The skin, enveloping the entire body, serves as a protective barrier against environmental factors. Its architecture and biomechanics are dynamic entities [1]. Therapeutically, the skin not only facilitates manual therapies but also harbors acupuncture stimulation points, which are used for treating a variety of conditions including musculoskeletal pain, reproductive disorders, and gastrointestinal diseases [2-5]. Acupuncture involves the puncturing and manipulation of specific skin points with needles to initiate healing processes for these conditions. Yet, the precise anatomical nature of acupoints remains elusive. The mast cell (MC) theory, one of the prevailing hypotheses, emerged from histological findings that MC density is greater at acupoints than at non-acupoints [6]. This theory posits that the penetration of an acupuncture needle into the epidermis mechanically stimulates the MCs residing in the dermis. These cells harbor a plethora of biological factors—histamine, serotonin, bradykinin, and proteases—within their intracellular granules. Upon stimulation, MCs degranulate, discharging these factors, which then exert a cascade of effects on nearby nerves and blood vessels, potentially triggering reactions within the nervous system, internal organs, and the immune and endocrine systems [6,7].

In rat models, Evans blue (EB)–reactive skin spots exhibiting neurogenic inflammation-induced extravasation have been proposed as acupoints, with approximately 70% of these spots aligning anatomically with known acupuncture points [8]. These EB-reactive spots release neuropeptides, such as substance P and CGRP (calcitonin gene-related peptide), inducing vasodilation and allowing fluid to seep from the vessels into the surrounding tissue, thereby increasing the electrical conductance of these spots relative to unstained skin areas [8,9].

We hypothesized that EB-reactive skin spots are rich in MCs. During our investigations to substantiate this hypothesis, we identified a novel nodal structure within the skin of rats. This article details the anatomical, histological, and gene expression characteristics of this newly discovered anatomical structure.

1. Animal preparation

All animal experiments adhered to the National Institutes of Health (NIH) guidelines for animal care and use. The Wonkwang University Animal Experiment Ethics Committee approved the experimental protocols (WKU18-14; WKU19-36; WKU20-71; WKU21-83), which were also in line with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. The researchers were not blinded to the group allocations. Animals were supplied by Charles River Technology (Gyeonggi-do, Korea), and experiments were conducted at the Korean Medicine Skin Research Laboratory of Wonkwang University (Iksan, Korea). Male Sprague Dawley rats, aged 7 to 9 weeks and weighing 220 ± 20 g, were housed at 21℃ to 23℃ with a 12-hour light-dark cycle (lights on at 6 AM) and ad libitum access to food and water. Anesthesia was induced via intramuscular injection of an anesthetic cocktail containing 0.4 ml alfaxalone (20 mg/kg, Jurox, Australia) and xylazine 0.1 ml (5 mg/kg, Bayer, Germany) into the hind limbs. After anesthesia induction, abdominal hair removal was performed with a razor blade, preceded by a soap solution application to minimize skin abrasions. The skin surface was then examined under a microscope for any bleeding, which, if present, halted further experimental steps to prevent artifactual results.

2. Characterization of EB-reactive skin spots

EB-reactive skin spots were visualized as per the established protocol [8]. A 5% EB solution was prepared by dissolving 0.01 g of EB powder (Sigma Chemical, USA) in 0.2 ml of phosphate-buffered saline (PBS, Life Technologies Corporation, USA). Subsequently, 0.20 to 0.25 ml (1 ml/kg) of this solution was injected into a tail vein to visualize EB-reactive skin spots. After 30 minutes, the nodal structures stained with EB were evaluated (Fig. 1A and 1B). The diameter and distance from adjacent nodes were measured from a scaled photographic image (Fig. 1C) using ImageJ (National Institutes of Health, Bethesda, MD, USA). The height, defined as the length from the top to the dermis–hypodermis interface, and the perimeter were quantified from 3D images reconstructed via X-ray microcomputed tomography (micro-CT) of ten isolated nodes from five rats (Fig. 1D), following previously described methods [10]. The isolated skin nodes were immediately flash-frozen in nitrogen gas. On the scanning day, frozen tissues were mounted on a micro-CT scanner stage (SkyScan 1172; Bruker, Kontich, Belgium) for imaging at 40 kV and 200 μA, with a pixel size of 0.95 μm, a 180° rotation angle, a 0.3° rotation step, and an exposure time of 1176 ms. Each node’s total scan time was approximately 2 hours. Projection images spanning 600 to 2000 sections were reconstructed using CTAn software (version 1.16, Bruker, Kontich, Belgium). The Skyscan suite (DataViewer and CTVox; Bruker, Kontich, Belgium) facilitated the analysis of reconstructed CT images. A 5% Alcian blue (AB) solution was prepared by dissolving 0.01 g of AB powder (Sigma Chemical) in 0.2 ml of PBS. Then, 0.20 to 0.25 ml (1 ml/kg) of this solution was injected into the tail vein to visualize reactive skin spots (Fig. 2). Hemacolor staining, hematoxylin and eosin (H&E) staining, and toluidine blue (TB) staining (Fig. 3) followed previously reported protocols [11]. For histological MC observation with TB staining, seven rats were sacrificed (Fig. 3J and 3K). From each rat, two node samples and two non-nodal skin areas were excised. Of the 14 node samples, a central slice was harvested from five nodes (Fig. 3J). Frozen skin tissues were sectioned at 6 µm thickness for microscopic examination after H&E and TB staining.

Figure 1. Skin nodes in the abdominal area of the rat. (A) Skin nodes and spots stained after Evans blue (EB) injection through a tail vein in the ventral area of the rat; A stained non-nodal spot found posterior from the xiphoid process was selected as the best matching anatomical location to an acupoint conception vessel (CV) 14. Scale bar: 4 mm; (A1) magnified image of (A) in the left side below the rib cage. EB-stained skin nodes were marked with arrows. The black arrows indicate skin nodes that are located along the kidney (KI 17-KI21), stomach (ST19-ST23), and spleen (SP14-SP16) meridians. The white arrow indicates a skin node located outside the known meridians. Four EB-reactive spots (CV11-CV14) are situated along the CV meridian. Scale bar: 4 mm; (A2) magnified image of (A) in the left-side inguinal area. Eleven stained skin nodes are indicated with black arrows. Six of those are along the line close to the liver channel. The location of the other five stained skin nodes is random. Scale bar: 4 mm; (A3) magnified image of (A) in the right-side inguinal area. Nine stained skin nodes are indicated with black arrows. Three of those are along the line close to the liver meridian. The location of the other six stained skin nodes is random. Multiple stained spots were observed in the inguinal area. Scale bar: 4 mm; (B) Skin spots in the hairless extremity stained after Evans blue injection through a tail vein in the rat, NA = non-acupoint; SP4 = spleen channel 4; (C) Shape of unstained skin nodes in the rat. Two skin nodes in milk color are pointed with black arrows. Scale between the lines: 1 mm; (D) The micro-CT image of a sagittal section of a node. The upper arrow indicates the top of the skin node. The middle arrow indicates an interface between dermis and hypodermis. The lower arrow indicates an interface between hypodermis and muscle layer of the skin node. Scale bar: 1 mm.

Figure 2. Skin nodes in the inguinal area stained with Alcian blue in the rat. (A) skin nodes in the left-hand side of the inguinal area before Alcian blue treatment. Six nodes marked with black arrows are not stained. An interval between the scale lines: 1 mm. Scale bar: 4 mm; (B) Image taken after injecting Alcian blue solution through a tail vein. Note the tip of the six nodes stained with injected Alcian blue. The scale of the image is the same as (A). Scale bar: 4 mm; (B1) Amplified view of (B). Scale bar: 4 mm.

Figure 3. Skin nodes in the stained rat abdominal area. (A) Skin nodes stained with the epidermal application of Hemacolor staining dyes, with arrows indicating each node. The vertical dotted line follows the linea alba. Scale bar: 1 cm. (B) An enlarged view of the nodes. Arrows indicate the same nodes as in 3A. Scale bar: 1 mm. (C) Magnified view of a single node filled with cells stained with Hemacolor dyes. Ex vivo microscopic view. Scale bar: 1 mm. (D) H&E staining image of a skin node (×40) with red arrows indicating hair follicles and black arrows indicating hairs. Scale bar: 500 µm. (E) H&E staining image of a skin node (×200) with arrows indicating the stratum corneum (red), stratum granulosum (black), and stratum basale (white). Scale bar: 100 µm. (F) H&E staining image of a skin node (×200) with arrows indicating erythrocytes stained in red. Scale bar: 100 µm. (G) A skin node stained with epidermally applied toluidine blue 1% solution in the abdominal area. Ex vivo microscopic observation (×40). Scale bar: 1 mm. (H) An amplified image of two mast cells in an ex vivo sample showing stained granules in the cells (×400). Scale bar: 10 µm. (I) An amplified image of a mast cell showing stained granules in the cell (×1000). Scale bar: 10 µm. (J) Mast cells stained with Toluidine blue in purple in a sagittal slice of a skin node (×40). A frozen skin node sample was sliced at a thickness of 6 µm for microscopic observation. Each arrow indicates a stained mast cell. Scale bar: 250 µm. (K) Mast cells stained with Toluidine blue in purple in a sagittal slice of a non-nodal skin area (×40). A frozen non-nodal skin sample was sliced at a thickness of 6 µm for microscopic observation. Each arrow indicates a stained mast cell. Scale bar: 250 µm.

3. RNA-seq analysis

For RNA-seq, EB infusion was omitted as skin nodes were visually discernible. Skin node samples, non-nodal skin, and inguinal lymph node (LN) tissues were excised using scissors. The excised skin nodes excluded the muscle layer, comprising only the epidermis, dermis, and part of the hypodermis. After excision, non-nodal area tissues outside a node’s perimeter were trimmed. Non-nodal skin samples were similarly excised from adjacent areas devoid of nodes.

Skin node samples were designated G1, non-nodal skin as G2, and LNs as G3. Tissue samples from these three categories (G1, G2, G3) were harvested from each of the 10 anesthetized rats and allocated into 30 labeled tubes (10 G1, 10 G2, 10 G3). Sample collection was conducted under stereomicroscopic guidance (SZX16, Olympus, Japan). The excised tissues were rinsed with 150 mM PBS and submerged in 0.3 ml TRIzol reagent (Molecular Research Center, OH, USA) for RNA isolation, then stored at –70℃ in a liquid nitrogen dewar. RNA extraction, library preparation, and RNA-seq data analysis were performed at Theragen Etex Company (Gyunggi do, Korea). Whole transcriptome sequencing was completed using RNA-seq technology. Total RNA from nodal tissue was extracted by pooling samples from the 10 G1-labeled tubes, using TRIzol LS reagent (Ambion, TX, USA). This procedure was replicated for G2 and G3 samples. Libraries were constructed automatically for all samples, with mRNA purification and fragmentation followed by double-stranded cDNA synthesis. The libraries’ final fragment sizes ranged from 350 to 450 bp, and their quality was assessed electrophoretically using an Agilent High Sensitivity DNA kit (Agilent Technologies, CA, USA) [12]. Transcriptome expression characteristics and individual gene expressions were statistically analyzed. After RNA-seq, data analysis proceeded as follows: sequencing output was saved in FastQC format. We evaluated the sequencing data’s quality by calculating the average quality of each base position, the average quality value of each read, Q20, Q30, GC content, base sequence components, junction content, and other metrics. High-quality reads were obtained by removing sequencing adapters and short sequences from the raw data. The STAR tool aligned the reads with the reference genome to determine their genomic positions [13]. The FastQC software package analyzed read distribution across various genomic features, including splice sites and strand-specificity. The randomness and integrity of the sequencing were assessed using a random distribution map. Then, splicing site saturation was assessed. Reads were aligned to ENSEMBL release 79 of the rat genome (Rnor_5.0; GCA_000001895.3) [14].

Tissue samples were collected from three different sites, skin nodes (G1), non-nodular skin area (G2), and inguinal lymph nodes (G3) for comparison. The gene expression level was comparatively measured by RNA-Sec analysis using an edgeR software package using FPKM (Fragments Per Kilobase Million) value as a unit [15]. The edgeR differential analysis results facilitated the filtering and clustering of significantly expressed genes into 14 functional categories [16]. The formula is FPKM = total exon Fragments/mapped reads (Millions) × exon length (KB) [17]. Only the genes with Fold Change > 2 and Normalized Data log₂ > 4 were selected for comparison by 14 different gene functions. Cuffdiff was used to compare the differentially expressed genes using a q-value < 0.05 as the cut-off.

4. Statistics

Data are presented as mean ± standard deviation. Differences between means were evaluated using the unpaired Student’s t-test (p < 0.05). The Cuffdiff tool was used to compare DEGs, with a q-value < 0.05 serving as the threshold for identifying significant DEGs.

1. Visualized EB reactive rat skin nodes

Following intravenous injection of the EB solution through the tail vein, multiple EB-reactive nodular skin spots were identified in the hair-shaved abdominal areas of the rats, as depicted in Fig. 1A. The EB-stained nodes were predominantly located beneath the left rib cage (Fig. 1A1) and on both sides of the inguinal region (Fig. 1A2 and 1A3). In addition to these protruding EB-stained nodes, there were also flat EB-stained spots, which appeared to be situated beneath the epidermis (Fig. 1A-1A3). These flat EB-stained skin spots were also observed in the hairless extremities, as shown in Fig. 1B. Within the area below the left rib cage, fourteen EB-stained nodes were discovered. Thirteen of these nodes aligned with the kidney (KI17-KI21), stomach (ST19-ST23), and spleen meridians (SP14-SP16), as illustrated in Fig. 1A1. One node adjacent to KI21, marked with a white arrow in Fig. 1A1, was located outside the established acupuncture meridians. Four EB-stained spots were situated along the conception vessel (CV) meridian (CV11-CV14), as shown in Fig. 1A1. Eleven EB-stained skin nodes, indicated by black arrows in Fig. 1A2, were found in the left inguinal area, with six along the liver channel line. The remaining five nodes were randomly placed. Numerous EB-stained spots were detected in this region, some along the inguinal ligament line but mostly unrelated to known acupuncture meridians. Nine EB-stained skin nodes, marked by black arrows in Fig. 1A3, were located in the right inguinal area, with three along the liver meridian line. The positions of the other six nodes appeared to be random.

The nodes were symmetrically distributed, with a mean 30.4 ± 6.3 (n = 5) on one side of the shaved abdominal area, but none along the midline. The mean distance between adjacent nodes was 3.29 ± 0.66 mm (n = 10), and the mean diameter of a nodes was 0.94 ± 0.09 mm (n = 10). These nodes were visible to the naked eye as milky-colored structures, even without the application of dye, as demonstrated in Fig. 1C. Measurements from 3D micro-CT scans revealed the average height and perimeter of the skin nodes to be 1.15 ± 0.04 mm (n = 10) and 2.79 ± 0.23 mm (n = 10), respectively (Fig. 1D). Careful examination of the micro-CT images revealed the protruded shape of the nodes, the interface between the dermis and hypodermis, and the boundary between the hypodermis and muscle layers, as indicated by arrows in Fig. 1D. However, no ductal structures within the nodes could be differentiated.

Inguinal area skin nodes in the rats were stained with AB solution administered via tail vein injection (Fig. 2). Topical application of AB solution did not result in staining; however, intravenous injection through the tail vein led to visible staining at the tips of six nodes (Fig. 2B), as confirmed in the magnified view (Fig. 2B1).

The cells within the skin nodes responded to Hemacolor staining when applied topically to the epidermal surface under anesthesia (Fig. 3A-C). Most cells within a node appeared as small, round, light-violet entities, likely white blood cells (Fig. 3C) [11]. Microscopic images of H&E staining within a node are shown in Fig. 3D-F. Clusters of numerous hair follicles were evident within nodes (Fig. 3D). An enlarged image (Fig. 3E) shows the epidermal cell layers, with the stratum corneum, stratum granulosum, and stratum basale indicated by red, black, and white arrows, respectively. Comparative measurements revealed no significant difference in the thicknesses of the epidermis and dermis between nodal and non-nodal areas. The mean epidermal thicknesses were 7.24 ± 1.15 µm (n = 5) for nodes and 8.63 ± 1.64 µm (n = 5) for non-nodal areas (p = 0.16), while the mean dermal thicknesses were 696.67 ± 94.75 µm (n = 5) for nodes and 586.80 ± 89.72 µm (n = 5) for non-nodal areas (p = 0.09). Red blood cells are indicated with arrows in Fig. 3F.

Fig. 3G-K depict TB staining results. TB solution applied epidermally stained multiple cells within a node, with MCs typically appearing purple (Fig. 3G) [11]. Fig. 3H shows purple-stained granules within two MCs, and Fig. 3I provides a higher-resolution image of an MC replete with granules. A count of stained MCs within a 1 × 1 mm area yielded a mean of 29.2 ± 2.95 (n = 5) cells in a node (Fig. 3J) and 15.8 ± 4.92 (n = 5) in non-nodal skin areas (Fig. 3K, p = 0.00079). A distinctive nodal feature observed was the distribution of hair follicles: a larger follicle surrounded by multiple smaller ones is depicted in Fig. 3J. This pattern of clustered follicles was common in skin nodes but not in non-nodal skin areas.

2. Comparative RNA expression of skin nodes

With the RNA-seq, we conducted DEG analysis to compare the gene expression functions across three groups: skin nodes (G1), non-nodular skin areas (G2), and LNs (G3). The LN group served as a non-skin tissue control. In G1, a variety of genes were abundantly expressed, showing significant differences when compared with G2 and G3, as illustrated in the scatter plots of Fig. 4. The distribution of dots—blue for genes more expressed in G1 and red for those less expressed—highlights the differences between the groups, with black dots indicating non-DEGs. Fig. 4A compares the expression profiles of 18,271 genes between G1 and G3, while Fig. 4B compares 17,725 genes between G1 and G2. A comprehensive list of all analyzed genes is provided in an accompanying Excel (Microsoft Corp., Redmond, WA, USA) spreadsheet as supplementary data.

Figure 4. Scatter plots of gene expression levels by RNA-seq analysis in three different rat body sites. Blue and red dots represent genes that are more and less expressed in the skin node, respectively. Black dots represent genes whose expression levels are not significantly different between the two groups. (A) Comparison of G1 to G3. (B) Comparison of G1 to G2. G1: skin node group; G2: non-nodular skin area group; G3: lymph node group. Val = log₂ ([FPKMgene + 1.0] / [FPKMaverage + 1.0]).

The gene expression pattern between G1 and G3 exhibited greater dispersion, with a Pearson correlation coefficient of 0.807 for the FPKM values of gene samples, as shown in Fig. 4A. This figure includes 750 upregulated genes (G3 > G1; red dots) and 798 downregulated genes (G3 < G1; blue dots). The dispersion between G1 and G2 was less pronounced, with a Pearson correlation coefficient of 0.909, featuring 797 upregulated genes (G2 > G1; red dots) and 79 downregulated genes (G2 < G1; blue dots), as depicted in Fig. 4B. Notably, 116 genes were exclusively expressed in G2, while 20 genes were unique to G1, as indicated in Fig. 4B.

Among the genes uniquely expressed in G1, with no expression in G2 and G3, the top five were further analyzed (Table 1). The genes exclusively expressed in G1 are listed in descending order of expression levels. AABR06014686.1, a gene of uncharacterized function, exhibited the highest expression level, with an FPKM value of 23584.9. The second-highest expression was observed for AABR06070504.1, a gene encoding protein FAM220A, with an FPKM value of 11.27. Urocortin 2 was the third. Genes exclusively expressed in G1, with no expression in G3, are listed in the lower section of the table, sorted by expression levels. The highest was a calcium-binding protein A14, with an FPKM value of 2470.43, followed by two keratin-associated proteins.

Table 1

Top 5 differentially expressed genes in G1 against G2 and G3 in rats.

Gene name	Description	FPKM (G1)	FPKM (G2)
AABR06014686.1	-	23584.9	0
AABR06070504.1	Protein FAM220A (Source:UniProtKB/TrEMBL;Acc:M0R495)	11.27	0
Ucn2	Urocortin 2	8.46	0
LOC100912610	Major urinary protein-like	6.63	0
Rn50_1_1736.2	40S ribosomal protein S29	2.63	0
Gene name	Description	FPKM (G1)	FPKM (G3)
S100a14	S100 calcium-binding protein A14	2470.43	0
Krtap8-1	Keratin-associated protein 8-1	1875.57	0
Krtap7-1	Keratin-associated protein 7-1	720.88	0
Stfa3	Stefin A3	696.13	0
Rn50_10_0706.4	WAP four-disulfide core domain protein 18	624.97	0

Gene functions were categorized into 14 sub-categories: aging, angiogenesis, apoptotic processes, cell cycle, cell death, cell differentiation, cell migration, DNA repair, extracellular matrix, immune response, inflammatory responses, neurogenesis, RNA splicing, and secretion. We compared gene expression between two groups to identify functions with higher expression levels in one group over the other. As summarized in Table 2, the number of genes predominantly expressed in skin nodes (G1) compared with non-nodular skin areas (G2) included 112 genes involved in cell differentiation, 52 in neurogenesis, 38 in cell migration, 27 in cell death, 26 in immune responses, 25 in inflammatory response, 19 in apoptosis, 18 in aging, 16 in cell cycle, 15 in secretion, 14 in angiogenesis, 12 in extracellular matrix, three in RNA splicing, and one in DNA repair. The most prominent gene groups were associated with cell differentiation and neurogenesis, which were also more prevalent in G1 compared with LNs (G3), with 365 and 193 upregulated genes, respectively. Genes significantly expressed in G1 were less expressed than in G2 across 13 of the 14 functions investigated, except for RNA splicing. Notably, genes associated with the extracellular matrix, which constitutes the structural material outside cells, were significantly less expressed in skin nodes (G1) than in non-nodular skin tissue (G2), with 80 genes downregulated. Conversely, relative to G3, genes related to the extracellular matrix were upregulated. In contrast, genes associated with immune and inflammatory responses were downregulated in G1, as indicated in Table 2.

Table 2

The number of genes significantly expressed in G1 compared to G2 and G3 by gene function category in rats.

Gene function	Total count	Significantly expressed genes in G1
		Compared to G2 (up/down)	Compared to G3 (up/down)
Cell differentiation	2,750	326 (112/214)	638 (365/273)
Neurogenesis	1,447	133 (52/81)	301 (193/108)
Cell cycle	807	33 (16/17)	132 (83/49)
Immune response	670	99 (26/73)	278 (55/223)
Cell death	659	73 (27/46)	188 (96/92)
Cell migration	624	101 (38/63)	200 (113/87)
Apoptotic process	593	58 (19/39)	171 (88/83)
Secretion	445	48 (15/33)	97 (44/53)
Aging	394	67 (18/49)	99 (57/42)
Inflammatory response	350	59 (25/34)	134 (54/79)
DNA repair	334	5 (1/4)	29 (8/21)
Extracellular matrix	294	92 (12/80)	78 (66/12)
RNA splicing	244	4 (3/1)	11 (4/7)
Angiogenesis	222	39 (14/25)	75 (45/30)

The expression levels of specific genes of interest were also examined. First, LYVE-1 (lymphatic vessel endothelial hyaluronan receptor 1), a marker gene for lymphatic vessels, was least expressed in skin nodes (FPKM values: G1: 3.1; G2: 12.7; G3: 27.7) [18]. Second, Krt10, the gene coding for keratin 10, which has been reported in the primo vascular system (PVS) on organ surfaces, was more expressed in skin nodes than in the comparison groups (FPKM values: G1: 5873.2; G2: 1324.7; G3: 1.0) [19].

In our attempt to validate the hypothesis that EB-reactive skin spots are MC-rich acupoints, we uncovered a novel anatomical structure within rat skin, which we have termed the “skin node.” EB dye is commonly utilized for preclinical and clinical evaluations due to its strong sulfonation reaction-based conjugation with serum albumin. Once administered intravenously, the majority of EB binds to albumin, leaving only a small fraction (0.11%-0.31%) unbound in the bloodstream [20]. Notably, EB’s slow excretion rate ensures its stability within the vascular circulation. An intriguing aspect of EB’s pharmacokinetics is its extravasation during neurogenic inflammation, a process driven by the vasodilatory effects of CGRP and substance P, released from the termini of afferent neurons [9,21]. Previous pathological animal model studies have reported serum-bound EB leakage from blood vessels, resulting in four to seven spots per rat, whereas control conditions typically yield less than one spot per rat [8].

In contrast, our study identified numerous EB-stained skin nodes across the abdominal region and hairless extremities of rats, even in the absence of a disease model, as illustrated in Fig. 1A and 1B. The cause of this discrepancy remains unclear, but it may be attributable to the hair removal process, which facilitated the observation of stained nodes.

Despite the use of EB staining to identify LNs in both animals and humans, several factors suggest that the skin nodes identified in our study are distinct from LNs [20]. Firstly, the skin nodes exhibited a high concentration of MCs stained with TB, as depicted in Fig. 3G-3K, whereas MCs typically have no LNs [22]. Secondly, the gene expression profile of skin nodes significantly diverged from that of LNs, as demonstrated in Fig. 4 and Table 1 and 2. Notably, the differential gene expression analysis summarized in Table 2 highlights cell differentiation and neurogenesis as the most significantly altered gene functions when compared with non-nodal skin (G2) and LNs (G3), reinforcing the notion that skin nodes are not LNs. Thirdly, the expression of LYVE-1, a lymphatic vessel marker gene, was markedly lower in skin nodes than in LNs.

The skin node identified in this study is not a sweat gland, as it has been well-established that rats lack sweat glands in the abdominal skin area [23]. Observations of a cluster of multiple hair follicles within the dermis layer of the skin node, as depicted in Fig. 3D and 3J, confirm that the size of a skin node is significantly larger than that of a hair follicle, ruling out the possibility that the skin node is merely a hair follicle. The skin is replete with various sensory nerve endings, such as the Pacinian corpuscle, which is sensitive to vibratory stimuli and forms an onion-like structure, and Meissner’s corpuscle, which functions as a pressure sensor. The location of these corpuscles in the dermal or hypodermal layers precludes the skin node’s protruded structure from being a nerve ending [24,25].

If the skin node is neither LN, sweat gland, hair follicle, nor nerve-ending corpuscle, its function remains to be elucidated. One hypothesis is that the skin node may be a MC-enriched site, potentially corresponding to an acupoint as proposed by the MC theory [6]. According to this theory, MCs, when activated by mechanical disturbance via an acupuncture needle, degranulate and release biological factors that influence nearby nerves and blood vessels, potentially triggering cascading effects on the nervous system, internal organs, and immune and endocrine systems [7]. Previous research has demonstrated a significantly higher concentration of MCs at acupoints compared with non-acupuncture sites in mice, rabbits, and humans [6,26,27]. Additional studies suggest that the activation of TRPV2, histamine H1, and adenosine A1 receptors in MCs may mediate the analgesic effects of acupuncture in mice [28]. Our findings indicate that the number of MCs in rat skin nodes is approximately double that in non-nodal skin areas, suggesting that skin nodes may indeed be MC-enriched sites, potentially serving as acupoints in the abdominal region of rats. The alignment of EB-reactive skin nodes with the kidney, stomach, and spleen meridians, as shown in Fig. 1A1, lends further credence to the notion that skin nodes could be MC-enriched acupoints. Another possibility is a connection to the PVS. The PVS, consisting of primo nodes (PNs) interconnected by primo vessels, is hypothesized to be the physical manifestation of acupuncture meridian channels [29]. Specifically, PNs in the skin are postulated to be the anatomical correlates of acupoints. PNs and primo vessels are characterized by a multi-ductal structure, distinguishing them from both blood and lymphatic vessels [30,31]. The abundant presence of MCs within PNs harvested from internal organ surfaces and lymphatic vessels in rats distinguishes PNs from LNs [11,22,32]. A recognized function of PNs is to facilitate extramedullary erythropoiesis, particularly under ischemic conditions [33]. A notable challenge to the PVS theory is the absence of PNs on the skin’s surface, specifically within the epidermal and dermal layers penetrated by acupuncture needles. Future research is necessary to determine whether the skin nodes identified in this study are indeed PNs within the skin. Nonetheless, several observations lend credibility to this hypothesis: firstly, skin nodes exhibit a high density of MCs, similar to PNs. Secondly, skin nodes absorb intravenously injected AB, as depicted in Fig. 2, a dye commonly used to stain the PVS due to its affinity for hyaluronic acid, which is prevalent within the PVS [34]. Lastly, skin nodes take up Hemacolor staining; this technique has been employed to color the PVS on organ surfaces and within the subcutaneous skin layer [11,35]. The distribution pattern of PNs along the CV meridian in the subcutaneous layer, as demonstrated by Lim et al. [35], corresponds with the EB-reactive skin spots along the CV meridian illustrated in Fig. 1A1. These findings suggest a potential connection between skin nodes and subcutaneous PNs.

This study’s limitations highlight areas for future research to validate the skin node as the anatomical basis of acupoints. Firstly, it is essential to explore species and sex differences, as the nodal structure has not been reported in the skin of other animals, such as dogs and horses, or in humans, despite the broader application of acupuncture beyond rats. Secondly, alternative hair removal methods, like depilatory creams, should be considered to prevent potential artifacts from blade-induced skin scratches, even though we meticulously used soap solutions to prevent shaving-related trauma. Thirdly, the observed increase in MCs within the skin node warrants further investigation into histochemical changes associated with acupuncture needling, given the established MC theory that acupuncture induces MC degranulation, triggering a cascade of cellular events with systemic effects [27,28]. Fourthly, the RNA-seq analysis results require validation through additional techniques such as real-time PCR, Western blotting, and immunohistochemistry. Finally, an in-depth examination of the skin node’s cellular architecture is necessary. Specifically, distinguishing cell types previously identified in the PVS—including mesothelial cells, chromaffin cells, immune cells, and p-microcells—within the skin node is crucial to ascertaining their identities as PNs [36].

In this study, we identified a novel anatomical structure within the skin of rats, characterized by its nodal shape and prominence above the skin surface. This skin node is distinguished by a markedly higher concentration of MCs relative to non-nodal areas and exhibits reactivity to dyes known to stain the PVS. Continued research into these MC-rich skin nodes is expected to enhance our comprehension of skin anatomy and physiology, as well as their potential clinical utility as acupoints. We advocate for further research to ascertain whether these MC-rich skin nodes constitute a component of the PVS, which is postulated to be the anatomical foundation of acupuncture meridians.

Supplementary data to this article can be found online at https://doi.org/10.51507/j.jams.2024.17.1.12.

Kiho Lee was a visiting scientist at Sungchul Kim’s lab at Wonkwang University, Iksan, South Korea. The concept development of the study was supported by a New Zealand - Korea Focal Point Programme grant from the Ministry of Business, Innovation and Employment to Kiho Lee (OBWT1101). Experimental works were carried out with the support of “Cooperative Research Programme for Agriculture Science and Technology Development (Project number: PJ 017015)” to Sungchul Kim from Rural Development Administration, Republic of Korea. The authors thank the late Professor Kwang-sup Soh for gifting microscopic instruments used in this study.

K.L. contributed to developing the concept, designing the experiments, anatomical observation, interpreting the results, and drafting the manuscript. J.Y.S. contributed to anatomical and histological observations and data archiving. E.C. contributed to designing and analyzing the RNA-Sec study. S.K. contributed to developing the concept, designing the experiments, interpreting the results, and supervising the experiments. All authors contributed to finalizing the manuscript.

Law data and materials analyzed and presented in this paper will be provided upon request to the corresponding author.

The authors declare no conflict of interest.