Evaluation of variations in urine laminin-γ2 monomer in non-muscle-invasive bladder cancer before and after transurethral resection of the bladder tumor

Abstract

Background

We have previously reported urine laminin-γ2 monomer (Lm-γ2m) as a diagnostic biomarker for patients with non-muscle-invasive bladder cancer (NMIBC). However, Lm-γ2m has not been evaluated as a prognostic biomarker for patients with NMIBC after transurethral resection of bladder tumor (TURBT). This study aimed to clarify whether urine Lm-γ2m offers a useful prognostic biomarker for patients with NMIBC after TURBT.

Methods

Participants comprised 41 patients with NMIBC. Urine Lm-γ2m was prospectively measured before and 3 months after TURBT and values were compared. Traditional urine tumor markers such as urine nuclear matrix protein 22 and cytology were also evaluated before and after TURBT. Patients were stratified by the presence of recurrence within 1 year after TURBT and values of urine Lm-γ2m before and after TURBT were analyzed.

Results

Mean urine Lm-γ2m values were significantly decreased after TURBT compared to before surgery. Mean urine Lm-γ2m before and after TURBT were higher in patients with recurrence within 1 year after surgery than in patients without recurrence. Urine Lm-γ2m before TURBT was also important to predict prognosis in patients with NMIBC treated using TURBT.

Conclusions

These results suggest urine Lm-γ2m as a potentially useful biomarker for NMIBC before and after TURBT.

Introduction

Around 70% of bladder cancers are non-muscle-invasive bladder cancers (NMIBCs), which are initially treated by endoscopic surgery and show good prognosis with bladder preservation [1]. However, the clinical features of NMIBC include frequent recurrence with temporal and spatial multiplicity after complete endoscopic resection [2]. Patients with NMIBC must therefore undergo repeated invasive examinations, including cystoscopy every 3 months for at least 2–3 years after endoscopic surgery, followed by repeated surgeries for recurrences, placing a significant burden on patients. Identification of biomarkers with high diagnostic accuracy for use at initial diagnosis and diagnosis of NMIBC recurrence is therefore urgently needed.

Laminin (Lm)−332 comprises three polypeptides, the Lm-α3, -β3 and -γ2 chains encoded by the LAMA3, LAMB3, and LAMC2 genes, respectively, and acts as an adhesive molecule in the basement membrane of normal epithelia [3]. Lm-γ2 expressed as a monomer forming in tumor cells and tissues has recently been reported to play a crucial role in the malignant progression of tumor cell processes such as growth, survival, motility, invasion and metastasis [4]. We have previously developed a fully automated chemiluminescent immunoassay using a specific monoclonal antibody against Lm-γ2 monomer (Lm-γ2m) and reported that measurement of Lm-γ2m in urine might provide an initial diagnostic biomarker, even in patients with low-grade NMIBC [5–7]. However, whether measurement of urine Lm-γ2m can be used as a biomarker in follow-up after transurethral resection of bladder tumor (TURBT) in patients with NMIBC has not been clarified.

In this study, we measured urine Lm-γ2m in patients with NMIBC before and after TURBT to prospectively investigate whether levels of urine Lm-γ2m decreased postoperatively compared to preoperatively and offered a predictive marker of intravesical recurrence after surgery.

Materials and methods

Study design and population

This prospective study enrolled 68 bladder cancer patients treated in the Department of Urology at Kochi Medical School Hospital or the Department of Urology at Kanagawa Cancer Center between February 2016 and February 2017 who provided urine for the study. All patients with bladder cancer showed clinically non-muscle-invasive tumors and N0 and M0 stage. All patients provided written informed consent, and the study protocol was approved by the review boards of both institutions (Kochi Medical School Hospital: ERB-106289; Kanagawa Cancer Center: 2019–22). This study was conducted in compliance with the Japanese ethical guidelines for medical and health research involving human subjects and the Declaration of Helsinki (revised in Brazil in 2013) [8]. Sixty-eight patients with bladder cancer were enrolled and 8 patients were excluded based on the exclusion criteria detailed below. Of the 60 patients in the full analysis set, seven patients each were excluded on the decision of an attending physician or due to missing values for urine markers, respectively; four patients were pathologically diagnosed with muscle-invasive bladder cancer after TURBT; and one patient died due to another cause. Finally, a total of 41 patients completed all analyses including urine Lm-γ2m/uCRN, NMP22 and urine cytology both before and three months after TURBT. Characteristics of the 41 patients with NMIBC evaluated are summarized in Table 1. Among the 41 NMIBC patients, median age was 72 years (range, 38–89 years). Pathological T stage was Ta in 75%, T1 in 15%, and Tis in 10%. Tumor grades were low in 27% and high in 73%. Tumor diameter was 3 cm or more in 4 patients (10%), and two or more tumors were present in 19 patients (46%). After TURBT, 25 patients received intravesical instillation of a single anthracycline-based chemotherapeutic and 11 patients received intravesical instillation of bacille de Calmette et Guérin (BCG). Six patients underwent a 2nd TURBT. None of the patients received systemic chemotherapy.

Table 1.

Characteristics of evaluated patients with NMIBC

Non-muscle invasive bladder cancer		Median (range) or n (%)
Total number		41
Median age		72 (38–89) years
Sex	male	33 (80)
	female	8 (20)
pT stage	Ta	31 (75)
	T1	6 (15)
	Tis	4 (10) *
	Tx	2 (5)
Grade	low	11 (27)
	high	30 (73)
Tumor size	≥ 3 cm	4 (10)
	< 3 cm	37 (90)
Tumor number	≥ 2	19 (46)
	< 2	22 (54)
Intravesical instillations	Anthracycline-based chemotherapeutics	25 (61)
	BCG	11 (27)
2nd TURBT		6 (15)

NMIBC non-muscle invasive bladder cancer, TURBT transurethral resection of bladder tumor, Ln-γ2m/uCRN laminin-γ2 monomer/urine creatinine, BCG bacille de Calmette et Guérin

^*Including two cases with carcinoma in situ only

Inclusion and exclusion criteria

Male and female patients between 20 and 90 years of age who met the following criteria were enrolled. Type of participant and target disease characteristics were primary non-muscle-invasive urothelial cancer in the bladder treated with TURBT. We excluded anyone characterized with the following conditions: 1) planned radical cystectomy; 2) recurrence of a tumor of the bladder within six months of previous TURBT; 3) history of prior active malignancy other than bladder cancer within the last 5 years; 4) suspected malignancy other than bladder cancer; 5) serious acute inflammatory diseases; 6) serious abnormalities in liver or kidney function; 7) lack of written informed consent; or 8) otherwise deemed by the physician in charge as unsuitable for inclusion in the study.

Endpoint

The primary endpoint of the study was the intra-individual variability of observed urine Lm-γ2m levels between before and three months after TURBT.

Clinical urine specimens

Briefly, 50 mL of self-voided urine specimen was centrifuged for 5 min at 1000 rpm after collection from the patient, then stored at −80 °C for 1–6 months. Specimens (250 μL) were then subjected to automated chemiluminescent immunoassay after centrifugation for 10 min at 15,000 rpm at 4 °C [6, 7]. Urine cytology (5 mL of 20 × concentrated urine), creatinine (5 mL), and NMP-22 (5 mL) were dispensed into the tubes designated by SRL. Freeze–thaw samples were centrifuged for 10 min at 10,000 × g before use for measurement.

Detection of urine Lm-γ2m by chemiluminescent immunoassay

Urine Lm-γ2m was determined by fully automated chemiluminescent immunoassay (CLIA) using the ARCHITECT system (Abbott Laboratories, North Chicago, IL), which employs a sandwich method with magnetic particles conjugated with a monoclonal antibody (2H2 mAb) that selectively recognizes Lm-γ2m and a rabbit polyclonal antibody recognizing acridinium-labeled Lm-γ2m. This sandwich CLIA can quantitatively measure urine Lm-γ2m with a concentration range of 0–20,000 pg/mL. Urine Lm-γ2m values were normalized by dividing by urine creatinine (uCRN), yielding urine Lm-γ2m/uCRN (ng/g·CRN × 100) [5, 6].

Urine cytology

Urine cytology was classified according to the Papanicolaou classification system and sorted into 7 classifications: 1) absence of atypical or abnormal cells; 2) atypical cytology, but no evidence of malignancy; 3a) probable benign atypia; 3) cytology suggestive of, but not conclusive for, malignancy; 3b) suspected malignancy; 4) cytology strongly suggestive of malignancy; and 5) cytology conclusive for malignancy [9]. In the present study, classes IV and V were grouped together as positive for malignancy.

Follow-up schedule

Patients were followed-up with the same schedule. Urine cytology and cystoscopy were performed preoperatively and at 3, 6, and 12 months postoperatively. The study was terminated when recurrence was confirmed. Cystoscopy-confirmed presence or absence of cancer or cancer recurrence, pathological stage, presence of carcinoma in situ (CIS), tumor grade, size and number were recorded. Radiographic imaging was routinely performed before and after TURBT to rule out the presence of upper urinary tract cancer.

Data management

All clinical and examination data were sent to the Data Center at the Integrated Center for Advanced Medical Technologies, Kochi Medical School using an Electronic Data Capture system. All data were managed by the data manager in the Data Center, then analyzed by an independent statistician.

Statistical analysis

Statistical analysis was performed using R version 3.5.2 (The R Foundation, Vienna, Austria). Urine Lm-γ2m/uCRN before and after TURBT was converted to logit, histograms were created, and t-tests were conducted. Univariate logistic regression analysis was performed for recurrence predictors in patients with NMIBC after TURBT. The optimal number of clusters in patients with NMIBC who had undergone TURBT was derived from plotted within-cluster sum-of-squares values. The k-means clustering used Euclidean distance, the Hartigan–Wong algorithm, and 10 iterations. The pROC and dcurves packages (The R Foundation) were used for analysis of receiver operating characteristic (ROC) curves and decision curve analysis, respectively. The predictive performances of urine Ln-γ2m/uCRN before and after TURBT were evaluated using ROC analysis, area under the ROC curve (AUC), and corresponding 95% confidence intervals (CIs) [10]. Diagnostic values were evaluated based on AUCs, with sensitivity and specificity evaluated at cut-off values, calculated as the maximum Youden index [11]. The influences of intravesical instillation of anthracycline-based chemotherapeutics and BCG, and 2nd TURBT, for decreasing urine Ln-γ2m/uCRN were evaluated by multivariate linear regression [12]. Values of P < 0.05 were considered statistically significant.

Results

Decreasing urine Lm-γ2m/uCRN values before and 3 months after TURBT in patients with NMIBC

Urine Lm-γ2m/uCRN was measured before and 3 months after complete resection of tumors. The mean (± standard error of the mean) postoperative log-transformed value of urine Lm-γ2m/uCRN (1.26 ± 0.21) was significantly lower than the preoperative value (−0.65 ± 0.23; P = 0.002 (Fig. 1). Actual mean pre- and postoperative urine values of Lm-γ2m/uCRN were 2.07 ± 1.50 and 0.36 ± 0.73 ng/g·CRN × 100.

Fig. 1.

Log-transformed values of urine Lm-γ2m/uCRN before and 3 months after TURBT. Median log values of urine Lm-γ2m/uCRN before and after surgery were −0.65 ± 0.23 and −1.26 ± 0.21, respectively. Urine Lm-γ2m/uCRN after TURBT was significantly lower than that before surgery (P = 0.002)

Change in urine Lm-γ2m/uCRN values before and 3 months after TURBT in patients with and without recurrence within 1 year after surgery.

Preoperative urine Lm-γ2m/uCRN was significantly higher in cases with recurrence within 1 year after surgery than in those without recurrence. Although no statistical difference was identified, postoperative urine Lm-γ2m/uCRN was also higher in cases with recurrence within 1 year after surgery than in those without recurrence. Urine Lm-γ2m/uCRN decreased after surgery in both cases with and without recurrence (Fig. 2).

Fig. 2.

Log-transformed values of urine Lm-γ2m/uCRN in patients with and without recurrence of NMIBC before and 3 months after TURBT. Log values of urine Lm-γ2m/uCRN both before and after surgery in patients with recurrence within 1 year after surgery were higher than those in patients without recurrence (−0.19 ± 0.65 versus −0.91 ± 0.21 before surgery and −0.87 ± 0.63 versus −1.39 ± 0.19 after surgery, respectively). Odds ratios for urine Lm-γ2m/uCRN in patients with and without recurrence within 1 year after TURBT were 1.71 (95% confidence interval, 1.03–3.12) before surgery and 1.38 (95% confidence interval, 0.79–2.69) after surgery

Predictive factors for NMIBC recurrence in the bladder within 1 year from univariate analysis.

We analyzed predictive factors including urine Lm-γ2m/uCRN, NMP22 and cytology (pathological T stage; presence of CIS; grade; size; and number of tumors) using univariate analyses. Odds ratios, 95%CIs and P-values are shown in Table 2. No patients showed positive urine cytology after TURBT (≥ class 4).

Table 2.

Univariate logistic regression analysis of recurrence predictors within 1 year after TURBT in patients with NMIBC

Predictors	Odds ratio (95%CI)	P-values
Urine Ln-γ2m/uCRN before TURBT	1.71 (1.03–3.12)	0.05
Urine Ln-γ2m/uCRN 3 months after TURBT	1.38 (0.79–2.69)	0.29
Urine NMP22 before TURBT	1.73 (1.11–2.91)	0.02
Urine NMP22 3 months after TURBT	1.41 (0.81–2.62)	0.23
Urine cytology before TURBT	0.00 (NA–NA)	0.99
Urine cytology 3 months after TURBT ≥ class 4 (reference 1, 2, 3, 3a, 3b)	-*	-*
pT1 (reference pTa, pTis, pTx)	2.56 (0.52–1.23)	0.26
With CIS	0.00 (NA–NA)	0.99
High-grade (reference Low-grade)	1.64 (0.33–12.26)	0.58
Tumor size ≥ 3 cm	3.63(0.34–34.39)	0.23
Tumor number ≥ 2	3.45 (0.75–19.10)	0.09

Ln-γ2m/uCRN laminin-γ2 monomer/urine creatinine, TURBT transurethral resection of bladder tumor, NMIBC non-muscle invasive bladder cancer, CI confidence interval, NMP nuclear matrix protein, CIS carcinoma in situ, NA not applicable

^*No cases

Clustering analysis using pre- and postoperative urine Lm-γ2m/uCRN in patients with or without recurrence within 1 year after TURBT.

The elbow method was used to determine the number of clusters from data and the optimal number of clusters was estimated to be two (Fig. 3a). Patients with high urine Lm-γ2m/uCRN before and after surgery, such as those in cluster group 1, showed a 44% rate of recurrence even if bladder cancers were completely resected. Patients in cluster group 2 who showed low urine Lm-γ2m/uCRN before and after surgery only showed a 12% rate of recurrence. Two patients with low urine Lm-γ2m/uCRN both before and after surgery experienced recurrence within 1 year in cluster group 1 (Fig. 3b).

Fig. 3.

The optimal number of clusters based on log-transformed values of urine Lm-γ2m/uCRN in patients with NMIBC who had undergone TURBT. Within-cluster sum of squares (WSS) values are plotted against the number of clusters, and the point at which a distinct "elbow “ or inflection occurs is identified. The optimal number of clusters was determined to be 2 (arrow ↓) based on the criterion that the "elbow" indicates the optimal number of clusters (a). Cluster group 1 is colored orange. Patients without recurrence within 1 year are shown with orange circles (●) and those with recurrence are shown with orange triangles (▲). Cluster group 2 is colored blue-green. Patients without recurrence within 1 year are shown with blue-green circles (●) and those with recurrence are shown with blue-green triangles (▲). Patients with high values for both pre- and postoperative urine Lm-γ2m/uCRN showed a greater risk of recurrence within 1 year after surgery in cluster group 2 (▲). Patients with low values for both pre- and postoperative urine Lm-γ2m/uCRN were less likely to experience recurrence within 1 year after surgery in cluster group 1 (●). A few exceptional cases did not reveal a relationship between recurrence and urine Lm-γ2m/uCRN (▲*, b)

Impact of intravesical instillation therapies and 2nd TURBT on TURBT-associated decrease in urine Lm-γ2m/uCRN.

We evaluated whether intravesical instillation of anthracycline-based chemotherapeutics and BCG, and 2nd TURBT affect the decrease in urine Lm-γ2m/uCRN after TURBT using multivariate linear regression analysis. Neither intravesical instillation therapies nor 2nd TURBT exerted significant impacts on the TURBT-associated decrease in urine Lm-γ2m/uCRN (Table 3).

Table 3.

Multivariate linear regression analysis associated with additional therapies after TURBT and difference in urine Lm-γ2m/uCRN between before and 3 months after the surgery

Additional therapies	Coefficient	95%CI	P value
1 st TURBT	0.95	0.37–1.57	< 0.01
Intravesical instillation of anthracycline-based chemotherapeutics	−0.44	−1.76–0.88	0.51
Intravesical instillation of BCG	−0.36	−1.15–0.42	0.36
2nd TURBT	−0.33	−1.25–0.59	0.47

TURBT transurethral resection of bladder tumor, Ln-γ2m/uCRN laminin-γ2 monomer/urine creatinine, CI confidence interval, BCG bacille de Calmette et Guérin

Diagnostic accuracies of urine Lm‐γ2m/CRN before and after TURBT

We evaluated the diagnostic accuracies of urine Lm-γ2m/uCRN before and 3 months after TURBT using ROC curve analysis. AUCs before and after TURBT were 0.69 (95%CI, 0.45–0.94) and 0.66 (95%CI, 0.42–0.90), respectively. Cut-off values for log urine Lm-γ2m/uCRN before and after TURBT were 0.33 and 0.29, demonstrating 60% and 50% sensitivity, and 90% and 90% specificity before and after TURBT, respectively. Actual measurements of urine Lm-γ2m/uCRN before and after TURBT were 1.390 and 1.336, respectively (Fig. 4).

Fig. 4.

Receiver operating characteristic (ROC) curves for urine Ln-γ2m/uCRN before (grayish line) and 3 months after (dashed black line) TURBT in patients with NMIBC. Areas under the ROC curves for urine Ln-γ2m/uCRN before and after TURBT are 0.69 (0.45–0.94) and 0.66 (0.42–0.90), respectively. Cut-off values derived from ROC curve analysis corresponding to the maximum Youden Index are 0.33 (●: grayish circle) and 0.29 (●: black circle), and demonstrated 60% and 50% sensitivity, and 90% and 90% specificity before and after TURBT, respectively

Discussion

Our previous findings demonstrated that measurement of urine Lm-γ2m/uCRN potentially represents a useful diagnostic marker for detecting NMIBC. The present prospective study demonstrated that urine Lm-γ2m/uCRN decreased following TURBT in NMIBC patients without being affected by intravesical instillation of anthracycline-based chemotherapeutics or BCG, or 2nd TURBT. Further, patients with high urine Lm-γ2m/uCRN both before and after surgery might be more likely to experience recurrence within 1 year after surgery. The data suggest that measurement of urine Lm-γ2m/uCRN might provide a prognostic marker for patients with NMIBC before and after TURBT.

Although Lm-γ2 represents a major component of the basement membrane proteins along with other laminin chains, Lm-γ2 as a monomer (i.e., Lm-γ2m) is involved in the malignant progression of tumor cells. Lm-γ2m acts as a paracrine epidermal growth factor-like growth factor through the action of proteolytic cleavage by matrix metalloproteinases [13–15]. Thus, urine Lm-γ2m might be not only a diagnostic, but also a biological and predictive marker of NMIBC. The present data demonstrated that urine Lm-γ2m/uCRN decreased significantly following surgical tumor resection, with high values of urine Lm-γ2m/uCRN when tumors were present and low values when tumors were absent after surgery. Urine Lm-γ2m/uCRN thus appears useful as a biomarker before and after the surgery in patients with NMIBC (Fig. 1). We evaluated whether urine Lm-γ2m/uCRN, NMP22, cytology, pathological T stage, carcinoma in situ, grade, and tumor size and number predicted recurrences within 1 year after TURBT using univariate logistic regression analysis. The odds ratio for urine Lm-γ2m/uCRN before TURBT was 1.71 (95%CI, 1.03–3.12), although that after TURBT was 1.38 (95%CI, 0.79–2.69). Preoperative urine Lm-γ2m/uCRN was the most powerful predictive factor (Table 2). Multivariate analysis was not performed because the number of patients was too few. The odds ratio for urine NMP22 before TURBT was 1.73 (95%CI, 1.11–2.91; P = 0.02) (Table 2). In the literature, initial level of NMP22 was not identified as a predictor of recurrence during long-term follow-up. Limitations of NMP-22 for detecting residual tumors before second TURBT in patients with NMIBC have also been reported [16]. While negative results for urine NMP22 have been reported, NMP22 offered better results than expected in the present study, with urine NMP22 levels along with urine Lm-γ2m/uCRN after TURBT being significantly decreased compared to levels before surgery (Supplementary Material 1) [17–19]. From our results, urine NMP22 must be re-evaluated as a urine marker for patients with NMIBC in clinical situations.

We also performed cluster analysis using the elbow method to determine the optimal number of clusters, using pre- and postoperative urine Lm-γ2m/uCRN in patients with or without recurrence within 1 year after TURBT. Cluster group 1 represented patients with low pre- and postoperative urine Lm-γ2m/uCRN, who were not more likely to develop recurrence within 1 year after surgery. Cluster group 2 presented patients with high pre- and postoperative urine Lm-γ2m/uCRN, who were more likely to experience recurrence within 1 year after surgery (Fig. 3b). However, two patients with recurrence of NMIBC within 1 year presented with low urine Lm-γ2m/uCRN both before and after surgery, and these patients would have required further investigation of potential Lm-γ2m-independent mechanisms of recurrence.

With the decrease in urine Lm-γ2m/uCRN after TURBT compared to before, cut-off values declined accordingly. Urine Lm-γ2m/uCRN was the indicator offering the best specificity before and after TURBT. Taken together, these findings suggested that measurement of urine Lm-γ2m/uCRN might be useful for assessing postoperative recurrence.

Limitations of the present study were the small sample size and short follow-up period of 1 year after surgery.

In conclusion, findings from the present investigation suggest that pre- and postoperative measurement of urine Lm-γ2m/uCRN is potentially useful for predicting NMIBC recurrence after TURBT. Large-scale prospective clinical studies involving a greater number of patients are warranted in future investigations.

Abbreviations

Lm-γ2m

Laminin-γ2 monomer

NMIBC

Non-muscle-invasive bladder cancer

TURBT

Transurethral resection of bladder tumor

Lm

Laminin

CLIA

Chemiluminescent immunoassay

uCRN

Urine creatinine

BCG

Bacille de Calmette et Guérin

ROC

Receiver operating characteristic

AUC

Area under the ROC curve

CI

Confidence interval

Authors’ contributions

K. Osaka contributed to the concept and drafted the manuscript. T. Kishida and S. Umemoto collected specimens and cared for the patients. T. Karashima contributed to the concept, drafted the manuscript and approved the final version of the manuscript. M. Nakagawa technically supported the measurement of Lm-γ2m. E. Yoshida and T. Yoshimura established the method for measuring Lm-γ2m. M. Sakaguchi managed and analyzed the data. K. Inoue, M. Seiki, N. Koshikawa and T. Shuin contributed to the concept and design of the study. All authors read and approved the final manuscript.

Funding

The present study was supported by a Translational Research Network Program Seeds B grant (to T. Shuin, M. Seiki and N. Koshikawa) by the Japan Agency for Medical Research and Development (AMED) under Grant Number 5lm0103002j0004 (B29), and an Adaptable & Seamless Technology Transfer Program through Target-driven R&D (A-STEP) grant (to N. Koshikawa, M. Seiki and T. Shuin) by AMED under Grant Number 17im0302428h0104. Research funding from Abbott Laboratories (North Chicago, IL) was provided to N. Koshikawa and T. Shuin.

Data availability

Requests for the study materials and dataset used to support the conclusions of this article should be directed to the corresponding author.

Declarations

Ethics approval and consent to participate

The study protocol was approved by our institutional review board (Kochi Medical School Hospital: ERB-106289; Kanagawa Cancer Center: 2019–22).

Consent for publication

All patients provided written informed consent and consented to publication.

Competing interests

T. Yoshimura, E. Yoshida and M. Nakagawa are employees of Abbott Japan LLC. K. Inoue received patent royalties and research funds from SBI Pharmaceuticals Co., Ltd., honoraria from Chugai Pharmaceutical Co., Ltd., and endowments from Astellas Pharma Inc., Takeda Pharmaceutical Co., Ltd., MC Medical Inc. and Ono Pharmaceutical Co., Ltd. M. Seiki and N. Koshikawa received research funds from Tokyo University. N. Koshikawa, T. Karashima and T. Shuin received research funds from Abbott Japan LLC. None of the other authors have any potential conflicts of interest.