Overexpression of high mobility group box 1 with poor prognosis in patients after radical prostatectomy


  • Tao Li and Yaping Gui contributed equally to the present study.

Tao Li, Tongji Hospital, Shanghai Tongji University School of Medicine – Department of Urology, 389 Xin Cun Road, Shanghai 200065, China. e-mail: quicktao@gmail.com


What's known on the subject? and What does the study add?

Recent studies have indicated that high mobility group box 1 (HMGB1) is related to the development and progression of human carcinomas. However, further studies were required to confirm the roles played by HMGB1 in clinical prostate cancer treatment.

We investigated the relationship between HMGB1 expression and the characteristics of prostate cancer, and also evaluated the significance of HMGB1 as a prognostic factor for biochemical recurrence-free survival after radical prostatectomy.


  • • To investigate high mobility group box 1 (HMGB1) expression in human prostate cancer (PC) cell lines and its prognostic significance after radical prostatectomy (RP).


  • • Quantitative reverse-transcription polymerase chain reaction and western blotting were used to detect HMGB1 mRNA and protein expression in PC cell lines.
  • • Immunohistochemistry coupled with the tissue microarray technique was performed to evaluate HMGB1 protein expression in 168 primary prostatectomy tissue samples.
  • • Clinicopathological features were compared between positive and negative HMGB1 protein expression groups.
  • • Kaplan–Meier and multivariate Cox analyses were applied to determine the prognostic value of HMGB1 protein expression on biochemical recurrence (BCR) for patients with PC who were undergoing RP.


  • • There were three PC cells (DU145, PC-3 and LNCaP) with overexpression of HMGB1 mRNA and protein compared to the non-transformed immortalized prostate cell RWPE-1.
  • • A total of 60.1% (101/168) of the PC samples appeared to have positive protein expression of HMGB1.
  • • HMGB1 protein expression was correlated with some clinicopathological parameters, such as pathological stage (pT) (P= 0.011), Gleason score, preoperative prostate-specific antigen concentration and BCR (P < 0.001, respectively).
  • • Positive HMGB1 immunostaining in patients with PC who were undergoing RP was significantly associated with poor median BCR-free survival (23.1 months vs 15.6 months) (P < 0.001).
  • • Multivariate analysis indicated that HMGB1 protein expression was an independent prognostic factor for BCR-free survival after RP (hazard ratio = 2.348, 95% confidence interval = 1.373–6.361, P= 0.001).


  • • Up-regulation of HMGB1 mRNA and protein concentrations was confirmed in PC cells.
  • • HMGB1 expression may contribute to the malignant progression of PC.
  • • HMGB1 presents as a novel prognostic factor for BCR after RP.

biochemical recurrence


glyceraldehyde 3-phosphate dehydrogenase


high mobility group box 1


prostate cancer


radical prostatectomy


quantitative RT-PCR


tissue microarray.


Prostate cancer (PC) is the most common genitourinary malignancy in men worldwide, and the incidence of the disease is gradually increasing annually [1]. More than 90% of PC cases are diagnosed as either localized or locally advanced disease [2]. To date, radical prostatectomy (RP) is the gold standard treatment for localized PC [3]. However, ≈25–50% post-RP patients on long-term follow-up experience biochemical recurrence (BCR), which is defined as a postoperative PSA concentration ≥ 0.2 ng/mL [4]. This is generally the earliest indicator of a recurrence [5].

High mobility group protein box 1 (HMGB1) is a non-histone chromatin protein and is highly conserved between species [6]. As a nuclear protein, HMGB1 regulates the transcription of many genes in the nucleus by promoting the formation of nucleoprotein complex [7]. Under certain conditions, HMGB1 is also present in the cytoplasm or secreted outside the cell to function as a signal molecule [8–10]. Recent evidence indicates that increased expression of HMGB1 is associated with hallmarks of cancer, such as proliferation, apoptosis, angiogenesis and metastasis. HMGB1 plays a critical role in the development and progression of many types of tumours [11].

Recent work by Gnanasekar et al. [10] has shown that the expression level of HMGB1 is up-regulated in PC cell lines [10]. Using an RNA interference approach, they also found that reducing HMGB1 expression levels led to a decrease in cellular viability. In the present study, we analyzed the expression of HMGB1 at cellular and tissue levels and investigated its clinicopathological and prognostic significance in patients with PC who were undergoing RP with a median long follow-up.



Tumour samples from 168 patients with PC were included in the present study. The patients undergong RP and regional lymph node dissection between January 2001 and July 2010 were from Tongji Hospital (a subsidiary of Shanghai Tongji University); they did not receive any treatment before surgery. The histopathological features of tumour specimens were classified according to the Gleason score system and the 2002 TNM classification system. Normal prostate tissue was obtained from patients with bladder cancer after total cystectomy in Tongji Hospital. The Research Ethics Committee of Tongji Hospital approved the protocol and verbal consent was obtained from all patients.


The Gleason scores of the PC samples were re-evaluated by two pathologists before the construction of the prostate TMA. The TMA was constructed with formalin-fixed, paraffin-embedded prostate tissue samples, and the areas of the invasive adenocarcinoma were identified in accordance with the corresponding haematoxylin and eosin stained slides. In a highly representative fashion, two replicate tumour samples (diameter 1 mm) were taken from the donor tissue blocks and arrayed into a recipient paraffin block (35 × 622 × 65 mm) using a tissue microarrayer (Beecher Instrument Inc., Sun Prairie, WI, USA), as described by Kononen et al. [12].


DU145, PC-3 and LNCaP cells were obtained from the Cell Bank, China Academy of Science (Shanghai, China). Cells were cultured in Dulbecco's modified Eagle's medium and RPMI-1640 medium (Invitrogen, Carlsbad, CA, USA), respectively, and supplemented with 10% fetal bovine serum. RWPE-1 cells were purchased from the ATCC. The RWPE-1 cell line was cultured in keratinocyte-serum free medium (catalogue number 10724-011; Invitrogen) supplemented with epidermal growth factor and bovine pituitary extract. No 10% fetal bovine serum was added to RWPE-1 cell lines. All cells were maintained at 37 °C in 5% CO2.


Total RNA was extracted from RWPE-1, DU145, PC-3 and LNCaP cells with Trizol reagent (Invitrogen) in accordance with the manufacturer's instructions. Total RNA (1 µg) was denatured at 65 °C for 5 min and then RT reactions were performed using a cDNA synthesis kit (Takara Bio Inc., Otsu, Japan) in accordance with the manufacturer's instructions. Real-time PCR was performed using SYBR® Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA) on a 7500 Fast Real-Time PCR System (Applied Biosystems). The primers for HMGB1 were: forward, 5′-AGAGCGGAGAGAGTGAGGAG-3′; reverse, 5′-GATCTCCTTTGCCCATGTTT-3′. The primers for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were: forward, 5′-TGACGCTGGGGCTGGCATTG-3′; reverse, 5′-GCTCTTGCTGGGGCTGGTGG-3′ (synthesized by Sangon Biotech, Shanghai, China). To estimate a difference in the expression of HMGB1, Ct values were normalized using GAPDH as an internal control. The relative mRNA expression was calculated using the 2–ΔΔCt method. The qRT-PCR experiments were repeated three times.


RWPE-1, DU145, PC-3 and LNCaP cells were lyzed in 1 × SDS lysis buffer (50 mm Tris–HCl, pH 6.8, 100 mm dithiothreitol, 2% SDS, 0.1% bromphenol blue, 10% glycerol) and boiled for 5–10 min. Standard western blotting analyses were performed to measure the GAPDH (Kangcheng, Shanghai, China). Secondary antibodies were purchased from Sigma-Aldrich (St Louis, MO, USA). The blot was developed with SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Rockford, IL, USA) and scanned with a LAS4000 mini Luminescent Image Analyzer (Fujifilm, Tokyo, Japan).


A specific rabbit polyclonal antibody was used to detect HMGB1 (Epitomics, Burlingame, CA, USA). The antibody was tested and optimized on the whole-tissue sections and test arrays. Once an appropriate dilution and incubation time was determined, two tissue array sections containing all patient samples were stained for HMGB1 via standard two-step immunohistochemistry.

The slides briefly were rehydrated in the graded alcohols after deparaffinization in xylene. The endogenous peroxidase was quenched with 0.3% hydrogen peroxide in methanol at room temperature (24 °C), and the sections were placed in a 0.01 M sodium citrate buffer (pH 6.0) at 120 °C for antigen retrieval. The primary antibody of the HMGB1 was applied at 1:100 dilution. The Envision System (Dako, Glostrup, Denmark) was used as the second antibody. Sections were visualized with 3,3′-diaminobenzidine, counterstained with haematoxylin, dehydrated, and mounted. Identical array sections stained in the absence of the primary antibody served as negative controls.

For evaluation of staining, a semi-quantitative assessment of the antibody staining on the TMAs was performed independently by two observers who were blinded to all clinicopathological variables. Each TMA spot was scanned to assign the scores. HMGB1 expression was estimated as the total HMGB1 immunostaining score, which was calculated as the sum of an extent of staining score and an intensity score. The extent of staining score reflects the fraction of positive staining cells (score 0, 0%; score 1, 1–25%; score 2, 26–50%; score 3, 51–75%; score 4, 76–100%). The intensity score represents the staining intensity (score 0, no staining signal; score 1, weak positive signal; score 2, moderate positive signal; score 3, strong positive signal). The sum of the intensity and the extent of staining scores was used as the final staining score (0–7) for HMGB1. For the purpose of statistical evaluation, tumours with a final staining score of ≥3 were considered as positive. This relatively simple, reproducible scoring method gives highly concordant results between independent evaluators and has been used in previous studies [13].


The BCR-free survival time was defined as the time from the date of surgery to that of the BCR, and the follow-up data were updated in July 2010. Any difference in the time to BCR of the PC after RP could be estimated in accordance with the Kaplan–Meier method. Using the log-rank test, we could analyze differences in survival. To investigate the relationship between HMGB1 expression and clinicopathological characteristics, the t-test, chi-squared analysis, the Mann–Whitney U-test and Fisher's exact test were employed in accordance with the test condition. A Cox proportional hazards model was performed to establish independent factor(s) for survival. P < 0.05 (two-sided) was considered statistically significant. Statistical analysis was performed using SPSS, version 14.0 (SPSS Inc., Chicago, IL, USA).



The transcriptional and translational expression of HMGB1 was evaluated in human PC cell lines DU145, PC-3 and LNCaP by qRT-PCR and western bloting, respectively. Compared with the non-transformed immortalized prostate cell RWPE-1, all three PC cells overexpressed HMGB1 at both the mRNA and protein level. Among three PC cell lines, DU145, a hormone-independent PC cell, showed the strongest HMGB1 expression (Fig. 1).

Figure 1.

Expression of high mobility group box 1 (HMGB1) mRNA and protein in prostate cancer cell lines. A, Real-time quantitative RT-PCR analysis of HMGB1 mRNA expression. B, Western blot analysis of HMGB1 protein expression in prostate cancer cell lines DU145, PC-3 and LNCap. The non-transformed immortalized prostate cell RWPE-1 was used as a control.


Immunochemistry was performed to analyze the protein expression of HMGB1 in 168 tumour samples from patients who were undergoing RP. Figure 2 shows the representative immunostaining of HMGB1 protein on tissue arrays. Compared with no HMGB1 protein staining in the normal prostate glandular epithelium (Fig. 2A,D), the protein expression of HMGB1 was detected mainly in the nucleus (Fig. 2B,E). However, in a few tumour samples, positive staining was found in the nucleus and cytoplasm (Fig. 2C,F). In total, 60.1% (101/168) of the PC samples appeared to show a positive protein expression of HMGB1. Clinicopathological features were compared between positive and negative HMGB1 protein expression groups. Among these 168 patients, there were nine who had lost their clinical data, including pathological stage (pT), lymph node status and capsular invasion. Using the t-test, chi-squared analysis, the Mann–Whitney U-test and Fisher's exact test, only some parameters were found to be significantly different between groups with a negative and positive expression of HMGB1 protein, such as pathological stage (pT), Gleason score, preoperative PSA concentration and BCR (P < 0.05) (Table 1).

Figure 2.

Immunohistochemical analysis of high mobility group box 1 (HMGB1) in prostate tumour samples. A, D, HMGB1 staining was negative in normal prostate glandular epithelium. B, E, Representative staining of HMGB1 on tissue arrays. C, F, In a few tumour samples, positive staining was found in both the nucleus and cytoplasm.

Table 1. Correlation between high mobility group box 1 (HMGB1) protein expression and clinicopathological features in patients with prostate cancer after radical prostatectomy
FeaturesHMGB1 protein expression P
  1. *P < 0.05; **P < 0.01. †Mean (t-test). ‡Chi-squared test. §Fisher's exact test. ¶Mann–Whitney U-test.

Age at surgery (years) (N= 168)0.692
 Median (range)69.0 (56.0–77.0)67.0 (48.0–77.0) 
Pathological stage (pT), n (N= 159)*‡0.011
Lymph node status, n (N= 159)§  0.082
Tumour margins, n (N= 168)§  0.331
Capsular invasion, n (N= 159)§  0.149
 No invasion6491
PSA concentration (ng/mL) before surgery (N= 168)**¶<0.001
 Median (range)16.40 (1.43–100)19.91 (2.61–87)
Gleason score (N= 168)**‡<0.001
 Gleason ≤ 62672
 Gleason ≥ 74129
Biochemical recurrence (N= 168)**‡<0.001
 Yes (≥0.2 ng/mL)2890
 No (<0.2 ng/mL)3911


A total 168 patients with intact follow-up information were included in the Kaplan–Meier analysis of the time to BCR. To determine whether groups with negative and positive HMGB1 protein expression were significantly different, the Kaplan–Meier method was used to estimate the probability of survival and also to compare the median BCR-free time and 5-year BCR-free survival rates between negative and positive HMGB1 expression groups. The mean (range) follow-up was 17.1 (2–68) months. During the present study, there were some patients with a short follow-up as a result of the rapid BCR after RP, although no patients were lost to follow-up. The results show that the 5-year BCR-free survival rates were 54.0% in the negative group and 6.0% in the positive group (P < 0.001), respectively. The median time to BCR-free survival in the negative group was 23.1 months vs 15.6 months in the positive group (P < 0.001) (Table 2). The log-rank test showed that the 5-year BCR-free survival rates and BCR-free time were significantly different between the negative and positive groups, indicating that positive HMGB1 protein expression was correlated with a shorter BCR-free survival time (P < 0.001) (Fig. 3).

Table 2. Biochemical recurrence (BCR)-free survival rates at 5 years and median BCR-free time for negative and positive high mobility group box 1 (HMGB1) protein expression groups of patients with prostate cancer after radical prostatectomy
HMGB1 expressionPatients (n)BCR-free survival rates at 5 years (%)BCR-free
Median (months)95% CI
  1. P < 0.001.

Figure 3.

Biochemical recurrence-free survival of negative and positive high mobility group box 1 (HMGB1) protein expression groups. The Kaplan–Meier method was used to draw survival curves for the HMGB1 expression groups.

A Cox proportional hazards model was applied to identify the independent predictors for BCR-free survival. Multivariate Cox analysis included PSA concentration before surgery, age, Gleason score, tumour margins, lymph node status, capsular invasion, pathological stage and HMGB1 expression, among which PSA concentration before surgery, tumour margins and HMGB1 expression showed a statistically significant difference. The results of the multivariate analysis (Table 3) indicated that HMGB1 protein expression was recognized as an independent prognostic factor for patients with PC who were undergoing RP (hazard ratio = 2.348, 95% CI = 1.373–6.361, P= 0.001).

Table 3. Multivariate analysis of biochemical recurrence-free survival for all patients with prostate cancer after radical prostatectomy
Comparison P Hazard ratio for recurrence95% CI
  1. HMGB1, high mobility group box 1.

Tumour margins0.0062.9551.413–3.901
HMGB1 expression0.0012.3481.373–6.361


Recent studies have indicated that HMGB1 relates to the development and progression of human carcinomas; for example, HMGB1 correlates with tumour invasion and has a significant impact on tumour progression and survival in patients with non-small cell lung cancer [14]. Several mechanisms have been proposed to account for the involvement of HMGB1 in tumourigenesis. It has been reported that HMGB1 activates signalling pathways involving protein kinase B (Akt), mitogen-activated protein kinases and nuclear factor-κB, which contribute to tumour growth [15–17].

As a nuclear protein, HMGB1 can facilitate the formation of nuclear hormone complexes and p53 or p73 transcriptional complexes [18,19]. HMGB1 also enhances the DNA-binding and transcriptional activity of steroid receptors, including androgen, oestrogen, progesterone, mineralocorticoid and glucocorticoid receptors; thus, HMGB1 may play a regulatory role in steroid receptor-mediated gene transcription in many hormone-dependent cancer cells [20]. In addition to its nuclear role, HMGB1 also functions as an extracellular signalling molecule during inflammation, cell differentiation, cell migration and tumour metastasis [6]. HMGB1 has a high affinity for several receptors, including the receptor for advanced glycation end products (RAGE), as well as Toll-like receptors TLR-2, TLR-4, TLR-9 and CD24, mediating the response to infection and injury, and thereby promoting inflammation [21,22].

Some previous studies of human colorectal carcinoma and squamous cell carcinoma of skin also indicate that a high HMGB1 protein concentration was linked to poor survival [23,24]. Among the many risk factors, inflammation plays an important role in the initiation and progression of PC [25,26]. HMGB1 (amphoterin) has been shown to be secreted from PC cells as a result of androgen deprivation [9] and its overexpression is associated with PC development [27]. Therefore, it is conceivable that HMGB1 may be an important mediator in prostate carcinogenesis.

Despite the combination of increasingly refined surgical techniques and a reduced incidence of surgical complications, the variable disease course leads to eventual recurrence in one-third of patients after RP [28]. Pound et al. [29] reported that no men had experienced a distant or local recurrence without BCR. Therefore, the utility of BCR predictors for guiding therapy improves the long-term survival time and disease-free survival time of PC after RP. Recent investigations have attempted to apply our improved understanding of tumour cell biology, aiming to identify additional tumour characteristics as possible prognostic factors. Such prognostically important factors may help determine the optimal treatment strategy based on the biological potential of the individual tumour [30]. In the present study, HMGB1 protein expression was inversely correlated with BCR-free survival. In particular, HMGB1 protein expression was confirmed to be a potential independent prognostic factor for predicting BCR in patients with PC after RP. Such information could allow the better planning of treatment strategies for patients with PC.

In conclusion, the present study shows that all three PC cell lines, DU145, PC-3 and LNCaP, expressed HMGB1 mRNA and protein; these results are consistent with previous studies [27]. Compared to the non-transformed immortalized prostate cell RWPE-1, all PC cell lines overexpressed translational and transcriptional HMGB1. An immunohistochemical staining approach coupled with the TMA technique showed that positive HMGB1 protein expression was correlated with pathological stage (pT), Gleason score, PSA concentration before surgery and BCR. Furthermore, we explored the clinical significance of HMGB1 expression. Positive HMGB1 protein expression in PC was significantly associated with poor BCR-free survival. Moreover, a multivariate analysis indicated that HMGB1 can be applied independently to predict the BCR for patients with PC after RP (P= 0.001).


We are grateful to Dr Wei Wang for critically reading the manuscript and for helpful suggestions, and also we thank Dr Jinke Chen for providing expert technical assistance. We are grateful to Dr Ashwin Aubeeluck for his help with the editing of the article in the English language. This work was supported by a China National Natural Science Foundation funded project (No. 81172426).


The authors declare that there are no conflicts of interest.