Immunohistochemical estimation of hypoxia in human obstructed bladder and correlation with clinical variables

Authors

  • George Koritsiadis,

    1. 1st Urology Department, Athens Medical School, Laiko Hospital, 1st Pathology Department, School of Medicine, National and Kapodistrian University of Athens, and
    Search for more papers by this author
  • Konstantinos Stravodimos,

    1. 1st Urology Department, Athens Medical School, Laiko Hospital, 1st Pathology Department, School of Medicine, National and Kapodistrian University of Athens, and
    Search for more papers by this author
  • George Koutalellis,

    1. 1st Urology Department, Athens Medical School, Laiko Hospital, 1st Pathology Department, School of Medicine, National and Kapodistrian University of Athens, and
    Search for more papers by this author
  • Georgios Agrogiannis,

    1. 1st Urology Department, Athens Medical School, Laiko Hospital, 1st Pathology Department, School of Medicine, National and Kapodistrian University of Athens, and
    Search for more papers by this author
  • Sotirios Koritsiadis,

    1. Urology Department, Nikaia General Hospital Piraeus, Greece
    Search for more papers by this author
  • Andreas Lazaris,

    1. 1st Urology Department, Athens Medical School, Laiko Hospital, 1st Pathology Department, School of Medicine, National and Kapodistrian University of Athens, and
    Search for more papers by this author
  • Constantinos Constantinides

    1. 1st Urology Department, Athens Medical School, Laiko Hospital, 1st Pathology Department, School of Medicine, National and Kapodistrian University of Athens, and
    Search for more papers by this author

George Koritsiadis, 1st Urology Department, Athens Medical School, Laiko Hospital, Athens, Greece.
e-mail: koritsiadisdc@yahoo.gr

Abstract

OBJECTIVE

To investigate the tissue distribution of ischaemia in human detrusor in patients with bladder outlet obstruction (BOO) and to correlate the results with clinical variables, as clinical BOO is a common problem in ageing men and ischaemia might be important in detrusor dysfunction.

PATIENTS AND METHODS

From September 2004 to October 2006, 70 patients were recruited, comprising 60 scheduled for surgery to treat benign prostatic hyperplasia (the study group) and 10 as controls. Detrusor tissue was retrieved and stained for hypoxia-inducible factor (HIF)-1α, a cellular marker of hypoxia.

RESULTS

The mean (sd) total number of cells immunoreactive to HIF-1α in the study group was 93.3 (48.09), and in the specimens from the control group only few rare cells showed weak immunoreactivity to HIF-1α (0–2). Positive cells were in different proportions between muscle bundles and submucosa, expressed mainly in stromal cells. The urothelium and detrusor muscle showed no immunoreactivity to HIF-1α. There was strong immunoreactivity in patients with prolonged BOO (<10 years), declining thereafter, and in those patients with urinary retention.

CONCLUSIONS

The urothelium and detrusor seem to be more resistant to hypoxic stress, while stromal cells perceive low oxygen tension. The bladder response to chronic hypoxia through HIF-1α expression is limited in time and might depend on the functional status of the detrusor.

Abbreviations
HIF

hypoxia-inducible factor

UR

urinary retention

BWT

bladder wall thickness

Vp

prostate volume

PVR

postvoid residual volume

US

ultrasonography

Qmax

maximum urinary flow rate

CA

carbonic anhydrase

VEGF

vascular endothelial growth factor

SPP

suprapubic adenectomy

INTRODUCTION

BOO is a common problem in ageing men, occurring in >80% of men in their eighth decade [1]; causes include BPH, urethral stricture, prostate cancer and detrusor sphincter dyssynergia [2]. With time, the obstructed bladder deteriorates due to alterations in its physiology and structure, and it is known that similar changes can also occur in the hypoxic bladder in experimental models [3]. Consequently, it was postulated that ischaemia has a major role in detrusor dysfunction [4] and this condition is characterized by a shift of metabolism towards anaerobic pathways [5].

In BOO there is a reduced blood flow to the detrusor muscle, which becomes hypoxic. This is more evident in the voiding phase, where high pressures are needed to overcome the resistance to flow [6]. In obstructed animals, blood flow to the bladder is inversely related to filling [7], leading to severe hypoxia and consequent further damage to the detrusor. Chronic BOO progresses from a compensated state, where emptying remains normal, to a decompensated state characterized by reduced flow, interrupted urination and finally a postvoid residual volume (PVR) [8]. The bladder responds initially by detrusor muscle hypertrophy and increasing intravesical pressure, to maintain constant flow. However, reduced bladder wall blood flow during the emptying phase results in further tissue hypoxia [6]. This observation comes from experimental models in animals, where obstruction is artificially induced from days to several weeks [1–7].

In humans the development of BOO is very insidious and progresses slowly, so the different stages of bladder deterioration cannot be fully investigated [4], thus there have been few reports to examine hypoxia in the obstructed human detrusor [9,10]. However, in vitro models were developed, and strips of detrusor smooth muscle used to investigate how BOO affects detrusor physiology [9]. Recently, with the use of laser Doppler ultrasonography (US), bladder blood flow was found to decrease during the later stages of filling, along with the increase in intravesical pressure [11]. Thus we investigated the tissue distribution of ischaemia in human detrusor of patients with BOO, and correlated the results with clinical variables.

PATIENTS AND METHODS

Seventy patients with BOO due to BPH were recruited between September 2004 and October 2006; 60 were defined as the study group and 10 served as controls. The study was approved by the ethical board and informed consent was obtained from all patients. Included were patients with BOO due to BPH, and those scheduled for TURP or suprapubic adenectomy (SPP) constituted the study group, while the control group comprised patients with superficial bladder cancer scheduled for TUR.

LUTS were classified according to the IPSS as minimal (score 0–7), moderate (8–19) and severe (>20). The duration of LUTS was also recorded. The maximum urinary flow rate (Qmax) and PVR were measured, the PSA level determined and prostate volume (Vp) estimated by abdominal US. The bladder wall thickness (BWT) was also determined by abdominal US using a baseline bladder capacity of 60% (i.e. 150–200 mL) with values of >2 mm being considered abnormal [12]. All patients had a detailed medical history taken, a physical examination and urine analysis, with laboratory blood tests.

The indications for surgery were: moderate to severe IPSS, a Qmax of <15 mL/s and a previous history of acute urinary retention (UR). An unsatisfactory level of quality of life was also considered an independent indication for surgery. TURP or suprapubic adenectomy was performed depending on prostate volume. Urodynamic studies were conducted in patients with equivocal flow rate results (as per the Abrams-Griffiths nomogram), and in patients aged <50 or >80 years.

Exclusion criteria included recurrent UTI, bladder lithiasis, abnormal haematocrit, previous surgery to the bladder or the prostate, and a history of prostate or bladder cancer. TRUS-guided biopsies of the prostate were taken when prostate cancer was suspected (PSA level >4 ng/mL or am abnormal DRE). Exclusion criteria in the control group included an IPSS of >7, recurrent UTI and previous TUR of the bladder.

Detrusor tissue was easily retrieved in patients who had SPP, whilst in patients who had TURP, a cold-cup biopsy was taken to avoid thermal damage to the specimens. Tissue was always retrieved from the dome of the bladder. Tissue samples from specimens were immersed and fixed in 10% buffered formaldehyde solution for 2 days, routinely processed for paraffin sections, and sections cut from the blocks and mounted on poly l-lysine-coated glass slides. Immunohistochemistry was applied on 4 µm thick sections. Slides were heated at 37 °C overnight. The primary antibody was anti-HIF-1α (Chemicon Inc., Temecula, CA, USA) applied at a dilution of 1:200. Antigen was retrieved by heating the slides in citrate-buffered solution in a microwave oven for 5 min in two cycles. Envision (Dako, Glostrup, Denmark) was used as secondary antibody. Finally, diaminobenzidine was applied as the chromogen and the slides were slightly counterstained with haematoxylin. In substitute negative controls the primary antibody was omitted and replaced by PBS. Sections from samples over-expressing HIF-1α were also stained with carbonic anhydrase (CA) IX, an additional marker of hypoxia (CA IX, clone M75, diluted 1:150) [13].

Two pathologists, unaware of the clinical data, assessed the staining; where the results were equivocal the slides were jointly re-examined for a final consensus. A minimum of 20 randomly selected, high-power fields through the whole section was examined.

The assessment of HIF-1α was based on a previously described method [14]. HIF-1α immunoreactivity was expressed in the nucleus and cytoplasm of stromal cells. The staining was assessed according to the number of positive cells and staining distribution. Specimens were grouped into high- and low-HIF-1α reactivity using a threshold of 80 reactive cells/slide, which represents the lowest 95% CI.

For statistical reasons both the absolute number of positive cells and the categorical nature of the staining was used. Descriptive statistics were used to present the study variables. The mean (sd, range) is given for continuous variables, while frequencies are presented for the categorical values. The chi-square test was used to test statistical significance in categorical variables and odds ratios to quantify the strength of association. The Mann–Whitney test for independent samples was used to assess differences in the mean values of continuous variables between the groups. The Spearman correlation coefficient (when appropriate) was used to examine the independence between categorical variables. All tests were two-sided and the level of significance was set at P ≤ 0.05.

RESULTS

The age of the patients in the study group was 69.2 (8.1, 48–80) years and in control group was 67.3 (5.1, 50–81) years, and hence not statistically different. In the study group the mean duration of BOO was 6 (4.8, 1–20 years); 40 of 60 (70%) men had BOO for <10 years and 20 (30%) for >10 years. The PSA level was 5.02 (3.1, 0.8–25) ng/mL and the mean PVR 191.2 (115.1, 0–300) mL; 35 of 60 (58%) men never had UR while 25 (42%) had at least one episode. Seventeen patients (29%) had a moderate IPSS and 43 (71%) had a severe IPSS. The BWT was <2 mm in 22 (37%) patients and >2 mm in 38 (63%). The PVR significantly correlated with the age of the patients (r = 0.293, P = 0.04). The age of patients with BOO for <10 years differed significantly from that in those with BOO for >10 years, at 67.1 (36.4, 40–82) vs 73.1  (6, 61–86) (P = 0.008).

The mean number of total cells immunoreactive to HIF-1α in the study group of 60 men was 93.3 (48.09). In the specimens from the control group only a few rare cells showed weak immunoreactivity to HIF-1α (<0–2) and it was mainly cytoplasmic (Fig. 1). The immunoreactivity in the study group was diffusely distributed among positive cells and was mainly nuclear and only weakly cytoplasmic. HIF-1α was expressed in stromal cells, including fibroblasts and rare macrophages; HIF-1α positive cells were of different proportions between muscle bundles and submucosa (Fig. 2a). The urothelium and detrusor muscle had no immunoreactivity to HIF-1α (Fig. 2b).

Figure 1.

The absence of hypoxic cells in control specimens (×100).

Figure 2.

a, Increased detection of HIF-1α in stromal cells among hyperplastic muscle cells (×100); b, absence of HIF-1α immunoreactivity in urothelium (×100).

The difference in the mean number of total cells between the study and the control group was statistically significant (P < 0.001). The evaluation of staining with CA IX gave similar results to HIF-1α, confirming the presence of hypoxia in stromal cells but not urothelium and detrusor muscle.

The clinical characteristics are summarized in (Table 1); HIF-1α immunoreactivity in men with BOO for <10 years differed significantly from those with BOO for >10 years (P = 0.018), being higher in the former (Fig. 3a). The risk of high immunoreactivity was four times greater in men with BOO for <10 years than in those with BOO for >10 years, with an odds ratio of 4.25 (95% CI 1.23–14.64).

Table 1.  The clinical characteristics and relation to HIF-1α expression with the probability of HIF-1α being highly immunoreactive
VariableOdds ratio (95% CI)P
IPSS1.429 (0.434–4.705)0.345
BOO <10 years4.245 (1.230–14.64)0.018
UR4.252 (1.214–14.88)0.02
BWT1.786 (0.403–7.906)0.44
Figure 3.

Box plots with outliers, showing HIF-1α immunoreactivity with: a, the duration of BOO; and b, the presence of UR.

In 25 patients with UR there was consistently higher immunoreactivity to HIF-1α than in those without UR (P = 0.020), and the risk of identifying a high expression of HIF-1α was four times higher in patients with UR (odds ratio 4.25, 95% CI 1.214–14.88). When adjusting for the duration of BOO, there was a higher probability of HIF-1α immunoreactivity in those with UR (odds ratio 8.33, 95% CI 1.48–46.93; P = 0.009). The duration of BOO differed significantly between those with and without UR, with the former being obstructed for a shorter period, at 3.95 (4) vs 6.24 (4.9) years (P = 0.045), and being older, although with marginal significance, at 71.14 (6.8) vs 66.8 (9.2) years (P = 0.081; Table 2, Fig. 3b).

Table 2.  The clinical and demographic characteristics of the patients according to UR
VariableNo URURP
  1. n, number of immunoreactive cells.

HIF-1α, n79.5 (51.2)92.6 (41.6)0.024
Vp, mL73.4 (28.1)80.5 (26.6)0.3
Age, years66.1 (7.7)72.3 (6.8)0.081
IPSS22.9 (5.6)19.2 (7.8)0.2
Years of BOO 6.24 (4.9) 3.95 (4.0)0.045

HIF-1α expression was higher in men with a BWT of >2 mm, at 102.6 (47, 20–175) cells, vs 93.3 (43, 25–150) cells in those with a BWT of <2 m, although this difference was not statistically significant (P = 0.4). There was no correlation between HIF-1α and any of the other variables assessed.

DISCUSSION

The most significant finding of the present study is that the obstructed bladder in the living human is more hypoxic than in controls, and responds by expressing HIF-1α. Previously only one report confirmed that human bladder under obstruction becomes hypoxic [15]. Several studies previously evaluated the role of low oxygen tension on malignant tissues, but only a few studies have evaluated the role of HIF-1α in benign conditions [16–19].

The response of human detrusor muscle to hypoxic insult was investigated in an in vitro cell culture model [9]. Those authors reported a time-dependent increase in HIF-1α expression (up to 72 h) with a doubling of vascular endothelial growth factor (VEGF), while spontaneous cell apoptosis and viability were not influenced. They concluded that detrusor muscle responds to hypoxic environment by maintaining cell viability. They also recognized that cultured cells differ from fresh detrusor tissue, whilst the response of submucosa and stromal cells was not considered.

We evaluated HIF-1α expression in bladder tissues; HIF-1α is an oxygen-dependent transcriptional activator, and activates genes involved in glucose transport and metabolism, up-regulates a gene involved in cell survival and apoptosis, and interferes with extracellular matrix metabolism and epithelial homeostasis [20]. HIF is a heterodimer composed of two subunits -1α and -1β. The first subunit is almost undetectable under normal oxygen tension because it is rapidly degraded, while the HIF-1β subunit is constitutively expressed and controlled in an oxygen-independent manner [21]. HIF-1α can also be stimulated by other cytokines in normoxic conditions [22], but it can also be activated in response to mechanical stress [23].

In animal models the loss of HIF-1α results in organ decompensation. In one of these studies, there was extensive muscle damage and reduced exercise time in genetically modified mice with loss of HIF-1α, relative to the wild-type mice, after prolonged exercise [24]. In a different study, an up-regulation of HIF-1α was a constant finding in athletes training in hypoxia, simultaneously with myoglobin synthesis, VEGF mRNA and glycolytic enzymes in muscle [25]. These observations considered together support the compensatory role of HIF-1α under hypoxic conditions in tissues under stress.

Although UR is not a marker of bladder decompensation, some investigators consider it as an early sign of bladder deterioration. If the resistance to flow is maintained, the structural and metabolic damage to the bladder becomes permanent and decompensation ensues because of tissue ischaemia [8]. Under hypoxic conditions, detrusor cells become completely dependent on lactate for ATP production, thus being rendered less sensitive to low oxygen tension [26]. Moreover, residual urine is a consequence of decreased aerobic metabolism. Hence, when the bladder depletes the energy reserves during the voiding phase, detrusor contraction fails prematurely and the result is a PVR. Also, it is known that HIF-1α in hypoxic tissues regulates the expression of all enzymes in the glycolytic pathway [20–24]. These observations are consistent with the present study, where high expression of HIF-1α was four times more likely in men with UR.

In a different study, immunostaining of cultured rabbit vascular smooth muscle cells showed reduced expression of HIF-1α in old vs young cells exposed to 6 h of hypoxia, and this was correlated with deficient VEGF up-regulation [27]. These authors concluded that there is an age-dependent defect in HIF activity responsible for the reduced expression of VEGF, and consequently impaired angiogenesis with ageing. In our study, HIF-1α expression did not correlate with age, but low expression was a constant finding in men with BOO for >10 years, who were also older.

There was a higher likelihood of expression of HIF-1α in stromal cells when the BWT was >2 mm, although this not statistically different from the expression in those with a BWT of <2 mm. Hypertrophic muscle has a higher metabolic demand and becomes more easily hypoxic under wall stress; a higher expression of HIF-1α would be expected in hypertrophied smooth muscle, but this was not detected. However Oelke et al.[12], using abdominal US, identified patients with BOO using a 2-mm threshold for BWT, but only after a complete urodynamic evaluation. We did not use urodynamics to categorize the present patients, so we cannot safely use HIF-1α expression to define the optimum threshold value for BWT in BOO.

In the present patients with BOO, tissue was retrieved from the dome, as it was reported that during filling the dome has the lowest perfusion [28], and we hypothesized that hypoxia would be more pronounced at this location in BOO. HIF-1α was expressed mainly in stromal cells between muscle bundles and in connective tissue beneath the mucosal layer, while urothelium and detrusor had no immunoreactivity. It seems that, in BOO, stromal cells are more likely to perceive low oxygen tension or high intravesical pressure, while urothelium and detrusor muscle seem to be more resistant to hypoxic stress, by contrast with what happens in experimental models [29]. It is likely that stromal cells interact with urothelium and detrusor through mediators induced by HIF-1α. Moreover, the diffuse staining suggests that hypoxia is uniform through the bladder wall and not a regional phenomenon, and that bladder dysfunction perhaps originates from these areas of hypoxia.

An interesting finding was the presence of HIF-1α both into the cytoplasm and nucleus. HIF-1α is mainly identified as a nuclear protein under hypoxic conditions. Zheng et al.[30] suggested that HIF-1α is present in both nuclear and cytoplasmic compartments under conditions of normoxia, hypoxia and re-oxygenation. Furthermore, proteosome-dependent degradation of the molecule occurs at a similar rate in both compartments.

Although two pathologists separately evaluated the slides in the present study, there was some interobserver variability, so it would be more accurate to estimate HIF-1α mRNA at tissue level. It would also be ideal to correlate HIF-1α expression with VEGF expression and urodynamic findings, as the IPSS is not as accurate in classifying bladder dysfunction. We also postulated that HIF-1α, apart from being an indicator of tissue hypoxia, is also a marker of organ compensation, but this evidence comes indirectly from studies in animal models or HIF-1α expression in striated muscle, conditions very different from what might happen in detrusor muscle. However, bladder decompensation is not fully understood and, to date, there has been no precise functional or biomedical definition of ‘compensation’ or ‘decompensation’ in BOO, so we could not define accurately those patients with detrusor failure.

In conclusion, we assessed human detrusor in its normal environment, and verified that in partial chronic BOO the detrusor is ischaemic and responds by expressing HIF-1α, to adapt to the new metabolic status. Hypoxia in the bladder is diffusely distributed and not a regional phenomenon. The urothelium and detrusor seem to be more resistant to hypoxic stress, whilst stromal cells perceive low oxygen tension. The HIF-1α response is limited in a time-dependent manner and might be related to the functional status of the detrusor. The bladder can possibly compensate for the first few years after that the adaptive response declines.

ACKNOWLEDGEMENTS

We thank Vic Manohar (MS, FRCS) for his contribution to the study, and Maria Kemerli and George Babaliaris for tissue preparation and immunohistochemistry.

CONFLICT OF INTEREST

None declared.

Ancillary