Abstract

   The cell transition from an inflammatory phase to a subsequent
   proliferative phase is crucial for wound healing, yet the driving
   mechanism remains unclear. By profiling lncRNA expression changes
   during human skin wound healing and screening lncRNA functions, we
   identify SNHG26 as a pivotal regulator in keratinocyte progenitors
   underpinning this phase transition. Snhg26-deficient mice exhibit
   impaired wound repair characterized by delayed re-epithelization
   accompanied by exacerbated inflammation. Single-cell transcriptome
   analysis combined with gain-of-function and loss-of-function of SNHG26
   in vitro and ex vivo reveals its specific role in facilitating
   inflammatory-to-proliferative state transition of keratinocyte
   progenitors. A mechanistic study unravels that SNHG26 interacts with
   and relocates the transcription factor ILF2 from inflammatory genomic
   loci, such as JUN, IL6, IL8, and CCL20, to the genomic locus of LAMB3.
   Collectively, our findings suggest that lncRNAs play cardinal roles in
   expediting tissue repair and regeneration and may constitute an
   invaluable reservoir of therapeutic targets in reparative medicine.

   Subject terms: Skin stem cells, Long non-coding RNAs, Trauma
     __________________________________________________________________

   The mechanism driving inflammatory-to-proliferative state transition
   during wound healing remains unclear. Here, the authors discover that
   SNHG26 interacts with ILF2, redirecting it from inflammatory genomic
   loci to the LAMB3 locus, thereby promoting wound healing.

Introduction

   Tissue repair requires three sequential and overlapping phases, i.e.,
   inflammation, proliferation, and remodeling^[58]1,[59]2. The transition
   from the inflammatory to the proliferative phase represents a key step
   during wound healing, as failure of this transition leads to impaired
   re-epithelialization accompanied by persistent inflammation, which has
   been recognized as a pathological hallmark of chronic
   wounds^[60]1,[61]3–[62]5. Therefore, gaining a deeper understanding of
   the mechanisms that drive this inflammation-proliferation phase
   transition may lead to innovative approaches to expedite tissue repair.

   Keratinocytes constitute the major cell type of the epidermis and play
   important roles in both the inflammatory and proliferative phases of
   wound healing. Upon skin injury, keratinocytes are activated and
   release a diverse range of inflammatory mediators that are crucial for
   the recruitment and activation of immune cells to the wound
   sites^[63]6–[64]8. In the following proliferative phase, the immune
   response of keratinocytes diminishes, and keratinocytes, especially
   stem/progenitor cells, play central roles in wound re-epithelialization
   via proliferation and migration^[65]9. The cell function switch from
   inflammation to proliferation or migration requires reprogrammed gene
   expression; however, the molecular machinery steering this phase
   transition is not well understood.

   Long noncoding RNAs (lncRNAs) are expanding the horizon in
   understanding the regulatory logic encoded in our genome^[66]10.
   Depending on their subcellular localization and specific interactions
   with DNA, RNA, and proteins, lncRNAs have been shown to modulate
   chromatin structure and function as well as the transcription of
   neighboring and distant genes and to affect RNA splicing, stability,
   and translation^[67]11. Notably, several lncRNAs have been shown to
   regulate keratinocyte progenitor proliferation (e.g., PRANCR^[68]12)
   and differentiation (e.g., TINCR and ANCR^[69]13,[70]14). We aimed to
   understand the roles of lncRNAs in the inflammatory-to-proliferative
   phase transition and identify lncRNA(s) crucial in wound repair.

   In this study, we tracked gene expression in human skin wound healing
   and identified 20 evolutionarily conserved lncRNAs with significant
   changes during the inflammatory and proliferative phases. Among these,
   small nucleolar RNA host gene 26 (SNHG26), a lncRNA derived from snoRNA
   host genes (SNHGs), emerged as a key regulator facilitating the
   transition from inflammation to proliferation in keratinocyte
   progenitors. Mechanistically, SNHG26 interacts with ILF2, shifting this
   transcription factor (TF) from inflammatory genomic loci to the LAMB3
   locus, thereby facilitating the transcriptional program switch from
   inflammation to re-epithelialization.

Results

LncRNA expression profiling and functional screening in human skin wound
healing

   To probe gene expression during human skin wound healing, we created
   full-thickness excisional wounds on the skin of healthy volunteers
   (n = 26) and then collected wound-edge tissues from the same donors at
   each healing stage, i.e., inflammation (day 1 wounds, NW1),
   proliferation (day 7 wounds, NW7), and remodeling (day 30 wounds, NW30)
   (Fig. [71]1a and Supplementary Data [72]1). We conducted
   whole-transcriptome analysis using polyadenylation-independent total
   RNA sequencing (RNA-seq) on matched skin, NW1, and NW7 samples from
   five donors, identifying 342 lncRNAs with significantly altered
   expression during wound repair (one-way ANOVA, p ≤ 0.005) (Fig. [73]1a
   and Supplementary Data [74]2). We focused on 20 lncRNAs conserved in
   humans and rodents^[75]15,[76]16, as evolutionary conservation suggests
   crucial functional roles^[77]17. After transfection of Lincode siRNAs
   targeting each of these lncRNAs, we assessed their impact on the
   inflammatory response, proliferation, and migration of human
   keratinocyte progenitors (Fig. [78]1b). Silencing SNHG26, RMRP, MALAT1,
   CRNDE, or SNHG6 increased TNFα-induced production of the inflammatory
   cytokines IL6 and CCL20. Moreover, downregulating SNHG26, SNHG6, or
   LOC101929709 decreased cell proliferation, while silencing SNHG26,
   WDFY3-AS2, CRNDE, SOX9-AS1, or GAS5 reduced keratinocyte migration.
   Notably, among these 20 conserved lncRNAs, knockdown of SNHG26
   exhibited the most pronounced proinflammatory, antiproliferative, and
   antimigratory effects on keratinocytes.

Fig. 1. LncRNA expression profiling and functional screening in human skin
wound healing.

   [79]Fig. 1
   [80]Open in a new tab

   a Profiling human skin lncRNA expression changes during wound healing
   by RNA-seq. b Functional screening of 20 differentially expressed and
   conserved lncRNAs in keratinocyte progenitors for cell proliferation,
   migration, and TNFα-induced inflammatory cytokine expression. c Human
   and mouse SNHG26 genomic loci. d, e qRT‒PCR analysis of SNHG26
   expression in human (n = 6 in Skin and NW1, n = 5 in NW7 and NW30) (d)
   and mouse acute wounds (n = 5) (e). f qRT-PCR analysis of SNHG26 in
   epidermal (EPI) CD45− and CD45+ cells and in dermal (Derm) CD90 + ,
   CD14 + , and CD3+ cells, which were isolated from the skin and day-7
   wounds of healthy donors (n = 5). g Visualization of SNHG26 expression
   in human acute wounds analyzed by spatial transcriptomics. h In situ
   hybridization analysis of SNHG26 in human skin and day-7 acute wounds.
   The asterisk (*) indicates the wound edge. Nuclei were stained with
   DAPI. Scale bar, 25 μm. i, j Northern blot analysis i and In situ
   hybridization analysis j of SNHG26 in cultured human keratinocyte
   progenitors (KC). Scale bar, 10 μm. k qRT-PCR analysis of SNHG26,
   MALAT1, and HPRT in nuclear (blue) and cytosolic (red) compartments of
   human keratinocyte progenitors (n = 3 biological replicates). Data are
   shown as mean ± SD (d–f, k). The data were analyzed by one-way ANOVA
   analysis (d, e) or unpaired two-tailed Student’s t test (f). Panel a
   was created with BioRender.com and released under a Creative Commons
   Attribution-NonCommercial-NoDerivs 4.0 International license
   ([81]https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en).

Characterization of SNHG26 expression in skin wound healing

   SNHG26 is an intergenic lncRNA encoded on human chromosome 7
   (GRCh37/hg19, chr7: 22,893,745-22,901,055) and mouse chromosome 5
   (GRCm38/mm10, chr5:23,850,597-23,855,038, annotated as 2700038G22Rik)
   (Fig. [82]1c, Supplementary Fig. [83]1a and [84]1b). Comparison of the
   RNA sequence of human and mouse SNHG26 showed poor sequence similarity
   (Supplementary Fig. [85]2). However, both the human and mouse SNHG26
   genes are located between the TOMM7 and FAM126A coding genes,
   suggesting interspecies syntenic conservation (Fig. [86]1c,
   Supplementary Fig. [87]1a and [88]1b).

   To investigate the expression of SNHG26 during human and mouse wound
   healing, we collected samples from the wound edges at various stages of
   healing including inflammation, proliferation, and remodeling phase.
   QRT-PCR analysis revealed that inflammation (S100A7 and S100A8),
   proliferation (MKI67), and remodeling (COL1A1 and COL3A1), which
   expression levels peak at the NW1, NW7, and NW30, respectively,
   supporting the associations between selected time points and phases of
   human wound healing (Supplementary Fig. [89]3a). Similarly, in the
   mouse wound model, we showed peaked expression of inflammatory genes
   (Il6 and Il1a), proliferative gene (Mki67), and remodeling process
   related gene (Eln) in NW3, NW7, and NW10 wounds, respectively, linking
   these time points with corresponding healing stage (Supplementary
   Fig. [90]3b–e). In both humans and mice, SNHG26 expression temporarily
   increased during the inflammatory and proliferative phases of wound
   healing, returning to baseline levels during remodeling phase
   (Fig. [91]1d, e). To identify the primary cell type(s) expressing
   SNHG26 in human skin and wounds, we isolated epidermal CD45− cells
   (primarily composed of keratinocytes) and CD45+ cells (leukocytes),
   dermal CD90+ cells (fibroblasts), CD14+ cells (macrophages), and CD3+
   cells (T cells) from matched skin and day 7 wounds of five healthy
   donors (Supplementary Data [92]1). SNHG26 was found in various skin
   cell types, but keratinocytes showed the most significant upregulation
   during wound healing (Fig. [93]1f). This observation was further
   validated by spatial transcriptomics (Fig. [94]1g, Supplementary
   Fig. [95]3f, g) and fluorescence in situ hybridization analysis (FISH,
   Fig. [96]1h) of human acute wounds, indicating that the injury-induced
   transient upregulation of SNHG26 expression predominantly occurred in
   wound-edge basal keratinocytes. Notably, the SNHG26 gene locus also
   contains a small nucleolar RNA SNORD93 in both humans and mice
   (Fig. [97]1c, Supplementary Fig. [98]1a and b), but only SNHG26 showed
   increased expression in keratinocytes during wound repair
   (Supplementary Fig. [99]3h and i). Moreover, SNHG26 expression was
   reduced in human diabetic foot ulcers and diabetic mouse wounds
   compared to acute wounds (Supplementary Fig. [100]3j–l), highlighting
   its relevance in wound healing and chronic wound pathology.

   We further characterized the molecular features of SNHG26 in
   keratinocyte progenitors. The Cap Analysis of Gene Expression (CAGE)
   project^[101]18 previously located the transcription start site (TSS)
   of SNHG26 at GRCh37/hg19 chr7:22893816-22893881 (Fig. [102]1c). Our 3′
   rapid amplification of complementary DNA ends (RACE) analysis indicated
   that the SNHG26 sequence matched the transcript ENST00000415611.9 from
   Ensembl, with a length of 3100 nt (Fig. [103]1c). This finding was
   further validated through Northern blot analysis (Fig. [104]1i).
   Furthermore, FISH and nucleus-cytoplasm fractionation assay revealed
   the nuclear localization of SNHG26 in both cultured keratinocyte
   progenitors and basal keratinocytes in human skin (Fig. [105]1h, j, k).
   Moreover, by separating the poly(A)^+ fraction from the poly(A)^− RNA
   fraction, we found SNHG26 to be a polyadenylated RNA (Supplementary
   Fig. [106]3m). With the Coding Potential Calculator 2
   algorithm^[107]19, we confirmed the non-coding potential of SNHG26
   (Supplementary Fig. [108]3n). Additionally, we determined that the
   half-life of SNHG26 in keratinocyte progenitors was 2.2 hours
   (Supplementary Fig. [109]3o). In summary, SNHG26 is a conserved,
   polyadenylated, nuclear lncRNA transiently upregulated in wound-edge
   basal keratinocytes.

Impaired wound healing in Snhg26-deficient mice

   Considering the potential role of SNHG26 in regulating keratinocyte
   inflammation, proliferation, and migration (Fig. [110]1b), we proceeded
   to investigate whether SNHG26 upregulation is indispensable for the
   wound healing process. To examine this, we generated Snhg26-knockout
   (KO) mice by deleting the 8 kb genomic locus (Supplementary
   Fig. [111]4a, b). These KO mice exhibited a generally normal phenotype
   with a slightly thinner epidermis (Supplementary Fig. [112]4c).
   However, upon skin injury, we observed a significant 40% delay
   (p = 0.0001) in wound healing in the Snhg26-KO mice compared to control
   mice (Fig. [113]2a). Recognizing that compensatory mechanisms might be
   at play in this model with constitutive Snhg26 deletion, we also
   intradermally injected Snhg26 antisense oligos (Snhg26-ASOs) at wound
   edges in wild-type (WT) mice (Supplementary Fig. [114]4d–f). These ASOs
   were designed to block the injury-induced upregulation of Snhg26 by
   facilitating RNaseH1-dependent cleavage of Snhg26 transcripts^[115]20.
   Notably, these ASOs effectively reduced SNHG26 expression in the
   epidermal compartment, while the dermal compartment remained unchanged
   (Supplementary Fig. [116]4f). Similar to the Snhg26-KO mice, mice
   treated with Snhg26-ASOs exhibited a 37.5% delay (p = 0.003) in wound
   healing compared to those receiving scrambled ASOs, emphasizing the
   crucial role of Snhg26 transcripts in wound repair (Fig. [117]2b).

Fig. 2. Impaired wound healing in Snhg26-deficient mice.

   [118]Fig. 2
   [119]Open in a new tab

   a, b Representative wound images and wound closure rates of Snhg26-KO
   mice (n = 5) vs. WT mice (n = 6) (a) and C57BL/6 mice treated with
   Snhg26-ASOs or control ASOs (n = 5/group) (b). Wound healing rate
   constant κ (day−1) was calculated using a one-phase decay model in
   GraphPad. c Gene Ontology analysis of differentially expressed genes in
   the wound-edge epidermis of Snhg26-KO mice vs. WT mice. d Gene set
   enrichment analysis of wound-edge epidermal gene expression changes in
   Snhg26-KO mice vs. WT mice. Shown are the false discovery rate (FDR)
   value and normalized enrichment score (NES). e qRT‒PCR analysis of
   Krt16 and Sprr1b expression in the skin and wound-edge epidermis of
   Snhg26-KO and WT mice (n = 5). f Hematoxylin and eosin staining of
   Snhg26-KO and WT mouse wounds. Areas of the wound tongue epithelium
   (marked with dotted lines) were quantified (n = 8). Scale bar, 200 μm.
   g qRT‒PCR analysis of Il6 and Il1b expression in mouse skin and
   wound-edge epidermis (n = 5). h Immunofluorescence analysis of CD68^+
   cells in mouse wounds (n = 4). The white dashed line demarcates the
   epidermis and dermis border. Scale bar, 20 μm. Data are shown as
   mean ± SD from two to three independent experiments (a, b, e–h). The
   data were analyzed by two-way ANOVA (a, b, e and g), two- sided
   Fisher’s exact test (c) or two-tailed Student’s t test (f and h).

   To uncover the gene expression changes linked to reduced healing in
   Snhg26-KO mice, we compared epidermal gene expression between KO
   (n = 3) and WT mice (n = 3). Microarray analysis identified 253 and 431
   differentially expressed genes (DEGs; fold-change ≥2 or ≤ -2, p
   value < 0.05) in the Snhg26-KO mice skin and day 3 wound edge,
   respectively (Supplementary Fig. [120]4g). Gene Ontology (GO) analysis
   revealed that the downregulated genes in Snhg26-KO skin epidermis were
   mainly associated with epidermal differentiation, while no enriched GO
   term was found among the upregulated DEGs (Supplementary Fig. [121]4h).
   In the wound-edge epidermis, Snhg26-KO mice exhibited upregulated
   inflammation-related genes and downregulated genes linked to skin
   development and cell migration (Fig. [122]2c). Gene set enrichment
   analysis (GSEA) showed that upregulated genes were associated with
   leukocyte migration, while downregulated genes were linked to
   epithelial cell proliferation and migration in the wound-edge epidermis
   of Snhg26-KO mice (Fig. [123]2d).

   We also noted reduced expression of Krt16 and Sprr1b in the wound-edge
   epidermis of Snhg26-KO mice compared to controls, as confirmed by
   qRT‒PCR analysis of additional skin and wound epidermal samples
   (Fig. [124]2e). Both Krt16 and Sprr1b are known for their roles in
   keratinocyte proliferation and migration^[125]21,[126]22. The decreased
   expression of these genes could potentially contribute to delayed
   re-epithelialization, as supported by histomorphometry analysis of
   newly formed epithelial tongues (Fig. [127]2f). Furthermore, we
   observed significant upregulation of proinflammatory cytokines,
   including Il6 and Il1b, in the wound-edge epidermis of Snhg26-KO mice
   (Fig. [128]2g) and increased macrophages (CD68 + ) in the dermal wound
   bed of Snhg26-KO mice (Fig. [129]2h). Similarly, mice treated with
   Snhg26-ASOs exhibited elevated levels of inflammatory cytokines (Il6,
   Ccl20, Ccl2, and Ccl5) and increased macrophages (CD68 + ) in the
   wounds (Supplementary Fig. [130]5).

   In summary, our study, utilizing Snhg26-KO mice and mice with transient
   Snhg26 inhibition at wound edges, provides conclusive evidence of
   Snhg26’s in vivo role in wound repair. Its deficiency not only delayed
   re-epithelialization but also exacerbated inflammatory response.

Single-cell transcriptome analysis of delayed wound repair in Snhg26-KO mice

   To gain deeper insights into the delayed wound repair in Snhg26-KO
   mice, we conducted single-cell RNA sequencing (scRNA-seq) analysis on
   skin and day 3-wound tissues from Snhg26-KO mice (n = 3) and control
   littermates (n = 3). Unsupervised clustering identified 23 cell
   clusters, including keratinocytes (basal stem and progenitors: C1-C4;
   spinous: C5; granular: C6; hair follicle keratinocytes: C7-C10),
   melanocytes (C11), fibroblasts (C12-C14), sebaceous gland cells (C15),
   vascular endothelial cells (C16-C17), muscle cells (C18-C19),
   macrophages (C20), T helper cells (C21), γδ T cells (C22), and
   Langerhans cells (C23) (Fig. [131]3a, Supplementary Fig. [132]6a). Each
   of these cell clusters contained cells from all the investigated
   samples (Supplementary Fig. [133]6b). Notably, our analysis revealed
   that Snhg26 was predominantly expressed in keratinocytes (Fig. [134]3b,
   c). While the cellular composition remained similar between the KO and
   WT mouse skin samples, notable differences were observed in the basal
   keratinocytes and macrophages in the wounds of Snhg26-KO mice compared
   to WT mice (Fig. [135]3d). We also calculated the number of DEGs in
   each cell cluster between Snhg26-KO and WT mice and found that
   keratinocytes showed the highest number of DEGs between the two groups
   (Fig. [136]3e), suggesting the critical role of SNHG26 in keratinocytes
   during wound healing.

Fig. 3. Single-cell transcriptomic analysis of impaired wound repair in
Snhg26-KO mice.

   [137]Fig. 3
   [138]Open in a new tab

   a UMAP plot and cluster annotation of 21491 cells from skin and wounds
   of Snhg26-KO and WT mice (n = 3/group). b UMAP plot showing the
   expression of Snhg26 in Snhg26-KO and WT mice. c Violin plot showing
   the expression of Snhg26 in different cell types in WT mice. d
   Comparison of skin and wound cell composition between the KO and WT
   mice. Red line indicates keratinocyte clusters with changed cell
   composition in the wound of KO mice. Blue line indicates immune cell
   clusters. e The proportion of DEGs [log2(FC) ≥ 0.3, -log10(FDR) > 2]
   within all expressed genes in individual cell clusters in wounds of
   Snhg26-KO and WT mice. f GO and KEGG analysis of the top 200 marker
   genes in basal keratinocyte cell clusters. g The proliferation,
   migration, and inflammatory scores were plotted for each keratinocyte
   cluster. h–j Violin plots showing keratinocyte migration (h),
   proliferation (i) and inflammation scores (j) in Snhg26-KO and WT mice
   wound basal keratinocytes. k Circle plots displaying the CCL signaling
   network in Snhg26-KO and WT mice wounds. The changed signals in the KO
   mice compared to the WT mice were highlighted with arrows. The data
   were analyzed by two- sided post hoc test (c), two- sided Fisher’s
   exact test (f) or two-sided Mann-Whitney U test (h–j).

   Consistent with a recent study on mouse wounds using scRNA-seq^[139]23,
   our results identified four basal stem and progenitor cell clusters: C1
   expressed Col17a1, a marker enriched in epidermal stem cells; C2
   consisted of proliferative basal cells expressing mitosis-related
   genes, such as Pcna and Mki67; C3 showed enrichment in
   glycolysis-related genes, important for keratinocyte migration^[140]24,
   and specifically expressed the keratinocyte activation markers Krt6b
   and Krt16^[141]25; and C4 expressed genes promoting cell cycle arrest
   (e.g., Ovol1) and inflammation (e.g., the TFs Fosl1 and
   Rel)^[142]26,[143]27 (Fig. [144]3f, Supplementary Fig. [145]7a, b,
   Supplementary Data [146]3). We further performed immunostaining for
   these four different basal cell populations in the wound edge: KRT15
   (Bas1), MKI67 (Bas2), GJB2 (Bas3), and KLF5 (Bas4). All four cell
   populations were successfully detected in the wound edge epidermis
   (Supplementary Fig. [147]7c). Functional characteristics of these
   keratinocyte progenitor clusters, including the proliferative C2,
   migratory C3, and inflammatory C4 progenitors, were more evident when
   we compared overall gene expression scores for these processes
   (Fig. [148]3g, Supplementary Data [149]4). Analyzing the functional
   scores of individual keratinocytes, we observed fewer proliferative or
   migratory keratinocytes but a higher number of inflammatory
   keratinocytes in the wounds of Snhg26-KO mice compared to WT mice
   (Fig. [150]3h–j). This observation was supported by GO analysis of the
   DEGs in keratinocyte progenitors between the Snhg26-KO and WT mice
   (Supplementary Fig. [151]8a–c). Furthermore, a cell-cell crosstalk
   analysis^[152]28 highlighted enhanced CCL and CXCL chemokine signaling
   from basal keratinocytes to macrophages in Snhg26-KO wounds
   (Fig. [153]3k, Supplementary Fig. [154]8d). These findings align with
   the more numbers of macrophages detected in these wounds by scRNA-seq
   (Fig. [155]3d) and IF staining (Fig. [156]2h). Moreover, we found that
   the expression of Krt16 and Sprr1b was significantly decreased, while
   the expression of Il6 and Il1b was increased in the keratinocytes of KO
   mice wounds compared with WT mice wounds (Supplementary
   Fig. [157]8e–h), which is consistence with the results as we detected
   by qRT-PCR (Fig. [158]2e and [159]g).

   Single-cell analysis allowed us to precisely discern the impact of
   Snhg26-KO in various skin cell types. In dermal fibroblasts, which play
   critical roles in skin wound healing, especially in modulating the
   inflammatory response^[160]29, we observed 198 DEGs (Fig. [161]3e). The
   downregulated genes in the Shng26-KO fibroblasts were associated with
   the TGF-β regulation of extracellular matrix, while the upregulated
   genes were related to TNF-α signaling via NF-kB pathway (Supplementary
   Fig. [162]8i). Cell-cell crosstalk analysis further confirmed the role
   of fibroblasts in wound inflammation. Specifically, fibroblasts (Fb1
   and Fb2) were identified as the primary cell types sending CCL and CXCL
   signals in mouse wounds (Fig. [163]3k and Supplementary Fig. [164]8d).
   In Snhg26-KO wounds, fibroblasts exhibited heightened CCL signals
   directed towards keratinocytes (Fig. [165]3k). Additionally, there was
   a decreased laminin signal from keratinocytes to fibroblasts in the
   Snhg26-KO mice (Supplementary Fig. [166]8j).

   In summary, Snhg26 disruption skewed keratinocyte progenitors towards
   an inflammatory state and away from proliferation or migration. While
   Snhg26 had a lesser impact on dermal fibroblasts, the changes in
   keratinocyte-fibroblast communication collectively contributed to
   delayed wound healing.

SNHG26 inhibits the human keratinocyte inflammatory response and promotes
re-epithelialization

   Next, we explored the physiological relevance of SNHG26 in human
   keratinocytes and skin wound healing. We silenced SNHG26 expression in
   human keratinocyte progenitors by transfecting SNHG26-ASOs followed by
   TNFα treatment to trigger an inflammatory response^[167]30. This ASO
   was designed to target the exons of SNHG26, and it led to a significant
   reduction in SNHG26 expression while leaving SNORD93 unaffected
   (Supplementary Fig. [168]9a, b). Microarray analysis of these cells
   revealed gene expression changes (ANOVA, P < 0.001) grouped into five
   patterns (Fig. [169]4a). Module 2 (M2) contained 898 upregulated genes
   related to inflammation. M3 comprised 123 TNFα-induced genes that were
   further upregulated by SNHG26 knockdown. M4 and M5 included 1157 and
   918 downregulated genes related to mitosis.

Fig. 4. SNHG26 inhibits the human keratinocyte inflammatory response and
promotes re-epithelialization.

   [170]Fig. 4
   [171]Open in a new tab

   a Transcriptome profiling of human keratinocyte progenitors with SNHG26
   knocked down followed by TNFα treatment. GO analysis was performed for
   genes in each expression module (right). b, c qRT‒PCR analysis of IL6,
   IL8, and CCL20 expression in human keratinocyte progenitors after
   SNHG26 knockdown (n = 4) (b) or overexpression (n = 3) (c) followed by
   TNFα treatment. d, e Scratch wound assay (n = 4) (d) and CyQUANT cell
   proliferation assay (n = 6) (e) of human keratinocyte progenitors with
   SNHG26 knockdown. Wound healing rate constant κ (hour−1) in (d) was
   calculated using a one-phase decay model in GraphPad. f, g Scratch
   wound assay (n = 5) (f) and colony formation assay (n = 3) (g) of human
   keratinocyte progenitors with SNHG26 overexpression. h, i Hematoxylin
   and eosin staining and quantification of human ex vivo wounds with
   topical application of SNHG26-ASOs (n = 3). Black arrows: the initial
   wound edges. Red arrows: newly formed epidermis in the front (day 5).
   Scale bar, 200 μm. j, k Representative images (top view) of ex vivo
   wounds stained with CellTracker™ Green CMFDA Dye (j) and quantification
   (n = 3) (k). The wounds were topically treated with SNHG26
   overexpression plasmids. Dashed white lines indicate the initial wound
   edges, and dashed red lines indicate newly formed epidermis. Data are
   shown as mean ± SD from two to three independent experiments (b–k). The
   data were analyzed by two- sided Fisher’s exact test (a), two-way ANOVA
   analysis (b–d, f, i, and k), or two-tailed Student’s t test (e and g).

   We confirmed expression changes in genes associated with inflammation,
   including IL6, IL8, and CCL20, which are induced in response to skin
   injury and play crucial roles in the inflammatory
   process^[172]31,[173]32. Silencing SNHG26 increased their expression
   under both basal and TNF-α-triggered inflammatory conditions
   (Fig. [174]4b). Conversely, nuclear overexpression of SNHG26 using a
   pZW1-snoVector reduced TNF-α-induced IL6, IL8, and CCL20 expression
   (Fig. [175]4c, Supplementary Fig. [176]9c)^[177]33. However,
   cytoplasmic overexpression of SNHG26 using a pcDNA3.1 vector had no
   discernible effect on these inflammatory genes (Supplementary
   Fig. [178]9d–g), highlighting the importance of subcellular
   localization in SNHG26’s function.

   Furthermore, we substantiated our findings by confirming that the
   silencing of SNHG26 reduced, whereas overexpression of SNHG26 increased
   the migration and proliferation of human keratinocyte progenitors. This
   was evident in scratch wound assays (Fig. [179]4d, f), cell
   proliferation assays (Fig. [180]4e), and colony formation assays
   (Fig. [181]4g, Supplementary Fig. [182]9h). Next, we extended our
   analysis to human ex vivo wound closure, which is an in vivo-like and
   clinically relevant model for analyzing human wound
   re-epithelization^[183]34,[184]35. We topically applied SNHG26-ASOs
   (Fig. [185]4h, i, Supplementary Fig. [186]9i, j) or its overexpression
   plasmid sno-SNHG26 (Fig. [187]4j, k) mixed in a lipid-based
   transfection reagent to the wounds. This intervention specifically
   modulated epidermal SNHG26 levels, leaving dermal levels unaffected
   (Supplementary Fig. [188]9k). We found that SNHG26-ASOs reduced wound
   re-epithelialization, while overexpression of SNHG26 enhanced this
   process (Fig. [189]4h–k).

   In addition, we conducted experiments to distinguish the roles of
   SNHG26 and SNORD93, both sharing the same genomic locus (Fig. [190]1c).
   Using CRISPR-Cas9 technology, we generated a SNHG26 knockout in
   hTERT-immortalized human keratinocyte cell line (Ker-CT) (Supplementary
   Fig. [191]9l, m). Subsequently, we reinstated the expression of either
   SNORD93 or SNHG26 through the transfection of their respective
   overexpressing plasmids (Supplementary Fig. [192]9n, o). Notably,
   SNORD93 had no impact on inflammatory responses or migration, whereas
   SNHG26 reintroduction exhibited anti-inflammatory and pro-migratory
   effects, confirming the specific role of SNHG26 (Supplementary
   Fig. [193]9p-s).

   Collectively, our mouse and human data jointly demonstrated the
   evolutionary conservation of SNHG26 in both expression and function: it
   is an injury-induced lncRNA in keratinocyte progenitors and plays
   pivotal roles in inhibiting cell inflammatory response while promoting
   cell proliferation and migration and wound re-epithelialization.

Silencing SNHG26 increase binding of ILF2 to inflammatory genomic loci

   Long non-coding RNAs (lncRNAs) regulate gene expression through both
   cis and trans mechanisms. To study if SNHG26 regulate the genes near
   its genomic locus, we investigated the expression of SNHG26-neighboring
   genes, Tomm7 and Fam126a, in both WT and SNHG26-KO mice under
   hemostatic and wounding conditions. We observed no significant changes
   in the expression levels of these genes between the WT and SNHG26-KO
   mice (Supplementary Fig. [194]10a, b). Similarly, knockdown of SNHG26
   in primary keratinocytes did not affect the expression of TOMM7 and
   FAM126A (Supplementary Fig. [195]10c, d). Thus, SNHG26 may exert it
   regulatory function by in trans.

   To elucidate SNHG26’s molecular mechanism, we performed RNA pull-down
   assay in human keratinocyte progenitors (Fig. [196]5a). An SDS‒PAGE
   analysis revealed a protein band at ~45 kDa specifically copurified
   with SNHG26 (Fig. [197]5b). The mass spectrometry (MS) analysis
   identified multiple proteins (Supplementary Data [198]5). We excluded
   keratins from this list, as they are commonly regarded as contaminants
   in MS studies^[199]36. We were particularly interested in proteins
   localized to the nucleus, given that SNHG26 is primarily expressed in
   this compartment. Therefore, we focused on interleukin enhance-binding
   factor 2 (ILF2), which is a transcription factor locating in the
   nucleus. We further confirmed the interaction between SNHG26 and ILF2
   with another set of independent samples by Western blotting
   (Fig. [200]5c). The reciprocal binding of SNHG26 and ILF2 protein was
   validated via RNA immunoprecipitation (RIP) using the anti-ILF2
   antibody, which showed that SNHG26, but not GAPDH, 18S rRNA or MALAT1,
   coprecipitated with ILF2 protein in human keratinocyte progenitors
   (Fig. [201]5d, e).

Fig. 5. Silencing SNHG26 increase binding of ILF2 to inflammatory genomic
loci.

   [202]Fig. 5
   [203]Open in a new tab

   a Schematic illustration showing SNHG26 pull-down assay. b SDS‒PAGE
   protein bands: the highlighted protein bands were analyzed by MS. c WB
   analysis of ILF2 copurified with SNHG26. d Schematic illustration
   showing the RNA immunoprecipitation (RIP) assay. e ILF2 RIP followed by
   qRT‒PCR analysis of copurified RNAs in human keratinocyte progenitors
   (n = 3). f qRT‒PCR analysis of ILF2 expression in human acute wounds
   (n = 7). g Heatmap showing global ILF2 occupancy in SNHG26 knockdown
   (SNHG26-ASO) and control (Ctrl-ASO) conditions. h, i ChIP‒qPCR analysis
   of ILF2 at the promoter of JUN in human keratinocyte progenitors after
   SNHG26 silencing (n = 3) (h) or treatment with TNFα for 0-60 minutes
   (n = 3) (i). j, k qRT‒PCR analysis and western blot of JUN expression
   in human keratinocyte progenitors after ILF2 silencing followed by TNFα
   stimulation (n = 3). l The expression of JUN in the basal keratinocytes
   of Snhg26-KO and WT mice wounds analyzed by scRNA-seq. m EnrichR
   analysis of the enrichment of TFs driving the upregulated genes in the
   wound-edge basal keratinocytes of Snhg26-KO mice compared to that of WT
   mice. n TFs with overrepresented binding sites among the upregulated
   genes in Snhg26-KO mouse wound keratinocytes, as identified by MetaCore
   analysis. o ChIP‒qPCR analysis of ILF2 at the promoters of inflammatory
   genes in human keratinocyte progenitors after TNFα stimulation (n = 3).
   p ELISA analysis of IL6, IL8, and CCL20 proteins produced by human
   keratinocyte progenitors with ILF2 silencing followed by TNFa
   stimulation (n = 3). q ChIP‒qPCR analysis of ILF2 occupation at the
   promoters of inflammatory genes in human keratinocyte progenitors with
   SNHG26 knockdown followed by TNFα stimulation (n = 3). Data are shown
   as mean ± SD from two to three independent experiments (c–e, h–k, o–q).
   The data were analyzed by one-way ANOVA (f, i and o), two-way ANOVA (e,
   j, p and q), two- sided Fisher’s exact test (m) or two-tailed Student’s
   t test (h and l). Panel a and d were created with BioRender.com
   released under a Creative Commons Attribution-NonCommercial-NoDerivs
   4.0 International license
   ([204]https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en).

   Using established methodologies^[205]37–[206]39, our quantitative
   estimations revealed that each human keratinocyte progenitor cell
   contains approximately 1.5-2 copies of SNHG26 transcripts and about
   104,125 ILF2 protein molecules (Supplementary Fig. [207]10e–g). During
   wound healing, SNHG26 expression increases 27.5-fold (Fig. [208]1f),
   resulting in an estimated 41-55 copies of SNHG26 RNA per wound
   keratinocyte cell. Notably, the stoichiometry of SNHG26 RNA to ILF2
   protein remains skewed, ranging from 1:1893 to 1:2540. In line with
   these findings, co-staining analysis showed SNHG26 RNA and ILF2 protein
   co-localized within a condensate-like structure in the cell nucleus,
   but the abundance of ILF2 proteins greatly exceeded that of SNHG26 RNA
   (Supplementary Fig. [209]10h–i). Interesting, we observed that the
   stability of SNHG26 was significantly decreased after ILF2 knockdown
   (Supplementary Fig. [210]10j). This result indicates that ILF2, or the
   nuclear condensates it forms, is crucial for maintaining the stability
   of SNHG26.

   To identify other protein components within these condensates, we
   employed a proximity-dependent biotin identification (BioID) assay,
   which label proteins in close proximity to ILF2 that fused to BioID2,
   and subsequently identifying them through mass spectrometry
   (MS)^[211]40. We successfully overexpressed ILF2-BioID2 fusion protein
   in human keratinocyte progenitors (Supplementary Fig. [212]11a). The MS
   identified 176 proteins with fold change ≥ 2 which may be involved in
   the IF2-interactome (Supplementary Data [213]6). We also excluded
   keratins from this list, as they are commonly regarded as contaminants
   in MS studies^[214]36. Gene enrichments analysis showed that the
   function of these proteins was related to skin development, cell-cell
   junction, and focal adhesion (Supplementary Fig. [215]11b–e).
   Furthermore, to identify the components in the SNHG26-ILF2 nuclear
   condensates, we also overlapped these proteins with the ones
   co-purified with SNHG26 in the RNA pull-down assay, identifying seven
   proteins in common and among them two proteins-ACTB, ANXA2 reside in
   the nucleus (Supplementary Fig. [216]11g). Therefore, our results
   indicated that these two proteins may be the additional components of
   ILF2/SHG26 condensates.

   ILF2 is a constitutively expressed nuclear protein that interacts with
   chromatin^[217]41 and is a TF for both mitotic and inflammatory
   genes^[218]42. It is highly expressed in the human epidermis, as shown
   in the Human Protein Atlas database^[219]43 (Supplementary
   Fig. [220]12a). ILF2 expression was upregulated in the inflammatory
   phase (NW1) of human skin wound healing (Fig. [221]5f, Supplementary
   Fig. [222]12b). To study the effects of SNHG26 knockdown on ILF2
   genomic occupancy, we performed ILF2 chromatin immunoprecipitation
   (ChIP)-sequencing in human keratinocyte progenitors. We identified 664
   peaks (p value < 0.05) enriched in keratinocytes treated with Ctrl-ASO
   and 589 peaks in cells with SNHG26-ASO (Fig. [223]5g, Supplementary
   Data [224]7). Focusing on ILF2 binding sites near gene promoters, we
   found that multiple cell growth-related pathways (e.g., EGFR1 pathway,
   DNA replication pre-initiation, G2/M transition, response to epidermal
   growth factor) were enriched in the Ctrl-ASO group. This indicates that
   ILF2 preferentially binds to cell growth-related gene loci in the
   presence of SNHG26 in keratinocytes (Supplementary Fig. [225]12d).
   However, when SNHG26 was knocked down, ILF2 left these cell
   growth-related gene loci (Fig. [226]5g). We found that MAPK signaling
   and cellular response to cytokine stimulus were enriched in the ILF2
   binding site-related genes in cells with SNHG26-ASO (Supplementary
   Fig. [227]12d).

   Particularly noteworthy is the prominent ILF2 binding locus at the JUN
   promoter, located within a 1-kilobase proximity to the TSS, which
   ranked as one of the top binding sites in cells with SNHG26 knockdown
   (Fig. [228]5g, Supplementary Fig. [229]13a, Supplementary Data [230]7).
   This result was further corroborated through ILF2 ChIP-qPCR analysis
   (Fig. [231]5h). Additionally, a ChIP‒qPCR analysis revealed that TNFα
   treatment triggered ILF2 binding to the JUN promoter within 30 minutes
   (Fig. [232]5i). However, knockdown of ILF2 completely abolished
   TNFα-induced JUN expression, emphasizing the crucial role played by
   ILF2 in driving JUN transcription (Fig. [233]5j, k, Supplementary
   Fig. [234]13b).

   Jun/AP1 is an immediate-early gene and a master TF that controls the
   expression of a wide range of inflammatory genes in epidermal
   cells^[235]44. Upon examining the scRNA-seq data of mouse wounds, we
   found that not only was Jun expression increased (Fig. [236]5l), but
   the genes it regulated were also enriched among the upregulated genes
   in the wound-edge basal keratinocytes of the Snhg26-KO mice
   (Fig. [237]5m). Moreover, Jun emerged as a central hub connected to
   multiple genes involved inflammatory signaling, e.g., Rel, Nfkbia,
   Mapk9, and Map3k8^[238]45–[239]47 in wound keratinocytes of Snhg26-KO
   mice (Supplementary Fig. [240]13c). Additionally, in the mouse
   wound-edge epidermis (Fig. [241]2c, d), c-Jun was identified as the
   Snhg26-regulated TF that controls the most genes with aberrant
   expression after Snhg26-KO (Fig. [242]5n). It’s worth noting that,
   unlike keratinocytes, JUN expression remained unaltered both in human
   fibroblasts transfected with SNHG26-ASO (Supplementary Fig. [243]13d)
   and in Snhg26-KO mice wounds (Supplementary Fig. [244]13e), indicating
   that SNHG26 likely exerts distinct functional mechanisms in fibroblasts
   compared to keratinocytes. TNF-α also induced ILF2 binding to the
   promoters of other immediate early response genes, i.e., IL6, IL8, and
   CCL20 (Fig. [245]5o). As a result, silencing ILF2 led to a reduction in
   their expression in keratinocytes (Fig. [246]5p, Supplementary
   Fig. [247]13f). Intriguingly, SNHG26 knockdown significantly increased
   the binding of ILF2 to the promoters of IL6, IL8, and CCL20
   (Fig. [248]5q).

   Thus, we concluded that SNHG26 interacts with ILF2 and prevents it from
   binding to inflammatory genomic loci, including the master TF JUN and
   several key cytokines and chemokines, consequently suppressing the
   inflammatory response in keratinocyte progenitors. Perturbation of
   SNHG26 impacts ILF2 binding site dynamics, shifting the focus from cell
   growth to inflammation-related processes within keratinocytes.

SNHG26 directs ILF2 protein to the LAMB3 genomic locus

   As SNHG26 block the interaction of ILF2 with the promoters of
   inflammatory genes, we asked whether SNHG26 may interact with chromatin
   and translocate ILF2 to another genomic locus. To assess this
   possibility, we map SNHG26 occupancy genome wide by Chromatin Isolation
   by RNA Purification sequencing (ChIRP-seq) in human keratinocyte
   progenitors^[249]48. This method allowed us to isolate approximately
   22% of SNHG26 RNA in human keratinocyte progenitors (Fig. [250]6a, b).
   Importantly, the SNHG26 probes did not retrieve GAPDH, nor did the LacZ
   probes retrieve SNHG26, confirming the method’s specificity
   (Fig. [251]6b). Through ChIRP-seq, we identified 43 SNHG26 occupancy
   sites across the genome (FDR < 0.05) (Fig. [252]6c, Supplementary
   Data [253]8). These sites were notably enriched within genic regions,
   particularly in regions annotated as promoters and introns
   (Fig. [254]6d). Notably, one of the most prominent SNHG26 RNA occupancy
   sites was located within its own genic region, further confirming the
   specificity of this ChIRP experiment (Fig. [255]6c, Supplementary
   Fig. [256]14a). Further analysis of the ChIRP-seq data using Multiple
   EM for Motif Elicitation (MEME)^[257]49 revealed that SNHG26 exhibited
   a preference for binding to a GA-rich polypurine DNA motif
   (Fig. [258]6e), indicating that SNHG26 accesses the genome through
   specific DNA sequences.

Fig. 6. SNHG26 directs ILF2 protein to the LAMB3 genomic locus.

   [259]Fig. 6
   [260]Open in a new tab

   a Schematic showing SNHG26 ChIRP assay followed by DNA sequencing. b
   qPCR analysis of the percentage of SNHG26 retrieved by its probes,
   GAPDH RNA was undetectable in the ChIRP assay (n = 3 biological
   replicates). c Heatmap showing global SNHG26 ChIRP-seq signal in human
   keratinocyte progenitors. d SNHG26 binding sites are enriched in genic
   regions, notably promoters and introns. e MEME analysis showing GA-rich
   homopurine motif enriched in SNHG26 binding sites. f Genome browser
   tracks showing SNHG26 occupancy at the LAMB3 locus. g ChIP‒qPCR
   analysis of ILF2 at the promoter of LAMB3 in human keratinocyte
   progenitors with SNHG26 silencing (n = 4). h qRT‒PCR analysis of LAMB3
   expression in keratinocyte progenitors with ILF2 silencing (n = 3).
   Data are shown as mean ± SD from one experiment (b) or two to three
   independent experiments (g, h). The data were analyzed by one-way ANOVA
   (g) or two-tailed Student’s t test (h).

   We focused on the LAMB3 gene due to its high-ranking among the top
   SNHG26 ChIRP-seq peaks (Fig. [261]6c, f). Silencing SNHG26 reduced ILF2
   binding to the LAMB3 gene, indicating that SNHG26 directs ILF2 to the
   LAMB3 locus by interacting with chromatin (Fig. [262]6g). Additionally,
   we demonstrated that ILF2 was crucial for driving LAMB3 expression, as
   silencing ILF2 significantly decreased the LAMB3 mRNA level in
   keratinocytes (Fig. [263]6h).

   LAMB3 encodes subunit beta 3 of Laminin 332, which is required for
   keratinocyte attachment to the basement membrane in the epidermis and
   crucial for keratinocyte proliferation and migration^[264]50.
   Consistent with previous research^[265]50, silencing LAMB3
   significantly impaired the migration, proliferation, and long-term
   growth of human keratinocyte progenitors (Fig. [266]7a–e, Supplementary
   Fig. [267]14b). Furthermore, LAMB3 expression increased at the
   proliferative phase of human skin wound healing (NW7), supporting its
   role in wound re-epithelialization (Fig. [268]7f). Our scRNA-seq
   analysis of mouse wounds indicated that Lamb3 was primarily expressed
   in basal keratinocytes and upregulated in wounds compared to the skin
   (Fig. [269]7g,h). However, Snhg26-KO mice exhibited significantly lower
   Lamb3 expression in wound-edge basal keratinocytes compared to WT mice
   (Fig. [270]7h–j). This finding aligns with the delayed
   re-epithelialization observed in Snhg26-KO mice (Fig. [271]2e, f).
   Moreover, we compared ILF2 ChIP-seq and SNHG26 ChIRP-seq results, to
   better understand the functional connection between SNHG26 and ILF2. We
   identified LAMB3, PLEC, LINC02386, and DISP3 as common genes bound by
   SNHG26 and ILF2 (Supplementary Fig. [272]14c). ILF2 ChIP-seq further
   show that SNHG26 KD reduces ILF2 binding to these genes. Interestingly,
   in addition to LAMB3, PLEC and DISP3 have also been reported to promote
   cell proliferation^[273]51,[274]52.

Fig. 7. LAMB3 promotes migration and proliferation of keratinocytes.

   [275]Fig. 7
   [276]Open in a new tab

   a, b Scratch wound assay of human keratinocyte progenitors with LAMB3
   knockdown (n = 4) or control (n = 5) (a) and representative photos are
   shown in (b). c–e CCK-8 cell proliferation assay (n = 6) c and colony
   formation assay (n = 6) d, e of human keratinocyte progenitors after
   LAMB3 knockdown. f LAMB3 expression in human acute wounds from 5 donors
   analyzed by RNA-seq ([277]GSE174661). g, h Violin plots showing Lamb3
   expression in each cluster of keratinocytes (g) and in the Snhg26-KO
   mice vs. WT mice (h) as analyzed by scRNA-seq. i qRT‒PCR analysis of
   Lamb3 expression in the skin and wound-edge epidermis of Snhg26-KO and
   WT mice (n = 5). j Immunofluorescence staining of LAMB3 in the
   wound-edge tissues of WT mice (n = 9) and SNHG26-KO mice (n = 11). The
   asterisk (*) indicates the wound edge. The arrow heads indicate the
   LAMB3+ basal cells. Basal layer keratinocyte expression of LAMBS was
   quantified (right panel). Scale bar, 100 μm. Data are shown as
   mean ± SD from one experiment (f) or two to three independent
   experiments (a–e, i, j). The data were analyzed by one-way ANOVA (f and
   h), two-way ANOVA (a, i), or two-tailed Student’s t test (c, e, j).

   In light of this multifaceted evidence, we propose a mechanism for the
   nuclear lncRNA SNHG26: it interacts with ILF2 and redirects it from
   inflammatory genomic loci, such as JUN, IL6, IL8, and CCL20, to the
   genomic locus of LAMB3, reshaping the gene expression program to
   facilitate the transition of keratinocyte progenitors from an
   inflammatory to a proliferative state.

Discussion

   In this study, we assessed the in vivo expression changes of lncRNAs
   throughout the human skin wound healing process, identifying SNHG26 as
   an evolutionarily conserved lncRNA induced by injury and crucial in
   driving the transition of keratinocyte progenitors from an inflammatory
   to a proliferative state (Supplementary Fig. [278]14d). Using a
   genetically engineered mouse model, we provide conclusive evidence
   regarding the in vivo role of SNHG26 in wound repair. It’s worth noting
   that only a limited number of lncRNA mouse models have been created to
   date, and few have been assessed for their impact on wound
   healing^[279]53. The deficiency of Snhg26 led to a shift in the
   transcriptional program, favoring the upregulation of inflammatory gene
   expression while simultaneously diminishing the expression of genes
   crucial for cell proliferation and migration in the basal keratinocytes
   at the wound edge. Furthermore, through experiments involving human
   keratinocyte progenitors and ex vivo wounds, we corroborated the
   conserved anti-inflammatory, pro-proliferative, and pro-migratory
   functions of SNHG26, confirming the clinical relevance of our findings.

   In terms of mechanism, we have discovered that SNHG26 interacts with
   the ILF2 protein, which forms a heterodimer complex with
   ILF3^[280]42,[281]54. This complex plays a pivotal role in maintaining
   embryonic stem cell pluripotency and regulating gene expression related
   to the cell cycle^[282]55. Furthermore, it competes with
   Staufen-mediated mRNA decay, thus sustaining mitotic mRNA stability
   post-transcriptionally^[283]54. Additionally, the ILF2/ILF3 complex is
   known to bind to the promoters of numerous inflammatory genes (e.g.,
   IL2 and IL13) and immediate early genes (e.g., EGR1, FOS, and JUN),
   thereby driving their transcriptional activity^[284]56–[285]58. Our
   study has unveiled that the injury-induced lncRNA SNHG26 directs this
   multifunctional TF, ILF2, away from inflammatory genomic loci (such as
   JUN, IL6, IL8, and CCL20) and redirects it to the LAMB3 gene, which is
   essential for cell proliferation and migration. This redirection
   enables keratinocyte progenitors to switch their cell states and thus
   perform different tasks at various stages of wound repair. Therefore,
   our study offers a fresh perspective on the established regulatory
   mechanisms governed by lncRNAs, shedding light on how they influence
   the functional specificity of transcription factors.

   Based on research performed in the past two decades, lncRNAs have been
   proposed to enable a supragenomic layer of gene regulation that
   integrates genomic and epigenomic levels of gene expression^[286]59.
   Recently, Unlike protein regulators, lncRNAs operate swiftly without
   necessitating translation, and their levels can be dynamically adjusted
   due to their lower stability compared to proteins. We hypothesize that
   these features equip lncRNAs with the ability to regulate gene
   expression rapidly and precisely, which is required for the swift
   transition of cell states during tissue repair. Consistent with this
   notion, our previous studies have shown that the lncRNA WAKMAR2^[287]60
   and microRNA-132^[288]61 could also facilitate the
   inflammation–proliferation phase transition of keratinocyte
   progenitors. These findings open up exciting possibilities for
   harnessing non-coding RNAs with broad implications for regulating cell
   state transitions and further study is warranted to identify these
   physiologically relevant regulatory RNAs.

Limitations of the study

   This study reveals a significant stoichiometric imbalance between
   SNHG26 RNA and its protein binding partner ILF2, with ratios ranging
   from 1:1893 to 1:2540 in human keratinocyte progenitors. While the
   stoichiometric challenges of lncRNAs in relation to their targets and
   cofactors are acknowledged, the underlying mechanisms remain
   unclear^[289]39,[290]62. Our co-staining analysis highlights the
   co-localization of SNHG26 RNA and ILF2 proteins within a
   condensate-like structure in the cell nucleus. This prompts further
   investigation into whether biomolecular condensate formation may
   facilitate factor accumulation and contribute to achieving a more
   balanced stoichiometry between SNHG26 and ILF2. Moreover, SNHG26
   exhibits broad expression across various cell types. Our single-cell
   analysis using the Snhg26 global KO mouse model has revealed
   significant gene expression alterations, not limited to keratinocytes
   but extending to other cell types such as fibroblasts and macrophages.
   Delving further into SNHG26’s functions in these diverse cell types
   will provide a comprehensive understanding of the pivotal roles played
   by this lncRNA.

Methods

Human samples

   26 healthy volunteers were enrolled at the Karolinska University
   Hospital, Stockholm, Sweden. To study the expression patten of lncRNAs
   in human skin wound healing, we made 2-3 full-thickness wounds by using
   3-mm biopsy punch at the skin of lower leg area or upper buttock area
   of each healthy donor. On day1, day7, or day30, the wound edge tissues
   were collected by using a 6-mm biopsy punch and snap-frozen for the
   following studies. Human skin samples used for ex vivo wound model were
   obtained from routine abdominal or breast reduction surgeries at the
   Department of Reconstructive Plastic Surgery, Karolinska University
   Hospital. Written informed consent was obtained from all donors for the
   collection and use of tissues for research. This study was approved by
   the Stockholm Regional Ethics Committee and conducted according to the
   Declaration of Helsinki’s principles.

Mice

   The Snhg26 knockout mice and their control littermates used in this
   study were generated by Cyagen Biosciences Company (Santa Clara, US).
   The C57BL/6 J wild type mice used for Snhg26-ASO treatment were
   obtained from Charles River Laboratories (Sulzfeld, Germany). All the
   mice were breed under pathogen-free conditions in Comparative Medicine
   Biomedicum (KMB) animal facility at Karolinska Institutet under
   standard laboratory conditions, which included free access to food and
   water, a 12-hour light/dark cycle, a stable temperature range of
   20-22 °C, and humidity maintained between 40-60%. The 8–10 weeks old
   mice with C57BL/6 J background were used in all the experiments. Both
   male and female mice were used. All the mouse experiments were approved
   by committee on animal experimentation of Swedish Board of Agriculture
   (Jordbruksverket).

Mice treatment

   The mice were wounded as previously described^[291]63. Briefly, all the
   mice were caged individually prior wound making. The back skin of
   8-weeks old Snhg26 knockout mice and their littermate controls at the
   telogen (resting) phase of the hair cycle were shaved with clippers
   followed by a treatment of depilatory cream. Two 4-mm full-thickness
   wounds extending through the panniculus carnosus were made on the
   dorsum on each side of midline by using biopsy punch. The mice were
   caged individually after surgical procedure to minimize the clawing
   from other mice. During the first two days, the animals received s.c.
   buprenorphine (0.03 mg/kg) twice a day for relieving pain and distress
   caused by the procedure. The wound area was photographed and measured
   every day until approximately 90% of wound area was healed. The wound
   area was quantified using ImageJ 1.49 software (National Institutes of
   Health) and adjusted for the area of the reference circle placed
   alongside the wounds. To investigate the healing rate in SNHG26-KO
   mice, we calculated the wound healing rate constant (day^-1) using a
   one-phase decay model in GraphPad, as described by Naik et al.
   ^[292]64.

   To study the effect of knockdown of SNHG26 on mice wound healing, the
   SNHG26 Antisense LNA GapmeRs (SNHG26-ASO, 0.25nmol each wound) or
   Control GapmeRs (Ctrl-ASO, 0.25 mol each wound) were mix with
   MaxSuppressor™ In Vivo RNA-LANCEr II (Bioo Scientific) according to
   manufacturer’s instruction in a total volume of 100 μl for each wound.
   The mixture was injected intradermally into 4 points around the
   wound-edges in the mice immediately after wounding.

   To study the Snhg26 expression change during wound healing in diabetic
   mice, two-4mm full thickness wounds were made on each of the db/db mice
   and WT mice as described above. The wound edge samples were collected
   on 3 days, 7 days, and 10 days after wounding. The expression of Snhg26
   was measured by qRT-PCR.

Bulk RNA sequencing analysis of human skin and wound samples

   Snap frozen skin and wound samples were homogenized by TissueLyser LT
   (Qiagen) and total RNA was extracted by using miRNeasy mini Kits
   (Qiagen). The RNA quantity and integrity were determined by Nanodrop
   one (Thermo Fisher) and Agilent 2100 Bioanalyzer (Agilent
   Technologies), respectively. Long RNA sequencing libraries were
   generated to study the lncRNA expression.

   Firstly, the ribosome RNA was depleted from 2 ug of total RNA by
   Epicentre Ribo-zero® rRNA Removal Kit (Epicentre). The RNA is then
   fragmented randomly in fragmentation buffer, followed by cDNA
   synthesis, and incorporating dUTPs in the second-strand synthesis step
   with NEB Next® Ultra^TM Directional RNA Library Prep Kit for Illumina®
   (NEB). We used SPRI beads purification method to select the insert size
   among 250bp-300bp (not including adapters) for library preparation. The
   libraries were sequenced on an Illumina Hiseq 4000 platform.

   HISAT2 was used to align the reads to the human reference genome
   (GRCh38). The expression of lncRNAs is assessed by Cuffdiff and the raw
   count of each sample was normalized to Fragments Per Kilobase of
   transcript sequence per Millions value (FPKM). Only RNAs with median of
   FPKM ≥ 0.5 in all the samples were kept for the following analysis.

   We process the file by a series of filter steps. Step1: Exon Number
   Filter: Filter the low expression level, low confidence single exon
   transcript, select the number of exon; Step2: Transcript Length Filter:
   Select transcript with length> 200 bp transcript; step3: Known
   Transcript Filter: filtering the transcripts which overlapped with
   database annotation exon region by Cuffcompare software, and add these
   annotation to lncRNA analysis; Step4: Coding Potential Filter: several
   popular software for coding potential analysis are adopted for coding
   potential filtering, including CPC^[293]65, CNCI(vertebrate)^[294]66,
   Pfam Analysis^[295]67, and the predicted lncRNAs come from the
   intersection of these methods. The Multiple Experiment Viewer (MeV)
   software was used to perform the differential expression analysis
   between the sample groups (One way ANOVA, p < 0.005). Gene ontology
   (GO) and KEGG pathway enrichment analysis were performed by using the
   online tool EnrichR ([296]https://maayanlab.cloud/Enrichr/)^[297]68.

Histological analysis

   Paraffin-embedded tissue sections (5 μm in thickness) were
   deparaffinized and rehydrated, followed by hematoxylin and eosin (H&E)
   staining. To evaluate the epithelial tongue area of wound edge, eight
   sections from the wounds of four wild type mice and four Snhg26-KO mice
   were analyzed by Image J.

   For immunofluorescence (IF), the sections were blocked with a blocking
   buffer (3% BSA, 0.5% Triton X-100 in PBS) for 30 minutes at room
   temperature. To check the macrophage infiltration, the rabbit
   anti-mouse CD68 antibody (1:10,000; MCA1957; AbD Serotec) was applied
   to the section for overnight at 4 °C. Matched IgG isotype controls were
   included for each staining. Primary antibodies were detected with
   Donkey anti-Rat IgG (H + L) Highly Cross-Adsorbed Secondary Antibody,
   Alexa Fluor™ 488 (1:500; A21208; Thermo Fisher). The detailed
   information of antibodies please see supplementary Supplementary
   Data [298]9. At last, the sections were mounted with ProLong™ Diamond
   Antifade Mountant with DAPI ([299]P36971; Thermo Fisher). The results
   were visualized with LSM800 confocal laser scanning microscope (Carl
   Zeiss).

Fluorescent in situ hybridization (FISH) and dual fluorescence staining

   To visualize the expression of SNHG26 in human skin and wound samples,
   the probes (C3 channel) for detecting human SNHG26 RNAs were purchased
   from Advanced Cell Diagnostics and smRNA-FISH was performed using the
   RNAscope® Multiplex Fluorescent Detection Kit v2 (323100, Advanced Cell
   Diagnostics) by following manufacturer’s instructions. Briefly, the
   fresh frozen tissue sections (8 μm) were fixed in cold 4% formaldehyde
   for 15 minutes followed by blocking with Hydrogen Peroxide for
   10 minutes at room temperature. After digested with RNAscope® Protease
   IV at room temperature for 30 minutes, the sections were incubated with
   SNHG26 probes for two hours at 40 °C in HybEZ™ II Hybridization System.
   Sequential hybridization procedures were performed with amplifiers and
   labeled probes (AMP1, AMP2, and AMP3). At last, TSA® plus fluorescein
   was assigned to C3 channel and mounted with ProLong™ Diamond Antifade
   Mountant with DAPI ([300]P36971; Thermo Fisher). The results were
   visualized with LSM800 confocal laser scanning microscope (Carl Zeiss).

   To visualize the expression of SNHG26 RNA and ILF2 protein in human
   primary keratinocytes, sequential RNA protein detection was performed
   by using RNA-Protein Co-detection Ancillary Kit (Advanced Cell
   Diagnostics, Cat: 323180) as described by the manufacture. We maintain
   human primary keratinocytes at early passages and subconfluent (under
   80% confluency) to prevent differentiation induced by confluency.
   Briefly, the cells were fixed with 4% formaldehyde for 15 minutes
   followed by target retrieval. The diluted ILF2 antibody (1:100) was add
   on the slide and incubated at 4 °C for overnight. After the second
   fixation with 4% formaldehyde and Protease Plus treatment, the sections
   were processed by RNAscope ® Multiplex Fluorescent staining protocol as
   described above. At last, the Alexa Fluor 488 conjugated secondary
   antibody was applied on the slide to visualize the ILF2 signal. The
   results were visualized with FV1000 confocal laser scanning microscope
   (Olympus).

Spatial transcriptomics (ST)

   The fresh frozen skin and wound tissues were embedded in Optimal
   Cutting Temperature compound (OCT, Sakura Tissue-TEK) on dry ice. The
   samples were then processed for ST analysis by using the Visium Spatial
   platform of 10x Genomics as per the manufacturer’s instructions.
   Briefly, the H&E staining was first performed to assess the morphology
   and quality of the tissues. After permeabilization, reverse
   transcription and second-strand synthesis were performed on the slides.
   cDNA Library preparation, cleaning up, and indexing were then
   conducted. The pooled libraries were sequenced on NovaSeq6000 S4-200
   (Illumina), generating ∼300 M reads per section. The raw ST data were
   processed using the standard Space Ranger pipeline (version 1.2) with
   the GRCh38 human reference genome and GENCODE v38 gene annotations, and
   visualized by using BBrowser (BioTuring).

Single cell RNA sequencing (scRNA-seq)

   Mice skin and wound edge tissues were incubated with 5U/ml Dispase II
   at 4 °C for overnight. The epidermis was gently separated from dermis
   by using tweezers and incubated in 0.025% Trypsin/ EDTA (Thermo Fisher)
   for 10-15 minutes at 37 °C. The dermal cell suspension was prepared by
   using the whole skin dissociation kit (Miltenyi Biotec) according to
   the manufacturer’s instructions. Equal amounts of epidermal and dermal
   cells were mixed. The red blood cells and dead cells were depleted by
   Red Blood Cell Removal Solution (Miltenyi Biotec) and Dead Cell Removal
   Kit (Miltenyi Biotec), respectively. 8000 cells were used to generate a
   scRNA-seq library with Chromium Next GEM Single Cell 3ʹ Reagent Kits
   v3.1 (10×Genomics) according to the manufacturer’s protocol. The
   libraries were subjected to high-throughput sequencing on the
   Novaseq6000 S4-200 (Illumina) platform, generating 50k read pairs per
   cell.

   Raw single-cell sequencing data were processed by using the standard
   10× Cell Ranger (v5.0.1) analysis workflow, including demultiplexing,
   aligning to the GRCm39 (mm39) mouse reference genome, barcode counting,
   and unique molecular identifier (UMI) quantification. The doublets of
   cells predicted by Scrublet^[301]69 were excluded. The filtered feature
   barcode matrices were used as input into a Seurat pipeline^[302]70. To
   this end, we removed mitochondrial genes, hemoglobin genes, ribosomal
   genes, genes expressed in less than 10 cells, as well as cells with
   less than 500 detected genes, less than 1000 UMIs and with more than
   10% mitochondrial gene expression. Finally, 22,271 cells were retained
   for the subsequent analyzes, including 10,609 Snhg26-KO cells (5282
   skin cells and 5327 wound cells) and 10,882 control cells (6281 skin
   cells and 4601 wound cells). The data were first normalized using the
   SCTransform^[303]71 function. Principal component analysis (PCA) was
   carried out on the top 4000 variable genes among all the samples, and
   the first 40 PCs determined by ElbowPlot were used in the RunHarmony
   function^[304]72 to mitigate potential batch effects among samples
   processed in different libraries. Uniform manifold approximation and
   projection (UMAP) plots were generated using the RunUMAP function with
   the first 40 harmonies. The clusters were obtained using the
   FindNeighbors and FindClusters functions with a resolution of 0.8. The
   cluster markers were identified using the function of FindAllMarker.
   The cell types were annotated according to the overlaps between the
   cluster markers and well-known signature genes of each cell type from
   previous studies.

   The migration, proliferation and inflammatory scores were calculated
   used Seurat’s AddModuleScore function. The genes used for calculating
   the migration score were generated from GO:0051549. The genes for the
   proliferation score were sourced from a previously reported list of
   common proliferation markers (ZWINT, E2F1, FEN1, FOXM1, H2AFZ, HMGB2,
   MCM2, MCM3, MCM4, MCM5, MCM6, MKI67, MYBL2, PCNA, PLK1, CCND1, AURKA,
   BUB1, TOP2A, TYMS, DEK, CCNB1, and CCNE1)^[305]73. The genes for
   calculating the inflammation scores were from a summarized datasets
   from GO:0090594 and literature^[306]8 (Supplementary Data [307]4). The
   CellChat package^[308]28 was used to infer the potential cell-cell
   communications.

Microarray analysis

   Human adult primary keratinocytes (C0055C; Thermo Fisher) were cultured
   in Epilife medium supplement with Human Keratinocyte Growth Supplement
   (HKGS, S0015, Thermo Fisher) and 1X Penicillin-Streptomycin. When cells
   grew to 60% confluent, 10 nM SNHG26-ASO was transfected into the cells
   for 24 hours followed by TNFα stimulation for another 3 hours. The
   total RNA was extracted by using miRNeasy mini Kits (Qiagen) following
   the manufacturer’s instructions. The transcriptome profiling was
   performed by using Affymetrix Clariom S microarray in Bioinformatics
   and Expression Analysis (BEA) core facility at the Karolinska
   Institute. The raw.cel files were imported to Transcriptome Analysis
   Console (TAC, Thermo Fisher) for visualization. Normalized expression
   data (log2 transformed value) were exported from TAC and subjected to
   MeV software to analyze the differential expressed genes (DEGs)
   (One-way ANOVA, p < 0.001). The expression modules were defined by the
   hierarchical clustering on the DEGs. Gene ontology enrichment analysis
   in each module was performed by using the online tool EnrichR
   ([309]https://maayanlab.cloud/Enrichr/)^[310]68.

Magnetic cell separation of different cell types from skin and wounds

   After being washed in PBS, the skin and wound biopsies skin and wound
   biopsies were incubated with Dispase II (5 U/mL, ThermoFisher
   Scientific) at 4 °C for overnight. On the next day, the epidermis was
   separated from the dermis and digested in 0.025% Trypsin-EDTA at 37 °C
   for 15 minutes. The CD45+ and CD45- cells were isolated from the
   epidermal cell suspension by using CD45 microbeads (Miltenyi Biotec,
   Bergisch Gladbach, Germany). The dermis was processed with a Whole Skin
   Dissociation Kit according to the manufacturer’s instructions (Miltenyi
   Biotec, Bergisch Gladbach, Germany). The dermis was digested with
   12.5 µL of Enzyme P, 50 µL of Enzyme D and 2.5 µL of Enzyme A at 37 °C
   for 3 hours followed by processing with Medicon tissue disruptor (BD
   Biosciences, San Diego, CA). The fibroblasts, macrophages, and T cells
   were separated with CD90 + , CD68 + , and CD3+ beads from the dermal
   cell suspension.

Gene deletion, silencing, and overexpression

   For silencing of SNHG26, 10 nM LNA-long
   RNA-GapmeR-antisense-oligonucleotide (SNHG26-ASO or Ctrl-ASO) (Qiagen)
   was transfected into human primary keratinocytes by using
   Lipofectamine™ 3000 (ThermoFisher Scientific). For the depletion of
   SNHG26 in Ker-CT human keratinocytes cell line (ATCC, Cat: CRL-4048),
   two sgRNAs targeting SNHG26 loci were cloned into lentiCRISPRv2
   plasmid, which was kindly provided by Prof. Yuping Lai and Prof. Dali
   Li in East China Normal University. The lentiCRISPRv2 plasmid was
   transfected into 293 T cells with the packaging plasmids pMD2G
   (AddGene) and psPAX2 (AddGene) for lentiviral production. Ker-CT cells
   were transduced with the virus supernatant for 12 hours and selected by
   1 μg ml^−1 of puromycin. The expression of SNHG26 was measured by
   qRT-PCR to confirm the SNHG26 knockout.

   For overexpression of SNHG26 and SNORD93, the cDNA sequence of these
   two genes were synthesized and cloned into pcDNA3.1 or PZW1-SnoVector
   (Sangon, Shanghai). The keratinocytes at 60-70% confluence were
   transfected with these plasmids for 48 hours with Lipofectamine™ 3000
   (ThermoFisher Scientific). The gene expression was measured by qRT-PCR.

Quantification of SNHG26 and ILF2 copy number

   To quantify the SNHG26 transcript, a standard curve by amplifying a
   dilution series of the PZW-Sno-SNHG26 plasmid was created by using
   real-time PCR. Subsequently, we measured the Ct value of SNHG26 in
   50,000 keratinocytes via qRT-PCR and calculated the copy number based
   on the standard curve.

   For the quantification of ILF2 protein, we conducted ILF2 Western
   blotting using a dilution series of ILF2 recombinant protein. The
   signal value of ILF2 was measured by Image J. A standard curve by
   plotting the signal values of the ILF2 immunoblot bands against the
   molarity of the ILF2 protein was then generated. The copy number of ILF
   protein in each keratinocyte was calculated by ILF2 band gray value and
   the standard curve.

Scratch assay

   Human adult primary keratinocytes were transfected with 10 nM
   SNHG26-ASO for 24 hours in Incucyte® Imagelock 96-well Plate
   (Sartorius). When cells grow to full confluent, the plate was placed in
   Essen® 96-well Wound Maker (Sartorius) to generate scratch wounds with
   approximately 800 µm width. After washing with PBS, EpiLife medium
   without any supplements was added to the cells. The plate was then
   placed into IncuCyte (Essen Bioscience) and imaged every two hours for
   48 hours. The images in each time point were analyzed with the
   integrated metrics in the IncuCyte ZOOM software.

Human ex vivo wound model

   The human skin samples were obtained from abdominal reduction surgeries
   or mastectomies at the Department of Reconstructive Plastic Surgery,
   Karolinska University Hospital. The wound was made by using 2-mm biopsy
   punch on the surface of the skin. The injured skin was excised using an
   8 mm biopsy punch and cultured in Dulbecco’s Modified Eagle Medium
   (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics
   (penicillin 100 U/ml and streptomycin 100 mg/ml; 15140122; Thermo
   Fisher). 0.2 nmol SNHG26-ASO GapmeR or control oligos (Qiagen) were
   dissolved in 30% pluronic F-127 gel (Sigma-Aldrich) and topically
   applied on the wounds immediately after the injury and also three days
   later. The biopsies were collected at Day3 and Day5 after injury for
   histological analysis.

   For the delivery of SNHG26 overexpression plasmid DNA, we used in
   vivo-jetPEI (Cat. 201-10 G, Polyplus-transfection, Illkirch, France)
   reagent as per the manufacturer’s guidelines. Specifically, 1.5 µg of
   plasmid DNA and 0.18 µl of in vivo-jetPEI were separately diluted in
   2.5 µl of 5% glucose solution. These two dilutions were then combined,
   and after a 15-minute incubation at room temperature, 5 µl of this
   mixture was topically administered to the wounds immediately after
   injury, as well as two and four days later. The progress of wound
   healing was monitored by using CellTracker Green CMFDA fluorescent dye
   (Cat. C2925, Invitrogen, Waltham, MA)^[311]74. 4 µl of the dye solution
   (50 µM) was added to each wound tissue and allowed to incubate for
   30 minutes at 37 °C with 5% CO[2]. Following the incubation, the
   tissues were rinsed with PBS and imaged using a fluorescence microscope
   (Nikon eclipse Ni-E). The wound samples were imaged three and five days
   after the initial injury, and the percentage of re-epithelialization
   was quantified by using the Image J software. The re-epithelization of
   each wound was evaluated by the healing rate (%)=(Area within initial
   wound edges - Area within newly formed epidermis)*100/ Area within
   initial wound edges.

RNA extraction and qRT-PCR

   The tissue samples were homogenized by using TissueLyser LT (Qiagen) in
   Trizol (Thermo Fisher) prior to RNA isolation. For cells, Trizol was
   directly added into the cell culture plate and incubated at room
   temperature for 5 minutes. The total RNA was extracted following the
   instructions of the manufacturer of the Trizol reagent. The RNA was
   dissolved in nuclease-free water and the concentration was determined
   by Nanodrop one.

   500 ng of total RNA was converted to cDNA using RevertAid First Strand
   cDNA Synthesis Kit (K1622, Thermo fisher). Quantitative PCR was
   performed by using SYBR™ Green PCR Master Mix (4309155, Applied
   Biosystems) or TaqMan™ Universal Master Mix II (4440038, Applied
   Biosystems). Target gene expression levels were normalized between
   samples to the internal control 18S rRNA in tissues and to GAPDH in
   cells. The fold change in gene expression was determined by the 2^-ΔΔCt
   method.

RNA pulldown and mass spectrometry

   SNHG26 from human keratinocyte progenitors was amplified and inserted
   into pcDNA3.1(+) plasmid through double restriction enzyme digestion
   and ligation method. Linearized pcDNA3.1(+) vector (Thermo Fisher)
   expressing SNHG26 and LacZ gene was used as a template to synthesize
   SNHG26 and LacZ RNA by using MEGAscript™ T7 Transcription Kit (AM1334,
   Thermo Fisher). The RNA was purified by using MEGAclear™ Transcription
   Clean-Up Kit (AM1908, Thermo Fisher). After the concentration was
   measured by Nanodrop One, the biotin labeled RNA was generated by
   Pierce™ RNA 3’ End Desthiobiotinylation Kit (20163, Thermo Fisher).
   Biotin labeled RNA was further incubated with Streptavidin Magnetic
   Beads to form Beads-RNA complex. Each binding reaction was performed
   with 50 pmol of biotinylated RNA-beads and 200 ug of cell lysates
   prepared from human keratinocyte progenitors by using Pierce™ IP Lysis
   Buffer (87787, Thermo Fisher). After extensive washing, the proteins on
   the beads were eluted by heating samples at 96 °C for 10 minutes in 1X
   Laemmli buffer containing 1% SDS. The eluted proteins were subjected to
   SDS-PAGE analysis and Coomassie Blue Staining. A gel slice
   corresponding to 37-50 kDa protein bands from both SNHG26 and LacZ
   pulldowns were excised and subjected to mass spectrometry analysis.

RNA Immunoprecipitation (RIP) assay

   The RIP assay was performed as previously described^[312]75. Briefly,
   2 × 10^7 human keratinocyte progenitors cell lysates were incubated
   with Pierce™ Protein A/G Magnetic Beads conjugated with mouse IgG or
   ILF2 antibody (sc-365283, Santa Cruz) at 4 °C for overnight.
   RNA-protein-antibody-Magnetic Beads complex was purified by Magnetic
   Separation. The immunoprecipitated RNAs were then extracted by Trizol.
   cDNA was synthesized using RevertAid First Strand cDNA Synthesis Kit
   (K1622, Thermo Fisher) and detected by qRT-PCR. Results were normalized
   to input RNA and shown as % enrichment.

ChIP-PCR and ChIP-sequencing

   The human keratinocyte progenitors were transfected with SNHG26-ASO or
   Ctrl-ASO. When grown to 90% confluent, the cells were cross-linked with
   1% methanol-free formaldehyde (28906, Thermo Fisher) for 10 minutes at
   room temperature. The cross-linking was quenched with 125 mM glycine
   and incubated at room temperature for 5 minutes. The ChIP assay was
   performed by MAGnify™ Chromatin Immunoprecipitation System (492024,
   Thermo Fisher). The cell Lysis Buffer with Protease Inhibitors was
   added to the cell pellet and incubated on ice for 5 minutes. The cell
   lysate was sonicated using Bioruptor UCD-200 (Diagenode) with the setup
   of “30 sec ON and 30 sec OFF” at the highest voltage to generate
   100-500 bp chromatin fragments. The sheared chromatin was diluted in a
   Dilution Buffer and incubated with Dynabeads® coupled ILF2 antibody and
   isotype control IgG at 4 °C for 2 hours. After extensive washing, the
   bead pellet was incubated with Reverse Crosslinking Buffer with
   Proteinase K (20 mg/ml) at 55 °C for 15 minutes. After discarding the
   beads, the supernatant was further heated at 65 °C for 15 minutes. The
   genomic DNA was isolated by using the DNA Purification Magnetic Beads,
   and then fragmented randomly. After electrophoresis, DNA fragments of
   the desired length were gel purified. Adapter ligation and DNA cluster
   preparation were performed and subjected to DNBSEQ-G400 platform
   sequencing with PE100.

   Raw sequencing data were processed by using Trimmomatic v0.36^[313]76
   to trim reads of low quality and shorter than 20 nucleotides. Clean
   reads were mapped to GRCh38 human reference genome using Bowtie2
   (v2.3.5.1)^[314]77 followed by removing PCR duplicates using Picard
   (v2.20.4) tools. The mapped reads were converted into bigwig format for
   IGV viewer^[315]78 using deepTools bamCoverage function^[316]79. Peaks
   were called using MACS2 (v2.2.6) with the default parameter^[317]80.
   Peaks from two biological replicates were merged and filtered using
   Bedtools^[318]81. The peak regions of each group were compared and
   visualized using the computeMatrix and plotHeatmap functions inside the
   deepTools. The ChIP-purified DNA was also subjected to qPCR analysis
   using primers targeted the promoter region of indicated genes
   (Supplementary Data [319]9). The ChIP-qPCR results were shown as a
   percentage of enrichment relative to input DNA.

ChIRP sequencing

   The ChIRP sequencing was performed as previously
   described^[320]48,[321]82. In this approach, we designed 28
   3’-biotinylated complementary DNA oligonucleotides to span the entire
   length of SNHG26 (3.1 kb). We used a similar set of probes targeting
   LacZ mRNA as a negative control. The SNHG26 and LacZ probes were
   designed at Biosearch Technologies ([322]www.biosearchtech.com) with
   the following criteria: number of probes = 1 probe /100 bp of RNA
   length; 2) percentage of GC = 45-50%; 3) Oligonucleotide length = 20.
   Sequences that had extensive complementarity to other sites in the
   genome or were repetitive were excluded. LacZ-specific probes were used
   as negative controls. To eliminate the artifacts due to the
   precipitation of nonspecific DNA fragments from off-target
   hybridization of the probes, we ranked all probes based on their
   relative positions along the target RNA and split them into two pools.
   As the two sets of probes shared no overlapping sequences, the only
   target they have in common is the RNA of interest and its associated
   chromatin. To obtain high confidence ChIRP-seq signal, we performed two
   independent ChIRP-seq runs with “even” and “odd” probes separately and
   focused our analysis exclusively on the overlap between their signals.
   The DNA probes labeled with BiotinTEG at 3’ end were ordered from
   Eurofins Scientific. 3×10^7 human keratinocyte progenitors were firstly
   cross-linked by incubating with 1% formaldehyde and quenched with
   125 mM glycine. After resuspending in the Swelling buffer, the nuclei
   were isolated and further cross-linked with 3% formaldehyde. The nuclei
   were resuspended in freshly supplemented Nuclear Lysis Buffer and
   sonicated for 3 hours at 4 °C with a Bioruptor. The chromatin was
   incubated with biotinylated probes pool at 37 °C for 4 hours in a
   hybridization oven. The RNA-Chromatin complex was captured with
   Dynabeads MyOne Streptavidin C1 magnetic beads. After extensive
   washing, the bead pellet was incubated with DNA Elution Buffer and the
   chromatin in the supernatant was reverse crosslinked with Proteinase K
   (20 mg/ml) at 50 °C for 45 minutes with gentle shaking. The DNA was
   purified by phenol-chloroform-isoamyl alcohol and subjected to NEBNext
   ChIP-seq library preparation. The paired-end sequencing libraries were
   sequenced by Illumina high-throughput sequencer (DNBSEQ PE100).

   ChIRP sequencing analysis was performed by using the same pipeline of
   ChIP-seq as described above. We merged the peaks from two biological
   replicates and filtered out the ENCODE Blacklist^[323]83 to construct
   the consensus peaks for each group using Bedtools^[324]81. The peaks
   were visualized using the computeMatrix and plotHeatmap functions
   inside the deepTools. The Motif Discovery function of MEME suite
   (version 5.5.4) was used to identify the de novo motif of SNHG26
   binding sites with the following parameters: the width of the expected
   motif was between 6 and 20 and the expected occurrence per sequence was
   set to zoops: zero or one site per sequence^[325]49,[326]84.

BioID assay and mass spectrometry

   MCS-BioID2-HA, kindly provided by Kyle Roux (Addgene plasmid # 74224),
   was used for the BioID assay as previously described in the
   literature^[327]40,[328]85,[329]86. ILF2 cDNA was synthesized and
   cloned into the MCS-BioID2-HA vector to produce the ILF2-BioID2 fusion
   protein. This recombinant plasmid, along with an empty vector, was
   transfected into human primary keratinocytes using Lipofectamine 3000
   (Thermo Fisher), following the manufacturer’s protocol. After 24 hours,
   the culture medium was supplemented with 50 µM biotin and incubation
   continued for another 16 hours. The cells were then washed with PBS and
   lysed with 540 µL of lysis buffer (8 M urea, 1 mM dithiothreitol (DTT),
   1x protease inhibitor in 50 mM Tris·Cl) per 10-cm dish. The lysates
   were sonicated for 1 minute using an Ultrasonic Sample Processing
   System (LICHEN, Shanghai) at a 30% duty cycle and an output level of 4.
   After sonication, Triton X-100 was added to achieve a final
   concentration of 1%, and the lysate was further centrifuged to collect
   the supernatant. The supernatant was then incubated with 100 µL of
   Dynabeads™ MyOne™ Streptavidin beads (Thermo Fisher) overnight.
   Following extensive washing, proteins bound to the beads were eluted
   and analyzed by mass spectrometry.

Quantification and statistical analysis

   A minimum of four mice were included per experimental group. Mice were
   randomly allocated to different experimental group and male mice were
   used. For cell experiments, each groups contains a minimum of three
   independent biological replicates. For quantification of histological
   staining, sections from at least four mice were analyzed by ImageJ
   software. All data were expressed as mean ±standard deviation (SD) or
   mean ± standard error of the mean (SEM). Statistical significance
   between groups was determined using either a two-tailed Student’s
   t-test or ANOVA analysis, by using GraphPad Prism 8 (GraphPad Software
   Inc, California, USA). For all statistical tests, P values < 0.05 were
   considered to be statistically significant.

Reporting summary

   Further information on research design is available in the [330]Nature
   Portfolio Reporting Summary linked to this article.

Supplementary information

   [331]Supplementary Information^ (3.8MB, pdf)
   [332]Peer Review File^ (4.7MB, pdf)
   [333]41467_2024_52783_MOESM3_ESM.pdf^ (107.2KB, pdf)

   Description of Additional Supplementary Files
   [334]Supplementary Data 1^ (10.8KB, xlsx)
   [335]Supplementary Data 2^ (74.2KB, xlsx)
   [336]Supplementary Data 3^ (39.5KB, xlsx)
   [337]Supplementary Data 4^ (15.4KB, xlsx)
   [338]Supplementary Data 5^ (15.4KB, xlsx)
   [339]Supplementary Data 6^ (23.6KB, xlsx)
   [340]Supplementary Data 7^ (262.3KB, xlsx)
   [341]Supplementary Data 8^ (21.1KB, xlsx)
   [342]Supplementary Data 9^ (18.8KB, xlsx)
   [343]Reporting Summary^ (2.9MB, pdf)

Source data

   [344]Source Data^ (4.5MB, xlsx)

Acknowledgements