Live imaging of skin immunity using two-photon microscopy: a short review

Abstract

Two-photon excitation microscopy has become an important technique in dermatologic research, providing high-resolution imaging of living skin and its immune responses. Unlike conventional histology, two-photon microscopy allows direct observation of living tissues in real time at cellular resolution. This method enables the visualization of keratinocyte organization, skin appendages such as hair follicles and sebaceous glands, and the vascular, lymphatic, and neural networks within the dermis. It has also revealed dynamic immune processes in the skin. Clinical application is still limited by safety concerns and the high cost of equipment. Nonetheless, strategies based on autofluorescence, melanin scattering, and second harmonic generation (SHG) have been explored to visualize collagen and fibrosis in the dermis. This review summarizes the structural, functional, and translational aspects of two-photon skin imaging. We outline the technical principles, applications in animal and human studies, and potential implications for dermatologic research and clinical diagnosis.

Introduction

The skin, positioned at the interface between the body and the external environment, is among the most accessible organs for microscopic investigation. This accessibility makes it an ideal subject for imaging-based research in both experimental models and clinical contexts. Until the early 2000 s, most of the structural and functional analyses of skin at the cellular level relied on histological approaches. Tissue specimens were typically obtained by biopsy and then subjected to fixation, embedding, thin sectioning, and staining, after which the architecture was examined under light or fluorescence microscopy. Although this workflow provided invaluable information about cellular composition and spatial organization, it could only deliver static snapshots, producing a single two-dimensional cross-section at one point in time. As a result, essential temporal processes such as immune cell migration, vascular leakage, or dynamic interactions between structural and immune cells were inevitably lost.

To address these limitations, live imaging techniques including fluorescence microscopy and confocal laser scanning microscopy were developed [1, 2]. These methods enabled intravital visualization of tissues and represented a step forward in functional studies of skin. However, they were constrained by two major drawbacks: limited depth of penetration and significant phototoxicity. In single-photon excitation, fluorophores not only in the focal plane but also along the entire light path are excited, which leads to strong out-of-focus fluorescence, rapid photobleaching, and local tissue heating. These effects restricted both the achievable depth and the duration of observation, thereby narrowing the scope of biological processes that could be studied in vivo.

The advent of two-photon excitation microscopy brought a major advance to this field [3, 4]. In this technique, fluorophores are excited only when two low-energy photons are absorbed simultaneously at the focal point of the objective lens. To achieve the necessary photon density, femtosecond-pulsed infrared lasers, typically operating at wavelengths of 700–1000 nm, are employed. This design confines excitation to a very small focal volume, ensuring that the fluorophores’ excitation occurs exclusively at the focal plane. As a result, background fluorescence, photobleaching, and phototoxicity are dramatically reduced. Furthermore, the use of longer-wavelength lasers enhances tissue penetration, allowing imaging at depths of 300–400 µm in skin and up to 1 mm in transparent tissues [5]. Compared to conventional fluorescence imaging, this represents a substantial improvement in both image quality and biological safety.

These properties confer two decisive advantages [6]. First, two-photon microscopy allows visualization of fine cellular structures even with relatively small amounts of fluorescent labeling, which would otherwise fade rapidly under conventional methods. Second, it permits long-term observation of living tissues with minimal photodamage, making it well suited for time-lapse imaging of dynamic biological processes. In the skin, this capability is particularly valuable, as immune cells migrate at speeds of up to10 µm per minute, vascular permeability can change within seconds, and interactions between keratinocytes, immune cells, and vascular structures unfold in real time [7]. By capturing such events directly, two-photon microscopy provides insights that were previously inaccessible.

The aim of this review is to introduce the principles and applications of two-photon live imaging in dermatology. We will begin with structural visualization of the epidermis, dermis, and appendageal structures. We then focus on immune responses, including contact dermatitis, vascular permeability, and inflammatory cell trafficking. In addition, we address recent attempts to adapt this technology for clinical use and highlight computational image-processing strategies, such as the conversion of fluorescence images into hematoxylin–eosin (H&E)-like representations, which may help bridge research imaging and routine diagnostics. By integrating these perspectives, this review aims to provide an overview of how two-photon imaging has advanced dermatologic science and to discuss future directions where this technology may further contribute to both experimental immunology and clinical dermatology.

In vivo imaging of skin structures by two photon microscopy

In experimental mouse models, the rapid advancement of intravital skin imaging has been supported by the development of a wide array of fluorescent reagents and genetically engineered reporter mice. Various dyes, lectins, and antibodies have been established to label cutaneous structures such as blood vessels, lymphatic vessels, nerves, keratinocytes, and skin appendages. In parallel, numerous transgenic mouse lines expressing fluorescent proteins in specific immune cell populations have enabled direct visualization of immune cell behavior in vivo. Together, these technological innovations have substantially expanded the scope of intravital imaging studies and have played a central role in advancing research on skin immunity (Table 1).

Table 1.

Fluorescent reagents and transgenic reporter mice for intravital skin imaging

	Reagents	Labeling target	Reference
Reagents	Dextran	Blood vessels Macrophages Langerhans cells	[8–10]
	Evans blue	Blood vessels	[11]
	Quantum dot	Blood vessels	[12]
	Lectins	Blood vessels keratinocytes	[13, 14]
	BODIPY	Sebaceous glands Adipocytes	[15]
	Anti-Lyve1 antibody	Lymphatic vessels	[16]
Transgenic mice	CAG-EGFP	Ubiquitous	[17]
	CAG-Kaede/KikGR	Ubiquitous (photoconvertible)	[18, 19]
	CD11c-EYFP	Dendritic cells	[20]
	Langerin-GFP	Langerhans cells	[21]
	Lysozyme M-EGFP	Neutrophils Monocyte, Macrophages	[11, 20]
	DPE-GFP	Macrophages Plasmacytoid dendritic cells, T cells	[11]
	CSF1R-EGFP	Neutrophils Monocyte, Macrophages	[20]
	IL-1b-DsRed	Langerhans cells	[22]
	Mcpt5-YFP	Mast cells	[23]
	CXCR6-EGFP	$\gamma \delta$T cells	[24]
	CD41-YFP	Platelets	[25]
	Tie2-GFP	Blood vessels	[26]
	Prox1-EGFP	Lymphatic vessels	[27]
	Thy1.2-EGFP	Neurons	[28]
	Keratin14-GFP	Keratinocytes	[29]

Epidermis

The skin is composed of the epidermis, dermis, and subcutaneous tissue. The epidermis is a stratified squamous epithelium in which keratinocytes adhere tightly to one another to form a compact barrier. While keratinocytes constitute the majority of the epidermal cell population, smaller numbers of immune cells, including Langerhans cells and T cells, are also present. Intravital two-photon microscopy has allowed these elements to be visualized directly in living skin, providing information that was previously inaccessible. This chapter describes how skin structures and cells can be visualized by intravital imaging. Intravital imaging of the epidermis can be interpreted with reference to a schematic illustration of epidermal architecture (Fig. 1A), which provides a spatial framework for understanding the organization of the epidermal layers and the imaging depths used in subsequent panels.

Fig. 1.

Intravital imaging of epidermal structures. A Schematic illustration of epidermal architecture. Dotted lines indicate the approximate depths at which images in B, C, and D were acquired. B, C Nuclear staining of the epidermis using Hoechst 33342. B The granular layer and C the basal layer. D Visualization of intraepidermal Langerhans cells and keratinocytes. Langerhans cells are shown in green, and keratinocytes are shown in red. Keratinocytes were labeled using isolectin B4. E Schematic illustration of skin structure. F Intravital imaging of sweat glands in the plantar skin of K5-Cre;R26-EGFP mice. The epidermis and sweat glands are labeled in green, and dermal collagen fibers are shown in red. G Vertical section image of a hair follicle. Dermal collagen fibers are shown in blue, and the hair follicle is outlined by a white dotted line. H Visualization of sebaceous glands labeled with BODIPY. Dermal collagen fibers are shown in blue. Scale bars = 50 µm

Nuclear staining reveals structural differences between epidermal layers (Fig. 1B, C). In the granular layer (Fig. 1B), nuclei appear relatively large, uniform in shape, and arranged at nearly equal intervals, whereas in the basal layer (Fig. 1C), nuclei are densely packed and irregular in shape. The highly ordered arrangement of keratinocytes in the granular layer has been compared to Kelvin’s tetrakaidecahedron, a polyhedron with 14 faces including two broad hexagons [30]. This efficient packing geometry explains how keratinocytes flatten and interlock to maintain epidermal barrier strength and integrity.

Within this architectural framework, keratinocytes and Langerhans cells can be simultaneously visualized by using fluorescent lectins in combination with genetically modified mouse models such as Langerin-GFP mice [13, 31] (Fig. 1D). Under these conditions, keratinocytes exhibit sponge-like or honeycomb-like arrangements, with Langerhans cells intercalated between them in a compressed fashion. When T cells or neutrophils infiltrate the epidermis, they extend dendritic processes as they migrate through the narrow intercellular spaces between keratinocytes. These live observations demonstrate that keratinocytes are not merely passive structural components but also function as dynamic scaffolds that support immune surveillance and facilitate cellular migration.

Skin appendages

Skin appendages have also been investigated in vivo using multiphoton imaging (Fig. 1E–H). In keratin-5 promoter-driven GFP mice [32], two-photon microscopy combined with lipid dyes such as BODIPY enables clear visualization of hair follicles, eccrine sweat glands, and sebaceous glands. In mice, eccrine sweat glands are limited to the footpads and appear as relatively large glomerular-like structures (Fig. 1F), while sebaceous glands exhibit mulberry-like morphologies in situ (Fig. 1G, and H). Importantly, these observations preserve the intact spatial relationship of appendages with surrounding tissues, eliminating the need for excision or sectioning.

Beyond descriptive anatomy, the ability to examine intact appendages provides a foundation for understanding functional interactions in vivo. Hair follicles represent sites of active epithelial–immune communication, while sebaceous glands regulate lipid secretion and barrier homeostasis [33, 34]. Two-photon imaging of these appendages has already revealed dynamic behaviors, such as immune cell accumulation around hair follicles during inflammatory responses [35]. This type of analysis, which retains the natural three-dimensional context, highlights the unique contribution of intravital imaging to dermatologic research.

Dermis

In contrast to the epidermis, the dermis contains relatively low cell density and is composed largely of extracellular matrix, collagen, blood vessels, lymphatic vessels, and nerves (Fig. 2). Two-photon microscopy allows these components to be clearly distinguished by combining fluorescent lectins with specific immunolabeling approaches [7, 36]. When observed in vivo, blood vessels and nerves often exhibit similar diameters and branching patterns, which makes them difficult to differentiate morphologically; however, targeted labeling with specific antibodies or lectins enables accurate distinction (Fig. 2B). By contrast, lymphatic vessels are characterized by larger calibers and fewer branches than blood vessels, allowing their identification based on morphology.

Fig. 2.

Intravital imaging of dermal structures. A Schematic illustration of the skin. Dotted lines indicate the approximate depths at which the images shown in B, C, and D were acquired. B Visualization of dermal neurons, blood vessels, and lymphatic vessels. Neurons were labeled with isolectin B4, and both blood and lymphatic vessels were stained with anti-CD31 antibody. C Intravital imaging of dermal blood vessels and sebaceous glands. Blood vessels were visualized using TRITC–dextran, and sebaceous glands were labeled with BODIPY. D Visualization of lymphatic vessels and dermal dendritic cells in the ear skin of CD11c-YFP mice. Lymphatic vessels were labeled in red by intravenous injection of anti-LYVE-1 antibody, and dermal dendritic cells are shown in yellow. Scale bars = 100 mm

These vascular and neural structures can also be visualized in relation to sebaceous glands or infiltrating immune cells, thereby clarifying their anatomical context (Fig. 2C, D). Such imaging has emphasized that dermal vessels, lymphatics, and nerves do not merely provide structural support but also act as dynamic scaffolds for immune cell migration and intercellular communication in both physiological and pathological states. By integrating these structural details with functional imaging, two-photon microscopy provides a framework for studying how dermal microenvironments regulate inflammation, wound healing, and barrier homeostasis in vivo [37].

In vivo imaging of cutaneous immune responses

General principles and resident immune cells in the skin

Live imaging has proven particularly valuable when applied to the study of dynamic immune responses in the skin [7]. Even under homeostatic conditions, the epidermis and dermis harbor resident immune populations, including dendritic cells, macrophages, and mast cells. When inflammation occurs, neutrophils and T cells rapidly infiltrate from the blood circulation into the skin. Since these cells migrate at speeds of approximately 5–10 µm per minute, their trajectories can be captured as dynamic sequences when imaged at intervals of a few minutes [38]. Two-photon microscopy has therefore provided important opportunities to observe immune surveillance in real time, revealing aspects of cell-mediated and humoral immunity that cannot be inferred from static histology.

In experimental animal models, specific immune cell populations can be reliably distinguished in vivo by combining genetic, molecular, and behavioral information. Transgenic reporter mice expressing fluorescent proteins under lineage-specific promoters, such as Langerin-GFP for Langerhans cells, CX3CR1-GFP for mononuclear phagocytes, and LysM-GFP for neutrophils, enable direct visualization of defined immune subsets (Table 1). Beyond labeling strategies, immune cell types can also be inferred from characteristic morphological and motility features, such as the dendritic morphology and relatively slow migration of Langerhans cells or the rapid, amoeboid movement of neutrophils. By integrating these parameters, intravital imaging in animal models allows accurate discrimination of immune cell populations within intact skin.

Contact dermatitis and perivascular antigen presentation

Contact dermatitis, also known as a contact hypersensitivity (CHS) response, serves as a prototypical model of T cell–driven type IV allergic responses [39, 40]. Following antigen penetration, sensitization occurs in the draining lymph nodes, where dendritic cells prime antigen-specific T cells. During the elicitation phase, these activated T cells infiltrate the skin to mediate inflammation. Although it was long suspected that antigen presentation might also take place locally within the skin during the elicitation phase, the responsible antigen-presenting cells and their organizational mechanisms were unclear. Intravital imaging has addressed these questions directly in murine models [8].

Under steady-state conditions, dermal dendritic cells distribute evenly throughout the dermis and actively migrate at speeds of around 5 µm per minute, continuously patrolling for antigens (Fig. 3A, left panel). Following antigen challenge, they gradually accumulate to form clusters, and strikingly, T cells converge into the same sites, creating perivascular foci of antigen presentation (Fig. 3A, right panel; Fig. 3B). These aggregates consistently appear around small blood vessels, where perivascular macrophages also contribute to cluster organization (Fig. 3C). Such findings demonstrate that antigen-presenting dendritic cells and antigen-receiving T cells interact locally in defined dermal microenvironments, expanding the classical view that antigen presentation is restricted to lymphoid tissues.

Fig. 3.

Intravital imaging of dermal immune cell dynamics following hapten challenge. A Intravital imaging of the ear skin of hapten-sensitized CD11c-YFP mice before (left) and 24 h after (right) topical hapten application. Yellow cells represent dermal dendritic cells. Lymphatic vessels were labeled with anti-LYVE-1 antibody. B Intravital imaging of the ear skin 24 h after hapten application. Green cells indicate dermal dendritic cells, red cells indicate macrophages, and yellow structures represent blood vessels. Areas where dendritic cells accumulate around blood vessels are outlined by white dotted lines. C Time-lapse intravital imaging of hapten-sensitized mice following topical hapten application, monitored up to 24 h after challenge. Green cells represent dendritic cells, red cells represent T cells, and blue signals indicate dermal collagen fibers. Regions where dendritic cells and T cells form perivascular clusters are outlined by white dotted lines. Scale bars = 100 µm

The clustering of immune cells in the dermis is not restricted to contact dermatitis. Similar perivascular aggregates have been reported in psoriasis, systemic lupus erythematosus, Kimura disease, and in the tumor microenvironment of malignant melanoma [41, 42]. This phenomenon resembles the tertiary lymphoid structures that arise at mucosal sites such as bronchus-associated or mucosa-associated lymphoid tissues during chronic antigen exposure. However, cutaneous aggregates differ in that they lack well-developed B cell follicles and high endothelial venules, indicating that the skin organizes antigen presentation through distinct strategies adapted to its function as a peripheral barrier [43].

Vascular permeability and type I allergic responses

Intravital imaging has also clarified the dynamics of vascular permeability [9, 44], which underlie type I allergic responses such as urticaria. In this setting, mast cell degranulation releases chemical mediators such as histamine, which acts on blood endothelial cells to loosen intercellular junctions. Intravital injection of fluorescent dextrans has shown that under steady-state conditions, molecules up to 40 kDa extravasate readily, whereas larger tracers such as 70 kDa or more remain confined to the blood circulation (Fig. 4A). This indicates that skin vessels maintain a physiological barrier close to the 70 kDa threshold. Following histamine administration, however, the barrier collapses, and leakage of even 2000 kDa dextrans into the dermis becomes evident (Fig. 4B). These changes occur predominantly in postcapillary venules, the classical sites of inflammatory leakage, and can be completely blocked by antihistamine pretreatment, confirming histamine’s central role in the process [9].

Fig. 4.

Intravital visualization of vascular permeability. A Evaluation of vascular permeability under steady-state conditions using fluorescently labeled dextran. FITC–dextran with molecular weights of 20 kDa, 40 kDa, and 70 kDa was intravenously injected into mice, and ear vasculature was imaged by intravital microscopy. B Increased vascular permeability following intravenous histamine injection. Mice were administered 150 kDa FITC–dextran, and ear vasculature was imaged before (left) and immediately after (right) histamine injection. Scale bars = 100 µm

These insights have reframed the classical understanding of tissue swelling [45]. Rather than being regarded merely as a pathological byproduct, tissue swelling should be considered a physiological mechanism for delivering water and macromolecules into inflamed tissues. Albumin (~ 66 kDa) leaks together with water, diluting and flushing out antigens/pathogens, while immunoglobulins (~ 150 kDa) infiltrate the tissue and contribute to humoral defense. In this sense, swelling represents an adaptive response that facilitates antigen/pathogen clearance and antibody access to peripheral tissues. Two-photon microscopy has therefore highlighted tissue swelling as an active participant in immune protection, transforming conventional views of type I allergic responses.

Clinical applications to human skin

Technical and safety considerations

When intravital imaging is applied to human skin, safety considerations are of paramount importance. Unlike animal experiments, in which genetic labeling and systemic administration of fluorescent dyes are routinely performed, human studies face strict regulatory limitations [36]. The output power of the laser must be carefully controlled to avoid tissue burns, and the excitation wavelength is generally restricted to the longer, less phototoxic range. Because genetic modification and broad systemic dye administration cannot be performed in human subjects [46], both the spatial resolution and imaging depth achievable in vivo are inherently more limited than those in experimental animal models. Under standard clinical laser settings, human skin can be imaged only to a depth of about 100–150 µm—the epidermis and superficial papillary dermis—because the skin is much thicker than that of experimental animals [4]. Several physical and optical factors further limit penetration depth and resolution. Melanin absorption reduces excitation light delivery to deeper layers, while optical scattering by dermal collagen progressively degrades photon coherence and signal intensity. In addition, clinical objective lenses with relatively low numerical aperture (NA) restrict achievable resolution, and refractive index mismatches at skin interfaces introduce optical aberrations that worsen with depth. Consequently, contrast in human studies relies primarily on autofluorescence, melanin-based scattering, and second harmonic generation (SHG) rather than on exogenous fluorescent probes [4, 47].

Because genetic reporters and systemic administration of cell-type–specific fluorescent probes are not feasible in clinical settings, the ability to discriminate specific immune cell subsets in vivo is inherently limited in human skin imaging. Although intrinsic signals and morphological features allow visualization of tissue architecture, superficial vasculature, and inflammatory cell accumulation, they do not permit definitive identification of individual immune cell types. As a result, current human imaging studies primarily focus on structural and functional changes at the tissue level rather than detailed cellular phenotyping.

Current clinical use and practical challenges

Because of these restrictions, the most common clinical applications of intravital optical imaging have been in the evaluation of tumors and pigmented lesions, where structural alterations are relatively pronounced and more readily detectable [46, 47]. By contrast, the diagnostic application of two-photon microscopy for inflammatory skin diseases remains challenging, as inflammatory processes involve dynamic cellular changes [48].

From a practical standpoint, insurance coverage strongly influences whether optical imaging technologies become integrated into clinical practice. Reflectance confocal microscopy has been approved for reimbursement in the United States and is now widely used in dermatology clinics, particularly for the assessment of skin tumors [49, 50]. In Japan, however, this technology has not been covered by insurance and has therefore not been broadly adopted. Commercially available bedside two-photon microscopes have been developed, but their cost remains prohibitive, restricting their use mainly to academic research centers rather than routine patient care [51].

Compared with confocal microscopy, two-photon excitation provides several important advantages. Long-wavelength excitation light produces brighter signals with reduced phototoxicity, point excitation minimizes collateral tissue damage, and superior penetration depth enables visualization of structures located deeper than those accessible with confocal instruments. However, spatial resolution remains lower than that achieved with confocal microscopy, and this limitation must be weighed against the advantages when considering diagnostic applications [5, 52].

In vivo collagen imaging using second harmonic generation (SHG)

A distinctive feature of two-photon microscopy is its ability to generate second harmonic generation (SHG) signals [53]. In human skin, SHG enables noninvasive visualization of dermal collagen fibers. This approach has proven especially relevant in fibrotic skin diseases such as systemic sclerosis, localized scleroderma, and chronic graft-versus-host disease, where increased collagen deposition and architectural disorganization are hallmark features [54, 55]. Beyond overt fibrosis, SHG imaging has also been applied to assess age-related dermal remodeling, demonstrating reduced collagen density and increased fiber fragmentation in photoaged skin [56]. Moreover, in wound healing models, longitudinal SHG imaging has captured the dynamic reorganization of collagen bundles during tissue repair, offering insights into the mechanisms underlying scar formation and remodeling [57]. Recent studies further indicate that SHG-derived metrics—including collagen orientation, density, and anisotropy—can serve as quantitative biomarkers for disease progression and therapeutic response, highlighting the translational potential of this technique for both research and clinical practice [58].

Image processing for diagnostic assistance

Virtual H&E for intravital imaging

Given the wealth of basic science data obtained in animal models, there is considerable anticipation that intravital imaging techniques can be translated into routine clinical dermatology [59]. To achieve this, however, the development of fluorescent probes and contrast agents that are safe for human application will be essential [60]. The ability to selectively label and monitor multiple cell types in vivo would greatly expand the diagnostic utility of this technology.

In addition, advances in computational image processing and automated diagnostic support are expected to further bridge the gap between laboratory research and clinical practice [61]. In routine dermatopathology, the standard workflow is based on vertical sections of skin biopsies stained with hematoxylin and eosin (H&E). For more than a century, diagnostic criteria and medical education have been grounded in the morphology and staining properties of H&E sections. Physicians are therefore deeply accustomed to recognizing nuclear detail, cytoplasmic staining, and tissue organization in this format. By contrast, the images generated by confocal or two-photon microscopy differ not only in orientation but also in visual appearance. These methods typically produce horizontal optical sections, often displayed as grayscale or monochromatic pseudo-colored images that emphasize fluorescence intensity rather than traditional morphology. Such differences create a steep interpretive barrier. For clinicians trained in classical histology, fluorescence images are unfamiliar and difficult to interpret, which hinders the widespread adoption of these technologies in clinical settings.

To address this barrier, researchers have developed computational approaches to transform fluorescence-based intravital imaging into H&E-like formats [62, 63]. The purpose of this transformation is not to add new biological information, but rather to recast the same data into a format that reproduces the color, texture, and contrast of conventional pathology, thereby generating virtual H&E slides familiar to clinicians. This reduces the cognitive gap between advanced imaging and traditional histology, facilitating clinical interpretation.

An illustrative example is shown in Fig. 5, where fluorescence images of murine skin obtained by two-photon microscopy were digitally recolored. Horizontal sections reveal keratinocyte layers, immune cell infiltrates, and appendageal structures that become more readily recognizable once rendered into H&E-like coloration (Fig. 5A). By further reconstructing horizontal (x–y) images into vertical (x–z axis) views, the spatial relationships of nuclei within the epidermis can be observed in an orientation comparable to that of conventional H&E-stained histological sections (Fig. 5B). Although the underlying raw data remain unchanged, this simple transformation greatly enhances interpretability and is particularly useful for clinicians unfamiliar with intravital imaging.

Fig. 5.

Example of converting fluorescence images into H&E-like coloration. A Conversion of a horizontal section image of murine ear skin into H&E-like coloration. Nuclei were stained with Hoechst 33342, and cytoplasm and stromal components were labeled with rhodamine. The left panel shows a pseudo-colored fluorescence image acquired by two-photon microscopy, and the right panel shows the corresponding image converted into H&E-like coloration. B Conversion of a vertical section image of murine skin. The upper panels show pseudo-colored fluorescence images acquired by two-photon microscopy, and the lower panels show the corresponding H&E-like color–converted images. Scale bars = 50 µm

Computational “virtual H&E” staining is not equivalent to conventional histopathology based on chemically stained tissue sections. Virtual H&E digitally approximates hematoxylin–eosin contrast but does not fully reproduce dye–tissue interactions; in particular, nuclear features lack the chromatin texture and resolution provided by chemical hematoxylin staining, and eosin-like coloration reflects an approximation of cytoplasmic protein density rather than true dye-binding specificity.

Validation studies have reported good overall correlation between virtual H&E images and conventional histology for gross tissue architecture and major pathological features; however, they also recognize diagnostic limitations, especially in the evaluation of subtle dysplasia and early neoplastic changes [61, 63]. Accordingly, at its current stage, virtual H&E should be regarded as a complementary or screening approach rather than a replacement for full-thickness vertical histological sections in routine diagnostic practice.

Toward optical biopsies and clinical integration

The broader implication of these developments is that computational image-processing tools could transform intravital imaging into a practical diagnostic modality. By coupling automated H&E-like conversion with real-time intravital microscopy, clinicians may be able to interpret “optical biopsies” directly at the bedside, enabling immediate tissue assessment without the delays of excision, fixation, and staining. Integrating machine-learning–based diagnostic support—for example, automated detection of atypical nuclei, inflammatory infiltrates, or dermal fibrosis—has the potential to reduce interpretive variability and improve diagnostic accuracy [64, 65]. Importantly, these approaches may also lessen reliance on invasive skin biopsies. In conditions that require repeated monitoring, such as psoriasis, atopic dermatitis, or cutaneous lymphoma, noninvasive, real-time optical biopsies could substantially ease patient burden. Beyond dermatology, similar strategies could be extended to other tissues accessible to optical imaging, including the oral mucosa, cervix, and gastrointestinal tract.

An additional consideration for clinical intravital imaging is the appropriate temporal scale of observation. Cutaneous immune responses encompass phenomena that occur over markedly different time frames. Rapid events unfolding over seconds to minutes, such as neutrophil swarming, vascular leakage, and immediate mast cell–mediated responses, can be effectively captured by short-term time-lapse imaging. Intermediate processes, including dendritic cell migration, T cell interactions, and antigen presentation dynamics, typically evolve over minutes to hours and require sustained but still limited observation periods. By contrast, long-term tissue changes such as fibrosis remodeling, wound healing, and chronic inflammatory progression occur over days to weeks and are better assessed through repeated imaging sessions rather than continuous observation. In clinical settings, these temporal considerations must be balanced against practical constraints, including motion artifacts, laser safety limits, and patient comfort, which restrict prolonged continuous imaging.

Conclusion and future perspectives

Two-photon intravital microscopy has opened new ways to observe living skin directly. It allows real-time observation of how immune cells move, how they interact with keratinocytes, and how tissue changes over time. Fluorescent reporters and functional tools, such as calcium indicators or cytokine reporter mice, will help reveal not only cell movements but also signaling processes inside the skin.

Further progress not only in imaging hardware—such as more stable lasers, improved optics, and better motion control—but also in methods to visualize specific tissue structures and cell types will make it possible to look deeper into tissue with clearer and more informative images. A key step for clinical use will be the development of safe fluorescent probes and contrast agents for humans. With these tools, two-photon microscopy could be used not only for research but also to monitor disease activity and treatment effects in skin disorders.

Looking ahead, the integration of intravital microscopy with other label-free approaches, such as second harmonic generation (SHG) for collagen imaging or autofluorescence lifetime analysis for metabolic readouts, may expand its clinical value. These techniques could provide complementary information on tissue remodeling, fibrosis, and cellular metabolism, enabling noninvasive and quantitative monitoring of disease progression. In the longer term, combining functional readouts with structural imaging may yield a more holistic picture of skin health and treatment response, moving beyond morphology alone.

There are still many challenges to be addressed, including safety, the high cost of equipment, and the need for more user-friendly systems. Nevertheless, the future direction is clear: intravital two-photon microscopy has the potential to complement, or even partially replace, conventional skin biopsies in the future. As imaging methods, labeling strategies, and computer analysis continue to improve, these technologies may change how we study and diagnose skin disease—moving from experimental settings toward real use in patient care.

Abbreviations

CHS

Contact hypersensitivity

GFP

Green fluorescent protein

H&E

Hematoxylin and eosin

NA

Numerical aperture

SHG

Second harmonic generation

Author’s contributions

GE was responsible for writing the manuscript.

Funding

None.

Data availability

Not applicable.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The author declares no competing interests.