Building an Insect Wing: Single-Cell and Spatial Transcriptomics of the Silkworm Wing Disc

Authors: IJHSB editorial board

Publish Date: 13.4.2026

Abstract

Adult insect wings do not form all at once. They arise from larval tissues called wing discs. In these tissues, cells proliferate, differentiate, and reorganize over time. A 2026 study of the silkworm Bombyx mori combined single-cell RNA sequencing, single-nucleus RNA sequencing, and spatial transcriptomics to construct a spatiotemporal atlas of wing-disc development across ten developmental timepoints. The study identified major cell populations and traced transitions between them. It also suggested that wing morphogenesis cells may act as central progenitors for epithelial and cuticle lineages. In addition, the authors showed that 20-hydroxyecdysone can rapidly accelerate later developmental gene programs. Together, these findings present wing formation as a staged, spatially organized, and hormonally regulated process. They also show how modern omics methods can clarify how an organ forms over time.

Keywords: wing disc; morphogenesis; spatial transcriptomics; single-cell RNA sequencing; 20-hydroxyecdysone; insect development

Introduction

A mature wing appears highly ordered. Its final structure seems precise and stable. Development, however, is a dynamic process. In holometabolous insects, adult appendages arise from imaginal discs. These are larval tissues that begin as small groups of undifferentiated cells and later undergo growth, patterning, and morphogenesis [1,2]. In the silkworm Bombyx mori, classical morphological work suggested that wing discs already exhibit an early differentiated state and carry a preliminary adult wing-vein pattern before metamorphosis is complete [1]. This makes the silkworm wing disc a particularly informative system for studying how an organ is progressively refined rather than built from a blank tissue.

Earlier approaches could not fully resolve tissue heterogeneity. A developing tissue contains cells in different states at the same time. Cells that appear similar under a microscope may already be following distinct developmental paths. Bulk transcriptomic methods reveal broad molecular changes, but they average signals across many cells. As a result, they may obscure important cell-specific transitions. Single-cell RNA sequencing helps address this problem by measuring gene expression in individual cells. Spatial transcriptomics adds positional information by showing where those cells are located in the tissue [3–5]. This is important because development depends not only on cell identity, but also on spatial organization.

These methods have expanded the questions that developmental biologists can ask. Researchers can now investigate which genes are active in specific cells, where those cells are positioned, how cell states change over time, and how signals influence those changes [3–5]. Development can therefore be studied as a coordinated process across both time and space.

A recent study by Liu et al. applied this framework to the silkworm Bombyx mori [6]. This is a useful model because its wing discs are relatively large and its developmental stages are well defined [1,6,7]. The study addressed a fundamental question: how does an insect wing build itself? The answer was not a single regulatory event. Instead, the authors described a layered developmental program that unfolds across time, tissue space, and endocrine state [6].

Methods

This Short Communication is a literature-based synthesis centered on the 2026 primary study by Liu et al. [6]. Supporting literature was identified through PubMed, Google Scholar, and publisher webpages using terms related to silkworm wing discs, imaginal discs, ecdysone signaling, wing morphogenesis, single-cell transcriptomics, and spatial transcriptomics. Priority was given to primary research articles and recent high-quality reviews [1–18]. The purpose is not to provide a full review of insect wing development. Rather, the goal is to explain the biological significance of one recent study in a concise and accessible form.

Results

A time-resolved atlas of wing formation

Liu et al. profiled silkworm wing discs across ten developmental timepoints spanning larval, wandering or spinning, and pupal stages [6]. The dataset included 11 libraries and, after quality control, 126,041 single-cell transcriptomes [6]. This dense temporal sampling is important. Development is not a single abrupt event. It consists of a sequence of state changes. By sampling across many timepoints, the study reconstructed developmental progression as a series of transitions rather than a simple before-and-after comparison [6].

The atlas identified 12 major cell types and their developmental relationships [6]. Among these, the most important were the wing morphogenesis (Wm) cells. The authors described these cells as central progenitors that appear to bifurcate toward epithelial and cuticle lineages under lineage-specific transcriptional control [6]. This is one of the main strengths of the study. It does not only classify cells into groups. It also proposes how one population may contribute to later tissue organization. The study therefore moves beyond a static description of cell types and toward a dynamic model of tissue formation [6].

The paper further showed that this developmental progression is not transcriptionally uniform. Time-resolved snRNA-seq suggested hierarchical reprogramming across stages, with Wm cells functioning as early signaling hubs during the transition toward later fates [6]. This point is important because it gives Wm cells a dual role. They are not only intermediates in a lineage model. They also appear to participate actively in the signaling logic of morphogenesis [6].

This conclusion becomes stronger when compared with earlier silkworm studies. A 2014 transcriptomic analysis of Bombyx mori wing discs during metamorphosis showed that the larva-to-pupa transition involves large-scale gene-expression changes [7]. These included upregulation of many 20-hydroxyecdysone-related genes, downregulation of juvenile hormone pathway genes, and strong expression of cuticle protein and chitin-related genes [7]. However, that earlier work used bulk transcriptomics. It could not identify which cells changed, in what order, or in which tissue regions. The 2026 study adds cellular and developmental resolution to a process that had previously been visible mainly at the tissue level [6,7].

The new atlas also gains meaning when placed beside earlier developmental genetics in Bombyx. In the wingless mutant flugellos, wing discs were shown to develop nearly normally until the fourth larval instar, but later fail to undergo normal epithelial invagination and tracheal migration [8]. These older observations already suggested that silkworm wing formation depends on tightly stage-specific morphogenetic transitions rather than simple tissue growth. The 2026 atlas provides the cellular framework that was missing from those earlier morphological descriptions [6,8].

Spatial transcriptomics turns clusters into tissue

An important feature of the study is its use of spatial transcriptomics. Single-cell data identify cell states, but spatial data show where those states are located in the tissue [6]. This is essential for understanding morphogenesis. A cell’s developmental behavior depends not only on its transcriptional profile, but also on its position, its neighboring cells, and the signals present in its local environment. Single-cell RNA sequencing defines identity. Spatial transcriptomics defines context. Together, these methods make tissue organization easier to interpret [3–5].

The spatial component of the study was also used carefully. scRNA-seq covered all sampled stages, but Stereo-seq was applied specifically to selected developmental stages, including L5D2, L5D3, and W2 [6]. This is an important detail. The study did not produce a full spatial map across all ten timepoints. Instead, it integrated dense temporal single-cell data with stage-specific spatial snapshots [6]. This design still allowed the authors to localize major cell populations and reconstruct spatial organization across key transitions.

Using this approach, Liu et al. showed that major cell populations occupy distinct anatomical regions in the developing wing disc [6]. Wm, epithelial, cuticle, matrix, immune, metabolic, and other cell types were not randomly distributed. Their spatial arrangement shifted with developmental stage, consistent with progressive tissue remodeling [6]. This strengthens the interpretation that morphogenesis is a spatially coordinated process rather than a collection of isolated transcriptional states.

This spatial logic is also consistent with earlier Bombyx developmental genetics. Positional cloning of the flugellos locus showed that the causative gene is fringe (fng), which encodes a glycosyltransferase involved in modulation of Notch signaling [9]. In wild-type silkworms, fng is actively expressed in wing discs, and wingless (wg) mRNA is localized at the dorsoventral boundary. In fl mutants, this boundary-associated expression is markedly repressed [9]. These findings provide a useful historical bridge between classical wing-patterning logic and the newer cell-state atlas. They show that spatial patterning in the silkworm wing disc has long been linked to localized regulatory activity, even before transcriptomic methods made those patterns globally measurable [6,9].

This is also why the study has significance beyond the silkworm model. Spatial transcriptomics has become an important method in modern biology because it reconnects gene expression with tissue architecture [4]. At the same time, comparative evaluations show that spatial platforms differ in capture strategy, spatial resolution, and signal behavior [5]. Their results must therefore be interpreted with caution. The silkworm study is compelling because it does not use spatial analysis as a descriptive add-on. Instead, it integrates spatial data with developmental timing, cell-state transitions, and hormone perturbation [5,6].

Transcription factors that shape Wm differentiation

Another strong part of the paper is its effort to move from descriptive clustering toward regulatory interpretation. Using SCENIC analysis, the authors identified transcription factors with stage-specific activity across Wm differentiation [6]. Among the most prominent were Rfx, Blimp-1, Dll, sqz, and Pur-alpha [6]. These factors were not active in the same way across time. Instead, they appeared to mark different phases and possible directions of Wm differentiation [6].

The paper proposes a useful functional distinction among these regulators. Genes regulated by Dll and sqz were associated mainly with wing morphogenesis and showed strong activity at early and late points in development, suggesting a role in directing Wm cells toward cuticle-related fate programs [6]. In contrast, Pur-alpha target genes were associated more strongly with epithelial differentiation [6]. Blimp-1 activity was enriched at the transition from late larval to wandering stages, suggesting that it may help promote the shift from Wm toward epithelial states [6]. Among these regulators, Rfx was highlighted as a particularly strong and cell-type-specific regulator of Wm identity [6].

This part of the study adds mechanistic depth. It suggests that the Wm population is not a vague transitional compartment. It is governed by a transcriptional program with identifiable regulators that may bias lineage choice [6]. That gives the model more explanatory power and makes the proposed differentiation logic easier to test in future work.

Functional validation strengthens the model

The most convincing mechanistic evidence in the paper comes from functional validation of Rfx. Liu et al. used RNA interference to reduce Rfx expression and then examined both molecular and morphological outcomes [6]. The knockdown disrupted the downstream regulatory network of Wm cells and altered the expression of several target genes involved in tissue organization and morphogenesis [6].

These transcriptional changes were accompanied by clear structural defects. Within 48 hours after RNAi treatment, wing discs showed lobulation in the central region, irregular epithelial invagination, and disorganized tracheal branching [6]. In adult moths, Rfx knockdown caused shortened veins, malformed marginal bristles, and jagged posterior wing edges [6]. The authors also reported that knockdown of the orthologous gene in the fall armyworm Spodoptera frugiperda produced similar defects, suggesting that the role of Rfx may be conserved across lepidopterans [6].

This is an important strength of the study. Many atlas papers remain correlational. Here, at least one central regulatory prediction was tested experimentally. That does not prove the entire lineage model. However, it does support the idea that Wm-associated regulators have real developmental consequences and are not only computational associations [6].

This functional layer also fits well with older and parallel Bombyx studies. The flugellos/fringe work established that disruption of a localized developmental regulator can block normal wing morphogenesis [9]. Later work showed that the miR-2 family can target both awd and fng, and that overexpression of miR-2 or functional disruption of those targets results in deformed adult wings [15]. Together, these studies strengthen the conclusion that the silkworm wing disc is not only transcriptionally dynamic but also genetically tractable, with multiple experimentally supported regulators of form [9,15].

20-hydroxyecdysone as a temporal regulator

The endocrine findings are among the most notable results of the study. In insects, 20-hydroxyecdysone is a major steroid hormone involved in molting and metamorphosis [7,12]. Previous work in Drosophila showed that signaling through the ecdysone receptor controls the switch to a post-critical-weight mode of wing-disc development [12]. More broadly, wing-disc studies have long treated endocrine regulation as a key component of developmental timing [2,12,13]. Liu et al. extend this view in a more direct and dynamic way [6].

Using time-resolved single-nucleus RNA sequencing after 20-hydroxyecdysone treatment, the authors described a phenomenon they termed “time-axis compression” [6]. Hormone exposure rapidly accelerated fate transitions and maturation of gene-expression programs. As a result, treated tissues began to resemble later natural developmental stages within hours [6]. In simple terms, the hormone appeared to fast-forward development at the transcriptomic level. This is stronger than a simple correlation between hormone presence and metamorphosis. It suggests that endocrine signals can alter where a tissue lies along its developmental trajectory [6].

It is important, however, to describe this result precisely. The study shows that 20-hydroxyecdysone treatment rapidly recapitulates aspects of natural developmental progression at the level of gene expression and inferred cell-state transition [6]. It does not show that the full complexity of natural wing morphogenesis can be recreated immediately in morphological terms. This distinction matters. It keeps the interpretation strong, but appropriately cautious.

The paper also reported that 20-hydroxyecdysone-related genes display clear spatiotemporal specificity across cell types [6]. This supports the idea that hormone response is not uniform across the tissue. Instead, endocrine signaling appears to intersect with cell identity and developmental stage. That makes hormone action part of the logic of pattern formation rather than a simple background trigger [6].

Earlier Bombyx work makes this endocrine interpretation more secure. In the flugellos mutant, wing discs fail to respond normally to 20-hydroxyecdysone during metamorphosis, linking defective morphogenesis to defective hormonal response [8,9]. In addition, promoter studies in the silkworm wing disc showed that the BmE74B promoter is activated by low concentrations of 20E in a stage-specific manner, and that putative ecdysone response elements are required for this activation [10]. Parallel work on the cuticle protein gene BMWCP10 showed that its promoter is directly and indirectly regulated by ecdysone in the wing disc, with a defined EcRE required for full hormone responsiveness [11]. These studies are important because they show that ecdysone action in Bombyx wing discs is not only a systemic developmental signal. It also operates through tissue-specific regulatory elements and stage-restricted promoter responses [10,11].

A five-stage model of wing development

To summarize their results, Liu et al. proposed a five-stage Gene Transition Model that integrates morphology, hormone levels, and transcriptomic change [6]. This model is useful both biologically and conceptually. Developmental omics studies often produce large numbers of clusters, trajectories, and pathways. A staged model provides a clearer structure. It suggests that wing formation can be divided into meaningful phases marked by increasing commitment and structural differentiation [6].

In this model, Stage 1 represents the developmental blueprint, Stage 2 the cellular foundation, Stage 3 structural formation, Stage 4 remodeling during the wandering or spinning period, and Stage 5 maturation and stability in the pupa-to-adult transition [6]. The model also links these stages to dynamic signaling pathways, including Notch, PARs, FGF, Hippo, BMP, Hedgehog, SEMA4, and Wnt signaling [6]. This integration of morphology, cell states, endocrine timing, and signaling pathways is one of the most ambitious features of the study.

This staged framework also makes the topic suitable for a student-oriented Short Communication. A wing is a familiar structure. A larval wing disc is less familiar and therefore intellectually engaging. The five-stage model adds narrative clarity. The hormone result adds mechanistic depth. Together, these features make the topic both current and accessible. The central story remains simple: a tissue begins in a relatively plastic state, progresses through organized transitions, and is accelerated by an endocrine signal that helps convert developmental potential into adult structure [6].

The overall developmental framework proposed by Liu et al. is summarized in Figure 1, which integrates temporal progression, major cell populations, inferred lineage relationships, spatial organization, key transcriptional regulators, and the accelerating effect of 20-hydroxyecdysone on transcriptomic maturation. Rather than presenting the atlas as a collection of isolated datasets, the figure highlights how these components converge into a single developmental model of wing formation [6].

Figure 1 is intended to guide the reader through the logic of the study from left to right. The early part of the figure should show the developmental timeline and the five-stage progression of the wing disc. The central part should position wing morphogenesis (Wm) cells as a pivotal transitional population and indicate their proposed differentiation toward epithelial and cuticle lineages. The spatial component should illustrate that these populations are not randomly distributed, but occupy distinct regions that change over developmental time. The regulatory layer should emphasize Rfx, Blimp-1, Dll, sqz, and Pur-alpha as candidate drivers of Wm fate progression. Finally, the endocrine component should show how 20-hydroxyecdysone accelerates transcriptomic maturation, thereby linking developmental timing to hormone-dependent gene regulation [6]. In this way, Figure 1 serves not only as a visual summary, but also as a conceptual map of how cell states, spatial organization, and endocrine signaling interact during silkworm wing morphogenesis.

Figure 1. Integrated model of silkworm wing-disc development across time, space, and endocrine regulation.
_{The figure summarizes the developmental timeline from larval to wandering or spinning and pupal stages, the major cell populations identified in the wing disc, the proposed differentiation of wing morphogenesis cells toward epithelial and cuticle lineages, the spatial organization of these populations at selected stages, key transcriptional regulators associated with Wm fate progression, and the accelerating effect of 20-hydroxyecdysone on transcriptomic maturation. Solid arrows indicate developmental progression supported by the study, whereas dashed arrows indicate inferred lineage relationships or regulatory influences. Selected earlier Bombyx findings, including the early differentiated state of the larval wing disc, the flugellos/fringe pathway, and stage-specific ecdysone-responsive promoter activity, provide historical context for the 2026 atlas.}

Discussion

The broader significance of this study lies in how it reframes organogenesis. A wing is not produced by one pathway acting uniformly across the tissue. It emerges from the interaction of cell identity, spatial organization, and developmental timing [3–6]. This is why the combination of single-cell and spatial methods is so informative. These tools show how development can be heterogeneous at the local level and coordinated at the tissue level. Complexity in a developing organ does not arise from disorder. It arises from structured differences that become integrated over time.

The silkworm study also fits into a wider development in lepidopteran biology. Recent single-cell and single-nucleus studies of butterfly wings have identified multiple wing cell types, scale-cell markers, sex-specific scale states, and evidence that lepidopteran scales arise from sensory organ precursors through a canonical lineage [14,17,18]. These studies do not address the same developmental question as the silkworm atlas. However, together they indicate that insect wings are becoming a strong comparative system for studying cell fate, tissue patterning, evolutionary diversification, and morphological specialization [14,17,18]. Research in this field is moving beyond adult wing description and toward the cellular logic of wing formation.

At the same time, the Bombyx literature now provides more than descriptive context. Classical morphology showed that silkworm wing discs are already partially differentiated before metamorphosis [1]. Mutant analysis linked failed wing development to disrupted morphogenetic response and defective hormone sensitivity [8,9]. Promoter studies demonstrated that hormone response in the wing disc is stage-specific and encoded at the cis-regulatory level [10,11]. More recent work further implicated the miR-2–awd/fng axis in wing morphogenesis and showed that the cytokine receptor DOME supports proper wing-disc development through JAK/STAT signaling [15,16]. When these studies are viewed together, the 2026 atlas appears not as an isolated breakthrough but as the highest-resolution layer in a longer Bombyx research tradition [1,6,8–11,15,16].

Some caution is still necessary. Trajectory models do not provide the same level of evidence as direct lineage tracing. Spatial transcriptomic maps are highly informative, but they are still shaped by technical limits such as platform-specific resolution and signal behavior [5]. In addition, most causal claims in the paper remain stronger for individual regulators, such as Rfx, than for the entire developmental network [5,6]. The study is therefore best read as a high-resolution model of wing development rather than a fully complete mechanism.

Another important point concerns biological scope. The value of this work does not depend on forcing direct parallels with vertebrate development or human medicine. Its importance lies in clarifying a general developmental principle: organs form through timed and spatially organized transitions among cell states [5,6]. At the same time, because wing development influences insect survival and performance, these findings may also interest applied entomology and insect biotechnology. The paper itself notes possible relevance for the manipulation of insect development in agriculture, although that applied prospect remains preliminary [6].

From this perspective, the wing becomes more than a finished appendage. It becomes a developmental record. Each structural feature reflects cells changing at the correct time and in the correct place. The 2026 silkworm atlas offers a way to read that record with unusual precision [6]. It presents development not as a single transformation, but as a sequence of coordinated events through which a larval disc becomes a structure capable of flight.

Conclusion

The silkworm wing-disc atlas described by Liu et al. presents development as a staged process organized across both time and space [6]. By combining single-cell analysis, spatial transcriptomics, hormone perturbation, and targeted functional validation, the study shows how a larval tissue is gradually transformed into an adult structure through coordinated cell-state transitions [6]. Its most important conclusion is that 20-hydroxyecdysone does not merely accompany metamorphosis. It also helps regulate the pace of developmental change [6,7,10–12]. Equally important, the study identifies Wm cells as central elements in this process and highlights transcriptional regulators such as Rfx, Blimp-1, Dll, and Pur-alpha as candidate drivers of fate resolution [6]. When read together with earlier Bombyx work on fringe, flugellos, and stage-specific ecdysone-responsive promoters, the paper also appears as part of a coherent developmental framework rather than a stand-alone atlas [8–11]. A completed wing is therefore more than a final form. It is the result of timed cellular decisions made in the correct spatial context and under the influence of appropriate signals.

References

Kango-Singh M, Singh A, Gopinathan KP. The wings of Bombyx mori develop from larval discs exhibiting an early differentiated state: a preliminary report. Journal of Biosciences. 2001;26(2):167–177. doi:10.1007/BF02703641.
Tripathi BK, Irvine KD. The wing imaginal disc. Genetics. 2022;220(4):iyac020. doi:10.1093/genetics/iyac020.
Everetts NJ, Worley MI, Yasutomi R, Yosef N, Hariharan IK. Single-cell transcriptomics of the Drosophila wing disc reveals instructive epithelium-to-myoblast interactions. eLife. 2021;10:e61276. doi:10.7554/eLife.61276.
Marx V. Method of the Year: spatially resolved transcriptomics. Nature Methods. 2021;18:9–14. doi:10.1038/s41592-020-01033-y.
You Y, Fu Y, Li L, et al. Systematic comparison of sequencing-based spatial transcriptomic methods. Nature Methods. 2024;21:1743–1754. doi:10.1038/s41592-024-02325-3.
Liu Q, He M, Chen H, et al. Single-cell and spatial transcriptomics define 20E-driven developmental reprogramming in silkworm wing disc. Nature Communications. 2026. doi:10.1038/s41467-026-69518-6.
Ou J, Deng HM, Zheng SC, et al. Transcriptomic analysis of developmental features of Bombyx mori wing disc during metamorphosis. BMC Genomics. 2014;15:820. doi:10.1186/1471-2164-15-820.
Fujiwara H, Hojyo T. Developmental profiles of wing imaginal discs of flügellos (fl), a wingless mutant of the silkworm, Bombyx mori. Development Genes and Evolution. 1997;207(1):12–18. doi:10.1007/s004270050087.
Sato K, Matsuoka Matsunaga T, Futahashi R, et al. Positional cloning of a Bombyx wingless locus flügellos (fl) reveals a crucial role for fringe that is specific for wing morphogenesis. Genetics. 2008;179(2):875–885. doi:10.1534/genetics.107.082784.
Kawasaki H, Iwanaga M, Wang HB. Stage-specific activation of the E74B promoter by low ecdysone concentrations in the wing disc of Bombyx mori. Gene. 2014;537(2):322–327. doi:10.1016/j.gene.2013.12.002.
Wang HB, Moriyama M, Iwanaga M, Kawasaki H. Ecdysone directly and indirectly regulates a cuticle protein gene, BMWCP10, in the wing disc of Bombyx mori. Insect Biochemistry and Molecular Biology. 2010;40(6):453–459. doi:10.1016/j.ibmb.2010.04.004.
Mirth CK, Truman JW, Riddiford LM. The ecdysone receptor controls the post-critical weight switch to nutrition-independent differentiation in Drosophila wing imaginal discs. Development. 2009;136(14):2345–2353.
Tobler A, Nijhout HF. A switch in the control of growth of the wing imaginal disks of Manduca sexta. PLoS ONE. 2010;5(5):e10723. doi:10.1371/journal.pone.0010723.
Prakash A, Dion E, Das Banerjee T, Monteiro A. The molecular basis of scale development highlighted by a single-cell atlas of Bicyclus anynana butterfly pupal forewings. Cell Reports. 2024;43:114147. doi:10.1016/j.celrep.2024.114147.
Ling L, Ge X, Li Z, et al. MiR-2 family targets awd and fng to regulate wing morphogenesis in Bombyx mori. RNA Biology. 2015;12(7):742–748. doi:10.1080/15476286.2015.1048957.
Wang Y, Zhou L, Liang W, et al. Cytokine receptor DOME controls wing disc development in Bombyx mori. Insect Biochemistry and Molecular Biology. 2022;148:103828. doi:10.1016/j.ibmb.2022.103828.
Loh LS, DeMarr KA, Tsimba M, et al. Lepidopteran scale cells derive from sensory organ precursors through a canonical lineage. Development. 2025;152(5):dev204501. doi:10.1242/dev.204501.
Loh LS, Hanly JJ, Carter A, et al. Single-nucleus transcriptomics of wing sexual dimorphism and scale cell specialization in sulphur butterflies. PLoS Biology. 2025;23(6):e3003233. doi:10.1371/journal.pbio.3003233.