Salivary KLK5 and uPA are potential biomarkers for malignant transformation of OLK and OLP

Abstract

BACKGROUND:

Oral squamous cell carcinoma (OSCC) usually originates from oral potentially malignant disorders (OPMD), such as oral leukoplakia (OLK) and oral lichen planus (OLP). Identifying biomarkers for the early diagnosis and evaluation of malignant transformation in OPMD could improve the survival rate of OSCC patients.

OBJECTIVE:

The present study aimed to screen for potential salivary biomarkers for evaluating the malignant transformation of OPMD.

METHODS:

Salivary proteases from OLK and OSCC patients or healthy donors and proteases in cultural medium from DOK and Cal-27 cells were detected with a human protease array kit. The concentrations of the salivary Kallikrein 5 (KLK5) and urokinase-type plasminogen activator (uPA) proteases were measured by ELISA. Receiver operating characteristics (ROC) to determine the potential value of these proteases in clinical diagnosis were calculated using SPSS software. Immunohistochemistry was used to detect the KLK5 and uPA expression in the oral organizations.

RESULTS:

The salivary protease spectrum was different among patients with OLK and OSCC and healthy donors. KLK5 and uPA levels in saliva tended to increase as the disease progressed (healthy $<$ OPMD [OLK and OLP] $<$ OSCC). ROC curves showed the optimum diagnostic cutoffs for KLK5 as a biomarker for OLK, OLP, and OSCC were 5.97, 6.03, and 9.45 pg/mL, respectively, while the cutoffs for uPA were 17.19, 17.26, and 20.96 pg/mL. Their combined analysis showed a higher sensitivity for the differential diagnosis of disease. Furthermore, higher levels of KLK5 and uPA were observed in OSCC tissues than in OLK and OLP.

CONCLUSIONS:

Salivary KLK5 and uPA are potential biomarkers for evaluating OLK and OLP malignant transformation and early diagnosis of OSCC.

Keywords

Saliva protease spectrum oral leukoplakia oral lichen planus oral squamous cell carcinoma kallikrein 5 urokinase-type plasminogen activator

1. Introduction

Oral squamous cell carcinoma (OSCC) is the most predominant type of oral cancer [1]. Although comprehensive strategies for its recognition and management have advanced in the past decade, the 5-year survival rate for OSCC has not significantly improved [2, 3]. This lack of improvement might be partly due to the lack of reliable diagnostic biomarkers or early detection methods, which delays the diagnosis of OSCC.

The histopathological development of OSCC occurs via oral potentially malignant disorders (OPMD) that go through stages, including epithelial hyperplasia and dysplasia, followed by OSCC stages consisting of early invasive and lymph node metastatic stages [4]. OPMD, which include oral leukoplakia (OLK), oral erythroplakia (OE), proliferative verrucous leukoplakia (PVL), oral submucous fibrosis (OSMF), oral lichen planus (OLP), and actinic cheilitis (AC) [5, 6], are morphologically altered tissues in which oral cancer is more likely to occur compared to the normal oral mucosa. As the most common OPMD, OLK is a pathological condition with white patches or plaques that could not be attributed to any clinically or histologically definitive lesion, while OLP is defined as a common chronic mucocutaneous disease with unknown etiology [7]. The mean malignant transformation rate of OLK for the total population amounts to 3.5% (405/11423), with a wide range of rates (0.13–34.0%) presented in a recent systematic review study [8]. OLP has a malignant transformation rate of 1.14% (95% CI $=$ 0.84–1.49%) [9]. Early resection reduces the risk of malignant transformation of OLK and OLP. However, the choice of conservative or surgical treatment depends on the estimated risk for cancer [10, 11]. Identifying biomarkers to evaluate the risk of cancerization of OLK and OLP is conducive to the early diagnosis of OSCC and the formulation of clinic therapeutic schedules [12, 13].

Saliva contains a variety of hormones, antibodies, microorganisms, proteins, enzymes, and cytokines. Its components reflect the physiological state of the body [14, 15]. With the development of salivaomics, an increased number of biomarkers related to various systemic and oral diseases, such as gastric cancer [16], type 2 diabetes [17], breast cancer [18], systemic lupus erythematosus [19], oral squamous cell carcinoma [20, 21, 22], periodontal disease [23], and caries [24, 25, 26], have been identified in saliva.

Proteases drive the life cycle of all proteins, ensuring the activation of newly synthesized proteins and degradation of unneeded proteins [27]. The abnormal release and activation of proteases can generate pathological changes in cells, tissues, and organs, especially in the formation and metastasis cancers [28]. Many kinds of proteases can be detected in human saliva, and changes in the salivary protease spectrum can be used as a potential biomarker for oral diseases [21, 29].

In this study, we analyzed the protease spectrum of saliva from healthy donors and patients with OLK and OSCC and in the supernatants from dysplastic oral keratinocytes (DOK, human mucosa precancerous lesion cell line) and Cal-27 (oral squamous cell carcinoma cell line) cells using human protease arrays. Based on the results of the protease array analysis, we measured the concentrations of the kallikrein 5 (KLK5) and urokinase-type plasminogen activator (uPA) proteases in the saliva of healthy donors and patients with OPMD (OLK and OLP) or OSCC by ELISA. The expression levels of these proteases were also evaluated in tissues by immunohistochemistry (IHC). The ELISA data were used to calculate the sensitivity and specificity for diagnosing OLK, OLP, and OSCC using receiver operating characteristic (ROC) curves. Our results demonstrated that salivary KLK5 and uPA are potential biomarkers for evaluating OLK and OLP malignant transformation and the early diagnosis of OSCC.

2. Materials and methods

2.1 Ethics statement

All volunteers were informed that their saliva would be collected to assay the concentration of proteases. All data, information, and samples were used for our research study only. Written informed consent for collecting saliva and tissues and identifying individual medical information was obtained from each subject. The Committee for Ethics Approval of the West China School of Stomatology, Sichuan University and State Key Laboratory of Oral Diseases approved this study (IRB Reference Number: WCHS-IRBD-2015-147).

2.2 Saliva collection

Stimulated saliva was collected with salivette tubes (Sarstedt AG & Co., Numbrecht, Germany). Before sample collection, participants were prohibited from eating and drinking for 2 h. They were then asked to gargle three times with 10 mL of clear water. After 30 min, they put the cotton from the salivette tube into their mouths without chewing for 10 min, and then spit the cotton back into the tubes. The tubes were kept on ice following collection. The samples were centrifuged at 10,000 $\times$ g for 10 min at 4 ${}^{\circ}$ C, and the supernatants were divided into sterile 0.5 mL EP tubes and stored at $-$ 80 ${}^{\circ}$ C. All samples were collected between 8:30 a.m. and 11:30 a.m.

A total of 120 saliva samples were collected from patients with OSCC ( $n=$ 20), OLK malignant transformation to OSCC (OLK-T) ( $n=$ 20), OLP malignant transformation to OSCC (OLP-T) ( $n=$ 20), OLK ( $n=$ 20), OLP ( $n=$ 20), and healthy controls ( $n=$ 20). OLK-T and OLP-T are defined as OSCC that developed due to malignant transformation of OLK and OLP. All OSCC patients were selected based on their pathological diagnosis (T1N0M0-T4N0M0), and saliva was collected before surgery. All volunteers were informed of the aims and procedures of the study and that their medical data would be used for this research study only. Patients were excluded from the study if they had systemic or salivary gland diseases, were smokers, or were currently taking non-steroidal anti-inflammatory drugs, antibiotics, or hormonal compounds. Patients who had received scaling and root planing or oral mucosal disease treatment in the previous six months were also excluded.

2.3 Cell culture and supernatant collection

Dysplastic oral keratinocytes (DOK) and Cal-27 cells were cultured as previously described [30]. To analyze secreted proteases, the cells were seeded in six-well plates at a density of 2 $\times$ 10 ${}^{5}$ cells/well and incubated in a humidified 5% CO ${}_{2}$ atmosphere at 37 ${}^{\circ}$ C overnight. The medium (Dulbecco’s modified Eagle’s medium (DMEM; Gibco) supplemented with 10% FBS, 100 IU/mL penicillin and 100 $\mu$ g/mL streptomycin) was replaced, and the cells were incubated for an additional 72 h. The medium was collected from the cells and centrifuged at 10,000 $\times$ g at 4 ${}^{\circ}$ C for 10 min. The supernatants were stored at $-$ 80 ${}^{\circ}C$ .

2.4 Protease array and image analysis

Twelve saliva samples ( $n=$ 4 each from patients with OLK or OSCC and healthy volunteers) were selected randomly from the 120 collected saliva samples and used for protease array analysis. Saliva from each volunteer (0.5 mL) was evaluated using a Human Protease Array Kit (R&D Systems, Inc Minneapolis, USA), according to the manufacturer’s instructions. Each sample was analyzed on two independent array membranes. For image analysis, the X-ray film (Carestream Health Inc. Wuhan, China) was scanned with a CanoScan LiDE 700F scanner (Canon, Beijing, China), and grayscale values for each blot were analyzed using ImageJ (National Institutes of Health, Bethesda, Maryland, USA). Briefly, the collected image was converted to grayscale (8 bit). The influence of the image background was eliminated by selecting a light background and setting the rolling ball radius to 50 pixels. The unit of length was set to pixel, and the “integrated density” was set as the indicator for the analysis. The images were inverted, so that the background pixel values were near zero. Each spot was circled and then measured. The negative control spots were used as the background value. The relative level (RL) of each salivary protease was calculated as follows:

$\displaystyle\text{RL}=\frac{\overline{\text{ID}}s-\overline{\text{ID}}\text{n% }}{\overline{\text{ID}}\text{rs}-\overline{\text{ID}}\text{n}}$

RL: Relative Level of Proteases; IDs: Integrated Density of samples; IDn: Integrated Density of negative control; IDrs: Integrated Density of reference spots.

2.5 Quantitative assessment of salivary proteases by ELISA

ELISA analysis was performed on 120 saliva samples from OSCC ( $n=$ 20), OLK-T ( $n=$ 20), OLP-T ( $n=$ 20), OLK ( $n=$ 20), and OLP ( $n=$ 20) patients and healthy controls ( $n=$ 20). Salivary KLK5 and uPA concentrations were measured using human KLK5 (JYM2052Hu) and human uPA (JYM1036Hu) ELISA kits (Wuhan Colorful Gene Biological Technology Co., Wuhan, China). The ELISAs were performed using the sandwich method, as recommended by the manufacturer. Briefly, the diluted standards and saliva samples (50 $\mu$ L) were added to the Microelisa stripplates, which were pre-coated with anti-human KLK5 or uPA antibody. Each sample was tested in triplicate. The plates were sealed, incubated at 37 ${}^{\circ}$ C for 30 min, and washed five times with wash buffer. HRP-conjugated KLK5- or uPA-specific antibody were added to each well except the blank control well. The sealed plates were incubated at 37 ${}^{\circ}$ C for 30 min and then washed five times with wash buffer. Chromogen Solution A (50 $\mu$ L) and Chromogen Solution B (50 $\mu$ L) were added to each well. After gently shaking, the plates were incubated at 37 ${}^{\circ}$ C for 10 min. The reaction was terminated with 50 $\mu$ L of stop solution. The optical density (OD) of each well was measured at 450 nm using a microtiter plate reader. The OD 450 nm value of the blank control well was set as zero. The standard curve and sample concentrations were calculated automatically by the microtiter plate reader.

The ELISA data were used to determine the optimal cutoff values, specificity, sensitivity, and area under the curve (AUC) with 95% confidence intervals (CI) with the ROC curve using SPSS 19.0 software.

2.6 Immunohistochemical analysis

A panel of 56 formalin-fixed, paraffin-embedded human OLK, OLP, and OSCC specimens were obtained from the West China Stomatological Hospital. For IHC analysis, all slides were deparaffinized twice in xylene (15 min each) and rehydrated through a series of graded alcohol solutions. Antigen retrieval was performed by heating the slides with boiled water for 20 min in 10 mmol/L sodium citrate buffer (pH 6.0). After cooling, endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 15 min and normal serum for 30 min at room temperature. Subsequently, the slides were incubated with rabbit anti-human uPA (cat. no. ET1703-26; 1:100; HUABIO) or KLK5 (cat.no. 10514-2-AP; 1:50; Proteintech) antibody at 4 ${}^{\circ}$ C overnight. For the negative controls, the primary antibody was replaced with antibody diluent. The following day, the slides were incubated with HRP-labeled goat rabbit secondary antibody for 30 min at 37 ${}^{\circ}$ C. Color development was performed using a DAB color development kit (ZSGB-BIO; China). The nuclei were stained with hematoxylin for 2 min at room temperature. Finally, the slices were dehydrated, cleaned, and mounted. Images were captured using the digital slice scanner Pannoramic MIDI (3DHistech, Hungary).

Image analysis was performed using ImageJ 1.48v software with the IHC Profiler plugin. For the ImageJ settings, the mode was set for cytoplasmic stained images, and the vector was H DAB. The staining intensity was evaluated according to the following ratings: 0, no staining; 1, weak staining; 2, medium staining; 3, strong staining. The H-score was calculated as follows: H-score $=$ 0 $\times$ (% negative tumor cells) $+$ 1 $\times$ (% weak staining) $+$ 2 $\times$ (% medium staining) $+$ 3 $\times$ (% strong staining). The H-scores ranged from 0 (100% negative tumor cells) to 300 (100% strong staining of tumor cells).

2.7 Statistical analysis

SPSS 19.0 software (Chicago, IL, USA) was used for data analysis. The statistical power for comparing KLK5 and uPA between groups was calculated using one-way analysis of variance F-tests in the means with G*Power (3.1.9.2) software. The 1- $\beta$ values for the experiments were 1.000, which indicated that our study was powered. The statistical significance of the comparisons was determined using one-way analysis of variance (ANOVA), followed by the post-hoc Tukey’s multiple comparisons test. $P<$ 0.05 indicated statistically significant differences. ROC curves were constructed to determine the optimal cutoff value, specificity, sensitivity, and area under the curve (AUC) with 95% confidence intervals (CI) for the potential biomarkers.

Table 1
Coordinates of proteases located on the human protease array membrane

	1/2	3/4	5/6	7/8	9/10	11/12	13/14	15/16	17/18	19/20
A	RS	ADAM8	ADAM9	ADAMTS1	ADAMTS13	Cathepsin A	Cathepsin B	Cathepsin C	Cathepsin D	RS
B		Cathepsin E	Cathepsin L	Cathepsin S	Cathepsin V	Cathepsin X/Z/P	DPPIV/CD26	Kallikrein 3/PSA	Kallikrein 5
C		Kallikrein 6	Kallikrein 7	Kallikrein 10	Kallikrein 11	Kallikrein 13	MMP-1	MMP-2	MMP-3
D		MMP-7	MMP-8	MMP-9	MMP-10	MMP-12	MMP-13	Neprilysin/CD10	Presenilin
E	RS	Proprotein	Proteinase 3	uPA/Urokinase	NC
		Convertase 9

RS: reference spot, NC: negative control.

3. Results

3.1 Saliva from OSCC and OLK patients contain more proteases than from healthy donors

We analyzed 12 saliva samples from healthy individuals and patients with OLK and OSCC using human protease array kits that detect 35 human proteases (Fig. 1A and Table 1). The salivary protease spectrum was quite different between the three groups. Saliva from healthy donors contained 17 proteases (Fig. 1B). In contrast, the saliva from OLK (Fig. 1C) and OSCC (Fig. 1D) patients contained 26 and 30 proteases, respectively. The relative levels of each protease calculated from scanned images are presented in Fig. 1G. Most of the salivary proteases (e.g., cathepsin V, KLK5, KLK7, KLK13, MMP-1, MMP-2, MMP-3, MMP-13, proteinase3, and uPA) increased gradually as the disease progressed from OLK to OSCC. In the saliva from OSCC patients, the relative KLK5 and uPA expression levels were 11.57- and 40.07-fold higher than those observed in healthy volunteers, respectively. Moreover, these expression levels were 2.88- and 11.86-fold higher than the levels in OLK patients, respectively. Furthermore, the relative KLK5 and uPA expression levels in the saliva of OLK patients were 4.02- and 3.38-fold higher than found in healthy volunteers. These results indicated that the identities and concentrations of salivary proteases, especially KLK5 and uPA, were associated with OSCC disease progression.

Figure 1.

Protease spectrum in saliva and cell culture supernatants. Saliva proteases were detected using human protease array kits. (A) Each dot on the membrane represents a single protease. (B) The salivary protease spectrum of healthy donors. (C) The salivary protease spectrum of OLK patients. (D) The salivary protease spectrum of OSCC patients. (E) The protease spectrum of DOK cell culture supernatant. (F) The protease spectrum of Cal-27 cell culture supernatant. (G) Quantitative analysis of saliva proteases. (H) Semi-quantitative analysis of supernatant proteases. Representative pictures for each type of sample are shown.

Table 2

Concentration of salivary KLK5 and uPA detected by ELISA

	Healthy	OLK	OLK-T	OLP	OLP- T	OSCC
KLK5 (pg/mL)	6.16 $\pm$ 0.35	8.787 $\pm$ 0.44	11.42 $\pm$ 0.77	9.09 $\pm$ 0.53	12.15 $\pm$ 0.70	12.11 $\pm$ 1.20
uPA (pg/mL)	16.47 $\pm$ 0.68	18.12 $\pm$ 0.70	20.26 $\pm$ 1.10	21.66 $\pm$ 0.57	26.75 $\pm$ 0.93	27.34 $\pm$ 1.00

OLK-T: OSCC which had undergone malignant transformation originating from OLK. OLP-T: OSCC which had undergone malignant transformation originating from OLP.

Table 3

Cutoff point, AUC, 95% CI, specificity and sensitivity of KLK5 and uPA for diagnosis of diseases

	Test result	AUC	Std.	Asymptotic	Asymptotic 95% Confidence interval		Cutoff	Specificity	Sensitivity
	variable(s)		Error ${}^{\text{a}}$	Sig. ${}^{\text{b}}$	Lower bound	Upper bound	(pg/mL)	(%)	(%)
KLK5	OLK	0.83	0.07	0.00	0.70	0.95	5.97	47.70	100.00
	OLK-T	0.95	0.03	0.00	0.88	1.00	8.36	55.20	94.70
	OLP	0.83	0.07	0.00	0.71	0.96	6.03	43.10	91.70
	OLP-T	0.98	0.02	0.00	0.94	1.00	8.80	69.50	94.70
	OSCC	0.90	0.05	0.00	0.81	0.99	9.45	63.20	84.20
uPA	OLK	0.85	0.06	0.00	0.73	0.97	17.19	47.90	94.70
	OLK-T	0.94	0.04	0.00	0.85	1.00	19.45	56.90	80.70
	OLP	0.90	0.05	0.00	0.80	0.99	17.26	63.20	100.00
	OLP-T	0.98	0.02	0.00	0.93	1.00	22.29	80.50	89.50
	OSCC	0.99	0.01	0.00	0.97	1.00	20.96	84.70	100.00

a: Standard error, b: Asymptotic significance.

3.2 OSCC cells release more proteases in the culture medium than oral mucosa precancerous lesion cell

The culture medium from DOK and Cal-27 cells were also analyzed with the human protease array kit. Cal-27 cells released 19 different proteases into the culture medium compared to 14 by DOK cells (Fig. 1E and F). Consistent with the saliva data, the relative expression levels of most proteases in the supernatant of Cal-27 cells (e.g., ADAM9, Cathepsin B, Cathepsin C, Cathepsin D, Cathepsin V, KLK5, KLK10, KLK11, MMP-3, MMP-7, MMP-10, MMP-13, and uPA) were higher than in the supernatant from DOK cells (Fig. 1H). In particular, the relative KLK5 and uPA expression levels in the Cal-27 medium were 12.86- and 499.41-fold higher than that in the DOK medium, respectively. These data suggested that KLK5 and uPA might be potential biomarkers to evaluate precancerous oral mucosa lesion cancerization.

Figure 2.

Quantitative analysis of patient salivary KLK5 and uPA levels. KLK5 and uPA levels were measured in the saliva from healthy donors and OLK, OLK-T, OLP, OLP-T, and OSCC patients by ELISA. Data are presented as the mean $\pm$ standard deviation (SD). a: $P<$ 0.05 compared to healthy controls, b: $P<$ 0.05 compared to OLK or OLP patients.

3.3 Salivary KLK5 and uPA are potential biomarkers for OLK and OLP malignant transformation and early diagnosis of OSCC

Based on the human protease array results from saliva and cell culture supernatants (Fig. 1), we evaluated the KLK5 and uPA concentrations in saliva samples from healthy individuals and patients with OLK, OLP, OLK-T, OLP-T, or OSCC. The levels of KLK5 and uPA gradually and significantly increased from healthy donors to patients with OLK or OLP to patients with OSCC (OLK-T, OLP-T, and OSCC) (Fig. 2 and Table 2). There were no significant differences between OLK-T, OLP-T and OSCC patients.

ROC curves showed the optimal KLK5 diagnostic cutoffs to be 5.97 pg/mL for OLK, 6.03 pg/mL for OLP, 8.36 pg/mL for OLK-T, 8.80 pg/mL for OLP-T, and 9.45 pg/mL for OSCC (Fig. 3A). The optimal uPA diagnostic cutoffs were 17.19 pg/mL for OLK, 17.26 pg/mL for OLP, 19.45 pg/mL for OLK-T, 22.29 pg/mL for OLP-T, and 20.96 pg/mL for OSCC (Fig. 3B). The corresponding AUC value, 95% CI, sensitivity, and specificity for KLK5 and uPA are presented in Table 3. Both KLK5 and uPA showed high sensitivity as a biomarker for the diagnosis of OLP, OLP-T, OLK, OLK-T, and OSCC, and their combination improved the diagnostic accuracy compared to either test alone, especially the specificity of diagnosis (Fig. 3C–G, Table 4).

Figure 3.

ROC curves for salivary KLK5 and uPA alone and in combination for the diagnosis and differential diagnosis of OLK, OLK-T, OLP, OLP-T, and OSCC. (A–B) ROC curves for KLK5 and uPA for patients with OLK, OLK-T, OLP, OLP-T, and OSCC versus healthy donors. (C–G) ROC curves for KLK5 and uPA alone and in combination for the diagnosis of OLK, OLP, OLK-T, OLP-T, and OSCC. (H–K) ROC curves for KLK5 and uPA alone and in combination for patients with OLK or OLP versus OSCC.

Table 4

AUC, 95% CI, specificity and sensitivity of KLK5, uPA and their combination for diagnosis of diseases

	Test result	AUC	Std.	Asymptotic	Asymptotic 95% Confidence interval		Specificity	Sensitivity
	variable(s)		Error ${}^{\text{a}}$	Sig. ${}^{\text{b}}$	Lower bound	Upper bound	(%)	(%)
OLK	KLK5	0.83	0.07	0.00	0.70	0.95	47.70	100.00
	uPA	0.65	0.09	0.13	0.46	0.83	47.90	94.70
	Comb	0.83	0.06	0.00	0.71	0.96	62.10	100.00
OLK-T	KLK5	0.83	0.07	0.00	0.71	0.96	55.20	94.70
	uPA	0.70	0.09	0.04	0.53	0.87	56.90	80.70
	Comb	0.83	0.06	0.00	0.71	0.96	72.10	100.00
OLP	KLK5	0.95	0.03	0.00	0.88	1.00	43.10	91.70
	uPA	0.81	0.07	0.00	0.67	0.95	63.20	100.00
	Comb	0.96	0.03	0.00	0.91	1.00	84.20	94.00
OLP-T	KLK5	0.98	0.02	0.00	0.94	1.00	69.50	94.70
	uPA	0.88	0.05	0.00	0.78	0.99	80.50	89.50
	Comb	0.99	0.01	0.00	0.97	1.00	89.50	100.00
OSCC	KLK5	0.90	0.05	0.00	0.81	0.99	63.20	84.20
	uPA	0.89	0.05	0.00	0.79	0.99	84.70	100.00
	Comb	0.95	0.03	0.00	0.88	1.00	94.20	89.50

a: Standard error, b: Asymptotic significance.

We found that raised salivary uPA concentrations could separately distinguish patients with OSCC from those with OLK or OLP, while salivary KLK5 alone could not. The individual tests had low AUC values (0.65–0.78) and low specificity (21.0–52.6%). When KLK5 and uPA were tested together, the differential diagnostic accuracy between OSCC and OLK/OLP improved, with AUC values of 0.85 (95% CI $=$ 0.73–0.97) for OLK and OLK-T, 0.88 (95% CI $=$ 0.76–0.99) for OLP and OLP-T, 0.89 (95% CI $=$ 0.79–0.99) for OLK and OSCC, and 0.86 (95% CI $=$ 0.74–0.98) for OLP and OSCC. Specificity also increased (67.90% for OLK and OLK-T, 63.2% for OLP and OLP-T, 67.5% for OLK and OSCC, and 65% for OLP and OSCC). (Fig. 3H–K, Table 5).

Table 5

Cutoff point, AUC, 95% CI, specificity and sensitivity of KLK5, uPA and their combination for discriminating OLK and OLP from OSCC

	Test result	AUC	Std.	Asymptotic	Asymptotic 95% Confidence interval		Cutoff	Specificity	Sensitivity
	variable(s)		Error ${}^{\text{a}}$	Sig. ${}^{\text{b}}$	Lower bound	Upper bound	(pg/mL)	(%)	(%)
OLK vs OLK-T	KLK5	0.74	0.08	0.01	0.58	0.90	8.26	34.80	94.70
	uPA	0.81	0.07	0.00	0.67	0.94	19.77	36.80	100.00
	Comb	0.85	0.06	0.00	0.73	0.97		67.90	89.50
OLP vs OLP-T	KLK5	0.78	0.07	0.00	0.64	0.93	7.25	36.80	100.00
	uPA	0.84	0.07	0.00	0.71	0.97	19.34	52.60	94.70
	Comb	0.88	0.06	0.00	0.76	0.99		63.20	90.50
OLK vs OSCC	KLK5	0.68	0.09	0.06	0.51	0.86	7.14	21.00	89.50
	uPA	0.86	0.06	0.00	0.74	0.97	20.57	47.40	94.30
	Comb	0.89	0.05	0.00	0.79	0.99		67.50	95.60
OLP vs OSCC	KLK5	0.65	0.09	0.11	0.47	0.83	7.49	21.10	84.20
	uPA	0.84	0.06	0.00	0.72	0.97	21.57	42.10	85.20
	Comb	0.86	0.06	0.00	0.74	0.98		65.00	75.40

a: Standard error, b: Asymptotic significance.

Altogether, the ROC analysis illustrated that salivary KLK5 and uPA were potential biomarkers for the diagnosis of OLK, OLP, and OSCC, and their combination was better for distinguishing OSCC from OLK and OLP.

3.4 OSCC expresses higher KLK5 and uPA levels than OLK and OLP

KLK5 and uPA expression levels were analyzed in 16 OLK, 5 OLP, and 35 OSCC specimens. As shown in Fig. 4A, both KLK5 and uPA were expressed at low levels in oral epithelial tissue from OLK and OLP patients. In contrast, OSCC tissues were more intensely stained. Semi-quantitative analysis showed that the KLK5 and uPA levels were significantly increased in OSCC compared to OLK and OLP (Fig. 4B–C; $P<$ 0.001). Moreover, the expression was higher in highly differentiated OSCC than in poorly differentiated OSCC ( $P<$ 0.05). These results suggested that KLK5 and uPA are involved in the development of OSCC and further indicated that they are potential biomarkers for evaluating the malignant transformation of OLK and OLP.

Figure 4.

KLK5 and uPA levels in OLK, OLP, and OSCC specimens. (A) KLK5 and uPA expression levels were analyzed by immunohistochemistry in OLK, OLP, and OSCC specimens; (B) H-score analysis of KLK5 in different oral tissues. (C) H-score analysis of uPA in different oral tissues. *** $P<$ 0.001, * $P<$ 0.05. Scale bar $=$ 200 $\mu$ m.

4. Discussion

The identification of biomarkers for diverse tumor types in the blood, urine, cerebrospinal fluid, or other body tissues has attracted significant attention in the last decade. In theory, tumor biomarkers can be used to screen, diagnose, stage, and monitor cancers [31]. Several proteases have been extensively studied as potential biomarkers, including uPA/uPAR and matrix metalloproteinases (MMPs) [32]. The abnormal activation or secretion of proteases can give rise to pathological processes in cells, tissues, and organs. Many proteases play important roles in the invasion and metastasis of cancers because of their ability to degrade the extracellular matrix (ECM) barrier surrounding a tumor [33]. Saliva is an important body fluid that can reflect the body’s physiological state. An increasing number of disease-related salivary biomarkers have been discovered, and the analysis of the salivary protease spectrum has been reported as a perfect way to evaluate oral health status [34, 35].

OLK and OLP are the most common OPMD. Many methods exist to prevent and treat OLK, such as drugs, conventional scalpel excision, photodynamic therapy, cryosurgery, and laser ablation, while the main method for treating OLP is medication [4]. Despite the continuous advancements in treatment technology, OLK and OLP still have a high malignant transformation rate [8, 9]. Recurrence and malignant transformation of OLK and OLP are major clinical obstacles. Identification of biomarkers with high specificity and sensitivity for identifying malignant transformation of OLK and OLP is a major focus in the field. Such markers could allow for surgical intervention at an early stage.

OLK, OLP, and OSCC are usually located on the surface of oral mucous and contact saliva directly. As a result, the saliva of those patients will contain all kinds of molecules secreted and released by tumor cells or cells of precancerous lesions, suggesting that it would be easier to find a specific and sensitive biomarker for oral diseases in saliva than in serum. Our results demonstrated that the salivary protease spectrum was quite different between healthy donors and OLK or OSCC patients. Indeed, both the number of proteases and their levels significantly increased with disease progression (i.e., health donors $<$ OLK $<$ OSCC). A similar phenomenon was observed with media collected from normal and oral squamous cell carcinoma cells. Our previous research demonstrated that the combined analysis of salivary protease MMP-9, KLK5, and cathepsin V was a potential biomarker for the diagnosis of OSCC [21], indicating that salivary protease spectrum analysis is a useful method to assess oral health status. In the present study, we found that salivary protease analysis could also be a useful method for evaluating the malignant transformation of OLK and OLP.

Furthermore, we selected two important proteases from the protease spectrum, KLK5 and uPA for further research as they were both clearly increased in the saliva samples and supernatants. Human kallikreins (KLKs) constitute a subgroup of 15 secreted trypsin or chymotrypsin-like serine proteases [36]. KLK5 is abundantly expressed in human OSCC and correlates with the formation of more aggressive OSCC tumors in a murine orthotopic OSCC model [37]. It may promote the metastatic dissemination of OSCC by promoting the loss of junctional integrity through the cleavage of desmoglein1 [38]. KLK5 can mediate PAR-2 activation, which regulates the expression of inflammation-associated mRNAs and microRNAs, thereby modulating the progression of oral tumors [39]. uPA belongs to the plasminogen activator system, which includes the plasminogen activators (uPA, tPA), the cell membrane receptor for uPA (uPAR), the plasminogen activator inhibitors (PAI-1, PAI-2), and plasmin [40, 41]. The uPA/uPAR complex plays a significant role in the invasive and metastatic potential of various tumors [42]. Active uPA increases the production of plasmin from plasminogen, leading to ECM degradation. This degradation facilitates cancer cell invasion into the surrounding tissue. Most head and neck cancer studies have shown that uPA/uPAR overexpression enhances tumor cell proliferation, migration, and invasion [43, 44]. The upregulation of uPA contributes to the promotion of oral tumor invasion by extracellular matrix metalloproteinase inducer (EMMPRIN) [44]. It also promotes ESCC cell proliferation, migration, and invasion via the PI3K/AKT and ERK signaling pathways [45]. Those studies indicated that KLK5 and uPA are important proteins in the development and metastasis of OSCC, but rare study of the two in precancerous lesions are performed.

In current study, we detected the contents of saliva KLK5 and uPA in healthy peoples, patients with OLK, OLP, OLK-T, OLP-T, and OSCC. Simultaneously, we tested the expression of them in OLK, OLP and OSCC histologic sections. The results confirmed that their expression levels progressively increased with the development of the OSCC. These data strongly demonstrated that KLK5 and uPA play an important role in the process of OLK and OLP cancerization. Furthermore, ROC curves analysis suggested that both KLK5 and uPA have high AUCs in the diagnosis of OLK and OLP, as well as OLK-T, OLP-T and OSCC. Based on the generated cutoff values for diagnosing OLK, OLP, OLK-T, OLP-T and OSCC, they all have high sensitivity but low specificity. However, the combined analysis of KLK5 and uPA not only significantly increased the specificity for the diagnosis of all five diseases but also allowed the differential diagnosis of OSCC from OLK and OLP. These results suggested that the combination of KLK5 and uPA could represent a potential biomarker for determining the malignant transformation of OLK or OLP into OSCC.

5. Conclusions

Our current study demonstrated that salivary KLK5 and uPA could be potential biomarkers for evaluating the malignant transformation of OLK and OLP and early diagnosis of OSCC. Importantly, their combination could aid in the differential diagnosis of OSCC from OLK and OLP. Unfortunately, our sample size was limited. The potential value of KLK5 and uPA as biomarkers requires a multicenter study involving a larger number of patients to verify their predictive efficacy.

Footnotes

Acknowledgments

This research was funded by the National Natural Science Foundation of China (No. 81372892) and the Interdisciplinary integration Foundation (2020KXK0405) from the State Key Laboratory of Oral Disease.

Conflict of interest

The authors declare that they have no conflicts of interest to disclose.

Author contributions

Conception: Yun Feng and Ping Zhang

Interpretation or analysis of data: Ping Zhang, Yingzhu Kang, Jiao Chen, Xiaoying Li, Min Luo, Hongli Chen, Bomiao Cui, Liwei Wang, Jingnan Wang, and Die Lv

Preparation of the manuscript: Ping Zhang and Yingzhu Kang

Revision for important intellectual content: Ping Zhang, Jiao Chen, and Xiaoying Li

References

Omori

Nishio

Masuda

Miyachi

Ueda

Nakano

Sato

Mimori

Taguchi

Hikasa

Nishina

Tashiro

Kiyono

Mak

T.W.

Nakao

Nakagawa

Maehama

and Suzuki

, YAP1 is a potent driver of the onset and progression of oral squamous cell carcinoma, Sci Adv 6 (2020), eaay3324.

Major

A.G.

Pitty

L.P.

and Farah

C.S.

, Cancer stem cell markers in head and neck squamous cell carcinoma, Stem Cells Int 2013 (2013), 319489.

Y.H.

W.S.

Lin

L.C.

Liu

C.S.

S.Y.

Wang

B.J.

Huang

B.M.

Yeh

Y.L.

Chiu

H.W.

Yang

W.L.

and Wang

Y.J.

, Bortezomib enhances radiosensitivity in oral cancer through inducing autophagy-mediated TRAF6 oncoprotein degradation, J Exp Clin Cancer Res 37 (2018), 91.

Maymone

M.B.C.

Greer

R.O.

Kesecker

Sahitya

P.C.

Burdine

L.K.

Cheng

A.D.

Maymone

A.C.

and Vashi

N.A.

, Premalignant and malignant oral mucosal lesions: Clinical and pathological findings, J Am Acad Dermatol 81 (2019), 59–71.

Sarode

S.C.

Sarode

G.S.

and Tupkari

J.V.

, Oral potentially malignant disorders: precising the definition, Oral Oncol 48 (2012), 759–760.

Mello

F.W.

Melo

Guerra

E.N.S.

Warnakulasuriya

Garnis

and Rivero

E.R.C.

, Oral potentially malignant disorders: A scoping review of prognostic biomarkers, Crit Rev Oncol Hematol 153 (2020), 102986.

Mortazavi

Safi

Baharvand

Jafari

Anbari

and Rahmani

, Oral white lesions: An updated clinical diagnostic decision tree, Dent J (Basel) 7 (2019).

Warnakulasuriya

and Ariyawardana

, Malignant transformation of oral leukoplakia: a systematic review of observational studies, J Oral Pathol Med 45 (2016), 155–166.

Gonzalez-Moles

M.A.

Ruiz-Avila

Gonzalez-Ruiz

Ayen

Gil-Montoya

J.A.

and Ramos-Garcia

, Malignant transformation risk of oral lichen planus: A systematic review and comprehensive meta-analysis, Oral Oncol 96 (2019), 121–130.

10.

Dionne

K.R.

Warnakulasuriya

Zain

R.B.

and Cheong

S.C.

, Potentially malignant disorders of the oral cavity: current practice and future directions in the clinic and laboratory, Int J Cancer 136 (2015), 503–515.

11.

Wetzel

S.L.

and Wollenberg

, Oral Potentially Malignant Disorders, Dent Clin North Am 64 (2020), 25–37.

12.

Mehrotra

and Gupta

D.K.

, Exciting new advances in oral cancer diagnosis: avenues to early detection, Head Neck Oncol 3 (2011), 33.

13.

Goossens

Nakagawa

Sun

and Hoshida

, Cancer biomarker discovery and validation, Transl Cancer Res 4 (2015), 256–269.

14.

Thomadaki

Helmerhorst

E.J.

Tian

Sun

Siqueira

W.L.

Walt

D.R.

and Oppenheim

F.G.

, Whole-saliva proteolysis and its impact on salivary diagnostics, J Dent Res 90 (2011), 1325–1330.

15.

Chojnowska

Baran

Wilinska

Sienicka

Cabaj-Wiater

and Knas

, Human saliva as a diagnostic material, Adv Med Sci 63 (2018), 185–191.

16.

Z.Z.

Wang

J.G.

and Zhang

X.L.

, Diagnostic model of saliva protein finger print analysis of patients with gastric cancer, World J Gastroenterol 15 (2009), 865–870.

17.

Rao

P.V.

Reddy

A.P.

Dasari

Krishnaprasad

Biggs

Roberts

C.T.

and Nagalla

S.R.

, Proteomic identification of salivary biomarkers of type-2 diabetes, J Proteome Res 8 (2009), 239–245.

18.

Streckfus

C.F.

Bigler

L.R.

and Zwick

, The use of surface-enhanced laser desorption/ionization time-of-flight mass spectrometry to detect putative breast cancer markers in saliva: a feasibility study, J Oral Pathol Med 35 (2006), 292–300.

19.

Abrao

A.L.

Falcao

D.P.

de Amorim

R.F.

Bezerra

A.C.

Pombeiro

G.A.

Guimaraes

L.J.

Fregni

Silva

L.P.

and da Mota

L.M.

, Salivary proteomics: A new adjuvant approach to the early diagnosis of familial juvenile systemic lupus erythematosus, Med Hypotheses 89 (2016), 97–100.

20.

Sannam Khan

Khurshid

Akhbar

and Faraz Moin

, Advances of Salivary Proteomics in Oral Squamous Cell Carcinoma (OSCC) Detection: An Update, Proteomes 4 (2016).

21.

Feng

Chen

Fan

Cui

Chen

Zhang

Jiang

Zhou

and Zhang

, Salivary protease spectrum biomarkers of oral cancer, Int J Oral Sci 11 (2019), 7.

22.

Khurshid

Zafar

M.S.

Khan

R.S.

Najeeb

Slowey

P.D.

and Rehman

I.U.

, Role of Salivary Biomarkers in Oral Cancer Detection, Adv Clin Chem 86 (2018), 23–70.

23.

Haigh

B.J.

Stewart

K.W.

Whelan

J.R.

Barnett

M.P.

Smolenski

G.A.

and Wheeler

T.T.

, Alterations in the salivary proteome associated with periodontitis, J Clin Periodontol 37 (2010), 241–247.

24.

Yang

T.Y.

Zhou

W.J.

S.T.

Yuan

W.W.

Luo

Zhang

J.H.

W.L.

Wang

X.Q.

Chen

Feng

Zhou

X.D.

and Zhang

, Role of saliva proteinase 3 in dental caries, Int J Oral Sci 7 (2015), 174–178.

25.

Khurshid

Warsi

Moin

S.F.

Slowey

P.D.

Latif

Zohaib

and Zafar

M.S.

, Biochemical analysis of oral fluids for disease detection, Adv Clin Chem 100 (2021), 205–253.

26.

Hamid

Adanir

Asiri

F.Y.I.

Abid

Zafar

M.S.

and Khurshid

, Salivary IgA as a Useful Biomarker for Dental Caries in Down’s Syndrome Patients: A Systematic Review and Meta-analysis, Eur J Dent 14 (2020), 665–671.

27.

Verhamme

I.M.

Leonard

S.E.

and Perkins

R.C.

, Proteases: Pivot Points in Functional Proteomics, Methods Mol Biol 1871 (2019), 313–392.

28.

Bond

J.S.

, Proteases: History, discovery, and roles in health and disease, J Biol Chem 294 (2019), 1643–1651.

29.

J.S.

Chen

Y.T.

Chiang

W.F.

Hsiao

Y.C.

Chu

L.J.

See

L.C.

C.S.

H.T.

Chen

H.W.

Chen

C.C.

Liao

W.C.

Chang

Y.T.

C.C.

Lin

C.Y.

Liu

S.Y.

Chiou

S.T.

Chia

S.L.

Chang

K.P.

Chien

C.Y.

Chang

S.W.

Chang

C.J.

Young

J.D.

Pao

C.C.

Chang

Y.S.

and Hartwell

L.H.

, Saliva protein biomarkers to detect oral squamous cell carcinoma in a high-risk population in Taiwan, Proc Natl Acad Sci U S A 113 (2016), 11549–11554.

30.

Cui

Chen

Luo

Wang

Chen

Kang

Wang

Zhou

Feng

and Zhang

, Protein kinase D3 regulates the expression of the immunosuppressive protein, PDL1, through STAT1/STAT3 signaling, Int J Oncol 56 (2020), 909–920.

31.

Wilson

J.L.

and Altman

R.B.

, Biomarkers: Delivering on the expectation of molecularly driven, quantitative health, Exp Biol Med (Maywood) 243 (2018), 313–322.

32.

Chiang

S.P.

Cabrera

R.M.

and Segall

J.E.

, Tumor cell intravasation, Am J Physiol Cell Physiol 311 (2016), C1–C14.

33.

Ricard-Blum

and Vallet

S.D.

, Proteases decode the extracellular matrix cryptome, Biochimie 122 (2016), 300–313.

34.

Kaczor-Urbanowicz

K.E.

Wei

Rao

S.L.

Kim

Shin

Cheng

Wong

D.T.W.

and Kim

, Clinical validity of saliva and novel technology for cancer detection, Biochim Biophys Acta Rev Cancer 1872 (2019), 49–59.

35.

Castagnola

Scarano

Passali

G.C.

Messana

Cabras

Iavarone

Di Cintio

Fiorita

De Corso

and Paludetti

, Salivary biomarkers and proteomics: future diagnostic and clinical utilities, Acta Otorhinolaryngol Ital 37 (2017), 94–101.

36.

Borgono

C.A.

and Diamandis

E.P.

, The emerging roles of human tissue kallikreins in cancer, Nat Rev Cancer 4 (2004), 876–890.

37.

Pettus

J.R.

Johnson

J.J.

Shi

Davis

J.W.

Koblinski

Ghosh

Liu

Ravosa

M.J.

Frazier

and Stack

M.S.

, Multiple kallikrein (KLK 5, 7, 8, and 10) expression in squamous cell carcinoma of the oral cavity, Histol Histopathol 24 (2009), 197–207.

38.

Jiang

Shi

Johnson

J.J.

Liu

and Stack

M.S.

, Kallikrein-5 promotes cleavage of desmoglein-1 and loss of cell-cell cohesion in oral squamous cell carcinoma, J Biol Chem 286 (2011), 9127–9135.

39.

Johnson

J.J.

Miller

D.L.

Jiang

Liu

Shi

Tarwater

Williams

Balsara

Sauter

E.R.

and Stack

M.S.

, Protease-activated Receptor-2 (PAR-2)-mediated Nf-kappaB Activation Suppresses Inflammation-associated Tumor Suppressor MicroRNAs in Oral Squamous Cell Carcinoma, J Biol Chem 291 (2016), 6936–6945.

40.

S.C.

Lin

C.W.

Yang

W.E.

Fan

W.L.

and Yang

S.F.

, The urokinase-type plasminogen activator (uPA) system as a biomarker and therapeutic target in human malignancies, Expert Opin Ther Targets 20 (2016), 551–566.

41.

Andreasen

P.A.

Kjoller

Christensen

and Duffy

M.J.

, The urokinase-type plasminogen activator system in cancer metastasis: a review, Int J Cancer 72 (1997), 1–22.

42.

McMahon

B.J.

and Kwaan

H.C.

, Components of the Plasminogen-Plasmin System as Biologic Markers for Cancer, Adv Exp Med Biol 867 (2015), 145–156.

43.

Magnussen

S.N.

Hadler-Olsen

Costea

D.E.

Berg

Jacobsen

C.C.

Mortensen

Salo

Martinez-Zubiaurre

Winberg

J.O.

Uhlin-Hansen

and Svineng

, Cleavage of the urokinase receptor (uPAR) on oral cancer cells: regulation by transforming growth factor – beta1 (TGF-beta1) and potential effects on migration and invasion, BMC Cancer 17 (2017), 350.

44.

Lescaille

Menashi

Cavelier-Balloy

Khayati

Quemener

Podgorniak

M.P.

Naimi

Calvo

Lebbe

and Mourah

, EMMPRIN/CD147 up-regulates urokinase-type plasminogen activator: implications in oral tumor progression, BMC Cancer 12 (2012), 115.

45.

Tian

Chen

Zhang

Wang

Han

Zhang

L.Y.

Nie

Zhao

Tan

Wang

Guan

X.Y.

and Hong

, Urokinase plasminogen activator secreted by cancer-associated fibroblasts induces tumor progression via PI3K/AKT and ERK signaling in esophageal squamous cell carcinoma, Oncotarget 8 (2017), 42300–42313.

Salivary KLK5 and uPA are potential biomarkers for malignant transformation of OLK and OLP

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Ethics statement

2.2 Saliva collection

2.3 Cell culture and supernatant collection

2.4 Protease array and image analysis

2.5 Quantitative assessment of salivary proteases by ELISA

2.6 Immunohistochemical analysis

2.7 Statistical analysis

Table 1 Coordinates of proteases located on the human protease array membrane

3.1 Saliva from OSCC and OLK patients contain more proteases than from healthy donors

5. Conclusions

Footnotes

Acknowledgments

Conflict of interest

Author contributions

References

Table 1
Coordinates of proteases located on the human protease array membrane