KLK15 alters connective tissues in hypermobile Ehlers-Danlos syndrome

Summary

Hypermobile Ehlers-Danlos syndrome (hEDS) is a debilitating multisystem condition characterized by joint hypermobility, chronic pain, and diverse comorbidities, yet its genetic basis remains undefined. Whole-exome sequencing (WES) of 200 patients with hEDS revealed rare and low frequency variants in 14 of 15 kallikrein (KLK) genes, including a recurrent KLK15 missense variant (p.Gly226Asp) segregating in multiple families. KLK15, a secreted serine protease, is expressed in connective and immune tissues and interacts with extracellular matrix (ECM) components, including fibronectin and lysyl oxidase (LOX). A KLK15 knock-in mouse model recapitulated hEDS features in tendons and cardiac valves and exhibited dysregulated cytokine profiles. The variant altered KLK15 and LOX compartmentalization within the ECM, consistent with a dominant-negative effect. These findings identify KLK15 as a contributor to hEDS and reveal broader roles for KLK protease-ECM-immune crosstalk in connective tissue regulation. This study reframes hEDS as a condition involving matrix remodeling and immune signaling beyond collagen defects.

Graphical abstract

Highlights

• •KLK15 identified as a novel disease gene in multiple families with hEDS
• •Knock-in mice have tendon, skin, and heart valve abnormalities similar to patients
• •Knock-in mice show altered immune signaling and extracellular matrix organization
• •Findings expand hEDS pathogenesis to proteolytic and immune dysregulation

Body system; Non-infectious disease; Biological process; Molecular genetics

Introduction

Ehlers-Danlos syndromes (EDS) comprise a heterogeneous group of heritable connective tissue disorders characterized by joint hypermobility, skin hyperextensibility, and tissue fragility.¹ While most EDS subtypes have known genetic etiologies involving genes encoding collagens or collagen-modifying enzymes, the most common subtype, hypermobile EDS (hEDS), remains genetically elusive. Patients with hEDS often experience widespread musculoskeletal pain, autonomic dysfunction, gastrointestinal dysmotility, and immune dysregulation, suggesting a multifactorial etiology that extends beyond structural connective tissue abnormalities.²

The clinical complexity of hEDS has prompted our investigations into candidate gene families that may influence diverse physiological systems involved in the disease. One such candidate is the kallikrein (KLK) gene family, a cluster of 15 serine protease-encoding genes located on chromosome 19 (Figure S1). These genes are among the most evolutionarily conserved and functionally versatile in the human genome.³ KLKs regulate proteolytic cascades that govern numerous biological processes, including extracellular matrix (ECM) remodeling,^4,5,6,7,8,9 inflammation,^10,11,12 vascular tone and autonomic regulation,¹³ connective tissue integrity,^12,14 and endocrine signaling.¹⁵ Their enzymatic activity, mediated by a conserved catalytic triad, enables precise substrate cleavage and is tightly controlled through hormonal signaling, transcriptional regulation, and post-translational modifications.

KLKs act at the interface of multiple systems disrupted in hEDS. For instance, KLKs can activate or degrade ECM components such as fibronectin (FN1), collagen, and elastin; modulate cytokines and chemokines that influence mast cell and fibroblast function; and regulate the release of vasoactive peptides such as bradykinin.^{15,16,17,18,19,20,21} Dysregulation of these pathways can result in hallmark features of hEDS, including connective tissue laxity, neuroinflammation, dysautonomia, and chronic pain. Furthermore, expression of members of the KLK gene family are enriched in tissues and cell types relevant to hEDS pathogenesis, including connective tissues, glandular epithelia, sensory neurons, and immune cells such as mast cells. Given their broad functional roles and widespread tissue distributions, the KLK gene family emerges as a biologically plausible contributor to hEDS. Despite their extensive involvement in physiological regulation, KLKs have been largely overlooked in the context of heritable connective tissue disorders. Investigating the genetic variation and functional impact of KLK family members in individuals with hEDS may provide novel insights into disease mechanisms and uncover potential therapeutic targets. The present study explores this hypothesis through integrated genetic, molecular, and functional analyses of KLK genes in patients with hEDS.

Results

Enrichment of KLK variants in hEDS

To investigate the potential involvement of KLK genes in hEDS, we performed whole-exome sequencing (WES) on a cohort of 207 clinically diagnosed individuals, based on the 2017 hEDS diagnostic criteria. One sample failed quality control (QC) and six individuals were carriers for known pathogenic variants in other connective tissue diseases. Of the remaining 200 individuals, no pathogenic variants were identified in genes previously associated with other connective tissue disorders, including the rarer subtypes of EDS. As previous reports have made claims for the involvement of the methylenetetrahydrofolate reductase (MTHFR) gene in hEDS, we also performed an initial screen for these variants in our cohort.²² We did not observe a significant enrichment of the common MTHFR polymorphisms C677T (p = 0.9864) or A1298C (p = 0.3156) when compared to GnomAD v2.1.1 non-Finnish European population. Thus, the high population frequency of these variants and lack of replication in our cohort render these common MTHFR variants as unlikely to cause hEDS.

Guided by a biologically plausible hypothesis that KLK gene variants may contribute to the pathogenesis of hEDS, we conducted a targeted analysis of the KLK gene cluster. After applying filtering criteria as described in the STAR Methods, we identified 49 unique rare variants (79 total alleles) across the cohort. Notably, 67 individuals (33.5%) carried at least one rare or low-frequency variant in a KLK gene (Table S1). Efforts to expand pedigrees in which identical variants were observed led to the identification of two families carrying the same missense KLK15 variant, p.Gly226Asp (c.677G>A; rs113529853), a missense substitution with a high predicted pathogenicity (Combined Annotation Dependent Depletion -CADD = 23.6) and low population frequency (Minor Allele Frequency - MAF = 0.002).

Familial segregation of KLK15^G226D/+

Family 1 was a four-generation family of European ancestry presenting with autosomal dominant hEDS (Figure 1A). Eleven family members were enrolled for genetic analysis, including five individuals who were clinically evaluated and met the 2017 clinical diagnostic criteria for hEDS²³ and three who were classified as probable hEDS. Due to the subjective nature of the criteria and changes in presentation with age, probable cases included either older adults with a clinical history and phenotype highly consistent with hEDS, or younger children who exhibited signs and symptoms of hEDS but were too young at the time of enrollment to fully meet the 2017 criteria (Figures 1A and S2).

Figure 1.

Kallikrein-15 enrichment and familial gene variant identification in hEDS

(A) Pedigree of a multigenerational family with autosomal dominant hEDS. Black circles and squares represent those with a clinical hEDS diagnosis. Gray indicates those who are probable-hEDS. Unaffected individuals are black and unknown phenotypes and those whose DNA was not available are marked with an asterisk. Circles and squares represent females and males, respectively. WES was performed on the proband (IV-1) and IV-4 (arrow and arrowhead).

(B) Chromatogram showing missense KLK15 variant (G/A).

(C) Pedigree of family 2 showing genotype of KLK15^G226D/+ allele (G/A) in affected I-1 and II-1.

(D) KLK15 variants identified in WES analyses of 198 hEDS individuals. Consequence of the variants in changing splicing or coding is represented as are the MAF (gnomAD v2.1.1), SIFT, PolyPhen, and CADD (v1.7) scores.

PCR genotyping and Sanger sequencing of the KLK15^G226D variant (Figure 1B) identified in the proband revealed a perfect phenotype-genotype correlation: none of the unaffected family members carried the variant, while all affected individuals did. To explore whether additional variants might contribute to disease within this family, WES was performed on the most distantly affected relative, a second cousin (IV-4). After filtering, between the proband (IV-1) and the second cousin (IV-4) (see STAR Methods), only four variants remained. Each variant was PCR-genotyped across all enrolled family members. Notably, the KLK15^G226D variant was the only shared variant across the family that exhibited perfect segregation with the disease phenotype, further supporting its role in hEDS. In the second family (family 2), a trio of European ancestry, including an affected mother and daughter, also underwent genotyping for the KLK15^G226D variant (Figures 1C and S2). The results confirmed segregation of the variant with the disease phenotype in this family as well. In addition to the recurrent p.Gly226Asp variant, we identified five other rare KLK15 variants in the WES cohort that passed filtering criteria, collectively resulting in a significant enrichment of KLK15 variants in patients with hEDS compared to controls (Figure 1D). Gene burden analysis using Fisher’s exact test revealed a markedly increased burden of KLK15 variants in the hEDS cohort relative to gnomAD controls (p = 5.5 × 10⁻⁵). The identification of the KLK15 p.G226D variant in two pedigrees led to the prioritization of this candidate variant through expression studies in addition to in vitro and in vivo analyses.

KLK15 expression

To correlate KLK15 with relevant tissue types affected in patients with hEDS, KLK15 mRNA expression was investigated using in silico and experimental approaches. Analysis of the GTEx dataset revealed that KLK15 is broadly expressed across various human tissues (Figure S3), with the highest levels observed in glandular tissues, such as the thyroid and adrenal glands, and notable expression in gastrointestinal tissues, including the colon and stomach. Most other tissues exhibited low, but detectable levels of KLK15 transcripts. Experimental validation through Reverse Transcriptase PCR (RT-PCR) confirmed the presence of KLK15 transcripts in human tissues, including the anterior cruciate ligament (ACL), dermal fibroblasts, and mast cells (Figures 2A and B). This expression pattern was conserved across species, as Klk15 transcripts were also detected in mouse tissues, including the thymus, skin, Achilles tendon, and adrenal gland (Figure 2C). RNAscope analysis further demonstrated widespread Klk15 expression throughout the murine adrenal gland, with skin-expression localized primarily to the epidermis and dermis (Figure 2D). These findings demonstrate that KLK15 is broadly but variably expressed, with distinct tissue-specific and cell-specific patterns conserved in both humans and mice, indicating its potential involvement in diverse physiological processes. Importantly, expression of KLK15 mRNA was detected in tissues that are known to be involved in hEDS pathologies and its various co-morbidities.

Figure 2.

Expression of KLK15

(A and B) RT-PCR showing KLK15 is expressed by human dermal fibroblasts, anterior cruciate ligament tissue (ACL), and in human mast cells.

(C) Expression of Klk15 mRNA is observed in subset of murine tissues including thymus, skin, and Achilles tendon.

(D) RNA scope showing positive signal throughout the murine adrenal gland and within the skin.

RT (±) is presence or absence (negative control) of reverse transcriptase enzyme. Scale bars, 100 μm.

KLK15 interactome

With expression of Klk15 being detected in the skin, combined with dermal phenotypes in patients with hEDS, we wished to further investigate if KLK15 is able to interact with relevant connective tissues proteins known to be altered in patients with hEDS. Thus, a yeast two-hybrid screen was performed using a human dermal fibroblast cDNA library with the processed human KLK15 protein as bait. The screen analyzed 194.7 million clones and identified a total of 128 clones, 91 of which were unique. Among these, 23 unique proteins were identified as interacting with KLK15 (Figure 3A). Notably, 10 of the 23 unique proteins (44%) were components of the ECM and were implicated in collagen processes. Some of the remaining proteins were linked to immune function, while others played roles in diverse cellular processes, including transcriptional regulation and repair, protein transport, and intracellular signaling. Of the identified interactors, six out of the top eight proteins with the highest confidence scores were ECM components. Validation experiments focused on the top two ECM-related hits, FN1 and lysyl oxidase (LOX), due to their well-established roles in collagen stabilization and involvement in connective tissue pathways. Overlapping clone sequences for FN1 and LOX identified by the two-hybrid screen enabled mapping of specific KLK15 interaction domains within these proteins (Figure 3B). For LOX, the interaction was mapped to the distal region of the signal sequence and the LOX domain, which are essential for its enzymatic activity. In FN1, KLK15 interacted with the amino terminus of the mature FN1 protein, a domain critical for FN1 dimerization and binding to fibrin and heparin. Recent studies have uniquely detected this KLK15-FN interaction motif, present as a potential cleavage fragment in blood samples from patients with hEDS and related hypermobility spectrum disorders (HSDs), suggesting a possible disease-relevant mechanism.²⁴ RNAscope analysis further demonstrated co-expression or spatial proximity of KLK15, FN1, and LOX mRNAs within the connective tissue of the skin, supporting their potential functional interplay (Figures 3C and 3D). Co-immunoprecipitation experiments validated the yeast two-hybrid results, confirming that KLK15 interacts with FN1 and LOX either directly or as part of a larger protein complex (Figures 3E and 3G).

Figure 3.

Identification of KLK15 interacting proteins

(A) Results from two-hybrid screens identifying a total of 23 novel protein partners with FN1 and LOX being highlighted as highest priority hits relevant to the extracellular matrix.

(B) Schematic of binding region where the KLK15 bait interacts with FN1 and LOX. s.s., signal sequence; SID, smallest interacting domain.

(C and D) RNAscope showing punctate signals from all three probes appear in close proximity for KLK15, Fibronectin, and LOX in the skin, suggesting co-expression in the same cells. Scale bars are 30 and 120 μm. C′ and D′ are 4× zoom images of (C) and (D), respectively.

(E) CoIP experiments demonstrating that KLK15 interacts with the secreted form of fibronectin.

(F and G) CoIP experiments demonstrating that KLK15 interacts with transfected LOX in both cell extracts and media samples.

Generation and characterization of a Klk15^G224D/+ mouse model

The discovery of KLK15 variants segregating in multiple hEDS families, coupled with its expression in affected tissues and immune cell types, and its interaction with ECM and immune proteins, supported a role for KLK15 in connective tissue biology and hEDS pathogenesis. These converging lines of evidence provided the rationale to generate a knock-in mouse model to directly assess the in vivo consequences of the disease-associated variant. Thus, CRISPR-Cas9 was used to generate knock-in mice carrying the KLK15 variant identified in the two affected families (Figure S4). The Achilles tendon, a tissue with notable Klk15 mRNA expression in mice (Figure 2C), was selected for analysis due to the incidence of Achilles tendon rupture in members of family 1 (Figure S2), making it an ideal model to assess structural and functional deficits associated with the variant. Freshly isolated Achilles tendons from Klk15^G224D/+ (n = 8) and control (n = 9) mice underwent mechanical testing. Microcomputed tomography (μCT) analysis confirmed that Klk15^G224D/+ tendons were anatomically comparable to controls in overall dimensions (Figures S5 and S6). Stress-strain curves revealed significant mechanical differences with Klk15^G224D/+ tendons exhibiting greater displacement, increased strain, and a reduced toe modulus, consistent with a more elastic tissue phenotype (Figures 4A–4E). This extended toe region occurred without significant changes in deformation at endpoints, aligning with previous studies on ligament injuries.^25,26 These findings suggest a structural collagen deficit as the underlying cause. To investigate a potential collagen phenotype further, ultrastructural analysis of Achilles tendons was performed using transmission electron microscopy (TEM) (Figures 4F–4H). Quantification of collagen fibrils across 20 independent regions from five control (Klk15^+/+) mice (n = 3,240 fibrils) and 24 regions from six Klk15^G224D/+ mice (n = 4,191 fibrils) revealed a 20% reduction in fibril diameter in mutant mice compared to controls (p < 0.0001; Figure 4G). Binning analysis in 10 nm increments showed a shift toward smaller fibril diameters in Klk15^G224D/+ tendons (Figure 4H), correlating with increased elasticity and reduced mechanical strength. These results align with previous studies linking reduced collagen fibril diameter to compromised biomechanical properties.^27,28

Figure 4.

Pathogenicity of the familial KLK15 variant

(A and B) Stress-strain curves from tensile testing of Klk15^+/+ (n = 9) and Klk15^G224D/+ (n = 8) tendons with a focus on the toe region (B).

(C–E) Statistical analysis of transition point displacement (unpaired t test, p = 0.0253, Cohen’s d = −1.21), transition strain (Mann-Whitney U test, p = 0.036, rank-biserial r = −0.051) and toe modulus (unpaired t test, p = 0.0218, Cohen’s d = 1.24) from tensile testing curves. Error bars represent mean ± standard deviation.

(F) Representative TEMs of Achilles tendon from Klk15^+/+ and Klk15^G224D/+ mice showing smaller collagen fibrils in the mutant tendons.

(G) Quantification of fibril diameters from wildtype (n = 3,240) and mutant (n = 4,191) tendons showing a significant decrease in average fibril diameter in Klk15^G224D/+ (84 vs. 101 nm). Statistical analysis was performed using a Mann-Whitney U test, ∗∗∗∗p < 0.0001, with an effect size (rank-biserial correlation, r) of −0.242. Bars represent data means.

(H) Distribution of fibril diameters in 10 nm increments showing a leftward shift in diameter.

(I) Echocardiography of 4-month-old wild-type (Klk15^+/+;n = 5) and mutant mice (Klk15^G224D/+; n = 6) showing mitral valve prolapse in 5/6 Klk15^G224D/+ mice above the level of the annulus (yellow line).

(J) Movats pentachrome stains revealed myxomatous mitral (arrow heads) and aortic leaflets (AV).

Areas of chondrodysplasia and proteoglycan accumulation are evident in the hinge region of the aortic valve (arrows).

Red, myocytes; blue, proteoglycans; black, elastin.

Given that mitral valve prolapse (MVP) is a common feature in patients with hEDS,² along with multiple individuals in family 1 diagnosed with MVP, we assessed cardiac phenotypes in Klk15^G224D/+ (n = 6) and control Klk15^+/+ (n = 5) mice. Echocardiography and histopathological analysis were performed to evaluate potential structural and functional cardiac abnormalities associated with the Klk15^G224D variant. While overall cardiac function was unchanged (Figure S7), 83% of mutant mice (5/6) exhibited valve dysfunction and MVP (Figure 4I). In contrast, no valve defects were detected in control mice (0/5). Histological analysis using Movat’s pentachrome stain revealed myxomatous degeneration of the mitral and aortic valves in 75% (3/4) of mutant mice, compared to none of the controls (0/4; Figure 4J). These findings in the Klk15^G224D/+ mouse model provide strong validation for the pathogenicity of the KLK15 variant identified in the hEDS cohort, linking it to both structural and functional deficits that are directly relevant to the clinical phenotypes observed in the affected individuals.

To explore potential mechanisms underlying the connective tissue defects, we leveraged our interactome data (Figure 3) to assess LOX expression and localization in the hEDS mouse model. Ultrastructural analyses in the mutant mice revealed alterations in collagen fibrils, suggesting potential defects in collagen cross-linking enzymes, providing rationale for a focused evaluation of LOX expression in the Klk15^G224D/+ mouse model. Consistent with the observed structural phenotype, we identified statistical differences in the distribution of both LOX and KLK15 within tail biopsies from control and mutant mice (Figure 5). In the Klk15^G224D/+ background, the ratio of soluble to insoluble LOX levels were significantly reduced, with a reciprocal increase in KLK15 levels compared to the subcellular localization of these proteins in control littermates. These findings demonstrate that the KLK15 variant not only undermines the structural and mechanical stability of connective tissues but also triggers molecular alterations, notably disrupting the expression and distribution of LOX within the connective tissue.

Figure 5.

Dominant impact of KLK15 variant on LOX and KLK15 expression

(A–C) Western analyses of tail biopsies from control (Klk15^+/+; n = 3) and mutant (Klk15^G224D/+; n = 5) mice probed for LOX expression in soluble or insoluble protein fractions. Significant shift in the ratio of LOX protein within the extracellular matrix is observed in the Klk15^G224D/+ biopsies compared to controls (Unpaired t test, p = 0.0073, Cohen’s d = −0.252).

(D–F) Western analyses of the same biopsies shown in (A–C) were probed for Klk15 showing an inverse relationship to LOX expression with significantly more Klk15 present in the insoluble fraction (unpaired t test, p = 0.008, Cohen’s d = −2.472) and a significant reduction in soluble to insoluble ratio (Welch’s t test, p = 0.0454, Cohen’s d = 3.112).

Error bars represent means ± standard deviation. These data suggest a dominant impact of the Klk15 variant on molecular changes in the connective tissues.

To provide additional support for the pathogenicity of Klk15^G224D variant, we performed cytokine profiling in serum samples from both mutant and wild-type mice using a multiplex panel of 111 analytes. This analysis identified eleven cytokines with statistically significant downregulation in mutant mice compared to controls (Figure S8). Changes were observed in FGF21, G-CSF, GM-CSF, IL-4, MIP3α, CXCL10, EGF, IL-6, IL-5, IL-27 p28, and IL-2. Each of these differentially expressed cytokines can regulate inflammatory pathways, suggesting that a key consequence of the Klk15 variant is altered immune signaling. The cytokine alterations observed in Klk15^G224D/+ mice align with immune manifestations reported in some patients with hEDS and may contribute to impaired tissue remodeling. This immune dysregulation not only links inflammatory signaling changes to structural abnormalities seen in tendon, valve, and tail, but may also exacerbate underlying connective tissue defects. These findings further support the pathogenic role of KLK15 variants in hEDS and provide evidence that dysregulated immune signaling maybe a chronic driver of hEDS pathogenesis.

Discussion

hEDS remains a challenging disorder to define at the molecular level, despite being the most common form of EDS. In contrast to rarer subtypes with established genetic etiologies involving collagen and collagen-modifying enzymes, the absence of a consistent genetic signature in hEDS has hindered diagnostic accuracy, mechanistic understanding, and therapeutic development. Our study represents a step forward by identifying KLK15, a secreted serine protease, as a disease-associated gene in hEDS and providing experimental evidence for its involvement in connective tissue dysfunction and immune dysregulation.

Investigation of the KLK gene family stemmed from striking parallels between the known biological functions of KLK proteins and the diverse phenotypic features of hEDS. KLKs regulate ECM remodeling, neurovascular signaling, immune modulation, and epithelial barrier integrity, all systems affected in hEDS. WES of 200 patients clinically diagnosed with hEDS revealed rare and predicted damaging variants in 14 of the 15 KLK genes, with a recurrent missense variant in KLK15 (p.Gly226Asp) segregating in two independent families. This level of convergence is notable in a genetically complex and heterogeneous condition such as hEDS and provides the potential for multiple KLK genes to be involved in hEDS.

Functional characterization of KLK15 provided multiple, independent lines of evidence supporting its pathogenic role. KLK15 was found to be broadly expressed in connective and glandular tissues, as well as in immune cells, including mast cells, cell types increasingly implicated in hEDS comorbidities. Protein interaction studies revealed molecular interactions between KLK15 and FN1 and LOX, both key mediators of ECM structure and stability. The FN1 interaction domain overlapped with regions recently identified as circulating biomarkers in patients with hEDS and HSD,²⁴ reinforcing the clinical relevance of this interaction.

While our data confirm a physical interaction between KLK15 and LOX, the possible consequences of this interaction have not yet been fully explored. LOX is known for its role in collagen crosslinking; however, emerging evidence has also demonstrated a connection to inflammation and immune signaling.²⁹ Given that LOX may have multiple roles, future studies should investigate LOX localization or activity in the context of immune cell function and cytokine secretion. We acknowledge that these observations do not establish a direct mechanistic link between KLK15 and LOX activity, and beyond the in vivo data discussed subsequently, future studies will be needed to fully understand the mechanisms at play.

The in vivo consequences of the p.Gly226Asp variant were evaluated in a CRISPR-Cas9-generated Klk15^G224D/+ knock-in mouse model. Our decision to analyze the Achilles tendon and the mitral valve was part of a deliberate effort to assess the systemic effects of the KLK15 variant across multiple connective tissues. Because hEDS is a multisystemic connective tissue disorder, it is critical to evaluate more than a single tissue type in order to capture a more complete phenotype-genotype correlation. In fact, our tissue selection was guided by the clinical phenotypes observed in affected family members, where the most prominent connective tissue manifestations were Achilles tendon rupture, MVP, and skin extensibility. These specific phenotypes naturally led us to focus on those three tissue types, allowing us to determine whether the variant exerts a broad, tissue-independent effect on connective tissue biology. Consistent with the patients, the mice exhibited phenotypes that closely mirrored clinical features seen in patients with hEDS, including altered tendon biomechanics, reduced collagen fibril diameter, and MVP. Notably, the Achilles tendons of mutant mice demonstrated increased elasticity and impaired mechanical strength, supported by ultrastructural evidence of collagen fibril thinning, findings consistent with prior reports of collagen abnormalities in hEDS tissues.³⁰ Moreover, the observation of valve degeneration in the majority of mutant mice strengthens the link between KLK15 dysfunction and cardiac involvement in hEDS, a feature seen in affected individuals from the KLK15-positive families.

Beyond structural defects, Klk15^G224D/+ mice also demonstrated significant molecular changes. First, the KLK15 variant exerted a dominant-negative effect in vivo, influencing the distribution of both KLK15 and LOX and depleting LOX from insoluble ECM compartments within the tail, providing a potential mechanism for altered collagen cross-linking and matrix integrity. Second, beyond ECM disruption, Klk15^G224D/+ mice exhibited systemic changes in immune signaling, as evidenced by the consistent downregulation of eleven circulating cytokines, many of which are key regulators of inflammation, tissue remodeling, and immune cell recruitment. While these immune findings provide valuable insights, further studies are needed to more precisely link KLK gene variants and LOX disruptions with immunological consequences. Regardless, the coordinated suppression in cytokine profiles suggests a dampened or dysregulated immune tone that may impair normal repair or homeostatic processes.

Additional links between KLKs, ECM remodeling, and immune signaling is further supported by prior studies showing that KLKs extend their function beyond structural roles to influence immune homeostasis, blood pressure regulation, and complement activation. For example, some KLKs cleave complement proteins C3 and C5, generating C3a and C5a, potent inflammatory mediators, and mast cell activators.^31,32,33 Dysregulation of this axis has been implicated in autoimmune and inflammatory conditions with features overlapping hEDS, including lupus, rheumatoid arthritis, multiple sclerosis, Sjögren’s syndrome, and hereditary angioedema.^{34,35,36,37,38,39} Additionally, KLKs have been shown to modulate transforming growth factor β signaling, stimulate cytokine production and pro-inflammatory signaling, and cleave immune-related proteins such as GDF15 and MIF, further supporting KLKs as modulators of immune responses.^40,41 The convergence of autoimmunity, mast cell activation, and autonomic dysfunction, hallmarks of POTS, MCAS, and rheumatologic overlap syndromes frequently seen in hEDS, suggests a shared upstream pathological mechanism potentially linked to the KLK system. Although our current data do not define a direct immunopathogenic mechanism, they support an indirect but biologically meaningful role for KLKs in modulating immune signaling. Clarifying this role will be critical for understanding the immunologic comorbidities of hEDS and for guiding future mechanistic and therapeutic studies.

The convergence of structural, molecular, and immunological cytokine abnormalities in this model highlights the multifactorial nature of hEDS pathogenesis and aligns with emerging perspectives that view hEDS as a complex multisystem disorder, not solely a structural collagenopathy. Taken together, these findings identify KLK15 as a novel disease-associated gene in hEDS, and more broadly, position the KLK family as a previously unrecognized contributor to connective tissue biology. This study shifts the paradigm of hEDS pathogenesis beyond collagen-centric mechanisms to introduce proteolytic signaling as a central regulatory axis. These insights open new avenues for biomarker development, mechanistic studies, and therapeutic exploration targeting KLK pathways.

Individuals with hEDS who carry rare or low frequency variants in the KLK gene family likely represent only a subset of the broader genetic landscape underlying this disorder. A key conclusion from this study is that hEDS is likely to be polygenic in most cases, involving both rare and common variants across the genome. Moreover, the interplay between these genetic risk factors, epigenetic regulation, and environmental influences will be critical to fully understand the complexity of hEDS pathogenesis. These findings underscore the need for large-scale genome-wide association studies, integrative multi-omic analyses, and longitudinal patient profiling to identify additional genetic modifiers and define how molecular networks are shaped by both inherited and external factors in hEDS. Nonetheless, insights from this study, coupled with increased physician awareness and refined diagnostic criteria, will be essential for uncovering the causes, triggers, and progression of hEDS. While structural connective tissue defects are well documented in hEDS, it remains unclear whether these abnormalities represent a primary pathogenic mechanism, a secondary consequence of immune dysregulation, or the result of disruption in another upstream regulatory pathway. Although the data presented here do not definitively resolve this question, they establish a foundational framework through which these hypotheses can be tested. Addressing this question at the molecular and cellular levels will be crucial, as rational therapeutic development depends on an accurate understanding of both disease onset and progression. Moreover, our results underscore the necessity of integrating genetic discoveries with functional molecular, cellular, and in vivo analyses, as genetic findings alone, without validation, will fall woefully short in fully explaining the complexities of this disease.

Limitations of the study

• (1)Genetic findings and relevance: The KLK15 p.Gly226Asp variant was identified in two multigenerational families, with clear segregation supporting its relevance despite limited generalizability. Additional rare KLK variants were observed, but functional validation was limited to KLK15, warranting further study across the KLK gene family.
• (2)Model and mechanistic limitations: The KLK15 knock-in mouse model replicates key hEDS features but may not reflect the full multisystemic phenotype. Similarly, all identified KLK15 binding partners have not yet been fully characterized in terms of downstream functional consequences, particularly in immune and autonomic systems.
• (3)Scope and diagnostic considerations: Proteomic and cytokine data provide insight into immune dysregulation but are limited by their cross-sectional design. Importantly, this study does not propose a diagnostic framework, and the absence of KLK15 variants should not preclude clinical diagnosis.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Russell A. Norris (norrisra@musc.edu).

Materials availability

Materials generated in this study will be made available as allowable by MUSC policy.

Data and code availability

• •Data reported in this paper will be shared by the lead contact upon request.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This work was supported by the National Institutes of Health, grant numbers: R01HL131546 (R.A.N.), P20GM103444 (R.A.N.), R01HL149696 (R.A.N.), F32AI181339 (C.G.), T32HL007260 (T.B., C.G., and K.M.), F31HL167482 (J.M.), F31152494 (K.M.), F31HL158243 (T.B.), T32GM132055 (C.G.), CA207729 (J.R.D.), R01DE021134 (H.Y.), U01DE031512 (H.Y.), P20GM121342 and (H.Y.); American Heart Association: 19TPA34850095 (R.A.N.); Ehlers-Danlos Society (R.A.N.); Connective Tissue Coalition (R.A.N.); Ines Mandl Foundation (R.A.N.); The Fullerton Foundation (R.A.N.); and the Maltz Foundation (R.A.N.). This work was supported in part through the computational and data resources and staff expertise provided by Scientific Computing and Data at the Icahn School of Medicine at Mount Sinai and supported by the Clinical and Translational Science Awards (CTSA) grant UL1TR004419 from the National Center for Advancing Translational Sciences (A.R.R.K. and I.K.). T.-E.B and N.B.-N. were supported by the European Research Council grant (ERC-Stg-ROSALIND-716628). The work at the Medical University of South Carolina was performed in a facility constructed with support from the National Institutes of Health, grant number C06 RR018823, from the Extramural Research Facilities Program of the National Center for Research Resources. Special thanks to the members of the Bobby Jones Chiari & Syringomyelia Foundation, including Dorothy Poppe, Mary Poppe, and Kaitlyn Esposito, for their support. We also express our sincerest gratitude to Jon Rodis and the Connective Tissue Coalition for their thoughtful conversations and continued support. Most importantly, we extend our deepest gratitude to the hEDS patients and study participants. Their partnership, generosity, and willingness to share their experiences have been instrumental in advancing this research, and without their contributions, these discoveries would not have been possible. Finally, the guidance and inspiration of the late Dr. Tom Borg were instrumental to this study. His visionary thinking and deep commitment to understanding the architecture of connective tissues, and the complex interplay between fibroblasts and non-fibroblasts in regulating biological processes, shaped much of the work presented here. This manuscript is dedicated to his enduring mentorship and the profound impact that he had over the past 30 years. His insistence on always asking, “So what?” continues to echo through our work and will remain a cornerstone of his scientific legacy. Graphical abstract was created with Biorender.com.

Author contributions

Conceptualization, C. Gensemer, T.P., T.B. V.D., R.A.N., N.B.-N., S.P., A.M., D.M., and S.A.S.; methodology, T.-E.B., A.R.R.K., M.L., and J.C.K.; investigation, C.A.F., S. Shapiro, F.H., R.P, R.M., S.A.K., A.A., J.G., H.Y., P.C., D.P.J., J.R.D., W.K., S. Severance, S.B., J.S., E.F., M.O., J.K., G.B., E.B., J.W., R. Byrd, D.B., S.L., K. Singleton, R.F., K. Stayer, S.D., E.D., E.R.Z., E.O., C.H., N.K., K.M., R. Biggs, I.K., J.M., L.G., K.B., A.W., C. Griggs, and M.G.; writing – original draft, C. Gensemer, T.P., V.D., M.G., and R.A.N.; writing – review and editing, C. Gensemer and R.A.N.; funding acquisition C. Gensemer, M.G., and R.A.N.; supervision R.A.N.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
HIS	Novus	cat# NBP2-61482
HA	Invitrogen	cat# MA5-27915
FN	Sigma	cat# F3648
Anti-Rabbit IgG Peroxidase	Sigma	cat# A9169-2 ML
LOX	Cell Signaling	cat# D8F2K
KLK15	LSbio	cat# C37113
Biological samples
Human Genomic DNA
Human ACL Tissue	Articular Engineering	cat# CDD-H-6800-F
Chemicals, peptides, and recombinant proteins
Fugene 4K	Promega	cat# E5911
Pierce ECL substrate	Thermo	cat# 32106
Tth111I	New England Biolabs	cat# R0185S
Critical commercial assays
iProof High Fidelity PCR Reagents	BioRad	cat# 1725331
RNeasy Kit	Qiagen	cat# 74104
qScript XLT One-Step RT-PCR	QuantaBio	cat# 76047-074
Proteome Profiler Mouse XL Cytokine Array Kit	R&D Systems	cat# ARY028
Deposited data
GTEx Portal	https://gtexportal.org/home/	ENSG00000174562.13
gnomAD	https://gnomad.broadinstitute.org/	v2.1.1
Experimental models: Cell lines
NIH3T3 Cells	ATCC	Cat# CRL-1658, RRID:CVCL_0594
LUVA Cells	Kerafast, Inc	Cat# EG1701-FP, RRID:CVCL_5G48
HDF Cells	ATCC	Cat# PCS-201-012
Experimental models: Organisms/strains
Mouse: Klk15G224D/+; c.671G>A (NM_174865.1)
Oligonucleotides
Klk15 Mouse RT PCR Primers; Forward: 5′ TGGCGACAAGGTGCTAGAAG-3′, Reverse: 5′-CGGGCAGGTTTGAAAAGTCG-3′
KLK15 Human RT PCR Primers; Forward Primer: 5′-AGTTGCTGGAAGGTGACGAG-3′, Reverse: 5′-TGGTTTCCCTGATCCACTCC-3′
Klk15 sgRNA: atcacaggggacatcgccccngg
Klk15 ssODN: caagtagctgcagacttttgtgtagacgcc aggcttggtggtagtatcacaggggacatcgtcccaggaga caatgccctgcagggcacccccacagaccaagggtcctc cggagtcaccctg
Klk15 Mouse Genotyping Primers; Forward:5′-CCCAG ATCCACTCCAAGTAGC-3′, Reverse: 5′-TGAGTTCTCT TGGGTTGTGTTG-3
POTEJ V1002M: 5′ GCTATGTTGCCCTGGACTTC 3 and 5′ ATTTGCGGTGGACAATGGAG 3’;
FRG2C D9N: 5′ AGGGACGTATAAAAGGCAGGTC 3′ and 5′ ACTAAGCCATTTCCCATCCCC 3’;
FRG2C L210M: 5′ CTCTGGTGAGTCTCTCACATGC 3′ and 5′ TCAGGGTGCTCCCAGCTTAG 3′
KLK15 G226D: 5′ TTCTGTTCCATGTCAGGCGG 3′ and 5′ TTCACTCAACCTGAGACCCC 3′
Recombinant DNA
KLK15 Clone	Origene	Cat# RC210131
LOX Clone	SIno	Cat# HG17796-CH
Software and algorithms
TRAPD	Guo MH, Plummer L, Chan Y-M, Hirschhorn JN, Lippincott MF. Burden testing of rare variants identified through exome sequencing using publicly available control data. American Journal of Human Genetics. 2018. 103(4):522–534.
Other
KLK15 Probe	Advanced Cell Diagnostics	cat#1252201-C3
LOX Probe	Advanced Cell Diagnostics	cat# 425311
FN Probe	Advanced Cell Diagnostics	cat# 316951-C2
Pierce IP Lysis Buffer	Thermo	cat# 87787
halt protease and phosphatase inhibitor cocktail	Thermo	cat# 78440
Anti-HA Magnetic Beads	Thermo	cat# 88836

Experimental model and study participant details

Animals

A knock-in mouse model carrying the orthologous mutation to the human KLK15 p.Gly226Asp variant (Klk15^G224D/+) was generated using CRISPR/Cas9 genome editing on a C57BL/6J background. Both male and female mice were used in the study, aged 2 to 5 months at the time of experimentation. Mice were housed in a specific-pathogen-free facility with a 12-h light/dark cycle and provided with standard chow and water ad libitum. Prior to tissue biopsy, mice were euthanized by isoflurane (Piramal), followed by cervical dislocation in line with the Guide for the Care and Use of Laboratory Animals (NIH publication no. 85–23, revised 1996). All experimental procedures were approved by and conducted in accordance with the Institutional Animal Care and Use Committee at the Medical University of South Carolina (protocol number: IACUC-2020-00956). Sex was balanced across experimental groups and littermates were used as experimental and control animals, when possible. The studies conducted did not show overt sex differences, however, statistical testing for sex-based effects was limited by sample size and should be explored in future studies.

Human subjects

Human participants in this study were individuals over the age of 12 and had a clinical diagnosis or a family member with a clinical diagnosis of hEDS. Both males and females were included, and gender, race, and ethnicity were self-reported. Participants were enrolled in the study under institutional review board (IRB)-approved protocols (Partners Healthcare, Boston, MA and MUSC, Charleston, SC) and participants provided written informed consent. The study was not designed to detect sex- or gender-based differences in phenotype or genetic findings, which is a limitation to the generalizability of the study results.

Characteristics of hEDS WES Cohort	n (%)
Total Participants	206 (100%)
Ethnicity
Not Hispanic or Latino	188 (91.3%)
Hispanic or Latino	10 (4.9%)
Unknown/Other/Unreported	8 (3.9%)
Race
White	195 (94.7%)
American Indian/Alaska Native	4 (1.9%)
Black or African American	1 (0.5%)
Asian	1 (0.5%)
Unknown/Other/Unreported	5 (2.4%)
Gender
Female	191 (92.7%)
Male	7 (3.4%)
Non-Binary	8 (3.9%)

Cell lines

Cell lines used in this study included primary adult human dermal fibroblasts (ATCC, PCS-201-012), NIH3T3 (ATCC Cat# CRL-1658, RRID:CVCL_0594) and human LUVA mast cells (Kerafast Cat# EG1701-FP, RRID:CVCL_5G48). Dermal fibroblasts were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. LUVA cells were maintained in StemPro-34 medium supplemented with glutaMAX, penicillin-streptomycin, and primocin. All cell cultures were maintained at 37°C with 5% CO₂. Cell lines were purchased, tested routinely for negative contamination of mycoplasma, as certified by the manufacturers, and were not further authenticated before use.

Method details

Expression analyses

Data used for the mRNA analyses of KLK15 described in this manuscript were obtained from https://gtexportal.org/home/, the GTEx Portal on May 1st, 2024. Bulk tissue mRNA expression revealed broad overlap of the KLK genes in various organ systems, although specific, hEDS relevant connective tissues were largely absent from this dataset (Figure S3), thus necessitating RT-PCR and RNAscope analyses. For RT-PCR analyses, mRNA was purified from freshly dissected mouse tissues, frozen human ACL (Articular Engineering CDD-H-6800-F), human dermal fibroblasts (ATCC, Inc), and human LUVA mast cells (Kerafast, Inc) according to RNeasy Kit (Qiagen, 74104) and cDNA synthesis was performed using qScript XLT One-Step RT-PCR Kit (QuantaBio 76047-074). Primers for mouse Klk15 were as follows: 5′-TGGCGACAAGGTGCTAGAAG-3′, 5′-CGGGCAGGTTTGAAAAGTCG-3’. Primers for human KLK15 were as follows: 5′-AGT TGC TGG AAG GTG ACG AG-3′, 5′-TGG TTT CCC TGA TCC ACT CC-3’. Negative controls included the RT-PCR reactions in the absence of reverse transcriptase. Positive bands were extracted (Qiagen, Inc), purified and confirmed as KLK15 by Sanger sequencing. RNAscope in situ hybridization was performed using the RNAscope Assay (Advanced Cell Diagnostics) following the manufacturer’s instructions. Briefly, tissue samples were fixed in 10% neutral-buffered formalin for 16–24 h at room temperature, followed by dehydration through a graded ethanol series and paraffin embedding. Sections (5 μm) were cut and mounted on Superfrost Plus slides. Tissue sections or cultured cells underwent pre-treatment steps, including target retrieval using a boiling antigen retrieval buffer and protease digestion (RNAscope Protease Plus) to optimize probe accessibility. Hybridization was performed using the RNAscope probe specific to KLK15 (catalog #1252201-C3), LOX (catalog #425311), or FN1 (catalog #316951-C2) at 40°C for 2 h in a HybEZ oven. After hybridization, signal amplification steps were carried out using the proprietary RNAscope amplification reagents. Detection was performed using the RNAscope fluorescent detection kit. Fluorophore-labeled amplification reagents were applied, and slides were counterstained with DAPI before mounting with ProLong Gold Antifade reagent. Imaging was conducted using a Leica SP8 confocal microscope.

Protein interactomes

Two hybrid screens

Yeast two-hybrid screening was performed by Hybrigenics Services, S.A.S., Evry, France, similar to our previous reports.⁴² Briefly, human KLK15 without the signal sequence (was PCR-amplified and cloned into pB27 as a C-terminal fusion to LexA (LexA-bait) and pB66 (N-GAL4-bait-C fusion). These constructs were used as a bait to screen a random-primed human fibroblast cDNA library_RP1. 194.7 million clones were screened, and a total of 193 His+ colonies were selected on medium lacking tryptophan, leucine and histidine. The prey fragments of the positive clones were amplified by PCR and sequenced at their 5′ and 3′ junctions. The resulting sequences were used to identify the corresponding interacting proteins in the GenBank database (NCBI) using a fully automated procedure. A confidence score (PBS, for Predicted Biological Score) was attributed to each interaction. The PBS is computed to assess the interaction reliability. This score represents the probability of an interaction to be non-specific. It is an e-value, primarily based on the comparison between the number of independent prey fragments found for an interaction and the chance of finding them at random (background noise). The value varies between 0 and 1. Several thresholds have been arbitrarily defined to rank the results in 4 categories from A (the highest confidence rank) to D. The PBS is also adjusted by integrating the PBS of other interactions from the database of >9,000 interaction studies at Hybrigenics in which interaction domains of the involved proteins have been found. These reciprocal interactions found in independent screens are technically very reliable and thus tagged as A, B, C, or D in order of interaction confidence.

Co-immunoprecipitation

NIH3T3cells (ATCC, Inc) were seeded at 2x10⁵/well on 6-well plates 24 h prior to transfection. Cells were co-transfected with HA-tagged KLK15 and HIS-tagged LOX plasmids using Fugene 4K transfection reagent (Promega). After 72 h, media was collected, and cells were lysed with Pierce IP Lysis Buffer (Thermo). Lysate and media were placed on ice with the addition of halt protease and phosphatase inhibitor cocktail (Thermo) to each sample. Lysate samples were sonicated briefly prior to IP. Anti-HA Magnetic Beads (Thermo) were washed in TBS-T and incubated at room temperature with media and lysate samples. Beads were collected with a magnetic stand, washed with TBS-T and eluted with 2× SDS-PAGE sample buffer with incubation at 98°. Western blots were probed for 1 h at room temperature with rabbit anti-HIS (Novus, NBP2-61482), rabbit anti-HA (Invitrogen, MA5-27915) and rabbit anti-FN (Sigma, F3648) at 1:200, followed by horseradish peroxidase (HRP)-conjugated secondary (1:5000) for 1 h at room temperature. Pierce ECL substrate (Thermo) was used for detection and imaging. Lysate and media co-IP experiments were performed in biological replicates.

Whole-exome sequencing

WES was performed on 207 patients with a clinical diagnosis of hEDS. Library preparations and sequencing reactions were performed through standard protocols. Briefly, genomic DNA sample were quantified using Qubit 2.0 Fluorometer (ThermoFisher Scientific, Waltham, MA, USA). Enrichment probes were designed against the region of interest and synthesized through Twist Biosciences – Twist Human Comprehensive Panel (South San Francisco, CA, USA). Library preparation was performed according to the manufacturer’s guidelines. Genomic DNA was fragmented by acoustic shearing with a Covaris S220 instrument. Fragmented DNAs were cleaned up and end repaired, as well as adenylated at the 3′ends. Adapters were ligated to the DNA fragments, and adapter-ligated DNA fragments were enriched with limited cycle PCR. Adapter-ligated DNA fragments were validated using Agilent TapeStation (Agilent Technologies, Palo Alto, CA, USA) and quantified using Qubit 2.0 Fluorometer. Adapter-ligated DNA fragments were hybridized with biotinylated baits. The hybrid DNAs were captured by streptavidin-coated binding beads. After extensive wash, the captured DNAs were amplified and indexed with Illumina indexing primers. Post-captured DNA libraries were validated using Agilent TapeStation (Agilent, Santa Clara, CA, USA) and quantified using Qubit 2.0 Fluorometer and Real-Time PCR (KAPA Biosystems, Wilmington, MA, USA).

The bioinformatics pipeline from FASTQ files to the variant analysis was performed suing the Franklin analysis platform (Genoox, Tel Aviv, Israel). The analysis was performed as follows: FASTQ files aligned to hg19 were transformed to BAM files using BWA. Duplicate reads were marked and filtered in the process. These processed BAM files were then put into two independent SNV calling tools: GATK haplotype caller, and FreeBayes in addition to a Franklin proprietary CNV caller. After removing 1 sample that failed QC and 6 samples that had pathogenic or likely pathogenic variants consistent with other connective tissue disease, 200 exomes were filtered for variants in all 15 KLK genes. High confidence variants were filtered for MAF <0.01 using GnomAD v2.1.1 exomes. In all cases, variants in known EDS genes, including TNXB, were ruled out. Previous reports have made claims for the involvement of MTHFR²² We did not observe a significant enrichment of the common MTHFR polymorphisms C677T (p = 0.9864) or A1298C (p = 0.3156) when compared to GnomAD v2.1.1 non-Finnish European population by Chi-square test. For an accurate synopsis of the impact of these specific MTHFR variants on human physiology, we direct the audience to statements released by the Centers for Disease Control (CDC: https://www.cdc.gov/ncbddd/folicacid/mthfr-gene-and-folic-acid.html) and American College of Medical Genetics recommendations.⁴³

For WES on expanded families: WES on the most distant relative (second cousin) in Family #1 was used to determine shared variants. As with the sporadic cohort, variants in other known EDS genes, such as TNXB, were ruled out and 4 total shared variants in our expanded Family #1 were identified that fit these filtering criteria. These variants were: POTEJ V1002M, FRG2C D9N, FRG2C L210M and KLK15 G226D. Sanger sequencing for each of these variants throughout the family was performed using the following primer sets: POTEJ V1002M: 5′ GCTATGTTGCCCTGGACTTC 3 and 5′ ATTTGCGGTGGACAATGGAG 3’; FRG2C D9N: 5′ AGGGACGTATAAAAGGCAGGTC 3′ and 5′ ACTAAGCCATTTCCCATCCCC 3’; FRG2C L210M: 5′ CTCTGGTGAGTCTCTCACATGC 3′ and 5′ TCAGGGTGCTCCCAGCTTAG 3’; KLK15 G226D: 5′ TTCTGTTCCATGTCAGGCGG 3′ and 5′ TTCACTCAACCTGAGACCCC 3′ and BioRad iProof High Fidelity PCR Reagents. PCR cleanup or gel extraction of fragments were generated and sent for sequencing of both strands using the same PCR oligomers. Sequence data were aligned using BLAST and analyzed for phenotype/genotype segregation. One variant: KLK15^G226D had a perfect phenotype/genotype transmission through Family # 1 and #2 and was functionally evaluated further as detailed below.

Associations between relevant KLK15 variants in our hEDS cohort were tested using a gene-based burden test implemented in TRAPD (Testing Rare vAriants using Public Data) package.⁴⁴ A total of 7 variants that survived filtering were identified in our cohort, with two unrelated individuals harboring the same variant. Five of the six alleles were found on gnomAD, v2.1.1.⁴⁵ Focusing on non-Finnish European populations, a total of 226 alleles were identified across 42,684 patients. These numbers were used to run a gene burden test using TRAPD.⁴⁴ For the dominant test, for each gene, we calculated the number of individuals carrying at least one qualifying variant. Coverage for the variants in gnomAD v2.1.1 was also considered in the analyses.

Generation of Klk15^G224D mouse

To model the polymorphism identified from our genetic studies, CRISPR-mediated genome editing technology in mouse embryos was utilized to generate a p. Gly224Asp (G224D) substitution in the homologous region of mouse Klk15 (NP_777354.1). The mouse KLK15 protein has two fewer amino acids than the human ortholog, making the human G226D variant equivalent to the G224D variant in the mouse. Substitution was achieved through a c.671G>A single nucleotide exchange in Klk15 (NM_174865.1) of zygotes from C57BL/6J mice. Concomitantly, this nucleotide substitution resulted in the creation of a novel Tth111I restriction site suitable for PCR-based genotyping. Synthetic guide RNA (sgRNA; atcacaggggacatcgccccngg) and single-stranded oligonucleotide (ssODN;5′caagtagctgcagacttttgtgtagacgccaggcttggtggtagtatcacaggggacatcgtcccaggagacaatgccctgcagggcacccccacagaccaagggtcctccggagtcaccctg; opposite strand showing single nucleotide exchange in bold and underlined) were procured from Integrated DNA Technologies, Inc. (IDT), which were designed and validated at the Genome Engineering and iPSC Center (GEiC), Washington University in St. Louis, St. Louis, MO; for design strategy and sequences, see Figure S4. CRISPR reagents were delivered by electroporation (EP) to single-cell embryos using 1 mm cuvettes and a Gene Pulser Xcell Eucaryotic system (Bio-Rad Laboratories, Inc.); the EP cocktails contained the following concentrations of reagents in Opti-MEM medium (Gibco/Fisher ThermoScientific): 4 μM Klk15 sgRNA, 4 μM Alt-R HiFi Cas9 Nuclease V3 (IDT),10 μM Klk15 ssODN. Electroporation conditions were as follows: square wave with 2 pulses of 30 V for 3 ms separated by 100 ms. Sequence analysis of 44 pups revealed four mice heterozygous for the G224D variant with a deletion or frameshift on the other allele, nine additional frameshift mice, and eight mice with deletions. The 4 founder Klk15^G224D/fs mice were bred to a second generation for true heterozygous (Klk15^G224D/+) pups that were used for breeding and the studies detailed below. Genotypes of the CRISPR-Cas9 Klk15 knock in mice were confirmed using the following primers and Tth111I restriction enzyme (New England Biolabs): 5′-CCCAGATCCACTCCAAGTAGC-3′ and 5′-TGAGTTCTCTTGGGTTGTGTTG-3’. All founder mice and those from F1 generations were genotyped and Sanger sequenced to confirm successful genome editing and germline transmission of the variant (Figure S4C).

Mechanical testing

Gastrocnemius muscle-Achilles tendon-calcaneus bone samples were first carefully dissected out from the mouse hind leg. Excess soft tissues surrounding the tendon were carefully cleaned off. Each sample was then scanned with μCT to determine the tendon cross-sectional area,⁴⁶ followed with a uniaxial tensile test to obtain the tendon mechanical properties.^25,26,46 Samples were stored in a moisture chamber during the sample handling. To determine the tendon cross-sectional area, gastrocnemius muscle-Achilles tendon-calcaneus bone samples were scanned with a Scanco μCT40 system (Scanco Medical, Wayne, PA, USA) at an energy setting of 55 kVP, 145 μA and an isotropic voxel resolution of 18 μm. ΜCT imaging of each sample was finished within 20 min. Tendon region in the cross-sectional image of each sample was segmented and quantified using ImageJ (version 1.52p, National Institutes of Health, Bethesda, MD). After μCT imaging, samples were immediately processed for the mechanical test. The gastrocnemius muscle fibers were carefully scraped off, leaving the tendon fibers intact. Both ends of each sample were gently glued on a thin sandpaper and gripped by custom tensile clamps attached to a mechanical testing system (Bose ELF3200, Bose, MN) with a 5 lbs. load cell. A 0.05 N tare load was first applied to each sample prior to the test. The sample initial length was determined by the clamp-clamp distance at the tare load. Samples were then preconditioned for 8 cycles between 0 and 5% strain followed with a ramp to failure at 0.01 mm/s. The force and displacement were recorded and analyzed to obtain the mechanical properties (e.g., elastic modulus) at both toe region and linear region using a bilinear fit. The transition point was defined as the point of intersection of the two linear fits of toe regions and linear region. The definition of toe region, linear region, transition point, and maximum point are shown in Figure 4. Toe modulus is the slope of the fitted line in the toe region while linear modulus (i.e., elastic modulus) is the slope of the fitted line in the linear region. Throughout the mechanical test, samples were kept hydrated with PBS drops. A total of 9 tendons from control KLK15^+/+ and 8 tendons from KLK15^G224D/+ mice were analyzed.

Transmission electron microscopy

Achilles tendons were freshly dissected from 10-week-old mice and fixed in 2% glutaraldehyde (TAAB, UK) in phosphate buffered saline (PBS) and processed for TEM. Briefly, samples were rinsed in PBS, and post-fixed in 1% osmium tetroxide (Merck, NJ) and 1.5% potassium ferrocyanide in PBS, dehydrated in a graded series of ethanol, and embedded in PolyBed 812 (Polysciences, PA) using acetonitrile as the transitional fluid. Ultrathin sections were stained with 2% uranyl acetate and Hanaichi’s lead citrate (0.15% lead nitrate, 0.15% sodium acetate, 1% sodium citrate dissolved in 41 mL water and 9 mL of 1 N sodium hydroxide (Fisher Scientific) and examined with a JEOL 1400+ electron microscope (JEOL, Japan). Consistent with prior studies on TEM analyses from tendon,^27,28,47,48 5 control, and 6 mutant mice were used for measurement of collagen fibril diameter and distribution. Micrographs (four per group) from non-overlapping regions of the tendon were taken from cross-sections. Thus, a total of 20 and 24 independent micrographs were analyzed per control and mutant tendons, respectively. Quantification of collagen fibril diameters were measured for each micrograph in entirety using the Photoshop measure tool. A total of 3,240 collagen fibrils from KLK15^+/+ and 4,191 collagen fibrils from KLK15^G224D/+ mice were analyzed blinded to genotype by two independent investigators. TEM processing and data acquisition were performed in collaboration with the University of South Carolina School of Medicine Instrumentation Resource Facility.

Echocardiography

Echocardiographic images of four-to-six-month hearts were acquired in the parasternal long axis and short axis views (40 MHz probe, Visualsonics 3100, Fujifilm, Ontario, Canada) from Klk15^+/+ (N = 5) and Klk15^G224D/+ (N = 6) mice. End-diastolic frames (coincident with the peak of the QRS complex) and end-systolic frames (smallest volume within a cardiac cycle) were used to determine chamber volumes. M-mode images from the short axis views were used to determine wall thicknesses and LV mass.

Movat’s pentachrome staining

Hearts from Klk15^+/+ (N = 4) and Klk15^G224D/+ mice (N = 4) at 4 months were fixed in zinc formalin (Sigma, Cat No: Z2902), embedded in paraffin, and sectioned at 5 μm. Deparaffinized sections were rehydrated through a graded series of ethanol’s to distilled water. Movat’s pentachrome staining was performed according to manufacturer’s manual (Poly Scientific R &D Corp, Cat No: K042). All samples were cover-slipped using Epredia cytoseal mountant (Epredia, Cat No: 83124). Images of mitral valve sections were captured throughout the regions of the mitral valve leaflets with Keyence BZ-X810 microscope (Keyence, IL). Files were transferred to Adobe Photoshop for labeling and figure preparation.

Tissue expression of lysyl oxidase and KLK15

Tail biopsies were harvested from control (N = 3) and Klk15^G224D/+ (N = 5) 3-month-old mice and solubilized in 1X RIPA followed by sonication. Samples were spun at 4°C, and soluble fraction was moved to clean tubes. Insoluble fractions and soluble lysate were solubilized in 200ul of 2X SDS-PAGE and boiled for 10 min. Equal amounts of protein were loaded onto 4–20% SDS-PAGE gels (Biorad) and run at 200V for 45 min. Gels were blotted onto nitrocellulose and Western analyses using antibodies against endogenous LOX (Cell Signaling, cat# D8F2K) and KLK15 (LSbio, cat#: C37113). All primary antibodies were used at a dilution of 1:1000, and appropriate HRP-labeled secondary antibodies were used at a concentration of 1:10000. Detection was performed using WestFemto (ThermoFisher, Inc) detection reagent and a chemidoc (Biorad, Inc).

Mouse serum collection

Blood was collected from mice and allowed to clot at room temperature for 30 min. After clotting, the samples were spun down at 2000 × g for 10 min at 4°C. The resulting supernatant was stored at −80°C.

Mouse cytokine arrays and image analysis

Mouse cytokine profiling was performed using a Proteome Profiler Mouse XL Cytokine Array Kit (R&D Systems catalog#: ARY028), which detects 111 mouse cytokines. Mouse XL Cytokine Arrays were incubated overnight at 4°C with 100 μL of mouse serum (n = 7 Klk15^+/+, 9 Klk15^G224D/+), and the procedure was performed according to the manufacturer’s instructions. Following incubation with a detection antibody cocktail, antibody conjugation, and recommended washes, the immunoblots on the membrane were developed with the chemiluminescent reagent, SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific catalog #: 34094). Unbiased quantitative analyses of the cytokine arrays were analyzed using FIJI software. The analyst was blinded to genotype for all samples. After uploading images into the software, a file of ROIs was applied to the image to identify standardized antibody spots. Automated measurements of integrated densities were recorded, and background correction was performed utilizing designated internal negative control reference spots, as recommended by the kit protocol. Samples with negative values were considered below the limit of detection and corrected to zero.

Quantification and statistical analysis

All statistical analyses were conducted using GraphPad Prism version 10.0 (GraphPad Software, Inc.). Normality was assessed with Shapiro-Wilk test, and variances were evaluated with an F-test. For datasets passing normality tests, an unpaired t-test was used. In cases of unequal variance, a Welch’s correction was applied. Data not meeting assumptions of normality were evaluated with Mann Whitney U test. A p-value < 0.05 was considered statistically significant. Graphs were generated using Prism to visualize data distribution showing all data points and the mean for each group with standard deviation. No multiple comparison corrections were applied due to the hypothesis-driven nature of all analyses. Effect size was determined with rank-biserial correlation (Mann Whitney U) and Cohen’s D with Hedges correction (Unpaired t test). The type of test used, p-value and effect size calculation is reported in figure legends.

Published: August 12, 2025