Skip to main content
NCBI home page
Biochemistry and Biophysics Reports logoLink to Biochemistry and Biophysics Reports
. 2026 Apr 24;46:102600. doi: 10.1016/j.bbrep.2026.102600

The role of HAI-1 in urothelial bladder cancer: Tissue expression, ectodomain shedding and clinical outcomes

Azad K Saeed a,b, Panagiotis Triantafyllakis c, Naheema S Gordon a, Nicholas D James d, Maurice P Zeegers e, KK Cheng f, Joep J de Jong g, Elai Davicioni h, Richard T Bryan a, Douglas G Ward a,⁎
PMCID: PMC13127185  PMID: 42064581

Abstract

Hepatocyte growth factor activator inhibitor type-1 (HAI-1) plays pivotal roles in epithelial integrity and tumour biology. Although implicated in various malignancies, its expression profile and prognostic value in bladder cancer (BC) remain incompletely defined. High levels of HAI-1 ectodomain in urine have previously been reported to be associated with poor prognosis in BC patients. This study aimed to determine the relationships between tissue and urine levels of HAI-1 and clinical outcomes in BC. This study used immunohistochemistry to measure HAI-1 expression across 770 BCs of all stages and grades. HAI-1 expression was scored on the basis of the percentage of positive cancer cells, subcellular localisation, and staining intensity. Additionally, HAI-1 (SPINT1) mRNA expression was compared with protein levels in tissue and urine. HAI-1 was highly expressed in low-grade, early-stage disease with strong membranous staining. Reduced overall HAI-1 expression, loss of membranous staining and increased cytoplasmic staining correlated with higher stage and grade and shorter survival. SPINT1 mRNA levels were positively correlated with membranous HAI-1 staining intensity (p = 0.005). Urinary levels of HAI-1 were negatively associated with the fraction of HAI-1 positive cancer cells and membranous staining intensity (p < 0.05). A positive correlation was observed between SPINT1 expression and urinary HAI-1 levels (p < 0.05). The Urobasal A subtype had lower urinary HAI-1 ectodomain levels than other subtypes. HAI-1 expression may serve as a biomarker of tumour differentiation and prognosis in BC. Increased ectodomain shedding into the urine, rather than increased expression, likely explains the higher urine HAI-1 levels seen in more aggressive tumours.

Keywords: HAI-1, Immunohistochemistry, Bladder cancer, Urine, SPINT1, Biomarker

Highlights

  • •

    Low-grade, early-stage bladder tumours express high levels of membranous HAI-1.

  • •

    Loss of membranous HAI-1 and increased cytoplasmic localisation are associated with higher stage, grade, and shorter survival.

  • •

    Urinary HAI-1 levels are inversely related to tumour HAI-1 expression, suggesting enhanced shedding in aggressive disease.

1. Introduction

Bladder cancer (BC) ranks among the top ten most commonly diagnosed malignancies worldwide [1]. The predominant histological form of BC is urothelial carcinoma. Approximately three-quarters of newly diagnosed BCs are classified as non-muscle-invasive bladder cancer (NMIBC) which, whilst not immediately life-threatening, has a high recurrence rate and an ongoing risk of progression [2]. The management of NMIBC requires appropriate intravesical treatment and long-term surveillance [3]. In contrast, muscle-invasive bladder cancer (MIBC) represents one quarter of new cases and requires radical therapeutic approaches such as cystectomy, systemic chemotherapy, radiotherapy, and immunotherapy [4,5]. Despite advances in treatment, there remains an urgent need for better prognostic and predictive biomarkers [6].

Hepatocyte growth factor activator inhibitor-1 (HAI-1) is a transmembrane protein containing two Kunitz-type serine protease inhibitory domains [7]. It plays a regulatory role in modulating the activity of key proteases, including matriptase and hepatocyte growth factor activator [8]. By inhibiting these enzymes, HAI-1 indirectly influences the activation of hepatocyte growth factor (HGF), a pivotal mediator of epithelial cell proliferation and motility [9]. Upon activation, HGF binds to its receptor, c-MET, initiating a cascade of intracellular signalling pathways that govern cellular proliferation, migration, survival, and invasion. HAI-1 is expressed in both normal [10] and malignant epithelia [11].

The biological significance of HAI-1 has been investigated in several malignancies, including prostate [12], pancreas [13], colon [14], liver [15], oesophagus [16], cervix [17], ovary [18], endometrium [19], and breast [20]. These studies have predominantly focused on evaluating HAI-1 expression in relation to tumour grade, stage, and patient survival. In BC only one study, based on a small patient cohort (n = 26), has examined the tissue expression of HAI-1 [21]. Thus, the distribution of HAI-1 across BC stages, grades and subtypes and its potential as a prognostic biomarker remain poorly defined.

HAI-1 functions not only as an inhibitor, but also as a key coordinator of protease maturation, intracellular trafficking, and site-specific activation. HAI-1 associates with matriptase within the secretory pathway, where it restrains premature enzymatic activity and enables proper delivery of the complex to the plasma membrane. When this regulatory relationship is perturbed, matriptase becomes mislocalised or aberrantly active [[22], [23], [24]]. At the cell surface, HAI-1 undergoes controlled ectodomain shedding, a process commonly linked to matriptase activation. Experimental data indicate that hormonal or microenvironmental stimuli can synchronise activation of matriptase with shedding of the HAI-1–matriptase complex, generating soluble HAI-1 species in the extracellular space [[25], [26], [27], [28]]. Although this shedding likely involves proteolytic cleavage within the extracellular region of HAI-1, the enzymes responsible and the regulatory mechanisms remain only partly characterised and seem to be highly context-dependent [28].

We previously reported elevated urinary HAI-1 as a poor prognostic indicator for high-risk NMIBC (HR-NMIBC) and MIBC [29]. We now explore the relationship between HAI-1 expression in tumour tissue and its urinary concentration to gain further insight into HAI-1's biological and clinical relevance in BC.

2. Materials and methods

2.1. Patient samples

We used 23 tissue microarrays (TMAs), containing duplicate tumour cores from 770 participants in the Bladder Cancer Prognosis Programme (BCPP) [30]. Patient recruitment occurred between 2005 and 2011 across ten hospitals in the West Midlands, UK, and participants gave written informed consent for the use of their biospecimens in BC biomedical research (ethics reference 06/MRE04/65). Tumour specimens were collected via transurethral resection prior to any therapeutic intervention. Pathological classification followed the TNM staging system and the 1973 World Health Organisation grading criteria, reflecting standard UK practice during the study period. All samples underwent immunohistochemical analysis for HAI-1 expression.

2.2. Immunostaining

Sections were dewaxed overnight at 56 °C, followed by sequential immersion in xylene and graded ethanol solutions for rehydration. After washing in distilled water and phosphate-buffered saline (PBS), antigen retrieval was achieved by heating the slides in citrate buffer (pH 6.0) using a microwave for 20 min at high temperature. Slides were then cooled to room temperature, washed in PBS and incubated in 3% hydrogen peroxide. Following washing in PBS containing 0.01% Tween® 20, tissue sections were circumscribed with a hydrophobic barrier pen. Sections were blocked with 20% normal donkey serum (Merck Life Science, UK) for 1 hr at room temperature and incubated with primary antibody against HAI-1 (polyclonal goat anti-human HAI-1 catalogue number: AF1048, Bio-techne, USA at 1:500 in PBS) overnight at 4 °C. Following three PBS washes, sections were incubated with donkey anti-goat IgG secondary antibody (VisUCyte HRP Polymer, catalogue number: VC004, Bio-techne, USA) for 30 min at room temperature, washed and visualised using DAB chromogen (Vector Laboratories, USA). Counterstaining was performed with Mayer's haematoxylin solution (Sigma Aldrich, USA). Antibody validation data is shown in Supplemental Fig. 1.

2.3. Pathological interpretation

Slides were scanned using a Zeiss Axio Slide Scanner, and two independent pathologists evaluated the tumour cores at 40 × magnification using ZEN 3.1 (blue edition) software, with consensus reached through joint review in discordant cases. To validate immunohistochemical scoring, DAB and haematoxylin signals were separated using the Deconvolution 2 plugin in ImageJ, ensuring accurate interpretation of staining patterns.

2.4. Immunohistochemical scoring

HAI-1 expression was scored using three criteria: the percentage of positively stained cancer cells, staining intensity, and subcellular localisation. The percentage of HAI-1 positive cells was scored: 0 (no detectable staining), 1 (≥1–<25%), 2 (≥25–<50%), 3 (≥50–<75%), and 4 (≥75%). These scores were also combined into 1-2 = low and 3-4 = high. Staining intensity for membranous and cytoplasmic staining was independently scored on a four-point scale, ranging from 0 (absent) to 3 (strong), with 0-1 considered weak staining and 2-3 intense staining. Subcellular localisation of HAI-1 was determined by identifying the major site of immunoreactivity within the cancer cells. Each case was categorised as cytoplasmic (score 1), membranous (score 2), or nuclear (score 3), based on the dominant staining pattern observed.

2.5. Urine HAI-1 measurement in BC patients

Urinary HAI-1 data for the BCPP cohort was previously reported (Snell et al., 2018). Briefly, quantification was performed by ELISA (Human HAI-1 kit, catalogue number: DY1048, Bio-techne, USA) using urine supernatants collected from 858 BCPP BC patients [29]. Urine samples were collected before patients underwent transurethral resection of bladder tumour(s) [30]. Urine samples were kept on ice and centrifuged at 2000 rpm for 10 min within 8 h of collection. Pellets and supernatants were stored separately at −80 °C [29]. All urinary HAI-1 levels are reported as ng urinary HAI-1 per mg of urinary creatinine and a threshold of 2 ng/mg creatinine was used to define low/high urine HAI-1.

2.6. Gene expression data

Microarray-based gene expression profiling data was generated for 455 of the tumours with paired tissue and urine HAI-1 data. From each block, 4 μm sections were prepared for histological review. A pathologist identified tumour-rich regions, and macro-dissection of adjacent unstained sections was performed for RNA extraction. Genome-wide gene expression was analysed using the Decipher Bladder Genomic Subtyping Classifier (GSC) oligonucleotide microarray assay in a CAP/CLIA laboratory (Veracyte, Inc San Diego, CA).

2.7. Statistical analysis

Statistical tests were carried out using GraphPad Prism (v10.5.0). One-way ANOVA was applied to assess differences across multiple variables. The relationship between two immunohistochemical scoring parameters was examined via the Mann–Whitney U test, and both Spearman and Pearson correlations were employed to explore associations between HAI-1 tissue scores and urinary HAI-1 and SPINT1 expression levels. Simple linear regression was used to evaluate the predictive capacity of tissue HAI-1 and SPINT1 expression for urinary HAI-1 concentrations. Prognostic implications of HAI-1 expression in tissue were assessed through univariate Kaplan–Meier survival analysis and log-rank testing, focusing on recurrence, progression and disease-specific survival. Statistical significance was defined at p < 0.05, with median values presented to reflect data distribution.

3. Results

3.1. Proportion of BC cells expressing HAI-1

HAI-1 immunostaining revealed a clear decrease in the proportion of positive cells with increasing stage and grade (Fig. 1a). NMIBCs most commonly scored 3 (49% of cases vs. 22% of MIBCs), whilst MIBCs most commonly scored 2 (54% of cases vs. 23% of NMIBCs, Supplementary Fig. 2a).

Fig. 1.

Fig. 1

Open in a new tab

HAI-1 expression across BC grades and stages. (a) HAI-1 positivity; (b) subcellular localisation of HAI-1; (c) membranous intensity; and (d) cytoplasmic intensity. Numbers of cases: pTa Grade 1 (n = 115), pTa Grade 2 (n = 192), pTa Grade 3 (n = 51), pT1 Grade 2 (n = 59), pT1 Grade 3 (n = 169), and pT2 Grade 3 (n = 184).

There was a progressive decline in positivity with increasing EAU NMIBC risk group: Score 4 was observed in 76% of low-risk cases, 33% of intermediate-risk cases, and 13% of high-risk cases, with scores 3 and 2 becoming more prevalent in higher-risk categories (Supplementary Fig. 3a).

3.2. Site of expression

Nuclear expression of HAI-1 expression was not observed in any BC. Low grade tumours predominantly exhibited expression in the plasma membrane. There was a clear shift from membranous to cytoplasmic staining with increasing grade, stage, and EAU NIMBC risk group (Fig. 1b and 2a and b). Overall, 58% of NMIBCs exhibited membranous expression, whilst 88% of MIBC samples were predominantly cytoplasmic (Supplementary Fig. 2b).

Fig. 2.

Fig. 2

Open in a new tab

HAI-1 localisation in BC. Representative sections showing HAI-1 localisation: (a) membranous and (b) cytoplasmic staining. (c) A urothelial section exhibiting dominant membranous HAI-1 expression throughout basal and intermediate layers, with umbrella cells showing pronounced cytoplasmic staining. (d) BC cells demonstrating disrupted adhesion and a shift toward enhanced cytoplasmic HAI-1 localisation.

Within NMIBC risk groups, membranous staining declined from 88% in low-risk to 40% in high-risk cases, with a corresponding rise in cytoplasmic expression (Supplementary Fig. 3b). We also observed that umbrella cells exhibited weak membranous and strong cytoplasmic HAI-1 staining, even in sections where other layers exhibited strong membranous expression, suggesting possible shedding of membranous HAI-1 into the urine. In contrast, high-grade tumours with disrupted cell cohesion predominantly displayed cytoplasmic localisation, indicating a shift in HAI-1 distribution associated with tumour progression and loss of epithelial integrity (Fig. 2c and d).

3.3. Alterations in membranous and cytoplasmic HAI-1 intensities

HAI-1 membranous staining intensity had an inverse relationship with tumour grade, stage, and EAU NMIBC risk group. Strong membranous expression (Score 3) was predominantly associated with low-grade tumours, notably present in 57% of grade 1 pTa grade 1 cases vs. 24% of grade 3 pTa cases (Fig. 1c). NMIBCs most frequently displayed moderate staining (38%), while MIBCs were dominated by weak (53%) or absent (29%) expression (Supplementary Fig. 2c). Strong membranous staining was prevalent in low-risk NMIBCs (52%) and progressively reduced in intermediate and high-risk NMIBCs (Supplementary Fig. 3c).

Cytoplasmic HAI-1 intensity demonstrated a distinct positive correlation with tumour aggressiveness. Weak cytoplasmic staining (Score 1) was common in low-grade and low-risk NMIBC (e.g., 55% in grade 1 pTa), whereas higher-grade tumours showed increased intensity, with 52% of pT2 grade 3 samples exhibiting moderate staining (Fig. 1d). MIBC cases had high cytoplasmic expression, with 66% showing moderate or strong staining (Supplementary Fig. 2d). Cytoplasmic intensity increased from low to high risk NMIBC, peaking in high-risk cases (56% moderate, 12% strong, Supplementary Fig. 3d).

3.4. HAI-1 expression across LUND molecular subtypes

HAI-1 expression varied across the LUND subtypes, paralleling patterns observed with tumour grade, stage, and risk group (Supplementary Fig. 4). The Urobasal A subtype exhibited the highest proportion of HAI-1–positive cells of any of the subtypes, with predominantly membranous localisation and high membranous staining intensity. The squamous cell carcinoma-like (SCCL) subtype exhibited the lowest proportion of positive cells and the lowest membrane staining.

3.5. Relationship between tissue HAI-1 expression and urine HAI-1 concentration

In the overall cohort, a significant inverse correlation was observed between the proportion of HAI-1–positive cells in tissue and urinary HAI-1 concentration (Spearman's ρ = −0.26, p < 0.0001; Supplementary Fig. 5a). Membranous staining intensity was also negatively correlated (ρ = −0.20, p < 0.0001; Supplementary Fig. 5c).

In high-risk NMIBC, the percentage of positive cells and cytoplasmic intensity scores negatively correlated with urine HAI-1 levels (p < 0.05; Supplementary Fig. 5e–h). None of the tissue-based HAI-1 metrics were significantly associated with urinary concentrations in MIBC (p > 0.05; Supplementary Fig. 54i–l).

To further explore the relationship between tissue HAI-1 expression and urinary HAI-1, each tissue parameter was dichotomised into low and high expression categories, and urine HAI-1 concentrations were compared between the categories. Across all patients (Fig. 3a), the proportion of HAI-1–positive cells showed a significant inverse association with urinary levels: low tissue expression (scores 1–2) corresponded to a higher median urinary concentration (1.69) compared with high tissue expression (scores 3–4; median 1.09; p < 0.0001). There was no significant difference in urinary HAI-1 between patients with predominantly membranous or cytoplasmic expression (p = 0.1278; Fig. 3b). In contrast, membranous intensity displayed a negative trend: weak staining (scores 0–1) was linked to elevated urinary HAI-1 (median 1.50), whereas strong staining (scores 2–3) corresponded to lower levels (median 1.09; p < 0.0001; Fig. 3c). The median urinary HAI-1 concentrations in patients with strong and weak cytoplasmic staining were 1.37 and 1.13, respectively (p > 0.05, Fig. 3d).

Fig. 3.

Fig. 3

Open in a new tab

Urinary HAI-1 levels compared across tissue HAI-1 expression categories. Panels (a–d) show Mann–Whitney U test results for all patients (n = 588) stratified by (a) percentage positivity, (b) localisation site, (c) membranous intensity, (d) cytoplasmic intensity, (e) urine levels comparison for NMIBC risk groups. Numbers of cases: Low-risk = 48, intermediate-risk = 151, high-risk = 230 and MIBC = 139.

In high-risk NMIBC, only the percentage of positive cells retained significant differences, with higher tissue HAI-1 associated with lower urinary HAI-1 levels (Supplementary Fig. 6a–d). In MIBC, none of the tissue-based parameters demonstrated significant associations with urinary HAI-1 (Supplementary Fig. 6e–h). Additionally, urine levels were compared according to NMIBC risk groups and MIBC. HR-NMIBC and MIBC patients exhibited significantly elevated urinary HAI-1 concentrations compared with patients in the low- and intermediate-risk groups (p < 0.05, Fig. 3e).

3.6. Prognostic value of tissue HAI-1

The associations between HAI-1 expression and disease specific survival (DSS) for the whole cohort are shown in Fig. 4. A low percentage of cancer cells expressing HAI-1-positive was significantly associated with shorter DSS (p < 0.0001, Fig. 4a). Membranous expression of HAI-1 was associated with longer DSS than cytoplasmic expression (p < 0.0001, Fig. 4b); membranous intensity scores further supported this trend, with stronger staining being associated with longer DSS (p < 0.0001, Fig. 4c) and higher cytoplasmic intensity being associated with shorter DSS (p < 0.01, Fig. 4d). These findings demonstrate the prognostic utility of HAI-1 tissue expression in BC. However, the relationship between HAI-1 expression and outcomes is most likely driven by the strong correlation of the IHC scores with stage and grade. When patients were stratified into NMIBC risk groups and MIBC, no significant associations were found between DSS and tissue HAI-1 expression. Supplementary Figs. 7–8 show overall survival (OS) and recurrence free survival (RFS) in HR-NMIBC.

Fig. 4.

Fig. 4

Open in a new tab

Tissue HAI-1 expression parameters and disease-specific survival in the all-patient cohort (n = 770). (a–d) Kaplan–Meier plots showing DSS stratified by: (a) percentage of HAI-1-positive cells, (b) subcellular localisation, (c) membranous staining intensity and (d) cytoplasmic staining intensity.

3.7. Correlation of SPINT1 gene expression, HAI-1 protein expression and urine HAI-1 ectodomain concentration

Median SPINT 1 levels were 0.24, 0.32, 0.36 and 0.26 in low, intermediate and high risk NMIBC and MIBC respectively (Fig. 5a). There was a positive correlation between SPINT1 expression and urinary HAI-1 concentrations across the whole patient cohort (r = 0.14, p = 0.0006; Supplementary Fig. 9a) and in high-risk NMIBC (r = 0.16, p = 0.02; Supplementary Fig. 9b) and MIBC (r = 0.29, p = 0.001; Supplementary Fig. 9c). The median urinary HAI-1 concentrations in all patients with low and high SPINT1 expression were 1.03 and 1.17 ng/mg creatinine (p = 0.005, Fig. 5b). Thus, although statistically significant, the effect of SPINT1 expression on urinary HAI-1 shedding is relatively modest.

Fig. 5.

Fig. 5

Open in a new tab

Relationship between SPINT1 expression, NMIBC risk group and urinary HAI-1 concentrations. (a) Comparison of SPINT1 transcript levels across NMIBC risk groups (n = 485, low-risk = 85, intermediate-risk = 185, and high-risk = 215) and MIBC (n = 124). (b) Urinary HAI-1 concentrations compared between all patients (n = 629) with low/high SPINT1 expression (median = 0.31).

HAI-1 membranous intensity demonstrated a weak positive correlation with tissue SPINT1 mRNA levels across the entire cohort (Supplementary Fig. 9a–d). There were no significant correlations in HR-NMIBC (Supplementary Fig. 9e–h)but in MIBC both the percentage of positive cells and membranous intensity scores were positively correlated with SPINT1 expression (Supplementary Fig. 9i–l).

3.8. Urine levels of HAI-1 and LUND subtypes

LUND subtyping was applied to the whole patient cohort, NMIBC risk groups and MIBC (Fig. 6a–c). In NMIBC, the lowest urine levels (median = 0.89) were recorded in Urobasal class A, whereas the highest amounts were found in SCCL (median = 1.98) and infiltrating (median = 1.94, Fig. 6b) classes (p > 0.05). The genomically unstable class in MIBC demonstrated the lowest urinary levels, with a median of 2.2 (p > 0.05, Fig. 6c). Urobasal class A had the lowest urinary levels in NMIBC and HR-NMIBC but there was no correlation between subtype and urine HAI-1 in MIBC (Fig. 6d–g).

Fig. 6.

Fig. 6

Open in a new tab

Urinary HAI-1 concentrations in relation to Lund molecular subtypes and clinical risk stratification. Panels illustrate: (a) distribution across all patients (n = 629); (b) NMIBC (n = 485); (c) MIBC (n = 124); (d) high-risk NMIBC (n = 215); (e) intermediate-risk NMIBC (n = 185); (f) low-risk NMIBC (n = 85); and (g) differential urine levels within the Urobasal A subtype across NMIBC risk groups (low, intermediate, and high).

4. Discussion

We have investigated the expression of HAI-1 across BC grades, stages and risk groups by IHC using scoring systems for the percentage of positive cells, the site of expression, membranous versus cytoplasmic expression, and membranous and cytoplasmic intensities. We observed a reduction in the percentage of positive cells with increasing grade, stage and risk group. In addition, strong membranous staining is seen in low-grade, early-stage tumours, but predominantly cytoplasmic localisation is seen in advanced disease.

Given a prior lack of consensus about the subcellular localisation of HAI-1 during cancer development, ambiguity has limited the interpretation of HAI-1's role in BC progression. With previous studies emphasising membranous staining [21], excluding cytoplasmic expression [12,15], or failing to define localisation criteria [[17], [18], [19]], the present study addresses these gaps by systematically analysing HAI-1 expression site and intensity across BC grades and stages.

Notably, 58% of NMIBC cases showed membranous staining, compared to just 12% of MIBC. These findings align with recent oesophageal carcinoma data, indicating that diminished membranous expression correlates with poorly differentiated cancers and an unfavourable prognosis [16]. The observed change in the site of expression may reflect changes in protein trafficking to the membrane and/or shedding of the extracellular domain from the cell surface by metalloproteases [13,25]. Our observations of cytosolic HAI-1 further emphasise the role of subcellular trafficking in dictating its function. Under normal conditions, HAI-1 moves in a regulated manner from the endoplasmic reticulum through the Golgi to the plasma membrane, typically alongside matriptase. Disruption of this pathway, whether through altered vesicle transport, impaired membrane anchoring, or increased internalisation, can lead to intracellular accumulation of HAI-1 [23]. Similar localisation shifts have been reported for other proteins, including Syndecan-1, where a transition from membranous to cytoplasmic distribution has been associated with unfavourable clinical outcomes in breast cancer [31]. Such redistribution may signal a failure to position HAI-1 at sites where protease activation must be controlled, thereby enabling excessive surface proteolysis and perturbing intracellular signalling. Indeed, mislocalised HAI-1 has been associated with deregulated matriptase activity, compromised epithelial integrity, and enhanced invasiveness [32]. In malignant cells, these trafficking abnormalities may be intensified by oncogenic pathways or altered membrane dynamics, effectively separating HAI-1 from its protease targets. Consequently, cytosolic HAI-1 may not simply be a by-product of dysregulation but an indicator of impaired protease control and a potential contributor to tumour cell adaptability and progression [16,33].

One previous study by Yamasaki et al. examined HAI-1 by IHC in 26 BC patients [21]. They reported that BCs with low HAI-1 expression and high matriptase expression are more aggressive. This is biologically plausible, as a low HAI-1: matriptase ratio should result in increased pericellular processing of molecules such as pro-HGF and PAR-2 and is consistent with studies in mice where HAI-1 knockout promotes cancer progression [34].

We found that the relatively indolent low-grade BCs have high plasma membrane levels of HAI-1, which likely effectively inhibits HAI-1's cognate proteases. In contrast, in aggressive high-grade BCs there is less membranous HAI-1 and most likely loss of normal regulation of pericellular serine protease activity. Thus, it is not surprising that membranous HAI-1 positivity is associated with better outcomes in our patient cohort comprised of all stages and grades of BC. However, this loss of expression and switch from membranous to cytoplasmic staining in high-grade disease is so ubiquitous that, in contrast to urine HAI-1 levels, the tissue data is not prognostic independently of stage and grade. Assessing levels of HAI-1-cognate protease complexes rather than total HAI-1 might offer greater prognostic insight, as it more directly reflects the protease–inhibitor balance that underpins tumour progression.

Urine levels of HAI-1 (shed ectodomain) were weakly positively correlated with SPINT1 mRNA expression but negatively correlated with tissue HAI-1 IHC scores. Additionally, SPINT1 expression was positively correlated with the membranous intensity scores. This suggests that the additional prognostic value of urine HAI-1 is because it acts as a surrogate of sheddase activity. Previous studies demonstrated that MMP7 can cleave the extracellular domain of HAI-1 in colon cancer cells [26], and MMP14 exerts a similar effect when co-expressed with HAI-1 in HEK293 cells [25]. Nonetheless, in the context of bladder carcinogenesis, the specific protease responsible for HAI-1 shedding into urine remains undescribed. We also showed that umbrella cells exhibited stronger cytoplasmic staining than intermediate cell layers (Fig. 2), suggesting that cytoplasmic localisation may be associated with urinary shedding.

To minimise confounding effects of tumour grade and stage, we stratified NMIBC and MIBC samples according to the LUND classification [35]. This taxonomy operates independently of pathological stage, enabling its application across all forms of urothelial carcinoma [36]. This approach revealed that Urobasal A tumours (associated with good outcomes) exhibited the lowest urinary HAI-1 levels in NMIBC [35]. Notably, when comparing Urobasal A subtypes across NMIBC risk categories, patients with HR-NMIBC demonstrated significantly higher urinary HAI-1 concentrations. These results further support the hypothesis that urinary HAI-1 levels correlate with tumour aggressiveness, underscoring its potential utility in clinical risk stratification.

The literature regarding SPINT1 expression during cancer progression remains inconclusive. While several studies have documented a negative association of HAI-1 expression in the development of malignancies such as prostate [37], ovarian [38], and endometrial cancers [19], others have demonstrated its upregulation in hepatocellular carcinoma [39] and thyroid cancer [40]. Our data indicate that SPINT1 expression is elevated in HR-NMIBC compared to other NMIBC risk groups, while a reduction in SPINT1 expression was observed in MIBC cases. This contrasts with findings by Isali et al. (2022), who reported SPINT1 upregulation in MIBC in comparison to control samples [41].

Despite the inclusion of 770 BC patients and the use of multiple HAI-1 scoring parameters, several limitations affect this study: 1) In several BC subgroups, particularly low- and intermediate-risk NMIBC, certain HAI-1 scores were under-represented or absent, restricting meaningful comparisons; 2) the number of clinical events within some subgroups was not sufficient to support robust survival analyses; 3) the categorical nature of immunohistochemical scoring poses analytical challenges.

5. Conclusion

In low-grade, early-stage, low-risk NMIBC and Urobasal A class, HAI-1 was highly expressed by most cancer cells, localised to the membrane with weak cytoplasmic staining, correlating with favourable outcomes. In higher grade and stage BCs, a shift toward a reduced percentage of positive cells, membranous expression and increased cytoplasmic localisation was observed. Tissue HAI-1 levels are prognostic in BC but not independent of stage and grade. The relationships between SPINT1 mRNA, tissue HAI-1 and urine HAI-1 suggest that the rate of ectodomain shedding determines urine levels rather than HAI-1 expression per se, and potentially indicates a role for sheddases in bladder carcinogenesis that warrants further investigation. Collectively, these findings highlight the complex regulation of HAI-1 in BC and its potential utility as a prognostic biomarker.

Ethics approval statement

This research has been reviewed and approved by an ethics committee ethics (reference [42]/MRE04/65); participants gave written informed consent for the use of their biospecimens in BC biomedical research.

Funding statement

This work was supported by a PhD studentship from the Higher Committee for Education Development (HCED). The funder had no role in study design, data collection, data analysis, manuscript preparation, or the decision to publish.

CRediT authorship contribution statement

Azad K. Saeed: Conceptualization, Data curation, Methodology, Writing – original draft. Panagiotis Triantafyllakis: Methodology. Naheema S. Gordon: Methodology. Nicholas D. James: Conceptualization. Maurice P. Zeegers: Conceptualization. K.K. Cheng: Conceptualization. Joep J. de Jong: Data curation. Elai Davicioni: Conceptualization. Richard T. Bryan: Supervision, Writing – review & editing. Douglas G. Ward: Supervision, Writing – review & editing.

Declaration of competing interest

Elai Davicioni is an employee of Veracyte Inc. Joep de Jong is a consultant for Veracyte Inc. The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This research was funded by the Higher Committee for Education Development (HCED) and the University of Sulaimani. Their steadfast commitment and continued encouragement have played a pivotal role in the progression of this work, and their support is acknowledged with sincere gratitude.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.bbrep.2026.102600.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

graphic file with name mmcfigs1.jpg

Open in a new tab

Supplementary Figure 1: Three control conditions were employed: no antibody control, absorption control using recombinant HAI-1 protein, and a complete staining protocol. a: Sections treated with no control and recombinant protein with HAI-1 antibody (b) exhibited minimal staining, confirming antibody specificity, whereas the positive control revealed robust HAI-1 expression HAI-1 expression in BC samples (c).

graphic file with name mmcfigs2.jpg

Open in a new tab

Supplementary Figure 2: Proportion of HAI-1 positive cells, HAI-1 localisation and scores of membranous and cytoplasmic intensities in NMIBC and MIBC. (a) distribution of HAI-1 positivity scores; (b) comparative shifts in HAI-1 localisation; (c) scores distribution of membranous intensity scores; (d) scores distribution of cytoplasmic intensity scores. Tumour subtypes included NMIBC (n = 586) and MIBC (n = 184).

graphic file with name mmcfigs3.jpg

Open in a new tab

Supplementary Figure 3: Proportion of HAI-1 positive cells, HAI-1 localisation and scores of membranous and cytoplasmic intensities across NMIBC risk groups. (a) distribution of HAI-1 positivity scores; (b) alterations in HAI-1 subcellular localisation; (c) scores distribution of membranous intensity; (d) scores distribution of cytoplasmic intensity scores. NMIBC risk groups (n = 558) included low-risk NMIBC (n = 58), intermediate-risk NMIBC (n = 195), and high-risk NMIBC (n = 305).

graphic file with name mmcfigs4.jpg

Open in a new tab

Supplementary Figure 4: Proportion of HAI-1 positive cells, HAI-1 localisation and scores of membranous and cytoplasmic intensities in LUND molecular subtypes (all cases, n = 599). (a) distribution of HAI-1 positivity scores; (b) alterations in HAI-1 subcellular localisation; (c) scores distribution of membranous intensity scores; (d) scores distribution of cytoplasmic intensity scores.

graphic file with name mmcfigs5.jpg

Open in a new tab

Supplementary Figure 5: Spearman correlation between tissue HAI-1 parameters and urinary HAI-1 levels. (a–d) All cases (n = 588): (a) percentage of HAI-1-positive cells, (b) subcellular localisation, (c) membranous intensity, and (d) cytoplasmic intensity. (e-h) High-risk NMIBC: same parameters as above (n = 230). (i-l) MIBC: same parameters as above (n = 139).

graphic file with name mmcfigs6.jpg

Open in a new tab

Supplementary Figure 6: Urinary HAI-1 levels compared across tissue HAI-1 expression categories. Panels (a–d) show Mann–Whitney U test results for HR-NMIBC (n = 230) stratified by (a) percentage positivity, (b) localisation site, (c) membranous intensity, and (d) cytoplasmic intensity. Panels (e-h) present the same comparisons for MIBC (n = 139).

graphic file with name mmcfigs7.jpg

Open in a new tab

Supplementary Figure 7: Univariate survival analysis of tissue HAI-1 expression parameters across the all-patient cohort (n = 770). (a–d) Kaplan–Meier plots showing overall survival stratified by: (a) percentage of HAI-1-positive cells, (b) subcellular localisation, (c) membranous staining intensity, and (d) cytoplasmic staining intensity.

graphic file with name mmcfigs8.jpg

Open in a new tab

Supplementary Figure 8: Univariate RFS analysis of tissue HAI-1 expression parameters in high-risk NMIBC patients (n = 305). (a–d) Kaplan–Meier plots illustrating RFS stratified by (a) percentage of HAI-1-positive cells, (b) subcellular localisation, (b) membranous intensity, and (d) cytoplasmic intensity.

graphic file with name mmcfigs9.jpg

Open in a new tab

Supplementary Figure 9: Pearson correlation analyses illustrating associations between urinary HAI-1 levels and SPINT1 expression in the entire cohort (a, n = 629), high-risk NMIBC subgroup (b, n = 215), and MIBC subgroup (c, n = 124).

graphic file with name mmcfigs10.jpg

Open in a new tab

Supplementary Figure 10: Spearman correlation between tissue HAI-1 parameters and SPINT1 expression levels. (a–d) All cases (n = 599): (a) percentage of HAI-1-positive cells, (b) subcellular localisation, (c) membranous intensity, and (d) cytoplasmic intensity. (e-h) High-risk NMIBC: same parameters as above (n = 225). (i-l) MIBC: same parameters as above (n = 136).

Data availability

Data will be made available on request.

References

  • 1.Zhang Y., et al. The global landscape of bladder cancer incidence and mortality in 2020 and projections to 2040. J. Glob. Health. 2023;13 doi: 10.7189/jogh.13.04109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Godlewski D., et al. Bladder cancer basic study and current clinical trials. Uro. 2024;4(3):145–196. [Google Scholar]
  • 3.Bedke J., et al. Urologic Oncology: Seminars and Original Investigations. Elsevier; 2023. Optimizing outcomes for high-risk, non-muscle-invasive bladder cancer: the evolving role of PD-(L) 1 inhibition. [DOI] [PubMed] [Google Scholar]
  • 4.van der Heijden A.G., Witjes J.A. Recurrence, progression, and follow-up in non–muscle-invasive bladder cancer. Eur. Urol. Suppl. 2009;8(7):556–562. [Google Scholar]
  • 5.Jubber I., et al. Epidemiology of bladder cancer in 2023: a systematic review of risk factors. Eur. Urol. 2023;84:176–190. doi: 10.1016/j.eururo.2023.03.029. [DOI] [PubMed] [Google Scholar]
  • 6.Wong V.K., et al. Imaging and management of bladder cancer. Cancers. 2021;13(6):1396. doi: 10.3390/cancers13061396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Hong Z., et al. Crystal structure of a two-domain fragment of hepatocyte growth factor activator inhibitor-1: functional interactions between the Kunitz-type inhibitor domain-1 and the neighboring polycystic kidney disease-like domain. J. Biol. Chem. 2016;291(27):14340–14355. doi: 10.1074/jbc.M115.707240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Wu S.-R., et al. The Kunitz domain I of hepatocyte growth factor activator inhibitor-2 inhibits matriptase activity and invasive ability of human prostate cancer cells. Sci. Rep. 2017;7(1) doi: 10.1038/s41598-017-15415-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Fukushima T., et al. Hepatocyte growth factor activator: a proteinase linking tissue injury with repair. Int. J. Mol. Sci. 2018;19(11):3435. doi: 10.3390/ijms19113435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Szymańska B., et al. Usability evaluation of urinary HAI-1, STMN-1 and TN-C in the diagnosis of bladder cancer. J. Clin. Med. 2025;14(11):3664. doi: 10.3390/jcm14113664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Oberst M., et al. Matriptase and HAI-1 are expressed by normal and malignant epithelial cells in vitro and in vivo. Am. J. Pathol. 2001;158(4):1301–1311. doi: 10.1016/S0002-9440(10)64081-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Yasuda K., et al. Expression of Hepatocyte growth factor activator inhibitor type-1 (HAI-1) in prostate cancer. Anticancer Res. 2013;33(2):575–581. [PubMed] [Google Scholar]
  • 13.Sakugawa C., et al. Prognostic significance of hepatocyte growth factor activator inhibitor type 1 (HAI-1) immunoreactivity in pancreatic ductal adenocarcinoma. BMC Res. Notes. 2017;10(1):1–7. doi: 10.1186/s13104-017-3014-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Ye J., et al. Down-regulation of HAI-1 is associated with poor-differentiation status of colorectal cancer. Hum. Cell. 2013;26:162–169. doi: 10.1007/s13577-013-0074-7. [DOI] [PubMed] [Google Scholar]
  • 15.Funagayama M., et al. Expression of hepatocyte growth factor activator inhibitor type 1 in human hepatocellular carcinoma and postoperative outcomes. World J. Surg. 2010;34:1563–1571. doi: 10.1007/s00268-010-0517-1. [DOI] [PubMed] [Google Scholar]
  • 16.Umekita Y., et al. Loss of tumor cell surface hepatocyte growth factor activator inhibitor-1 predicts worse prognosis in esophageal squamous cell carcinoma. Pathol. Res. Pract. 2025;266 doi: 10.1016/j.prp.2025.155809. [DOI] [PubMed] [Google Scholar]
  • 17.Nakamura K., et al. The role of hepatocyte growth factor activator inhibitor-1 (HAI-1) as a prognostic indicator in cervical cancer. Int. J. Oncol. 2009;35(2):239–248. [PubMed] [Google Scholar]
  • 18.Nakamura K., et al. Expression of hepatocyte growth factor activator inhibitors (HAI-1 and HAI-2) in ovarian cancer. Int. J. Oncol. 2009;34(2):345–353. [PubMed] [Google Scholar]
  • 19.Nakamura K., et al. The role of hepatocyte growth factor activator inhibitor (HAI)‐1 and HAI‐2 in endometrial cancer. Int. J. Cancer. 2011;128(11):2613–2624. doi: 10.1002/ijc.25606. [DOI] [PubMed] [Google Scholar]
  • 20.Parr C., et al. The hepatocyte growth factor regulatory factors in human breast cancer. Clin. Cancer Res. 2004;10(1):202–211. doi: 10.1158/1078-0432.ccr-0553-3. [DOI] [PubMed] [Google Scholar]
  • 21.Yamasaki K., et al. Matriptase-induced phosphorylation of MET is significantly associated with poor prognosis in invasive bladder cancer; an immunohistochemical analysis. Int. J. Mol. Sci. 2018;19(12):3708. doi: 10.3390/ijms19123708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Oberst M.D., et al. HAI-1 regulates activation and expression of matriptase, a membrane-bound serine protease. Am. J. Physiol. Cell Physiol. 2005;289(2):C462–C470. doi: 10.1152/ajpcell.00076.2005. [DOI] [PubMed] [Google Scholar]
  • 23.Chiu Y.-L., et al. Differential subcellular distribution renders HAI-2 a less effective protease inhibitor than HAI-1 in the control of extracellular matriptase proteolytic activity. Genes Dis. 2022;9(4):1049–1061. doi: 10.1016/j.gendis.2020.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Zhang Q., et al. Overexpression of Matriptase‐1 inhibitor HAI‐1 as an effective strategy to reduce proteolytic clipping in Chinese hamster ovary cells. Biotechnol. J. 2025;20(10) doi: 10.1002/biot.70141. [DOI] [PubMed] [Google Scholar]
  • 25.Domoto T., et al. Cleavage of hepatocyte growth factor activator inhibitor‐1 by membrane‐type MMP‐1 activates matriptase. Cancer Sci. 2012;103(3):448–454. doi: 10.1111/j.1349-7006.2011.02162.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Ishikawa T., et al. Matrix metalloproteinase-7 induces homotypic tumor cell aggregation via proteolytic cleavage of the membrane-bound Kunitz-type inhibitor HAI-1. J. Biol. Chem. 2017;292(50):20769–20784. doi: 10.1074/jbc.M117.796789. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Kiyomiya K.-i., et al. Matriptase activation and shedding with HAI-1 is induced by steroid sex hormones in human prostate cancer cells, but not in breast cancer cells. Am. J. Physiol. Cell Physiol. 2006;291(1):C40–C49. doi: 10.1152/ajpcell.00351.2005. [DOI] [PubMed] [Google Scholar]
  • 28.Kataoka H., et al. Hepatocyte growth factor activator inhibitors (HAI‐1 and HAI‐2): emerging key players in epithelial integrity and cancer. Pathol. Int. 2018;68(3):145–158. doi: 10.1111/pin.12647. [DOI] [PubMed] [Google Scholar]
  • 29.Snell K.I., et al. Exploring the roles of urinary HAI-1, EpCAM & EGFR in bladder cancer prognosis & risk stratification. Oncotarget. 2018;9(38) doi: 10.18632/oncotarget.25397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Zeegers M.P., et al. The West Midlands bladder cancer prognosis programme: rationale and design. BJU Int. 2010;105(6):784–788. doi: 10.1111/j.1464-410X.2009.08849.x. [DOI] [PubMed] [Google Scholar]
  • 31.Kind S., et al. A shift from membranous and stromal syndecan‐1 (CD138) expression to cytoplasmic CD138 expression is associated with poor prognosis in breast cancer. Mol. Carcinog. 2019;58(12):2306–2315. doi: 10.1002/mc.23119. [DOI] [PubMed] [Google Scholar]
  • 32.Yamashita F., et al. Role of the polycystic kidney disease domain in matriptase chaperone activity and localization of hepatocyte growth factor activator inhibitor‐1. FEBS J. 2022;289(12):3422–3439. doi: 10.1111/febs.16348. [DOI] [PubMed] [Google Scholar]
  • 33.Chen C.-Y., Lin T.-N., Huang H.-P. Membrane-anchored serine protease inhibitors: physiological functions, mechanisms, and roles in cancer. Int. J. Mol. Sci. 2026;27(4):2000. doi: 10.3390/ijms27042000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ye J., et al. Loss of hepatocyte growth factor activator inhibitor type 1 participates in metastatic spreading of human pancreatic cancer cells in a mouse orthotopic transplantation model. Cancer Sci. 2014;105(1):44–51. doi: 10.1111/cas.12306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Sjödahl G., et al. A molecular taxonomy for urothelial carcinoma. Clin. Cancer Res. 2012;18(12):3377–3386. doi: 10.1158/1078-0432.CCR-12-0077-T. [DOI] [PubMed] [Google Scholar]
  • 36.Höglund M., et al. The lund taxonomy for bladder cancer classification–from gene expression clustering to cancer cell molecular phenotypes, and back again. J. Pathol. 2023;259(4):369–375. doi: 10.1002/path.6062. [DOI] [PubMed] [Google Scholar]
  • 37.Saleem M., et al. A novel biomarker for staging human prostate adenocarcinoma: overexpression of matriptase with concomitant loss of its inhibitor, hepatocyte growth factor activator inhibitor-1. Cancer Epidemiol. Biomark. Prev. 2006;15(2):217–227. doi: 10.1158/1055-9965.EPI-05-0737. [DOI] [PubMed] [Google Scholar]
  • 38.Oberst M.D., et al. Expression of the serine protease matriptase and its inhibitor HAI-1 in epithelial ovarian cancer: correlation with clinical outcome and tumor clinicopathological parameters. Clin. Cancer Res. 2002;8(4):1101–1107. [PubMed] [Google Scholar]
  • 39.Nagata K., et al. Expression of hepatocyte growth factor activator and hepatocyte growth factor activator inhibitor type 1 in human hepatocellular carcinoma. Biochem. Biophys. Res. Commun. 2001;289(1):205–211. doi: 10.1006/bbrc.2001.5916. [DOI] [PubMed] [Google Scholar]
  • 40.Liu C.-L., et al. Expression of serine peptidase inhibitor Kunitz type 1 in differentiated thyroid cancer. Histochem. Cell Biol. 2018;149(6):635–644. doi: 10.1007/s00418-018-1660-2. [DOI] [PubMed] [Google Scholar]
  • 41.Isali I., et al. Urologic Oncology: Seminars and Original Investigations. Elsevier; 2022. Gene network profiling in muscle-invasive bladder cancer: a systematic review and meta-analysis. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data will be made available on request.

Articles from Biochemistry and Biophysics Reports are provided here courtesy of Elsevier

RESOURCES