Introduction

Psoriasis is a common and complex chronic inflammatory skin disease that is characterized by epidermal hyperplasia, aberrant differentiation of keratinocytes, exaggerated angiogenesis, and dermal infiltration of inflammatory cells (Perera et al, 2012; Boehncke & Schön, 2015; Lebwohl, 2018). It affects 2–3% of the global population (Crow, 2012). The imiquimod (IMQ)-induced mouse model largely mimics the psoriasis phenotype in humans (van der Fits et al, 2009) and has been widely used for the pre-clinical study of psoriasis. Recent studies using the IMQ-induced mouse model have identified the involvement of the IL-23/IL-17 axis in psoriasis (Boehncke & Schön, 2015; Kopp et al, 2015; Burkett & Kuchroo, 2016; Lebwohl, 2019). Activated macrophages and dendritic cells (DCs) generate IL-23, TNF-α, IL-6, and IL-1β to stimulate and polarize helper T (Th) cells to transition into Th1, Th17, and Th22 cell subsets, which, respectively, produce IFN-γ, IL-17A/F, and IL-22. These pro-inflammatory cytokines then promote the proliferation and activation of keratinocyte cells, which in turn produce IL-23 and aggravate the inflammation and psoriasis progression. In psoriatic patients, the major TNF-α-producing cells are the slan+ macrophages (Brunner et al, 2013). The reduction in these cells during psoriasis remission underlines their critical role in psoriasis development, maintenance, or both (Brunner et al, 2013), emphasizing an important role for macrophages in the development of psoriasis. However, the cellular and molecular mechanisms of psoriasis underlying its development remain unclear.

The Src homology-2 domain-containing protein tyrosine phosphatase-2 (SHP2) is a ubiquitously expressed cytoplasmic tyrosine phosphatase containing two SH2 domains and one catalytic protein tyrosine phosphatase (PTP) domain (Feng et al, 1993). It is encoded by PTPN11 and has critical functions in cell proliferation, differentiation, and survival (Qu, 2002). SHP2 is inactive in the basal state, but when combined with a phosphotyrosine (pY) protein, it is activated as a phosphatase to elicit downstream signaling (Barford & Neel, 1998). SHP2 missense mutations account for approximately 50% of the cases of Noonan syndrome (Keilhack et al, 2005), a common autosomal dominant disorder characterized by multiple, variably penetrant defects. Intriguingly, a 55-year-old woman with many cardinal physical characteristics of Noonan syndrome also suffered simultaneously from annular pustular psoriasis (Catharino et al, 2016), suggesting a potential role of SHP2 in psoriasis; however, this possibility has not been investigated either in vitro or in vivo.

Other possible effectors of psoriasis are the toll-like receptors (TLRs), which are essential sensors of a variety of viral and microbial infections and endogenously derived damage-associated molecular patterns (DAMPs) (Bryant et al, 2015). TLRs recruit adaptor proteins and activate transcription factors to produce inflammatory cytokines that promote the initiation of adaptive immune responses (Newman et al, 2016). The TLR family consists of 10 members in humans and 13 in mice, which are expressed on the cell surface and intracellularly (Barbalat et al, 2011). The intracellular TLRs, including TLR3, TLR7, TLR8, and TLR9 in humans, recognize nucleic acids (Blasius & Beutler, 2010). TLR7 recognizes single-stranded RNA and traffics it to endosomes (Wang et al, 2020). Imiquimod (IMQ) is a TLR7 ligand that has been used for the topical treatment of genital and perianal warts caused by the human papillomavirus (Beutner & Tyring, 1997). Interestingly, topical IMQ treatment of patients with actinic keratoses and superficial basal cell carcinomas can exacerbate even previously well-controlled psoriasis (Gilliet et al, 2004). Such IMQ-exacerbated psoriasis occurs both in the topically treated areas and at distant, previously unaffected skin sites (Wu et al, 2004). These findings suggest, but have not yet confirmed, that TLR7 activation may promote psoriasis.

In this study, we performed single-cell RNA sequencing (scRNA-seq), a powerful and unbiased method of examining the pathological processes of human disease (Shalek & Benson, 2017), to profile the different transcriptomics of the skin microenvironment at a single-cell level with the SHP2 allosteric inhibitor SHP099. Unsupervised analyses revealed that SHP099 impaired skin inflammation in myeloid cells (especially macrophages). Ablation of SHP2 in macrophages weakened NF-κB activation, subsequently resulting in amelioration of IMQ-induced psoriasis-like skin inflammation in mice. Mechanistically, SHP2 dephosphorylated TLR7 at Tyr1024 to promote its ubiquitination in macrophages and trafficking to the endosome. Overall, our findings identify SHP2 as a critical regulator of psoriasis and as a potential therapeutic target for the treatment of psoriasis-related skin diseases.

Results SHP2 expression increased in both human psoriatic patients and IMQ-induced psoriasis-like mice

To assess whether SHP2 may play a functional role in psoriasis, we first analyzed the expression and activity of SHP2 in human samples collected from healthy donors (normal controls) and psoriatic patients. At the mRNA level, SHP2 (encoded by the PTPN11 gene) was significantly higher in the skin lesions (Fig 1A) and human peripheral blood mononuclear cells (PBMCs) (Fig 1B) of psoriatic patients than those of normal controls. At the protein level, skin biopsies (Fig 1C) and PBMCs (Fig 1D) from psoriatic patients also showed a considerable increase in SHP2 compared to normal controls. Consistently, enzymatic assays revealed increased SHP2 activity in PBMCs from psoriatic patients versus normal controls (Fig 1E). Additionally, the p-ERK levels, which were shown to be positively regulated by SHP2, were also higher in the skin lesions of psoriatic patients than those of normal controls (Fig 1F). Similarly, SHP2 expression was increased in the dorsal skin of the IMQ-induced psoriasis-like murine model, both at the mRNA and protein levels (Fig 1G and H). Collectively, data from human specimens and murine models indicate an increase in SHP2 functioning in psoriasis.

Figure 1. SHP2 expression increased in both human psoriatic patients and IMQ-induced psoriasis-like mice

Expression of PTPN11 (gene encoding SHP2) in skin lesions in psoriatic patients compared with skin from healthy donors based on microarray data (No. GSE14905). Expression of PTPN11 in human PBMCs of psoriatic patients (n = 14) and normal controls (n = 16). Representative SHP2 staining in skin sections of psoriatic patients (n = 13) and normal controls (n = 5). Scale bars: 200 μm. Western blot analysis of PBMCs lysates derived from psoriatic patients and normal controls. The catalytic activity of SHP2 was measured in human PBMCs lysates derived from psoriatic patients (n = 25) and normal controls (n = 25). Representative p-ERK staining of skin sections of psoriatic patients and normal controls. Scale bars: 200 μm. Quantitative PCR analysis of Ptpn11 mRNA levels in the IMQ-treated or non-treated dorsal back of C57BL/6 mice at day 5 (n = 6/group). Data were normalized to Gapdh expression. Representative histological sections of IMQ-treated or non-treated dorsal back of C57BL/6 mice at day 5. Scale bar: 100 μm. Quantitative PCR analysis of mRNA levels in human PBMCs derived from psoriatic patients (n = 20) and normal healthy controls (n = 20).

Data information: Data are represented as mean ± SEM. P values are determined by two-tailed Mann–Whitney U test (A and B) or two-tailed unpaired Student’s t-test (E, G, and I). *P < 0.05, **P < 0.01.

Source data are available online for this figure.

Next, we treated the PBMCs derived from human donors with SHP099, a potent allosteric inhibitor of SHP2 (Chen et al, 2016). As expected, when compared to normal control donors, the PBMCs obtained from psoriatic patients expressed significantly higher levels of psoriasis-related cytokines; however, when these PBMCs were treated with 10 μM SHP099, these cytokines decreased remarkably, to a level comparable to that in the normal controls (Fig 1I). Furthermore, SHP099 also markedly reduced IMQ-induced levels of psoriasis-related cytokines in human PBMCs from healthy donors (Appendix Fig S1), without affecting their basal levels (Appendix Fig S2). Taken together, our data demonstrated that SHP2 expression increased in both human psoriatic patients and IMQ-induced psoriasis-like mice.

SHP2 inhibitor SHP099 attenuated the psoriasis-like phenotype in the IMQ-induced murine model

To determine whether SHP2 activity affects psoriasis pathogenesis, we treated the murine psoriatic model with SHP099. SHP099 significantly suppressed IMQ-induced swelling, epidermal acanthosis, keratinocytes proliferation, and dermal inflammatory cell infiltration (Fig 2A), accompanied by a drastic decrease in the clinical scores (Fig 2B). SHP099 did not affect the skin condition of normal mice (Appendix Fig S3).

Figure 2. SHP2 inhibitor SHP099 attenuated the psoriasis-like phenotype in the IMQ-induced murine model

C57BL/6 mice (n = 6/group) were treated with indicated dose of SHP099 or vehicle for 4 days.

A. Phenotypic presentation (top), H&E staining (middle), and statistical data (bottom) (mean ± SEM) of dorsal skin. Scale bar: 100 μm. B. Clinical scores plotted with mean ± SEM. *Denotes statistical significance when compared with the IMQ group. C, D. Quantitative PCR analysis of mRNA encoding IL-23/IL-17A axis cytokines (C) and other psoriasis-related cytokines (D) in the dorsal skin of C57BL/6 mice treated with indicated dose of SHP099 or vehicle for 4 days. Results were normalized to Gapdh expression. E. ELISA quantification of protein levels of cytokines in mouse serum.

Data information: Data are represented as mean ± SEM. P values are determined by two-tailed unpaired Student’s t-test (A, C–E) or Tukey multiple-comparison test (B). *P < 0.05, **P < 0.01.

Source data are available online for this figure.

Consistent with the ameliorated psoriasis-like phenotype, the mRNA levels of genes related to the IL-23/IL-17 axis, such as Il23a, Il17a, and Il22 (Fig 2C), and the expression of Tnfa, Il6, and Il1b was decreased in the skin (Fig 2D). The IL-23 and IL-17A serum levels were also significantly reduced by SHP099 (Fig 2E). SHP099 had a moderate effect on the IL-23-induced psoriasis-like mouse model, although not as strong as on the IMQ model. Total ear thickness clearly decreased, but psoriasis-related inflammatory cytokines were only slightly affected (Fig EV1). Taken together, these data demonstrated that SHP099 ameliorated the IMQ-induced and IL-23-induced psoriasis progression in mice.

Click here to expand this figure.

Figure EV1. SHP099 attenuated the psoriasis-like phenotype in the IL-23-induced murine model

C57BL/6 mice (n = 5/group) were subjected to IL-23-induced psoriasis-like skin inflammation and were treated with 10 mg/kg SHP099 or PBS for 4 days.

A, B. Ear thickness (A), H&E staining, and statistic results (B) of ear skin from SHP099 treated with or without C57BL/6 mice injected intradermally with rmIL-23 or PBS for 4 days. Ear thickness was measured daily. C. Quantitative PCR analysis of mRNA encoding IL-23/IL-17A axis cytokines and other psoriasis-related cytokines in the ear skin. Results were normalized to Gapdh expression.

Data information: Data are represented as mean ± SEM. P values are determined by Tukey multiple-comparison test (A–C). *P < 0.05, **P < 0.01, ns, not significant.

Source data are available online for this figure.

scRNA-seq analysis revealed that SHP099 reprogrammed IMQ-induced inflammation in mouse skin

To clarify the mechanism by which the small-molecule allosteric inhibitor SHP099 improves psoriasis, we used scRNA-seq to better understand the pathological processes and determine transcriptomic changes. We dissociated dorsal skin from untreated (Sham) mice, IMQ-induced (IMQ) mice, and IMQ-induced mice treated with 10 mg/kg SHP099 (IMQ+SHP099) in a single-cell suspension and compared unsorted cells from the three groups using scRNA-seq analysis (Fig 3A). A total of 49,543 cells met the preprocessing threshold (Appendix Fig S4) and were used for downstream analysis. Uniform manifold approximation and projection (UMAP) plotting revealed 6 clusters, including endothelial cells, fibroblasts, keratinocytes, lymphocytes, myeloid cells, and Schwann cells (Fig 3B), which were further separated into 25 subpopulations for more detailed profiling of rare subsets (Fig 3C). Cluster identities were determined by the expression of unique marker genes using heatmaps (Fig 3D and Appendix Fig S5), bar charts (Fig 3E), and feature plots (Appendix Fig S6), revealing successful capture of major cell subsets. Consistent with Fig 2, the inflammation-associated genes significantly increased in IMQ-induced mouse skin compared to sham controls, and treatment with SHP099 reduced the expression of inflammation-associated genes (Fig 3F). Obviously, SHP099 treatment normalized myeloid cells and endothelial cells in psoriasis-like mouse skin lesions. Overall, these results indicated that the identified cell types and differentially expressed marker genes corresponded to most major known cell types, suggesting that SHP099 improves the IMQ-induced psoriasis-like phenotype by inhibiting the expression of inflammation-associated genes.

Figure 3. scRNA-seq analysis revealed that SHP099 reprogrammed IMQ-induced inflammation in mouse skin

Workflow for single-cell profiling of dorsal skin from the sham, IMQ, and IMQ+SHP099 three groups. The number of cells used in the integration was shown for each sample group (bottom). Each group consists of two samples, including two mice per sample in the sham group and three mice in the model and SHP099 administration groups. UMAP visualization of all cells. Clusters are colored and labeled according to their inferred cell type identities. UMAP plot of cells colored by cell subsets. Heatmap displayed the top 20 differentially expressed genes for each cluster (left). Selected genes for each cluster are highlighted (middle). Feature plots of expression distribution for cluster-specific marker genes (right). Relative expression of selected cluster-specific genes in each cluster. High-density bar charts showed the distribution of the normalized expression levels of genes. Violin plot showed the relative expression of 47 inflammation-associated genes based on sample groups and cell clusters.

Source data are available online for this figure.

SHP099 regulated genes associated with inflammation in myeloid cells in mouse skin

To further explore the biological significances of transcriptional changes in the myeloid cells cluster, we first separated the myeloid cells into six clusters: Plac8+ macrophages, Pf4+ macrophages, Langerhans cells, neutrophils, basophils, and proliferating dendritic cells (Fig 4A). Among these subclusters, macrophages accounted for the largest proportion (Fig 4B). We performed pathway enrichment analysis using upregulated genes in the myeloid cells and found that biological pathways, such as C-type lectin receptor signaling, NF-κB signaling, and TNF signaling pathways, were highly enriched (Fig 4C and D; Appendix Figs S7 and S8). To identify IMQ response genes, we examined the overlap of upregulated genes using differential expressed genes (DEGs) analysis of IMQ versus normal and IMQ versus SHP099 in myeloid cells (Fig 4E). We mapped these upregulated genes (n = 108) into functional protein association networks from the STRING database (Szklarczyk et al, 2019), and identified Itgam, Ccl2, Itgb2, Mmp9, Fn1, Lyz2, Cybb, Cd14, C3, and Clec4d as the top 10 hub genes (Fig 4F and Appendix Fig S9). Interestingly, these genes were mainly expressed in macrophages and Langerhans cells (Fig 4G), implying that they may function as master regulators in a cell type-specific manner.

Figure 4. SHP099 regulated genes associated with inflammation in myeloid cells in mouse skin

UMAP plot of myeloid cells subclusters. Bar plots showing the percentage (%) of cell types in each sample group. Representative KEGG pathways enriched in upregulated genes in myeloid cells. Gene set enrichment analysis (GSEA) of the NF-κB signaling pathway. Volcano plots of DEGs that were upregulated (red) or downregulated (blue) in myeloid cells. Examples of DEGs were labeled. Gene networks of DEGs that were upregulated in the IMQ group compared with the sham and SHP099 groups. Expression patterns of the top 10 hub genes in (F). Heatmap (top) displays the average expression in the three groups and in different myeloid cell subsets. Boxplots (bottom) show the corresponding expression distribution. SHP2 deficiency in myeloid cells alleviated the psoriasis-like phenotype in the IMQ-induced murine model

To determine the causal role of SHP2 in psoriasis, we generated myeloid cell lineages (monocytes, mature macrophages, and granulocytes)-specific SHP2 conditional knockout mice, called M-Shp2−/− mice (Lyz2-cre:Shp2fl/fl mice), and confirmed a pretty efficiency of Shp2 deletion in bone marrow-derived macrophages (BMDMs), but not in bone marrow-derived dendritic cells (BMDCs), as shown in Appendix Fig S10A. We treated the animals’ dorsal skin with IMQ, and after 4 days of treatment, the wild-type mice exhibited swelling, epidermal acanthosis, and greatly increased proliferation of keratinocytes, dermal inflammatory cell infiltration, and inflammatory cytokines. These pathological changes were remarkably ameliorated in M-Shp2−/− littermates (Fig 5A–D). Although Lyz2-cre is also expressed in neutrophils, given the high knockdown efficiency of Shp2 in macrophages and our group demonstration that SHP2 deletion in neutrophils alleviates psoriasis-like skin inflammation in mice (preprint: Ding et al, 2021), our work focuses on the role of SHP2 in macrophages. To further examine the effect of SHP2 deletion on myeloid cells on psoriasis, we generated dendritic cells-specific SHP2 conditional knockout mice, called DC-Shp2−/− mice (Itgax-cre:Shp2fl/fl mice), and examined the effect of Shp2 deletion in DCs and macrophages, as shown in Appendix Fig S10B–E. Upon IMQ challenge, psoriasis-like pathological changes were not significantly different in DC-Shp2−/− littermates (Fig 5E–H), suggesting that DC-derived SHP2 did not affect psoriasis progression. A similar decrease in IMQ-activated inflammatory cytokines (e.g., Il23a, Tnfa, Il6, and Il1b) was also observed in the BMDMs (Appendix Fig S11A) and peritoneal macrophages (PMs; Appendix Fig S11B) of the M-Shp2−/− mice. Notably, such a decrease in the pro-inflammatory cytokines in M-Shp2−/− mice was recapitulated by SHP2-deficient THP-1 cells compared to the control cells (Appendix Fig S11C). Also, when stimulated with IL-36, peritoneal macrophages derived from M-Shp2−/− mice also produced fewer psoriasis-related cytokines than those derived from the wild-type littermates (Appendix Fig S11D).

Figure 5. SHP2 deficiency in myeloid cells alleviated the psoriasis-like phenotype in the IMQ-induced murine model

A–H. Representative images (A and E), histological sections of dorsal back (B and F), and clinical scores (C and G) and quantitative PCR analysis of mRNA levels in dorsal skin (D and H) from wild-type (n = 6) and M-Shp2−/− mice or DC-Shp2−/− mice (n = 6) treated with IMQ for 4 days. Scale bar: 100 μm. Left: H&E (hematoxylin and eosin) staining data; right: statistical data (mean ± SEM). Results were normalized to Gapdh expression.

Data information: Data are represented as mean ± SEM. P values were calculated by Tukey multiple-comparison test (B, D, F, H) or Bonferroni multiple-comparison test (C, G). *P < 0.05, **P < 0.01, ns, not significant.

Source data are available online for this figure.

Additionally, there was a substantial increase in CD68+ macrophages (Appendix Fig S12A), neutrophils (Appendix Fig S13A), and DCs (Appendix Fig S13B) infiltration in the skin lesions of psoriatic patients compared to normal controls, consistent with Fig 3. Noticeably, there is a clear co-localization of CD68 and SHP2 occurred in the psoriatic dermis (Appendix Fig S12B). A similar result was observed in the IMQ-induced psoriasis-like murine model (Appendix Fig S12C), suggesting that SHP2 was highly expressed in the infiltrated macrophages in the psoriatic skin. To ground our observation in a spatial sense, we performed spatial transcriptome sequencing analysis in normal skin and psoriasis patient lesion skin using the section-specific barcodes. We obtained two skin sections from normal control and one lesion skin from psoriasis patient (Appendix Fig S14A and B), and we focused on cluster 1 and cluster 3 located in the dermis (Appendix Fig S14C). PTPN11 was most strongly correlated with macrophages in lesional psoriatic skin (Appendix Fig S14D). And the expression of CD68, which was co-expressed with PTPN11, was also upregulated in the skin lesions of psoriatic (Appendix Fig S14E and F). There are many kinds of cells inside each spot, and the differences between the cells in the spots in Cluster 1 and Cluster 3 of normal control and psoriatic patient are obvious. Normal control had only a few macrophages, whereas psoriatic patient had a large infiltration of macrophages (Appendix Fig S14G), which is consistent with the results of Appendix Fig S12A. Strong cellular interactions, especially between macrophages and fibroblasts, were present in Cluster 1 and Cluster 3 of psoriatic patient (Appendix Fig S14H). The above demonstrates that PTPN11 in macrophages plays an important role in psoriasis. Taken together, our data demonstrated that SHP2 deficiency in myeloid cells promoted resistance to IMQ-induced psoriasis in mice.

SHP2 deficiency in macrophages mitigated psoriasis by suppressing NF-κB activation

RNA sequencing of peritoneal macrophages isolated from wild-type and M-Shp2−/− mice and then treated with IMQ for 4 h in vitro revealed cytokines enrichment related to NF-κB signaling (Fig 6A and B). The decrease in several genes in the NF-κB signaling in macrophages from M-Shp2−/− mice was confirmed by quantitative PCR (qPCR; Fig 6C). Comparisons between IMQ-stimulated BMDMs and PMs from wild-type versus M-Shp2−/− mice also demonstrated significantly lower phosphorylation levels of IKKα/β and its downstream p65 in macrophages from M-Shp2−/− mice, specifically after 15–60 min of IMQ stimulation (Fig 6D). However, after IMQ stimulation, p-p38 and p-IRF3 levels were not differentially expressed in the IMQ-stimulated PMs from M-Shp2−/− mice and wild-type mice (Appendix Fig S15). In parallel, analysis of the skin lesions collected from IMQ-induced mice (Fig 6E) and psoriatic patients (Appendix Fig S16) also revealed significantly higher levels of p-p65 protein in the infiltrated cells. Despite the small difference in proportion of p65+ within the CD68+ cells between the normal and psoriasis, both the infiltration of macrophages and the expression of p-p65 were significantly higher in the lesions of psoriatic patients than in healthy individuals, which contributes to the progression of psoriasis. RNA sequencing of peritoneal macrophages confirmed a significant reduction in the mRNA expression of NF-κB signaling-related cytokines in response to IMQ stimulation in vitro due to SHP099 (Fig EV2). This result was consistent with Fig 4. Collectively, these data suggested that the loss of SHP2 in macrophages ameliorated the disease severity by attenuating NF-κB activation. Furthermore, we performed scRNA-seq on unsorted cells from normal skin and lesional psoriatic skin. After performing unsupervised clustering and a UMAP plot analyses (Appendix Fig S17A), cluster identities were determined by the expression of established markers (Appendix Fig S17B), which indicates successful capture of major skin cell subsets. Heatmap showed that PTPN11 was positively more correlated with the NF-κB pathway in macrophages than dendritic cells from psoriasis lesions (Appendix Fig S17C), suggesting that PTPN11 exacerbates psoriasis mainly in macrophages by activating the NF-κB pathway.

Figure 6. SHP2 deficiency in macrophages mitigated psoriasis by suppressing NF-κB activation

Heat map showing mRNA expression in peritoneal macrophages derived from wild-type and M-Shp2−/− mice after IMQ treatment based on RNA sequencing (n = 3/group). Colors represent high (red) and low (blue) intensity. Genes labeled in green indicate they are in the NF-κB signaling. Volcano plot image of upregulated (in red) and downregulated (in green) genes (M-Shp2−/− mice compared to wild-type mice) in peritoneal macrophages from wild-type and M-Shp2−/− mice after IMQ treatment (n = 3/group). Expression levels of representative psoriasis-related genes decreased in the IMQ model of M-Shp2−/− mice compared to wild-type mice (n = 3/group). BMDMs and PMs derived from wild-type and M-Shp2−/− mice were stimulated by IMQ (10 µg/ml) for indicated times. Whole cell lysates were subjected to Western blotting. Representative p-p65 staining of skin sections from mouse. Scale bar: 100 μm.

Data information: Data are represented as mean ± SEM. P values are determined by Tukey multiple-comparison test (C). **P < 0.01, ns, not significant.

Source data are available online for this figure.

Click here to expand this figure.

Figure EV2. SHP2-allosteric inhibitor SHP099 prevented psoriasis by downregulating NF-κB activation

A. Volcano plot image of upregulated genes (red) and downregulated genes (green) (treated with SHP099 compared to untreated with SHP099) from peritoneal macrophages derived from C57BL/6 mice untreated or pretreated with SHP099 (10 µM) for 2 h and then stimulated by IMQ (10 µg/ml) for 4 h (n = 3/group). B, C. Peritoneal macrophages derived from C57BL/6 mice were untreated or pretreated with SHP099 (10 µM) for 2 h and then stimulated by IMQ (10 µg/ml) for 4 h. Expression levels of indicated genes decreased in the SHP099 group compared to medium control.

Data information: Data are represented as mean ± SEM. P values are determined by Tukey multiple-comparison test (B, C). *P < 0.05, **P < 0.01.

Source data are available online for this figure.

SHP2 interacted with TLR7 and promoted TLR7 trafficking to endosomes

To further determine the molecular mechanism underlying SHP2-modulated psoriasis, we identified SHP2-interacting proteins. Specifically, we overexpressed HA-tagged SHP2 in THP-1 cells, immunoprecipitated SHP2 with HA antibody, and performed mass spectrometry. Of the mass spectrometry-identified SHP2-interacting proteins, we focused on TLR7 because it has been reported to participate in the development of psoriasis (Kim et al, 2018). Next, we performed co-immunoprecipitation (co-IP), which confirmed that TLR7 indeed interacted with SHP2 in HEK293T cells overexpressing HA-tagged SHP2 and GFP-tagged TLR7 (Fig 7A). The endogenous interaction between TLR7 and SHP2 was augmented in THP-1 cells (Fig 7B) and BMDMs (Fig 7C) after IMQ stimulation. Agreeing with the co-IP results, confocal microscopy revealed a significant increase in the co-localization of TLR7 and SHP2 in THP-1 cells following IMQ treatment (Fig 7D).

Figure 7. SHP2 interacted with TLR7 and promoted TLR7 trafficking to endosomes

A. HEK293T cells were transfected with plasmids expressing HA-tagged SHP2 and GFP-tagged TLR7. After 48 h, cells were left unstimulated or stimulated by IMQ (10 µg/ml) for 30 min. Cell lysates were immunoprecipitated by anti-HA or anti-GFP and probed by indicated antibodies. B. PMA-differentiated THP-1 cells were unstimulated or stimulated by IMQ (10 µg/ml) for 30 min. Cell lysates were immunoprecipitated by anti-TLR7 or anti-SHP2 and probed by indicated antibodies. C. BMDMs derived from wild-type mice were stimulated by IMQ (10 µg/ml) for 30 min. Cell lysates were immunoprecipitated by anti-SHP2 and probed by indicated antibodies. D. Representative images of confocal microscopy performed with PMA-differentiated THP-1 cells treated with IMQ (10 µg/ml) for indicated durations. Scale bar: 20 µm. E. Immunoblotting of HEK293T cells co-transfected for 48 h with GFP-TLR7, plus HA-SHP2, HA-SHP2 mutant vectors, followed by immunoprecipitation with anti-GFP beads. F. PMA-differentiated THP-1 cells were infected with vector or SHP2 lentivirus and then treated with or without R848 (10 µg/ml). Subcellular fractionation was performed, and cytoplasmic and cell membrane proteins were probed by respective antibodies. G, H. Immunoblot analysis of TLR7 and SHP2 in the isolated ER, Golgi, and endosome from PMA-differentiated THP-1 cells with either SHP2 overexpressed (G) and SHP2 deficiency (H). I. Skin sections from psoriatic patients and healthy controls were immune-stained for TLR7 together with a marker of the early endosome (EEA1) or late endosome/lysosome (LAMP1) prior to analysis by confocal microscopy, showing TLR7 localization in different organelles. Scale bar: 50 μm.

Source data are available online for this figure.

SHP2 possesses two SH2 domains at the N terminus, a PTP domain and a phosphotyrosine-containing tail. To determine the structural basis of its scaffolding function, we generated a series of SHP2 mutants, with the PTP, SH2, or C-terminal tail domain deleted. Deletion of the PTP or SH2 domain almost completely abolished the association of SHP2 and TLR7 (Fig 7E), suggesting that the PTP and SH2 domains of SHP2 interact with TLR7.

Next, we interrogated the molecular effect of SHP2 interaction with TLR7. Resiquimod (R848) is a TLR7/8 agonist that induces consistent effects with IMQ (Appendix Fig S18). Analysis of isolated cytoplasm and cell membranes confirmed the increase in TLR7 on the cell membranes when SHP2 was overexpressed (Fig 7F). Simultaneously, we observed an accumulation of TLR7 on the cell membrane, which was more obvious in cells overexpressing SHP2 and eliminated when SHP2 was knocked down (Appendix Fig S19). Collectively, these data suggested that SHP2 promotes TLR7 localization to the cell membrane.

To obtain further details on the TLR7 membrane localization, we isolated the endoplasmic reticulum (ER), Golgi, endosomes, and plasma membranes from THP-1 cells. SHP2 overexpression caused an increased level of TLR7 only in the endosomes but not in the ER or Golgi (Fig 7G). Conversely, in SHP2-deficient cells, TLR7 decreased in the endosomes (Fig 7H). Detection of subcellular localization of TLR7 in the human skin revealed comparable levels of TLR7 in the ER and Golgi (marked by calreticulin and giantin, respectively) between psoriatic patients and normal controls (Fig 7I and Appendix Fig S20). By contrast, TLR7 was significantly increased in the endosomes of the skin from psoriatic patients, where it was co-located with EEA1 and LAMP1 (Fig 7I). SHP099 also significantly decreased TLR7 in the endosomes, where it co-localized with EEA1 and LAMP1 (Appendix Fig S21). Therefore, the data in Fig 7 indicated that SHP2 promotes the trafficking of TLR7 to the endosomes, particularly in the context of psoriasis.

SHP2 enhanced activation of TLR7/NF-κB signaling in a phosphatase-dependent manner

Given that SHP2 is a tyrosine phosphatase, we queried whether SHP2 dephosphorylated tyrosine residues on TLR7. In HEK293T cells co-transfected with GFP-TLR7 and HA-SHP2 (wild-type, WT) vectors and then stimulated by IMQ, we detected phosphor-tyrosine (p-Tyr) in TLR7 pulled down by anti-GFP (Fig 8A, left lane). When the WT SHP2 was replaced with a gain-of-function mutant (SHP2D61A), the p-Tyr in TLR7 was inhibited (Fig 8A, middle lane); however, the loss-of-function mutant SHP2C459S intensified the p-Tyr in TLR7 (Fig 8A, right lane).

Figure 8. SHP2 enhanced activation of TLR7/NF-κB signaling in a phosphatase-dependent manner

A. Immunoblotting of HEK293T cells co-transfected for 48 h with GFP-TLR7, plus HA-tagged SHP2 or SHP2 mutant vectors, followed by immunoprecipitation with anti-GFP. B, E. Immunoblotting of TLR7 in the endosome isolated from HEK293T cells that were transfected with GFP-TLR7 or GFP-TLR7 mutant vectors. C, F. Immunoblotting of p-p65 in lysates of HEK293T cells transfected with GFP-TLR7 and TLR7 mutant vectors and then treated with IMQ (10 µg/ml) for 30 min. D, G. HEK293T cells transfected with NF-κB-luciferase reporter and indicated TLR7 mutant vectors and then treated with IMQ (10 µg/ml) for 24 h were harvested for luciferase assay. H. Immunofluorescence staining of HEK293T cells transfected with GFP-TLR7 and TLR7 mutant vectors, prior to treatment of R848 (10 µg/ml) for 30 min and labeled with specific antibodies. Scale bar: 50 μm. I. Representative p-TLR7 Y1024 staining of skin sections from human. Scale bar: 100 μm.

Data information: Data are represented as mean ± SEM. P values are determined by two-tailed unpaired Student’s t-test (D, G). **P < 0.01, ns, not significant.

Source data are available online for this figure.

TLR7 has two predicted tyrosine residues (Y807 and Y881) in the linker region (http://www.phosphosite.org/) and four tyrosine residues (Y891, Y897, Y972, and Y1024) in the TIR domain. We thus constructed mutant plasmids of TLR7 by replacing tyrosine with alanine and used them to transfect HEK293T cells with these plasmids. Western blots of endosomal proteins from HEK293T cells transfected with GFP-TLR7 or mutant vectors revealed that several mutations, including Y891A, Y897A, Y972A, and Y1024A, increased TLR7 expression relative to cells expressing wild-type TLR7 (Fig 8B). Immunoblotting assays showed that compared to wild-type TLR7, Y897A, Y972A, and Y1024A mutants enhanced TLR7 responses with p-p65 as a proxy (Fig 8C). NF-κB activation, assessed using a luciferase assay, also increased in the HEK293T cells overexpressing Y897A, Y972A, and Y1024A mutants of TLR7 (Fig 8D).

To further explore the effects of TLR7 phosphorylation on