara). The treated sections were counterstained with hematoxylin, dehydrated in ethanol, and cleared in xylene.

2.2. Morphometric analysis of DC‐rich areas

Using DCsign‐immunostained sections corresponding to the maximum cross‐sectional area of the node, we measured (1) the entire area of nodes, (2) the proportional area of a DC cluster, and (3) the proportional length of a subcapsular linear cluster of DCs in the nodal circumference (Figure 2). After manual tracing of the node and lesions, scanned images (Adobe Photoshop) of each site obtained using a 1× objective lens were processed with ImageJ software (version 1.45; U.S. National Institutes of Health). All statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS; version 26.0). In addition, we counted the number of DC rosettes per section (subsection 3.1.3). Correlations of DC parameters with age, gender, and smoking index were analyzed using Spearman's test, where a coefficient r > 0.4 was considered statistically significant.

FIGURE 2.

Linear clusters of dendritic cells (DCs) along the subcapsular sinus in the entire view of the node at the ultra‐low magnification: five nodes from five patients without metastasis. Panel a (lower paratracheal node), a 70‐year‐old woman, non‐smoker; panel b (lower paratracheal node), a 63‐year‐old man, smoker; panel c (interlobar node), a 63‐year‐old man, non‐smoker; panel d (interlobar node), a 71‐year‐old woman, smoker; panel e (subcarinal node), a 61‐year‐old man, smoker. Green lines indicate subcapsular linear clusters of DCs: the proportional lengths in circumference of the node are 44.9% (panel a), 41.0% (panel b), 88.5% (panel c), 45.0% (panel d) and 49.1% (panel e). Panels f and g are higher magnification views of the lower right‐hand side angles of panels c and e, respectively. Tringles indicate clusters of anthracotic macrophages: nodes of smokers tended to carry greater numbers of anthracotic macrophages when compared between near‐aged persons (panel a vs. panel d), but even non‐smokers are likely to have them in the node (panel c). Sectional area of nodes varied considerably even between near‐aged persons (panels b, c, and e). Panels a–e (or panels f and g) were prepared at the same magnification (scale bars: 1 mm in panel a, 0.1 mm in panel f).

2.3. Photographs

Most histology photographs were taken with a Nikon Eclipse 80, whereas ultra‐low magnification photographs (objective lens less than 1×) were obtained using a high‐grade flat scanner with translucent illumination (GTX970; Epson).

3. RESULTS

3.1. Regional lymph nodes from patients without cancer metastasis

3.1.1. Selection of antibodies for identifying DCs

The superficial cortex as well as the subcapsular sinus showed strong reactivity of DCsign (CD209) and fascin (Figure 1a,b), whereas no or weak reactivity was detected for the antibody against CD83 (Figure 1c). Usually, fascin and CD83 did not react with DCsign‐positive cell clusters in the medullary sinus. Therefore, in the present study, we judged that, after long fixation, DCsign was the most suitable DC marker. We used the tonsil for the positive control, but no or few CD83‐positive cells were found in the tonsil.

DCsign was reported to be expressed in common among nodal DCs and a subpopulation of macrophages (see Section 1). A critically important point was an evaluation of DCsign‐ and CD68‐double positive cells. However, in the present condition of immunohistochemistry, a cell population of the double‐positive was few in number (Figure 1a vs. d). We considered that their influence was limited in our evaluation of a difference in overall distributions of DCs and macrophages. Conversely, the present immunohistochemistry was unlikely to favor for identification of moDCs (see Section 1).

3.1.2. Localization of DCsign‐positive DCs and CD68‐positive macrophages

To emphasize subcapsular clusters of DCs, we show lower‐magnification views in Figure 2. The proportion (%) of the length of the subcapsular linear cluster of DCs along the entire circumferential length of the node ranged from 5% to 85% (mean 34.0%; Table 3). In most cases, subcapsular linear clusters were solitary or fragmented, irrespective of whether the other clusters were large or small. Therefore, the subcapsular cluster usually did not cover the entire surface of the superficial cortex. The subcapsular cluster was shorter in older patients (p = 0.009).

TABLE 3.

Site‐dependent difference in sectional areas of the node and dendritic cells (DCs) cluster in patients without metastasis.

	Nodal area a (mean), mm2	DCs/nodal area b (mean), %	SC‐DCs/cercum c (mean), %	Rosette d (mean)
Lower paratrach	126.0	6.7	34.0	41.2
Subcarinal	108.3	12.0	46.0	26.2
Hilar	53.3	7.4	18.0	7.2
Intrapulmonary	70.5	7.0	35.3	16.1
Total	83.2	8.4	34.0	20.5

^a Total of the maximum sectional areas of 1–8 nodes at the site.

^b Proportion of the sectional area of the DCs cluster in the total sectional area of the nodes at the site.

^c Proportion of the total length of the subcapsular DCs cluster in the circumferential length of nodes at the site.

^d Numbers of DC rosettes in the maximum sectional areas of 1–8 nodes at the site.

Anthracotic macrophages with phagocytotic carbon particles—characteristics of the human thoracic node—tended to occupy a large area of the medullary sinus in smokers' nodes relative to those of similar‐age non‐smokers (Figure 2a vs. d). However, even nodes from non‐smokers were likely to show severe anthracosis (Figure 2c). Likewise, the maximum cross‐sectional area of a node varied considerably between similar‐age persons irrespective of whether they were smokers or non‐smokers. The total cross‐sectional areas of multiple nodes obtained from one site (16.5–233.5 mm²) were largest in the lower paratracheal site and smallest in the hilar site (mean, 83.2 mm²; Table 3). Irrespective of how much carbon particles were contained, anthoracotic macrophages were negative in TdT‐mediated dUTP‐biotin nick end‐labeling (TUNEL) method (figure, not shown).

Dendritic cell clusters were present not only in the subcapsular area but also in a border area between the cortex and medullary sinus (Figures 3 and 4). Likewise, a linear subcapsular cluster of DCs (Figures 3a,b and 4a,b) was not in the subcapsular sinus itself, but rather was located along the outer surface of the superficial cortex (or along the cortical aspect of the sinus). Notably, where the subcapsular DC cluster extended into the cortex along the trabecula, it became lined or surrounded by SMA‐positive, endothelium‐like cells (Figure 3d,h). The subcapsular cluster of DCs was continuous with another linear cluster along the trabecular and medullary cord that extended between the subcapsular area and medullary sinus (Figures 3b and 4c). DCs were not enriched throughout the entire paracortex, but instead were concentrated in process‐like thin medullary cords inserted into the paracortex (the so‐called paracortical sinus). DC clusters occupied 0.7%–25.2% of the nodal cross‐sectional area (mean areas of multiple nodes at one site, 8.4%; Table 3). On the immunostaining of elastin, the nodal capsule was clearly discriminated from either the subcapsular sinus or the DC cluster (figure, not shown).

FIGURE 3.

Distribution of dendritic cells (DCs) and macrophages in a lower paratracheal node from a 70‐year‐old woman without metastasis (non‐smoker). Panels a and e are adjacent sections. Panels a–c display reactivity of DCsign, while panels e–g exhibit that of CD68. Two squares in panel a (or e) are shown in panels b and c (or f and g), respectively. DCs provide clusters in the subcapsular area, a border area between the medullary sinus and cortex and along the trabecula. DCsign‐negative, pale‐colored area in panels b and c is occupied by macrophages (panels f and g). A few macrophage is seen in the DC‐rich area in panel d corresponds to an area shown in panels b and f, while panel h shows an area shown in panels c and g. Smooth muscle actin (SMA)‐positive, endothelial‐like cells are seen along the DC clusters (arrows in panels d and h). Panels a and e (or panels b–d and f–h) were prepared at the same magnification (scale bar: 1 mm in panel a, 0.1 mm in panel b).

FIGURE 4.

A complementary distribution of dendritic cells (DCs) and macrophages in interlobar node and lower paratracheal nodes from a 66‐year‐old man without metastasis (smoker). Panels a and e (or panels c and g) are adjacent sections. Panels a, b, e, and f, an interlobar node; panels c, d, g, and h, a lower paratracheal node. Panels a–d display reactivity of DCsign, while panels e–h exhibit that of CD68. A square in panel a, c, e, or g is shown in panel b, d, f, or h. An insert at the lower angle of panel b displays the elastin‐positive capsule using the near section. DCs provide clusters in the subcapsular area and the medullary cord as well as along the medullary sinus. DCsign‐negative areas in panels b and d are occupied by macrophages (panels f and h). Panels a, c, e, and g (or panels b, d, f, and h) were prepared at the same magnification (scale bar: 1 mm in panel a, 0.1 mm in panel b).

No or few subcapsular macrophages were detected in most specimens (Figures 3e and 4f). Instead, macrophage clusters were seen in medullary cords that were inserted into and divided the paracortical areas, and the medullary sinus. In addition, the superficial cortex sometimes contained large, round clusters of macrophages (i.e., sinus histiocytes). Notably, anthracotic macrophages were distant from the sinus endothelium, in contrast to DCs, which were localized along the endothelium. Overall, DCs and macrophages showed quite different, even complementary, distributions, especially in subcapsular and paracortical areas.

3.1.3. Mesh‐like DC rosettes

Paracortex or T‐lymphocyte areas (Figure 5a,b) contained 50–300‐μm diameter, DCsign‐positive, mesh‐like rosettes (Figure 5c) that were connected with DCsign‐positive lymph sinuses (the so‐called paracortical sinus; Figures 5e and 6). In contrast, factor VIII‐positive blood vessels were not attached to rosettes (Figures 5f and 6b,f). As was the case for large clusters of DCs in the trabeculae and cortex, rosettes were also surrounded by an SMA‐positive, endothelium‐like membrane or the lymph endothelium itself. DC rosettes were often solitary but sometimes progressed to large DC clusters in the medullary cord or a border area between the cortex and medullary sinus. Solitary rosettes were surrounded by SMA‐positive, endothelium‐like cells (Figure 6c,g) that were not sinus endothelial cells, as they stained negative with silver impregnation (Figure 6d,h). Rosettes likely contained both T and B lymphocytes. Immunostaining of elastin also demonstrated factor VIII‐positive vascular wall (figure, not shown). Notably, some of the rosettes were TUNEL‐positive (Figure 7) in contrast to the nearby cluster of DCs.

FIGURE 5.

Dendritic cell (DC) rosettes and lymphatic endothelial lining in a paracortical area facing the medullary sinus of a lower paratracheal node from a 66 years old man without metastasis (smoker). The same specimen as shown in Figure 4. Panels a–d (or panels e and f) are adjacent sections. The paracortex or T‐lymphocyte area (panels a and b) contains not only DCsign‐negative/factor VIII‐positive blood vessels (arrows in panels c and d) but DCsign‐positive/factor VIII‐negative lymphatic sinusoid endothelium (arrowhead in panels c–f). DCsign‐positive, mesh‐like rosettes (arrowheads in panel c) are also seen and one of them is connected to the sinusoidal endothelium. Panel g, a near section of panel e, is a higher magnification view of one of the rosettes. All panels were prepared at the same magnification (scale bars in panels a and g, 0.1 mm).

FIGURE 6.

Dendritic cell (DC) rosette and its connection to a lymphatic sinus in a lower paratracheal node from a 70 years old woman without metastasis (non‐smoker). A specimen same as in Figure 3. Panels a and e, DCsign; panels b and f, factor VIII; panels c and g, smooth muscle actin (SMA); panels d and h, silver staining. Panels a–c (or panels e–g) are adjacent sections. Panels a–d are 0.1 mm distant from panels e–h. Higher magnification views of two squares in panels c and g are shown in panels d and h, respectively. A rosette of DCs (arrowhead in panels a–d) is surrounded by SMA‐positive, endothelial‐like cells (arrows in panels c and g). The endothelial‐like cells are negative for silver staining (arrows in panels d and h). Panels a–c and e–g (or panels d and h) were prepared at the same magnification (scale bars in panels a and d: 0.1 mm).

FIGURE 7.

TdT‐mediated dUTP‐biotin nick end‐labeling (TUNEL)‐positive dendritic cell (DC) rosettes in a lower paratracheal node from a 70‐year‐old woman without metastasis (non‐smoker). A specimen same as in Figures 3 and 6. In the cortex of the node, panels a and b display a site distant from that shown in panels c and d. Panels a and c, DCsign; panels b and d, TUNEL method. In contrast to the DCsign‐positive sinus and cluster (panels a and c), two rosettes (arrowhead) are TUNEL‐positive (panels b and d). All panels were prepared at the same magnification (scale bar in panel a: 0.1 mm).

The number of rosettes per section ranged from 0 to 57 (mean, 20.5) and tended to be higher in the lower paratracheal site (Table 3). Rosettes were absent in seven nodes from two patients (both female non‐smokers); in one of the two patients, no rosettes were detected in any of the three nodes. The number of rosettes was greater in nodes with a larger area of DC clusters (p = 0.002). The number of rosettes per DC‐rich area was weakly correlated with smoking index (p = 0.032), with rosette number tending to increase in smokers.

3.2. Regional lymph nodes with cancer metastasis

As was the case in nodes without metastasis, the intranodal distributions of DCs and macrophages were usually quite different and even complementary. However, macrophages were sometimes abundant in DC clusters (30 of 137 nodes; 5 of 12 patients) with metastasis. Mesh‐like rosettes of DCs were also often seen in the 137 nodes with metastasis (0–85 per section; mean, 14.2; Table 4). Rosettes were absent in 25 nodes from four patients (two female non‐smokers, a male non‐smoker, and a male smoker), and in three of these four patients, all nodes examined appeared to carry no rosettes. The total cross‐sectional areas of multiple nodes obtained from one site (14.3–428.4 mm²) were largest in the subcarinal site and smallest in the hilar site (mean, 117.5 mm²; Table 4). Overall, no significant difference in nodal cross‐sectional area, DC cluster area or proportion in the node, rosette number, or length of the subcapsular linear cluster was found between nodes with and without metastasis. Likewise, none of these parameters significantly correlated with gender, age, or smoking index.

TABLE 4.

Site‐dependent difference in sectional areas of the node and dendritic cells (DCs) cluster in patients with metastasis.

	Nodal area a (mean), mm2	DCs/nodal area b (mean), %	SC‐DCs/cercum c (mean), %	Rosette d (mean)
Lower paratrach	115.9	7.1	25.6	17.4
Subcarinal	277.8	4.5	20.0	22.0
Hilar	59.5	12.6	44.0	22.6
Intrapulmonary	85.6	6.1	25.0	6.3
Total	117.5	7.1	27.3	14.2

^a Total of the maximum sectional areas of 2–8 nodes at the site.

^b Proportion of the sectional area of the DCs cluster in the total sectional area of the nodes at the site.

^c Proportion of the total length of the subcapsular DCs cluster in the circumferential length of nodes at the site.

^d Numbers of DC rosettes in the maximum sectional areas of 2–8 nodes at the site.

4. DISCUSSION

To our knowledge, the present study is the first systematic report of the distribution of human nodal DCs that includes the entirety of the tissue. The materials used in the present study did not originate from donation for education and research of anatomy but instead were obtained from lung cancer patients. However, differences in histological findings between specimens with and without cancer metastasis appeared to be limited to whether macrophages were present together with abundant DCs. Conversely, nodes from 23 patients without metastasis were most likely to exhibit “usual” morphologies of nodal DCs typical of healthy men and women over 50 years old. Surgically obtained specimens have two advantages: (1) they are more suitable for immunostaining than specimens from donation for anatomy; and (2) they are accompanied by detailed information on lifestyle, such as smoking. Smoking is the most likely lifestyle factor to impact thoracic node histology, but studies comparing smokers and non‐smokers in this context are limited. Using surgically obtained thoracic nodes and a morphometric approach, Jin et al. (2022) demonstrated that smoking does not reliably cause hyalinization (an onion peel‐like fibrosis), but does increase secondary follicles.

As emphasized in previous research (Angel et al., 2009; Engering et al., 2004), DCs are readily detected in the outer zone of the T lymphocyte‐rich paracortex. However, this description tends to create the false impression that DCs are intermingled with T lymphocytes in this region to facilitate antigen presentation. Rather than being “in” the T lymphocyte area, DCs are arranged along the subcapsular, medullary, and other lymphatic sinuses. In particular, these sinuses provide thin processes that are inserted deeply into T lymphocyte areas. We verified that all or most regions of the sinus endothelium are positive for DCsign. Park et al. (2014) classified process‐like sinuses in the human nodal cortex according to LYVE‐1 reactivity into (1) the “trabecular sinus”, lined by LYVE1‐negative endothelial cells; and (2) the “paracortical sinus”, with an LYVE1‐positive endothelial lining. This “cortex versus sinus” dichotomy is not applicable to human nodes, where DCs showed a perisinusoidal distribution, at least in thoracic nodes.

In contrast to rodent models, the subcapsular sinus of the human thoracic node contains no or few CD68‐positive macrophages (Jin et al., 2022; Murakami & Taniguchi, 2004). Instead, we found a linear cluster of DCs along the inner aspect of the subcapsular sinus endothelium (or the outer or superficial aspect of the cortex)—striking features of this study. Engering et al. (2004) considered subcapsular DCs to be immature cells that are “waiting” for antigens from afferent lymph vessels. According to the present morphometric analysis, the subcapsular cluster of DCs tended to be longer in younger specimens but showed no difference in length between specimens with or without cancer metastasis. Likewise, we found no correlation between cluster length and smoking index. One of the characteristics of human thoracic nodes is the presence of anthracotic macrophages filled with carbon particles; these macrophages do not accumulate in the sinus but instead localize to the cortex outside of the sinus endothelium (Jin et al., 2022; Taniguchi et al., 2003), a localization that might result from subcapsular clustering of DCs.

Another striking feature of the present study is the clear discrepancy between territories of DCs and macrophages, which were present in what could be called a “complementary” distribution. In general, DCs and macrophages express abundant common antigens, including MHC class II, CD11 (integrin alpha), CD80, and CD83 (reviewed by Rogers et al., 2014). Even DCsign (CD209) is a likely common antigen (moDCs; see Section 1). Therefore, rather than constituting a discrepancy per se, this distribution of DCs and macrophages appears to reflect a colocalization process that supports the related functions of these two cell types. However, according to Morandi et al. (2014), under “steady‐state” conditions, IL‐27 exerts immunosuppressive activity by crippling antigen processing by DCs, which express functional IL‐27 receptors. Because macrophages are the major source of IL‐27 in lymph nodes, the present study suggests that, nodes without metastasis, populated by “steady state” macrophages, might avoid colocalization by DCs through the action of suppressive molecules like IL‐27.

In the paracortex and medullary sinus, we observed a mesh‐like rosette of DCs connected to the DCsign‐positive endothelial lining of the lymphatic sinus. Although we did not perform a 3D reconstruction, observations of semiserial sections of 20–500‐μm‐thick paraffin blocks suggest that the peripheral end of a trabecular DC cluster likely gives rise to a rosette. Actually, fewer rosettes were seen in nodes with a lower proportion of DC clusters per section. Thus, rosettes might progress to evolve into large DC clusters along a sinus, although solitary rosettes lacked apparent communication with the cluster. Conversely, the rosette is also likely to be a degenerative morphology since some of the rosettes were TUNEL‐positive. Although the rosette number tending to increase in smokers, we did not have evidences of correlation between the TUNEL‐positive rosette and smoking index.

Park et al. (2014) reported the expression of DCsign in lymphatic endothelial cells in human nodes. Both the close topographical relationship and the common expression of DCsign suggest coordination of function between DCs and the sinus endothelium in the recruitment and transfer of antigens. Finally, silver staining suggested that most of the SMA‐positive endothelium‐like cells around the rosette and along the DC cluster are unlikely to be endothelial cells of the lymphatic sinus or blood vessels. Indeed, the anti‐SMA antibody used in the current study stains the blood vessel endothelium in human nodes (Jin et al., 2022). However, most of the SMA‐positive lining might be composed of myofibrillar cells or a derivative of the sinus endothelium. LYVE‐1 immunostaining could be useful for confirming the identity of these cells. Nevertheless, we did not deny some of the SMA‐positive lining really corresponding to the sinus endothelium. Overall, further studies using high‐dimensional techniques is necessary to reveal cooperation between the nodal DC and macrophage at the sinus endothelium.

4.1. Study limitations

The fact that CD68 was the only macrophage marker used is a considerable limitation of this study. CD169 immunostaining could discriminate macrophage subpopulations in DC clusters. Another limitation is that all sample materials were obtained from persons over 50 years old. Therefore, we are unable to discard the possibility that the clear discrepancy between clusters of macrophages and DCs might result from decreased antigen presentation with age.

AUTHOR CONTRIBUTIONS

Masaya Aoki: correct data, and draft manuscript; Zhe‐Wu Jin: correct and analysis data, manuscript editing, and funding acquisition; Kazuhiro Ueda: data collection, analysis data, and manuscript editing; Go Kamimura: data analysis and revise manuscript; Aya Takeda‐Harada: data analysis and revise manuscript; Gen Murakami: design study, data analysis and manuscript writing; Masami Sato: data analysis and revise manuscript.

Aoki, M. , Jin, Z.‐W. , Ueda, K. , Kamimura, G. , Takeda‐Harada, A. , Murakami, G. et al. (2023) Localization of macrophages and dendritic cells in human thoracic lymph nodes: An immunohistochemical study using surgically obtained specimens. Journal of Anatomy, 243, 504–516. Available from: 10.1111/joa.13870

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.