SHP2 promotes sarcoidosis severity by inhibiting SKP2-targeted ubiquitination of TBET in CD8⁺ T cells

Abstract

Sarcoidosis is an interstitial lung disease (ILD) characterized by IFNγ and TBET dysregulation. Although one-third of patients progress from granulomatous inflammation to severe lung damage, the molecular mechanisms underlying this process remain unclear. Here, we report that pharmacological inhibition of phosphorylated SH2 containing protein tyrosine phosphatase-2 (pSHP2), a facilitator of aberrant IFNγ levels, decreases large granuloma formation and macrophage infiltration in the lungs of mice with sarcoidosis-like disease. Positive treatment outcomes were dependent on the effective enhancement of TBET ubiquitination within CD8⁺ T cells. Mechanistically, our findings identified a previously unknown post-translational modification (PTM) pathway in which the E3 F-box protein S-phase kinase-associated protein 2 (SKP2) targets TBET for ubiquitination in T cells under normal conditions. However, this pathway was disrupted by aberrant pSHP2 signaling in CD8⁺ T cells from patients with progressive pulmonary sarcoidosis as well as end-stage disease. Ex vivo inhibition of pSHP2 in CD8⁺ T cells from end-stage sarcoidosis patients enhanced TBET ubiquitination while suppressing IFNγ and collagen synthesis. Therefore, our functional data provide new mechanistic insights into the SHP2-dependent post-translational regulation of TBET and identify SHP2 inhibition as a potential therapeutic intervention against severe sarcoidosis. Broadly, this study also implicates the possibility of using SHP099 as a therapeutic measure against human diseases linked to TBET or ubiquitination.

One-sentence summary:

Aberrant SHP2 activity in severe sarcoidosis impairs SKP2-mediated ubiquitination of TBET in CD8⁺ T cells.

INTRODUCTION

Adequate production of interferon gamma (IFNγ) is essential for effective immunity, but excessive amounts of this and other pro-inflammatory cytokines can lead to aggressive inflammation to induce immune dysregulation and tissue damage [1]. Accordingly, enhanced production of IFNγ has been proposed as a hallmark of severity in several pulmonary disorders, including sarcoidosis [2]. Sarcoidosis is an idiopathic disease marked by granulomatous inflammation. Amongst the majority of patients who develop pulmonary involvement, one-third will experience progressive remodeling of the lung architecture resulting in severe organ damage and fibrosis [3]. Unfortunately, direct targeting of IFNγ to improve disease outcome has proven ineffective in a spectrum of clinical trials. This emphasizes the urgent need of identifying signaling pathways that facilitate IFNγ dysregulation for designing innovative therapeutics against severe sarcoidosis as well as other IFNγ-driven disorders.

While the role of the non-receptor tyrosine phosphatase Src homolog-2 domain-containing phosphatase 2 (SHP2), encoded by the PTPN11 gene, in eosinophils, fibroblasts, airway epithelial cells and endothelial cells has been explored in inflammatory lung disease, its function in T lymphocytes within this context remains less known [4]. SHP2 blockade was shown to ameliorate the severity of systemic lupus erythematosus by diminishing IFNγ production in double-negative T lymphocytes, but augment CD4⁺ T cell-specific IFNγ in colitis-associated colon carcinoma [5, 6]. Though not in the context of the lung, these seemingly opposing functions emphasize the significance of SHP2 as an essential mediator of IFNγ in T lymphocytes during distinct inflammatory conditions. Likewise, SHP2 activity in fibroblasts promotes lung injury in systemic sclerosis but confers resistance to bleomycin-induced lung damage [7–9], yet how its activity in T lymphocytes impacts severe lung disease is unknown [4]. Previous reports have implicated key regulators of SHP2 activity (e.g., TCR, PD-1) in progressive sarcoidosis, suggesting a likely role for SHP2 in disease pathogenesis [10]. TBET, a member of the T-box family of transcription factors (TF), is another known facilitator of IFNγ production. TBET activity overrides repressive modifications to augment IFNγ expression by directly binding to its promoter [11]. Additionally, concurrent deficiency of TBET and T lymphocytes in mice promotes resistance to bleomycin-induced pulmonary fibrosis [12]. TBET upregulation in pulmonary sarcoidosis has been confirmed by different studies and, analogous to IFNγ, is primarily restricted to T lymphocyte-enriched areas [13, 14]. Although, the molecular mechanisms employed by IFNγ to drive sarcoidosis severity remain unclear, these studies identify SHP2 and TBET as vital players regulating both IFNγ and severe inflammatory lung disease.

To date, a link establishing a direct relationship between SHP2 and TBET has not been reported. In this study, we used a combination of in vivo, in vitro and ex vivo techniques to elucidate the role of SHP2 in severe sarcoidosis. Our data uncovers how phosphorylated SHP2 (pSHP2) modulates ubiquitinated TBET to promote IFNγ production in CD8⁺ T cells from sarcoidosis patients. We found that SHP2 inhibition, using the small molecule inhibitor SHP099, triggers favorable changes in vivo and ex vivo. Specifically, SHP099 treatment reduced large granulomas and infiltrating macrophages in a mouse model of sarcoidosis-like disease by enhancing TBET ubiquitination in CD8⁺ T cells. SHP099 also increased Lysine (K) 48- and K63-linked ubiquitination of TBET while simultaneously normalizing excessive CD8⁺ T cell receptor (TCR)-mediated IFNγ and COL1A1 synthesis ex vivo in end-stage sarcoidosis precision-cut lung slices (PCLS). Thus, the present study reveals the targeting of this pathway as a promising strategy for the prevention or amelioration of severe disease in sarcoidosis patients. Moreover, it provides a better understanding of the post-translational mechanisms regulating TBET in T lymphocytes.

RESULTS

SHP099 inhibits large granulomas in the lungs of Tsc2^fl/flLyz2-Cre mice

The Tsc2^fl/flLyz2-Cre model develops sarcoid-like granulomas independent of exposure to specific foreign antigens, such as those derived from P. acnes or Mycobacterium. This is noteworthy because a unifying etiological factor for sarcoidosis has not been identified. The genetic basis of the model also corresponds with a genetic predisposition to progressive human sarcoidosis [15]; therefore making it relevant to our assessment of progressive disease mechanisms. Thus, to explore the role of SHP2 in sarcoidosis, we treated Tsc2^fl/flLyz2-Cre mice with the small molecule inhibitor SHP099. SHP099 maintains SHP2 in an auto-inhibitory conformation, thereby preventing accessibility to its active site and phosphatase activity [16]. Subsequent to treatment, we harvested the lungs for immunohistochemistry and flow cytometric (FC) analysis (Figure 1a). Our results indicated that SHP099 significantly minimized the total number of macrophage (Mac2⁺) aggregates in the lungs of Tsc2^fl/flLyz2-Cre mice compared to vehicle-treated mice (Figure 1b, c). Interestingly, responders to treatment were mice whom developed larger macrophage aggregates (>0.05 mm²) than non-responder mice (<0.05 mm², Figure 1b, c), suggesting an association between treatment and disease severity. Changes in cell infiltration were confirmed by reduced frequency of F4/80⁺ macrophages in lungs after treatment (SF 1a, b).

Figure 1. Reduction of macrophage aggregates in Tsc2^fl/flLyz2-Cre mice.

a) Schematic of experimental design. b) IHC staining for the macrophage marker Mac2 (red), counter stained with Mayer’s Hematoxylin, in lung tissue sections of untreated Tsc2^fl/fl (n=8) and vehicle- (n=7) or SHP099-treated Tsc2^fl/flLyz2-Cre mice, including responders (n=4) and nonresponders (n=5) to treatment. Second panel depicts large granulomas (>0.05 mm²). c) Number of large granulomas in the lungs of treated and untreated mice. d-e) Representative histograms and quantification of ubiquitinated (Ub) TBET in untreated, vehicle - and SHP099-treated mice. f-h) Representative histograms and dot plots depicting pSHP2 at the activation site Y542 and TBET in CD45⁺ immune cells devoid of CD8⁺ T cells (CD8^- leukocytes) and CD3⁺CD8⁺ T cells (n=6) sorted from Tsc2^fl/flLyz2-Cre mice. i) Pearson’s correlation analysis between TBET ubiquitination within CD8⁺ T cells and large granulomas in lung tissue sections of Tsc2^fl/flLyz2-Cre mice after SHP099 (n=9). Data for all figures expressed as mean ± SEM. Asterisks indicate statistical significance where *P<0.05, **P<0.01 and ***P<0.001.

Inflammation can highjack ubiquitination, a post-translational modification (PTM) mechanism, to regulate protein activity and promote continuous inflammation [17]. To assess whether a relationship exists between SHP2 and TBET, we examined TBET regulatory mechanisms in Tsc2^fl/flLyz2-Cre mice. Positive treatment outcomes were only detected in mice with enhanced TBET ubiquitination levels within CD8⁺ T cells (Figure 1d, e). In its inactive state, SHP2 maintains an intramolecular interaction between its C-terminal phosphotyrosine domain and N-terminal SH2 domain to prevent substrate accessibility to its active site. Phosphorylation at tyrosine 542 (Y542) alleviates this auto-inhibition, causing the protein to unfold and become enzymatically active [18]. Thus, we assessed phosphorylated SHP2 (pSHP2) in CD3⁺CD8⁺ T cells against immune cells devoid of CD8⁺ T cells (CD8^- leukocytes). The frequency of CD8^- leukocytes with SHP2 activity in Tsc2^fl/flLyz2-Cre mice was less than 1% (Figure 1f, h; blue bars), suggesting that pSHP2-dependent regulation of TBET ubiquitination is absent to moderate within other leukocytes compared to CD8⁺ T cells. Further, the percent of CD8⁺ T cells expressing TBET in Tsc2^fl/flLyz2-Cre mice was 8 times greater than in other immune cells (Figure 1g, h; purple bars), indicating that TBET-dependent responses in Tsc2^fl/flLyz2-Cre mice are more pronounced among CD8⁺ T cells. Validating our findings, a strong significant inverse correlation was detected between the number of large macrophage aggregates and TBET ubiquitination within CD8⁺ T cells (Figure 1i). A strong significant inverse correlation was also detected between TBET ubiquitination within CD8⁺ T cells, but not CD8^- leukocytes, and overall granuloma numbers (>0.05mm² and 0.02–0.05mm²) in Tsc2^fl/flLyz2-Cre mice after treatment (SF1c, d). In contrast, no significant associations (or trends toward significance) were detected between TBET ubiquitination levels within CD8^- leukocytes and the number of large granulomas in treated or untreated mice (SF1e). Likewise, no significant correlations were detected in treated or untreated groups for ubiquitinated TBET within CD8^- leukocytes, or CD8⁺ T cells, and percent of infiltrating macrophages in Tsc2^fl/flLyz2-Cre mice (SF1f, g); thus indicating that unlike macrophage aggregation, macrophage infiltration in the lungs of Tsc2^fl/flLyz2-Cre mice is TBET independent.

Since the aforementioned data suggested that SHP099 may not target this pathway in other immune cells, we tested the response of mouse alveolar macrophages to 10–50 μM SHP099. Significant differences in TBET ubiquitination were not detected within alveolar macrophages upon 10–20 μM SHP099 (SF1h). However, exposure to 50 μM SHP099 significantly enhanced the levels of TBET ubiquitination within these cells (SF1h), suggesting that SHP099 may also target macrophage functionality in Tsc2^fl/flLyz2-Cre mice. SHP2 has been shown to activate the mTORC1/S6K1 pathway in skeletal myoblasts [19]. Since Tsc2^fl/flLyz2-Cre mice develop non-caseating granulomas in the lungs due to a myeloid cell-specific deletion of tuberous sclerosis complex 2 (Tsc2)[15], a suppressor of mTORC1 signaling [20], we assessed macrophage functionality in response to SHP099 by evaluating mTORC1 activity through its downstream signaling target and effector ribosomal protein S6 [21]. No significant changes in the phosphorylation of S6 (pS6) were detected in the lungs of responder mice (SF1i, j). Similarly, no significant differences in the number or percent of pS6⁺ macrophages (Mac2⁺) were detected per mm² of lung tissue in responder mice compared to vehicle-treated mice (SF1k–m). Thus, our findings suggest that if SHP099 elicits an effect on alveolar macrophages in Tsc2^fl/flLyz2-Cre mice it is minimal compared to its effects on CD8⁺ T cells.

Collectively, these results suggest that SHP2 regulates the PTM of TBET in CD8⁺ T cells to induce granuloma formation in vivo.

TBX21/TBET is strongest footprint amongst key SHP2-targetted transcriptional regulators within sarcoidosis CD8⁺ T cells

To independently validate our in vivo findings in humans, we used an archived transcriptomic dataset (GSE42832) [22] to identify genes differentially expressed between CD8⁺ T cells from sarcoidosis patients and healthy controls (Table S1). We then retrieved a SHP2 pathway transcriptional signature (Table S2) from the Molecular Signatures Database (MSIGDB) [23] and used the hypergeometric test to evaluate the distribution of this set of genes within sarcoidosis v control CD8⁺ T cell differentially expressed genes. Consistent with the activation of the SHP2 pathway in sarcoidosis CD8⁺ T cells, the SHP2 signature was significantly enriched amongst sarcoidosis CD8⁺ T cell-induced genes (ENR = 4, p = 9E-03) but not repressed genes (Figure 2a; Table S1, column E). Further validating our results, a strong induction of IFNG, in addition to other genes with previously established roles in sarcoidosis severity including JAK3 and STAT1 [24], was noted within CD8⁺ T cells in the sarcoidosis dataset (Figure 2a).

Figure 2. TBX21/TBET has strongest footprint amongst SHP2-targetted transcriptional regulators.

a) Volcano plot showing enrichment per hypergeometric test of SHP2 pathway genes (yellow) among transcripts significantly induced (red) in sarcoidosis v control CD8⁺ T cells. Sources were GSE42832 for sarcoidosis dataset and Molecular Signatures Database for SHP2 pathway genes. See Supplemental Table (ST) 1 for underlying numerical data. b) Regulatory network plot of signaling node ChIP-Seq HCT intersections with SHP2 pathway genes induced in sarcoidosis CD8⁺ T cells. Nodes with strongest and most significant HCT footprints are distributed in top right quadrant. See ST2. c) Pearson’s correlation between percent of CD8⁺ T cells in sarcoidosis patients (blood) and corresponding force vital capacity (FVC, n=39). d) IFNγ (pg/ml) produced by patient CD8⁺ and CD4⁺ T cells without stimulation (basal: CD8, n=21; CD4, n=38) and upon TCR stimulation (stimulated: CD8, n=36, CD4, n=27). e) Flow cytometric (FC) analysis of the mean fluorescent intensity (MFI) of IFNγ-producing CD8⁺ and CD4⁺ T cells isolated from the bronchoalveolar lavage (BAL) of patients (basal: CD8, n=10, CD4, n=10; stimulated: CD8, n=11, CD4, n=11). f) IFNγ (pg/ml) secreted by activated CD8⁺ T cells from the blood of patients with active pulmonary sarcoidosis and continued decline in lung function (progressive sarcoidosis, P; n=17) or patients with inactive disease and normalization of lung function (non-progressive sarcoidosis, NP; n=8). g) Pearson’s correlation between IFNγ (pg/ml) secreted by stimulated CD8⁺ T cells and corresponding FVC (n=11). h-i) Microarray analysis of differentially expressed HLA-A, B, C and TAP1/2 genes in lung explants from patients with non-progressive (n=8) or fibrotic pulmonary sarcoidosis (end-stage sarcoidosis, n=7). j-k) Validation of antigen experienced CD8⁺ T cells in blood and BAL of sarcoidosis (S) patients compared to healthy (Ctrl) cells using the memory maker CD45RO (blood: Ctrl, n=20, S, n=44; BAL: Ctrl, n=4, S, n=36).

High confidence transcriptional target (HCT) intersection analysis identifies regulatory footprints for cellular signaling nodes (e.g., TFs) within a gene set of interest [25–29]. To identify candidate signaling nodes regulated by SHP2 signaling, we subsequently performed HCT intersection analysis on a set of eight genes representing the intersection between MSIGDB SHP2 pathway genes and genes significantly induced within CD8⁺ T cells in sarcoidosis from the GSE42832 dataset (IFNG, STAT1, LMO4, NTRK1, SOS1, JAK3, PTPN11, AFDN). Figure 2b plots the node intersection log odds ratio (OR) versus the intersection log10 (-log10 p), such that nodes with the largest and most significant regulatory footprints are distributed towards the top right. Validating our analysis, we observed strong footprints for NFATC2 and RELB, both previously linked to studies in sarcoidosis [30, 31] as well as SHP2 signaling [32, 33] (Figure 2b, Table S2). The node displaying the strongest confidence intersection with SHP2 pathway genes was TBX21/TBET (OR=35, p=4E-05; Figure 2b, Table S2), highly suggestive of a SHP2-TBET pathway within sarcoidosis CD8⁺ T cells. Providing further validation for this finding was a significant footprint for PRDM1/Blimp, a well characterized TBET-associated protein in CD8⁺ T cells [34] (Figure 2b, Table S2).

CD8⁺ T cell receptor (TCR)-mediated IFNγ production and induction of antigen processing molecules in severe sarcoidosis

Because effective SHP099 treatment responses in Tsc2^fl/flLyz2-Cre mice resulted from significant changes in TBET ubiquitination within CD8⁺ T cells, we explored a potential relationship between CD8⁺ T cells and sarcoidosis severity. Our data revealed a positive correlation between the percent of CD8⁺ T cells and the force vital capacity (FVC, a measure of lung function) of patients (Figure 2c), suggesting reduced numbers of CD8⁺ T cells are disadvantageous to patients and increased numbers are associated with higher lung function. An enrichment of CD8⁺ TCR-signaling pathway genes has been associated with pulmonary sarcoidosis [35]. SHP2 has also been shown to regulate the production of IFNγ [36]. We, therefore, evaluated a potential relation between TCR-signaling in CD8⁺ T cells and disease severity in sarcoidosis. For an accurate assessment, we examined samples derived from the blood as well as the lungs. Consistent with previous observations using PMA/Ionomycin stimulation [37–40], we found that TCR-activated patient CD8⁺ T cells produced greater amounts of IFNγ than healthy CD8⁺ T cells (Ctrl, SF2a); but no differences were detected under basal (unstimulated) conditions (SF2a). No differences in IFNγ were observed in cells stimulated with either anti-CD3 alone or in combination with anti-CD28 (SF2b, c). However, in contrast to other stimuli [37–40], TCR engagement induced 3-times more IFNγ in patient CD8⁺ T cells (mean=13,929) than CD4⁺ T cells (mean=4,682; Figure 2d). These observations paralleled our findings in cells from the bronchoalveolar lavage (BAL) of patients (Figure 2e). Additionally, CD8⁺ T cells isolated from subjects with active pulmonary disease and continued decline in lung function, i.e. progressive patients (P), produced greater levels of IFNγ than cells isolated from patients with FVC normalization, i.e. non-progressive patients (NP) (Figure 2f). Importantly, our findings demonstrated a significant inverse correlation between the amount of CD8⁺ TCR-induced IFNγ and FVC (Figure 2g), implying increased disease severity in subjects exhibiting high levels of CD8⁺ T cell-derived IFNγ.

IFNγ induces the expression of MHC-I and associated peptide transporters (i.e. TAP1 and TAP2) [41, 42]. We therefore examined the expression of these molecules in end-stage sarcoidosis. Microarray expression analysis in transbronchial biopsies from patients with fibrotic sarcoidosis (end-stage sarcoidosis) revealed a significant increase in HLA-A, HLA-B, HLA-C, TAP1 and TAP2 genes compared to biopsies from patients with non-progressive sarcoidosis (Figure 2h, i) [43–46]. To validate these findings, we assessed the levels of antigen-experienced CD8⁺ T cells and the chronic activation marker programmed death-1 (PD-1). Our data indicated an increased frequency of CD8⁺ T cells expressing the memory cell marker CD45RO in the blood and BAL of sarcoidosis patients (Figure 2j, k). Likewise, PD-1 surface expression on patient CD8⁺ T cells was significantly elevated (SF2d). Collectively, these data support a role for a TCR-induced IFNγ signature in patient CD8⁺ T cells that may instigate a worsened disease prognosis.

Enhanced phosphorylated SHP2 in patients with severe sarcoidosis

Since our in vivo data suggested a role for SHP2 in sarcoidosis, we examined its levels in patients. No differences in the expression of PTPN11 were detected upon TCR stimulation between healthy (Ctrl) and sarcoidosis patient CD8⁺ T cells isolated from the blood (Figure 3a). In contrast, ELISA assay revealed elevated levels of pSHP2 in cell lysates of activated patient CD8⁺ T cells compared to controls (Figure 3b). Akin to the findings in Figure 2, no differences in pSHP2 were detected in CD8⁺ T cells stimulated with anti-CD3 only or a combination of anti-CD3 and -CD28 (SF3a, b). Our data also indicated that patient CD8⁺ T cells with pSHP2 expressed greater amounts of IFNγ (Figure 3c) with relevance to active pulmonary disease and continued decline in lung function (Figure 3d). These data suggested a link between pSHP2 and end-stage sarcoidosis. To test this, we examined the levels of pSHP2 in fibrotic, end-stage sarcoidosis lung explants by immunohistochemistry (IHC) and western blotting. Enhanced pSHP2 was detected in sarcoidosis lungs compared to control, non-diseased lungs (Figure 3e, f), suggesting this phenotype carries over to areas of severe pulmonary disease. Immunofluorescence staining to visualize SHP2 and CD8 co-localization in end-stage sarcoidosis lung samples demonstrated co-staining within areas of cell aggregation (Figure 3g). To summarize, we found an increase in phosphorylated SHP2 in sarcoidosis patient CD8⁺ T cells upon TCR engagement and its activity appears to be relevant to disease progression and severity.

Figure 3. Enhanced SHP2 activity in sarcoidosis.

a) Relative PTPN11 mRNA normalized to β-actin in activated CD8⁺ T cells from healthy (Ctrl, n=6) and sarcoidosis (S, n=9) subjects (blood). b) ELISA analysis of intracellular pSHP2 (Y542) in healthy (Ctrl, n=10) and sarcoidosis (S, n=8) CD8⁺ T cell lysates (blood) upon TCR stimulation. Graph displaying pSHP2/total SHP2 protein ratios. c) IFNγ expression in patient CD8⁺ T cells with (+) and without (−) active SHP2 (n=10). d) Percent and MFI for pSHP2 (Y542) in stimulated CD8⁺ T cells isolated from the blood of patients (P, n=17; NP, n=7). e) IHC staining for pSHP2 in non-diseased controls (n=2) and end-stage sarcoidosis (n=3) lung explants. Images at 20x magnification. f) Immunoblot of pSHP2 at Y542 (normalized to both β-actin and SHP2) in non-diseased (Ctrl, n=7) and end-stage sarcoidosis (S, n=4) lung tissues. g) IF co-staining (yellow-orange) of SHP2 (green) and CD8α (red) within end-stage sarcoidosis lung explants (scale bar=100 μm). h) 0 (n=11), 20 (n=11) and 50 (n=7) μM SHP099 administration on activated CD8⁺ T cell-specific IFNγ (blood of patients).

pSHP2-dependent regulation of the CD8⁺ T cell phenotype in sarcoidosis patients

Since our results suggested that patient CD8⁺ T cells with phosphorylated SHP2 were primarily responsible for the production of IFNγ (Figure 3c), we evaluated whether induction of IFNγ in these cells is dependent on SHP2 activation. Pharmacological inhibition of SHP2 activity in patient CD8⁺ T cells with SHP099 led to a significant decline in IFNγ (Figure 3h). To test whether this phenotype could be induced in T cells via SHP2 manipulation, Jurkat T cells (JTCs) were transfected with a GST-tagged SHP2 plasmid and then subjected to TCR stimulation. Effective transfection was confirmed by RT-qPCR (SF3c). Subsequently, enhanced pSHP2 was noted simultaneously with increased IFNγ in JTCs (SF3d, e). These results suggest that excessive pro-inflammatory responses exhibited by TCR-activated patient CD8⁺ T cells is pSHP2-dependent, which is amenable to therapeutic targeting.

SHP2-mediated regulation of TBET ubiquitination in sarcoidosis CD8⁺ T cells

Our in vivo findings implied a role for SHP2 in the PTM of TBET, a direct facilitator of IFNγ production [11]. Hence, to elucidate the mechanism by which SHP2 regulates IFNγ in activated patient CD8⁺ T cells we examined TBET levels in relation to SHP2 activity. Our results indicated elevated TBET in patient CD8⁺ T cells expressing pSHP2 (Figure 4a), an effect that was negated following SHP099 administration (Figure 4b). Elevated TBET in activated CD8⁺ T cells isolated from the blood of patients was coupled with lower levels of ubiquitinated TBET (Figure 4c), implying enhanced TBET protein levels are the result of impaired PTM of TBET. To validate that reduced TBET ubiquitination in patient CD8⁺ T cells was the result of SHP2 activity, we assessed the effect of SHP099 on ubiquitinated TBET. Our results revealed a dose-dependent increase in ubiquitinated TBET in patient CD8⁺ T cells following inhibition of SHP2 phosphatase activity (Figure 4d).

Figure 4. Direct SKP2-TBET protein interactions in sarcoidosis.

a) Percent TBET in patient CD8⁺ T cells (blood) with (+) and without (−) active SHP2 (n=10). b) TBET in activated CD8⁺ T cells isolated from blood of patients upon SHP099 (0, 20 μM: n=18; 50 μM, n=7). c) Percent of ubiquitinated TBET in CD8⁺ T cells from the blood of healthy (Ctrl, n=10) and sarcoidosis (S, n=18) subjects after TCR stimulation. d) TBET ubiquitination restored in TCR-stimulated patient CD8⁺ T cells (blood) following SHP099 (0, 20 μM, n=18; 50 μM, n=7). e) FC analysis of SKP2 in activated CD8⁺ T cells (blood; Ctrl, n=7; S, n=12). f-g) Representative images for IHC staining and quantification of SKP2 positive cells (brown) in non-diseased (control) and end-stage sarcoidosis lung explants (n=2). Images at 4x magnification. h) Microarray analysis of SKP2 in patients with non-progressive (n=8) or fibrotic, end-stage sarcoidosis (n=7). i) Co-immunoprecipitation (co-IP) analysis of endogenous SKP2, TBET and SHP2 proteins in activated CD8⁺ T cells from blood of patients (n=3). j) IF co-staining (yellow) of SKP2 (green) and TBET (red) in the nucleus and quantification of positive cells in control (non-diseased) and end-stage sarcoidosis lung explants (n=2; scale bar=20 μm).

The ubiquitin proteasome system (UPS) is a PTM mechanism involving a series of reactions catalyzed by 3 enzymes (i.e., E1-E3). E3s mediate the final step by recognizing and catalyzing the transfer of ubiquitin to target proteins [47]. Based on findings from a prior investigation [48], we identified aberrancies in the E3 ubiquitin ligase SKP2 in T cells. It is currently unknown whether SHP2 plays a role in the post-translational regulation of SKP2. Since SKP2 levels have not been reported in sarcoidosis, we assessed its expression in severe disease. Low SKP2 in TCR-activated, blood-derived patient CD8⁺ T cells (Figure 4e) corresponded with a significant decline in SKP2 and SKP2 in end-stage sarcoidosis lung explants (Figure 4f–h). Our data also demonstrated a dose-dependent reduction of SKP2 in SHP2 transfected JTCs (SF3f) that paralleled a dose-dependent increase in TBET (SF3g).

SHP2 inhibits SKP2 E3 ligase activity to extend TBET stability

SKP2 has been shown to promote chronic inflammation in cancer cells by targeting TFs for ubiquitin-mediated degradation [49, 50], but whether SKP2 mediates the ubiquitination of TFs in immune cells remains unexplored. Endogenous co-IP (i.e., no introduction of plasmids) confirmed SKP2 and TBET protein interactions were present physiologically in JTCs (SF4a–c). Probing for SKP2 in Jurkat T whole cell lysates (WCL) demonstrated that SKP2 co-precipitated with TBET (SF4a), suggesting direct SKP2-TBET protein interactions. These interactions were confirmed with the inverse co-IP (SF4b). Our data also indicated analogous protein interactions between SHP2 and SKP2 (SF4c). Most importantly, endogenous co-IP revealed direct protein contact between SHP2 and SKP2 as well as SKP2 and TBET in activated patient CD8⁺ T cells (Figure 4i). Immunofluorescence analysis in control, non-diseased and end-stage sarcoidosis lung explants also detected SKP2-TBET interactions (Figure 4j, yellow). Higher co-localization was detected between SKP2 and TBET in control than sarcoidosis lungs (Figure 4j, white arrow heads), suggesting that SKP2-TBET interactions are a normal occurrence under physiological conditions.

SKP2 exerts its function through its E3 ubiquitin ligase activity [51–53]. Since an interaction between SKP2 and TBET has not been reported, we assessed the effect of SKP2 E3 ligase activity on TBET. JTCs were transfected with a combination of TBET, ubiquitin, wild type (WT) SKP2, mutant SKP2 (ΔF) or empty vector (EV) plasmids. In the presence of WT SKP2 with normal E3 ligase activity, TBET protein levels were significantly reduced by about 75% (red line) from control conditions (EV, blue line, SF5). In contrast, in the presence of a mutant form of SKP2 (ΔF, purple line) only lacking E3 ligase activity, TBET protein levels resembled EV controls (SF5) suggesting SKP2 E3 ligase activity has a direct impact on TBET protein levels. To evaluate how SHP2 affects this interaction, we also transfected JTCs with TBET, ubiquitin, WT SHP2 and WT SKP2. The inclusion of WT SHP2 together with WT SKP2 (green line) prevented TBET protein level reductions previously observed in conditions with WT SKP2 plasmids (red, ~75%) and mimicked the TBET levels observed for SKP2 ΔF (SF5). This data suggests that akin to SKP2 ΔF, SHP2 impedes SKP2 E3 ligase activity.

TBET is a substrate for K48- and K63-linked SKP2-mediated polyubiquitination

To further characterize SKP2 and TBET interactions, we assessed whether TBET is a substrate for SKP2-mediated ubiquitination. We co-transfected HEK293T cells with plasmids expressing Myc-SKP2, Flag-TBET, HA-ubiquitin or an empty vector (EV) (SF6a). Our results indicated increased ubiquitination of TBET in the presence of SKP2 (SF6a, 4^th row) in contrast to cells co-transfected with either EV, Flag-TBET or HA-ubiquitin only (SF6a, 1^st-3^rd rows, respectively). These results suggest that TBET is a target for SKP2-mediated ubiquitination. We then transfected HEK293T cells with plasmids containing either WT or mutant (K48 or K63) ubiquitin. Each mutant ubiquitin plasmid carried one lysine residue at the specified position. Analysis of conventional residues in the ubiquitin molecule identified K48 and K63 as the functionally significant residues for SKP2-mediated ubiquitination of TBET (SF6a, 6^th and 10^th rows, respectively), since ubiquitination of TBET still occurred in the presence of K48- and K63-only mutants (SF6a). To better assess the relevance of K48 and K63 in the ubiquitination of TBET, we generated KR mutants in which lysine was replaced with arginine at the specified position (K48R and K63R, SF6a). Despite the presence of SKP2, we found that both K48R and K63R mutants attenuated SKP2-induced TBET ubiquitination (SF6a, 8^th and 12^th rows), suggesting that both K48 and K63 are indispensable for SKP2-mediated ubiquitination of TBET. To assess whether K48- and K63-linked ubiquitination mediated the proteasomal or lysosomal degradation of TBET, we tested TBET protein changes in response to MG132 (proteasome inhibitor) and chloroquine (lysosome inhibitor). Since our data suggests this process is significantly disrupted in patient CD8⁺ T cells, samples from lungs of control subjects were used for analysis. MG132 treatment enhanced the levels of TBET in TCR-stimulated CD8⁺ T cells (SF6b), whereas exposure of cells to chloroquine did not elicit changes (SF6c), indicating that TBET is degraded via the ubiquitin-proteasome pathway.

Inhibition of SKP2 E3 ligase activity is essential for SHP2 induction of IFNγ

To assess the functional effects of SKP2 in this pathway, we developed a SKP2 knockdown (KD) Jurkat T cell line by small interfering RNA (siRNA). SKP2 KD increased IFNγ production (SF7a). The efficacy of the SKP2 KD was confirmed by RT-qPCR and western blotting (SF7b, c). High IFNγ in JTCs corresponded with elevated TBET levels (SF7c, d, blue bars). We next evaluated the effect of blocking SKP2 E3 ubiquitin ligase activity on IFNγ and TBET ubiquitination in primary cells. Healthy CD8⁺ T cells were treated with the small molecule inhibitor SZL P1-41 which binds to the F-box domain of SKP2, thereby blocking the formation of the SKP2-SCF complex which is essential for executing its E3 ligase activity. SZL P1-41 treatment enhanced IFNγ (SF7e) and increased TBET by suppressing the amount of ubiquitinated TBET in healthy CD8⁺ T cells (SF7f, g). Taken together, these data suggest that SKP2 controls the expression of IFNγ through direct TBET ubiquitylation.

Because SHP2 is required to induce the activated phenotype in patient CD8⁺ T cells (Figure 3), we next sought to assess whether SKP2 inactivation was essential for SHP2 to execute its functional effects. We first evaluated SKP2 total protein and E3 ubiquitin ligase activity in JTCs in response to SHP2 overexpression. We found a significant decline in total SKP2 with a concurrent increase in p27, a direct downstream target of SKP2 E3 ubiquitin ligase activity (SF8a). We then transfected JTCs with Myc-SKP2. After confirming elevated total SKP2 and decrease p27 (SF8b, c) by western blotting as well as increased SKP2 mRNA (SF8c), we proceeded to reevaluate IFNγ and TBET. In contrast to JTCs with SKP2 KD or inhibition of SKP2 E3 ligase activity (SF7), IFNγ was significantly diminished in JTCs with enhanced SKP2 (SF8d). To ascertain SKP2 regulation of IFNγ through TBET, we also quantified TBET protein levels. Western blot indicated dose-dependent reductions in TBET (SF8b, e). This was further confirmed by FC (SF8f). Collectively, our data suggest that induction of excessive IFNγ in CD8⁺ T cells is the result of SHP2-dependent inhibition of SKP2 E3 ubiquitin ligase activity.

Ex vivo SHP099 treatment restores excessive CD8⁺ T cell effector responses concomitantly with COL1A1 expression

To assess the role of SHP099 in a setting that closely resembles in vivo human conditions, we used precision-cut lung slices (PCLS) obtained from non-disease control and fibrotic, end-stage sarcoidosis lung explants. PCLS are 3-dimensional living tissue slices that retain cell-cell proximity and interactions observed in the human lung microenvironment. Analogous to our observations in vitro, FC analysis indicated reduced percentages of CD8⁺ T cells in our patient cohort (Figure 5a, b). Increased pSHP2 was also detected in CD8⁺ T cells, but not CD4⁺ T cells, within sarcoidosis PCLS (Figure 5b). These results suggests that, at least within the context of sarcoidosis, this pathway is specific to CD8⁺ T cells and inconsequential to CD4⁺ T cells. Immunoblotting indicated significantly elevated TBET protein levels in sarcoidosis samples (Figure 5c). Administration of 50 μM of SHP099 significantly increased K48- and K63-linked ubiquitinated TBET (Figures 5d–g) and reduced CD8⁺ T cell-specific IFNγ by about 50% within CD8⁺ T cells, without affecting CD4⁺ T cell-derived IFNγ, in sarcoidosis PCLS (Figure 5h) while simultaneously normalizing extracellular COL1A1 synthesis (Figure 5i). Assessment of SHP099 on SHP2 pathway-related proteins in activated patient CD8⁺ T cells revealed a 50% reduction in pSHP2 and TBET (Figure 5j). In contrast, SKP2 protein was significantly enhanced (Figure 5j) whereas low p27 protein levels were detected in CD8⁺ T cells following SHP099 administration (Figure 5j). Further, 0, 10, 20 and 50 μM SHP099 exposure did not elicit significant changes in TBET ubiquitination within human lung fibroblasts or myofibroblasts (Figure 5k). Since the majority of these cells indicated a TBET ubiquitination level of 0% (Figure 5k), the MFI was n/a indicating a lack of shifting in fluorescence intensity. This data indicates that therapeutic efficacy ex vivo was not a result of direct targeting of TBET ubiquitination within fibroblasts or myofibroblasts.

Figure 5. Concurrent reduction of CD8⁺ T cell-specific IFNγ and normalization of COL1A1 in fibrotic, end-stage sarcoidosis PCLS.

a) Frequency of CD3⁺CD8⁺ T cells in control, non-diseased (Ctrl, n=6) and end-stage sarcoidosis (S, n=4) precision cut lung slices (PCLS). b) FC analysis of pSHP2 in CD4⁺ and CD8⁺ T cells within PCLS (Ctrl, n=6; S, n=4). Each dot in a and b represents data averaged from 3–6 PCLS experimental replicates gathered from a total of 3 independent experiments. c) Immunoblot analysis of normalized TBET protein to β-actin (Ctrl, n=4; S, n=3). d-g) MFI of Lysine (K)48- and K63-linked ubiquitination of TBET within CD8⁺ T cells in untreated (NT), vehicle (DMSO)- or SHP099-treated sarcoidosis PCLS. Matching shapes in f and g indicate PCLS replicates from the same sarcoidosis donor (n=3). h) Percent of CD4⁺ and CD8⁺ T cell-specific IFNγ in treated sarcoidosis PCLS. i) Normalization of COL1A1 mRNA to GAPDH in non-disease control (ctrl, n=3) and sarcoidosis (n=4) PCLS after SHP099. To account for variability, each dot in i represents an average of 3 experimental replicates from 3 independent experiments. j) Immunoblot images and quantification for patient CD8⁺ T cells treated with SHP099 and probed for pSHP2, TBET, SKP2, p27 (direct downstream target of SKP2 E3 ligase activity) and β-actin (n=5). k) Percent TBET ubiquitination within human lung fibroblasts and myofibroblasts (24h stimulation with TGF-β) exposed to 10–50 μM of SHP099 (n=5).

Collectively, our results suggest that pSHP2 regulates the production of CD8⁺ T cell-specific IFNγ in severe sarcoidosis to promote lung injury in patients and imply SHP099 as a potential treatment of severe granulomatous disease.

In vivo validation of the SHP2/SKP2/TBET pathway in Tsc2^fl/flLyz2-Cre mice

Substantiating our findings in humans, the lungs of Tsc2^fl/flLyz2-Cre mice displayed significantly fewer CD8⁺ T cells than Tsc2^fl/fl mice (Figure 6a). Similarly, our results indicated enhanced PD-1 receptor expression and elevated SHP2 activity in CD8⁺ T cells from Tsc2^fl/flLyz2-Cre mice (Figure 6b, c). Further, CD8⁺ T cells from Tsc2^fl/flLyz2-Cre mice with SHP2 phosphatase activity expressed greater amounts of TBET and IFNγ than cells with inactive SHP2 (Figure 6d, e). IF detected enhanced ubiquitin molecules within areas of SKP2 and TBET co-localization (Figure 6f, g; orange arrow heads), while IHC staining indicated SKP2 enrichment (Figure 6h), within lung tissue sections of Tsc2^fl/flLyz2-Cre mice after SHP099 administration. Thus, our results indicate that aberrant SHP2 signaling within CD8⁺ T cells plays a significant role in the formation of lung granulomas.

Figure 6. In vivo validation of pSHP2-dependent processes in CD8⁺ T cells.

a) Percent CD8⁺ T cells in the lungs of Tsc2^fl/fl mice (n=8) and Tsc2^fl/flLyz2-Cre mice (n=7). b) PD-1 surface expression (n=7) and c) SHP2 activity within Tsc2^fl/flLyz2-Cre mice (n=7) and Tsc2^fl/fl mice (n=8) with corresponding histograms. d-e) TBET and IFNγ expression in CD8⁺ T cells from Tsc2^fl/flLyz2-Cre mice with SHP2 phosphatase activity and inactivity (n=7). f-g) IF images demonstrating ubiquitin intensity (orange arrow heads) in the lungs of vehicle- or SHP099-treated Tsc2^fl/flLyz2-Cre mice (n=4; scale bar=25 μm) h) IHC staining for SKP2 in lung tissue sections of vehicle- (n=7) and SHP099-treated Tsc2^fl/flLyz2-Cre mice (n=4).

DISCUSSION

The aim of our study was to investigate the function of SHP2, a moderator of inflammation, in severe sarcoidosis by assessing the clinical utility of SHP099. Since granulomas in end-stage sarcoidosis are associated with fibrogenesis, we assessed the potential clinical utility of SHP099 by evaluating its effects on granuloma formation and collagen synthesis. Figures 1b, c demonstrated nearly complete inhibition of large granulomas by SHP099 in Tsc2^fl/flLyz2-Cre mice; whereas Figure 5i confirmed the normalization of collagen in end-stage sarcoidosis PCLS after SHP099 administration. Our in vivo data indicated that inhibition of large granulomas after treatment resulted from specific TBET PTMs within CD8⁺ T cells. Independent validation in human sarcoidosis corroborated these findings by identifying TBX21/TBET as having the strongest footprint amongst a key set of induced SHP2-transcriptional regulators within CD8⁺ T cells. Although TBET protein stability is controlled by proteasomal degradation [54], the precise E3 ubiquitin ligase responsible for TBET ubiquitylation has not been identified. Our results indicated that SKP2 E3 ubiquitin ligase activity mediates K48-/K63-linked polyubiquitination of TBET in normal, healthy T cells. However, SHP2 hyperactivity in patient CD8⁺ T cells impedes this process to promote continued expression of IFNγ. Conversely, SHP099 treatment enhanced TBET ubiquitination while concomitantly minimizing IFNγ expression in CD8⁺ T cells. Thus, our findings revealed a previously-unknown mechanism for the regulation of TBET that is governed by SHP2. Consequently, in addition to supporting the use of the TSC2^fl/flLyz2-Cre model to investigate progressive disease-relevant mechanisms, our study carries therapeutic implications for targeting pSHP2 in severe sarcoidosis and provides mechanistic insight into a signaling pathway for the PTM of TBET.

Paralleling our findings in vivo, activated CD8⁺ T cells from end-stage pulmonary sarcoidosis patients displayed elevated SHP2 activity. These cells also expressed greater amounts of TBET and IFNγ than CD8⁺ T cells with SHP2 inactivity, and CD4⁺ T cells. Our results coincide with previous observations that sarcoidosis CD4⁺ T cells display an exhausted phenotype and therefore produce minimal quantities of IFNγ upon TCR stimulation [10]. Although not in the context of granulomatous disorders, several published studies have linked SHP2 to biomarkers currently used in the diagnosis of sarcoidosis, thus supporting a potential role for enhanced SHP2 activity in sarcoidosis. For example, the angiotensin-converting enzyme (ACE), elevated in the sera of sarcoidosis patients [55], was shown to trigger the activation of SHP2 in vascular smooth muscle cells [56]. Similarly, interleukin-2 (IL-2) induces the phosphorylation of SHP2 through IL-2 receptor signaling, a marker associated with sarcoidosis activity [57], in the T cell leukemia cell line ILT-Mat [58]. Further, while an overall significant decline in CD8⁺ T cells was not detected in the blood of all patients, a significant correlation between reduced frequencies of CD8⁺ T cells and worse lung function was noted. Thus, given that fewer CD8⁺ T cells were present in the lungs of Tsc2^fl/flLyz2-Cre compared to Tsc2^fl/fl mice validates that this occurrence is physiologically relevant to granuloma formation. Previous reports have supported a plausible role for CD8⁺ T cells in sarcoidosis pathophysiology. For example, an overall reduced presence of CD8⁺ T cells in the lungs of sarcoidosis versus Löfgren’s Syndrome, an acute manifestation of the disease, was reported. The same investigation found more common variants between blood- and lung-derived CD8⁺ T cells than CD4⁺ T cells, suggesting a CD8⁺ T cell-specific genetic predisposition to unremitting sarcoidosis [59]. Additionally, the enhanced presence of perforin-, granzyme B- and granulysin-producing CD8⁺ T cells in sarcoidosis subjects compared to controls and patients with Löfgren’s Syndrome was also reported [60]. HLA-DR⁺ CD8⁺ T cells in sarcoidosis BAL has been proposed as a marker to improve diagnostic accuracy [61]. Collectively, our findings together with the aforementioned published reports support an activated CD8⁺ T cell phenotype in severe sarcoidosis that has physiological relevance to granuloma formation.

Given the preceding, the discovery that pSHP2 induced excessive functionality in patient CD8⁺ T cells displaying high levels of an exhaustion marker at first seemed contradictory. Further confounding these observations are previous studies demonstrating that TBET directly represses PDCD1 and is therefore downregulated, and inversely proportional to PD-1, in exhausted CD8⁺ T cells [62]. Yet, our data consistently indicated high levels of TBET in patient CD8⁺ T cells upon TCR stimulation. Fortunately, during the course of this study, Collier et al. published an elegant report contrasting the phenotypic and genotypic similarities that exist in chronically stimulated CD8⁺ T cells across conditions [63]. Although chronically activated CD8⁺ T cells across diseases share common features (e.g., PD-1 upregulation), the persistence of antigen as a result of chronic inflammation in autoimmune disorders promotes the generation of autoreactive CD8⁺ T cells that breach tolerance mechanisms [64]. This, in turn, facilitates the release of high amounts of effector molecules (e.g., IFNγ, granzyme B, perforin) that subsequently injure the host’s own tissues [63]. Our results indicated that CD8⁺ T cells from the blood and lungs of patients are antigen-experienced. This and elevated PD-1 receptor expression imply that these cells have been chronically exposed to antigen. Previous reports have identified autoantigens in the BAL and sera of patients as well as imbalanced proportions of B and T_FH cells in pulmonary sarcoidosis [65, 66]. Further, dysregulation of TBET, in spite of its levels being up- or downregulated, is a key inducer of autoreactive CD8⁺ T cells [67, 68]. Taken together, our findings in conjunction with previous reports introduces the plausibility of pSHP2 driving the production of autoreactive CD8⁺ T cells through TBET dysregulation in severe sarcoidosis. We are currently investigating this hypothesis.

Our results also revealed an upregulation of HLA-A, B and C as well as TAP-1 and TAP-2 genes in fibrotic, end-stage sarcoidosis. This year an enrichment for HLA-A, -B and –C was reported in “immune-interacting fibroblasts” amongst sarcoidosis granulomas [69]. Due to the specificity of these molecules for CD8⁺ T cells, our results together with the latter study emphasize the significance of CD8⁺ T cells in severe sarcoidosis. Further, the capacity of IFNγ to modulate the expression of MHC-I molecules has been known since the 1990s [41]. More recently, it was demonstrated that SHP2 activity in tumor cells deregulates the levels of MHC-I through an IFNγ-dependent pathway [36]. IFNγ induction of MHC-I-associated peptide transporters, such as TAP1, in macrophages was shown to be STAT1-dependent, as STAT1 knockout mice fail to induce TAP1 expression even in the presence of IFNγ [42]. An abundance of STAT1 in the lungs and lymph nodes of sarcoidosis patients has been reported [70]. More importantly, JAK/STAT signaling was recently identified as a molecular signature for sarcoidosis severity [24]. Therefore, augmented levels of each of these factors (i.e., SHP2, IFNγ, STAT1, MHC-I, TAP1/2) in end-stage sarcoidosis could be the result of an interconnected molecular network working to enhance the recruitment of CD8⁺ T cells in severe disease. Given these observations, another unexpected finding was that a decline in CD8⁺ T cells was observed in lung explants of end-stage pulmonary sarcoidosis patients. Future investigations to clarify the genetic basis of this deficiency are warranted, and whether restoring CD8⁺ T cell numbers can detain or reverse severe disease in patients. In the abovementioned study, SHP2 inhibition in tumor cells augmented anti-tumor immunity by enhancing the recruitment of cytotoxic T cells [36]. Future examinations will assess whether enhanced SHP2 activity in sarcoidosis impacts the migration and recruitment of CD8⁺ T cells.

Endogenous co-IP analysis indicated direct SHP2-SKP2 interactions within patient CD8⁺ T cells. Direct SHP2-SKP2 interactions were further validated in JTCs. Inclusion of WT SHP2 together with WT SKP2 plasmids in JTCs prevented TBET protein level reductions to those previously observed in conditions with WT SKP2 alone (≤75%) and mimicked the TBET levels observed for a mutant form of SKP2 only lacking E3 ligase activity (SKP2 ΔF). This data suggests that akin to SKP2 ΔF, SHP2 directly impedes SKP2 E3 ligase activity. Further, SHP099 treatment in patient CD8⁺ T cells significantly enhanced SKP2 protein levels while reducing its downstream E3 ligase target, p27; and SHP099 administration in vivo increased SKP2 protein levels in the lungs of Tsc2^fl/flLyz2-Cre mice. In contrast, SHP2 overexpression in JTCs led to significant declines in SKP2 while simultaneously increasing p27. Thus, this data suggests that in addition to directly inhibiting SKP2 E3 ligase activity, SHP2 negatively impacts the total protein levels of SKP2 within CD8⁺ T cells. Recent discoveries have indicated that PTMs, such as phosphorylation, can regulate the ubiquitin code by directly targeting ubiquitin molecules and/or machinery [71]. Thus, understanding the precise molecular binding dynamics of SHP2 to SKP2 and its effect on ubiquitin is essential for determining whether stabilizing or destabilizing such interactions would be beneficial in human diseases linked to ubiquitination [71]. Hence, future studies will evaluate the structural binding dynamics of SHP2-SKP2 and its effects on ubiquitin molecules in drug design.

Direct SKP2 and TBET interactions have not been previously reported. Immunofluorescence-based methods demonstrated co-localization between SKP2 and TBET in non-diseased lungs, implying these protein interactions may be physiological to different cell types and pertinent to various conditions. In contrast, reduced co-localization was noted in end-stage sarcoidosis lung explants, suggesting SKP2-TBET interactions are impaired in patients with severe disease. Endogenous co-IP analysis validated these results by showing direct SKP2-TBET interactions in patient CD8⁺ and JTCs. Further analysis demonstrated that SKP2 exerted its function on TBET through its E3 ubiquitin ligase activity. In vivo analysis also indicated co-localization of SKP2 and TBET within areas of high ubiquitin intensity after SHP099 treatment. Additionally, our results demonstrated that deficiency or pharmacological inactivation of SKP2 E3 ubiquitin ligase activity with SZL P1-41 stimulated high levels of TBET and IFNγ in JTCs and primary CD8⁺ T cells, whereas induction of SKP2 had an opposing effect. Coinciding with our findings, SKP2 was recently found to interact with USP18, a negative regulator of type I and III interferons, in HeLa cells [72]; and although IFNγ is a type II interferon, our results suggest that SKP2 may partake in the regulation of all interferons.

Our investigation revealed that SKP2-dependent ubiquitination of TBET was through K48-/K63-branched linkages. However, a significant decline in K48- and K63-linked ubiquitinated TBET was detected ex vivo in patient CD8⁺ T cells with SHP2 hyperactivity compared to non-diseased lungs, indicating that this process is impaired in patients with severe sarcoidosis. The function of K48-/K63-branched linkages was first described in the context of robust inflammation [73]. While ubiquitin chains consisting of only one type of linkage can regulate both proteasome-dependent and -independent processes, branched polyubiquitin chains have been shown to preferentially trigger proteasome-mediated degradation [74, 75]. Consistent with these reports, our results indicated that SKP2-mediated K48/K63-branched chains in CD8⁺ T cells supported the degradation of TBET through the UPS. However, we also noted that 3 out of 4 subjects showed an increase in TBET protein levels after chloroquine exposure. Thus, the mechanisms by which TBET protein is degraded may be subject- and/or disease-dependent.

Our study exhibited limitations. Since mTORC1 is an essential regulator of alveolar macrophage function [76, 77], our results suggested that SHP099 did not elicit a direct effect on these cells in vivo. However, it remains unclear whether SHP099 had an indirect effect on the inflammatory state of macrophages within Tsc2^fl/flLyz2-Cre mice. Previous studies have indicated an altered inflammatory state in macrophages in response to SHP099 [78]. Hence, future studies are needed to determine the effects of SHP099 on the inflammatory state of macrophages within this model. Another limitation of our study was the inability to verify high pSHP2 and reduced ubiquitinated TBET in the lungs of Tsc2^fl/flLyz2-Cre mice prior to SHP099 exposure for post-treatment comparisons. Thirdly, we were unable to validate our findings ex vivo in a large cohort due to sample limitations. To bypass this limitation, however, we attempted to confirm our findings in vitro, ex vivo and in vivo using various methodologies. Fourthly, we did not examine the inhibitory effects of SHP099 on parallel pathways through which SHP2 has been shown to also regulate IFNγ (e.g., JAK/STAT and ERK/MAPK signaling) [6, 36]. Nevertheless, the aforementioned limitations do not detract from the importance of our findings and striking therapeutic implications.

In conclusion, our investigation provides new insight into the post-translational regulatory mechanism of TBET by identifying SKP2 as its E3 ligase via K63/K48-linked polyubiquitination, which has not been previously reported in any cell type (SF8g). Further, whether CD8⁺ T cells play a significant role in sarcoidosis remains a highly disputed topic. Many groups believe that these cells do not play a major role but, prior to our investigation, sufficient evidence to support or dispute this theory was not available. Thus, our investigation also provides strong molecular and immunological evidence supporting the significance of CD8⁺ T cells in severe pulmonary sarcoidosis. Lastly, our investigation offers strong evidence that SHP099 could be employed to suppress TBET and IFNγ expression in sarcoidosis CD8⁺ T cells. However, our in vivo studies in mice also demonstrated the existence of a CD8⁺ T cell population that was resistant to SHP099 treatment. Therefore, our findings provide a new paradigm to explore the genetic and molecular factors that could further characterize these two subpopulations in humans and their distinct contributions to sarcoidosis pathogenesis.

MATERIALS AND METHODS

Study Design

The aim of our study was to assess the function of SHP2 in severe sarcoidosis. Early analysis revealed enhanced SHP2 signaling within sarcoidosis CD8⁺ T cells, thus our subsequent objective was to evaluate the role of aberrant SHP2 signaling within sarcoidosis CD8⁺ T cells and its effects in severe disease. Our predefined endpoints included the number of large granulomas (>0.05 mm²), levels of collagen, IFNγ, TBET and ubiquitinated TBET. Females and males of all races, contingent on sample availability, were included in this study. To ascertain statistical significance was not a result of outliers, statistical evaluation was performed with and without the inclusion of outliers. However, all data points, with matching statics, were included in final figures. Data acquisition and analysis (e.g., quantification) was performed blinded. Once completed, samples were de-identified and assigned to appropriate groups (e.g., control v sarcoidosis) for statistical assessment. Due to sample limitations, sample size (n) is indicated in each figure legend. Published ATS/ERS/WASOG clinical and radiographic criteria were used to diagnose sarcoidosis [79, 80]. All study participants provided written informed consent approved by the IRB. All mouse studies were approved by the Medical University of Vienna and Austrian Ethics Committee for animal experiments (GZ.BMWF.2020-0.547.514).

Statistical Analysis

An unpaired two-tailed Student’s t-test was used for statistical analysis between untreated cohorts. A paired Student’s t-test was used to compare differences between untreated and SHP099-treated cohorts. Multiple group comparisons were performed using a one-way ANOVA with Tukey’s post hoc test. Associations were evaluated by Spearman’s correlation. Statistical analysis for all figures was carried out using GraphPad Prism. A P<0.05 was considered statistically significant.

Table 1.

Demographics for peripheral blood control and sarcoidosis populations

	Controls	Sarcoidosis
Number	21	44
Gender (female; male)	13; 8	24; 20
Age, median (max, min)	32 (58, 26)	55 (77, 20)
Race	7 AA; 14 C	20 AA; 24 C

AA= African American, C= Caucasian

Table 2.

Demographics for bronchoalveolar lavage (BAL) control and sarcoidosis populations

	Controls	Sarcoidosis
Number	4	36
Gender (female; male)	2; 2	19; 17
Age, median (max, min)	33 (44, 22)	47 (65, 25)
Race	4 C	20 AA; 16 C

AA= African American, C= Caucasian

Table 3.

Demographics for control and end-stage sarcoidosis lung tissue, lung suspensions and PCLS

	Controls	Sarcoidosis
Number	8	4
Gender (female; male)	4; 4	2; 2
Age, median (max, min)	50.5 (56, 29)	62 (69, 37)
Race	4 AA; 4 C	2 AA; 2 C

AA= African American, C= Caucasian