EGFR mimotope alleviates renal fibrosis by promoting EGFR/RPL6/PCBP2 ternary complex formation

Abstract

Epidermal growth factor receptor (EGFR) signalling plays a crucial role in renal fibrosis. We previously demonstrated that EGFR mimotope vaccination could alleviate renal fibrosis by restoring major histocompatibility complex class IB (MHC-IB) expression in renal tubular epithelial cells and promoting MHC-IB-enriched exosome release. Intriguingly, the mechanism by which the EGFR mimotope affects MHC-IB remains unknown. In this study, we demonstrated that administering purified EGFR mimotope-induced antibodies (EM-pAb) can effectively upregulate ribosomal protein ligand (RPL)-6, poly (rC) binding protein 2 (PCBP2), and MHC-IB in tubular epithelial cells. Mechanistically, the EM-pAb promoted EGFR/ribosomal protein ligand 6 (RPL6)/poly (rC) binding protein 2 (PCBP2) functional ternary complex formation, re-established the binding capability of PCBP2 to MHC-IB β-chain, and further regulated the NK cell population. RPL6 was confirmed as the key orchestrator of the ternary complex formation. Rpl6^CKI mice further revealed the protective role of RPL6 in the animal model of renal fibrosis. RPL6 and kidney function showed a significant correlation in patients with chronic kidney disease. Overall, our findings uncover the signalling pathway by which EGFR modulates MHC-IB and validate the regulatory role of EM-pAb in this signalling pathway, indicating the therapeutic potential of RPL6 in chronic kidney disease.

Introduction

A mimotope is a peptide vaccine composed of a synthetic or engineered peptide sequence that structurally mimics the natural antigenic epitopes expressed on the cell surface. The active immunisation induced by a mimotope can target the native epitope and regulate downstream signalling¹. To date, mimotopes have been widely employed in immunology and related fields for exploring pathogenic immune responses, mapping antibody-antigen interactions, and designing targeted therapeutics for various diseases^2–7.

In the initial phase of our mimotope investigation, we used phage-display libraries to obtain two EGFR mimotopes capable of generating high antibody titres and recognizing EGFR⁸. EGFR signalling is a classical cascade central to embryonic development and the maintenance of epithelial tissue integrity. EGFR signalling dysfunction results in exacerbation of malignancies, allergies, infections, and autoimmune diseases⁹. In subsequent investigations, we expanded the application of EGFR mimotopes to two autoimmune diseases. In rheumatoid arthritis, EGFR mimotopes were used as a diagnostic tool to screen patients with early-stage aggressive rheumatoid arthritis¹⁰. Further, vaccination with the EGFR mimotope alleviated autoimmune responses in an experimental model of the Sjögren syndrome. Notably, macrophages were significantly decreased in the exocrine glands of mice receiving the EGFR mimotope treatment¹¹.

Renal fibrosis is an inevitable pathological characteristic of chronic kidney disease (CKD). Persistent chronic fibrosis leads to aggressive tissue destruction and ultimately, loss of kidney function. Therefore, CKD is a worldwide health problem presenting tremendous socioeconomic challenges^12,13. Investigation of CKD pathogenesis revealed that the EGFR signalling cascade and macrophages are both major drivers promoting renal fibrosis progression^14,15. As expected, vaccination with the EGFR mimotope alleviated experimental renal fibrosis. EGFR signalling and macrophage activation were successfully blocked by the EGFR mimotope¹⁶.

To clarify the mechanism by which the EGFR mimotope inhibits macrophage activation, we performed further experiments and found that the EGFR mimotope induced renal tubular epithelial cells (TECs) to release major histocompatibility complex (MHC)-IB-carrying exosomes, which inhibited macrophage activation. RNA sequencing analysis showed that only three genes, ribosomal protein ligand 6 (RPL6), poly (rC) binding protein 2 (PCBP2), and histocompatibility complex class IB (MHC-IB), were upregulated in the EGFR mimotope-treated TECs. Extended experiments revealed that MHC-IB expression on the surface of TECs and TEC-exosomes is controlled by RPL6¹⁷. In the present study, we aimed to elucidate the underlying modulatory effects of the EGFR mimotope on these molecules.

Results

EM-pAb injection ameliorated renal fibrosis and promoted RPL6 expression

Mimotope-induced active immunisation can trigger multiple immune responses. EM-pAb, the main physiological response component from EGFR mimotope immunisation, was purified and verified for its effectiveness. The EM-pAb treatment procedure is shown in Fig. 1A. The protective effect of EM-pAb against renal fibrosis was observed using Masson trichrome staining (Fig. 1B). The accumulation of fibronectin, α-SMA, and Collagen I was decreased in EM-pAb-treated kidney tissues (Fig. 1C).

Fig. 1. EM-pAb injection ameliorated renal fibrosis and promoted RPL6 expression.

A The UUO surgery and the EM-pAb treatment procedure are shown. The purified EGFR mimotope pAb (EM-pAb) was administered intravenously on days -1 (10 mg/kg) and 2 (5 mg/kg). B The protective effect of EM-pAb against renal fibrosis was observed using Masson trichrome staining. C Western blotting assay showed that the accumulation of fibronectin, α-SMA, and Collagen I was decreased in EM-pAb-treated kidney tissues. D The expression of RPL6 was detected with immunofluorescence staining. We confirmed that EM-pAb injection upregulated RPL6 expression in kidney tubular cells. (Mean ± SEM; ***p < 0.001; ** p < 0.005; *p < 0.05; representative image. n = 6 per group. Western blotting assay was performed duplicate.).

In a previous study, we found that EM immunisation could only upregulate three genes, RPL6, PCBP2, and MHC-IB, in HK-2 cells and that RPL6 plays a crucial role in MHC-IB expression¹⁷. Consequently, we confirmed that EM-pAb injection upregulated RPL6 expression in kidney tubular cells (Fig. 1D).

Overexpression of RPL6 in tubular epithelial cells alleviated renal fibrosis in conditional knock-in mice

Rpl6^CKI mice were generated to investigate the role of RPL6 in renal fibrosis. We performed a unilateral ureteric obstruction (UUO) surgery to establish the animal model of renal fibrosis. Compared with WT mice, Rpl6^CKI mice showed a lower extent of fibrosis and matrix accumulation. When siRNA against RPL6 was administered to Rpl6^CKI mice, fibrosis was exacerbated (Fig. 2).

Fig. 2. Overexpression of RPL6 in tubular epithelial cells alleviated renal fibrosis.

A Rpl6^CKI mice were generated to investigate the role of RPL6 in renal fibrosis. Rpl6^flox/flox mice were crossed with Cdh16^cre transgenic mice to obtain Rpl6^CKI mice with RPL6 knock-in in kidney tubular epithelial cells. The procedure of the UUO surgery and the in vivo siRNA RPL6 treatments are shown. In vivo siRNA RPL6 was administered intravenously on days -1 (40 nmol), 2 (20 nmol), and 5 (20 nmol). B Masson trichrome staining and immunohistochemistry showed that as compared with WT mice, Rpl6^CKI mice showed a lower extent of fibrosis and matrix accumulation. When siRNA against RPL6 was administered to Rpl6^CKI mice, fibrosis was exacerbated. No significant difference was observed between the WT sham group and the Rpl6^CKI sham group. C The similar results were obtained by western blotting assay. (Mean ± SEM; ***p < 0.001; **p < 0.005; *p < 0.05; n.s.: non-significant. n = 6 in wild type sham group; n = 6 in wild type UUO group; n = 3 in Rpl6^CKI sham group; n = 6 in Rpl6^CKI UUO group; n = 5 in Rpl6^CKI UUO + siRNA group. Western blotting assay was performed triplicate.).

RPL6, PCBP2, and MHC-IB were upregulated and showed consistent expression patterns in Rpl6^CKI tubular epithelial cells

The tissue expression of RPL6, PCBP2, and MHC-IB was determined using TSA staining. The expression of RPL6 and MHC-IB was significantly decreased in WT mice that underwent UUO surgery. In contrast, the levels of RPL6, PCBP2, and MHC-IB were significantly higher in Rpl6^CKI mice that underwent surgery. Further siRNA administration induced an expected decrease in RPL6, PCBP2, and MHC-IB expression. Tissues from the Rpl6^CKI sham group showed higher levels of RPL6 and PCBP2 than those in tissues from the WT sham group.

Next, we analysed the co-expression of RPL6, PCBP2, and MHC-IB in tubular epithelial cells. Overlay curves showed that in Rpl6^CKI sham and Rpl6^CKI UUO mice, consistent expression of RPL6, PCBP2, and MHC-IB was markedly observed in AQP-1⁺ tubular epithelial cells. However, RPL6 knockdown disrupted this consistency (Fig. 3).

Fig. 3. Expressions of RPL6, PCBP2 and MHC-IB were upregulated and exhibited consistency in the Rpl6^CKI kidney tissues.

The tissue expression of RPL6, PCBP2, and MHC-IB was determined using TSA staining. The expression of RPL6 and MHC-IB was significantly decreased in WT mice that underwent UUO surgery. In contrast, the levels of RPL6, PCBP2, and MHC-IB were significantly higher in Rpl6^CKI mice that underwent surgery. Further siRNA administration induced an expected decrease in RPL6, PCBP2, and MHC-IB expression. Tissues from the Rpl6^CKI sham group showed higher levels of RPL6 and PCBP2 than those in tissues from the WT sham group. Overlay curves showed that in Rpl6^CKI sham and Rpl6^CKI UUO mice, consistent expression of RPL6, PCBP2, and MHC-IB was markedly observed in AQP-1⁺ tubular epithelial cells. However, RPL6 knockdown disrupted this consistency. (Mean ± SEM; ***p < 0.001; **p < 0.005; *p < 0.05; n.s.: non-significant. n = 6 in wild-type sham group, n = 6 in wild-type UUO group, n = 3 in Rpl6^CKI sham group, n = 6 in Rpl6^CKI UUO group, and n = 5 in Rpl6^CKI UUO + siRNA group. The overlap curves are calculated based on the area within the white border circle in the image).

EM-pAb upregulated RPL6, PCBP2, β2-microglobulin, and MHC-IB in HK-2 cells

Next, we used EGF to activate EGFR and observed the changes in RPL6, PCBP2, β2-microglobulin, and MHC-IB expression in HK-2 cells. MHC-I is a trimeric complex composed of an α-chain and the invariant light chain, β2-microglobulin. We set multiple time intervals and EGF concentration ranges to collect the cell samples. The results showed that the expression of RPL6, PCBP2, β2-microglobulin, and MHC-IB was significantly decreased when HK-2 cells were treated with 100 ng/ml EGF for 72 h (Fig. 4A, B). Control pAb or EM-pAb was used to determine the optimal effective concentration that could rescue the reduced expression of these proteins. Overall, these data indicated that EM-pAb treatment could effectively increase the expression of these proteins at 2 μg/ml, whereas control pAb treatment could not rescue the expression of these proteins up to 10 μg/ml (Fig. 4C).

Fig. 4. EM-pAb upregulated RPL6, PCBP2, β2-microglobulin and MHC-IB in HK-2 cells.

A HK-2 cells were treated with 100 ng/mL EGF and harvested at multiple time intervals. The results showed that after treatment for 72 h, the expression of RPL6, PCBP2, β2-microglobulin (the β-chain of MHC-IB), and MHC-IB began to decrease significantly. B HK-2 cells were treated with different EGF concentrations for 72 h. The results showed that when HK-2 cells incubated with the EGF concentration above 50 ng/mL, the expression of RPL6, PCBP2, β2-microglobulin, and MHC-IB began to decrease significantly. C The HK-2 cells were treated with 100 ng/mL EGF for 72 h. The control pAb or EM-pAb were used to determine the optimal effective concentration that could rescue the reduced expression of these proteins. The data indicated that EM-pAb treatment could effectively increase the expression of these proteins at 2 μg/ml. D HK-2 cells were treated with siRNA RPL6 (20nmol) for 72 h and rescued by using control pAb or EM-pAb. RPL6 knockdown dramatically reduced the expression of RPL6, β2-microglobulin, and MHC-IB, but no significant changes were observed in PCBP2. Neither control pAb nor EM-pAb treatment rescued the expression of RPL6, β2-microglobulin and MHC-IB. E HK-2 cells received control pAb or EM-pAb without EGF or siRNA treatment. Only the expression of PCBP2 increased in EM-pAb-treated HK-2 cells, with no significant changes observed in RPL6, β2-microglobulin, and MHC-IB. F Surface MHC-IB and β2-microglobulin were analysis using flow cytometry. The flow cytometry results were parallel to those of the western blotting assay. RPL6 knockdown dramatically reduced the expression of β2-microglobulin and MHC-IB at the cell surface. Neither control pAb nor EM-pAb treatment rescued the expression of β2-microglobulin and MHC-IB. In contrast, EM-pAb treatment rescued EGF-induced downregulation of β2-microglobulin and MHC-IB. (Mean ± SEM; ***p < 0.001; **p < 0.005; *, p < 0.05; n.s.: non-significant; Western blotting assay was performed in triplicate).

As direct knockdown of RPL6 could promote fibrosis in an animal model, we detected the expression of these proteins in HK-2 cells treated with siRNA RPL6 and rescued them using control-pAb or EM-pAb. Intriguingly, RPL6 knockdown dramatically reduced the expression of RPL6, β2-microglobulin, and MHC-IB, but no significant changes were observed in PCBP2. Moreover, neither the control pAb nor EM-pAb treatment rescued the expression of RPL6, β2-microglobulin, or MHC-IB (Fig. 4D).

We further investigated whether EM-pAb increased the expression of these proteins in HK-2 cells without EGF or siRNA treatment. Contrary to the results obtained with RPL6 knockdown, only the expression of PCBP2 increased in EM-pAb-treated HK-2 cells, with no significant changes observed in RPL6, β2-microglobulin, and MHC-IB (Fig. 4E).

Surface MHC-IB and β2-microglobulin were analysed using flow cytometry. The flow cytometry results were parallel to those of the western blotting assay. RPL6 knockdown dramatically reduced the expression of β2-microglobulin and MHC-IB at the cell surface. Neither CP-pAb nor EM-pAb treatment rescued the expression of β2-microglobulin and MHC-IB. In contrast, EM-pAb treatment rescued EGF-induced downregulation of β2-microglobulin and MHC-IB (Fig. 4F).

EM-pAb promoted RPL6-dependent EGFR/RPL6/PCBP2 complex formation and binding to the RNA of MHC-IB β-chain

Considering that RPL6 and PCBP2 are constituents of the protein-synthesizing machinery, we investigated whether RPL6 and PCBP2 would form a complex and bind to the MHC-IB RNA. The protein expression data obtained after EGF stimulation or RPL6 knockdown indicated that RPL6 may play a crucial role. An immunoprecipitation assay showed that EGFR, RPL6, and PCBP2 formed complexes in HK-2 cells. Both EGF stimulation and RPL6 knockdown downregulated complex formation. However, in cells treated with RPL6 siRNA, EM-pAb could not restore complex formation, whereas EM-pAb could do so in EGF-treated cells (Fig. 5A). The result of the RNA pull down assay with the β2-microglobulin transcript was similar to that of immunoprecipitation. In cells treated with RPL6 siRNA, PCBP2 showed no significant change in any group. In contrast, EGF stimulation reduced PCBP2 binding to the RNA of β2-microglobulin, while EM-pAb recovered this binding in EGF-treated cells (Fig. 5B). PCBP2 failed to bind to the transcript of the MHC-IB α-chain under either RPL6 knockdown or EGF stimulation (Fig. 5C).

Fig. 5. EM-pAb promoted RPL6-dependent EGFR/RPL6/PCBP2 complex formation and binding to the β2-microglobulin transcript.

A An immunoprecipitation assay was performed to observe whether EGFR, RPL6, and PCBP2 could form complex in HK-2 cells. Both EGF stimulation and RPL6 knockdown downregulated complex formation. However, in cells treated with RPL6 siRNA, EM-pAb could not restore complex formation, whereas EM-pAb could do so in EGF-treated cells. B RNA pull-down assay was performed to observe the binding of PCBP2 to the β2-microglobulin transcript. The result was similar to that of immunoprecipitation. In cells treated with RPL6 siRNA, PCBP2 showed no significant change in any group. In contrast, EGF stimulation reduced PCBP2 binding to the RNA of β2-microglobulin, while EM-pAb recovered this binding in EGF-treated cells. C PCBP2 failed to bind to the transcript of the MHC-IB α-chain under either RPL6 knockdown or EGF stimulation. (Immunoprecipitation assay was performed in triplicate; RNA pull down assay was performed duplicate).

EM-pAb regulated EGFR phosphorylation, internalisation, and downstream classical signalling

We first examined Akt and Erk signalling in HK‑2 cells under different treatment conditions. EGF stimulation markedly increased the phosphorylation of both Akt and Erk in HK‑2 cells. EM-pAb effectively attenuated the EGF‑induced phosphorylation of Akt and Erk. These observations were consistent with our initial expectations. However, distinct signalling patterns emerged in cells transfected with siRNA targeting RPL6. RPL6 knockdown significantly elevated Akt phosphorylation, and this effect was strongly enhanced by control pAb, whereas EM-pAb reduced Akt phosphorylation. Total Akt levels were not significantly altered by RPL6 knockdown but were downregulated by both control pAb and EM-pAb. In contrast, RPL6 knockdown did not affect total Erk or phosphorylated Erk levels, whereas both control pAb and EM-pAb significantly reduced the total Erk and phosphorylated Erk levels (Fig. 6A).

Fig. 6. EM-pAb regulated EGFR phosphorylation, internalization and downstream classic signaling.

After treatment with EGF (100 ng/ml) for 72 h, HK-2 cells were washed with cold PBS twice and solubilized in lysis buffer. A Akt/Erk signaling was detected by western blotting assay. EGF stimulation markedly increased the phosphorylation of both Akt and Erk in HK‑2 cells. EM-pAb effectively attenuated the EGF‑induced phosphorylation of Akt and Erk. RPL6 knockdown significantly elevated Akt phosphorylation, and this effect was strongly enhanced by control pAb, whereas EM-pAb reduced Akt phosphorylation. Total Akt levels were not significantly altered by RPL6 knockdown but were downregulated by both control pAb and EM-pAb. In contrast, RPL6 knockdown did not affect total Erk or phosphorylated Erk levels, whereas both control pAb and EM-pAb significantly reduced the total Erk and phosphorylated Erk levels. B Only EGFR Tyr845 maintained phosphorylation, whereas EM-pAb effectively reduced the Tyr845 phosphorylation level. C EGFR internalisation was observed using confocal microscopy. EGFR showed diffuse plasma membrane localisation under basal conditions. EGF treatment significantly diminished diffuse EGFR staining at the cell surface, while inducing prominent intracellular aggregation in the cytoplasm. Additional application of EM-pAb altered the distribution of EGFR. A small part of EM-pAb binding was observed at the plasma membrane, whereas the majority of EM-pAb was detected in the cytoplasm and was concentrated around the nucleus. EGFR distribution was consistent with that of the majority of EM-pAb. (Mean ± SEM; ***p < 0.001; **p < 0.005; *p < 0.05; n.s.: non-significant; Western blotting assay and EGFR phosphorylation analysis were performed in triplicate; EGFR internalization observation was performed duplicate.).

EGFR phosphorylation was detected using the RayBio® Human EGFR Phosphorylation Antibody Array 1. After HK-2 cells were treated with EGF (100 ng/ml) for 72 h, only EGFR Tyr845 maintained phosphorylation, whereas EM-pAb effectively reduced the Tyr845 phosphorylation level (Fig. 6B).

EGFR internalisation was observed using confocal microscopy. EGFR showed diffuse plasma membrane localisation under basal conditions. EGF treatment significantly diminished diffuse EGFR staining at the cell surface, while inducing prominent intracellular aggregation in the cytoplasm. Additional application of EM-pAb altered the distribution of EGFR. A small part of EM-pAb binding was observed at the plasma membrane, whereas the majority of EM-pAb was detected in the cytoplasm and was concentrated around the nucleus. EGFR distribution was consistent with that of the majority of EM-pAb (Fig. 6C).

Overexpression of RPL6 attracted the NK population and inhibited macrophages in Rpl6^CKI kidney tissues

MHC class I molecules are mainly recognised by CD8⁺ and NK cells¹⁸. In our previous study, we noted that EM-induced MHC-IB exosome treatment recruits NCR1⁺ cells to kidney tissues but does not affect the T cell population¹⁷. The NK cell population was then considered for subsequent experiments. NCR1 is expressed in murine tissue-resident and mature NK cells. These NK cell populations interact with MHC-I via the Ly49 inhibitory receptor family. Thus, a high level of MHC-I increases the suppression of NK cell activity and promotes self-tolerance¹⁹.

NCR1⁺ NK cells and F4/80⁺ macrophages were detected in the kidneys of mice from each group. NCR1⁺ NK cells were significantly increased in Rpl6^CKI mice, but were disrupted by in vivo siRNA administration. The distribution of F4/80⁺ macrophages in the tissues showed an inverse trend compared with that of NCR1⁺ NK cells.

Next, we analysed the consistency of Ly49 and MHC-IB expression in NCR1⁺ NK and AQP-1⁺ tubular epithelial cells. The overlay curves showed that in WT sham mice, Ly49 expression was consistently observed in NCR1⁺ NK cells, whereas MHC-IB expression was consistently observed in AQP-1⁺ tubular epithelial cells. The same trend was observed in Rpl6^CKI sham mice. In WT UUO mice, Ly49 expression seemed to be consistent with that of AQP-1. No significant consistency was observed between NCR1 and MHC-IB. In Rpl6^CKI mice undergoing surgery, the expression patterns of MHC-IB and Ly49 in the AQP-1⁺ tubular epithelial cells and NCR1⁺ NK cells were inverted. The two curves exhibited remarkable consistency. However, application of siRNA resulted in consistent curves (Fig. 7).

Fig. 7. Overexpression of RPL6 in tubular epithelial cells attracted NK population and inhibited macrophages.

NCR1⁺ NK cells and F4/80⁺ macrophages were detected in the kidneys of mice from each group. NCR1⁺ NK cells were significantly increased in Rpl6^CKI mice, but were disrupted by in vivo siRNA administration. The distribution of F4/80⁺ macrophages in the tissues showed an inverse trend compared with that of NCR1⁺ NK cells. The overlay curves showed that in WT sham mice, Ly49 expression was consistently observed in NCR1⁺ NK cells, whereas MHC-IB expression was consistently observed in AQP-1⁺ tubular epithelial cells. The same trend was observed in Rpl6^CKI sham mice. In WT UUO mice, Ly49 expression seemed to be consistent with that of AQP-1. No significant consistency was observed between NCR1 and MHC-IB. In Rpl6^CKI mice undergoing surgery, the expression patterns of MHC-IB and Ly49 in the AQP-1⁺ tubular epithelial cells and NCR1⁺ NK cells were inverted. The two curves exhibited remarkable consistency. However, application of siRNA resulted in consistent curves. (Mean ± SEM; ***p < 0.001; **p < 0.005; *p < 0.05; n.s.: non-significant. n = 6 in wild-type sham group, n = 6 in wild-type UUO group, n = 3 in Rpl6^CKI sham group, n = 6 in Rpl6^CKI UUO group, and n = 5 in Rpl6^CKI UUO + siRNA group. The overlap curves are calculated based on the area with the asterisk in the image).

RPL6 was correlated with CKD progression and fibrosis

Clinical specimens were collected to verify the correlation between RPL6 expression and CKD progression. Masson trichrome and immunofluorescence staining showed that progressive silencing of the RPL6 signal emerged as a hallmark of advanced renal fibrosis (Fig. 8A). In statistical analysis, the RPL6 level showed a significant inverse correlation with the renal functional parameters in patients with CKD. The RPL6 level in HK-2 cells exposed to CKD stage III sera was higher than that in CKD stage V sera. In contrast, the RPL6 concentration in urine samples was positively correlated with CKD progression. High RPL6 levels were observed in the urine of patients with CKD stage V (Figs. 8B, C).

Fig. 8. RPL6 correlated with CKD progression and fibrosis.

A Clinical specimens were collected to verify the correlation between RPL6 expression and CKD progression. Five kidney biopsy samples were obtained to compare the fibrosis and RPL6 expression. Masson trichrome and immunofluorescence staining showed that progressive silencing of the RPL6 signal emerged as a hallmark of advanced renal fibrosis. B, C The sera and urine samples were used to determine the correlation between RPL6 level and kidney function of patients with CKD. In statistical analysis, the RPL6 level showed a significant inverse correlation with the renal functional parameters in patients with CKD. The RPL6 level in HK-2 cells exposed to CKD stage III sera was higher than that in CKD stage V sera. In contrast, the RPL6 concentration in urine samples was positively correlated with CKD progression. High RPL6 levels were observed in the urine of patients with CKD stage V. (Mean ± SEM; ***p < 0.001; **p < 0.005; *p < 0.05; n.s.: non-significant; Western blotting assay was performed in duplicate).

EM-pAb modulated the functional phenotype of NK cells and cytokine profiles

As shown in Fig. 9A (Created in BioRender. NOEr×, _. (2026) https://BioRender.com/4how1ug), HK-2 cells were treated with the sera of patients with CKD, followed by incubation and analysis of the NK cell population. Consequently, we analysed the CD56^brightCD16^- and CD56^dimCD16⁺ NK populations and further stained NKG2A in CD56^brightCD16^-. Cytokines from NK cells were measured to indirectly evaluate the function of NK cells. An additional group of HK-2 cells exposed to CKD stage V was administered EM-pAb.

Fig. 9. EM-pAb modulated the functional phenotype of NK cells, release of cytokines, expression of RPL6, and cell viability.

A HK-2 cells were treated with the sera of patients with CKD, followed by incubation and analysis of the NK cell population. (Created in BioRender. NOEr×, _. (2026) https:// BioRender.com/4how1ug) B As indicated by the experimental data, the frequency of NK cells within the CD56^brightCD16^- gate was the highest in the EM-pAb administration group, followed by that in the stage V group, and was the lowest in the stage III group. Extended NKG2A staining revealed that the EM-pAb group induced a significant increase in the CD56^brightCD16^-NKG2A⁺ subset. In contrast, the frequency of CD56^dimCD16⁺NK cells was significantly lower in the EM-pAb administration group. C We performed CBA experiments to analyse the characteristics of NK cytokines. An extreme increase in IFN-γ was observed in the EM-pAb-administered group, whereas IL-1 and IL-10 exhibited a slight but statistically significant increase. D Lirilumab, a fully human IgG4 anti-KIR, was used to block the interaction between KIR and MHC-I. Application of Lirilumab significantly reduced the CD56^bightCD16^-NKG2A⁺ subset and IFN-γ release. However, no significant change was observed in the frequency of CD56^dimCD16⁺NK cells. E We examined the changes in RPL6 expression in HK-2 cells exposed to the sera of patients with CKD at different stages. RPL6 expression was lower in HK-2 cells exposed to stage V CKD sera than in those exposed to stage III CKD sera. EM-pAb administration restored the expression of RPL6 in cells exposed to stage V CKD sera. F HK-2 cells seeded in 96-well plates were used for viability evaluation under different treatment conditions. In HK-2 cells receiving sera from patients with stage III CKD, EM-pAb promoted HK-2 cell viability irrespective of co-culture with NK cells. In contrast, when HK-2 cells were exposed to the sera of patients with stage IV/V, EM-pAb could only recover the HK-2 cell viability reduced by NK cells. (Mean ± SEM; ***p < 0.001; **p < 0.005; *p < 0.05; n.s.: non-significant. Western blotting assay and cell viability assay were detected in duplicate).

As indicated by the experimental data, the frequency of NK cells within the CD56^brightCD16^- gate was the highest in the EM-pAb administration group, followed by that in the stage V group, and was the lowest in the stage III group. Extended NKG2A staining revealed that the EM-pAb group induced a significant increase in the CD56^bightCD16^-NKG2A⁺ subset. In contrast, the frequency of CD56^dimCD16⁺NK cells was significantly lower in the EM-pAb administration group (Fig. 9B). We performed CBA experiments to analyse the characteristics of NK cytokines. An extreme increase in IFN-γ was observed in the EM-pAb-administered group, whereas IL-1 and IL-10 exhibited a slight but statistically significant increase (Fig. 9C). Application of Lirilumab significantly reduced the CD56^bightCD16^-NKG2A⁺ subset and IFN-γ release. However, no significant change was observed in the frequency of CD56^dimCD16⁺NK cells (Fig. 9D).

EM-pAb upregulated RPL6 in the HK-2 cells treated with sera from patients with CKD

Subsequently, we examined the changes in RPL6 expression in HK-2 cells exposed to the sera of patients with CKD at different stages. RPL6 expression was lower in HK-2 cells exposed to stage V CKD sera than in those exposed to stage III CKD sera. EM-pAb administration restored the expression of RPL6 in cells exposed to stage V CKD sera (Fig. 9E).

EM-pAb recovered HK-2 cell viability

HK-2 cells seeded in 96-well plates were used for viability evaluation under different treatment conditions. In HK-2 cells receiving sera from patients with stage III CKD, EM-pAb promoted HK-2 cell viability irrespective of co-culture with NK cells. In contrast, when HK-2 cells were exposed to the sera of patients with stage IV/V, EM-pAb could only recover the HK-2 cell viability reduced by NK cells (Fig. 9F).

Discussion

The major findings of the current study were as follows: (i) EM-pAb promoted the formation of the RPL6-dependent EGFR/RPL6/PCBP2 ternary complex and further modulated MHC-IB expression and the NK cell population; (ii) RPL6 showed a protective effect in experimental renal fibrosis and correlated with kidney function in CKD specimens.

To date, investigations into the relationship between EGFR and MHC-I have predominantly focused on Brucella abortus infection and oncological research^20–26. This relationship remains to be examined in renal fibrosis. Our data suggest a new direct interaction between EGFR, RPL6, and PCBP2, which modulates MHC-IB in tubular epithelial cells. This discovery contributes novel dimensions to the epithelial-immune crosstalk in renal fibrosis.

RPL6 is a member of the ribosomal protein family whose role remains poorly characterized. Limited studies have indicated that RPL6 may be a critical regulatory factor involved in immunosurveillance. Wei et al. reported that RPL6 depletion could reduce ubiquitin-dependent influenza A virus-encoded peptide presentation²⁷. In another study, prime-and-trap vaccination targeting RPL6 activated liver-resident CD8 TRM cells and induced effective protection against malaria²⁸. In our preliminary study, we serendipitously found that RPL6 was increased in HK-2 cells exposed to EM-sera, and that RPL6 knockdown undermined EM-induced HK-2 exosome treatment in an animal model, suggesting a potential protective role of RPL6 in renal fibrosis. In the present study, we generated Rpl6^CKI mice and experimentally validated our hypothesis regarding the anti-fibrotic role of RPL6.

The current investigation of EGFR mimotope application depends on disease pathologies in which EGFR signalling exerts critical regulatory functions. Demonstrating that an EGFR mimotope could modulate MHC-IB expression via the EGFR/RPL6/PCBP2 ternary complex would undoubtedly expand the therapeutic scope of the EGFR mimotope strategy. Therefore, we purified EM-pAb and examined its modulatory effects on the EGFR/RPL6/PCBP2 ternary complex. As expected, EM-pAb ameliorated experimental renal fibrosis and mechanistically promoted formation of the ternary complex as well as upregulated MHC-IB expression. Notably, normal intracellular expression of RPL6 was necessary for the regulatory function of EM-pAb, suggesting that RPL6 was the key orchestrator in ternary complex formation. Further, we assumed that RPL6 may be an important regulator for Akt signals based on the results of siRNA interference. Under physiological conditions, RPL6 could interact with EGFR and maintain its stability at the plasma membrane. This interaction may impede the binding of PI3K to EGFR and further inhibit the downstream AKT signal; however, this interaction did not affect the binding of Ras and EGFR or the downstream ERK signal. Consequently, upon RPL6 depletion, EGFR may undergo spontaneous activation at the membrane and activate PI3K/Akt signalling, without influencing Ras/Erk activity. The EGFR internalisation assay results were intriguing. A small part of EM-pAb binding was observed at the plasma membrane, whereas the majority of EM-pAb was detected in the cytoplasm and concentrated around the nucleus. This change may be related to the EGFR/RPL6/PCBP2 ternary complex formation. As EM-pAb could inhibit the phosphorylation of EGFR Tyr845, we speculate that EM-pAb may exert both blocking and agonistic effects on EGFR.

In our previous study, we noticed that the NK cell infiltration in kidney tissues increased following EGFR mimotope treatment. Consequently, we conducted a preliminary investigation on how the NK cell population changed following their exposure to EM-pAb-treated tubular cells. The results showed that the EM-pAb-treated HK-2 cells significantly increased the frequency of the CD56^bightCD16^-NKG2A⁺ subset and IFN-γ release, and that applying the KIR antibody could effectively disrupt these alterations in the NK population. Although we cannot exclude other possible mechanisms, EGFR/RPL6/PCBP2 ternary complex formation-dependent MHC expression is most likely responsible for the observed changes in the NK population. However, our experimental data revealed findings that are partially in contrast to those reported in literature. In a study by Law et al., high expression levels of CD56^bright NK cells were found to accumulate in kidney tissues and contribute to renal fibrosis by releasing IFN-γ²⁹. In contrast, recent studies also reported that mesenchymal stromal cells treated with IFN-γ and TNF-α could demonstrate significant anti-fibrotic effects^30–32. We thus considered that the dual roles of the CD56^bightCD16^- population and IFN-γ in renal fibrosis may rely on the dynamic alterations in the fibrotic-immune microenvironment.

In conclusion, EM-pAb treatment ameliorates renal fibrosis by re-establishing epithelial-immune crosstalk. RPL6 can be considered as both a clinical biomarker and therapeutic target. Overall, this work provides further insights into the functional mechanisms of the EGFR mimotope and expands its translational and clinical potential.

Methods

Peptide preparation, animal immunisation, and antibody purification

The EGFR mimotope (WHTEILKSYPHEGGGSGGGS)⁸ was synthesised chemically with a purity of > 98%. A peptide segment (AIPLVVPFYSHSGGGSGGGS) in the M13 phage from the phage display library (E8110S, NEB) served as a control. Animal immunisation and antibody purification were conducted by UniBio Biotechnology Co. Ltd. (Shanghai, China). The concentration of rabbit antibodies induced by the EGFR mimotope (EM-pAb) and control peptide (control pAb) was 1 mg/mL.

Cells and animals

Ethical approval involving animals was obtained from the Ethics Committee of East China Normal University according to the Animal Administration Guidelines of China (No. m20231103).

Rpl6^flox/flox mice were generated by Cyagen Biosciences Co., Ltd. (Suzhou, China) using the CRISPR-Cas9 gene-targeting platform. Knock-in mice were created at the ROSA26 locus (located on mouse chromosome 6) on the C57BL/6 J background. The CAG promoter-loxP-PGK-Neo-6xSV40pA-loxP-Kozak-Mouse Rpl6 CDS-FLAG tag-WPRE-BGH pA cassette was inserted between exons 1 and 2 of ROSA26. Homology arms were generated via PCR using the BAC clone as a template. Cas9 and gRNA were co-injected into fertilised eggs using a targeting vector for mouse production. Targeted mice were identified using PCR and gene sequencing. Rpl6^flox/flox mice were then crossed with Cdh16^cre transgenic mice to obtain Rpl6^CKI mice with RPL6 knock-in in kidney tubular epithelial cells. Cdh16^cre transgenic mice were obtained from Cyagen Biosciences Co. Ltd. (Suzhou, China). HK-2 cells (human renal proximal tubular epithelial cells [TECs]; CRL-2190) were obtained from the American Type Culture Collection. Cells were maintained in Dulbecco’s modified Eagle medium/nutrient mixture F-12 (SH30023.01, HyClone) supplemented with 10% foetal bovine serum (10099141 C, ThermoFisher), 100 U/mL penicillin, and 100 mg/mL streptomycin at 37 °C and 5% CO₂.

Construction of animal models and treatment

Rpl6^CKI and wild-type (WT) male C57BL/6 J mice were maintained in a specific pathogen-free facility at the Animal Science Center of East China Normal University. Unilateral ureteric obstruction (UUO) surgery was performed by the Animal Science Center of East China Normal University to generate a renal fibrosis model³³. The mice were divided into different groups according to the subsequent experiments: the WT sham (n = 6 mice), WT UUO (n = 6 mice), WT UUO + EGFR mimotope (EM-pAb, n = 6 mice), Sham Rpl6^CKI (n = 3 mice), UUO Rpl6^CKI (n = 6 mice), and UUO Rpl6^CKI + in vivo siRNA (n = 5 mice) groups. The animals were anaesthetised with 2% isoflurane inhalation and euthanised with 5% CO₂ gradual inhalation. The procedure for the EM-pAb treatment of UUO is shown in Fig. 1A. The EM-pAb was administered intravenously on days -1 (10 mg/kg) and 2 (5 mg/kg). The procedure for the in vivo treatment of UUO using small interfering RNA (siRNA) against RPL6 is shown in Fig. 2A. The siRNA sequence was GCTGGCTAGTGGCTTATTA, which was chemically modified with 2′-O-methyl and a 5′-terminal cholesterol conjugate for animal application¹⁷. In vivo siRNA RPL6 was administered intravenously on days -1 (40 nmol), 2 (20 nmol), and 5 (20 nmol). The siRNA was produced by RIBOBIO Co., Ltd. (Guangzhou, China).

Histology analysis

Paraffin-embedded kidney specimens were prepared and subjected to Masson trichrome staining to evaluate fibrosis. The levels of fibronectin (2413, Abcam),α-SMA (7817, Abcam), and collagen I (260043, Abcam) were determined using immunohistochemistry. RPL6 (PA5-30217, ThermoFisher), PCBP2 (2531, Abclonal), MHC-IB (PA5-88804, ThermoFisher), AQP-1 (168387, Abcam), NCR1 (233558, Abcam), and LY49 (NBP1-76983, NOVUS) were stained using a tyramide signal amplification (TSA) kit (50013; Absin, Shanghai, China).

Cell treatment

HK-2 cells were plated in 6-well plates and cultured until they reached 40% confluence. The cells were then cultured in FBS-free medium for 24 h, followed by treatment with EGF, siRNA, or pAb.

EGFR phosphorylation analysis

EGFR phosphorylation was analysed using the RayBio® Human EGFR Phosphorylation Antibody Array 1. Briefly, after treatment with EGF (100 ng/ml) for 72 h, HK-2 cells were washed twice with cold PBS and solubilised in lysis buffer. Cell lysates were gently rocked at 2–8 °C for 30 min, followed by centrifugation at 14,000 × g for 10 min. Then, 1 ml of lysates were transferred into each array membrane, and incubated at room temperature for 2 h. Samples from each container were decanted and washed six times with washing buffer. The array membrane was then incubated with a Biotin-conjugated anti-EGFR antibody with gentle shaking at room temperature for 2 h. After the array membrane was washed another six times, HRP-conjugated streptavidin was added to each membrane with gentle shaking at room temperature for 2 h. Finally, the membranes were scanned using a chemiluminescence/fluorescence image analysis system (Tanon, Shanghai, China).

Immunofluorescence staining

HK-2 cells were fixed and permeabilised using the PrimeFlow™ Fixation/Permeabilization Buffer Set (ThermoFisher, 88-18001; 88-18009), followed by blocking with 10% goat sera at room temperature for 2 h. The cells were then incubated with anti-EGFR antibodies (80543, Santa Cruz Biotechnology) at room temperature for 2 h. Cells were washed three times with cold PBS and incubated with Alexa Fluor™ 488 anti-mouse antibodies (ThermoFisher, A-11001). Cells were washed another three times with cold PBS and Alexa Fluor™ 647 anti-rabbit antibodies (ThermoFisher, A-21244) were added. The stained HK-2 cells were observed using a confocal laser scanning microscope (STELLARIS 5, Leica). The images were obtained using Leica Application Suite X.

Immunoprecipitation

Immunoprecipitation was performed using an IP kit (955; Absin, Shanghai, China). The cells were incubated with the lysis buffer in the presence of 1% protease and phosphatase inhibitor cocktail with gentle shaking for 2 h at 4 °C. The cell lysate was then centrifuged at 14,000 × g for 15 min at 4 °C. Protein A/G agarose beads were pretreated for 2 h at 4 °C with constant rotation to diminish nonspecific binding. For immunoprecipitation, cell lysates were incubated with control IgG or anti-EGFR antibodies (26038, Cell Signaling Technology; 80543, Santa Cruz Biotechnology) overnight at 4 °C with constant rotation. Protein A/G agarose beads were added and the incubation continued for an additional 2 h. The beads were then washed six times with washing buffer. The precipitated proteins were boiled in 1× loading buffer and subjected to western blot analysis.

RNA pull-down

The β2-microglobulin (GenBank: NM_004048.4) and MHC-IB (GenBank: NM_005514.8) RNA templates were prepared using plasmid (PUC57) construction. Plasmids and primers were provided by FitGene Biotechnology Co. Ltd. (Guangzhou, China). The primer sequences used for β2-microglobulin RNA were: TAATACGACTCACTATAGGGATTCCTGAAGCTGACAGC (sense forward); TCAAGTTTTATGATTTATTTAACTTGTGG (sense reverse); ATTCCTGAAGCTGACAGC (antisense forward); and TAATACGACTCACTATAGGGTCAAGTTTTATGATTTATTTAACTTGTGG (antisense reverse). The primer sequences used for MHC-IB RNA (α-chain) were: TAATACGACTCACTATAGGGAGAGTCTCCTCAGACGCC (sense forward); AGCGCGCTCCAGCTTGTCCTT (sense reverse); AGAGTCTCCTCAGACGCC (antisense forward); and TAATACGACTCACTATAGGGAGCGCGCTCCAGCTTGTCCTT (antisense reverse). After amplification, the PCR products were excised and purified using a gel extraction kit (28506; QIAGEN). The RNA was then labelled with biotin and purified using an RNeasy Mini Kit (74904, QIAGEN). For RNA secondary structure formation, 3 μg of biotinylated RNA in RNA structure buffer was heated at 95 °C for 2 min, placed in an ice bath for 3 min, and equilibrated at room temperature for 30 min. Streptavidin beads were washed thrice with RIP lysis buffer and resuspended in 50 μL buffer. Structured biotinylated RNA was incubated with streptavidin beads at RT for 1 h. This was followed by centrifugation at 12,000 × g for 15 min at 4 °C. The beads were washed thrice and resuspended. Cell lysates were prepared using RIP lysis buffer containing RNase/protease inhibitors, followed by incubation with RNA-bead complexes at room temperature for 1 h. The beads were then washed five times. Biotin elution buffer was used to capture the bound proteins by shaking at room temperature for 15 min. Finally, elution was performed for subsequent Western blot analysis.

Western blotting analysis

Protein expression in cells and animal tissues was detected using Western blot analysis. Cell proteins were extracted using radioimmunoprecipitation assay buffer (RIPA buffer) (9229; Absin, Shanghai, China) containing a protease inhibitor (M5293, AbMole, Shanghai, China). Tissue protein samples were extracted using the Minute^TM total protein Extraction Kit for Animal Tissue (SD001; Invitrogen, Beijing, China). Equal amounts of protein were separated using SDS-polyacrylamide gel electrophoresis and electrophoretically transferred to PVDF membranes (ISEQ00010, Merck Millipore). After incubation with 5% bovine serum albumin or milk at room temperature for 2 h, the membranes were incubated overnight at 4 °C with primary antibodies against RPL6 (126100, Abcam), PCBP2 (83017, Cell Signaling Technology), MHC-IB (PA5-88804, ThermoFisher), β2-microglobulin (126100, Abcam), Fibronectin (2413, Abcam), p-Akt (4060, Cell Signaling Technology), Akt (4091, Cell Signaling Technology), p-Erk (4377, Cell Signaling Technology), Erk (44695, Cell Signaling Technology), α-SMA (7817, Abcam), Collagen I (25310, Abclonal), Hsp70 (5439, Abcam), and tubulin-α (32293, Santa Cruz Biotechnology). The bound antibodies were then incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody (7074/7076; Cell Signaling Technology) for 1 h at room temperature. The membranes were then scanned using a chemiluminescence/fluorescence image analysis system (Tanon, Shanghai, China).

CKD population

The study complied with the Declaration of Helsinki, and all participants provided written informed consent. The study was approved by the ethics committee of Shanghai Fifth People’s Hospital (No. 2018-123×2). CKD was defined as eGFR <60 ml/min per 1.73 m² or UACR ≥ 30 mg/g based on the Kidney Disease Improving Global Outcomes (KDIGO) 2012 definitions³⁴. Twenty-eight patients with CKD were recruited from the Department of Nephrology of the Shanghai Fifth People’s Hospital. The clinical data and laboratory parameters of the patients with CKD are shown in Supplementary Fig. 1.

The CKD stage was categorised based on KDIGO 2012 as follows: stage 1, GFR ≥ 90 ml/min per 1.73 m²; stage 2, GFR 60 to <90 ml/min per 1.73 m²; stage 3, GFR 30 to < 60 ml/min per 1.73 m²; stage 4, GFR 15 to < 30 ml/min per 1.73 m²; and stage 5, GFR < 15 ml/min per 1.73 m². eGFR was calculated using the calibrated serum creatinine values by applying the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation³⁵.

Enzyme-linked immunosorbent assay

The RPL6 concentration in the urine of patients was detected using an enzyme-linked immunosorbent assay (ELISA) kit (JLC_Y10323, Gelatins, Shanghai), according to the manufacturer’s instructions.

Flow cytometry, cytometric bead array and cell viability detection

To detect β2-microglobulin and MHC-IB expression in HK-2 cells, the cells were fixed with 4% paraformaldehyde in PBS for 10 min, followed by incubation with an APC-conjugated anti-human β2-microglobulin antibody (316311, Biolegend) and an anti-MHC-1 antibody (ab70328, Abcam) for 1 h at room temperature. A PE-conjugated anti-mouse antibody (12-4010-82, eBioscience) was used to detect the anti-MHC-1 antibody. The cells were washed thrice and analysed using the BD FACSCanto™ II system.

For NK cell populations, the sample preparation procedure is illustrated in Fig. 8A. Briefly, sera from patients with CKD were added to the culture medium at a dilution of 1:100. HK-2 cells were incubated in medium with or without pAb (2 μg/ml). After 72 h, the medium was discarded, and HK-2 cells were washed thrice with ice-cold PBS. Fresh medium containing natural killer (NK) cells from healthy individuals was added to the culture dish. NK cells were isolated using the Dynabeads™ Untouched™ human NK kit (130-092-657, ThermoFisher). The fresh medium comprised Dulbecco’s modified Eagle’s medium/nutrient mixture F-12 and CTS NK-Xpander Medium (A5019001, ThermoFisher). The 1:1 mixture was supplemented with 1800 IU/mL IL-2 (HZ-1015; Proteintech) and 5% human serum. Lirilumab (IPH2102, MCE), a fully human IgG4 anti-KIR2DL1, -L2, -L3, -S1, -S2, was used to block the interaction between KIR and MHC-I. The E:T ratio was 1.25:1. After 24 h, NK cells were collected and incubated with antibodies against CD3 (317341, BioLegend), CD16 (302022, BioLegend), CD56 (362549, BioLegend), and NKG2A (375139, BioLegend) for 1 h at room temperature. NK cell populations were then analysed using the BD FACSCanto II system. Cytokines in the supernatant were detected using a cytometric bead array. The retained HK-2 cells were prepared for Western blotting analysis. Further, HK-2 cells were cultured in 96-well plates and treated as described above; cell viability was then detected using the Cell Counting Kit-8 (CCK-8; HY-K0301, MedChemExpress).

Statistical analysis

Image Pro Plus 6.0 software (Media Cybernetics) was used to measure the positive value for immunohistochemistry experiments. The Intensity (SUM) or IOD (SUM) of staining regions was calculated in each image. ImageJ software (version 1.53e; NIH) was used to analyse the positive area (IntDen) or overlay curve (Gray value/Distance) in fluorescence staining and western blotting. Uncropped blots are shown in Supplementary Fig. 2. For flow cytometry analysis, the cell frequency was obtained using BD FACS Celesta software. The ELISA and cell viability values were obtained using a Microplate Photometer (infinite M200pro, TECAN). The EGFR phosphorylation level was calculated according to the formula: Y845 (normalised signal) = Positive (1) / Positive (2) × Y845.

GraphPad Prism 10.0 (GraphPad Software) software was used for statistical analysis. The data are expressed as the mean ± standard error of the mean (SEM). Correlation coefficients between laboratory parameters and RPL6 were analysed as adjusted Spearman’s correlation (rs). Differences between groups were evaluated for statistical significance using one-way analysis of variance (ANOVA) with Tukey’s post hoc test or two-way ANOVA, and statistical significance was set at p < 0.05.

Author contributions

Lin Yang, Qing Wu, and Yu Chen conducted all the experiments and prepared the manuscript. Di Zhao and Yuxuan Zhou participated in the in vitro experiments. Xiaojing Zhong and ZhenHao Yang participated in the clinical data collection and analysis. Lin Yang and Jianying Niu conceived the study and interpreted data.

Data availability

No novel protein sequences, DNA/RNA sequences, high-throughput sequencing data or other sequencing/structure data were generated in this study. The siRNA sequences, truncated RNA probes for RNA pull-down assay, and peptide sequences screened from the phage library are artificially designed, which have been listed in the manuscript. The datasets generated and/or analyzed during the current study are not publicly available due to reasons of sensitivity but are available from the corresponding author on request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41541-026-01513-w.